818 files changed, 62349 insertions, 23359 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index f5120046ff80..e3b4b0f02b3d 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -8,6 +8,7 @@ menuconfig DRM
 	tristate "Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)"
 	depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU && HAS_DMA
 	select HDMI
+	select FB_CMDLINE
 	select I2C
 	select I2C_ALGOBIT
 	select DMA_SHARED_BUFFER
@@ -24,12 +25,6 @@ config DRM_MIPI_DSI
 	bool
 	depends on DRM
 
-config DRM_USB
-	tristate
-	depends on DRM
-	depends on USB_SUPPORT && USB_ARCH_HAS_HCD
-	select USB
-
 config DRM_KMS_HELPER
 	tristate
 	depends on DRM
@@ -114,6 +109,8 @@ config DRM_RADEON
 	select POWER_SUPPLY
 	select HWMON
 	select BACKLIGHT_CLASS_DEVICE
+	select INTERVAL_TREE
+	select MMU_NOTIFIER
 	help
 	  Choose this option if you have an ATI Radeon graphics card.  There
 	  are both PCI and AGP versions.  You don't need to choose this to
@@ -201,3 +198,5 @@ source "drivers/gpu/drm/msm/Kconfig"
 source "drivers/gpu/drm/tegra/Kconfig"
 
 source "drivers/gpu/drm/panel/Kconfig"
+
+source "drivers/gpu/drm/sti/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index dd2ba4269740..9292a761ea6d 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -4,10 +4,10 @@
 
 ccflags-y := -Iinclude/drm
 
-drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
+drm-y       :=	drm_auth.o drm_bufs.o drm_cache.o \
 		drm_context.o drm_dma.o \
-		drm_drv.o drm_fops.o drm_gem.o drm_ioctl.o drm_irq.o \
-		drm_lock.o drm_memory.o drm_stub.o drm_vm.o \
+		drm_fops.o drm_gem.o drm_ioctl.o drm_irq.o \
+		drm_lock.o drm_memory.o drm_drv.o drm_vm.o \
 		drm_agpsupport.o drm_scatter.o drm_pci.o \
 		drm_platform.o drm_sysfs.o drm_hashtab.o drm_mm.o \
 		drm_crtc.o drm_modes.o drm_edid.o \
@@ -20,11 +20,10 @@ drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
 drm-$(CONFIG_PCI) += ati_pcigart.o
 drm-$(CONFIG_DRM_PANEL) += drm_panel.o
-
-drm-usb-y   := drm_usb.o
+drm-$(CONFIG_OF) += drm_of.o
 
 drm_kms_helper-y := drm_crtc_helper.o drm_dp_helper.o drm_probe_helper.o \
-		drm_plane_helper.o
+		drm_plane_helper.o drm_dp_mst_topology.o
 drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
 drm_kms_helper-$(CONFIG_DRM_KMS_FB_HELPER) += drm_fb_helper.o
 drm_kms_helper-$(CONFIG_DRM_KMS_CMA_HELPER) += drm_fb_cma_helper.o
@@ -35,7 +34,6 @@ CFLAGS_drm_trace_points.o := -I$(src)
 
 obj-$(CONFIG_DRM)	+= drm.o
 obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o
-obj-$(CONFIG_DRM_USB)   += drm_usb.o
 obj-$(CONFIG_DRM_TTM)	+= ttm/
 obj-$(CONFIG_DRM_TDFX)	+= tdfx/
 obj-$(CONFIG_DRM_R128)	+= r128/
@@ -63,6 +61,7 @@ obj-$(CONFIG_DRM_QXL) += qxl/
 obj-$(CONFIG_DRM_BOCHS) += bochs/
 obj-$(CONFIG_DRM_MSM) += msm/
 obj-$(CONFIG_DRM_TEGRA) += tegra/
+obj-$(CONFIG_DRM_STI) += sti/
 obj-y			+= i2c/
 obj-y			+= panel/
 obj-y			+= bridge/
diff --git a/drivers/gpu/drm/armada/armada_510.c b/drivers/gpu/drm/armada/armada_510.c
index 59948eff6095..ad3d2ebf95c9 100644
--- a/drivers/gpu/drm/armada/armada_510.c
+++ b/drivers/gpu/drm/armada/armada_510.c
@@ -15,20 +15,19 @@
 #include "armada_drm.h"
 #include "armada_hw.h"
 
-static int armada510_init(struct armada_private *priv, struct device *dev)
+static int armada510_crtc_init(struct armada_crtc *dcrtc, struct device *dev)
 {
-	priv->extclk[0] = devm_clk_get(dev, "ext_ref_clk_1");
+	struct clk *clk;
 
-	if (IS_ERR(priv->extclk[0]) && PTR_ERR(priv->extclk[0]) == -ENOENT)
-		priv->extclk[0] = ERR_PTR(-EPROBE_DEFER);
+	clk = devm_clk_get(dev, "ext_ref_clk1");
+	if (IS_ERR(clk))
+		return PTR_ERR(clk) == -ENOENT ? -EPROBE_DEFER : PTR_ERR(clk);
 
-	return PTR_RET(priv->extclk[0]);
-}
+	dcrtc->extclk[0] = clk;
 
-static int armada510_crtc_init(struct armada_crtc *dcrtc)
-{
 	/* Lower the watermark so to eliminate jitter at higher bandwidths */
 	armada_updatel(0x20, (1 << 11) | 0xff, dcrtc->base + LCD_CFG_RDREG4F);
+
 	return 0;
 }
 
@@ -45,8 +44,7 @@ static int armada510_crtc_init(struct armada_crtc *dcrtc)
 static int armada510_crtc_compute_clock(struct armada_crtc *dcrtc,
 	const struct drm_display_mode *mode, uint32_t *sclk)
 {
-	struct armada_private *priv = dcrtc->crtc.dev->dev_private;
-	struct clk *clk = priv->extclk[0];
+	struct clk *clk = dcrtc->extclk[0];
 	int ret;
 
 	if (dcrtc->num == 1)
@@ -81,7 +79,6 @@ static int armada510_crtc_compute_clock(struct armada_crtc *dcrtc,
 const struct armada_variant armada510_ops = {
 	.has_spu_adv_reg = true,
 	.spu_adv_reg = ADV_HWC32ENABLE | ADV_HWC32ARGB | ADV_HWC32BLEND,
-	.init = armada510_init,
-	.crtc_init = armada510_crtc_init,
-	.crtc_compute_clock = armada510_crtc_compute_clock,
+	.init = armada510_crtc_init,
+	.compute_clock = armada510_crtc_compute_clock,
 };
diff --git a/drivers/gpu/drm/armada/armada_crtc.c b/drivers/gpu/drm/armada/armada_crtc.c
index 3aedf9e993e6..e4a1490b42c2 100644
--- a/drivers/gpu/drm/armada/armada_crtc.c
+++ b/drivers/gpu/drm/armada/armada_crtc.c
@@ -7,6 +7,9 @@
  * published by the Free Software Foundation.
  */
 #include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include "armada_crtc.h"
@@ -257,7 +260,7 @@ static void armada_drm_vblank_off(struct armada_crtc *dcrtc)
 	 * Tell the DRM core that vblank IRQs aren't going to happen for
 	 * a while.  This cleans up any pending vblank events for us.
 	 */
-	drm_vblank_off(dev, dcrtc->num);
+	drm_crtc_vblank_off(&dcrtc->crtc);
 
 	/* Handle any pending flip event. */
 	spin_lock_irq(&dev->event_lock);
@@ -286,6 +289,8 @@ static void armada_drm_crtc_dpms(struct drm_crtc *crtc, int dpms)
 		armada_drm_crtc_update(dcrtc);
 		if (dpms_blanked(dpms))
 			armada_drm_vblank_off(dcrtc);
+		else
+			drm_crtc_vblank_on(&dcrtc->crtc);
 	}
 }
 
@@ -332,24 +337,23 @@ static void armada_drm_crtc_commit(struct drm_crtc *crtc)
 static bool armada_drm_crtc_mode_fixup(struct drm_crtc *crtc,
 	const struct drm_display_mode *mode, struct drm_display_mode *adj)
 {
-	struct armada_private *priv = crtc->dev->dev_private;
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
 	int ret;
 
 	/* We can't do interlaced modes if we don't have the SPU_ADV_REG */
-	if (!priv->variant->has_spu_adv_reg &&
+	if (!dcrtc->variant->has_spu_adv_reg &&
 	    adj->flags & DRM_MODE_FLAG_INTERLACE)
 		return false;
 
 	/* Check whether the display mode is possible */
-	ret = priv->variant->crtc_compute_clock(dcrtc, adj, NULL);
+	ret = dcrtc->variant->compute_clock(dcrtc, adj, NULL);
 	if (ret)
 		return false;
 
 	return true;
 }
 
-void armada_drm_crtc_irq(struct armada_crtc *dcrtc, u32 stat)
+static void armada_drm_crtc_irq(struct armada_crtc *dcrtc, u32 stat)
 {
 	struct armada_vbl_event *e, *n;
 	void __iomem *base = dcrtc->base;
@@ -410,6 +414,27 @@ void armada_drm_crtc_irq(struct armada_crtc *dcrtc, u32 stat)
 	}
 }
 
+static irqreturn_t armada_drm_irq(int irq, void *arg)
+{
+	struct armada_crtc *dcrtc = arg;
+	u32 v, stat = readl_relaxed(dcrtc->base + LCD_SPU_IRQ_ISR);
+
+	/*
+	 * This is rediculous - rather than writing bits to clear, we
+	 * have to set the actual status register value.  This is racy.
+	 */
+	writel_relaxed(0, dcrtc->base + LCD_SPU_IRQ_ISR);
+
+	/* Mask out those interrupts we haven't enabled */
+	v = stat & dcrtc->irq_ena;
+
+	if (v & (VSYNC_IRQ|GRA_FRAME_IRQ|DUMB_FRAMEDONE)) {
+		armada_drm_crtc_irq(dcrtc, stat);
+		return IRQ_HANDLED;
+	}
+	return IRQ_NONE;
+}
+
 /* These are locked by dev->vbl_lock */
 void armada_drm_crtc_disable_irq(struct armada_crtc *dcrtc, u32 mask)
 {
@@ -470,7 +495,6 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_display_mode *mode, struct drm_display_mode *adj,
 	int x, int y, struct drm_framebuffer *old_fb)
 {
-	struct armada_private *priv = crtc->dev->dev_private;
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
 	struct armada_regs regs[17];
 	uint32_t lm, rm, tm, bm, val, sclk;
@@ -504,7 +528,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	/* Wait for pending flips to complete */
 	wait_event(dcrtc->frame_wait, !dcrtc->frame_work);
 
-	drm_vblank_pre_modeset(crtc->dev, dcrtc->num);
+	drm_crtc_vblank_off(crtc);
 
 	crtc->mode = *adj;
 
@@ -515,7 +539,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	}
 
 	/* Now compute the divider for real */
-	priv->variant->crtc_compute_clock(dcrtc, adj, &sclk);
+	dcrtc->variant->compute_clock(dcrtc, adj, &sclk);
 
 	/* Ensure graphic fifo is enabled */
 	armada_reg_queue_mod(regs, i, 0, CFG_PDWN64x66, LCD_SPU_SRAM_PARA1);
@@ -537,7 +561,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	dcrtc->v[1].spu_v_porch = tm << 16 | bm;
 	val = adj->crtc_hsync_start;
 	dcrtc->v[1].spu_adv_reg = val << 20 | val | ADV_VSYNCOFFEN |
-		priv->variant->spu_adv_reg;
+		dcrtc->variant->spu_adv_reg;
 
 	if (interlaced) {
 		/* Odd interlaced frame */
@@ -546,7 +570,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 		dcrtc->v[0].spu_v_porch = dcrtc->v[1].spu_v_porch + 1;
 		val = adj->crtc_hsync_start - adj->crtc_htotal / 2;
 		dcrtc->v[0].spu_adv_reg = val << 20 | val | ADV_VSYNCOFFEN |
-			priv->variant->spu_adv_reg;
+			dcrtc->variant->spu_adv_reg;
 	} else {
 		dcrtc->v[0] = dcrtc->v[1];
 	}
@@ -561,7 +585,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 	armada_reg_queue_set(regs, i, dcrtc->v[0].spu_v_h_total,
 			   LCD_SPUT_V_H_TOTAL);
 
-	if (priv->variant->has_spu_adv_reg) {
+	if (dcrtc->variant->has_spu_adv_reg) {
 		armada_reg_queue_mod(regs, i, dcrtc->v[0].spu_adv_reg,
 				     ADV_VSYNC_L_OFF | ADV_VSYNC_H_OFF |
 				     ADV_VSYNCOFFEN, LCD_SPU_ADV_REG);
@@ -595,7 +619,7 @@ static int armada_drm_crtc_mode_set(struct drm_crtc *crtc,
 
 	armada_drm_crtc_update(dcrtc);
 
-	drm_vblank_post_modeset(crtc->dev, dcrtc->num);
+	drm_crtc_vblank_on(crtc);
 	armada_drm_crtc_finish_fb(dcrtc, old_fb, dpms_blanked(dcrtc->dpms));
 
 	return 0;
@@ -805,12 +829,11 @@ static int armada_drm_crtc_cursor_set(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
-	struct armada_private *priv = crtc->dev->dev_private;
 	struct armada_gem_object *obj = NULL;
 	int ret;
 
 	/* If no cursor support, replicate drm's return value */
-	if (!priv->variant->has_spu_adv_reg)
+	if (!dcrtc->variant->has_spu_adv_reg)
 		return -ENXIO;
 
 	if (handle && w > 0 && h > 0) {
@@ -858,11 +881,10 @@ static int armada_drm_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 {
 	struct drm_device *dev = crtc->dev;
 	struct armada_crtc *dcrtc = drm_to_armada_crtc(crtc);
-	struct armada_private *priv = crtc->dev->dev_private;
 	int ret;
 
 	/* If no cursor support, replicate drm's return value */
-	if (!priv->variant->has_spu_adv_reg)
+	if (!dcrtc->variant->has_spu_adv_reg)
 		return -EFAULT;
 
 	mutex_lock(&dev->struct_mutex);
@@ -888,6 +910,10 @@ static void armada_drm_crtc_destroy(struct drm_crtc *crtc)
 	if (!IS_ERR(dcrtc->clk))
 		clk_disable_unprepare(dcrtc->clk);
 
+	writel_relaxed(0, dcrtc->base + LCD_SPU_IRQ_ENA);
+
+	of_node_put(dcrtc->crtc.port);
+
 	kfree(dcrtc);
 }
 
@@ -921,18 +947,15 @@ static int armada_drm_crtc_page_flip(struct drm_crtc *crtc,
 	armada_reg_queue_end(work->regs, i);
 
 	/*
-	 * Hold the old framebuffer for the work - DRM appears to drop our
-	 * reference to the old framebuffer in drm_mode_page_flip_ioctl().
+	 * Ensure that we hold a reference on the new framebuffer.
+	 * This has to match the behaviour in mode_set.
 	 */
-	drm_framebuffer_reference(work->old_fb);
+	drm_framebuffer_reference(fb);
 
 	ret = armada_drm_crtc_queue_frame_work(dcrtc, work);
 	if (ret) {
-		/*
-		 * Undo our reference above; DRM does not drop the reference
-		 * to this object on error, so that's okay.
-		 */
-		drm_framebuffer_unreference(work->old_fb);
+		/* Undo our reference above */
+		drm_framebuffer_unreference(fb);
 		kfree(work);
 		return ret;
 	}
@@ -1027,19 +1050,20 @@ static int armada_drm_crtc_create_properties(struct drm_device *dev)
 	return 0;
 }
 
-int armada_drm_crtc_create(struct drm_device *dev, unsigned num,
-	struct resource *res)
+int armada_drm_crtc_create(struct drm_device *drm, struct device *dev,
+	struct resource *res, int irq, const struct armada_variant *variant,
+	struct device_node *port)
 {
-	struct armada_private *priv = dev->dev_private;
+	struct armada_private *priv = drm->dev_private;
 	struct armada_crtc *dcrtc;
 	void __iomem *base;
 	int ret;
 
-	ret = armada_drm_crtc_create_properties(dev);
+	ret = armada_drm_crtc_create_properties(drm);
 	if (ret)
 		return ret;
 
-	base = devm_ioremap_resource(dev->dev, res);
+	base = devm_ioremap_resource(dev, res);
 	if (IS_ERR(base))
 		return PTR_ERR(base);
 
@@ -1049,8 +1073,12 @@ int armada_drm_crtc_create(struct drm_device *dev, unsigned num,
 		return -ENOMEM;
 	}
 
+	if (dev != drm->dev)
+		dev_set_drvdata(dev, dcrtc);
+
+	dcrtc->variant = variant;
 	dcrtc->base = base;
-	dcrtc->num = num;
+	dcrtc->num = drm->mode_config.num_crtc;
 	dcrtc->clk = ERR_PTR(-EINVAL);
 	dcrtc->csc_yuv_mode = CSC_AUTO;
 	dcrtc->csc_rgb_mode = CSC_AUTO;
@@ -1072,9 +1100,18 @@ int armada_drm_crtc_create(struct drm_device *dev, unsigned num,
 		       CFG_PDWN64x66, dcrtc->base + LCD_SPU_SRAM_PARA1);
 	writel_relaxed(0x2032ff81, dcrtc->base + LCD_SPU_DMA_CTRL1);
 	writel_relaxed(0x00000000, dcrtc->base + LCD_SPU_GRA_OVSA_HPXL_VLN);
+	writel_relaxed(dcrtc->irq_ena, dcrtc->base + LCD_SPU_IRQ_ENA);
+	writel_relaxed(0, dcrtc->base + LCD_SPU_IRQ_ISR);
 
-	if (priv->variant->crtc_init) {
-		ret = priv->variant->crtc_init(dcrtc);
+	ret = devm_request_irq(dev, irq, armada_drm_irq, 0, "armada_drm_crtc",
+			       dcrtc);
+	if (ret < 0) {
+		kfree(dcrtc);
+		return ret;
+	}
+
+	if (dcrtc->variant->init) {
+		ret = dcrtc->variant->init(dcrtc, dev);
 		if (ret) {
 			kfree(dcrtc);
 			return ret;
@@ -1086,7 +1123,8 @@ int armada_drm_crtc_create(struct drm_device *dev, unsigned num,
 
 	priv->dcrtc[dcrtc->num] = dcrtc;
 
-	drm_crtc_init(dev, &dcrtc->crtc, &armada_crtc_funcs);
+	dcrtc->crtc.port = port;
+	drm_crtc_init(drm, &dcrtc->crtc, &armada_crtc_funcs);
 	drm_crtc_helper_add(&dcrtc->crtc, &armada_crtc_helper_funcs);
 
 	drm_object_attach_property(&dcrtc->crtc.base, priv->csc_yuv_prop,
@@ -1094,5 +1132,107 @@ int armada_drm_crtc_create(struct drm_device *dev, unsigned num,
 	drm_object_attach_property(&dcrtc->crtc.base, priv->csc_rgb_prop,
 				   dcrtc->csc_rgb_mode);
 
-	return armada_overlay_plane_create(dev, 1 << dcrtc->num);
+	return armada_overlay_plane_create(drm, 1 << dcrtc->num);
+}
+
+static int
+armada_lcd_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct drm_device *drm = data;
+	struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	int irq = platform_get_irq(pdev, 0);
+	const struct armada_variant *variant;
+	struct device_node *port = NULL;
+
+	if (irq < 0)
+		return irq;
+
+	if (!dev->of_node) {
+		const struct platform_device_id *id;
+
+		id = platform_get_device_id(pdev);
+		if (!id)
+			return -ENXIO;
+
+		variant = (const struct armada_variant *)id->driver_data;
+	} else {
+		const struct of_device_id *match;
+		struct device_node *np, *parent = dev->of_node;
+
+		match = of_match_device(dev->driver->of_match_table, dev);
+		if (!match)
+			return -ENXIO;
+
+		np = of_get_child_by_name(parent, "ports");
+		if (np)
+			parent = np;
+		port = of_get_child_by_name(parent, "port");
+		of_node_put(np);
+		if (!port) {
+			dev_err(dev, "no port node found in %s\n",
+				parent->full_name);
+			return -ENXIO;
+		}
+
+		variant = match->data;
+	}
+
+	return armada_drm_crtc_create(drm, dev, res, irq, variant, port);
+}
+
+static void
+armada_lcd_unbind(struct device *dev, struct device *master, void *data)
+{
+	struct armada_crtc *dcrtc = dev_get_drvdata(dev);
+
+	armada_drm_crtc_destroy(&dcrtc->crtc);
 }
+
+static const struct component_ops armada_lcd_ops = {
+	.bind = armada_lcd_bind,
+	.unbind = armada_lcd_unbind,
+};
+
+static int armada_lcd_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &armada_lcd_ops);
+}
+
+static int armada_lcd_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &armada_lcd_ops);
+	return 0;
+}
+
+static struct of_device_id armada_lcd_of_match[] = {
+	{
+		.compatible	= "marvell,dove-lcd",
+		.data		= &armada510_ops,
+	},
+	{}
+};
+MODULE_DEVICE_TABLE(of, armada_lcd_of_match);
+
+static const struct platform_device_id armada_lcd_platform_ids[] = {
+	{
+		.name		= "armada-lcd",
+		.driver_data	= (unsigned long)&armada510_ops,
+	}, {
+		.name		= "armada-510-lcd",
+		.driver_data	= (unsigned long)&armada510_ops,
+	},
+	{ },
+};
+MODULE_DEVICE_TABLE(platform, armada_lcd_platform_ids);
+
+struct platform_driver armada_lcd_platform_driver = {
+	.probe	= armada_lcd_probe,
+	.remove	= armada_lcd_remove,
+	.driver = {
+		.name	= "armada-lcd",
+		.owner	=  THIS_MODULE,
+		.of_match_table = armada_lcd_of_match,
+	},
+	.id_table = armada_lcd_platform_ids,
+};
diff --git a/drivers/gpu/drm/armada/armada_crtc.h b/drivers/gpu/drm/armada/armada_crtc.h
index 9c10a07e7492..98102a5a9af5 100644
--- a/drivers/gpu/drm/armada/armada_crtc.h
+++ b/drivers/gpu/drm/armada/armada_crtc.h
@@ -32,12 +32,15 @@ struct armada_regs {
 	armada_reg_queue_mod(_r, _i, 0, 0, ~0)
 
 struct armada_frame_work;
+struct armada_variant;
 
 struct armada_crtc {
 	struct drm_crtc		crtc;
+	const struct armada_variant *variant;
 	unsigned		num;
 	void __iomem		*base;
 	struct clk		*clk;
+	struct clk		*extclk[2];
 	struct {
 		uint32_t	spu_v_h_total;
 		uint32_t	spu_v_porch;
@@ -72,12 +75,16 @@ struct armada_crtc {
 };
 #define drm_to_armada_crtc(c) container_of(c, struct armada_crtc, crtc)
 
-int armada_drm_crtc_create(struct drm_device *, unsigned, struct resource *);
+struct device_node;
+int armada_drm_crtc_create(struct drm_device *, struct device *,
+	struct resource *, int, const struct armada_variant *,
+	struct device_node *);
 void armada_drm_crtc_gamma_set(struct drm_crtc *, u16, u16, u16, int);
 void armada_drm_crtc_gamma_get(struct drm_crtc *, u16 *, u16 *, u16 *, int);
-void armada_drm_crtc_irq(struct armada_crtc *, u32);
 void armada_drm_crtc_disable_irq(struct armada_crtc *, u32);
 void armada_drm_crtc_enable_irq(struct armada_crtc *, u32);
 void armada_drm_crtc_update_regs(struct armada_crtc *, struct armada_regs *);
 
+extern struct platform_driver armada_lcd_platform_driver;
+
 #endif
diff --git a/drivers/gpu/drm/armada/armada_drm.h b/drivers/gpu/drm/armada/armada_drm.h
index a72cae03b99b..ea63c6c7c66f 100644
--- a/drivers/gpu/drm/armada/armada_drm.h
+++ b/drivers/gpu/drm/armada/armada_drm.h
@@ -59,26 +59,23 @@ void armada_drm_vbl_event_remove_unlocked(struct armada_crtc *,
 struct armada_private;
 
 struct armada_variant {
-	bool	has_spu_adv_reg;
+	bool has_spu_adv_reg;
 	uint32_t spu_adv_reg;
-	int (*init)(struct armada_private *, struct device *);
-	int (*crtc_init)(struct armada_crtc *);
-	int (*crtc_compute_clock)(struct armada_crtc *,
-				  const struct drm_display_mode *,
-				  uint32_t *);
+	int (*init)(struct armada_crtc *, struct device *);
+	int (*compute_clock)(struct armada_crtc *,
+			     const struct drm_display_mode *,
+			     uint32_t *);
 };
 
 /* Variant ops */
 extern const struct armada_variant armada510_ops;
 
 struct armada_private {
-	const struct armada_variant *variant;
 	struct work_struct	fb_unref_work;
 	DECLARE_KFIFO(fb_unref, struct drm_framebuffer *, 8);
 	struct drm_fb_helper	*fbdev;
 	struct armada_crtc	*dcrtc[2];
 	struct drm_mm		linear;
-	struct clk		*extclk[2];
 	struct drm_property	*csc_yuv_prop;
 	struct drm_property	*csc_rgb_prop;
 	struct drm_property	*colorkey_prop;
diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c
index 8ab3cd1a8cdb..908e5316eac4 100644
--- a/drivers/gpu/drm/armada/armada_drv.c
+++ b/drivers/gpu/drm/armada/armada_drv.c
@@ -6,7 +6,9 @@
  * published by the Free Software Foundation.
  */
 #include <linux/clk.h>
+#include <linux/component.h>
 #include <linux/module.h>
+#include <linux/of_graph.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include "armada_crtc.h"
@@ -52,6 +54,11 @@ static const struct armada_drm_slave_config tda19988_config = {
 };
 #endif
 
+static bool is_componentized(struct device *dev)
+{
+	return dev->of_node || dev->platform_data;
+}
+
 static void armada_drm_unref_work(struct work_struct *work)
 {
 	struct armada_private *priv =
@@ -85,6 +92,7 @@ void armada_drm_queue_unref_work(struct drm_device *dev,
 static int armada_drm_load(struct drm_device *dev, unsigned long flags)
 {
 	const struct platform_device_id *id;
+	const struct armada_variant *variant;
 	struct armada_private *priv;
 	struct resource *res[ARRAY_SIZE(priv->dcrtc)];
 	struct resource *mem = NULL;
@@ -107,7 +115,7 @@ static int armada_drm_load(struct drm_device *dev, unsigned long flags)
 			return -EINVAL;
 	}
 
-	if (!res[0] || !mem)
+	if (!mem)
 		return -ENXIO;
 
 	if (!devm_request_mem_region(dev->dev, mem->start,
@@ -128,11 +136,7 @@ static int armada_drm_load(struct drm_device *dev, unsigned long flags)
 	if (!id)
 		return -ENXIO;
 
-	priv->variant = (struct armada_variant *)id->driver_data;
-
-	ret = priv->variant->init(priv, dev->dev);
-	if (ret)
-		return ret;
+	variant = (const struct armada_variant *)id->driver_data;
 
 	INIT_WORK(&priv->fb_unref_work, armada_drm_unref_work);
 	INIT_KFIFO(priv->fb_unref);
@@ -155,40 +159,51 @@ static int armada_drm_load(struct drm_device *dev, unsigned long flags)
 
 	/* Create all LCD controllers */
 	for (n = 0; n < ARRAY_SIZE(priv->dcrtc); n++) {
+		int irq;
+
 		if (!res[n])
 			break;
 
-		ret = armada_drm_crtc_create(dev, n, res[n]);
+		irq = platform_get_irq(dev->platformdev, n);
+		if (irq < 0)
+			goto err_kms;
+
+		ret = armada_drm_crtc_create(dev, dev->dev, res[n], irq,
+					     variant, NULL);
 		if (ret)
 			goto err_kms;
 	}
 
+	if (is_componentized(dev->dev)) {
+		ret = component_bind_all(dev->dev, dev);
+		if (ret)
+			goto err_kms;
+	} else {
 #ifdef CONFIG_DRM_ARMADA_TDA1998X
-	ret = armada_drm_connector_slave_create(dev, &tda19988_config);
-	if (ret)
-		goto err_kms;
+		ret = armada_drm_connector_slave_create(dev, &tda19988_config);
+		if (ret)
+			goto err_kms;
 #endif
+	}
 
-	ret = drm_vblank_init(dev, n);
-	if (ret)
-		goto err_kms;
-
-	ret = drm_irq_install(dev, platform_get_irq(dev->platformdev, 0));
+	ret = drm_vblank_init(dev, dev->mode_config.num_crtc);
 	if (ret)
-		goto err_kms;
+		goto err_comp;
 
+	dev->irq_enabled = true;
 	dev->vblank_disable_allowed = 1;
 
 	ret = armada_fbdev_init(dev);
 	if (ret)
-		goto err_irq;
+		goto err_comp;
 
 	drm_kms_helper_poll_init(dev);
 
 	return 0;
 
- err_irq:
-	drm_irq_uninstall(dev);
+ err_comp:
+	if (is_componentized(dev->dev))
+		component_unbind_all(dev->dev, dev);
  err_kms:
 	drm_mode_config_cleanup(dev);
 	drm_mm_takedown(&priv->linear);
@@ -203,7 +218,10 @@ static int armada_drm_unload(struct drm_device *dev)
 
 	drm_kms_helper_poll_fini(dev);
 	armada_fbdev_fini(dev);
-	drm_irq_uninstall(dev);
+
+	if (is_componentized(dev->dev))
+		component_unbind_all(dev->dev, dev);
+
 	drm_mode_config_cleanup(dev);
 	drm_mm_takedown(&priv->linear);
 	flush_work(&priv->fb_unref_work);
@@ -259,52 +277,6 @@ static void armada_drm_disable_vblank(struct drm_device *dev, int crtc)
 	armada_drm_crtc_disable_irq(priv->dcrtc[crtc], VSYNC_IRQ_ENA);
 }
 
-static irqreturn_t armada_drm_irq_handler(int irq, void *arg)
-{
-	struct drm_device *dev = arg;
-	struct armada_private *priv = dev->dev_private;
-	struct armada_crtc *dcrtc = priv->dcrtc[0];
-	uint32_t v, stat = readl_relaxed(dcrtc->base + LCD_SPU_IRQ_ISR);
-	irqreturn_t handled = IRQ_NONE;
-
-	/*
-	 * This is rediculous - rather than writing bits to clear, we
-	 * have to set the actual status register value.  This is racy.
-	 */
-	writel_relaxed(0, dcrtc->base + LCD_SPU_IRQ_ISR);
-
-	/* Mask out those interrupts we haven't enabled */
-	v = stat & dcrtc->irq_ena;
-
-	if (v & (VSYNC_IRQ|GRA_FRAME_IRQ|DUMB_FRAMEDONE)) {
-		armada_drm_crtc_irq(dcrtc, stat);
-		handled = IRQ_HANDLED;
-	}
-
-	return handled;
-}
-
-static int armada_drm_irq_postinstall(struct drm_device *dev)
-{
-	struct armada_private *priv = dev->dev_private;
-	struct armada_crtc *dcrtc = priv->dcrtc[0];
-
-	spin_lock_irq(&dev->vbl_lock);
-	writel_relaxed(dcrtc->irq_ena, dcrtc->base + LCD_SPU_IRQ_ENA);
-	writel(0, dcrtc->base + LCD_SPU_IRQ_ISR);
-	spin_unlock_irq(&dev->vbl_lock);
-
-	return 0;
-}
-
-static void armada_drm_irq_uninstall(struct drm_device *dev)
-{
-	struct armada_private *priv = dev->dev_private;
-	struct armada_crtc *dcrtc = priv->dcrtc[0];
-
-	writel(0, dcrtc->base + LCD_SPU_IRQ_ENA);
-}
-
 static struct drm_ioctl_desc armada_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(ARMADA_GEM_CREATE, armada_gem_create_ioctl,
 		DRM_UNLOCKED),
@@ -337,12 +309,10 @@ static struct drm_driver armada_drm_driver = {
 	.postclose		= NULL,
 	.lastclose		= armada_drm_lastclose,
 	.unload			= armada_drm_unload,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= armada_drm_enable_vblank,
 	.disable_vblank		= armada_drm_disable_vblank,
-	.irq_handler		= armada_drm_irq_handler,
-	.irq_postinstall	= armada_drm_irq_postinstall,
-	.irq_uninstall		= armada_drm_irq_uninstall,
 #ifdef CONFIG_DEBUG_FS
 	.debugfs_init		= armada_drm_debugfs_init,
 	.debugfs_cleanup	= armada_drm_debugfs_cleanup,
@@ -367,14 +337,135 @@ static struct drm_driver armada_drm_driver = {
 	.fops			= &armada_drm_fops,
 };
 
+static int armada_drm_bind(struct device *dev)
+{
+	return drm_platform_init(&armada_drm_driver, to_platform_device(dev));
+}
+
+static void armada_drm_unbind(struct device *dev)
+{
+	drm_put_dev(dev_get_drvdata(dev));
+}
+
+static int compare_of(struct device *dev, void *data)
+{
+	return dev->of_node == data;
+}
+
+static int compare_dev_name(struct device *dev, void *data)
+{
+	const char *name = data;
+	return !strcmp(dev_name(dev), name);
+}
+
+static void armada_add_endpoints(struct device *dev,
+	struct component_match **match, struct device_node *port)
+{
+	struct device_node *ep, *remote;
+
+	for_each_child_of_node(port, ep) {
+		remote = of_graph_get_remote_port_parent(ep);
+		if (!remote || !of_device_is_available(remote)) {
+			of_node_put(remote);
+			continue;
+		} else if (!of_device_is_available(remote->parent)) {
+			dev_warn(dev, "parent device of %s is not available\n",
+				 remote->full_name);
+			of_node_put(remote);
+			continue;
+		}
+
+		component_match_add(dev, match, compare_of, remote);
+		of_node_put(remote);
+	}
+}
+
+static int armada_drm_find_components(struct device *dev,
+	struct component_match **match)
+{
+	struct device_node *port;
+	int i;
+
+	if (dev->of_node) {
+		struct device_node *np = dev->of_node;
+
+		for (i = 0; ; i++) {
+			port = of_parse_phandle(np, "ports", i);
+			if (!port)
+				break;
+
+			component_match_add(dev, match, compare_of, port);
+			of_node_put(port);
+		}
+
+		if (i == 0) {
+			dev_err(dev, "missing 'ports' property\n");
+			return -ENODEV;
+		}
+
+		for (i = 0; ; i++) {
+			port = of_parse_phandle(np, "ports", i);
+			if (!port)
+				break;
+
+			armada_add_endpoints(dev, match, port);
+			of_node_put(port);
+		}
+	} else if (dev->platform_data) {
+		char **devices = dev->platform_data;
+		struct device *d;
+
+		for (i = 0; devices[i]; i++)
+			component_match_add(dev, match, compare_dev_name,
+					    devices[i]);
+
+		if (i == 0) {
+			dev_err(dev, "missing 'ports' property\n");
+			return -ENODEV;
+		}
+
+		for (i = 0; devices[i]; i++) {
+			d = bus_find_device_by_name(&platform_bus_type, NULL,
+					devices[i]);
+			if (d && d->of_node) {
+				for_each_child_of_node(d->of_node, port)
+					armada_add_endpoints(dev, match, port);
+			}
+			put_device(d);
+		}
+	}
+
+	return 0;
+}
+
+static const struct component_master_ops armada_master_ops = {
+	.bind = armada_drm_bind,
+	.unbind = armada_drm_unbind,
+};
+
 static int armada_drm_probe(struct platform_device *pdev)
 {
-	return drm_platform_init(&armada_drm_driver, pdev);
+	if (is_componentized(&pdev->dev)) {
+		struct component_match *match = NULL;
+		int ret;
+
+		ret = armada_drm_find_components(&pdev->dev, &match);
+		if (ret < 0)
+			return ret;
+
+		return component_master_add_with_match(&pdev->dev,
+				&armada_master_ops, match);
+	} else {
+		return drm_platform_init(&armada_drm_driver, pdev);
+	}
 }
 
 static int armada_drm_remove(struct platform_device *pdev)
 {
-	drm_put_dev(platform_get_drvdata(pdev));
+	if (is_componentized(&pdev->dev))
+		component_master_del(&pdev->dev, &armada_master_ops);
+	else
+		drm_put_dev(platform_get_drvdata(pdev));
 	return 0;
 }
 
@@ -402,14 +493,24 @@ static struct platform_driver armada_drm_platform_driver = {
 
 static int __init armada_drm_init(void)
 {
+	int ret;
+
 	armada_drm_driver.num_ioctls = ARRAY_SIZE(armada_ioctls);
-	return platform_driver_register(&armada_drm_platform_driver);
+
+	ret = platform_driver_register(&armada_lcd_platform_driver);
+	if (ret)
+		return ret;
+	ret = platform_driver_register(&armada_drm_platform_driver);
+	if (ret)
+		platform_driver_unregister(&armada_lcd_platform_driver);
+	return ret;
 }
 module_init(armada_drm_init);
 
 static void __exit armada_drm_exit(void)
 {
 	platform_driver_unregister(&armada_drm_platform_driver);
+	platform_driver_unregister(&armada_lcd_platform_driver);
 }
 module_exit(armada_drm_exit);
 
diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c
index fd166f532ab9..7838e731b0de 100644
--- a/drivers/gpu/drm/armada/armada_fbdev.c
+++ b/drivers/gpu/drm/armada/armada_fbdev.c
@@ -131,7 +131,7 @@ static int armada_fb_probe(struct drm_fb_helper *fbh,
 	return ret;
 }
 
-static struct drm_fb_helper_funcs armada_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs armada_fb_helper_funcs = {
 	.gamma_set	= armada_drm_crtc_gamma_set,
 	.gamma_get	= armada_drm_crtc_gamma_get,
 	.fb_probe	= armada_fb_probe,
@@ -149,7 +149,7 @@ int armada_fbdev_init(struct drm_device *dev)
 
 	priv->fbdev = fbh;
 
-	fbh->funcs = &armada_fb_helper_funcs;
+	drm_fb_helper_prepare(dev, fbh, &armada_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(dev, fbh, 1, 1);
 	if (ret) {
diff --git a/drivers/gpu/drm/armada/armada_gem.h b/drivers/gpu/drm/armada/armada_gem.h
index 00b6cd461a03..b000ea3a829a 100644
--- a/drivers/gpu/drm/armada/armada_gem.h
+++ b/drivers/gpu/drm/armada/armada_gem.h
@@ -8,6 +8,8 @@
 #ifndef ARMADA_GEM_H
 #define ARMADA_GEM_H
 
+#include <drm/drm_gem.h>
+
 /* GEM */
 struct armada_gem_object {
 	struct drm_gem_object	obj;
diff --git a/drivers/gpu/drm/armada/armada_output.c b/drivers/gpu/drm/armada/armada_output.c
index d685a5421485..abbc309fe539 100644
--- a/drivers/gpu/drm/armada/armada_output.c
+++ b/drivers/gpu/drm/armada/armada_output.c
@@ -48,7 +48,7 @@ static void armada_drm_connector_destroy(struct drm_connector *conn)
 {
 	struct armada_connector *dconn = drm_to_armada_conn(conn);
 
-	drm_sysfs_connector_remove(conn);
+	drm_connector_unregister(conn);
 	drm_connector_cleanup(conn);
 	kfree(dconn);
 }
@@ -141,7 +141,7 @@ int armada_output_create(struct drm_device *dev,
 	if (ret)
 		goto err_conn;
 
-	ret = drm_sysfs_connector_add(&dconn->conn);
+	ret = drm_connector_register(&dconn->conn);
 	if (ret)
 		goto err_sysfs;
 
diff --git a/drivers/gpu/drm/ast/ast_dp501.c b/drivers/gpu/drm/ast/ast_dp501.c
index 5da4b62285fa..76f07f38b941 100644
--- a/drivers/gpu/drm/ast/ast_dp501.c
+++ b/drivers/gpu/drm/ast/ast_dp501.c
@@ -379,11 +379,39 @@ static bool ast_init_dvo(struct drm_device *dev)
 	return true;
 }
 
+
+static void ast_init_analog(struct drm_device *dev)
+{
+	struct ast_private *ast = dev->dev_private;
+	u32 data;
+
+	/*
+	 * Set DAC source to VGA mode in SCU2C via the P2A
+	 * bridge. First configure the P2U to target the SCU
+	 * in case it isn't at this stage.
+	 */
+	ast_write32(ast, 0xf004, 0x1e6e0000);
+	ast_write32(ast, 0xf000, 0x1);
+
+	/* Then unlock the SCU with the magic password */
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+	ast_write32(ast, 0x12000, 0x1688a8a8);
+
+	/* Finally, clear bits [17:16] of SCU2c */
+	data = ast_read32(ast, 0x1202c);
+	data &= 0xfffcffff;
+	ast_write32(ast, 0, data);
+
+	/* Disable DVO */
+	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xcf, 0x00);
+}
+
 void ast_init_3rdtx(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 	u8 jreg;
-	u32 data;
+
 	if (ast->chip == AST2300 || ast->chip == AST2400) {
 		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xd1, 0xff);
 		switch (jreg & 0x0e) {
@@ -399,12 +427,8 @@ void ast_init_3rdtx(struct drm_device *dev)
 		default:
 			if (ast->tx_chip_type == AST_TX_SIL164)
 				ast_init_dvo(dev);
-			else {
-				ast_write32(ast, 0x12000, 0x1688a8a8);
-				data = ast_read32(ast, 0x1202c);
-				data &= 0xfffcffff;
-				ast_write32(ast, 0, data);
-			}
+			else
+				ast_init_analog(dev);
 		}
 	}
 }
diff --git a/drivers/gpu/drm/ast/ast_drv.c b/drivers/gpu/drm/ast/ast_drv.c
index 44074fbcf7ff..9a32d9dfdd26 100644
--- a/drivers/gpu/drm/ast/ast_drv.c
+++ b/drivers/gpu/drm/ast/ast_drv.c
@@ -51,7 +51,7 @@ static struct drm_driver driver;
 	.subdevice = PCI_ANY_ID,		\
 	.driver_data = (unsigned long) info }
 
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+static const struct pci_device_id pciidlist[] = {
 	AST_VGA_DEVICE(PCI_CHIP_AST2000, NULL),
 	AST_VGA_DEVICE(PCI_CHIP_AST2100, NULL),
 	/*	AST_VGA_DEVICE(PCI_CHIP_AST1180, NULL), - don't bind to 1180 for now */
@@ -199,6 +199,7 @@ static struct drm_driver driver = {
 
 	.load = ast_driver_load,
 	.unload = ast_driver_unload,
+	.set_busid = drm_pci_set_busid,
 
 	.fops = &ast_fops,
 	.name = DRIVER_NAME,
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 5d6a87573c33..86205a28e56b 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -36,6 +36,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 
@@ -125,8 +127,9 @@ struct ast_gem_object;
 
 #define AST_IO_AR_PORT_WRITE		(0x40)
 #define AST_IO_MISC_PORT_WRITE		(0x42)
+#define AST_IO_VGA_ENABLE_PORT		(0x43)
 #define AST_IO_SEQ_PORT			(0x44)
-#define AST_DAC_INDEX_READ		(0x3c7)
+#define AST_IO_DAC_INDEX_READ		(0x47)
 #define AST_IO_DAC_INDEX_WRITE		(0x48)
 #define AST_IO_DAC_DATA		        (0x49)
 #define AST_IO_GR_PORT			(0x4E)
@@ -134,6 +137,8 @@ struct ast_gem_object;
 #define AST_IO_INPUT_STATUS1_READ	(0x5A)
 #define AST_IO_MISC_PORT_READ		(0x4C)
 
+#define AST_IO_MM_OFFSET		(0x380)
+
 #define __ast_read(x) \
 static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
 u##x val = 0;\
@@ -316,7 +321,7 @@ struct ast_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_ast_bo(gobj) container_of((gobj), struct ast_bo, gem)
@@ -362,7 +367,7 @@ static inline int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
 {
 	int ret;
 
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, NULL);
 	if (ret) {
 		if (ret != -ERESTARTSYS && ret != -EBUSY)
 			DRM_ERROR("reserve failed %p\n", bo);
@@ -381,6 +386,9 @@ int ast_bo_push_sysram(struct ast_bo *bo);
 int ast_mmap(struct file *filp, struct vm_area_struct *vma);
 
 /* ast post */
+void ast_enable_vga(struct drm_device *dev);
+void ast_enable_mmio(struct drm_device *dev);
+bool ast_is_vga_enabled(struct drm_device *dev);
 void ast_post_gpu(struct drm_device *dev);
 u32 ast_mindwm(struct ast_private *ast, u32 r);
 void ast_moutdwm(struct ast_private *ast, u32 r, u32 v);
diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index a28640f47c27..5c60ae524c45 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -186,7 +186,8 @@ static int astfb_create_object(struct ast_fbdev *afbdev,
 static int astfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
-	struct ast_fbdev *afbdev = (struct ast_fbdev *)helper;
+	struct ast_fbdev *afbdev =
+		container_of(helper, struct ast_fbdev, helper);
 	struct drm_device *dev = afbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct drm_framebuffer *fb;
@@ -287,7 +288,7 @@ static void ast_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 	*blue = ast_crtc->lut_b[regno] << 8;
 }
 
-static struct drm_fb_helper_funcs ast_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs ast_fb_helper_funcs = {
 	.gamma_set = ast_fb_gamma_set,
 	.gamma_get = ast_fb_gamma_get,
 	.fb_probe = astfb_create,
@@ -328,8 +329,10 @@ int ast_fbdev_init(struct drm_device *dev)
 		return -ENOMEM;
 
 	ast->fbdev = afbdev;
-	afbdev->helper.funcs = &ast_fb_helper_funcs;
 	spin_lock_init(&afbdev->dirty_lock);
+
+	drm_fb_helper_prepare(dev, &afbdev->helper, &ast_fb_helper_funcs);
+
 	ret = drm_fb_helper_init(dev, &afbdev->helper,
 				 1, 1);
 	if (ret) {
diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c
index a2cc6be97983..035dacc93382 100644
--- a/drivers/gpu/drm/ast/ast_main.c
+++ b/drivers/gpu/drm/ast/ast_main.c
@@ -63,10 +63,11 @@ uint8_t ast_get_index_reg_mask(struct ast_private *ast,
 }
 
 
-static int ast_detect_chip(struct drm_device *dev)
+static int ast_detect_chip(struct drm_device *dev, bool *need_post)
 {
 	struct ast_private *ast = dev->dev_private;
 	uint32_t data, jreg;
+	ast_open_key(ast);
 
 	if (dev->pdev->device == PCI_CHIP_AST1180) {
 		ast->chip = AST1100;
@@ -104,11 +105,26 @@ static int ast_detect_chip(struct drm_device *dev)
 			}
 			ast->vga2_clone = false;
 		} else {
-			ast->chip = 2000;
+			ast->chip = AST2000;
 			DRM_INFO("AST 2000 detected\n");
 		}
 	}
 
+	/*
+	 * If VGA isn't enabled, we need to enable now or subsequent
+	 * access to the scratch registers will fail. We also inform
+	 * our caller that it needs to POST the chip
+	 * (Assumption: VGA not enabled -> need to POST)
+	 */
+	if (!ast_is_vga_enabled(dev)) {
+		ast_enable_vga(dev);
+		ast_enable_mmio(dev);
+		DRM_INFO("VGA not enabled on entry, requesting chip POST\n");
+		*need_post = true;
+	} else
+		*need_post = false;
+
+	/* Check if we support wide screen */
 	switch (ast->chip) {
 	case AST1180:
 		ast->support_wide_screen = true;
@@ -124,6 +140,7 @@ static int ast_detect_chip(struct drm_device *dev)
 			ast->support_wide_screen = true;
 		else {
 			ast->support_wide_screen = false;
+			/* Read SCU7c (silicon revision register) */
 			ast_write32(ast, 0xf004, 0x1e6e0000);
 			ast_write32(ast, 0xf000, 0x1);
 			data = ast_read32(ast, 0x1207c);
@@ -136,11 +153,29 @@ static int ast_detect_chip(struct drm_device *dev)
 		break;
 	}
 
+	/* Check 3rd Tx option (digital output afaik) */
 	ast->tx_chip_type = AST_TX_NONE;
-	jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xff);
-	if (jreg & 0x80)
-		ast->tx_chip_type = AST_TX_SIL164;
+
+	/*
+	 * VGACRA3 Enhanced Color Mode Register, check if DVO is already
+	 * enabled, in that case, assume we have a SIL164 TMDS transmitter
+	 *
+	 * Don't make that assumption if we the chip wasn't enabled and
+	 * is at power-on reset, otherwise we'll incorrectly "detect" a
+	 * SIL164 when there is none.
+	 */
+	if (!*need_post) {
+		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa3, 0xff);
+		if (jreg & 0x80)
+			ast->tx_chip_type = AST_TX_SIL164;
+	}
+
 	if ((ast->chip == AST2300) || (ast->chip == AST2400)) {
+		/*
+		 * On AST2300 and 2400, look the configuration set by the SoC in
+		 * the SOC scratch register #1 bits 11:8 (interestingly marked
+		 * as "reserved" in the spec)
+		 */
 		jreg = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xd1, 0xff);
 		switch (jreg) {
 		case 0x04:
@@ -161,6 +196,17 @@ static int ast_detect_chip(struct drm_device *dev)
 		}
 	}
 
+	/* Print stuff for diagnostic purposes */
+	switch(ast->tx_chip_type) {
+	case AST_TX_SIL164:
+		DRM_INFO("Using Sil164 TMDS transmitter\n");
+		break;
+	case AST_TX_DP501:
+		DRM_INFO("Using DP501 DisplayPort transmitter\n");
+		break;
+	default:
+		DRM_INFO("Analog VGA only\n");
+	}
 	return 0;
 }
 
@@ -345,6 +391,7 @@ static u32 ast_get_vram_info(struct drm_device *dev)
 int ast_driver_load(struct drm_device *dev, unsigned long flags)
 {
 	struct ast_private *ast;
+	bool need_post;
 	int ret = 0;
 
 	ast = kzalloc(sizeof(struct ast_private), GFP_KERNEL);
@@ -359,13 +406,27 @@ int ast_driver_load(struct drm_device *dev, unsigned long flags)
 		ret = -EIO;
 		goto out_free;
 	}
-	ast->ioregs = pci_iomap(dev->pdev, 2, 0);
+
+	/*
+	 * If we don't have IO space at all, use MMIO now and
+	 * assume the chip has MMIO enabled by default (rev 0x20
+	 * and higher).
+	 */
+	if (!(pci_resource_flags(dev->pdev, 2) & IORESOURCE_IO)) {
+		DRM_INFO("platform has no IO space, trying MMIO\n");
+		ast->ioregs = ast->regs + AST_IO_MM_OFFSET;
+	}
+
+	/* "map" IO regs if the above hasn't done so already */
 	if (!ast->ioregs) {
-		ret = -EIO;
-		goto out_free;
+		ast->ioregs = pci_iomap(dev->pdev, 2, 0);
+		if (!ast->ioregs) {
+			ret = -EIO;
+			goto out_free;
+		}
 	}
 
-	ast_detect_chip(dev);
+	ast_detect_chip(dev, &need_post);
 
 	if (ast->chip != AST1180) {
 		ast_get_dram_info(dev);
@@ -373,6 +434,9 @@ int ast_driver_load(struct drm_device *dev, unsigned long flags)
 		DRM_INFO("dram %d %d %d %08x\n", ast->mclk, ast->dram_type, ast->dram_bus_width, ast->vram_size);
 	}
 
+	if (need_post)
+		ast_post_gpu(dev);
+
 	ret = ast_mm_init(ast);
 	if (ret)
 		goto out_free;
diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index 114aee941d46..9dc0fd5c1ea4 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -80,6 +80,8 @@ static bool ast_get_vbios_mode_info(struct drm_crtc *crtc, struct drm_display_mo
 	struct ast_private *ast = crtc->dev->dev_private;
 	u32 refresh_rate_index = 0, mode_id, color_index, refresh_rate;
 	u32 hborder, vborder;
+	bool check_sync;
+	struct ast_vbios_enhtable *best = NULL;
 
 	switch (crtc->primary->fb->bits_per_pixel) {
 	case 8:
@@ -141,14 +143,34 @@ static bool ast_get_vbios_mode_info(struct drm_crtc *crtc, struct drm_display_mo
 	}
 
 	refresh_rate = drm_mode_vrefresh(mode);
-	while (vbios_mode->enh_table->refresh_rate < refresh_rate) {
-		vbios_mode->enh_table++;
-		if ((vbios_mode->enh_table->refresh_rate > refresh_rate) ||
-		    (vbios_mode->enh_table->refresh_rate == 0xff)) {
-			vbios_mode->enh_table--;
-			break;
+	check_sync = vbios_mode->enh_table->flags & WideScreenMode;
+	do {
+		struct ast_vbios_enhtable *loop = vbios_mode->enh_table;
+
+		while (loop->refresh_rate != 0xff) {
+			if ((check_sync) &&
+			    (((mode->flags & DRM_MODE_FLAG_NVSYNC)  &&
+			      (loop->flags & PVSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_PVSYNC)  &&
+			      (loop->flags & NVSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_NHSYNC)  &&
+			      (loop->flags & PHSync))  ||
+			     ((mode->flags & DRM_MODE_FLAG_PHSYNC)  &&
+			      (loop->flags & NHSync)))) {
+				loop++;
+				continue;
+			}
+			if (loop->refresh_rate <= refresh_rate
+			    && (!best || loop->refresh_rate > best->refresh_rate))
+				best = loop;
+			loop++;
 		}
-	}
+		if (best || !check_sync)
+			break;
+		check_sync = 0;
+	} while (1);
+	if (best)
+		vbios_mode->enh_table = best;
 
 	hborder = (vbios_mode->enh_table->flags & HBorder) ? 8 : 0;
 	vborder = (vbios_mode->enh_table->flags & VBorder) ? 8 : 0;
@@ -419,8 +441,10 @@ static void ast_set_sync_reg(struct drm_device *dev, struct drm_display_mode *mo
 	struct ast_private *ast = dev->dev_private;
 	u8 jreg;
 
-	jreg = ast_io_read8(ast, AST_IO_MISC_PORT_READ);
-	jreg |= (vbios_mode->enh_table->flags & SyncNN);
+	jreg  = ast_io_read8(ast, AST_IO_MISC_PORT_READ);
+	jreg &= ~0xC0;
+	if (vbios_mode->enh_table->flags & NVSync) jreg |= 0x80;
+	if (vbios_mode->enh_table->flags & NHSync) jreg |= 0x40;
 	ast_io_write8(ast, AST_IO_MISC_PORT_WRITE, jreg);
 }
 
@@ -667,17 +691,9 @@ static void ast_encoder_destroy(struct drm_encoder *encoder)
 static struct drm_encoder *ast_best_single_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -829,7 +845,7 @@ static void ast_connector_destroy(struct drm_connector *connector)
 {
 	struct ast_connector *ast_connector = to_ast_connector(connector);
 	ast_i2c_destroy(ast_connector->i2c);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -871,7 +887,7 @@ static int ast_connector_init(struct drm_device *dev)
 	connector->interlace_allowed = 0;
 	connector->doublescan_allowed = 0;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 
@@ -1088,8 +1104,8 @@ static u32 copy_cursor_image(u8 *src, u8 *dst, int width, int height)
 			srcdata32[1].ul = *((u32 *)(srcxor + 4)) & 0xf0f0f0f0;
 			data32.b[0] = srcdata32[0].b[1] | (srcdata32[0].b[0] >> 4);
 			data32.b[1] = srcdata32[0].b[3] | (srcdata32[0].b[2] >> 4);
-			data32.b[2] = srcdata32[0].b[1] | (srcdata32[1].b[0] >> 4);
-			data32.b[3] = srcdata32[0].b[3] | (srcdata32[1].b[2] >> 4);
+			data32.b[2] = srcdata32[1].b[1] | (srcdata32[1].b[0] >> 4);
+			data32.b[3] = srcdata32[1].b[3] | (srcdata32[1].b[2] >> 4);
 
 			writel(data32.ul, dstxor);
 			csum += data32.ul;
diff --git a/drivers/gpu/drm/ast/ast_post.c b/drivers/gpu/drm/ast/ast_post.c
index 38d437f3a267..810c51d92b99 100644
--- a/drivers/gpu/drm/ast/ast_post.c
+++ b/drivers/gpu/drm/ast/ast_post.c
@@ -33,18 +33,23 @@
 
 static void ast_init_dram_2300(struct drm_device *dev);
 
-static void
-ast_enable_vga(struct drm_device *dev)
+void ast_enable_vga(struct drm_device *dev)
+{
+	struct ast_private *ast = dev->dev_private;
+
+	ast_io_write8(ast, AST_IO_VGA_ENABLE_PORT, 0x01);
+	ast_io_write8(ast, AST_IO_MISC_PORT_WRITE, 0x01);
+}
+
+void ast_enable_mmio(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 
-	ast_io_write8(ast, 0x43, 0x01);
-	ast_io_write8(ast, 0x42, 0x01);
+	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xa1, 0xff, 0x04);
 }
 
-#if 0 /* will use later */
-static bool
-ast_is_vga_enabled(struct drm_device *dev)
+
+bool ast_is_vga_enabled(struct drm_device *dev)
 {
 	struct ast_private *ast = dev->dev_private;
 	u8 ch;
@@ -52,7 +57,7 @@ ast_is_vga_enabled(struct drm_device *dev)
 	if (ast->chip == AST1180) {
 		/* TODO 1180 */
 	} else {
-		ch = ast_io_read8(ast, 0x43);
+		ch = ast_io_read8(ast, AST_IO_VGA_ENABLE_PORT);
 		if (ch) {
 			ast_open_key(ast);
 			ch = ast_get_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff);
@@ -61,7 +66,6 @@ ast_is_vga_enabled(struct drm_device *dev)
 	}
 	return 0;
 }
-#endif
 
 static const u8 extreginfo[] = { 0x0f, 0x04, 0x1c, 0xff };
 static const u8 extreginfo_ast2300a0[] = { 0x0f, 0x04, 0x1c, 0xff };
@@ -371,6 +375,7 @@ void ast_post_gpu(struct drm_device *dev)
 	pci_write_config_dword(ast->dev->pdev, 0x04, reg);
 
 	ast_enable_vga(dev);
+	ast_enable_mmio(dev);
 	ast_open_key(ast);
 	ast_set_def_ext_reg(dev);
 
diff --git a/drivers/gpu/drm/ast/ast_tables.h b/drivers/gpu/drm/ast/ast_tables.h
index 4c761dcea972..3608d5aa7451 100644
--- a/drivers/gpu/drm/ast/ast_tables.h
+++ b/drivers/gpu/drm/ast/ast_tables.h
@@ -35,14 +35,18 @@
 #define HalfDCLK                0x00000002
 #define DoubleScanMode          0x00000004
 #define LineCompareOff          0x00000008
-#define SyncPP                  0x00000000
-#define SyncPN                  0x00000040
-#define SyncNP                  0x00000080
-#define SyncNN                  0x000000C0
 #define HBorder                 0x00000020
 #define VBorder                 0x00000010
 #define WideScreenMode		0x00000100
 #define NewModeInfo		0x00000200
+#define NHSync			0x00000400
+#define PHSync			0x00000800
+#define NVSync			0x00001000
+#define PVSync			0x00002000
+#define SyncPP			(PVSync | PHSync)
+#define SyncPN			(PVSync | NHSync)
+#define SyncNP			(NVSync | PHSync)
+#define SyncNN			(NVSync | NHSync)
 
 /* DCLK Index */
 #define VCLK25_175     		0x00
@@ -72,6 +76,7 @@
 #define VCLK119     		0x17
 #define VCLK85_5     		0x18
 #define VCLK97_75     		0x19
+#define VCLK118_25			0x1A
 
 static struct ast_vbios_dclk_info dclk_table[] = {
 	{0x2C, 0xE7, 0x03},					/* 00: VCLK25_175	*/
@@ -99,6 +104,8 @@ static struct ast_vbios_dclk_info dclk_table[] = {
 	{0x25, 0x65, 0x80},					/* 16: VCLK88.75    */
 	{0x77, 0x58, 0x80},					/* 17: VCLK119      */
 	{0x32, 0x67, 0x80},				    /* 18: VCLK85_5     */
+	{0x6a, 0x6d, 0x80},					/* 19: VCLK97_75	*/
+	{0x3b, 0x2c, 0x81},					/* 1A: VCLK118_25	*/
 };
 
 static struct ast_vbios_stdtable vbios_stdtable[] = {
@@ -245,8 +252,10 @@ static struct ast_vbios_enhtable res_1360x768[] = {
 static struct ast_vbios_enhtable res_1600x900[] = {
 	{1760, 1600, 48, 32, 926,  900, 3, 5, VCLK97_75,	/* 60Hz CVT RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x3A },
-	{1760, 1600, 48, 32, 926,  900, 3, 5, VCLK97_75,	/* end */
-	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x3A }
+	{2112, 1600, 88,168, 934,  900, 3, 5, VCLK118_25,	/* 60Hz CVT */
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x3A },
+	{2112, 1600, 88,168, 934,  900, 3, 5, VCLK118_25,	/* 60Hz CVT */
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x3A },
 };
 
 static struct ast_vbios_enhtable res_1920x1080[] = {
@@ -260,11 +269,11 @@ static struct ast_vbios_enhtable res_1920x1080[] = {
 /* 16:10 */
 static struct ast_vbios_enhtable res_1280x800[] = {
 	{1440, 1280, 48, 32,  823,  800, 3, 6, VCLK71,	/* 60Hz RB */
-	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 35 },
+	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x35 },
 	{1680, 1280, 72,128,  831,  800, 3, 6, VCLK83_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x35 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x35 },
 	{1680, 1280, 72,128,  831,  800, 3, 6, VCLK83_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x35 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x35 },
 
 };
 
@@ -272,24 +281,24 @@ static struct ast_vbios_enhtable res_1440x900[] = {
 	{1600, 1440, 48, 32,  926,  900, 3, 6, VCLK88_75,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x36 },
 	{1904, 1440, 80,152,  934,  900, 3, 6, VCLK106_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x36 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x36 },
 	{1904, 1440, 80,152,  934,  900, 3, 6, VCLK106_5,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x36 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x36 },
 };
 
 static struct ast_vbios_enhtable res_1680x1050[] = {
 	{1840, 1680, 48, 32, 1080, 1050, 3, 6, VCLK119,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x37 },
 	{2240, 1680,104,176, 1089, 1050, 3, 6, VCLK146_25,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x37 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 2, 0x37 },
 	{2240, 1680,104,176, 1089, 1050, 3, 6, VCLK146_25,	/* 60Hz */
-	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x37 },
+	 (SyncPN | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 2, 0x37 },
 };
 
 static struct ast_vbios_enhtable res_1920x1200[] = {
-	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz */
+	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz RB*/
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 60, 1, 0x34 },
-	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz */
+	{2080, 1920, 48, 32, 1235, 1200, 3, 6, VCLK154,	/* 60Hz RB */
 	 (SyncNP | Charx8Dot | LineCompareOff | WideScreenMode | NewModeInfo), 0xFF, 1, 0x34 },
 };
 
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index b8246227bab0..08f82eae6939 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -293,18 +293,22 @@ void ast_mm_fini(struct ast_private *ast)
 void ast_ttm_placement(struct ast_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
+
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int ast_bo_create(struct drm_device *dev, int size, int align,
@@ -335,7 +339,7 @@ int ast_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&ast->ttm.bdev, &astbo->bo, size,
 			  ttm_bo_type_device, &astbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, ast_bo_ttm_destroy);
+			  NULL, NULL, ast_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -360,7 +364,7 @@ int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	ast_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -383,7 +387,7 @@ int ast_bo_unpin(struct ast_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -407,7 +411,7 @@ int ast_bo_push_sysram(struct ast_bo *bo)
 
 	ast_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -423,7 +427,7 @@ int ast_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct ast_private *ast;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	ast = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/ati_pcigart.c b/drivers/gpu/drm/ati_pcigart.c
index c399dea27a3b..6c4d4b6eba80 100644
--- a/drivers/gpu/drm/ati_pcigart.c
+++ b/drivers/gpu/drm/ati_pcigart.c
@@ -34,6 +34,8 @@
 #include <linux/export.h>
 #include <drm/drmP.h>
 
+#include <drm/ati_pcigart.h>
+
 # define ATI_PCIGART_PAGE_SIZE		4096	/**< PCI GART page size */
 
 static int drm_ati_alloc_pcigart_table(struct drm_device *dev,
diff --git a/drivers/gpu/drm/bochs/bochs.h b/drivers/gpu/drm/bochs/bochs.h
index 7eb52dd44b01..71f2687fc3cc 100644
--- a/drivers/gpu/drm/bochs/bochs.h
+++ b/drivers/gpu/drm/bochs/bochs.h
@@ -7,6 +7,8 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
 
+#include <drm/drm_gem.h>
+
 #include <ttm/ttm_bo_driver.h>
 #include <ttm/ttm_page_alloc.h>
 
@@ -99,7 +101,7 @@ struct bochs_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 
diff --git a/drivers/gpu/drm/bochs/bochs_drv.c b/drivers/gpu/drm/bochs/bochs_drv.c
index 9c13df29fd20..98837bde2d25 100644
--- a/drivers/gpu/drm/bochs/bochs_drv.c
+++ b/drivers/gpu/drm/bochs/bochs_drv.c
@@ -82,6 +82,7 @@ static struct drm_driver bochs_driver = {
 	.driver_features	= DRIVER_GEM | DRIVER_MODESET,
 	.load			= bochs_load,
 	.unload			= bochs_unload,
+	.set_busid		= drm_pci_set_busid,
 	.fops			= &bochs_fops,
 	.name			= "bochs-drm",
 	.desc			= "bochs dispi vga interface (qemu stdvga)",
@@ -97,6 +98,7 @@ static struct drm_driver bochs_driver = {
 /* ---------------------------------------------------------------------- */
 /* pm interface                                                           */
 
+#ifdef CONFIG_PM_SLEEP
 static int bochs_pm_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
@@ -131,6 +133,7 @@ static int bochs_pm_resume(struct device *dev)
 	drm_kms_helper_poll_enable(drm_dev);
 	return 0;
 }
+#endif
 
 static const struct dev_pm_ops bochs_pm_ops = {
 	SET_SYSTEM_SLEEP_PM_OPS(bochs_pm_suspend,
@@ -175,7 +178,7 @@ static void bochs_pci_remove(struct pci_dev *pdev)
 	drm_put_dev(dev);
 }
 
-static DEFINE_PCI_DEVICE_TABLE(bochs_pci_tbl) = {
+static const struct pci_device_id bochs_pci_tbl[] = {
 	{
 		.vendor      = 0x1234,
 		.device      = 0x1111,
diff --git a/drivers/gpu/drm/bochs/bochs_fbdev.c b/drivers/gpu/drm/bochs/bochs_fbdev.c
index 561b84474122..fe95d31cd110 100644
--- a/drivers/gpu/drm/bochs/bochs_fbdev.c
+++ b/drivers/gpu/drm/bochs/bochs_fbdev.c
@@ -72,7 +72,7 @@ static int bochsfb_create(struct drm_fb_helper *helper,
 
 	bo = gem_to_bochs_bo(gobj);
 
-	ret = ttm_bo_reserve(&bo->bo, true, false, false, 0);
+	ret = ttm_bo_reserve(&bo->bo, true, false, false, NULL);
 	if (ret)
 		return ret;
 
@@ -179,7 +179,7 @@ void bochs_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 	*blue  = regno;
 }
 
-static struct drm_fb_helper_funcs bochs_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs bochs_fb_helper_funcs = {
 	.gamma_set = bochs_fb_gamma_set,
 	.gamma_get = bochs_fb_gamma_get,
 	.fb_probe = bochsfb_create,
@@ -189,7 +189,8 @@ int bochs_fbdev_init(struct bochs_device *bochs)
 {
 	int ret;
 
-	bochs->fb.helper.funcs = &bochs_fb_helper_funcs;
+	drm_fb_helper_prepare(bochs->dev, &bochs->fb.helper,
+			      &bochs_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(bochs->dev, &bochs->fb.helper,
 				 1, 1);
diff --git a/drivers/gpu/drm/bochs/bochs_kms.c b/drivers/gpu/drm/bochs/bochs_kms.c
index dcf2e55f4ae9..6b7efcf363d6 100644
--- a/drivers/gpu/drm/bochs/bochs_kms.c
+++ b/drivers/gpu/drm/bochs/bochs_kms.c
@@ -53,7 +53,7 @@ static int bochs_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	if (old_fb) {
 		bochs_fb = to_bochs_framebuffer(old_fb);
 		bo = gem_to_bochs_bo(bochs_fb->obj);
-		ret = ttm_bo_reserve(&bo->bo, true, false, false, 0);
+		ret = ttm_bo_reserve(&bo->bo, true, false, false, NULL);
 		if (ret) {
 			DRM_ERROR("failed to reserve old_fb bo\n");
 		} else {
@@ -67,7 +67,7 @@ static int bochs_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 
 	bochs_fb = to_bochs_framebuffer(crtc->primary->fb);
 	bo = gem_to_bochs_bo(bochs_fb->obj);
-	ret = ttm_bo_reserve(&bo->bo, true, false, false, 0);
+	ret = ttm_bo_reserve(&bo->bo, true, false, false, NULL);
 	if (ret)
 		return ret;
 
@@ -216,18 +216,9 @@ static struct drm_encoder *
 bochs_connector_best_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id,
-					   DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -259,6 +250,7 @@ static void bochs_connector_init(struct drm_device *dev)
 			   DRM_MODE_CONNECTOR_VIRTUAL);
 	drm_connector_helper_add(connector,
 				 &bochs_connector_connector_helper_funcs);
+	drm_connector_register(connector);
 }
 
 
diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c
index b9a695d92792..66286ff518d4 100644
--- a/drivers/gpu/drm/bochs/bochs_mm.c
+++ b/drivers/gpu/drm/bochs/bochs_mm.c
@@ -257,20 +257,26 @@ void bochs_mm_fini(struct bochs_device *bochs)
 
 static void bochs_ttm_placement(struct bochs_bo *bo, int domain)
 {
+	unsigned i;
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM) {
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED
+		bo->placements[c++].flags = TTM_PL_FLAG_WC
+			| TTM_PL_FLAG_UNCACHED
 			| TTM_PL_FLAG_VRAM;
 	}
 	if (domain & TTM_PL_FLAG_SYSTEM) {
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING
+			| TTM_PL_FLAG_SYSTEM;
 	}
 	if (!c) {
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING
+			| TTM_PL_FLAG_SYSTEM;
+	}
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
 	}
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
@@ -294,7 +300,7 @@ int bochs_bo_pin(struct bochs_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	bochs_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -319,7 +325,7 @@ int bochs_bo_unpin(struct bochs_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -333,7 +339,7 @@ int bochs_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct bochs_device *bochs;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	bochs = file_priv->minor->dev->dev_private;
@@ -371,7 +377,7 @@ static int bochs_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&bochs->ttm.bdev, &bochsbo->bo, size,
 			  ttm_bo_type_device, &bochsbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, bochs_bo_ttm_destroy);
+			  NULL, NULL, bochs_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -387,7 +393,7 @@ int bochs_gem_create(struct drm_device *dev, u32 size, bool iskernel,
 
 	*obj = NULL;
 
-	size = ALIGN(size, PAGE_SIZE);
+	size = PAGE_ALIGN(size);
 	if (size == 0)
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/bridge/ptn3460.c b/drivers/gpu/drm/bridge/ptn3460.c
index 98fd17ae4916..d466696ed5e8 100644
--- a/drivers/gpu/drm/bridge/ptn3460.c
+++ b/drivers/gpu/drm/bridge/ptn3460.c
@@ -328,7 +328,7 @@ int ptn3460_init(struct drm_device *dev, struct drm_encoder *encoder,
 	}
 	drm_connector_helper_add(&ptn_bridge->connector,
 			&ptn3460_connector_helper_funcs);
-	drm_sysfs_connector_add(&ptn_bridge->connector);
+	drm_connector_register(&ptn_bridge->connector);
 	drm_mode_connector_attach_encoder(&ptn_bridge->connector, encoder);
 
 	return 0;
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.c b/drivers/gpu/drm/cirrus/cirrus_drv.c
index 08ce520f61a5..c2a1cba1e984 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.c
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.c
@@ -29,9 +29,11 @@ module_param_named(modeset, cirrus_modeset, int, 0400);
 static struct drm_driver driver;
 
 /* only bind to the cirrus chip in qemu */
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+static const struct pci_device_id pciidlist[] = {
 	{ PCI_VENDOR_ID_CIRRUS, PCI_DEVICE_ID_CIRRUS_5446, 0x1af4, 0x1100, 0,
 	  0, 0 },
+	{ PCI_VENDOR_ID_CIRRUS, PCI_DEVICE_ID_CIRRUS_5446, PCI_VENDOR_ID_XEN,
+	  0x0001, 0, 0, 0 },
 	{0,}
 };
 
@@ -76,6 +78,7 @@ static void cirrus_pci_remove(struct pci_dev *pdev)
 	drm_put_dev(dev);
 }
 
+#ifdef CONFIG_PM_SLEEP
 static int cirrus_pm_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
@@ -110,6 +113,7 @@ static int cirrus_pm_resume(struct device *dev)
 	drm_kms_helper_poll_enable(drm_dev);
 	return 0;
 }
+#endif
 
 static const struct file_operations cirrus_driver_fops = {
 	.owner = THIS_MODULE,
@@ -126,6 +130,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM,
 	.load = cirrus_driver_load,
 	.unload = cirrus_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.fops = &cirrus_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 117d3eca5e37..d44e69daa239 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -21,6 +21,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #define DRIVER_AUTHOR		"Matthew Garrett"
 
 #define DRIVER_NAME		"cirrus"
@@ -163,7 +165,7 @@ struct cirrus_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_cirrus_bo(gobj) container_of((gobj), struct cirrus_bo, gem)
@@ -241,7 +243,7 @@ static inline int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
 {
 	int ret;
 
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, NULL);
 	if (ret) {
 		if (ret != -ERESTARTSYS && ret != -EBUSY)
 			DRM_ERROR("reserve failed %p\n", bo);
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 32bbba0a787b..d231b1c317af 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -160,7 +160,8 @@ static int cirrusfb_create_object(struct cirrus_fbdev *afbdev,
 static int cirrusfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct cirrus_fbdev *gfbdev = (struct cirrus_fbdev *)helper;
+	struct cirrus_fbdev *gfbdev =
+		container_of(helper, struct cirrus_fbdev, helper);
 	struct drm_device *dev = gfbdev->helper.dev;
 	struct cirrus_device *cdev = gfbdev->helper.dev->dev_private;
 	struct fb_info *info;
@@ -288,7 +289,7 @@ static int cirrus_fbdev_destroy(struct drm_device *dev,
 	return 0;
 }
 
-static struct drm_fb_helper_funcs cirrus_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs cirrus_fb_helper_funcs = {
 	.gamma_set = cirrus_crtc_fb_gamma_set,
 	.gamma_get = cirrus_crtc_fb_gamma_get,
 	.fb_probe = cirrusfb_create,
@@ -306,9 +307,11 @@ int cirrus_fbdev_init(struct cirrus_device *cdev)
 		return -ENOMEM;
 
 	cdev->mode_info.gfbdev = gfbdev;
-	gfbdev->helper.funcs = &cirrus_fb_helper_funcs;
 	spin_lock_init(&gfbdev->dirty_lock);
 
+	drm_fb_helper_prepare(cdev->dev, &gfbdev->helper,
+			      &cirrus_fb_helper_funcs);
+
 	ret = drm_fb_helper_init(cdev->dev, &gfbdev->helper,
 				 cdev->num_crtc, CIRRUSFB_CONN_LIMIT);
 	if (ret) {
diff --git a/drivers/gpu/drm/cirrus/cirrus_mode.c b/drivers/gpu/drm/cirrus/cirrus_mode.c
index 49332c5fe35b..c7c5a9d91fa0 100644
--- a/drivers/gpu/drm/cirrus/cirrus_mode.c
+++ b/drivers/gpu/drm/cirrus/cirrus_mode.c
@@ -509,19 +509,9 @@ static struct drm_encoder *cirrus_connector_best_encoder(struct drm_connector
 						  *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj =
-		    drm_mode_object_find(connector->dev, enc_id,
-					 DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -565,6 +555,7 @@ static struct drm_connector *cirrus_vga_init(struct drm_device *dev)
 
 	drm_connector_helper_add(connector, &cirrus_vga_connector_helper_funcs);
 
+	drm_connector_register(connector);
 	return connector;
 }
 
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 92e6b7786097..dfffd528517a 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -298,18 +298,21 @@ void cirrus_mm_fini(struct cirrus_device *cirrus)
 void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
@@ -340,7 +343,7 @@ int cirrus_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&cirrus->ttm.bdev, &cirrusbo->bo, size,
 			  ttm_bo_type_device, &cirrusbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, cirrus_bo_ttm_destroy);
+			  NULL, NULL, cirrus_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -365,7 +368,7 @@ int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	cirrus_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -392,7 +395,7 @@ int cirrus_bo_push_sysram(struct cirrus_bo *bo)
 
 	cirrus_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -408,7 +411,7 @@ int cirrus_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct cirrus_device *cirrus;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	cirrus = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/drm_agpsupport.c b/drivers/gpu/drm/drm_agpsupport.c
index dde205cef384..4b2b4aa5033b 100644
--- a/drivers/gpu/drm/drm_agpsupport.c
+++ b/drivers/gpu/drm/drm_agpsupport.c
@@ -34,6 +34,7 @@
 #include <drm/drmP.h>
 #include <linux/module.h>
 #include <linux/slab.h>
+#include "drm_legacy.h"
 
 #if __OS_HAS_AGP
 
diff --git a/drivers/gpu/drm/drm_auth.c b/drivers/gpu/drm/drm_auth.c
index 3cedae12b3c1..fc8e8aaa34fb 100644
--- a/drivers/gpu/drm/drm_auth.c
+++ b/drivers/gpu/drm/drm_auth.c
@@ -34,6 +34,13 @@
  */
 
 #include <drm/drmP.h>
+#include "drm_internal.h"
+
+struct drm_magic_entry {
+	struct list_head head;
+	struct drm_hash_item hash_item;
+	struct drm_file *priv;
+};
 
 /**
  * Find the file with the given magic number.
diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c
index 68175b54504b..569064a00693 100644
--- a/drivers/gpu/drm/drm_bufs.c
+++ b/drivers/gpu/drm/drm_bufs.c
@@ -1,18 +1,13 @@
-/**
- * \file drm_bufs.c
- * Generic buffer template
- *
- * \author Rickard E. (Rik) Faith <faith@valinux.com>
- * \author Gareth Hughes <gareth@valinux.com>
- */
-
 /*
- * Created: Thu Nov 23 03:10:50 2000 by gareth@valinux.com
+ * Legacy: Generic DRM Buffer Management
  *
  * Copyright 1999, 2000 Precision Insight, Inc., Cedar Park, Texas.
  * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
  * All Rights Reserved.
  *
+ * Author: Rickard E. (Rik) Faith <faith@valinux.com>
+ * Author: Gareth Hughes <gareth@valinux.com>
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
@@ -39,6 +34,7 @@
 #include <linux/export.h>
 #include <asm/shmparam.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 static struct drm_map_list *drm_find_matching_map(struct drm_device *dev,
 						  struct drm_local_map *map)
@@ -365,9 +361,9 @@ static int drm_addmap_core(struct drm_device * dev, resource_size_t offset,
 	return 0;
 }
 
-int drm_addmap(struct drm_device * dev, resource_size_t offset,
-	       unsigned int size, enum drm_map_type type,
-	       enum drm_map_flags flags, struct drm_local_map ** map_ptr)
+int drm_legacy_addmap(struct drm_device * dev, resource_size_t offset,
+		      unsigned int size, enum drm_map_type type,
+		      enum drm_map_flags flags, struct drm_local_map **map_ptr)
 {
 	struct drm_map_list *list;
 	int rc;
@@ -377,8 +373,7 @@ int drm_addmap(struct drm_device * dev, resource_size_t offset,
 		*map_ptr = list->map;
 	return rc;
 }
-
-EXPORT_SYMBOL(drm_addmap);
+EXPORT_SYMBOL(drm_legacy_addmap);
 
 /**
  * Ioctl to specify a range of memory that is available for mapping by a
@@ -391,8 +386,8 @@ EXPORT_SYMBOL(drm_addmap);
  * \return zero on success or a negative value on error.
  *
  */
-int drm_addmap_ioctl(struct drm_device *dev, void *data,
-		     struct drm_file *file_priv)
+int drm_legacy_addmap_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *file_priv)
 {
 	struct drm_map *map = data;
 	struct drm_map_list *maplist;
@@ -429,9 +424,9 @@ int drm_addmap_ioctl(struct drm_device *dev, void *data,
  * its being used, and free any associate resource (such as MTRR's) if it's not
  * being on use.
  *
- * \sa drm_addmap
+ * \sa drm_legacy_addmap
  */
-int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
+int drm_legacy_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 {
 	struct drm_map_list *r_list = NULL, *list_t;
 	drm_dma_handle_t dmah;
@@ -478,26 +473,26 @@ int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 		dmah.vaddr = map->handle;
 		dmah.busaddr = map->offset;
 		dmah.size = map->size;
-		__drm_pci_free(dev, &dmah);
+		__drm_legacy_pci_free(dev, &dmah);
 		break;
 	}
 	kfree(map);
 
 	return 0;
 }
-EXPORT_SYMBOL(drm_rmmap_locked);
+EXPORT_SYMBOL(drm_legacy_rmmap_locked);
 
-int drm_rmmap(struct drm_device *dev, struct drm_local_map *map)
+int drm_legacy_rmmap(struct drm_device *dev, struct drm_local_map *map)
 {
 	int ret;
 
 	mutex_lock(&dev->struct_mutex);
-	ret = drm_rmmap_locked(dev, map);
+	ret = drm_legacy_rmmap_locked(dev, map);
 	mutex_unlock(&dev->struct_mutex);
 
 	return ret;
 }
-EXPORT_SYMBOL(drm_rmmap);
+EXPORT_SYMBOL(drm_legacy_rmmap);
 
 /* The rmmap ioctl appears to be unnecessary.  All mappings are torn down on
  * the last close of the device, and this is necessary for cleanup when things
@@ -514,8 +509,8 @@ EXPORT_SYMBOL(drm_rmmap);
  * \param arg pointer to a struct drm_map structure.
  * \return zero on success or a negative value on error.
  */
-int drm_rmmap_ioctl(struct drm_device *dev, void *data,
-		    struct drm_file *file_priv)
+int drm_legacy_rmmap_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file_priv)
 {
 	struct drm_map *request = data;
 	struct drm_local_map *map = NULL;
@@ -546,7 +541,7 @@ int drm_rmmap_ioctl(struct drm_device *dev, void *data,
 		return 0;
 	}
 
-	ret = drm_rmmap_locked(dev, map);
+	ret = drm_legacy_rmmap_locked(dev, map);
 
 	mutex_unlock(&dev->struct_mutex);
 
@@ -599,7 +594,8 @@ static void drm_cleanup_buf_error(struct drm_device * dev,
  * reallocates the buffer list of the same size order to accommodate the new
  * buffers.
  */
-int drm_addbufs_agp(struct drm_device * dev, struct drm_buf_desc * request)
+int drm_legacy_addbufs_agp(struct drm_device *dev,
+			   struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_entry *entry;
@@ -759,10 +755,11 @@ int drm_addbufs_agp(struct drm_device * dev, struct drm_buf_desc * request)
 	atomic_dec(&dev->buf_alloc);
 	return 0;
 }
-EXPORT_SYMBOL(drm_addbufs_agp);
+EXPORT_SYMBOL(drm_legacy_addbufs_agp);
 #endif				/* __OS_HAS_AGP */
 
-int drm_addbufs_pci(struct drm_device * dev, struct drm_buf_desc * request)
+int drm_legacy_addbufs_pci(struct drm_device *dev,
+			   struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int count;
@@ -964,9 +961,10 @@ int drm_addbufs_pci(struct drm_device * dev, struct drm_buf_desc * request)
 	return 0;
 
 }
-EXPORT_SYMBOL(drm_addbufs_pci);
+EXPORT_SYMBOL(drm_legacy_addbufs_pci);
 
-static int drm_addbufs_sg(struct drm_device * dev, struct drm_buf_desc * request)
+static int drm_legacy_addbufs_sg(struct drm_device *dev,
+				 struct drm_buf_desc *request)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_entry *entry;
@@ -1135,8 +1133,8 @@ static int drm_addbufs_sg(struct drm_device * dev, struct drm_buf_desc * request
  * addbufs_sg() or addbufs_pci() for AGP, scatter-gather or consistent
  * PCI memory respectively.
  */
-int drm_addbufs(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
+int drm_legacy_addbufs(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_buf_desc *request = data;
 	int ret;
@@ -1149,15 +1147,15 @@ int drm_addbufs(struct drm_device *dev, void *data,
 
 #if __OS_HAS_AGP
 	if (request->flags & _DRM_AGP_BUFFER)
-		ret = drm_addbufs_agp(dev, request);
+		ret = drm_legacy_addbufs_agp(dev, request);
 	else
 #endif
 	if (request->flags & _DRM_SG_BUFFER)
-		ret = drm_addbufs_sg(dev, request);
+		ret = drm_legacy_addbufs_sg(dev, request);
 	else if (request->flags & _DRM_FB_BUFFER)
 		ret = -EINVAL;
 	else
-		ret = drm_addbufs_pci(dev, request);
+		ret = drm_legacy_addbufs_pci(dev, request);
 
 	return ret;
 }
@@ -1179,8 +1177,8 @@ int drm_addbufs(struct drm_device *dev, void *data,
  * lock, preventing of allocating more buffers after this call. Information
  * about each requested buffer is then copied into user space.
  */
-int drm_infobufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_infobufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_info *request = data;
@@ -1217,7 +1215,6 @@ int drm_infobufs(struct drm_device *dev, void *data,
 				struct drm_buf_desc __user *to =
 				    &request->list[count];
 				struct drm_buf_entry *from = &dma->bufs[i];
-				struct drm_freelist *list = &dma->bufs[i].freelist;
 				if (copy_to_user(&to->count,
 						 &from->buf_count,
 						 sizeof(from->buf_count)) ||
@@ -1225,19 +1222,19 @@ int drm_infobufs(struct drm_device *dev, void *data,
 						 &from->buf_size,
 						 sizeof(from->buf_size)) ||
 				    copy_to_user(&to->low_mark,
-						 &list->low_mark,
-						 sizeof(list->low_mark)) ||
+						 &from->low_mark,
+						 sizeof(from->low_mark)) ||
 				    copy_to_user(&to->high_mark,
-						 &list->high_mark,
-						 sizeof(list->high_mark)))
+						 &from->high_mark,
+						 sizeof(from->high_mark)))
 					return -EFAULT;
 
 				DRM_DEBUG("%d %d %d %d %d\n",
 					  i,
 					  dma->bufs[i].buf_count,
 					  dma->bufs[i].buf_size,
-					  dma->bufs[i].freelist.low_mark,
-					  dma->bufs[i].freelist.high_mark);
+					  dma->bufs[i].low_mark,
+					  dma->bufs[i].high_mark);
 				++count;
 			}
 		}
@@ -1261,8 +1258,8 @@ int drm_infobufs(struct drm_device *dev, void *data,
  *
  * \note This ioctl is deprecated and mostly never used.
  */
-int drm_markbufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_markbufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_desc *request = data;
@@ -1290,8 +1287,8 @@ int drm_markbufs(struct drm_device *dev, void *data,
 	if (request->high_mark < 0 || request->high_mark > entry->buf_count)
 		return -EINVAL;
 
-	entry->freelist.low_mark = request->low_mark;
-	entry->freelist.high_mark = request->high_mark;
+	entry->low_mark = request->low_mark;
+	entry->high_mark = request->high_mark;
 
 	return 0;
 }
@@ -1308,8 +1305,8 @@ int drm_markbufs(struct drm_device *dev, void *data,
  * Calls free_buffer() for each used buffer.
  * This function is primarily used for debugging.
  */
-int drm_freebufs(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_freebufs(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	struct drm_buf_free *request = data;
@@ -1341,7 +1338,7 @@ int drm_freebufs(struct drm_device *dev, void *data,
 				  task_pid_nr(current));
 			return -EINVAL;
 		}
-		drm_free_buffer(dev, buf);
+		drm_legacy_free_buffer(dev, buf);
 	}
 
 	return 0;
@@ -1361,8 +1358,8 @@ int drm_freebufs(struct drm_device *dev, void *data,
  * offset equal to 0, which drm_mmap() interpretes as PCI buffers and calls
  * drm_mmap_dma().
  */
-int drm_mapbufs(struct drm_device *dev, void *data,
-	        struct drm_file *file_priv)
+int drm_legacy_mapbufs(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int retcode = 0;
@@ -1449,7 +1446,7 @@ int drm_mapbufs(struct drm_device *dev, void *data,
 	return retcode;
 }
 
-int drm_dma_ioctl(struct drm_device *dev, void *data,
+int drm_legacy_dma_ioctl(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	if (drm_core_check_feature(dev, DRIVER_MODESET))
@@ -1461,7 +1458,7 @@ int drm_dma_ioctl(struct drm_device *dev, void *data,
 		return -EINVAL;
 }
 
-struct drm_local_map *drm_getsarea(struct drm_device *dev)
+struct drm_local_map *drm_legacy_getsarea(struct drm_device *dev)
 {
 	struct drm_map_list *entry;
 
@@ -1473,4 +1470,4 @@ struct drm_local_map *drm_getsarea(struct drm_device *dev)
 	}
 	return NULL;
 }
-EXPORT_SYMBOL(drm_getsarea);
+EXPORT_SYMBOL(drm_legacy_getsarea);
diff --git a/drivers/gpu/drm/drm_context.c b/drivers/gpu/drm/drm_context.c
index a4b017b6849e..9b23525c0ed0 100644
--- a/drivers/gpu/drm/drm_context.c
+++ b/drivers/gpu/drm/drm_context.c
@@ -1,18 +1,13 @@
-/**
- * \file drm_context.c
- * IOCTLs for generic contexts
- *
- * \author Rickard E. (Rik) Faith <faith@valinux.com>
- * \author Gareth Hughes <gareth@valinux.com>
- */
-
 /*
- * Created: Fri Nov 24 18:31:37 2000 by gareth@valinux.com
+ * Legacy: Generic DRM Contexts
  *
  * Copyright 1999, 2000 Precision Insight, Inc., Cedar Park, Texas.
  * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
  * All Rights Reserved.
  *
+ * Author: Rickard E. (Rik) Faith <faith@valinux.com>
+ * Author: Gareth Hughes <gareth@valinux.com>
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
@@ -33,14 +28,14 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-/*
- * ChangeLog:
- *  2001-11-16	Torsten Duwe <duwe@caldera.de>
- *		added context constructor/destructor hooks,
- *		needed by SiS driver's memory management.
- */
-
 #include <drm/drmP.h>
+#include "drm_legacy.h"
+
+struct drm_ctx_list {
+	struct list_head head;
+	drm_context_t handle;
+	struct drm_file *tag;
+};
 
 /******************************************************************/
 /** \name Context bitmap support */
@@ -56,7 +51,7 @@
  * in drm_device::ctx_idr, while holding the drm_device::struct_mutex
  * lock.
  */
-void drm_ctxbitmap_free(struct drm_device * dev, int ctx_handle)
+void drm_legacy_ctxbitmap_free(struct drm_device * dev, int ctx_handle)
 {
 	mutex_lock(&dev->struct_mutex);
 	idr_remove(&dev->ctx_idr, ctx_handle);
@@ -72,7 +67,7 @@ void drm_ctxbitmap_free(struct drm_device * dev, int ctx_handle)
  * Allocate a new idr from drm_device::ctx_idr while holding the
  * drm_device::struct_mutex lock.
  */
-static int drm_ctxbitmap_next(struct drm_device * dev)
+static int drm_legacy_ctxbitmap_next(struct drm_device * dev)
 {
 	int ret;
 
@@ -90,7 +85,7 @@ static int drm_ctxbitmap_next(struct drm_device * dev)
  *
  * Initialise the drm_device::ctx_idr
  */
-int drm_ctxbitmap_init(struct drm_device * dev)
+int drm_legacy_ctxbitmap_init(struct drm_device * dev)
 {
 	idr_init(&dev->ctx_idr);
 	return 0;
@@ -104,13 +99,43 @@ int drm_ctxbitmap_init(struct drm_device * dev)
  * Free all idr members using drm_ctx_sarea_free helper function
  * while holding the drm_device::struct_mutex lock.
  */
-void drm_ctxbitmap_cleanup(struct drm_device * dev)
+void drm_legacy_ctxbitmap_cleanup(struct drm_device * dev)
 {
 	mutex_lock(&dev->struct_mutex);
 	idr_destroy(&dev->ctx_idr);
 	mutex_unlock(&dev->struct_mutex);
 }
 
+/**
+ * drm_ctxbitmap_flush() - Flush all contexts owned by a file
+ * @dev: DRM device to operate on
+ * @file: Open file to flush contexts for
+ *
+ * This iterates over all contexts on @dev and drops them if they're owned by
+ * @file. Note that after this call returns, new contexts might be added if
+ * the file is still alive.
+ */
+void drm_legacy_ctxbitmap_flush(struct drm_device *dev, struct drm_file *file)
+{
+	struct drm_ctx_list *pos, *tmp;
+
+	mutex_lock(&dev->ctxlist_mutex);
+
+	list_for_each_entry_safe(pos, tmp, &dev->ctxlist, head) {
+		if (pos->tag == file &&
+		    pos->handle != DRM_KERNEL_CONTEXT) {
+			if (dev->driver->context_dtor)
+				dev->driver->context_dtor(dev, pos->handle);
+
+			drm_legacy_ctxbitmap_free(dev, pos->handle);
+			list_del(&pos->head);
+			kfree(pos);
+		}
+	}
+
+	mutex_unlock(&dev->ctxlist_mutex);
+}
+
 /*@}*/
 
 /******************************************************************/
@@ -129,8 +154,8 @@ void drm_ctxbitmap_cleanup(struct drm_device * dev)
  * Gets the map from drm_device::ctx_idr with the handle specified and
  * returns its handle.
  */
-int drm_getsareactx(struct drm_device *dev, void *data,
-		    struct drm_file *file_priv)
+int drm_legacy_getsareactx(struct drm_device *dev, void *data,
+			   struct drm_file *file_priv)
 {
 	struct drm_ctx_priv_map *request = data;
 	struct drm_local_map *map;
@@ -173,8 +198,8 @@ int drm_getsareactx(struct drm_device *dev, void *data,
  * Searches the mapping specified in \p arg and update the entry in
  * drm_device::ctx_idr with it.
  */
-int drm_setsareactx(struct drm_device *dev, void *data,
-		    struct drm_file *file_priv)
+int drm_legacy_setsareactx(struct drm_device *dev, void *data,
+			   struct drm_file *file_priv)
 {
 	struct drm_ctx_priv_map *request = data;
 	struct drm_local_map *map = NULL;
@@ -273,8 +298,8 @@ static int drm_context_switch_complete(struct drm_device *dev,
  * \param arg user argument pointing to a drm_ctx_res structure.
  * \return zero on success or a negative number on failure.
  */
-int drm_resctx(struct drm_device *dev, void *data,
-	       struct drm_file *file_priv)
+int drm_legacy_resctx(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
 {
 	struct drm_ctx_res *res = data;
 	struct drm_ctx ctx;
@@ -304,16 +329,16 @@ int drm_resctx(struct drm_device *dev, void *data,
  *
  * Get a new handle for the context and copy to userspace.
  */
-int drm_addctx(struct drm_device *dev, void *data,
-	       struct drm_file *file_priv)
+int drm_legacy_addctx(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
 {
 	struct drm_ctx_list *ctx_entry;
 	struct drm_ctx *ctx = data;
 
-	ctx->handle = drm_ctxbitmap_next(dev);
+	ctx->handle = drm_legacy_ctxbitmap_next(dev);
 	if (ctx->handle == DRM_KERNEL_CONTEXT) {
 		/* Skip kernel's context and get a new one. */
-		ctx->handle = drm_ctxbitmap_next(dev);
+		ctx->handle = drm_legacy_ctxbitmap_next(dev);
 	}
 	DRM_DEBUG("%d\n", ctx->handle);
 	if (ctx->handle == -1) {
@@ -348,7 +373,8 @@ int drm_addctx(struct drm_device *dev, void *data,
  * \param arg user argument pointing to a drm_ctx structure.
  * \return zero on success or a negative number on failure.
  */
-int drm_getctx(struct drm_device *dev, void *data, struct drm_file *file_priv)
+int drm_legacy_getctx(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
 {
 	struct drm_ctx *ctx = data;
 
@@ -369,8 +395,8 @@ int drm_getctx(struct drm_device *dev, void *data, struct drm_file *file_priv)
  *
  * Calls context_switch().
  */
-int drm_switchctx(struct drm_device *dev, void *data,
-		  struct drm_file *file_priv)
+int drm_legacy_switchctx(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv)
 {
 	struct drm_ctx *ctx = data;
 
@@ -389,8 +415,8 @@ int drm_switchctx(struct drm_device *dev, void *data,
  *
  * Calls context_switch_complete().
  */
-int drm_newctx(struct drm_device *dev, void *data,
-	       struct drm_file *file_priv)
+int drm_legacy_newctx(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
 {
 	struct drm_ctx *ctx = data;
 
@@ -411,8 +437,8 @@ int drm_newctx(struct drm_device *dev, void *data,
  *
  * If not the special kernel context, calls ctxbitmap_free() to free the specified context.
  */
-int drm_rmctx(struct drm_device *dev, void *data,
-	      struct drm_file *file_priv)
+int drm_legacy_rmctx(struct drm_device *dev, void *data,
+		     struct drm_file *file_priv)
 {
 	struct drm_ctx *ctx = data;
 
@@ -420,7 +446,7 @@ int drm_rmctx(struct drm_device *dev, void *data,
 	if (ctx->handle != DRM_KERNEL_CONTEXT) {
 		if (dev->driver->context_dtor)
 			dev->driver->context_dtor(dev, ctx->handle);
-		drm_ctxbitmap_free(dev, ctx->handle);
+		drm_legacy_ctxbitmap_free(dev, ctx->handle);
 	}
 
 	mutex_lock(&dev->ctxlist_mutex);
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index fe94cc10cd35..e79c8d3700d8 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -40,101 +40,11 @@
 #include <drm/drm_modeset_lock.h>
 
 #include "drm_crtc_internal.h"
+#include "drm_internal.h"
 
-/**
- * drm_modeset_lock_all - take all modeset locks
- * @dev: drm device
- *
- * This function takes all modeset locks, suitable where a more fine-grained
- * scheme isn't (yet) implemented. Locks must be dropped with
- * drm_modeset_unlock_all.
- */
-void drm_modeset_lock_all(struct drm_device *dev)
-{
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_modeset_acquire_ctx *ctx;
-	int ret;
-
-	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
-	if (WARN_ON(!ctx))
-		return;
-
-	mutex_lock(&config->mutex);
-
-	drm_modeset_acquire_init(ctx, 0);
-
-retry:
-	ret = drm_modeset_lock(&config->connection_mutex, ctx);
-	if (ret)
-		goto fail;
-	ret = drm_modeset_lock_all_crtcs(dev, ctx);
-	if (ret)
-		goto fail;
-
-	WARN_ON(config->acquire_ctx);
-
-	/* now we hold the locks, so now that it is safe, stash the
-	 * ctx for drm_modeset_unlock_all():
-	 */
-	config->acquire_ctx = ctx;
-
-	drm_warn_on_modeset_not_all_locked(dev);
-
-	return;
-
-fail:
-	if (ret == -EDEADLK) {
-		drm_modeset_backoff(ctx);
-		goto retry;
-	}
-}
-EXPORT_SYMBOL(drm_modeset_lock_all);
-
-/**
- * drm_modeset_unlock_all - drop all modeset locks
- * @dev: device
- *
- * This function drop all modeset locks taken by drm_modeset_lock_all.
- */
-void drm_modeset_unlock_all(struct drm_device *dev)
-{
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_modeset_acquire_ctx *ctx = config->acquire_ctx;
-
-	if (WARN_ON(!ctx))
-		return;
-
-	config->acquire_ctx = NULL;
-	drm_modeset_drop_locks(ctx);
-	drm_modeset_acquire_fini(ctx);
-
-	kfree(ctx);
-
-	mutex_unlock(&dev->mode_config.mutex);
-}
-EXPORT_SYMBOL(drm_modeset_unlock_all);
-
-/**
- * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked
- * @dev: device
- *
- * Useful as a debug assert.
- */
-void drm_warn_on_modeset_not_all_locked(struct drm_device *dev)
-{
-	struct drm_crtc *crtc;
-
-	/* Locking is currently fubar in the panic handler. */
-	if (oops_in_progress)
-		return;
-
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
-		WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
-
-	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
-	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
-}
-EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
+static struct drm_framebuffer *add_framebuffer_internal(struct drm_device *dev,
+							struct drm_mode_fb_cmd2 *r,
+							struct drm_file *file_priv);
 
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
@@ -178,6 +88,12 @@ static const struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
 	{ DRM_MODE_SCALE_ASPECT, "Full aspect" },
 };
 
+static const struct drm_prop_enum_list drm_aspect_ratio_enum_list[] = {
+	{ DRM_MODE_PICTURE_ASPECT_NONE, "Automatic" },
+	{ DRM_MODE_PICTURE_ASPECT_4_3, "4:3" },
+	{ DRM_MODE_PICTURE_ASPECT_16_9, "16:9" },
+};
+
 /*
  * Non-global properties, but "required" for certain connectors.
  */
@@ -357,6 +273,32 @@ const char *drm_get_format_name(uint32_t format)
 }
 EXPORT_SYMBOL(drm_get_format_name);
 
+/*
+ * Internal function to assign a slot in the object idr and optionally
+ * register the object into the idr.
+ */
+static int drm_mode_object_get_reg(struct drm_device *dev,
+				   struct drm_mode_object *obj,
+				   uint32_t obj_type,
+				   bool register_obj)
+{
+	int ret;
+
+	mutex_lock(&dev->mode_config.idr_mutex);
+	ret = idr_alloc(&dev->mode_config.crtc_idr, register_obj ? obj : NULL, 1, 0, GFP_KERNEL);
+	if (ret >= 0) {
+		/*
+		 * Set up the object linking under the protection of the idr
+		 * lock so that other users can't see inconsistent state.
+		 */
+		obj->id = ret;
+		obj->type = obj_type;
+	}
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	return ret < 0 ? ret : 0;
+}
+
 /**
  * drm_mode_object_get - allocate a new modeset identifier
  * @dev: DRM device
@@ -375,21 +317,15 @@ EXPORT_SYMBOL(drm_get_format_name);
 int drm_mode_object_get(struct drm_device *dev,
 			struct drm_mode_object *obj, uint32_t obj_type)
 {
-	int ret;
+	return drm_mode_object_get_reg(dev, obj, obj_type, true);
+}
 
+static void drm_mode_object_register(struct drm_device *dev,
+				     struct drm_mode_object *obj)
+{
 	mutex_lock(&dev->mode_config.idr_mutex);
-	ret = idr_alloc(&dev->mode_config.crtc_idr, obj, 1, 0, GFP_KERNEL);
-	if (ret >= 0) {
-		/*
-		 * Set up the object linking under the protection of the idr
-		 * lock so that other users can't see inconsistent state.
-		 */
-		obj->id = ret;
-		obj->type = obj_type;
-	}
+	idr_replace(&dev->mode_config.crtc_idr, obj, obj->id);
 	mutex_unlock(&dev->mode_config.idr_mutex);
-
-	return ret < 0 ? ret : 0;
 }
 
 /**
@@ -416,8 +352,12 @@ static struct drm_mode_object *_object_find(struct drm_device *dev,
 
 	mutex_lock(&dev->mode_config.idr_mutex);
 	obj = idr_find(&dev->mode_config.crtc_idr, id);
-	if (!obj || (type != DRM_MODE_OBJECT_ANY && obj->type != type) ||
-	    (obj->id != id))
+	if (obj && type != DRM_MODE_OBJECT_ANY && obj->type != type)
+		obj = NULL;
+	if (obj && obj->id != id)
+		obj = NULL;
+	/* don't leak out unref'd fb's */
+	if (obj && (obj->type == DRM_MODE_OBJECT_FB))
 		obj = NULL;
 	mutex_unlock(&dev->mode_config.idr_mutex);
 
@@ -444,9 +384,6 @@ struct drm_mode_object *drm_mode_object_find(struct drm_device *dev,
 	 * function.*/
 	WARN_ON(type == DRM_MODE_OBJECT_FB);
 	obj = _object_find(dev, id, type);
-	/* don't leak out unref'd fb's */
-	if (obj && (obj->type == DRM_MODE_OBJECT_FB))
-		obj = NULL;
 	return obj;
 }
 EXPORT_SYMBOL(drm_mode_object_find);
@@ -484,9 +421,6 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb,
 	if (ret)
 		goto out;
 
-	/* Grab the idr reference. */
-	drm_framebuffer_reference(fb);
-
 	dev->mode_config.num_fb++;
 	list_add(&fb->head, &dev->mode_config.fb_list);
 out:
@@ -496,10 +430,34 @@ out:
 }
 EXPORT_SYMBOL(drm_framebuffer_init);
 
+/* dev->mode_config.fb_lock must be held! */
+static void __drm_framebuffer_unregister(struct drm_device *dev,
+					 struct drm_framebuffer *fb)
+{
+	mutex_lock(&dev->mode_config.idr_mutex);
+	idr_remove(&dev->mode_config.crtc_idr, fb->base.id);
+	mutex_unlock(&dev->mode_config.idr_mutex);
+
+	fb->base.id = 0;
+}
+
 static void drm_framebuffer_free(struct kref *kref)
 {
 	struct drm_framebuffer *fb =
 			container_of(kref, struct drm_framebuffer, refcount);
+	struct drm_device *dev = fb->dev;
+
+	/*
+	 * The lookup idr holds a weak reference, which has not necessarily been
+	 * removed at this point. Check for that.
+	 */
+	mutex_lock(&dev->mode_config.fb_lock);
+	if (fb->base.id) {
+		/* Mark fb as reaped and drop idr ref. */
+		__drm_framebuffer_unregister(dev, fb);
+	}
+	mutex_unlock(&dev->mode_config.fb_lock);
+
 	fb->funcs->destroy(fb);
 }
 
@@ -536,8 +494,10 @@ struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev,
 
 	mutex_lock(&dev->mode_config.fb_lock);
 	fb = __drm_framebuffer_lookup(dev, id);
-	if (fb)
-		drm_framebuffer_reference(fb);
+	if (fb) {
+		if (!kref_get_unless_zero(&fb->refcount))
+			fb = NULL;
+	}
 	mutex_unlock(&dev->mode_config.fb_lock);
 
 	return fb;
@@ -581,19 +541,6 @@ static void __drm_framebuffer_unreference(struct drm_framebuffer *fb)
 	kref_put(&fb->refcount, drm_framebuffer_free_bug);
 }
 
-/* dev->mode_config.fb_lock must be held! */
-static void __drm_framebuffer_unregister(struct drm_device *dev,
-					 struct drm_framebuffer *fb)
-{
-	mutex_lock(&dev->mode_config.idr_mutex);
-	idr_remove(&dev->mode_config.crtc_idr, fb->base.id);
-	mutex_unlock(&dev->mode_config.idr_mutex);
-
-	fb->base.id = 0;
-
-	__drm_framebuffer_unreference(fb);
-}
-
 /**
  * drm_framebuffer_unregister_private - unregister a private fb from the lookup idr
  * @fb: fb to unregister
@@ -723,7 +670,7 @@ DEFINE_WW_CLASS(crtc_ww_class);
  */
 int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 			      struct drm_plane *primary,
-			      void *cursor,
+			      struct drm_plane *cursor,
 			      const struct drm_crtc_funcs *funcs)
 {
 	struct drm_mode_config *config = &dev->mode_config;
@@ -733,11 +680,7 @@ int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 	crtc->funcs = funcs;
 	crtc->invert_dimensions = false;
 
-	drm_modeset_lock_all(dev);
 	drm_modeset_lock_init(&crtc->mutex);
-	/* dropped by _unlock_all(): */
-	drm_modeset_lock(&crtc->mutex, config->acquire_ctx);
-
 	ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC);
 	if (ret)
 		goto out;
@@ -748,11 +691,13 @@ int drm_crtc_init_with_planes(struct drm_device *dev, struct drm_crtc *crtc,
 	config->num_crtc++;
 
 	crtc->primary = primary;
+	crtc->cursor = cursor;
 	if (primary)
 		primary->possible_crtcs = 1 << drm_crtc_index(crtc);
+	if (cursor)
+		cursor->possible_crtcs = 1 << drm_crtc_index(crtc);
 
  out:
-	drm_modeset_unlock_all(dev);
 
 	return ret;
 }
@@ -819,6 +764,59 @@ static void drm_mode_remove(struct drm_connector *connector,
 }
 
 /**
+ * drm_connector_get_cmdline_mode - reads the user's cmdline mode
+ * @connector: connector to quwery
+ * @mode: returned mode
+ *
+ * The kernel supports per-connector configration of its consoles through
+ * use of the video= parameter. This function parses that option and
+ * extracts the user's specified mode (or enable/disable status) for a
+ * particular connector. This is typically only used during the early fbdev
+ * setup.
+ */
+static void drm_connector_get_cmdline_mode(struct drm_connector *connector)
+{
+	struct drm_cmdline_mode *mode = &connector->cmdline_mode;
+	char *option = NULL;
+
+	if (fb_get_options(connector->name, &option))
+		return;
+
+	if (!drm_mode_parse_command_line_for_connector(option,
+						       connector,
+						       mode))
+		return;
+
+	if (mode->force) {
+		const char *s;
+
+		switch (mode->force) {
+		case DRM_FORCE_OFF:
+			s = "OFF";
+			break;
+		case DRM_FORCE_ON_DIGITAL:
+			s = "ON - dig";
+			break;
+		default:
+		case DRM_FORCE_ON:
+			s = "ON";
+			break;
+		}
+
+		DRM_INFO("forcing %s connector %s\n", connector->name, s);
+		connector->force = mode->force;
+	}
+
+	DRM_DEBUG_KMS("cmdline mode for connector %s %dx%d@%dHz%s%s%s\n",
+		      connector->name,
+		      mode->xres, mode->yres,
+		      mode->refresh_specified ? mode->refresh : 60,
+		      mode->rb ? " reduced blanking" : "",
+		      mode->margins ? " with margins" : "",
+		      mode->interlace ?  " interlaced" : "");
+}
+
+/**
  * drm_connector_init - Init a preallocated connector
  * @dev: DRM device
  * @connector: the connector to init
@@ -842,7 +840,7 @@ int drm_connector_init(struct drm_device *dev,
 
 	drm_modeset_lock_all(dev);
 
-	ret = drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR);
+	ret = drm_mode_object_get_reg(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR, false);
 	if (ret)
 		goto out_unlock;
 
@@ -870,6 +868,8 @@ int drm_connector_init(struct drm_device *dev,
 	connector->edid_blob_ptr = NULL;
 	connector->status = connector_status_unknown;
 
+	drm_connector_get_cmdline_mode(connector);
+
 	list_add_tail(&connector->head, &dev->mode_config.connector_list);
 	dev->mode_config.num_connector++;
 
@@ -881,6 +881,8 @@ int drm_connector_init(struct drm_device *dev,
 	drm_object_attach_property(&connector->base,
 				      dev->mode_config.dpms_property, 0);
 
+	connector->debugfs_entry = NULL;
+
 out_put:
 	if (ret)
 		drm_mode_object_put(dev, &connector->base);
@@ -921,6 +923,72 @@ void drm_connector_cleanup(struct drm_connector *connector)
 EXPORT_SYMBOL(drm_connector_cleanup);
 
 /**
+ * drm_connector_index - find the index of a registered connector
+ * @connector: connector to find index for
+ *
+ * Given a registered connector, return the index of that connector within a DRM
+ * device's list of connectors.
+ */
+unsigned int drm_connector_index(struct drm_connector *connector)
+{
+	unsigned int index = 0;
+	struct drm_connector *tmp;
+
+	list_for_each_entry(tmp, &connector->dev->mode_config.connector_list, head) {
+		if (tmp == connector)
+			return index;
+
+		index++;
+	}
+
+	BUG();
+}
+EXPORT_SYMBOL(drm_connector_index);
+
+/**
+ * drm_connector_register - register a connector
+ * @connector: the connector to register
+ *
+ * Register userspace interfaces for a connector
+ *
+ * Returns:
+ * Zero on success, error code on failure.
+ */
+int drm_connector_register(struct drm_connector *connector)
+{
+	int ret;
+
+	drm_mode_object_register(connector->dev, &connector->base);
+
+	ret = drm_sysfs_connector_add(connector);
+	if (ret)
+		return ret;
+
+	ret = drm_debugfs_connector_add(connector);
+	if (ret) {
+		drm_sysfs_connector_remove(connector);
+		return ret;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_connector_register);
+
+/**
+ * drm_connector_unregister - unregister a connector
+ * @connector: the connector to unregister
+ *
+ * Unregister userspace interfaces for a connector
+ */
+void drm_connector_unregister(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_debugfs_connector_remove(connector);
+}
+EXPORT_SYMBOL(drm_connector_unregister);
+
+
+/**
  * drm_connector_unplug_all - unregister connector userspace interfaces
  * @dev: drm device
  *
@@ -934,7 +1002,7 @@ void drm_connector_unplug_all(struct drm_device *dev)
 
 	/* taking the mode config mutex ends up in a clash with sysfs */
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
-		drm_sysfs_connector_remove(connector);
+		drm_connector_unregister(connector);
 
 }
 EXPORT_SYMBOL(drm_connector_unplug_all);
@@ -1182,6 +1250,29 @@ void drm_plane_cleanup(struct drm_plane *plane)
 EXPORT_SYMBOL(drm_plane_cleanup);
 
 /**
+ * drm_plane_index - find the index of a registered plane
+ * @plane: plane to find index for
+ *
+ * Given a registered plane, return the index of that CRTC within a DRM
+ * device's list of planes.
+ */
+unsigned int drm_plane_index(struct drm_plane *plane)
+{
+	unsigned int index = 0;
+	struct drm_plane *tmp;
+
+	list_for_each_entry(tmp, &plane->dev->mode_config.plane_list, head) {
+		if (tmp == plane)
+			return index;
+
+		index++;
+	}
+
+	BUG();
+}
+EXPORT_SYMBOL(drm_plane_index);
+
+/**
  * drm_plane_force_disable - Forcibly disable a plane
  * @plane: plane to disable
  *
@@ -1192,19 +1283,21 @@ EXPORT_SYMBOL(drm_plane_cleanup);
  */
 void drm_plane_force_disable(struct drm_plane *plane)
 {
-	struct drm_framebuffer *old_fb = plane->fb;
 	int ret;
 
-	if (!old_fb)
+	if (!plane->fb)
 		return;
 
+	plane->old_fb = plane->fb;
 	ret = plane->funcs->disable_plane(plane);
 	if (ret) {
 		DRM_ERROR("failed to disable plane with busy fb\n");
+		plane->old_fb = NULL;
 		return;
 	}
 	/* disconnect the plane from the fb and crtc: */
-	__drm_framebuffer_unreference(old_fb);
+	__drm_framebuffer_unreference(plane->old_fb);
+	plane->old_fb = NULL;
 	plane->fb = NULL;
 	plane->crtc = NULL;
 }
@@ -1214,6 +1307,7 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev)
 {
 	struct drm_property *edid;
 	struct drm_property *dpms;
+	struct drm_property *dev_path;
 
 	/*
 	 * Standard properties (apply to all connectors)
@@ -1228,6 +1322,12 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev)
 				   ARRAY_SIZE(drm_dpms_enum_list));
 	dev->mode_config.dpms_property = dpms;
 
+	dev_path = drm_property_create(dev,
+				       DRM_MODE_PROP_BLOB |
+				       DRM_MODE_PROP_IMMUTABLE,
+				       "PATH", 0);
+	dev->mode_config.path_property = dev_path;
+
 	return 0;
 }
 
@@ -1384,6 +1484,33 @@ int drm_mode_create_scaling_mode_property(struct drm_device *dev)
 EXPORT_SYMBOL(drm_mode_create_scaling_mode_property);
 
 /**
+ * drm_mode_create_aspect_ratio_property - create aspect ratio property
+ * @dev: DRM device
+ *
+ * Called by a driver the first time it's needed, must be attached to desired
+ * connectors.
+ *
+ * Returns:
+ * Zero on success, errno on failure.
+ */
+int drm_mode_create_aspect_ratio_property(struct drm_device *dev)
+{
+	if (dev->mode_config.aspect_ratio_property)
+		return 0;
+
+	dev->mode_config.aspect_ratio_property =
+		drm_property_create_enum(dev, 0, "aspect ratio",
+				drm_aspect_ratio_enum_list,
+				ARRAY_SIZE(drm_aspect_ratio_enum_list));
+
+	if (dev->mode_config.aspect_ratio_property == NULL)
+		return -ENOMEM;
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_mode_create_aspect_ratio_property);
+
+/**
  * drm_mode_create_dirty_property - create dirty property
  * @dev: DRM device
  *
@@ -1470,6 +1597,15 @@ int drm_mode_group_init_legacy_group(struct drm_device *dev,
 }
 EXPORT_SYMBOL(drm_mode_group_init_legacy_group);
 
+void drm_reinit_primary_mode_group(struct drm_device *dev)
+{
+	drm_modeset_lock_all(dev);
+	drm_mode_group_destroy(&dev->primary->mode_group);
+	drm_mode_group_init_legacy_group(dev, &dev->primary->mode_group);
+	drm_modeset_unlock_all(dev);
+}
+EXPORT_SYMBOL(drm_reinit_primary_mode_group);
+
 /**
  * drm_crtc_convert_to_umode - convert a drm_display_mode into a modeinfo
  * @out: drm_mode_modeinfo struct to return to the user
@@ -2118,63 +2254,38 @@ out:
 	return ret;
 }
 
-/**
- * drm_mode_setplane - configure a plane's configuration
- * @dev: DRM device
- * @data: ioctl data*
- * @file_priv: DRM file info
+/*
+ * setplane_internal - setplane handler for internal callers
  *
- * Set plane configuration, including placement, fb, scaling, and other factors.
- * Or pass a NULL fb to disable.
+ * Note that we assume an extra reference has already been taken on fb.  If the
+ * update fails, this reference will be dropped before return; if it succeeds,
+ * the previous framebuffer (if any) will be unreferenced instead.
  *
- * Returns:
- * Zero on success, errno on failure.
+ * src_{x,y,w,h} are provided in 16.16 fixed point format
  */
-int drm_mode_setplane(struct drm_device *dev, void *data,
-		      struct drm_file *file_priv)
+static int __setplane_internal(struct drm_plane *plane,
+			       struct drm_crtc *crtc,
+			       struct drm_framebuffer *fb,
+			       int32_t crtc_x, int32_t crtc_y,
+			       uint32_t crtc_w, uint32_t crtc_h,
+			       /* src_{x,y,w,h} values are 16.16 fixed point */
+			       uint32_t src_x, uint32_t src_y,
+			       uint32_t src_w, uint32_t src_h)
 {
-	struct drm_mode_set_plane *plane_req = data;
-	struct drm_plane *plane;
-	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb = NULL, *old_fb = NULL;
 	int ret = 0;
 	unsigned int fb_width, fb_height;
 	int i;
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		return -EINVAL;
-
-	/*
-	 * First, find the plane, crtc, and fb objects.  If not available,
-	 * we don't bother to call the driver.
-	 */
-	plane = drm_plane_find(dev, plane_req->plane_id);
-	if (!plane) {
-		DRM_DEBUG_KMS("Unknown plane ID %d\n",
-			      plane_req->plane_id);
-		return -ENOENT;
-	}
-
 	/* No fb means shut it down */
-	if (!plane_req->fb_id) {
-		drm_modeset_lock_all(dev);
-		old_fb = plane->fb;
+	if (!fb) {
+		plane->old_fb = plane->fb;
 		ret = plane->funcs->disable_plane(plane);
 		if (!ret) {
 			plane->crtc = NULL;
 			plane->fb = NULL;
 		} else {
-			old_fb = NULL;
+			plane->old_fb = NULL;
 		}
-		drm_modeset_unlock_all(dev);
-		goto out;
-	}
-
-	crtc = drm_crtc_find(dev, plane_req->crtc_id);
-	if (!crtc) {
-		DRM_DEBUG_KMS("Unknown crtc ID %d\n",
-			      plane_req->crtc_id);
-		ret = -ENOENT;
 		goto out;
 	}
 
@@ -2185,14 +2296,6 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 		goto out;
 	}
 
-	fb = drm_framebuffer_lookup(dev, plane_req->fb_id);
-	if (!fb) {
-		DRM_DEBUG_KMS("Unknown framebuffer ID %d\n",
-			      plane_req->fb_id);
-		ret = -ENOENT;
-		goto out;
-	}
-
 	/* Check whether this plane supports the fb pixel format. */
 	for (i = 0; i < plane->format_count; i++)
 		if (fb->pixel_format == plane->format_types[i])
@@ -2208,62 +2311,141 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 	fb_height = fb->height << 16;
 
 	/* Make sure source coordinates are inside the fb. */
-	if (plane_req->src_w > fb_width ||
-	    plane_req->src_x > fb_width - plane_req->src_w ||
-	    plane_req->src_h > fb_height ||
-	    plane_req->src_y > fb_height - plane_req->src_h) {
+	if (src_w > fb_width ||
+	    src_x > fb_width - src_w ||
+	    src_h > fb_height ||
+	    src_y > fb_height - src_h) {
 		DRM_DEBUG_KMS("Invalid source coordinates "
 			      "%u.%06ux%u.%06u+%u.%06u+%u.%06u\n",
-			      plane_req->src_w >> 16,
-			      ((plane_req->src_w & 0xffff) * 15625) >> 10,
-			      plane_req->src_h >> 16,
-			      ((plane_req->src_h & 0xffff) * 15625) >> 10,
-			      plane_req->src_x >> 16,
-			      ((plane_req->src_x & 0xffff) * 15625) >> 10,
-			      plane_req->src_y >> 16,
-			      ((plane_req->src_y & 0xffff) * 15625) >> 10);
+			      src_w >> 16, ((src_w & 0xffff) * 15625) >> 10,
+			      src_h >> 16, ((src_h & 0xffff) * 15625) >> 10,
+			      src_x >> 16, ((src_x & 0xffff) * 15625) >> 10,
+			      src_y >> 16, ((src_y & 0xffff) * 15625) >> 10);
 		ret = -ENOSPC;
 		goto out;
 	}
 
-	/* Give drivers some help against integer overflows */
-	if (plane_req->crtc_w > INT_MAX ||
-	    plane_req->crtc_x > INT_MAX - (int32_t) plane_req->crtc_w ||
-	    plane_req->crtc_h > INT_MAX ||
-	    plane_req->crtc_y > INT_MAX - (int32_t) plane_req->crtc_h) {
-		DRM_DEBUG_KMS("Invalid CRTC coordinates %ux%u+%d+%d\n",
-			      plane_req->crtc_w, plane_req->crtc_h,
-			      plane_req->crtc_x, plane_req->crtc_y);
-		ret = -ERANGE;
-		goto out;
-	}
-
-	drm_modeset_lock_all(dev);
-	old_fb = plane->fb;
+	plane->old_fb = plane->fb;
 	ret = plane->funcs->update_plane(plane, crtc, fb,
-					 plane_req->crtc_x, plane_req->crtc_y,
-					 plane_req->crtc_w, plane_req->crtc_h,
-					 plane_req->src_x, plane_req->src_y,
-					 plane_req->src_w, plane_req->src_h);
+					 crtc_x, crtc_y, crtc_w, crtc_h,
+					 src_x, src_y, src_w, src_h);
 	if (!ret) {
 		plane->crtc = crtc;
 		plane->fb = fb;
 		fb = NULL;
 	} else {
-		old_fb = NULL;
+		plane->old_fb = NULL;
 	}
-	drm_modeset_unlock_all(dev);
 
 out:
 	if (fb)
 		drm_framebuffer_unreference(fb);
-	if (old_fb)
-		drm_framebuffer_unreference(old_fb);
+	if (plane->old_fb)
+		drm_framebuffer_unreference(plane->old_fb);
+	plane->old_fb = NULL;
+
+	return ret;
+}
+
+static int setplane_internal(struct drm_plane *plane,
+			     struct drm_crtc *crtc,
+			     struct drm_framebuffer *fb,
+			     int32_t crtc_x, int32_t crtc_y,
+			     uint32_t crtc_w, uint32_t crtc_h,
+			     /* src_{x,y,w,h} values are 16.16 fixed point */
+			     uint32_t src_x, uint32_t src_y,
+			     uint32_t src_w, uint32_t src_h)
+{
+	int ret;
+
+	drm_modeset_lock_all(plane->dev);
+	ret = __setplane_internal(plane, crtc, fb,
+				  crtc_x, crtc_y, crtc_w, crtc_h,
+				  src_x, src_y, src_w, src_h);
+	drm_modeset_unlock_all(plane->dev);
 
 	return ret;
 }
 
 /**
+ * drm_mode_setplane - configure a plane's configuration
+ * @dev: DRM device
+ * @data: ioctl data*
+ * @file_priv: DRM file info
+ *
+ * Set plane configuration, including placement, fb, scaling, and other factors.
+ * Or pass a NULL fb to disable (planes may be disabled without providing a
+ * valid crtc).
+ *
+ * Returns:
+ * Zero on success, errno on failure.
+ */
+int drm_mode_setplane(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
+{
+	struct drm_mode_set_plane *plane_req = data;
+	struct drm_mode_object *obj;
+	struct drm_plane *plane;
+	struct drm_crtc *crtc = NULL;
+	struct drm_framebuffer *fb = NULL;
+
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		return -EINVAL;
+
+	/* Give drivers some help against integer overflows */
+	if (plane_req->crtc_w > INT_MAX ||
+	    plane_req->crtc_x > INT_MAX - (int32_t) plane_req->crtc_w ||
+	    plane_req->crtc_h > INT_MAX ||
+	    plane_req->crtc_y > INT_MAX - (int32_t) plane_req->crtc_h) {
+		DRM_DEBUG_KMS("Invalid CRTC coordinates %ux%u+%d+%d\n",
+			      plane_req->crtc_w, plane_req->crtc_h,
+			      plane_req->crtc_x, plane_req->crtc_y);
+		return -ERANGE;
+	}
+
+	/*
+	 * First, find the plane, crtc, and fb objects.  If not available,
+	 * we don't bother to call the driver.
+	 */
+	obj = drm_mode_object_find(dev, plane_req->plane_id,
+				   DRM_MODE_OBJECT_PLANE);
+	if (!obj) {
+		DRM_DEBUG_KMS("Unknown plane ID %d\n",
+			      plane_req->plane_id);
+		return -ENOENT;
+	}
+	plane = obj_to_plane(obj);
+
+	if (plane_req->fb_id) {
+		fb = drm_framebuffer_lookup(dev, plane_req->fb_id);
+		if (!fb) {
+			DRM_DEBUG_KMS("Unknown framebuffer ID %d\n",
+				      plane_req->fb_id);
+			return -ENOENT;
+		}
+
+		obj = drm_mode_object_find(dev, plane_req->crtc_id,
+					   DRM_MODE_OBJECT_CRTC);
+		if (!obj) {
+			DRM_DEBUG_KMS("Unknown crtc ID %d\n",
+				      plane_req->crtc_id);
+			return -ENOENT;
+		}
+		crtc = obj_to_crtc(obj);
+	}
+
+	/*
+	 * setplane_internal will take care of deref'ing either the old or new
+	 * framebuffer depending on success.
+	 */
+	return setplane_internal(plane, crtc, fb,
+				 plane_req->crtc_x, plane_req->crtc_y,
+				 plane_req->crtc_w, plane_req->crtc_h,
+				 plane_req->src_x, plane_req->src_y,
+				 plane_req->src_w, plane_req->src_h);
+}
+
+/**
  * drm_mode_set_config_internal - helper to call ->set_config
  * @set: modeset config to set
  *
@@ -2286,7 +2468,7 @@ int drm_mode_set_config_internal(struct drm_mode_set *set)
 	 * crtcs. Atomic modeset will have saner semantics ...
 	 */
 	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head)
-		tmp->old_fb = tmp->primary->fb;
+		tmp->primary->old_fb = tmp->primary->fb;
 
 	fb = set->fb;
 
@@ -2299,8 +2481,9 @@ int drm_mode_set_config_internal(struct drm_mode_set *set)
 	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head) {
 		if (tmp->primary->fb)
 			drm_framebuffer_reference(tmp->primary->fb);
-		if (tmp->old_fb)
-			drm_framebuffer_unreference(tmp->old_fb);
+		if (tmp->primary->old_fb)
+			drm_framebuffer_unreference(tmp->primary->old_fb);
+		tmp->primary->old_fb = NULL;
 	}
 
 	return ret;
@@ -2509,6 +2692,101 @@ out:
 	return ret;
 }
 
+/**
+ * drm_mode_cursor_universal - translate legacy cursor ioctl call into a
+ *     universal plane handler call
+ * @crtc: crtc to update cursor for
+ * @req: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
+ *
+ * Legacy cursor ioctl's work directly with driver buffer handles.  To
+ * translate legacy ioctl calls into universal plane handler calls, we need to
+ * wrap the native buffer handle in a drm_framebuffer.
+ *
+ * Note that we assume any handle passed to the legacy ioctls was a 32-bit ARGB
+ * buffer with a pitch of 4*width; the universal plane interface should be used
+ * directly in cases where the hardware can support other buffer settings and
+ * userspace wants to make use of these capabilities.
+ *
+ * Returns:
+ * Zero on success, errno on failure.
+ */
+static int drm_mode_cursor_universal(struct drm_crtc *crtc,
+				     struct drm_mode_cursor2 *req,
+				     struct drm_file *file_priv)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_framebuffer *fb = NULL;
+	struct drm_mode_fb_cmd2 fbreq = {
+		.width = req->width,
+		.height = req->height,
+		.pixel_format = DRM_FORMAT_ARGB8888,
+		.pitches = { req->width * 4 },
+		.handles = { req->handle },
+	};
+	int32_t crtc_x, crtc_y;
+	uint32_t crtc_w = 0, crtc_h = 0;
+	uint32_t src_w = 0, src_h = 0;
+	int ret = 0;
+
+	BUG_ON(!crtc->cursor);
+	WARN_ON(crtc->cursor->crtc != crtc && crtc->cursor->crtc != NULL);
+
+	/*
+	 * Obtain fb we'll be using (either new or existing) and take an extra
+	 * reference to it if fb != null.  setplane will take care of dropping
+	 * the reference if the plane update fails.
+	 */
+	if (req->flags & DRM_MODE_CURSOR_BO) {
+		if (req->handle) {
+			fb = add_framebuffer_internal(dev, &fbreq, file_priv);
+			if (IS_ERR(fb)) {
+				DRM_DEBUG_KMS("failed to wrap cursor buffer in drm framebuffer\n");
+				return PTR_ERR(fb);
+			}
+
+			drm_framebuffer_reference(fb);
+		} else {
+			fb = NULL;
+		}
+	} else {
+		fb = crtc->cursor->fb;
+		if (fb)
+			drm_framebuffer_reference(fb);
+	}
+
+	if (req->flags & DRM_MODE_CURSOR_MOVE) {
+		crtc_x = req->x;
+		crtc_y = req->y;
+	} else {
+		crtc_x = crtc->cursor_x;
+		crtc_y = crtc->cursor_y;
+	}
+
+	if (fb) {
+		crtc_w = fb->width;
+		crtc_h = fb->height;
+		src_w = fb->width << 16;
+		src_h = fb->height << 16;
+	}
+
+	/*
+	 * setplane_internal will take care of deref'ing either the old or new
+	 * framebuffer depending on success.
+	 */
+	ret = __setplane_internal(crtc->cursor, crtc, fb,
+				crtc_x, crtc_y, crtc_w, crtc_h,
+				0, 0, src_w, src_h);
+
+	/* Update successful; save new cursor position, if necessary */
+	if (ret == 0 && req->flags & DRM_MODE_CURSOR_MOVE) {
+		crtc->cursor_x = req->x;
+		crtc->cursor_y = req->y;
+	}
+
+	return ret;
+}
+
 static int drm_mode_cursor_common(struct drm_device *dev,
 				  struct drm_mode_cursor2 *req,
 				  struct drm_file *file_priv)
@@ -2528,7 +2806,16 @@ static int drm_mode_cursor_common(struct drm_device *dev,
 		return -ENOENT;
 	}
 
-	drm_modeset_lock(&crtc->mutex, NULL);
+	/*
+	 * If this crtc has a universal cursor plane, call that plane's update
+	 * handler rather than using legacy cursor handlers.
+	 */
+	drm_modeset_lock_crtc(crtc);
+	if (crtc->cursor) {
+		ret = drm_mode_cursor_universal(crtc, req, file_priv);
+		goto out;
+	}
+
 	if (req->flags & DRM_MODE_CURSOR_BO) {
 		if (!crtc->funcs->cursor_set && !crtc->funcs->cursor_set2) {
 			ret = -ENXIO;
@@ -2552,7 +2839,7 @@ static int drm_mode_cursor_common(struct drm_device *dev,
 		}
 	}
 out:
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 
 	return ret;
 
@@ -2827,56 +3114,38 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
 	return 0;
 }
 
-/**
- * drm_mode_addfb2 - add an FB to the graphics configuration
- * @dev: drm device for the ioctl
- * @data: data pointer for the ioctl
- * @file_priv: drm file for the ioctl call
- *
- * Add a new FB to the specified CRTC, given a user request with format. This is
- * the 2nd version of the addfb ioctl, which supports multi-planar framebuffers
- * and uses fourcc codes as pixel format specifiers.
- *
- * Called by the user via ioctl.
- *
- * Returns:
- * Zero on success, errno on failure.
- */
-int drm_mode_addfb2(struct drm_device *dev,
-		    void *data, struct drm_file *file_priv)
+static struct drm_framebuffer *add_framebuffer_internal(struct drm_device *dev,
+							struct drm_mode_fb_cmd2 *r,
+							struct drm_file *file_priv)
 {
-	struct drm_mode_fb_cmd2 *r = data;
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_framebuffer *fb;
 	int ret;
 
-	if (!drm_core_check_feature(dev, DRIVER_MODESET))
-		return -EINVAL;
-
 	if (r->flags & ~DRM_MODE_FB_INTERLACED) {
 		DRM_DEBUG_KMS("bad framebuffer flags 0x%08x\n", r->flags);
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 	}
 
 	if ((config->min_width > r->width) || (r->width > config->max_width)) {
 		DRM_DEBUG_KMS("bad framebuffer width %d, should be >= %d && <= %d\n",
 			  r->width, config->min_width, config->max_width);
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 	}
 	if ((config->min_height > r->height) || (r->height > config->max_height)) {
 		DRM_DEBUG_KMS("bad framebuffer height %d, should be >= %d && <= %d\n",
 			  r->height, config->min_height, config->max_height);
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 	}
 
 	ret = framebuffer_check(r);
 	if (ret)
-		return ret;
+		return ERR_PTR(ret);
 
 	fb = dev->mode_config.funcs->fb_create(dev, file_priv, r);
 	if (IS_ERR(fb)) {
 		DRM_DEBUG_KMS("could not create framebuffer\n");
-		return PTR_ERR(fb);
+		return fb;
 	}
 
 	mutex_lock(&file_priv->fbs_lock);
@@ -2885,8 +3154,37 @@ int drm_mode_addfb2(struct drm_device *dev,
 	DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id);
 	mutex_unlock(&file_priv->fbs_lock);
 
+	return fb;
+}
 
-	return ret;
+/**
+ * drm_mode_addfb2 - add an FB to the graphics configuration
+ * @dev: drm device for the ioctl
+ * @data: data pointer for the ioctl
+ * @file_priv: drm file for the ioctl call
+ *
+ * Add a new FB to the specified CRTC, given a user request with format. This is
+ * the 2nd version of the addfb ioctl, which supports multi-planar framebuffers
+ * and uses fourcc codes as pixel format specifiers.
+ *
+ * Called by the user via ioctl.
+ *
+ * Returns:
+ * Zero on success, errno on failure.
+ */
+int drm_mode_addfb2(struct drm_device *dev,
+		    void *data, struct drm_file *file_priv)
+{
+	struct drm_framebuffer *fb;
+
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		return -EINVAL;
+
+	fb = add_framebuffer_internal(dev, data, file_priv);
+	if (IS_ERR(fb))
+		return PTR_ERR(fb);
+
+	return 0;
 }
 
 /**
@@ -3102,7 +3400,16 @@ void drm_fb_release(struct drm_file *priv)
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_framebuffer *fb, *tfb;
 
-	mutex_lock(&priv->fbs_lock);
+	/*
+	 * When the file gets released that means no one else can access the fb
+	 * list any more, so no need to grab fpriv->fbs_lock. And we need to to
+	 * avoid upsetting lockdep since the universal cursor code adds a
+	 * framebuffer while holding mutex locks.
+	 *
+	 * Note that a real deadlock between fpriv->fbs_lock and the modeset
+	 * locks is impossible here since no one else but this function can get
+	 * at it any more.
+	 */
 	list_for_each_entry_safe(fb, tfb, &priv->fbs, filp_head) {
 
 		mutex_lock(&dev->mode_config.fb_lock);
@@ -3115,7 +3422,6 @@ void drm_fb_release(struct drm_file *priv)
 		/* This will also drop the fpriv->fbs reference. */
 		drm_framebuffer_remove(fb);
 	}
-	mutex_unlock(&priv->fbs_lock);
 }
 
 /**
@@ -3176,7 +3482,7 @@ fail:
 EXPORT_SYMBOL(drm_property_create);
 
 /**
- * drm_property_create - create a new enumeration property type
+ * drm_property_create_enum - create a new enumeration property type
  * @dev: drm device
  * @flags: flags specifying the property type
  * @name: name of the property
@@ -3222,14 +3528,15 @@ struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags,
 EXPORT_SYMBOL(drm_property_create_enum);
 
 /**
- * drm_property_create - create a new bitmask property type
+ * drm_property_create_bitmask - create a new bitmask property type
  * @dev: drm device
  * @flags: flags specifying the property type
  * @name: name of the property
  * @props: enumeration lists with property bitflags
- * @num_values: number of pre-defined values
+ * @num_props: size of the @props array
+ * @supported_bits: bitmask of all supported enumeration values
  *
- * This creates a new generic drm property which can then be attached to a drm
+ * This creates a new bitmask drm property which can then be attached to a drm
  * object with drm_object_attach_property. The returned property object must be
  * freed with drm_property_destroy.
  *
@@ -3242,19 +3549,28 @@ EXPORT_SYMBOL(drm_property_create_enum);
 struct drm_property *drm_property_create_bitmask(struct drm_device *dev,
 					 int flags, const char *name,
 					 const struct drm_prop_enum_list *props,
-					 int num_values)
+					 int num_props,
+					 uint64_t supported_bits)
 {
 	struct drm_property *property;
-	int i, ret;
+	int i, ret, index = 0;
+	int num_values = hweight64(supported_bits);
 
 	flags |= DRM_MODE_PROP_BITMASK;
 
 	property = drm_property_create(dev, flags, name, num_values);
 	if (!property)
 		return NULL;
+	for (i = 0; i < num_props; i++) {
+		if (!(supported_bits & (1ULL << props[i].type)))
+			continue;
 
-	for (i = 0; i < num_values; i++) {
-		ret = drm_property_add_enum(property, i,
+		if (WARN_ON(index >= num_values)) {
+			drm_property_destroy(dev, property);
+			return NULL;
+		}
+
+		ret = drm_property_add_enum(property, index++,
 				      props[i].type,
 				      props[i].name);
 		if (ret) {
@@ -3284,7 +3600,7 @@ static struct drm_property *property_create_range(struct drm_device *dev,
 }
 
 /**
- * drm_property_create - create a new ranged property type
+ * drm_property_create_range - create a new ranged property type
  * @dev: drm device
  * @flags: flags specifying the property type
  * @name: name of the property
@@ -3703,6 +4019,25 @@ done:
 	return ret;
 }
 
+int drm_mode_connector_set_path_property(struct drm_connector *connector,
+					 char *path)
+{
+	struct drm_device *dev = connector->dev;
+	int ret, size;
+	size = strlen(path) + 1;
+
+	connector->path_blob_ptr = drm_property_create_blob(connector->dev,
+							    size, path);
+	if (!connector->path_blob_ptr)
+		return -EINVAL;
+
+	ret = drm_object_property_set_value(&connector->base,
+					    dev->mode_config.path_property,
+					    connector->path_blob_ptr->base.id);
+	return ret;
+}
+EXPORT_SYMBOL(drm_mode_connector_set_path_property);
+
 /**
  * drm_mode_connector_update_edid_property - update the edid property of a connector
  * @connector: drm connector
@@ -3720,6 +4055,10 @@ int drm_mode_connector_update_edid_property(struct drm_connector *connector,
 	struct drm_device *dev = connector->dev;
 	int ret, size;
 
+	/* ignore requests to set edid when overridden */
+	if (connector->override_edid)
+		return 0;
+
 	if (connector->edid_blob_ptr)
 		drm_property_destroy_blob(dev, connector->edid_blob_ptr);
 
@@ -3857,12 +4196,25 @@ static int drm_mode_crtc_set_obj_prop(struct drm_mode_object *obj,
 	return ret;
 }
 
-static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj,
-				      struct drm_property *property,
-				      uint64_t value)
+/**
+ * drm_mode_plane_set_obj_prop - set the value of a property
+ * @plane: drm plane object to set property value for
+ * @property: property to set
+ * @value: value the property should be set to
+ *
+ * This functions sets a given property on a given plane object. This function
+ * calls the driver's ->set_property callback and changes the software state of
+ * the property if the callback succeeds.
+ *
+ * Returns:
+ * Zero on success, error code on failure.
+ */
+int drm_mode_plane_set_obj_prop(struct drm_plane *plane,
+				struct drm_property *property,
+				uint64_t value)
 {
 	int ret = -EINVAL;
-	struct drm_plane *plane = obj_to_plane(obj);
+	struct drm_mode_object *obj = &plane->base;
 
 	if (plane->funcs->set_property)
 		ret = plane->funcs->set_property(plane, property, value);
@@ -3871,6 +4223,7 @@ static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj,
 
 	return ret;
 }
+EXPORT_SYMBOL(drm_mode_plane_set_obj_prop);
 
 /**
  * drm_mode_getproperty_ioctl - get the current value of a object's property
@@ -4009,7 +4362,8 @@ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data,
 		ret = drm_mode_crtc_set_obj_prop(arg_obj, property, arg->value);
 		break;
 	case DRM_MODE_OBJECT_PLANE:
-		ret = drm_mode_plane_set_obj_prop(arg_obj, property, arg->value);
+		ret = drm_mode_plane_set_obj_prop(obj_to_plane(arg_obj),
+						  property, arg->value);
 		break;
 	}
 
@@ -4229,7 +4583,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 {
 	struct drm_mode_crtc_page_flip *page_flip = data;
 	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb = NULL, *old_fb = NULL;
+	struct drm_framebuffer *fb = NULL;
 	struct drm_pending_vblank_event *e = NULL;
 	unsigned long flags;
 	int ret = -EINVAL;
@@ -4245,7 +4599,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 	if (!crtc)
 		return -ENOENT;
 
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc);
 	if (crtc->primary->fb == NULL) {
 		/* The framebuffer is currently unbound, presumably
 		 * due to a hotplug event, that userspace has not
@@ -4301,7 +4655,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			(void (*) (struct drm_pending_event *)) kfree;
 	}
 
-	old_fb = crtc->primary->fb;
+	crtc->primary->old_fb = crtc->primary->fb;
 	ret = crtc->funcs->page_flip(crtc, fb, e, page_flip->flags);
 	if (ret) {
 		if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) {
@@ -4311,7 +4665,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 			kfree(e);
 		}
 		/* Keep the old fb, don't unref it. */
-		old_fb = NULL;
+		crtc->primary->old_fb = NULL;
 	} else {
 		/*
 		 * Warn if the driver hasn't properly updated the crtc->fb
@@ -4327,9 +4681,10 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,
 out:
 	if (fb)
 		drm_framebuffer_unreference(fb);
-	if (old_fb)
-		drm_framebuffer_unreference(old_fb);
-	drm_modeset_unlock(&crtc->mutex);
+	if (crtc->primary->old_fb)
+		drm_framebuffer_unreference(crtc->primary->old_fb);
+	crtc->primary->old_fb = NULL;
+	drm_modeset_unlock_crtc(crtc);
 
 	return ret;
 }
@@ -4345,9 +4700,14 @@ out:
 void drm_mode_config_reset(struct drm_device *dev)
 {
 	struct drm_crtc *crtc;
+	struct drm_plane *plane;
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
 
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+		if (plane->funcs->reset)
+			plane->funcs->reset(plane);
+
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 		if (crtc->funcs->reset)
 			crtc->funcs->reset(crtc);
@@ -4396,8 +4756,9 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev,
 		return -EINVAL;
 
 	/* overflow checks for 32bit size calculations */
+	/* NOTE: DIV_ROUND_UP() can overflow */
 	cpp = DIV_ROUND_UP(args->bpp, 8);
-	if (cpp > 0xffffffffU / args->width)
+	if (!cpp || cpp > 0xffffffffU / args->width)
 		return -EINVAL;
 	stride = cpp * args->width;
 	if (args->height > 0xffffffffU / stride)
@@ -4680,6 +5041,36 @@ int drm_format_vert_chroma_subsampling(uint32_t format)
 EXPORT_SYMBOL(drm_format_vert_chroma_subsampling);
 
 /**
+ * drm_rotation_simplify() - Try to simplify the rotation
+ * @rotation: Rotation to be simplified
+ * @supported_rotations: Supported rotations
+ *
+ * Attempt to simplify the rotation to a form that is supported.
+ * Eg. if the hardware supports everything except DRM_REFLECT_X
+ * one could call this function like this:
+ *
+ * drm_rotation_simplify(rotation, BIT(DRM_ROTATE_0) |
+ *                       BIT(DRM_ROTATE_90) | BIT(DRM_ROTATE_180) |
+ *                       BIT(DRM_ROTATE_270) | BIT(DRM_REFLECT_Y));
+ *
+ * to eliminate the DRM_ROTATE_X flag. Depending on what kind of
+ * transforms the hardware supports, this function may not
+ * be able to produce a supported transform, so the caller should
+ * check the result afterwards.
+ */
+unsigned int drm_rotation_simplify(unsigned int rotation,
+				   unsigned int supported_rotations)
+{
+	if (rotation & ~supported_rotations) {
+		rotation ^= BIT(DRM_REFLECT_X) | BIT(DRM_REFLECT_Y);
+		rotation = (rotation & ~0xf) | BIT((ffs(rotation & 0xf) + 1) % 4);
+	}
+
+	return rotation;
+}
+EXPORT_SYMBOL(drm_rotation_simplify);
+
+/**
  * drm_mode_config_init - initialize DRM mode_configuration structure
  * @dev: DRM device
  *
@@ -4797,3 +5188,21 @@ void drm_mode_config_cleanup(struct drm_device *dev)
 	drm_modeset_lock_fini(&dev->mode_config.connection_mutex);
 }
 EXPORT_SYMBOL(drm_mode_config_cleanup);
+
+struct drm_property *drm_mode_create_rotation_property(struct drm_device *dev,
+						       unsigned int supported_rotations)
+{
+	static const struct drm_prop_enum_list props[] = {
+		{ DRM_ROTATE_0,   "rotate-0" },
+		{ DRM_ROTATE_90,  "rotate-90" },
+		{ DRM_ROTATE_180, "rotate-180" },
+		{ DRM_ROTATE_270, "rotate-270" },
+		{ DRM_REFLECT_X,  "reflect-x" },
+		{ DRM_REFLECT_Y,  "reflect-y" },
+	};
+
+	return drm_property_create_bitmask(dev, 0, "rotation",
+					   props, ARRAY_SIZE(props),
+					   supported_rotations);
+}
+EXPORT_SYMBOL(drm_mode_create_rotation_property);
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index 78b37f3febd3..6c65a0a28fbd 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -818,6 +818,7 @@ void drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb,
 	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &fb->depth,
 				    &fb->bits_per_pixel);
 	fb->pixel_format = mode_cmd->pixel_format;
+	fb->flags = mode_cmd->flags;
 }
 EXPORT_SYMBOL(drm_helper_mode_fill_fb_struct);
 
diff --git a/drivers/gpu/drm/drm_debugfs.c b/drivers/gpu/drm/drm_debugfs.c
index b4b51d46f339..3bcf8e6a85b3 100644
--- a/drivers/gpu/drm/drm_debugfs.c
+++ b/drivers/gpu/drm/drm_debugfs.c
@@ -35,6 +35,8 @@
 #include <linux/slab.h>
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include <drm/drm_edid.h>
+#include "drm_internal.h"
 
 #if defined(CONFIG_DEBUG_FS)
 
@@ -48,9 +50,7 @@ static const struct drm_info_list drm_debugfs_list[] = {
 	{"clients", drm_clients_info, 0},
 	{"bufs", drm_bufs_info, 0},
 	{"gem_names", drm_gem_name_info, DRIVER_GEM},
-#if DRM_DEBUG_CODE
 	{"vma", drm_vma_info, 0},
-#endif
 };
 #define DRM_DEBUGFS_ENTRIES ARRAY_SIZE(drm_debugfs_list)
 
@@ -237,5 +237,186 @@ int drm_debugfs_cleanup(struct drm_minor *minor)
 	return 0;
 }
 
+static int connector_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	const char *status;
+
+	switch (connector->force) {
+	case DRM_FORCE_ON:
+		status = "on\n";
+		break;
+
+	case DRM_FORCE_ON_DIGITAL:
+		status = "digital\n";
+		break;
+
+	case DRM_FORCE_OFF:
+		status = "off\n";
+		break;
+
+	case DRM_FORCE_UNSPECIFIED:
+		status = "unspecified\n";
+		break;
+
+	default:
+		return 0;
+	}
+
+	seq_puts(m, status);
+
+	return 0;
+}
+
+static int connector_open(struct inode *inode, struct file *file)
+{
+	struct drm_connector *dev = inode->i_private;
+
+	return single_open(file, connector_show, dev);
+}
+
+static ssize_t connector_write(struct file *file, const char __user *ubuf,
+			       size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_connector *connector = m->private;
+	char buf[12];
+
+	if (len > sizeof(buf) - 1)
+		return -EINVAL;
+
+	if (copy_from_user(buf, ubuf, len))
+		return -EFAULT;
+
+	buf[len] = '\0';
+
+	if (!strcmp(buf, "on"))
+		connector->force = DRM_FORCE_ON;
+	else if (!strcmp(buf, "digital"))
+		connector->force = DRM_FORCE_ON_DIGITAL;
+	else if (!strcmp(buf, "off"))
+		connector->force = DRM_FORCE_OFF;
+	else if (!strcmp(buf, "unspecified"))
+		connector->force = DRM_FORCE_UNSPECIFIED;
+	else
+		return -EINVAL;
+
+	return len;
+}
+
+static int edid_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct drm_property_blob *edid = connector->edid_blob_ptr;
+
+	if (connector->override_edid && edid)
+		seq_write(m, edid->data, edid->length);
+
+	return 0;
+}
+
+static int edid_open(struct inode *inode, struct file *file)
+{
+	struct drm_connector *dev = inode->i_private;
+
+	return single_open(file, edid_show, dev);
+}
+
+static ssize_t edid_write(struct file *file, const char __user *ubuf,
+			  size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_connector *connector = m->private;
+	char *buf;
+	struct edid *edid;
+	int ret;
+
+	buf = memdup_user(ubuf, len);
+	if (IS_ERR(buf))
+		return PTR_ERR(buf);
+
+	edid = (struct edid *) buf;
+
+	if (len == 5 && !strncmp(buf, "reset", 5)) {
+		connector->override_edid = false;
+		ret = drm_mode_connector_update_edid_property(connector, NULL);
+	} else if (len < EDID_LENGTH ||
+		   EDID_LENGTH * (1 + edid->extensions) > len)
+		ret = -EINVAL;
+	else {
+		connector->override_edid = false;
+		ret = drm_mode_connector_update_edid_property(connector, edid);
+		if (!ret)
+			connector->override_edid = true;
+	}
+
+	kfree(buf);
+
+	return (ret) ? ret : len;
+}
+
+static const struct file_operations drm_edid_fops = {
+	.owner = THIS_MODULE,
+	.open = edid_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = edid_write
+};
+
+
+static const struct file_operations drm_connector_fops = {
+	.owner = THIS_MODULE,
+	.open = connector_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = connector_write
+};
+
+int drm_debugfs_connector_add(struct drm_connector *connector)
+{
+	struct drm_minor *minor = connector->dev->primary;
+	struct dentry *root, *ent;
+
+	if (!minor->debugfs_root)
+		return -1;
+
+	root = debugfs_create_dir(connector->name, minor->debugfs_root);
+	if (!root)
+		return -ENOMEM;
+
+	connector->debugfs_entry = root;
+
+	/* force */
+	ent = debugfs_create_file("force", S_IRUGO | S_IWUSR, root, connector,
+				  &drm_connector_fops);
+	if (!ent)
+		goto error;
+
+	/* edid */
+	ent = debugfs_create_file("edid_override", S_IRUGO | S_IWUSR, root,
+				  connector, &drm_edid_fops);
+	if (!ent)
+		goto error;
+
+	return 0;
+
+error:
+	debugfs_remove_recursive(connector->debugfs_entry);
+	connector->debugfs_entry = NULL;
+	return -ENOMEM;
+}
+
+void drm_debugfs_connector_remove(struct drm_connector *connector)
+{
+	if (!connector->debugfs_entry)
+		return;
+
+	debugfs_remove_recursive(connector->debugfs_entry);
+
+	connector->debugfs_entry = NULL;
+}
+
 #endif /* CONFIG_DEBUG_FS */
 
diff --git a/drivers/gpu/drm/drm_dma.c b/drivers/gpu/drm/drm_dma.c
index 8a140a953754..ea481800ef56 100644
--- a/drivers/gpu/drm/drm_dma.c
+++ b/drivers/gpu/drm/drm_dma.c
@@ -35,6 +35,7 @@
 
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 /**
  * Initialize the DMA data.
@@ -124,7 +125,7 @@ void drm_legacy_dma_takedown(struct drm_device *dev)
  *
  * Resets the fields of \p buf.
  */
-void drm_free_buffer(struct drm_device *dev, struct drm_buf * buf)
+void drm_legacy_free_buffer(struct drm_device *dev, struct drm_buf * buf)
 {
 	if (!buf)
 		return;
@@ -142,8 +143,8 @@ void drm_free_buffer(struct drm_device *dev, struct drm_buf * buf)
  *
  * Frees each buffer associated with \p file_priv not already on the hardware.
  */
-void drm_core_reclaim_buffers(struct drm_device *dev,
-			      struct drm_file *file_priv)
+void drm_legacy_reclaim_buffers(struct drm_device *dev,
+				struct drm_file *file_priv)
 {
 	struct drm_device_dma *dma = dev->dma;
 	int i;
@@ -154,7 +155,7 @@ void drm_core_reclaim_buffers(struct drm_device *dev,
 		if (dma->buflist[i]->file_priv == file_priv) {
 			switch (dma->buflist[i]->list) {
 			case DRM_LIST_NONE:
-				drm_free_buffer(dev, dma->buflist[i]);
+				drm_legacy_free_buffer(dev, dma->buflist[i]);
 				break;
 			case DRM_LIST_WAIT:
 				dma->buflist[i]->list = DRM_LIST_RECLAIM;
@@ -166,5 +167,3 @@ void drm_core_reclaim_buffers(struct drm_device *dev,
 		}
 	}
 }
-
-EXPORT_SYMBOL(drm_core_reclaim_buffers);
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
new file mode 100644
index 000000000000..070f913d2dba
--- /dev/null
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -0,0 +1,2753 @@
+/*
+ * Copyright © 2014 Red Hat
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission.  The copyright holders make no representations
+ * about the suitability of this software for any purpose.  It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/seq_file.h>
+#include <linux/i2c.h>
+#include <drm/drm_dp_mst_helper.h>
+#include <drm/drmP.h>
+
+#include <drm/drm_fixed.h>
+
+/**
+ * DOC: dp mst helper
+ *
+ * These functions contain parts of the DisplayPort 1.2a MultiStream Transport
+ * protocol. The helpers contain a topology manager and bandwidth manager.
+ * The helpers encapsulate the sending and received of sideband msgs.
+ */
+static bool dump_dp_payload_table(struct drm_dp_mst_topology_mgr *mgr,
+				  char *buf);
+static int test_calc_pbn_mode(void);
+
+static void drm_dp_put_port(struct drm_dp_mst_port *port);
+
+static int drm_dp_dpcd_write_payload(struct drm_dp_mst_topology_mgr *mgr,
+				     int id,
+				     struct drm_dp_payload *payload);
+
+static int drm_dp_send_dpcd_write(struct drm_dp_mst_topology_mgr *mgr,
+				  struct drm_dp_mst_port *port,
+				  int offset, int size, u8 *bytes);
+
+static int drm_dp_send_link_address(struct drm_dp_mst_topology_mgr *mgr,
+				    struct drm_dp_mst_branch *mstb);
+static int drm_dp_send_enum_path_resources(struct drm_dp_mst_topology_mgr *mgr,
+					   struct drm_dp_mst_branch *mstb,
+					   struct drm_dp_mst_port *port);
+static bool drm_dp_validate_guid(struct drm_dp_mst_topology_mgr *mgr,
+				 u8 *guid);
+
+static int drm_dp_mst_register_i2c_bus(struct drm_dp_aux *aux);
+static void drm_dp_mst_unregister_i2c_bus(struct drm_dp_aux *aux);
+static void drm_dp_mst_kick_tx(struct drm_dp_mst_topology_mgr *mgr);
+/* sideband msg handling */
+static u8 drm_dp_msg_header_crc4(const uint8_t *data, size_t num_nibbles)
+{
+	u8 bitmask = 0x80;
+	u8 bitshift = 7;
+	u8 array_index = 0;
+	int number_of_bits = num_nibbles * 4;
+	u8 remainder = 0;
+
+	while (number_of_bits != 0) {
+		number_of_bits--;
+		remainder <<= 1;
+		remainder |= (data[array_index] & bitmask) >> bitshift;
+		bitmask >>= 1;
+		bitshift--;
+		if (bitmask == 0) {
+			bitmask = 0x80;
+			bitshift = 7;
+			array_index++;
+		}
+		if ((remainder & 0x10) == 0x10)
+			remainder ^= 0x13;
+	}
+
+	number_of_bits = 4;
+	while (number_of_bits != 0) {
+		number_of_bits--;
+		remainder <<= 1;
+		if ((remainder & 0x10) != 0)
+			remainder ^= 0x13;
+	}
+
+	return remainder;
+}
+
+static u8 drm_dp_msg_data_crc4(const uint8_t *data, u8 number_of_bytes)
+{
+	u8 bitmask = 0x80;
+	u8 bitshift = 7;
+	u8 array_index = 0;
+	int number_of_bits = number_of_bytes * 8;
+	u16 remainder = 0;
+
+	while (number_of_bits != 0) {
+		number_of_bits--;
+		remainder <<= 1;
+		remainder |= (data[array_index] & bitmask) >> bitshift;
+		bitmask >>= 1;
+		bitshift--;
+		if (bitmask == 0) {
+			bitmask = 0x80;
+			bitshift = 7;
+			array_index++;
+		}
+		if ((remainder & 0x100) == 0x100)
+			remainder ^= 0xd5;
+	}
+
+	number_of_bits = 8;
+	while (number_of_bits != 0) {
+		number_of_bits--;
+		remainder <<= 1;
+		if ((remainder & 0x100) != 0)
+			remainder ^= 0xd5;
+	}
+
+	return remainder & 0xff;
+}
+static inline u8 drm_dp_calc_sb_hdr_size(struct drm_dp_sideband_msg_hdr *hdr)
+{
+	u8 size = 3;
+	size += (hdr->lct / 2);
+	return size;
+}
+
+static void drm_dp_encode_sideband_msg_hdr(struct drm_dp_sideband_msg_hdr *hdr,
+					   u8 *buf, int *len)
+{
+	int idx = 0;
+	int i;
+	u8 crc4;
+	buf[idx++] = ((hdr->lct & 0xf) << 4) | (hdr->lcr & 0xf);
+	for (i = 0; i < (hdr->lct / 2); i++)
+		buf[idx++] = hdr->rad[i];
+	buf[idx++] = (hdr->broadcast << 7) | (hdr->path_msg << 6) |
+		(hdr->msg_len & 0x3f);
+	buf[idx++] = (hdr->somt << 7) | (hdr->eomt << 6) | (hdr->seqno << 4);
+
+	crc4 = drm_dp_msg_header_crc4(buf, (idx * 2) - 1);
+	buf[idx - 1] |= (crc4 & 0xf);
+
+	*len = idx;
+}
+
+static bool drm_dp_decode_sideband_msg_hdr(struct drm_dp_sideband_msg_hdr *hdr,
+					   u8 *buf, int buflen, u8 *hdrlen)
+{
+	u8 crc4;
+	u8 len;
+	int i;
+	u8 idx;
+	if (buf[0] == 0)
+		return false;
+	len = 3;
+	len += ((buf[0] & 0xf0) >> 4) / 2;
+	if (len > buflen)
+		return false;
+	crc4 = drm_dp_msg_header_crc4(buf, (len * 2) - 1);
+
+	if ((crc4 & 0xf) != (buf[len - 1] & 0xf)) {
+		DRM_DEBUG_KMS("crc4 mismatch 0x%x 0x%x\n", crc4, buf[len - 1]);
+		return false;
+	}
+
+	hdr->lct = (buf[0] & 0xf0) >> 4;
+	hdr->lcr = (buf[0] & 0xf);
+	idx = 1;
+	for (i = 0; i < (hdr->lct / 2); i++)
+		hdr->rad[i] = buf[idx++];
+	hdr->broadcast = (buf[idx] >> 7) & 0x1;
+	hdr->path_msg = (buf[idx] >> 6) & 0x1;
+	hdr->msg_len = buf[idx] & 0x3f;
+	idx++;
+	hdr->somt = (buf[idx] >> 7) & 0x1;
+	hdr->eomt = (buf[idx] >> 6) & 0x1;
+	hdr->seqno = (buf[idx] >> 4) & 0x1;
+	idx++;
+	*hdrlen = idx;
+	return true;
+}
+
+static void drm_dp_encode_sideband_req(struct drm_dp_sideband_msg_req_body *req,
+				       struct drm_dp_sideband_msg_tx *raw)
+{
+	int idx = 0;
+	int i;
+	u8 *buf = raw->msg;
+	buf[idx++] = req->req_type & 0x7f;
+
+	switch (req->req_type) {
+	case DP_ENUM_PATH_RESOURCES:
+		buf[idx] = (req->u.port_num.port_number & 0xf) << 4;
+		idx++;
+		break;
+	case DP_ALLOCATE_PAYLOAD:
+		buf[idx] = (req->u.allocate_payload.port_number & 0xf) << 4 |
+			(req->u.allocate_payload.number_sdp_streams & 0xf);
+		idx++;
+		buf[idx] = (req->u.allocate_payload.vcpi & 0x7f);
+		idx++;
+		buf[idx] = (req->u.allocate_payload.pbn >> 8);
+		idx++;
+		buf[idx] = (req->u.allocate_payload.pbn & 0xff);
+		idx++;
+		for (i = 0; i < req->u.allocate_payload.number_sdp_streams / 2; i++) {
+			buf[idx] = ((req->u.allocate_payload.sdp_stream_sink[i * 2] & 0xf) << 4) |
+				(req->u.allocate_payload.sdp_stream_sink[i * 2 + 1] & 0xf);
+			idx++;
+		}
+		if (req->u.allocate_payload.number_sdp_streams & 1) {
+			i = req->u.allocate_payload.number_sdp_streams - 1;
+			buf[idx] = (req->u.allocate_payload.sdp_stream_sink[i] & 0xf) << 4;
+			idx++;
+		}
+		break;
+	case DP_QUERY_PAYLOAD:
+		buf[idx] = (req->u.query_payload.port_number & 0xf) << 4;
+		idx++;
+		buf[idx] = (req->u.query_payload.vcpi & 0x7f);
+		idx++;
+		break;
+	case DP_REMOTE_DPCD_READ:
+		buf[idx] = (req->u.dpcd_read.port_number & 0xf) << 4;
+		buf[idx] |= ((req->u.dpcd_read.dpcd_address & 0xf0000) >> 16) & 0xf;
+		idx++;
+		buf[idx] = (req->u.dpcd_read.dpcd_address & 0xff00) >> 8;
+		idx++;
+		buf[idx] = (req->u.dpcd_read.dpcd_address & 0xff);
+		idx++;
+		buf[idx] = (req->u.dpcd_read.num_bytes);
+		idx++;
+		break;
+
+	case DP_REMOTE_DPCD_WRITE:
+		buf[idx] = (req->u.dpcd_write.port_number & 0xf) << 4;
+		buf[idx] |= ((req->u.dpcd_write.dpcd_address & 0xf0000) >> 16) & 0xf;
+		idx++;
+		buf[idx] = (req->u.dpcd_write.dpcd_address & 0xff00) >> 8;
+		idx++;
+		buf[idx] = (req->u.dpcd_write.dpcd_address & 0xff);
+		idx++;
+		buf[idx] = (req->u.dpcd_write.num_bytes);
+		idx++;
+		memcpy(&buf[idx], req->u.dpcd_write.bytes, req->u.dpcd_write.num_bytes);
+		idx += req->u.dpcd_write.num_bytes;
+		break;
+	case DP_REMOTE_I2C_READ:
+		buf[idx] = (req->u.i2c_read.port_number & 0xf) << 4;
+		buf[idx] |= (req->u.i2c_read.num_transactions & 0x3);
+		idx++;
+		for (i = 0; i < (req->u.i2c_read.num_transactions & 0x3); i++) {
+			buf[idx] = req->u.i2c_read.transactions[i].i2c_dev_id & 0x7f;
+			idx++;
+			buf[idx] = req->u.i2c_read.transactions[i].num_bytes;
+			idx++;
+			memcpy(&buf[idx], req->u.i2c_read.transactions[i].bytes, req->u.i2c_read.transactions[i].num_bytes);
+			idx += req->u.i2c_read.transactions[i].num_bytes;
+
+			buf[idx] = (req->u.i2c_read.transactions[i].no_stop_bit & 0x1) << 5;
+			buf[idx] |= (req->u.i2c_read.transactions[i].i2c_transaction_delay & 0xf);
+			idx++;
+		}
+		buf[idx] = (req->u.i2c_read.read_i2c_device_id) & 0x7f;
+		idx++;
+		buf[idx] = (req->u.i2c_read.num_bytes_read);
+		idx++;
+		break;
+
+	case DP_REMOTE_I2C_WRITE:
+		buf[idx] = (req->u.i2c_write.port_number & 0xf) << 4;
+		idx++;
+		buf[idx] = (req->u.i2c_write.write_i2c_device_id) & 0x7f;
+		idx++;
+		buf[idx] = (req->u.i2c_write.num_bytes);
+		idx++;
+		memcpy(&buf[idx], req->u.i2c_write.bytes, req->u.i2c_write.num_bytes);
+		idx += req->u.i2c_write.num_bytes;
+		break;
+	}
+	raw->cur_len = idx;
+}
+
+static void drm_dp_crc_sideband_chunk_req(u8 *msg, u8 len)
+{
+	u8 crc4;
+	crc4 = drm_dp_msg_data_crc4(msg, len);
+	msg[len] = crc4;
+}
+
+static void drm_dp_encode_sideband_reply(struct drm_dp_sideband_msg_reply_body *rep,
+					 struct drm_dp_sideband_msg_tx *raw)
+{
+	int idx = 0;
+	u8 *buf = raw->msg;
+
+	buf[idx++] = (rep->reply_type & 0x1) << 7 | (rep->req_type & 0x7f);
+
+	raw->cur_len = idx;
+}
+
+/* this adds a chunk of msg to the builder to get the final msg */
+static bool drm_dp_sideband_msg_build(struct drm_dp_sideband_msg_rx *msg,
+				      u8 *replybuf, u8 replybuflen, bool hdr)
+{
+	int ret;
+	u8 crc4;
+
+	if (hdr) {
+		u8 hdrlen;
+		struct drm_dp_sideband_msg_hdr recv_hdr;
+		ret = drm_dp_decode_sideband_msg_hdr(&recv_hdr, replybuf, replybuflen, &hdrlen);
+		if (ret == false) {
+			print_hex_dump(KERN_DEBUG, "failed hdr", DUMP_PREFIX_NONE, 16, 1, replybuf, replybuflen, false);
+			return false;
+		}
+
+		/* get length contained in this portion */
+		msg->curchunk_len = recv_hdr.msg_len;
+		msg->curchunk_hdrlen = hdrlen;
+
+		/* we have already gotten an somt - don't bother parsing */
+		if (recv_hdr.somt && msg->have_somt)
+			return false;
+
+		if (recv_hdr.somt) {
+			memcpy(&msg->initial_hdr, &recv_hdr, sizeof(struct drm_dp_sideband_msg_hdr));
+			msg->have_somt = true;
+		}
+		if (recv_hdr.eomt)
+			msg->have_eomt = true;
+
+		/* copy the bytes for the remainder of this header chunk */
+		msg->curchunk_idx = min(msg->curchunk_len, (u8)(replybuflen - hdrlen));
+		memcpy(&msg->chunk[0], replybuf + hdrlen, msg->curchunk_idx);
+	} else {
+		memcpy(&msg->chunk[msg->curchunk_idx], replybuf, replybuflen);
+		msg->curchunk_idx += replybuflen;
+	}
+
+	if (msg->curchunk_idx >= msg->curchunk_len) {
+		/* do CRC */
+		crc4 = drm_dp_msg_data_crc4(msg->chunk, msg->curchunk_len - 1);
+		/* copy chunk into bigger msg */
+		memcpy(&msg->msg[msg->curlen], msg->chunk, msg->curchunk_len - 1);
+		msg->curlen += msg->curchunk_len - 1;
+	}
+	return true;
+}
+
+static bool drm_dp_sideband_parse_link_address(struct drm_dp_sideband_msg_rx *raw,
+					       struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	int i;
+	memcpy(repmsg->u.link_addr.guid, &raw->msg[idx], 16);
+	idx += 16;
+	repmsg->u.link_addr.nports = raw->msg[idx] & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	for (i = 0; i < repmsg->u.link_addr.nports; i++) {
+		if (raw->msg[idx] & 0x80)
+			repmsg->u.link_addr.ports[i].input_port = 1;
+
+		repmsg->u.link_addr.ports[i].peer_device_type = (raw->msg[idx] >> 4) & 0x7;
+		repmsg->u.link_addr.ports[i].port_number = (raw->msg[idx] & 0xf);
+
+		idx++;
+		if (idx > raw->curlen)
+			goto fail_len;
+		repmsg->u.link_addr.ports[i].mcs = (raw->msg[idx] >> 7) & 0x1;
+		repmsg->u.link_addr.ports[i].ddps = (raw->msg[idx] >> 6) & 0x1;
+		if (repmsg->u.link_addr.ports[i].input_port == 0)
+			repmsg->u.link_addr.ports[i].legacy_device_plug_status = (raw->msg[idx] >> 5) & 0x1;
+		idx++;
+		if (idx > raw->curlen)
+			goto fail_len;
+		if (repmsg->u.link_addr.ports[i].input_port == 0) {
+			repmsg->u.link_addr.ports[i].dpcd_revision = (raw->msg[idx]);
+			idx++;
+			if (idx > raw->curlen)
+				goto fail_len;
+			memcpy(repmsg->u.link_addr.ports[i].peer_guid, &raw->msg[idx], 16);
+			idx += 16;
+			if (idx > raw->curlen)
+				goto fail_len;
+			repmsg->u.link_addr.ports[i].num_sdp_streams = (raw->msg[idx] >> 4) & 0xf;
+			repmsg->u.link_addr.ports[i].num_sdp_stream_sinks = (raw->msg[idx] & 0xf);
+			idx++;
+
+		}
+		if (idx > raw->curlen)
+			goto fail_len;
+	}
+
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("link address reply parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_remote_dpcd_read(struct drm_dp_sideband_msg_rx *raw,
+						   struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	repmsg->u.remote_dpcd_read_ack.port_number = raw->msg[idx] & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.remote_dpcd_read_ack.num_bytes = raw->msg[idx];
+	if (idx > raw->curlen)
+		goto fail_len;
+
+	memcpy(repmsg->u.remote_dpcd_read_ack.bytes, &raw->msg[idx], repmsg->u.remote_dpcd_read_ack.num_bytes);
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("link address reply parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_remote_dpcd_write(struct drm_dp_sideband_msg_rx *raw,
+						      struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	repmsg->u.remote_dpcd_write_ack.port_number = raw->msg[idx] & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_remote_i2c_read_ack(struct drm_dp_sideband_msg_rx *raw,
+						      struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+
+	repmsg->u.remote_i2c_read_ack.port_number = (raw->msg[idx] & 0xf);
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.remote_i2c_read_ack.num_bytes = raw->msg[idx];
+	idx++;
+	/* TODO check */
+	memcpy(repmsg->u.remote_i2c_read_ack.bytes, &raw->msg[idx], repmsg->u.remote_i2c_read_ack.num_bytes);
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("remote i2c reply parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_enum_path_resources_ack(struct drm_dp_sideband_msg_rx *raw,
+							  struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	repmsg->u.path_resources.port_number = (raw->msg[idx] >> 4) & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.path_resources.full_payload_bw_number = (raw->msg[idx] << 8) | (raw->msg[idx+1]);
+	idx += 2;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.path_resources.avail_payload_bw_number = (raw->msg[idx] << 8) | (raw->msg[idx+1]);
+	idx += 2;
+	if (idx > raw->curlen)
+		goto fail_len;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("enum resource parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_allocate_payload_ack(struct drm_dp_sideband_msg_rx *raw,
+							  struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	repmsg->u.allocate_payload.port_number = (raw->msg[idx] >> 4) & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.allocate_payload.vcpi = raw->msg[idx];
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.allocate_payload.allocated_pbn = (raw->msg[idx] << 8) | (raw->msg[idx+1]);
+	idx += 2;
+	if (idx > raw->curlen)
+		goto fail_len;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("allocate payload parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_query_payload_ack(struct drm_dp_sideband_msg_rx *raw,
+						    struct drm_dp_sideband_msg_reply_body *repmsg)
+{
+	int idx = 1;
+	repmsg->u.query_payload.port_number = (raw->msg[idx] >> 4) & 0xf;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+	repmsg->u.query_payload.allocated_pbn = (raw->msg[idx] << 8) | (raw->msg[idx + 1]);
+	idx += 2;
+	if (idx > raw->curlen)
+		goto fail_len;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("query payload parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_reply(struct drm_dp_sideband_msg_rx *raw,
+					struct drm_dp_sideband_msg_reply_body *msg)
+{
+	memset(msg, 0, sizeof(*msg));
+	msg->reply_type = (raw->msg[0] & 0x80) >> 7;
+	msg->req_type = (raw->msg[0] & 0x7f);
+
+	if (msg->reply_type) {
+		memcpy(msg->u.nak.guid, &raw->msg[1], 16);
+		msg->u.nak.reason = raw->msg[17];
+		msg->u.nak.nak_data = raw->msg[18];
+		return false;
+	}
+
+	switch (msg->req_type) {
+	case DP_LINK_ADDRESS:
+		return drm_dp_sideband_parse_link_address(raw, msg);
+	case DP_QUERY_PAYLOAD:
+		return drm_dp_sideband_parse_query_payload_ack(raw, msg);
+	case DP_REMOTE_DPCD_READ:
+		return drm_dp_sideband_parse_remote_dpcd_read(raw, msg);
+	case DP_REMOTE_DPCD_WRITE:
+		return drm_dp_sideband_parse_remote_dpcd_write(raw, msg);
+	case DP_REMOTE_I2C_READ:
+		return drm_dp_sideband_parse_remote_i2c_read_ack(raw, msg);
+	case DP_ENUM_PATH_RESOURCES:
+		return drm_dp_sideband_parse_enum_path_resources_ack(raw, msg);
+	case DP_ALLOCATE_PAYLOAD:
+		return drm_dp_sideband_parse_allocate_payload_ack(raw, msg);
+	default:
+		DRM_ERROR("Got unknown reply 0x%02x\n", msg->req_type);
+		return false;
+	}
+}
+
+static bool drm_dp_sideband_parse_connection_status_notify(struct drm_dp_sideband_msg_rx *raw,
+							   struct drm_dp_sideband_msg_req_body *msg)
+{
+	int idx = 1;
+
+	msg->u.conn_stat.port_number = (raw->msg[idx] & 0xf0) >> 4;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+
+	memcpy(msg->u.conn_stat.guid, &raw->msg[idx], 16);
+	idx += 16;
+	if (idx > raw->curlen)
+		goto fail_len;
+
+	msg->u.conn_stat.legacy_device_plug_status = (raw->msg[idx] >> 6) & 0x1;
+	msg->u.conn_stat.displayport_device_plug_status = (raw->msg[idx] >> 5) & 0x1;
+	msg->u.conn_stat.message_capability_status = (raw->msg[idx] >> 4) & 0x1;
+	msg->u.conn_stat.input_port = (raw->msg[idx] >> 3) & 0x1;
+	msg->u.conn_stat.peer_device_type = (raw->msg[idx] & 0x7);
+	idx++;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("connection status reply parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_resource_status_notify(struct drm_dp_sideband_msg_rx *raw,
+							   struct drm_dp_sideband_msg_req_body *msg)
+{
+	int idx = 1;
+
+	msg->u.resource_stat.port_number = (raw->msg[idx] & 0xf0) >> 4;
+	idx++;
+	if (idx > raw->curlen)
+		goto fail_len;
+
+	memcpy(msg->u.resource_stat.guid, &raw->msg[idx], 16);
+	idx += 16;
+	if (idx > raw->curlen)
+		goto fail_len;
+
+	msg->u.resource_stat.available_pbn = (raw->msg[idx] << 8) | (raw->msg[idx + 1]);
+	idx++;
+	return true;
+fail_len:
+	DRM_DEBUG_KMS("resource status reply parse length fail %d %d\n", idx, raw->curlen);
+	return false;
+}
+
+static bool drm_dp_sideband_parse_req(struct drm_dp_sideband_msg_rx *raw,
+				      struct drm_dp_sideband_msg_req_body *msg)
+{
+	memset(msg, 0, sizeof(*msg));
+	msg->req_type = (raw->msg[0] & 0x7f);
+
+	switch (msg->req_type) {
+	case DP_CONNECTION_STATUS_NOTIFY:
+		return drm_dp_sideband_parse_connection_status_notify(raw, msg);
+	case DP_RESOURCE_STATUS_NOTIFY:
+		return drm_dp_sideband_parse_resource_status_notify(raw, msg);
+	default:
+		DRM_ERROR("Got unknown request 0x%02x\n", msg->req_type);
+		return false;
+	}
+}
+
+static int build_dpcd_write(struct drm_dp_sideband_msg_tx *msg, u8 port_num, u32 offset, u8 num_bytes, u8 *bytes)
+{
+	struct drm_dp_sideband_msg_req_body req;
+
+	req.req_type = DP_REMOTE_DPCD_WRITE;
+	req.u.dpcd_write.port_number = port_num;
+	req.u.dpcd_write.dpcd_address = offset;
+	req.u.dpcd_write.num_bytes = num_bytes;
+	req.u.dpcd_write.bytes = bytes;
+	drm_dp_encode_sideband_req(&req, msg);
+
+	return 0;
+}
+
+static int build_link_address(struct drm_dp_sideband_msg_tx *msg)
+{
+	struct drm_dp_sideband_msg_req_body req;
+
+	req.req_type = DP_LINK_ADDRESS;
+	drm_dp_encode_sideband_req(&req, msg);
+	return 0;
+}
+
+static int build_enum_path_resources(struct drm_dp_sideband_msg_tx *msg, int port_num)
+{
+	struct drm_dp_sideband_msg_req_body req;
+
+	req.req_type = DP_ENUM_PATH_RESOURCES;
+	req.u.port_num.port_number = port_num;
+	drm_dp_encode_sideband_req(&req, msg);
+	msg->path_msg = true;
+	return 0;
+}
+
+static int build_allocate_payload(struct drm_dp_sideband_msg_tx *msg, int port_num,
+				  u8 vcpi, uint16_t pbn)
+{
+	struct drm_dp_sideband_msg_req_body req;
+	memset(&req, 0, sizeof(req));
+	req.req_type = DP_ALLOCATE_PAYLOAD;
+	req.u.allocate_payload.port_number = port_num;
+	req.u.allocate_payload.vcpi = vcpi;
+	req.u.allocate_payload.pbn = pbn;
+	drm_dp_encode_sideband_req(&req, msg);
+	msg->path_msg = true;
+	return 0;
+}
+
+static int drm_dp_mst_assign_payload_id(struct drm_dp_mst_topology_mgr *mgr,
+					struct drm_dp_vcpi *vcpi)
+{
+	int ret, vcpi_ret;
+
+	mutex_lock(&mgr->payload_lock);
+	ret = find_first_zero_bit(&mgr->payload_mask, mgr->max_payloads + 1);
+	if (ret > mgr->max_payloads) {
+		ret = -EINVAL;
+		DRM_DEBUG_KMS("out of payload ids %d\n", ret);
+		goto out_unlock;
+	}
+
+	vcpi_ret = find_first_zero_bit(&mgr->vcpi_mask, mgr->max_payloads + 1);
+	if (vcpi_ret > mgr->max_payloads) {
+		ret = -EINVAL;
+		DRM_DEBUG_KMS("out of vcpi ids %d\n", ret);
+		goto out_unlock;
+	}
+
+	set_bit(ret, &mgr->payload_mask);
+	set_bit(vcpi_ret, &mgr->vcpi_mask);
+	vcpi->vcpi = vcpi_ret + 1;
+	mgr->proposed_vcpis[ret - 1] = vcpi;
+out_unlock:
+	mutex_unlock(&mgr->payload_lock);
+	return ret;
+}
+
+static void drm_dp_mst_put_payload_id(struct drm_dp_mst_topology_mgr *mgr,
+				      int vcpi)
+{
+	int i;
+	if (vcpi == 0)
+		return;
+
+	mutex_lock(&mgr->payload_lock);
+	DRM_DEBUG_KMS("putting payload %d\n", vcpi);
+	clear_bit(vcpi - 1, &mgr->vcpi_mask);
+
+	for (i = 0; i < mgr->max_payloads; i++) {
+		if (mgr->proposed_vcpis[i])
+			if (mgr->proposed_vcpis[i]->vcpi == vcpi) {
+				mgr->proposed_vcpis[i] = NULL;
+				clear_bit(i + 1, &mgr->payload_mask);
+			}
+	}
+	mutex_unlock(&mgr->payload_lock);
+}
+
+static bool check_txmsg_state(struct drm_dp_mst_topology_mgr *mgr,
+			      struct drm_dp_sideband_msg_tx *txmsg)
+{
+	bool ret;
+	mutex_lock(&mgr->qlock);
+	ret = (txmsg->state == DRM_DP_SIDEBAND_TX_RX ||
+	       txmsg->state == DRM_DP_SIDEBAND_TX_TIMEOUT);
+	mutex_unlock(&mgr->qlock);
+	return ret;
+}
+
+static int drm_dp_mst_wait_tx_reply(struct drm_dp_mst_branch *mstb,
+				    struct drm_dp_sideband_msg_tx *txmsg)
+{
+	struct drm_dp_mst_topology_mgr *mgr = mstb->mgr;
+	int ret;
+
+	ret = wait_event_timeout(mgr->tx_waitq,
+				 check_txmsg_state(mgr, txmsg),
+				 (4 * HZ));
+	mutex_lock(&mstb->mgr->qlock);
+	if (ret > 0) {
+		if (txmsg->state == DRM_DP_SIDEBAND_TX_TIMEOUT) {
+			ret = -EIO;
+			goto out;
+		}
+	} else {
+		DRM_DEBUG_KMS("timedout msg send %p %d %d\n", txmsg, txmsg->state, txmsg->seqno);
+
+		/* dump some state */
+		ret = -EIO;
+
+		/* remove from q */
+		if (txmsg->state == DRM_DP_SIDEBAND_TX_QUEUED ||
+		    txmsg->state == DRM_DP_SIDEBAND_TX_START_SEND) {
+			list_del(&txmsg->next);
+		}
+
+		if (txmsg->state == DRM_DP_SIDEBAND_TX_START_SEND ||
+		    txmsg->state == DRM_DP_SIDEBAND_TX_SENT) {
+			mstb->tx_slots[txmsg->seqno] = NULL;
+		}
+	}
+out:
+	mutex_unlock(&mgr->qlock);
+
+	return ret;
+}
+
+static struct drm_dp_mst_branch *drm_dp_add_mst_branch_device(u8 lct, u8 *rad)
+{
+	struct drm_dp_mst_branch *mstb;
+
+	mstb = kzalloc(sizeof(*mstb), GFP_KERNEL);
+	if (!mstb)
+		return NULL;
+
+	mstb->lct = lct;
+	if (lct > 1)
+		memcpy(mstb->rad, rad, lct / 2);
+	INIT_LIST_HEAD(&mstb->ports);
+	kref_init(&mstb->kref);
+	return mstb;
+}
+
+static void drm_dp_destroy_mst_branch_device(struct kref *kref)
+{
+	struct drm_dp_mst_branch *mstb = container_of(kref, struct drm_dp_mst_branch, kref);
+	struct drm_dp_mst_port *port, *tmp;
+	bool wake_tx = false;
+
+	cancel_work_sync(&mstb->mgr->work);
+
+	/*
+	 * destroy all ports - don't need lock
+	 * as there are no more references to the mst branch
+	 * device at this point.
+	 */
+	list_for_each_entry_safe(port, tmp, &mstb->ports, next) {
+		list_del(&port->next);
+		drm_dp_put_port(port);
+	}
+
+	/* drop any tx slots msg */
+	mutex_lock(&mstb->mgr->qlock);
+	if (mstb->tx_slots[0]) {
+		mstb->tx_slots[0]->state = DRM_DP_SIDEBAND_TX_TIMEOUT;
+		mstb->tx_slots[0] = NULL;
+		wake_tx = true;
+	}
+	if (mstb->tx_slots[1]) {
+		mstb->tx_slots[1]->state = DRM_DP_SIDEBAND_TX_TIMEOUT;
+		mstb->tx_slots[1] = NULL;
+		wake_tx = true;
+	}
+	mutex_unlock(&mstb->mgr->qlock);
+
+	if (wake_tx)
+		wake_up(&mstb->mgr->tx_waitq);
+	kfree(mstb);
+}
+
+static void drm_dp_put_mst_branch_device(struct drm_dp_mst_branch *mstb)
+{
+	kref_put(&mstb->kref, drm_dp_destroy_mst_branch_device);
+}
+
+
+static void drm_dp_port_teardown_pdt(struct drm_dp_mst_port *port, int old_pdt)
+{
+	switch (old_pdt) {
+	case DP_PEER_DEVICE_DP_LEGACY_CONV:
+	case DP_PEER_DEVICE_SST_SINK:
+		/* remove i2c over sideband */
+		drm_dp_mst_unregister_i2c_bus(&port->aux);
+		break;
+	case DP_PEER_DEVICE_MST_BRANCHING:
+		drm_dp_put_mst_branch_device(port->mstb);
+		port->mstb = NULL;
+		break;
+	}
+}
+
+static void drm_dp_destroy_port(struct kref *kref)
+{
+	struct drm_dp_mst_port *port = container_of(kref, struct drm_dp_mst_port, kref);
+	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
+	if (!port->input) {
+		port->vcpi.num_slots = 0;
+		if (port->connector)
+			(*port->mgr->cbs->destroy_connector)(mgr, port->connector);
+		drm_dp_port_teardown_pdt(port, port->pdt);
+
+		if (!port->input && port->vcpi.vcpi > 0)
+			drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi);
+	}
+	kfree(port);
+
+	(*mgr->cbs->hotplug)(mgr);
+}
+
+static void drm_dp_put_port(struct drm_dp_mst_port *port)
+{
+	kref_put(&port->kref, drm_dp_destroy_port);
+}
+
+static struct drm_dp_mst_branch *drm_dp_mst_get_validated_mstb_ref_locked(struct drm_dp_mst_branch *mstb, struct drm_dp_mst_branch *to_find)
+{
+	struct drm_dp_mst_port *port;
+	struct drm_dp_mst_branch *rmstb;
+	if (to_find == mstb) {
+		kref_get(&mstb->kref);
+		return mstb;
+	}
+	list_for_each_entry(port, &mstb->ports, next) {
+		if (port->mstb) {
+			rmstb = drm_dp_mst_get_validated_mstb_ref_locked(port->mstb, to_find);
+			if (rmstb)
+				return rmstb;
+		}
+	}
+	return NULL;
+}
+
+static struct drm_dp_mst_branch *drm_dp_get_validated_mstb_ref(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_branch *mstb)
+{
+	struct drm_dp_mst_branch *rmstb = NULL;
+	mutex_lock(&mgr->lock);
+	if (mgr->mst_primary)
+		rmstb = drm_dp_mst_get_validated_mstb_ref_locked(mgr->mst_primary, mstb);
+	mutex_unlock(&mgr->lock);
+	return rmstb;
+}
+
+static struct drm_dp_mst_port *drm_dp_mst_get_port_ref_locked(struct drm_dp_mst_branch *mstb, struct drm_dp_mst_port *to_find)
+{
+	struct drm_dp_mst_port *port, *mport;
+
+	list_for_each_entry(port, &mstb->ports, next) {
+		if (port == to_find) {
+			kref_get(&port->kref);
+			return port;
+		}
+		if (port->mstb) {
+			mport = drm_dp_mst_get_port_ref_locked(port->mstb, to_find);
+			if (mport)
+				return mport;
+		}
+	}
+	return NULL;
+}
+
+static struct drm_dp_mst_port *drm_dp_get_validated_port_ref(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+{
+	struct drm_dp_mst_port *rport = NULL;
+	mutex_lock(&mgr->lock);
+	if (mgr->mst_primary)
+		rport = drm_dp_mst_get_port_ref_locked(mgr->mst_primary, port);
+	mutex_unlock(&mgr->lock);
+	return rport;
+}
+
+static struct drm_dp_mst_port *drm_dp_get_port(struct drm_dp_mst_branch *mstb, u8 port_num)
+{
+	struct drm_dp_mst_port *port;
+
+	list_for_each_entry(port, &mstb->ports, next) {
+		if (port->port_num == port_num) {
+			kref_get(&port->kref);
+			return port;
+		}
+	}
+
+	return NULL;
+}
+
+/*
+ * calculate a new RAD for this MST branch device
+ * if parent has an LCT of 2 then it has 1 nibble of RAD,
+ * if parent has an LCT of 3 then it has 2 nibbles of RAD,
+ */
+static u8 drm_dp_calculate_rad(struct drm_dp_mst_port *port,
+				 u8 *rad)
+{
+	int lct = port->parent->lct;
+	int shift = 4;
+	int idx = lct / 2;
+	if (lct > 1) {
+		memcpy(rad, port->parent->rad, idx);
+		shift = (lct % 2) ? 4 : 0;
+	} else
+		rad[0] = 0;
+
+	rad[idx] |= port->port_num << shift;
+	return lct + 1;
+}
+
+/*
+ * return sends link address for new mstb
+ */
+static bool drm_dp_port_setup_pdt(struct drm_dp_mst_port *port)
+{
+	int ret;
+	u8 rad[6], lct;
+	bool send_link = false;
+	switch (port->pdt) {
+	case DP_PEER_DEVICE_DP_LEGACY_CONV:
+	case DP_PEER_DEVICE_SST_SINK:
+		/* add i2c over sideband */
+		ret = drm_dp_mst_register_i2c_bus(&port->aux);
+		break;
+	case DP_PEER_DEVICE_MST_BRANCHING:
+		lct = drm_dp_calculate_rad(port, rad);
+
+		port->mstb = drm_dp_add_mst_branch_device(lct, rad);
+		port->mstb->mgr = port->mgr;
+		port->mstb->port_parent = port;
+
+		send_link = true;
+		break;
+	}
+	return send_link;
+}
+
+static void drm_dp_check_port_guid(struct drm_dp_mst_branch *mstb,
+				   struct drm_dp_mst_port *port)
+{
+	int ret;
+	if (port->dpcd_rev >= 0x12) {
+		port->guid_valid = drm_dp_validate_guid(mstb->mgr, port->guid);
+		if (!port->guid_valid) {
+			ret = drm_dp_send_dpcd_write(mstb->mgr,
+						     port,
+						     DP_GUID,
+						     16, port->guid);
+			port->guid_valid = true;
+		}
+	}
+}
+
+static void build_mst_prop_path(struct drm_dp_mst_port *port,
+				struct drm_dp_mst_branch *mstb,
+				char *proppath)
+{
+	int i;
+	char temp[8];
+	snprintf(proppath, 255, "mst:%d", mstb->mgr->conn_base_id);
+	for (i = 0; i < (mstb->lct - 1); i++) {
+		int shift = (i % 2) ? 0 : 4;
+		int port_num = mstb->rad[i / 2] >> shift;
+		snprintf(temp, 8, "-%d", port_num);
+		strncat(proppath, temp, 255);
+	}
+	snprintf(temp, 8, "-%d", port->port_num);
+	strncat(proppath, temp, 255);
+}
+
+static void drm_dp_add_port(struct drm_dp_mst_branch *mstb,
+			    struct device *dev,
+			    struct drm_dp_link_addr_reply_port *port_msg)
+{
+	struct drm_dp_mst_port *port;
+	bool ret;
+	bool created = false;
+	int old_pdt = 0;
+	int old_ddps = 0;
+	port = drm_dp_get_port(mstb, port_msg->port_number);
+	if (!port) {
+		port = kzalloc(sizeof(*port), GFP_KERNEL);
+		if (!port)
+			return;
+		kref_init(&port->kref);
+		port->parent = mstb;
+		port->port_num = port_msg->port_number;
+		port->mgr = mstb->mgr;
+		port->aux.name = "DPMST";
+		port->aux.dev = dev;
+		created = true;
+	} else {
+		old_pdt = port->pdt;
+		old_ddps = port->ddps;
+	}
+
+	port->pdt = port_msg->peer_device_type;
+	port->input = port_msg->input_port;
+	port->mcs = port_msg->mcs;
+	port->ddps = port_msg->ddps;
+	port->ldps = port_msg->legacy_device_plug_status;
+	port->dpcd_rev = port_msg->dpcd_revision;
+	port->num_sdp_streams = port_msg->num_sdp_streams;
+	port->num_sdp_stream_sinks = port_msg->num_sdp_stream_sinks;
+	memcpy(port->guid, port_msg->peer_guid, 16);
+
+	/* manage mstb port lists with mgr lock - take a reference
+	   for this list */
+	if (created) {
+		mutex_lock(&mstb->mgr->lock);
+		kref_get(&port->kref);
+		list_add(&port->next, &mstb->ports);
+		mutex_unlock(&mstb->mgr->lock);
+	}
+
+	if (old_ddps != port->ddps) {
+		if (port->ddps) {
+			drm_dp_check_port_guid(mstb, port);
+			if (!port->input)
+				drm_dp_send_enum_path_resources(mstb->mgr, mstb, port);
+		} else {
+			port->guid_valid = false;
+			port->available_pbn = 0;
+			}
+	}
+
+	if (old_pdt != port->pdt && !port->input) {
+		drm_dp_port_teardown_pdt(port, old_pdt);
+
+		ret = drm_dp_port_setup_pdt(port);
+		if (ret == true) {
+			drm_dp_send_link_address(mstb->mgr, port->mstb);
+			port->mstb->link_address_sent = true;
+		}
+	}
+
+	if (created && !port->input) {
+		char proppath[255];
+		build_mst_prop_path(port, mstb, proppath);
+		port->connector = (*mstb->mgr->cbs->add_connector)(mstb->mgr, port, proppath);
+	}
+
+	/* put reference to this port */
+	drm_dp_put_port(port);
+}
+
+static void drm_dp_update_port(struct drm_dp_mst_branch *mstb,
+			       struct drm_dp_connection_status_notify *conn_stat)
+{
+	struct drm_dp_mst_port *port;
+	int old_pdt;
+	int old_ddps;
+	bool dowork = false;
+	port = drm_dp_get_port(mstb, conn_stat->port_number);
+	if (!port)
+		return;
+
+	old_ddps = port->ddps;
+	old_pdt = port->pdt;
+	port->pdt = conn_stat->peer_device_type;
+	port->mcs = conn_stat->message_capability_status;
+	port->ldps = conn_stat->legacy_device_plug_status;
+	port->ddps = conn_stat->displayport_device_plug_status;
+
+	if (old_ddps != port->ddps) {
+		if (port->ddps) {
+			drm_dp_check_port_guid(mstb, port);
+			dowork = true;
+		} else {
+			port->guid_valid = false;
+			port->available_pbn = 0;
+		}
+	}
+	if (old_pdt != port->pdt && !port->input) {
+		drm_dp_port_teardown_pdt(port, old_pdt);
+
+		if (drm_dp_port_setup_pdt(port))
+			dowork = true;
+	}
+
+	drm_dp_put_port(port);
+	if (dowork)
+		queue_work(system_long_wq, &mstb->mgr->work);
+
+}
+
+static struct drm_dp_mst_branch *drm_dp_get_mst_branch_device(struct drm_dp_mst_topology_mgr *mgr,
+							       u8 lct, u8 *rad)
+{
+	struct drm_dp_mst_branch *mstb;
+	struct drm_dp_mst_port *port;
+	int i;
+	/* find the port by iterating down */
+	mstb = mgr->mst_primary;
+
+	for (i = 0; i < lct - 1; i++) {
+		int shift = (i % 2) ? 0 : 4;
+		int port_num = rad[i / 2] >> shift;
+
+		list_for_each_entry(port, &mstb->ports, next) {
+			if (port->port_num == port_num) {
+				if (!port->mstb) {
+					DRM_ERROR("failed to lookup MSTB with lct %d, rad %02x\n", lct, rad[0]);
+					return NULL;
+				}
+
+				mstb = port->mstb;
+				break;
+			}
+		}
+	}
+	kref_get(&mstb->kref);
+	return mstb;
+}
+
+static void drm_dp_check_and_send_link_address(struct drm_dp_mst_topology_mgr *mgr,
+					       struct drm_dp_mst_branch *mstb)
+{
+	struct drm_dp_mst_port *port;
+
+	if (!mstb->link_address_sent) {
+		drm_dp_send_link_address(mgr, mstb);
+		mstb->link_address_sent = true;
+	}
+	list_for_each_entry(port, &mstb->ports, next) {
+		if (port->input)
+			continue;
+
+		if (!port->ddps)
+			continue;
+
+		if (!port->available_pbn)
+			drm_dp_send_enum_path_resources(mgr, mstb, port);
+
+		if (port->mstb)
+			drm_dp_check_and_send_link_address(mgr, port->mstb);
+	}
+}
+
+static void drm_dp_mst_link_probe_work(struct work_struct *work)
+{
+	struct drm_dp_mst_topology_mgr *mgr = container_of(work, struct drm_dp_mst_topology_mgr, work);
+
+	drm_dp_check_and_send_link_address(mgr, mgr->mst_primary);
+
+}
+
+static bool drm_dp_validate_guid(struct drm_dp_mst_topology_mgr *mgr,
+				 u8 *guid)
+{
+	static u8 zero_guid[16];
+
+	if (!memcmp(guid, zero_guid, 16)) {
+		u64 salt = get_jiffies_64();
+		memcpy(&guid[0], &salt, sizeof(u64));
+		memcpy(&guid[8], &salt, sizeof(u64));
+		return false;
+	}
+	return true;
+}
+
+#if 0
+static int build_dpcd_read(struct drm_dp_sideband_msg_tx *msg, u8 port_num, u32 offset, u8 num_bytes)
+{
+	struct drm_dp_sideband_msg_req_body req;
+
+	req.req_type = DP_REMOTE_DPCD_READ;
+	req.u.dpcd_read.port_number = port_num;
+	req.u.dpcd_read.dpcd_address = offset;
+	req.u.dpcd_read.num_bytes = num_bytes;
+	drm_dp_encode_sideband_req(&req, msg);
+
+	return 0;
+}
+#endif
+
+static int drm_dp_send_sideband_msg(struct drm_dp_mst_topology_mgr *mgr,
+				    bool up, u8 *msg, int len)
+{
+	int ret;
+	int regbase = up ? DP_SIDEBAND_MSG_UP_REP_BASE : DP_SIDEBAND_MSG_DOWN_REQ_BASE;
+	int tosend, total, offset;
+	int retries = 0;
+
+retry:
+	total = len;
+	offset = 0;
+	do {
+		tosend = min3(mgr->max_dpcd_transaction_bytes, 16, total);
+
+		ret = drm_dp_dpcd_write(mgr->aux, regbase + offset,
+					&msg[offset],
+					tosend);
+		if (ret != tosend) {
+			if (ret == -EIO && retries < 5) {
+				retries++;
+				goto retry;
+			}
+			DRM_DEBUG_KMS("failed to dpcd write %d %d\n", tosend, ret);
+			WARN(1, "fail\n");
+
+			return -EIO;
+		}
+		offset += tosend;
+		total -= tosend;
+	} while (total > 0);
+	return 0;
+}
+
+static int set_hdr_from_dst_qlock(struct drm_dp_sideband_msg_hdr *hdr,
+				  struct drm_dp_sideband_msg_tx *txmsg)
+{
+	struct drm_dp_mst_branch *mstb = txmsg->dst;
+
+	/* both msg slots are full */
+	if (txmsg->seqno == -1) {
+		if (mstb->tx_slots[0] && mstb->tx_slots[1]) {
+			DRM_DEBUG_KMS("%s: failed to find slot\n", __func__);
+			return -EAGAIN;
+		}
+		if (mstb->tx_slots[0] == NULL && mstb->tx_slots[1] == NULL) {
+			txmsg->seqno = mstb->last_seqno;
+			mstb->last_seqno ^= 1;
+		} else if (mstb->tx_slots[0] == NULL)
+			txmsg->seqno = 0;
+		else
+			txmsg->seqno = 1;
+		mstb->tx_slots[txmsg->seqno] = txmsg;
+	}
+	hdr->broadcast = 0;
+	hdr->path_msg = txmsg->path_msg;
+	hdr->lct = mstb->lct;
+	hdr->lcr = mstb->lct - 1;
+	if (mstb->lct > 1)
+		memcpy(hdr->rad, mstb->rad, mstb->lct / 2);
+	hdr->seqno = txmsg->seqno;
+	return 0;
+}
+/*
+ * process a single block of the next message in the sideband queue
+ */
+static int process_single_tx_qlock(struct drm_dp_mst_topology_mgr *mgr,
+				   struct drm_dp_sideband_msg_tx *txmsg,
+				   bool up)
+{
+	u8 chunk[48];
+	struct drm_dp_sideband_msg_hdr hdr;
+	int len, space, idx, tosend;
+	int ret;
+
+	memset(&hdr, 0, sizeof(struct drm_dp_sideband_msg_hdr));
+
+	if (txmsg->state == DRM_DP_SIDEBAND_TX_QUEUED) {
+		txmsg->seqno = -1;
+		txmsg->state = DRM_DP_SIDEBAND_TX_START_SEND;
+	}
+
+	/* make hdr from dst mst - for replies use seqno
+	   otherwise assign one */
+	ret = set_hdr_from_dst_qlock(&hdr, txmsg);
+	if (ret < 0)
+		return ret;
+
+	/* amount left to send in this message */
+	len = txmsg->cur_len - txmsg->cur_offset;
+
+	/* 48 - sideband msg size - 1 byte for data CRC, x header bytes */
+	space = 48 - 1 - drm_dp_calc_sb_hdr_size(&hdr);
+
+	tosend = min(len, space);
+	if (len == txmsg->cur_len)
+		hdr.somt = 1;
+	if (space >= len)
+		hdr.eomt = 1;
+
+
+	hdr.msg_len = tosend + 1;
+	drm_dp_encode_sideband_msg_hdr(&hdr, chunk, &idx);
+	memcpy(&chunk[idx], &txmsg->msg[txmsg->cur_offset], tosend);
+	/* add crc at end */
+	drm_dp_crc_sideband_chunk_req(&chunk[idx], tosend);
+	idx += tosend + 1;
+
+	ret = drm_dp_send_sideband_msg(mgr, up, chunk, idx);
+	if (ret) {
+		DRM_DEBUG_KMS("sideband msg failed to send\n");
+		return ret;
+	}
+
+	txmsg->cur_offset += tosend;
+	if (txmsg->cur_offset == txmsg->cur_len) {
+		txmsg->state = DRM_DP_SIDEBAND_TX_SENT;
+		return 1;
+	}
+	return 0;
+}
+
+/* must be called holding qlock */
+static void process_single_down_tx_qlock(struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct drm_dp_sideband_msg_tx *txmsg;
+	int ret;
+
+	/* construct a chunk from the first msg in the tx_msg queue */
+	if (list_empty(&mgr->tx_msg_downq)) {
+		mgr->tx_down_in_progress = false;
+		return;
+	}
+	mgr->tx_down_in_progress = true;
+
+	txmsg = list_first_entry(&mgr->tx_msg_downq, struct drm_dp_sideband_msg_tx, next);
+	ret = process_single_tx_qlock(mgr, txmsg, false);
+	if (ret == 1) {
+		/* txmsg is sent it should be in the slots now */
+		list_del(&txmsg->next);
+	} else if (ret) {
+		DRM_DEBUG_KMS("failed to send msg in q %d\n", ret);
+		list_del(&txmsg->next);
+		if (txmsg->seqno != -1)
+			txmsg->dst->tx_slots[txmsg->seqno] = NULL;
+		txmsg->state = DRM_DP_SIDEBAND_TX_TIMEOUT;
+		wake_up(&mgr->tx_waitq);
+	}
+	if (list_empty(&mgr->tx_msg_downq)) {
+		mgr->tx_down_in_progress = false;
+		return;
+	}
+}
+
+/* called holding qlock */
+static void process_single_up_tx_qlock(struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct drm_dp_sideband_msg_tx *txmsg;
+	int ret;
+
+	/* construct a chunk from the first msg in the tx_msg queue */
+	if (list_empty(&mgr->tx_msg_upq)) {
+		mgr->tx_up_in_progress = false;
+		return;
+	}
+
+	txmsg = list_first_entry(&mgr->tx_msg_upq, struct drm_dp_sideband_msg_tx, next);
+	ret = process_single_tx_qlock(mgr, txmsg, true);
+	if (ret == 1) {
+		/* up txmsgs aren't put in slots - so free after we send it */
+		list_del(&txmsg->next);
+		kfree(txmsg);
+	} else if (ret)
+		DRM_DEBUG_KMS("failed to send msg in q %d\n", ret);
+	mgr->tx_up_in_progress = true;
+}
+
+static void drm_dp_queue_down_tx(struct drm_dp_mst_topology_mgr *mgr,
+				 struct drm_dp_sideband_msg_tx *txmsg)
+{
+	mutex_lock(&mgr->qlock);
+	list_add_tail(&txmsg->next, &mgr->tx_msg_downq);
+	if (!mgr->tx_down_in_progress)
+		process_single_down_tx_qlock(mgr);
+	mutex_unlock(&mgr->qlock);
+}
+
+static int drm_dp_send_link_address(struct drm_dp_mst_topology_mgr *mgr,
+				    struct drm_dp_mst_branch *mstb)
+{
+	int len;
+	struct drm_dp_sideband_msg_tx *txmsg;
+	int ret;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg)
+		return -ENOMEM;
+
+	txmsg->dst = mstb;
+	len = build_link_address(txmsg);
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+	if (ret > 0) {
+		int i;
+
+		if (txmsg->reply.reply_type == 1)
+			DRM_DEBUG_KMS("link address nak received\n");
+		else {
+			DRM_DEBUG_KMS("link address reply: %d\n", txmsg->reply.u.link_addr.nports);
+			for (i = 0; i < txmsg->reply.u.link_addr.nports; i++) {
+				DRM_DEBUG_KMS("port %d: input %d, pdt: %d, pn: %d, dpcd_rev: %02x, mcs: %d, ddps: %d, ldps %d, sdp %d/%d\n", i,
+				       txmsg->reply.u.link_addr.ports[i].input_port,
+				       txmsg->reply.u.link_addr.ports[i].peer_device_type,
+				       txmsg->reply.u.link_addr.ports[i].port_number,
+				       txmsg->reply.u.link_addr.ports[i].dpcd_revision,
+				       txmsg->reply.u.link_addr.ports[i].mcs,
+				       txmsg->reply.u.link_addr.ports[i].ddps,
+				       txmsg->reply.u.link_addr.ports[i].legacy_device_plug_status,
+				       txmsg->reply.u.link_addr.ports[i].num_sdp_streams,
+				       txmsg->reply.u.link_addr.ports[i].num_sdp_stream_sinks);
+			}
+			for (i = 0; i < txmsg->reply.u.link_addr.nports; i++) {
+				drm_dp_add_port(mstb, mgr->dev, &txmsg->reply.u.link_addr.ports[i]);
+			}
+			(*mgr->cbs->hotplug)(mgr);
+		}
+	} else
+		DRM_DEBUG_KMS("link address failed %d\n", ret);
+
+	kfree(txmsg);
+	return 0;
+}
+
+static int drm_dp_send_enum_path_resources(struct drm_dp_mst_topology_mgr *mgr,
+					   struct drm_dp_mst_branch *mstb,
+					   struct drm_dp_mst_port *port)
+{
+	int len;
+	struct drm_dp_sideband_msg_tx *txmsg;
+	int ret;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg)
+		return -ENOMEM;
+
+	txmsg->dst = mstb;
+	len = build_enum_path_resources(txmsg, port->port_num);
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+	if (ret > 0) {
+		if (txmsg->reply.reply_type == 1)
+			DRM_DEBUG_KMS("enum path resources nak received\n");
+		else {
+			if (port->port_num != txmsg->reply.u.path_resources.port_number)
+				DRM_ERROR("got incorrect port in response\n");
+			DRM_DEBUG_KMS("enum path resources %d: %d %d\n", txmsg->reply.u.path_resources.port_number, txmsg->reply.u.path_resources.full_payload_bw_number,
+			       txmsg->reply.u.path_resources.avail_payload_bw_number);
+			port->available_pbn = txmsg->reply.u.path_resources.avail_payload_bw_number;
+		}
+	}
+
+	kfree(txmsg);
+	return 0;
+}
+
+static int drm_dp_payload_send_msg(struct drm_dp_mst_topology_mgr *mgr,
+				   struct drm_dp_mst_port *port,
+				   int id,
+				   int pbn)
+{
+	struct drm_dp_sideband_msg_tx *txmsg;
+	struct drm_dp_mst_branch *mstb;
+	int len, ret;
+
+	mstb = drm_dp_get_validated_mstb_ref(mgr, port->parent);
+	if (!mstb)
+		return -EINVAL;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg) {
+		ret = -ENOMEM;
+		goto fail_put;
+	}
+
+	txmsg->dst = mstb;
+	len = build_allocate_payload(txmsg, port->port_num,
+				     id,
+				     pbn);
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+	if (ret > 0) {
+		if (txmsg->reply.reply_type == 1) {
+			ret = -EINVAL;
+		} else
+			ret = 0;
+	}
+	kfree(txmsg);
+fail_put:
+	drm_dp_put_mst_branch_device(mstb);
+	return ret;
+}
+
+static int drm_dp_create_payload_step1(struct drm_dp_mst_topology_mgr *mgr,
+				       int id,
+				       struct drm_dp_payload *payload)
+{
+	int ret;
+
+	ret = drm_dp_dpcd_write_payload(mgr, id, payload);
+	if (ret < 0) {
+		payload->payload_state = 0;
+		return ret;
+	}
+	payload->payload_state = DP_PAYLOAD_LOCAL;
+	return 0;
+}
+
+static int drm_dp_create_payload_step2(struct drm_dp_mst_topology_mgr *mgr,
+				       struct drm_dp_mst_port *port,
+				       int id,
+				       struct drm_dp_payload *payload)
+{
+	int ret;
+	ret = drm_dp_payload_send_msg(mgr, port, id, port->vcpi.pbn);
+	if (ret < 0)
+		return ret;
+	payload->payload_state = DP_PAYLOAD_REMOTE;
+	return ret;
+}
+
+static int drm_dp_destroy_payload_step1(struct drm_dp_mst_topology_mgr *mgr,
+					struct drm_dp_mst_port *port,
+					int id,
+					struct drm_dp_payload *payload)
+{
+	DRM_DEBUG_KMS("\n");
+	/* its okay for these to fail */
+	if (port) {
+		drm_dp_payload_send_msg(mgr, port, id, 0);
+	}
+
+	drm_dp_dpcd_write_payload(mgr, id, payload);
+	payload->payload_state = DP_PAYLOAD_DELETE_LOCAL;
+	return 0;
+}
+
+static int drm_dp_destroy_payload_step2(struct drm_dp_mst_topology_mgr *mgr,
+					int id,
+					struct drm_dp_payload *payload)
+{
+	payload->payload_state = 0;
+	return 0;
+}
+
+/**
+ * drm_dp_update_payload_part1() - Execute payload update part 1
+ * @mgr: manager to use.
+ *
+ * This iterates over all proposed virtual channels, and tries to
+ * allocate space in the link for them. For 0->slots transitions,
+ * this step just writes the VCPI to the MST device. For slots->0
+ * transitions, this writes the updated VCPIs and removes the
+ * remote VC payloads.
+ *
+ * after calling this the driver should generate ACT and payload
+ * packets.
+ */
+int drm_dp_update_payload_part1(struct drm_dp_mst_topology_mgr *mgr)
+{
+	int i, j;
+	int cur_slots = 1;
+	struct drm_dp_payload req_payload;
+	struct drm_dp_mst_port *port;
+
+	mutex_lock(&mgr->payload_lock);
+	for (i = 0; i < mgr->max_payloads; i++) {
+		/* solve the current payloads - compare to the hw ones
+		   - update the hw view */
+		req_payload.start_slot = cur_slots;
+		if (mgr->proposed_vcpis[i]) {
+			port = container_of(mgr->proposed_vcpis[i], struct drm_dp_mst_port, vcpi);
+			req_payload.num_slots = mgr->proposed_vcpis[i]->num_slots;
+		} else {
+			port = NULL;
+			req_payload.num_slots = 0;
+		}
+
+		if (mgr->payloads[i].start_slot != req_payload.start_slot) {
+			mgr->payloads[i].start_slot = req_payload.start_slot;
+		}
+		/* work out what is required to happen with this payload */
+		if (mgr->payloads[i].num_slots != req_payload.num_slots) {
+
+			/* need to push an update for this payload */
+			if (req_payload.num_slots) {
+				drm_dp_create_payload_step1(mgr, mgr->proposed_vcpis[i]->vcpi, &req_payload);
+				mgr->payloads[i].num_slots = req_payload.num_slots;
+			} else if (mgr->payloads[i].num_slots) {
+				mgr->payloads[i].num_slots = 0;
+				drm_dp_destroy_payload_step1(mgr, port, port->vcpi.vcpi, &mgr->payloads[i]);
+				req_payload.payload_state = mgr->payloads[i].payload_state;
+				mgr->payloads[i].start_slot = 0;
+			}
+			mgr->payloads[i].payload_state = req_payload.payload_state;
+		}
+		cur_slots += req_payload.num_slots;
+	}
+
+	for (i = 0; i < mgr->max_payloads; i++) {
+		if (mgr->payloads[i].payload_state == DP_PAYLOAD_DELETE_LOCAL) {
+			DRM_DEBUG_KMS("removing payload %d\n", i);
+			for (j = i; j < mgr->max_payloads - 1; j++) {
+				memcpy(&mgr->payloads[j], &mgr->payloads[j + 1], sizeof(struct drm_dp_payload));
+				mgr->proposed_vcpis[j] = mgr->proposed_vcpis[j + 1];
+				if (mgr->proposed_vcpis[j] && mgr->proposed_vcpis[j]->num_slots) {
+					set_bit(j + 1, &mgr->payload_mask);
+				} else {
+					clear_bit(j + 1, &mgr->payload_mask);
+				}
+			}
+			memset(&mgr->payloads[mgr->max_payloads - 1], 0, sizeof(struct drm_dp_payload));
+			mgr->proposed_vcpis[mgr->max_payloads - 1] = NULL;
+			clear_bit(mgr->max_payloads, &mgr->payload_mask);
+
+		}
+	}
+	mutex_unlock(&mgr->payload_lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_dp_update_payload_part1);
+
+/**
+ * drm_dp_update_payload_part2() - Execute payload update part 2
+ * @mgr: manager to use.
+ *
+ * This iterates over all proposed virtual channels, and tries to
+ * allocate space in the link for them. For 0->slots transitions,
+ * this step writes the remote VC payload commands. For slots->0
+ * this just resets some internal state.
+ */
+int drm_dp_update_payload_part2(struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct drm_dp_mst_port *port;
+	int i;
+	int ret = 0;
+	mutex_lock(&mgr->payload_lock);
+	for (i = 0; i < mgr->max_payloads; i++) {
+
+		if (!mgr->proposed_vcpis[i])
+			continue;
+
+		port = container_of(mgr->proposed_vcpis[i], struct drm_dp_mst_port, vcpi);
+
+		DRM_DEBUG_KMS("payload %d %d\n", i, mgr->payloads[i].payload_state);
+		if (mgr->payloads[i].payload_state == DP_PAYLOAD_LOCAL) {
+			ret = drm_dp_create_payload_step2(mgr, port, mgr->proposed_vcpis[i]->vcpi, &mgr->payloads[i]);
+		} else if (mgr->payloads[i].payload_state == DP_PAYLOAD_DELETE_LOCAL) {
+			ret = drm_dp_destroy_payload_step2(mgr, mgr->proposed_vcpis[i]->vcpi, &mgr->payloads[i]);
+		}
+		if (ret) {
+			mutex_unlock(&mgr->payload_lock);
+			return ret;
+		}
+	}
+	mutex_unlock(&mgr->payload_lock);
+	return 0;
+}
+EXPORT_SYMBOL(drm_dp_update_payload_part2);
+
+#if 0 /* unused as of yet */
+static int drm_dp_send_dpcd_read(struct drm_dp_mst_topology_mgr *mgr,
+				 struct drm_dp_mst_port *port,
+				 int offset, int size)
+{
+	int len;
+	struct drm_dp_sideband_msg_tx *txmsg;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg)
+		return -ENOMEM;
+
+	len = build_dpcd_read(txmsg, port->port_num, 0, 8);
+	txmsg->dst = port->parent;
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	return 0;
+}
+#endif
+
+static int drm_dp_send_dpcd_write(struct drm_dp_mst_topology_mgr *mgr,
+				  struct drm_dp_mst_port *port,
+				  int offset, int size, u8 *bytes)
+{
+	int len;
+	int ret;
+	struct drm_dp_sideband_msg_tx *txmsg;
+	struct drm_dp_mst_branch *mstb;
+
+	mstb = drm_dp_get_validated_mstb_ref(mgr, port->parent);
+	if (!mstb)
+		return -EINVAL;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg) {
+		ret = -ENOMEM;
+		goto fail_put;
+	}
+
+	len = build_dpcd_write(txmsg, port->port_num, offset, size, bytes);
+	txmsg->dst = mstb;
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+	if (ret > 0) {
+		if (txmsg->reply.reply_type == 1) {
+			ret = -EINVAL;
+		} else
+			ret = 0;
+	}
+	kfree(txmsg);
+fail_put:
+	drm_dp_put_mst_branch_device(mstb);
+	return ret;
+}
+
+static int drm_dp_encode_up_ack_reply(struct drm_dp_sideband_msg_tx *msg, u8 req_type)
+{
+	struct drm_dp_sideband_msg_reply_body reply;
+
+	reply.reply_type = 1;
+	reply.req_type = req_type;
+	drm_dp_encode_sideband_reply(&reply, msg);
+	return 0;
+}
+
+static int drm_dp_send_up_ack_reply(struct drm_dp_mst_topology_mgr *mgr,
+				    struct drm_dp_mst_branch *mstb,
+				    int req_type, int seqno, bool broadcast)
+{
+	struct drm_dp_sideband_msg_tx *txmsg;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg)
+		return -ENOMEM;
+
+	txmsg->dst = mstb;
+	txmsg->seqno = seqno;
+	drm_dp_encode_up_ack_reply(txmsg, req_type);
+
+	mutex_lock(&mgr->qlock);
+	list_add_tail(&txmsg->next, &mgr->tx_msg_upq);
+	if (!mgr->tx_up_in_progress) {
+		process_single_up_tx_qlock(mgr);
+	}
+	mutex_unlock(&mgr->qlock);
+	return 0;
+}
+
+static int drm_dp_get_vc_payload_bw(int dp_link_bw, int dp_link_count)
+{
+	switch (dp_link_bw) {
+	case DP_LINK_BW_1_62:
+		return 3 * dp_link_count;
+	case DP_LINK_BW_2_7:
+		return 5 * dp_link_count;
+	case DP_LINK_BW_5_4:
+		return 10 * dp_link_count;
+	}
+	BUG();
+}
+
+/**
+ * drm_dp_mst_topology_mgr_set_mst() - Set the MST state for a topology manager
+ * @mgr: manager to set state for
+ * @mst_state: true to enable MST on this connector - false to disable.
+ *
+ * This is called by the driver when it detects an MST capable device plugged
+ * into a DP MST capable port, or when a DP MST capable device is unplugged.
+ */
+int drm_dp_mst_topology_mgr_set_mst(struct drm_dp_mst_topology_mgr *mgr, bool mst_state)
+{
+	int ret = 0;
+	struct drm_dp_mst_branch *mstb = NULL;
+
+	mutex_lock(&mgr->lock);
+	if (mst_state == mgr->mst_state)
+		goto out_unlock;
+
+	mgr->mst_state = mst_state;
+	/* set the device into MST mode */
+	if (mst_state) {
+		WARN_ON(mgr->mst_primary);
+
+		/* get dpcd info */
+		ret = drm_dp_dpcd_read(mgr->aux, DP_DPCD_REV, mgr->dpcd, DP_RECEIVER_CAP_SIZE);
+		if (ret != DP_RECEIVER_CAP_SIZE) {
+			DRM_DEBUG_KMS("failed to read DPCD\n");
+			goto out_unlock;
+		}
+
+		mgr->pbn_div = drm_dp_get_vc_payload_bw(mgr->dpcd[1], mgr->dpcd[2] & DP_MAX_LANE_COUNT_MASK);
+		mgr->total_pbn = 2560;
+		mgr->total_slots = DIV_ROUND_UP(mgr->total_pbn, mgr->pbn_div);
+		mgr->avail_slots = mgr->total_slots;
+
+		/* add initial branch device at LCT 1 */
+		mstb = drm_dp_add_mst_branch_device(1, NULL);
+		if (mstb == NULL) {
+			ret = -ENOMEM;
+			goto out_unlock;
+		}
+		mstb->mgr = mgr;
+
+		/* give this the main reference */
+		mgr->mst_primary = mstb;
+		kref_get(&mgr->mst_primary->kref);
+
+		{
+			struct drm_dp_payload reset_pay;
+			reset_pay.start_slot = 0;
+			reset_pay.num_slots = 0x3f;
+			drm_dp_dpcd_write_payload(mgr, 0, &reset_pay);
+		}
+
+		ret = drm_dp_dpcd_writeb(mgr->aux, DP_MSTM_CTRL,
+					 DP_MST_EN | DP_UP_REQ_EN | DP_UPSTREAM_IS_SRC);
+		if (ret < 0) {
+			goto out_unlock;
+		}
+
+
+		/* sort out guid */
+		ret = drm_dp_dpcd_read(mgr->aux, DP_GUID, mgr->guid, 16);
+		if (ret != 16) {
+			DRM_DEBUG_KMS("failed to read DP GUID %d\n", ret);
+			goto out_unlock;
+		}
+
+		mgr->guid_valid = drm_dp_validate_guid(mgr, mgr->guid);
+		if (!mgr->guid_valid) {
+			ret = drm_dp_dpcd_write(mgr->aux, DP_GUID, mgr->guid, 16);
+			mgr->guid_valid = true;
+		}
+
+		queue_work(system_long_wq, &mgr->work);
+
+		ret = 0;
+	} else {
+		/* disable MST on the device */
+		mstb = mgr->mst_primary;
+		mgr->mst_primary = NULL;
+		/* this can fail if the device is gone */
+		drm_dp_dpcd_writeb(mgr->aux, DP_MSTM_CTRL, 0);
+		ret = 0;
+		memset(mgr->payloads, 0, mgr->max_payloads * sizeof(struct drm_dp_payload));
+		mgr->payload_mask = 0;
+		set_bit(0, &mgr->payload_mask);
+		mgr->vcpi_mask = 0;
+	}
+
+out_unlock:
+	mutex_unlock(&mgr->lock);
+	if (mstb)
+		drm_dp_put_mst_branch_device(mstb);
+	return ret;
+
+}
+EXPORT_SYMBOL(drm_dp_mst_topology_mgr_set_mst);
+
+/**
+ * drm_dp_mst_topology_mgr_suspend() - suspend the MST manager
+ * @mgr: manager to suspend
+ *
+ * This function tells the MST device that we can't handle UP messages
+ * anymore. This should stop it from sending any since we are suspended.
+ */
+void drm_dp_mst_topology_mgr_suspend(struct drm_dp_mst_topology_mgr *mgr)
+{
+	mutex_lock(&mgr->lock);
+	drm_dp_dpcd_writeb(mgr->aux, DP_MSTM_CTRL,
+			   DP_MST_EN | DP_UPSTREAM_IS_SRC);
+	mutex_unlock(&mgr->lock);
+}
+EXPORT_SYMBOL(drm_dp_mst_topology_mgr_suspend);
+
+/**
+ * drm_dp_mst_topology_mgr_resume() - resume the MST manager
+ * @mgr: manager to resume
+ *
+ * This will fetch DPCD and see if the device is still there,
+ * if it is, it will rewrite the MSTM control bits, and return.
+ *
+ * if the device fails this returns -1, and the driver should do
+ * a full MST reprobe, in case we were undocked.
+ */
+int drm_dp_mst_topology_mgr_resume(struct drm_dp_mst_topology_mgr *mgr)
+{
+	int ret = 0;
+
+	mutex_lock(&mgr->lock);
+
+	if (mgr->mst_primary) {
+		int sret;
+		sret = drm_dp_dpcd_read(mgr->aux, DP_DPCD_REV, mgr->dpcd, DP_RECEIVER_CAP_SIZE);
+		if (sret != DP_RECEIVER_CAP_SIZE) {
+			DRM_DEBUG_KMS("dpcd read failed - undocked during suspend?\n");
+			ret = -1;
+			goto out_unlock;
+		}
+
+		ret = drm_dp_dpcd_writeb(mgr->aux, DP_MSTM_CTRL,
+					 DP_MST_EN | DP_UP_REQ_EN | DP_UPSTREAM_IS_SRC);
+		if (ret < 0) {
+			DRM_DEBUG_KMS("mst write failed - undocked during suspend?\n");
+			ret = -1;
+			goto out_unlock;
+		}
+		ret = 0;
+	} else
+		ret = -1;
+
+out_unlock:
+	mutex_unlock(&mgr->lock);
+	return ret;
+}
+EXPORT_SYMBOL(drm_dp_mst_topology_mgr_resume);
+
+static void drm_dp_get_one_sb_msg(struct drm_dp_mst_topology_mgr *mgr, bool up)
+{
+	int len;
+	u8 replyblock[32];
+	int replylen, origlen, curreply;
+	int ret;
+	struct drm_dp_sideband_msg_rx *msg;
+	int basereg = up ? DP_SIDEBAND_MSG_UP_REQ_BASE : DP_SIDEBAND_MSG_DOWN_REP_BASE;
+	msg = up ? &mgr->up_req_recv : &mgr->down_rep_recv;
+
+	len = min(mgr->max_dpcd_transaction_bytes, 16);
+	ret = drm_dp_dpcd_read(mgr->aux, basereg,
+			       replyblock, len);
+	if (ret != len) {
+		DRM_DEBUG_KMS("failed to read DPCD down rep %d %d\n", len, ret);
+		return;
+	}
+	ret = drm_dp_sideband_msg_build(msg, replyblock, len, true);
+	if (!ret) {
+		DRM_DEBUG_KMS("sideband msg build failed %d\n", replyblock[0]);
+		return;
+	}
+	replylen = msg->curchunk_len + msg->curchunk_hdrlen;
+
+	origlen = replylen;
+	replylen -= len;
+	curreply = len;
+	while (replylen > 0) {
+		len = min3(replylen, mgr->max_dpcd_transaction_bytes, 16);
+		ret = drm_dp_dpcd_read(mgr->aux, basereg + curreply,
+				    replyblock, len);
+		if (ret != len) {
+			DRM_DEBUG_KMS("failed to read a chunk\n");
+		}
+		ret = drm_dp_sideband_msg_build(msg, replyblock, len, false);
+		if (ret == false)
+			DRM_DEBUG_KMS("failed to build sideband msg\n");
+		curreply += len;
+		replylen -= len;
+	}
+}
+
+static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
+{
+	int ret = 0;
+
+	drm_dp_get_one_sb_msg(mgr, false);
+
+	if (mgr->down_rep_recv.have_eomt) {
+		struct drm_dp_sideband_msg_tx *txmsg;
+		struct drm_dp_mst_branch *mstb;
+		int slot = -1;
+		mstb = drm_dp_get_mst_branch_device(mgr,
+						    mgr->down_rep_recv.initial_hdr.lct,
+						    mgr->down_rep_recv.initial_hdr.rad);
+
+		if (!mstb) {
+			DRM_DEBUG_KMS("Got MST reply from unknown device %d\n", mgr->down_rep_recv.initial_hdr.lct);
+			memset(&mgr->down_rep_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
+			return 0;
+		}
+
+		/* find the message */
+		slot = mgr->down_rep_recv.initial_hdr.seqno;
+		mutex_lock(&mgr->qlock);
+		txmsg = mstb->tx_slots[slot];
+		/* remove from slots */
+		mutex_unlock(&mgr->qlock);
+
+		if (!txmsg) {
+			DRM_DEBUG_KMS("Got MST reply with no msg %p %d %d %02x %02x\n",
+			       mstb,
+			       mgr->down_rep_recv.initial_hdr.seqno,
+			       mgr->down_rep_recv.initial_hdr.lct,
+				      mgr->down_rep_recv.initial_hdr.rad[0],
+				      mgr->down_rep_recv.msg[0]);
+			drm_dp_put_mst_branch_device(mstb);
+			memset(&mgr->down_rep_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
+			return 0;
+		}
+
+		drm_dp_sideband_parse_reply(&mgr->down_rep_recv, &txmsg->reply);
+		if (txmsg->reply.reply_type == 1) {
+			DRM_DEBUG_KMS("Got NAK reply: req 0x%02x, reason 0x%02x, nak data 0x%02x\n", txmsg->reply.req_type, txmsg->reply.u.nak.reason, txmsg->reply.u.nak.nak_data);
+		}
+
+		memset(&mgr->down_rep_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
+		drm_dp_put_mst_branch_device(mstb);
+
+		mutex_lock(&mgr->qlock);
+		txmsg->state = DRM_DP_SIDEBAND_TX_RX;
+		mstb->tx_slots[slot] = NULL;
+		mutex_unlock(&mgr->qlock);
+
+		wake_up(&mgr->tx_waitq);
+	}
+	return ret;
+}
+
+static int drm_dp_mst_handle_up_req(struct drm_dp_mst_topology_mgr *mgr)
+{
+	int ret = 0;
+	drm_dp_get_one_sb_msg(mgr, true);
+
+	if (mgr->up_req_recv.have_eomt) {
+		struct drm_dp_sideband_msg_req_body msg;
+		struct drm_dp_mst_branch *mstb;
+		bool seqno;
+		mstb = drm_dp_get_mst_branch_device(mgr,
+						    mgr->up_req_recv.initial_hdr.lct,
+						    mgr->up_req_recv.initial_hdr.rad);
+		if (!mstb) {
+			DRM_DEBUG_KMS("Got MST reply from unknown device %d\n", mgr->up_req_recv.initial_hdr.lct);
+			memset(&mgr->up_req_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
+			return 0;
+		}
+
+		seqno = mgr->up_req_recv.initial_hdr.seqno;
+		drm_dp_sideband_parse_req(&mgr->up_req_recv, &msg);
+
+		if (msg.req_type == DP_CONNECTION_STATUS_NOTIFY) {
+			drm_dp_send_up_ack_reply(mgr, mstb, msg.req_type, seqno, false);
+			drm_dp_update_port(mstb, &msg.u.conn_stat);
+			DRM_DEBUG_KMS("Got CSN: pn: %d ldps:%d ddps: %d mcs: %d ip: %d pdt: %d\n", msg.u.conn_stat.port_number, msg.u.conn_stat.legacy_device_plug_status, msg.u.conn_stat.displayport_device_plug_status, msg.u.conn_stat.message_capability_status, msg.u.conn_stat.input_port, msg.u.conn_stat.peer_device_type);
+			(*mgr->cbs->hotplug)(mgr);
+
+		} else if (msg.req_type == DP_RESOURCE_STATUS_NOTIFY) {
+			drm_dp_send_up_ack_reply(mgr, mstb, msg.req_type, seqno, false);
+			DRM_DEBUG_KMS("Got RSN: pn: %d avail_pbn %d\n", msg.u.resource_stat.port_number, msg.u.resource_stat.available_pbn);
+		}
+
+		drm_dp_put_mst_branch_device(mstb);
+		memset(&mgr->up_req_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
+	}
+	return ret;
+}
+
+/**
+ * drm_dp_mst_hpd_irq() - MST hotplug IRQ notify
+ * @mgr: manager to notify irq for.
+ * @esi: 4 bytes from SINK_COUNT_ESI
+ * @handled: whether the hpd interrupt was consumed or not
+ *
+ * This should be called from the driver when it detects a short IRQ,
+ * along with the value of the DEVICE_SERVICE_IRQ_VECTOR_ESI0. The
+ * topology manager will process the sideband messages received as a result
+ * of this.
+ */
+int drm_dp_mst_hpd_irq(struct drm_dp_mst_topology_mgr *mgr, u8 *esi, bool *handled)
+{
+	int ret = 0;
+	int sc;
+	*handled = false;
+	sc = esi[0] & 0x3f;
+
+	if (sc != mgr->sink_count) {
+		mgr->sink_count = sc;
+		*handled = true;
+	}
+
+	if (esi[1] & DP_DOWN_REP_MSG_RDY) {
+		ret = drm_dp_mst_handle_down_rep(mgr);
+		*handled = true;
+	}
+
+	if (esi[1] & DP_UP_REQ_MSG_RDY) {
+		ret |= drm_dp_mst_handle_up_req(mgr);
+		*handled = true;
+	}
+
+	drm_dp_mst_kick_tx(mgr);
+	return ret;
+}
+EXPORT_SYMBOL(drm_dp_mst_hpd_irq);
+
+/**
+ * drm_dp_mst_detect_port() - get connection status for an MST port
+ * @mgr: manager for this port
+ * @port: unverified pointer to a port
+ *
+ * This returns the current connection state for a port. It validates the
+ * port pointer still exists so the caller doesn't require a reference
+ */
+enum drm_connector_status drm_dp_mst_detect_port(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+{
+	enum drm_connector_status status = connector_status_disconnected;
+
+	/* we need to search for the port in the mgr in case its gone */
+	port = drm_dp_get_validated_port_ref(mgr, port);
+	if (!port)
+		return connector_status_disconnected;
+
+	if (!port->ddps)
+		goto out;
+
+	switch (port->pdt) {
+	case DP_PEER_DEVICE_NONE:
+	case DP_PEER_DEVICE_MST_BRANCHING:
+		break;
+
+	case DP_PEER_DEVICE_SST_SINK:
+		status = connector_status_connected;
+		break;
+	case DP_PEER_DEVICE_DP_LEGACY_CONV:
+		if (port->ldps)
+			status = connector_status_connected;
+		break;
+	}
+out:
+	drm_dp_put_port(port);
+	return status;
+}
+EXPORT_SYMBOL(drm_dp_mst_detect_port);
+
+/**
+ * drm_dp_mst_get_edid() - get EDID for an MST port
+ * @connector: toplevel connector to get EDID for
+ * @mgr: manager for this port
+ * @port: unverified pointer to a port.
+ *
+ * This returns an EDID for the port connected to a connector,
+ * It validates the pointer still exists so the caller doesn't require a
+ * reference.
+ */
+struct edid *drm_dp_mst_get_edid(struct drm_connector *connector, struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+{
+	struct edid *edid = NULL;
+
+	/* we need to search for the port in the mgr in case its gone */
+	port = drm_dp_get_validated_port_ref(mgr, port);
+	if (!port)
+		return NULL;
+
+	edid = drm_get_edid(connector, &port->aux.ddc);
+	drm_dp_put_port(port);
+	return edid;
+}
+EXPORT_SYMBOL(drm_dp_mst_get_edid);
+
+/**
+ * drm_dp_find_vcpi_slots() - find slots for this PBN value
+ * @mgr: manager to use
+ * @pbn: payload bandwidth to convert into slots.
+ */
+int drm_dp_find_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr,
+			   int pbn)
+{
+	int num_slots;
+
+	num_slots = DIV_ROUND_UP(pbn, mgr->pbn_div);
+
+	if (num_slots > mgr->avail_slots)
+		return -ENOSPC;
+	return num_slots;
+}
+EXPORT_SYMBOL(drm_dp_find_vcpi_slots);
+
+static int drm_dp_init_vcpi(struct drm_dp_mst_topology_mgr *mgr,
+			    struct drm_dp_vcpi *vcpi, int pbn)
+{
+	int num_slots;
+	int ret;
+
+	num_slots = DIV_ROUND_UP(pbn, mgr->pbn_div);
+
+	if (num_slots > mgr->avail_slots)
+		return -ENOSPC;
+
+	vcpi->pbn = pbn;
+	vcpi->aligned_pbn = num_slots * mgr->pbn_div;
+	vcpi->num_slots = num_slots;
+
+	ret = drm_dp_mst_assign_payload_id(mgr, vcpi);
+	if (ret < 0)
+		return ret;
+	return 0;
+}
+
+/**
+ * drm_dp_mst_allocate_vcpi() - Allocate a virtual channel
+ * @mgr: manager for this port
+ * @port: port to allocate a virtual channel for.
+ * @pbn: payload bandwidth number to request
+ * @slots: returned number of slots for this PBN.
+ */
+bool drm_dp_mst_allocate_vcpi(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port, int pbn, int *slots)
+{
+	int ret;
+
+	port = drm_dp_get_validated_port_ref(mgr, port);
+	if (!port)
+		return false;
+
+	if (port->vcpi.vcpi > 0) {
+		DRM_DEBUG_KMS("payload: vcpi %d already allocated for pbn %d - requested pbn %d\n", port->vcpi.vcpi, port->vcpi.pbn, pbn);
+		if (pbn == port->vcpi.pbn) {
+			*slots = port->vcpi.num_slots;
+			return true;
+		}
+	}
+
+	ret = drm_dp_init_vcpi(mgr, &port->vcpi, pbn);
+	if (ret) {
+		DRM_DEBUG_KMS("failed to init vcpi %d %d %d\n", DIV_ROUND_UP(pbn, mgr->pbn_div), mgr->avail_slots, ret);
+		goto out;
+	}
+	DRM_DEBUG_KMS("initing vcpi for %d %d\n", pbn, port->vcpi.num_slots);
+	*slots = port->vcpi.num_slots;
+
+	drm_dp_put_port(port);
+	return true;
+out:
+	return false;
+}
+EXPORT_SYMBOL(drm_dp_mst_allocate_vcpi);
+
+/**
+ * drm_dp_mst_reset_vcpi_slots() - Reset number of slots to 0 for VCPI
+ * @mgr: manager for this port
+ * @port: unverified pointer to a port.
+ *
+ * This just resets the number of slots for the ports VCPI for later programming.
+ */
+void drm_dp_mst_reset_vcpi_slots(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+{
+	port = drm_dp_get_validated_port_ref(mgr, port);
+	if (!port)
+		return;
+	port->vcpi.num_slots = 0;
+	drm_dp_put_port(port);
+}
+EXPORT_SYMBOL(drm_dp_mst_reset_vcpi_slots);
+
+/**
+ * drm_dp_mst_deallocate_vcpi() - deallocate a VCPI
+ * @mgr: manager for this port
+ * @port: unverified port to deallocate vcpi for
+ */
+void drm_dp_mst_deallocate_vcpi(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port)
+{
+	port = drm_dp_get_validated_port_ref(mgr, port);
+	if (!port)
+		return;
+
+	drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi);
+	port->vcpi.num_slots = 0;
+	port->vcpi.pbn = 0;
+	port->vcpi.aligned_pbn = 0;
+	port->vcpi.vcpi = 0;
+	drm_dp_put_port(port);
+}
+EXPORT_SYMBOL(drm_dp_mst_deallocate_vcpi);
+
+static int drm_dp_dpcd_write_payload(struct drm_dp_mst_topology_mgr *mgr,
+				     int id, struct drm_dp_payload *payload)
+{
+	u8 payload_alloc[3], status;
+	int ret;
+	int retries = 0;
+
+	drm_dp_dpcd_writeb(mgr->aux, DP_PAYLOAD_TABLE_UPDATE_STATUS,
+			   DP_PAYLOAD_TABLE_UPDATED);
+
+	payload_alloc[0] = id;
+	payload_alloc[1] = payload->start_slot;
+	payload_alloc[2] = payload->num_slots;
+
+	ret = drm_dp_dpcd_write(mgr->aux, DP_PAYLOAD_ALLOCATE_SET, payload_alloc, 3);
+	if (ret != 3) {
+		DRM_DEBUG_KMS("failed to write payload allocation %d\n", ret);
+		goto fail;
+	}
+
+retry:
+	ret = drm_dp_dpcd_readb(mgr->aux, DP_PAYLOAD_TABLE_UPDATE_STATUS, &status);
+	if (ret < 0) {
+		DRM_DEBUG_KMS("failed to read payload table status %d\n", ret);
+		goto fail;
+	}
+
+	if (!(status & DP_PAYLOAD_TABLE_UPDATED)) {
+		retries++;
+		if (retries < 20) {
+			usleep_range(10000, 20000);
+			goto retry;
+		}
+		DRM_DEBUG_KMS("status not set after read payload table status %d\n", status);
+		ret = -EINVAL;
+		goto fail;
+	}
+	ret = 0;
+fail:
+	return ret;
+}
+
+
+/**
+ * drm_dp_check_act_status() - Check ACT handled status.
+ * @mgr: manager to use
+ *
+ * Check the payload status bits in the DPCD for ACT handled completion.
+ */
+int drm_dp_check_act_status(struct drm_dp_mst_topology_mgr *mgr)
+{
+	u8 status;
+	int ret;
+	int count = 0;
+
+	do {
+		ret = drm_dp_dpcd_readb(mgr->aux, DP_PAYLOAD_TABLE_UPDATE_STATUS, &status);
+
+		if (ret < 0) {
+			DRM_DEBUG_KMS("failed to read payload table status %d\n", ret);
+			goto fail;
+		}
+
+		if (status & DP_PAYLOAD_ACT_HANDLED)
+			break;
+		count++;
+		udelay(100);
+
+	} while (count < 30);
+
+	if (!(status & DP_PAYLOAD_ACT_HANDLED)) {
+		DRM_DEBUG_KMS("failed to get ACT bit %d after %d retries\n", status, count);
+		ret = -EINVAL;
+		goto fail;
+	}
+	return 0;
+fail:
+	return ret;
+}
+EXPORT_SYMBOL(drm_dp_check_act_status);
+
+/**
+ * drm_dp_calc_pbn_mode() - Calculate the PBN for a mode.
+ * @clock: dot clock for the mode
+ * @bpp: bpp for the mode.
+ *
+ * This uses the formula in the spec to calculate the PBN value for a mode.
+ */
+int drm_dp_calc_pbn_mode(int clock, int bpp)
+{
+	fixed20_12 pix_bw;
+	fixed20_12 fbpp;
+	fixed20_12 result;
+	fixed20_12 margin, tmp;
+	u32 res;
+
+	pix_bw.full = dfixed_const(clock);
+	fbpp.full = dfixed_const(bpp);
+	tmp.full = dfixed_const(8);
+	fbpp.full = dfixed_div(fbpp, tmp);
+
+	result.full = dfixed_mul(pix_bw, fbpp);
+	margin.full = dfixed_const(54);
+	tmp.full = dfixed_const(64);
+	margin.full = dfixed_div(margin, tmp);
+	result.full = dfixed_div(result, margin);
+
+	margin.full = dfixed_const(1006);
+	tmp.full = dfixed_const(1000);
+	margin.full = dfixed_div(margin, tmp);
+	result.full = dfixed_mul(result, margin);
+
+	result.full = dfixed_div(result, tmp);
+	result.full = dfixed_ceil(result);
+	res = dfixed_trunc(result);
+	return res;
+}
+EXPORT_SYMBOL(drm_dp_calc_pbn_mode);
+
+static int test_calc_pbn_mode(void)
+{
+	int ret;
+	ret = drm_dp_calc_pbn_mode(154000, 30);
+	if (ret != 689)
+		return -EINVAL;
+	ret = drm_dp_calc_pbn_mode(234000, 30);
+	if (ret != 1047)
+		return -EINVAL;
+	return 0;
+}
+
+/* we want to kick the TX after we've ack the up/down IRQs. */
+static void drm_dp_mst_kick_tx(struct drm_dp_mst_topology_mgr *mgr)
+{
+	queue_work(system_long_wq, &mgr->tx_work);
+}
+
+static void drm_dp_mst_dump_mstb(struct seq_file *m,
+				 struct drm_dp_mst_branch *mstb)
+{
+	struct drm_dp_mst_port *port;
+	int tabs = mstb->lct;
+	char prefix[10];
+	int i;
+
+	for (i = 0; i < tabs; i++)
+		prefix[i] = '\t';
+	prefix[i] = '\0';
+
+	seq_printf(m, "%smst: %p, %d\n", prefix, mstb, mstb->num_ports);
+	list_for_each_entry(port, &mstb->ports, next) {
+		seq_printf(m, "%sport: %d: ddps: %d ldps: %d, %p, conn: %p\n", prefix, port->port_num, port->ddps, port->ldps, port, port->connector);
+		if (port->mstb)
+			drm_dp_mst_dump_mstb(m, port->mstb);
+	}
+}
+
+static bool dump_dp_payload_table(struct drm_dp_mst_topology_mgr *mgr,
+				  char *buf)
+{
+	int ret;
+	int i;
+	for (i = 0; i < 4; i++) {
+		ret = drm_dp_dpcd_read(mgr->aux, DP_PAYLOAD_TABLE_UPDATE_STATUS + (i * 16), &buf[i * 16], 16);
+		if (ret != 16)
+			break;
+	}
+	if (i == 4)
+		return true;
+	return false;
+}
+
+/**
+ * drm_dp_mst_dump_topology(): dump topology to seq file.
+ * @m: seq_file to dump output to
+ * @mgr: manager to dump current topology for.
+ *
+ * helper to dump MST topology to a seq file for debugfs.
+ */
+void drm_dp_mst_dump_topology(struct seq_file *m,
+			      struct drm_dp_mst_topology_mgr *mgr)
+{
+	int i;
+	struct drm_dp_mst_port *port;
+	mutex_lock(&mgr->lock);
+	if (mgr->mst_primary)
+		drm_dp_mst_dump_mstb(m, mgr->mst_primary);
+
+	/* dump VCPIs */
+	mutex_unlock(&mgr->lock);
+
+	mutex_lock(&mgr->payload_lock);
+	seq_printf(m, "vcpi: %lx %lx\n", mgr->payload_mask, mgr->vcpi_mask);
+
+	for (i = 0; i < mgr->max_payloads; i++) {
+		if (mgr->proposed_vcpis[i]) {
+			port = container_of(mgr->proposed_vcpis[i], struct drm_dp_mst_port, vcpi);
+			seq_printf(m, "vcpi %d: %d %d %d\n", i, port->port_num, port->vcpi.vcpi, port->vcpi.num_slots);
+		} else
+			seq_printf(m, "vcpi %d:unsed\n", i);
+	}
+	for (i = 0; i < mgr->max_payloads; i++) {
+		seq_printf(m, "payload %d: %d, %d, %d\n",
+			   i,
+			   mgr->payloads[i].payload_state,
+			   mgr->payloads[i].start_slot,
+			   mgr->payloads[i].num_slots);
+
+
+	}
+	mutex_unlock(&mgr->payload_lock);
+
+	mutex_lock(&mgr->lock);
+	if (mgr->mst_primary) {
+		u8 buf[64];
+		bool bret;
+		int ret;
+		ret = drm_dp_dpcd_read(mgr->aux, DP_DPCD_REV, buf, DP_RECEIVER_CAP_SIZE);
+		seq_printf(m, "dpcd: ");
+		for (i = 0; i < DP_RECEIVER_CAP_SIZE; i++)
+			seq_printf(m, "%02x ", buf[i]);
+		seq_printf(m, "\n");
+		ret = drm_dp_dpcd_read(mgr->aux, DP_FAUX_CAP, buf, 2);
+		seq_printf(m, "faux/mst: ");
+		for (i = 0; i < 2; i++)
+			seq_printf(m, "%02x ", buf[i]);
+		seq_printf(m, "\n");
+		ret = drm_dp_dpcd_read(mgr->aux, DP_MSTM_CTRL, buf, 1);
+		seq_printf(m, "mst ctrl: ");
+		for (i = 0; i < 1; i++)
+			seq_printf(m, "%02x ", buf[i]);
+		seq_printf(m, "\n");
+
+		bret = dump_dp_payload_table(mgr, buf);
+		if (bret == true) {
+			seq_printf(m, "payload table: ");
+			for (i = 0; i < 63; i++)
+				seq_printf(m, "%02x ", buf[i]);
+			seq_printf(m, "\n");
+		}
+
+	}
+
+	mutex_unlock(&mgr->lock);
+
+}
+EXPORT_SYMBOL(drm_dp_mst_dump_topology);
+
+static void drm_dp_tx_work(struct work_struct *work)
+{
+	struct drm_dp_mst_topology_mgr *mgr = container_of(work, struct drm_dp_mst_topology_mgr, tx_work);
+
+	mutex_lock(&mgr->qlock);
+	if (mgr->tx_down_in_progress)
+		process_single_down_tx_qlock(mgr);
+	mutex_unlock(&mgr->qlock);
+}
+
+/**
+ * drm_dp_mst_topology_mgr_init - initialise a topology manager
+ * @mgr: manager struct to initialise
+ * @dev: device providing this structure - for i2c addition.
+ * @aux: DP helper aux channel to talk to this device
+ * @max_dpcd_transaction_bytes: hw specific DPCD transaction limit
+ * @max_payloads: maximum number of payloads this GPU can source
+ * @conn_base_id: the connector object ID the MST device is connected to.
+ *
+ * Return 0 for success, or negative error code on failure
+ */
+int drm_dp_mst_topology_mgr_init(struct drm_dp_mst_topology_mgr *mgr,
+				 struct device *dev, struct drm_dp_aux *aux,
+				 int max_dpcd_transaction_bytes,
+				 int max_payloads, int conn_base_id)
+{
+	mutex_init(&mgr->lock);
+	mutex_init(&mgr->qlock);
+	mutex_init(&mgr->payload_lock);
+	INIT_LIST_HEAD(&mgr->tx_msg_upq);
+	INIT_LIST_HEAD(&mgr->tx_msg_downq);
+	INIT_WORK(&mgr->work, drm_dp_mst_link_probe_work);
+	INIT_WORK(&mgr->tx_work, drm_dp_tx_work);
+	init_waitqueue_head(&mgr->tx_waitq);
+	mgr->dev = dev;
+	mgr->aux = aux;
+	mgr->max_dpcd_transaction_bytes = max_dpcd_transaction_bytes;
+	mgr->max_payloads = max_payloads;
+	mgr->conn_base_id = conn_base_id;
+	mgr->payloads = kcalloc(max_payloads, sizeof(struct drm_dp_payload), GFP_KERNEL);
+	if (!mgr->payloads)
+		return -ENOMEM;
+	mgr->proposed_vcpis = kcalloc(max_payloads, sizeof(struct drm_dp_vcpi *), GFP_KERNEL);
+	if (!mgr->proposed_vcpis)
+		return -ENOMEM;
+	set_bit(0, &mgr->payload_mask);
+	test_calc_pbn_mode();
+	return 0;
+}
+EXPORT_SYMBOL(drm_dp_mst_topology_mgr_init);
+
+/**
+ * drm_dp_mst_topology_mgr_destroy() - destroy topology manager.
+ * @mgr: manager to destroy
+ */
+void drm_dp_mst_topology_mgr_destroy(struct drm_dp_mst_topology_mgr *mgr)
+{
+	mutex_lock(&mgr->payload_lock);
+	kfree(mgr->payloads);
+	mgr->payloads = NULL;
+	kfree(mgr->proposed_vcpis);
+	mgr->proposed_vcpis = NULL;
+	mutex_unlock(&mgr->payload_lock);
+	mgr->dev = NULL;
+	mgr->aux = NULL;
+}
+EXPORT_SYMBOL(drm_dp_mst_topology_mgr_destroy);
+
+/* I2C device */
+static int drm_dp_mst_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs,
+			       int num)
+{
+	struct drm_dp_aux *aux = adapter->algo_data;
+	struct drm_dp_mst_port *port = container_of(aux, struct drm_dp_mst_port, aux);
+	struct drm_dp_mst_branch *mstb;
+	struct drm_dp_mst_topology_mgr *mgr = port->mgr;
+	unsigned int i;
+	bool reading = false;
+	struct drm_dp_sideband_msg_req_body msg;
+	struct drm_dp_sideband_msg_tx *txmsg = NULL;
+	int ret;
+
+	mstb = drm_dp_get_validated_mstb_ref(mgr, port->parent);
+	if (!mstb)
+		return -EREMOTEIO;
+
+	/* construct i2c msg */
+	/* see if last msg is a read */
+	if (msgs[num - 1].flags & I2C_M_RD)
+		reading = true;
+
+	if (!reading) {
+		DRM_DEBUG_KMS("Unsupported I2C transaction for MST device\n");
+		ret = -EIO;
+		goto out;
+	}
+
+	msg.req_type = DP_REMOTE_I2C_READ;
+	msg.u.i2c_read.num_transactions = num - 1;
+	msg.u.i2c_read.port_number = port->port_num;
+	for (i = 0; i < num - 1; i++) {
+		msg.u.i2c_read.transactions[i].i2c_dev_id = msgs[i].addr;
+		msg.u.i2c_read.transactions[i].num_bytes = msgs[i].len;
+		msg.u.i2c_read.transactions[i].bytes = msgs[i].buf;
+	}
+	msg.u.i2c_read.read_i2c_device_id = msgs[num - 1].addr;
+	msg.u.i2c_read.num_bytes_read = msgs[num - 1].len;
+
+	txmsg = kzalloc(sizeof(*txmsg), GFP_KERNEL);
+	if (!txmsg) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	txmsg->dst = mstb;
+	drm_dp_encode_sideband_req(&msg, txmsg);
+
+	drm_dp_queue_down_tx(mgr, txmsg);
+
+	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
+	if (ret > 0) {
+
+		if (txmsg->reply.reply_type == 1) { /* got a NAK back */
+			ret = -EREMOTEIO;
+			goto out;
+		}
+		if (txmsg->reply.u.remote_i2c_read_ack.num_bytes != msgs[num - 1].len) {
+			ret = -EIO;
+			goto out;
+		}
+		memcpy(msgs[num - 1].buf, txmsg->reply.u.remote_i2c_read_ack.bytes, msgs[num - 1].len);
+		ret = num;
+	}
+out:
+	kfree(txmsg);
+	drm_dp_put_mst_branch_device(mstb);
+	return ret;
+}
+
+static u32 drm_dp_mst_i2c_functionality(struct i2c_adapter *adapter)
+{
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
+	       I2C_FUNC_SMBUS_READ_BLOCK_DATA |
+	       I2C_FUNC_SMBUS_BLOCK_PROC_CALL |
+	       I2C_FUNC_10BIT_ADDR;
+}
+
+static const struct i2c_algorithm drm_dp_mst_i2c_algo = {
+	.functionality = drm_dp_mst_i2c_functionality,
+	.master_xfer = drm_dp_mst_i2c_xfer,
+};
+
+/**
+ * drm_dp_mst_register_i2c_bus() - register an I2C adapter for I2C-over-AUX
+ * @aux: DisplayPort AUX channel
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+static int drm_dp_mst_register_i2c_bus(struct drm_dp_aux *aux)
+{
+	aux->ddc.algo = &drm_dp_mst_i2c_algo;
+	aux->ddc.algo_data = aux;
+	aux->ddc.retries = 3;
+
+	aux->ddc.class = I2C_CLASS_DDC;
+	aux->ddc.owner = THIS_MODULE;
+	aux->ddc.dev.parent = aux->dev;
+	aux->ddc.dev.of_node = aux->dev->of_node;
+
+	strlcpy(aux->ddc.name, aux->name ? aux->name : dev_name(aux->dev),
+		sizeof(aux->ddc.name));
+
+	return i2c_add_adapter(&aux->ddc);
+}
+
+/**
+ * drm_dp_mst_unregister_i2c_bus() - unregister an I2C-over-AUX adapter
+ * @aux: DisplayPort AUX channel
+ */
+static void drm_dp_mst_unregister_i2c_bus(struct drm_dp_aux *aux)
+{
+	i2c_del_adapter(&aux->ddc);
+}
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 8218078b6133..bc3da32d4585 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -1,31 +1,11 @@
-/**
- * \file drm_drv.c
- * Generic driver template
- *
- * \author Rickard E. (Rik) Faith <faith@valinux.com>
- * \author Gareth Hughes <gareth@valinux.com>
- *
- * To use this template, you must at least define the following (samples
- * given for the MGA driver):
- *
- * \code
- * #define DRIVER_AUTHOR	"VA Linux Systems, Inc."
- *
- * #define DRIVER_NAME		"mga"
- * #define DRIVER_DESC		"Matrox G200/G400"
- * #define DRIVER_DATE		"20001127"
- *
- * #define drm_x		mga_##x
- * \endcode
- */
-
 /*
- * Created: Thu Nov 23 03:10:50 2000 by gareth@valinux.com
+ * Created: Fri Jan 19 10:48:35 2001 by faith@acm.org
  *
- * Copyright 1999, 2000 Precision Insight, Inc., Cedar Park, Texas.
- * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
+ * Copyright 2001 VA Linux Systems, Inc., Sunnyvale, California.
  * All Rights Reserved.
  *
+ * Author Rickard E. (Rik) Faith <faith@valinux.com>
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
@@ -40,432 +20,886 @@
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
  * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
  */
 
 #include <linux/debugfs.h>
+#include <linux/fs.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/mount.h>
 #include <linux/slab.h>
-#include <linux/export.h>
 #include <drm/drmP.h>
 #include <drm/drm_core.h>
+#include "drm_legacy.h"
+#include "drm_internal.h"
 
+unsigned int drm_debug = 0;	/* 1 to enable debug output */
+EXPORT_SYMBOL(drm_debug);
 
-static int drm_version(struct drm_device *dev, void *data,
-		       struct drm_file *file_priv);
-
-#define DRM_IOCTL_DEF(ioctl, _func, _flags) \
-	[DRM_IOCTL_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0, .name = #ioctl}
-
-/** Ioctl table */
-static const struct drm_ioctl_desc drm_ioctls[] = {
-	DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_IRQ_BUSID, drm_irq_by_busid, DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_CLIENT_CAP, drm_setclientcap, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_VERSION, drm_setversion, DRM_MASTER),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_UNIQUE, drm_setunique, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_BLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_UNBLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AUTH_MAGIC, drm_authmagic, DRM_AUTH|DRM_MASTER),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_MAP, drm_addmap_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_MAP, drm_rmmap_ioctl, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_SAREA_CTX, drm_setsareactx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_SAREA_CTX, drm_getsareactx, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SET_MASTER, drm_setmaster_ioctl, DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_DROP_MASTER, drm_dropmaster_ioctl, DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_CTX, drm_addctx, DRM_AUTH|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_CTX, drm_rmctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_MOD_CTX, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_GET_CTX, drm_getctx, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_SWITCH_CTX, drm_switchctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_NEW_CTX, drm_newctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RES_CTX, drm_resctx, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_RM_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_LOCK, drm_lock, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_UNLOCK, drm_unlock, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_FINISH, drm_noop, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_ADD_BUFS, drm_addbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_MARK_BUFS, drm_markbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_INFO_BUFS, drm_infobufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_MAP_BUFS, drm_mapbufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_FREE_BUFS, drm_freebufs, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_DMA, drm_dma_ioctl, DRM_AUTH),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_CONTROL, drm_control, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-#if __OS_HAS_AGP
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ACQUIRE, drm_agp_acquire_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_RELEASE, drm_agp_release_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ENABLE, drm_agp_enable_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_INFO, drm_agp_info_ioctl, DRM_AUTH),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ALLOC, drm_agp_alloc_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_FREE, drm_agp_free_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_BIND, drm_agp_bind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_AGP_UNBIND, drm_agp_unbind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-#endif
-
-	DRM_IOCTL_DEF(DRM_IOCTL_SG_ALLOC, drm_sg_alloc, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF(DRM_IOCTL_SG_FREE, drm_sg_free, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_WAIT_VBLANK, drm_wait_vblank, DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODESET_CTL, drm_modeset_ctl, 0),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_UPDATE_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
-	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
-
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-};
+MODULE_AUTHOR(CORE_AUTHOR);
+MODULE_DESCRIPTION(CORE_DESC);
+MODULE_LICENSE("GPL and additional rights");
+MODULE_PARM_DESC(debug, "Enable debug output");
+MODULE_PARM_DESC(vblankoffdelay, "Delay until vblank irq auto-disable [msecs] (0: never disable, <0: disable immediately)");
+MODULE_PARM_DESC(timestamp_precision_usec, "Max. error on timestamps [usecs]");
+MODULE_PARM_DESC(timestamp_monotonic, "Use monotonic timestamps");
 
-#define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
+module_param_named(debug, drm_debug, int, 0600);
 
-/** File operations structure */
-static const struct file_operations drm_stub_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_stub_open,
-	.llseek = noop_llseek,
-};
+static DEFINE_SPINLOCK(drm_minor_lock);
+static struct idr drm_minors_idr;
 
-static int __init drm_core_init(void)
+struct class *drm_class;
+static struct dentry *drm_debugfs_root;
+
+void drm_err(const char *func, const char *format, ...)
 {
-	int ret = -ENOMEM;
+	struct va_format vaf;
+	va_list args;
 
-	drm_global_init();
-	drm_connector_ida_init();
-	idr_init(&drm_minors_idr);
+	va_start(args, format);
 
-	if (register_chrdev(DRM_MAJOR, "drm", &drm_stub_fops))
-		goto err_p1;
+	vaf.fmt = format;
+	vaf.va = &args;
 
-	drm_class = drm_sysfs_create(THIS_MODULE, "drm");
-	if (IS_ERR(drm_class)) {
-		printk(KERN_ERR "DRM: Error creating drm class.\n");
-		ret = PTR_ERR(drm_class);
-		goto err_p2;
+	printk(KERN_ERR "[" DRM_NAME ":%s] *ERROR* %pV", func, &vaf);
+
+	va_end(args);
+}
+EXPORT_SYMBOL(drm_err);
+
+void drm_ut_debug_printk(const char *function_name, const char *format, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, format);
+	vaf.fmt = format;
+	vaf.va = &args;
+
+	printk(KERN_DEBUG "[" DRM_NAME ":%s] %pV", function_name, &vaf);
+
+	va_end(args);
+}
+EXPORT_SYMBOL(drm_ut_debug_printk);
+
+#define DRM_MAGIC_HASH_ORDER  4  /**< Size of key hash table. Must be power of 2. */
+
+struct drm_master *drm_master_create(struct drm_minor *minor)
+{
+	struct drm_master *master;
+
+	master = kzalloc(sizeof(*master), GFP_KERNEL);
+	if (!master)
+		return NULL;
+
+	kref_init(&master->refcount);
+	spin_lock_init(&master->lock.spinlock);
+	init_waitqueue_head(&master->lock.lock_queue);
+	if (drm_ht_create(&master->magiclist, DRM_MAGIC_HASH_ORDER)) {
+		kfree(master);
+		return NULL;
 	}
+	INIT_LIST_HEAD(&master->magicfree);
+	master->minor = minor;
 
-	drm_debugfs_root = debugfs_create_dir("dri", NULL);
-	if (!drm_debugfs_root) {
-		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
-		ret = -1;
-		goto err_p3;
+	return master;
+}
+
+struct drm_master *drm_master_get(struct drm_master *master)
+{
+	kref_get(&master->refcount);
+	return master;
+}
+EXPORT_SYMBOL(drm_master_get);
+
+static void drm_master_destroy(struct kref *kref)
+{
+	struct drm_master *master = container_of(kref, struct drm_master, refcount);
+	struct drm_device *dev = master->minor->dev;
+	struct drm_map_list *r_list, *list_temp;
+
+	mutex_lock(&dev->struct_mutex);
+	if (dev->driver->master_destroy)
+		dev->driver->master_destroy(dev, master);
+
+	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head) {
+		if (r_list->master == master) {
+			drm_legacy_rmmap_locked(dev, r_list->map);
+			r_list = NULL;
+		}
 	}
 
-	DRM_INFO("Initialized %s %d.%d.%d %s\n",
-		 CORE_NAME, CORE_MAJOR, CORE_MINOR, CORE_PATCHLEVEL, CORE_DATE);
-	return 0;
-err_p3:
-	drm_sysfs_destroy();
-err_p2:
-	unregister_chrdev(DRM_MAJOR, "drm");
+	if (master->unique) {
+		kfree(master->unique);
+		master->unique = NULL;
+		master->unique_len = 0;
+	}
 
-	idr_destroy(&drm_minors_idr);
-err_p1:
-	return ret;
+	drm_ht_remove(&master->magiclist);
+
+	mutex_unlock(&dev->struct_mutex);
+	kfree(master);
 }
 
-static void __exit drm_core_exit(void)
+void drm_master_put(struct drm_master **master)
 {
-	debugfs_remove(drm_debugfs_root);
-	drm_sysfs_destroy();
+	kref_put(&(*master)->refcount, drm_master_destroy);
+	*master = NULL;
+}
+EXPORT_SYMBOL(drm_master_put);
 
-	unregister_chrdev(DRM_MAJOR, "drm");
+int drm_setmaster_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
+{
+	int ret = 0;
 
-	drm_connector_ida_destroy();
-	idr_destroy(&drm_minors_idr);
+	mutex_lock(&dev->master_mutex);
+	if (file_priv->is_master)
+		goto out_unlock;
+
+	if (file_priv->minor->master) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	if (!file_priv->master) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	file_priv->minor->master = drm_master_get(file_priv->master);
+	file_priv->is_master = 1;
+	if (dev->driver->master_set) {
+		ret = dev->driver->master_set(dev, file_priv, false);
+		if (unlikely(ret != 0)) {
+			file_priv->is_master = 0;
+			drm_master_put(&file_priv->minor->master);
+		}
+	}
+
+out_unlock:
+	mutex_unlock(&dev->master_mutex);
+	return ret;
 }
 
-module_init(drm_core_init);
-module_exit(drm_core_exit);
+int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv)
+{
+	int ret = -EINVAL;
 
-/**
- * Copy and IOCTL return string to user space
+	mutex_lock(&dev->master_mutex);
+	if (!file_priv->is_master)
+		goto out_unlock;
+
+	if (!file_priv->minor->master)
+		goto out_unlock;
+
+	ret = 0;
+	if (dev->driver->master_drop)
+		dev->driver->master_drop(dev, file_priv, false);
+	drm_master_put(&file_priv->minor->master);
+	file_priv->is_master = 0;
+
+out_unlock:
+	mutex_unlock(&dev->master_mutex);
+	return ret;
+}
+
+/*
+ * DRM Minors
+ * A DRM device can provide several char-dev interfaces on the DRM-Major. Each
+ * of them is represented by a drm_minor object. Depending on the capabilities
+ * of the device-driver, different interfaces are registered.
+ *
+ * Minors can be accessed via dev->$minor_name. This pointer is either
+ * NULL or a valid drm_minor pointer and stays valid as long as the device is
+ * valid. This means, DRM minors have the same life-time as the underlying
+ * device. However, this doesn't mean that the minor is active. Minors are
+ * registered and unregistered dynamically according to device-state.
  */
-static int drm_copy_field(char *buf, size_t *buf_len, const char *value)
+
+static struct drm_minor **drm_minor_get_slot(struct drm_device *dev,
+					     unsigned int type)
+{
+	switch (type) {
+	case DRM_MINOR_LEGACY:
+		return &dev->primary;
+	case DRM_MINOR_RENDER:
+		return &dev->render;
+	case DRM_MINOR_CONTROL:
+		return &dev->control;
+	default:
+		return NULL;
+	}
+}
+
+static int drm_minor_alloc(struct drm_device *dev, unsigned int type)
+{
+	struct drm_minor *minor;
+	unsigned long flags;
+	int r;
+
+	minor = kzalloc(sizeof(*minor), GFP_KERNEL);
+	if (!minor)
+		return -ENOMEM;
+
+	minor->type = type;
+	minor->dev = dev;
+
+	idr_preload(GFP_KERNEL);
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	r = idr_alloc(&drm_minors_idr,
+		      NULL,
+		      64 * type,
+		      64 * (type + 1),
+		      GFP_NOWAIT);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+	idr_preload_end();
+
+	if (r < 0)
+		goto err_free;
+
+	minor->index = r;
+
+	minor->kdev = drm_sysfs_minor_alloc(minor);
+	if (IS_ERR(minor->kdev)) {
+		r = PTR_ERR(minor->kdev);
+		goto err_index;
+	}
+
+	*drm_minor_get_slot(dev, type) = minor;
+	return 0;
+
+err_index:
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	idr_remove(&drm_minors_idr, minor->index);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+err_free:
+	kfree(minor);
+	return r;
+}
+
+static void drm_minor_free(struct drm_device *dev, unsigned int type)
+{
+	struct drm_minor **slot, *minor;
+	unsigned long flags;
+
+	slot = drm_minor_get_slot(dev, type);
+	minor = *slot;
+	if (!minor)
+		return;
+
+	drm_mode_group_destroy(&minor->mode_group);
+	put_device(minor->kdev);
+
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	idr_remove(&drm_minors_idr, minor->index);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+
+	kfree(minor);
+	*slot = NULL;
+}
+
+static int drm_minor_register(struct drm_device *dev, unsigned int type)
 {
-	int len;
+	struct drm_minor *minor;
+	unsigned long flags;
+	int ret;
 
-	/* don't overflow userbuf */
-	len = strlen(value);
-	if (len > *buf_len)
-		len = *buf_len;
+	DRM_DEBUG("\n");
 
-	/* let userspace know exact length of driver value (which could be
-	 * larger than the userspace-supplied buffer) */
-	*buf_len = strlen(value);
+	minor = *drm_minor_get_slot(dev, type);
+	if (!minor)
+		return 0;
 
-	/* finally, try filling in the userbuf */
-	if (len && buf)
-		if (copy_to_user(buf, value, len))
-			return -EFAULT;
+	ret = drm_debugfs_init(minor, minor->index, drm_debugfs_root);
+	if (ret) {
+		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
+		return ret;
+	}
+
+	ret = device_add(minor->kdev);
+	if (ret)
+		goto err_debugfs;
+
+	/* replace NULL with @minor so lookups will succeed from now on */
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	idr_replace(&drm_minors_idr, minor, minor->index);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+
+	DRM_DEBUG("new minor registered %d\n", minor->index);
 	return 0;
+
+err_debugfs:
+	drm_debugfs_cleanup(minor);
+	return ret;
+}
+
+static void drm_minor_unregister(struct drm_device *dev, unsigned int type)
+{
+	struct drm_minor *minor;
+	unsigned long flags;
+
+	minor = *drm_minor_get_slot(dev, type);
+	if (!minor || !device_is_registered(minor->kdev))
+		return;
+
+	/* replace @minor with NULL so lookups will fail from now on */
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	idr_replace(&drm_minors_idr, NULL, minor->index);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+
+	device_del(minor->kdev);
+	dev_set_drvdata(minor->kdev, NULL); /* safety belt */
+	drm_debugfs_cleanup(minor);
 }
 
 /**
- * Get version information
+ * drm_minor_acquire - Acquire a DRM minor
+ * @minor_id: Minor ID of the DRM-minor
+ *
+ * Looks up the given minor-ID and returns the respective DRM-minor object. The
+ * refence-count of the underlying device is increased so you must release this
+ * object with drm_minor_release().
  *
- * \param inode device inode.
- * \param filp file pointer.
- * \param cmd command.
- * \param arg user argument, pointing to a drm_version structure.
- * \return zero on success or negative number on failure.
+ * As long as you hold this minor, it is guaranteed that the object and the
+ * minor->dev pointer will stay valid! However, the device may get unplugged and
+ * unregistered while you hold the minor.
  *
- * Fills in the version information in \p arg.
+ * Returns:
+ * Pointer to minor-object with increased device-refcount, or PTR_ERR on
+ * failure.
  */
-static int drm_version(struct drm_device *dev, void *data,
-		       struct drm_file *file_priv)
+struct drm_minor *drm_minor_acquire(unsigned int minor_id)
 {
-	struct drm_version *version = data;
-	int err;
+	struct drm_minor *minor;
+	unsigned long flags;
+
+	spin_lock_irqsave(&drm_minor_lock, flags);
+	minor = idr_find(&drm_minors_idr, minor_id);
+	if (minor)
+		drm_dev_ref(minor->dev);
+	spin_unlock_irqrestore(&drm_minor_lock, flags);
+
+	if (!minor) {
+		return ERR_PTR(-ENODEV);
+	} else if (drm_device_is_unplugged(minor->dev)) {
+		drm_dev_unref(minor->dev);
+		return ERR_PTR(-ENODEV);
+	}
 
-	version->version_major = dev->driver->major;
-	version->version_minor = dev->driver->minor;
-	version->version_patchlevel = dev->driver->patchlevel;
-	err = drm_copy_field(version->name, &version->name_len,
-			dev->driver->name);
-	if (!err)
-		err = drm_copy_field(version->date, &version->date_len,
-				dev->driver->date);
-	if (!err)
-		err = drm_copy_field(version->desc, &version->desc_len,
-				dev->driver->desc);
+	return minor;
+}
 
-	return err;
+/**
+ * drm_minor_release - Release DRM minor
+ * @minor: Pointer to DRM minor object
+ *
+ * Release a minor that was previously acquired via drm_minor_acquire().
+ */
+void drm_minor_release(struct drm_minor *minor)
+{
+	drm_dev_unref(minor->dev);
 }
 
 /**
- * drm_ioctl_permit - Check ioctl permissions against caller
+ * drm_put_dev - Unregister and release a DRM device
+ * @dev: DRM device
+ *
+ * Called at module unload time or when a PCI device is unplugged.
  *
- * @flags: ioctl permission flags.
- * @file_priv: Pointer to struct drm_file identifying the caller.
+ * Use of this function is discouraged. It will eventually go away completely.
+ * Please use drm_dev_unregister() and drm_dev_unref() explicitly instead.
  *
- * Checks whether the caller is allowed to run an ioctl with the
- * indicated permissions. If so, returns zero. Otherwise returns an
- * error code suitable for ioctl return.
+ * Cleans up all DRM device, calling drm_lastclose().
  */
-static int drm_ioctl_permit(u32 flags, struct drm_file *file_priv)
+void drm_put_dev(struct drm_device *dev)
 {
-	/* ROOT_ONLY is only for CAP_SYS_ADMIN */
-	if (unlikely((flags & DRM_ROOT_ONLY) && !capable(CAP_SYS_ADMIN)))
-		return -EACCES;
-
-	/* AUTH is only for authenticated or render client */
-	if (unlikely((flags & DRM_AUTH) && !drm_is_render_client(file_priv) &&
-		     !file_priv->authenticated))
-		return -EACCES;
-
-	/* MASTER is only for master or control clients */
-	if (unlikely((flags & DRM_MASTER) && !file_priv->is_master &&
-		     !drm_is_control_client(file_priv)))
-		return -EACCES;
-
-	/* Control clients must be explicitly allowed */
-	if (unlikely(!(flags & DRM_CONTROL_ALLOW) &&
-		     drm_is_control_client(file_priv)))
-		return -EACCES;
-
-	/* Render clients must be explicitly allowed */
-	if (unlikely(!(flags & DRM_RENDER_ALLOW) &&
-		     drm_is_render_client(file_priv)))
-		return -EACCES;
+	DRM_DEBUG("\n");
 
-	return 0;
+	if (!dev) {
+		DRM_ERROR("cleanup called no dev\n");
+		return;
+	}
+
+	drm_dev_unregister(dev);
+	drm_dev_unref(dev);
+}
+EXPORT_SYMBOL(drm_put_dev);
+
+void drm_unplug_dev(struct drm_device *dev)
+{
+	/* for a USB device */
+	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
+	drm_minor_unregister(dev, DRM_MINOR_RENDER);
+	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
+
+	mutex_lock(&drm_global_mutex);
+
+	drm_device_set_unplugged(dev);
+
+	if (dev->open_count == 0) {
+		drm_put_dev(dev);
+	}
+	mutex_unlock(&drm_global_mutex);
+}
+EXPORT_SYMBOL(drm_unplug_dev);
+
+/*
+ * DRM internal mount
+ * We want to be able to allocate our own "struct address_space" to control
+ * memory-mappings in VRAM (or stolen RAM, ...). However, core MM does not allow
+ * stand-alone address_space objects, so we need an underlying inode. As there
+ * is no way to allocate an independent inode easily, we need a fake internal
+ * VFS mount-point.
+ *
+ * The drm_fs_inode_new() function allocates a new inode, drm_fs_inode_free()
+ * frees it again. You are allowed to use iget() and iput() to get references to
+ * the inode. But each drm_fs_inode_new() call must be paired with exactly one
+ * drm_fs_inode_free() call (which does not have to be the last iput()).
+ * We use drm_fs_inode_*() to manage our internal VFS mount-point and share it
+ * between multiple inode-users. You could, technically, call
+ * iget() + drm_fs_inode_free() directly after alloc and sometime later do an
+ * iput(), but this way you'd end up with a new vfsmount for each inode.
+ */
+
+static int drm_fs_cnt;
+static struct vfsmount *drm_fs_mnt;
+
+static const struct dentry_operations drm_fs_dops = {
+	.d_dname	= simple_dname,
+};
+
+static const struct super_operations drm_fs_sops = {
+	.statfs		= simple_statfs,
+};
+
+static struct dentry *drm_fs_mount(struct file_system_type *fs_type, int flags,
+				   const char *dev_name, void *data)
+{
+	return mount_pseudo(fs_type,
+			    "drm:",
+			    &drm_fs_sops,
+			    &drm_fs_dops,
+			    0x010203ff);
+}
+
+static struct file_system_type drm_fs_type = {
+	.name		= "drm",
+	.owner		= THIS_MODULE,
+	.mount		= drm_fs_mount,
+	.kill_sb	= kill_anon_super,
+};
+
+static struct inode *drm_fs_inode_new(void)
+{
+	struct inode *inode;
+	int r;
+
+	r = simple_pin_fs(&drm_fs_type, &drm_fs_mnt, &drm_fs_cnt);
+	if (r < 0) {
+		DRM_ERROR("Cannot mount pseudo fs: %d\n", r);
+		return ERR_PTR(r);
+	}
+
+	inode = alloc_anon_inode(drm_fs_mnt->mnt_sb);
+	if (IS_ERR(inode))
+		simple_release_fs(&drm_fs_mnt, &drm_fs_cnt);
+
+	return inode;
+}
+
+static void drm_fs_inode_free(struct inode *inode)
+{
+	if (inode) {
+		iput(inode);
+		simple_release_fs(&drm_fs_mnt, &drm_fs_cnt);
+	}
 }
 
 /**
- * Called whenever a process performs an ioctl on /dev/drm.
+ * drm_dev_alloc - Allocate new DRM device
+ * @driver: DRM driver to allocate device for
+ * @parent: Parent device object
+ *
+ * Allocate and initialize a new DRM device. No device registration is done.
+ * Call drm_dev_register() to advertice the device to user space and register it
+ * with other core subsystems.
  *
- * \param inode device inode.
- * \param file_priv DRM file private.
- * \param cmd command.
- * \param arg user argument.
- * \return zero on success or negative number on failure.
+ * The initial ref-count of the object is 1. Use drm_dev_ref() and
+ * drm_dev_unref() to take and drop further ref-counts.
  *
- * Looks up the ioctl function in the ::ioctls table, checking for root
- * previleges if so required, and dispatches to the respective function.
+ * RETURNS:
+ * Pointer to new DRM device, or NULL if out of memory.
  */
-long drm_ioctl(struct file *filp,
-	      unsigned int cmd, unsigned long arg)
+struct drm_device *drm_dev_alloc(struct drm_driver *driver,
+				 struct device *parent)
 {
-	struct drm_file *file_priv = filp->private_data;
 	struct drm_device *dev;
-	const struct drm_ioctl_desc *ioctl = NULL;
-	drm_ioctl_t *func;
-	unsigned int nr = DRM_IOCTL_NR(cmd);
-	int retcode = -EINVAL;
-	char stack_kdata[128];
-	char *kdata = NULL;
-	unsigned int usize, asize;
-
-	dev = file_priv->minor->dev;
-
-	if (drm_device_is_unplugged(dev))
-		return -ENODEV;
-
-	if ((nr >= DRM_CORE_IOCTL_COUNT) &&
-	    ((nr < DRM_COMMAND_BASE) || (nr >= DRM_COMMAND_END)))
-		goto err_i1;
-	if ((nr >= DRM_COMMAND_BASE) && (nr < DRM_COMMAND_END) &&
-	    (nr < DRM_COMMAND_BASE + dev->driver->num_ioctls)) {
-		u32 drv_size;
-		ioctl = &dev->driver->ioctls[nr - DRM_COMMAND_BASE];
-		drv_size = _IOC_SIZE(ioctl->cmd_drv);
-		usize = asize = _IOC_SIZE(cmd);
-		if (drv_size > asize)
-			asize = drv_size;
-		cmd = ioctl->cmd_drv;
+	int ret;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return NULL;
+
+	kref_init(&dev->ref);
+	dev->dev = parent;
+	dev->driver = driver;
+
+	INIT_LIST_HEAD(&dev->filelist);
+	INIT_LIST_HEAD(&dev->ctxlist);
+	INIT_LIST_HEAD(&dev->vmalist);
+	INIT_LIST_HEAD(&dev->maplist);
+	INIT_LIST_HEAD(&dev->vblank_event_list);
+
+	spin_lock_init(&dev->buf_lock);
+	spin_lock_init(&dev->event_lock);
+	mutex_init(&dev->struct_mutex);
+	mutex_init(&dev->ctxlist_mutex);
+	mutex_init(&dev->master_mutex);
+
+	dev->anon_inode = drm_fs_inode_new();
+	if (IS_ERR(dev->anon_inode)) {
+		ret = PTR_ERR(dev->anon_inode);
+		DRM_ERROR("Cannot allocate anonymous inode: %d\n", ret);
+		goto err_free;
 	}
-	else if ((nr >= DRM_COMMAND_END) || (nr < DRM_COMMAND_BASE)) {
-		u32 drv_size;
-
-		ioctl = &drm_ioctls[nr];
 
-		drv_size = _IOC_SIZE(ioctl->cmd);
-		usize = asize = _IOC_SIZE(cmd);
-		if (drv_size > asize)
-			asize = drv_size;
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		ret = drm_minor_alloc(dev, DRM_MINOR_CONTROL);
+		if (ret)
+			goto err_minors;
+	}
 
-		cmd = ioctl->cmd;
-	} else
-		goto err_i1;
+	if (drm_core_check_feature(dev, DRIVER_RENDER)) {
+		ret = drm_minor_alloc(dev, DRM_MINOR_RENDER);
+		if (ret)
+			goto err_minors;
+	}
 
-	DRM_DEBUG("pid=%d, dev=0x%lx, auth=%d, %s\n",
-		  task_pid_nr(current),
-		  (long)old_encode_dev(file_priv->minor->kdev->devt),
-		  file_priv->authenticated, ioctl->name);
+	ret = drm_minor_alloc(dev, DRM_MINOR_LEGACY);
+	if (ret)
+		goto err_minors;
 
-	/* Do not trust userspace, use our own definition */
-	func = ioctl->func;
+	if (drm_ht_create(&dev->map_hash, 12))
+		goto err_minors;
 
-	if (unlikely(!func)) {
-		DRM_DEBUG("no function\n");
-		retcode = -EINVAL;
-		goto err_i1;
+	ret = drm_legacy_ctxbitmap_init(dev);
+	if (ret) {
+		DRM_ERROR("Cannot allocate memory for context bitmap.\n");
+		goto err_ht;
 	}
 
-	retcode = drm_ioctl_permit(ioctl->flags, file_priv);
-	if (unlikely(retcode))
-		goto err_i1;
-
-	if (cmd & (IOC_IN | IOC_OUT)) {
-		if (asize <= sizeof(stack_kdata)) {
-			kdata = stack_kdata;
-		} else {
-			kdata = kmalloc(asize, GFP_KERNEL);
-			if (!kdata) {
-				retcode = -ENOMEM;
-				goto err_i1;
-			}
+	if (drm_core_check_feature(dev, DRIVER_GEM)) {
+		ret = drm_gem_init(dev);
+		if (ret) {
+			DRM_ERROR("Cannot initialize graphics execution manager (GEM)\n");
+			goto err_ctxbitmap;
 		}
-		if (asize > usize)
-			memset(kdata + usize, 0, asize - usize);
 	}
 
-	if (cmd & IOC_IN) {
-		if (copy_from_user(kdata, (void __user *)arg,
-				   usize) != 0) {
-			retcode = -EFAULT;
-			goto err_i1;
-		}
-	} else if (cmd & IOC_OUT) {
-		memset(kdata, 0, usize);
-	}
+	return dev;
+
+err_ctxbitmap:
+	drm_legacy_ctxbitmap_cleanup(dev);
+err_ht:
+	drm_ht_remove(&dev->map_hash);
+err_minors:
+	drm_minor_free(dev, DRM_MINOR_LEGACY);
+	drm_minor_free(dev, DRM_MINOR_RENDER);
+	drm_minor_free(dev, DRM_MINOR_CONTROL);
+	drm_fs_inode_free(dev->anon_inode);
+err_free:
+	mutex_destroy(&dev->master_mutex);
+	kfree(dev);
+	return NULL;
+}
+EXPORT_SYMBOL(drm_dev_alloc);
+
+static void drm_dev_release(struct kref *ref)
+{
+	struct drm_device *dev = container_of(ref, struct drm_device, ref);
+
+	if (drm_core_check_feature(dev, DRIVER_GEM))
+		drm_gem_destroy(dev);
+
+	drm_legacy_ctxbitmap_cleanup(dev);
+	drm_ht_remove(&dev->map_hash);
+	drm_fs_inode_free(dev->anon_inode);
+
+	drm_minor_free(dev, DRM_MINOR_LEGACY);
+	drm_minor_free(dev, DRM_MINOR_RENDER);
+	drm_minor_free(dev, DRM_MINOR_CONTROL);
+
+	mutex_destroy(&dev->master_mutex);
+	kfree(dev->unique);
+	kfree(dev);
+}
+
+/**
+ * drm_dev_ref - Take reference of a DRM device
+ * @dev: device to take reference of or NULL
+ *
+ * This increases the ref-count of @dev by one. You *must* already own a
+ * reference when calling this. Use drm_dev_unref() to drop this reference
+ * again.
+ *
+ * This function never fails. However, this function does not provide *any*
+ * guarantee whether the device is alive or running. It only provides a
+ * reference to the object and the memory associated with it.
+ */
+void drm_dev_ref(struct drm_device *dev)
+{
+	if (dev)
+		kref_get(&dev->ref);
+}
+EXPORT_SYMBOL(drm_dev_ref);
+
+/**
+ * drm_dev_unref - Drop reference of a DRM device
+ * @dev: device to drop reference of or NULL
+ *
+ * This decreases the ref-count of @dev by one. The device is destroyed if the
+ * ref-count drops to zero.
+ */
+void drm_dev_unref(struct drm_device *dev)
+{
+	if (dev)
+		kref_put(&dev->ref, drm_dev_release);
+}
+EXPORT_SYMBOL(drm_dev_unref);
+
+/**
+ * drm_dev_register - Register DRM device
+ * @dev: Device to register
+ * @flags: Flags passed to the driver's .load() function
+ *
+ * Register the DRM device @dev with the system, advertise device to user-space
+ * and start normal device operation. @dev must be allocated via drm_dev_alloc()
+ * previously.
+ *
+ * Never call this twice on any device!
+ *
+ * RETURNS:
+ * 0 on success, negative error code on failure.
+ */
+int drm_dev_register(struct drm_device *dev, unsigned long flags)
+{
+	int ret;
+
+	mutex_lock(&drm_global_mutex);
+
+	ret = drm_minor_register(dev, DRM_MINOR_CONTROL);
+	if (ret)
+		goto err_minors;
 
-	if (ioctl->flags & DRM_UNLOCKED)
-		retcode = func(dev, kdata, file_priv);
-	else {
-		mutex_lock(&drm_global_mutex);
-		retcode = func(dev, kdata, file_priv);
-		mutex_unlock(&drm_global_mutex);
+	ret = drm_minor_register(dev, DRM_MINOR_RENDER);
+	if (ret)
+		goto err_minors;
+
+	ret = drm_minor_register(dev, DRM_MINOR_LEGACY);
+	if (ret)
+		goto err_minors;
+
+	if (dev->driver->load) {
+		ret = dev->driver->load(dev, flags);
+		if (ret)
+			goto err_minors;
 	}
 
-	if (cmd & IOC_OUT) {
-		if (copy_to_user((void __user *)arg, kdata,
-				 usize) != 0)
-			retcode = -EFAULT;
+	/* setup grouping for legacy outputs */
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		ret = drm_mode_group_init_legacy_group(dev,
+				&dev->primary->mode_group);
+		if (ret)
+			goto err_unload;
 	}
 
-      err_i1:
-	if (!ioctl)
-		DRM_DEBUG("invalid ioctl: pid=%d, dev=0x%lx, auth=%d, cmd=0x%02x, nr=0x%02x\n",
-			  task_pid_nr(current),
-			  (long)old_encode_dev(file_priv->minor->kdev->devt),
-			  file_priv->authenticated, cmd, nr);
-
-	if (kdata != stack_kdata)
-		kfree(kdata);
-	if (retcode)
-		DRM_DEBUG("ret = %d\n", retcode);
-	return retcode;
+	ret = 0;
+	goto out_unlock;
+
+err_unload:
+	if (dev->driver->unload)
+		dev->driver->unload(dev);
+err_minors:
+	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
+	drm_minor_unregister(dev, DRM_MINOR_RENDER);
+	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
+out_unlock:
+	mutex_unlock(&drm_global_mutex);
+	return ret;
+}
+EXPORT_SYMBOL(drm_dev_register);
+
+/**
+ * drm_dev_unregister - Unregister DRM device
+ * @dev: Device to unregister
+ *
+ * Unregister the DRM device from the system. This does the reverse of
+ * drm_dev_register() but does not deallocate the device. The caller must call
+ * drm_dev_unref() to drop their final reference.
+ */
+void drm_dev_unregister(struct drm_device *dev)
+{
+	struct drm_map_list *r_list, *list_temp;
+
+	drm_lastclose(dev);
+
+	if (dev->driver->unload)
+		dev->driver->unload(dev);
+
+	if (dev->agp)
+		drm_pci_agp_destroy(dev);
+
+	drm_vblank_cleanup(dev);
+
+	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head)
+		drm_legacy_rmmap(dev, r_list->map);
+
+	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
+	drm_minor_unregister(dev, DRM_MINOR_RENDER);
+	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
 }
-EXPORT_SYMBOL(drm_ioctl);
+EXPORT_SYMBOL(drm_dev_unregister);
 
 /**
- * drm_ioctl_flags - Check for core ioctl and return ioctl permission flags
+ * drm_dev_set_unique - Set the unique name of a DRM device
+ * @dev: device of which to set the unique name
+ * @fmt: format string for unique name
+ *
+ * Sets the unique name of a DRM device using the specified format string and
+ * a variable list of arguments. Drivers can use this at driver probe time if
+ * the unique name of the devices they drive is static.
  *
- * @nr: Ioctl number.
- * @flags: Where to return the ioctl permission flags
+ * Return: 0 on success or a negative error code on failure.
  */
-bool drm_ioctl_flags(unsigned int nr, unsigned int *flags)
+int drm_dev_set_unique(struct drm_device *dev, const char *fmt, ...)
 {
-	if ((nr >= DRM_COMMAND_END && nr < DRM_CORE_IOCTL_COUNT) ||
-	    (nr < DRM_COMMAND_BASE)) {
-		*flags = drm_ioctls[nr].flags;
-		return true;
+	va_list ap;
+
+	kfree(dev->unique);
+
+	va_start(ap, fmt);
+	dev->unique = kvasprintf(GFP_KERNEL, fmt, ap);
+	va_end(ap);
+
+	return dev->unique ? 0 : -ENOMEM;
+}
+EXPORT_SYMBOL(drm_dev_set_unique);
+
+/*
+ * DRM Core
+ * The DRM core module initializes all global DRM objects and makes them
+ * available to drivers. Once setup, drivers can probe their respective
+ * devices.
+ * Currently, core management includes:
+ *  - The "DRM-Global" key/value database
+ *  - Global ID management for connectors
+ *  - DRM major number allocation
+ *  - DRM minor management
+ *  - DRM sysfs class
+ *  - DRM debugfs root
+ *
+ * Furthermore, the DRM core provides dynamic char-dev lookups. For each
+ * interface registered on a DRM device, you can request minor numbers from DRM
+ * core. DRM core takes care of major-number management and char-dev
+ * registration. A stub ->open() callback forwards any open() requests to the
+ * registered minor.
+ */
+
+static int drm_stub_open(struct inode *inode, struct file *filp)
+{
+	const struct file_operations *new_fops;
+	struct drm_minor *minor;
+	int err;
+
+	DRM_DEBUG("\n");
+
+	mutex_lock(&drm_global_mutex);
+	minor = drm_minor_acquire(iminor(inode));
+	if (IS_ERR(minor)) {
+		err = PTR_ERR(minor);
+		goto out_unlock;
+	}
+
+	new_fops = fops_get(minor->dev->driver->fops);
+	if (!new_fops) {
+		err = -ENODEV;
+		goto out_release;
 	}
 
-	return false;
+	replace_fops(filp, new_fops);
+	if (filp->f_op->open)
+		err = filp->f_op->open(inode, filp);
+	else
+		err = 0;
+
+out_release:
+	drm_minor_release(minor);
+out_unlock:
+	mutex_unlock(&drm_global_mutex);
+	return err;
 }
-EXPORT_SYMBOL(drm_ioctl_flags);
+
+static const struct file_operations drm_stub_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_stub_open,
+	.llseek = noop_llseek,
+};
+
+static int __init drm_core_init(void)
+{
+	int ret = -ENOMEM;
+
+	drm_global_init();
+	drm_connector_ida_init();
+	idr_init(&drm_minors_idr);
+
+	if (register_chrdev(DRM_MAJOR, "drm", &drm_stub_fops))
+		goto err_p1;
+
+	drm_class = drm_sysfs_create(THIS_MODULE, "drm");
+	if (IS_ERR(drm_class)) {
+		printk(KERN_ERR "DRM: Error creating drm class.\n");
+		ret = PTR_ERR(drm_class);
+		goto err_p2;
+	}
+
+	drm_debugfs_root = debugfs_create_dir("dri", NULL);
+	if (!drm_debugfs_root) {
+		DRM_ERROR("Cannot create /sys/kernel/debug/dri\n");
+		ret = -1;
+		goto err_p3;
+	}
+
+	DRM_INFO("Initialized %s %d.%d.%d %s\n",
+		 CORE_NAME, CORE_MAJOR, CORE_MINOR, CORE_PATCHLEVEL, CORE_DATE);
+	return 0;
+err_p3:
+	drm_sysfs_destroy();
+err_p2:
+	unregister_chrdev(DRM_MAJOR, "drm");
+
+	idr_destroy(&drm_minors_idr);
+err_p1:
+	return ret;
+}
+
+static void __exit drm_core_exit(void)
+{
+	debugfs_remove(drm_debugfs_root);
+	drm_sysfs_destroy();
+
+	unregister_chrdev(DRM_MAJOR, "drm");
+
+	drm_connector_ida_destroy();
+	idr_destroy(&drm_minors_idr);
+}
+
+module_init(drm_core_init);
+module_exit(drm_core_exit);
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index dfa9769b26b5..3bf999134bcc 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -632,27 +632,27 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
 			DRM_MODE_FLAG_INTERLACE),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 6 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 6 - 720(1440)x480i@60Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 7 - 1440x480i@60Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 7 - 720(1440)x480i@60Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 8 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
+	/* 8 - 720(1440)x240@60Hz */
+	{ DRM_MODE("720x240", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 9 - 1440x240@60Hz */
-	{ DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478,
-		   1602, 1716, 0, 240, 244, 247, 262, 0,
+	/* 9 - 720(1440)x240@60Hz */
+	{ DRM_MODE("720x240", DRM_MODE_TYPE_DRIVER, 13500, 720, 739,
+		   801, 858, 0, 240, 244, 247, 262, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 60, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -714,27 +714,27 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC |
 			DRM_MODE_FLAG_INTERLACE),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 21 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 21 - 720(1440)x576i@50Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 22 - 1440x576i@50Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 22 - 720(1440)x576i@50Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 23 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
+	/* 23 - 720(1440)x288@50Hz */
+	{ DRM_MODE("720x288", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 24 - 1440x288@50Hz */
-	{ DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464,
-		   1590, 1728, 0, 288, 290, 293, 312, 0,
+	/* 24 - 720(1440)x288@50Hz */
+	{ DRM_MODE("720x288", DRM_MODE_TYPE_DRIVER, 13500, 720, 732,
+		   795, 864, 0, 288, 290, 293, 312, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 50, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -837,17 +837,17 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   796, 864, 0, 576, 581, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 44 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 44 - 720(1440)x576i@100Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK),
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 45 - 1440x576i@100Hz */
-	{ DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 45 - 720(1440)x576i@100Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 27000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
-			DRM_MODE_FLAG_DBLCLK),
+			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 100, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
 	/* 46 - 1920x1080i@120Hz */
 	{ DRM_MODE("1920x1080i", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008,
@@ -870,15 +870,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   798, 858, 0, 480, 489, 495, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 50 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 50 - 720(1440)x480i@120Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 27000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 51 - 1440x480i@120Hz */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 51 - 720(1440)x480i@120Hz */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 27000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 120, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -892,15 +892,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   796, 864, 0, 576, 581, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 54 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 54 - 720(1440)x576i@200Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 55 - 1440x576i@200Hz */
-	{ DRM_MODE("1440x576i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464,
-		   1590, 1728, 0, 576, 580, 586, 625, 0,
+	/* 55 - 720(1440)x576i@200Hz */
+	{ DRM_MODE("720x576i", DRM_MODE_TYPE_DRIVER, 54000, 720, 732,
+		   795, 864, 0, 576, 580, 586, 625, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 200, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -914,15 +914,15 @@ static const struct drm_display_mode edid_cea_modes[] = {
 		   798, 858, 0, 480, 489, 495, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
-	/* 58 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 58 - 720(1440)x480i@240 */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 54000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_4_3, },
-	/* 59 - 1440x480i@240 */
-	{ DRM_MODE("1440x480i", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478,
-		   1602, 1716, 0, 480, 488, 494, 525, 0,
+	/* 59 - 720(1440)x480i@240 */
+	{ DRM_MODE("720x480i", DRM_MODE_TYPE_DRIVER, 54000, 720, 739,
+		   801, 858, 0, 480, 488, 494, 525, 0,
 		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC |
 			DRM_MODE_FLAG_INTERLACE | DRM_MODE_FLAG_DBLCLK),
 	  .vrefresh = 240, .picture_aspect_ratio = HDMI_PICTURE_ASPECT_16_9, },
@@ -2103,7 +2103,8 @@ static int
 add_inferred_modes(struct drm_connector *connector, struct edid *edid)
 {
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	if (version_greater(edid, 1, 0))
@@ -2169,7 +2170,8 @@ add_established_modes(struct drm_connector *connector, struct edid *edid)
 		((edid->established_timings.mfg_rsvd & 0x80) << 9);
 	int i, modes = 0;
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	for (i = 0; i <= EDID_EST_TIMINGS; i++) {
@@ -2227,7 +2229,8 @@ add_standard_modes(struct drm_connector *connector, struct edid *edid)
 {
 	int i, modes = 0;
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	for (i = 0; i < EDID_STD_TIMINGS; i++) {
@@ -2313,7 +2316,8 @@ static int
 add_cvt_modes(struct drm_connector *connector, struct edid *edid)
 {	
 	struct detailed_mode_closure closure = {
-		connector, edid, 0, 0, 0
+		.connector = connector,
+		.edid = edid,
 	};
 
 	if (version_greater(edid, 1, 2))
@@ -2357,11 +2361,10 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid,
 		   u32 quirks)
 {
 	struct detailed_mode_closure closure = {
-		connector,
-		edid,
-		1,
-		quirks,
-		0
+		.connector = connector,
+		.edid = edid,
+		.preferred = 1,
+		.quirks = quirks,
 	};
 
 	if (closure.preferred && !version_greater(edid, 1, 3))
@@ -3305,6 +3308,7 @@ struct drm_connector *drm_select_eld(struct drm_encoder *encoder,
 	struct drm_device *dev = encoder->dev;
 
 	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
 		if (connector->encoder == encoder && connector->eld[0])
@@ -3432,10 +3436,10 @@ EXPORT_SYMBOL(drm_rgb_quant_range_selectable);
 /**
  * drm_assign_hdmi_deep_color_info - detect whether monitor supports
  * hdmi deep color modes and update drm_display_info if so.
- *
  * @edid: monitor EDID information
  * @info: Updated with maximum supported deep color bpc and color format
  *        if deep color supported.
+ * @connector: DRM connector, used only for debug output
  *
  * Parse the CEA extension according to CEA-861-B.
  * Return true if HDMI deep color supported, false if not or unknown.
@@ -3775,8 +3779,14 @@ drm_hdmi_avi_infoframe_from_display_mode(struct hdmi_avi_infoframe *frame,
 
 	frame->picture_aspect = HDMI_PICTURE_ASPECT_NONE;
 
-	/* Populate picture aspect ratio from CEA mode list */
-	if (frame->video_code > 0)
+	/*
+	 * Populate picture aspect ratio from either
+	 * user input (if specified) or from the CEA mode list.
+	 */
+	if (mode->picture_aspect_ratio == HDMI_PICTURE_ASPECT_4_3 ||
+		mode->picture_aspect_ratio == HDMI_PICTURE_ASPECT_16_9)
+		frame->picture_aspect = mode->picture_aspect_ratio;
+	else if (frame->video_code > 0)
 		frame->picture_aspect = drm_get_cea_aspect_ratio(
 						frame->video_code);
 
diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c
index f27c883be391..cc0ae047ed3b 100644
--- a/drivers/gpu/drm/drm_fb_cma_helper.c
+++ b/drivers/gpu/drm/drm_fb_cma_helper.c
@@ -327,7 +327,7 @@ err_drm_gem_cma_free_object:
 	return ret;
 }
 
-static struct drm_fb_helper_funcs drm_fb_cma_helper_funcs = {
+static const struct drm_fb_helper_funcs drm_fb_cma_helper_funcs = {
 	.fb_probe = drm_fbdev_cma_create,
 };
 
@@ -354,9 +354,10 @@ struct drm_fbdev_cma *drm_fbdev_cma_init(struct drm_device *dev,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	fbdev_cma->fb_helper.funcs = &drm_fb_cma_helper_funcs;
 	helper = &fbdev_cma->fb_helper;
 
+	drm_fb_helper_prepare(dev, helper, &drm_fb_cma_helper_funcs);
+
 	ret = drm_fb_helper_init(dev, helper, num_crtc, max_conn_count);
 	if (ret < 0) {
 		dev_err(dev->dev, "Failed to initialize drm fb helper.\n");
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index d5d8cea1a679..0c0c39bac23d 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -49,10 +49,11 @@ static LIST_HEAD(kernel_fb_helper_list);
  * helper functions used by many drivers to implement the kernel mode setting
  * interfaces.
  *
- * Initialization is done as a three-step process with drm_fb_helper_init(),
- * drm_fb_helper_single_add_all_connectors() and drm_fb_helper_initial_config().
- * Drivers with fancier requirements than the default behaviour can override the
- * second step with their own code.  Teardown is done with drm_fb_helper_fini().
+ * Initialization is done as a four-step process with drm_fb_helper_prepare(),
+ * drm_fb_helper_init(), drm_fb_helper_single_add_all_connectors() and
+ * drm_fb_helper_initial_config(). Drivers with fancier requirements than the
+ * default behaviour can override the third step with their own code.
+ * Teardown is done with drm_fb_helper_fini().
  *
  * At runtime drivers should restore the fbdev console by calling
  * drm_fb_helper_restore_fbdev_mode() from their ->lastclose callback. They
@@ -63,6 +64,19 @@ static LIST_HEAD(kernel_fb_helper_list);
  *
  * All other functions exported by the fb helper library can be used to
  * implement the fbdev driver interface by the driver.
+ *
+ * It is possible, though perhaps somewhat tricky, to implement race-free
+ * hotplug detection using the fbdev helpers. The drm_fb_helper_prepare()
+ * helper must be called first to initialize the minimum required to make
+ * hotplug detection work. Drivers also need to make sure to properly set up
+ * the dev->mode_config.funcs member. After calling drm_kms_helper_poll_init()
+ * it is safe to enable interrupts and start processing hotplug events. At the
+ * same time, drivers should initialize all modeset objects such as CRTCs,
+ * encoders and connectors. To finish up the fbdev helper initialization, the
+ * drm_fb_helper_init() function is called. To probe for all attached displays
+ * and set up an initial configuration using the detected hardware, drivers
+ * should call drm_fb_helper_single_add_all_connectors() followed by
+ * drm_fb_helper_initial_config().
  */
 
 /**
@@ -105,59 +119,57 @@ fail:
 }
 EXPORT_SYMBOL(drm_fb_helper_single_add_all_connectors);
 
-static int drm_fb_helper_parse_command_line(struct drm_fb_helper *fb_helper)
+int drm_fb_helper_add_one_connector(struct drm_fb_helper *fb_helper, struct drm_connector *connector)
 {
-	struct drm_fb_helper_connector *fb_helper_conn;
-	int i;
+	struct drm_fb_helper_connector **temp;
+	struct drm_fb_helper_connector *fb_helper_connector;
 
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		struct drm_cmdline_mode *mode;
-		struct drm_connector *connector;
-		char *option = NULL;
+	WARN_ON(!mutex_is_locked(&fb_helper->dev->mode_config.mutex));
+	if (fb_helper->connector_count + 1 > fb_helper->connector_info_alloc_count) {
+		temp = krealloc(fb_helper->connector_info, sizeof(struct drm_fb_helper_connector *) * (fb_helper->connector_count + 1), GFP_KERNEL);
+		if (!temp)
+			return -ENOMEM;
 
-		fb_helper_conn = fb_helper->connector_info[i];
-		connector = fb_helper_conn->connector;
-		mode = &fb_helper_conn->cmdline_mode;
+		fb_helper->connector_info_alloc_count = fb_helper->connector_count + 1;
+		fb_helper->connector_info = temp;
+	}
 
-		/* do something on return - turn off connector maybe */
-		if (fb_get_options(connector->name, &option))
-			continue;
 
-		if (drm_mode_parse_command_line_for_connector(option,
-							      connector,
-							      mode)) {
-			if (mode->force) {
-				const char *s;
-				switch (mode->force) {
-				case DRM_FORCE_OFF:
-					s = "OFF";
-					break;
-				case DRM_FORCE_ON_DIGITAL:
-					s = "ON - dig";
-					break;
-				default:
-				case DRM_FORCE_ON:
-					s = "ON";
-					break;
-				}
-
-				DRM_INFO("forcing %s connector %s\n",
-					 connector->name, s);
-				connector->force = mode->force;
-			}
+	fb_helper_connector = kzalloc(sizeof(struct drm_fb_helper_connector), GFP_KERNEL);
+	if (!fb_helper_connector)
+		return -ENOMEM;
 
-			DRM_DEBUG_KMS("cmdline mode for connector %s %dx%d@%dHz%s%s%s\n",
-				      connector->name,
-				      mode->xres, mode->yres,
-				      mode->refresh_specified ? mode->refresh : 60,
-				      mode->rb ? " reduced blanking" : "",
-				      mode->margins ? " with margins" : "",
-				      mode->interlace ?  " interlaced" : "");
-		}
+	fb_helper_connector->connector = connector;
+	fb_helper->connector_info[fb_helper->connector_count++] = fb_helper_connector;
+	return 0;
+}
+EXPORT_SYMBOL(drm_fb_helper_add_one_connector);
+
+int drm_fb_helper_remove_one_connector(struct drm_fb_helper *fb_helper,
+				       struct drm_connector *connector)
+{
+	struct drm_fb_helper_connector *fb_helper_connector;
+	int i, j;
+
+	WARN_ON(!mutex_is_locked(&fb_helper->dev->mode_config.mutex));
+
+	for (i = 0; i < fb_helper->connector_count; i++) {
+		if (fb_helper->connector_info[i]->connector == connector)
+			break;
+	}
+
+	if (i == fb_helper->connector_count)
+		return -EINVAL;
+	fb_helper_connector = fb_helper->connector_info[i];
 
+	for (j = i + 1; j < fb_helper->connector_count; j++) {
+		fb_helper->connector_info[j - 1] = fb_helper->connector_info[j];
 	}
+	fb_helper->connector_count--;
+	kfree(fb_helper_connector);
 	return 0;
 }
+EXPORT_SYMBOL(drm_fb_helper_remove_one_connector);
 
 static void drm_fb_helper_save_lut_atomic(struct drm_crtc *crtc, struct drm_fb_helper *helper)
 {
@@ -199,9 +211,6 @@ int drm_fb_helper_debug_enter(struct fb_info *info)
 	struct drm_crtc_helper_funcs *funcs;
 	int i;
 
-	if (list_empty(&kernel_fb_helper_list))
-		return false;
-
 	list_for_each_entry(helper, &kernel_fb_helper_list, kernel_fb_list) {
 		for (i = 0; i < helper->crtc_count; i++) {
 			struct drm_mode_set *mode_set =
@@ -282,10 +291,17 @@ static bool restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 
 	drm_warn_on_modeset_not_all_locked(dev);
 
-	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
 		if (plane->type != DRM_PLANE_TYPE_PRIMARY)
 			drm_plane_force_disable(plane);
 
+		if (dev->mode_config.rotation_property) {
+			drm_mode_plane_set_obj_prop(plane,
+						    dev->mode_config.rotation_property,
+						    BIT(DRM_ROTATE_0));
+		}
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
 		struct drm_crtc *crtc = mode_set->crtc;
@@ -356,11 +372,11 @@ static bool drm_fb_helper_force_kernel_mode(void)
 		if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 			continue;
 
-		/* NOTE: we use lockless flag below to avoid grabbing other
-		 * modeset locks.  So just trylock the underlying mutex
-		 * directly:
+		/*
+		 * NOTE: Use trylock mode to avoid deadlocks and sleeping in
+		 * panic context.
 		 */
-		if (!mutex_trylock(&dev->mode_config.mutex)) {
+		if (__drm_modeset_lock_all(dev, true) != 0) {
 			error = true;
 			continue;
 		}
@@ -369,7 +385,7 @@ static bool drm_fb_helper_force_kernel_mode(void)
 		if (ret)
 			error = true;
 
-		mutex_unlock(&dev->mode_config.mutex);
+		drm_modeset_unlock_all(dev);
 	}
 	return error;
 }
@@ -531,6 +547,24 @@ static void drm_fb_helper_crtc_free(struct drm_fb_helper *helper)
 }
 
 /**
+ * drm_fb_helper_prepare - setup a drm_fb_helper structure
+ * @dev: DRM device
+ * @helper: driver-allocated fbdev helper structure to set up
+ * @funcs: pointer to structure of functions associate with this helper
+ *
+ * Sets up the bare minimum to make the framebuffer helper usable. This is
+ * useful to implement race-free initialization of the polling helpers.
+ */
+void drm_fb_helper_prepare(struct drm_device *dev, struct drm_fb_helper *helper,
+			   const struct drm_fb_helper_funcs *funcs)
+{
+	INIT_LIST_HEAD(&helper->kernel_fb_list);
+	helper->funcs = funcs;
+	helper->dev = dev;
+}
+EXPORT_SYMBOL(drm_fb_helper_prepare);
+
+/**
  * drm_fb_helper_init - initialize a drm_fb_helper structure
  * @dev: drm device
  * @fb_helper: driver-allocated fbdev helper structure to initialize
@@ -542,8 +576,7 @@ static void drm_fb_helper_crtc_free(struct drm_fb_helper *helper)
  * nor register the fbdev. This is only done in drm_fb_helper_initial_config()
  * to allow driver writes more control over the exact init sequence.
  *
- * Drivers must set fb_helper->funcs before calling
- * drm_fb_helper_initial_config().
+ * Drivers must call drm_fb_helper_prepare() before calling this function.
  *
  * RETURNS:
  * Zero if everything went ok, nonzero otherwise.
@@ -558,10 +591,6 @@ int drm_fb_helper_init(struct drm_device *dev,
 	if (!max_conn_count)
 		return -EINVAL;
 
-	fb_helper->dev = dev;
-
-	INIT_LIST_HEAD(&fb_helper->kernel_fb_list);
-
 	fb_helper->crtc_info = kcalloc(crtc_count, sizeof(struct drm_fb_helper_crtc), GFP_KERNEL);
 	if (!fb_helper->crtc_info)
 		return -ENOMEM;
@@ -572,6 +601,7 @@ int drm_fb_helper_init(struct drm_device *dev,
 		kfree(fb_helper->crtc_info);
 		return -ENOMEM;
 	}
+	fb_helper->connector_info_alloc_count = dev->mode_config.num_connector;
 	fb_helper->connector_count = 0;
 
 	for (i = 0; i < crtc_count; i++) {
@@ -936,7 +966,7 @@ static int drm_fb_helper_single_fb_probe(struct drm_fb_helper *fb_helper,
 		struct drm_fb_helper_connector *fb_helper_conn = fb_helper->connector_info[i];
 		struct drm_cmdline_mode *cmdline_mode;
 
-		cmdline_mode = &fb_helper_conn->cmdline_mode;
+		cmdline_mode = &fb_helper_conn->connector->cmdline_mode;
 
 		if (cmdline_mode->bpp_specified) {
 			switch (cmdline_mode->bpp) {
@@ -1056,7 +1086,6 @@ void drm_fb_helper_fill_fix(struct fb_info *info, uint32_t pitch,
 	info->fix.ypanstep = 1; /* doing it in hw */
 	info->fix.ywrapstep = 0;
 	info->fix.accel = FB_ACCEL_NONE;
-	info->fix.type_aux = 0;
 
 	info->fix.line_length = pitch;
 	return;
@@ -1184,9 +1213,7 @@ EXPORT_SYMBOL(drm_has_preferred_mode);
 
 static bool drm_has_cmdline_mode(struct drm_fb_helper_connector *fb_connector)
 {
-	struct drm_cmdline_mode *cmdline_mode;
-	cmdline_mode = &fb_connector->cmdline_mode;
-	return cmdline_mode->specified;
+	return fb_connector->connector->cmdline_mode.specified;
 }
 
 struct drm_display_mode *drm_pick_cmdline_mode(struct drm_fb_helper_connector *fb_helper_conn,
@@ -1196,7 +1223,7 @@ struct drm_display_mode *drm_pick_cmdline_mode(struct drm_fb_helper_connector *f
 	struct drm_display_mode *mode = NULL;
 	bool prefer_non_interlace;
 
-	cmdline_mode = &fb_helper_conn->cmdline_mode;
+	cmdline_mode = &fb_helper_conn->connector->cmdline_mode;
 	if (cmdline_mode->specified == false)
 		return mode;
 
@@ -1581,8 +1608,6 @@ bool drm_fb_helper_initial_config(struct drm_fb_helper *fb_helper, int bpp_sel)
 	struct drm_device *dev = fb_helper->dev;
 	int count = 0;
 
-	drm_fb_helper_parse_command_line(fb_helper);
-
 	mutex_lock(&dev->mode_config.mutex);
 	count = drm_fb_helper_probe_connector_modes(fb_helper,
 						    dev->mode_config.max_width,
@@ -1613,8 +1638,10 @@ EXPORT_SYMBOL(drm_fb_helper_initial_config);
  * either the output polling work or a work item launched from the driver's
  * hotplug interrupt).
  *
- * Note that the driver must ensure that this is only called _after_ the fb has
- * been fully set up, i.e. after the call to drm_fb_helper_initial_config.
+ * Note that drivers may call this even before calling
+ * drm_fb_helper_initial_config but only aftert drm_fb_helper_init. This allows
+ * for a race-free fbcon setup and will make sure that the fbdev emulation will
+ * not miss any hotplug events.
  *
  * RETURNS:
  * 0 on success and a non-zero error code otherwise.
@@ -1624,11 +1651,8 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	struct drm_device *dev = fb_helper->dev;
 	u32 max_width, max_height;
 
-	if (!fb_helper->fb)
-		return 0;
-
 	mutex_lock(&fb_helper->dev->mode_config.mutex);
-	if (!drm_fb_helper_is_bound(fb_helper)) {
+	if (!fb_helper->fb || !drm_fb_helper_is_bound(fb_helper)) {
 		fb_helper->delayed_hotplug = true;
 		mutex_unlock(&fb_helper->dev->mode_config.mutex);
 		return 0;
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 021fe5d11df5..ed7bc68f7e87 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -38,10 +38,11 @@
 #include <linux/poll.h>
 #include <linux/slab.h>
 #include <linux/module.h>
+#include "drm_legacy.h"
+#include "drm_internal.h"
 
 /* from BKL pushdown */
 DEFINE_MUTEX(drm_global_mutex);
-EXPORT_SYMBOL(drm_global_mutex);
 
 static int drm_open_helper(struct file *filp, struct drm_minor *minor);
 
@@ -112,55 +113,12 @@ err_undo:
 EXPORT_SYMBOL(drm_open);
 
 /**
- * File \c open operation.
- *
- * \param inode device inode.
- * \param filp file pointer.
- *
- * Puts the dev->fops corresponding to the device minor number into
- * \p filp, call the \c open method, and restore the file operations.
- */
-int drm_stub_open(struct inode *inode, struct file *filp)
-{
-	struct drm_device *dev;
-	struct drm_minor *minor;
-	int err = -ENODEV;
-	const struct file_operations *new_fops;
-
-	DRM_DEBUG("\n");
-
-	mutex_lock(&drm_global_mutex);
-	minor = drm_minor_acquire(iminor(inode));
-	if (IS_ERR(minor))
-		goto out_unlock;
-
-	dev = minor->dev;
-	new_fops = fops_get(dev->driver->fops);
-	if (!new_fops)
-		goto out_release;
-
-	replace_fops(filp, new_fops);
-	if (filp->f_op->open)
-		err = filp->f_op->open(inode, filp);
-
-out_release:
-	drm_minor_release(minor);
-out_unlock:
-	mutex_unlock(&drm_global_mutex);
-	return err;
-}
-
-/**
  * Check whether DRI will run on this CPU.
  *
  * \return non-zero if the DRI will run on this CPU, or zero otherwise.
  */
 static int drm_cpu_valid(void)
 {
-#if defined(__i386__)
-	if (boot_cpu_data.x86 == 3)
-		return 0;	/* No cmpxchg on a 386 */
-#endif
 #if defined(__sparc__) && !defined(__sparc_v9__)
 	return 0;		/* No cmpxchg before v9 sparc. */
 #endif
@@ -203,8 +161,7 @@ static int drm_open_helper(struct file *filp, struct drm_minor *minor)
 	priv->minor = minor;
 
 	/* for compatibility root is always authenticated */
-	priv->always_authenticated = capable(CAP_SYS_ADMIN);
-	priv->authenticated = priv->always_authenticated;
+	priv->authenticated = capable(CAP_SYS_ADMIN);
 	priv->lock_count = 0;
 
 	INIT_LIST_HEAD(&priv->lhead);
@@ -214,7 +171,7 @@ static int drm_open_helper(struct file *filp, struct drm_minor *minor)
 	init_waitqueue_head(&priv->event_wait);
 	priv->event_space = 4096; /* set aside 4k for event buffer */
 
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_open(dev, priv);
 
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
@@ -299,7 +256,7 @@ out_close:
 out_prime_destroy:
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
 		drm_prime_destroy_file_private(&priv->prime);
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_release(dev, priv);
 	put_pid(priv->pid);
 	kfree(priv);
@@ -311,11 +268,11 @@ static void drm_master_release(struct drm_device *dev, struct file *filp)
 {
 	struct drm_file *file_priv = filp->private_data;
 
-	if (drm_i_have_hw_lock(dev, file_priv)) {
+	if (drm_legacy_i_have_hw_lock(dev, file_priv)) {
 		DRM_DEBUG("File %p released, freeing lock for context %d\n",
 			  filp, _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
-		drm_lock_free(&file_priv->master->lock,
-			      _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
+		drm_legacy_lock_free(&file_priv->master->lock,
+				     _DRM_LOCKING_CONTEXT(file_priv->master->lock.hw_lock->lock));
 	}
 }
 
@@ -373,8 +330,6 @@ static void drm_legacy_dev_reinit(struct drm_device *dev)
  */
 int drm_lastclose(struct drm_device * dev)
 {
-	struct drm_vma_entry *vma, *vma_temp;
-
 	DRM_DEBUG("\n");
 
 	if (dev->driver->lastclose)
@@ -389,13 +344,7 @@ int drm_lastclose(struct drm_device * dev)
 	drm_agp_clear(dev);
 
 	drm_legacy_sg_cleanup(dev);
-
-	/* Clear vma list (only built for debugging) */
-	list_for_each_entry_safe(vma, vma_temp, &dev->vmalist, head) {
-		list_del(&vma->head);
-		kfree(vma);
-	}
-
+	drm_legacy_vma_flush(dev);
 	drm_legacy_dma_takedown(dev);
 
 	mutex_unlock(&dev->struct_mutex);
@@ -429,6 +378,10 @@ int drm_release(struct inode *inode, struct file *filp)
 
 	DRM_DEBUG("open_count = %d\n", dev->open_count);
 
+	mutex_lock(&dev->struct_mutex);
+	list_del(&file_priv->lhead);
+	mutex_unlock(&dev->struct_mutex);
+
 	if (dev->driver->preclose)
 		dev->driver->preclose(dev, file_priv);
 
@@ -451,54 +404,28 @@ int drm_release(struct inode *inode, struct file *filp)
 		drm_master_release(dev, filp);
 
 	if (drm_core_check_feature(dev, DRIVER_HAVE_DMA))
-		drm_core_reclaim_buffers(dev, file_priv);
+		drm_legacy_reclaim_buffers(dev, file_priv);
 
 	drm_events_release(file_priv);
 
-	if (dev->driver->driver_features & DRIVER_MODESET)
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
 		drm_fb_release(file_priv);
 
-	if (dev->driver->driver_features & DRIVER_GEM)
+	if (drm_core_check_feature(dev, DRIVER_GEM))
 		drm_gem_release(dev, file_priv);
 
-	mutex_lock(&dev->ctxlist_mutex);
-	if (!list_empty(&dev->ctxlist)) {
-		struct drm_ctx_list *pos, *n;
-
-		list_for_each_entry_safe(pos, n, &dev->ctxlist, head) {
-			if (pos->tag == file_priv &&
-			    pos->handle != DRM_KERNEL_CONTEXT) {
-				if (dev->driver->context_dtor)
-					dev->driver->context_dtor(dev,
-								  pos->handle);
-
-				drm_ctxbitmap_free(dev, pos->handle);
-
-				list_del(&pos->head);
-				kfree(pos);
-			}
-		}
-	}
-	mutex_unlock(&dev->ctxlist_mutex);
+	drm_legacy_ctxbitmap_flush(dev, file_priv);
 
 	mutex_lock(&dev->master_mutex);
 
 	if (file_priv->is_master) {
 		struct drm_master *master = file_priv->master;
-		struct drm_file *temp;
-
-		mutex_lock(&dev->struct_mutex);
-		list_for_each_entry(temp, &dev->filelist, lhead) {
-			if ((temp->master == file_priv->master) &&
-			    (temp != file_priv))
-				temp->authenticated = temp->always_authenticated;
-		}
 
 		/**
 		 * Since the master is disappearing, so is the
 		 * possibility to lock.
 		 */
-
+		mutex_lock(&dev->struct_mutex);
 		if (master->lock.hw_lock) {
 			if (dev->sigdata.lock == master->lock.hw_lock)
 				dev->sigdata.lock = NULL;
@@ -522,10 +449,6 @@ int drm_release(struct inode *inode, struct file *filp)
 	file_priv->is_master = 0;
 	mutex_unlock(&dev->master_mutex);
 
-	mutex_lock(&dev->struct_mutex);
-	list_del(&file_priv->lhead);
-	mutex_unlock(&dev->struct_mutex);
-
 	if (dev->driver->postclose)
 		dev->driver->postclose(dev, file_priv);
 
@@ -533,6 +456,8 @@ int drm_release(struct inode *inode, struct file *filp)
 	if (drm_core_check_feature(dev, DRIVER_PRIME))
 		drm_prime_destroy_file_private(&file_priv->prime);
 
+	WARN_ON(!list_empty(&file_priv->event_list));
+
 	put_pid(file_priv->pid);
 	kfree(file_priv);
 
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index f7d71190aad5..f6ca51259fa3 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -38,6 +38,8 @@
 #include <linux/dma-buf.h>
 #include <drm/drmP.h>
 #include <drm/drm_vma_manager.h>
+#include <drm/drm_gem.h>
+#include "drm_internal.h"
 
 /** @file drm_gem.c
  *
@@ -146,7 +148,7 @@ int drm_gem_object_init(struct drm_device *dev,
 EXPORT_SYMBOL(drm_gem_object_init);
 
 /**
- * drm_gem_object_init - initialize an allocated private GEM object
+ * drm_gem_private_object_init - initialize an allocated private GEM object
  * @dev: drm_device the object should be initialized for
  * @obj: drm_gem_object to initialize
  * @size: object size
@@ -441,18 +443,31 @@ EXPORT_SYMBOL(drm_gem_create_mmap_offset);
  * drm_gem_get_pages - helper to allocate backing pages for a GEM object
  * from shmem
  * @obj: obj in question
- * @gfpmask: gfp mask of requested pages
+ *
+ * This reads the page-array of the shmem-backing storage of the given gem
+ * object. An array of pages is returned. If a page is not allocated or
+ * swapped-out, this will allocate/swap-in the required pages. Note that the
+ * whole object is covered by the page-array and pinned in memory.
+ *
+ * Use drm_gem_put_pages() to release the array and unpin all pages.
+ *
+ * This uses the GFP-mask set on the shmem-mapping (see mapping_set_gfp_mask()).
+ * If you require other GFP-masks, you have to do those allocations yourself.
+ *
+ * Note that you are not allowed to change gfp-zones during runtime. That is,
+ * shmem_read_mapping_page_gfp() must be called with the same gfp_zone(gfp) as
+ * set during initialization. If you have special zone constraints, set them
+ * after drm_gem_init_object() via mapping_set_gfp_mask(). shmem-core takes care
+ * to keep pages in the required zone during swap-in.
  */
-struct page **drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask)
+struct page **drm_gem_get_pages(struct drm_gem_object *obj)
 {
-	struct inode *inode;
 	struct address_space *mapping;
 	struct page *p, **pages;
 	int i, npages;
 
 	/* This is the shared memory object that backs the GEM resource */
-	inode = file_inode(obj->filp);
-	mapping = inode->i_mapping;
+	mapping = file_inode(obj->filp)->i_mapping;
 
 	/* We already BUG_ON() for non-page-aligned sizes in
 	 * drm_gem_object_init(), so we should never hit this unless
@@ -466,10 +481,8 @@ struct page **drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask)
 	if (pages == NULL)
 		return ERR_PTR(-ENOMEM);
 
-	gfpmask |= mapping_gfp_mask(mapping);
-
 	for (i = 0; i < npages; i++) {
-		p = shmem_read_mapping_page_gfp(mapping, i, gfpmask);
+		p = shmem_read_mapping_page(mapping, i);
 		if (IS_ERR(p))
 			goto fail;
 		pages[i] = p;
@@ -479,7 +492,7 @@ struct page **drm_gem_get_pages(struct drm_gem_object *obj, gfp_t gfpmask)
 		 * __GFP_DMA32 to be set in mapping_gfp_mask(inode->i_mapping)
 		 * so shmem can relocate pages during swapin if required.
 		 */
-		BUG_ON((gfpmask & __GFP_DMA32) &&
+		BUG_ON((mapping_gfp_mask(mapping) & __GFP_DMA32) &&
 				(page_to_pfn(p) >= 0x00100000UL));
 	}
 
@@ -568,7 +581,7 @@ drm_gem_close_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_close *args = data;
 	int ret;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	ret = drm_gem_handle_delete(file_priv, args->handle);
@@ -595,7 +608,7 @@ drm_gem_flink_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_object *obj;
 	int ret;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
@@ -648,7 +661,7 @@ drm_gem_open_ioctl(struct drm_device *dev, void *data,
 	int ret;
 	u32 handle;
 
-	if (!(dev->driver->driver_features & DRIVER_GEM))
+	if (!drm_core_check_feature(dev, DRIVER_GEM))
 		return -ENODEV;
 
 	mutex_lock(&dev->object_name_lock);
@@ -876,7 +889,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 					   vma_pages(vma));
 	if (!node) {
 		mutex_unlock(&dev->struct_mutex);
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	} else if (!drm_vma_node_is_allowed(node, filp)) {
 		mutex_unlock(&dev->struct_mutex);
 		return -EACCES;
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index 05c97c5350a1..0316310e2cc4 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -316,7 +316,8 @@ out:
 EXPORT_SYMBOL_GPL(drm_gem_cma_prime_get_sg_table);
 
 struct drm_gem_object *
-drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
+drm_gem_cma_prime_import_sg_table(struct drm_device *dev,
+				  struct dma_buf_attachment *attach,
 				  struct sg_table *sgt)
 {
 	struct drm_gem_cma_object *cma_obj;
@@ -325,14 +326,14 @@ drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
 		return ERR_PTR(-EINVAL);
 
 	/* Create a CMA GEM buffer. */
-	cma_obj = __drm_gem_cma_create(dev, size);
+	cma_obj = __drm_gem_cma_create(dev, attach->dmabuf->size);
 	if (IS_ERR(cma_obj))
-		return ERR_PTR(PTR_ERR(cma_obj));
+		return ERR_CAST(cma_obj);
 
 	cma_obj->paddr = sg_dma_address(sgt->sgl);
 	cma_obj->sgt = sgt;
 
-	DRM_DEBUG_PRIME("dma_addr = %pad, size = %zu\n", &cma_obj->paddr, size);
+	DRM_DEBUG_PRIME("dma_addr = %pad, size = %zu\n", &cma_obj->paddr, attach->dmabuf->size);
 
 	return &cma_obj->base;
 }
diff --git a/drivers/gpu/drm/drm_hashtab.c b/drivers/gpu/drm/drm_hashtab.c
index 7e4bae760e27..c3b80fd65d62 100644
--- a/drivers/gpu/drm/drm_hashtab.c
+++ b/drivers/gpu/drm/drm_hashtab.c
@@ -125,7 +125,7 @@ int drm_ht_insert_item(struct drm_open_hash *ht, struct drm_hash_item *item)
 		parent = &entry->head;
 	}
 	if (parent) {
-		hlist_add_after_rcu(parent, &item->head);
+		hlist_add_behind_rcu(&item->head, parent);
 	} else {
 		hlist_add_head_rcu(&item->head, h_list);
 	}
diff --git a/drivers/gpu/drm/drm_info.c b/drivers/gpu/drm/drm_info.c
index 86feedd5e6f6..51efebd434f3 100644
--- a/drivers/gpu/drm/drm_info.c
+++ b/drivers/gpu/drm/drm_info.c
@@ -35,6 +35,9 @@
 
 #include <linux/seq_file.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
+
+#include "drm_legacy.h"
 
 /**
  * Called when "/proc/dri/.../name" is read.
@@ -132,7 +135,7 @@ int drm_bufs_info(struct seq_file *m, void *data)
 				   i,
 				   dma->bufs[i].buf_size,
 				   dma->bufs[i].buf_count,
-				   atomic_read(&dma->bufs[i].freelist.count),
+				   0,
 				   dma->bufs[i].seg_count,
 				   seg_pages,
 				   seg_pages * PAGE_SIZE / 1024);
@@ -183,15 +186,32 @@ int drm_clients_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	struct drm_file *priv;
 
+	seq_printf(m,
+		   "%20s %5s %3s master a %5s %10s\n",
+		   "command",
+		   "pid",
+		   "dev",
+		   "uid",
+		   "magic");
+
+	/* dev->filelist is sorted youngest first, but we want to present
+	 * oldest first (i.e. kernel, servers, clients), so walk backwardss.
+	 */
 	mutex_lock(&dev->struct_mutex);
-	seq_printf(m, "a dev	pid    uid	magic\n\n");
-	list_for_each_entry(priv, &dev->filelist, lhead) {
-		seq_printf(m, "%c %3d %5d %5d %10u\n",
-			   priv->authenticated ? 'y' : 'n',
-			   priv->minor->index,
+	list_for_each_entry_reverse(priv, &dev->filelist, lhead) {
+		struct task_struct *task;
+
+		rcu_read_lock(); /* locks pid_task()->comm */
+		task = pid_task(priv->pid, PIDTYPE_PID);
+		seq_printf(m, "%20s %5d %3d   %c    %c %5d %10u\n",
+			   task ? task->comm : "<unknown>",
 			   pid_vnr(priv->pid),
+			   priv->minor->index,
+			   priv->is_master ? 'y' : 'n',
+			   priv->authenticated ? 'y' : 'n',
 			   from_kuid_munged(seq_user_ns(m), priv->uid),
 			   priv->magic);
+		rcu_read_unlock();
 	}
 	mutex_unlock(&dev->struct_mutex);
 	return 0;
@@ -223,62 +243,3 @@ int drm_gem_name_info(struct seq_file *m, void *data)
 
 	return 0;
 }
-
-#if DRM_DEBUG_CODE
-
-int drm_vma_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_vma_entry *pt;
-	struct vm_area_struct *vma;
-	unsigned long vma_count = 0;
-#if defined(__i386__)
-	unsigned int pgprot;
-#endif
-
-	mutex_lock(&dev->struct_mutex);
-	list_for_each_entry(pt, &dev->vmalist, head)
-		vma_count++;
-
-	seq_printf(m, "vma use count: %lu, high_memory = %pK, 0x%pK\n",
-		   vma_count, high_memory,
-		   (void *)(unsigned long)virt_to_phys(high_memory));
-
-	list_for_each_entry(pt, &dev->vmalist, head) {
-		vma = pt->vma;
-		if (!vma)
-			continue;
-		seq_printf(m,
-			   "\n%5d 0x%pK-0x%pK %c%c%c%c%c%c 0x%08lx000",
-			   pt->pid,
-			   (void *)vma->vm_start, (void *)vma->vm_end,
-			   vma->vm_flags & VM_READ ? 'r' : '-',
-			   vma->vm_flags & VM_WRITE ? 'w' : '-',
-			   vma->vm_flags & VM_EXEC ? 'x' : '-',
-			   vma->vm_flags & VM_MAYSHARE ? 's' : 'p',
-			   vma->vm_flags & VM_LOCKED ? 'l' : '-',
-			   vma->vm_flags & VM_IO ? 'i' : '-',
-			   vma->vm_pgoff);
-
-#if defined(__i386__)
-		pgprot = pgprot_val(vma->vm_page_prot);
-		seq_printf(m, " %c%c%c%c%c%c%c%c%c",
-			   pgprot & _PAGE_PRESENT ? 'p' : '-',
-			   pgprot & _PAGE_RW ? 'w' : 'r',
-			   pgprot & _PAGE_USER ? 'u' : 's',
-			   pgprot & _PAGE_PWT ? 't' : 'b',
-			   pgprot & _PAGE_PCD ? 'u' : 'c',
-			   pgprot & _PAGE_ACCESSED ? 'a' : '-',
-			   pgprot & _PAGE_DIRTY ? 'd' : '-',
-			   pgprot & _PAGE_PSE ? 'm' : 'k',
-			   pgprot & _PAGE_GLOBAL ? 'g' : 'l');
-#endif
-		seq_printf(m, "\n");
-	}
-	mutex_unlock(&dev->struct_mutex);
-	return 0;
-}
-
-#endif
-
diff --git a/drivers/gpu/drm/drm_internal.h b/drivers/gpu/drm/drm_internal.h
new file mode 100644
index 000000000000..7cc0a3516871
--- /dev/null
+++ b/drivers/gpu/drm/drm_internal.h
@@ -0,0 +1,132 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *   Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/* drm_irq.c */
+extern unsigned int drm_timestamp_monotonic;
+
+/* drm_fops.c */
+extern struct mutex drm_global_mutex;
+int drm_lastclose(struct drm_device *dev);
+
+/* drm_pci.c */
+int drm_pci_set_unique(struct drm_device *dev,
+		       struct drm_master *master,
+		       struct drm_unique *u);
+int drm_irq_by_busid(struct drm_device *dev, void *data,
+		     struct drm_file *file_priv);
+
+/* drm_vm.c */
+int drm_vma_info(struct seq_file *m, void *data);
+void drm_vm_open_locked(struct drm_device *dev, struct vm_area_struct *vma);
+void drm_vm_close_locked(struct drm_device *dev, struct vm_area_struct *vma);
+
+/* drm_prime.c */
+int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv);
+int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv);
+
+void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv);
+void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv);
+void drm_prime_remove_buf_handle_locked(struct drm_prime_file_private *prime_fpriv,
+					struct dma_buf *dma_buf);
+
+/* drm_info.c */
+int drm_name_info(struct seq_file *m, void *data);
+int drm_vm_info(struct seq_file *m, void *data);
+int drm_bufs_info(struct seq_file *m, void *data);
+int drm_vblank_info(struct seq_file *m, void *data);
+int drm_clients_info(struct seq_file *m, void* data);
+int drm_gem_name_info(struct seq_file *m, void *data);
+
+/* drm_irq.c */
+int drm_control(struct drm_device *dev, void *data,
+		struct drm_file *file_priv);
+int drm_modeset_ctl(struct drm_device *dev, void *data,
+		    struct drm_file *file_priv);
+
+/* drm_auth.c */
+int drm_getmagic(struct drm_device *dev, void *data,
+		 struct drm_file *file_priv);
+int drm_authmagic(struct drm_device *dev, void *data,
+		  struct drm_file *file_priv);
+int drm_remove_magic(struct drm_master *master, drm_magic_t magic);
+
+/* drm_sysfs.c */
+extern struct class *drm_class;
+
+struct class *drm_sysfs_create(struct module *owner, char *name);
+void drm_sysfs_destroy(void);
+struct device *drm_sysfs_minor_alloc(struct drm_minor *minor);
+int drm_sysfs_connector_add(struct drm_connector *connector);
+void drm_sysfs_connector_remove(struct drm_connector *connector);
+
+/* drm_gem.c */
+int drm_gem_init(struct drm_device *dev);
+void drm_gem_destroy(struct drm_device *dev);
+int drm_gem_handle_create_tail(struct drm_file *file_priv,
+			       struct drm_gem_object *obj,
+			       u32 *handlep);
+int drm_gem_close_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_gem_flink_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_gem_open_ioctl(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv);
+void drm_gem_open(struct drm_device *dev, struct drm_file *file_private);
+void drm_gem_release(struct drm_device *dev, struct drm_file *file_private);
+
+/* drm_drv.c */
+int drm_setmaster_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv);
+struct drm_master *drm_master_create(struct drm_minor *minor);
+
+/* drm_debugfs.c */
+#if defined(CONFIG_DEBUG_FS)
+int drm_debugfs_init(struct drm_minor *minor, int minor_id,
+		     struct dentry *root);
+int drm_debugfs_cleanup(struct drm_minor *minor);
+int drm_debugfs_connector_add(struct drm_connector *connector);
+void drm_debugfs_connector_remove(struct drm_connector *connector);
+#else
+static inline int drm_debugfs_init(struct drm_minor *minor, int minor_id,
+				   struct dentry *root)
+{
+	return 0;
+}
+
+static inline int drm_debugfs_cleanup(struct drm_minor *minor)
+{
+	return 0;
+}
+
+static inline int drm_debugfs_connector_add(struct drm_connector *connector)
+{
+	return 0;
+}
+static inline void drm_debugfs_connector_remove(struct drm_connector *connector)
+{
+}
+#endif
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index 69c61f392e66..00587a1e3c83 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -1,11 +1,3 @@
-/**
- * \file drm_ioctl.c
- * IOCTL processing for DRM
- *
- * \author Rickard E. (Rik) Faith <faith@valinux.com>
- * \author Gareth Hughes <gareth@valinux.com>
- */
-
 /*
  * Created: Fri Jan  8 09:01:26 1999 by faith@valinux.com
  *
@@ -13,6 +5,9 @@
  * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
  * All Rights Reserved.
  *
+ * Author Rickard E. (Rik) Faith <faith@valinux.com>
+ * Author Gareth Hughes <gareth@valinux.com>
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
@@ -35,6 +30,8 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_core.h>
+#include "drm_legacy.h"
+#include "drm_internal.h"
 
 #include <linux/pci.h>
 #include <linux/export.h>
@@ -42,6 +39,9 @@
 #include <asm/mtrr.h>
 #endif
 
+static int drm_version(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv);
+
 /**
  * Get the bus id.
  *
@@ -53,7 +53,7 @@
  *
  * Copies the bus id from drm_device::unique into user space.
  */
-int drm_getunique(struct drm_device *dev, void *data,
+static int drm_getunique(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_unique *u = data;
@@ -75,7 +75,6 @@ drm_unset_busid(struct drm_device *dev,
 	kfree(master->unique);
 	master->unique = NULL;
 	master->unique_len = 0;
-	master->unique_size = 0;
 }
 
 /**
@@ -93,7 +92,7 @@ drm_unset_busid(struct drm_device *dev,
  * version 1.1 or greater. Also note that KMS is all version 1.1 and later and
  * UMS was only ever supported on pci devices.
  */
-int drm_setunique(struct drm_device *dev, void *data,
+static int drm_setunique(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_unique *u = data;
@@ -131,15 +130,15 @@ static int drm_set_busid(struct drm_device *dev, struct drm_file *file_priv)
 	if (master->unique != NULL)
 		drm_unset_busid(dev, master);
 
-	if (dev->driver->bus && dev->driver->bus->set_busid) {
-		ret = dev->driver->bus->set_busid(dev, master);
+	if (dev->driver->set_busid) {
+		ret = dev->driver->set_busid(dev, master);
 		if (ret) {
 			drm_unset_busid(dev, master);
 			return ret;
 		}
 	} else {
 		if (WARN(dev->unique == NULL,
-			 "No drm_bus.set_busid() implementation provided by "
+			 "No drm_driver.set_busid() implementation provided by "
 			 "%ps. Use drm_dev_set_unique() to set the unique "
 			 "name explicitly.", dev->driver))
 			return -EINVAL;
@@ -165,7 +164,7 @@ static int drm_set_busid(struct drm_device *dev, struct drm_file *file_priv)
  * Searches for the mapping with the specified offset and copies its information
  * into userspace
  */
-int drm_getmap(struct drm_device *dev, void *data,
+static int drm_getmap(struct drm_device *dev, void *data,
 	       struct drm_file *file_priv)
 {
 	struct drm_map *map = data;
@@ -226,7 +225,7 @@ int drm_getmap(struct drm_device *dev, void *data,
  * Searches for the client with the specified index and copies its information
  * into userspace
  */
-int drm_getclient(struct drm_device *dev, void *data,
+static int drm_getclient(struct drm_device *dev, void *data,
 		  struct drm_file *file_priv)
 {
 	struct drm_client *client = data;
@@ -266,7 +265,7 @@ int drm_getclient(struct drm_device *dev, void *data,
  *
  * \return zero on success or a negative number on failure.
  */
-int drm_getstats(struct drm_device *dev, void *data,
+static int drm_getstats(struct drm_device *dev, void *data,
 		 struct drm_file *file_priv)
 {
 	struct drm_stats *stats = data;
@@ -280,7 +279,7 @@ int drm_getstats(struct drm_device *dev, void *data,
 /**
  * Get device/driver capabilities
  */
-int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
+static int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_get_cap *req = data;
 
@@ -330,7 +329,7 @@ int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 /**
  * Set device/driver capabilities
  */
-int
+static int
 drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_set_client_cap *req = data;
@@ -342,8 +341,6 @@ drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 		file_priv->stereo_allowed = req->value;
 		break;
 	case DRM_CLIENT_CAP_UNIVERSAL_PLANES:
-		if (!drm_universal_planes)
-			return -EINVAL;
 		if (req->value > 1)
 			return -EINVAL;
 		file_priv->universal_planes = req->value;
@@ -366,7 +363,7 @@ drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
  *
  * Sets the requested interface version
  */
-int drm_setversion(struct drm_device *dev, void *data, struct drm_file *file_priv)
+static int drm_setversion(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_set_version *sv = data;
 	int if_version, retcode = 0;
@@ -417,3 +414,358 @@ int drm_noop(struct drm_device *dev, void *data,
 	return 0;
 }
 EXPORT_SYMBOL(drm_noop);
+
+/**
+ * Copy and IOCTL return string to user space
+ */
+static int drm_copy_field(char __user *buf, size_t *buf_len, const char *value)
+{
+	int len;
+
+	/* don't overflow userbuf */
+	len = strlen(value);
+	if (len > *buf_len)
+		len = *buf_len;
+
+	/* let userspace know exact length of driver value (which could be
+	 * larger than the userspace-supplied buffer) */
+	*buf_len = strlen(value);
+
+	/* finally, try filling in the userbuf */
+	if (len && buf)
+		if (copy_to_user(buf, value, len))
+			return -EFAULT;
+	return 0;
+}
+
+/**
+ * Get version information
+ *
+ * \param inode device inode.
+ * \param filp file pointer.
+ * \param cmd command.
+ * \param arg user argument, pointing to a drm_version structure.
+ * \return zero on success or negative number on failure.
+ *
+ * Fills in the version information in \p arg.
+ */
+static int drm_version(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
+{
+	struct drm_version *version = data;
+	int err;
+
+	version->version_major = dev->driver->major;
+	version->version_minor = dev->driver->minor;
+	version->version_patchlevel = dev->driver->patchlevel;
+	err = drm_copy_field(version->name, &version->name_len,
+			dev->driver->name);
+	if (!err)
+		err = drm_copy_field(version->date, &version->date_len,
+				dev->driver->date);
+	if (!err)
+		err = drm_copy_field(version->desc, &version->desc_len,
+				dev->driver->desc);
+
+	return err;
+}
+
+/**
+ * drm_ioctl_permit - Check ioctl permissions against caller
+ *
+ * @flags: ioctl permission flags.
+ * @file_priv: Pointer to struct drm_file identifying the caller.
+ *
+ * Checks whether the caller is allowed to run an ioctl with the
+ * indicated permissions. If so, returns zero. Otherwise returns an
+ * error code suitable for ioctl return.
+ */
+static int drm_ioctl_permit(u32 flags, struct drm_file *file_priv)
+{
+	/* ROOT_ONLY is only for CAP_SYS_ADMIN */
+	if (unlikely((flags & DRM_ROOT_ONLY) && !capable(CAP_SYS_ADMIN)))
+		return -EACCES;
+
+	/* AUTH is only for authenticated or render client */
+	if (unlikely((flags & DRM_AUTH) && !drm_is_render_client(file_priv) &&
+		     !file_priv->authenticated))
+		return -EACCES;
+
+	/* MASTER is only for master or control clients */
+	if (unlikely((flags & DRM_MASTER) && !file_priv->is_master &&
+		     !drm_is_control_client(file_priv)))
+		return -EACCES;
+
+	/* Control clients must be explicitly allowed */
+	if (unlikely(!(flags & DRM_CONTROL_ALLOW) &&
+		     drm_is_control_client(file_priv)))
+		return -EACCES;
+
+	/* Render clients must be explicitly allowed */
+	if (unlikely(!(flags & DRM_RENDER_ALLOW) &&
+		     drm_is_render_client(file_priv)))
+		return -EACCES;
+
+	return 0;
+}
+
+#define DRM_IOCTL_DEF(ioctl, _func, _flags) \
+	[DRM_IOCTL_NR(ioctl)] = {.cmd = ioctl, .func = _func, .flags = _flags, .cmd_drv = 0, .name = #ioctl}
+
+/** Ioctl table */
+static const struct drm_ioctl_desc drm_ioctls[] = {
+	DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_IRQ_BUSID, drm_irq_by_busid, DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_CLIENT_CAP, drm_setclientcap, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_VERSION, drm_setversion, DRM_MASTER),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_UNIQUE, drm_setunique, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_BLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_UNBLOCK, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AUTH_MAGIC, drm_authmagic, DRM_AUTH|DRM_MASTER),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_MAP, drm_legacy_addmap_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_MAP, drm_legacy_rmmap_ioctl, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_SAREA_CTX, drm_legacy_setsareactx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_SAREA_CTX, drm_legacy_getsareactx, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SET_MASTER, drm_setmaster_ioctl, DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_DROP_MASTER, drm_dropmaster_ioctl, DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_CTX, drm_legacy_addctx, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_CTX, drm_legacy_rmctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_MOD_CTX, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_GET_CTX, drm_legacy_getctx, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_SWITCH_CTX, drm_legacy_switchctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_NEW_CTX, drm_legacy_newctx, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RES_CTX, drm_legacy_resctx, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_RM_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_LOCK, drm_legacy_lock, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_UNLOCK, drm_legacy_unlock, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_FINISH, drm_noop, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_ADD_BUFS, drm_legacy_addbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_MARK_BUFS, drm_legacy_markbufs, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_INFO_BUFS, drm_legacy_infobufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_MAP_BUFS, drm_legacy_mapbufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_FREE_BUFS, drm_legacy_freebufs, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_DMA, drm_legacy_dma_ioctl, DRM_AUTH),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_CONTROL, drm_control, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+#if __OS_HAS_AGP
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ACQUIRE, drm_agp_acquire_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_RELEASE, drm_agp_release_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ENABLE, drm_agp_enable_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_INFO, drm_agp_info_ioctl, DRM_AUTH),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_ALLOC, drm_agp_alloc_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_FREE, drm_agp_free_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_BIND, drm_agp_bind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_AGP_UNBIND, drm_agp_unbind_ioctl, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+#endif
+
+	DRM_IOCTL_DEF(DRM_IOCTL_SG_ALLOC, drm_legacy_sg_alloc, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF(DRM_IOCTL_SG_FREE, drm_legacy_sg_free, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_WAIT_VBLANK, drm_wait_vblank, DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODESET_CTL, drm_modeset_ctl, 0),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_UPDATE_DRAW, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+};
+
+#define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
+
+/**
+ * Called whenever a process performs an ioctl on /dev/drm.
+ *
+ * \param inode device inode.
+ * \param file_priv DRM file private.
+ * \param cmd command.
+ * \param arg user argument.
+ * \return zero on success or negative number on failure.
+ *
+ * Looks up the ioctl function in the ::ioctls table, checking for root
+ * previleges if so required, and dispatches to the respective function.
+ */
+long drm_ioctl(struct file *filp,
+	      unsigned int cmd, unsigned long arg)
+{
+	struct drm_file *file_priv = filp->private_data;
+	struct drm_device *dev;
+	const struct drm_ioctl_desc *ioctl = NULL;
+	drm_ioctl_t *func;
+	unsigned int nr = DRM_IOCTL_NR(cmd);
+	int retcode = -EINVAL;
+	char stack_kdata[128];
+	char *kdata = NULL;
+	unsigned int usize, asize;
+
+	dev = file_priv->minor->dev;
+
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
+	if ((nr >= DRM_CORE_IOCTL_COUNT) &&
+	    ((nr < DRM_COMMAND_BASE) || (nr >= DRM_COMMAND_END)))
+		goto err_i1;
+	if ((nr >= DRM_COMMAND_BASE) && (nr < DRM_COMMAND_END) &&
+	    (nr < DRM_COMMAND_BASE + dev->driver->num_ioctls)) {
+		u32 drv_size;
+		ioctl = &dev->driver->ioctls[nr - DRM_COMMAND_BASE];
+		drv_size = _IOC_SIZE(ioctl->cmd_drv);
+		usize = asize = _IOC_SIZE(cmd);
+		if (drv_size > asize)
+			asize = drv_size;
+		cmd = ioctl->cmd_drv;
+	}
+	else if ((nr >= DRM_COMMAND_END) || (nr < DRM_COMMAND_BASE)) {
+		u32 drv_size;
+
+		ioctl = &drm_ioctls[nr];
+
+		drv_size = _IOC_SIZE(ioctl->cmd);
+		usize = asize = _IOC_SIZE(cmd);
+		if (drv_size > asize)
+			asize = drv_size;
+
+		cmd = ioctl->cmd;
+	} else
+		goto err_i1;
+
+	DRM_DEBUG("pid=%d, dev=0x%lx, auth=%d, %s\n",
+		  task_pid_nr(current),
+		  (long)old_encode_dev(file_priv->minor->kdev->devt),
+		  file_priv->authenticated, ioctl->name);
+
+	/* Do not trust userspace, use our own definition */
+	func = ioctl->func;
+
+	if (unlikely(!func)) {
+		DRM_DEBUG("no function\n");
+		retcode = -EINVAL;
+		goto err_i1;
+	}
+
+	retcode = drm_ioctl_permit(ioctl->flags, file_priv);
+	if (unlikely(retcode))
+		goto err_i1;
+
+	if (cmd & (IOC_IN | IOC_OUT)) {
+		if (asize <= sizeof(stack_kdata)) {
+			kdata = stack_kdata;
+		} else {
+			kdata = kmalloc(asize, GFP_KERNEL);
+			if (!kdata) {
+				retcode = -ENOMEM;
+				goto err_i1;
+			}
+		}
+		if (asize > usize)
+			memset(kdata + usize, 0, asize - usize);
+	}
+
+	if (cmd & IOC_IN) {
+		if (copy_from_user(kdata, (void __user *)arg,
+				   usize) != 0) {
+			retcode = -EFAULT;
+			goto err_i1;
+		}
+	} else if (cmd & IOC_OUT) {
+		memset(kdata, 0, usize);
+	}
+
+	if (ioctl->flags & DRM_UNLOCKED)
+		retcode = func(dev, kdata, file_priv);
+	else {
+		mutex_lock(&drm_global_mutex);
+		retcode = func(dev, kdata, file_priv);
+		mutex_unlock(&drm_global_mutex);
+	}
+
+	if (cmd & IOC_OUT) {
+		if (copy_to_user((void __user *)arg, kdata,
+				 usize) != 0)
+			retcode = -EFAULT;
+	}
+
+      err_i1:
+	if (!ioctl)
+		DRM_DEBUG("invalid ioctl: pid=%d, dev=0x%lx, auth=%d, cmd=0x%02x, nr=0x%02x\n",
+			  task_pid_nr(current),
+			  (long)old_encode_dev(file_priv->minor->kdev->devt),
+			  file_priv->authenticated, cmd, nr);
+
+	if (kdata != stack_kdata)
+		kfree(kdata);
+	if (retcode)
+		DRM_DEBUG("ret = %d\n", retcode);
+	return retcode;
+}
+EXPORT_SYMBOL(drm_ioctl);
+
+/**
+ * drm_ioctl_flags - Check for core ioctl and return ioctl permission flags
+ *
+ * @nr: Ioctl number.
+ * @flags: Where to return the ioctl permission flags
+ */
+bool drm_ioctl_flags(unsigned int nr, unsigned int *flags)
+{
+	if ((nr >= DRM_COMMAND_END && nr < DRM_CORE_IOCTL_COUNT) ||
+	    (nr < DRM_COMMAND_BASE)) {
+		*flags = drm_ioctls[nr].flags;
+		return true;
+	}
+
+	return false;
+}
+EXPORT_SYMBOL(drm_ioctl_flags);
diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c
index 08ba1209228e..5ef03c216a27 100644
--- a/drivers/gpu/drm/drm_irq.c
+++ b/drivers/gpu/drm/drm_irq.c
@@ -34,6 +34,7 @@
 
 #include <drm/drmP.h>
 #include "drm_trace.h"
+#include "drm_internal.h"
 
 #include <linux/interrupt.h>	/* For task queue support */
 #include <linux/slab.h>
@@ -55,12 +56,91 @@
  */
 #define DRM_REDUNDANT_VBLIRQ_THRESH_NS 1000000
 
+static bool
+drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
+			  struct timeval *tvblank, unsigned flags);
+
+static unsigned int drm_timestamp_precision = 20;  /* Default to 20 usecs. */
+
 /*
- * Clear vblank timestamp buffer for a crtc.
+ * Default to use monotonic timestamps for wait-for-vblank and page-flip
+ * complete events.
+ */
+unsigned int drm_timestamp_monotonic = 1;
+
+static int drm_vblank_offdelay = 5000;    /* Default to 5000 msecs. */
+
+module_param_named(vblankoffdelay, drm_vblank_offdelay, int, 0600);
+module_param_named(timestamp_precision_usec, drm_timestamp_precision, int, 0600);
+module_param_named(timestamp_monotonic, drm_timestamp_monotonic, int, 0600);
+
+/**
+ * drm_update_vblank_count - update the master vblank counter
+ * @dev: DRM device
+ * @crtc: counter to update
+ *
+ * Call back into the driver to update the appropriate vblank counter
+ * (specified by @crtc).  Deal with wraparound, if it occurred, and
+ * update the last read value so we can deal with wraparound on the next
+ * call if necessary.
+ *
+ * Only necessary when going from off->on, to account for frames we
+ * didn't get an interrupt for.
+ *
+ * Note: caller must hold dev->vbl_lock since this reads & writes
+ * device vblank fields.
  */
-static void clear_vblank_timestamps(struct drm_device *dev, int crtc)
+static void drm_update_vblank_count(struct drm_device *dev, int crtc)
 {
-	memset(dev->vblank[crtc].time, 0, sizeof(dev->vblank[crtc].time));
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+	u32 cur_vblank, diff, tslot;
+	bool rc;
+	struct timeval t_vblank;
+
+	/*
+	 * Interrupts were disabled prior to this call, so deal with counter
+	 * wrap if needed.
+	 * NOTE!  It's possible we lost a full dev->max_vblank_count events
+	 * here if the register is small or we had vblank interrupts off for
+	 * a long time.
+	 *
+	 * We repeat the hardware vblank counter & timestamp query until
+	 * we get consistent results. This to prevent races between gpu
+	 * updating its hardware counter while we are retrieving the
+	 * corresponding vblank timestamp.
+	 */
+	do {
+		cur_vblank = dev->driver->get_vblank_counter(dev, crtc);
+		rc = drm_get_last_vbltimestamp(dev, crtc, &t_vblank, 0);
+	} while (cur_vblank != dev->driver->get_vblank_counter(dev, crtc));
+
+	/* Deal with counter wrap */
+	diff = cur_vblank - vblank->last;
+	if (cur_vblank < vblank->last) {
+		diff += dev->max_vblank_count;
+
+		DRM_DEBUG("last_vblank[%d]=0x%x, cur_vblank=0x%x => diff=0x%x\n",
+			  crtc, vblank->last, cur_vblank, diff);
+	}
+
+	DRM_DEBUG("updating vblank count on crtc %d, missed %d\n",
+		  crtc, diff);
+
+	if (diff == 0)
+		return;
+
+	/* Reinitialize corresponding vblank timestamp if high-precision query
+	 * available. Skip this step if query unsupported or failed. Will
+	 * reinitialize delayed at next vblank interrupt in that case.
+	 */
+	if (rc) {
+		tslot = atomic_read(&vblank->count) + diff;
+		vblanktimestamp(dev, crtc, tslot) = t_vblank;
+	}
+
+	smp_mb__before_atomic();
+	atomic_add(diff, &vblank->count);
+	smp_mb__after_atomic();
 }
 
 /*
@@ -71,10 +151,11 @@ static void clear_vblank_timestamps(struct drm_device *dev, int crtc)
  */
 static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 	u32 vblcount;
 	s64 diff_ns;
-	int vblrc;
+	bool vblrc;
 	struct timeval tvblank;
 	int count = DRM_TIMESTAMP_MAXRETRIES;
 
@@ -84,8 +165,28 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 */
 	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
 
+	/*
+	 * If the vblank interrupt was already disbled update the count
+	 * and timestamp to maintain the appearance that the counter
+	 * has been ticking all along until this time. This makes the
+	 * count account for the entire time between drm_vblank_on() and
+	 * drm_vblank_off().
+	 *
+	 * But only do this if precise vblank timestamps are available.
+	 * Otherwise we might read a totally bogus timestamp since drivers
+	 * lacking precise timestamp support rely upon sampling the system clock
+	 * at vblank interrupt time. Which obviously won't work out well if the
+	 * vblank interrupt is disabled.
+	 */
+	if (!vblank->enabled &&
+	    drm_get_last_vbltimestamp(dev, crtc, &tvblank, 0)) {
+		drm_update_vblank_count(dev, crtc);
+		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+		return;
+	}
+
 	dev->driver->disable_vblank(dev, crtc);
-	dev->vblank[crtc].enabled = false;
+	vblank->enabled = false;
 
 	/* No further vblank irq's will be processed after
 	 * this point. Get current hardware vblank count and
@@ -100,9 +201,9 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * delayed gpu counter increment.
 	 */
 	do {
-		dev->vblank[crtc].last = dev->driver->get_vblank_counter(dev, crtc);
+		vblank->last = dev->driver->get_vblank_counter(dev, crtc);
 		vblrc = drm_get_last_vbltimestamp(dev, crtc, &tvblank, 0);
-	} while (dev->vblank[crtc].last != dev->driver->get_vblank_counter(dev, crtc) && (--count) && vblrc);
+	} while (vblank->last != dev->driver->get_vblank_counter(dev, crtc) && (--count) && vblrc);
 
 	if (!count)
 		vblrc = 0;
@@ -110,7 +211,7 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	/* Compute time difference to stored timestamp of last vblank
 	 * as updated by last invocation of drm_handle_vblank() in vblank irq.
 	 */
-	vblcount = atomic_read(&dev->vblank[crtc].count);
+	vblcount = atomic_read(&vblank->count);
 	diff_ns = timeval_to_ns(&tvblank) -
 		  timeval_to_ns(&vblanktimestamp(dev, crtc, vblcount));
 
@@ -126,14 +227,18 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc)
 	 * available. In that case we can't account for this and just
 	 * hope for the best.
 	 */
-	if ((vblrc > 0) && (abs64(diff_ns) > 1000000)) {
-		atomic_inc(&dev->vblank[crtc].count);
+	if (vblrc && (abs64(diff_ns) > 1000000)) {
+		/* Store new timestamp in ringbuffer. */
+		vblanktimestamp(dev, crtc, vblcount + 1) = tvblank;
+
+		/* Increment cooked vblank count. This also atomically commits
+		 * the timestamp computed above.
+		 */
+		smp_mb__before_atomic();
+		atomic_inc(&vblank->count);
 		smp_mb__after_atomic();
 	}
 
-	/* Invalidate all timestamps while vblank irq's are off. */
-	clear_vblank_timestamps(dev, crtc);
-
 	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
 }
 
@@ -164,14 +269,20 @@ static void vblank_disable_fn(unsigned long arg)
 void drm_vblank_cleanup(struct drm_device *dev)
 {
 	int crtc;
+	unsigned long irqflags;
 
 	/* Bail if the driver didn't call drm_vblank_init() */
 	if (dev->num_crtcs == 0)
 		return;
 
 	for (crtc = 0; crtc < dev->num_crtcs; crtc++) {
-		del_timer_sync(&dev->vblank[crtc].disable_timer);
-		vblank_disable_fn((unsigned long)&dev->vblank[crtc]);
+		struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+		del_timer_sync(&vblank->disable_timer);
+
+		spin_lock_irqsave(&dev->vbl_lock, irqflags);
+		vblank_disable_and_save(dev, crtc);
+		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 	}
 
 	kfree(dev->vblank);
@@ -204,11 +315,13 @@ int drm_vblank_init(struct drm_device *dev, int num_crtcs)
 		goto err;
 
 	for (i = 0; i < num_crtcs; i++) {
-		dev->vblank[i].dev = dev;
-		dev->vblank[i].crtc = i;
-		init_waitqueue_head(&dev->vblank[i].queue);
-		setup_timer(&dev->vblank[i].disable_timer, vblank_disable_fn,
-			    (unsigned long)&dev->vblank[i]);
+		struct drm_vblank_crtc *vblank = &dev->vblank[i];
+
+		vblank->dev = dev;
+		vblank->crtc = i;
+		init_waitqueue_head(&vblank->queue);
+		setup_timer(&vblank->disable_timer, vblank_disable_fn,
+			    (unsigned long)vblank);
 	}
 
 	DRM_INFO("Supports vblank timestamp caching Rev 2 (21.10.2013).\n");
@@ -224,7 +337,7 @@ int drm_vblank_init(struct drm_device *dev, int num_crtcs)
 	return 0;
 
 err:
-	drm_vblank_cleanup(dev);
+	dev->num_crtcs = 0;
 	return ret;
 }
 EXPORT_SYMBOL(drm_vblank_init);
@@ -360,9 +473,11 @@ int drm_irq_uninstall(struct drm_device *dev)
 	if (dev->num_crtcs) {
 		spin_lock_irqsave(&dev->vbl_lock, irqflags);
 		for (i = 0; i < dev->num_crtcs; i++) {
-			wake_up(&dev->vblank[i].queue);
-			dev->vblank[i].enabled = false;
-			dev->vblank[i].last =
+			struct drm_vblank_crtc *vblank = &dev->vblank[i];
+
+			wake_up(&vblank->queue);
+			vblank->enabled = false;
+			vblank->last =
 				dev->driver->get_vblank_counter(dev, i);
 		}
 		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
@@ -617,7 +732,7 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 	 * within vblank area, counting down the number of lines until
 	 * start of scanout.
 	 */
-	invbl = vbl_status & DRM_SCANOUTPOS_INVBL;
+	invbl = vbl_status & DRM_SCANOUTPOS_IN_VBLANK;
 
 	/* Convert scanout position into elapsed time at raw_time query
 	 * since start of scanout at first display scanline. delta_ns
@@ -647,7 +762,7 @@ int drm_calc_vbltimestamp_from_scanoutpos(struct drm_device *dev, int crtc,
 
 	vbl_status = DRM_VBLANKTIME_SCANOUTPOS_METHOD;
 	if (invbl)
-		vbl_status |= DRM_VBLANKTIME_INVBL;
+		vbl_status |= DRM_VBLANKTIME_IN_VBLANK;
 
 	return vbl_status;
 }
@@ -679,10 +794,11 @@ static struct timeval get_drm_timestamp(void)
  * call, i.e., it isn't very precisely locked to the true vblank.
  *
  * Returns:
- * Non-zero if timestamp is considered to be very precise, zero otherwise.
+ * True if timestamp is considered to be very precise, false otherwise.
  */
-u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
-			      struct timeval *tvblank, unsigned flags)
+static bool
+drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
+			  struct timeval *tvblank, unsigned flags)
 {
 	int ret;
 
@@ -694,7 +810,7 @@ u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
 		ret = dev->driver->get_vblank_timestamp(dev, crtc, &max_error,
 							tvblank, flags);
 		if (ret > 0)
-			return (u32) ret;
+			return true;
 	}
 
 	/* GPU high precision timestamp query unsupported or failed.
@@ -702,9 +818,8 @@ u32 drm_get_last_vbltimestamp(struct drm_device *dev, int crtc,
 	 */
 	*tvblank = get_drm_timestamp();
 
-	return 0;
+	return false;
 }
-EXPORT_SYMBOL(drm_get_last_vbltimestamp);
 
 /**
  * drm_vblank_count - retrieve "cooked" vblank counter value
@@ -720,7 +835,11 @@ EXPORT_SYMBOL(drm_get_last_vbltimestamp);
  */
 u32 drm_vblank_count(struct drm_device *dev, int crtc)
 {
-	return atomic_read(&dev->vblank[crtc].count);
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return 0;
+	return atomic_read(&vblank->count);
 }
 EXPORT_SYMBOL(drm_vblank_count);
 
@@ -740,18 +859,22 @@ EXPORT_SYMBOL(drm_vblank_count);
 u32 drm_vblank_count_and_time(struct drm_device *dev, int crtc,
 			      struct timeval *vblanktime)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	u32 cur_vblank;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return 0;
+
 	/* Read timestamp from slot of _vblank_time ringbuffer
 	 * that corresponds to current vblank count. Retry if
 	 * count has incremented during readout. This works like
 	 * a seqlock.
 	 */
 	do {
-		cur_vblank = atomic_read(&dev->vblank[crtc].count);
+		cur_vblank = atomic_read(&vblank->count);
 		*vblanktime = vblanktimestamp(dev, crtc, cur_vblank);
 		smp_rmb();
-	} while (cur_vblank != atomic_read(&dev->vblank[crtc].count));
+	} while (cur_vblank != atomic_read(&vblank->count));
 
 	return cur_vblank;
 }
@@ -800,83 +923,20 @@ void drm_send_vblank_event(struct drm_device *dev, int crtc,
 EXPORT_SYMBOL(drm_send_vblank_event);
 
 /**
- * drm_update_vblank_count - update the master vblank counter
- * @dev: DRM device
- * @crtc: counter to update
- *
- * Call back into the driver to update the appropriate vblank counter
- * (specified by @crtc).  Deal with wraparound, if it occurred, and
- * update the last read value so we can deal with wraparound on the next
- * call if necessary.
- *
- * Only necessary when going from off->on, to account for frames we
- * didn't get an interrupt for.
- *
- * Note: caller must hold dev->vbl_lock since this reads & writes
- * device vblank fields.
- */
-static void drm_update_vblank_count(struct drm_device *dev, int crtc)
-{
-	u32 cur_vblank, diff, tslot, rc;
-	struct timeval t_vblank;
-
-	/*
-	 * Interrupts were disabled prior to this call, so deal with counter
-	 * wrap if needed.
-	 * NOTE!  It's possible we lost a full dev->max_vblank_count events
-	 * here if the register is small or we had vblank interrupts off for
-	 * a long time.
-	 *
-	 * We repeat the hardware vblank counter & timestamp query until
-	 * we get consistent results. This to prevent races between gpu
-	 * updating its hardware counter while we are retrieving the
-	 * corresponding vblank timestamp.
-	 */
-	do {
-		cur_vblank = dev->driver->get_vblank_counter(dev, crtc);
-		rc = drm_get_last_vbltimestamp(dev, crtc, &t_vblank, 0);
-	} while (cur_vblank != dev->driver->get_vblank_counter(dev, crtc));
-
-	/* Deal with counter wrap */
-	diff = cur_vblank - dev->vblank[crtc].last;
-	if (cur_vblank < dev->vblank[crtc].last) {
-		diff += dev->max_vblank_count;
-
-		DRM_DEBUG("last_vblank[%d]=0x%x, cur_vblank=0x%x => diff=0x%x\n",
-			  crtc, dev->vblank[crtc].last, cur_vblank, diff);
-	}
-
-	DRM_DEBUG("enabling vblank interrupts on crtc %d, missed %d\n",
-		  crtc, diff);
-
-	/* Reinitialize corresponding vblank timestamp if high-precision query
-	 * available. Skip this step if query unsupported or failed. Will
-	 * reinitialize delayed at next vblank interrupt in that case.
-	 */
-	if (rc) {
-		tslot = atomic_read(&dev->vblank[crtc].count) + diff;
-		vblanktimestamp(dev, crtc, tslot) = t_vblank;
-	}
-
-	smp_mb__before_atomic();
-	atomic_add(diff, &dev->vblank[crtc].count);
-	smp_mb__after_atomic();
-}
-
-/**
  * drm_vblank_enable - enable the vblank interrupt on a CRTC
  * @dev: DRM device
  * @crtc: CRTC in question
  */
 static int drm_vblank_enable(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	int ret = 0;
 
 	assert_spin_locked(&dev->vbl_lock);
 
 	spin_lock(&dev->vblank_time_lock);
 
-	if (!dev->vblank[crtc].enabled) {
+	if (!vblank->enabled) {
 		/*
 		 * Enable vblank irqs under vblank_time_lock protection.
 		 * All vblank count & timestamp updates are held off
@@ -887,9 +947,9 @@ static int drm_vblank_enable(struct drm_device *dev, int crtc)
 		ret = dev->driver->enable_vblank(dev, crtc);
 		DRM_DEBUG("enabling vblank on crtc %d, ret: %d\n", crtc, ret);
 		if (ret)
-			atomic_dec(&dev->vblank[crtc].refcount);
+			atomic_dec(&vblank->refcount);
 		else {
-			dev->vblank[crtc].enabled = true;
+			vblank->enabled = true;
 			drm_update_vblank_count(dev, crtc);
 		}
 	}
@@ -914,16 +974,20 @@ static int drm_vblank_enable(struct drm_device *dev, int crtc)
  */
 int drm_vblank_get(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 	int ret = 0;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return -EINVAL;
+
 	spin_lock_irqsave(&dev->vbl_lock, irqflags);
 	/* Going from 0->1 means we have to enable interrupts again */
-	if (atomic_add_return(1, &dev->vblank[crtc].refcount) == 1) {
+	if (atomic_add_return(1, &vblank->refcount) == 1) {
 		ret = drm_vblank_enable(dev, crtc);
 	} else {
-		if (!dev->vblank[crtc].enabled) {
-			atomic_dec(&dev->vblank[crtc].refcount);
+		if (!vblank->enabled) {
+			atomic_dec(&vblank->refcount);
 			ret = -EINVAL;
 		}
 	}
@@ -963,13 +1027,23 @@ EXPORT_SYMBOL(drm_crtc_vblank_get);
  */
 void drm_vblank_put(struct drm_device *dev, int crtc)
 {
-	BUG_ON(atomic_read(&dev->vblank[crtc].refcount) == 0);
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
+	BUG_ON(atomic_read(&vblank->refcount) == 0);
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
 
 	/* Last user schedules interrupt disable */
-	if (atomic_dec_and_test(&dev->vblank[crtc].refcount) &&
-	    (drm_vblank_offdelay > 0))
-		mod_timer(&dev->vblank[crtc].disable_timer,
-			  jiffies + ((drm_vblank_offdelay * HZ)/1000));
+	if (atomic_dec_and_test(&vblank->refcount)) {
+		if (drm_vblank_offdelay == 0)
+			return;
+		else if (dev->vblank_disable_immediate || drm_vblank_offdelay < 0)
+			vblank_disable_fn((unsigned long)vblank);
+		else
+			mod_timer(&vblank->disable_timer,
+				  jiffies + ((drm_vblank_offdelay * HZ)/1000));
+	}
 }
 EXPORT_SYMBOL(drm_vblank_put);
 
@@ -989,6 +1063,50 @@ void drm_crtc_vblank_put(struct drm_crtc *crtc)
 EXPORT_SYMBOL(drm_crtc_vblank_put);
 
 /**
+ * drm_wait_one_vblank - wait for one vblank
+ * @dev: DRM device
+ * @crtc: crtc index
+ *
+ * This waits for one vblank to pass on @crtc, using the irq driver interfaces.
+ * It is a failure to call this when the vblank irq for @crtc is disabled, e.g.
+ * due to lack of driver support or because the crtc is off.
+ */
+void drm_wait_one_vblank(struct drm_device *dev, int crtc)
+{
+	int ret;
+	u32 last;
+
+	ret = drm_vblank_get(dev, crtc);
+	if (WARN(ret, "vblank not available on crtc %i, ret=%i\n", crtc, ret))
+		return;
+
+	last = drm_vblank_count(dev, crtc);
+
+	ret = wait_event_timeout(dev->vblank[crtc].queue,
+				 last != drm_vblank_count(dev, crtc),
+				 msecs_to_jiffies(100));
+
+	WARN(ret == 0, "vblank wait timed out on crtc %i\n", crtc);
+
+	drm_vblank_put(dev, crtc);
+}
+EXPORT_SYMBOL(drm_wait_one_vblank);
+
+/**
+ * drm_crtc_wait_one_vblank - wait for one vblank
+ * @crtc: DRM crtc
+ *
+ * This waits for one vblank to pass on @crtc, using the irq driver interfaces.
+ * It is a failure to call this when the vblank irq for @crtc is disabled, e.g.
+ * due to lack of driver support or because the crtc is off.
+ */
+void drm_crtc_wait_one_vblank(struct drm_crtc *crtc)
+{
+	drm_wait_one_vblank(crtc->dev, drm_crtc_index(crtc));
+}
+EXPORT_SYMBOL(drm_crtc_wait_one_vblank);
+
+/**
  * drm_vblank_off - disable vblank events on a CRTC
  * @dev: DRM device
  * @crtc: CRTC in question
@@ -1004,19 +1122,34 @@ EXPORT_SYMBOL(drm_crtc_vblank_put);
  */
 void drm_vblank_off(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	struct drm_pending_vblank_event *e, *t;
 	struct timeval now;
 	unsigned long irqflags;
 	unsigned int seq;
 
-	spin_lock_irqsave(&dev->vbl_lock, irqflags);
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
+	spin_lock_irqsave(&dev->event_lock, irqflags);
+
+	spin_lock(&dev->vbl_lock);
 	vblank_disable_and_save(dev, crtc);
-	wake_up(&dev->vblank[crtc].queue);
+	wake_up(&vblank->queue);
+
+	/*
+	 * Prevent subsequent drm_vblank_get() from re-enabling
+	 * the vblank interrupt by bumping the refcount.
+	 */
+	if (!vblank->inmodeset) {
+		atomic_inc(&vblank->refcount);
+		vblank->inmodeset = 1;
+	}
+	spin_unlock(&dev->vbl_lock);
 
 	/* Send any queued vblank events, lest the natives grow disquiet */
 	seq = drm_vblank_count_and_time(dev, crtc, &now);
 
-	spin_lock(&dev->event_lock);
 	list_for_each_entry_safe(e, t, &dev->vblank_event_list, base.link) {
 		if (e->pipe != crtc)
 			continue;
@@ -1027,9 +1160,7 @@ void drm_vblank_off(struct drm_device *dev, int crtc)
 		drm_vblank_put(dev, e->pipe);
 		send_vblank_event(dev, e, seq, &now);
 	}
-	spin_unlock(&dev->event_lock);
-
-	spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
+	spin_unlock_irqrestore(&dev->event_lock, irqflags);
 }
 EXPORT_SYMBOL(drm_vblank_off);
 
@@ -1066,11 +1197,35 @@ EXPORT_SYMBOL(drm_crtc_vblank_off);
  */
 void drm_vblank_on(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
 	spin_lock_irqsave(&dev->vbl_lock, irqflags);
-	/* re-enable interrupts if there's are users left */
-	if (atomic_read(&dev->vblank[crtc].refcount) != 0)
+	/* Drop our private "prevent drm_vblank_get" refcount */
+	if (vblank->inmodeset) {
+		atomic_dec(&vblank->refcount);
+		vblank->inmodeset = 0;
+	}
+
+	/*
+	 * sample the current counter to avoid random jumps
+	 * when drm_vblank_enable() applies the diff
+	 *
+	 * -1 to make sure user will never see the same
+	 * vblank counter value before and after a modeset
+	 */
+	vblank->last =
+		(dev->driver->get_vblank_counter(dev, crtc) - 1) &
+		dev->max_vblank_count;
+	/*
+	 * re-enable interrupts if there are users left, or the
+	 * user wishes vblank interrupts to be enabled all the time.
+	 */
+	if (atomic_read(&vblank->refcount) != 0 ||
+	    (!dev->vblank_disable_immediate && drm_vblank_offdelay == 0))
 		WARN_ON(drm_vblank_enable(dev, crtc));
 	spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 }
@@ -1118,9 +1273,15 @@ EXPORT_SYMBOL(drm_crtc_vblank_on);
  */
 void drm_vblank_pre_modeset(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
+
 	/* vblank is not initialized (IRQ not installed ?), or has been freed */
 	if (!dev->num_crtcs)
 		return;
+
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return;
+
 	/*
 	 * To avoid all the problems that might happen if interrupts
 	 * were enabled/disabled around or between these calls, we just
@@ -1128,10 +1289,10 @@ void drm_vblank_pre_modeset(struct drm_device *dev, int crtc)
 	 * to avoid corrupting the count if multiple, mismatch calls occur),
 	 * so that interrupts remain enabled in the interim.
 	 */
-	if (!dev->vblank[crtc].inmodeset) {
-		dev->vblank[crtc].inmodeset = 0x1;
+	if (!vblank->inmodeset) {
+		vblank->inmodeset = 0x1;
 		if (drm_vblank_get(dev, crtc) == 0)
-			dev->vblank[crtc].inmodeset |= 0x2;
+			vblank->inmodeset |= 0x2;
 	}
 }
 EXPORT_SYMBOL(drm_vblank_pre_modeset);
@@ -1146,21 +1307,22 @@ EXPORT_SYMBOL(drm_vblank_pre_modeset);
  */
 void drm_vblank_post_modeset(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	unsigned long irqflags;
 
 	/* vblank is not initialized (IRQ not installed ?), or has been freed */
 	if (!dev->num_crtcs)
 		return;
 
-	if (dev->vblank[crtc].inmodeset) {
+	if (vblank->inmodeset) {
 		spin_lock_irqsave(&dev->vbl_lock, irqflags);
 		dev->vblank_disable_allowed = true;
 		spin_unlock_irqrestore(&dev->vbl_lock, irqflags);
 
-		if (dev->vblank[crtc].inmodeset & 0x2)
+		if (vblank->inmodeset & 0x2)
 			drm_vblank_put(dev, crtc);
 
-		dev->vblank[crtc].inmodeset = 0;
+		vblank->inmodeset = 0;
 	}
 }
 EXPORT_SYMBOL(drm_vblank_post_modeset);
@@ -1212,6 +1374,7 @@ static int drm_queue_vblank_event(struct drm_device *dev, int pipe,
 				  union drm_wait_vblank *vblwait,
 				  struct drm_file *file_priv)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
 	struct drm_pending_vblank_event *e;
 	struct timeval now;
 	unsigned long flags;
@@ -1235,6 +1398,18 @@ static int drm_queue_vblank_event(struct drm_device *dev, int pipe,
 
 	spin_lock_irqsave(&dev->event_lock, flags);
 
+	/*
+	 * drm_vblank_off() might have been called after we called
+	 * drm_vblank_get(). drm_vblank_off() holds event_lock
+	 * around the vblank disable, so no need for further locking.
+	 * The reference from drm_vblank_get() protects against
+	 * vblank disable from another source.
+	 */
+	if (!vblank->enabled) {
+		ret = -EINVAL;
+		goto err_unlock;
+	}
+
 	if (file_priv->event_space < sizeof e->event) {
 		ret = -EBUSY;
 		goto err_unlock;
@@ -1295,6 +1470,7 @@ err_put:
 int drm_wait_vblank(struct drm_device *dev, void *data,
 		    struct drm_file *file_priv)
 {
+	struct drm_vblank_crtc *vblank;
 	union drm_wait_vblank *vblwait = data;
 	int ret;
 	unsigned int flags, seq, crtc, high_crtc;
@@ -1324,6 +1500,8 @@ int drm_wait_vblank(struct drm_device *dev, void *data,
 	if (crtc >= dev->num_crtcs)
 		return -EINVAL;
 
+	vblank = &dev->vblank[crtc];
+
 	ret = drm_vblank_get(dev, crtc);
 	if (ret) {
 		DRM_DEBUG("failed to acquire vblank counter, %d\n", ret);
@@ -1356,11 +1534,11 @@ int drm_wait_vblank(struct drm_device *dev, void *data,
 
 	DRM_DEBUG("waiting on vblank count %d, crtc %d\n",
 		  vblwait->request.sequence, crtc);
-	dev->vblank[crtc].last_wait = vblwait->request.sequence;
-	DRM_WAIT_ON(ret, dev->vblank[crtc].queue, 3 * HZ,
+	vblank->last_wait = vblwait->request.sequence;
+	DRM_WAIT_ON(ret, vblank->queue, 3 * HZ,
 		    (((drm_vblank_count(dev, crtc) -
 		       vblwait->request.sequence) <= (1 << 23)) ||
-		     !dev->vblank[crtc].enabled ||
+		     !vblank->enabled ||
 		     !dev->irq_enabled));
 
 	if (ret != -EINTR) {
@@ -1385,12 +1563,11 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
 {
 	struct drm_pending_vblank_event *e, *t;
 	struct timeval now;
-	unsigned long flags;
 	unsigned int seq;
 
-	seq = drm_vblank_count_and_time(dev, crtc, &now);
+	assert_spin_locked(&dev->event_lock);
 
-	spin_lock_irqsave(&dev->event_lock, flags);
+	seq = drm_vblank_count_and_time(dev, crtc, &now);
 
 	list_for_each_entry_safe(e, t, &dev->vblank_event_list, base.link) {
 		if (e->pipe != crtc)
@@ -1406,8 +1583,6 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
 		send_vblank_event(dev, e, seq, &now);
 	}
 
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-
 	trace_drm_vblank_event(crtc, seq);
 }
 
@@ -1421,6 +1596,7 @@ static void drm_handle_vblank_events(struct drm_device *dev, int crtc)
  */
 bool drm_handle_vblank(struct drm_device *dev, int crtc)
 {
+	struct drm_vblank_crtc *vblank = &dev->vblank[crtc];
 	u32 vblcount;
 	s64 diff_ns;
 	struct timeval tvblank;
@@ -1429,15 +1605,21 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	if (!dev->num_crtcs)
 		return false;
 
+	if (WARN_ON(crtc >= dev->num_crtcs))
+		return false;
+
+	spin_lock_irqsave(&dev->event_lock, irqflags);
+
 	/* Need timestamp lock to prevent concurrent execution with
 	 * vblank enable/disable, as this would cause inconsistent
 	 * or corrupted timestamps and vblank counts.
 	 */
-	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
+	spin_lock(&dev->vblank_time_lock);
 
 	/* Vblank irq handling disabled. Nothing to do. */
-	if (!dev->vblank[crtc].enabled) {
-		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+	if (!vblank->enabled) {
+		spin_unlock(&dev->vblank_time_lock);
+		spin_unlock_irqrestore(&dev->event_lock, irqflags);
 		return false;
 	}
 
@@ -1446,7 +1628,7 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 	 */
 
 	/* Get current timestamp and count. */
-	vblcount = atomic_read(&dev->vblank[crtc].count);
+	vblcount = atomic_read(&vblank->count);
 	drm_get_last_vbltimestamp(dev, crtc, &tvblank, DRM_CALLED_FROM_VBLIRQ);
 
 	/* Compute time difference to timestamp of last vblank */
@@ -1470,17 +1652,20 @@ bool drm_handle_vblank(struct drm_device *dev, int crtc)
 		 * the timestamp computed above.
 		 */
 		smp_mb__before_atomic();
-		atomic_inc(&dev->vblank[crtc].count);
+		atomic_inc(&vblank->count);
 		smp_mb__after_atomic();
 	} else {
 		DRM_DEBUG("crtc %d: Redundant vblirq ignored. diff_ns = %d\n",
 			  crtc, (int) diff_ns);
 	}
 
-	wake_up(&dev->vblank[crtc].queue);
+	spin_unlock(&dev->vblank_time_lock);
+
+	wake_up(&vblank->queue);
 	drm_handle_vblank_events(dev, crtc);
 
-	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
+	spin_unlock_irqrestore(&dev->event_lock, irqflags);
+
 	return true;
 }
 EXPORT_SYMBOL(drm_handle_vblank);
diff --git a/drivers/gpu/drm/drm_legacy.h b/drivers/gpu/drm/drm_legacy.h
new file mode 100644
index 000000000000..c1dc61473db5
--- /dev/null
+++ b/drivers/gpu/drm/drm_legacy.h
@@ -0,0 +1,113 @@
+#ifndef __DRM_LEGACY_H__
+#define __DRM_LEGACY_H__
+
+/*
+ * Copyright (c) 2014 David Herrmann <dh.herrmann@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * This file contains legacy interfaces that modern drm drivers
+ * should no longer be using. They cannot be removed as legacy
+ * drivers use them, and removing them are API breaks.
+ */
+#include <linux/list.h>
+#include <drm/drm_legacy.h>
+
+struct agp_memory;
+struct drm_device;
+struct drm_file;
+
+/*
+ * Generic DRM Contexts
+ */
+
+#define DRM_KERNEL_CONTEXT		0
+#define DRM_RESERVED_CONTEXTS		1
+
+int drm_legacy_ctxbitmap_init(struct drm_device *dev);
+void drm_legacy_ctxbitmap_cleanup(struct drm_device *dev);
+void drm_legacy_ctxbitmap_free(struct drm_device *dev, int ctx_handle);
+void drm_legacy_ctxbitmap_flush(struct drm_device *dev, struct drm_file *file);
+
+int drm_legacy_resctx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_addctx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_getctx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_switchctx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_newctx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_rmctx(struct drm_device *d, void *v, struct drm_file *f);
+
+int drm_legacy_setsareactx(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_getsareactx(struct drm_device *d, void *v, struct drm_file *f);
+
+/*
+ * Generic Buffer Management
+ */
+
+#define DRM_MAP_HASH_OFFSET 0x10000000
+
+int drm_legacy_addmap_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_rmmap_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_addbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_infobufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_markbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_freebufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_mapbufs(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_dma_ioctl(struct drm_device *d, void *v, struct drm_file *f);
+
+void drm_legacy_vma_flush(struct drm_device *d);
+
+/*
+ * AGP Support
+ */
+
+struct drm_agp_mem {
+	unsigned long handle;
+	struct agp_memory *memory;
+	unsigned long bound;
+	int pages;
+	struct list_head head;
+};
+
+/*
+ * Generic Userspace Locking-API
+ */
+
+int drm_legacy_i_have_hw_lock(struct drm_device *d, struct drm_file *f);
+int drm_legacy_lock(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_unlock(struct drm_device *d, void *v, struct drm_file *f);
+int drm_legacy_lock_free(struct drm_lock_data *lock, unsigned int ctx);
+
+/* DMA support */
+int drm_legacy_dma_setup(struct drm_device *dev);
+void drm_legacy_dma_takedown(struct drm_device *dev);
+void drm_legacy_free_buffer(struct drm_device *dev,
+			    struct drm_buf * buf);
+void drm_legacy_reclaim_buffers(struct drm_device *dev,
+				struct drm_file *filp);
+
+/* Scatter Gather Support */
+void drm_legacy_sg_cleanup(struct drm_device *dev);
+int drm_legacy_sg_alloc(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+int drm_legacy_sg_free(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv);
+
+#endif /* __DRM_LEGACY_H__ */
diff --git a/drivers/gpu/drm/drm_lock.c b/drivers/gpu/drm/drm_lock.c
index f6452682141b..f861361a635e 100644
--- a/drivers/gpu/drm/drm_lock.c
+++ b/drivers/gpu/drm/drm_lock.c
@@ -35,6 +35,8 @@
 
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
+#include "drm_internal.h"
 
 static int drm_notifier(void *priv);
 
@@ -51,7 +53,8 @@ static int drm_lock_take(struct drm_lock_data *lock_data, unsigned int context);
  *
  * Add the current task to the lock wait queue, and attempt to take to lock.
  */
-int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
+int drm_legacy_lock(struct drm_device *dev, void *data,
+		    struct drm_file *file_priv)
 {
 	DECLARE_WAITQUEUE(entry, current);
 	struct drm_lock *lock = data;
@@ -119,7 +122,7 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 		sigaddset(&dev->sigmask, SIGTTOU);
 		dev->sigdata.context = lock->context;
 		dev->sigdata.lock = master->lock.hw_lock;
-		block_all_signals(drm_notifier, &dev->sigdata, &dev->sigmask);
+		block_all_signals(drm_notifier, dev, &dev->sigmask);
 	}
 
 	if (dev->driver->dma_quiescent && (lock->flags & _DRM_LOCK_QUIESCENT))
@@ -145,7 +148,7 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv)
  *
  * Transfer and free the lock.
  */
-int drm_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
+int drm_legacy_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 {
 	struct drm_lock *lock = data;
 	struct drm_master *master = file_priv->master;
@@ -156,7 +159,7 @@ int drm_unlock(struct drm_device *dev, void *data, struct drm_file *file_priv)
 		return -EINVAL;
 	}
 
-	if (drm_lock_free(&master->lock, lock->context)) {
+	if (drm_legacy_lock_free(&master->lock, lock->context)) {
 		/* FIXME: Should really bail out here. */
 	}
 
@@ -249,7 +252,7 @@ static int drm_lock_transfer(struct drm_lock_data *lock_data,
  * Marks the lock as not held, via the \p cmpxchg instruction. Wakes any task
  * waiting on the lock queue.
  */
-int drm_lock_free(struct drm_lock_data *lock_data, unsigned int context)
+int drm_legacy_lock_free(struct drm_lock_data *lock_data, unsigned int context)
 {
 	unsigned int old, new, prev;
 	volatile unsigned int *lock = &lock_data->hw_lock->lock;
@@ -285,26 +288,27 @@ int drm_lock_free(struct drm_lock_data *lock_data, unsigned int context)
  * If the lock is not held, then let the signal proceed as usual.  If the lock
  * is held, then set the contended flag and keep the signal blocked.
  *
- * \param priv pointer to a drm_sigdata structure.
+ * \param priv pointer to a drm_device structure.
  * \return one if the signal should be delivered normally, or zero if the
  * signal should be blocked.
  */
 static int drm_notifier(void *priv)
 {
-	struct drm_sigdata *s = (struct drm_sigdata *) priv;
+	struct drm_device *dev = priv;
+	struct drm_hw_lock *lock = dev->sigdata.lock;
 	unsigned int old, new, prev;
 
 	/* Allow signal delivery if lock isn't held */
-	if (!s->lock || !_DRM_LOCK_IS_HELD(s->lock->lock)
-	    || _DRM_LOCKING_CONTEXT(s->lock->lock) != s->context)
+	if (!lock || !_DRM_LOCK_IS_HELD(lock->lock)
+	    || _DRM_LOCKING_CONTEXT(lock->lock) != dev->sigdata.context)
 		return 1;
 
 	/* Otherwise, set flag to force call to
 	   drmUnlock */
 	do {
-		old = s->lock->lock;
+		old = lock->lock;
 		new = old | _DRM_LOCK_CONT;
-		prev = cmpxchg(&s->lock->lock, old, new);
+		prev = cmpxchg(&lock->lock, old, new);
 	} while (prev != old);
 	return 0;
 }
@@ -322,7 +326,7 @@ static int drm_notifier(void *priv)
  * having to worry about starvation.
  */
 
-void drm_idlelock_take(struct drm_lock_data *lock_data)
+void drm_legacy_idlelock_take(struct drm_lock_data *lock_data)
 {
 	int ret;
 
@@ -339,9 +343,9 @@ void drm_idlelock_take(struct drm_lock_data *lock_data)
 	}
 	spin_unlock_bh(&lock_data->spinlock);
 }
-EXPORT_SYMBOL(drm_idlelock_take);
+EXPORT_SYMBOL(drm_legacy_idlelock_take);
 
-void drm_idlelock_release(struct drm_lock_data *lock_data)
+void drm_legacy_idlelock_release(struct drm_lock_data *lock_data)
 {
 	unsigned int old, prev;
 	volatile unsigned int *lock = &lock_data->hw_lock->lock;
@@ -359,9 +363,10 @@ void drm_idlelock_release(struct drm_lock_data *lock_data)
 	}
 	spin_unlock_bh(&lock_data->spinlock);
 }
-EXPORT_SYMBOL(drm_idlelock_release);
+EXPORT_SYMBOL(drm_legacy_idlelock_release);
 
-int drm_i_have_hw_lock(struct drm_device *dev, struct drm_file *file_priv)
+int drm_legacy_i_have_hw_lock(struct drm_device *dev,
+			      struct drm_file *file_priv)
 {
 	struct drm_master *master = file_priv->master;
 	return (file_priv->lock_count && master->lock.hw_lock &&
diff --git a/drivers/gpu/drm/drm_memory.c b/drivers/gpu/drm/drm_memory.c
index 00c67c0f2381..a521ef6ff807 100644
--- a/drivers/gpu/drm/drm_memory.c
+++ b/drivers/gpu/drm/drm_memory.c
@@ -36,8 +36,20 @@
 #include <linux/highmem.h>
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 #if __OS_HAS_AGP
+
+#ifdef HAVE_PAGE_AGP
+# include <asm/agp.h>
+#else
+# ifdef __powerpc__
+#  define PAGE_AGP	__pgprot(_PAGE_KERNEL | _PAGE_NO_CACHE)
+# else
+#  define PAGE_AGP	PAGE_KERNEL
+# endif
+#endif
+
 static void *agp_remap(unsigned long offset, unsigned long size,
 		       struct drm_device * dev)
 {
@@ -108,25 +120,25 @@ static inline void *agp_remap(unsigned long offset, unsigned long size,
 
 #endif				/* agp */
 
-void drm_core_ioremap(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremap(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (dev->agp && dev->agp->cant_use_aperture && map->type == _DRM_AGP)
 		map->handle = agp_remap(map->offset, map->size, dev);
 	else
 		map->handle = ioremap(map->offset, map->size);
 }
-EXPORT_SYMBOL(drm_core_ioremap);
+EXPORT_SYMBOL(drm_legacy_ioremap);
 
-void drm_core_ioremap_wc(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremap_wc(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (dev->agp && dev->agp->cant_use_aperture && map->type == _DRM_AGP)
 		map->handle = agp_remap(map->offset, map->size, dev);
 	else
 		map->handle = ioremap_wc(map->offset, map->size);
 }
-EXPORT_SYMBOL(drm_core_ioremap_wc);
+EXPORT_SYMBOL(drm_legacy_ioremap_wc);
 
-void drm_core_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
+void drm_legacy_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
 {
 	if (!map->handle || !map->size)
 		return;
@@ -136,4 +148,4 @@ void drm_core_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
 	else
 		iounmap(map->handle);
 }
-EXPORT_SYMBOL(drm_core_ioremapfree);
+EXPORT_SYMBOL(drm_legacy_ioremapfree);
diff --git a/drivers/gpu/drm/drm_mipi_dsi.c b/drivers/gpu/drm/drm_mipi_dsi.c
index e633df2f68d8..eb6dfe52cab2 100644
--- a/drivers/gpu/drm/drm_mipi_dsi.c
+++ b/drivers/gpu/drm/drm_mipi_dsi.c
@@ -201,16 +201,15 @@ EXPORT_SYMBOL(mipi_dsi_detach);
 /**
  * mipi_dsi_dcs_write - send DCS write command
  * @dsi: DSI device
- * @channel: virtual channel
  * @data: pointer to the command followed by parameters
  * @len: length of @data
  */
-int mipi_dsi_dcs_write(struct mipi_dsi_device *dsi, unsigned int channel,
-		       const void *data, size_t len)
+ssize_t mipi_dsi_dcs_write(struct mipi_dsi_device *dsi, const void *data,
+			    size_t len)
 {
 	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
 	struct mipi_dsi_msg msg = {
-		.channel = channel,
+		.channel = dsi->channel,
 		.tx_buf = data,
 		.tx_len = len
 	};
@@ -232,6 +231,9 @@ int mipi_dsi_dcs_write(struct mipi_dsi_device *dsi, unsigned int channel,
 		break;
 	}
 
+	if (dsi->mode_flags & MIPI_DSI_MODE_LPM)
+		msg.flags = MIPI_DSI_MSG_USE_LPM;
+
 	return ops->transfer(dsi->host, &msg);
 }
 EXPORT_SYMBOL(mipi_dsi_dcs_write);
@@ -239,19 +241,18 @@ EXPORT_SYMBOL(mipi_dsi_dcs_write);
 /**
  * mipi_dsi_dcs_read - send DCS read request command
  * @dsi: DSI device
- * @channel: virtual channel
  * @cmd: DCS read command
  * @data: pointer to read buffer
  * @len: length of @data
  *
  * Function returns number of read bytes or error code.
  */
-ssize_t mipi_dsi_dcs_read(struct mipi_dsi_device *dsi, unsigned int channel,
-			  u8 cmd, void *data, size_t len)
+ssize_t mipi_dsi_dcs_read(struct mipi_dsi_device *dsi, u8 cmd, void *data,
+			  size_t len)
 {
 	const struct mipi_dsi_host_ops *ops = dsi->host->ops;
 	struct mipi_dsi_msg msg = {
-		.channel = channel,
+		.channel = dsi->channel,
 		.type = MIPI_DSI_DCS_READ,
 		.tx_buf = &cmd,
 		.tx_len = 1,
@@ -262,6 +263,9 @@ ssize_t mipi_dsi_dcs_read(struct mipi_dsi_device *dsi, unsigned int channel,
 	if (!ops || !ops->transfer)
 		return -ENOSYS;
 
+	if (dsi->mode_flags & MIPI_DSI_MODE_LPM)
+		msg.flags = MIPI_DSI_MSG_USE_LPM;
+
 	return ops->transfer(dsi->host, &msg);
 }
 EXPORT_SYMBOL(mipi_dsi_dcs_read);
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index bedf1894e17e..d1b7d2006529 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -1259,6 +1259,7 @@ drm_mode_create_from_cmdline_mode(struct drm_device *dev,
 	if (!mode)
 		return NULL;
 
+	mode->type |= DRM_MODE_TYPE_USERDEF;
 	drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
 	return mode;
 }
diff --git a/drivers/gpu/drm/drm_modeset_lock.c b/drivers/gpu/drm/drm_modeset_lock.c
index 0dc57d5ecd10..474e4d12a2d8 100644
--- a/drivers/gpu/drm/drm_modeset_lock.c
+++ b/drivers/gpu/drm/drm_modeset_lock.c
@@ -35,7 +35,7 @@
  * of extra utility/tracking out of our acquire-ctx.  This is provided
  * by drm_modeset_lock / drm_modeset_acquire_ctx.
  *
- * For basic principles of ww_mutex, see: Documentation/ww-mutex-design.txt
+ * For basic principles of ww_mutex, see: Documentation/locking/ww-mutex-design.txt
  *
  * The basic usage pattern is to:
  *
@@ -57,6 +57,212 @@
 
 
 /**
+ * __drm_modeset_lock_all - internal helper to grab all modeset locks
+ * @dev: DRM device
+ * @trylock: trylock mode for atomic contexts
+ *
+ * This is a special version of drm_modeset_lock_all() which can also be used in
+ * atomic contexts. Then @trylock must be set to true.
+ *
+ * Returns:
+ * 0 on success or negative error code on failure.
+ */
+int __drm_modeset_lock_all(struct drm_device *dev,
+			   bool trylock)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_modeset_acquire_ctx *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx),
+		      trylock ? GFP_ATOMIC : GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	if (trylock) {
+		if (!mutex_trylock(&config->mutex))
+			return -EBUSY;
+	} else {
+		mutex_lock(&config->mutex);
+	}
+
+	drm_modeset_acquire_init(ctx, 0);
+	ctx->trylock_only = trylock;
+
+retry:
+	ret = drm_modeset_lock(&config->connection_mutex, ctx);
+	if (ret)
+		goto fail;
+	ret = drm_modeset_lock_all_crtcs(dev, ctx);
+	if (ret)
+		goto fail;
+
+	WARN_ON(config->acquire_ctx);
+
+	/* now we hold the locks, so now that it is safe, stash the
+	 * ctx for drm_modeset_unlock_all():
+	 */
+	config->acquire_ctx = ctx;
+
+	drm_warn_on_modeset_not_all_locked(dev);
+
+	return 0;
+
+fail:
+	if (ret == -EDEADLK) {
+		drm_modeset_backoff(ctx);
+		goto retry;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL(__drm_modeset_lock_all);
+
+/**
+ * drm_modeset_lock_all - take all modeset locks
+ * @dev: drm device
+ *
+ * This function takes all modeset locks, suitable where a more fine-grained
+ * scheme isn't (yet) implemented. Locks must be dropped with
+ * drm_modeset_unlock_all.
+ */
+void drm_modeset_lock_all(struct drm_device *dev)
+{
+	WARN_ON(__drm_modeset_lock_all(dev, false) != 0);
+}
+EXPORT_SYMBOL(drm_modeset_lock_all);
+
+/**
+ * drm_modeset_unlock_all - drop all modeset locks
+ * @dev: device
+ *
+ * This function drop all modeset locks taken by drm_modeset_lock_all.
+ */
+void drm_modeset_unlock_all(struct drm_device *dev)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_modeset_acquire_ctx *ctx = config->acquire_ctx;
+
+	if (WARN_ON(!ctx))
+		return;
+
+	config->acquire_ctx = NULL;
+	drm_modeset_drop_locks(ctx);
+	drm_modeset_acquire_fini(ctx);
+
+	kfree(ctx);
+
+	mutex_unlock(&dev->mode_config.mutex);
+}
+EXPORT_SYMBOL(drm_modeset_unlock_all);
+
+/**
+ * drm_modeset_lock_crtc - lock crtc with hidden acquire ctx
+ * @crtc: drm crtc
+ *
+ * This function locks the given crtc using a hidden acquire context. This is
+ * necessary so that drivers internally using the atomic interfaces can grab
+ * further locks with the lock acquire context.
+ */
+void drm_modeset_lock_crtc(struct drm_crtc *crtc)
+{
+	struct drm_modeset_acquire_ctx *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (WARN_ON(!ctx))
+		return;
+
+	drm_modeset_acquire_init(ctx, 0);
+
+retry:
+	ret = drm_modeset_lock(&crtc->mutex, ctx);
+	if (ret)
+		goto fail;
+
+	WARN_ON(crtc->acquire_ctx);
+
+	/* now we hold the locks, so now that it is safe, stash the
+	 * ctx for drm_modeset_unlock_crtc():
+	 */
+	crtc->acquire_ctx = ctx;
+
+	return;
+
+fail:
+	if (ret == -EDEADLK) {
+		drm_modeset_backoff(ctx);
+		goto retry;
+	}
+}
+EXPORT_SYMBOL(drm_modeset_lock_crtc);
+
+/**
+ * drm_modeset_legacy_acquire_ctx - find acquire ctx for legacy ioctls
+ * @crtc: drm crtc
+ *
+ * Legacy ioctl operations like cursor updates or page flips only have per-crtc
+ * locking, and store the acquire ctx in the corresponding crtc. All other
+ * legacy operations take all locks and use a global acquire context. This
+ * function grabs the right one.
+ */
+struct drm_modeset_acquire_ctx *
+drm_modeset_legacy_acquire_ctx(struct drm_crtc *crtc)
+{
+	if (crtc->acquire_ctx)
+		return crtc->acquire_ctx;
+
+	WARN_ON(!crtc->dev->mode_config.acquire_ctx);
+
+	return crtc->dev->mode_config.acquire_ctx;
+}
+EXPORT_SYMBOL(drm_modeset_legacy_acquire_ctx);
+
+/**
+ * drm_modeset_unlock_crtc - drop crtc lock
+ * @crtc: drm crtc
+ *
+ * This drops the crtc lock acquire with drm_modeset_lock_crtc() and all other
+ * locks acquired through the hidden context.
+ */
+void drm_modeset_unlock_crtc(struct drm_crtc *crtc)
+{
+	struct drm_modeset_acquire_ctx *ctx = crtc->acquire_ctx;
+
+	if (WARN_ON(!ctx))
+		return;
+
+	crtc->acquire_ctx = NULL;
+	drm_modeset_drop_locks(ctx);
+	drm_modeset_acquire_fini(ctx);
+
+	kfree(ctx);
+}
+EXPORT_SYMBOL(drm_modeset_unlock_crtc);
+
+/**
+ * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked
+ * @dev: device
+ *
+ * Useful as a debug assert.
+ */
+void drm_warn_on_modeset_not_all_locked(struct drm_device *dev)
+{
+	struct drm_crtc *crtc;
+
+	/* Locking is currently fubar in the panic handler. */
+	if (oops_in_progress)
+		return;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
+		WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
+
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+}
+EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
+
+/**
  * drm_modeset_acquire_init - initialize acquire context
  * @ctx: the acquire context
  * @flags: for future
@@ -108,7 +314,12 @@ static inline int modeset_lock(struct drm_modeset_lock *lock,
 
 	WARN_ON(ctx->contended);
 
-	if (interruptible && slow) {
+	if (ctx->trylock_only) {
+		if (!ww_mutex_trylock(&lock->mutex))
+			return -EBUSY;
+		else
+			return 0;
+	} else if (interruptible && slow) {
 		ret = ww_mutex_lock_slow_interruptible(&lock->mutex, &ctx->ww_ctx);
 	} else if (interruptible) {
 		ret = ww_mutex_lock_interruptible(&lock->mutex, &ctx->ww_ctx);
diff --git a/drivers/gpu/drm/drm_of.c b/drivers/gpu/drm/drm_of.c
new file mode 100644
index 000000000000..16150a00c237
--- /dev/null
+++ b/drivers/gpu/drm/drm_of.c
@@ -0,0 +1,67 @@
+#include <linux/export.h>
+#include <linux/list.h>
+#include <linux/of_graph.h>
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_of.h>
+
+/**
+ * drm_crtc_port_mask - find the mask of a registered CRTC by port OF node
+ * @dev: DRM device
+ * @port: port OF node
+ *
+ * Given a port OF node, return the possible mask of the corresponding
+ * CRTC within a device's list of CRTCs.  Returns zero if not found.
+ */
+static uint32_t drm_crtc_port_mask(struct drm_device *dev,
+				   struct device_node *port)
+{
+	unsigned int index = 0;
+	struct drm_crtc *tmp;
+
+	list_for_each_entry(tmp, &dev->mode_config.crtc_list, head) {
+		if (tmp->port == port)
+			return 1 << index;
+
+		index++;
+	}
+
+	return 0;
+}
+
+/**
+ * drm_of_find_possible_crtcs - find the possible CRTCs for an encoder port
+ * @dev: DRM device
+ * @port: encoder port to scan for endpoints
+ *
+ * Scan all endpoints attached to a port, locate their attached CRTCs,
+ * and generate the DRM mask of CRTCs which may be attached to this
+ * encoder.
+ *
+ * See Documentation/devicetree/bindings/graph.txt for the bindings.
+ */
+uint32_t drm_of_find_possible_crtcs(struct drm_device *dev,
+				    struct device_node *port)
+{
+	struct device_node *remote_port, *ep = NULL;
+	uint32_t possible_crtcs = 0;
+
+	do {
+		ep = of_graph_get_next_endpoint(port, ep);
+		if (!ep)
+			break;
+
+		remote_port = of_graph_get_remote_port(ep);
+		if (!remote_port) {
+			of_node_put(ep);
+			return 0;
+		}
+
+		possible_crtcs |= drm_crtc_port_mask(dev, remote_port);
+
+		of_node_put(remote_port);
+	} while (1);
+
+	return possible_crtcs;
+}
+EXPORT_SYMBOL(drm_of_find_possible_crtcs);
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 020cfd934854..fd29f03645b8 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -27,6 +27,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/export.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 /**
  * drm_pci_alloc - Allocate a PCI consistent memory block, for DMA.
@@ -81,7 +82,7 @@ EXPORT_SYMBOL(drm_pci_alloc);
  *
  * This function is for internal use in the Linux-specific DRM core code.
  */
-void __drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
+void __drm_legacy_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
 {
 	unsigned long addr;
 	size_t sz;
@@ -105,7 +106,7 @@ void __drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
  */
 void drm_pci_free(struct drm_device * dev, drm_dma_handle_t * dmah)
 {
-	__drm_pci_free(dev, dmah);
+	__drm_legacy_pci_free(dev, dmah);
 	kfree(dmah);
 }
 
@@ -127,34 +128,20 @@ static int drm_get_pci_domain(struct drm_device *dev)
 	return pci_domain_nr(dev->pdev->bus);
 }
 
-static int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
+int drm_pci_set_busid(struct drm_device *dev, struct drm_master *master)
 {
-	int len, ret;
-	master->unique_len = 40;
-	master->unique_size = master->unique_len;
-	master->unique = kmalloc(master->unique_size, GFP_KERNEL);
-	if (master->unique == NULL)
+	master->unique = kasprintf(GFP_KERNEL, "pci:%04x:%02x:%02x.%d",
+					drm_get_pci_domain(dev),
+					dev->pdev->bus->number,
+					PCI_SLOT(dev->pdev->devfn),
+					PCI_FUNC(dev->pdev->devfn));
+	if (!master->unique)
 		return -ENOMEM;
 
-
-	len = snprintf(master->unique, master->unique_len,
-		       "pci:%04x:%02x:%02x.%d",
-		       drm_get_pci_domain(dev),
-		       dev->pdev->bus->number,
-		       PCI_SLOT(dev->pdev->devfn),
-		       PCI_FUNC(dev->pdev->devfn));
-
-	if (len >= master->unique_len) {
-		DRM_ERROR("buffer overflow");
-		ret = -EINVAL;
-		goto err;
-	} else
-		master->unique_len = len;
-
+	master->unique_len = strlen(master->unique);
 	return 0;
-err:
-	return ret;
 }
+EXPORT_SYMBOL(drm_pci_set_busid);
 
 int drm_pci_set_unique(struct drm_device *dev,
 		       struct drm_master *master,
@@ -163,8 +150,7 @@ int drm_pci_set_unique(struct drm_device *dev,
 	int domain, bus, slot, func, ret;
 
 	master->unique_len = u->unique_len;
-	master->unique_size = u->unique_len + 1;
-	master->unique = kmalloc(master->unique_size, GFP_KERNEL);
+	master->unique = kmalloc(master->unique_len + 1, GFP_KERNEL);
 	if (!master->unique) {
 		ret = -ENOMEM;
 		goto err;
@@ -269,10 +255,6 @@ void drm_pci_agp_destroy(struct drm_device *dev)
 	}
 }
 
-static struct drm_bus drm_pci_bus = {
-	.set_busid = drm_pci_set_busid,
-};
-
 /**
  * drm_get_pci_dev - Register a PCI device with the DRM subsystem
  * @pdev: PCI device
@@ -353,8 +335,6 @@ int drm_pci_init(struct drm_driver *driver, struct pci_driver *pdriver)
 
 	DRM_DEBUG("\n");
 
-	driver->bus = &drm_pci_bus;
-
 	if (driver->driver_features & DRIVER_MODESET)
 		return pci_register_driver(pdriver);
 
diff --git a/drivers/gpu/drm/drm_plane_helper.c b/drivers/gpu/drm/drm_plane_helper.c
index 6d133149cc74..827ec1a3040b 100644
--- a/drivers/gpu/drm/drm_plane_helper.c
+++ b/drivers/gpu/drm/drm_plane_helper.c
@@ -335,9 +335,10 @@ struct drm_plane *drm_primary_helper_create_plane(struct drm_device *dev,
 	}
 
 	/* possible_crtc's will be filled in later by crtc_init */
-	ret = drm_plane_init(dev, primary, 0, &drm_primary_helper_funcs,
-			     formats, num_formats,
-			     DRM_PLANE_TYPE_PRIMARY);
+	ret = drm_universal_plane_init(dev, primary, 0,
+				       &drm_primary_helper_funcs,
+				       formats, num_formats,
+				       DRM_PLANE_TYPE_PRIMARY);
 	if (ret) {
 		kfree(primary);
 		primary = NULL;
diff --git a/drivers/gpu/drm/drm_platform.c b/drivers/gpu/drm/drm_platform.c
index d5b76f148c12..5314c9d5fef4 100644
--- a/drivers/gpu/drm/drm_platform.c
+++ b/drivers/gpu/drm/drm_platform.c
@@ -68,42 +68,23 @@ err_free:
 	return ret;
 }
 
-static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master)
+int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master)
 {
-	int len, ret, id;
-
-	master->unique_len = 13 + strlen(dev->platformdev->name);
-	master->unique_size = master->unique_len;
-	master->unique = kmalloc(master->unique_len + 1, GFP_KERNEL);
-
-	if (master->unique == NULL)
-		return -ENOMEM;
+	int id;
 
 	id = dev->platformdev->id;
-
-	/* if only a single instance of the platform device, id will be
-	 * set to -1.. use 0 instead to avoid a funny looking bus-id:
-	 */
-	if (id == -1)
+	if (id < 0)
 		id = 0;
 
-	len = snprintf(master->unique, master->unique_len,
-			"platform:%s:%02d", dev->platformdev->name, id);
-
-	if (len > master->unique_len) {
-		DRM_ERROR("Unique buffer overflowed\n");
-		ret = -EINVAL;
-		goto err;
-	}
+	master->unique = kasprintf(GFP_KERNEL, "platform:%s:%02d",
+						dev->platformdev->name, id);
+	if (!master->unique)
+		return -ENOMEM;
 
+	master->unique_len = strlen(master->unique);
 	return 0;
-err:
-	return ret;
 }
-
-static struct drm_bus drm_platform_bus = {
-	.set_busid = drm_platform_set_busid,
-};
+EXPORT_SYMBOL(drm_platform_set_busid);
 
 /**
  * drm_platform_init - Register a platform device with the DRM subsystem
@@ -120,7 +101,6 @@ int drm_platform_init(struct drm_driver *driver, struct platform_device *platfor
 {
 	DRM_DEBUG("\n");
 
-	driver->bus = &drm_platform_bus;
 	return drm_get_platform_dev(platform_device, driver);
 }
 EXPORT_SYMBOL(drm_platform_init);
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 99d578bad17e..78ca30808422 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -29,6 +29,9 @@
 #include <linux/export.h>
 #include <linux/dma-buf.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
+
+#include "drm_internal.h"
 
 /*
  * DMA-BUF/GEM Object references and lifetime overview:
@@ -522,7 +525,7 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 		goto fail_detach;
 	}
 
-	obj = dev->driver->gem_prime_import_sg_table(dev, dma_buf->size, sgt);
+	obj = dev->driver->gem_prime_import_sg_table(dev, attach, sgt);
 	if (IS_ERR(obj)) {
 		ret = PTR_ERR(obj);
 		goto fail_unmap;
diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
index d22676b89cbb..6857e9ad6339 100644
--- a/drivers/gpu/drm/drm_probe_helper.c
+++ b/drivers/gpu/drm/drm_probe_helper.c
@@ -82,6 +82,22 @@ static void drm_mode_validate_flag(struct drm_connector *connector,
 	return;
 }
 
+static int drm_helper_probe_add_cmdline_mode(struct drm_connector *connector)
+{
+	struct drm_display_mode *mode;
+
+	if (!connector->cmdline_mode.specified)
+		return 0;
+
+	mode = drm_mode_create_from_cmdline_mode(connector->dev,
+						 &connector->cmdline_mode);
+	if (mode == NULL)
+		return 0;
+
+	drm_mode_probed_add(connector, mode);
+	return 1;
+}
+
 static int drm_helper_probe_single_connector_modes_merge_bits(struct drm_connector *connector,
 							      uint32_t maxX, uint32_t maxY, bool merge_type_bits)
 {
@@ -130,10 +146,18 @@ static int drm_helper_probe_single_connector_modes_merge_bits(struct drm_connect
 	count = drm_load_edid_firmware(connector);
 	if (count == 0)
 #endif
-		count = (*connector_funcs->get_modes)(connector);
+	{
+		if (connector->override_edid) {
+			struct edid *edid = (struct edid *) connector->edid_blob_ptr->data;
+
+			count = drm_add_edid_modes(connector, edid);
+		} else
+			count = (*connector_funcs->get_modes)(connector);
+	}
 
 	if (count == 0 && connector->status == connector_status_connected)
 		count = drm_add_modes_noedid(connector, 1024, 768);
+	count += drm_helper_probe_add_cmdline_mode(connector);
 	if (count == 0)
 		goto prune;
 
diff --git a/drivers/gpu/drm/drm_rect.c b/drivers/gpu/drm/drm_rect.c
index 7047ca025787..631f5afd451c 100644
--- a/drivers/gpu/drm/drm_rect.c
+++ b/drivers/gpu/drm/drm_rect.c
@@ -293,3 +293,143 @@ void drm_rect_debug_print(const struct drm_rect *r, bool fixed_point)
 		DRM_DEBUG_KMS("%dx%d%+d%+d\n", w, h, r->x1, r->y1);
 }
 EXPORT_SYMBOL(drm_rect_debug_print);
+
+/**
+ * drm_rect_rotate - Rotate the rectangle
+ * @r: rectangle to be rotated
+ * @width: Width of the coordinate space
+ * @height: Height of the coordinate space
+ * @rotation: Transformation to be applied
+ *
+ * Apply @rotation to the coordinates of rectangle @r.
+ *
+ * @width and @height combined with @rotation define
+ * the location of the new origin.
+ *
+ * @width correcsponds to the horizontal and @height
+ * to the vertical axis of the untransformed coordinate
+ * space.
+ */
+void drm_rect_rotate(struct drm_rect *r,
+		     int width, int height,
+		     unsigned int rotation)
+{
+	struct drm_rect tmp;
+
+	if (rotation & (BIT(DRM_REFLECT_X) | BIT(DRM_REFLECT_Y))) {
+		tmp = *r;
+
+		if (rotation & BIT(DRM_REFLECT_X)) {
+			r->x1 = width - tmp.x2;
+			r->x2 = width - tmp.x1;
+		}
+
+		if (rotation & BIT(DRM_REFLECT_Y)) {
+			r->y1 = height - tmp.y2;
+			r->y2 = height - tmp.y1;
+		}
+	}
+
+	switch (rotation & 0xf) {
+	case BIT(DRM_ROTATE_0):
+		break;
+	case BIT(DRM_ROTATE_90):
+		tmp = *r;
+		r->x1 = tmp.y1;
+		r->x2 = tmp.y2;
+		r->y1 = width - tmp.x2;
+		r->y2 = width - tmp.x1;
+		break;
+	case BIT(DRM_ROTATE_180):
+		tmp = *r;
+		r->x1 = width - tmp.x2;
+		r->x2 = width - tmp.x1;
+		r->y1 = height - tmp.y2;
+		r->y2 = height - tmp.y1;
+		break;
+	case BIT(DRM_ROTATE_270):
+		tmp = *r;
+		r->x1 = height - tmp.y2;
+		r->x2 = height - tmp.y1;
+		r->y1 = tmp.x1;
+		r->y2 = tmp.x2;
+		break;
+	default:
+		break;
+	}
+}
+EXPORT_SYMBOL(drm_rect_rotate);
+
+/**
+ * drm_rect_rotate_inv - Inverse rotate the rectangle
+ * @r: rectangle to be rotated
+ * @width: Width of the coordinate space
+ * @height: Height of the coordinate space
+ * @rotation: Transformation whose inverse is to be applied
+ *
+ * Apply the inverse of @rotation to the coordinates
+ * of rectangle @r.
+ *
+ * @width and @height combined with @rotation define
+ * the location of the new origin.
+ *
+ * @width correcsponds to the horizontal and @height
+ * to the vertical axis of the original untransformed
+ * coordinate space, so that you never have to flip
+ * them when doing a rotatation and its inverse.
+ * That is, if you do:
+ *
+ * drm_rotate(&r, width, height, rotation);
+ * drm_rotate_inv(&r, width, height, rotation);
+ *
+ * you will always get back the original rectangle.
+ */
+void drm_rect_rotate_inv(struct drm_rect *r,
+			 int width, int height,
+			 unsigned int rotation)
+{
+	struct drm_rect tmp;
+
+	switch (rotation & 0xf) {
+	case BIT(DRM_ROTATE_0):
+		break;
+	case BIT(DRM_ROTATE_90):
+		tmp = *r;
+		r->x1 = width - tmp.y2;
+		r->x2 = width - tmp.y1;
+		r->y1 = tmp.x1;
+		r->y2 = tmp.x2;
+		break;
+	case BIT(DRM_ROTATE_180):
+		tmp = *r;
+		r->x1 = width - tmp.x2;
+		r->x2 = width - tmp.x1;
+		r->y1 = height - tmp.y2;
+		r->y2 = height - tmp.y1;
+		break;
+	case BIT(DRM_ROTATE_270):
+		tmp = *r;
+		r->x1 = tmp.y1;
+		r->x2 = tmp.y2;
+		r->y1 = height - tmp.x2;
+		r->y2 = height - tmp.x1;
+		break;
+	default:
+		break;
+	}
+
+	if (rotation & (BIT(DRM_REFLECT_X) | BIT(DRM_REFLECT_Y))) {
+		tmp = *r;
+
+		if (rotation & BIT(DRM_REFLECT_X)) {
+			r->x1 = width - tmp.x2;
+			r->x2 = width - tmp.x1;
+		}
+
+		if (rotation & BIT(DRM_REFLECT_Y)) {
+			r->y1 = height - tmp.y2;
+			r->y2 = height - tmp.y1;
+		}
+	}
+}
+EXPORT_SYMBOL(drm_rect_rotate_inv);
diff --git a/drivers/gpu/drm/drm_scatter.c b/drivers/gpu/drm/drm_scatter.c
index 1c78406f6e71..4f0f3b36d537 100644
--- a/drivers/gpu/drm/drm_scatter.c
+++ b/drivers/gpu/drm/drm_scatter.c
@@ -34,6 +34,7 @@
 #include <linux/vmalloc.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
+#include "drm_legacy.h"
 
 #define DEBUG_SCATTER 0
 
@@ -78,8 +79,8 @@ void drm_legacy_sg_cleanup(struct drm_device *dev)
 # define ScatterHandle(x) (unsigned int)(x)
 #endif
 
-int drm_sg_alloc(struct drm_device *dev, void *data,
-		 struct drm_file *file_priv)
+int drm_legacy_sg_alloc(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
 {
 	struct drm_scatter_gather *request = data;
 	struct drm_sg_mem *entry;
@@ -194,8 +195,8 @@ int drm_sg_alloc(struct drm_device *dev, void *data,
 	return -ENOMEM;
 }
 
-int drm_sg_free(struct drm_device *dev, void *data,
-		struct drm_file *file_priv)
+int drm_legacy_sg_free(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
 {
 	struct drm_scatter_gather *request = data;
 	struct drm_sg_mem *entry;
diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
deleted file mode 100644
index 14d16464000a..000000000000
--- a/drivers/gpu/drm/drm_stub.c
+++ /dev/null
@@ -1,805 +0,0 @@
-/*
- * Created: Fri Jan 19 10:48:35 2001 by faith@acm.org
- *
- * Copyright 2001 VA Linux Systems, Inc., Sunnyvale, California.
- * All Rights Reserved.
- *
- * Author Rickard E. (Rik) Faith <faith@valinux.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- */
-
-#include <linux/fs.h>
-#include <linux/module.h>
-#include <linux/moduleparam.h>
-#include <linux/mount.h>
-#include <linux/slab.h>
-#include <drm/drmP.h>
-#include <drm/drm_core.h>
-
-unsigned int drm_debug = 0;	/* 1 to enable debug output */
-EXPORT_SYMBOL(drm_debug);
-
-unsigned int drm_rnodes = 0;	/* 1 to enable experimental render nodes API */
-EXPORT_SYMBOL(drm_rnodes);
-
-/* 1 to allow user space to request universal planes (experimental) */
-unsigned int drm_universal_planes = 0;
-EXPORT_SYMBOL(drm_universal_planes);
-
-unsigned int drm_vblank_offdelay = 5000;    /* Default to 5000 msecs. */
-EXPORT_SYMBOL(drm_vblank_offdelay);
-
-unsigned int drm_timestamp_precision = 20;  /* Default to 20 usecs. */
-EXPORT_SYMBOL(drm_timestamp_precision);
-
-/*
- * Default to use monotonic timestamps for wait-for-vblank and page-flip
- * complete events.
- */
-unsigned int drm_timestamp_monotonic = 1;
-
-MODULE_AUTHOR(CORE_AUTHOR);
-MODULE_DESCRIPTION(CORE_DESC);
-MODULE_LICENSE("GPL and additional rights");
-MODULE_PARM_DESC(debug, "Enable debug output");
-MODULE_PARM_DESC(rnodes, "Enable experimental render nodes API");
-MODULE_PARM_DESC(vblankoffdelay, "Delay until vblank irq auto-disable [msecs]");
-MODULE_PARM_DESC(timestamp_precision_usec, "Max. error on timestamps [usecs]");
-MODULE_PARM_DESC(timestamp_monotonic, "Use monotonic timestamps");
-
-module_param_named(debug, drm_debug, int, 0600);
-module_param_named(rnodes, drm_rnodes, int, 0600);
-module_param_named(universal_planes, drm_universal_planes, int, 0600);
-module_param_named(vblankoffdelay, drm_vblank_offdelay, int, 0600);
-module_param_named(timestamp_precision_usec, drm_timestamp_precision, int, 0600);
-module_param_named(timestamp_monotonic, drm_timestamp_monotonic, int, 0600);
-
-static DEFINE_SPINLOCK(drm_minor_lock);
-struct idr drm_minors_idr;
-
-struct class *drm_class;
-struct dentry *drm_debugfs_root;
-
-int drm_err(const char *func, const char *format, ...)
-{
-	struct va_format vaf;
-	va_list args;
-	int r;
-
-	va_start(args, format);
-
-	vaf.fmt = format;
-	vaf.va = &args;
-
-	r = printk(KERN_ERR "[" DRM_NAME ":%s] *ERROR* %pV", func, &vaf);
-
-	va_end(args);
-
-	return r;
-}
-EXPORT_SYMBOL(drm_err);
-
-void drm_ut_debug_printk(const char *function_name, const char *format, ...)
-{
-	struct va_format vaf;
-	va_list args;
-
-	va_start(args, format);
-	vaf.fmt = format;
-	vaf.va = &args;
-
-	printk(KERN_DEBUG "[" DRM_NAME ":%s] %pV", function_name, &vaf);
-
-	va_end(args);
-}
-EXPORT_SYMBOL(drm_ut_debug_printk);
-
-struct drm_master *drm_master_create(struct drm_minor *minor)
-{
-	struct drm_master *master;
-
-	master = kzalloc(sizeof(*master), GFP_KERNEL);
-	if (!master)
-		return NULL;
-
-	kref_init(&master->refcount);
-	spin_lock_init(&master->lock.spinlock);
-	init_waitqueue_head(&master->lock.lock_queue);
-	if (drm_ht_create(&master->magiclist, DRM_MAGIC_HASH_ORDER)) {
-		kfree(master);
-		return NULL;
-	}
-	INIT_LIST_HEAD(&master->magicfree);
-	master->minor = minor;
-
-	return master;
-}
-
-struct drm_master *drm_master_get(struct drm_master *master)
-{
-	kref_get(&master->refcount);
-	return master;
-}
-EXPORT_SYMBOL(drm_master_get);
-
-static void drm_master_destroy(struct kref *kref)
-{
-	struct drm_master *master = container_of(kref, struct drm_master, refcount);
-	struct drm_magic_entry *pt, *next;
-	struct drm_device *dev = master->minor->dev;
-	struct drm_map_list *r_list, *list_temp;
-
-	mutex_lock(&dev->struct_mutex);
-	if (dev->driver->master_destroy)
-		dev->driver->master_destroy(dev, master);
-
-	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head) {
-		if (r_list->master == master) {
-			drm_rmmap_locked(dev, r_list->map);
-			r_list = NULL;
-		}
-	}
-
-	if (master->unique) {
-		kfree(master->unique);
-		master->unique = NULL;
-		master->unique_len = 0;
-	}
-
-	list_for_each_entry_safe(pt, next, &master->magicfree, head) {
-		list_del(&pt->head);
-		drm_ht_remove_item(&master->magiclist, &pt->hash_item);
-		kfree(pt);
-	}
-
-	drm_ht_remove(&master->magiclist);
-
-	mutex_unlock(&dev->struct_mutex);
-	kfree(master);
-}
-
-void drm_master_put(struct drm_master **master)
-{
-	kref_put(&(*master)->refcount, drm_master_destroy);
-	*master = NULL;
-}
-EXPORT_SYMBOL(drm_master_put);
-
-int drm_setmaster_ioctl(struct drm_device *dev, void *data,
-			struct drm_file *file_priv)
-{
-	int ret = 0;
-
-	mutex_lock(&dev->master_mutex);
-	if (file_priv->is_master)
-		goto out_unlock;
-
-	if (file_priv->minor->master) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	if (!file_priv->master) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	file_priv->minor->master = drm_master_get(file_priv->master);
-	file_priv->is_master = 1;
-	if (dev->driver->master_set) {
-		ret = dev->driver->master_set(dev, file_priv, false);
-		if (unlikely(ret != 0)) {
-			file_priv->is_master = 0;
-			drm_master_put(&file_priv->minor->master);
-		}
-	}
-
-out_unlock:
-	mutex_unlock(&dev->master_mutex);
-	return ret;
-}
-
-int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
-			 struct drm_file *file_priv)
-{
-	int ret = -EINVAL;
-
-	mutex_lock(&dev->master_mutex);
-	if (!file_priv->is_master)
-		goto out_unlock;
-
-	if (!file_priv->minor->master)
-		goto out_unlock;
-
-	ret = 0;
-	if (dev->driver->master_drop)
-		dev->driver->master_drop(dev, file_priv, false);
-	drm_master_put(&file_priv->minor->master);
-	file_priv->is_master = 0;
-
-out_unlock:
-	mutex_unlock(&dev->master_mutex);
-	return ret;
-}
-
-/*
- * DRM Minors
- * A DRM device can provide several char-dev interfaces on the DRM-Major. Each
- * of them is represented by a drm_minor object. Depending on the capabilities
- * of the device-driver, different interfaces are registered.
- *
- * Minors can be accessed via dev->$minor_name. This pointer is either
- * NULL or a valid drm_minor pointer and stays valid as long as the device is
- * valid. This means, DRM minors have the same life-time as the underlying
- * device. However, this doesn't mean that the minor is active. Minors are
- * registered and unregistered dynamically according to device-state.
- */
-
-static struct drm_minor **drm_minor_get_slot(struct drm_device *dev,
-					     unsigned int type)
-{
-	switch (type) {
-	case DRM_MINOR_LEGACY:
-		return &dev->primary;
-	case DRM_MINOR_RENDER:
-		return &dev->render;
-	case DRM_MINOR_CONTROL:
-		return &dev->control;
-	default:
-		return NULL;
-	}
-}
-
-static int drm_minor_alloc(struct drm_device *dev, unsigned int type)
-{
-	struct drm_minor *minor;
-
-	minor = kzalloc(sizeof(*minor), GFP_KERNEL);
-	if (!minor)
-		return -ENOMEM;
-
-	minor->type = type;
-	minor->dev = dev;
-
-	*drm_minor_get_slot(dev, type) = minor;
-	return 0;
-}
-
-static void drm_minor_free(struct drm_device *dev, unsigned int type)
-{
-	struct drm_minor **slot;
-
-	slot = drm_minor_get_slot(dev, type);
-	if (*slot) {
-		drm_mode_group_destroy(&(*slot)->mode_group);
-		kfree(*slot);
-		*slot = NULL;
-	}
-}
-
-static int drm_minor_register(struct drm_device *dev, unsigned int type)
-{
-	struct drm_minor *new_minor;
-	unsigned long flags;
-	int ret;
-	int minor_id;
-
-	DRM_DEBUG("\n");
-
-	new_minor = *drm_minor_get_slot(dev, type);
-	if (!new_minor)
-		return 0;
-
-	idr_preload(GFP_KERNEL);
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	minor_id = idr_alloc(&drm_minors_idr,
-			     NULL,
-			     64 * type,
-			     64 * (type + 1),
-			     GFP_NOWAIT);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-	idr_preload_end();
-
-	if (minor_id < 0)
-		return minor_id;
-
-	new_minor->index = minor_id;
-
-	ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root);
-	if (ret) {
-		DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n");
-		goto err_id;
-	}
-
-	ret = drm_sysfs_device_add(new_minor);
-	if (ret) {
-		DRM_ERROR("DRM: Error sysfs_device_add.\n");
-		goto err_debugfs;
-	}
-
-	/* replace NULL with @minor so lookups will succeed from now on */
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	idr_replace(&drm_minors_idr, new_minor, new_minor->index);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-
-	DRM_DEBUG("new minor assigned %d\n", minor_id);
-	return 0;
-
-err_debugfs:
-	drm_debugfs_cleanup(new_minor);
-err_id:
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	idr_remove(&drm_minors_idr, minor_id);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-	new_minor->index = 0;
-	return ret;
-}
-
-static void drm_minor_unregister(struct drm_device *dev, unsigned int type)
-{
-	struct drm_minor *minor;
-	unsigned long flags;
-
-	minor = *drm_minor_get_slot(dev, type);
-	if (!minor || !minor->kdev)
-		return;
-
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	idr_remove(&drm_minors_idr, minor->index);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-	minor->index = 0;
-
-	drm_debugfs_cleanup(minor);
-	drm_sysfs_device_remove(minor);
-}
-
-/**
- * drm_minor_acquire - Acquire a DRM minor
- * @minor_id: Minor ID of the DRM-minor
- *
- * Looks up the given minor-ID and returns the respective DRM-minor object. The
- * refence-count of the underlying device is increased so you must release this
- * object with drm_minor_release().
- *
- * As long as you hold this minor, it is guaranteed that the object and the
- * minor->dev pointer will stay valid! However, the device may get unplugged and
- * unregistered while you hold the minor.
- *
- * Returns:
- * Pointer to minor-object with increased device-refcount, or PTR_ERR on
- * failure.
- */
-struct drm_minor *drm_minor_acquire(unsigned int minor_id)
-{
-	struct drm_minor *minor;
-	unsigned long flags;
-
-	spin_lock_irqsave(&drm_minor_lock, flags);
-	minor = idr_find(&drm_minors_idr, minor_id);
-	if (minor)
-		drm_dev_ref(minor->dev);
-	spin_unlock_irqrestore(&drm_minor_lock, flags);
-
-	if (!minor) {
-		return ERR_PTR(-ENODEV);
-	} else if (drm_device_is_unplugged(minor->dev)) {
-		drm_dev_unref(minor->dev);
-		return ERR_PTR(-ENODEV);
-	}
-
-	return minor;
-}
-
-/**
- * drm_minor_release - Release DRM minor
- * @minor: Pointer to DRM minor object
- *
- * Release a minor that was previously acquired via drm_minor_acquire().
- */
-void drm_minor_release(struct drm_minor *minor)
-{
-	drm_dev_unref(minor->dev);
-}
-
-/**
- * drm_put_dev - Unregister and release a DRM device
- * @dev: DRM device
- *
- * Called at module unload time or when a PCI device is unplugged.
- *
- * Use of this function is discouraged. It will eventually go away completely.
- * Please use drm_dev_unregister() and drm_dev_unref() explicitly instead.
- *
- * Cleans up all DRM device, calling drm_lastclose().
- */
-void drm_put_dev(struct drm_device *dev)
-{
-	DRM_DEBUG("\n");
-
-	if (!dev) {
-		DRM_ERROR("cleanup called no dev\n");
-		return;
-	}
-
-	drm_dev_unregister(dev);
-	drm_dev_unref(dev);
-}
-EXPORT_SYMBOL(drm_put_dev);
-
-void drm_unplug_dev(struct drm_device *dev)
-{
-	/* for a USB device */
-	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
-	drm_minor_unregister(dev, DRM_MINOR_RENDER);
-	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
-
-	mutex_lock(&drm_global_mutex);
-
-	drm_device_set_unplugged(dev);
-
-	if (dev->open_count == 0) {
-		drm_put_dev(dev);
-	}
-	mutex_unlock(&drm_global_mutex);
-}
-EXPORT_SYMBOL(drm_unplug_dev);
-
-/*
- * DRM internal mount
- * We want to be able to allocate our own "struct address_space" to control
- * memory-mappings in VRAM (or stolen RAM, ...). However, core MM does not allow
- * stand-alone address_space objects, so we need an underlying inode. As there
- * is no way to allocate an independent inode easily, we need a fake internal
- * VFS mount-point.
- *
- * The drm_fs_inode_new() function allocates a new inode, drm_fs_inode_free()
- * frees it again. You are allowed to use iget() and iput() to get references to
- * the inode. But each drm_fs_inode_new() call must be paired with exactly one
- * drm_fs_inode_free() call (which does not have to be the last iput()).
- * We use drm_fs_inode_*() to manage our internal VFS mount-point and share it
- * between multiple inode-users. You could, technically, call
- * iget() + drm_fs_inode_free() directly after alloc and sometime later do an
- * iput(), but this way you'd end up with a new vfsmount for each inode.
- */
-
-static int drm_fs_cnt;
-static struct vfsmount *drm_fs_mnt;
-
-static const struct dentry_operations drm_fs_dops = {
-	.d_dname	= simple_dname,
-};
-
-static const struct super_operations drm_fs_sops = {
-	.statfs		= simple_statfs,
-};
-
-static struct dentry *drm_fs_mount(struct file_system_type *fs_type, int flags,
-				   const char *dev_name, void *data)
-{
-	return mount_pseudo(fs_type,
-			    "drm:",
-			    &drm_fs_sops,
-			    &drm_fs_dops,
-			    0x010203ff);
-}
-
-static struct file_system_type drm_fs_type = {
-	.name		= "drm",
-	.owner		= THIS_MODULE,
-	.mount		= drm_fs_mount,
-	.kill_sb	= kill_anon_super,
-};
-
-static struct inode *drm_fs_inode_new(void)
-{
-	struct inode *inode;
-	int r;
-
-	r = simple_pin_fs(&drm_fs_type, &drm_fs_mnt, &drm_fs_cnt);
-	if (r < 0) {
-		DRM_ERROR("Cannot mount pseudo fs: %d\n", r);
-		return ERR_PTR(r);
-	}
-
-	inode = alloc_anon_inode(drm_fs_mnt->mnt_sb);
-	if (IS_ERR(inode))
-		simple_release_fs(&drm_fs_mnt, &drm_fs_cnt);
-
-	return inode;
-}
-
-static void drm_fs_inode_free(struct inode *inode)
-{
-	if (inode) {
-		iput(inode);
-		simple_release_fs(&drm_fs_mnt, &drm_fs_cnt);
-	}
-}
-
-/**
- * drm_dev_alloc - Allocate new DRM device
- * @driver: DRM driver to allocate device for
- * @parent: Parent device object
- *
- * Allocate and initialize a new DRM device. No device registration is done.
- * Call drm_dev_register() to advertice the device to user space and register it
- * with other core subsystems.
- *
- * The initial ref-count of the object is 1. Use drm_dev_ref() and
- * drm_dev_unref() to take and drop further ref-counts.
- *
- * RETURNS:
- * Pointer to new DRM device, or NULL if out of memory.
- */
-struct drm_device *drm_dev_alloc(struct drm_driver *driver,
-				 struct device *parent)
-{
-	struct drm_device *dev;
-	int ret;
-
-	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
-	if (!dev)
-		return NULL;
-
-	kref_init(&dev->ref);
-	dev->dev = parent;
-	dev->driver = driver;
-
-	INIT_LIST_HEAD(&dev->filelist);
-	INIT_LIST_HEAD(&dev->ctxlist);
-	INIT_LIST_HEAD(&dev->vmalist);
-	INIT_LIST_HEAD(&dev->maplist);
-	INIT_LIST_HEAD(&dev->vblank_event_list);
-
-	spin_lock_init(&dev->buf_lock);
-	spin_lock_init(&dev->event_lock);
-	mutex_init(&dev->struct_mutex);
-	mutex_init(&dev->ctxlist_mutex);
-	mutex_init(&dev->master_mutex);
-
-	dev->anon_inode = drm_fs_inode_new();
-	if (IS_ERR(dev->anon_inode)) {
-		ret = PTR_ERR(dev->anon_inode);
-		DRM_ERROR("Cannot allocate anonymous inode: %d\n", ret);
-		goto err_free;
-	}
-
-	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
-		ret = drm_minor_alloc(dev, DRM_MINOR_CONTROL);
-		if (ret)
-			goto err_minors;
-	}
-
-	if (drm_core_check_feature(dev, DRIVER_RENDER) && drm_rnodes) {
-		ret = drm_minor_alloc(dev, DRM_MINOR_RENDER);
-		if (ret)
-			goto err_minors;
-	}
-
-	ret = drm_minor_alloc(dev, DRM_MINOR_LEGACY);
-	if (ret)
-		goto err_minors;
-
-	if (drm_ht_create(&dev->map_hash, 12))
-		goto err_minors;
-
-	ret = drm_ctxbitmap_init(dev);
-	if (ret) {
-		DRM_ERROR("Cannot allocate memory for context bitmap.\n");
-		goto err_ht;
-	}
-
-	if (driver->driver_features & DRIVER_GEM) {
-		ret = drm_gem_init(dev);
-		if (ret) {
-			DRM_ERROR("Cannot initialize graphics execution manager (GEM)\n");
-			goto err_ctxbitmap;
-		}
-	}
-
-	return dev;
-
-err_ctxbitmap:
-	drm_ctxbitmap_cleanup(dev);
-err_ht:
-	drm_ht_remove(&dev->map_hash);
-err_minors:
-	drm_minor_free(dev, DRM_MINOR_LEGACY);
-	drm_minor_free(dev, DRM_MINOR_RENDER);
-	drm_minor_free(dev, DRM_MINOR_CONTROL);
-	drm_fs_inode_free(dev->anon_inode);
-err_free:
-	mutex_destroy(&dev->master_mutex);
-	kfree(dev);
-	return NULL;
-}
-EXPORT_SYMBOL(drm_dev_alloc);
-
-static void drm_dev_release(struct kref *ref)
-{
-	struct drm_device *dev = container_of(ref, struct drm_device, ref);
-
-	if (dev->driver->driver_features & DRIVER_GEM)
-		drm_gem_destroy(dev);
-
-	drm_ctxbitmap_cleanup(dev);
-	drm_ht_remove(&dev->map_hash);
-	drm_fs_inode_free(dev->anon_inode);
-
-	drm_minor_free(dev, DRM_MINOR_LEGACY);
-	drm_minor_free(dev, DRM_MINOR_RENDER);
-	drm_minor_free(dev, DRM_MINOR_CONTROL);
-
-	mutex_destroy(&dev->master_mutex);
-	kfree(dev->unique);
-	kfree(dev);
-}
-
-/**
- * drm_dev_ref - Take reference of a DRM device
- * @dev: device to take reference of or NULL
- *
- * This increases the ref-count of @dev by one. You *must* already own a
- * reference when calling this. Use drm_dev_unref() to drop this reference
- * again.
- *
- * This function never fails. However, this function does not provide *any*
- * guarantee whether the device is alive or running. It only provides a
- * reference to the object and the memory associated with it.
- */
-void drm_dev_ref(struct drm_device *dev)
-{
-	if (dev)
-		kref_get(&dev->ref);
-}
-EXPORT_SYMBOL(drm_dev_ref);
-
-/**
- * drm_dev_unref - Drop reference of a DRM device
- * @dev: device to drop reference of or NULL
- *
- * This decreases the ref-count of @dev by one. The device is destroyed if the
- * ref-count drops to zero.
- */
-void drm_dev_unref(struct drm_device *dev)
-{
-	if (dev)
-		kref_put(&dev->ref, drm_dev_release);
-}
-EXPORT_SYMBOL(drm_dev_unref);
-
-/**
- * drm_dev_register - Register DRM device
- * @dev: Device to register
- * @flags: Flags passed to the driver's .load() function
- *
- * Register the DRM device @dev with the system, advertise device to user-space
- * and start normal device operation. @dev must be allocated via drm_dev_alloc()
- * previously.
- *
- * Never call this twice on any device!
- *
- * RETURNS:
- * 0 on success, negative error code on failure.
- */
-int drm_dev_register(struct drm_device *dev, unsigned long flags)
-{
-	int ret;
-
-	mutex_lock(&drm_global_mutex);
-
-	ret = drm_minor_register(dev, DRM_MINOR_CONTROL);
-	if (ret)
-		goto err_minors;
-
-	ret = drm_minor_register(dev, DRM_MINOR_RENDER);
-	if (ret)
-		goto err_minors;
-
-	ret = drm_minor_register(dev, DRM_MINOR_LEGACY);
-	if (ret)
-		goto err_minors;
-
-	if (dev->driver->load) {
-		ret = dev->driver->load(dev, flags);
-		if (ret)
-			goto err_minors;
-	}
-
-	/* setup grouping for legacy outputs */
-	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
-		ret = drm_mode_group_init_legacy_group(dev,
-				&dev->primary->mode_group);
-		if (ret)
-			goto err_unload;
-	}
-
-	ret = 0;
-	goto out_unlock;
-
-err_unload:
-	if (dev->driver->unload)
-		dev->driver->unload(dev);
-err_minors:
-	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
-	drm_minor_unregister(dev, DRM_MINOR_RENDER);
-	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
-out_unlock:
-	mutex_unlock(&drm_global_mutex);
-	return ret;
-}
-EXPORT_SYMBOL(drm_dev_register);
-
-/**
- * drm_dev_unregister - Unregister DRM device
- * @dev: Device to unregister
- *
- * Unregister the DRM device from the system. This does the reverse of
- * drm_dev_register() but does not deallocate the device. The caller must call
- * drm_dev_unref() to drop their final reference.
- */
-void drm_dev_unregister(struct drm_device *dev)
-{
-	struct drm_map_list *r_list, *list_temp;
-
-	drm_lastclose(dev);
-
-	if (dev->driver->unload)
-		dev->driver->unload(dev);
-
-	if (dev->agp)
-		drm_pci_agp_destroy(dev);
-
-	drm_vblank_cleanup(dev);
-
-	list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head)
-		drm_rmmap(dev, r_list->map);
-
-	drm_minor_unregister(dev, DRM_MINOR_LEGACY);
-	drm_minor_unregister(dev, DRM_MINOR_RENDER);
-	drm_minor_unregister(dev, DRM_MINOR_CONTROL);
-}
-EXPORT_SYMBOL(drm_dev_unregister);
-
-/**
- * drm_dev_set_unique - Set the unique name of a DRM device
- * @dev: device of which to set the unique name
- * @fmt: format string for unique name
- *
- * Sets the unique name of a DRM device using the specified format string and
- * a variable list of arguments. Drivers can use this at driver probe time if
- * the unique name of the devices they drive is static.
- *
- * Return: 0 on success or a negative error code on failure.
- */
-int drm_dev_set_unique(struct drm_device *dev, const char *fmt, ...)
-{
-	va_list ap;
-
-	kfree(dev->unique);
-
-	va_start(ap, fmt);
-	dev->unique = kvasprintf(GFP_KERNEL, fmt, ap);
-	va_end(ap);
-
-	return dev->unique ? 0 : -ENOMEM;
-}
-EXPORT_SYMBOL(drm_dev_set_unique);
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 369b26278e76..cc3d6d6d67e0 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -21,6 +21,7 @@
 #include <drm/drm_sysfs.h>
 #include <drm/drm_core.h>
 #include <drm/drmP.h>
+#include "drm_internal.h"
 
 #define to_drm_minor(d) dev_get_drvdata(d)
 #define to_drm_connector(d) dev_get_drvdata(d)
@@ -438,7 +439,6 @@ err_out_files:
 out:
 	return ret;
 }
-EXPORT_SYMBOL(drm_sysfs_connector_add);
 
 /**
  * drm_sysfs_connector_remove - remove an connector device from sysfs
@@ -468,7 +468,6 @@ void drm_sysfs_connector_remove(struct drm_connector *connector)
 	device_unregister(connector->kdev);
 	connector->kdev = NULL;
 }
-EXPORT_SYMBOL(drm_sysfs_connector_remove);
 
 /**
  * drm_sysfs_hotplug_event - generate a DRM uevent
@@ -495,71 +494,55 @@ static void drm_sysfs_release(struct device *dev)
 }
 
 /**
- * drm_sysfs_device_add - adds a class device to sysfs for a character driver
- * @dev: DRM device to be added
- * @head: DRM head in question
+ * drm_sysfs_minor_alloc() - Allocate sysfs device for given minor
+ * @minor: minor to allocate sysfs device for
  *
- * Add a DRM device to the DRM's device model class.  We use @dev's PCI device
- * as the parent for the Linux device, and make sure it has a file containing
- * the driver we're using (for userspace compatibility).
+ * This allocates a new sysfs device for @minor and returns it. The device is
+ * not registered nor linked. The caller has to use device_add() and
+ * device_del() to register and unregister it.
+ *
+ * Note that dev_get_drvdata() on the new device will return the minor.
+ * However, the device does not hold a ref-count to the minor nor to the
+ * underlying drm_device. This is unproblematic as long as you access the
+ * private data only in sysfs callbacks. device_del() disables those
+ * synchronously, so they cannot be called after you cleanup a minor.
  */
-int drm_sysfs_device_add(struct drm_minor *minor)
+struct device *drm_sysfs_minor_alloc(struct drm_minor *minor)
 {
-	char *minor_str;
+	const char *minor_str;
+	struct device *kdev;
 	int r;
 
 	if (minor->type == DRM_MINOR_CONTROL)
 		minor_str = "controlD%d";
-        else if (minor->type == DRM_MINOR_RENDER)
-                minor_str = "renderD%d";
-        else
-                minor_str = "card%d";
-
-	minor->kdev = kzalloc(sizeof(*minor->kdev), GFP_KERNEL);
-	if (!minor->kdev) {
-		r = -ENOMEM;
-		goto error;
-	}
-
-	device_initialize(minor->kdev);
-	minor->kdev->devt = MKDEV(DRM_MAJOR, minor->index);
-	minor->kdev->class = drm_class;
-	minor->kdev->type = &drm_sysfs_device_minor;
-	minor->kdev->parent = minor->dev->dev;
-	minor->kdev->release = drm_sysfs_release;
-	dev_set_drvdata(minor->kdev, minor);
-
-	r = dev_set_name(minor->kdev, minor_str, minor->index);
+	else if (minor->type == DRM_MINOR_RENDER)
+		minor_str = "renderD%d";
+	else
+		minor_str = "card%d";
+
+	kdev = kzalloc(sizeof(*kdev), GFP_KERNEL);
+	if (!kdev)
+		return ERR_PTR(-ENOMEM);
+
+	device_initialize(kdev);
+	kdev->devt = MKDEV(DRM_MAJOR, minor->index);
+	kdev->class = drm_class;
+	kdev->type = &drm_sysfs_device_minor;
+	kdev->parent = minor->dev->dev;
+	kdev->release = drm_sysfs_release;
+	dev_set_drvdata(kdev, minor);
+
+	r = dev_set_name(kdev, minor_str, minor->index);
 	if (r < 0)
-		goto error;
-
-	r = device_add(minor->kdev);
-	if (r < 0)
-		goto error;
-
-	return 0;
+		goto err_free;
 
-error:
-	DRM_ERROR("device create failed %d\n", r);
-	put_device(minor->kdev);
-	return r;
-}
+	return kdev;
 
-/**
- * drm_sysfs_device_remove - remove DRM device
- * @dev: DRM device to remove
- *
- * This call unregisters and cleans up a class device that was created with a
- * call to drm_sysfs_device_add()
- */
-void drm_sysfs_device_remove(struct drm_minor *minor)
-{
-	if (minor->kdev)
-		device_unregister(minor->kdev);
-	minor->kdev = NULL;
+err_free:
+	put_device(kdev);
+	return ERR_PTR(r);
 }
 
-
 /**
  * drm_class_device_register - Register a struct device in the drm class.
  *
diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c
deleted file mode 100644
index f2fe94aab901..000000000000
--- a/drivers/gpu/drm/drm_usb.c
+++ /dev/null
@@ -1,88 +0,0 @@
-#include <drm/drmP.h>
-#include <drm/drm_usb.h>
-#include <linux/usb.h>
-#include <linux/module.h>
-
-int drm_get_usb_dev(struct usb_interface *interface,
-		    const struct usb_device_id *id,
-		    struct drm_driver *driver)
-{
-	struct drm_device *dev;
-	int ret;
-
-	DRM_DEBUG("\n");
-
-	dev = drm_dev_alloc(driver, &interface->dev);
-	if (!dev)
-		return -ENOMEM;
-
-	dev->usbdev = interface_to_usbdev(interface);
-	usb_set_intfdata(interface, dev);
-
-	ret = drm_dev_register(dev, 0);
-	if (ret)
-		goto err_free;
-
-	DRM_INFO("Initialized %s %d.%d.%d %s on minor %d\n",
-		 driver->name, driver->major, driver->minor, driver->patchlevel,
-		 driver->date, dev->primary->index);
-
-	return 0;
-
-err_free:
-	drm_dev_unref(dev);
-	return ret;
-
-}
-EXPORT_SYMBOL(drm_get_usb_dev);
-
-static int drm_usb_set_busid(struct drm_device *dev,
-			       struct drm_master *master)
-{
-	return 0;
-}
-
-static struct drm_bus drm_usb_bus = {
-	.set_busid = drm_usb_set_busid,
-};
-
-/**
- * drm_usb_init - Register matching USB devices with the DRM subsystem
- * @driver: DRM device driver
- * @udriver: USB device driver
- *
- * Registers one or more devices matched by a USB driver with the DRM
- * subsystem.
- *
- * Return: 0 on success or a negative error code on failure.
- */
-int drm_usb_init(struct drm_driver *driver, struct usb_driver *udriver)
-{
-	int res;
-	DRM_DEBUG("\n");
-
-	driver->bus = &drm_usb_bus;
-
-	res = usb_register(udriver);
-	return res;
-}
-EXPORT_SYMBOL(drm_usb_init);
-
-/**
- * drm_usb_exit - Unregister matching USB devices from the DRM subsystem
- * @driver: DRM device driver
- * @udriver: USB device driver
- *
- * Unregisters one or more devices matched by a USB driver from the DRM
- * subsystem.
- */
-void drm_usb_exit(struct drm_driver *driver,
-		  struct usb_driver *udriver)
-{
-	usb_deregister(udriver);
-}
-EXPORT_SYMBOL(drm_usb_exit);
-
-MODULE_AUTHOR("David Airlie");
-MODULE_DESCRIPTION("USB DRM support");
-MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index 24e045c4f531..4a2c328959e5 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -35,10 +35,19 @@
 
 #include <drm/drmP.h>
 #include <linux/export.h>
+#include <linux/seq_file.h>
 #if defined(__ia64__)
 #include <linux/efi.h>
 #include <linux/slab.h>
 #endif
+#include <asm/pgtable.h>
+#include "drm_legacy.h"
+
+struct drm_vma_entry {
+	struct list_head head;
+	struct vm_area_struct *vma;
+	pid_t pid;
+};
 
 static void drm_vm_open(struct vm_area_struct *vma);
 static void drm_vm_close(struct vm_area_struct *vma);
@@ -48,15 +57,11 @@ static pgprot_t drm_io_prot(struct drm_local_map *map,
 {
 	pgprot_t tmp = vm_get_page_prot(vma->vm_flags);
 
-#if defined(__i386__) || defined(__x86_64__)
+#if defined(__i386__) || defined(__x86_64__) || defined(__powerpc__)
 	if (map->type == _DRM_REGISTERS && !(map->flags & _DRM_WRITE_COMBINING))
 		tmp = pgprot_noncached(tmp);
 	else
 		tmp = pgprot_writecombine(tmp);
-#elif defined(__powerpc__)
-	pgprot_val(tmp) |= _PAGE_NO_CACHE;
-	if (map->type == _DRM_REGISTERS)
-		pgprot_val(tmp) |= _PAGE_GUARDED;
 #elif defined(__ia64__)
 	if (efi_range_is_wc(vma->vm_start, vma->vm_end -
 				    vma->vm_start))
@@ -263,7 +268,7 @@ static void drm_vm_shm_close(struct vm_area_struct *vma)
 				dmah.vaddr = map->handle;
 				dmah.busaddr = map->offset;
 				dmah.size = map->size;
-				__drm_pci_free(dev, &dmah);
+				__drm_legacy_pci_free(dev, &dmah);
 				break;
 			}
 			kfree(map);
@@ -412,7 +417,6 @@ void drm_vm_open_locked(struct drm_device *dev,
 		list_add(&vma_entry->head, &dev->vmalist);
 	}
 }
-EXPORT_SYMBOL_GPL(drm_vm_open_locked);
 
 static void drm_vm_open(struct vm_area_struct *vma)
 {
@@ -532,7 +536,7 @@ static resource_size_t drm_core_get_reg_ofs(struct drm_device *dev)
  * according to the mapping type and remaps the pages. Finally sets the file
  * pointer and calls vm_open().
  */
-int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
+static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *priv = filp->private_data;
 	struct drm_device *dev = priv->minor->dev;
@@ -646,7 +650,7 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 	return 0;
 }
 
-int drm_mmap(struct file *filp, struct vm_area_struct *vma)
+int drm_legacy_mmap(struct file *filp, struct vm_area_struct *vma)
 {
 	struct drm_file *priv = filp->private_data;
 	struct drm_device *dev = priv->minor->dev;
@@ -661,4 +665,69 @@ int drm_mmap(struct file *filp, struct vm_area_struct *vma)
 
 	return ret;
 }
-EXPORT_SYMBOL(drm_mmap);
+EXPORT_SYMBOL(drm_legacy_mmap);
+
+void drm_legacy_vma_flush(struct drm_device *dev)
+{
+	struct drm_vma_entry *vma, *vma_temp;
+
+	/* Clear vma list (only needed for legacy drivers) */
+	list_for_each_entry_safe(vma, vma_temp, &dev->vmalist, head) {
+		list_del(&vma->head);
+		kfree(vma);
+	}
+}
+
+int drm_vma_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_vma_entry *pt;
+	struct vm_area_struct *vma;
+	unsigned long vma_count = 0;
+#if defined(__i386__)
+	unsigned int pgprot;
+#endif
+
+	mutex_lock(&dev->struct_mutex);
+	list_for_each_entry(pt, &dev->vmalist, head)
+		vma_count++;
+
+	seq_printf(m, "vma use count: %lu, high_memory = %pK, 0x%pK\n",
+		   vma_count, high_memory,
+		   (void *)(unsigned long)virt_to_phys(high_memory));
+
+	list_for_each_entry(pt, &dev->vmalist, head) {
+		vma = pt->vma;
+		if (!vma)
+			continue;
+		seq_printf(m,
+			   "\n%5d 0x%pK-0x%pK %c%c%c%c%c%c 0x%08lx000",
+			   pt->pid,
+			   (void *)vma->vm_start, (void *)vma->vm_end,
+			   vma->vm_flags & VM_READ ? 'r' : '-',
+			   vma->vm_flags & VM_WRITE ? 'w' : '-',
+			   vma->vm_flags & VM_EXEC ? 'x' : '-',
+			   vma->vm_flags & VM_MAYSHARE ? 's' : 'p',
+			   vma->vm_flags & VM_LOCKED ? 'l' : '-',
+			   vma->vm_flags & VM_IO ? 'i' : '-',
+			   vma->vm_pgoff);
+
+#if defined(__i386__)
+		pgprot = pgprot_val(vma->vm_page_prot);
+		seq_printf(m, " %c%c%c%c%c%c%c%c%c",
+			   pgprot & _PAGE_PRESENT ? 'p' : '-',
+			   pgprot & _PAGE_RW ? 'w' : 'r',
+			   pgprot & _PAGE_USER ? 'u' : 's',
+			   pgprot & _PAGE_PWT ? 't' : 'b',
+			   pgprot & _PAGE_PCD ? 'u' : 'c',
+			   pgprot & _PAGE_ACCESSED ? 'a' : '-',
+			   pgprot & _PAGE_DIRTY ? 'd' : '-',
+			   pgprot & _PAGE_PSE ? 'm' : 'k',
+			   pgprot & _PAGE_GLOBAL ? 'g' : 'l');
+#endif
+		seq_printf(m, "\n");
+	}
+	mutex_unlock(&dev->struct_mutex);
+	return 0;
+}
diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig
index 178d2a9672a8..7f9f6f9e9b7e 100644
--- a/drivers/gpu/drm/exynos/Kconfig
+++ b/drivers/gpu/drm/exynos/Kconfig
@@ -28,6 +28,7 @@ config DRM_EXYNOS_FIMD
 	bool "Exynos DRM FIMD"
 	depends on DRM_EXYNOS && !FB_S3C
 	select FB_MODE_HELPERS
+	select MFD_SYSCON
 	help
 	  Choose this option if you want to use Exynos FIMD for DRM.
 
@@ -52,6 +53,7 @@ config DRM_EXYNOS_DP
 	bool "EXYNOS DRM DP driver support"
 	depends on DRM_EXYNOS_FIMD && ARCH_EXYNOS && (DRM_PTN3460=n || DRM_PTN3460=y || DRM_PTN3460=DRM_EXYNOS)
 	default DRM_EXYNOS
+	select DRM_PANEL
 	help
 	  This enables support for DP device.
 
diff --git a/drivers/gpu/drm/exynos/exynos_dp_core.c b/drivers/gpu/drm/exynos/exynos_dp_core.c
index a8ffc8c1477b..6adb1e5cfb08 100644
--- a/drivers/gpu/drm/exynos/exynos_dp_core.c
+++ b/drivers/gpu/drm/exynos/exynos_dp_core.c
@@ -16,7 +16,6 @@
 #include <linux/clk.h>
 #include <linux/io.h>
 #include <linux/interrupt.h>
-#include <linux/delay.h>
 #include <linux/of.h>
 #include <linux/of_gpio.h>
 #include <linux/gpio.h>
@@ -28,6 +27,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_panel.h>
 #include <drm/bridge/ptn3460.h>
 
 #include "exynos_drm_drv.h"
@@ -41,7 +41,7 @@ struct bridge_init {
 	struct device_node *node;
 };
 
-static int exynos_dp_init_dp(struct exynos_dp_device *dp)
+static void exynos_dp_init_dp(struct exynos_dp_device *dp)
 {
 	exynos_dp_reset(dp);
 
@@ -58,8 +58,6 @@ static int exynos_dp_init_dp(struct exynos_dp_device *dp)
 
 	exynos_dp_init_hpd(dp);
 	exynos_dp_init_aux(dp);
-
-	return 0;
 }
 
 static int exynos_dp_detect_hpd(struct exynos_dp_device *dp)
@@ -331,8 +329,8 @@ static int exynos_dp_link_start(struct exynos_dp_device *dp)
 		return retval;
 
 	for (lane = 0; lane < lane_count; lane++)
-		buf[lane] = DP_TRAIN_PRE_EMPHASIS_0 |
-			    DP_TRAIN_VOLTAGE_SWING_400;
+		buf[lane] = DP_TRAIN_PRE_EMPH_LEVEL_0 |
+			    DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 
 	retval = exynos_dp_write_bytes_to_dpcd(dp, DP_TRAINING_LANE0_SET,
 			lane_count, buf);
@@ -875,10 +873,24 @@ static irqreturn_t exynos_dp_irq_handler(int irq, void *arg)
 static void exynos_dp_hotplug(struct work_struct *work)
 {
 	struct exynos_dp_device *dp;
-	int ret;
 
 	dp = container_of(work, struct exynos_dp_device, hotplug_work);
 
+	if (dp->drm_dev)
+		drm_helper_hpd_irq_event(dp->drm_dev);
+}
+
+static void exynos_dp_commit(struct exynos_drm_display *display)
+{
+	struct exynos_dp_device *dp = display->ctx;
+	int ret;
+
+	/* Keep the panel disabled while we configure video */
+	if (dp->panel) {
+		if (drm_panel_disable(dp->panel))
+			DRM_ERROR("failed to disable the panel\n");
+	}
+
 	ret = exynos_dp_detect_hpd(dp);
 	if (ret) {
 		/* Cable has been disconnected, we're done */
@@ -909,6 +921,12 @@ static void exynos_dp_hotplug(struct work_struct *work)
 	ret = exynos_dp_config_video(dp);
 	if (ret)
 		dev_err(dp->dev, "unable to config video\n");
+
+	/* Safe to enable the panel now */
+	if (dp->panel) {
+		if (drm_panel_enable(dp->panel))
+			DRM_ERROR("failed to enable the panel\n");
+	}
 }
 
 static enum drm_connector_status exynos_dp_detect(
@@ -919,6 +937,8 @@ static enum drm_connector_status exynos_dp_detect(
 
 static void exynos_dp_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
 }
 
 static struct drm_connector_funcs exynos_dp_connector_funcs = {
@@ -933,15 +953,18 @@ static int exynos_dp_get_modes(struct drm_connector *connector)
 	struct exynos_dp_device *dp = ctx_from_connector(connector);
 	struct drm_display_mode *mode;
 
+	if (dp->panel)
+		return drm_panel_get_modes(dp->panel);
+
 	mode = drm_mode_create(connector->dev);
 	if (!mode) {
 		DRM_ERROR("failed to create a new display mode.\n");
 		return 0;
 	}
 
-	drm_display_mode_from_videomode(&dp->panel.vm, mode);
-	mode->width_mm = dp->panel.width_mm;
-	mode->height_mm = dp->panel.height_mm;
+	drm_display_mode_from_videomode(&dp->priv.vm, mode);
+	mode->width_mm = dp->priv.width_mm;
+	mode->height_mm = dp->priv.height_mm;
 	connector->display_info.width_mm = mode->width_mm;
 	connector->display_info.height_mm = mode->height_mm;
 
@@ -1018,10 +1041,13 @@ static int exynos_dp_create_connector(struct exynos_drm_display *display,
 	}
 
 	drm_connector_helper_add(connector, &exynos_dp_connector_helper_funcs);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
-	return 0;
+	if (dp->panel)
+		ret = drm_panel_attach(dp->panel, &dp->connector);
+
+	return ret;
 }
 
 static void exynos_dp_phy_init(struct exynos_dp_device *dp)
@@ -1050,26 +1076,50 @@ static void exynos_dp_phy_exit(struct exynos_dp_device *dp)
 	}
 }
 
-static void exynos_dp_poweron(struct exynos_dp_device *dp)
+static void exynos_dp_poweron(struct exynos_drm_display *display)
 {
+	struct exynos_dp_device *dp = display->ctx;
+
 	if (dp->dpms_mode == DRM_MODE_DPMS_ON)
 		return;
 
+	if (dp->panel) {
+		if (drm_panel_prepare(dp->panel)) {
+			DRM_ERROR("failed to setup the panel\n");
+			return;
+		}
+	}
+
 	clk_prepare_enable(dp->clock);
 	exynos_dp_phy_init(dp);
 	exynos_dp_init_dp(dp);
 	enable_irq(dp->irq);
+	exynos_dp_commit(display);
 }
 
-static void exynos_dp_poweroff(struct exynos_dp_device *dp)
+static void exynos_dp_poweroff(struct exynos_drm_display *display)
 {
+	struct exynos_dp_device *dp = display->ctx;
+
 	if (dp->dpms_mode != DRM_MODE_DPMS_ON)
 		return;
 
+	if (dp->panel) {
+		if (drm_panel_disable(dp->panel)) {
+			DRM_ERROR("failed to disable the panel\n");
+			return;
+		}
+	}
+
 	disable_irq(dp->irq);
 	flush_work(&dp->hotplug_work);
 	exynos_dp_phy_exit(dp);
 	clk_disable_unprepare(dp->clock);
+
+	if (dp->panel) {
+		if (drm_panel_unprepare(dp->panel))
+			DRM_ERROR("failed to turnoff the panel\n");
+	}
 }
 
 static void exynos_dp_dpms(struct exynos_drm_display *display, int mode)
@@ -1078,12 +1128,12 @@ static void exynos_dp_dpms(struct exynos_drm_display *display, int mode)
 
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
-		exynos_dp_poweron(dp);
+		exynos_dp_poweron(display);
 		break;
 	case DRM_MODE_DPMS_STANDBY:
 	case DRM_MODE_DPMS_SUSPEND:
 	case DRM_MODE_DPMS_OFF:
-		exynos_dp_poweroff(dp);
+		exynos_dp_poweroff(display);
 		break;
 	default:
 		break;
@@ -1094,6 +1144,7 @@ static void exynos_dp_dpms(struct exynos_drm_display *display, int mode)
 static struct exynos_drm_display_ops exynos_dp_display_ops = {
 	.create_connector = exynos_dp_create_connector,
 	.dpms = exynos_dp_dpms,
+	.commit = exynos_dp_commit,
 };
 
 static struct exynos_drm_display exynos_dp_display = {
@@ -1201,7 +1252,7 @@ static int exynos_dp_dt_parse_panel(struct exynos_dp_device *dp)
 {
 	int ret;
 
-	ret = of_get_videomode(dp->dev->of_node, &dp->panel.vm,
+	ret = of_get_videomode(dp->dev->of_node, &dp->priv.vm,
 			OF_USE_NATIVE_MODE);
 	if (ret) {
 		DRM_ERROR("failed: of_get_videomode() : %d\n", ret);
@@ -1215,16 +1266,10 @@ static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 	struct platform_device *pdev = to_platform_device(dev);
 	struct drm_device *drm_dev = data;
 	struct resource *res;
-	struct exynos_dp_device *dp;
+	struct exynos_dp_device *dp = exynos_dp_display.ctx;
 	unsigned int irq_flags;
-
 	int ret = 0;
 
-	dp = devm_kzalloc(&pdev->dev, sizeof(struct exynos_dp_device),
-				GFP_KERNEL);
-	if (!dp)
-		return -ENOMEM;
-
 	dp->dev = &pdev->dev;
 	dp->dpms_mode = DRM_MODE_DPMS_OFF;
 
@@ -1236,9 +1281,11 @@ static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 	if (ret)
 		return ret;
 
-	ret = exynos_dp_dt_parse_panel(dp);
-	if (ret)
-		return ret;
+	if (!dp->panel) {
+		ret = exynos_dp_dt_parse_panel(dp);
+		if (ret)
+			return ret;
+	}
 
 	dp->clock = devm_clk_get(&pdev->dev, "dp");
 	if (IS_ERR(dp->clock)) {
@@ -1298,7 +1345,6 @@ static int exynos_dp_bind(struct device *dev, struct device *master, void *data)
 	disable_irq(dp->irq);
 
 	dp->drm_dev = drm_dev;
-	exynos_dp_display.ctx = dp;
 
 	platform_set_drvdata(pdev, &exynos_dp_display);
 
@@ -1309,13 +1355,8 @@ static void exynos_dp_unbind(struct device *dev, struct device *master,
 				void *data)
 {
 	struct exynos_drm_display *display = dev_get_drvdata(dev);
-	struct exynos_dp_device *dp = display->ctx;
-	struct drm_encoder *encoder = dp->encoder;
 
 	exynos_dp_dpms(display, DRM_MODE_DPMS_OFF);
-
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&dp->connector);
 }
 
 static const struct component_ops exynos_dp_ops = {
@@ -1325,6 +1366,9 @@ static const struct component_ops exynos_dp_ops = {
 
 static int exynos_dp_probe(struct platform_device *pdev)
 {
+	struct device *dev = &pdev->dev;
+	struct device_node *panel_node;
+	struct exynos_dp_device *dp;
 	int ret;
 
 	ret = exynos_drm_component_add(&pdev->dev, EXYNOS_DEVICE_TYPE_CONNECTOR,
@@ -1332,6 +1376,21 @@ static int exynos_dp_probe(struct platform_device *pdev)
 	if (ret)
 		return ret;
 
+	dp = devm_kzalloc(&pdev->dev, sizeof(struct exynos_dp_device),
+				GFP_KERNEL);
+	if (!dp)
+		return -ENOMEM;
+
+	panel_node = of_parse_phandle(dev->of_node, "panel", 0);
+	if (panel_node) {
+		dp->panel = of_drm_find_panel(panel_node);
+		of_node_put(panel_node);
+		if (!dp->panel)
+			return -EPROBE_DEFER;
+	}
+
+	exynos_dp_display.ctx = dp;
+
 	ret = component_add(&pdev->dev, &exynos_dp_ops);
 	if (ret)
 		exynos_drm_component_del(&pdev->dev,
@@ -1376,6 +1435,7 @@ static const struct of_device_id exynos_dp_match[] = {
 	{ .compatible = "samsung,exynos5-dp" },
 	{},
 };
+MODULE_DEVICE_TABLE(of, exynos_dp_match);
 
 struct platform_driver dp_driver = {
 	.probe		= exynos_dp_probe,
@@ -1390,4 +1450,4 @@ struct platform_driver dp_driver = {
 
 MODULE_AUTHOR("Jingoo Han <jg1.han@samsung.com>");
 MODULE_DESCRIPTION("Samsung SoC DP Driver");
-MODULE_LICENSE("GPL");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/exynos/exynos_dp_core.h b/drivers/gpu/drm/exynos/exynos_dp_core.h
index 02cc4f9ab903..a1aee6931bd7 100644
--- a/drivers/gpu/drm/exynos/exynos_dp_core.h
+++ b/drivers/gpu/drm/exynos/exynos_dp_core.h
@@ -149,6 +149,7 @@ struct exynos_dp_device {
 	struct drm_device	*drm_dev;
 	struct drm_connector	connector;
 	struct drm_encoder	*encoder;
+	struct drm_panel	*panel;
 	struct clk		*clock;
 	unsigned int		irq;
 	void __iomem		*reg_base;
@@ -162,7 +163,7 @@ struct exynos_dp_device {
 	int			dpms_mode;
 	int			hpd_gpio;
 
-	struct exynos_drm_panel_info panel;
+	struct exynos_drm_panel_info priv;
 };
 
 /* exynos_dp_reg.c */
diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c
index 9a16dbe121d1..ba9b3d5ed672 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_connector.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c
@@ -117,20 +117,7 @@ static struct drm_encoder *exynos_drm_best_encoder(
 	struct drm_device *dev = connector->dev;
 	struct exynos_drm_connector *exynos_connector =
 					to_exynos_connector(connector);
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
-	obj = drm_mode_object_find(dev, exynos_connector->encoder_id,
-				   DRM_MODE_OBJECT_ENCODER);
-	if (!obj) {
-		DRM_DEBUG_KMS("Unknown ENCODER ID %d\n",
-				exynos_connector->encoder_id);
-		return NULL;
-	}
-
-	encoder = obj_to_encoder(obj);
-
-	return encoder;
+	return drm_encoder_find(dev, exynos_connector->encoder_id);
 }
 
 static struct drm_connector_helper_funcs exynos_connector_helper_funcs = {
@@ -185,7 +172,7 @@ static void exynos_drm_connector_destroy(struct drm_connector *connector)
 	struct exynos_drm_connector *exynos_connector =
 		to_exynos_connector(connector);
 
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(exynos_connector);
 }
@@ -230,7 +217,7 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev,
 	drm_connector_init(dev, connector, &exynos_connector_funcs, type);
 	drm_connector_helper_add(connector, &exynos_connector_helper_funcs);
 
-	err = drm_sysfs_connector_add(connector);
+	err = drm_connector_register(connector);
 	if (err)
 		goto err_connector;
 
@@ -250,7 +237,7 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev,
 	return connector;
 
 err_sysfs:
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 err_connector:
 	drm_connector_cleanup(connector);
 	kfree(exynos_connector);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
index 95c9435d0266..45026e693225 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
@@ -32,7 +32,6 @@ enum exynos_crtc_mode {
  * Exynos specific crtc structure.
  *
  * @drm_crtc: crtc object.
- * @drm_plane: pointer of private plane object for this crtc
  * @manager: the manager associated with this crtc
  * @pipe: a crtc index created at load() with a new crtc object creation
  *	and the crtc object would be set to private->crtc array
@@ -46,7 +45,6 @@ enum exynos_crtc_mode {
  */
 struct exynos_drm_crtc {
 	struct drm_crtc			drm_crtc;
-	struct drm_plane		*plane;
 	struct exynos_drm_manager	*manager;
 	unsigned int			pipe;
 	unsigned int			dpms;
@@ -69,15 +67,20 @@ static void exynos_drm_crtc_dpms(struct drm_crtc *crtc, int mode)
 
 	if (mode > DRM_MODE_DPMS_ON) {
 		/* wait for the completion of page flip. */
-		wait_event(exynos_crtc->pending_flip_queue,
-				atomic_read(&exynos_crtc->pending_flip) == 0);
-		drm_vblank_off(crtc->dev, exynos_crtc->pipe);
+		if (!wait_event_timeout(exynos_crtc->pending_flip_queue,
+				!atomic_read(&exynos_crtc->pending_flip),
+				HZ/20))
+			atomic_set(&exynos_crtc->pending_flip, 0);
+		drm_crtc_vblank_off(crtc);
 	}
 
 	if (manager->ops->dpms)
 		manager->ops->dpms(manager, mode);
 
 	exynos_crtc->dpms = mode;
+
+	if (mode == DRM_MODE_DPMS_ON)
+		drm_crtc_vblank_on(crtc);
 }
 
 static void exynos_drm_crtc_prepare(struct drm_crtc *crtc)
@@ -92,12 +95,12 @@ static void exynos_drm_crtc_commit(struct drm_crtc *crtc)
 
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
 
-	exynos_plane_commit(exynos_crtc->plane);
+	exynos_plane_commit(crtc->primary);
 
 	if (manager->ops->commit)
 		manager->ops->commit(manager);
 
-	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_ON);
+	exynos_plane_dpms(crtc->primary, DRM_MODE_DPMS_ON);
 }
 
 static bool
@@ -121,10 +124,9 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 	struct exynos_drm_manager *manager = exynos_crtc->manager;
-	struct drm_plane *plane = exynos_crtc->plane;
+	struct drm_framebuffer *fb = crtc->primary->fb;
 	unsigned int crtc_w;
 	unsigned int crtc_h;
-	int ret;
 
 	/*
 	 * copy the mode data adjusted by mode_fixup() into crtc->mode
@@ -132,29 +134,21 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	 */
 	memcpy(&crtc->mode, adjusted_mode, sizeof(*adjusted_mode));
 
-	crtc_w = crtc->primary->fb->width - x;
-	crtc_h = crtc->primary->fb->height - y;
+	crtc_w = fb->width - x;
+	crtc_h = fb->height - y;
 
 	if (manager->ops->mode_set)
 		manager->ops->mode_set(manager, &crtc->mode);
 
-	ret = exynos_plane_mode_set(plane, crtc, crtc->primary->fb, 0, 0, crtc_w, crtc_h,
-				    x, y, crtc_w, crtc_h);
-	if (ret)
-		return ret;
-
-	plane->crtc = crtc;
-	plane->fb = crtc->primary->fb;
-	drm_framebuffer_reference(plane->fb);
-
-	return 0;
+	return exynos_plane_mode_set(crtc->primary, crtc, fb, 0, 0,
+				     crtc_w, crtc_h, x, y, crtc_w, crtc_h);
 }
 
 static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 					  struct drm_framebuffer *old_fb)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
-	struct drm_plane *plane = exynos_crtc->plane;
+	struct drm_framebuffer *fb = crtc->primary->fb;
 	unsigned int crtc_w;
 	unsigned int crtc_h;
 	int ret;
@@ -165,11 +159,11 @@ static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 		return -EPERM;
 	}
 
-	crtc_w = crtc->primary->fb->width - x;
-	crtc_h = crtc->primary->fb->height - y;
+	crtc_w = fb->width - x;
+	crtc_h = fb->height - y;
 
-	ret = exynos_plane_mode_set(plane, crtc, crtc->primary->fb, 0, 0, crtc_w, crtc_h,
-				    x, y, crtc_w, crtc_h);
+	ret = exynos_plane_mode_set(crtc->primary, crtc, fb, 0, 0,
+				    crtc_w, crtc_h, x, y, crtc_w, crtc_h);
 	if (ret)
 		return ret;
 
@@ -259,6 +253,7 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 			spin_lock_irq(&dev->event_lock);
 			drm_vblank_put(dev, exynos_crtc->pipe);
 			list_del(&event->base.link);
+			atomic_set(&exynos_crtc->pending_flip, 0);
 			spin_unlock_irq(&dev->event_lock);
 
 			goto out;
@@ -301,8 +296,7 @@ static int exynos_drm_crtc_set_property(struct drm_crtc *crtc,
 			exynos_drm_crtc_commit(crtc);
 			break;
 		case CRTC_MODE_BLANK:
-			exynos_plane_dpms(exynos_crtc->plane,
-					  DRM_MODE_DPMS_OFF);
+			exynos_plane_dpms(crtc->primary, DRM_MODE_DPMS_OFF);
 			break;
 		default:
 			break;
@@ -348,8 +342,10 @@ static void exynos_drm_crtc_attach_mode_property(struct drm_crtc *crtc)
 int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 {
 	struct exynos_drm_crtc *exynos_crtc;
+	struct drm_plane *plane;
 	struct exynos_drm_private *private = manager->drm_dev->dev_private;
 	struct drm_crtc *crtc;
+	int ret;
 
 	exynos_crtc = kzalloc(sizeof(*exynos_crtc), GFP_KERNEL);
 	if (!exynos_crtc)
@@ -361,11 +357,11 @@ int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 	exynos_crtc->dpms = DRM_MODE_DPMS_OFF;
 	exynos_crtc->manager = manager;
 	exynos_crtc->pipe = manager->pipe;
-	exynos_crtc->plane = exynos_plane_init(manager->drm_dev,
-				1 << manager->pipe, true);
-	if (!exynos_crtc->plane) {
-		kfree(exynos_crtc);
-		return -ENOMEM;
+	plane = exynos_plane_init(manager->drm_dev, 1 << manager->pipe,
+				  DRM_PLANE_TYPE_PRIMARY);
+	if (IS_ERR(plane)) {
+		ret = PTR_ERR(plane);
+		goto err_plane;
 	}
 
 	manager->crtc = &exynos_crtc->drm_crtc;
@@ -373,12 +369,22 @@ int exynos_drm_crtc_create(struct exynos_drm_manager *manager)
 
 	private->crtc[manager->pipe] = crtc;
 
-	drm_crtc_init(manager->drm_dev, crtc, &exynos_crtc_funcs);
+	ret = drm_crtc_init_with_planes(manager->drm_dev, crtc, plane, NULL,
+					&exynos_crtc_funcs);
+	if (ret < 0)
+		goto err_crtc;
+
 	drm_crtc_helper_add(crtc, &exynos_crtc_helper_funcs);
 
 	exynos_drm_crtc_attach_mode_property(crtc);
 
 	return 0;
+
+err_crtc:
+	plane->funcs->destroy(plane);
+err_plane:
+	kfree(exynos_crtc);
+	return ret;
 }
 
 int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int pipe)
@@ -508,3 +514,11 @@ int exynos_drm_crtc_get_pipe_from_type(struct drm_device *drm_dev,
 
 	return -EPERM;
 }
+
+void exynos_drm_crtc_te_handler(struct drm_crtc *crtc)
+{
+	struct exynos_drm_manager *manager = to_exynos_crtc(crtc)->manager;
+
+	if (manager->ops->te_handler)
+		manager->ops->te_handler(manager);
+}
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.h b/drivers/gpu/drm/exynos/exynos_drm_crtc.h
index 9f74b10a8a01..690dcddab725 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.h
@@ -36,4 +36,11 @@ void exynos_drm_crtc_plane_disable(struct drm_crtc *crtc, int zpos);
 int exynos_drm_crtc_get_pipe_from_type(struct drm_device *drm_dev,
 					unsigned int out_type);
 
+/*
+ * This function calls the crtc device(manager)'s te_handler() callback
+ * to trigger to transfer video image at the tearing effect synchronization
+ * signal.
+ */
+void exynos_drm_crtc_te_handler(struct drm_crtc *crtc);
+
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dpi.c b/drivers/gpu/drm/exynos/exynos_drm_dpi.c
index 9e530f205ad2..3dc678ed9949 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dpi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dpi.c
@@ -48,7 +48,7 @@ exynos_dpi_detect(struct drm_connector *connector, bool force)
 
 static void exynos_dpi_connector_destroy(struct drm_connector *connector)
 {
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 }
 
@@ -117,7 +117,7 @@ static int exynos_dpi_create_connector(struct exynos_drm_display *display,
 	}
 
 	drm_connector_helper_add(connector, &exynos_dpi_connector_helper_funcs);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	return 0;
@@ -125,14 +125,18 @@ static int exynos_dpi_create_connector(struct exynos_drm_display *display,
 
 static void exynos_dpi_poweron(struct exynos_dpi *ctx)
 {
-	if (ctx->panel)
+	if (ctx->panel) {
+		drm_panel_prepare(ctx->panel);
 		drm_panel_enable(ctx->panel);
+	}
 }
 
 static void exynos_dpi_poweroff(struct exynos_dpi *ctx)
 {
-	if (ctx->panel)
+	if (ctx->panel) {
 		drm_panel_disable(ctx->panel);
+		drm_panel_unprepare(ctx->panel);
+	}
 }
 
 static void exynos_dpi_dpms(struct exynos_drm_display *display, int mode)
@@ -334,12 +338,12 @@ err_del_component:
 
 int exynos_dpi_remove(struct device *dev)
 {
-	struct drm_encoder *encoder = exynos_dpi_display.encoder;
 	struct exynos_dpi *ctx = exynos_dpi_display.ctx;
 
 	exynos_dpi_dpms(&exynos_dpi_display, DRM_MODE_DPMS_OFF);
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&ctx->connector);
+
+	if (ctx->panel)
+		drm_panel_detach(ctx->panel);
 
 	exynos_drm_component_del(dev, EXYNOS_DEVICE_TYPE_CONNECTOR);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index ab7d182063c3..e5c4c6c8c967 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -15,7 +15,6 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
-#include <linux/anon_inodes.h>
 #include <linux/component.h>
 
 #include <drm/exynos_drm.h>
@@ -39,8 +38,6 @@
 #define DRIVER_MAJOR	1
 #define DRIVER_MINOR	0
 
-#define VBLANK_OFF_DELAY	50000
-
 static struct platform_device *exynos_drm_pdev;
 
 static DEFINE_MUTEX(drm_component_lock);
@@ -88,44 +85,63 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 		struct drm_plane *plane;
 		unsigned long possible_crtcs = (1 << MAX_CRTC) - 1;
 
-		plane = exynos_plane_init(dev, possible_crtcs, false);
-		if (!plane)
-			goto err_mode_config_cleanup;
-	}
-
-	/* init kms poll for handling hpd */
-	drm_kms_helper_poll_init(dev);
+		plane = exynos_plane_init(dev, possible_crtcs,
+					  DRM_PLANE_TYPE_OVERLAY);
+		if (!IS_ERR(plane))
+			continue;
 
-	ret = drm_vblank_init(dev, MAX_CRTC);
-	if (ret)
+		ret = PTR_ERR(plane);
 		goto err_mode_config_cleanup;
+	}
 
 	/* setup possible_clones. */
 	exynos_drm_encoder_setup(dev);
 
-	drm_vblank_offdelay = VBLANK_OFF_DELAY;
-
 	platform_set_drvdata(dev->platformdev, dev);
 
 	/* Try to bind all sub drivers. */
 	ret = component_bind_all(dev->dev, dev);
 	if (ret)
-		goto err_cleanup_vblank;
+		goto err_mode_config_cleanup;
+
+	ret = drm_vblank_init(dev, dev->mode_config.num_crtc);
+	if (ret)
+		goto err_unbind_all;
 
 	/* Probe non kms sub drivers and virtual display driver. */
 	ret = exynos_drm_device_subdrv_probe(dev);
 	if (ret)
-		goto err_unbind_all;
+		goto err_cleanup_vblank;
+
+	/*
+	 * enable drm irq mode.
+	 * - with irq_enabled = true, we can use the vblank feature.
+	 *
+	 * P.S. note that we wouldn't use drm irq handler but
+	 *	just specific driver own one instead because
+	 *	drm framework supports only one irq handler.
+	 */
+	dev->irq_enabled = true;
+
+	/*
+	 * with vblank_disable_allowed = true, vblank interrupt will be disabled
+	 * by drm timer once a current process gives up ownership of
+	 * vblank event.(after drm_vblank_put function is called)
+	 */
+	dev->vblank_disable_allowed = true;
+
+	/* init kms poll for handling hpd */
+	drm_kms_helper_poll_init(dev);
 
 	/* force connectors detection */
 	drm_helper_hpd_irq_event(dev);
 
 	return 0;
 
-err_unbind_all:
-	component_unbind_all(dev->dev, dev);
 err_cleanup_vblank:
 	drm_vblank_cleanup(dev);
+err_unbind_all:
+	component_unbind_all(dev->dev, dev);
 err_mode_config_cleanup:
 	drm_mode_config_cleanup(dev);
 	drm_release_iommu_mapping(dev);
@@ -140,23 +156,19 @@ static int exynos_drm_unload(struct drm_device *dev)
 	exynos_drm_device_subdrv_remove(dev);
 
 	exynos_drm_fbdev_fini(dev);
-	drm_vblank_cleanup(dev);
 	drm_kms_helper_poll_fini(dev);
-	drm_mode_config_cleanup(dev);
 
+	drm_vblank_cleanup(dev);
+	component_unbind_all(dev->dev, dev);
+	drm_mode_config_cleanup(dev);
 	drm_release_iommu_mapping(dev);
-	kfree(dev->dev_private);
 
-	component_unbind_all(dev->dev, dev);
+	kfree(dev->dev_private);
 	dev->dev_private = NULL;
 
 	return 0;
 }
 
-static const struct file_operations exynos_drm_gem_fops = {
-	.mmap = exynos_drm_gem_mmap_buffer,
-};
-
 static int exynos_drm_suspend(struct drm_device *dev, pm_message_t state)
 {
 	struct drm_connector *connector;
@@ -182,8 +194,12 @@ static int exynos_drm_resume(struct drm_device *dev)
 
 	drm_modeset_lock_all(dev);
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (connector->funcs->dpms)
-			connector->funcs->dpms(connector, connector->dpms);
+		if (connector->funcs->dpms) {
+			int dpms = connector->dpms;
+
+			connector->dpms = DRM_MODE_DPMS_OFF;
+			connector->funcs->dpms(connector, dpms);
+		}
 	}
 	drm_modeset_unlock_all(dev);
 
@@ -195,7 +211,6 @@ static int exynos_drm_resume(struct drm_device *dev)
 static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv;
-	struct file *anon_filp;
 	int ret;
 
 	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
@@ -208,21 +223,8 @@ static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 	if (ret)
 		goto err_file_priv_free;
 
-	anon_filp = anon_inode_getfile("exynos_gem", &exynos_drm_gem_fops,
-					NULL, 0);
-	if (IS_ERR(anon_filp)) {
-		ret = PTR_ERR(anon_filp);
-		goto err_subdrv_close;
-	}
-
-	anon_filp->f_mode = FMODE_READ | FMODE_WRITE;
-	file_priv->anon_filp = anon_filp;
-
 	return ret;
 
-err_subdrv_close:
-	exynos_drm_subdrv_close(dev, file);
-
 err_file_priv_free:
 	kfree(file_priv);
 	file->driver_priv = NULL;
@@ -238,7 +240,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 {
 	struct exynos_drm_private *private = dev->dev_private;
-	struct drm_exynos_file_private *file_priv;
 	struct drm_pending_vblank_event *v, *vt;
 	struct drm_pending_event *e, *et;
 	unsigned long flags;
@@ -264,10 +265,6 @@ static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 	}
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
-	file_priv = file->driver_priv;
-	if (file_priv->anon_filp)
-		fput(file_priv->anon_filp);
-
 	kfree(file->driver_priv);
 	file->driver_priv = NULL;
 }
@@ -286,11 +283,6 @@ static const struct vm_operations_struct exynos_drm_gem_vm_ops = {
 static const struct drm_ioctl_desc exynos_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_CREATE, exynos_drm_gem_create_ioctl,
 			DRM_UNLOCKED | DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_MAP_OFFSET,
-			exynos_drm_gem_map_offset_ioctl, DRM_UNLOCKED |
-			DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_MMAP,
-			exynos_drm_gem_mmap_ioctl, DRM_UNLOCKED | DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(EXYNOS_GEM_GET,
 			exynos_drm_gem_get_ioctl, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(EXYNOS_VIDI_CONNECTION,
@@ -334,6 +326,7 @@ static struct drm_driver exynos_drm_driver = {
 	.preclose		= exynos_drm_preclose,
 	.lastclose		= exynos_drm_lastclose,
 	.postclose		= exynos_drm_postclose,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= exynos_drm_crtc_enable_vblank,
 	.disable_vblank		= exynos_drm_crtc_disable_vblank,
@@ -362,7 +355,7 @@ static int exynos_drm_sys_suspend(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	pm_message_t message;
 
-	if (pm_runtime_suspended(dev))
+	if (pm_runtime_suspended(dev) || !drm_dev)
 		return 0;
 
 	message.event = PM_EVENT_SUSPEND;
@@ -373,7 +366,7 @@ static int exynos_drm_sys_resume(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 
-	if (pm_runtime_suspended(dev))
+	if (pm_runtime_suspended(dev) || !drm_dev)
 		return 0;
 
 	return exynos_drm_resume(drm_dev);
@@ -489,21 +482,26 @@ void exynos_drm_component_del(struct device *dev,
 	mutex_unlock(&drm_component_lock);
 }
 
-static int compare_of(struct device *dev, void *data)
+static int compare_dev(struct device *dev, void *data)
 {
 	return dev == (struct device *)data;
 }
 
-static int exynos_drm_add_components(struct device *dev, struct master *m)
+static struct component_match *exynos_drm_match_add(struct device *dev)
 {
+	struct component_match *match = NULL;
 	struct component_dev *cdev;
 	unsigned int attach_cnt = 0;
 
 	mutex_lock(&drm_component_lock);
 
-	list_for_each_entry(cdev, &drm_component_list, list) {
-		int ret;
+	/* Do not retry to probe if there is no any kms driver regitered. */
+	if (list_empty(&drm_component_list)) {
+		mutex_unlock(&drm_component_lock);
+		return ERR_PTR(-ENODEV);
+	}
 
+	list_for_each_entry(cdev, &drm_component_list, list) {
 		/*
 		 * Add components to master only in case that crtc and
 		 * encoder/connector device objects exist.
@@ -518,16 +516,10 @@ static int exynos_drm_add_components(struct device *dev, struct master *m)
 		/*
 		 * fimd and dpi modules have same device object so add
 		 * only crtc device object in this case.
-		 *
-		 * TODO. if dpi module follows driver-model driver then
-		 * below codes can be removed.
 		 */
 		if (cdev->crtc_dev == cdev->conn_dev) {
-			ret = component_master_add_child(m, compare_of,
-					cdev->crtc_dev);
-			if (ret < 0)
-				return ret;
-
+			component_match_add(dev, &match, compare_dev,
+						cdev->crtc_dev);
 			goto out_lock;
 		}
 
@@ -537,11 +529,8 @@ static int exynos_drm_add_components(struct device *dev, struct master *m)
 		 * connector/encoder need pipe number of crtc when they
 		 * are created.
 		 */
-		ret = component_master_add_child(m, compare_of, cdev->crtc_dev);
-		ret |= component_master_add_child(m, compare_of,
-							cdev->conn_dev);
-		if (ret < 0)
-			return ret;
+		component_match_add(dev, &match, compare_dev, cdev->crtc_dev);
+		component_match_add(dev, &match, compare_dev, cdev->conn_dev);
 
 out_lock:
 		mutex_lock(&drm_component_lock);
@@ -549,7 +538,7 @@ out_lock:
 
 	mutex_unlock(&drm_component_lock);
 
-	return attach_cnt ? 0 : -ENODEV;
+	return attach_cnt ? match : ERR_PTR(-EPROBE_DEFER);
 }
 
 static int exynos_drm_bind(struct device *dev)
@@ -563,13 +552,13 @@ static void exynos_drm_unbind(struct device *dev)
 }
 
 static const struct component_master_ops exynos_drm_ops = {
-	.add_components = exynos_drm_add_components,
 	.bind		= exynos_drm_bind,
 	.unbind		= exynos_drm_unbind,
 };
 
 static int exynos_drm_platform_probe(struct platform_device *pdev)
 {
+	struct component_match *match;
 	int ret;
 
 	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
@@ -602,10 +591,21 @@ static int exynos_drm_platform_probe(struct platform_device *pdev)
 		goto err_unregister_mixer_drv;
 #endif
 
+	match = exynos_drm_match_add(&pdev->dev);
+	if (IS_ERR(match)) {
+		ret = PTR_ERR(match);
+		goto err_unregister_hdmi_drv;
+	}
+
+	ret = component_master_add_with_match(&pdev->dev, &exynos_drm_ops,
+						match);
+	if (ret < 0)
+		goto err_unregister_hdmi_drv;
+
 #ifdef CONFIG_DRM_EXYNOS_G2D
 	ret = platform_driver_register(&g2d_driver);
 	if (ret < 0)
-		goto err_unregister_hdmi_drv;
+		goto err_del_component_master;
 #endif
 
 #ifdef CONFIG_DRM_EXYNOS_FIMC
@@ -636,11 +636,7 @@ static int exynos_drm_platform_probe(struct platform_device *pdev)
 		goto err_unregister_ipp_drv;
 #endif
 
-	ret = component_master_add(&pdev->dev, &exynos_drm_ops);
-	if (ret < 0)
-		DRM_DEBUG_KMS("re-tried by last sub driver probed later.\n");
-
-	return 0;
+	return ret;
 
 #ifdef CONFIG_DRM_EXYNOS_IPP
 err_unregister_ipp_drv:
@@ -665,9 +661,11 @@ err_unregister_g2d_drv:
 
 #ifdef CONFIG_DRM_EXYNOS_G2D
 	platform_driver_unregister(&g2d_driver);
-err_unregister_hdmi_drv:
+err_del_component_master:
 #endif
+	component_master_del(&pdev->dev, &exynos_drm_ops);
 
+err_unregister_hdmi_drv:
 #ifdef CONFIG_DRM_EXYNOS_HDMI
 	platform_driver_unregister(&hdmi_driver);
 err_unregister_mixer_drv:
@@ -748,6 +746,18 @@ static int exynos_drm_init(void)
 {
 	int ret;
 
+	/*
+	 * Register device object only in case of Exynos SoC.
+	 *
+	 * Below codes resolves temporarily infinite loop issue incurred
+	 * by Exynos drm driver when using multi-platform kernel.
+	 * So these codes will be replaced with more generic way later.
+	 */
+	if (!of_machine_is_compatible("samsung,exynos3") &&
+			!of_machine_is_compatible("samsung,exynos4") &&
+			!of_machine_is_compatible("samsung,exynos5"))
+		return -ENODEV;
+
 	exynos_drm_pdev = platform_device_register_simple("exynos-drm", -1,
 								NULL, 0);
 	if (IS_ERR(exynos_drm_pdev))
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index 06cde4506278..d22e640f59a0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -40,8 +40,6 @@ struct drm_device;
 struct exynos_drm_overlay;
 struct drm_connector;
 
-extern unsigned int drm_vblank_offdelay;
-
 /* This enumerates device type. */
 enum exynos_drm_device_type {
 	EXYNOS_DEVICE_TYPE_NONE,
@@ -188,6 +186,8 @@ struct exynos_drm_display {
  * @win_commit: apply hardware specific overlay data to registers.
  * @win_enable: enable hardware specific overlay.
  * @win_disable: disable hardware specific overlay.
+ * @te_handler: trigger to transfer video image at the tearing effect
+ *	synchronization signal if there is a page flip request.
  */
 struct exynos_drm_manager;
 struct exynos_drm_manager_ops {
@@ -206,6 +206,7 @@ struct exynos_drm_manager_ops {
 	void (*win_commit)(struct exynos_drm_manager *mgr, int zpos);
 	void (*win_enable)(struct exynos_drm_manager *mgr, int zpos);
 	void (*win_disable)(struct exynos_drm_manager *mgr, int zpos);
+	void (*te_handler)(struct exynos_drm_manager *mgr);
 };
 
 /*
@@ -236,15 +237,9 @@ struct exynos_drm_g2d_private {
 	struct list_head	userptr_list;
 };
 
-struct exynos_drm_ipp_private {
-	struct device	*dev;
-	struct list_head	event_list;
-};
-
 struct drm_exynos_file_private {
 	struct exynos_drm_g2d_private	*g2d_priv;
-	struct exynos_drm_ipp_private	*ipp_priv;
-	struct file			*anon_filp;
+	struct device			*ipp_dev;
 };
 
 /*
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dsi.c b/drivers/gpu/drm/exynos/exynos_drm_dsi.c
index 6302aa64f6c1..acf7e9e39dcd 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dsi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dsi.c
@@ -16,7 +16,10 @@
 #include <drm/drm_panel.h>
 
 #include <linux/clk.h>
+#include <linux/gpio/consumer.h>
 #include <linux/irq.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.h>
 #include <linux/phy/phy.h>
 #include <linux/regulator/consumer.h>
 #include <linux/component.h>
@@ -24,6 +27,7 @@
 #include <video/mipi_display.h>
 #include <video/videomode.h>
 
+#include "exynos_drm_crtc.h"
 #include "exynos_drm_drv.h"
 
 /* returns true iff both arguments logically differs */
@@ -54,9 +58,12 @@
 
 /* FIFO memory AC characteristic register */
 #define DSIM_PLLCTRL_REG	0x4c	/* PLL control register */
-#define DSIM_PLLTMR_REG		0x50	/* PLL timer register */
 #define DSIM_PHYACCHR_REG	0x54	/* D-PHY AC characteristic register */
 #define DSIM_PHYACCHR1_REG	0x58	/* D-PHY AC characteristic register1 */
+#define DSIM_PHYCTRL_REG	0x5c
+#define DSIM_PHYTIMING_REG	0x64
+#define DSIM_PHYTIMING1_REG	0x68
+#define DSIM_PHYTIMING2_REG	0x6c
 
 /* DSIM_STATUS */
 #define DSIM_STOP_STATE_DAT(x)		(((x) & 0xf) << 0)
@@ -107,6 +114,8 @@
 #define DSIM_SYNC_INFORM		(1 << 27)
 #define DSIM_EOT_DISABLE		(1 << 28)
 #define DSIM_MFLUSH_VS			(1 << 29)
+/* This flag is valid only for exynos3250/3472/4415/5260/5430 */
+#define DSIM_CLKLANE_STOP		(1 << 30)
 
 /* DSIM_ESCMODE */
 #define DSIM_TX_TRIGGER_RST		(1 << 4)
@@ -200,6 +209,24 @@
 #define DSIM_PLL_M(x)			((x) << 4)
 #define DSIM_PLL_S(x)			((x) << 1)
 
+/* DSIM_PHYCTRL */
+#define DSIM_PHYCTRL_ULPS_EXIT(x)	(((x) & 0x1ff) << 0)
+
+/* DSIM_PHYTIMING */
+#define DSIM_PHYTIMING_LPX(x)		((x) << 8)
+#define DSIM_PHYTIMING_HS_EXIT(x)	((x) << 0)
+
+/* DSIM_PHYTIMING1 */
+#define DSIM_PHYTIMING1_CLK_PREPARE(x)	((x) << 24)
+#define DSIM_PHYTIMING1_CLK_ZERO(x)	((x) << 16)
+#define DSIM_PHYTIMING1_CLK_POST(x)	((x) << 8)
+#define DSIM_PHYTIMING1_CLK_TRAIL(x)	((x) << 0)
+
+/* DSIM_PHYTIMING2 */
+#define DSIM_PHYTIMING2_HS_PREPARE(x)	((x) << 16)
+#define DSIM_PHYTIMING2_HS_ZERO(x)	((x) << 8)
+#define DSIM_PHYTIMING2_HS_TRAIL(x)	((x) << 0)
+
 #define DSI_MAX_BUS_WIDTH		4
 #define DSI_NUM_VIRTUAL_CHANNELS	4
 #define DSI_TX_FIFO_SIZE		2048
@@ -233,6 +260,13 @@ struct exynos_dsi_transfer {
 #define DSIM_STATE_INITIALIZED		BIT(1)
 #define DSIM_STATE_CMD_LPM		BIT(2)
 
+struct exynos_dsi_driver_data {
+	unsigned int plltmr_reg;
+
+	unsigned int has_freqband:1;
+	unsigned int has_clklane_stop:1;
+};
+
 struct exynos_dsi {
 	struct mipi_dsi_host dsi_host;
 	struct drm_connector connector;
@@ -247,6 +281,7 @@ struct exynos_dsi {
 	struct clk *bus_clk;
 	struct regulator_bulk_data supplies[2];
 	int irq;
+	int te_gpio;
 
 	u32 pll_clk_rate;
 	u32 burst_clk_rate;
@@ -262,11 +297,48 @@ struct exynos_dsi {
 
 	spinlock_t transfer_lock; /* protects transfer_list */
 	struct list_head transfer_list;
+
+	struct exynos_dsi_driver_data *driver_data;
 };
 
 #define host_to_dsi(host) container_of(host, struct exynos_dsi, dsi_host)
 #define connector_to_dsi(c) container_of(c, struct exynos_dsi, connector)
 
+static struct exynos_dsi_driver_data exynos3_dsi_driver_data = {
+	.plltmr_reg = 0x50,
+	.has_freqband = 1,
+	.has_clklane_stop = 1,
+};
+
+static struct exynos_dsi_driver_data exynos4_dsi_driver_data = {
+	.plltmr_reg = 0x50,
+	.has_freqband = 1,
+	.has_clklane_stop = 1,
+};
+
+static struct exynos_dsi_driver_data exynos5_dsi_driver_data = {
+	.plltmr_reg = 0x58,
+};
+
+static struct of_device_id exynos_dsi_of_match[] = {
+	{ .compatible = "samsung,exynos3250-mipi-dsi",
+	  .data = &exynos3_dsi_driver_data },
+	{ .compatible = "samsung,exynos4210-mipi-dsi",
+	  .data = &exynos4_dsi_driver_data },
+	{ .compatible = "samsung,exynos5410-mipi-dsi",
+	  .data = &exynos5_dsi_driver_data },
+	{ }
+};
+
+static inline struct exynos_dsi_driver_data *exynos_dsi_get_driver_data(
+						struct platform_device *pdev)
+{
+	const struct of_device_id *of_id =
+			of_match_device(exynos_dsi_of_match, &pdev->dev);
+
+	return (struct exynos_dsi_driver_data *)of_id->data;
+}
+
 static void exynos_dsi_wait_for_reset(struct exynos_dsi *dsi)
 {
 	if (wait_for_completion_timeout(&dsi->completed, msecs_to_jiffies(300)))
@@ -340,14 +412,9 @@ static unsigned long exynos_dsi_pll_find_pms(struct exynos_dsi *dsi,
 static unsigned long exynos_dsi_set_pll(struct exynos_dsi *dsi,
 					unsigned long freq)
 {
-	static const unsigned long freq_bands[] = {
-		100 * MHZ, 120 * MHZ, 160 * MHZ, 200 * MHZ,
-		270 * MHZ, 320 * MHZ, 390 * MHZ, 450 * MHZ,
-		510 * MHZ, 560 * MHZ, 640 * MHZ, 690 * MHZ,
-		770 * MHZ, 870 * MHZ, 950 * MHZ,
-	};
+	struct exynos_dsi_driver_data *driver_data = dsi->driver_data;
 	unsigned long fin, fout;
-	int timeout, band;
+	int timeout;
 	u8 p, s;
 	u16 m;
 	u32 reg;
@@ -366,27 +433,39 @@ static unsigned long exynos_dsi_set_pll(struct exynos_dsi *dsi,
 	if (!fout) {
 		dev_err(dsi->dev,
 			"failed to find PLL PMS for requested frequency\n");
-		return -EFAULT;
+		return 0;
 	}
+	dev_dbg(dsi->dev, "PLL freq %lu, (p %d, m %d, s %d)\n", fout, p, m, s);
 
-	for (band = 0; band < ARRAY_SIZE(freq_bands); ++band)
-		if (fout < freq_bands[band])
-			break;
+	writel(500, dsi->reg_base + driver_data->plltmr_reg);
 
-	dev_dbg(dsi->dev, "PLL freq %lu, (p %d, m %d, s %d), band %d\n", fout,
-		p, m, s, band);
+	reg = DSIM_PLL_EN | DSIM_PLL_P(p) | DSIM_PLL_M(m) | DSIM_PLL_S(s);
 
-	writel(500, dsi->reg_base + DSIM_PLLTMR_REG);
+	if (driver_data->has_freqband) {
+		static const unsigned long freq_bands[] = {
+			100 * MHZ, 120 * MHZ, 160 * MHZ, 200 * MHZ,
+			270 * MHZ, 320 * MHZ, 390 * MHZ, 450 * MHZ,
+			510 * MHZ, 560 * MHZ, 640 * MHZ, 690 * MHZ,
+			770 * MHZ, 870 * MHZ, 950 * MHZ,
+		};
+		int band;
+
+		for (band = 0; band < ARRAY_SIZE(freq_bands); ++band)
+			if (fout < freq_bands[band])
+				break;
+
+		dev_dbg(dsi->dev, "band %d\n", band);
+
+		reg |= DSIM_FREQ_BAND(band);
+	}
 
-	reg = DSIM_FREQ_BAND(band) | DSIM_PLL_EN
-			| DSIM_PLL_P(p) | DSIM_PLL_M(m) | DSIM_PLL_S(s);
 	writel(reg, dsi->reg_base + DSIM_PLLCTRL_REG);
 
 	timeout = 1000;
 	do {
 		if (timeout-- == 0) {
 			dev_err(dsi->dev, "PLL failed to stabilize\n");
-			return -EFAULT;
+			return 0;
 		}
 		reg = readl(dsi->reg_base + DSIM_STATUS_REG);
 	} while ((reg & DSIM_PLL_STABLE) == 0);
@@ -433,6 +512,59 @@ static int exynos_dsi_enable_clock(struct exynos_dsi *dsi)
 	return 0;
 }
 
+static void exynos_dsi_set_phy_ctrl(struct exynos_dsi *dsi)
+{
+	struct exynos_dsi_driver_data *driver_data = dsi->driver_data;
+	u32 reg;
+
+	if (driver_data->has_freqband)
+		return;
+
+	/* B D-PHY: D-PHY Master & Slave Analog Block control */
+	reg = DSIM_PHYCTRL_ULPS_EXIT(0x0af);
+	writel(reg, dsi->reg_base + DSIM_PHYCTRL_REG);
+
+	/*
+	 * T LPX: Transmitted length of any Low-Power state period
+	 * T HS-EXIT: Time that the transmitter drives LP-11 following a HS
+	 *	burst
+	 */
+	reg = DSIM_PHYTIMING_LPX(0x06) | DSIM_PHYTIMING_HS_EXIT(0x0b);
+	writel(reg, dsi->reg_base + DSIM_PHYTIMING_REG);
+
+	/*
+	 * T CLK-PREPARE: Time that the transmitter drives the Clock Lane LP-00
+	 *	Line state immediately before the HS-0 Line state starting the
+	 *	HS transmission
+	 * T CLK-ZERO: Time that the transmitter drives the HS-0 state prior to
+	 *	transmitting the Clock.
+	 * T CLK_POST: Time that the transmitter continues to send HS clock
+	 *	after the last associated Data Lane has transitioned to LP Mode
+	 *	Interval is defined as the period from the end of T HS-TRAIL to
+	 *	the beginning of T CLK-TRAIL
+	 * T CLK-TRAIL: Time that the transmitter drives the HS-0 state after
+	 *	the last payload clock bit of a HS transmission burst
+	 */
+	reg = DSIM_PHYTIMING1_CLK_PREPARE(0x07) |
+			DSIM_PHYTIMING1_CLK_ZERO(0x27) |
+			DSIM_PHYTIMING1_CLK_POST(0x0d) |
+			DSIM_PHYTIMING1_CLK_TRAIL(0x08);
+	writel(reg, dsi->reg_base + DSIM_PHYTIMING1_REG);
+
+	/*
+	 * T HS-PREPARE: Time that the transmitter drives the Data Lane LP-00
+	 *	Line state immediately before the HS-0 Line state starting the
+	 *	HS transmission
+	 * T HS-ZERO: Time that the transmitter drives the HS-0 state prior to
+	 *	transmitting the Sync sequence.
+	 * T HS-TRAIL: Time that the transmitter drives the flipped differential
+	 *	state after last payload data bit of a HS transmission burst
+	 */
+	reg = DSIM_PHYTIMING2_HS_PREPARE(0x09) | DSIM_PHYTIMING2_HS_ZERO(0x0d) |
+			DSIM_PHYTIMING2_HS_TRAIL(0x0b);
+	writel(reg, dsi->reg_base + DSIM_PHYTIMING2_REG);
+}
+
 static void exynos_dsi_disable_clock(struct exynos_dsi *dsi)
 {
 	u32 reg;
@@ -449,6 +581,7 @@ static void exynos_dsi_disable_clock(struct exynos_dsi *dsi)
 
 static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 {
+	struct exynos_dsi_driver_data *driver_data = dsi->driver_data;
 	int timeout;
 	u32 reg;
 	u32 lanes_mask;
@@ -468,13 +601,20 @@ static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 	/* DSI configuration */
 	reg = 0;
 
+	/*
+	 * The first bit of mode_flags specifies display configuration.
+	 * If this bit is set[= MIPI_DSI_MODE_VIDEO], dsi will support video
+	 * mode, otherwise it will support command mode.
+	 */
 	if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO) {
 		reg |= DSIM_VIDEO_MODE;
 
+		/*
+		 * The user manual describes that following bits are ignored in
+		 * command mode.
+		 */
 		if (!(dsi->mode_flags & MIPI_DSI_MODE_VSYNC_FLUSH))
 			reg |= DSIM_MFLUSH_VS;
-		if (!(dsi->mode_flags & MIPI_DSI_MODE_EOT_PACKET))
-			reg |= DSIM_EOT_DISABLE;
 		if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE)
 			reg |= DSIM_SYNC_INFORM;
 		if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_BURST)
@@ -491,6 +631,9 @@ static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 			reg |= DSIM_HSA_MODE;
 	}
 
+	if (!(dsi->mode_flags & MIPI_DSI_MODE_EOT_PACKET))
+		reg |= DSIM_EOT_DISABLE;
+
 	switch (dsi->format) {
 	case MIPI_DSI_FMT_RGB888:
 		reg |= DSIM_MAIN_PIX_FORMAT_RGB888;
@@ -520,6 +663,20 @@ static int exynos_dsi_init_link(struct exynos_dsi *dsi)
 	reg |= DSIM_LANE_EN(lanes_mask);
 	writel(reg, dsi->reg_base + DSIM_CONFIG_REG);
 
+	/*
+	 * Use non-continuous clock mode if the periparal wants and
+	 * host controller supports
+	 *
+	 * In non-continous clock mode, host controller will turn off
+	 * the HS clock between high-speed transmissions to reduce
+	 * power consumption.
+	 */
+	if (driver_data->has_clklane_stop &&
+			dsi->mode_flags & MIPI_DSI_CLOCK_NON_CONTINUOUS) {
+		reg |= DSIM_CLKLANE_STOP;
+		writel(reg, dsi->reg_base + DSIM_CONFIG_REG);
+	}
+
 	/* Check clock and data lane state are stop state */
 	timeout = 100;
 	do {
@@ -944,17 +1101,90 @@ static irqreturn_t exynos_dsi_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+static irqreturn_t exynos_dsi_te_irq_handler(int irq, void *dev_id)
+{
+	struct exynos_dsi *dsi = (struct exynos_dsi *)dev_id;
+	struct drm_encoder *encoder = dsi->encoder;
+
+	if (dsi->state & DSIM_STATE_ENABLED)
+		exynos_drm_crtc_te_handler(encoder->crtc);
+
+	return IRQ_HANDLED;
+}
+
+static void exynos_dsi_enable_irq(struct exynos_dsi *dsi)
+{
+	enable_irq(dsi->irq);
+
+	if (gpio_is_valid(dsi->te_gpio))
+		enable_irq(gpio_to_irq(dsi->te_gpio));
+}
+
+static void exynos_dsi_disable_irq(struct exynos_dsi *dsi)
+{
+	if (gpio_is_valid(dsi->te_gpio))
+		disable_irq(gpio_to_irq(dsi->te_gpio));
+
+	disable_irq(dsi->irq);
+}
+
 static int exynos_dsi_init(struct exynos_dsi *dsi)
 {
-	exynos_dsi_enable_clock(dsi);
 	exynos_dsi_reset(dsi);
-	enable_irq(dsi->irq);
+	exynos_dsi_enable_irq(dsi);
+	exynos_dsi_enable_clock(dsi);
 	exynos_dsi_wait_for_reset(dsi);
+	exynos_dsi_set_phy_ctrl(dsi);
 	exynos_dsi_init_link(dsi);
 
 	return 0;
 }
 
+static int exynos_dsi_register_te_irq(struct exynos_dsi *dsi)
+{
+	int ret;
+
+	dsi->te_gpio = of_get_named_gpio(dsi->panel_node, "te-gpios", 0);
+	if (!gpio_is_valid(dsi->te_gpio)) {
+		dev_err(dsi->dev, "no te-gpios specified\n");
+		ret = dsi->te_gpio;
+		goto out;
+	}
+
+	ret = gpio_request_one(dsi->te_gpio, GPIOF_IN, "te_gpio");
+	if (ret) {
+		dev_err(dsi->dev, "gpio request failed with %d\n", ret);
+		goto out;
+	}
+
+	/*
+	 * This TE GPIO IRQ should not be set to IRQ_NOAUTOEN, because panel
+	 * calls drm_panel_init() first then calls mipi_dsi_attach() in probe().
+	 * It means that te_gpio is invalid when exynos_dsi_enable_irq() is
+	 * called by drm_panel_init() before panel is attached.
+	 */
+	ret = request_threaded_irq(gpio_to_irq(dsi->te_gpio),
+					exynos_dsi_te_irq_handler, NULL,
+					IRQF_TRIGGER_RISING, "TE", dsi);
+	if (ret) {
+		dev_err(dsi->dev, "request interrupt failed with %d\n", ret);
+		gpio_free(dsi->te_gpio);
+		goto out;
+	}
+
+out:
+	return ret;
+}
+
+static void exynos_dsi_unregister_te_irq(struct exynos_dsi *dsi)
+{
+	if (gpio_is_valid(dsi->te_gpio)) {
+		free_irq(gpio_to_irq(dsi->te_gpio), dsi);
+		gpio_free(dsi->te_gpio);
+		dsi->te_gpio = -ENOENT;
+	}
+}
+
 static int exynos_dsi_host_attach(struct mipi_dsi_host *host,
 				  struct mipi_dsi_device *device)
 {
@@ -968,6 +1198,19 @@ static int exynos_dsi_host_attach(struct mipi_dsi_host *host,
 	if (dsi->connector.dev)
 		drm_helper_hpd_irq_event(dsi->connector.dev);
 
+	/*
+	 * This is a temporary solution and should be made by more generic way.
+	 *
+	 * If attached panel device is for command mode one, dsi should register
+	 * TE interrupt handler.
+	 */
+	if (!(dsi->mode_flags & MIPI_DSI_MODE_VIDEO)) {
+		int ret = exynos_dsi_register_te_irq(dsi);
+
+		if (ret)
+			return ret;
+	}
+
 	return 0;
 }
 
@@ -976,6 +1219,8 @@ static int exynos_dsi_host_detach(struct mipi_dsi_host *host,
 {
 	struct exynos_dsi *dsi = host_to_dsi(host);
 
+	exynos_dsi_unregister_te_irq(dsi);
+
 	dsi->panel_node = NULL;
 
 	if (dsi->connector.dev)
@@ -1089,7 +1334,7 @@ static void exynos_dsi_poweroff(struct exynos_dsi *dsi)
 
 		exynos_dsi_disable_clock(dsi);
 
-		disable_irq(dsi->irq);
+		exynos_dsi_disable_irq(dsi);
 	}
 
 	dsi->state &= ~DSIM_STATE_CMD_LPM;
@@ -1115,7 +1360,7 @@ static int exynos_dsi_enable(struct exynos_dsi *dsi)
 	if (ret < 0)
 		return ret;
 
-	ret = drm_panel_enable(dsi->panel);
+	ret = drm_panel_prepare(dsi->panel);
 	if (ret < 0) {
 		exynos_dsi_poweroff(dsi);
 		return ret;
@@ -1124,6 +1369,14 @@ static int exynos_dsi_enable(struct exynos_dsi *dsi)
 	exynos_dsi_set_display_mode(dsi);
 	exynos_dsi_set_display_enable(dsi, true);
 
+	ret = drm_panel_enable(dsi->panel);
+	if (ret < 0) {
+		exynos_dsi_set_display_enable(dsi, false);
+		drm_panel_unprepare(dsi->panel);
+		exynos_dsi_poweroff(dsi);
+		return ret;
+	}
+
 	dsi->state |= DSIM_STATE_ENABLED;
 
 	return 0;
@@ -1134,8 +1387,9 @@ static void exynos_dsi_disable(struct exynos_dsi *dsi)
 	if (!(dsi->state & DSIM_STATE_ENABLED))
 		return;
 
-	exynos_dsi_set_display_enable(dsi, false);
 	drm_panel_disable(dsi->panel);
+	exynos_dsi_set_display_enable(dsi, false);
+	drm_panel_unprepare(dsi->panel);
 	exynos_dsi_poweroff(dsi);
 
 	dsi->state &= ~DSIM_STATE_ENABLED;
@@ -1187,6 +1441,9 @@ exynos_dsi_detect(struct drm_connector *connector, bool force)
 
 static void exynos_dsi_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+	connector->dev = NULL;
 }
 
 static struct drm_connector_funcs exynos_dsi_connector_funcs = {
@@ -1246,7 +1503,7 @@ static int exynos_dsi_create_connector(struct exynos_drm_display *display,
 	}
 
 	drm_connector_helper_add(connector, &exynos_dsi_connector_helper_funcs);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	return 0;
@@ -1278,6 +1535,7 @@ static struct exynos_drm_display exynos_dsi_display = {
 	.type = EXYNOS_DISPLAY_TYPE_LCD,
 	.ops = &exynos_dsi_display_ops,
 };
+MODULE_DEVICE_TABLE(of, exynos_dsi_of_match);
 
 /* of_* functions will be removed after merge of of_graph patches */
 static struct device_node *
@@ -1402,14 +1660,10 @@ static void exynos_dsi_unbind(struct device *dev, struct device *master,
 				void *data)
 {
 	struct exynos_dsi *dsi = exynos_dsi_display.ctx;
-	struct drm_encoder *encoder = dsi->encoder;
 
 	exynos_dsi_dpms(&exynos_dsi_display, DRM_MODE_DPMS_OFF);
 
 	mipi_dsi_host_unregister(&dsi->dsi_host);
-
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&dsi->connector);
 }
 
 static const struct component_ops exynos_dsi_component_ops = {
@@ -1435,6 +1689,9 @@ static int exynos_dsi_probe(struct platform_device *pdev)
 		goto err_del_component;
 	}
 
+	/* To be checked as invalid one */
+	dsi->te_gpio = -ENOENT;
+
 	init_completion(&dsi->completed);
 	spin_lock_init(&dsi->transfer_lock);
 	INIT_LIST_HEAD(&dsi->transfer_list);
@@ -1443,6 +1700,7 @@ static int exynos_dsi_probe(struct platform_device *pdev)
 	dsi->dsi_host.dev = &pdev->dev;
 
 	dsi->dev = &pdev->dev;
+	dsi->driver_data = exynos_dsi_get_driver_data(pdev);
 
 	ret = exynos_dsi_parse_dt(dsi);
 	if (ret)
@@ -1525,11 +1783,6 @@ static int exynos_dsi_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static struct of_device_id exynos_dsi_of_match[] = {
-	{ .compatible = "samsung,exynos4210-mipi-dsi" },
-	{ }
-};
-
 struct platform_driver dsi_driver = {
 	.probe = exynos_dsi_probe,
 	.remove = exynos_dsi_remove,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 65a22cad7b36..d346d1e6eda0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -165,6 +165,7 @@ exynos_drm_framebuffer_init(struct drm_device *dev,
 
 	ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs);
 	if (ret) {
+		kfree(exynos_fb);
 		DRM_ERROR("failed to initialize framebuffer\n");
 		return ERR_PTR(ret);
 	}
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index d771b467cf0c..e12ea90c6237 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -123,6 +123,7 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper,
 
 	fbi->screen_base = buffer->kvaddr + offset;
 	fbi->screen_size = size;
+	fbi->fix.smem_len = size;
 
 	return 0;
 }
@@ -225,7 +226,7 @@ out:
 	return ret;
 }
 
-static struct drm_fb_helper_funcs exynos_drm_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs exynos_drm_fb_helper_funcs = {
 	.fb_probe =	exynos_drm_fbdev_create,
 };
 
@@ -266,7 +267,8 @@ int exynos_drm_fbdev_init(struct drm_device *dev)
 		return -ENOMEM;
 
 	private->fb_helper = helper = &fbdev->drm_fb_helper;
-	helper->funcs = &exynos_drm_fb_helper_funcs;
+
+	drm_fb_helper_prepare(dev, helper, &exynos_drm_fb_helper_funcs);
 
 	num_crtc = dev->mode_config.num_crtc;
 
@@ -352,9 +354,6 @@ void exynos_drm_fbdev_fini(struct drm_device *dev)
 
 	fbdev = to_exynos_fbdev(private->fb_helper);
 
-	if (fbdev->exynos_gem_obj)
-		exynos_drm_gem_destroy(fbdev->exynos_gem_obj);
-
 	exynos_drm_fbdev_destroy(dev, private->fb_helper);
 	kfree(fbdev);
 	private->fb_helper = NULL;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index 831dde9034c6..68d38eb6774d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -336,9 +336,6 @@ static bool fimc_check_ovf(struct fimc_context *ctx)
 		fimc_set_bits(ctx, EXYNOS_CIWDOFST,
 			EXYNOS_CIWDOFST_CLROVFIY | EXYNOS_CIWDOFST_CLROVFICB |
 			EXYNOS_CIWDOFST_CLROVFICR);
-		fimc_clear_bits(ctx, EXYNOS_CIWDOFST,
-			EXYNOS_CIWDOFST_CLROVFIY | EXYNOS_CIWDOFST_CLROVFICB |
-			EXYNOS_CIWDOFST_CLROVFICR);
 
 		dev_err(ippdrv->dev, "occurred overflow at %d, status 0x%x.\n",
 			ctx->id, status);
@@ -718,24 +715,24 @@ static int fimc_src_set_addr(struct device *dev,
 	case IPP_BUF_ENQUEUE:
 		config = &property->config[EXYNOS_DRM_OPS_SRC];
 		fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_Y],
-			EXYNOS_CIIYSA(buf_id));
+			EXYNOS_CIIYSA0);
 
 		if (config->fmt == DRM_FORMAT_YVU420) {
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CR],
-				EXYNOS_CIICBSA(buf_id));
+				EXYNOS_CIICBSA0);
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CB],
-				EXYNOS_CIICRSA(buf_id));
+				EXYNOS_CIICRSA0);
 		} else {
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CB],
-				EXYNOS_CIICBSA(buf_id));
+				EXYNOS_CIICBSA0);
 			fimc_write(ctx, buf_info->base[EXYNOS_DRM_PLANAR_CR],
-				EXYNOS_CIICRSA(buf_id));
+				EXYNOS_CIICRSA0);
 		}
 		break;
 	case IPP_BUF_DEQUEUE:
-		fimc_write(ctx, 0x0, EXYNOS_CIIYSA(buf_id));
-		fimc_write(ctx, 0x0, EXYNOS_CIICBSA(buf_id));
-		fimc_write(ctx, 0x0, EXYNOS_CIICRSA(buf_id));
+		fimc_write(ctx, 0x0, EXYNOS_CIIYSA0);
+		fimc_write(ctx, 0x0, EXYNOS_CIICBSA0);
+		fimc_write(ctx, 0x0, EXYNOS_CIICRSA0);
 		break;
 	default:
 		/* bypass */
@@ -1122,67 +1119,34 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 	return 0;
 }
 
-static int fimc_dst_get_buf_count(struct fimc_context *ctx)
-{
-	u32 cfg, buf_num;
-
-	cfg = fimc_read(ctx, EXYNOS_CIFCNTSEQ);
-
-	buf_num = hweight32(cfg);
-
-	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
-
-	return buf_num;
-}
-
-static int fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
+static void fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
 		enum drm_exynos_ipp_buf_type buf_type)
 {
-	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
-	bool enable;
-	u32 cfg;
-	u32 mask = 0x00000001 << buf_id;
-	int ret = 0;
 	unsigned long flags;
+	u32 buf_num;
+	u32 cfg;
 
 	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	spin_lock_irqsave(&ctx->lock, flags);
 
-	/* mask register set */
 	cfg = fimc_read(ctx, EXYNOS_CIFCNTSEQ);
 
-	switch (buf_type) {
-	case IPP_BUF_ENQUEUE:
-		enable = true;
-		break;
-	case IPP_BUF_DEQUEUE:
-		enable = false;
-		break;
-	default:
-		dev_err(ippdrv->dev, "invalid buf ctrl parameter.\n");
-		ret =  -EINVAL;
-		goto err_unlock;
-	}
+	if (buf_type == IPP_BUF_ENQUEUE)
+		cfg |= (1 << buf_id);
+	else
+		cfg &= ~(1 << buf_id);
 
-	/* sequence id */
-	cfg &= ~mask;
-	cfg |= (enable << buf_id);
 	fimc_write(ctx, cfg, EXYNOS_CIFCNTSEQ);
 
-	/* interrupt enable */
-	if (buf_type == IPP_BUF_ENQUEUE &&
-	    fimc_dst_get_buf_count(ctx) >= FIMC_BUF_START)
-		fimc_mask_irq(ctx, true);
+	buf_num = hweight32(cfg);
 
-	/* interrupt disable */
-	if (buf_type == IPP_BUF_DEQUEUE &&
-	    fimc_dst_get_buf_count(ctx) <= FIMC_BUF_STOP)
+	if (buf_type == IPP_BUF_ENQUEUE && buf_num >= FIMC_BUF_START)
+		fimc_mask_irq(ctx, true);
+	else if (buf_type == IPP_BUF_DEQUEUE && buf_num <= FIMC_BUF_STOP)
 		fimc_mask_irq(ctx, false);
 
-err_unlock:
 	spin_unlock_irqrestore(&ctx->lock, flags);
-	return ret;
 }
 
 static int fimc_dst_set_addr(struct device *dev,
@@ -1240,7 +1204,9 @@ static int fimc_dst_set_addr(struct device *dev,
 		break;
 	}
 
-	return fimc_dst_set_buf_seq(ctx, buf_id, buf_type);
+	fimc_dst_set_buf_seq(ctx, buf_id, buf_type);
+
+	return 0;
 }
 
 static struct exynos_drm_ipp_ops fimc_dst_ops = {
@@ -1291,14 +1257,11 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 
 	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
-	if (fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE) < 0) {
-		DRM_ERROR("failed to dequeue.\n");
-		return IRQ_HANDLED;
-	}
+	fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE);
 
 	event_work->ippdrv = ippdrv;
 	event_work->buf_id[EXYNOS_DRM_OPS_DST] = buf_id;
-	queue_work(ippdrv->event_workq, (struct work_struct *)event_work);
+	queue_work(ippdrv->event_workq, &event_work->work);
 
 	return IRQ_HANDLED;
 }
@@ -1590,11 +1553,8 @@ static int fimc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 
 	fimc_clear_bits(ctx, EXYNOS_CIOCTRL, EXYNOS_CIOCTRL_WEAVE_MASK);
 
-	if (cmd == IPP_CMD_M2M) {
-		fimc_set_bits(ctx, EXYNOS_MSCTRL, EXYNOS_MSCTRL_ENVID);
-
+	if (cmd == IPP_CMD_M2M)
 		fimc_set_bits(ctx, EXYNOS_MSCTRL, EXYNOS_MSCTRL_ENVID);
-	}
 
 	return 0;
 }
@@ -1887,6 +1847,7 @@ static const struct of_device_id fimc_of_match[] = {
 	{ .compatible = "samsung,exynos4212-fimc" },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, fimc_of_match);
 
 struct platform_driver fimc_driver = {
 	.probe		= fimc_probe,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 33161ad38201..085b066a9993 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -20,6 +20,8 @@
 #include <linux/of_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/component.h>
+#include <linux/mfd/syscon.h>
+#include <linux/regmap.h>
 
 #include <video/of_display_timing.h>
 #include <video/of_videomode.h>
@@ -61,6 +63,24 @@
 /* color key value register for hardware window 1 ~ 4. */
 #define WKEYCON1_BASE(x)		((WKEYCON1 + 0x140) + ((x - 1) * 8))
 
+/* I80 / RGB trigger control register */
+#define TRIGCON				0x1A4
+#define TRGMODE_I80_RGB_ENABLE_I80	(1 << 0)
+#define SWTRGCMD_I80_RGB_ENABLE		(1 << 1)
+
+/* display mode change control register except exynos4 */
+#define VIDOUT_CON			0x000
+#define VIDOUT_CON_F_I80_LDI0		(0x2 << 8)
+
+/* I80 interface control for main LDI register */
+#define I80IFCONFAx(x)			(0x1B0 + (x) * 4)
+#define I80IFCONFBx(x)			(0x1B8 + (x) * 4)
+#define LCD_CS_SETUP(x)			((x) << 16)
+#define LCD_WR_SETUP(x)			((x) << 12)
+#define LCD_WR_ACTIVE(x)		((x) << 8)
+#define LCD_WR_HOLD(x)			((x) << 4)
+#define I80IFEN_ENABLE			(1 << 0)
+
 /* FIMD has totally five hardware windows. */
 #define WINDOWS_NR	5
 
@@ -68,10 +88,14 @@
 
 struct fimd_driver_data {
 	unsigned int timing_base;
+	unsigned int lcdblk_offset;
+	unsigned int lcdblk_vt_shift;
+	unsigned int lcdblk_bypass_shift;
 
 	unsigned int has_shadowcon:1;
 	unsigned int has_clksel:1;
 	unsigned int has_limited_fmt:1;
+	unsigned int has_vidoutcon:1;
 };
 
 static struct fimd_driver_data s3c64xx_fimd_driver_data = {
@@ -80,14 +104,29 @@ static struct fimd_driver_data s3c64xx_fimd_driver_data = {
 	.has_limited_fmt = 1,
 };
 
+static struct fimd_driver_data exynos3_fimd_driver_data = {
+	.timing_base = 0x20000,
+	.lcdblk_offset = 0x210,
+	.lcdblk_bypass_shift = 1,
+	.has_shadowcon = 1,
+	.has_vidoutcon = 1,
+};
+
 static struct fimd_driver_data exynos4_fimd_driver_data = {
 	.timing_base = 0x0,
+	.lcdblk_offset = 0x210,
+	.lcdblk_vt_shift = 10,
+	.lcdblk_bypass_shift = 1,
 	.has_shadowcon = 1,
 };
 
 static struct fimd_driver_data exynos5_fimd_driver_data = {
 	.timing_base = 0x20000,
+	.lcdblk_offset = 0x214,
+	.lcdblk_vt_shift = 24,
+	.lcdblk_bypass_shift = 15,
 	.has_shadowcon = 1,
+	.has_vidoutcon = 1,
 };
 
 struct fimd_win_data {
@@ -112,15 +151,22 @@ struct fimd_context {
 	struct clk			*bus_clk;
 	struct clk			*lcd_clk;
 	void __iomem			*regs;
+	struct regmap			*sysreg;
 	struct drm_display_mode		mode;
 	struct fimd_win_data		win_data[WINDOWS_NR];
 	unsigned int			default_win;
 	unsigned long			irq_flags;
+	u32				vidcon0;
 	u32				vidcon1;
+	u32				vidout_con;
+	u32				i80ifcon;
+	bool				i80_if;
 	bool				suspended;
 	int				pipe;
 	wait_queue_head_t		wait_vsync_queue;
 	atomic_t			wait_vsync_event;
+	atomic_t			win_updated;
+	atomic_t			triggering;
 
 	struct exynos_drm_panel_info panel;
 	struct fimd_driver_data *driver_data;
@@ -130,12 +176,15 @@ struct fimd_context {
 static const struct of_device_id fimd_driver_dt_match[] = {
 	{ .compatible = "samsung,s3c6400-fimd",
 	  .data = &s3c64xx_fimd_driver_data },
+	{ .compatible = "samsung,exynos3250-fimd",
+	  .data = &exynos3_fimd_driver_data },
 	{ .compatible = "samsung,exynos4210-fimd",
 	  .data = &exynos4_fimd_driver_data },
 	{ .compatible = "samsung,exynos5250-fimd",
 	  .data = &exynos5_fimd_driver_data },
 	{},
 };
+MODULE_DEVICE_TABLE(of, fimd_driver_dt_match);
 
 static inline struct fimd_driver_data *drm_fimd_get_driver_data(
 	struct platform_device *pdev)
@@ -165,7 +214,6 @@ static void fimd_wait_for_vblank(struct exynos_drm_manager *mgr)
 		DRM_DEBUG_KMS("vblank wait timed out.\n");
 }
 
-
 static void fimd_clear_channel(struct exynos_drm_manager *mgr)
 {
 	struct fimd_context *ctx = mgr->ctx;
@@ -175,17 +223,31 @@ static void fimd_clear_channel(struct exynos_drm_manager *mgr)
 
 	/* Check if any channel is enabled. */
 	for (win = 0; win < WINDOWS_NR; win++) {
-		u32 val = readl(ctx->regs + SHADOWCON);
-		if (val & SHADOWCON_CHx_ENABLE(win)) {
-			val &= ~SHADOWCON_CHx_ENABLE(win);
-			writel(val, ctx->regs + SHADOWCON);
+		u32 val = readl(ctx->regs + WINCON(win));
+
+		if (val & WINCONx_ENWIN) {
+			/* wincon */
+			val &= ~WINCONx_ENWIN;
+			writel(val, ctx->regs + WINCON(win));
+
+			/* unprotect windows */
+			if (ctx->driver_data->has_shadowcon) {
+				val = readl(ctx->regs + SHADOWCON);
+				val &= ~SHADOWCON_CHx_ENABLE(win);
+				writel(val, ctx->regs + SHADOWCON);
+			}
 			ch_enabled = 1;
 		}
 	}
 
 	/* Wait for vsync, as disable channel takes effect at next vsync */
-	if (ch_enabled)
+	if (ch_enabled) {
+		unsigned int state = ctx->suspended;
+
+		ctx->suspended = 0;
 		fimd_wait_for_vblank(mgr);
+		ctx->suspended = state;
+	}
 }
 
 static int fimd_mgr_initialize(struct exynos_drm_manager *mgr,
@@ -198,23 +260,6 @@ static int fimd_mgr_initialize(struct exynos_drm_manager *mgr,
 	mgr->drm_dev = ctx->drm_dev = drm_dev;
 	mgr->pipe = ctx->pipe = priv->pipe++;
 
-	/*
-	 * enable drm irq mode.
-	 * - with irq_enabled = true, we can use the vblank feature.
-	 *
-	 * P.S. note that we wouldn't use drm irq handler but
-	 *	just specific driver own one instead because
-	 *	drm framework supports only one irq handler.
-	 */
-	drm_dev->irq_enabled = true;
-
-	/*
-	 * with vblank_disable_allowed = true, vblank interrupt will be disabled
-	 * by drm timer once a current process gives up ownership of
-	 * vblank event.(after drm_vblank_put function is called)
-	 */
-	drm_dev->vblank_disable_allowed = true;
-
 	/* attach this sub driver to iommu mapping if supported. */
 	if (is_drm_iommu_supported(ctx->drm_dev)) {
 		/*
@@ -243,6 +288,14 @@ static u32 fimd_calc_clkdiv(struct fimd_context *ctx,
 	unsigned long ideal_clk = mode->htotal * mode->vtotal * mode->vrefresh;
 	u32 clkdiv;
 
+	if (ctx->i80_if) {
+		/*
+		 * The frame done interrupt should be occurred prior to the
+		 * next TE signal.
+		 */
+		ideal_clk *= 2;
+	}
+
 	/* Find the clock divider value that gets us closest to ideal_clk */
 	clkdiv = DIV_ROUND_UP(clk_get_rate(ctx->lcd_clk), ideal_clk);
 
@@ -271,11 +324,10 @@ static void fimd_commit(struct exynos_drm_manager *mgr)
 {
 	struct fimd_context *ctx = mgr->ctx;
 	struct drm_display_mode *mode = &ctx->mode;
-	struct fimd_driver_data *driver_data;
-	u32 val, clkdiv, vidcon1;
-	int vsync_len, vbpd, vfpd, hsync_len, hbpd, hfpd;
+	struct fimd_driver_data *driver_data = ctx->driver_data;
+	void *timing_base = ctx->regs + driver_data->timing_base;
+	u32 val, clkdiv;
 
-	driver_data = ctx->driver_data;
 	if (ctx->suspended)
 		return;
 
@@ -283,33 +335,65 @@ static void fimd_commit(struct exynos_drm_manager *mgr)
 	if (mode->htotal == 0 || mode->vtotal == 0)
 		return;
 
-	/* setup polarity values */
-	vidcon1 = ctx->vidcon1;
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		vidcon1 |= VIDCON1_INV_VSYNC;
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		vidcon1 |= VIDCON1_INV_HSYNC;
-	writel(vidcon1, ctx->regs + driver_data->timing_base + VIDCON1);
-
-	/* setup vertical timing values. */
-	vsync_len = mode->crtc_vsync_end - mode->crtc_vsync_start;
-	vbpd = mode->crtc_vtotal - mode->crtc_vsync_end;
-	vfpd = mode->crtc_vsync_start - mode->crtc_vdisplay;
-
-	val = VIDTCON0_VBPD(vbpd - 1) |
-		VIDTCON0_VFPD(vfpd - 1) |
-		VIDTCON0_VSPW(vsync_len - 1);
-	writel(val, ctx->regs + driver_data->timing_base + VIDTCON0);
-
-	/* setup horizontal timing values.  */
-	hsync_len = mode->crtc_hsync_end - mode->crtc_hsync_start;
-	hbpd = mode->crtc_htotal - mode->crtc_hsync_end;
-	hfpd = mode->crtc_hsync_start - mode->crtc_hdisplay;
-
-	val = VIDTCON1_HBPD(hbpd - 1) |
-		VIDTCON1_HFPD(hfpd - 1) |
-		VIDTCON1_HSPW(hsync_len - 1);
-	writel(val, ctx->regs + driver_data->timing_base + VIDTCON1);
+	if (ctx->i80_if) {
+		val = ctx->i80ifcon | I80IFEN_ENABLE;
+		writel(val, timing_base + I80IFCONFAx(0));
+
+		/* disable auto frame rate */
+		writel(0, timing_base + I80IFCONFBx(0));
+
+		/* set video type selection to I80 interface */
+		if (ctx->sysreg && regmap_update_bits(ctx->sysreg,
+					driver_data->lcdblk_offset,
+					0x3 << driver_data->lcdblk_vt_shift,
+					0x1 << driver_data->lcdblk_vt_shift)) {
+			DRM_ERROR("Failed to update sysreg for I80 i/f.\n");
+			return;
+		}
+	} else {
+		int vsync_len, vbpd, vfpd, hsync_len, hbpd, hfpd;
+		u32 vidcon1;
+
+		/* setup polarity values */
+		vidcon1 = ctx->vidcon1;
+		if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+			vidcon1 |= VIDCON1_INV_VSYNC;
+		if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+			vidcon1 |= VIDCON1_INV_HSYNC;
+		writel(vidcon1, ctx->regs + driver_data->timing_base + VIDCON1);
+
+		/* setup vertical timing values. */
+		vsync_len = mode->crtc_vsync_end - mode->crtc_vsync_start;
+		vbpd = mode->crtc_vtotal - mode->crtc_vsync_end;
+		vfpd = mode->crtc_vsync_start - mode->crtc_vdisplay;
+
+		val = VIDTCON0_VBPD(vbpd - 1) |
+			VIDTCON0_VFPD(vfpd - 1) |
+			VIDTCON0_VSPW(vsync_len - 1);
+		writel(val, ctx->regs + driver_data->timing_base + VIDTCON0);
+
+		/* setup horizontal timing values.  */
+		hsync_len = mode->crtc_hsync_end - mode->crtc_hsync_start;
+		hbpd = mode->crtc_htotal - mode->crtc_hsync_end;
+		hfpd = mode->crtc_hsync_start - mode->crtc_hdisplay;
+
+		val = VIDTCON1_HBPD(hbpd - 1) |
+			VIDTCON1_HFPD(hfpd - 1) |
+			VIDTCON1_HSPW(hsync_len - 1);
+		writel(val, ctx->regs + driver_data->timing_base + VIDTCON1);
+	}
+
+	if (driver_data->has_vidoutcon)
+		writel(ctx->vidout_con, timing_base + VIDOUT_CON);
+
+	/* set bypass selection */
+	if (ctx->sysreg && regmap_update_bits(ctx->sysreg,
+				driver_data->lcdblk_offset,
+				0x1 << driver_data->lcdblk_bypass_shift,
+				0x1 << driver_data->lcdblk_bypass_shift)) {
+		DRM_ERROR("Failed to update sysreg for bypass setting.\n");
+		return;
+	}
 
 	/* setup horizontal and vertical display size. */
 	val = VIDTCON2_LINEVAL(mode->vdisplay - 1) |
@@ -322,7 +406,8 @@ static void fimd_commit(struct exynos_drm_manager *mgr)
 	 * fields of register with prefix '_F' would be updated
 	 * at vsync(same as dma start)
 	 */
-	val = VIDCON0_ENVID | VIDCON0_ENVID_F;
+	val = ctx->vidcon0;
+	val |= VIDCON0_ENVID | VIDCON0_ENVID_F;
 
 	if (ctx->driver_data->has_clksel)
 		val |= VIDCON0_CLKSEL_LCD;
@@ -660,6 +745,9 @@ static void fimd_win_commit(struct exynos_drm_manager *mgr, int zpos)
 	}
 
 	win_data->enabled = true;
+
+	if (ctx->i80_if)
+		atomic_set(&ctx->win_updated, 1);
 }
 
 static void fimd_win_disable(struct exynos_drm_manager *mgr, int zpos)
@@ -838,6 +926,58 @@ static void fimd_dpms(struct exynos_drm_manager *mgr, int mode)
 	}
 }
 
+static void fimd_trigger(struct device *dev)
+{
+	struct exynos_drm_manager *mgr = get_fimd_manager(dev);
+	struct fimd_context *ctx = mgr->ctx;
+	struct fimd_driver_data *driver_data = ctx->driver_data;
+	void *timing_base = ctx->regs + driver_data->timing_base;
+	u32 reg;
+
+	atomic_set(&ctx->triggering, 1);
+
+	reg = readl(ctx->regs + VIDINTCON0);
+	reg |= (VIDINTCON0_INT_ENABLE | VIDINTCON0_INT_I80IFDONE |
+						VIDINTCON0_INT_SYSMAINCON);
+	writel(reg, ctx->regs + VIDINTCON0);
+
+	reg = readl(timing_base + TRIGCON);
+	reg |= (TRGMODE_I80_RGB_ENABLE_I80 | SWTRGCMD_I80_RGB_ENABLE);
+	writel(reg, timing_base + TRIGCON);
+}
+
+static void fimd_te_handler(struct exynos_drm_manager *mgr)
+{
+	struct fimd_context *ctx = mgr->ctx;
+
+	/* Checks the crtc is detached already from encoder */
+	if (ctx->pipe < 0 || !ctx->drm_dev)
+		return;
+
+	 /*
+	 * Skips to trigger if in triggering state, because multiple triggering
+	 * requests can cause panel reset.
+	 */
+	if (atomic_read(&ctx->triggering))
+		return;
+
+	/*
+	 * If there is a page flip request, triggers and handles the page flip
+	 * event so that current fb can be updated into panel GRAM.
+	 */
+	if (atomic_add_unless(&ctx->win_updated, -1, 0))
+		fimd_trigger(ctx->dev);
+
+	/* Wakes up vsync event queue */
+	if (atomic_read(&ctx->wait_vsync_event)) {
+		atomic_set(&ctx->wait_vsync_event, 0);
+		wake_up(&ctx->wait_vsync_queue);
+
+		if (!atomic_read(&ctx->triggering))
+			drm_handle_vblank(ctx->drm_dev, ctx->pipe);
+	}
+}
+
 static struct exynos_drm_manager_ops fimd_manager_ops = {
 	.dpms = fimd_dpms,
 	.mode_fixup = fimd_mode_fixup,
@@ -849,6 +989,7 @@ static struct exynos_drm_manager_ops fimd_manager_ops = {
 	.win_mode_set = fimd_win_mode_set,
 	.win_commit = fimd_win_commit,
 	.win_disable = fimd_win_disable,
+	.te_handler = fimd_te_handler,
 };
 
 static struct exynos_drm_manager fimd_manager = {
@@ -859,26 +1000,40 @@ static struct exynos_drm_manager fimd_manager = {
 static irqreturn_t fimd_irq_handler(int irq, void *dev_id)
 {
 	struct fimd_context *ctx = (struct fimd_context *)dev_id;
-	u32 val;
+	u32 val, clear_bit;
 
 	val = readl(ctx->regs + VIDINTCON1);
 
-	if (val & VIDINTCON1_INT_FRAME)
-		/* VSYNC interrupt */
-		writel(VIDINTCON1_INT_FRAME, ctx->regs + VIDINTCON1);
+	clear_bit = ctx->i80_if ? VIDINTCON1_INT_I80 : VIDINTCON1_INT_FRAME;
+	if (val & clear_bit)
+		writel(clear_bit, ctx->regs + VIDINTCON1);
 
 	/* check the crtc is detached already from encoder */
 	if (ctx->pipe < 0 || !ctx->drm_dev)
 		goto out;
 
-	drm_handle_vblank(ctx->drm_dev, ctx->pipe);
-	exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
+	if (ctx->i80_if) {
+		/* unset I80 frame done interrupt */
+		val = readl(ctx->regs + VIDINTCON0);
+		val &= ~(VIDINTCON0_INT_I80IFDONE | VIDINTCON0_INT_SYSMAINCON);
+		writel(val, ctx->regs + VIDINTCON0);
+
+		/* exit triggering mode */
+		atomic_set(&ctx->triggering, 0);
 
-	/* set wait vsync event to zero and wake up queue. */
-	if (atomic_read(&ctx->wait_vsync_event)) {
-		atomic_set(&ctx->wait_vsync_event, 0);
-		wake_up(&ctx->wait_vsync_queue);
+		drm_handle_vblank(ctx->drm_dev, ctx->pipe);
+		exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
+	} else {
+		drm_handle_vblank(ctx->drm_dev, ctx->pipe);
+		exynos_drm_crtc_finish_pageflip(ctx->drm_dev, ctx->pipe);
+
+		/* set wait vsync event to zero and wake up queue. */
+		if (atomic_read(&ctx->wait_vsync_event)) {
+			atomic_set(&ctx->wait_vsync_event, 0);
+			wake_up(&ctx->wait_vsync_queue);
+		}
 	}
+
 out:
 	return IRQ_HANDLED;
 }
@@ -902,7 +1057,6 @@ static void fimd_unbind(struct device *dev, struct device *master,
 {
 	struct exynos_drm_manager *mgr = dev_get_drvdata(dev);
 	struct fimd_context *ctx = fimd_manager.ctx;
-	struct drm_crtc *crtc = mgr->crtc;
 
 	fimd_dpms(mgr, DRM_MODE_DPMS_OFF);
 
@@ -910,8 +1064,6 @@ static void fimd_unbind(struct device *dev, struct device *master,
 		exynos_dpi_remove(dev);
 
 	fimd_mgr_remove(mgr);
-
-	crtc->funcs->destroy(crtc);
 }
 
 static const struct component_ops fimd_component_ops = {
@@ -923,6 +1075,7 @@ static int fimd_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct fimd_context *ctx;
+	struct device_node *i80_if_timings;
 	struct resource *res;
 	int ret = -EINVAL;
 
@@ -944,12 +1097,51 @@ static int fimd_probe(struct platform_device *pdev)
 
 	ctx->dev = dev;
 	ctx->suspended = true;
+	ctx->driver_data = drm_fimd_get_driver_data(pdev);
 
 	if (of_property_read_bool(dev->of_node, "samsung,invert-vden"))
 		ctx->vidcon1 |= VIDCON1_INV_VDEN;
 	if (of_property_read_bool(dev->of_node, "samsung,invert-vclk"))
 		ctx->vidcon1 |= VIDCON1_INV_VCLK;
 
+	i80_if_timings = of_get_child_by_name(dev->of_node, "i80-if-timings");
+	if (i80_if_timings) {
+		u32 val;
+
+		ctx->i80_if = true;
+
+		if (ctx->driver_data->has_vidoutcon)
+			ctx->vidout_con |= VIDOUT_CON_F_I80_LDI0;
+		else
+			ctx->vidcon0 |= VIDCON0_VIDOUT_I80_LDI0;
+		/*
+		 * The user manual describes that this "DSI_EN" bit is required
+		 * to enable I80 24-bit data interface.
+		 */
+		ctx->vidcon0 |= VIDCON0_DSI_EN;
+
+		if (of_property_read_u32(i80_if_timings, "cs-setup", &val))
+			val = 0;
+		ctx->i80ifcon = LCD_CS_SETUP(val);
+		if (of_property_read_u32(i80_if_timings, "wr-setup", &val))
+			val = 0;
+		ctx->i80ifcon |= LCD_WR_SETUP(val);
+		if (of_property_read_u32(i80_if_timings, "wr-active", &val))
+			val = 1;
+		ctx->i80ifcon |= LCD_WR_ACTIVE(val);
+		if (of_property_read_u32(i80_if_timings, "wr-hold", &val))
+			val = 0;
+		ctx->i80ifcon |= LCD_WR_HOLD(val);
+	}
+	of_node_put(i80_if_timings);
+
+	ctx->sysreg = syscon_regmap_lookup_by_phandle(dev->of_node,
+							"samsung,sysreg");
+	if (IS_ERR(ctx->sysreg)) {
+		dev_warn(dev, "failed to get system register.\n");
+		ctx->sysreg = NULL;
+	}
+
 	ctx->bus_clk = devm_clk_get(dev, "fimd");
 	if (IS_ERR(ctx->bus_clk)) {
 		dev_err(dev, "failed to get bus clock\n");
@@ -972,7 +1164,8 @@ static int fimd_probe(struct platform_device *pdev)
 		goto err_del_component;
 	}
 
-	res = platform_get_resource_byname(pdev, IORESOURCE_IRQ, "vsync");
+	res = platform_get_resource_byname(pdev, IORESOURCE_IRQ,
+					   ctx->i80_if ? "lcd_sys" : "vsync");
 	if (!res) {
 		dev_err(dev, "irq request failed.\n");
 		ret = -ENXIO;
@@ -986,7 +1179,6 @@ static int fimd_probe(struct platform_device *pdev)
 		goto err_del_component;
 	}
 
-	ctx->driver_data = drm_fimd_get_driver_data(pdev);
 	init_waitqueue_head(&ctx->wait_vsync_queue);
 	atomic_set(&ctx->wait_vsync_event, 0);
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index 800158714473..6ff8599f6cbf 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -302,9 +302,12 @@ static void g2d_fini_cmdlist(struct g2d_data *g2d)
 	struct exynos_drm_subdrv *subdrv = &g2d->subdrv;
 
 	kfree(g2d->cmdlist_node);
-	dma_free_attrs(subdrv->drm_dev->dev, G2D_CMDLIST_POOL_SIZE,
-			g2d->cmdlist_pool_virt,
-			g2d->cmdlist_pool, &g2d->cmdlist_dma_attrs);
+
+	if (g2d->cmdlist_pool_virt && g2d->cmdlist_pool) {
+		dma_free_attrs(subdrv->drm_dev->dev, G2D_CMDLIST_POOL_SIZE,
+				g2d->cmdlist_pool_virt,
+				g2d->cmdlist_pool, &g2d->cmdlist_dma_attrs);
+	}
 }
 
 static struct g2d_cmdlist_node *g2d_get_cmdlist(struct g2d_data *g2d)
@@ -1042,8 +1045,23 @@ err:
 int exynos_g2d_get_ver_ioctl(struct drm_device *drm_dev, void *data,
 			     struct drm_file *file)
 {
+	struct drm_exynos_file_private *file_priv = file->driver_priv;
+	struct exynos_drm_g2d_private *g2d_priv = file_priv->g2d_priv;
+	struct device *dev;
+	struct g2d_data *g2d;
 	struct drm_exynos_g2d_get_ver *ver = data;
 
+	if (!g2d_priv)
+		return -ENODEV;
+
+	dev = g2d_priv->dev;
+	if (!dev)
+		return -ENODEV;
+
+	g2d = dev_get_drvdata(dev);
+	if (!g2d)
+		return -EFAULT;
+
 	ver->major = G2D_HW_MAJOR_VER;
 	ver->minor = G2D_HW_MINOR_VER;
 
@@ -1056,7 +1074,7 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data,
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
 	struct exynos_drm_g2d_private *g2d_priv = file_priv->g2d_priv;
-	struct device *dev = g2d_priv->dev;
+	struct device *dev;
 	struct g2d_data *g2d;
 	struct drm_exynos_g2d_set_cmdlist *req = data;
 	struct drm_exynos_g2d_cmd *cmd;
@@ -1067,6 +1085,10 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data,
 	int size;
 	int ret;
 
+	if (!g2d_priv)
+		return -ENODEV;
+
+	dev = g2d_priv->dev;
 	if (!dev)
 		return -ENODEV;
 
@@ -1223,13 +1245,17 @@ int exynos_g2d_exec_ioctl(struct drm_device *drm_dev, void *data,
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
 	struct exynos_drm_g2d_private *g2d_priv = file_priv->g2d_priv;
-	struct device *dev = g2d_priv->dev;
+	struct device *dev;
 	struct g2d_data *g2d;
 	struct drm_exynos_g2d_exec *req = data;
 	struct g2d_runqueue_node *runqueue_node;
 	struct list_head *run_cmdlist;
 	struct list_head *event_list;
 
+	if (!g2d_priv)
+		return -ENODEV;
+
+	dev = g2d_priv->dev;
 	if (!dev)
 		return -ENODEV;
 
@@ -1544,8 +1570,10 @@ static const struct dev_pm_ops g2d_pm_ops = {
 
 static const struct of_device_id exynos_g2d_match[] = {
 	{ .compatible = "samsung,exynos5250-g2d" },
+	{ .compatible = "samsung,exynos4212-g2d" },
 	{},
 };
+MODULE_DEVICE_TABLE(of, exynos_g2d_match);
 
 struct platform_driver g2d_driver = {
 	.probe		= g2d_probe,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 163a054922cb..0d5b9698d384 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -301,7 +301,6 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 					unsigned int gem_handle,
 					struct drm_file *filp)
 {
-	struct exynos_drm_gem_obj *exynos_gem_obj;
 	struct drm_gem_object *obj;
 
 	obj = drm_gem_object_lookup(dev, filp, gem_handle);
@@ -310,8 +309,6 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 		return;
 	}
 
-	exynos_gem_obj = to_exynos_gem_obj(obj);
-
 	drm_gem_object_unreference_unlocked(obj);
 
 	/*
@@ -321,40 +318,16 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 	drm_gem_object_unreference_unlocked(obj);
 }
 
-int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
-				    struct drm_file *file_priv)
-{
-	struct drm_exynos_gem_map_off *args = data;
-
-	DRM_DEBUG_KMS("handle = 0x%x, offset = 0x%lx\n",
-			args->handle, (unsigned long)args->offset);
-
-	if (!(dev->driver->driver_features & DRIVER_GEM)) {
-		DRM_ERROR("does not support GEM.\n");
-		return -ENODEV;
-	}
-
-	return exynos_drm_gem_dumb_map_offset(file_priv, dev, args->handle,
-			&args->offset);
-}
-
-int exynos_drm_gem_mmap_buffer(struct file *filp,
+int exynos_drm_gem_mmap_buffer(struct exynos_drm_gem_obj *exynos_gem_obj,
 				      struct vm_area_struct *vma)
 {
-	struct drm_gem_object *obj = filp->private_data;
-	struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj);
-	struct drm_device *drm_dev = obj->dev;
+	struct drm_device *drm_dev = exynos_gem_obj->base.dev;
 	struct exynos_drm_gem_buf *buffer;
 	unsigned long vm_size;
 	int ret;
 
-	WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex));
-
-	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_private_data = obj;
-	vma->vm_ops = drm_dev->driver->gem_vm_ops;
-
-	update_vm_cache_attr(exynos_gem_obj, vma);
+	vma->vm_flags &= ~VM_PFNMAP;
+	vma->vm_pgoff = 0;
 
 	vm_size = vma->vm_end - vma->vm_start;
 
@@ -376,60 +349,6 @@ int exynos_drm_gem_mmap_buffer(struct file *filp,
 		return ret;
 	}
 
-	/*
-	 * take a reference to this mapping of the object. And this reference
-	 * is unreferenced by the corresponding vm_close call.
-	 */
-	drm_gem_object_reference(obj);
-
-	drm_vm_open_locked(drm_dev, vma);
-
-	return 0;
-}
-
-int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv)
-{
-	struct drm_exynos_file_private *exynos_file_priv;
-	struct drm_exynos_gem_mmap *args = data;
-	struct drm_gem_object *obj;
-	struct file *anon_filp;
-	unsigned long addr;
-
-	if (!(dev->driver->driver_features & DRIVER_GEM)) {
-		DRM_ERROR("does not support GEM.\n");
-		return -ENODEV;
-	}
-
-	mutex_lock(&dev->struct_mutex);
-
-	obj = drm_gem_object_lookup(dev, file_priv, args->handle);
-	if (!obj) {
-		DRM_ERROR("failed to lookup gem object.\n");
-		mutex_unlock(&dev->struct_mutex);
-		return -EINVAL;
-	}
-
-	exynos_file_priv = file_priv->driver_priv;
-	anon_filp = exynos_file_priv->anon_filp;
-	anon_filp->private_data = obj;
-
-	addr = vm_mmap(anon_filp, 0, args->size, PROT_READ | PROT_WRITE,
-			MAP_SHARED, 0);
-
-	drm_gem_object_unreference(obj);
-
-	if (IS_ERR_VALUE(addr)) {
-		mutex_unlock(&dev->struct_mutex);
-		return (int)addr;
-	}
-
-	mutex_unlock(&dev->struct_mutex);
-
-	args->mapped = addr;
-
-	DRM_DEBUG_KMS("mapped = 0x%lx\n", (unsigned long)args->mapped);
-
 	return 0;
 }
 
@@ -713,16 +632,20 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	exynos_gem_obj = to_exynos_gem_obj(obj);
 
 	ret = check_gem_flags(exynos_gem_obj->flags);
-	if (ret) {
-		drm_gem_vm_close(vma);
-		drm_gem_free_mmap_offset(obj);
-		return ret;
-	}
-
-	vma->vm_flags &= ~VM_PFNMAP;
-	vma->vm_flags |= VM_MIXEDMAP;
+	if (ret)
+		goto err_close_vm;
 
 	update_vm_cache_attr(exynos_gem_obj, vma);
 
+	ret = exynos_drm_gem_mmap_buffer(exynos_gem_obj, vma);
+	if (ret)
+		goto err_close_vm;
+
+	return ret;
+
+err_close_vm:
+	drm_gem_vm_close(vma);
+	drm_gem_free_mmap_offset(obj);
+
 	return ret;
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.h b/drivers/gpu/drm/exynos/exynos_drm_gem.h
index 1592c0ba7de8..ec58fe9c40df 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.h
@@ -12,6 +12,8 @@
 #ifndef _EXYNOS_DRM_GEM_H_
 #define _EXYNOS_DRM_GEM_H_
 
+#include <drm/drm_gem.h>
+
 #define to_exynos_gem_obj(x)	container_of(x,\
 			struct exynos_drm_gem_obj, base)
 
@@ -111,20 +113,6 @@ void exynos_drm_gem_put_dma_addr(struct drm_device *dev,
 					unsigned int gem_handle,
 					struct drm_file *filp);
 
-/* get buffer offset to map to user space. */
-int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
-				    struct drm_file *file_priv);
-
-/*
- * mmap the physically continuous memory that a gem object contains
- * to user space.
- */
-int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv);
-
-int exynos_drm_gem_mmap_buffer(struct file *filp,
-				      struct vm_area_struct *vma);
-
 /* map user space allocated by malloc to pages. */
 int exynos_drm_gem_userptr_ioctl(struct drm_device *dev, void *data,
 				      struct drm_file *file_priv);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
index 9e3ff1672965..c6a013fc321c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
@@ -1326,8 +1326,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 			buf_id[EXYNOS_DRM_OPS_SRC];
 		event_work->buf_id[EXYNOS_DRM_OPS_DST] =
 			buf_id[EXYNOS_DRM_OPS_DST];
-		queue_work(ippdrv->event_workq,
-			(struct work_struct *)event_work);
+		queue_work(ippdrv->event_workq, &event_work->work);
 	}
 
 	return IRQ_HANDLED;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
index a1888e128f1d..00d74b18f7cb 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
@@ -75,7 +75,6 @@ struct drm_exynos_ipp_mem_node {
 	u32	prop_id;
 	u32	buf_id;
 	struct drm_exynos_ipp_buf_info	buf_info;
-	struct drm_file		*filp;
 };
 
 /*
@@ -129,9 +128,6 @@ void exynos_platform_device_ipp_unregister(void)
 
 int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 {
-	if (!ippdrv)
-		return -EINVAL;
-
 	mutex_lock(&exynos_drm_ippdrv_lock);
 	list_add_tail(&ippdrv->drv_list, &exynos_drm_ippdrv_list);
 	mutex_unlock(&exynos_drm_ippdrv_lock);
@@ -141,9 +137,6 @@ int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 
 int exynos_drm_ippdrv_unregister(struct exynos_drm_ippdrv *ippdrv)
 {
-	if (!ippdrv)
-		return -EINVAL;
-
 	mutex_lock(&exynos_drm_ippdrv_lock);
 	list_del(&ippdrv->drv_list);
 	mutex_unlock(&exynos_drm_ippdrv_lock);
@@ -151,20 +144,15 @@ int exynos_drm_ippdrv_unregister(struct exynos_drm_ippdrv *ippdrv)
 	return 0;
 }
 
-static int ipp_create_id(struct idr *id_idr, struct mutex *lock, void *obj,
-		u32 *idp)
+static int ipp_create_id(struct idr *id_idr, struct mutex *lock, void *obj)
 {
 	int ret;
 
-	/* do the allocation under our mutexlock */
 	mutex_lock(lock);
 	ret = idr_alloc(id_idr, obj, 1, 0, GFP_KERNEL);
 	mutex_unlock(lock);
-	if (ret < 0)
-		return ret;
 
-	*idp = ret;
-	return 0;
+	return ret;
 }
 
 static void ipp_remove_id(struct idr *id_idr, struct mutex *lock, u32 id)
@@ -178,35 +166,25 @@ static void *ipp_find_obj(struct idr *id_idr, struct mutex *lock, u32 id)
 {
 	void *obj;
 
-	DRM_DEBUG_KMS("id[%d]\n", id);
-
 	mutex_lock(lock);
-
-	/* find object using handle */
 	obj = idr_find(id_idr, id);
-	if (!obj) {
-		DRM_ERROR("failed to find object.\n");
-		mutex_unlock(lock);
-		return ERR_PTR(-ENODEV);
-	}
-
 	mutex_unlock(lock);
 
 	return obj;
 }
 
-static inline bool ipp_check_dedicated(struct exynos_drm_ippdrv *ippdrv,
-		enum drm_exynos_ipp_cmd	cmd)
+static int ipp_check_driver(struct exynos_drm_ippdrv *ippdrv,
+			    struct drm_exynos_ipp_property *property)
 {
-	/*
-	 * check dedicated flag and WB, OUTPUT operation with
-	 * power on state.
-	 */
-	if (ippdrv->dedicated || (!ipp_is_m2m_cmd(cmd) &&
-	    !pm_runtime_suspended(ippdrv->dev)))
-		return true;
+	if (ippdrv->dedicated || (!ipp_is_m2m_cmd(property->cmd) &&
+				  !pm_runtime_suspended(ippdrv->dev)))
+		return -EBUSY;
 
-	return false;
+	if (ippdrv->check_property &&
+	    ippdrv->check_property(ippdrv->dev, property))
+		return -EINVAL;
+
+	return 0;
 }
 
 static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
@@ -214,62 +192,30 @@ static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
 {
 	struct exynos_drm_ippdrv *ippdrv;
 	u32 ipp_id = property->ipp_id;
-
-	DRM_DEBUG_KMS("ipp_id[%d]\n", ipp_id);
+	int ret;
 
 	if (ipp_id) {
-		/* find ipp driver using idr */
-		ippdrv = ipp_find_obj(&ctx->ipp_idr, &ctx->ipp_lock,
-			ipp_id);
-		if (IS_ERR(ippdrv)) {
-			DRM_ERROR("not found ipp%d driver.\n", ipp_id);
-			return ippdrv;
+		ippdrv = ipp_find_obj(&ctx->ipp_idr, &ctx->ipp_lock, ipp_id);
+		if (!ippdrv) {
+			DRM_DEBUG("ipp%d driver not found\n", ipp_id);
+			return ERR_PTR(-ENODEV);
 		}
 
-		/*
-		 * WB, OUTPUT opertion not supported multi-operation.
-		 * so, make dedicated state at set property ioctl.
-		 * when ipp driver finished operations, clear dedicated flags.
-		 */
-		if (ipp_check_dedicated(ippdrv, property->cmd)) {
-			DRM_ERROR("already used choose device.\n");
-			return ERR_PTR(-EBUSY);
-		}
-
-		/*
-		 * This is necessary to find correct device in ipp drivers.
-		 * ipp drivers have different abilities,
-		 * so need to check property.
-		 */
-		if (ippdrv->check_property &&
-		    ippdrv->check_property(ippdrv->dev, property)) {
-			DRM_ERROR("not support property.\n");
-			return ERR_PTR(-EINVAL);
+		ret = ipp_check_driver(ippdrv, property);
+		if (ret < 0) {
+			DRM_DEBUG("ipp%d driver check error %d\n", ipp_id, ret);
+			return ERR_PTR(ret);
 		}
 
 		return ippdrv;
 	} else {
-		/*
-		 * This case is search all ipp driver for finding.
-		 * user application don't set ipp_id in this case,
-		 * so ipp subsystem search correct driver in driver list.
-		 */
 		list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
-			if (ipp_check_dedicated(ippdrv, property->cmd)) {
-				DRM_DEBUG_KMS("used device.\n");
-				continue;
-			}
-
-			if (ippdrv->check_property &&
-			    ippdrv->check_property(ippdrv->dev, property)) {
-				DRM_DEBUG_KMS("not support property.\n");
-				continue;
-			}
-
-			return ippdrv;
+			ret = ipp_check_driver(ippdrv, property);
+			if (ret == 0)
+				return ippdrv;
 		}
 
-		DRM_ERROR("not support ipp driver operations.\n");
+		DRM_DEBUG("cannot find driver suitable for given property.\n");
 	}
 
 	return ERR_PTR(-ENODEV);
@@ -308,8 +254,7 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 		struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv = file_priv->ipp_priv;
-	struct device *dev = priv->dev;
+	struct device *dev = file_priv->ipp_dev;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_prop_list *prop_list = data;
 	struct exynos_drm_ippdrv *ippdrv;
@@ -346,10 +291,10 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 		 */
 		ippdrv = ipp_find_obj(&ctx->ipp_idr, &ctx->ipp_lock,
 						prop_list->ipp_id);
-		if (IS_ERR(ippdrv)) {
+		if (!ippdrv) {
 			DRM_ERROR("not found ipp%d driver.\n",
 					prop_list->ipp_id);
-			return PTR_ERR(ippdrv);
+			return -ENODEV;
 		}
 
 		*prop_list = ippdrv->prop_list;
@@ -373,44 +318,6 @@ static void ipp_print_property(struct drm_exynos_ipp_property *property,
 		sz->hsize, sz->vsize, config->flip, config->degree);
 }
 
-static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
-{
-	struct exynos_drm_ippdrv *ippdrv;
-	struct drm_exynos_ipp_cmd_node *c_node;
-	u32 prop_id = property->prop_id;
-
-	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
-
-	ippdrv = ipp_find_drv_by_handle(prop_id);
-	if (IS_ERR(ippdrv)) {
-		DRM_ERROR("failed to get ipp driver.\n");
-		return -EINVAL;
-	}
-
-	/*
-	 * Find command node using command list in ippdrv.
-	 * when we find this command no using prop_id.
-	 * return property information set in this command node.
-	 */
-	mutex_lock(&ippdrv->cmd_lock);
-	list_for_each_entry(c_node, &ippdrv->cmd_list, list) {
-		if ((c_node->property.prop_id == prop_id) &&
-		    (c_node->state == IPP_STATE_STOP)) {
-			mutex_unlock(&ippdrv->cmd_lock);
-			DRM_DEBUG_KMS("found cmd[%d]ippdrv[0x%x]\n",
-				property->cmd, (int)ippdrv);
-
-			c_node->property = *property;
-			return 0;
-		}
-	}
-	mutex_unlock(&ippdrv->cmd_lock);
-
-	DRM_ERROR("failed to search property.\n");
-
-	return -EINVAL;
-}
-
 static struct drm_exynos_ipp_cmd_work *ipp_create_cmd_work(void)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work;
@@ -432,7 +339,7 @@ static struct drm_exynos_ipp_event_work *ipp_create_event_work(void)
 	if (!event_work)
 		return ERR_PTR(-ENOMEM);
 
-	INIT_WORK((struct work_struct *)event_work, ipp_sched_event);
+	INIT_WORK(&event_work->work, ipp_sched_event);
 
 	return event_work;
 }
@@ -441,12 +348,12 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv = file_priv->ipp_priv;
-	struct device *dev = priv->dev;
+	struct device *dev = file_priv->ipp_dev;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_property *property = data;
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
+	u32 prop_id;
 	int ret, i;
 
 	if (!ctx) {
@@ -459,6 +366,8 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
+	prop_id = property->prop_id;
+
 	/*
 	 * This is log print for user application property.
 	 * user application set various property.
@@ -467,14 +376,24 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		ipp_print_property(property, i);
 
 	/*
-	 * set property ioctl generated new prop_id.
-	 * but in this case already asigned prop_id using old set property.
-	 * e.g PAUSE state. this case supports find current prop_id and use it
-	 * instead of allocation.
+	 * In case prop_id is not zero try to set existing property.
 	 */
-	if (property->prop_id) {
-		DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
-		return ipp_find_and_set_property(property);
+	if (prop_id) {
+		c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock, prop_id);
+
+		if (!c_node || c_node->filp != file) {
+			DRM_DEBUG_KMS("prop_id[%d] not found\n", prop_id);
+			return -EINVAL;
+		}
+
+		if (c_node->state != IPP_STATE_STOP) {
+			DRM_DEBUG_KMS("prop_id[%d] not stopped\n", prop_id);
+			return -EINVAL;
+		}
+
+		c_node->property = *property;
+
+		return 0;
 	}
 
 	/* find ipp driver using ipp id */
@@ -489,21 +408,20 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	if (!c_node)
 		return -ENOMEM;
 
-	/* create property id */
-	ret = ipp_create_id(&ctx->prop_idr, &ctx->prop_lock, c_node,
-		&property->prop_id);
-	if (ret) {
+	ret = ipp_create_id(&ctx->prop_idr, &ctx->prop_lock, c_node);
+	if (ret < 0) {
 		DRM_ERROR("failed to create id.\n");
 		goto err_clear;
 	}
+	property->prop_id = ret;
 
 	DRM_DEBUG_KMS("created prop_id[%d]cmd[%d]ippdrv[0x%x]\n",
 		property->prop_id, property->cmd, (int)ippdrv);
 
 	/* stored property information and ippdrv in private data */
-	c_node->priv = priv;
 	c_node->property = *property;
 	c_node->state = IPP_STATE_IDLE;
+	c_node->filp = file;
 
 	c_node->start_work = ipp_create_cmd_work();
 	if (IS_ERR(c_node->start_work)) {
@@ -534,7 +452,6 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		INIT_LIST_HEAD(&c_node->mem_list[i]);
 
 	INIT_LIST_HEAD(&c_node->event_list);
-	list_splice_init(&priv->event_list, &c_node->event_list);
 	mutex_lock(&ippdrv->cmd_lock);
 	list_add_tail(&c_node->list, &ippdrv->cmd_list);
 	mutex_unlock(&ippdrv->cmd_lock);
@@ -556,148 +473,55 @@ err_clear:
 	return ret;
 }
 
-static void ipp_clean_cmd_node(struct ipp_context *ctx,
-				struct drm_exynos_ipp_cmd_node *c_node)
-{
-	/* delete list */
-	list_del(&c_node->list);
-
-	ipp_remove_id(&ctx->prop_idr, &ctx->prop_lock,
-			c_node->property.prop_id);
-
-	/* destroy mutex */
-	mutex_destroy(&c_node->lock);
-	mutex_destroy(&c_node->mem_lock);
-	mutex_destroy(&c_node->event_lock);
-
-	/* free command node */
-	kfree(c_node->start_work);
-	kfree(c_node->stop_work);
-	kfree(c_node->event_work);
-	kfree(c_node);
-}
-
-static int ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
-{
-	struct drm_exynos_ipp_property *property = &c_node->property;
-	struct drm_exynos_ipp_mem_node *m_node;
-	struct list_head *head;
-	int ret, i, count[EXYNOS_DRM_OPS_MAX] = { 0, };
-
-	for_each_ipp_ops(i) {
-		/* source/destination memory list */
-		head = &c_node->mem_list[i];
-
-		/* find memory node entry */
-		list_for_each_entry(m_node, head, list) {
-			DRM_DEBUG_KMS("%s,count[%d]m_node[0x%x]\n",
-				i ? "dst" : "src", count[i], (int)m_node);
-			count[i]++;
-		}
-	}
-
-	DRM_DEBUG_KMS("min[%d]max[%d]\n",
-		min(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]),
-		max(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]));
-
-	/*
-	 * M2M operations should be need paired memory address.
-	 * so, need to check minimum count about src, dst.
-	 * other case not use paired memory, so use maximum count
-	 */
-	if (ipp_is_m2m_cmd(property->cmd))
-		ret = min(count[EXYNOS_DRM_OPS_SRC],
-			count[EXYNOS_DRM_OPS_DST]);
-	else
-		ret = max(count[EXYNOS_DRM_OPS_SRC],
-			count[EXYNOS_DRM_OPS_DST]);
-
-	return ret;
-}
-
-static struct drm_exynos_ipp_mem_node
-		*ipp_find_mem_node(struct drm_exynos_ipp_cmd_node *c_node,
-		struct drm_exynos_ipp_queue_buf *qbuf)
-{
-	struct drm_exynos_ipp_mem_node *m_node;
-	struct list_head *head;
-	int count = 0;
-
-	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
-
-	/* source/destination memory list */
-	head = &c_node->mem_list[qbuf->ops_id];
-
-	/* find memory node from memory list */
-	list_for_each_entry(m_node, head, list) {
-		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
-
-		/* compare buffer id */
-		if (m_node->buf_id == qbuf->buf_id)
-			return m_node;
-	}
-
-	return NULL;
-}
-
-static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
+static int ipp_put_mem_node(struct drm_device *drm_dev,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_mem_node *m_node)
 {
-	struct exynos_drm_ipp_ops *ops = NULL;
-	int ret = 0;
+	int i;
 
 	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
-		DRM_ERROR("invalid queue node.\n");
+		DRM_ERROR("invalid dequeue node.\n");
 		return -EFAULT;
 	}
 
 	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
-	/* get operations callback */
-	ops = ippdrv->ops[m_node->ops_id];
-	if (!ops) {
-		DRM_ERROR("not support ops.\n");
-		return -EFAULT;
+	/* put gem buffer */
+	for_each_ipp_planar(i) {
+		unsigned long handle = m_node->buf_info.handles[i];
+		if (handle)
+			exynos_drm_gem_put_dma_addr(drm_dev, handle,
+							c_node->filp);
 	}
 
-	/* set address and enable irq */
-	if (ops->set_addr) {
-		ret = ops->set_addr(ippdrv->dev, &m_node->buf_info,
-			m_node->buf_id, IPP_BUF_ENQUEUE);
-		if (ret) {
-			DRM_ERROR("failed to set addr.\n");
-			return ret;
-		}
-	}
+	list_del(&m_node->list);
+	kfree(m_node);
 
-	return ret;
+	return 0;
 }
 
 static struct drm_exynos_ipp_mem_node
 		*ipp_get_mem_node(struct drm_device *drm_dev,
-		struct drm_file *file,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_queue_buf *qbuf)
 {
 	struct drm_exynos_ipp_mem_node *m_node;
-	struct drm_exynos_ipp_buf_info buf_info;
-	void *addr;
+	struct drm_exynos_ipp_buf_info *buf_info;
 	int i;
 
 	m_node = kzalloc(sizeof(*m_node), GFP_KERNEL);
 	if (!m_node)
 		return ERR_PTR(-ENOMEM);
 
-	/* clear base address for error handling */
-	memset(&buf_info, 0x0, sizeof(buf_info));
+	buf_info = &m_node->buf_info;
 
 	/* operations, buffer id */
 	m_node->ops_id = qbuf->ops_id;
 	m_node->prop_id = qbuf->prop_id;
 	m_node->buf_id = qbuf->buf_id;
+	INIT_LIST_HEAD(&m_node->list);
 
 	DRM_DEBUG_KMS("m_node[0x%x]ops_id[%d]\n", (int)m_node, qbuf->ops_id);
 	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]\n", qbuf->prop_id, m_node->buf_id);
@@ -707,61 +531,47 @@ static struct drm_exynos_ipp_mem_node
 
 		/* get dma address by handle */
 		if (qbuf->handle[i]) {
+			dma_addr_t *addr;
+
 			addr = exynos_drm_gem_get_dma_addr(drm_dev,
-					qbuf->handle[i], file);
+					qbuf->handle[i], c_node->filp);
 			if (IS_ERR(addr)) {
 				DRM_ERROR("failed to get addr.\n");
-				goto err_clear;
+				ipp_put_mem_node(drm_dev, c_node, m_node);
+				return ERR_PTR(-EFAULT);
 			}
 
-			buf_info.handles[i] = qbuf->handle[i];
-			buf_info.base[i] = *(dma_addr_t *) addr;
-			DRM_DEBUG_KMS("i[%d]base[0x%x]hd[0x%x]\n",
-				i, buf_info.base[i], (int)buf_info.handles[i]);
+			buf_info->handles[i] = qbuf->handle[i];
+			buf_info->base[i] = *addr;
+			DRM_DEBUG_KMS("i[%d]base[0x%x]hd[0x%lx]\n", i,
+				      buf_info->base[i], buf_info->handles[i]);
 		}
 	}
 
-	m_node->filp = file;
-	m_node->buf_info = buf_info;
 	mutex_lock(&c_node->mem_lock);
 	list_add_tail(&m_node->list, &c_node->mem_list[qbuf->ops_id]);
 	mutex_unlock(&c_node->mem_lock);
 
 	return m_node;
-
-err_clear:
-	kfree(m_node);
-	return ERR_PTR(-EFAULT);
 }
 
-static int ipp_put_mem_node(struct drm_device *drm_dev,
-		struct drm_exynos_ipp_cmd_node *c_node,
-		struct drm_exynos_ipp_mem_node *m_node)
+static void ipp_clean_mem_nodes(struct drm_device *drm_dev,
+			       struct drm_exynos_ipp_cmd_node *c_node, int ops)
 {
-	int i;
-
-	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
+	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
+	struct list_head *head = &c_node->mem_list[ops];
 
-	if (!m_node) {
-		DRM_ERROR("invalid dequeue node.\n");
-		return -EFAULT;
-	}
+	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
+	list_for_each_entry_safe(m_node, tm_node, head, list) {
+		int ret;
 
-	/* put gem buffer */
-	for_each_ipp_planar(i) {
-		unsigned long handle = m_node->buf_info.handles[i];
-		if (handle)
-			exynos_drm_gem_put_dma_addr(drm_dev, handle,
-							m_node->filp);
+		ret = ipp_put_mem_node(drm_dev, c_node, m_node);
+		if (ret)
+			DRM_ERROR("failed to put m_node.\n");
 	}
 
-	/* delete list in queue */
-	list_del(&m_node->list);
-	kfree(m_node);
-
-	return 0;
+	mutex_unlock(&c_node->mem_lock);
 }
 
 static void ipp_free_event(struct drm_pending_event *event)
@@ -770,7 +580,6 @@ static void ipp_free_event(struct drm_pending_event *event)
 }
 
 static int ipp_get_event(struct drm_device *drm_dev,
-		struct drm_file *file,
 		struct drm_exynos_ipp_cmd_node *c_node,
 		struct drm_exynos_ipp_queue_buf *qbuf)
 {
@@ -782,7 +591,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	e = kzalloc(sizeof(*e), GFP_KERNEL);
 	if (!e) {
 		spin_lock_irqsave(&drm_dev->event_lock, flags);
-		file->event_space += sizeof(e->event);
+		c_node->filp->event_space += sizeof(e->event);
 		spin_unlock_irqrestore(&drm_dev->event_lock, flags);
 		return -ENOMEM;
 	}
@@ -794,7 +603,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	e->event.prop_id = qbuf->prop_id;
 	e->event.buf_id[EXYNOS_DRM_OPS_DST] = qbuf->buf_id;
 	e->base.event = &e->event.base;
-	e->base.file_priv = file;
+	e->base.file_priv = c_node->filp;
 	e->base.destroy = ipp_free_event;
 	mutex_lock(&c_node->event_lock);
 	list_add_tail(&e->base.link, &c_node->event_list);
@@ -839,6 +648,115 @@ out_unlock:
 	return;
 }
 
+static void ipp_clean_cmd_node(struct ipp_context *ctx,
+				struct drm_exynos_ipp_cmd_node *c_node)
+{
+	int i;
+
+	/* cancel works */
+	cancel_work_sync(&c_node->start_work->work);
+	cancel_work_sync(&c_node->stop_work->work);
+	cancel_work_sync(&c_node->event_work->work);
+
+	/* put event */
+	ipp_put_event(c_node, NULL);
+
+	for_each_ipp_ops(i)
+		ipp_clean_mem_nodes(ctx->subdrv.drm_dev, c_node, i);
+
+	/* delete list */
+	list_del(&c_node->list);
+
+	ipp_remove_id(&ctx->prop_idr, &ctx->prop_lock,
+			c_node->property.prop_id);
+
+	/* destroy mutex */
+	mutex_destroy(&c_node->lock);
+	mutex_destroy(&c_node->mem_lock);
+	mutex_destroy(&c_node->event_lock);
+
+	/* free command node */
+	kfree(c_node->start_work);
+	kfree(c_node->stop_work);
+	kfree(c_node->event_work);
+	kfree(c_node);
+}
+
+static bool ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
+{
+	switch (c_node->property.cmd) {
+	case IPP_CMD_WB:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
+	case IPP_CMD_OUTPUT:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]);
+	case IPP_CMD_M2M:
+	default:
+		return !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_SRC]) &&
+		       !list_empty(&c_node->mem_list[EXYNOS_DRM_OPS_DST]);
+	}
+}
+
+static struct drm_exynos_ipp_mem_node
+		*ipp_find_mem_node(struct drm_exynos_ipp_cmd_node *c_node,
+		struct drm_exynos_ipp_queue_buf *qbuf)
+{
+	struct drm_exynos_ipp_mem_node *m_node;
+	struct list_head *head;
+	int count = 0;
+
+	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
+
+	/* source/destination memory list */
+	head = &c_node->mem_list[qbuf->ops_id];
+
+	/* find memory node from memory list */
+	list_for_each_entry(m_node, head, list) {
+		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
+
+		/* compare buffer id */
+		if (m_node->buf_id == qbuf->buf_id)
+			return m_node;
+	}
+
+	return NULL;
+}
+
+static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
+		struct drm_exynos_ipp_cmd_node *c_node,
+		struct drm_exynos_ipp_mem_node *m_node)
+{
+	struct exynos_drm_ipp_ops *ops = NULL;
+	int ret = 0;
+
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
+
+	if (!m_node) {
+		DRM_ERROR("invalid queue node.\n");
+		return -EFAULT;
+	}
+
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
+
+	/* get operations callback */
+	ops = ippdrv->ops[m_node->ops_id];
+	if (!ops) {
+		DRM_ERROR("not support ops.\n");
+		return -EFAULT;
+	}
+
+	/* set address and enable irq */
+	if (ops->set_addr) {
+		ret = ops->set_addr(ippdrv->dev, &m_node->buf_info,
+			m_node->buf_id, IPP_BUF_ENQUEUE);
+		if (ret) {
+			DRM_ERROR("failed to set addr.\n");
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
 static void ipp_handle_cmd_work(struct device *dev,
 		struct exynos_drm_ippdrv *ippdrv,
 		struct drm_exynos_ipp_cmd_work *cmd_work,
@@ -848,7 +766,7 @@ static void ipp_handle_cmd_work(struct device *dev,
 
 	cmd_work->ippdrv = ippdrv;
 	cmd_work->c_node = c_node;
-	queue_work(ctx->cmd_workq, (struct work_struct *)cmd_work);
+	queue_work(ctx->cmd_workq, &cmd_work->work);
 }
 
 static int ipp_queue_buf_with_run(struct device *dev,
@@ -930,8 +848,7 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 		struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv = file_priv->ipp_priv;
-	struct device *dev = priv->dev;
+	struct device *dev = file_priv->ipp_dev;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_queue_buf *qbuf = data;
 	struct drm_exynos_ipp_cmd_node *c_node;
@@ -955,16 +872,16 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 	/* find command node */
 	c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock,
 		qbuf->prop_id);
-	if (IS_ERR(c_node)) {
+	if (!c_node || c_node->filp != file) {
 		DRM_ERROR("failed to get command node.\n");
-		return PTR_ERR(c_node);
+		return -ENODEV;
 	}
 
 	/* buffer control */
 	switch (qbuf->buf_type) {
 	case IPP_BUF_ENQUEUE:
 		/* get memory node */
-		m_node = ipp_get_mem_node(drm_dev, file, c_node, qbuf);
+		m_node = ipp_get_mem_node(drm_dev, c_node, qbuf);
 		if (IS_ERR(m_node)) {
 			DRM_ERROR("failed to get m_node.\n");
 			return PTR_ERR(m_node);
@@ -977,7 +894,7 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 		 */
 		if (qbuf->ops_id == EXYNOS_DRM_OPS_DST) {
 			/* get event for destination buffer */
-			ret = ipp_get_event(drm_dev, file, c_node, qbuf);
+			ret = ipp_get_event(drm_dev, c_node, qbuf);
 			if (ret) {
 				DRM_ERROR("failed to get event.\n");
 				goto err_clean_node;
@@ -1062,9 +979,8 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 		struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv = file_priv->ipp_priv;
 	struct exynos_drm_ippdrv *ippdrv = NULL;
-	struct device *dev = priv->dev;
+	struct device *dev = file_priv->ipp_dev;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_cmd_ctrl *cmd_ctrl = data;
 	struct drm_exynos_ipp_cmd_work *cmd_work;
@@ -1091,9 +1007,9 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 
 	c_node = ipp_find_obj(&ctx->prop_idr, &ctx->prop_lock,
 		cmd_ctrl->prop_id);
-	if (IS_ERR(c_node)) {
+	if (!c_node || c_node->filp != file) {
 		DRM_ERROR("invalid command node list.\n");
-		return PTR_ERR(c_node);
+		return -ENODEV;
 	}
 
 	if (!exynos_drm_ipp_check_valid(ippdrv->dev, cmd_ctrl->ctrl,
@@ -1198,7 +1114,6 @@ static int ipp_set_property(struct exynos_drm_ippdrv *ippdrv,
 	/* reset h/w block */
 	if (ippdrv->reset &&
 	    ippdrv->reset(ippdrv->dev)) {
-		DRM_ERROR("failed to reset.\n");
 		return -EINVAL;
 	}
 
@@ -1216,30 +1131,24 @@ static int ipp_set_property(struct exynos_drm_ippdrv *ippdrv,
 		/* set format */
 		if (ops->set_fmt) {
 			ret = ops->set_fmt(ippdrv->dev, config->fmt);
-			if (ret) {
-				DRM_ERROR("not support format.\n");
+			if (ret)
 				return ret;
-			}
 		}
 
 		/* set transform for rotation, flip */
 		if (ops->set_transf) {
 			ret = ops->set_transf(ippdrv->dev, config->degree,
 				config->flip, &swap);
-			if (ret) {
-				DRM_ERROR("not support tranf.\n");
-				return -EINVAL;
-			}
+			if (ret)
+				return ret;
 		}
 
 		/* set size */
 		if (ops->set_size) {
 			ret = ops->set_size(ippdrv->dev, swap, &config->pos,
 				&config->sz);
-			if (ret) {
-				DRM_ERROR("not support size.\n");
+			if (ret)
 				return ret;
-			}
 		}
 	}
 
@@ -1283,11 +1192,6 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 
 			m_node = list_first_entry(head,
 				struct drm_exynos_ipp_mem_node, list);
-			if (!m_node) {
-				DRM_ERROR("failed to get node.\n");
-				ret = -EFAULT;
-				goto err_unlock;
-			}
 
 			DRM_DEBUG_KMS("m_node[0x%x]\n", (int)m_node);
 
@@ -1353,80 +1257,39 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		struct exynos_drm_ippdrv *ippdrv,
 		struct drm_exynos_ipp_cmd_node *c_node)
 {
-	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
 	struct drm_exynos_ipp_property *property = &c_node->property;
-	struct list_head *head;
-	int ret = 0, i;
+	int i;
 
 	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
-	/* put event */
-	ipp_put_event(c_node, NULL);
-
-	mutex_lock(&c_node->mem_lock);
+	/* stop operations */
+	if (ippdrv->stop)
+		ippdrv->stop(ippdrv->dev, property->cmd);
 
 	/* check command */
 	switch (property->cmd) {
 	case IPP_CMD_M2M:
-		for_each_ipp_ops(i) {
-			/* source/destination memory list */
-			head = &c_node->mem_list[i];
-
-			list_for_each_entry_safe(m_node, tm_node,
-				head, list) {
-				ret = ipp_put_mem_node(drm_dev, c_node,
-					m_node);
-				if (ret) {
-					DRM_ERROR("failed to put m_node.\n");
-					goto err_clear;
-				}
-			}
-		}
+		for_each_ipp_ops(i)
+			ipp_clean_mem_nodes(drm_dev, c_node, i);
 		break;
 	case IPP_CMD_WB:
-		/* destination memory list */
-		head = &c_node->mem_list[EXYNOS_DRM_OPS_DST];
-
-		list_for_each_entry_safe(m_node, tm_node, head, list) {
-			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
-			if (ret) {
-				DRM_ERROR("failed to put m_node.\n");
-				goto err_clear;
-			}
-		}
+		ipp_clean_mem_nodes(drm_dev, c_node, EXYNOS_DRM_OPS_DST);
 		break;
 	case IPP_CMD_OUTPUT:
-		/* source memory list */
-		head = &c_node->mem_list[EXYNOS_DRM_OPS_SRC];
-
-		list_for_each_entry_safe(m_node, tm_node, head, list) {
-			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
-			if (ret) {
-				DRM_ERROR("failed to put m_node.\n");
-				goto err_clear;
-			}
-		}
+		ipp_clean_mem_nodes(drm_dev, c_node, EXYNOS_DRM_OPS_SRC);
 		break;
 	default:
 		DRM_ERROR("invalid operations.\n");
-		ret = -EINVAL;
-		goto err_clear;
+		return -EINVAL;
 	}
 
-err_clear:
-	mutex_unlock(&c_node->mem_lock);
-
-	/* stop operations */
-	if (ippdrv->stop)
-		ippdrv->stop(ippdrv->dev, property->cmd);
-
-	return ret;
+	return 0;
 }
 
 void ipp_sched_cmd(struct work_struct *work)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work =
-		(struct drm_exynos_ipp_cmd_work *)work;
+		container_of(work, struct drm_exynos_ipp_cmd_work, work);
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
 	struct drm_exynos_ipp_property *property;
@@ -1545,11 +1408,6 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 
 			m_node = list_first_entry(head,
 				struct drm_exynos_ipp_mem_node, list);
-			if (!m_node) {
-				DRM_ERROR("empty memory node.\n");
-				ret = -ENOMEM;
-				goto err_mem_unlock;
-			}
 
 			tbuf_id[i] = m_node->buf_id;
 			DRM_DEBUG_KMS("%s buf_id[%d]\n",
@@ -1586,11 +1444,6 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 
 		m_node = list_first_entry(head,
 			struct drm_exynos_ipp_mem_node, list);
-		if (!m_node) {
-			DRM_ERROR("empty memory node.\n");
-			ret = -ENOMEM;
-			goto err_mem_unlock;
-		}
 
 		tbuf_id[EXYNOS_DRM_OPS_SRC] = m_node->buf_id;
 
@@ -1649,7 +1502,7 @@ err_event_unlock:
 void ipp_sched_event(struct work_struct *work)
 {
 	struct drm_exynos_ipp_event_work *event_work =
-		(struct drm_exynos_ipp_event_work *)work;
+		container_of(work, struct drm_exynos_ipp_event_work, work);
 	struct exynos_drm_ippdrv *ippdrv;
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int ret;
@@ -1704,21 +1557,17 @@ static int ipp_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 
 	/* get ipp driver entry */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
-		u32 ipp_id;
-
 		ippdrv->drm_dev = drm_dev;
 
-		ret = ipp_create_id(&ctx->ipp_idr, &ctx->ipp_lock, ippdrv,
-				    &ipp_id);
-		if (ret || ipp_id == 0) {
+		ret = ipp_create_id(&ctx->ipp_idr, &ctx->ipp_lock, ippdrv);
+		if (ret < 0) {
 			DRM_ERROR("failed to create id.\n");
 			goto err;
 		}
+		ippdrv->prop_list.ipp_id = ret;
 
 		DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]ipp_id[%d]\n",
-			count++, (int)ippdrv, ipp_id);
-
-		ippdrv->prop_list.ipp_id = ipp_id;
+			count++, (int)ippdrv, ret);
 
 		/* store parent device for node */
 		ippdrv->parent_dev = dev;
@@ -1756,11 +1605,11 @@ err:
 
 static void ipp_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	struct exynos_drm_ippdrv *ippdrv;
+	struct exynos_drm_ippdrv *ippdrv, *t;
 	struct ipp_context *ctx = get_ipp_context(dev);
 
 	/* get ipp driver entry */
-	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
+	list_for_each_entry_safe(ippdrv, t, &exynos_drm_ippdrv_list, drv_list) {
 		if (is_drm_iommu_supported(drm_dev))
 			drm_iommu_detach_device(drm_dev, ippdrv->dev);
 
@@ -1776,17 +1625,10 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 		struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv;
-
-	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-	priv->dev = dev;
-	file_priv->ipp_priv = priv;
 
-	INIT_LIST_HEAD(&priv->event_list);
+	file_priv->ipp_dev = dev;
 
-	DRM_DEBUG_KMS("done priv[0x%x]\n", (int)priv);
+	DRM_DEBUG_KMS("done priv[0x%x]\n", (int)dev);
 
 	return 0;
 }
@@ -1794,15 +1636,11 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 		struct drm_file *file)
 {
-	struct drm_exynos_file_private *file_priv = file->driver_priv;
-	struct exynos_drm_ipp_private *priv = file_priv->ipp_priv;
 	struct exynos_drm_ippdrv *ippdrv = NULL;
 	struct ipp_context *ctx = get_ipp_context(dev);
 	struct drm_exynos_ipp_cmd_node *c_node, *tc_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("for priv[0x%x]\n", (int)priv);
-
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		mutex_lock(&ippdrv->cmd_lock);
 		list_for_each_entry_safe(c_node, tc_node,
@@ -1810,7 +1648,7 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 			DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n",
 				count++, (int)ippdrv);
 
-			if (c_node->priv == priv) {
+			if (c_node->filp == file) {
 				/*
 				 * userland goto unnormal state. process killed.
 				 * and close the file.
@@ -1832,7 +1670,6 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 		mutex_unlock(&ippdrv->cmd_lock);
 	}
 
-	kfree(priv);
 	return;
 }
 
@@ -1927,63 +1764,12 @@ static int ipp_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static int ipp_power_ctrl(struct ipp_context *ctx, bool enable)
-{
-	DRM_DEBUG_KMS("enable[%d]\n", enable);
-
-	return 0;
-}
-
-#ifdef CONFIG_PM_SLEEP
-static int ipp_suspend(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	if (pm_runtime_suspended(dev))
-		return 0;
-
-	return ipp_power_ctrl(ctx, false);
-}
-
-static int ipp_resume(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	if (!pm_runtime_suspended(dev))
-		return ipp_power_ctrl(ctx, true);
-
-	return 0;
-}
-#endif
-
-#ifdef CONFIG_PM_RUNTIME
-static int ipp_runtime_suspend(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	return ipp_power_ctrl(ctx, false);
-}
-
-static int ipp_runtime_resume(struct device *dev)
-{
-	struct ipp_context *ctx = get_ipp_context(dev);
-
-	return ipp_power_ctrl(ctx, true);
-}
-#endif
-
-static const struct dev_pm_ops ipp_pm_ops = {
-	SET_SYSTEM_SLEEP_PM_OPS(ipp_suspend, ipp_resume)
-	SET_RUNTIME_PM_OPS(ipp_runtime_suspend, ipp_runtime_resume, NULL)
-};
-
 struct platform_driver ipp_driver = {
 	.probe		= ipp_probe,
 	.remove		= ipp_remove,
 	.driver		= {
 		.name	= "exynos-drm-ipp",
 		.owner	= THIS_MODULE,
-		.pm	= &ipp_pm_ops,
 	},
 };
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.h b/drivers/gpu/drm/exynos/exynos_drm_ipp.h
index 7aaeaae757c2..2a61547a39d0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.h
@@ -48,7 +48,6 @@ struct drm_exynos_ipp_cmd_work {
 /*
  * A structure of command node.
  *
- * @priv: IPP private information.
  * @list: list head to command queue information.
  * @event_list: list head of event.
  * @mem_list: list head to source,destination memory queue information.
@@ -62,9 +61,9 @@ struct drm_exynos_ipp_cmd_work {
  * @stop_work: stop command work structure.
  * @event_work: event work structure.
  * @state: state of command node.
+ * @filp: associated file pointer.
  */
 struct drm_exynos_ipp_cmd_node {
-	struct exynos_drm_ipp_private *priv;
 	struct list_head	list;
 	struct list_head	event_list;
 	struct list_head	mem_list[EXYNOS_DRM_OPS_MAX];
@@ -78,6 +77,7 @@ struct drm_exynos_ipp_cmd_node {
 	struct drm_exynos_ipp_cmd_work *stop_work;
 	struct drm_exynos_ipp_event_work *event_work;
 	enum drm_exynos_ipp_state	state;
+	struct drm_file	*filp;
 };
 
 /*
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c
index 8371cbd7631d..c7045a663763 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c
@@ -139,6 +139,8 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 			overlay->crtc_x, overlay->crtc_y,
 			overlay->crtc_width, overlay->crtc_height);
 
+	plane->crtc = crtc;
+
 	exynos_drm_crtc_plane_mode_set(crtc, overlay);
 
 	return 0;
@@ -187,8 +189,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (ret < 0)
 		return ret;
 
-	plane->crtc = crtc;
-
 	exynos_plane_commit(plane);
 	exynos_plane_dpms(plane, DRM_MODE_DPMS_ON);
 
@@ -254,25 +254,26 @@ static void exynos_plane_attach_zpos_property(struct drm_plane *plane)
 }
 
 struct drm_plane *exynos_plane_init(struct drm_device *dev,
-				    unsigned long possible_crtcs, bool priv)
+				    unsigned long possible_crtcs,
+				    enum drm_plane_type type)
 {
 	struct exynos_plane *exynos_plane;
 	int err;
 
 	exynos_plane = kzalloc(sizeof(struct exynos_plane), GFP_KERNEL);
 	if (!exynos_plane)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
-	err = drm_plane_init(dev, &exynos_plane->base, possible_crtcs,
-			      &exynos_plane_funcs, formats, ARRAY_SIZE(formats),
-			      priv);
+	err = drm_universal_plane_init(dev, &exynos_plane->base, possible_crtcs,
+				       &exynos_plane_funcs, formats,
+				       ARRAY_SIZE(formats), type);
 	if (err) {
 		DRM_ERROR("failed to initialize plane\n");
 		kfree(exynos_plane);
-		return NULL;
+		return ERR_PTR(err);
 	}
 
-	if (priv)
+	if (type == DRM_PLANE_TYPE_PRIMARY)
 		exynos_plane->overlay.zpos = DEFAULT_ZPOS;
 	else
 		exynos_plane_attach_zpos_property(&exynos_plane->base);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.h b/drivers/gpu/drm/exynos/exynos_drm_plane.h
index 84d464c90d3d..0d1986b115f8 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.h
@@ -17,4 +17,5 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 void exynos_plane_commit(struct drm_plane *plane);
 void exynos_plane_dpms(struct drm_plane *plane, int mode);
 struct drm_plane *exynos_plane_init(struct drm_device *dev,
-				    unsigned long possible_crtcs, bool priv);
+				    unsigned long possible_crtcs,
+				    enum drm_plane_type type);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index f01fbb6dc1f0..b6a37d4f5b13 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -156,8 +156,7 @@ static irqreturn_t rotator_irq_handler(int irq, void *arg)
 		event_work->ippdrv = ippdrv;
 		event_work->buf_id[EXYNOS_DRM_OPS_DST] =
 			rot->cur_buf_id[EXYNOS_DRM_OPS_DST];
-		queue_work(ippdrv->event_workq,
-			(struct work_struct *)event_work);
+		queue_work(ippdrv->event_workq, &event_work->work);
 	} else {
 		DRM_ERROR("the SFR is set illegally\n");
 	}
@@ -691,6 +690,7 @@ static const struct of_device_id exynos_rotator_match[] = {
 	},
 	{},
 };
+MODULE_DEVICE_TABLE(of, exynos_rotator_match);
 
 static int rotator_probe(struct platform_device *pdev)
 {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
index 2fb8705d6461..50faf913e574 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_vidi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
@@ -303,23 +303,6 @@ static int vidi_mgr_initialize(struct exynos_drm_manager *mgr,
 	mgr->drm_dev = ctx->drm_dev = drm_dev;
 	mgr->pipe = ctx->pipe = priv->pipe++;
 
-	/*
-	 * enable drm irq mode.
-	 * - with irq_enabled = 1, we can use the vblank feature.
-	 *
-	 * P.S. note that we wouldn't use drm irq handler but
-	 *	just specific driver own one instead because
-	 *	drm framework supports only one irq handler.
-	 */
-	drm_dev->irq_enabled = 1;
-
-	/*
-	 * with vblank_disable_allowed = 1, vblank interrupt will be disabled
-	 * by drm timer once a current process gives up ownership of
-	 * vblank event.(after drm_vblank_put function is called)
-	 */
-	drm_dev->vblank_disable_allowed = 1;
-
 	return 0;
 }
 
@@ -562,7 +545,7 @@ static int vidi_create_connector(struct exynos_drm_display *display,
 	}
 
 	drm_connector_helper_add(connector, &vidi_connector_helper_funcs);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	return 0;
@@ -647,8 +630,6 @@ static int vidi_remove(struct platform_device *pdev)
 {
 	struct exynos_drm_manager *mgr = platform_get_drvdata(pdev);
 	struct vidi_context *ctx = mgr->ctx;
-	struct drm_encoder *encoder = ctx->encoder;
-	struct drm_crtc *crtc = mgr->crtc;
 
 	if (ctx->raw_edid != (struct edid *)fake_edid_info) {
 		kfree(ctx->raw_edid);
@@ -657,10 +638,6 @@ static int vidi_remove(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
-	crtc->funcs->destroy(crtc);
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&ctx->connector);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index aa259b0a873a..563a19e62eb2 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -84,6 +84,7 @@ struct hdmi_resources {
 	struct clk			*sclk_hdmiphy;
 	struct clk			*mout_hdmi;
 	struct regulator_bulk_data	*regul_bulk;
+	struct regulator		*reg_hdmi_en;
 	int				regul_count;
 };
 
@@ -592,6 +593,13 @@ static struct hdmi_driver_data exynos4212_hdmi_driver_data = {
 	.is_apb_phy	= 0,
 };
 
+static struct hdmi_driver_data exynos4210_hdmi_driver_data = {
+	.type		= HDMI_TYPE13,
+	.phy_confs	= hdmiphy_v13_configs,
+	.phy_conf_count	= ARRAY_SIZE(hdmiphy_v13_configs),
+	.is_apb_phy	= 0,
+};
+
 static struct hdmi_driver_data exynos5_hdmi_driver_data = {
 	.type		= HDMI_TYPE14,
 	.phy_confs	= hdmiphy_v13_configs,
@@ -1032,6 +1040,8 @@ static enum drm_connector_status hdmi_detect(struct drm_connector *connector,
 
 static void hdmi_connector_destroy(struct drm_connector *connector)
 {
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
 }
 
 static struct drm_connector_funcs hdmi_connector_funcs = {
@@ -1129,7 +1139,7 @@ static int hdmi_create_connector(struct exynos_drm_display *display,
 	}
 
 	drm_connector_helper_add(connector, &hdmi_connector_helper_funcs);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	return 0;
@@ -1241,14 +1251,13 @@ static void hdmi_reg_acr(struct hdmi_context *hdata, u8 *acr)
 
 static void hdmi_audio_init(struct hdmi_context *hdata)
 {
-	u32 sample_rate, bits_per_sample, frame_size_code;
+	u32 sample_rate, bits_per_sample;
 	u32 data_num, bit_ch, sample_frq;
 	u32 val;
 	u8 acr[7];
 
 	sample_rate = 44100;
 	bits_per_sample = 16;
-	frame_size_code = 0;
 
 	switch (bits_per_sample) {
 	case 20:
@@ -2168,7 +2177,6 @@ static int hdmi_resources_init(struct hdmi_context *hdata)
 	struct device *dev = hdata->dev;
 	struct hdmi_resources *res = &hdata->res;
 	static char *supply[] = {
-		"hdmi-en",
 		"vdd",
 		"vdd_osc",
 		"vdd_pll",
@@ -2228,6 +2236,20 @@ static int hdmi_resources_init(struct hdmi_context *hdata)
 	}
 	res->regul_count = ARRAY_SIZE(supply);
 
+	res->reg_hdmi_en = devm_regulator_get(dev, "hdmi-en");
+	if (IS_ERR(res->reg_hdmi_en) && PTR_ERR(res->reg_hdmi_en) != -ENOENT) {
+		DRM_ERROR("failed to get hdmi-en regulator\n");
+		return PTR_ERR(res->reg_hdmi_en);
+	}
+	if (!IS_ERR(res->reg_hdmi_en)) {
+		ret = regulator_enable(res->reg_hdmi_en);
+		if (ret) {
+			DRM_ERROR("failed to enable hdmi-en regulator\n");
+			return ret;
+		}
+	} else
+		res->reg_hdmi_en = NULL;
+
 	return ret;
 fail:
 	DRM_ERROR("HDMI resource init - failed\n");
@@ -2263,6 +2285,9 @@ static struct of_device_id hdmi_match_types[] = {
 		.compatible = "samsung,exynos5-hdmi",
 		.data = &exynos5_hdmi_driver_data,
 	}, {
+		.compatible = "samsung,exynos4210-hdmi",
+		.data = &exynos4210_hdmi_driver_data,
+	}, {
 		.compatible = "samsung,exynos4212-hdmi",
 		.data = &exynos4212_hdmi_driver_data,
 	}, {
@@ -2272,6 +2297,7 @@ static struct of_device_id hdmi_match_types[] = {
 		/* end node */
 	}
 };
+MODULE_DEVICE_TABLE (of, hdmi_match_types);
 
 static int hdmi_bind(struct device *dev, struct device *master, void *data)
 {
@@ -2286,12 +2312,6 @@ static int hdmi_bind(struct device *dev, struct device *master, void *data)
 
 static void hdmi_unbind(struct device *dev, struct device *master, void *data)
 {
-	struct exynos_drm_display *display = get_hdmi_display(dev);
-	struct drm_encoder *encoder = display->encoder;
-	struct hdmi_context *hdata = display->ctx;
-
-	encoder->funcs->destroy(encoder);
-	drm_connector_cleanup(&hdata->connector);
 }
 
 static const struct component_ops hdmi_component_ops = {
@@ -2494,7 +2514,11 @@ static int hdmi_remove(struct platform_device *pdev)
 
 	cancel_delayed_work_sync(&hdata->hotplug_work);
 
-	put_device(&hdata->hdmiphy_port->dev);
+	if (hdata->res.reg_hdmi_en)
+		regulator_disable(hdata->res.reg_hdmi_en);
+
+	if (hdata->hdmiphy_port)
+		put_device(&hdata->hdmiphy_port->dev);
 	put_device(&hdata->ddc_adpt->dev);
 
 	pm_runtime_disable(&pdev->dev);
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 7529946d0a74..a41c84ee3a2d 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -76,7 +76,7 @@ struct mixer_resources {
 	struct clk		*vp;
 	struct clk		*sclk_mixer;
 	struct clk		*sclk_hdmi;
-	struct clk		*sclk_dac;
+	struct clk		*mout_mixer;
 };
 
 enum mixer_version_id {
@@ -93,6 +93,7 @@ struct mixer_context {
 	bool			interlace;
 	bool			powered;
 	bool			vp_enabled;
+	bool			has_sclk;
 	u32			int_en;
 
 	struct mutex		mixer_mutex;
@@ -106,6 +107,7 @@ struct mixer_context {
 struct mixer_drv_data {
 	enum mixer_version_id	version;
 	bool					is_vp_enabled;
+	bool					has_sclk;
 };
 
 static const u8 filter_y_horiz_tap8[] = {
@@ -363,6 +365,11 @@ static void mixer_cfg_layer(struct mixer_context *ctx, int win, bool enable)
 			vp_reg_writemask(res, VP_ENABLE, val, VP_ENABLE_ON);
 			mixer_reg_writemask(res, MXR_CFG, val,
 				MXR_CFG_VP_ENABLE);
+
+			/* control blending of graphic layer 0 */
+			mixer_reg_writemask(res, MXR_GRAPHIC_CFG(0), val,
+					MXR_GRP_CFG_BLEND_PRE_MUL |
+					MXR_GRP_CFG_PIXEL_BLEND_EN);
 		}
 		break;
 	}
@@ -809,19 +816,23 @@ static int vp_resources_init(struct mixer_context *mixer_ctx)
 		dev_err(dev, "failed to get clock 'vp'\n");
 		return -ENODEV;
 	}
-	mixer_res->sclk_mixer = devm_clk_get(dev, "sclk_mixer");
-	if (IS_ERR(mixer_res->sclk_mixer)) {
-		dev_err(dev, "failed to get clock 'sclk_mixer'\n");
-		return -ENODEV;
-	}
-	mixer_res->sclk_dac = devm_clk_get(dev, "sclk_dac");
-	if (IS_ERR(mixer_res->sclk_dac)) {
-		dev_err(dev, "failed to get clock 'sclk_dac'\n");
-		return -ENODEV;
-	}
 
-	if (mixer_res->sclk_hdmi)
-		clk_set_parent(mixer_res->sclk_mixer, mixer_res->sclk_hdmi);
+	if (mixer_ctx->has_sclk) {
+		mixer_res->sclk_mixer = devm_clk_get(dev, "sclk_mixer");
+		if (IS_ERR(mixer_res->sclk_mixer)) {
+			dev_err(dev, "failed to get clock 'sclk_mixer'\n");
+			return -ENODEV;
+		}
+		mixer_res->mout_mixer = devm_clk_get(dev, "mout_mixer");
+		if (IS_ERR(mixer_res->mout_mixer)) {
+			dev_err(dev, "failed to get clock 'mout_mixer'\n");
+			return -ENODEV;
+		}
+
+		if (mixer_res->sclk_hdmi && mixer_res->mout_mixer)
+			clk_set_parent(mixer_res->mout_mixer,
+				       mixer_res->sclk_hdmi);
+	}
 
 	res = platform_get_resource(mixer_ctx->pdev, IORESOURCE_MEM, 1);
 	if (res == NULL) {
@@ -1082,7 +1093,8 @@ static void mixer_poweron(struct exynos_drm_manager *mgr)
 	clk_prepare_enable(res->mixer);
 	if (ctx->vp_enabled) {
 		clk_prepare_enable(res->vp);
-		clk_prepare_enable(res->sclk_mixer);
+		if (ctx->has_sclk)
+			clk_prepare_enable(res->sclk_mixer);
 	}
 
 	mutex_lock(&ctx->mixer_mutex);
@@ -1121,7 +1133,8 @@ static void mixer_poweroff(struct exynos_drm_manager *mgr)
 	clk_disable_unprepare(res->mixer);
 	if (ctx->vp_enabled) {
 		clk_disable_unprepare(res->vp);
-		clk_disable_unprepare(res->sclk_mixer);
+		if (ctx->has_sclk)
+			clk_disable_unprepare(res->sclk_mixer);
 	}
 
 	pm_runtime_put_sync(ctx->dev);
@@ -1189,9 +1202,15 @@ static struct mixer_drv_data exynos5250_mxr_drv_data = {
 	.is_vp_enabled = 0,
 };
 
+static struct mixer_drv_data exynos4212_mxr_drv_data = {
+	.version = MXR_VER_0_0_0_16,
+	.is_vp_enabled = 1,
+};
+
 static struct mixer_drv_data exynos4210_mxr_drv_data = {
 	.version = MXR_VER_0_0_0_16,
 	.is_vp_enabled = 1,
+	.has_sclk = 1,
 };
 
 static struct platform_device_id mixer_driver_types[] = {
@@ -1208,6 +1227,12 @@ static struct platform_device_id mixer_driver_types[] = {
 
 static struct of_device_id mixer_match_types[] = {
 	{
+		.compatible = "samsung,exynos4210-mixer",
+		.data	= &exynos4210_mxr_drv_data,
+	}, {
+		.compatible = "samsung,exynos4212-mixer",
+		.data	= &exynos4212_mxr_drv_data,
+	}, {
 		.compatible = "samsung,exynos5-mixer",
 		.data	= &exynos5250_mxr_drv_data,
 	}, {
@@ -1220,6 +1245,7 @@ static struct of_device_id mixer_match_types[] = {
 		/* end node */
 	}
 };
+MODULE_DEVICE_TABLE(of, mixer_match_types);
 
 static int mixer_bind(struct device *dev, struct device *manager, void *data)
 {
@@ -1251,6 +1277,7 @@ static int mixer_bind(struct device *dev, struct device *manager, void *data)
 	ctx->pdev = pdev;
 	ctx->dev = dev;
 	ctx->vp_enabled = drv->is_vp_enabled;
+	ctx->has_sclk = drv->has_sclk;
 	ctx->mxr_ver = drv->version;
 	init_waitqueue_head(&ctx->wait_vsync_queue);
 	atomic_set(&ctx->wait_vsync_event, 0);
@@ -1275,15 +1302,12 @@ static int mixer_bind(struct device *dev, struct device *manager, void *data)
 static void mixer_unbind(struct device *dev, struct device *master, void *data)
 {
 	struct exynos_drm_manager *mgr = dev_get_drvdata(dev);
-	struct drm_crtc *crtc = mgr->crtc;
 
 	dev_info(dev, "remove successful\n");
 
 	mixer_mgr_remove(mgr);
 
 	pm_runtime_disable(dev);
-
-	crtc->funcs->destroy(crtc);
 }
 
 static const struct component_ops mixer_component_ops = {
diff --git a/drivers/gpu/drm/gma500/cdv_intel_crt.c b/drivers/gpu/drm/gma500/cdv_intel_crt.c
index c18268cd516e..248c33a35ebf 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_crt.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_crt.c
@@ -192,7 +192,7 @@ static void cdv_intel_crt_destroy(struct drm_connector *connector)
 	struct gma_encoder *gma_encoder = gma_attached_encoder(connector);
 
 	psb_intel_i2c_destroy(gma_encoder->ddc_bus);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -304,7 +304,7 @@ void cdv_intel_crt_init(struct drm_device *dev,
 	drm_connector_helper_add(connector,
 					&cdv_intel_crt_connector_helper_funcs);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	return;
 failed_ddc:
diff --git a/drivers/gpu/drm/gma500/cdv_intel_dp.c b/drivers/gpu/drm/gma500/cdv_intel_dp.c
index 9ff30c2efadb..9f158eab517a 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_dp.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_dp.c
@@ -1089,7 +1089,7 @@ static char	*link_train_names[] = {
 };
 #endif
 
-#define CDV_DP_VOLTAGE_MAX	    DP_TRAIN_VOLTAGE_SWING_1200
+#define CDV_DP_VOLTAGE_MAX	    DP_TRAIN_VOLTAGE_SWING_LEVEL_3
 /*
 static uint8_t
 cdv_intel_dp_pre_emphasis_max(uint8_t voltage_swing)
@@ -1276,7 +1276,7 @@ cdv_intel_dp_set_vswing_premph(struct gma_encoder *encoder, uint8_t signal_level
 		cdv_sb_write(dev, ddi_reg->VSwing2, dp_vswing_premph_table[index]);
 
 	/* ;gfx_dpio_set_reg(0x814c, 0x40802040) */
-	if ((vswing + premph) == DP_TRAIN_VOLTAGE_SWING_1200)
+	if ((vswing + premph) == DP_TRAIN_VOLTAGE_SWING_LEVEL_3)
 		cdv_sb_write(dev, ddi_reg->VSwing3, 0x70802040);
 	else
 		cdv_sb_write(dev, ddi_reg->VSwing3, 0x40802040);
@@ -1713,7 +1713,7 @@ cdv_intel_dp_destroy(struct drm_connector *connector)
 		}
 	}
 	i2c_del_adapter(&intel_dp->adapter);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -1847,7 +1847,7 @@ cdv_intel_dp_init(struct drm_device *dev, struct psb_intel_mode_device *mode_dev
 	connector->interlace_allowed = false;
 	connector->doublescan_allowed = false;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	/* Set up the DDC bus. */
 	switch (output_reg) {
diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
index b99084b3f706..4268bf210034 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c
@@ -248,7 +248,7 @@ static void cdv_hdmi_destroy(struct drm_connector *connector)
 
 	if (gma_encoder->i2c_bus)
 		psb_intel_i2c_destroy(gma_encoder->i2c_bus);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -356,7 +356,7 @@ void cdv_hdmi_init(struct drm_device *dev,
 
 	hdmi_priv->hdmi_i2c_adapter = &(gma_encoder->i2c_bus->adapter);
 	hdmi_priv->dev = dev;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return;
 
 failed_ddc:
diff --git a/drivers/gpu/drm/gma500/cdv_intel_lvds.c b/drivers/gpu/drm/gma500/cdv_intel_lvds.c
index 8ecc920fc26d..0b770396548c 100644
--- a/drivers/gpu/drm/gma500/cdv_intel_lvds.c
+++ b/drivers/gpu/drm/gma500/cdv_intel_lvds.c
@@ -446,7 +446,7 @@ static void cdv_intel_lvds_destroy(struct drm_connector *connector)
 
 	if (gma_encoder->i2c_bus)
 		psb_intel_i2c_destroy(gma_encoder->i2c_bus);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -774,7 +774,7 @@ void cdv_intel_lvds_init(struct drm_device *dev,
 
 out:
 	mutex_unlock(&dev->mode_config.mutex);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return;
 
 failed_find:
diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c
index e7fcc148f333..ddd90ddbc200 100644
--- a/drivers/gpu/drm/gma500/framebuffer.c
+++ b/drivers/gpu/drm/gma500/framebuffer.c
@@ -540,7 +540,8 @@ static void psbfb_gamma_get(struct drm_crtc *crtc, u16 *red,
 static int psbfb_probe(struct drm_fb_helper *helper,
 				struct drm_fb_helper_surface_size *sizes)
 {
-	struct psb_fbdev *psb_fbdev = (struct psb_fbdev *)helper;
+	struct psb_fbdev *psb_fbdev =
+		container_of(helper, struct psb_fbdev, psb_fb_helper);
 	struct drm_device *dev = psb_fbdev->psb_fb_helper.dev;
 	struct drm_psb_private *dev_priv = dev->dev_private;
 	int bytespp;
@@ -561,7 +562,7 @@ static int psbfb_probe(struct drm_fb_helper *helper,
 	return psbfb_create(psb_fbdev, sizes);
 }
 
-static struct drm_fb_helper_funcs psb_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs psb_fb_helper_funcs = {
 	.gamma_set = psbfb_gamma_set,
 	.gamma_get = psbfb_gamma_get,
 	.fb_probe = psbfb_probe,
@@ -600,7 +601,8 @@ int psb_fbdev_init(struct drm_device *dev)
 	}
 
 	dev_priv->fbdev = fbdev;
-	fbdev->psb_fb_helper.funcs = &psb_fb_helper_funcs;
+
+	drm_fb_helper_prepare(dev, &fbdev->psb_fb_helper, &psb_fb_helper_funcs);
 
 	drm_fb_helper_init(dev, &fbdev->psb_fb_helper, dev_priv->ops->crtcs,
 							INTELFB_CONN_LIMIT);
diff --git a/drivers/gpu/drm/gma500/gtt.c b/drivers/gpu/drm/gma500/gtt.c
index 592d205a0089..ce015db59dc6 100644
--- a/drivers/gpu/drm/gma500/gtt.c
+++ b/drivers/gpu/drm/gma500/gtt.c
@@ -206,7 +206,7 @@ static int psb_gtt_attach_pages(struct gtt_range *gt)
 
 	WARN_ON(gt->pages);
 
-	pages = drm_gem_get_pages(&gt->gem, 0);
+	pages = drm_gem_get_pages(&gt->gem);
 	if (IS_ERR(pages))
 		return PTR_ERR(pages);
 
diff --git a/drivers/gpu/drm/gma500/gtt.h b/drivers/gpu/drm/gma500/gtt.h
index f5860a739bd8..cdbb350c9d5d 100644
--- a/drivers/gpu/drm/gma500/gtt.h
+++ b/drivers/gpu/drm/gma500/gtt.h
@@ -21,6 +21,7 @@
 #define _PSB_GTT_H_
 
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
 
 /* This wants cleaning up with respect to the psb_dev and un-needed stuff */
 struct psb_gtt {
diff --git a/drivers/gpu/drm/gma500/intel_bios.c b/drivers/gpu/drm/gma500/intel_bios.c
index d3497348c4d5..63bde4e86c6a 100644
--- a/drivers/gpu/drm/gma500/intel_bios.c
+++ b/drivers/gpu/drm/gma500/intel_bios.c
@@ -116,30 +116,30 @@ parse_edp(struct drm_psb_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->preemphasis) {
 	case 0:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_0;
 		break;
 	case 1:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_1;
 		break;
 	case 2:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_2;
 		break;
 	case 3:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_3;
 		break;
 	}
 	switch (edp_link_params->vswing) {
 	case 0:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 		break;
 	case 1:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
 		break;
 	case 2:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 		break;
 	case 3:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 		break;
 	}
 	DRM_DEBUG_KMS("VBT reports EDP: VSwing  %d, Preemph %d\n",
diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
index 6e91b20ce2e5..abf2248da61e 100644
--- a/drivers/gpu/drm/gma500/mdfld_dsi_output.c
+++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c
@@ -318,7 +318,7 @@ static void mdfld_dsi_connector_destroy(struct drm_connector *connector)
 
 	if (!dsi_connector)
 		return;
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	sender = dsi_connector->pkg_sender;
 	mdfld_dsi_pkg_sender_destroy(sender);
@@ -597,7 +597,7 @@ void mdfld_dsi_output_init(struct drm_device *dev,
 	dsi_config->encoder = encoder;
 	encoder->base.type = (pipe == 0) ? INTEL_OUTPUT_MIPI :
 		INTEL_OUTPUT_MIPI2;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return;
 
 	/*TODO: add code to destroy outputs on error*/
diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi.c b/drivers/gpu/drm/gma500/oaktrail_hdmi.c
index cf018ddcc5a6..54f73f50571a 100644
--- a/drivers/gpu/drm/gma500/oaktrail_hdmi.c
+++ b/drivers/gpu/drm/gma500/oaktrail_hdmi.c
@@ -665,7 +665,7 @@ void oaktrail_hdmi_init(struct drm_device *dev,
 	connector->display_info.subpixel_order = SubPixelHorizontalRGB;
 	connector->interlace_allowed = false;
 	connector->doublescan_allowed = false;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	dev_info(dev->dev, "HDMI initialised.\n");
 
 	return;
@@ -674,7 +674,7 @@ failed_connector:
 	kfree(gma_encoder);
 }
 
-static DEFINE_PCI_DEVICE_TABLE(hdmi_ids) = {
+static const struct pci_device_id hdmi_ids[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x080d) },
 	{ 0 }
 };
diff --git a/drivers/gpu/drm/gma500/oaktrail_lvds.c b/drivers/gpu/drm/gma500/oaktrail_lvds.c
index 9b099468a5db..0d39da6e8b7a 100644
--- a/drivers/gpu/drm/gma500/oaktrail_lvds.c
+++ b/drivers/gpu/drm/gma500/oaktrail_lvds.c
@@ -404,7 +404,7 @@ void oaktrail_lvds_init(struct drm_device *dev,
 out:
 	mutex_unlock(&dev->mode_config.mutex);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return;
 
 failed_find:
diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index 6e8fe9ec02b5..6ec3a905fdd2 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -54,7 +54,7 @@ static int psb_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent);
  * PowerVR SGX545    - Cedartrail - Intel GMA 3650, Intel Atom D2550, D2700,
  *                                  N2800
  */
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+static const struct pci_device_id pciidlist[] = {
 	{ 0x8086, 0x8108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &psb_chip_ops },
 	{ 0x8086, 0x8109, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &psb_chip_ops },
 #if defined(CONFIG_DRM_GMA600)
@@ -476,6 +476,7 @@ static struct drm_driver driver = {
 	.unload = psb_driver_unload,
 	.lastclose = psb_driver_lastclose,
 	.preclose = psb_driver_preclose,
+	.set_busid = drm_pci_set_busid,
 
 	.num_ioctls = ARRAY_SIZE(psb_ioctls),
 	.device_is_agp = psb_driver_device_is_agp,
diff --git a/drivers/gpu/drm/gma500/psb_intel_lvds.c b/drivers/gpu/drm/gma500/psb_intel_lvds.c
index d7778d0472c1..88aad95bde09 100644
--- a/drivers/gpu/drm/gma500/psb_intel_lvds.c
+++ b/drivers/gpu/drm/gma500/psb_intel_lvds.c
@@ -563,7 +563,7 @@ void psb_intel_lvds_destroy(struct drm_connector *connector)
 
 	if (lvds_priv->ddc_bus)
 		psb_intel_i2c_destroy(lvds_priv->ddc_bus);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -829,7 +829,7 @@ void psb_intel_lvds_init(struct drm_device *dev,
 	 */
 out:
 	mutex_unlock(&dev->mode_config.mutex);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return;
 
 failed_find:
diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
index deeb0829b129..0be96fdb5e28 100644
--- a/drivers/gpu/drm/gma500/psb_intel_sdvo.c
+++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c
@@ -1682,7 +1682,7 @@ static void psb_intel_sdvo_destroy(struct drm_connector *connector)
 				     psb_intel_sdvo_connector->tv_format);
 
 	psb_intel_sdvo_destroy_enhance_property(connector);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -2071,7 +2071,7 @@ psb_intel_sdvo_connector_init(struct psb_intel_sdvo_connector *connector,
 	connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB;
 
 	gma_connector_attach_encoder(&connector->base, &encoder->base);
-	drm_sysfs_connector_add(&connector->base.base);
+	drm_connector_register(&connector->base.base);
 }
 
 static void
diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c
index ac357b02bd35..d4762799351d 100644
--- a/drivers/gpu/drm/i2c/tda998x_drv.c
+++ b/drivers/gpu/drm/i2c/tda998x_drv.c
@@ -15,8 +15,7 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-
-
+#include <linux/component.h>
 #include <linux/hdmi.h>
 #include <linux/module.h>
 #include <linux/irq.h>
@@ -730,12 +729,9 @@ tda998x_configure_audio(struct tda998x_priv *priv,
 
 /* DRM encoder functions */
 
-static void
-tda998x_encoder_set_config(struct drm_encoder *encoder, void *params)
+static void tda998x_encoder_set_config(struct tda998x_priv *priv,
+				       const struct tda998x_encoder_params *p)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
-	struct tda998x_encoder_params *p = params;
-
 	priv->vip_cntrl_0 = VIP_CNTRL_0_SWAP_A(p->swap_a) |
 			    (p->mirr_a ? VIP_CNTRL_0_MIRR_A : 0) |
 			    VIP_CNTRL_0_SWAP_B(p->swap_b) |
@@ -752,11 +748,8 @@ tda998x_encoder_set_config(struct drm_encoder *encoder, void *params)
 	priv->params = *p;
 }
 
-static void
-tda998x_encoder_dpms(struct drm_encoder *encoder, int mode)
+static void tda998x_encoder_dpms(struct tda998x_priv *priv, int mode)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
-
 	/* we only care about on or off: */
 	if (mode != DRM_MODE_DPMS_ON)
 		mode = DRM_MODE_DPMS_OFF;
@@ -806,9 +799,8 @@ tda998x_encoder_mode_fixup(struct drm_encoder *encoder,
 	return true;
 }
 
-static int
-tda998x_encoder_mode_valid(struct drm_encoder *encoder,
-			  struct drm_display_mode *mode)
+static int tda998x_encoder_mode_valid(struct tda998x_priv *priv,
+				      struct drm_display_mode *mode)
 {
 	if (mode->clock > 150000)
 		return MODE_CLOCK_HIGH;
@@ -820,11 +812,10 @@ tda998x_encoder_mode_valid(struct drm_encoder *encoder,
 }
 
 static void
-tda998x_encoder_mode_set(struct drm_encoder *encoder,
-			struct drm_display_mode *mode,
-			struct drm_display_mode *adjusted_mode)
+tda998x_encoder_mode_set(struct tda998x_priv *priv,
+			 struct drm_display_mode *mode,
+			 struct drm_display_mode *adjusted_mode)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
 	uint16_t ref_pix, ref_line, n_pix, n_line;
 	uint16_t hs_pix_s, hs_pix_e;
 	uint16_t vs1_pix_s, vs1_pix_e, vs1_line_s, vs1_line_e;
@@ -1012,20 +1003,16 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder,
 }
 
 static enum drm_connector_status
-tda998x_encoder_detect(struct drm_encoder *encoder,
-		      struct drm_connector *connector)
+tda998x_encoder_detect(struct tda998x_priv *priv)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
 	uint8_t val = cec_read(priv, REG_CEC_RXSHPDLEV);
 
 	return (val & CEC_RXSHPDLEV_HPD) ? connector_status_connected :
 			connector_status_disconnected;
 }
 
-static int
-read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk)
+static int read_edid_block(struct tda998x_priv *priv, uint8_t *buf, int blk)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
 	uint8_t offset, segptr;
 	int ret, i;
 
@@ -1079,10 +1066,8 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk)
 	return 0;
 }
 
-static uint8_t *
-do_get_edid(struct drm_encoder *encoder)
+static uint8_t *do_get_edid(struct tda998x_priv *priv)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
 	int j, valid_extensions = 0;
 	uint8_t *block, *new;
 	bool print_bad_edid = drm_debug & DRM_UT_KMS;
@@ -1094,7 +1079,7 @@ do_get_edid(struct drm_encoder *encoder)
 		reg_clear(priv, REG_TX4, TX4_PD_RAM);
 
 	/* base block fetch */
-	if (read_edid_block(encoder, block, 0))
+	if (read_edid_block(priv, block, 0))
 		goto fail;
 
 	if (!drm_edid_block_valid(block, 0, print_bad_edid))
@@ -1111,7 +1096,7 @@ do_get_edid(struct drm_encoder *encoder)
 
 	for (j = 1; j <= block[0x7e]; j++) {
 		uint8_t *ext_block = block + (valid_extensions + 1) * EDID_LENGTH;
-		if (read_edid_block(encoder, ext_block, j))
+		if (read_edid_block(priv, ext_block, j))
 			goto fail;
 
 		if (!drm_edid_block_valid(ext_block, j, print_bad_edid))
@@ -1144,11 +1129,10 @@ fail:
 }
 
 static int
-tda998x_encoder_get_modes(struct drm_encoder *encoder,
-			 struct drm_connector *connector)
+tda998x_encoder_get_modes(struct tda998x_priv *priv,
+			  struct drm_connector *connector)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
-	struct edid *edid = (struct edid *)do_get_edid(encoder);
+	struct edid *edid = (struct edid *)do_get_edid(priv);
 	int n = 0;
 
 	if (edid) {
@@ -1161,18 +1145,14 @@ tda998x_encoder_get_modes(struct drm_encoder *encoder,
 	return n;
 }
 
-static int
-tda998x_encoder_create_resources(struct drm_encoder *encoder,
-				struct drm_connector *connector)
+static void tda998x_encoder_set_polling(struct tda998x_priv *priv,
+					struct drm_connector *connector)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
-
 	if (priv->hdmi->irq)
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 	else
 		connector->polled = DRM_CONNECTOR_POLL_CONNECT |
 			DRM_CONNECTOR_POLL_DISCONNECT;
-	return 0;
 }
 
 static int
@@ -1185,66 +1165,97 @@ tda998x_encoder_set_property(struct drm_encoder *encoder,
 	return 0;
 }
 
-static void
-tda998x_encoder_destroy(struct drm_encoder *encoder)
+static void tda998x_destroy(struct tda998x_priv *priv)
 {
-	struct tda998x_priv *priv = to_tda998x_priv(encoder);
-
 	/* disable all IRQs and free the IRQ handler */
 	cec_write(priv, REG_CEC_RXSHPDINTENA, 0);
 	reg_clear(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD);
 	if (priv->hdmi->irq)
 		free_irq(priv->hdmi->irq, priv);
 
-	if (priv->cec)
-		i2c_unregister_device(priv->cec);
+	i2c_unregister_device(priv->cec);
+}
+
+/* Slave encoder support */
+
+static void
+tda998x_encoder_slave_set_config(struct drm_encoder *encoder, void *params)
+{
+	tda998x_encoder_set_config(to_tda998x_priv(encoder), params);
+}
+
+static void tda998x_encoder_slave_destroy(struct drm_encoder *encoder)
+{
+	struct tda998x_priv *priv = to_tda998x_priv(encoder);
+
+	tda998x_destroy(priv);
 	drm_i2c_encoder_destroy(encoder);
 	kfree(priv);
 }
 
-static struct drm_encoder_slave_funcs tda998x_encoder_funcs = {
-	.set_config = tda998x_encoder_set_config,
-	.destroy = tda998x_encoder_destroy,
-	.dpms = tda998x_encoder_dpms,
-	.save = tda998x_encoder_save,
-	.restore = tda998x_encoder_restore,
-	.mode_fixup = tda998x_encoder_mode_fixup,
-	.mode_valid = tda998x_encoder_mode_valid,
-	.mode_set = tda998x_encoder_mode_set,
-	.detect = tda998x_encoder_detect,
-	.get_modes = tda998x_encoder_get_modes,
-	.create_resources = tda998x_encoder_create_resources,
-	.set_property = tda998x_encoder_set_property,
-};
+static void tda998x_encoder_slave_dpms(struct drm_encoder *encoder, int mode)
+{
+	tda998x_encoder_dpms(to_tda998x_priv(encoder), mode);
+}
 
-/* I2C driver functions */
+static int tda998x_encoder_slave_mode_valid(struct drm_encoder *encoder,
+					    struct drm_display_mode *mode)
+{
+	return tda998x_encoder_mode_valid(to_tda998x_priv(encoder), mode);
+}
 
-static int
-tda998x_probe(struct i2c_client *client, const struct i2c_device_id *id)
+static void
+tda998x_encoder_slave_mode_set(struct drm_encoder *encoder,
+			       struct drm_display_mode *mode,
+			       struct drm_display_mode *adjusted_mode)
 {
-	return 0;
+	tda998x_encoder_mode_set(to_tda998x_priv(encoder), mode, adjusted_mode);
+}
+
+static enum drm_connector_status
+tda998x_encoder_slave_detect(struct drm_encoder *encoder,
+			     struct drm_connector *connector)
+{
+	return tda998x_encoder_detect(to_tda998x_priv(encoder));
+}
+
+static int tda998x_encoder_slave_get_modes(struct drm_encoder *encoder,
+					   struct drm_connector *connector)
+{
+	return tda998x_encoder_get_modes(to_tda998x_priv(encoder), connector);
 }
 
 static int
-tda998x_remove(struct i2c_client *client)
+tda998x_encoder_slave_create_resources(struct drm_encoder *encoder,
+				       struct drm_connector *connector)
 {
+	tda998x_encoder_set_polling(to_tda998x_priv(encoder), connector);
 	return 0;
 }
 
-static int
-tda998x_encoder_init(struct i2c_client *client,
-		    struct drm_device *dev,
-		    struct drm_encoder_slave *encoder_slave)
+static struct drm_encoder_slave_funcs tda998x_encoder_slave_funcs = {
+	.set_config = tda998x_encoder_slave_set_config,
+	.destroy = tda998x_encoder_slave_destroy,
+	.dpms = tda998x_encoder_slave_dpms,
+	.save = tda998x_encoder_save,
+	.restore = tda998x_encoder_restore,
+	.mode_fixup = tda998x_encoder_mode_fixup,
+	.mode_valid = tda998x_encoder_slave_mode_valid,
+	.mode_set = tda998x_encoder_slave_mode_set,
+	.detect = tda998x_encoder_slave_detect,
+	.get_modes = tda998x_encoder_slave_get_modes,
+	.create_resources = tda998x_encoder_slave_create_resources,
+	.set_property = tda998x_encoder_set_property,
+};
+
+/* I2C driver functions */
+
+static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv)
 {
-	struct tda998x_priv *priv;
 	struct device_node *np = client->dev.of_node;
 	u32 video;
 	int rev_lo, rev_hi, ret;
 
-	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-
 	priv->vip_cntrl_0 = VIP_CNTRL_0_SWAP_A(2) | VIP_CNTRL_0_SWAP_B(3);
 	priv->vip_cntrl_1 = VIP_CNTRL_1_SWAP_C(0) | VIP_CNTRL_1_SWAP_D(1);
 	priv->vip_cntrl_2 = VIP_CNTRL_2_SWAP_E(4) | VIP_CNTRL_2_SWAP_F(5);
@@ -1252,17 +1263,11 @@ tda998x_encoder_init(struct i2c_client *client,
 	priv->current_page = 0xff;
 	priv->hdmi = client;
 	priv->cec = i2c_new_dummy(client->adapter, 0x34);
-	if (!priv->cec) {
-		kfree(priv);
+	if (!priv->cec)
 		return -ENODEV;
-	}
 
-	priv->encoder = &encoder_slave->base;
 	priv->dpms = DRM_MODE_DPMS_OFF;
 
-	encoder_slave->slave_priv = priv;
-	encoder_slave->slave_funcs = &tda998x_encoder_funcs;
-
 	/* wake up the device: */
 	cec_write(priv, REG_CEC_ENAMODS,
 			CEC_ENAMODS_EN_RXSENS | CEC_ENAMODS_EN_HDMI);
@@ -1365,12 +1370,231 @@ fail:
 	 */
 	if (priv->cec)
 		i2c_unregister_device(priv->cec);
-	kfree(priv);
-	encoder_slave->slave_priv = NULL;
-	encoder_slave->slave_funcs = NULL;
 	return -ENXIO;
 }
 
+static int tda998x_encoder_init(struct i2c_client *client,
+				struct drm_device *dev,
+				struct drm_encoder_slave *encoder_slave)
+{
+	struct tda998x_priv *priv;
+	int ret;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->encoder = &encoder_slave->base;
+
+	ret = tda998x_create(client, priv);
+	if (ret) {
+		kfree(priv);
+		return ret;
+	}
+
+	encoder_slave->slave_priv = priv;
+	encoder_slave->slave_funcs = &tda998x_encoder_slave_funcs;
+
+	return 0;
+}
+
+struct tda998x_priv2 {
+	struct tda998x_priv base;
+	struct drm_encoder encoder;
+	struct drm_connector connector;
+};
+
+#define conn_to_tda998x_priv2(x) \
+	container_of(x, struct tda998x_priv2, connector);
+
+#define enc_to_tda998x_priv2(x) \
+	container_of(x, struct tda998x_priv2, encoder);
+
+static void tda998x_encoder2_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct tda998x_priv2 *priv = enc_to_tda998x_priv2(encoder);
+
+	tda998x_encoder_dpms(&priv->base, mode);
+}
+
+static void tda998x_encoder_prepare(struct drm_encoder *encoder)
+{
+	tda998x_encoder2_dpms(encoder, DRM_MODE_DPMS_OFF);
+}
+
+static void tda998x_encoder_commit(struct drm_encoder *encoder)
+{
+	tda998x_encoder2_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static void tda998x_encoder2_mode_set(struct drm_encoder *encoder,
+				      struct drm_display_mode *mode,
+				      struct drm_display_mode *adjusted_mode)
+{
+	struct tda998x_priv2 *priv = enc_to_tda998x_priv2(encoder);
+
+	tda998x_encoder_mode_set(&priv->base, mode, adjusted_mode);
+}
+
+static const struct drm_encoder_helper_funcs tda998x_encoder_helper_funcs = {
+	.dpms = tda998x_encoder2_dpms,
+	.save = tda998x_encoder_save,
+	.restore = tda998x_encoder_restore,
+	.mode_fixup = tda998x_encoder_mode_fixup,
+	.prepare = tda998x_encoder_prepare,
+	.commit = tda998x_encoder_commit,
+	.mode_set = tda998x_encoder2_mode_set,
+};
+
+static void tda998x_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct tda998x_priv2 *priv = enc_to_tda998x_priv2(encoder);
+
+	tda998x_destroy(&priv->base);
+	drm_encoder_cleanup(encoder);
+}
+
+static const struct drm_encoder_funcs tda998x_encoder_funcs = {
+	.destroy = tda998x_encoder_destroy,
+};
+
+static int tda998x_connector_get_modes(struct drm_connector *connector)
+{
+	struct tda998x_priv2 *priv = conn_to_tda998x_priv2(connector);
+
+	return tda998x_encoder_get_modes(&priv->base, connector);
+}
+
+static int tda998x_connector_mode_valid(struct drm_connector *connector,
+					struct drm_display_mode *mode)
+{
+	struct tda998x_priv2 *priv = conn_to_tda998x_priv2(connector);
+
+	return tda998x_encoder_mode_valid(&priv->base, mode);
+}
+
+static struct drm_encoder *
+tda998x_connector_best_encoder(struct drm_connector *connector)
+{
+	struct tda998x_priv2 *priv = conn_to_tda998x_priv2(connector);
+
+	return &priv->encoder;
+}
+
+static
+const struct drm_connector_helper_funcs tda998x_connector_helper_funcs = {
+	.get_modes = tda998x_connector_get_modes,
+	.mode_valid = tda998x_connector_mode_valid,
+	.best_encoder = tda998x_connector_best_encoder,
+};
+
+static enum drm_connector_status
+tda998x_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct tda998x_priv2 *priv = conn_to_tda998x_priv2(connector);
+
+	return tda998x_encoder_detect(&priv->base);
+}
+
+static void tda998x_connector_destroy(struct drm_connector *connector)
+{
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+}
+
+static const struct drm_connector_funcs tda998x_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = tda998x_connector_detect,
+	.destroy = tda998x_connector_destroy,
+};
+
+static int tda998x_bind(struct device *dev, struct device *master, void *data)
+{
+	struct tda998x_encoder_params *params = dev->platform_data;
+	struct i2c_client *client = to_i2c_client(dev);
+	struct drm_device *drm = data;
+	struct tda998x_priv2 *priv;
+	int ret;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	dev_set_drvdata(dev, priv);
+
+	priv->base.encoder = &priv->encoder;
+	priv->connector.interlace_allowed = 1;
+	priv->encoder.possible_crtcs = 1 << 0;
+
+	ret = tda998x_create(client, &priv->base);
+	if (ret)
+		return ret;
+
+	if (!dev->of_node && params)
+		tda998x_encoder_set_config(&priv->base, params);
+
+	tda998x_encoder_set_polling(&priv->base, &priv->connector);
+
+	drm_encoder_helper_add(&priv->encoder, &tda998x_encoder_helper_funcs);
+	ret = drm_encoder_init(drm, &priv->encoder, &tda998x_encoder_funcs,
+			       DRM_MODE_ENCODER_TMDS);
+	if (ret)
+		goto err_encoder;
+
+	drm_connector_helper_add(&priv->connector,
+				 &tda998x_connector_helper_funcs);
+	ret = drm_connector_init(drm, &priv->connector,
+				 &tda998x_connector_funcs,
+				 DRM_MODE_CONNECTOR_HDMIA);
+	if (ret)
+		goto err_connector;
+
+	ret = drm_connector_register(&priv->connector);
+	if (ret)
+		goto err_sysfs;
+
+	priv->connector.encoder = &priv->encoder;
+	drm_mode_connector_attach_encoder(&priv->connector, &priv->encoder);
+
+	return 0;
+
+err_sysfs:
+	drm_connector_cleanup(&priv->connector);
+err_connector:
+	drm_encoder_cleanup(&priv->encoder);
+err_encoder:
+	tda998x_destroy(&priv->base);
+	return ret;
+}
+
+static void tda998x_unbind(struct device *dev, struct device *master,
+			   void *data)
+{
+	struct tda998x_priv2 *priv = dev_get_drvdata(dev);
+
+	drm_connector_cleanup(&priv->connector);
+	drm_encoder_cleanup(&priv->encoder);
+	tda998x_destroy(&priv->base);
+}
+
+static const struct component_ops tda998x_ops = {
+	.bind = tda998x_bind,
+	.unbind = tda998x_unbind,
+};
+
+static int
+tda998x_probe(struct i2c_client *client, const struct i2c_device_id *id)
+{
+	return component_add(&client->dev, &tda998x_ops);
+}
+
+static int tda998x_remove(struct i2c_client *client)
+{
+	component_del(&client->dev, &tda998x_ops);
+	return 0;
+}
+
 #ifdef CONFIG_OF
 static const struct of_device_id tda998x_dt_ids[] = {
 	{ .compatible = "nxp,tda998x", },
diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c
index e88bac1d781f..d91856779beb 100644
--- a/drivers/gpu/drm/i810/i810_dma.c
+++ b/drivers/gpu/drm/i810/i810_dma.c
@@ -213,7 +213,7 @@ static int i810_dma_cleanup(struct drm_device *dev)
 		    (drm_i810_private_t *) dev->dev_private;
 
 		if (dev_priv->ring.virtual_start)
-			drm_core_ioremapfree(&dev_priv->ring.map, dev);
+			drm_legacy_ioremapfree(&dev_priv->ring.map, dev);
 		if (dev_priv->hw_status_page) {
 			pci_free_consistent(dev->pdev, PAGE_SIZE,
 					    dev_priv->hw_status_page,
@@ -227,7 +227,7 @@ static int i810_dma_cleanup(struct drm_device *dev)
 			drm_i810_buf_priv_t *buf_priv = buf->dev_private;
 
 			if (buf_priv->kernel_virtual && buf->total)
-				drm_core_ioremapfree(&buf_priv->map, dev);
+				drm_legacy_ioremapfree(&buf_priv->map, dev);
 		}
 	}
 	return 0;
@@ -306,7 +306,7 @@ static int i810_freelist_init(struct drm_device *dev, drm_i810_private_t *dev_pr
 		buf_priv->map.flags = 0;
 		buf_priv->map.mtrr = 0;
 
-		drm_core_ioremap(&buf_priv->map, dev);
+		drm_legacy_ioremap(&buf_priv->map, dev);
 		buf_priv->kernel_virtual = buf_priv->map.handle;
 
 	}
@@ -334,7 +334,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 		DRM_ERROR("can not find sarea!\n");
 		return -EINVAL;
 	}
-	dev_priv->mmio_map = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio_map = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio_map) {
 		dev->dev_private = (void *)dev_priv;
 		i810_dma_cleanup(dev);
@@ -342,7 +342,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		dev->dev_private = (void *)dev_priv;
 		i810_dma_cleanup(dev);
@@ -363,7 +363,7 @@ static int i810_dma_initialize(struct drm_device *dev,
 	dev_priv->ring.map.flags = 0;
 	dev_priv->ring.map.mtrr = 0;
 
-	drm_core_ioremap(&dev_priv->ring.map, dev);
+	drm_legacy_ioremap(&dev_priv->ring.map, dev);
 
 	if (dev_priv->ring.map.handle == NULL) {
 		dev->dev_private = (void *)dev_priv;
@@ -393,15 +393,14 @@ static int i810_dma_initialize(struct drm_device *dev,
 
 	/* Program Hardware Status Page */
 	dev_priv->hw_status_page =
-	    pci_alloc_consistent(dev->pdev, PAGE_SIZE,
-				 &dev_priv->dma_status_page);
+		pci_zalloc_consistent(dev->pdev, PAGE_SIZE,
+				      &dev_priv->dma_status_page);
 	if (!dev_priv->hw_status_page) {
 		dev->dev_private = (void *)dev_priv;
 		i810_dma_cleanup(dev);
 		DRM_ERROR("Can not allocate hardware status page\n");
 		return -ENOMEM;
 	}
-	memset(dev_priv->hw_status_page, 0, PAGE_SIZE);
 	DRM_DEBUG("hw status page @ %p\n", dev_priv->hw_status_page);
 
 	I810_WRITE(0x02080, dev_priv->dma_status_page);
@@ -1216,9 +1215,9 @@ void i810_driver_preclose(struct drm_device *dev, struct drm_file *file_priv)
 	}
 
 	if (file_priv->master && file_priv->master->lock.hw_lock) {
-		drm_idlelock_take(&file_priv->master->lock);
+		drm_legacy_idlelock_take(&file_priv->master->lock);
 		i810_driver_reclaim_buffers(dev, file_priv);
-		drm_idlelock_release(&file_priv->master->lock);
+		drm_legacy_idlelock_release(&file_priv->master->lock);
 	} else {
 		/* master disappeared, clean up stuff anyway and hope nothing
 		 * goes wrong */
diff --git a/drivers/gpu/drm/i810/i810_drv.c b/drivers/gpu/drm/i810/i810_drv.c
index 441ccf8f5bdc..44f4a131c8dd 100644
--- a/drivers/gpu/drm/i810/i810_drv.c
+++ b/drivers/gpu/drm/i810/i810_drv.c
@@ -47,7 +47,7 @@ static const struct file_operations i810_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -63,6 +63,7 @@ static struct drm_driver driver = {
 	.load = i810_driver_load,
 	.lastclose = i810_driver_lastclose,
 	.preclose = i810_driver_preclose,
+	.set_busid = drm_pci_set_busid,
 	.device_is_agp = i810_driver_device_is_agp,
 	.dma_quiescent = i810_driver_dma_quiescent,
 	.ioctls = i810_ioctls,
diff --git a/drivers/gpu/drm/i810/i810_drv.h b/drivers/gpu/drm/i810/i810_drv.h
index d4d16eddd651..93ec5dc4e7d3 100644
--- a/drivers/gpu/drm/i810/i810_drv.h
+++ b/drivers/gpu/drm/i810/i810_drv.h
@@ -32,6 +32,8 @@
 #ifndef _I810_DRV_H_
 #define _I810_DRV_H_
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/i915/Kconfig b/drivers/gpu/drm/i915/Kconfig
index 437e1824d0bf..4e39ab34eb1c 100644
--- a/drivers/gpu/drm/i915/Kconfig
+++ b/drivers/gpu/drm/i915/Kconfig
@@ -69,15 +69,3 @@ config DRM_I915_PRELIMINARY_HW_SUPPORT
 	  option changes the default for that module option.
 
 	  If in doubt, say "N".
-
-config DRM_I915_UMS
-	bool "Enable userspace modesetting on Intel hardware (DEPRECATED)"
-	depends on DRM_I915 && BROKEN
-	default n
-	help
-	  Choose this option if you still need userspace modesetting.
-
-	  Userspace modesetting is deprecated for quite some time now, so
-	  enable this only if you have ancient versions of the DDX drivers.
-
-	  If in doubt, say "N".
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index cad1683d8bb5..c1dd485aeb6c 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -31,6 +31,7 @@ i915-y += i915_cmd_parser.o \
 	  i915_gpu_error.o \
 	  i915_irq.o \
 	  i915_trace_points.o \
+	  intel_lrc.o \
 	  intel_ringbuffer.o \
 	  intel_uncore.o
 
@@ -59,6 +60,7 @@ i915-y += dvo_ch7017.o \
 	  intel_crt.o \
 	  intel_ddi.o \
 	  intel_dp.o \
+	  intel_dp_mst.o \
 	  intel_dsi_cmd.o \
 	  intel_dsi.o \
 	  intel_dsi_pll.o \
diff --git a/drivers/gpu/drm/i915/dvo_ns2501.c b/drivers/gpu/drm/i915/dvo_ns2501.c
index 74f2af7c2d3e..441630434d34 100644
--- a/drivers/gpu/drm/i915/dvo_ns2501.c
+++ b/drivers/gpu/drm/i915/dvo_ns2501.c
@@ -60,16 +60,297 @@
 
 #define NS2501_REGC 0x0c
 
+enum {
+	MODE_640x480,
+	MODE_800x600,
+	MODE_1024x768,
+};
+
+struct ns2501_reg {
+	 uint8_t offset;
+	 uint8_t value;
+};
+
+/*
+ * Magic values based on what the BIOS on
+ * Fujitsu-Siemens Lifebook S6010 programs (1024x768 panel).
+ */
+static const struct ns2501_reg regs_1024x768[][86] = {
+	[MODE_640x480] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x11, },
+		[5] = { .offset = 0x1c, .value = 0x54, },
+		[6] = { .offset = 0x1d, .value = 0x03, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0x90, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0x0f, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x16, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x02, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0xff, },
+		[20] = { .offset = 0x81, .value = 0x07, },
+		[21] = { .offset = 0x82, .value = 0x3d, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x0a, },
+		[33] = { .offset = 0x9c, .value = 0x24, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x00, },
+		[43] = { .offset = 0xb9, .value = 0xa0, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x00, },
+		[47] = { .offset = 0x11, .value = 0xa0, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x10, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x04, },
+		[67] = { .offset = 0xaa, .value = 0x70, },
+		[68] = { .offset = 0xab, .value = 0x4f, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x84, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x30, }, /* 0x31 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x05, },
+	},
+	[MODE_800x600] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x19, },
+		[5] = { .offset = 0x1c, .value = 0x64, },
+		[6] = { .offset = 0x1d, .value = 0x02, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0xd7, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0xf8, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x1a, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x73, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0x27, },
+		[20] = { .offset = 0x81, .value = 0x03, },
+		[21] = { .offset = 0x82, .value = 0x41, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x06, },
+		[33] = { .offset = 0x9c, .value = 0x23, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x30, },
+		[43] = { .offset = 0xb9, .value = 0xc8, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x20, },
+		[47] = { .offset = 0x11, .value = 0xc8, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x04, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x83, },
+		[67] = { .offset = 0xaa, .value = 0x40, },
+		[68] = { .offset = 0xab, .value = 0x32, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x80, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x30, }, /* 0x31 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x07, },
+	},
+	[MODE_1024x768] = {
+		[0] = { .offset = 0x0a, .value = 0x81, },
+		[1] = { .offset = 0x18, .value = 0x07, },
+		[2] = { .offset = 0x19, .value = 0x00, },
+		[3] = { .offset = 0x1a, .value = 0x00, },
+		[4] = { .offset = 0x1b, .value = 0x11, },
+		[5] = { .offset = 0x1c, .value = 0x54, },
+		[6] = { .offset = 0x1d, .value = 0x03, },
+		[7] = { .offset = 0x1e, .value = 0x02, },
+		[8] = { .offset = 0xf3, .value = 0x90, },
+		[9] = { .offset = 0xf9, .value = 0x00, },
+		[10] = { .offset = 0xc1, .value = 0x90, },
+		[11] = { .offset = 0xc2, .value = 0x00, },
+		[12] = { .offset = 0xc3, .value = 0x0f, },
+		[13] = { .offset = 0xc4, .value = 0x03, },
+		[14] = { .offset = 0xc5, .value = 0x16, },
+		[15] = { .offset = 0xc6, .value = 0x00, },
+		[16] = { .offset = 0xc7, .value = 0x02, },
+		[17] = { .offset = 0xc8, .value = 0x02, },
+		[18] = { .offset = 0xf4, .value = 0x00, },
+		[19] = { .offset = 0x80, .value = 0xff, },
+		[20] = { .offset = 0x81, .value = 0x07, },
+		[21] = { .offset = 0x82, .value = 0x3d, },
+		[22] = { .offset = 0x83, .value = 0x05, },
+		[23] = { .offset = 0x94, .value = 0x00, },
+		[24] = { .offset = 0x95, .value = 0x00, },
+		[25] = { .offset = 0x96, .value = 0x05, },
+		[26] = { .offset = 0x97, .value = 0x00, },
+		[27] = { .offset = 0x9a, .value = 0x88, },
+		[28] = { .offset = 0x9b, .value = 0x00, },
+		[29] = { .offset = 0x98, .value = 0x00, },
+		[30] = { .offset = 0x99, .value = 0x00, },
+		[31] = { .offset = 0xf7, .value = 0x88, },
+		[32] = { .offset = 0xf8, .value = 0x0a, },
+		[33] = { .offset = 0x9c, .value = 0x24, },
+		[34] = { .offset = 0x9d, .value = 0x00, },
+		[35] = { .offset = 0x9e, .value = 0x25, },
+		[36] = { .offset = 0x9f, .value = 0x03, },
+		[37] = { .offset = 0xa0, .value = 0x28, },
+		[38] = { .offset = 0xa1, .value = 0x01, },
+		[39] = { .offset = 0xa2, .value = 0x28, },
+		[40] = { .offset = 0xa3, .value = 0x05, },
+		[41] = { .offset = 0xb6, .value = 0x09, },
+		[42] = { .offset = 0xb8, .value = 0x00, },
+		[43] = { .offset = 0xb9, .value = 0xa0, },
+		[44] = { .offset = 0xba, .value = 0x00, },
+		[45] = { .offset = 0xbb, .value = 0x20, },
+		[46] = { .offset = 0x10, .value = 0x00, },
+		[47] = { .offset = 0x11, .value = 0xa0, },
+		[48] = { .offset = 0x12, .value = 0x02, },
+		[49] = { .offset = 0x20, .value = 0x00, },
+		[50] = { .offset = 0x22, .value = 0x00, },
+		[51] = { .offset = 0x23, .value = 0x00, },
+		[52] = { .offset = 0x24, .value = 0x00, },
+		[53] = { .offset = 0x25, .value = 0x00, },
+		[54] = { .offset = 0x8c, .value = 0x10, },
+		[55] = { .offset = 0x8d, .value = 0x02, },
+		[56] = { .offset = 0x8e, .value = 0x10, },
+		[57] = { .offset = 0x8f, .value = 0x00, },
+		[58] = { .offset = 0x90, .value = 0xff, },
+		[59] = { .offset = 0x91, .value = 0x07, },
+		[60] = { .offset = 0x92, .value = 0xa0, },
+		[61] = { .offset = 0x93, .value = 0x02, },
+		[62] = { .offset = 0xa5, .value = 0x00, },
+		[63] = { .offset = 0xa6, .value = 0x00, },
+		[64] = { .offset = 0xa7, .value = 0x00, },
+		[65] = { .offset = 0xa8, .value = 0x00, },
+		[66] = { .offset = 0xa9, .value = 0x04, },
+		[67] = { .offset = 0xaa, .value = 0x70, },
+		[68] = { .offset = 0xab, .value = 0x4f, },
+		[69] = { .offset = 0xac, .value = 0x00, },
+		[70] = { .offset = 0xa4, .value = 0x84, },
+		[71] = { .offset = 0x7e, .value = 0x18, },
+		[72] = { .offset = 0x84, .value = 0x00, },
+		[73] = { .offset = 0x85, .value = 0x00, },
+		[74] = { .offset = 0x86, .value = 0x00, },
+		[75] = { .offset = 0x87, .value = 0x00, },
+		[76] = { .offset = 0x88, .value = 0x00, },
+		[77] = { .offset = 0x89, .value = 0x00, },
+		[78] = { .offset = 0x8a, .value = 0x00, },
+		[79] = { .offset = 0x8b, .value = 0x00, },
+		[80] = { .offset = 0x26, .value = 0x00, },
+		[81] = { .offset = 0x27, .value = 0x00, },
+		[82] = { .offset = 0xad, .value = 0x00, },
+		[83] = { .offset = 0x08, .value = 0x34, }, /* 0x35 */
+		[84] = { .offset = 0x41, .value = 0x00, },
+		[85] = { .offset = 0xc0, .value = 0x01, },
+	},
+};
+
+static const struct ns2501_reg regs_init[] = {
+	[0] = { .offset = 0x35, .value = 0xff, },
+	[1] = { .offset = 0x34, .value = 0x00, },
+	[2] = { .offset = 0x08, .value = 0x30, },
+};
+
 struct ns2501_priv {
-	//I2CDevRec d;
 	bool quiet;
-	int reg_8_shadow;
-	int reg_8_set;
-	// Shadow registers for i915
-	int dvoc;
-	int pll_a;
-	int srcdim;
-	int fw_blc;
+	const struct ns2501_reg *regs;
 };
 
 #define NSPTR(d) ((NS2501Ptr)(d->DriverPrivate.ptr))
@@ -205,11 +486,9 @@ static bool ns2501_init(struct intel_dvo_device *dvo,
 		goto out;
 	}
 	ns->quiet = false;
-	ns->reg_8_set = 0;
-	ns->reg_8_shadow =
-	    NS2501_8_PD | NS2501_8_BPAS | NS2501_8_VEN | NS2501_8_HEN;
 
 	DRM_DEBUG_KMS("init ns2501 dvo controller successfully!\n");
+
 	return true;
 
 out:
@@ -242,9 +521,9 @@ static enum drm_mode_status ns2501_mode_valid(struct intel_dvo_device *dvo,
 	 * of the panel in here so we could always accept it
 	 * by disabling the scaler.
 	 */
-	if ((mode->hdisplay == 800 && mode->vdisplay == 600) ||
-	    (mode->hdisplay == 640 && mode->vdisplay == 480) ||
-	    (mode->hdisplay == 1024 && mode->vdisplay == 768)) {
+	if ((mode->hdisplay == 640 && mode->vdisplay == 480 && mode->clock == 25175) ||
+	    (mode->hdisplay == 800 && mode->vdisplay == 600 && mode->clock == 40000) ||
+	    (mode->hdisplay == 1024 && mode->vdisplay == 768 && mode->clock == 65000)) {
 		return MODE_OK;
 	} else {
 		return MODE_ONE_SIZE;	/* Is this a reasonable error? */
@@ -255,180 +534,30 @@ static void ns2501_mode_set(struct intel_dvo_device *dvo,
 			    struct drm_display_mode *mode,
 			    struct drm_display_mode *adjusted_mode)
 {
-	bool ok;
-	int retries = 10;
 	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
+	int mode_idx, i;
 
 	DRM_DEBUG_KMS
 	    ("set mode (hdisplay=%d,htotal=%d,vdisplay=%d,vtotal=%d).\n",
 	     mode->hdisplay, mode->htotal, mode->vdisplay, mode->vtotal);
 
-	/*
-	 * Where do I find the native resolution for which scaling is not required???
-	 *
-	 * First trigger the DVO on as otherwise the chip does not appear on the i2c
-	 * bus.
-	 */
-	do {
-		ok = true;
-
-		if (mode->hdisplay == 800 && mode->vdisplay == 600) {
-			/* mode 277 */
-			ns->reg_8_shadow &= ~NS2501_8_BPAS;
-			DRM_DEBUG_KMS("switching to 800x600\n");
-
-			/*
-			 * No, I do not know where this data comes from.
-			 * It is just what the video bios left in the DVO, so
-			 * I'm just copying it here over.
-			 * This also means that I cannot support any other modes
-			 * except the ones supported by the bios.
-			 */
-			ok &= ns2501_writeb(dvo, 0x11, 0xc8);	// 0xc7 also works.
-			ok &= ns2501_writeb(dvo, 0x1b, 0x19);
-			ok &= ns2501_writeb(dvo, 0x1c, 0x62);	// VBIOS left 0x64 here, but 0x62 works nicer
-			ok &= ns2501_writeb(dvo, 0x1d, 0x02);
-
-			ok &= ns2501_writeb(dvo, 0x34, 0x03);
-			ok &= ns2501_writeb(dvo, 0x35, 0xff);
+	if (mode->hdisplay == 640 && mode->vdisplay == 480)
+		mode_idx = MODE_640x480;
+	else if (mode->hdisplay == 800 && mode->vdisplay == 600)
+		mode_idx = MODE_800x600;
+	else if (mode->hdisplay == 1024 && mode->vdisplay == 768)
+		mode_idx = MODE_1024x768;
+	else
+		return;
 
-			ok &= ns2501_writeb(dvo, 0x80, 0x27);
-			ok &= ns2501_writeb(dvo, 0x81, 0x03);
-			ok &= ns2501_writeb(dvo, 0x82, 0x41);
-			ok &= ns2501_writeb(dvo, 0x83, 0x05);
+	/* Hopefully doing it every time won't hurt... */
+	for (i = 0; i < ARRAY_SIZE(regs_init); i++)
+		ns2501_writeb(dvo, regs_init[i].offset, regs_init[i].value);
 
-			ok &= ns2501_writeb(dvo, 0x8d, 0x02);
-			ok &= ns2501_writeb(dvo, 0x8e, 0x04);
-			ok &= ns2501_writeb(dvo, 0x8f, 0x00);
+	ns->regs = regs_1024x768[mode_idx];
 
-			ok &= ns2501_writeb(dvo, 0x90, 0xfe);	/* vertical. VBIOS left 0xff here, but 0xfe works better */
-			ok &= ns2501_writeb(dvo, 0x91, 0x07);
-			ok &= ns2501_writeb(dvo, 0x94, 0x00);
-			ok &= ns2501_writeb(dvo, 0x95, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x96, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x99, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9a, 0x88);
-
-			ok &= ns2501_writeb(dvo, 0x9c, 0x23);	/* Looks like first and last line of the image. */
-			ok &= ns2501_writeb(dvo, 0x9d, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9e, 0x25);
-			ok &= ns2501_writeb(dvo, 0x9f, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0xa4, 0x80);
-
-			ok &= ns2501_writeb(dvo, 0xb6, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0xb9, 0xc8);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xba, 0x00);	/* horizontal? */
-
-			ok &= ns2501_writeb(dvo, 0xc0, 0x05);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xc1, 0xd7);
-
-			ok &= ns2501_writeb(dvo, 0xc2, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc3, 0xf8);
-
-			ok &= ns2501_writeb(dvo, 0xc4, 0x03);
-			ok &= ns2501_writeb(dvo, 0xc5, 0x1a);
-
-			ok &= ns2501_writeb(dvo, 0xc6, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc7, 0x73);
-			ok &= ns2501_writeb(dvo, 0xc8, 0x02);
-
-		} else if (mode->hdisplay == 640 && mode->vdisplay == 480) {
-			/* mode 274 */
-			DRM_DEBUG_KMS("switching to 640x480\n");
-			/*
-			 * No, I do not know where this data comes from.
-			 * It is just what the video bios left in the DVO, so
-			 * I'm just copying it here over.
-			 * This also means that I cannot support any other modes
-			 * except the ones supported by the bios.
-			 */
-			ns->reg_8_shadow &= ~NS2501_8_BPAS;
-
-			ok &= ns2501_writeb(dvo, 0x11, 0xa0);
-			ok &= ns2501_writeb(dvo, 0x1b, 0x11);
-			ok &= ns2501_writeb(dvo, 0x1c, 0x54);
-			ok &= ns2501_writeb(dvo, 0x1d, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0x34, 0x03);
-			ok &= ns2501_writeb(dvo, 0x35, 0xff);
-
-			ok &= ns2501_writeb(dvo, 0x80, 0xff);
-			ok &= ns2501_writeb(dvo, 0x81, 0x07);
-			ok &= ns2501_writeb(dvo, 0x82, 0x3d);
-			ok &= ns2501_writeb(dvo, 0x83, 0x05);
-
-			ok &= ns2501_writeb(dvo, 0x8d, 0x02);
-			ok &= ns2501_writeb(dvo, 0x8e, 0x10);
-			ok &= ns2501_writeb(dvo, 0x8f, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x90, 0xff);	/* vertical */
-			ok &= ns2501_writeb(dvo, 0x91, 0x07);
-			ok &= ns2501_writeb(dvo, 0x94, 0x00);
-			ok &= ns2501_writeb(dvo, 0x95, 0x00);
-
-			ok &= ns2501_writeb(dvo, 0x96, 0x05);
-
-			ok &= ns2501_writeb(dvo, 0x99, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9a, 0x88);
-
-			ok &= ns2501_writeb(dvo, 0x9c, 0x24);
-			ok &= ns2501_writeb(dvo, 0x9d, 0x00);
-			ok &= ns2501_writeb(dvo, 0x9e, 0x25);
-			ok &= ns2501_writeb(dvo, 0x9f, 0x03);
-
-			ok &= ns2501_writeb(dvo, 0xa4, 0x84);
-
-			ok &= ns2501_writeb(dvo, 0xb6, 0x09);
-
-			ok &= ns2501_writeb(dvo, 0xb9, 0xa0);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xba, 0x00);	/* horizontal? */
-
-			ok &= ns2501_writeb(dvo, 0xc0, 0x05);	/* horizontal? */
-			ok &= ns2501_writeb(dvo, 0xc1, 0x90);
-
-			ok &= ns2501_writeb(dvo, 0xc2, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc3, 0x0f);
-
-			ok &= ns2501_writeb(dvo, 0xc4, 0x03);
-			ok &= ns2501_writeb(dvo, 0xc5, 0x16);
-
-			ok &= ns2501_writeb(dvo, 0xc6, 0x00);
-			ok &= ns2501_writeb(dvo, 0xc7, 0x02);
-			ok &= ns2501_writeb(dvo, 0xc8, 0x02);
-
-		} else if (mode->hdisplay == 1024 && mode->vdisplay == 768) {
-			/* mode 280 */
-			DRM_DEBUG_KMS("switching to 1024x768\n");
-			/*
-			 * This might or might not work, actually. I'm silently
-			 * assuming here that the native panel resolution is
-			 * 1024x768. If not, then this leaves the scaler disabled
-			 * generating a picture that is likely not the expected.
-			 *
-			 * Problem is that I do not know where to take the panel
-			 * dimensions from.
-			 *
-			 * Enable the bypass, scaling not required.
-			 *
-			 * The scaler registers are irrelevant here....
-			 *
-			 */
-			ns->reg_8_shadow |= NS2501_8_BPAS;
-			ok &= ns2501_writeb(dvo, 0x37, 0x44);
-		} else {
-			/*
-			 * Data not known. Bummer!
-			 * Hopefully, the code should not go here
-			 * as mode_OK delivered no other modes.
-			 */
-			ns->reg_8_shadow |= NS2501_8_BPAS;
-		}
-		ok &= ns2501_writeb(dvo, NS2501_REG8, ns->reg_8_shadow);
-	} while (!ok && retries--);
+	for (i = 0; i < 84; i++)
+		ns2501_writeb(dvo, ns->regs[i].offset, ns->regs[i].value);
 }
 
 /* set the NS2501 power state */
@@ -439,60 +568,46 @@ static bool ns2501_get_hw_state(struct intel_dvo_device *dvo)
 	if (!ns2501_readb(dvo, NS2501_REG8, &ch))
 		return false;
 
-	if (ch & NS2501_8_PD)
-		return true;
-	else
-		return false;
+	return ch & NS2501_8_PD;
 }
 
 /* set the NS2501 power state */
 static void ns2501_dpms(struct intel_dvo_device *dvo, bool enable)
 {
-	bool ok;
-	int retries = 10;
 	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
-	unsigned char ch;
 
 	DRM_DEBUG_KMS("Trying set the dpms of the DVO to %i\n", enable);
 
-	ch = ns->reg_8_shadow;
+	if (enable) {
+		if (WARN_ON(ns->regs[83].offset != 0x08 ||
+			    ns->regs[84].offset != 0x41 ||
+			    ns->regs[85].offset != 0xc0))
+			return;
 
-	if (enable)
-		ch |= NS2501_8_PD;
-	else
-		ch &= ~NS2501_8_PD;
-
-	if (ns->reg_8_set == 0 || ns->reg_8_shadow != ch) {
-		ns->reg_8_set = 1;
-		ns->reg_8_shadow = ch;
-
-		do {
-			ok = true;
-			ok &= ns2501_writeb(dvo, NS2501_REG8, ch);
-			ok &=
-			    ns2501_writeb(dvo, 0x34,
-					  enable ? 0x03 : 0x00);
-			ok &=
-			    ns2501_writeb(dvo, 0x35,
-					  enable ? 0xff : 0x00);
-		} while (!ok && retries--);
-	}
-}
+		ns2501_writeb(dvo, 0xc0, ns->regs[85].value | 0x08);
 
-static void ns2501_dump_regs(struct intel_dvo_device *dvo)
-{
-	uint8_t val;
-
-	ns2501_readb(dvo, NS2501_FREQ_LO, &val);
-	DRM_DEBUG_KMS("NS2501_FREQ_LO: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_FREQ_HI, &val);
-	DRM_DEBUG_KMS("NS2501_FREQ_HI: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REG8, &val);
-	DRM_DEBUG_KMS("NS2501_REG8: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REG9, &val);
-	DRM_DEBUG_KMS("NS2501_REG9: 0x%02x\n", val);
-	ns2501_readb(dvo, NS2501_REGC, &val);
-	DRM_DEBUG_KMS("NS2501_REGC: 0x%02x\n", val);
+		ns2501_writeb(dvo, 0x41, ns->regs[84].value);
+
+		ns2501_writeb(dvo, 0x34, 0x01);
+		msleep(15);
+
+		ns2501_writeb(dvo, 0x08, 0x35);
+		if (!(ns->regs[83].value & NS2501_8_BPAS))
+			ns2501_writeb(dvo, 0x08, 0x31);
+		msleep(200);
+
+		ns2501_writeb(dvo, 0x34, 0x03);
+
+		ns2501_writeb(dvo, 0xc0, ns->regs[85].value);
+	} else {
+		ns2501_writeb(dvo, 0x34, 0x01);
+		msleep(200);
+
+		ns2501_writeb(dvo, 0x08, 0x34);
+		msleep(15);
+
+		ns2501_writeb(dvo, 0x34, 0x00);
+	}
 }
 
 static void ns2501_destroy(struct intel_dvo_device *dvo)
@@ -512,6 +627,5 @@ struct intel_dvo_dev_ops ns2501_ops = {
 	.mode_set = ns2501_mode_set,
 	.dpms = ns2501_dpms,
 	.get_hw_state = ns2501_get_hw_state,
-	.dump_regs = ns2501_dump_regs,
 	.destroy = ns2501_destroy,
 };
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index 9d7954366bd2..593b657d3e59 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -426,6 +426,9 @@ static const u32 gen7_render_regs[] = {
 	GEN7_SO_WRITE_OFFSET(1),
 	GEN7_SO_WRITE_OFFSET(2),
 	GEN7_SO_WRITE_OFFSET(3),
+	GEN7_L3SQCREG1,
+	GEN7_L3CNTLREG2,
+	GEN7_L3CNTLREG3,
 };
 
 static const u32 gen7_blt_regs[] = {
@@ -706,11 +709,13 @@ int i915_cmd_parser_init_ring(struct intel_engine_cs *ring)
 	BUG_ON(!validate_cmds_sorted(ring, cmd_tables, cmd_table_count));
 	BUG_ON(!validate_regs_sorted(ring));
 
-	ret = init_hash_table(ring, cmd_tables, cmd_table_count);
-	if (ret) {
-		DRM_ERROR("CMD: cmd_parser_init failed!\n");
-		fini_hash_table(ring);
-		return ret;
+	if (hash_empty(ring->cmd_hash)) {
+		ret = init_hash_table(ring, cmd_tables, cmd_table_count);
+		if (ret) {
+			DRM_ERROR("CMD: cmd_parser_init failed!\n");
+			fini_hash_table(ring);
+			return ret;
+		}
 	}
 
 	ring->needs_cmd_parser = true;
@@ -839,8 +844,6 @@ finish:
  */
 bool i915_needs_cmd_parser(struct intel_engine_cs *ring)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-
 	if (!ring->needs_cmd_parser)
 		return false;
 
@@ -849,7 +852,7 @@ bool i915_needs_cmd_parser(struct intel_engine_cs *ring)
 	 * disabled. That will cause all of the parser's PPGTT checks to
 	 * fail. For now, disable parsing when PPGTT is off.
 	 */
-	if (!dev_priv->mm.aliasing_ppgtt)
+	if (USES_PPGTT(ring->dev))
 		return false;
 
 	return (i915.enable_cmd_parser == 1);
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index b8c689202c40..063b44817e08 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -136,7 +136,7 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 		   obj->last_read_seqno,
 		   obj->last_write_seqno,
 		   obj->last_fenced_seqno,
-		   i915_cache_level_str(obj->cache_level),
+		   i915_cache_level_str(to_i915(obj->base.dev), obj->cache_level),
 		   obj->dirty ? " dirty" : "",
 		   obj->madv == I915_MADV_DONTNEED ? " purgeable" : "");
 	if (obj->base.name)
@@ -170,11 +170,13 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 	}
 	if (obj->ring != NULL)
 		seq_printf(m, " (%s)", obj->ring->name);
+	if (obj->frontbuffer_bits)
+		seq_printf(m, " (frontbuffer: 0x%03x)", obj->frontbuffer_bits);
 }
 
 static void describe_ctx(struct seq_file *m, struct intel_context *ctx)
 {
-	seq_putc(m, ctx->is_initialized ? 'I' : 'i');
+	seq_putc(m, ctx->legacy_hw_ctx.initialized ? 'I' : 'i');
 	seq_putc(m, ctx->remap_slice ? 'R' : 'r');
 	seq_putc(m, ' ');
 }
@@ -331,7 +333,7 @@ static int per_file_stats(int id, void *ptr, void *data)
 			}
 
 			ppgtt = container_of(vma->vm, struct i915_hw_ppgtt, base);
-			if (ppgtt->ctx && ppgtt->ctx->file_priv != stats->file_priv)
+			if (ppgtt->file_priv != stats->file_priv)
 				continue;
 
 			if (obj->ring) /* XXX per-vma statistic */
@@ -513,8 +515,14 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
 	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long flags;
 	struct intel_crtc *crtc;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
 
 	for_each_intel_crtc(dev, crtc) {
 		const char pipe = pipe_name(crtc->pipe);
@@ -527,6 +535,8 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 			seq_printf(m, "No flip due on pipe %c (plane %c)\n",
 				   pipe, plane);
 		} else {
+			u32 addr;
+
 			if (atomic_read(&work->pending) < INTEL_FLIP_COMPLETE) {
 				seq_printf(m, "Flip queued on pipe %c (plane %c)\n",
 					   pipe, plane);
@@ -534,28 +544,42 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 				seq_printf(m, "Flip pending (waiting for vsync) on pipe %c (plane %c)\n",
 					   pipe, plane);
 			}
+			if (work->flip_queued_ring) {
+				seq_printf(m, "Flip queued on %s at seqno %u, next seqno %u [current breadcrumb %u], completed? %d\n",
+					   work->flip_queued_ring->name,
+					   work->flip_queued_seqno,
+					   dev_priv->next_seqno,
+					   work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+					   i915_seqno_passed(work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+							     work->flip_queued_seqno));
+			} else
+				seq_printf(m, "Flip not associated with any ring\n");
+			seq_printf(m, "Flip queued on frame %d, (was ready on frame %d), now %d\n",
+				   work->flip_queued_vblank,
+				   work->flip_ready_vblank,
+				   drm_vblank_count(dev, crtc->pipe));
 			if (work->enable_stall_check)
 				seq_puts(m, "Stall check enabled, ");
 			else
 				seq_puts(m, "Stall check waiting for page flip ioctl, ");
 			seq_printf(m, "%d prepares\n", atomic_read(&work->pending));
 
-			if (work->old_fb_obj) {
-				struct drm_i915_gem_object *obj = work->old_fb_obj;
-				if (obj)
-					seq_printf(m, "Old framebuffer gtt_offset 0x%08lx\n",
-						   i915_gem_obj_ggtt_offset(obj));
-			}
+			if (INTEL_INFO(dev)->gen >= 4)
+				addr = I915_HI_DISPBASE(I915_READ(DSPSURF(crtc->plane)));
+			else
+				addr = I915_READ(DSPADDR(crtc->plane));
+			seq_printf(m, "Current scanout address 0x%08x\n", addr);
+
 			if (work->pending_flip_obj) {
-				struct drm_i915_gem_object *obj = work->pending_flip_obj;
-				if (obj)
-					seq_printf(m, "New framebuffer gtt_offset 0x%08lx\n",
-						   i915_gem_obj_ggtt_offset(obj));
+				seq_printf(m, "New framebuffer address 0x%08lx\n", (long)work->gtt_offset);
+				seq_printf(m, "MMIO update completed? %d\n",  addr == work->gtt_offset);
 			}
 		}
 		spin_unlock_irqrestore(&dev->event_lock, flags);
 	}
 
+	mutex_unlock(&dev->struct_mutex);
+
 	return 0;
 }
 
@@ -641,7 +665,6 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 	intel_runtime_pm_get(dev_priv);
 
 	if (IS_CHERRYVIEW(dev)) {
-		int i;
 		seq_printf(m, "Master Interrupt Control:\t%08x\n",
 			   I915_READ(GEN8_MASTER_IRQ));
 
@@ -653,7 +676,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(VLV_IIR_RW));
 		seq_printf(m, "Display IMR:\t%08x\n",
 			   I915_READ(VLV_IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -693,7 +716,13 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 				   i, I915_READ(GEN8_GT_IER(i)));
 		}
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
+			if (!intel_display_power_enabled(dev_priv,
+						POWER_DOMAIN_PIPE(pipe))) {
+				seq_printf(m, "Pipe %c power disabled\n",
+					   pipe_name(pipe));
+				continue;
+			}
 			seq_printf(m, "Pipe %c IMR:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(GEN8_DE_PIPE_IMR(pipe)));
@@ -734,7 +763,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(VLV_IIR_RW));
 		seq_printf(m, "Display IMR:\t%08x\n",
 			   I915_READ(VLV_IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -770,7 +799,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(IIR));
 		seq_printf(m, "Interrupt mask:      %08x\n",
 			   I915_READ(IMR));
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			seq_printf(m, "Pipe %c stat:         %08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
@@ -918,7 +947,7 @@ static ssize_t i915_error_state_read(struct file *file, char __user *userbuf,
 	ssize_t ret_count = 0;
 	int ret;
 
-	ret = i915_error_state_buf_init(&error_str, count, *pos);
+	ret = i915_error_state_buf_init(&error_str, to_i915(error_priv->dev), count, *pos);
 	if (ret)
 		return ret;
 
@@ -985,29 +1014,6 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_next_seqno_fops,
 			i915_next_seqno_get, i915_next_seqno_set,
 			"0x%llx\n");
 
-static int i915_rstdby_delays(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u16 crstanddelay;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-	intel_runtime_pm_get(dev_priv);
-
-	crstanddelay = I915_READ16(CRSTANDVID);
-
-	intel_runtime_pm_put(dev_priv);
-	mutex_unlock(&dev->struct_mutex);
-
-	seq_printf(m, "w/ctx: %d, w/o ctx: %d\n", (crstanddelay >> 8) & 0x3f, (crstanddelay & 0x3f));
-
-	return 0;
-}
-
 static int i915_frequency_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1029,7 +1035,8 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 			   MEMSTAT_VID_SHIFT);
 		seq_printf(m, "Current P-state: %d\n",
 			   (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT);
-	} else if ((IS_GEN6(dev) || IS_GEN7(dev)) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_GEN6(dev) || (IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) ||
+		   IS_BROADWELL(dev)) {
 		u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 		u32 rp_state_limits = I915_READ(GEN6_RP_STATE_LIMITS);
 		u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
@@ -1037,6 +1044,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		u32 rpstat, cagf, reqf;
 		u32 rpupei, rpcurup, rpprevup;
 		u32 rpdownei, rpcurdown, rpprevdown;
+		u32 pm_ier, pm_imr, pm_isr, pm_iir, pm_mask;
 		int max_freq;
 
 		/* RPSTAT1 is in the GT power well */
@@ -1048,7 +1056,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 
 		reqf = I915_READ(GEN6_RPNSWREQ);
 		reqf &= ~GEN6_TURBO_DISABLE;
-		if (IS_HASWELL(dev))
+		if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 			reqf >>= 24;
 		else
 			reqf >>= 25;
@@ -1065,7 +1073,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		rpdownei = I915_READ(GEN6_RP_CUR_DOWN_EI);
 		rpcurdown = I915_READ(GEN6_RP_CUR_DOWN);
 		rpprevdown = I915_READ(GEN6_RP_PREV_DOWN);
-		if (IS_HASWELL(dev))
+		if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 			cagf = (rpstat & HSW_CAGF_MASK) >> HSW_CAGF_SHIFT;
 		else
 			cagf = (rpstat & GEN6_CAGF_MASK) >> GEN6_CAGF_SHIFT;
@@ -1074,12 +1082,21 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 		mutex_unlock(&dev->struct_mutex);
 
+		if (IS_GEN6(dev) || IS_GEN7(dev)) {
+			pm_ier = I915_READ(GEN6_PMIER);
+			pm_imr = I915_READ(GEN6_PMIMR);
+			pm_isr = I915_READ(GEN6_PMISR);
+			pm_iir = I915_READ(GEN6_PMIIR);
+			pm_mask = I915_READ(GEN6_PMINTRMSK);
+		} else {
+			pm_ier = I915_READ(GEN8_GT_IER(2));
+			pm_imr = I915_READ(GEN8_GT_IMR(2));
+			pm_isr = I915_READ(GEN8_GT_ISR(2));
+			pm_iir = I915_READ(GEN8_GT_IIR(2));
+			pm_mask = I915_READ(GEN6_PMINTRMSK);
+		}
 		seq_printf(m, "PM IER=0x%08x IMR=0x%08x ISR=0x%08x IIR=0x%08x, MASK=0x%08x\n",
-			   I915_READ(GEN6_PMIER),
-			   I915_READ(GEN6_PMIMR),
-			   I915_READ(GEN6_PMISR),
-			   I915_READ(GEN6_PMIIR),
-			   I915_READ(GEN6_PMINTRMSK));
+			   pm_ier, pm_imr, pm_isr, pm_iir, pm_mask);
 		seq_printf(m, "GT_PERF_STATUS: 0x%08x\n", gt_perf_status);
 		seq_printf(m, "Render p-state ratio: %d\n",
 			   (gt_perf_status & 0xff00) >> 8);
@@ -1121,20 +1138,21 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		seq_printf(m, "Max overclocked frequency: %dMHz\n",
 			   dev_priv->rps.max_freq * GT_FREQUENCY_MULTIPLIER);
 	} else if (IS_VALLEYVIEW(dev)) {
-		u32 freq_sts, val;
+		u32 freq_sts;
 
 		mutex_lock(&dev_priv->rps.hw_lock);
 		freq_sts = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
 		seq_printf(m, "PUNIT_REG_GPU_FREQ_STS: 0x%08x\n", freq_sts);
 		seq_printf(m, "DDR freq: %d MHz\n", dev_priv->mem_freq);
 
-		val = valleyview_rps_max_freq(dev_priv);
 		seq_printf(m, "max GPU freq: %d MHz\n",
-			   vlv_gpu_freq(dev_priv, val));
+			   vlv_gpu_freq(dev_priv, dev_priv->rps.max_freq));
 
-		val = valleyview_rps_min_freq(dev_priv);
 		seq_printf(m, "min GPU freq: %d MHz\n",
-			   vlv_gpu_freq(dev_priv, val));
+			   vlv_gpu_freq(dev_priv, dev_priv->rps.min_freq));
+
+		seq_printf(m, "efficient (RPe) frequency: %d MHz\n",
+			   vlv_gpu_freq(dev_priv, dev_priv->rps.efficient_freq));
 
 		seq_printf(m, "current GPU freq: %d MHz\n",
 			   vlv_gpu_freq(dev_priv, (freq_sts >> 8) & 0xff));
@@ -1148,61 +1166,6 @@ out:
 	return ret;
 }
 
-static int i915_delayfreq_table(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 delayfreq;
-	int ret, i;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-	intel_runtime_pm_get(dev_priv);
-
-	for (i = 0; i < 16; i++) {
-		delayfreq = I915_READ(PXVFREQ_BASE + i * 4);
-		seq_printf(m, "P%02dVIDFREQ: 0x%08x (VID: %d)\n", i, delayfreq,
-			   (delayfreq & PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT);
-	}
-
-	intel_runtime_pm_put(dev_priv);
-
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
-static inline int MAP_TO_MV(int map)
-{
-	return 1250 - (map * 25);
-}
-
-static int i915_inttoext_table(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 inttoext;
-	int ret, i;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-	intel_runtime_pm_get(dev_priv);
-
-	for (i = 1; i <= 32; i++) {
-		inttoext = I915_READ(INTTOEXT_BASE_ILK + i * 4);
-		seq_printf(m, "INTTOEXT%02d: 0x%08x\n", i, inttoext);
-	}
-
-	intel_runtime_pm_put(dev_priv);
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
 static int ironlake_drpc_info(struct seq_file *m)
 {
 	struct drm_info_node *node = m->private;
@@ -1432,7 +1395,7 @@ static int i915_drpc_info(struct seq_file *m, void *unused)
 
 	if (IS_VALLEYVIEW(dev))
 		return vlv_drpc_info(m);
-	else if (IS_GEN6(dev) || IS_GEN7(dev))
+	else if (INTEL_INFO(dev)->gen >= 6)
 		return gen6_drpc_info(m);
 	else
 		return ironlake_drpc_info(m);
@@ -1500,6 +1463,47 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_fbc_fc_get(void *data, u64 *val)
+{
+	struct drm_device *dev = data;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
+		return -ENODEV;
+
+	drm_modeset_lock_all(dev);
+	*val = dev_priv->fbc.false_color;
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
+static int i915_fbc_fc_set(void *data, u64 val)
+{
+	struct drm_device *dev = data;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 reg;
+
+	if (INTEL_INFO(dev)->gen < 7 || !HAS_FBC(dev))
+		return -ENODEV;
+
+	drm_modeset_lock_all(dev);
+
+	reg = I915_READ(ILK_DPFC_CONTROL);
+	dev_priv->fbc.false_color = val;
+
+	I915_WRITE(ILK_DPFC_CONTROL, val ?
+		   (reg | FBC_CTL_FALSE_COLOR) :
+		   (reg & ~FBC_CTL_FALSE_COLOR));
+
+	drm_modeset_unlock_all(dev);
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_fbc_fc_fops,
+			i915_fbc_fc_get, i915_fbc_fc_set,
+			"%llu\n");
+
 static int i915_ips_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1513,10 +1517,17 @@ static int i915_ips_status(struct seq_file *m, void *unused)
 
 	intel_runtime_pm_get(dev_priv);
 
-	if (IS_BROADWELL(dev) || I915_READ(IPS_CTL) & IPS_ENABLE)
-		seq_puts(m, "enabled\n");
-	else
-		seq_puts(m, "disabled\n");
+	seq_printf(m, "Enabled by kernel parameter: %s\n",
+		   yesno(i915.enable_ips));
+
+	if (INTEL_INFO(dev)->gen >= 8) {
+		seq_puts(m, "Currently: unknown\n");
+	} else {
+		if (I915_READ(IPS_CTL) & IPS_ENABLE)
+			seq_puts(m, "Currently: enabled\n");
+		else
+			seq_puts(m, "Currently: disabled\n");
+	}
 
 	intel_runtime_pm_put(dev_priv);
 
@@ -1620,26 +1631,6 @@ out:
 	return ret;
 }
 
-static int i915_gfxec(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-	intel_runtime_pm_get(dev_priv);
-
-	seq_printf(m, "GFXEC: %ld\n", (unsigned long)I915_READ(0x112f4));
-	intel_runtime_pm_put(dev_priv);
-
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
 static int i915_opregion(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1677,9 +1668,6 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 
 #ifdef CONFIG_DRM_I915_FBDEV
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (ret)
-		return ret;
 
 	ifbdev = dev_priv->fbdev;
 	fb = to_intel_framebuffer(ifbdev->helper.fb);
@@ -1692,7 +1680,6 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 		   atomic_read(&fb->base.refcount.refcount));
 	describe_obj(m, fb->obj);
 	seq_putc(m, '\n');
-	mutex_unlock(&dev->mode_config.mutex);
 #endif
 
 	mutex_lock(&dev->mode_config.fb_lock);
@@ -1714,6 +1701,14 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
 	return 0;
 }
 
+static void describe_ctx_ringbuf(struct seq_file *m,
+				 struct intel_ringbuffer *ringbuf)
+{
+	seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u, last head: %d)",
+		   ringbuf->space, ringbuf->head, ringbuf->tail,
+		   ringbuf->last_retired_head);
+}
+
 static int i915_context_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = m->private;
@@ -1723,7 +1718,7 @@ static int i915_context_status(struct seq_file *m, void *unused)
 	struct intel_context *ctx;
 	int ret, i;
 
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
@@ -1740,20 +1735,172 @@ static int i915_context_status(struct seq_file *m, void *unused)
 	}
 
 	list_for_each_entry(ctx, &dev_priv->context_list, link) {
-		if (ctx->obj == NULL)
+		if (!i915.enable_execlists &&
+		    ctx->legacy_hw_ctx.rcs_state == NULL)
 			continue;
 
 		seq_puts(m, "HW context ");
 		describe_ctx(m, ctx);
-		for_each_ring(ring, dev_priv, i)
+		for_each_ring(ring, dev_priv, i) {
 			if (ring->default_context == ctx)
-				seq_printf(m, "(default context %s) ", ring->name);
+				seq_printf(m, "(default context %s) ",
+					   ring->name);
+		}
+
+		if (i915.enable_execlists) {
+			seq_putc(m, '\n');
+			for_each_ring(ring, dev_priv, i) {
+				struct drm_i915_gem_object *ctx_obj =
+					ctx->engine[i].state;
+				struct intel_ringbuffer *ringbuf =
+					ctx->engine[i].ringbuf;
+
+				seq_printf(m, "%s: ", ring->name);
+				if (ctx_obj)
+					describe_obj(m, ctx_obj);
+				if (ringbuf)
+					describe_ctx_ringbuf(m, ringbuf);
+				seq_putc(m, '\n');
+			}
+		} else {
+			describe_obj(m, ctx->legacy_hw_ctx.rcs_state);
+		}
+
+		seq_putc(m, '\n');
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int i915_dump_lrc(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	struct intel_context *ctx;
+	int ret, i;
+
+	if (!i915.enable_execlists) {
+		seq_printf(m, "Logical Ring Contexts are disabled\n");
+		return 0;
+	}
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	list_for_each_entry(ctx, &dev_priv->context_list, link) {
+		for_each_ring(ring, dev_priv, i) {
+			struct drm_i915_gem_object *ctx_obj = ctx->engine[i].state;
+
+			if (ring->default_context == ctx)
+				continue;
+
+			if (ctx_obj) {
+				struct page *page = i915_gem_object_get_page(ctx_obj, 1);
+				uint32_t *reg_state = kmap_atomic(page);
+				int j;
+
+				seq_printf(m, "CONTEXT: %s %u\n", ring->name,
+						intel_execlists_ctx_id(ctx_obj));
+
+				for (j = 0; j < 0x600 / sizeof(u32) / 4; j += 4) {
+					seq_printf(m, "\t[0x%08lx] 0x%08x 0x%08x 0x%08x 0x%08x\n",
+					i915_gem_obj_ggtt_offset(ctx_obj) + 4096 + (j * 4),
+					reg_state[j], reg_state[j + 1],
+					reg_state[j + 2], reg_state[j + 3]);
+				}
+				kunmap_atomic(reg_state);
+
+				seq_putc(m, '\n');
+			}
+		}
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
+static int i915_execlists(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	u32 status_pointer;
+	u8 read_pointer;
+	u8 write_pointer;
+	u32 status;
+	u32 ctx_id;
+	struct list_head *cursor;
+	int ring_id, i;
+	int ret;
+
+	if (!i915.enable_execlists) {
+		seq_puts(m, "Logical Ring Contexts are disabled\n");
+		return 0;
+	}
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	for_each_ring(ring, dev_priv, ring_id) {
+		struct intel_ctx_submit_request *head_req = NULL;
+		int count = 0;
+		unsigned long flags;
+
+		seq_printf(m, "%s\n", ring->name);
+
+		status = I915_READ(RING_EXECLIST_STATUS(ring));
+		ctx_id = I915_READ(RING_EXECLIST_STATUS(ring) + 4);
+		seq_printf(m, "\tExeclist status: 0x%08X, context: %u\n",
+			   status, ctx_id);
+
+		status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
+		seq_printf(m, "\tStatus pointer: 0x%08X\n", status_pointer);
+
+		read_pointer = ring->next_context_status_buffer;
+		write_pointer = status_pointer & 0x07;
+		if (read_pointer > write_pointer)
+			write_pointer += 6;
+		seq_printf(m, "\tRead pointer: 0x%08X, write pointer 0x%08X\n",
+			   read_pointer, write_pointer);
+
+		for (i = 0; i < 6; i++) {
+			status = I915_READ(RING_CONTEXT_STATUS_BUF(ring) + 8*i);
+			ctx_id = I915_READ(RING_CONTEXT_STATUS_BUF(ring) + 8*i + 4);
+
+			seq_printf(m, "\tStatus buffer %d: 0x%08X, context: %u\n",
+				   i, status, ctx_id);
+		}
+
+		spin_lock_irqsave(&ring->execlist_lock, flags);
+		list_for_each(cursor, &ring->execlist_queue)
+			count++;
+		head_req = list_first_entry_or_null(&ring->execlist_queue,
+				struct intel_ctx_submit_request, execlist_link);
+		spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+		seq_printf(m, "\t%d requests in queue\n", count);
+		if (head_req) {
+			struct drm_i915_gem_object *ctx_obj;
+
+			ctx_obj = head_req->ctx->engine[ring_id].state;
+			seq_printf(m, "\tHead request id: %u\n",
+				   intel_execlists_ctx_id(ctx_obj));
+			seq_printf(m, "\tHead request tail: %u\n",
+				   head_req->tail);
+		}
 
-		describe_obj(m, ctx->obj);
 		seq_putc(m, '\n');
 	}
 
-	mutex_unlock(&dev->mode_config.mutex);
+	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
 }
@@ -1858,12 +2005,18 @@ static int per_file_ctx(int id, void *ptr, void *data)
 {
 	struct intel_context *ctx = ptr;
 	struct seq_file *m = data;
-	struct i915_hw_ppgtt *ppgtt = ctx_to_ppgtt(ctx);
+	struct i915_hw_ppgtt *ppgtt = ctx->ppgtt;
+
+	if (!ppgtt) {
+		seq_printf(m, "  no ppgtt for context %d\n",
+			   ctx->user_handle);
+		return 0;
+	}
 
 	if (i915_gem_context_is_default(ctx))
 		seq_puts(m, "  default context:\n");
 	else
-		seq_printf(m, "  context %d:\n", ctx->id);
+		seq_printf(m, "  context %d:\n", ctx->user_handle);
 	ppgtt->debug_dump(ppgtt, m);
 
 	return 0;
@@ -1918,8 +2071,7 @@ static void gen6_ppgtt_info(struct seq_file *m, struct drm_device *dev)
 		seq_printf(m, "pd gtt offset: 0x%08x\n", ppgtt->pd_offset);
 
 		ppgtt->debug_dump(ppgtt, m);
-	} else
-		return;
+	}
 
 	list_for_each_entry_reverse(file, &dev->filelist, lhead) {
 		struct drm_i915_file_private *file_priv = file->driver_priv;
@@ -1976,17 +2128,25 @@ static int i915_edp_psr_status(struct seq_file *m, void *data)
 
 	intel_runtime_pm_get(dev_priv);
 
+	mutex_lock(&dev_priv->psr.lock);
 	seq_printf(m, "Sink_Support: %s\n", yesno(dev_priv->psr.sink_support));
 	seq_printf(m, "Source_OK: %s\n", yesno(dev_priv->psr.source_ok));
+	seq_printf(m, "Enabled: %s\n", yesno((bool)dev_priv->psr.enabled));
+	seq_printf(m, "Active: %s\n", yesno(dev_priv->psr.active));
+	seq_printf(m, "Busy frontbuffer bits: 0x%03x\n",
+		   dev_priv->psr.busy_frontbuffer_bits);
+	seq_printf(m, "Re-enable work scheduled: %s\n",
+		   yesno(work_busy(&dev_priv->psr.work.work)));
 
 	enabled = HAS_PSR(dev) &&
 		I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE;
-	seq_printf(m, "Enabled: %s\n", yesno(enabled));
+	seq_printf(m, "HW Enabled & Active bit: %s\n", yesno(enabled));
 
 	if (HAS_PSR(dev))
 		psrperf = I915_READ(EDP_PSR_PERF_CNT(dev)) &
 			EDP_PSR_PERF_CNT_MASK;
 	seq_printf(m, "Performance_Counter: %u\n", psrperf);
+	mutex_unlock(&dev_priv->psr.lock);
 
 	intel_runtime_pm_put(dev_priv);
 	return 0;
@@ -2072,7 +2232,7 @@ static int i915_pc8_status(struct seq_file *m, void *unused)
 
 	seq_printf(m, "GPU idle: %s\n", yesno(!dev_priv->mm.busy));
 	seq_printf(m, "IRQs disabled: %s\n",
-		   yesno(dev_priv->pm.irqs_disabled));
+		   yesno(!intel_irqs_enabled(dev_priv)));
 
 	return 0;
 }
@@ -2126,6 +2286,8 @@ static const char *power_domain_str(enum intel_display_power_domain domain)
 		return "VGA";
 	case POWER_DOMAIN_AUDIO:
 		return "AUDIO";
+	case POWER_DOMAIN_PLLS:
+		return "PLLS";
 	case POWER_DOMAIN_INIT:
 		return "INIT";
 	default:
@@ -2223,9 +2385,12 @@ static void intel_crtc_info(struct seq_file *m, struct intel_crtc *intel_crtc)
 	struct drm_crtc *crtc = &intel_crtc->base;
 	struct intel_encoder *intel_encoder;
 
-	seq_printf(m, "\tfb: %d, pos: %dx%d, size: %dx%d\n",
-		   crtc->primary->fb->base.id, crtc->x, crtc->y,
-		   crtc->primary->fb->width, crtc->primary->fb->height);
+	if (crtc->primary->fb)
+		seq_printf(m, "\tfb: %d, pos: %dx%d, size: %dx%d\n",
+			   crtc->primary->fb->base.id, crtc->x, crtc->y,
+			   crtc->primary->fb->width, crtc->primary->fb->height);
+	else
+		seq_puts(m, "\tprimary plane disabled\n");
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder)
 		intel_encoder_info(m, intel_crtc, intel_encoder);
 }
@@ -2287,13 +2452,15 @@ static void intel_connector_info(struct seq_file *m,
 		seq_printf(m, "\tCEA rev: %d\n",
 			   connector->display_info.cea_rev);
 	}
-	if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
-	    intel_encoder->type == INTEL_OUTPUT_EDP)
-		intel_dp_info(m, intel_connector);
-	else if (intel_encoder->type == INTEL_OUTPUT_HDMI)
-		intel_hdmi_info(m, intel_connector);
-	else if (intel_encoder->type == INTEL_OUTPUT_LVDS)
-		intel_lvds_info(m, intel_connector);
+	if (intel_encoder) {
+		if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
+		    intel_encoder->type == INTEL_OUTPUT_EDP)
+			intel_dp_info(m, intel_connector);
+		else if (intel_encoder->type == INTEL_OUTPUT_HDMI)
+			intel_hdmi_info(m, intel_connector);
+		else if (intel_encoder->type == INTEL_OUTPUT_LVDS)
+			intel_lvds_info(m, intel_connector);
+	}
 
 	seq_printf(m, "\tmodes:\n");
 	list_for_each_entry(mode, &connector->modes, head)
@@ -2347,17 +2514,17 @@ static int i915_display_info(struct seq_file *m, void *unused)
 		bool active;
 		int x, y;
 
-		seq_printf(m, "CRTC %d: pipe: %c, active: %s\n",
+		seq_printf(m, "CRTC %d: pipe: %c, active=%s (size=%dx%d)\n",
 			   crtc->base.base.id, pipe_name(crtc->pipe),
-			   yesno(crtc->active));
+			   yesno(crtc->active), crtc->config.pipe_src_w, crtc->config.pipe_src_h);
 		if (crtc->active) {
 			intel_crtc_info(m, crtc);
 
 			active = cursor_position(dev, crtc->pipe, &x, &y);
-			seq_printf(m, "\tcursor visible? %s, position (%d, %d), addr 0x%08x, active? %s\n",
+			seq_printf(m, "\tcursor visible? %s, position (%d, %d), size %dx%d, addr 0x%08x, active? %s\n",
 				   yesno(crtc->cursor_base),
-				   x, y, crtc->cursor_addr,
-				   yesno(active));
+				   x, y, crtc->cursor_width, crtc->cursor_height,
+				   crtc->cursor_addr, yesno(active));
 		}
 
 		seq_printf(m, "\tunderrun reporting: cpu=%s pch=%s \n",
@@ -2377,12 +2544,166 @@ static int i915_display_info(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_semaphore_status(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	int num_rings = hweight32(INTEL_INFO(dev)->ring_mask);
+	int i, j, ret;
+
+	if (!i915_semaphore_is_enabled(dev)) {
+		seq_puts(m, "Semaphores are disabled\n");
+		return 0;
+	}
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+	intel_runtime_pm_get(dev_priv);
+
+	if (IS_BROADWELL(dev)) {
+		struct page *page;
+		uint64_t *seqno;
+
+		page = i915_gem_object_get_page(dev_priv->semaphore_obj, 0);
+
+		seqno = (uint64_t *)kmap_atomic(page);
+		for_each_ring(ring, dev_priv, i) {
+			uint64_t offset;
+
+			seq_printf(m, "%s\n", ring->name);
+
+			seq_puts(m, "  Last signal:");
+			for (j = 0; j < num_rings; j++) {
+				offset = i * I915_NUM_RINGS + j;
+				seq_printf(m, "0x%08llx (0x%02llx) ",
+					   seqno[offset], offset * 8);
+			}
+			seq_putc(m, '\n');
+
+			seq_puts(m, "  Last wait:  ");
+			for (j = 0; j < num_rings; j++) {
+				offset = i + (j * I915_NUM_RINGS);
+				seq_printf(m, "0x%08llx (0x%02llx) ",
+					   seqno[offset], offset * 8);
+			}
+			seq_putc(m, '\n');
+
+		}
+		kunmap_atomic(seqno);
+	} else {
+		seq_puts(m, "  Last signal:");
+		for_each_ring(ring, dev_priv, i)
+			for (j = 0; j < num_rings; j++)
+				seq_printf(m, "0x%08x\n",
+					   I915_READ(ring->semaphore.mbox.signal[j]));
+		seq_putc(m, '\n');
+	}
+
+	seq_puts(m, "\nSync seqno:\n");
+	for_each_ring(ring, dev_priv, i) {
+		for (j = 0; j < num_rings; j++) {
+			seq_printf(m, "  0x%08x ", ring->semaphore.sync_seqno[j]);
+		}
+		seq_putc(m, '\n');
+	}
+	seq_putc(m, '\n');
+
+	intel_runtime_pm_put(dev_priv);
+	mutex_unlock(&dev->struct_mutex);
+	return 0;
+}
+
+static int i915_shared_dplls_info(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	drm_modeset_lock_all(dev);
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+
+		seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->name, pll->id);
+		seq_printf(m, " refcount: %i, active: %i, on: %s\n", pll->refcount,
+			   pll->active, yesno(pll->on));
+		seq_printf(m, " tracked hardware state:\n");
+		seq_printf(m, " dpll:    0x%08x\n", pll->hw_state.dpll);
+		seq_printf(m, " dpll_md: 0x%08x\n", pll->hw_state.dpll_md);
+		seq_printf(m, " fp0:     0x%08x\n", pll->hw_state.fp0);
+		seq_printf(m, " fp1:     0x%08x\n", pll->hw_state.fp1);
+		seq_printf(m, " wrpll:   0x%08x\n", pll->hw_state.wrpll);
+	}
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
+static int i915_wa_registers(struct seq_file *m, void *unused)
+{
+	int i;
+	int ret;
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	intel_runtime_pm_get(dev_priv);
+
+	seq_printf(m, "Workarounds applied: %d\n", dev_priv->num_wa_regs);
+	for (i = 0; i < dev_priv->num_wa_regs; ++i) {
+		u32 addr, mask;
+
+		addr = dev_priv->intel_wa_regs[i].addr;
+		mask = dev_priv->intel_wa_regs[i].mask;
+		dev_priv->intel_wa_regs[i].value = I915_READ(addr) | mask;
+		if (dev_priv->intel_wa_regs[i].addr)
+			seq_printf(m, "0x%X: 0x%08X, mask: 0x%08X\n",
+				   dev_priv->intel_wa_regs[i].addr,
+				   dev_priv->intel_wa_regs[i].value,
+				   dev_priv->intel_wa_regs[i].mask);
+	}
+
+	intel_runtime_pm_put(dev_priv);
+	mutex_unlock(&dev->struct_mutex);
+
+	return 0;
+}
+
 struct pipe_crc_info {
 	const char *name;
 	struct drm_device *dev;
 	enum pipe pipe;
 };
 
+static int i915_dp_mst_info(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_encoder *encoder;
+	struct intel_encoder *intel_encoder;
+	struct intel_digital_port *intel_dig_port;
+	drm_modeset_lock_all(dev);
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		intel_encoder = to_intel_encoder(encoder);
+		if (intel_encoder->type != INTEL_OUTPUT_DISPLAYPORT)
+			continue;
+		intel_dig_port = enc_to_dig_port(encoder);
+		if (!intel_dig_port->dp.can_mst)
+			continue;
+
+		drm_dp_mst_dump_topology(m, &intel_dig_port->dp.mst_mgr);
+	}
+	drm_modeset_unlock_all(dev);
+	return 0;
+}
+
 static int i915_pipe_crc_open(struct inode *inode, struct file *filep)
 {
 	struct pipe_crc_info *info = inode->i_private;
@@ -2616,8 +2937,7 @@ static int i9xx_pipe_crc_auto_source(struct drm_device *dev, enum pipe pipe,
 	*source = INTEL_PIPE_CRC_SOURCE_PIPE;
 
 	drm_modeset_lock_all(dev);
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (!encoder->base.crtc)
 			continue;
 
@@ -2849,7 +3169,60 @@ static int ilk_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
 	return 0;
 }
 
-static int ivb_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
+static void hsw_trans_edp_pipe_A_crc_wa(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc =
+		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_A]);
+
+	drm_modeset_lock_all(dev);
+	/*
+	 * If we use the eDP transcoder we need to make sure that we don't
+	 * bypass the pfit, since otherwise the pipe CRC source won't work. Only
+	 * relevant on hsw with pipe A when using the always-on power well
+	 * routing.
+	 */
+	if (crtc->config.cpu_transcoder == TRANSCODER_EDP &&
+	    !crtc->config.pch_pfit.enabled) {
+		crtc->config.pch_pfit.force_thru = true;
+
+		intel_display_power_get(dev_priv,
+					POWER_DOMAIN_PIPE_PANEL_FITTER(PIPE_A));
+
+		dev_priv->display.crtc_disable(&crtc->base);
+		dev_priv->display.crtc_enable(&crtc->base);
+	}
+	drm_modeset_unlock_all(dev);
+}
+
+static void hsw_undo_trans_edp_pipe_A_crc_wa(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc =
+		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_A]);
+
+	drm_modeset_lock_all(dev);
+	/*
+	 * If we use the eDP transcoder we need to make sure that we don't
+	 * bypass the pfit, since otherwise the pipe CRC source won't work. Only
+	 * relevant on hsw with pipe A when using the always-on power well
+	 * routing.
+	 */
+	if (crtc->config.pch_pfit.force_thru) {
+		crtc->config.pch_pfit.force_thru = false;
+
+		dev_priv->display.crtc_disable(&crtc->base);
+		dev_priv->display.crtc_enable(&crtc->base);
+
+		intel_display_power_put(dev_priv,
+					POWER_DOMAIN_PIPE_PANEL_FITTER(PIPE_A));
+	}
+	drm_modeset_unlock_all(dev);
+}
+
+static int ivb_pipe_crc_ctl_reg(struct drm_device *dev,
+				enum pipe pipe,
+				enum intel_pipe_crc_source *source,
 				uint32_t *val)
 {
 	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO)
@@ -2863,6 +3236,9 @@ static int ivb_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
 		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_SPRITE_IVB;
 		break;
 	case INTEL_PIPE_CRC_SOURCE_PF:
+		if (IS_HASWELL(dev) && pipe == PIPE_A)
+			hsw_trans_edp_pipe_A_crc_wa(dev);
+
 		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PF_IVB;
 		break;
 	case INTEL_PIPE_CRC_SOURCE_NONE:
@@ -2895,11 +3271,11 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 	else if (INTEL_INFO(dev)->gen < 5)
 		ret = i9xx_pipe_crc_ctl_reg(dev, pipe, &source, &val);
 	else if (IS_VALLEYVIEW(dev))
-		ret = vlv_pipe_crc_ctl_reg(dev,pipe, &source, &val);
+		ret = vlv_pipe_crc_ctl_reg(dev, pipe, &source, &val);
 	else if (IS_GEN5(dev) || IS_GEN6(dev))
 		ret = ilk_pipe_crc_ctl_reg(&source, &val);
 	else
-		ret = ivb_pipe_crc_ctl_reg(&source, &val);
+		ret = ivb_pipe_crc_ctl_reg(dev, pipe, &source, &val);
 
 	if (ret != 0)
 		return ret;
@@ -2929,11 +3305,16 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 	/* real source -> none transition */
 	if (source == INTEL_PIPE_CRC_SOURCE_NONE) {
 		struct intel_pipe_crc_entry *entries;
+		struct intel_crtc *crtc =
+			to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 
 		DRM_DEBUG_DRIVER("stopping CRCs for pipe %c\n",
 				 pipe_name(pipe));
 
-		intel_wait_for_vblank(dev, pipe);
+		drm_modeset_lock(&crtc->base.mutex, NULL);
+		if (crtc->active)
+			intel_wait_for_vblank(dev, pipe);
+		drm_modeset_unlock(&crtc->base.mutex);
 
 		spin_lock_irq(&pipe_crc->lock);
 		entries = pipe_crc->entries;
@@ -2946,6 +3327,8 @@ static int pipe_crc_set_source(struct drm_device *dev, enum pipe pipe,
 			g4x_undo_pipe_scramble_reset(dev, pipe);
 		else if (IS_VALLEYVIEW(dev))
 			vlv_undo_pipe_scramble_reset(dev, pipe);
+		else if (IS_HASWELL(dev) && pipe == PIPE_A)
+			hsw_undo_trans_edp_pipe_A_crc_wa(dev);
 	}
 
 	return 0;
@@ -3177,7 +3560,7 @@ static int pri_wm_latency_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
 
-	if (!HAS_PCH_SPLIT(dev))
+	if (HAS_GMCH_DISPLAY(dev))
 		return -ENODEV;
 
 	return single_open(file, pri_wm_latency_show, dev);
@@ -3187,7 +3570,7 @@ static int spr_wm_latency_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
 
-	if (!HAS_PCH_SPLIT(dev))
+	if (HAS_GMCH_DISPLAY(dev))
 		return -ENODEV;
 
 	return single_open(file, spr_wm_latency_show, dev);
@@ -3197,7 +3580,7 @@ static int cur_wm_latency_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
 
-	if (!HAS_PCH_SPLIT(dev))
+	if (HAS_GMCH_DISPLAY(dev))
 		return -ENODEV;
 
 	return single_open(file, cur_wm_latency_show, dev);
@@ -3443,9 +3826,6 @@ i915_drop_caches_set(void *data, u64 val)
 {
 	struct drm_device *dev = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_object *obj, *next;
-	struct i915_address_space *vm;
-	struct i915_vma *vma, *x;
 	int ret;
 
 	DRM_DEBUG("Dropping caches: 0x%08llx\n", val);
@@ -3465,29 +3845,11 @@ i915_drop_caches_set(void *data, u64 val)
 	if (val & (DROP_RETIRE | DROP_ACTIVE))
 		i915_gem_retire_requests(dev);
 
-	if (val & DROP_BOUND) {
-		list_for_each_entry(vm, &dev_priv->vm_list, global_link) {
-			list_for_each_entry_safe(vma, x, &vm->inactive_list,
-						 mm_list) {
-				if (vma->pin_count)
-					continue;
+	if (val & DROP_BOUND)
+		i915_gem_shrink(dev_priv, LONG_MAX, I915_SHRINK_BOUND);
 
-				ret = i915_vma_unbind(vma);
-				if (ret)
-					goto unlock;
-			}
-		}
-	}
-
-	if (val & DROP_UNBOUND) {
-		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
-					 global_list)
-			if (obj->pages_pin_count == 0) {
-				ret = i915_gem_object_put_pages(obj);
-				if (ret)
-					goto unlock;
-			}
-	}
+	if (val & DROP_UNBOUND)
+		i915_gem_shrink(dev_priv, LONG_MAX, I915_SHRINK_UNBOUND);
 
 unlock:
 	mutex_unlock(&dev->struct_mutex);
@@ -3506,7 +3868,7 @@ i915_max_freq_get(void *data, u64 *val)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
+	if (INTEL_INFO(dev)->gen < 6)
 		return -ENODEV;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -3532,7 +3894,7 @@ i915_max_freq_set(void *data, u64 val)
 	u32 rp_state_cap, hw_max, hw_min;
 	int ret;
 
-	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
+	if (INTEL_INFO(dev)->gen < 6)
 		return -ENODEV;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -3549,8 +3911,8 @@ i915_max_freq_set(void *data, u64 val)
 	if (IS_VALLEYVIEW(dev)) {
 		val = vlv_freq_opcode(dev_priv, val);
 
-		hw_max = valleyview_rps_max_freq(dev_priv);
-		hw_min = valleyview_rps_min_freq(dev_priv);
+		hw_max = dev_priv->rps.max_freq;
+		hw_min = dev_priv->rps.min_freq;
 	} else {
 		do_div(val, GT_FREQUENCY_MULTIPLIER);
 
@@ -3587,7 +3949,7 @@ i915_min_freq_get(void *data, u64 *val)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
+	if (INTEL_INFO(dev)->gen < 6)
 		return -ENODEV;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -3613,7 +3975,7 @@ i915_min_freq_set(void *data, u64 val)
 	u32 rp_state_cap, hw_max, hw_min;
 	int ret;
 
-	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
+	if (INTEL_INFO(dev)->gen < 6)
 		return -ENODEV;
 
 	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
@@ -3630,8 +3992,8 @@ i915_min_freq_set(void *data, u64 val)
 	if (IS_VALLEYVIEW(dev)) {
 		val = vlv_freq_opcode(dev_priv, val);
 
-		hw_max = valleyview_rps_max_freq(dev_priv);
-		hw_min = valleyview_rps_min_freq(dev_priv);
+		hw_max = dev_priv->rps.max_freq;
+		hw_min = dev_priv->rps.min_freq;
 	} else {
 		do_div(val, GT_FREQUENCY_MULTIPLIER);
 
@@ -3799,20 +4161,18 @@ static const struct drm_info_list i915_debugfs_list[] = {
 	{"i915_gem_hws_blt", i915_hws_info, 0, (void *)BCS},
 	{"i915_gem_hws_bsd", i915_hws_info, 0, (void *)VCS},
 	{"i915_gem_hws_vebox", i915_hws_info, 0, (void *)VECS},
-	{"i915_rstdby_delays", i915_rstdby_delays, 0},
 	{"i915_frequency_info", i915_frequency_info, 0},
-	{"i915_delayfreq_table", i915_delayfreq_table, 0},
-	{"i915_inttoext_table", i915_inttoext_table, 0},
 	{"i915_drpc_info", i915_drpc_info, 0},
 	{"i915_emon_status", i915_emon_status, 0},
 	{"i915_ring_freq_table", i915_ring_freq_table, 0},
-	{"i915_gfxec", i915_gfxec, 0},
 	{"i915_fbc_status", i915_fbc_status, 0},
 	{"i915_ips_status", i915_ips_status, 0},
 	{"i915_sr_status", i915_sr_status, 0},
 	{"i915_opregion", i915_opregion, 0},
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
 	{"i915_context_status", i915_context_status, 0},
+	{"i915_dump_lrc", i915_dump_lrc, 0},
+	{"i915_execlists", i915_execlists, 0},
 	{"i915_gen6_forcewake_count", i915_gen6_forcewake_count_info, 0},
 	{"i915_swizzle_info", i915_swizzle_info, 0},
 	{"i915_ppgtt_info", i915_ppgtt_info, 0},
@@ -3823,6 +4183,10 @@ static const struct drm_info_list i915_debugfs_list[] = {
 	{"i915_pc8_status", i915_pc8_status, 0},
 	{"i915_power_domain_info", i915_power_domain_info, 0},
 	{"i915_display_info", i915_display_info, 0},
+	{"i915_semaphore_status", i915_semaphore_status, 0},
+	{"i915_shared_dplls_info", i915_shared_dplls_info, 0},
+	{"i915_dp_mst_info", i915_dp_mst_info, 0},
+	{"i915_wa_registers", i915_wa_registers, 0},
 };
 #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list)
 
@@ -3844,6 +4208,7 @@ static const struct i915_debugfs_files {
 	{"i915_pri_wm_latency", &i915_pri_wm_latency_fops},
 	{"i915_spr_wm_latency", &i915_spr_wm_latency_fops},
 	{"i915_cur_wm_latency", &i915_cur_wm_latency_fops},
+	{"i915_fbc_false_color", &i915_fbc_fc_fops},
 };
 
 void intel_display_crc_init(struct drm_device *dev)
@@ -3851,7 +4216,7 @@ void intel_display_crc_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe];
 
 		pipe_crc->opened = false;
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index d44344140627..1403b01e8216 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -28,9 +28,11 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/async.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
+#include <drm/drm_legacy.h>
 #include "intel_drv.h"
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
@@ -138,7 +140,7 @@ static void i915_free_hws(struct drm_device *dev)
 	I915_WRITE(HWS_PGA, 0x1ffff000);
 }
 
-void i915_kernel_lost_context(struct drm_device * dev)
+void i915_kernel_lost_context(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_master_private *master_priv;
@@ -166,7 +168,7 @@ void i915_kernel_lost_context(struct drm_device * dev)
 		master_priv->sarea_priv->perf_boxes |= I915_BOX_RING_EMPTY;
 }
 
-static int i915_dma_cleanup(struct drm_device * dev)
+static int i915_dma_cleanup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
@@ -190,13 +192,13 @@ static int i915_dma_cleanup(struct drm_device * dev)
 	return 0;
 }
 
-static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init)
+static int i915_initialize(struct drm_device *dev, drm_i915_init_t *init)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
 	int ret;
 
-	master_priv->sarea = drm_getsarea(dev);
+	master_priv->sarea = drm_legacy_getsarea(dev);
 	if (master_priv->sarea) {
 		master_priv->sarea_priv = (drm_i915_sarea_t *)
 			((u8 *)master_priv->sarea->handle + init->sarea_priv_offset);
@@ -235,7 +237,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init)
 	return 0;
 }
 
-static int i915_dma_resume(struct drm_device * dev)
+static int i915_dma_resume(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring = LP_RING(dev_priv);
@@ -359,7 +361,7 @@ static int validate_cmd(int cmd)
 	return 0;
 }
 
-static int i915_emit_cmds(struct drm_device * dev, int *buffer, int dwords)
+static int i915_emit_cmds(struct drm_device *dev, int *buffer, int dwords)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i, ret;
@@ -369,6 +371,7 @@ static int i915_emit_cmds(struct drm_device * dev, int *buffer, int dwords)
 
 	for (i = 0; i < dwords;) {
 		int sz = validate_cmd(buffer[i]);
+
 		if (sz == 0 || i + sz > dwords)
 			return -EINVAL;
 		i += sz;
@@ -453,7 +456,7 @@ static void i915_emit_breadcrumb(struct drm_device *dev)
 	}
 }
 
-static int i915_dispatch_cmdbuffer(struct drm_device * dev,
+static int i915_dispatch_cmdbuffer(struct drm_device *dev,
 				   drm_i915_cmdbuffer_t *cmd,
 				   struct drm_clip_rect *cliprects,
 				   void *cmdbuf)
@@ -487,8 +490,8 @@ static int i915_dispatch_cmdbuffer(struct drm_device * dev,
 	return 0;
 }
 
-static int i915_dispatch_batchbuffer(struct drm_device * dev,
-				     drm_i915_batchbuffer_t * batch,
+static int i915_dispatch_batchbuffer(struct drm_device *dev,
+				     drm_i915_batchbuffer_t *batch,
 				     struct drm_clip_rect *cliprects)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -549,7 +552,7 @@ static int i915_dispatch_batchbuffer(struct drm_device * dev,
 	return 0;
 }
 
-static int i915_dispatch_flip(struct drm_device * dev)
+static int i915_dispatch_flip(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_master_private *master_priv =
@@ -755,7 +758,7 @@ fail_batch_free:
 	return ret;
 }
 
-static int i915_emit_irq(struct drm_device * dev)
+static int i915_emit_irq(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
@@ -781,7 +784,7 @@ static int i915_emit_irq(struct drm_device * dev)
 	return dev_priv->dri1.counter;
 }
 
-static int i915_wait_irq(struct drm_device * dev, int irq_nr)
+static int i915_wait_irq(struct drm_device *dev, int irq_nr)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv;
@@ -998,7 +1001,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = HAS_WT(dev);
 		break;
 	case I915_PARAM_HAS_ALIASING_PPGTT:
-		value = dev_priv->mm.aliasing_ppgtt || USES_FULL_PPGTT(dev);
+		value = USES_PPGTT(dev);
 		break;
 	case I915_PARAM_HAS_WAIT_TIMEOUT:
 		value = 1;
@@ -1266,6 +1269,7 @@ static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	pm_message_t pmm = { .event = PM_EVENT_SUSPEND };
+
 	if (state == VGA_SWITCHEROO_ON) {
 		pr_info("switched on\n");
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
@@ -1334,6 +1338,13 @@ static int i915_load_modeset_init(struct drm_device *dev)
 
 	intel_power_domains_init_hw(dev_priv);
 
+	/*
+	 * We enable some interrupt sources in our postinstall hooks, so mark
+	 * interrupts as enabled _before_ actually enabling them to avoid
+	 * special cases in our ordering checks.
+	 */
+	dev_priv->pm._irqs_disabled = false;
+
 	ret = drm_irq_install(dev, dev->pdev->irq);
 	if (ret)
 		goto cleanup_gem_stolen;
@@ -1346,8 +1357,6 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	if (ret)
 		goto cleanup_irq;
 
-	INIT_WORK(&dev_priv->console_resume_work, intel_console_resume);
-
 	intel_modeset_gem_init(dev);
 
 	/* Always safe in the mode setting case. */
@@ -1373,10 +1382,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	 * scanning against hotplug events. Hence do this first and ignore the
 	 * tiny window where we will loose hotplug notifactions.
 	 */
-	intel_fbdev_initial_config(dev);
-
-	/* Only enable hotplug handling once the fbdev is fully set up. */
-	dev_priv->enable_hotplug_processing = true;
+	async_schedule(intel_fbdev_initial_config, dev_priv);
 
 	drm_kms_helper_poll_init(dev);
 
@@ -1387,7 +1393,6 @@ cleanup_gem:
 	i915_gem_cleanup_ringbuffer(dev);
 	i915_gem_context_fini(dev);
 	mutex_unlock(&dev->struct_mutex);
-	WARN_ON(dev_priv->mm.aliasing_ppgtt);
 cleanup_irq:
 	drm_irq_uninstall(dev);
 cleanup_gem_stolen:
@@ -1425,15 +1430,16 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
 }
 
 #if IS_ENABLED(CONFIG_FB)
-static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
+static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 {
 	struct apertures_struct *ap;
 	struct pci_dev *pdev = dev_priv->dev->pdev;
 	bool primary;
+	int ret;
 
 	ap = alloc_apertures(1);
 	if (!ap)
-		return;
+		return -ENOMEM;
 
 	ap->ranges[0].base = dev_priv->gtt.mappable_base;
 	ap->ranges[0].size = dev_priv->gtt.mappable_end;
@@ -1441,13 +1447,16 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 	primary =
 		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
 
-	remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
+	ret = remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
 
 	kfree(ap);
+
+	return ret;
 }
 #else
-static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
+static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 {
+	return 0;
 }
 #endif
 
@@ -1492,10 +1501,11 @@ static void i915_dump_device_info(struct drm_i915_private *dev_priv)
 #define SEP_EMPTY
 #define PRINT_FLAG(name) info->name ? #name "," : ""
 #define SEP_COMMA ,
-	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x flags="
+	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x rev=0x%02x flags="
 			 DEV_INFO_FOR_EACH_FLAG(PRINT_S, SEP_EMPTY),
 			 info->gen,
 			 dev_priv->dev->pdev->device,
+			 dev_priv->dev->pdev->revision,
 			 DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_COMMA));
 #undef PRINT_S
 #undef SEP_EMPTY
@@ -1525,10 +1535,10 @@ static void intel_device_info_runtime_init(struct drm_device *dev)
 	info = (struct intel_device_info *)&dev_priv->info;
 
 	if (IS_VALLEYVIEW(dev))
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			info->num_sprites[pipe] = 2;
 	else
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			info->num_sprites[pipe] = 1;
 
 	if (i915.disable_display) {
@@ -1594,18 +1604,20 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (dev_priv == NULL)
 		return -ENOMEM;
 
-	dev->dev_private = (void *)dev_priv;
+	dev->dev_private = dev_priv;
 	dev_priv->dev = dev;
 
-	/* copy initial configuration to dev_priv->info */
+	/* Setup the write-once "constant" device info */
 	device_info = (struct intel_device_info *)&dev_priv->info;
-	*device_info = *info;
+	memcpy(device_info, info, sizeof(dev_priv->info));
+	device_info->device_id = dev->pdev->device;
 
 	spin_lock_init(&dev_priv->irq_lock);
 	spin_lock_init(&dev_priv->gpu_error.lock);
 	spin_lock_init(&dev_priv->backlight_lock);
 	spin_lock_init(&dev_priv->uncore.lock);
 	spin_lock_init(&dev_priv->mm.object_stat_lock);
+	spin_lock_init(&dev_priv->mmio_flip_lock);
 	mutex_init(&dev_priv->dpio_lock);
 	mutex_init(&dev_priv->modeset_restore_lock);
 
@@ -1664,7 +1676,11 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 			goto out_gtt;
 		}
 
-		i915_kick_out_firmware_fb(dev_priv);
+		ret = i915_kick_out_firmware_fb(dev_priv);
+		if (ret) {
+			DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
+			goto out_gtt;
+		}
 	}
 
 	pci_set_master(dev->pdev);
@@ -1717,6 +1733,13 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out_mtrrfree;
 	}
 
+	dev_priv->dp_wq = alloc_ordered_workqueue("i915-dp", 0);
+	if (dev_priv->dp_wq == NULL) {
+		DRM_ERROR("Failed to create our dp workqueue.\n");
+		ret = -ENOMEM;
+		goto out_freewq;
+	}
+
 	intel_irq_init(dev);
 	intel_uncore_sanitize(dev);
 
@@ -1792,12 +1815,14 @@ out_gem_unload:
 	intel_teardown_gmbus(dev);
 	intel_teardown_mchbar(dev);
 	pm_qos_remove_request(&dev_priv->pm_qos);
+	destroy_workqueue(dev_priv->dp_wq);
+out_freewq:
 	destroy_workqueue(dev_priv->wq);
 out_mtrrfree:
 	arch_phys_wc_del(dev_priv->gtt.mtrr);
 	io_mapping_free(dev_priv->gtt.mappable);
 out_gtt:
-	dev_priv->gtt.base.cleanup(&dev_priv->gtt.base);
+	i915_global_gtt_cleanup(dev);
 out_regs:
 	intel_uncore_fini(dev);
 	pci_iounmap(dev->pdev, dev_priv->regs);
@@ -1844,7 +1869,6 @@ int i915_driver_unload(struct drm_device *dev)
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		intel_fbdev_fini(dev);
 		intel_modeset_cleanup(dev);
-		cancel_work_sync(&dev_priv->console_resume_work);
 
 		/*
 		 * free the memory space allocated for the child device
@@ -1877,7 +1901,6 @@ int i915_driver_unload(struct drm_device *dev)
 		mutex_lock(&dev->struct_mutex);
 		i915_gem_cleanup_ringbuffer(dev);
 		i915_gem_context_fini(dev);
-		WARN_ON(dev_priv->mm.aliasing_ppgtt);
 		mutex_unlock(&dev->struct_mutex);
 		i915_gem_cleanup_stolen(dev);
 
@@ -1885,17 +1908,16 @@ int i915_driver_unload(struct drm_device *dev)
 			i915_free_hws(dev);
 	}
 
-	WARN_ON(!list_empty(&dev_priv->vm_list));
-
 	drm_vblank_cleanup(dev);
 
 	intel_teardown_gmbus(dev);
 	intel_teardown_mchbar(dev);
 
+	destroy_workqueue(dev_priv->dp_wq);
 	destroy_workqueue(dev_priv->wq);
 	pm_qos_remove_request(&dev_priv->pm_qos);
 
-	dev_priv->gtt.base.cleanup(&dev_priv->gtt.base);
+	i915_global_gtt_cleanup(dev);
 
 	intel_uncore_fini(dev);
 	if (dev_priv->regs != NULL)
@@ -1933,7 +1955,7 @@ int i915_driver_open(struct drm_device *dev, struct drm_file *file)
  * and DMA structures, since the kernel won't be using them, and clea
  * up any GEM state.
  */
-void i915_driver_lastclose(struct drm_device * dev)
+void i915_driver_lastclose(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -1954,12 +1976,15 @@ void i915_driver_lastclose(struct drm_device * dev)
 	i915_dma_cleanup(dev);
 }
 
-void i915_driver_preclose(struct drm_device * dev, struct drm_file *file_priv)
+void i915_driver_preclose(struct drm_device *dev, struct drm_file *file)
 {
 	mutex_lock(&dev->struct_mutex);
-	i915_gem_context_close(dev, file_priv);
-	i915_gem_release(dev, file_priv);
+	i915_gem_context_close(dev, file);
+	i915_gem_release(dev, file);
 	mutex_unlock(&dev->struct_mutex);
+
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		intel_modeset_preclose(dev, file);
 }
 
 void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
@@ -2031,7 +2056,7 @@ int i915_max_ioctl = ARRAY_SIZE(i915_ioctls);
  * manage the gtt, we need to claim that all intel devices are agp.  For
  * otherwise the drm core refuses to initialize the agp support code.
  */
-int i915_driver_device_is_agp(struct drm_device * dev)
+int i915_driver_device_is_agp(struct drm_device *dev)
 {
 	return 1;
 }
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 651e65e051c0..2318b4c7a8f8 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -28,6 +28,7 @@
  */
 
 #include <linux/device.h>
+#include <linux/acpi.h>
 #include <drm/drmP.h>
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
@@ -46,8 +47,6 @@ static struct drm_driver driver;
 			  PIPE_C_OFFSET, PIPE_EDP_OFFSET }, \
 	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
 			   TRANSCODER_C_OFFSET, TRANSCODER_EDP_OFFSET }, \
-	.dpll_offsets = { DPLL_A_OFFSET, DPLL_B_OFFSET }, \
-	.dpll_md_offsets = { DPLL_A_MD_OFFSET, DPLL_B_MD_OFFSET }, \
 	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET }
 
 #define GEN_CHV_PIPEOFFSETS \
@@ -55,10 +54,6 @@ static struct drm_driver driver;
 			  CHV_PIPE_C_OFFSET }, \
 	.trans_offsets = { TRANSCODER_A_OFFSET, TRANSCODER_B_OFFSET, \
 			   CHV_TRANSCODER_C_OFFSET, }, \
-	.dpll_offsets = { DPLL_A_OFFSET, DPLL_B_OFFSET, \
-			  CHV_DPLL_C_OFFSET }, \
-	.dpll_md_offsets = { DPLL_A_MD_OFFSET, DPLL_B_MD_OFFSET, \
-			     CHV_DPLL_C_MD_OFFSET }, \
 	.palette_offsets = { PALETTE_A_OFFSET, PALETTE_B_OFFSET, \
 			     CHV_PALETTE_C_OFFSET }
 
@@ -308,6 +303,7 @@ static const struct intel_device_info intel_broadwell_d_info = {
 	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
 	.has_llc = 1,
 	.has_ddi = 1,
+	.has_fpga_dbg = 1,
 	.has_fbc = 1,
 	GEN_DEFAULT_PIPEOFFSETS,
 	IVB_CURSOR_OFFSETS,
@@ -319,6 +315,7 @@ static const struct intel_device_info intel_broadwell_m_info = {
 	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING,
 	.has_llc = 1,
 	.has_ddi = 1,
+	.has_fpga_dbg = 1,
 	.has_fbc = 1,
 	GEN_DEFAULT_PIPEOFFSETS,
 	IVB_CURSOR_OFFSETS,
@@ -330,6 +327,7 @@ static const struct intel_device_info intel_broadwell_gt3d_info = {
 	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
 	.has_llc = 1,
 	.has_ddi = 1,
+	.has_fpga_dbg = 1,
 	.has_fbc = 1,
 	GEN_DEFAULT_PIPEOFFSETS,
 	IVB_CURSOR_OFFSETS,
@@ -341,6 +339,7 @@ static const struct intel_device_info intel_broadwell_gt3m_info = {
 	.ring_mask = RENDER_RING | BSD_RING | BLT_RING | VEBOX_RING | BSD2_RING,
 	.has_llc = 1,
 	.has_ddi = 1,
+	.has_fpga_dbg = 1,
 	.has_fbc = 1,
 	GEN_DEFAULT_PIPEOFFSETS,
 	IVB_CURSOR_OFFSETS,
@@ -482,6 +481,10 @@ bool i915_semaphore_is_enabled(struct drm_device *dev)
 	if (i915.semaphores >= 0)
 		return i915.semaphores;
 
+	/* TODO: make semaphores and Execlists play nicely together */
+	if (i915.enable_execlists)
+		return false;
+
 	/* Until we get further testing... */
 	if (IS_GEN8(dev))
 		return false;
@@ -495,12 +498,45 @@ bool i915_semaphore_is_enabled(struct drm_device *dev)
 	return true;
 }
 
+void intel_hpd_cancel_work(struct drm_i915_private *dev_priv)
+{
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	dev_priv->long_hpd_port_mask = 0;
+	dev_priv->short_hpd_port_mask = 0;
+	dev_priv->hpd_event_bits = 0;
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	cancel_work_sync(&dev_priv->dig_port_work);
+	cancel_work_sync(&dev_priv->hotplug_work);
+	cancel_delayed_work_sync(&dev_priv->hotplug_reenable_work);
+}
+
+static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct drm_encoder *encoder;
+
+	drm_modeset_lock_all(dev);
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
+
+		if (intel_encoder->suspend)
+			intel_encoder->suspend(intel_encoder);
+	}
+	drm_modeset_unlock_all(dev);
+}
+
+static int intel_suspend_complete(struct drm_i915_private *dev_priv);
+static int intel_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume);
+
 static int i915_drm_freeze(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc;
-
-	intel_runtime_pm_get(dev_priv);
+	pci_power_t opregion_target_state;
 
 	/* ignore lid events during suspend */
 	mutex_lock(&dev_priv->modeset_restore_lock);
@@ -526,21 +562,26 @@ static int i915_drm_freeze(struct drm_device *dev)
 			return error;
 		}
 
-		drm_irq_uninstall(dev);
-		dev_priv->enable_hotplug_processing = false;
-
-		intel_disable_gt_powersave(dev);
-
 		/*
 		 * Disable CRTCs directly since we want to preserve sw state
-		 * for _thaw.
+		 * for _thaw. Also, power gate the CRTC power wells.
 		 */
 		drm_modeset_lock_all(dev);
-		for_each_crtc(dev, crtc) {
-			dev_priv->display.crtc_disable(crtc);
-		}
+		for_each_crtc(dev, crtc)
+			intel_crtc_control(crtc, false);
 		drm_modeset_unlock_all(dev);
 
+		intel_dp_mst_suspend(dev);
+
+		flush_delayed_work(&dev_priv->rps.delayed_resume_work);
+
+		intel_runtime_pm_disable_interrupts(dev);
+		intel_hpd_cancel_work(dev_priv);
+
+		intel_suspend_encoders(dev_priv);
+
+		intel_suspend_gt_powersave(dev);
+
 		intel_modeset_suspend_hw(dev);
 	}
 
@@ -548,15 +589,22 @@ static int i915_drm_freeze(struct drm_device *dev)
 
 	i915_save_state(dev);
 
+	opregion_target_state = PCI_D3cold;
+#if IS_ENABLED(CONFIG_ACPI_SLEEP)
+	if (acpi_target_system_state() < ACPI_STATE_S3)
+		opregion_target_state = PCI_D1;
+#endif
+	intel_opregion_notify_adapter(dev, opregion_target_state);
+
+	intel_uncore_forcewake_reset(dev, false);
 	intel_opregion_fini(dev);
-	intel_uncore_fini(dev);
 
-	console_lock();
-	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED);
-	console_unlock();
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED, true);
 
 	dev_priv->suspend_count++;
 
+	intel_display_set_init_power(dev_priv, false);
+
 	return 0;
 }
 
@@ -590,27 +638,20 @@ int i915_suspend(struct drm_device *dev, pm_message_t state)
 	return 0;
 }
 
-void intel_console_resume(struct work_struct *work)
-{
-	struct drm_i915_private *dev_priv =
-		container_of(work, struct drm_i915_private,
-			     console_resume_work);
-	struct drm_device *dev = dev_priv->dev;
-
-	console_lock();
-	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
-	console_unlock();
-}
-
 static int i915_drm_thaw_early(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
 
-	intel_uncore_early_sanitize(dev);
+	ret = intel_resume_prepare(dev_priv, false);
+	if (ret)
+		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
+
+	intel_uncore_early_sanitize(dev, true);
 	intel_uncore_sanitize(dev);
 	intel_power_domains_init_hw(dev_priv);
 
-	return 0;
+	return ret;
 }
 
 static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
@@ -639,11 +680,19 @@ static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
 		}
 		mutex_unlock(&dev->struct_mutex);
 
-		/* We need working interrupts for modeset enabling ... */
-		drm_irq_install(dev, dev->pdev->irq);
+		intel_runtime_pm_restore_interrupts(dev);
 
 		intel_modeset_init_hw(dev);
 
+		{
+			unsigned long irqflags;
+			spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+			if (dev_priv->display.hpd_irq_setup)
+				dev_priv->display.hpd_irq_setup(dev);
+			spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		}
+
+		intel_dp_mst_resume(dev);
 		drm_modeset_lock_all(dev);
 		intel_modeset_setup_hw_state(dev, true);
 		drm_modeset_unlock_all(dev);
@@ -655,30 +704,20 @@ static int __i915_drm_thaw(struct drm_device *dev, bool restore_gtt_mappings)
 		 * notifications.
 		 * */
 		intel_hpd_init(dev);
-		dev_priv->enable_hotplug_processing = true;
 		/* Config may have changed between suspend and resume */
 		drm_helper_hpd_irq_event(dev);
 	}
 
 	intel_opregion_init(dev);
 
-	/*
-	 * The console lock can be pretty contented on resume due
-	 * to all the printk activity.  Try to keep it out of the hot
-	 * path of resume if possible.
-	 */
-	if (console_trylock()) {
-		intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
-		console_unlock();
-	} else {
-		schedule_work(&dev_priv->console_resume_work);
-	}
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING, false);
 
 	mutex_lock(&dev_priv->modeset_restore_lock);
 	dev_priv->modeset_restore = MODESET_DONE;
 	mutex_unlock(&dev_priv->modeset_restore_lock);
 
-	intel_runtime_pm_put(dev_priv);
+	intel_opregion_notify_adapter(dev, PCI_D0);
+
 	return 0;
 }
 
@@ -805,7 +844,13 @@ int i915_reset(struct drm_device *dev)
 			!dev_priv->ums.mm_suspended) {
 		dev_priv->ums.mm_suspended = 0;
 
+		/* Used to prevent gem_check_wedged returning -EAGAIN during gpu reset */
+		dev_priv->gpu_error.reload_in_reset = true;
+
 		ret = i915_gem_init_hw(dev);
+
+		dev_priv->gpu_error.reload_in_reset = false;
+
 		mutex_unlock(&dev->struct_mutex);
 		if (ret) {
 			DRM_ERROR("Failed hw init on reset %d\n", ret);
@@ -826,8 +871,6 @@ int i915_reset(struct drm_device *dev)
 		 */
 		if (INTEL_INFO(dev)->gen > 5)
 			intel_reset_gt_powersave(dev);
-
-		intel_hpd_init(dev);
 	} else {
 		mutex_unlock(&dev->struct_mutex);
 	}
@@ -887,6 +930,8 @@ static int i915_pm_suspend_late(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+	int ret;
 
 	/*
 	 * We have a suspedn ordering issue with the snd-hda driver also
@@ -900,10 +945,16 @@ static int i915_pm_suspend_late(struct device *dev)
 	if (drm_dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, PCI_D3hot);
+	ret = intel_suspend_complete(dev_priv);
 
-	return 0;
+	if (ret)
+		DRM_ERROR("Suspend complete failed: %d\n", ret);
+	else {
+		pci_disable_device(pdev);
+		pci_set_power_state(pdev, PCI_D3hot);
+	}
+
+	return ret;
 }
 
 static int i915_pm_resume_early(struct device *dev)
@@ -935,6 +986,15 @@ static int i915_pm_freeze(struct device *dev)
 	return i915_drm_freeze(drm_dev);
 }
 
+static int i915_pm_freeze_late(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct drm_i915_private *dev_priv = drm_dev->dev_private;
+
+	return intel_suspend_complete(dev_priv);
+}
+
 static int i915_pm_thaw_early(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
@@ -959,23 +1019,26 @@ static int i915_pm_poweroff(struct device *dev)
 	return i915_drm_freeze(drm_dev);
 }
 
-static int hsw_runtime_suspend(struct drm_i915_private *dev_priv)
+static int hsw_suspend_complete(struct drm_i915_private *dev_priv)
 {
 	hsw_enable_pc8(dev_priv);
 
 	return 0;
 }
 
-static int snb_runtime_resume(struct drm_i915_private *dev_priv)
+static int snb_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume)
 {
 	struct drm_device *dev = dev_priv->dev;
 
-	intel_init_pch_refclk(dev);
+	if (rpm_resume)
+		intel_init_pch_refclk(dev);
 
 	return 0;
 }
 
-static int hsw_runtime_resume(struct drm_i915_private *dev_priv)
+static int hsw_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume)
 {
 	hsw_disable_pc8(dev_priv);
 
@@ -1271,7 +1334,7 @@ static void vlv_check_no_gt_access(struct drm_i915_private *dev_priv)
 	I915_WRITE(VLV_GTLC_PW_STATUS, VLV_GTLC_ALLOWWAKEERR);
 }
 
-static int vlv_runtime_suspend(struct drm_i915_private *dev_priv)
+static int vlv_suspend_complete(struct drm_i915_private *dev_priv)
 {
 	u32 mask;
 	int err;
@@ -1311,7 +1374,8 @@ err1:
 	return err;
 }
 
-static int vlv_runtime_resume(struct drm_i915_private *dev_priv)
+static int vlv_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume)
 {
 	struct drm_device *dev = dev_priv->dev;
 	int err;
@@ -1336,8 +1400,10 @@ static int vlv_runtime_resume(struct drm_i915_private *dev_priv)
 
 	vlv_check_no_gt_access(dev_priv);
 
-	intel_init_clock_gating(dev);
-	i915_gem_restore_fences(dev);
+	if (rpm_resume) {
+		intel_init_clock_gating(dev);
+		i915_gem_restore_fences(dev);
+	}
 
 	return ret;
 }
@@ -1352,7 +1418,9 @@ static int intel_runtime_suspend(struct device *device)
 	if (WARN_ON_ONCE(!(dev_priv->rps.enabled && intel_enable_rc6(dev))))
 		return -ENODEV;
 
-	WARN_ON(!HAS_RUNTIME_PM(dev));
+	if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev)))
+		return -ENODEV;
+
 	assert_force_wake_inactive(dev_priv);
 
 	DRM_DEBUG_KMS("Suspending device\n");
@@ -1389,17 +1457,7 @@ static int intel_runtime_suspend(struct device *device)
 	cancel_work_sync(&dev_priv->rps.work);
 	intel_runtime_pm_disable_interrupts(dev);
 
-	if (IS_GEN6(dev)) {
-		ret = 0;
-	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
-		ret = hsw_runtime_suspend(dev_priv);
-	} else if (IS_VALLEYVIEW(dev)) {
-		ret = vlv_runtime_suspend(dev_priv);
-	} else {
-		ret = -ENODEV;
-		WARN_ON(1);
-	}
-
+	ret = intel_suspend_complete(dev_priv);
 	if (ret) {
 		DRM_ERROR("Runtime suspend failed, disabling it (%d)\n", ret);
 		intel_runtime_pm_restore_interrupts(dev);
@@ -1411,13 +1469,29 @@ static int intel_runtime_suspend(struct device *device)
 	dev_priv->pm.suspended = true;
 
 	/*
-	 * current versions of firmware which depend on this opregion
-	 * notification have repurposed the D1 definition to mean
-	 * "runtime suspended" vs. what you would normally expect (D3)
-	 * to distinguish it from notifications that might be sent
-	 * via the suspend path.
+	 * FIXME: We really should find a document that references the arguments
+	 * used below!
 	 */
-	intel_opregion_notify_adapter(dev, PCI_D1);
+	if (IS_HASWELL(dev)) {
+		/*
+		 * current versions of firmware which depend on this opregion
+		 * notification have repurposed the D1 definition to mean
+		 * "runtime suspended" vs. what you would normally expect (D3)
+		 * to distinguish it from notifications that might be sent via
+		 * the suspend path.
+		 */
+		intel_opregion_notify_adapter(dev, PCI_D1);
+	} else {
+		/*
+		 * On Broadwell, if we use PCI_D1 the PCH DDI ports will stop
+		 * being detected, and the call we do at intel_runtime_resume()
+		 * won't be able to restore them. Since PCI_D3hot matches the
+		 * actual specification and appears to be working, use it. Let's
+		 * assume the other non-Haswell platforms will stay the same as
+		 * Broadwell.
+		 */
+		intel_opregion_notify_adapter(dev, PCI_D3hot);
+	}
 
 	DRM_DEBUG_KMS("Device suspended\n");
 	return 0;
@@ -1430,24 +1504,15 @@ static int intel_runtime_resume(struct device *device)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	WARN_ON(!HAS_RUNTIME_PM(dev));
+	if (WARN_ON_ONCE(!HAS_RUNTIME_PM(dev)))
+		return -ENODEV;
 
 	DRM_DEBUG_KMS("Resuming device\n");
 
 	intel_opregion_notify_adapter(dev, PCI_D0);
 	dev_priv->pm.suspended = false;
 
-	if (IS_GEN6(dev)) {
-		ret = snb_runtime_resume(dev_priv);
-	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
-		ret = hsw_runtime_resume(dev_priv);
-	} else if (IS_VALLEYVIEW(dev)) {
-		ret = vlv_runtime_resume(dev_priv);
-	} else {
-		WARN_ON(1);
-		ret = -ENODEV;
-	}
-
+	ret = intel_resume_prepare(dev_priv, true);
 	/*
 	 * No point of rolling back things in case of an error, as the best
 	 * we can do is to hope that things will still work (and disable RPM).
@@ -1466,12 +1531,55 @@ static int intel_runtime_resume(struct device *device)
 	return ret;
 }
 
+/*
+ * This function implements common functionality of runtime and system
+ * suspend sequence.
+ */
+static int intel_suspend_complete(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	int ret;
+
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		ret = hsw_suspend_complete(dev_priv);
+	else if (IS_VALLEYVIEW(dev))
+		ret = vlv_suspend_complete(dev_priv);
+	else
+		ret = 0;
+
+	return ret;
+}
+
+/*
+ * This function implements common functionality of runtime and system
+ * resume sequence. Variable rpm_resume used for implementing different
+ * code paths.
+ */
+static int intel_resume_prepare(struct drm_i915_private *dev_priv,
+				bool rpm_resume)
+{
+	struct drm_device *dev = dev_priv->dev;
+	int ret;
+
+	if (IS_GEN6(dev))
+		ret = snb_resume_prepare(dev_priv, rpm_resume);
+	else if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		ret = hsw_resume_prepare(dev_priv, rpm_resume);
+	else if (IS_VALLEYVIEW(dev))
+		ret = vlv_resume_prepare(dev_priv, rpm_resume);
+	else
+		ret = 0;
+
+	return ret;
+}
+
 static const struct dev_pm_ops i915_pm_ops = {
 	.suspend = i915_pm_suspend,
 	.suspend_late = i915_pm_suspend_late,
 	.resume_early = i915_pm_resume_early,
 	.resume = i915_pm_resume,
 	.freeze = i915_pm_freeze,
+	.freeze_late = i915_pm_freeze_late,
 	.thaw_early = i915_pm_thaw_early,
 	.thaw = i915_pm_thaw,
 	.poweroff = i915_pm_poweroff,
@@ -1515,6 +1623,7 @@ static struct drm_driver driver = {
 	.lastclose = i915_driver_lastclose,
 	.preclose = i915_driver_preclose,
 	.postclose = i915_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 
 	/* Used in place of i915_pm_ops for non-DRIVER_MODESET */
 	.suspend = i915_suspend,
@@ -1606,6 +1715,8 @@ static void __exit i915_exit(void)
 module_init(i915_init);
 module_exit(i915_exit);
 
-MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_AUTHOR("Tungsten Graphics, Inc.");
+MODULE_AUTHOR("Intel Corporation");
+
 MODULE_DESCRIPTION(DRIVER_DESC);
 MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 1f7700897dfc..16a6f6d187a1 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -35,11 +35,15 @@
 #include "i915_reg.h"
 #include "intel_bios.h"
 #include "intel_ringbuffer.h"
+#include "intel_lrc.h"
 #include "i915_gem_gtt.h"
+#include "i915_gem_render_state.h"
 #include <linux/io-mapping.h>
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 #include <drm/intel-gtt.h>
+#include <drm/drm_legacy.h> /* for struct drm_dma_handle */
+#include <drm/drm_gem.h>
 #include <linux/backlight.h>
 #include <linux/hashtable.h>
 #include <linux/intel-iommu.h>
@@ -49,11 +53,9 @@
 /* General customization:
  */
 
-#define DRIVER_AUTHOR		"Tungsten Graphics, Inc."
-
 #define DRIVER_NAME		"i915"
 #define DRIVER_DESC		"Intel Graphics"
-#define DRIVER_DATE		"20080730"
+#define DRIVER_DATE		"20140905"
 
 enum pipe {
 	INVALID_PIPE = -1,
@@ -129,6 +131,7 @@ enum intel_display_power_domain {
 	POWER_DOMAIN_PORT_OTHER,
 	POWER_DOMAIN_VGA,
 	POWER_DOMAIN_AUDIO,
+	POWER_DOMAIN_PLLS,
 	POWER_DOMAIN_INIT,
 
 	POWER_DOMAIN_NUM,
@@ -161,7 +164,10 @@ enum hpd_pin {
 	 I915_GEM_DOMAIN_INSTRUCTION | \
 	 I915_GEM_DOMAIN_VERTEX)
 
-#define for_each_pipe(p) for ((p) = 0; (p) < INTEL_INFO(dev)->num_pipes; (p)++)
+#define for_each_pipe(__dev_priv, __p) \
+	for ((__p) = 0; (__p) < INTEL_INFO(__dev_priv)->num_pipes; (__p)++)
+#define for_each_plane(pipe, p) \
+	for ((p) = 0; (p) < INTEL_INFO(dev)->num_sprites[(pipe)] + 1; (p)++)
 #define for_each_sprite(p, s) for ((s) = 0; (s) < INTEL_INFO(dev)->num_sprites[(p)]; (s)++)
 
 #define for_each_crtc(dev, crtc) \
@@ -170,6 +176,11 @@ enum hpd_pin {
 #define for_each_intel_crtc(dev, intel_crtc) \
 	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list, base.head)
 
+#define for_each_intel_encoder(dev, intel_encoder)		\
+	list_for_each_entry(intel_encoder,			\
+			    &(dev)->mode_config.encoder_list,	\
+			    base.head)
+
 #define for_each_encoder_on_crtc(dev, __crtc, intel_encoder) \
 	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
 		if ((intel_encoder)->base.crtc == (__crtc))
@@ -178,22 +189,33 @@ enum hpd_pin {
 	list_for_each_entry((intel_connector), &(dev)->mode_config.connector_list, base.head) \
 		if ((intel_connector)->base.encoder == (__encoder))
 
+#define for_each_power_domain(domain, mask)				\
+	for ((domain) = 0; (domain) < POWER_DOMAIN_NUM; (domain)++)	\
+		if ((1 << (domain)) & (mask))
+
 struct drm_i915_private;
+struct i915_mm_struct;
 struct i915_mmu_object;
 
 enum intel_dpll_id {
 	DPLL_ID_PRIVATE = -1, /* non-shared dpll in use */
 	/* real shared dpll ids must be >= 0 */
-	DPLL_ID_PCH_PLL_A,
-	DPLL_ID_PCH_PLL_B,
+	DPLL_ID_PCH_PLL_A = 0,
+	DPLL_ID_PCH_PLL_B = 1,
+	DPLL_ID_WRPLL1 = 0,
+	DPLL_ID_WRPLL2 = 1,
 };
 #define I915_NUM_PLLS 2
 
 struct intel_dpll_hw_state {
+	/* i9xx, pch plls */
 	uint32_t dpll;
 	uint32_t dpll_md;
 	uint32_t fp0;
 	uint32_t fp1;
+
+	/* hsw, bdw */
+	uint32_t wrpll;
 };
 
 struct intel_shared_dpll {
@@ -204,6 +226,8 @@ struct intel_shared_dpll {
 	/* should match the index in the dev_priv->shared_dplls array */
 	enum intel_dpll_id id;
 	struct intel_dpll_hw_state hw_state;
+	/* The mode_set hook is optional and should be used together with the
+	 * intel_prepare_shared_dpll function. */
 	void (*mode_set)(struct drm_i915_private *dev_priv,
 			 struct intel_shared_dpll *pll);
 	void (*enable)(struct drm_i915_private *dev_priv,
@@ -228,12 +252,6 @@ void intel_link_compute_m_n(int bpp, int nlanes,
 			    int pixel_clock, int link_clock,
 			    struct intel_link_m_n *m_n);
 
-struct intel_ddi_plls {
-	int spll_refcount;
-	int wrpll1_refcount;
-	int wrpll2_refcount;
-};
-
 /* Interface history:
  *
  * 1.1: Original.
@@ -272,8 +290,10 @@ struct intel_opregion {
 struct intel_overlay;
 struct intel_overlay_error_state;
 
+struct drm_local_map;
+
 struct drm_i915_master_private {
-	drm_local_map_t *sarea;
+	struct drm_local_map *sarea;
 	struct _drm_i915_sarea *sarea_priv;
 };
 #define I915_FENCE_REG_NONE -1
@@ -310,6 +330,7 @@ struct drm_i915_error_state {
 	u32 eir;
 	u32 pgtbl_er;
 	u32 ier;
+	u32 gtier[4];
 	u32 ccid;
 	u32 derrmr;
 	u32 forcewake;
@@ -324,6 +345,7 @@ struct drm_i915_error_state {
 	u64 fence[I915_MAX_NUM_FENCES];
 	struct intel_overlay_error_state *overlay;
 	struct intel_display_error_state *display;
+	struct drm_i915_error_object *semaphore_obj;
 
 	struct drm_i915_error_ring {
 		bool valid;
@@ -381,6 +403,7 @@ struct drm_i915_error_state {
 		pid_t pid;
 		char comm[TASK_COMM_LEN];
 	} ring[I915_NUM_RINGS];
+
 	struct drm_i915_error_buffer {
 		u32 size;
 		u32 name;
@@ -399,6 +422,7 @@ struct drm_i915_error_state {
 	} **active_bo, **pinned_bo;
 
 	u32 *active_bo_count, *pinned_bo_count;
+	u32 vm_count;
 };
 
 struct intel_connector;
@@ -435,8 +459,8 @@ struct drm_i915_display_funcs {
 	void (*update_wm)(struct drm_crtc *crtc);
 	void (*update_sprite_wm)(struct drm_plane *plane,
 				 struct drm_crtc *crtc,
-				 uint32_t sprite_width, int pixel_size,
-				 bool enable, bool scaled);
+				 uint32_t sprite_width, uint32_t sprite_height,
+				 int pixel_size, bool enable, bool scaled);
 	void (*modeset_global_resources)(struct drm_device *dev);
 	/* Returns the active state of the crtc, and if the crtc is active,
 	 * fills out the pipe-config with the hw state. */
@@ -544,6 +568,7 @@ struct intel_uncore {
 
 struct intel_device_info {
 	u32 display_mmio_offset;
+	u16 device_id;
 	u8 num_pipes:3;
 	u8 num_sprites[I915_MAX_PIPES];
 	u8 gen;
@@ -552,8 +577,6 @@ struct intel_device_info {
 	/* Register offsets for the various display pipes and transcoders */
 	int pipe_offsets[I915_MAX_TRANSCODERS];
 	int trans_offsets[I915_MAX_TRANSCODERS];
-	int dpll_offsets[I915_MAX_PIPES];
-	int dpll_md_offsets[I915_MAX_PIPES];
 	int palette_offsets[I915_MAX_PIPES];
 	int cursor_offsets[I915_MAX_PIPES];
 };
@@ -586,30 +609,60 @@ struct i915_ctx_hang_stats {
 };
 
 /* This must match up with the value previously used for execbuf2.rsvd1. */
-#define DEFAULT_CONTEXT_ID 0
+#define DEFAULT_CONTEXT_HANDLE 0
+/**
+ * struct intel_context - as the name implies, represents a context.
+ * @ref: reference count.
+ * @user_handle: userspace tracking identity for this context.
+ * @remap_slice: l3 row remapping information.
+ * @file_priv: filp associated with this context (NULL for global default
+ *	       context).
+ * @hang_stats: information about the role of this context in possible GPU
+ *		hangs.
+ * @vm: virtual memory space used by this context.
+ * @legacy_hw_ctx: render context backing object and whether it is correctly
+ *                initialized (legacy ring submission mechanism only).
+ * @link: link in the global list of contexts.
+ *
+ * Contexts are memory images used by the hardware to store copies of their
+ * internal state.
+ */
 struct intel_context {
 	struct kref ref;
-	int id;
-	bool is_initialized;
+	int user_handle;
 	uint8_t remap_slice;
 	struct drm_i915_file_private *file_priv;
-	struct intel_engine_cs *last_ring;
-	struct drm_i915_gem_object *obj;
 	struct i915_ctx_hang_stats hang_stats;
-	struct i915_address_space *vm;
+	struct i915_hw_ppgtt *ppgtt;
+
+	/* Legacy ring buffer submission */
+	struct {
+		struct drm_i915_gem_object *rcs_state;
+		bool initialized;
+	} legacy_hw_ctx;
+
+	/* Execlists */
+	bool rcs_initialized;
+	struct {
+		struct drm_i915_gem_object *state;
+		struct intel_ringbuffer *ringbuf;
+	} engine[I915_NUM_RINGS];
 
 	struct list_head link;
 };
 
 struct i915_fbc {
 	unsigned long size;
+	unsigned threshold;
 	unsigned int fb_id;
 	enum plane plane;
 	int y;
 
-	struct drm_mm_node *compressed_fb;
+	struct drm_mm_node compressed_fb;
 	struct drm_mm_node *compressed_llb;
 
+	bool false_color;
+
 	struct intel_fbc_work {
 		struct delayed_work work;
 		struct drm_crtc *crtc;
@@ -635,9 +688,15 @@ struct i915_drrs {
 	struct intel_connector *connector;
 };
 
+struct intel_dp;
 struct i915_psr {
+	struct mutex lock;
 	bool sink_support;
 	bool source_ok;
+	struct intel_dp *enabled;
+	bool active;
+	struct delayed_work work;
+	unsigned busy_frontbuffer_bits;
 };
 
 enum intel_pch {
@@ -657,6 +716,7 @@ enum intel_sbi_destination {
 #define QUIRK_LVDS_SSC_DISABLE (1<<1)
 #define QUIRK_INVERT_BRIGHTNESS (1<<2)
 #define QUIRK_BACKLIGHT_PRESENT (1<<3)
+#define QUIRK_PIPEB_FORCE (1<<4)
 
 struct intel_fbdev;
 struct intel_fbc_work;
@@ -880,6 +940,12 @@ struct vlv_s0ix_state {
 	u32 clock_gate_dis2;
 };
 
+struct intel_rps_ei {
+	u32 cz_clock;
+	u32 render_c0;
+	u32 media_c0;
+};
+
 struct intel_gen6_power_mgmt {
 	/* work and pm_iir are protected by dev_priv->irq_lock */
 	struct work_struct work;
@@ -903,6 +969,9 @@ struct intel_gen6_power_mgmt {
 	u8 efficient_freq;	/* AKA RPe. Pre-determined balanced frequency */
 	u8 rp1_freq;		/* "less than" RP0 power/freqency */
 	u8 rp0_freq;		/* Non-overclocked max frequency. */
+	u32 cz_freq;
+
+	u32 ei_interrupt_count;
 
 	int last_adj;
 	enum { LOW_POWER, BETWEEN, HIGH_POWER } power;
@@ -910,6 +979,9 @@ struct intel_gen6_power_mgmt {
 	bool enabled;
 	struct delayed_work delayed_resume_work;
 
+	/* manual wa residency calculations */
+	struct intel_rps_ei up_ei, down_ei;
+
 	/*
 	 * Protects RPS/RC6 register access and PCU communication.
 	 * Must be taken after struct_mutex if nested.
@@ -1104,6 +1176,7 @@ struct i915_gem_mm {
 };
 
 struct drm_i915_error_state_buf {
+	struct drm_i915_private *i915;
 	unsigned bytes;
 	unsigned size;
 	int err;
@@ -1176,6 +1249,9 @@ struct i915_gpu_error {
 
 	/* For missed irq/seqno simulation. */
 	unsigned int test_irq_rings;
+
+	/* Used to prevent gem_check_wedged returning -EAGAIN during gpu reset   */
+	bool reload_in_reset;
 };
 
 enum modeset_restore {
@@ -1185,6 +1261,12 @@ enum modeset_restore {
 };
 
 struct ddi_vbt_port_info {
+	/*
+	 * This is an index in the HDMI/DVI DDI buffer translation table.
+	 * The special value HDMI_LEVEL_SHIFT_UNKNOWN means the VBT didn't
+	 * populate this field.
+	 */
+#define HDMI_LEVEL_SHIFT_UNKNOWN	0xff
 	uint8_t hdmi_level_shift;
 
 	uint8_t supports_dvi:1;
@@ -1230,6 +1312,7 @@ struct intel_vbt_data {
 		u16 pwm_freq_hz;
 		bool present;
 		bool active_low_pwm;
+		u8 min_brightness;	/* min_brightness/255 of max */
 	} backlight;
 
 	/* MIPI DSI */
@@ -1299,7 +1382,7 @@ struct ilk_wm_values {
  */
 struct i915_runtime_pm {
 	bool suspended;
-	bool irqs_disabled;
+	bool _irqs_disabled;
 };
 
 enum intel_pipe_crc_source {
@@ -1332,6 +1415,17 @@ struct intel_pipe_crc {
 	wait_queue_head_t wq;
 };
 
+struct i915_frontbuffer_tracking {
+	struct mutex lock;
+
+	/*
+	 * Tracking bits for delayed frontbuffer flushing du to gpu activity or
+	 * scheduled flips.
+	 */
+	unsigned busy_bits;
+	unsigned flip_bits;
+};
+
 struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -1363,14 +1457,18 @@ struct drm_i915_private {
 
 	struct pci_dev *bridge_dev;
 	struct intel_engine_cs ring[I915_NUM_RINGS];
+	struct drm_i915_gem_object *semaphore_obj;
 	uint32_t last_seqno, next_seqno;
 
-	drm_dma_handle_t *status_page_dmah;
+	struct drm_dma_handle *status_page_dmah;
 	struct resource mch_res;
 
 	/* protects the irq masks */
 	spinlock_t irq_lock;
 
+	/* protects the mmio flip data */
+	spinlock_t mmio_flip_lock;
+
 	bool display_irqs_enabled;
 
 	/* To control wakeup latency, e.g. for irq-driven dp aux transfers. */
@@ -1390,7 +1488,6 @@ struct drm_i915_private {
 	u32 pipestat_irq_mask[I915_MAX_PIPES];
 
 	struct work_struct hotplug_work;
-	bool enable_hotplug_processing;
 	struct {
 		unsigned long hpd_last_jiffies;
 		int hpd_cnt;
@@ -1401,7 +1498,7 @@ struct drm_i915_private {
 		} hpd_mark;
 	} hpd_stats[HPD_NUM_PINS];
 	u32 hpd_event_bits;
-	struct timer_list hotplug_reenable_timer;
+	struct delayed_work hotplug_reenable_work;
 
 	struct i915_fbc fbc;
 	struct i915_drrs drrs;
@@ -1417,6 +1514,9 @@ struct drm_i915_private {
 	/* LVDS info */
 	bool no_aux_handshake;
 
+	/* protects panel power sequencer state */
+	struct mutex pps_mutex;
+
 	struct drm_i915_fence_reg fence_regs[I915_MAX_NUM_FENCES]; /* assume 965 */
 	int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
@@ -1449,9 +1549,8 @@ struct drm_i915_private {
 	struct i915_gtt gtt; /* VM representing the global address space */
 
 	struct i915_gem_mm mm;
-#if defined(CONFIG_MMU_NOTIFIER)
-	DECLARE_HASHTABLE(mmu_notifiers, 7);
-#endif
+	DECLARE_HASHTABLE(mm_structs, 7);
+	struct mutex mm_lock;
 
 	/* Kernel Modesetting */
 
@@ -1467,14 +1566,30 @@ struct drm_i915_private {
 
 	int num_shared_dpll;
 	struct intel_shared_dpll shared_dplls[I915_NUM_PLLS];
-	struct intel_ddi_plls ddi_plls;
 	int dpio_phy_iosf_port[I915_NUM_PHYS_VLV];
 
+	/*
+	 * workarounds are currently applied at different places and
+	 * changes are being done to consolidate them so exact count is
+	 * not clear at this point, use a max value for now.
+	 */
+#define I915_MAX_WA_REGS  16
+	struct {
+		u32 addr;
+		u32 value;
+		/* bitmask representing WA bits */
+		u32 mask;
+	} intel_wa_regs[I915_MAX_WA_REGS];
+	u32 num_wa_regs;
+
 	/* Reclocking support */
 	bool render_reclock_avail;
 	bool lvds_downclock_avail;
 	/* indicates the reduced downclock for LVDS*/
 	int lvds_downclock;
+
+	struct i915_frontbuffer_tracking fb_tracking;
+
 	u16 orig_clock;
 
 	bool mchbar_need_disable;
@@ -1502,14 +1617,9 @@ struct drm_i915_private {
 #ifdef CONFIG_DRM_I915_FBDEV
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
+	struct work_struct fbdev_suspend_work;
 #endif
 
-	/*
-	 * The console may be contended at resume, but we don't
-	 * want it to block on it.
-	 */
-	struct work_struct console_resume_work;
-
 	struct drm_property *broadcast_rgb_property;
 	struct drm_property *force_audio_property;
 
@@ -1541,12 +1651,42 @@ struct drm_i915_private {
 
 	struct i915_runtime_pm pm;
 
+	struct intel_digital_port *hpd_irq_port[I915_MAX_PORTS];
+	u32 long_hpd_port_mask;
+	u32 short_hpd_port_mask;
+	struct work_struct dig_port_work;
+
+	/*
+	 * if we get a HPD irq from DP and a HPD irq from non-DP
+	 * the non-DP HPD could block the workqueue on a mode config
+	 * mutex getting, that userspace may have taken. However
+	 * userspace is waiting on the DP workqueue to run which is
+	 * blocked behind the non-DP one.
+	 */
+	struct workqueue_struct *dp_wq;
+
+	uint32_t bios_vgacntr;
+
 	/* Old dri1 support infrastructure, beware the dragons ya fools entering
 	 * here! */
 	struct i915_dri1_state dri1;
 	/* Old ums support infrastructure, same warning applies. */
 	struct i915_ums_state ums;
 
+	/* Abstract the submission mechanism (legacy ringbuffer or execlists) away */
+	struct {
+		int (*do_execbuf)(struct drm_device *dev, struct drm_file *file,
+				  struct intel_engine_cs *ring,
+				  struct intel_context *ctx,
+				  struct drm_i915_gem_execbuffer2 *args,
+				  struct list_head *vmas,
+				  struct drm_i915_gem_object *batch_obj,
+				  u64 exec_start, u32 flags);
+		int (*init_rings)(struct drm_device *dev);
+		void (*cleanup_ring)(struct intel_engine_cs *ring);
+		void (*stop_ring)(struct intel_engine_cs *ring);
+	} gt;
+
 	/*
 	 * NOTE: This is the dri1/ums dungeon, don't add stuff here. Your patch
 	 * will be rejected. Instead look for a better place.
@@ -1592,6 +1732,28 @@ struct drm_i915_gem_object_ops {
 	void (*release)(struct drm_i915_gem_object *);
 };
 
+/*
+ * Frontbuffer tracking bits. Set in obj->frontbuffer_bits while a gem bo is
+ * considered to be the frontbuffer for the given plane interface-vise. This
+ * doesn't mean that the hw necessarily already scans it out, but that any
+ * rendering (by the cpu or gpu) will land in the frontbuffer eventually.
+ *
+ * We have one bit per pipe and per scanout plane type.
+ */
+#define INTEL_FRONTBUFFER_BITS_PER_PIPE 4
+#define INTEL_FRONTBUFFER_BITS \
+	(INTEL_FRONTBUFFER_BITS_PER_PIPE * I915_MAX_PIPES)
+#define INTEL_FRONTBUFFER_PRIMARY(pipe) \
+	(1 << (INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe)))
+#define INTEL_FRONTBUFFER_CURSOR(pipe) \
+	(1 << (1 +(INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe))))
+#define INTEL_FRONTBUFFER_SPRITE(pipe) \
+	(1 << (2 +(INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe))))
+#define INTEL_FRONTBUFFER_OVERLAY(pipe) \
+	(1 << (3 +(INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe))))
+#define INTEL_FRONTBUFFER_ALL_MASK(pipe) \
+	(0xf << (INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe)))
+
 struct drm_i915_gem_object {
 	struct drm_gem_object base;
 
@@ -1662,17 +1824,18 @@ struct drm_i915_gem_object {
 	unsigned int pin_display:1;
 
 	/*
-	 * Is the GPU currently using a fence to access this buffer,
+	 * Is the object to be mapped as read-only to the GPU
+	 * Only honoured if hardware has relevant pte bit
 	 */
-	unsigned int pending_fenced_gpu_access:1;
-	unsigned int fenced_gpu_access:1;
-
+	unsigned long gt_ro:1;
 	unsigned int cache_level:3;
 
 	unsigned int has_aliasing_ppgtt_mapping:1;
 	unsigned int has_global_gtt_mapping:1;
 	unsigned int has_dma_mapping:1;
 
+	unsigned int frontbuffer_bits:INTEL_FRONTBUFFER_BITS;
+
 	struct sg_table *pages;
 	int pages_pin_count;
 
@@ -1702,7 +1865,7 @@ struct drm_i915_gem_object {
 	struct drm_file *pin_filp;
 
 	/** for phy allocated objects */
-	drm_dma_handle_t *phys_handle;
+	struct drm_dma_handle *phys_handle;
 
 	union {
 		struct i915_gem_userptr {
@@ -1711,14 +1874,18 @@ struct drm_i915_gem_object {
 			unsigned workers :4;
 #define I915_GEM_USERPTR_MAX_WORKERS 15
 
-			struct mm_struct *mm;
-			struct i915_mmu_object *mn;
+			struct i915_mm_struct *mm;
+			struct i915_mmu_object *mmu_object;
 			struct work_struct *work;
 		} userptr;
 	};
 };
 #define to_intel_bo(x) container_of(x, struct drm_i915_gem_object, base)
 
+void i915_gem_track_fb(struct drm_i915_gem_object *old,
+		       struct drm_i915_gem_object *new,
+		       unsigned frontbuffer_bits);
+
 /**
  * Request queue structure.
  *
@@ -1864,51 +2031,63 @@ struct drm_i915_cmd_table {
 	int count;
 };
 
-#define INTEL_INFO(dev)	(&to_i915(dev)->info)
-
-#define IS_I830(dev)		((dev)->pdev->device == 0x3577)
-#define IS_845G(dev)		((dev)->pdev->device == 0x2562)
+/* Note that the (struct drm_i915_private *) cast is just to shut up gcc. */
+#define __I915__(p) ({ \
+	struct drm_i915_private *__p; \
+	if (__builtin_types_compatible_p(typeof(*p), struct drm_i915_private)) \
+		__p = (struct drm_i915_private *)p; \
+	else if (__builtin_types_compatible_p(typeof(*p), struct drm_device)) \
+		__p = to_i915((struct drm_device *)p); \
+	else \
+		BUILD_BUG(); \
+	__p; \
+})
+#define INTEL_INFO(p) 	(&__I915__(p)->info)
+#define INTEL_DEVID(p)	(INTEL_INFO(p)->device_id)
+
+#define IS_I830(dev)		(INTEL_DEVID(dev) == 0x3577)
+#define IS_845G(dev)		(INTEL_DEVID(dev) == 0x2562)
 #define IS_I85X(dev)		(INTEL_INFO(dev)->is_i85x)
-#define IS_I865G(dev)		((dev)->pdev->device == 0x2572)
+#define IS_I865G(dev)		(INTEL_DEVID(dev) == 0x2572)
 #define IS_I915G(dev)		(INTEL_INFO(dev)->is_i915g)
-#define IS_I915GM(dev)		((dev)->pdev->device == 0x2592)
-#define IS_I945G(dev)		((dev)->pdev->device == 0x2772)
+#define IS_I915GM(dev)		(INTEL_DEVID(dev) == 0x2592)
+#define IS_I945G(dev)		(INTEL_DEVID(dev) == 0x2772)
 #define IS_I945GM(dev)		(INTEL_INFO(dev)->is_i945gm)
 #define IS_BROADWATER(dev)	(INTEL_INFO(dev)->is_broadwater)
 #define IS_CRESTLINE(dev)	(INTEL_INFO(dev)->is_crestline)
-#define IS_GM45(dev)		((dev)->pdev->device == 0x2A42)
+#define IS_GM45(dev)		(INTEL_DEVID(dev) == 0x2A42)
 #define IS_G4X(dev)		(INTEL_INFO(dev)->is_g4x)
-#define IS_PINEVIEW_G(dev)	((dev)->pdev->device == 0xa001)
-#define IS_PINEVIEW_M(dev)	((dev)->pdev->device == 0xa011)
+#define IS_PINEVIEW_G(dev)	(INTEL_DEVID(dev) == 0xa001)
+#define IS_PINEVIEW_M(dev)	(INTEL_DEVID(dev) == 0xa011)
 #define IS_PINEVIEW(dev)	(INTEL_INFO(dev)->is_pineview)
 #define IS_G33(dev)		(INTEL_INFO(dev)->is_g33)
-#define IS_IRONLAKE_M(dev)	((dev)->pdev->device == 0x0046)
+#define IS_IRONLAKE_M(dev)	(INTEL_DEVID(dev) == 0x0046)
 #define IS_IVYBRIDGE(dev)	(INTEL_INFO(dev)->is_ivybridge)
-#define IS_IVB_GT1(dev)		((dev)->pdev->device == 0x0156 || \
-				 (dev)->pdev->device == 0x0152 || \
-				 (dev)->pdev->device == 0x015a)
-#define IS_SNB_GT1(dev)		((dev)->pdev->device == 0x0102 || \
-				 (dev)->pdev->device == 0x0106 || \
-				 (dev)->pdev->device == 0x010A)
+#define IS_IVB_GT1(dev)		(INTEL_DEVID(dev) == 0x0156 || \
+				 INTEL_DEVID(dev) == 0x0152 || \
+				 INTEL_DEVID(dev) == 0x015a)
+#define IS_SNB_GT1(dev)		(INTEL_DEVID(dev) == 0x0102 || \
+				 INTEL_DEVID(dev) == 0x0106 || \
+				 INTEL_DEVID(dev) == 0x010A)
 #define IS_VALLEYVIEW(dev)	(INTEL_INFO(dev)->is_valleyview)
 #define IS_CHERRYVIEW(dev)	(INTEL_INFO(dev)->is_valleyview && IS_GEN8(dev))
 #define IS_HASWELL(dev)	(INTEL_INFO(dev)->is_haswell)
 #define IS_BROADWELL(dev)	(!INTEL_INFO(dev)->is_valleyview && IS_GEN8(dev))
 #define IS_MOBILE(dev)		(INTEL_INFO(dev)->is_mobile)
 #define IS_HSW_EARLY_SDV(dev)	(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0xFF00) == 0x0C00)
+				 (INTEL_DEVID(dev) & 0xFF00) == 0x0C00)
 #define IS_BDW_ULT(dev)		(IS_BROADWELL(dev) && \
-				 (((dev)->pdev->device & 0xf) == 0x2  || \
-				 ((dev)->pdev->device & 0xf) == 0x6 || \
-				 ((dev)->pdev->device & 0xf) == 0xe))
+				 ((INTEL_DEVID(dev) & 0xf) == 0x2  || \
+				 (INTEL_DEVID(dev) & 0xf) == 0x6 || \
+				 (INTEL_DEVID(dev) & 0xf) == 0xe))
 #define IS_HSW_ULT(dev)		(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0xFF00) == 0x0A00)
+				 (INTEL_DEVID(dev) & 0xFF00) == 0x0A00)
 #define IS_ULT(dev)		(IS_HSW_ULT(dev) || IS_BDW_ULT(dev))
 #define IS_HSW_GT3(dev)		(IS_HASWELL(dev) && \
-				 ((dev)->pdev->device & 0x00F0) == 0x0020)
+				 (INTEL_DEVID(dev) & 0x00F0) == 0x0020)
 /* ULX machines are also considered ULT. */
-#define IS_HSW_ULX(dev)		((dev)->pdev->device == 0x0A0E || \
-				 (dev)->pdev->device == 0x0A1E)
+#define IS_HSW_ULX(dev)		(INTEL_DEVID(dev) == 0x0A0E || \
+				 INTEL_DEVID(dev) == 0x0A1E)
 #define IS_PRELIMINARY_HW(intel_info) ((intel_info)->is_preliminary)
 
 /*
@@ -1940,12 +2119,11 @@ struct drm_i915_cmd_table {
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
 #define HAS_HW_CONTEXTS(dev)	(INTEL_INFO(dev)->gen >= 6)
-#define HAS_ALIASING_PPGTT(dev)	(INTEL_INFO(dev)->gen >= 6 && \
-				 (!IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)))
-#define HAS_PPGTT(dev)		(INTEL_INFO(dev)->gen >= 7 \
-				 && !IS_GEN8(dev))
-#define USES_PPGTT(dev)		intel_enable_ppgtt(dev, false)
-#define USES_FULL_PPGTT(dev)	intel_enable_ppgtt(dev, true)
+#define HAS_LOGICAL_RING_CONTEXTS(dev)	(INTEL_INFO(dev)->gen >= 8)
+#define HAS_ALIASING_PPGTT(dev)	(INTEL_INFO(dev)->gen >= 6)
+#define HAS_PPGTT(dev)		(INTEL_INFO(dev)->gen >= 7 && !IS_GEN8(dev))
+#define USES_PPGTT(dev)		(i915.enable_ppgtt)
+#define USES_FULL_PPGTT(dev)	(i915.enable_ppgtt == 2)
 
 #define HAS_OVERLAY(dev)		(INTEL_INFO(dev)->has_overlay)
 #define OVERLAY_NEEDS_PHYSICAL(dev)	(INTEL_INFO(dev)->overlay_needs_physical)
@@ -1998,6 +2176,8 @@ struct drm_i915_cmd_table {
 #define HAS_PCH_NOP(dev) (INTEL_PCH_TYPE(dev) == PCH_NOP)
 #define HAS_PCH_SPLIT(dev) (INTEL_PCH_TYPE(dev) != PCH_NONE)
 
+#define HAS_GMCH_DISPLAY(dev) (INTEL_INFO(dev)->gen < 5 || IS_VALLEYVIEW(dev))
+
 /* DPF == dynamic parity feature */
 #define HAS_L3_DPF(dev) (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
 #define NUM_L3_SLICES(dev) (IS_HSW_GT3(dev) ? 2 : HAS_L3_DPF(dev))
@@ -2027,6 +2207,7 @@ struct i915_params {
 	int enable_rc6;
 	int enable_fbc;
 	int enable_ppgtt;
+	int enable_execlists;
 	int enable_psr;
 	unsigned int preliminary_hw_support;
 	int disable_power_well;
@@ -2040,6 +2221,8 @@ struct i915_params {
 	bool reset;
 	bool disable_display;
 	bool disable_vtd_wa;
+	int use_mmio_flip;
+	bool mmio_debug;
 };
 extern struct i915_params i915 __read_mostly;
 
@@ -2048,12 +2231,12 @@ void i915_update_dri1_breadcrumb(struct drm_device *dev);
 extern void i915_kernel_lost_context(struct drm_device * dev);
 extern int i915_driver_load(struct drm_device *, unsigned long flags);
 extern int i915_driver_unload(struct drm_device *);
-extern int i915_driver_open(struct drm_device *dev, struct drm_file *file_priv);
+extern int i915_driver_open(struct drm_device *dev, struct drm_file *file);
 extern void i915_driver_lastclose(struct drm_device * dev);
 extern void i915_driver_preclose(struct drm_device *dev,
-				 struct drm_file *file_priv);
+				 struct drm_file *file);
 extern void i915_driver_postclose(struct drm_device *dev,
-				  struct drm_file *file_priv);
+				  struct drm_file *file);
 extern int i915_driver_device_is_agp(struct drm_device * dev);
 #ifdef CONFIG_COMPAT
 extern long i915_compat_ioctl(struct file *filp, unsigned int cmd,
@@ -2069,8 +2252,7 @@ extern unsigned long i915_mch_val(struct drm_i915_private *dev_priv);
 extern unsigned long i915_gfx_val(struct drm_i915_private *dev_priv);
 extern void i915_update_gfx_val(struct drm_i915_private *dev_priv);
 int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool on);
-
-extern void intel_console_resume(struct work_struct *work);
+void intel_hpd_cancel_work(struct drm_i915_private *dev_priv);
 
 /* i915_irq.c */
 void i915_queue_hangcheck(struct drm_device *dev);
@@ -2084,10 +2266,12 @@ extern void intel_irq_init(struct drm_device *dev);
 extern void intel_hpd_init(struct drm_device *dev);
 
 extern void intel_uncore_sanitize(struct drm_device *dev);
-extern void intel_uncore_early_sanitize(struct drm_device *dev);
+extern void intel_uncore_early_sanitize(struct drm_device *dev,
+					bool restore_forcewake);
 extern void intel_uncore_init(struct drm_device *dev);
 extern void intel_uncore_check_errors(struct drm_device *dev);
 extern void intel_uncore_fini(struct drm_device *dev);
+extern void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore);
 
 void
 i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
@@ -2117,6 +2301,20 @@ int i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
 int i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv);
+void i915_gem_execbuffer_move_to_active(struct list_head *vmas,
+					struct intel_engine_cs *ring);
+void i915_gem_execbuffer_retire_commands(struct drm_device *dev,
+					 struct drm_file *file,
+					 struct intel_engine_cs *ring,
+					 struct drm_i915_gem_object *obj);
+int i915_gem_ringbuffer_submission(struct drm_device *dev,
+				   struct drm_file *file,
+				   struct intel_engine_cs *ring,
+				   struct intel_context *ctx,
+				   struct drm_i915_gem_execbuffer2 *args,
+				   struct list_head *vmas,
+				   struct drm_i915_gem_object *batch_obj,
+				   u64 exec_start, u32 flags);
 int i915_gem_execbuffer(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 int i915_gem_execbuffer2(struct drm_device *dev, void *data,
@@ -2151,6 +2349,12 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 int i915_gem_wait_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
 void i915_gem_load(struct drm_device *dev);
+unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv,
+			      long target,
+			      unsigned flags);
+#define I915_SHRINK_PURGEABLE 0x1
+#define I915_SHRINK_UNBOUND 0x2
+#define I915_SHRINK_BOUND 0x4
 void *i915_gem_object_alloc(struct drm_device *dev);
 void i915_gem_object_free(struct drm_i915_gem_object *obj);
 void i915_gem_object_init(struct drm_i915_gem_object *obj,
@@ -2235,6 +2439,8 @@ bool i915_gem_retire_requests(struct drm_device *dev);
 void i915_gem_retire_requests_ring(struct intel_engine_cs *ring);
 int __must_check i915_gem_check_wedge(struct i915_gpu_error *error,
 				      bool interruptible);
+int __must_check i915_gem_check_olr(struct intel_engine_cs *ring, u32 seqno);
+
 static inline bool i915_reset_in_progress(struct i915_gpu_error *error)
 {
 	return unlikely(atomic_read(&error->reset_counter)
@@ -2267,6 +2473,7 @@ void i915_gem_reset(struct drm_device *dev);
 bool i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force);
 int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_init(struct drm_device *dev);
+int i915_gem_init_rings(struct drm_device *dev);
 int __must_check i915_gem_init_hw(struct drm_device *dev);
 int i915_gem_l3_remap(struct intel_engine_cs *ring, int slice);
 void i915_gem_init_swizzling(struct drm_device *dev);
@@ -2337,7 +2544,7 @@ static inline bool i915_gem_obj_is_pinned(struct drm_i915_gem_object *obj) {
 }
 
 /* Some GGTT VM helpers */
-#define obj_to_ggtt(obj) \
+#define i915_obj_to_ggtt(obj) \
 	(&((struct drm_i915_private *)(obj)->base.dev->dev_private)->gtt.base)
 static inline bool i915_is_ggtt(struct i915_address_space *vm)
 {
@@ -2346,21 +2553,30 @@ static inline bool i915_is_ggtt(struct i915_address_space *vm)
 	return vm == ggtt;
 }
 
+static inline struct i915_hw_ppgtt *
+i915_vm_to_ppgtt(struct i915_address_space *vm)
+{
+	WARN_ON(i915_is_ggtt(vm));
+
+	return container_of(vm, struct i915_hw_ppgtt, base);
+}
+
+
 static inline bool i915_gem_obj_ggtt_bound(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_bound(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_bound(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline unsigned long
 i915_gem_obj_ggtt_offset(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_offset(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_offset(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline unsigned long
 i915_gem_obj_ggtt_size(struct drm_i915_gem_object *obj)
 {
-	return i915_gem_obj_size(obj, obj_to_ggtt(obj));
+	return i915_gem_obj_size(obj, i915_obj_to_ggtt(obj));
 }
 
 static inline int __must_check
@@ -2368,7 +2584,8 @@ i915_gem_obj_ggtt_pin(struct drm_i915_gem_object *obj,
 		      uint32_t alignment,
 		      unsigned flags)
 {
-	return i915_gem_object_pin(obj, obj_to_ggtt(obj), alignment, flags | PIN_GLOBAL);
+	return i915_gem_object_pin(obj, i915_obj_to_ggtt(obj),
+				   alignment, flags | PIN_GLOBAL);
 }
 
 static inline int
@@ -2380,7 +2597,6 @@ i915_gem_object_ggtt_unbind(struct drm_i915_gem_object *obj)
 void i915_gem_object_ggtt_unpin(struct drm_i915_gem_object *obj);
 
 /* i915_gem_context.c */
-#define ctx_to_ppgtt(ctx) container_of((ctx)->vm, struct i915_hw_ppgtt, base)
 int __must_check i915_gem_context_init(struct drm_device *dev);
 void i915_gem_context_fini(struct drm_device *dev);
 void i915_gem_context_reset(struct drm_device *dev);
@@ -2392,6 +2608,8 @@ int i915_switch_context(struct intel_engine_cs *ring,
 struct intel_context *
 i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id);
 void i915_gem_context_free(struct kref *ctx_ref);
+struct drm_i915_gem_object *
+i915_gem_alloc_context_obj(struct drm_device *dev, size_t size);
 static inline void i915_gem_context_reference(struct intel_context *ctx)
 {
 	kref_get(&ctx->ref);
@@ -2404,7 +2622,7 @@ static inline void i915_gem_context_unreference(struct intel_context *ctx)
 
 static inline bool i915_gem_context_is_default(const struct intel_context *c)
 {
-	return c->id == DEFAULT_CONTEXT_ID;
+	return c->user_handle == DEFAULT_CONTEXT_HANDLE;
 }
 
 int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
@@ -2412,8 +2630,6 @@ int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 				   struct drm_file *file);
 
-/* i915_gem_render_state.c */
-int i915_gem_render_state_init(struct intel_engine_cs *ring);
 /* i915_gem_evict.c */
 int __must_check i915_gem_evict_something(struct drm_device *dev,
 					  struct i915_address_space *vm,
@@ -2435,7 +2651,7 @@ static inline void i915_gem_chipset_flush(struct drm_device *dev)
 
 /* i915_gem_stolen.c */
 int i915_gem_init_stolen(struct drm_device *dev);
-int i915_gem_stolen_setup_compression(struct drm_device *dev, int size);
+int i915_gem_stolen_setup_compression(struct drm_device *dev, int size, int fb_cpp);
 void i915_gem_stolen_cleanup_compression(struct drm_device *dev);
 void i915_gem_cleanup_stolen(struct drm_device *dev);
 struct drm_i915_gem_object *
@@ -2445,7 +2661,6 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 					       u32 stolen_offset,
 					       u32 gtt_offset,
 					       u32 size);
-void i915_gem_object_release_stolen(struct drm_i915_gem_object *obj);
 
 /* i915_gem_tiling.c */
 static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
@@ -2482,6 +2697,7 @@ void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
 int i915_error_state_to_str(struct drm_i915_error_state_buf *estr,
 			    const struct i915_error_state_file_priv *error);
 int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb,
+			      struct drm_i915_private *i915,
 			      size_t count, loff_t pos);
 static inline void i915_error_state_buf_release(
 	struct drm_i915_error_state_buf *eb)
@@ -2496,7 +2712,7 @@ void i915_error_state_put(struct i915_error_state_file_priv *error_priv);
 void i915_destroy_error_state(struct drm_device *dev);
 
 void i915_get_extra_instdone(struct drm_device *dev, uint32_t *instdone);
-const char *i915_cache_level_str(int type);
+const char *i915_cache_level_str(struct drm_i915_private *i915, int type);
 
 /* i915_cmd_parser.c */
 int i915_cmd_parser_get_version(void);
@@ -2588,13 +2804,14 @@ extern void intel_modeset_setup_hw_state(struct drm_device *dev,
 extern void i915_redisable_vga(struct drm_device *dev);
 extern void i915_redisable_vga_power_on(struct drm_device *dev);
 extern bool intel_fbc_enabled(struct drm_device *dev);
+extern void gen8_fbc_sw_flush(struct drm_device *dev, u32 value);
 extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
 extern void intel_init_pch_refclk(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
 extern void valleyview_set_rps(struct drm_device *dev, u8 val);
-extern int valleyview_rps_max_freq(struct drm_i915_private *dev_priv);
-extern int valleyview_rps_min_freq(struct drm_i915_private *dev_priv);
+extern void intel_set_memory_cxsr(struct drm_i915_private *dev_priv,
+				  bool enable);
 extern void intel_detect_pch(struct drm_device *dev);
 extern int intel_trans_dp_port_sel(struct drm_crtc *crtc);
 extern int intel_enable_rc6(const struct drm_device *dev);
@@ -2605,6 +2822,8 @@ int i915_reg_read_ioctl(struct drm_device *dev, void *data,
 int i915_get_reset_stats_ioctl(struct drm_device *dev, void *data,
 			       struct drm_file *file);
 
+void intel_notify_mmio_flip(struct intel_engine_cs *ring);
+
 /* overlay */
 extern struct intel_overlay_error_state *intel_overlay_capture_error_state(struct drm_device *dev);
 extern void intel_overlay_print_error_state(struct drm_i915_error_state_buf *e,
@@ -2700,10 +2919,10 @@ int vlv_freq_opcode(struct drm_i915_private *dev_priv, int val);
 
 static inline uint32_t i915_vgacntrl_reg(struct drm_device *dev)
 {
-	if (HAS_PCH_SPLIT(dev))
-		return CPU_VGACNTRL;
-	else if (IS_VALLEYVIEW(dev))
+	if (IS_VALLEYVIEW(dev))
 		return VLV_VGACNTRL;
+	else if (INTEL_INFO(dev)->gen >= 5)
+		return CPU_VGACNTRL;
 	else
 		return VGACNTRL;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index f247d922e44a..28f91df2604d 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -60,7 +60,6 @@ static unsigned long i915_gem_shrinker_scan(struct shrinker *shrinker,
 static int i915_gem_shrinker_oom(struct notifier_block *nb,
 				 unsigned long event,
 				 void *ptr);
-static unsigned long i915_gem_purge(struct drm_i915_private *dev_priv, long target);
 static unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv);
 
 static bool cpu_cache_is_coherent(struct drm_device *dev,
@@ -1085,7 +1084,13 @@ i915_gem_check_wedge(struct i915_gpu_error *error,
 		if (i915_terminally_wedged(error))
 			return -EIO;
 
-		return -EAGAIN;
+		/*
+		 * Check if GPU Reset is in progress - we need intel_ring_begin
+		 * to work properly to reinit the hw state while the gpu is
+		 * still marked as reset-in-progress. Handle this with a flag.
+		 */
+		if (!error->reload_in_reset)
+			return -EAGAIN;
 	}
 
 	return 0;
@@ -1095,7 +1100,7 @@ i915_gem_check_wedge(struct i915_gpu_error *error,
  * Compare seqno against outstanding lazy request. Emit a request if they are
  * equal.
  */
-static int
+int
 i915_gem_check_olr(struct intel_engine_cs *ring, u32 seqno)
 {
 	int ret;
@@ -1161,14 +1166,14 @@ static int __wait_seqno(struct intel_engine_cs *ring, u32 seqno,
 	s64 before, now;
 	int ret;
 
-	WARN(dev_priv->pm.irqs_disabled, "IRQs disabled\n");
+	WARN(!intel_irqs_enabled(dev_priv), "IRQs disabled");
 
 	if (i915_seqno_passed(ring->get_seqno(ring, true), seqno))
 		return 0;
 
 	timeout_expire = timeout ? jiffies + nsecs_to_jiffies((u64)*timeout) : 0;
 
-	if (INTEL_INFO(dev)->gen >= 6 && can_wait_boost(file_priv)) {
+	if (INTEL_INFO(dev)->gen >= 6 && ring->id == RCS && can_wait_boost(file_priv)) {
 		gen6_rps_boost(dev_priv);
 		if (file_priv)
 			mod_delayed_work(dev_priv->wq,
@@ -1560,14 +1565,29 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	if (ret)
 		goto unpin;
 
-	obj->fault_mappable = true;
-
+	/* Finally, remap it using the new GTT offset */
 	pfn = dev_priv->gtt.mappable_base + i915_gem_obj_ggtt_offset(obj);
 	pfn >>= PAGE_SHIFT;
-	pfn += page_offset;
 
-	/* Finally, remap it using the new GTT offset */
-	ret = vm_insert_pfn(vma, (unsigned long)vmf->virtual_address, pfn);
+	if (!obj->fault_mappable) {
+		unsigned long size = min_t(unsigned long,
+					   vma->vm_end - vma->vm_start,
+					   obj->base.size);
+		int i;
+
+		for (i = 0; i < size >> PAGE_SHIFT; i++) {
+			ret = vm_insert_pfn(vma,
+					    (unsigned long)vma->vm_start + i * PAGE_SIZE,
+					    pfn + i);
+			if (ret)
+				break;
+		}
+
+		obj->fault_mappable = true;
+	} else
+		ret = vm_insert_pfn(vma,
+				    (unsigned long)vmf->virtual_address,
+				    pfn + page_offset);
 unpin:
 	i915_gem_object_ggtt_unpin(obj);
 unlock:
@@ -1575,10 +1595,13 @@ unlock:
 out:
 	switch (ret) {
 	case -EIO:
-		/* If this -EIO is due to a gpu hang, give the reset code a
-		 * chance to clean up the mess. Otherwise return the proper
-		 * SIGBUS. */
-		if (i915_terminally_wedged(&dev_priv->gpu_error)) {
+		/*
+		 * We eat errors when the gpu is terminally wedged to avoid
+		 * userspace unduly crashing (gl has no provisions for mmaps to
+		 * fail). But any other -EIO isn't ours (e.g. swap in failure)
+		 * and so needs to be reported.
+		 */
+		if (!i915_terminally_wedged(&dev_priv->gpu_error)) {
 			ret = VM_FAULT_SIGBUS;
 			break;
 		}
@@ -1717,7 +1740,11 @@ static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj)
 	 * offsets on purgeable objects by truncating it and marking it purged,
 	 * which prevents userspace from ever using that object again.
 	 */
-	i915_gem_purge(dev_priv, obj->base.size >> PAGE_SHIFT);
+	i915_gem_shrink(dev_priv,
+			obj->base.size >> PAGE_SHIFT,
+			I915_SHRINK_BOUND |
+			I915_SHRINK_UNBOUND |
+			I915_SHRINK_PURGEABLE);
 	ret = drm_gem_create_mmap_offset(&obj->base);
 	if (ret != -ENOSPC)
 		goto out;
@@ -1914,12 +1941,11 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-static unsigned long
-__i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
-		  bool purgeable_only)
+unsigned long
+i915_gem_shrink(struct drm_i915_private *dev_priv,
+		long target, unsigned flags)
 {
-	struct list_head still_in_list;
-	struct drm_i915_gem_object *obj;
+	const bool purgeable_only = flags & I915_SHRINK_PURGEABLE;
 	unsigned long count = 0;
 
 	/*
@@ -1941,62 +1967,68 @@ __i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
 	 * dev->struct_mutex and so we won't ever be able to observe an
 	 * object on the bound_list with a reference count equals 0.
 	 */
-	INIT_LIST_HEAD(&still_in_list);
-	while (count < target && !list_empty(&dev_priv->mm.unbound_list)) {
-		obj = list_first_entry(&dev_priv->mm.unbound_list,
-				       typeof(*obj), global_list);
-		list_move_tail(&obj->global_list, &still_in_list);
+	if (flags & I915_SHRINK_UNBOUND) {
+		struct list_head still_in_list;
 
-		if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
-			continue;
+		INIT_LIST_HEAD(&still_in_list);
+		while (count < target && !list_empty(&dev_priv->mm.unbound_list)) {
+			struct drm_i915_gem_object *obj;
 
-		drm_gem_object_reference(&obj->base);
+			obj = list_first_entry(&dev_priv->mm.unbound_list,
+					       typeof(*obj), global_list);
+			list_move_tail(&obj->global_list, &still_in_list);
 
-		if (i915_gem_object_put_pages(obj) == 0)
-			count += obj->base.size >> PAGE_SHIFT;
+			if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
+				continue;
+
+			drm_gem_object_reference(&obj->base);
 
-		drm_gem_object_unreference(&obj->base);
+			if (i915_gem_object_put_pages(obj) == 0)
+				count += obj->base.size >> PAGE_SHIFT;
+
+			drm_gem_object_unreference(&obj->base);
+		}
+		list_splice(&still_in_list, &dev_priv->mm.unbound_list);
 	}
-	list_splice(&still_in_list, &dev_priv->mm.unbound_list);
 
-	INIT_LIST_HEAD(&still_in_list);
-	while (count < target && !list_empty(&dev_priv->mm.bound_list)) {
-		struct i915_vma *vma, *v;
+	if (flags & I915_SHRINK_BOUND) {
+		struct list_head still_in_list;
 
-		obj = list_first_entry(&dev_priv->mm.bound_list,
-				       typeof(*obj), global_list);
-		list_move_tail(&obj->global_list, &still_in_list);
+		INIT_LIST_HEAD(&still_in_list);
+		while (count < target && !list_empty(&dev_priv->mm.bound_list)) {
+			struct drm_i915_gem_object *obj;
+			struct i915_vma *vma, *v;
 
-		if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
-			continue;
+			obj = list_first_entry(&dev_priv->mm.bound_list,
+					       typeof(*obj), global_list);
+			list_move_tail(&obj->global_list, &still_in_list);
 
-		drm_gem_object_reference(&obj->base);
+			if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
+				continue;
 
-		list_for_each_entry_safe(vma, v, &obj->vma_list, vma_link)
-			if (i915_vma_unbind(vma))
-				break;
+			drm_gem_object_reference(&obj->base);
 
-		if (i915_gem_object_put_pages(obj) == 0)
-			count += obj->base.size >> PAGE_SHIFT;
+			list_for_each_entry_safe(vma, v, &obj->vma_list, vma_link)
+				if (i915_vma_unbind(vma))
+					break;
 
-		drm_gem_object_unreference(&obj->base);
+			if (i915_gem_object_put_pages(obj) == 0)
+				count += obj->base.size >> PAGE_SHIFT;
+
+			drm_gem_object_unreference(&obj->base);
+		}
+		list_splice(&still_in_list, &dev_priv->mm.bound_list);
 	}
-	list_splice(&still_in_list, &dev_priv->mm.bound_list);
 
 	return count;
 }
 
 static unsigned long
-i915_gem_purge(struct drm_i915_private *dev_priv, long target)
-{
-	return __i915_gem_shrink(dev_priv, target, true);
-}
-
-static unsigned long
 i915_gem_shrink_all(struct drm_i915_private *dev_priv)
 {
 	i915_gem_evict_everything(dev_priv->dev);
-	return __i915_gem_shrink(dev_priv, LONG_MAX, false);
+	return i915_gem_shrink(dev_priv, LONG_MAX,
+			       I915_SHRINK_BOUND | I915_SHRINK_UNBOUND);
 }
 
 static int
@@ -2043,7 +2075,11 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	for (i = 0; i < page_count; i++) {
 		page = shmem_read_mapping_page_gfp(mapping, i, gfp);
 		if (IS_ERR(page)) {
-			i915_gem_purge(dev_priv, page_count);
+			i915_gem_shrink(dev_priv,
+					page_count,
+					I915_SHRINK_BOUND |
+					I915_SHRINK_UNBOUND |
+					I915_SHRINK_PURGEABLE);
 			page = shmem_read_mapping_page_gfp(mapping, i, gfp);
 		}
 		if (IS_ERR(page)) {
@@ -2051,16 +2087,10 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 			 * our own buffer, now let the real VM do its job and
 			 * go down in flames if truly OOM.
 			 */
-			gfp &= ~(__GFP_NORETRY | __GFP_NOWARN | __GFP_NO_KSWAPD);
-			gfp |= __GFP_IO | __GFP_WAIT;
-
 			i915_gem_shrink_all(dev_priv);
-			page = shmem_read_mapping_page_gfp(mapping, i, gfp);
+			page = shmem_read_mapping_page(mapping, i);
 			if (IS_ERR(page))
 				goto err_pages;
-
-			gfp |= __GFP_NORETRY | __GFP_NOWARN | __GFP_NO_KSWAPD;
-			gfp &= ~(__GFP_IO | __GFP_WAIT);
 		}
 #ifdef CONFIG_SWIOTLB
 		if (swiotlb_nr_tbl()) {
@@ -2151,8 +2181,6 @@ static void
 i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 			       struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 seqno = intel_ring_get_seqno(ring);
 
 	BUG_ON(ring == NULL);
@@ -2171,19 +2199,6 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 	list_move_tail(&obj->ring_list, &ring->active_list);
 
 	obj->last_read_seqno = seqno;
-
-	if (obj->fenced_gpu_access) {
-		obj->last_fenced_seqno = seqno;
-
-		/* Bump MRU to take account of the delayed flush */
-		if (obj->fence_reg != I915_FENCE_REG_NONE) {
-			struct drm_i915_fence_reg *reg;
-
-			reg = &dev_priv->fence_regs[obj->fence_reg];
-			list_move_tail(&reg->lru_list,
-				       &dev_priv->mm.fence_list);
-		}
-	}
 }
 
 void i915_vma_move_to_active(struct i915_vma *vma,
@@ -2209,6 +2224,8 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
 			list_move_tail(&vma->mm_list, &vm->inactive_list);
 	}
 
+	intel_fb_obj_flush(obj, true);
+
 	list_del_init(&obj->ring_list);
 	obj->ring = NULL;
 
@@ -2217,7 +2234,6 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
 	obj->base.write_domain = 0;
 
 	obj->last_fenced_seqno = 0;
-	obj->fenced_gpu_access = false;
 
 	obj->active = 0;
 	drm_gem_object_unreference(&obj->base);
@@ -2315,10 +2331,21 @@ int __i915_add_request(struct intel_engine_cs *ring,
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_request *request;
+	struct intel_ringbuffer *ringbuf;
 	u32 request_ring_position, request_start;
 	int ret;
 
-	request_start = intel_ring_get_tail(ring);
+	request = ring->preallocated_lazy_request;
+	if (WARN_ON(request == NULL))
+		return -ENOMEM;
+
+	if (i915.enable_execlists) {
+		struct intel_context *ctx = request->ctx;
+		ringbuf = ctx->engine[ring->id].ringbuf;
+	} else
+		ringbuf = ring->buffer;
+
+	request_start = intel_ring_get_tail(ringbuf);
 	/*
 	 * Emit any outstanding flushes - execbuf can fail to emit the flush
 	 * after having emitted the batchbuffer command. Hence we need to fix
@@ -2326,24 +2353,32 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 * is that the flush _must_ happen before the next request, no matter
 	 * what.
 	 */
-	ret = intel_ring_flush_all_caches(ring);
-	if (ret)
-		return ret;
-
-	request = ring->preallocated_lazy_request;
-	if (WARN_ON(request == NULL))
-		return -ENOMEM;
+	if (i915.enable_execlists) {
+		ret = logical_ring_flush_all_caches(ringbuf);
+		if (ret)
+			return ret;
+	} else {
+		ret = intel_ring_flush_all_caches(ring);
+		if (ret)
+			return ret;
+	}
 
 	/* Record the position of the start of the request so that
 	 * should we detect the updated seqno part-way through the
 	 * GPU processing the request, we never over-estimate the
 	 * position of the head.
 	 */
-	request_ring_position = intel_ring_get_tail(ring);
+	request_ring_position = intel_ring_get_tail(ringbuf);
 
-	ret = ring->add_request(ring);
-	if (ret)
-		return ret;
+	if (i915.enable_execlists) {
+		ret = ring->emit_request(ringbuf);
+		if (ret)
+			return ret;
+	} else {
+		ret = ring->add_request(ring);
+		if (ret)
+			return ret;
+	}
 
 	request->seqno = intel_ring_get_seqno(ring);
 	request->ring = ring;
@@ -2358,12 +2393,14 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 */
 	request->batch_obj = obj;
 
-	/* Hold a reference to the current context so that we can inspect
-	 * it later in case a hangcheck error event fires.
-	 */
-	request->ctx = ring->last_context;
-	if (request->ctx)
-		i915_gem_context_reference(request->ctx);
+	if (!i915.enable_execlists) {
+		/* Hold a reference to the current context so that we can inspect
+		 * it later in case a hangcheck error event fires.
+		 */
+		request->ctx = ring->last_context;
+		if (request->ctx)
+			i915_gem_context_reference(request->ctx);
+	}
 
 	request->emitted_jiffies = jiffies;
 	list_add_tail(&request->list, &ring->request_list);
@@ -2534,6 +2571,18 @@ static void i915_gem_reset_ring_cleanup(struct drm_i915_private *dev_priv,
 		i915_gem_free_request(request);
 	}
 
+	while (!list_empty(&ring->execlist_queue)) {
+		struct intel_ctx_submit_request *submit_req;
+
+		submit_req = list_first_entry(&ring->execlist_queue,
+				struct intel_ctx_submit_request,
+				execlist_link);
+		list_del(&submit_req->execlist_link);
+		intel_runtime_pm_put(dev_priv);
+		i915_gem_context_unreference(submit_req->ctx);
+		kfree(submit_req);
+	}
+
 	/* These may not have been flush before the reset, do so now */
 	kfree(ring->preallocated_lazy_request);
 	ring->preallocated_lazy_request = NULL;
@@ -2618,6 +2667,7 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 
 	while (!list_empty(&ring->request_list)) {
 		struct drm_i915_gem_request *request;
+		struct intel_ringbuffer *ringbuf;
 
 		request = list_first_entry(&ring->request_list,
 					   struct drm_i915_gem_request,
@@ -2627,12 +2677,24 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 			break;
 
 		trace_i915_gem_request_retire(ring, request->seqno);
+
+		/* This is one of the few common intersection points
+		 * between legacy ringbuffer submission and execlists:
+		 * we need to tell them apart in order to find the correct
+		 * ringbuffer to which the request belongs to.
+		 */
+		if (i915.enable_execlists) {
+			struct intel_context *ctx = request->ctx;
+			ringbuf = ctx->engine[ring->id].ringbuf;
+		} else
+			ringbuf = ring->buffer;
+
 		/* We know the GPU must have read the request to have
 		 * sent us the seqno + interrupt, so use the position
 		 * of tail of the request to update the last known position
 		 * of the GPU head.
 		 */
-		ring->buffer->last_retired_head = request->tail;
+		ringbuf->last_retired_head = request->tail;
 
 		i915_gem_free_request(request);
 	}
@@ -2822,6 +2884,8 @@ i915_gem_object_sync(struct drm_i915_gem_object *obj,
 	idx = intel_ring_sync_index(from, to);
 
 	seqno = obj->last_read_seqno;
+	/* Optimization: Avoid semaphore sync when we are sure we already
+	 * waited for an object with higher seqno */
 	if (seqno <= from->semaphore.sync_seqno[idx])
 		return 0;
 
@@ -2892,6 +2956,9 @@ int i915_vma_unbind(struct i915_vma *vma)
 	 * cause memory corruption through use-after-free.
 	 */
 
+	/* Throw away the active reference before moving to the unbound list */
+	i915_gem_object_retire(obj);
+
 	if (i915_is_ggtt(vma->vm)) {
 		i915_gem_object_finish_gtt(obj);
 
@@ -2905,20 +2972,19 @@ int i915_vma_unbind(struct i915_vma *vma)
 
 	vma->unbind_vma(vma);
 
-	i915_gem_gtt_finish_object(obj);
-
 	list_del_init(&vma->mm_list);
-	/* Avoid an unnecessary call to unbind on rebind. */
 	if (i915_is_ggtt(vma->vm))
-		obj->map_and_fenceable = true;
+		obj->map_and_fenceable = false;
 
 	drm_mm_remove_node(&vma->node);
 	i915_gem_vma_destroy(vma);
 
 	/* Since the unbound list is global, only move to that list if
 	 * no more VMAs exist. */
-	if (list_empty(&obj->vma_list))
+	if (list_empty(&obj->vma_list)) {
+		i915_gem_gtt_finish_object(obj);
 		list_move_tail(&obj->global_list, &dev_priv->mm.unbound_list);
+	}
 
 	/* And finally now the object is completely decoupled from this vma,
 	 * we can drop its hold on the backing storage and allow it to be
@@ -2937,9 +3003,11 @@ int i915_gpu_idle(struct drm_device *dev)
 
 	/* Flush everything onto the inactive list. */
 	for_each_ring(ring, dev_priv, i) {
-		ret = i915_switch_context(ring, ring->default_context);
-		if (ret)
-			return ret;
+		if (!i915.enable_execlists) {
+			ret = i915_switch_context(ring, ring->default_context);
+			if (ret)
+				return ret;
+		}
 
 		ret = intel_ring_idle(ring);
 		if (ret)
@@ -3153,7 +3221,6 @@ i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
 		obj->last_fenced_seqno = 0;
 	}
 
-	obj->fenced_gpu_access = false;
 	return 0;
 }
 
@@ -3260,6 +3327,9 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 			return 0;
 		}
 	} else if (enable) {
+		if (WARN_ON(!obj->map_and_fenceable))
+			return -EINVAL;
+
 		reg = i915_find_fence_reg(dev);
 		if (IS_ERR(reg))
 			return PTR_ERR(reg);
@@ -3281,17 +3351,20 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-static bool i915_gem_valid_gtt_space(struct drm_device *dev,
-				     struct drm_mm_node *gtt_space,
+static bool i915_gem_valid_gtt_space(struct i915_vma *vma,
 				     unsigned long cache_level)
 {
+	struct drm_mm_node *gtt_space = &vma->node;
 	struct drm_mm_node *other;
 
-	/* On non-LLC machines we have to be careful when putting differing
-	 * types of snoopable memory together to avoid the prefetcher
-	 * crossing memory domains and dying.
+	/*
+	 * On some machines we have to be careful when putting differing types
+	 * of snoopable memory together to avoid the prefetcher crossing memory
+	 * domains and dying. During vm initialisation, we decide whether or not
+	 * these constraints apply and set the drm_mm.color_adjust
+	 * appropriately.
 	 */
-	if (HAS_LLC(dev))
+	if (vma->vm->mm.color_adjust == NULL)
 		return true;
 
 	if (!drm_mm_node_allocated(gtt_space))
@@ -3429,8 +3502,7 @@ search_free:
 
 		goto err_free_vma;
 	}
-	if (WARN_ON(!i915_gem_valid_gtt_space(dev, &vma->node,
-					      obj->cache_level))) {
+	if (WARN_ON(!i915_gem_valid_gtt_space(vma, obj->cache_level))) {
 		ret = -EINVAL;
 		goto err_remove_node;
 	}
@@ -3530,6 +3602,8 @@ i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj)
 	old_write_domain = obj->base.write_domain;
 	obj->base.write_domain = 0;
 
+	intel_fb_obj_flush(obj, false);
+
 	trace_i915_gem_object_change_domain(obj,
 					    obj->base.read_domains,
 					    old_write_domain);
@@ -3551,6 +3625,8 @@ i915_gem_object_flush_cpu_write_domain(struct drm_i915_gem_object *obj,
 	old_write_domain = obj->base.write_domain;
 	obj->base.write_domain = 0;
 
+	intel_fb_obj_flush(obj, false);
+
 	trace_i915_gem_object_change_domain(obj,
 					    obj->base.read_domains,
 					    old_write_domain);
@@ -3566,11 +3642,12 @@ int
 i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct i915_vma *vma = i915_gem_obj_to_ggtt(obj);
 	uint32_t old_write_domain, old_read_domains;
 	int ret;
 
 	/* Not valid to be called on unbound objects. */
-	if (!i915_gem_obj_bound_any(obj))
+	if (vma == NULL)
 		return -EINVAL;
 
 	if (obj->base.write_domain == I915_GEM_DOMAIN_GTT)
@@ -3604,18 +3681,17 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 		obj->dirty = 1;
 	}
 
+	if (write)
+		intel_fb_obj_invalidate(obj, NULL);
+
 	trace_i915_gem_object_change_domain(obj,
 					    old_read_domains,
 					    old_write_domain);
 
 	/* And bump the LRU for this access */
-	if (i915_gem_object_is_inactive(obj)) {
-		struct i915_vma *vma = i915_gem_obj_to_ggtt(obj);
-		if (vma)
-			list_move_tail(&vma->mm_list,
-				       &dev_priv->gtt.base.inactive_list);
-
-	}
+	if (i915_gem_object_is_inactive(obj))
+		list_move_tail(&vma->mm_list,
+			       &dev_priv->gtt.base.inactive_list);
 
 	return 0;
 }
@@ -3636,7 +3712,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 	}
 
 	list_for_each_entry_safe(vma, next, &obj->vma_list, vma_link) {
-		if (!i915_gem_valid_gtt_space(dev, &vma->node, cache_level)) {
+		if (!i915_gem_valid_gtt_space(vma, cache_level)) {
 			ret = i915_vma_unbind(vma);
 			if (ret)
 				return ret;
@@ -3779,9 +3855,6 @@ static bool is_pin_display(struct drm_i915_gem_object *obj)
 {
 	struct i915_vma *vma;
 
-	if (list_empty(&obj->vma_list))
-		return false;
-
 	vma = i915_gem_obj_to_ggtt(obj);
 	if (!vma)
 		return false;
@@ -3940,6 +4013,9 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 		obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	}
 
+	if (write)
+		intel_fb_obj_invalidate(obj, NULL);
+
 	trace_i915_gem_object_change_domain(obj,
 					    old_read_domains,
 					    old_write_domain);
@@ -4305,8 +4381,6 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 
 	obj->fence_reg = I915_FENCE_REG_NONE;
 	obj->madv = I915_MADV_WILLNEED;
-	/* Avoid an unnecessary call to unbind on the first bind. */
-	obj->map_and_fenceable = true;
 
 	i915_gem_info_add_obj(obj->base.dev->dev_private, obj->base.size);
 }
@@ -4428,13 +4502,14 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 	if (obj->stolen)
 		i915_gem_object_unpin_pages(obj);
 
+	WARN_ON(obj->frontbuffer_bits);
+
 	if (WARN_ON(obj->pages_pin_count))
 		obj->pages_pin_count = 0;
 	if (discard_backing_storage(obj))
 		obj->madv = I915_MADV_DONTNEED;
 	i915_gem_object_put_pages(obj);
 	i915_gem_object_free_mmap_offset(obj);
-	i915_gem_object_release_stolen(obj);
 
 	BUG_ON(obj->pages);
 
@@ -4466,12 +4541,18 @@ struct i915_vma *i915_gem_obj_to_vma(struct drm_i915_gem_object *obj,
 
 void i915_gem_vma_destroy(struct i915_vma *vma)
 {
+	struct i915_address_space *vm = NULL;
 	WARN_ON(vma->node.allocated);
 
 	/* Keep the vma as a placeholder in the execbuffer reservation lists */
 	if (!list_empty(&vma->exec_list))
 		return;
 
+	vm = vma->vm;
+
+	if (!i915_is_ggtt(vm))
+		i915_ppgtt_put(i915_vm_to_ppgtt(vm));
+
 	list_del(&vma->vma_link);
 
 	kfree(vma);
@@ -4485,7 +4566,7 @@ i915_gem_stop_ringbuffers(struct drm_device *dev)
 	int i;
 
 	for_each_ring(ring, dev_priv, i)
-		intel_stop_ring_buffer(ring);
+		dev_priv->gt.stop_ring(ring);
 }
 
 int
@@ -4521,7 +4602,7 @@ i915_gem_suspend(struct drm_device *dev)
 
 	del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
 	cancel_delayed_work_sync(&dev_priv->mm.retire_work);
-	cancel_delayed_work_sync(&dev_priv->mm.idle_work);
+	flush_delayed_work(&dev_priv->mm.idle_work);
 
 	return 0;
 
@@ -4602,11 +4683,46 @@ intel_enable_blt(struct drm_device *dev)
 	return true;
 }
 
-static int i915_gem_init_rings(struct drm_device *dev)
+static void init_unused_ring(struct drm_device *dev, u32 base)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE(RING_CTL(base), 0);
+	I915_WRITE(RING_HEAD(base), 0);
+	I915_WRITE(RING_TAIL(base), 0);
+	I915_WRITE(RING_START(base), 0);
+}
+
+static void init_unused_rings(struct drm_device *dev)
+{
+	if (IS_I830(dev)) {
+		init_unused_ring(dev, PRB1_BASE);
+		init_unused_ring(dev, SRB0_BASE);
+		init_unused_ring(dev, SRB1_BASE);
+		init_unused_ring(dev, SRB2_BASE);
+		init_unused_ring(dev, SRB3_BASE);
+	} else if (IS_GEN2(dev)) {
+		init_unused_ring(dev, SRB0_BASE);
+		init_unused_ring(dev, SRB1_BASE);
+	} else if (IS_GEN3(dev)) {
+		init_unused_ring(dev, PRB1_BASE);
+		init_unused_ring(dev, PRB2_BASE);
+	}
+}
+
+int i915_gem_init_rings(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
+	/*
+	 * At least 830 can leave some of the unused rings
+	 * "active" (ie. head != tail) after resume which
+	 * will prevent c3 entry. Makes sure all unused rings
+	 * are totally idle.
+	 */
+	init_unused_rings(dev);
+
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -4685,7 +4801,7 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	i915_gem_init_swizzling(dev);
 
-	ret = i915_gem_init_rings(dev);
+	ret = dev_priv->gt.init_rings(dev);
 	if (ret)
 		return ret;
 
@@ -4703,6 +4819,14 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (ret && ret != -EIO) {
 		DRM_ERROR("Context enable failed %d\n", ret);
 		i915_gem_cleanup_ringbuffer(dev);
+
+		return ret;
+	}
+
+	ret = i915_ppgtt_init_hw(dev);
+	if (ret && ret != -EIO) {
+		DRM_ERROR("PPGTT enable failed %d\n", ret);
+		i915_gem_cleanup_ringbuffer(dev);
 	}
 
 	return ret;
@@ -4713,6 +4837,9 @@ int i915_gem_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
+	i915.enable_execlists = intel_sanitize_enable_execlists(dev,
+			i915.enable_execlists);
+
 	mutex_lock(&dev->struct_mutex);
 
 	if (IS_VALLEYVIEW(dev)) {
@@ -4723,7 +4850,24 @@ int i915_gem_init(struct drm_device *dev)
 			DRM_DEBUG_DRIVER("allow wake ack timed out\n");
 	}
 
-	i915_gem_init_userptr(dev);
+	if (!i915.enable_execlists) {
+		dev_priv->gt.do_execbuf = i915_gem_ringbuffer_submission;
+		dev_priv->gt.init_rings = i915_gem_init_rings;
+		dev_priv->gt.cleanup_ring = intel_cleanup_ring_buffer;
+		dev_priv->gt.stop_ring = intel_stop_ring_buffer;
+	} else {
+		dev_priv->gt.do_execbuf = intel_execlists_submission;
+		dev_priv->gt.init_rings = intel_logical_rings_init;
+		dev_priv->gt.cleanup_ring = intel_logical_ring_cleanup;
+		dev_priv->gt.stop_ring = intel_logical_ring_stop;
+	}
+
+	ret = i915_gem_init_userptr(dev);
+	if (ret) {
+		mutex_unlock(&dev->struct_mutex);
+		return ret;
+	}
+
 	i915_gem_init_global_gtt(dev);
 
 	ret = i915_gem_context_init(dev);
@@ -4758,7 +4902,7 @@ i915_gem_cleanup_ringbuffer(struct drm_device *dev)
 	int i;
 
 	for_each_ring(ring, dev_priv, i)
-		intel_cleanup_ring_buffer(ring);
+		dev_priv->gt.cleanup_ring(ring);
 }
 
 int
@@ -4912,6 +5056,8 @@ i915_gem_load(struct drm_device *dev)
 
 	dev_priv->mm.oom_notifier.notifier_call = i915_gem_shrinker_oom;
 	register_oom_notifier(&dev_priv->mm.oom_notifier);
+
+	mutex_init(&dev_priv->fb_tracking.lock);
 }
 
 void i915_gem_release(struct drm_device *dev, struct drm_file *file)
@@ -4973,6 +5119,23 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file)
 	return ret;
 }
 
+void i915_gem_track_fb(struct drm_i915_gem_object *old,
+		       struct drm_i915_gem_object *new,
+		       unsigned frontbuffer_bits)
+{
+	if (old) {
+		WARN_ON(!mutex_is_locked(&old->base.dev->struct_mutex));
+		WARN_ON(!(old->frontbuffer_bits & frontbuffer_bits));
+		old->frontbuffer_bits &= ~frontbuffer_bits;
+	}
+
+	if (new) {
+		WARN_ON(!mutex_is_locked(&new->base.dev->struct_mutex));
+		WARN_ON(new->frontbuffer_bits & frontbuffer_bits);
+		new->frontbuffer_bits |= frontbuffer_bits;
+	}
+}
+
 static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task)
 {
 	if (!mutex_is_locked(mutex))
@@ -5051,16 +5214,15 @@ unsigned long i915_gem_obj_offset(struct drm_i915_gem_object *o,
 	struct drm_i915_private *dev_priv = o->base.dev->dev_private;
 	struct i915_vma *vma;
 
-	if (!dev_priv->mm.aliasing_ppgtt ||
-	    vm == &dev_priv->mm.aliasing_ppgtt->base)
-		vm = &dev_priv->gtt.base;
+	WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
 
-	BUG_ON(list_empty(&o->vma_list));
 	list_for_each_entry(vma, &o->vma_list, vma_link) {
 		if (vma->vm == vm)
 			return vma->node.start;
 
 	}
+	WARN(1, "%s vma for this object not found.\n",
+	     i915_is_ggtt(vm) ? "global" : "ppgtt");
 	return -1;
 }
 
@@ -5093,9 +5255,7 @@ unsigned long i915_gem_obj_size(struct drm_i915_gem_object *o,
 	struct drm_i915_private *dev_priv = o->base.dev->dev_private;
 	struct i915_vma *vma;
 
-	if (!dev_priv->mm.aliasing_ppgtt ||
-	    vm == &dev_priv->mm.aliasing_ppgtt->base)
-		vm = &dev_priv->gtt.base;
+	WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base);
 
 	BUG_ON(list_empty(&o->vma_list));
 
@@ -5118,11 +5278,16 @@ i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)
 	if (!i915_gem_shrinker_lock(dev, &unlock))
 		return SHRINK_STOP;
 
-	freed = i915_gem_purge(dev_priv, sc->nr_to_scan);
+	freed = i915_gem_shrink(dev_priv,
+				sc->nr_to_scan,
+				I915_SHRINK_BOUND |
+				I915_SHRINK_UNBOUND |
+				I915_SHRINK_PURGEABLE);
 	if (freed < sc->nr_to_scan)
-		freed += __i915_gem_shrink(dev_priv,
-					   sc->nr_to_scan - freed,
-					   false);
+		freed += i915_gem_shrink(dev_priv,
+					 sc->nr_to_scan - freed,
+					 I915_SHRINK_BOUND |
+					 I915_SHRINK_UNBOUND);
 	if (unlock)
 		mutex_unlock(&dev->struct_mutex);
 
@@ -5141,8 +5306,11 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	bool was_interruptible;
 	bool unlock;
 
-	while (!i915_gem_shrinker_lock(dev, &unlock) && --timeout)
+	while (!i915_gem_shrinker_lock(dev, &unlock) && --timeout) {
 		schedule_timeout_killable(1);
+		if (fatal_signal_pending(current))
+			return NOTIFY_DONE;
+	}
 	if (timeout == 0) {
 		pr_err("Unable to purge GPU memory due lock contention.\n");
 		return NOTIFY_DONE;
@@ -5197,14 +5365,8 @@ struct i915_vma *i915_gem_obj_to_ggtt(struct drm_i915_gem_object *obj)
 {
 	struct i915_vma *vma;
 
-	/* This WARN has probably outlived its usefulness (callers already
-	 * WARN if they don't find the GGTT vma they expect). When removing,
-	 * remember to remove the pre-check in is_pin_display() as well */
-	if (WARN_ON(list_empty(&obj->vma_list)))
-		return NULL;
-
 	vma = list_first_entry(&obj->vma_list, typeof(*vma), vma_link);
-	if (vma->vm != obj_to_ggtt(obj))
+	if (vma->vm != i915_obj_to_ggtt(obj))
 		return NULL;
 
 	return vma;
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index a5ddf3bce9c3..a5221d8f1580 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -96,50 +96,6 @@
 #define GEN6_CONTEXT_ALIGN (64<<10)
 #define GEN7_CONTEXT_ALIGN 4096
 
-static void do_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt)
-{
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_address_space *vm = &ppgtt->base;
-
-	if (ppgtt == dev_priv->mm.aliasing_ppgtt ||
-	    (list_empty(&vm->active_list) && list_empty(&vm->inactive_list))) {
-		ppgtt->base.cleanup(&ppgtt->base);
-		return;
-	}
-
-	/*
-	 * Make sure vmas are unbound before we take down the drm_mm
-	 *
-	 * FIXME: Proper refcounting should take care of this, this shouldn't be
-	 * needed at all.
-	 */
-	if (!list_empty(&vm->active_list)) {
-		struct i915_vma *vma;
-
-		list_for_each_entry(vma, &vm->active_list, mm_list)
-			if (WARN_ON(list_empty(&vma->vma_link) ||
-				    list_is_singular(&vma->vma_link)))
-				break;
-
-		i915_gem_evict_vm(&ppgtt->base, true);
-	} else {
-		i915_gem_retire_requests(dev);
-		i915_gem_evict_vm(&ppgtt->base, false);
-	}
-
-	ppgtt->base.cleanup(&ppgtt->base);
-}
-
-static void ppgtt_release(struct kref *kref)
-{
-	struct i915_hw_ppgtt *ppgtt =
-		container_of(kref, struct i915_hw_ppgtt, ref);
-
-	do_ppgtt_cleanup(ppgtt);
-	kfree(ppgtt);
-}
-
 static size_t get_context_alignment(struct drm_device *dev)
 {
 	if (IS_GEN6(dev))
@@ -179,48 +135,52 @@ static int get_context_size(struct drm_device *dev)
 void i915_gem_context_free(struct kref *ctx_ref)
 {
 	struct intel_context *ctx = container_of(ctx_ref,
-						   typeof(*ctx), ref);
-	struct i915_hw_ppgtt *ppgtt = NULL;
+						 typeof(*ctx), ref);
 
-	if (ctx->obj) {
-		/* We refcount even the aliasing PPGTT to keep the code symmetric */
-		if (USES_PPGTT(ctx->obj->base.dev))
-			ppgtt = ctx_to_ppgtt(ctx);
+	if (i915.enable_execlists)
+		intel_lr_context_free(ctx);
 
-		/* XXX: Free up the object before tearing down the address space, in
-		 * case we're bound in the PPGTT */
-		drm_gem_object_unreference(&ctx->obj->base);
-	}
+	i915_ppgtt_put(ctx->ppgtt);
 
-	if (ppgtt)
-		kref_put(&ppgtt->ref, ppgtt_release);
+	if (ctx->legacy_hw_ctx.rcs_state)
+		drm_gem_object_unreference(&ctx->legacy_hw_ctx.rcs_state->base);
 	list_del(&ctx->link);
 	kfree(ctx);
 }
 
-static struct i915_hw_ppgtt *
-create_vm_for_ctx(struct drm_device *dev, struct intel_context *ctx)
+struct drm_i915_gem_object *
+i915_gem_alloc_context_obj(struct drm_device *dev, size_t size)
 {
-	struct i915_hw_ppgtt *ppgtt;
+	struct drm_i915_gem_object *obj;
 	int ret;
 
-	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
-	if (!ppgtt)
+	obj = i915_gem_alloc_object(dev, size);
+	if (obj == NULL)
 		return ERR_PTR(-ENOMEM);
 
-	ret = i915_gem_init_ppgtt(dev, ppgtt);
-	if (ret) {
-		kfree(ppgtt);
-		return ERR_PTR(ret);
+	/*
+	 * Try to make the context utilize L3 as well as LLC.
+	 *
+	 * On VLV we don't have L3 controls in the PTEs so we
+	 * shouldn't touch the cache level, especially as that
+	 * would make the object snooped which might have a
+	 * negative performance impact.
+	 */
+	if (INTEL_INFO(dev)->gen >= 7 && !IS_VALLEYVIEW(dev)) {
+		ret = i915_gem_object_set_cache_level(obj, I915_CACHE_L3_LLC);
+		/* Failure shouldn't ever happen this early */
+		if (WARN_ON(ret)) {
+			drm_gem_object_unreference(&obj->base);
+			return ERR_PTR(ret);
+		}
 	}
 
-	ppgtt->ctx = ctx;
-	return ppgtt;
+	return obj;
 }
 
 static struct intel_context *
 __create_hw_context(struct drm_device *dev,
-		  struct drm_i915_file_private *file_priv)
+		    struct drm_i915_file_private *file_priv)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_context *ctx;
@@ -234,40 +194,26 @@ __create_hw_context(struct drm_device *dev,
 	list_add_tail(&ctx->link, &dev_priv->context_list);
 
 	if (dev_priv->hw_context_size) {
-		ctx->obj = i915_gem_alloc_object(dev, dev_priv->hw_context_size);
-		if (ctx->obj == NULL) {
-			ret = -ENOMEM;
+		struct drm_i915_gem_object *obj =
+				i915_gem_alloc_context_obj(dev, dev_priv->hw_context_size);
+		if (IS_ERR(obj)) {
+			ret = PTR_ERR(obj);
 			goto err_out;
 		}
-
-		/*
-		 * Try to make the context utilize L3 as well as LLC.
-		 *
-		 * On VLV we don't have L3 controls in the PTEs so we
-		 * shouldn't touch the cache level, especially as that
-		 * would make the object snooped which might have a
-		 * negative performance impact.
-		 */
-		if (INTEL_INFO(dev)->gen >= 7 && !IS_VALLEYVIEW(dev)) {
-			ret = i915_gem_object_set_cache_level(ctx->obj,
-							      I915_CACHE_L3_LLC);
-			/* Failure shouldn't ever happen this early */
-			if (WARN_ON(ret))
-				goto err_out;
-		}
+		ctx->legacy_hw_ctx.rcs_state = obj;
 	}
 
 	/* Default context will never have a file_priv */
 	if (file_priv != NULL) {
 		ret = idr_alloc(&file_priv->context_idr, ctx,
-				DEFAULT_CONTEXT_ID, 0, GFP_KERNEL);
+				DEFAULT_CONTEXT_HANDLE, 0, GFP_KERNEL);
 		if (ret < 0)
 			goto err_out;
 	} else
-		ret = DEFAULT_CONTEXT_ID;
+		ret = DEFAULT_CONTEXT_HANDLE;
 
 	ctx->file_priv = file_priv;
-	ctx->id = ret;
+	ctx->user_handle = ret;
 	/* NB: Mark all slices as needing a remap so that when the context first
 	 * loads it will restore whatever remap state already exists. If there
 	 * is no remap info, it will be a NOP. */
@@ -287,11 +233,9 @@ err_out:
  */
 static struct intel_context *
 i915_gem_create_context(struct drm_device *dev,
-			struct drm_i915_file_private *file_priv,
-			bool create_vm)
+			struct drm_i915_file_private *file_priv)
 {
 	const bool is_global_default_ctx = file_priv == NULL;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_context *ctx;
 	int ret = 0;
 
@@ -301,7 +245,7 @@ i915_gem_create_context(struct drm_device *dev,
 	if (IS_ERR(ctx))
 		return ctx;
 
-	if (is_global_default_ctx && ctx->obj) {
+	if (is_global_default_ctx && ctx->legacy_hw_ctx.rcs_state) {
 		/* We may need to do things with the shrinker which
 		 * require us to immediately switch back to the default
 		 * context. This can cause a problem as pinning the
@@ -309,7 +253,7 @@ i915_gem_create_context(struct drm_device *dev,
 		 * be available. To avoid this we always pin the default
 		 * context.
 		 */
-		ret = i915_gem_obj_ggtt_pin(ctx->obj,
+		ret = i915_gem_obj_ggtt_pin(ctx->legacy_hw_ctx.rcs_state,
 					    get_context_alignment(dev), 0);
 		if (ret) {
 			DRM_DEBUG_DRIVER("Couldn't pin %d\n", ret);
@@ -317,40 +261,24 @@ i915_gem_create_context(struct drm_device *dev,
 		}
 	}
 
-	if (create_vm) {
-		struct i915_hw_ppgtt *ppgtt = create_vm_for_ctx(dev, ctx);
+	if (USES_FULL_PPGTT(dev)) {
+		struct i915_hw_ppgtt *ppgtt = i915_ppgtt_create(dev, file_priv);
 
 		if (IS_ERR_OR_NULL(ppgtt)) {
 			DRM_DEBUG_DRIVER("PPGTT setup failed (%ld)\n",
 					 PTR_ERR(ppgtt));
 			ret = PTR_ERR(ppgtt);
 			goto err_unpin;
-		} else
-			ctx->vm = &ppgtt->base;
-
-		/* This case is reserved for the global default context and
-		 * should only happen once. */
-		if (is_global_default_ctx) {
-			if (WARN_ON(dev_priv->mm.aliasing_ppgtt)) {
-				ret = -EEXIST;
-				goto err_unpin;
-			}
-
-			dev_priv->mm.aliasing_ppgtt = ppgtt;
 		}
-	} else if (USES_PPGTT(dev)) {
-		/* For platforms which only have aliasing PPGTT, we fake the
-		 * address space and refcounting. */
-		ctx->vm = &dev_priv->mm.aliasing_ppgtt->base;
-		kref_get(&dev_priv->mm.aliasing_ppgtt->ref);
-	} else
-		ctx->vm = &dev_priv->gtt.base;
+
+		ctx->ppgtt = ppgtt;
+	}
 
 	return ctx;
 
 err_unpin:
-	if (is_global_default_ctx && ctx->obj)
-		i915_gem_object_ggtt_unpin(ctx->obj);
+	if (is_global_default_ctx && ctx->legacy_hw_ctx.rcs_state)
+		i915_gem_object_ggtt_unpin(ctx->legacy_hw_ctx.rcs_state);
 err_destroy:
 	i915_gem_context_unreference(ctx);
 	return ERR_PTR(ret);
@@ -361,30 +289,23 @@ void i915_gem_context_reset(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	/* Prevent the hardware from restoring the last context (which hung) on
-	 * the next switch */
+	/* In execlists mode we will unreference the context when the execlist
+	 * queue is cleared and the requests destroyed.
+	 */
+	if (i915.enable_execlists)
+		return;
+
 	for (i = 0; i < I915_NUM_RINGS; i++) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
-		struct intel_context *dctx = ring->default_context;
-
-		/* Do a fake switch to the default context */
-		if (ring->last_context == dctx)
-			continue;
+		struct intel_context *lctx = ring->last_context;
 
-		if (!ring->last_context)
-			continue;
+		if (lctx) {
+			if (lctx->legacy_hw_ctx.rcs_state && i == RCS)
+				i915_gem_object_ggtt_unpin(lctx->legacy_hw_ctx.rcs_state);
 
-		if (dctx->obj && i == RCS) {
-			WARN_ON(i915_gem_obj_ggtt_pin(dctx->obj,
-						      get_context_alignment(dev), 0));
-			/* Fake a finish/inactive */
-			dctx->obj->base.write_domain = 0;
-			dctx->obj->active = 0;
+			i915_gem_context_unreference(lctx);
+			ring->last_context = NULL;
 		}
-
-		i915_gem_context_unreference(ring->last_context);
-		i915_gem_context_reference(dctx);
-		ring->last_context = dctx;
 	}
 }
 
@@ -399,7 +320,11 @@ int i915_gem_context_init(struct drm_device *dev)
 	if (WARN_ON(dev_priv->ring[RCS].default_context))
 		return 0;
 
-	if (HAS_HW_CONTEXTS(dev)) {
+	if (i915.enable_execlists) {
+		/* NB: intentionally left blank. We will allocate our own
+		 * backing objects as we need them, thank you very much */
+		dev_priv->hw_context_size = 0;
+	} else if (HAS_HW_CONTEXTS(dev)) {
 		dev_priv->hw_context_size = round_up(get_context_size(dev), 4096);
 		if (dev_priv->hw_context_size > (1<<20)) {
 			DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n",
@@ -408,18 +333,23 @@ int i915_gem_context_init(struct drm_device *dev)
 		}
 	}
 
-	ctx = i915_gem_create_context(dev, NULL, USES_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, NULL);
 	if (IS_ERR(ctx)) {
 		DRM_ERROR("Failed to create default global context (error %ld)\n",
 			  PTR_ERR(ctx));
 		return PTR_ERR(ctx);
 	}
 
-	/* NB: RCS will hold a ref for all rings */
-	for (i = 0; i < I915_NUM_RINGS; i++)
-		dev_priv->ring[i].default_context = ctx;
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct intel_engine_cs *ring = &dev_priv->ring[i];
+
+		/* NB: RCS will hold a ref for all rings */
+		ring->default_context = ctx;
+	}
 
-	DRM_DEBUG_DRIVER("%s context support initialized\n", dev_priv->hw_context_size ? "HW" : "fake");
+	DRM_DEBUG_DRIVER("%s context support initialized\n",
+			i915.enable_execlists ? "LR" :
+			dev_priv->hw_context_size ? "HW" : "fake");
 	return 0;
 }
 
@@ -429,7 +359,7 @@ void i915_gem_context_fini(struct drm_device *dev)
 	struct intel_context *dctx = dev_priv->ring[RCS].default_context;
 	int i;
 
-	if (dctx->obj) {
+	if (dctx->legacy_hw_ctx.rcs_state) {
 		/* The only known way to stop the gpu from accessing the hw context is
 		 * to reset it. Do this as the very last operation to avoid confusing
 		 * other code, leading to spurious errors. */
@@ -444,13 +374,13 @@ void i915_gem_context_fini(struct drm_device *dev)
 		WARN_ON(!dev_priv->ring[RCS].last_context);
 		if (dev_priv->ring[RCS].last_context == dctx) {
 			/* Fake switch to NULL context */
-			WARN_ON(dctx->obj->active);
-			i915_gem_object_ggtt_unpin(dctx->obj);
+			WARN_ON(dctx->legacy_hw_ctx.rcs_state->active);
+			i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
 			i915_gem_context_unreference(dctx);
 			dev_priv->ring[RCS].last_context = NULL;
 		}
 
-		i915_gem_object_ggtt_unpin(dctx->obj);
+		i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
 	}
 
 	for (i = 0; i < I915_NUM_RINGS; i++) {
@@ -471,19 +401,11 @@ int i915_gem_context_enable(struct drm_i915_private *dev_priv)
 	struct intel_engine_cs *ring;
 	int ret, i;
 
-	/* This is the only place the aliasing PPGTT gets enabled, which means
-	 * it has to happen before we bail on reset */
-	if (dev_priv->mm.aliasing_ppgtt) {
-		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-		ppgtt->enable(ppgtt);
-	}
+	BUG_ON(!dev_priv->ring[RCS].default_context);
 
-	/* FIXME: We should make this work, even in reset */
-	if (i915_reset_in_progress(&dev_priv->gpu_error))
+	if (i915.enable_execlists)
 		return 0;
 
-	BUG_ON(!dev_priv->ring[RCS].default_context);
-
 	for_each_ring(ring, dev_priv, i) {
 		ret = i915_switch_context(ring, ring->default_context);
 		if (ret)
@@ -509,7 +431,7 @@ int i915_gem_context_open(struct drm_device *dev, struct drm_file *file)
 	idr_init(&file_priv->context_idr);
 
 	mutex_lock(&dev->struct_mutex);
-	ctx = i915_gem_create_context(dev, file_priv, USES_FULL_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, file_priv);
 	mutex_unlock(&dev->struct_mutex);
 
 	if (IS_ERR(ctx)) {
@@ -545,6 +467,7 @@ mi_set_context(struct intel_engine_cs *ring,
 	       struct intel_context *new_context,
 	       u32 hw_flags)
 {
+	u32 flags = hw_flags | MI_MM_SPACE_GTT;
 	int ret;
 
 	/* w/a: If Flush TLB Invalidation Mode is enabled, driver must do a TLB
@@ -558,6 +481,10 @@ mi_set_context(struct intel_engine_cs *ring,
 			return ret;
 	}
 
+	/* These flags are for resource streamer on HSW+ */
+	if (!IS_HASWELL(ring->dev) && INTEL_INFO(ring->dev)->gen < 8)
+		flags |= (MI_SAVE_EXT_STATE_EN | MI_RESTORE_EXT_STATE_EN);
+
 	ret = intel_ring_begin(ring, 6);
 	if (ret)
 		return ret;
@@ -570,11 +497,8 @@ mi_set_context(struct intel_engine_cs *ring,
 
 	intel_ring_emit(ring, MI_NOOP);
 	intel_ring_emit(ring, MI_SET_CONTEXT);
-	intel_ring_emit(ring, i915_gem_obj_ggtt_offset(new_context->obj) |
-			MI_MM_SPACE_GTT |
-			MI_SAVE_EXT_STATE_EN |
-			MI_RESTORE_EXT_STATE_EN |
-			hw_flags);
+	intel_ring_emit(ring, i915_gem_obj_ggtt_offset(new_context->legacy_hw_ctx.rcs_state) |
+			flags);
 	/*
 	 * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
 	 * WaMiSetContext_Hang:snb,ivb,vlv
@@ -596,22 +520,21 @@ static int do_switch(struct intel_engine_cs *ring,
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct intel_context *from = ring->last_context;
-	struct i915_hw_ppgtt *ppgtt = ctx_to_ppgtt(to);
 	u32 hw_flags = 0;
 	bool uninitialized = false;
 	int ret, i;
 
 	if (from != NULL && ring == &dev_priv->ring[RCS]) {
-		BUG_ON(from->obj == NULL);
-		BUG_ON(!i915_gem_obj_is_pinned(from->obj));
+		BUG_ON(from->legacy_hw_ctx.rcs_state == NULL);
+		BUG_ON(!i915_gem_obj_is_pinned(from->legacy_hw_ctx.rcs_state));
 	}
 
-	if (from == to && from->last_ring == ring && !to->remap_slice)
+	if (from == to && !to->remap_slice)
 		return 0;
 
 	/* Trying to pin first makes error handling easier. */
 	if (ring == &dev_priv->ring[RCS]) {
-		ret = i915_gem_obj_ggtt_pin(to->obj,
+		ret = i915_gem_obj_ggtt_pin(to->legacy_hw_ctx.rcs_state,
 					    get_context_alignment(ring->dev), 0);
 		if (ret)
 			return ret;
@@ -624,8 +547,8 @@ static int do_switch(struct intel_engine_cs *ring,
 	 */
 	from = ring->last_context;
 
-	if (USES_FULL_PPGTT(ring->dev)) {
-		ret = ppgtt->switch_mm(ppgtt, ring, false);
+	if (to->ppgtt) {
+		ret = to->ppgtt->switch_mm(to->ppgtt, ring);
 		if (ret)
 			goto unpin_out;
 	}
@@ -644,17 +567,17 @@ static int do_switch(struct intel_engine_cs *ring,
 	 *
 	 * XXX: We need a real interface to do this instead of trickery.
 	 */
-	ret = i915_gem_object_set_to_gtt_domain(to->obj, false);
+	ret = i915_gem_object_set_to_gtt_domain(to->legacy_hw_ctx.rcs_state, false);
 	if (ret)
 		goto unpin_out;
 
-	if (!to->obj->has_global_gtt_mapping) {
-		struct i915_vma *vma = i915_gem_obj_to_vma(to->obj,
+	if (!to->legacy_hw_ctx.rcs_state->has_global_gtt_mapping) {
+		struct i915_vma *vma = i915_gem_obj_to_vma(to->legacy_hw_ctx.rcs_state,
 							   &dev_priv->gtt.base);
-		vma->bind_vma(vma, to->obj->cache_level, GLOBAL_BIND);
+		vma->bind_vma(vma, to->legacy_hw_ctx.rcs_state->cache_level, GLOBAL_BIND);
 	}
 
-	if (!to->is_initialized || i915_gem_context_is_default(to))
+	if (!to->legacy_hw_ctx.initialized || i915_gem_context_is_default(to))
 		hw_flags |= MI_RESTORE_INHIBIT;
 
 	ret = mi_set_context(ring, to, hw_flags);
@@ -680,8 +603,8 @@ static int do_switch(struct intel_engine_cs *ring,
 	 * MI_SET_CONTEXT instead of when the next seqno has completed.
 	 */
 	if (from != NULL) {
-		from->obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
-		i915_vma_move_to_active(i915_gem_obj_to_ggtt(from->obj), ring);
+		from->legacy_hw_ctx.rcs_state->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
+		i915_vma_move_to_active(i915_gem_obj_to_ggtt(from->legacy_hw_ctx.rcs_state), ring);
 		/* As long as MI_SET_CONTEXT is serializing, ie. it flushes the
 		 * whole damn pipeline, we don't need to explicitly mark the
 		 * object dirty. The only exception is that the context must be
@@ -689,23 +612,28 @@ static int do_switch(struct intel_engine_cs *ring,
 		 * able to defer doing this until we know the object would be
 		 * swapped, but there is no way to do that yet.
 		 */
-		from->obj->dirty = 1;
-		BUG_ON(from->obj->ring != ring);
+		from->legacy_hw_ctx.rcs_state->dirty = 1;
+		BUG_ON(from->legacy_hw_ctx.rcs_state->ring != ring);
 
 		/* obj is kept alive until the next request by its active ref */
-		i915_gem_object_ggtt_unpin(from->obj);
+		i915_gem_object_ggtt_unpin(from->legacy_hw_ctx.rcs_state);
 		i915_gem_context_unreference(from);
 	}
 
-	uninitialized = !to->is_initialized && from == NULL;
-	to->is_initialized = true;
+	uninitialized = !to->legacy_hw_ctx.initialized && from == NULL;
+	to->legacy_hw_ctx.initialized = true;
 
 done:
 	i915_gem_context_reference(to);
 	ring->last_context = to;
-	to->last_ring = ring;
 
 	if (uninitialized) {
+		if (ring->init_context) {
+			ret = ring->init_context(ring);
+			if (ret)
+				DRM_ERROR("ring init context: %d\n", ret);
+		}
+
 		ret = i915_gem_render_state_init(ring);
 		if (ret)
 			DRM_ERROR("init render state: %d\n", ret);
@@ -715,7 +643,7 @@ done:
 
 unpin_out:
 	if (ring->id == RCS)
-		i915_gem_object_ggtt_unpin(to->obj);
+		i915_gem_object_ggtt_unpin(to->legacy_hw_ctx.rcs_state);
 	return ret;
 }
 
@@ -726,17 +654,22 @@ unpin_out:
  *
  * The context life cycle is simple. The context refcount is incremented and
  * decremented by 1 and create and destroy. If the context is in use by the GPU,
- * it will have a refoucnt > 1. This allows us to destroy the context abstract
+ * it will have a refcount > 1. This allows us to destroy the context abstract
  * object while letting the normal object tracking destroy the backing BO.
+ *
+ * This function should not be used in execlists mode.  Instead the context is
+ * switched by writing to the ELSP and requests keep a reference to their
+ * context.
  */
 int i915_switch_context(struct intel_engine_cs *ring,
 			struct intel_context *to)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 
+	WARN_ON(i915.enable_execlists);
 	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
 
-	if (to->obj == NULL) { /* We have the fake context */
+	if (to->legacy_hw_ctx.rcs_state == NULL) { /* We have the fake context */
 		if (to != ring->last_context) {
 			i915_gem_context_reference(to);
 			if (ring->last_context)
@@ -749,9 +682,9 @@ int i915_switch_context(struct intel_engine_cs *ring,
 	return do_switch(ring, to);
 }
 
-static bool hw_context_enabled(struct drm_device *dev)
+static bool contexts_enabled(struct drm_device *dev)
 {
-	return to_i915(dev)->hw_context_size;
+	return i915.enable_execlists || to_i915(dev)->hw_context_size;
 }
 
 int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
@@ -762,19 +695,19 @@ int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 	struct intel_context *ctx;
 	int ret;
 
-	if (!hw_context_enabled(dev))
+	if (!contexts_enabled(dev))
 		return -ENODEV;
 
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
 
-	ctx = i915_gem_create_context(dev, file_priv, USES_FULL_PPGTT(dev));
+	ctx = i915_gem_create_context(dev, file_priv);
 	mutex_unlock(&dev->struct_mutex);
 	if (IS_ERR(ctx))
 		return PTR_ERR(ctx);
 
-	args->ctx_id = ctx->id;
+	args->ctx_id = ctx->user_handle;
 	DRM_DEBUG_DRIVER("HW context %d created\n", args->ctx_id);
 
 	return 0;
@@ -788,7 +721,7 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 	struct intel_context *ctx;
 	int ret;
 
-	if (args->ctx_id == DEFAULT_CONTEXT_ID)
+	if (args->ctx_id == DEFAULT_CONTEXT_HANDLE)
 		return -ENOENT;
 
 	ret = i915_mutex_lock_interruptible(dev);
@@ -801,7 +734,7 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 		return PTR_ERR(ctx);
 	}
 
-	idr_remove(&ctx->file_priv->context_idr, ctx->id);
+	idr_remove(&ctx->file_priv->context_idr, ctx->user_handle);
 	i915_gem_context_unreference(ctx);
 	mutex_unlock(&dev->struct_mutex);
 
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index bbf4b12d842e..886ff2ee7a28 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -243,7 +243,7 @@ int
 i915_gem_evict_everything(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_address_space *vm;
+	struct i915_address_space *vm, *v;
 	bool lists_empty = true;
 	int ret;
 
@@ -270,7 +270,7 @@ i915_gem_evict_everything(struct drm_device *dev)
 	i915_gem_retire_requests(dev);
 
 	/* Having flushed everything, unbind() should never raise an error */
-	list_for_each_entry(vm, &dev_priv->vm_list, global_link)
+	list_for_each_entry_safe(vm, v, &dev_priv->vm_list, global_link)
 		WARN_ON(i915_gem_evict_vm(vm, false));
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 3a30133f93e8..1a0611bb576b 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -35,6 +35,7 @@
 
 #define  __EXEC_OBJECT_HAS_PIN (1<<31)
 #define  __EXEC_OBJECT_HAS_FENCE (1<<30)
+#define  __EXEC_OBJECT_NEEDS_MAP (1<<29)
 #define  __EXEC_OBJECT_NEEDS_BIAS (1<<28)
 
 #define BATCH_OFFSET_BIAS (256*1024)
@@ -94,7 +95,6 @@ eb_lookup_vmas(struct eb_vmas *eb,
 	       struct i915_address_space *vm,
 	       struct drm_file *file)
 {
-	struct drm_i915_private *dev_priv = vm->dev->dev_private;
 	struct drm_i915_gem_object *obj;
 	struct list_head objects;
 	int i, ret;
@@ -129,20 +129,6 @@ eb_lookup_vmas(struct eb_vmas *eb,
 	i = 0;
 	while (!list_empty(&objects)) {
 		struct i915_vma *vma;
-		struct i915_address_space *bind_vm = vm;
-
-		if (exec[i].flags & EXEC_OBJECT_NEEDS_GTT &&
-		    USES_FULL_PPGTT(vm->dev)) {
-			ret = -EINVAL;
-			goto err;
-		}
-
-		/* If we have secure dispatch, or the userspace assures us that
-		 * they know what they're doing, use the GGTT VM.
-		 */
-		if (((args->flags & I915_EXEC_SECURE) &&
-		    (i == (args->buffer_count - 1))))
-			bind_vm = &dev_priv->gtt.base;
 
 		obj = list_first_entry(&objects,
 				       struct drm_i915_gem_object,
@@ -156,7 +142,7 @@ eb_lookup_vmas(struct eb_vmas *eb,
 		 * from the (obj, vm) we don't run the risk of creating
 		 * duplicated vmas for the same vm.
 		 */
-		vma = i915_gem_obj_lookup_or_create_vma(obj, bind_vm);
+		vma = i915_gem_obj_lookup_or_create_vma(obj, vm);
 		if (IS_ERR(vma)) {
 			DRM_DEBUG("Failed to lookup VMA\n");
 			ret = PTR_ERR(vma);
@@ -307,7 +293,7 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
 	struct drm_device *dev = obj->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint64_t delta = reloc->delta + target_offset;
-	uint32_t __iomem *reloc_entry;
+	uint64_t offset;
 	void __iomem *reloc_page;
 	int ret;
 
@@ -320,25 +306,24 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
 		return ret;
 
 	/* Map the page containing the relocation we're going to perform.  */
-	reloc->offset += i915_gem_obj_ggtt_offset(obj);
+	offset = i915_gem_obj_ggtt_offset(obj);
+	offset += reloc->offset;
 	reloc_page = io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
-			reloc->offset & PAGE_MASK);
-	reloc_entry = (uint32_t __iomem *)
-		(reloc_page + offset_in_page(reloc->offset));
-	iowrite32(lower_32_bits(delta), reloc_entry);
+					      offset & PAGE_MASK);
+	iowrite32(lower_32_bits(delta), reloc_page + offset_in_page(offset));
 
 	if (INTEL_INFO(dev)->gen >= 8) {
-		reloc_entry += 1;
+		offset += sizeof(uint32_t);
 
-		if (offset_in_page(reloc->offset + sizeof(uint32_t)) == 0) {
+		if (offset_in_page(offset) == 0) {
 			io_mapping_unmap_atomic(reloc_page);
-			reloc_page = io_mapping_map_atomic_wc(
-					dev_priv->gtt.mappable,
-					reloc->offset + sizeof(uint32_t));
-			reloc_entry = reloc_page;
+			reloc_page =
+				io_mapping_map_atomic_wc(dev_priv->gtt.mappable,
+							 offset);
 		}
 
-		iowrite32(upper_32_bits(delta), reloc_entry);
+		iowrite32(upper_32_bits(delta),
+			  reloc_page + offset_in_page(offset));
 	}
 
 	io_mapping_unmap_atomic(reloc_page);
@@ -535,34 +520,18 @@ i915_gem_execbuffer_relocate(struct eb_vmas *eb)
 }
 
 static int
-need_reloc_mappable(struct i915_vma *vma)
-{
-	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	return entry->relocation_count && !use_cpu_reloc(vma->obj) &&
-		i915_is_ggtt(vma->vm);
-}
-
-static int
 i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 				struct intel_engine_cs *ring,
 				bool *need_reloc)
 {
 	struct drm_i915_gem_object *obj = vma->obj;
 	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
-	bool need_fence;
 	uint64_t flags;
 	int ret;
 
 	flags = 0;
-
-	need_fence =
-		has_fenced_gpu_access &&
-		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-		obj->tiling_mode != I915_TILING_NONE;
-	if (need_fence || need_reloc_mappable(vma))
+	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP)
 		flags |= PIN_MAPPABLE;
-
 	if (entry->flags & EXEC_OBJECT_NEEDS_GTT)
 		flags |= PIN_GLOBAL;
 	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS)
@@ -574,17 +543,13 @@ i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 
 	entry->flags |= __EXEC_OBJECT_HAS_PIN;
 
-	if (has_fenced_gpu_access) {
-		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
-			ret = i915_gem_object_get_fence(obj);
-			if (ret)
-				return ret;
-
-			if (i915_gem_object_pin_fence(obj))
-				entry->flags |= __EXEC_OBJECT_HAS_FENCE;
+	if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
+		ret = i915_gem_object_get_fence(obj);
+		if (ret)
+			return ret;
 
-			obj->pending_fenced_gpu_access = true;
-		}
+		if (i915_gem_object_pin_fence(obj))
+			entry->flags |= __EXEC_OBJECT_HAS_FENCE;
 	}
 
 	if (entry->offset != vma->node.start) {
@@ -601,26 +566,40 @@ i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 }
 
 static bool
-eb_vma_misplaced(struct i915_vma *vma, bool has_fenced_gpu_access)
+need_reloc_mappable(struct i915_vma *vma)
 {
 	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
-	struct drm_i915_gem_object *obj = vma->obj;
-	bool need_fence, need_mappable;
 
-	need_fence =
-		has_fenced_gpu_access &&
-		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-		obj->tiling_mode != I915_TILING_NONE;
-	need_mappable = need_fence || need_reloc_mappable(vma);
+	if (entry->relocation_count == 0)
+		return false;
+
+	if (!i915_is_ggtt(vma->vm))
+		return false;
 
-	WARN_ON((need_mappable || need_fence) &&
+	/* See also use_cpu_reloc() */
+	if (HAS_LLC(vma->obj->base.dev))
+		return false;
+
+	if (vma->obj->base.write_domain == I915_GEM_DOMAIN_CPU)
+		return false;
+
+	return true;
+}
+
+static bool
+eb_vma_misplaced(struct i915_vma *vma)
+{
+	struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP &&
 	       !i915_is_ggtt(vma->vm));
 
 	if (entry->alignment &&
 	    vma->node.start & (entry->alignment - 1))
 		return true;
 
-	if (need_mappable && !obj->map_and_fenceable)
+	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP && !obj->map_and_fenceable)
 		return true;
 
 	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS &&
@@ -642,9 +621,6 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
 	int retry;
 
-	if (list_empty(vmas))
-		return 0;
-
 	i915_gem_retire_requests_ring(ring);
 
 	vm = list_first_entry(vmas, struct i915_vma, exec_list)->vm;
@@ -658,20 +634,21 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 		obj = vma->obj;
 		entry = vma->exec_entry;
 
+		if (!has_fenced_gpu_access)
+			entry->flags &= ~EXEC_OBJECT_NEEDS_FENCE;
 		need_fence =
-			has_fenced_gpu_access &&
 			entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
 			obj->tiling_mode != I915_TILING_NONE;
 		need_mappable = need_fence || need_reloc_mappable(vma);
 
-		if (need_mappable)
+		if (need_mappable) {
+			entry->flags |= __EXEC_OBJECT_NEEDS_MAP;
 			list_move(&vma->exec_list, &ordered_vmas);
-		else
+		} else
 			list_move_tail(&vma->exec_list, &ordered_vmas);
 
 		obj->base.pending_read_domains = I915_GEM_GPU_DOMAINS & ~I915_GEM_DOMAIN_COMMAND;
 		obj->base.pending_write_domain = 0;
-		obj->pending_fenced_gpu_access = false;
 	}
 	list_splice(&ordered_vmas, vmas);
 
@@ -696,7 +673,7 @@ i915_gem_execbuffer_reserve(struct intel_engine_cs *ring,
 			if (!drm_mm_node_allocated(&vma->node))
 				continue;
 
-			if (eb_vma_misplaced(vma, has_fenced_gpu_access))
+			if (eb_vma_misplaced(vma))
 				ret = i915_vma_unbind(vma);
 			else
 				ret = i915_gem_execbuffer_reserve_vma(vma, ring, need_relocs);
@@ -744,9 +721,6 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	int i, total, ret;
 	unsigned count = args->buffer_count;
 
-	if (WARN_ON(list_empty(&eb->vmas)))
-		return 0;
-
 	vm = list_first_entry(&eb->vmas, struct i915_vma, exec_list)->vm;
 
 	/* We may process another execbuffer during the unlock... */
@@ -890,18 +864,24 @@ i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec)
 }
 
 static int
-validate_exec_list(struct drm_i915_gem_exec_object2 *exec,
+validate_exec_list(struct drm_device *dev,
+		   struct drm_i915_gem_exec_object2 *exec,
 		   int count)
 {
-	int i;
 	unsigned relocs_total = 0;
 	unsigned relocs_max = UINT_MAX / sizeof(struct drm_i915_gem_relocation_entry);
+	unsigned invalid_flags;
+	int i;
+
+	invalid_flags = __EXEC_OBJECT_UNKNOWN_FLAGS;
+	if (USES_FULL_PPGTT(dev))
+		invalid_flags |= EXEC_OBJECT_NEEDS_GTT;
 
 	for (i = 0; i < count; i++) {
 		char __user *ptr = to_user_ptr(exec[i].relocs_ptr);
 		int length; /* limited by fault_in_pages_readable() */
 
-		if (exec[i].flags & __EXEC_OBJECT_UNKNOWN_FLAGS)
+		if (exec[i].flags & invalid_flags)
 			return -EINVAL;
 
 		/* First check for malicious input causing overflow in
@@ -938,7 +918,7 @@ i915_gem_validate_context(struct drm_device *dev, struct drm_file *file,
 	struct intel_context *ctx = NULL;
 	struct i915_ctx_hang_stats *hs;
 
-	if (ring->id != RCS && ctx_id != DEFAULT_CONTEXT_ID)
+	if (ring->id != RCS && ctx_id != DEFAULT_CONTEXT_HANDLE)
 		return ERR_PTR(-EINVAL);
 
 	ctx = i915_gem_context_get(file->driver_priv, ctx_id);
@@ -951,16 +931,26 @@ i915_gem_validate_context(struct drm_device *dev, struct drm_file *file,
 		return ERR_PTR(-EIO);
 	}
 
+	if (i915.enable_execlists && !ctx->engine[ring->id].state) {
+		int ret = intel_lr_context_deferred_create(ctx, ring);
+		if (ret) {
+			DRM_DEBUG("Could not create LRC %u: %d\n", ctx_id, ret);
+			return ERR_PTR(ret);
+		}
+	}
+
 	return ctx;
 }
 
-static void
+void
 i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 				   struct intel_engine_cs *ring)
 {
+	u32 seqno = intel_ring_get_seqno(ring);
 	struct i915_vma *vma;
 
 	list_for_each_entry(vma, vmas, exec_list) {
+		struct drm_i915_gem_exec_object2 *entry = vma->exec_entry;
 		struct drm_i915_gem_object *obj = vma->obj;
 		u32 old_read = obj->base.read_domains;
 		u32 old_write = obj->base.write_domain;
@@ -969,26 +959,31 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 		if (obj->base.write_domain == 0)
 			obj->base.pending_read_domains |= obj->base.read_domains;
 		obj->base.read_domains = obj->base.pending_read_domains;
-		obj->fenced_gpu_access = obj->pending_fenced_gpu_access;
 
 		i915_vma_move_to_active(vma, ring);
 		if (obj->base.write_domain) {
 			obj->dirty = 1;
-			obj->last_write_seqno = intel_ring_get_seqno(ring);
-			/* check for potential scanout */
-			if (i915_gem_obj_ggtt_bound(obj) &&
-			    i915_gem_obj_to_ggtt(obj)->pin_count)
-				intel_mark_fb_busy(obj, ring);
+			obj->last_write_seqno = seqno;
+
+			intel_fb_obj_invalidate(obj, ring);
 
 			/* update for the implicit flush after a batch */
 			obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
 		}
+		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
+			obj->last_fenced_seqno = seqno;
+			if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
+				struct drm_i915_private *dev_priv = to_i915(ring->dev);
+				list_move_tail(&dev_priv->fence_regs[obj->fence_reg].lru_list,
+					       &dev_priv->mm.fence_list);
+			}
+		}
 
 		trace_i915_gem_object_change_domain(obj, old_read, old_write);
 	}
 }
 
-static void
+void
 i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 				    struct drm_file *file,
 				    struct intel_engine_cs *ring,
@@ -1028,6 +1023,163 @@ i915_reset_gen7_sol_offsets(struct drm_device *dev,
 	return 0;
 }
 
+int
+i915_gem_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags)
+{
+	struct drm_clip_rect *cliprects = NULL;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u64 exec_len;
+	int instp_mode;
+	u32 instp_mask;
+	int i, ret = 0;
+
+	if (args->num_cliprects != 0) {
+		if (ring != &dev_priv->ring[RCS]) {
+			DRM_DEBUG("clip rectangles are only valid with the render ring\n");
+			return -EINVAL;
+		}
+
+		if (INTEL_INFO(dev)->gen >= 5) {
+			DRM_DEBUG("clip rectangles are only valid on pre-gen5\n");
+			return -EINVAL;
+		}
+
+		if (args->num_cliprects > UINT_MAX / sizeof(*cliprects)) {
+			DRM_DEBUG("execbuf with %u cliprects\n",
+				  args->num_cliprects);
+			return -EINVAL;
+		}
+
+		cliprects = kcalloc(args->num_cliprects,
+				    sizeof(*cliprects),
+				    GFP_KERNEL);
+		if (cliprects == NULL) {
+			ret = -ENOMEM;
+			goto error;
+		}
+
+		if (copy_from_user(cliprects,
+				   to_user_ptr(args->cliprects_ptr),
+				   sizeof(*cliprects)*args->num_cliprects)) {
+			ret = -EFAULT;
+			goto error;
+		}
+	} else {
+		if (args->DR4 == 0xffffffff) {
+			DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
+			args->DR4 = 0;
+		}
+
+		if (args->DR1 || args->DR4 || args->cliprects_ptr) {
+			DRM_DEBUG("0 cliprects but dirt in cliprects fields\n");
+			return -EINVAL;
+		}
+	}
+
+	ret = i915_gem_execbuffer_move_to_gpu(ring, vmas);
+	if (ret)
+		goto error;
+
+	ret = i915_switch_context(ring, ctx);
+	if (ret)
+		goto error;
+
+	instp_mode = args->flags & I915_EXEC_CONSTANTS_MASK;
+	instp_mask = I915_EXEC_CONSTANTS_MASK;
+	switch (instp_mode) {
+	case I915_EXEC_CONSTANTS_REL_GENERAL:
+	case I915_EXEC_CONSTANTS_ABSOLUTE:
+	case I915_EXEC_CONSTANTS_REL_SURFACE:
+		if (instp_mode != 0 && ring != &dev_priv->ring[RCS]) {
+			DRM_DEBUG("non-0 rel constants mode on non-RCS\n");
+			ret = -EINVAL;
+			goto error;
+		}
+
+		if (instp_mode != dev_priv->relative_constants_mode) {
+			if (INTEL_INFO(dev)->gen < 4) {
+				DRM_DEBUG("no rel constants on pre-gen4\n");
+				ret = -EINVAL;
+				goto error;
+			}
+
+			if (INTEL_INFO(dev)->gen > 5 &&
+			    instp_mode == I915_EXEC_CONSTANTS_REL_SURFACE) {
+				DRM_DEBUG("rel surface constants mode invalid on gen5+\n");
+				ret = -EINVAL;
+				goto error;
+			}
+
+			/* The HW changed the meaning on this bit on gen6 */
+			if (INTEL_INFO(dev)->gen >= 6)
+				instp_mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
+		}
+		break;
+	default:
+		DRM_DEBUG("execbuf with unknown constants: %d\n", instp_mode);
+		ret = -EINVAL;
+		goto error;
+	}
+
+	if (ring == &dev_priv->ring[RCS] &&
+			instp_mode != dev_priv->relative_constants_mode) {
+		ret = intel_ring_begin(ring, 4);
+		if (ret)
+			goto error;
+
+		intel_ring_emit(ring, MI_NOOP);
+		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+		intel_ring_emit(ring, INSTPM);
+		intel_ring_emit(ring, instp_mask << 16 | instp_mode);
+		intel_ring_advance(ring);
+
+		dev_priv->relative_constants_mode = instp_mode;
+	}
+
+	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
+		ret = i915_reset_gen7_sol_offsets(dev, ring);
+		if (ret)
+			goto error;
+	}
+
+	exec_len = args->batch_len;
+	if (cliprects) {
+		for (i = 0; i < args->num_cliprects; i++) {
+			ret = i915_emit_box(dev, &cliprects[i],
+					    args->DR1, args->DR4);
+			if (ret)
+				goto error;
+
+			ret = ring->dispatch_execbuffer(ring,
+							exec_start, exec_len,
+							flags);
+			if (ret)
+				goto error;
+		}
+	} else {
+		ret = ring->dispatch_execbuffer(ring,
+						exec_start, exec_len,
+						flags);
+		if (ret)
+			return ret;
+	}
+
+	trace_i915_gem_ring_dispatch(ring, intel_ring_get_seqno(ring), flags);
+
+	i915_gem_execbuffer_move_to_active(vmas, ring);
+	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
+
+error:
+	kfree(cliprects);
+	return ret;
+}
+
 /**
  * Find one BSD ring to dispatch the corresponding BSD command.
  * The Ring ID is returned.
@@ -1087,20 +1239,19 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct eb_vmas *eb;
 	struct drm_i915_gem_object *batch_obj;
-	struct drm_clip_rect *cliprects = NULL;
 	struct intel_engine_cs *ring;
 	struct intel_context *ctx;
 	struct i915_address_space *vm;
 	const u32 ctx_id = i915_execbuffer2_get_context_id(*args);
-	u64 exec_start = args->batch_start_offset, exec_len;
-	u32 mask, flags;
-	int ret, mode, i;
+	u64 exec_start = args->batch_start_offset;
+	u32 flags;
+	int ret;
 	bool need_relocs;
 
 	if (!i915_gem_check_execbuffer(args))
 		return -EINVAL;
 
-	ret = validate_exec_list(exec, args->buffer_count);
+	ret = validate_exec_list(dev, exec, args->buffer_count);
 	if (ret)
 		return ret;
 
@@ -1138,87 +1289,11 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
-	mode = args->flags & I915_EXEC_CONSTANTS_MASK;
-	mask = I915_EXEC_CONSTANTS_MASK;
-	switch (mode) {
-	case I915_EXEC_CONSTANTS_REL_GENERAL:
-	case I915_EXEC_CONSTANTS_ABSOLUTE:
-	case I915_EXEC_CONSTANTS_REL_SURFACE:
-		if (mode != 0 && ring != &dev_priv->ring[RCS]) {
-			DRM_DEBUG("non-0 rel constants mode on non-RCS\n");
-			return -EINVAL;
-		}
-
-		if (mode != dev_priv->relative_constants_mode) {
-			if (INTEL_INFO(dev)->gen < 4) {
-				DRM_DEBUG("no rel constants on pre-gen4\n");
-				return -EINVAL;
-			}
-
-			if (INTEL_INFO(dev)->gen > 5 &&
-			    mode == I915_EXEC_CONSTANTS_REL_SURFACE) {
-				DRM_DEBUG("rel surface constants mode invalid on gen5+\n");
-				return -EINVAL;
-			}
-
-			/* The HW changed the meaning on this bit on gen6 */
-			if (INTEL_INFO(dev)->gen >= 6)
-				mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
-		}
-		break;
-	default:
-		DRM_DEBUG("execbuf with unknown constants: %d\n", mode);
-		return -EINVAL;
-	}
-
 	if (args->buffer_count < 1) {
 		DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
-	if (args->num_cliprects != 0) {
-		if (ring != &dev_priv->ring[RCS]) {
-			DRM_DEBUG("clip rectangles are only valid with the render ring\n");
-			return -EINVAL;
-		}
-
-		if (INTEL_INFO(dev)->gen >= 5) {
-			DRM_DEBUG("clip rectangles are only valid on pre-gen5\n");
-			return -EINVAL;
-		}
-
-		if (args->num_cliprects > UINT_MAX / sizeof(*cliprects)) {
-			DRM_DEBUG("execbuf with %u cliprects\n",
-				  args->num_cliprects);
-			return -EINVAL;
-		}
-
-		cliprects = kcalloc(args->num_cliprects,
-				    sizeof(*cliprects),
-				    GFP_KERNEL);
-		if (cliprects == NULL) {
-			ret = -ENOMEM;
-			goto pre_mutex_err;
-		}
-
-		if (copy_from_user(cliprects,
-				   to_user_ptr(args->cliprects_ptr),
-				   sizeof(*cliprects)*args->num_cliprects)) {
-			ret = -EFAULT;
-			goto pre_mutex_err;
-		}
-	} else {
-		if (args->DR4 == 0xffffffff) {
-			DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
-			args->DR4 = 0;
-		}
-
-		if (args->DR1 || args->DR4 || args->cliprects_ptr) {
-			DRM_DEBUG("0 cliprects but dirt in cliprects fields\n");
-			return -EINVAL;
-		}
-	}
-
 	intel_runtime_pm_get(dev_priv);
 
 	ret = i915_mutex_lock_interruptible(dev);
@@ -1240,8 +1315,9 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	i915_gem_context_reference(ctx);
 
-	vm = ctx->vm;
-	if (!USES_FULL_PPGTT(dev))
+	if (ctx->ppgtt)
+		vm = &ctx->ppgtt->base;
+	else
 		vm = &dev_priv->gtt.base;
 
 	eb = eb_create(args);
@@ -1308,77 +1384,36 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	/* snb/ivb/vlv conflate the "batch in ppgtt" bit with the "non-secure
 	 * batch" bit. Hence we need to pin secure batches into the global gtt.
 	 * hsw should have this fixed, but bdw mucks it up again. */
-	if (flags & I915_DISPATCH_SECURE &&
-	    !batch_obj->has_global_gtt_mapping) {
-		/* When we have multiple VMs, we'll need to make sure that we
-		 * allocate space first */
-		struct i915_vma *vma = i915_gem_obj_to_ggtt(batch_obj);
-		BUG_ON(!vma);
-		vma->bind_vma(vma, batch_obj->cache_level, GLOBAL_BIND);
-	}
-
-	if (flags & I915_DISPATCH_SECURE)
-		exec_start += i915_gem_obj_ggtt_offset(batch_obj);
-	else
-		exec_start += i915_gem_obj_offset(batch_obj, vm);
-
-	ret = i915_gem_execbuffer_move_to_gpu(ring, &eb->vmas);
-	if (ret)
-		goto err;
-
-	ret = i915_switch_context(ring, ctx);
-	if (ret)
-		goto err;
-
-	if (ring == &dev_priv->ring[RCS] &&
-	    mode != dev_priv->relative_constants_mode) {
-		ret = intel_ring_begin(ring, 4);
-		if (ret)
-				goto err;
-
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit(ring, INSTPM);
-		intel_ring_emit(ring, mask << 16 | mode);
-		intel_ring_advance(ring);
-
-		dev_priv->relative_constants_mode = mode;
-	}
-
-	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
-		ret = i915_reset_gen7_sol_offsets(dev, ring);
-		if (ret)
-			goto err;
-	}
-
-
-	exec_len = args->batch_len;
-	if (cliprects) {
-		for (i = 0; i < args->num_cliprects; i++) {
-			ret = i915_emit_box(dev, &cliprects[i],
-					    args->DR1, args->DR4);
-			if (ret)
-				goto err;
-
-			ret = ring->dispatch_execbuffer(ring,
-							exec_start, exec_len,
-							flags);
-			if (ret)
-				goto err;
-		}
-	} else {
-		ret = ring->dispatch_execbuffer(ring,
-						exec_start, exec_len,
-						flags);
+	if (flags & I915_DISPATCH_SECURE) {
+		/*
+		 * So on first glance it looks freaky that we pin the batch here
+		 * outside of the reservation loop. But:
+		 * - The batch is already pinned into the relevant ppgtt, so we
+		 *   already have the backing storage fully allocated.
+		 * - No other BO uses the global gtt (well contexts, but meh),
+		 *   so we don't really have issues with mutliple objects not
+		 *   fitting due to fragmentation.
+		 * So this is actually safe.
+		 */
+		ret = i915_gem_obj_ggtt_pin(batch_obj, 0, 0);
 		if (ret)
 			goto err;
-	}
 
-	trace_i915_gem_ring_dispatch(ring, intel_ring_get_seqno(ring), flags);
+		exec_start += i915_gem_obj_ggtt_offset(batch_obj);
+	} else
+		exec_start += i915_gem_obj_offset(batch_obj, vm);
 
-	i915_gem_execbuffer_move_to_active(&eb->vmas, ring);
-	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
+	ret = dev_priv->gt.do_execbuf(dev, file, ring, ctx, args,
+				      &eb->vmas, batch_obj, exec_start, flags);
 
+	/*
+	 * FIXME: We crucially rely upon the active tracking for the (ppgtt)
+	 * batch vma for correctness. For less ugly and less fragility this
+	 * needs to be adjusted to also track the ggtt batch vma properly as
+	 * active.
+	 */
+	if (flags & I915_DISPATCH_SECURE)
+		i915_gem_object_ggtt_unpin(batch_obj);
 err:
 	/* the request owns the ref now */
 	i915_gem_context_unreference(ctx);
@@ -1387,8 +1422,6 @@ err:
 	mutex_unlock(&dev->struct_mutex);
 
 pre_mutex_err:
-	kfree(cliprects);
-
 	/* intel_gpu_busy should also get a ref, so it will free when the device
 	 * is really idle. */
 	intel_runtime_pm_put(dev_priv);
@@ -1525,7 +1558,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 	ret = i915_gem_do_execbuffer(dev, data, file, args, exec2_list);
 	if (!ret) {
 		/* Copy the new buffer offsets back to the user's exec list. */
-		struct drm_i915_gem_exec_object2 *user_exec_list =
+		struct drm_i915_gem_exec_object2 __user *user_exec_list =
 				   to_user_ptr(args->buffers_ptr);
 		int i;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 8b3cde703364..728938f02341 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -33,17 +33,6 @@
 static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv);
 static void chv_setup_private_ppat(struct drm_i915_private *dev_priv);
 
-bool intel_enable_ppgtt(struct drm_device *dev, bool full)
-{
-	if (i915.enable_ppgtt == 0)
-		return false;
-
-	if (i915.enable_ppgtt == 1 && full)
-		return false;
-
-	return true;
-}
-
 static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt)
 {
 	if (enable_ppgtt == 0 || !HAS_ALIASING_PPGTT(dev))
@@ -63,6 +52,13 @@ static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt)
 	}
 #endif
 
+	/* Early VLV doesn't have this */
+	if (IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev) &&
+	    dev->pdev->revision < 0xb) {
+		DRM_DEBUG_DRIVER("disabling PPGTT on pre-B3 step VLV\n");
+		return 0;
+	}
+
 	return HAS_ALIASING_PPGTT(dev) ? 1 : 0;
 }
 
@@ -71,7 +67,6 @@ static void ppgtt_bind_vma(struct i915_vma *vma,
 			   enum i915_cache_level cache_level,
 			   u32 flags);
 static void ppgtt_unbind_vma(struct i915_vma *vma);
-static int gen8_ppgtt_enable(struct i915_hw_ppgtt *ppgtt);
 
 static inline gen8_gtt_pte_t gen8_pte_encode(dma_addr_t addr,
 					     enum i915_cache_level level,
@@ -110,7 +105,7 @@ static inline gen8_ppgtt_pde_t gen8_pde_encode(struct drm_device *dev,
 
 static gen6_gtt_pte_t snb_pte_encode(dma_addr_t addr,
 				     enum i915_cache_level level,
-				     bool valid)
+				     bool valid, u32 unused)
 {
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
@@ -132,7 +127,7 @@ static gen6_gtt_pte_t snb_pte_encode(dma_addr_t addr,
 
 static gen6_gtt_pte_t ivb_pte_encode(dma_addr_t addr,
 				     enum i915_cache_level level,
-				     bool valid)
+				     bool valid, u32 unused)
 {
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
@@ -156,7 +151,7 @@ static gen6_gtt_pte_t ivb_pte_encode(dma_addr_t addr,
 
 static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr,
 				     enum i915_cache_level level,
-				     bool valid)
+				     bool valid, u32 flags)
 {
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
@@ -164,7 +159,8 @@ static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr,
 	/* Mark the page as writeable.  Other platforms don't have a
 	 * setting for read-only/writable, so this matches that behavior.
 	 */
-	pte |= BYT_PTE_WRITEABLE;
+	if (!(flags & PTE_READ_ONLY))
+		pte |= BYT_PTE_WRITEABLE;
 
 	if (level != I915_CACHE_NONE)
 		pte |= BYT_PTE_SNOOPED_BY_CPU_CACHES;
@@ -174,7 +170,7 @@ static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr,
 
 static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
 				     enum i915_cache_level level,
-				     bool valid)
+				     bool valid, u32 unused)
 {
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= HSW_PTE_ADDR_ENCODE(addr);
@@ -187,7 +183,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr,
 
 static gen6_gtt_pte_t iris_pte_encode(dma_addr_t addr,
 				      enum i915_cache_level level,
-				      bool valid)
+				      bool valid, u32 unused)
 {
 	gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0;
 	pte |= HSW_PTE_ADDR_ENCODE(addr);
@@ -208,19 +204,12 @@ static gen6_gtt_pte_t iris_pte_encode(dma_addr_t addr,
 
 /* Broadwell Page Directory Pointer Descriptors */
 static int gen8_write_pdp(struct intel_engine_cs *ring, unsigned entry,
-			   uint64_t val, bool synchronous)
+			   uint64_t val)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	int ret;
 
 	BUG_ON(entry >= 4);
 
-	if (synchronous) {
-		I915_WRITE(GEN8_RING_PDP_UDW(ring, entry), val >> 32);
-		I915_WRITE(GEN8_RING_PDP_LDW(ring, entry), (u32)val);
-		return 0;
-	}
-
 	ret = intel_ring_begin(ring, 6);
 	if (ret)
 		return ret;
@@ -237,8 +226,7 @@ static int gen8_write_pdp(struct intel_engine_cs *ring, unsigned entry,
 }
 
 static int gen8_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
 	int i, ret;
 
@@ -247,7 +235,7 @@ static int gen8_mm_switch(struct i915_hw_ppgtt *ppgtt,
 
 	for (i = used_pd - 1; i >= 0; i--) {
 		dma_addr_t addr = ppgtt->pd_dma_addr[i];
-		ret = gen8_write_pdp(ring, i, addr, synchronous);
+		ret = gen8_write_pdp(ring, i, addr);
 		if (ret)
 			return ret;
 	}
@@ -301,7 +289,7 @@ static void gen8_ppgtt_clear_range(struct i915_address_space *vm,
 static void gen8_ppgtt_insert_entries(struct i915_address_space *vm,
 				      struct sg_table *pages,
 				      uint64_t start,
-				      enum i915_cache_level cache_level)
+				      enum i915_cache_level cache_level, u32 unused)
 {
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
@@ -395,9 +383,6 @@ static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
 
-	list_del(&vm->global_link);
-	drm_mm_takedown(&vm->mm);
-
 	gen8_ppgtt_unmap_pages(ppgtt);
 	gen8_ppgtt_free(ppgtt);
 }
@@ -607,7 +592,6 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt, uint64_t size)
 		kunmap_atomic(pd_vaddr);
 	}
 
-	ppgtt->enable = gen8_ppgtt_enable;
 	ppgtt->switch_mm = gen8_mm_switch;
 	ppgtt->base.clear_range = gen8_ppgtt_clear_range;
 	ppgtt->base.insert_entries = gen8_ppgtt_insert_entries;
@@ -639,7 +623,7 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 	uint32_t pd_entry;
 	int pte, pde;
 
-	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true);
+	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true, 0);
 
 	pd_addr = (gen6_gtt_pte_t __iomem *)dev_priv->gtt.gsm +
 		ppgtt->pd_offset / sizeof(gen6_gtt_pte_t);
@@ -716,29 +700,10 @@ static uint32_t get_pd_offset(struct i915_hw_ppgtt *ppgtt)
 }
 
 static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			 struct intel_engine_cs *ring,
-			 bool synchronous)
+			 struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/* If we're in reset, we can assume the GPU is sufficiently idle to
-	 * manually frob these bits. Ideally we could use the ring functions,
-	 * except our error handling makes it quite difficult (can't use
-	 * intel_ring_begin, ring->flush, or intel_ring_advance)
-	 *
-	 * FIXME: We should try not to special case reset
-	 */
-	if (synchronous ||
-	    i915_reset_in_progress(&dev_priv->gpu_error)) {
-		WARN_ON(ppgtt != dev_priv->mm.aliasing_ppgtt);
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
-		POSTING_READ(RING_PP_DIR_BASE(ring));
-		return 0;
-	}
-
 	/* NB: TLBs must be flushed and invalidated before a switch */
 	ret = ring->flush(ring, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
@@ -760,29 +725,10 @@ static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 }
 
 static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = ppgtt->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	/* If we're in reset, we can assume the GPU is sufficiently idle to
-	 * manually frob these bits. Ideally we could use the ring functions,
-	 * except our error handling makes it quite difficult (can't use
-	 * intel_ring_begin, ring->flush, or intel_ring_advance)
-	 *
-	 * FIXME: We should try not to special case reset
-	 */
-	if (synchronous ||
-	    i915_reset_in_progress(&dev_priv->gpu_error)) {
-		WARN_ON(ppgtt != dev_priv->mm.aliasing_ppgtt);
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
-		POSTING_READ(RING_PP_DIR_BASE(ring));
-		return 0;
-	}
-
 	/* NB: TLBs must be flushed and invalidated before a switch */
 	ret = ring->flush(ring, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
 	if (ret)
@@ -811,14 +757,11 @@ static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 }
 
 static int gen6_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			  struct intel_engine_cs *ring,
-			  bool synchronous)
+			  struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (!synchronous)
-		return 0;
 
 	I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
 	I915_WRITE(RING_PP_DIR_BASE(ring), get_pd_offset(ppgtt));
@@ -828,39 +771,20 @@ static int gen6_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	return 0;
 }
 
-static int gen8_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen8_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
-	int j, ret;
+	int j;
 
 	for_each_ring(ring, dev_priv, j) {
 		I915_WRITE(RING_MODE_GEN7(ring),
 			   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-		/* We promise to do a switch later with FULL PPGTT. If this is
-		 * aliasing, this is the one and only switch we'll do */
-		if (USES_FULL_PPGTT(dev))
-			continue;
-
-		ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			goto err_out;
 	}
-
-	return 0;
-
-err_out:
-	for_each_ring(ring, dev_priv, j)
-		I915_WRITE(RING_MODE_GEN7(ring),
-			   _MASKED_BIT_DISABLE(GFX_PPGTT_ENABLE));
-	return ret;
 }
 
-static int gen7_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen7_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
 	uint32_t ecochk, ecobits;
@@ -879,31 +803,16 @@ static int gen7_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
 	I915_WRITE(GAM_ECOCHK, ecochk);
 
 	for_each_ring(ring, dev_priv, i) {
-		int ret;
 		/* GFX_MODE is per-ring on gen7+ */
 		I915_WRITE(RING_MODE_GEN7(ring),
 			   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-		/* We promise to do a switch later with FULL PPGTT. If this is
-		 * aliasing, this is the one and only switch we'll do */
-		if (USES_FULL_PPGTT(dev))
-			continue;
-
-		ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			return ret;
 	}
-
-	return 0;
 }
 
-static int gen6_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
+static void gen6_ppgtt_enable(struct drm_device *dev)
 {
-	struct drm_device *dev = ppgtt->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_engine_cs *ring;
 	uint32_t ecochk, gab_ctl, ecobits;
-	int i;
 
 	ecobits = I915_READ(GAC_ECO_BITS);
 	I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_SNB_BIT |
@@ -916,14 +825,6 @@ static int gen6_ppgtt_enable(struct i915_hw_ppgtt *ppgtt)
 	I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT | ECOCHK_PPGTT_CACHE64B);
 
 	I915_WRITE(GFX_MODE, _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
-
-	for_each_ring(ring, dev_priv, i) {
-		int ret = ppgtt->switch_mm(ppgtt, ring, true);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
 }
 
 /* PPGTT support for Sandybdrige/Gen6 and later */
@@ -941,7 +842,7 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
 	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
 	unsigned last_pte, i;
 
-	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true);
+	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true, 0);
 
 	while (num_entries) {
 		last_pte = first_pte + num_entries;
@@ -964,7 +865,7 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
 static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
 				      struct sg_table *pages,
 				      uint64_t start,
-				      enum i915_cache_level cache_level)
+				      enum i915_cache_level cache_level, u32 flags)
 {
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
@@ -981,7 +882,8 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
 
 		pt_vaddr[act_pte] =
 			vm->pte_encode(sg_page_iter_dma_address(&sg_iter),
-				       cache_level, true);
+				       cache_level, true, flags);
+
 		if (++act_pte == I915_PPGTT_PT_ENTRIES) {
 			kunmap_atomic(pt_vaddr);
 			pt_vaddr = NULL;
@@ -1020,8 +922,6 @@ static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
 	struct i915_hw_ppgtt *ppgtt =
 		container_of(vm, struct i915_hw_ppgtt, base);
 
-	list_del(&vm->global_link);
-	drm_mm_takedown(&ppgtt->base.mm);
 	drm_mm_remove_node(&ppgtt->node);
 
 	gen6_ppgtt_unmap_pages(ppgtt);
@@ -1142,13 +1042,10 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 
 	ppgtt->base.pte_encode = dev_priv->gtt.base.pte_encode;
 	if (IS_GEN6(dev)) {
-		ppgtt->enable = gen6_ppgtt_enable;
 		ppgtt->switch_mm = gen6_mm_switch;
 	} else if (IS_HASWELL(dev)) {
-		ppgtt->enable = gen7_ppgtt_enable;
 		ppgtt->switch_mm = hsw_mm_switch;
 	} else if (IS_GEN7(dev)) {
-		ppgtt->enable = gen7_ppgtt_enable;
 		ppgtt->switch_mm = gen7_mm_switch;
 	} else
 		BUG();
@@ -1179,47 +1076,126 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 			 ppgtt->node.size >> 20,
 			 ppgtt->node.start / PAGE_SIZE);
 
+	gen6_write_pdes(ppgtt);
+	DRM_DEBUG("Adding PPGTT at offset %x\n",
+		  ppgtt->pd_offset << 10);
+
 	return 0;
 }
 
-int i915_gem_init_ppgtt(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
+static int __hw_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret = 0;
 
 	ppgtt->base.dev = dev;
 	ppgtt->base.scratch = dev_priv->gtt.base.scratch;
 
 	if (INTEL_INFO(dev)->gen < 8)
-		ret = gen6_ppgtt_init(ppgtt);
+		return gen6_ppgtt_init(ppgtt);
 	else if (IS_GEN8(dev))
-		ret = gen8_ppgtt_init(ppgtt, dev_priv->gtt.base.total);
+		return gen8_ppgtt_init(ppgtt, dev_priv->gtt.base.total);
 	else
 		BUG();
+}
+int i915_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret = 0;
 
-	if (!ret) {
-		struct drm_i915_private *dev_priv = dev->dev_private;
+	ret = __hw_ppgtt_init(dev, ppgtt);
+	if (ret == 0) {
 		kref_init(&ppgtt->ref);
 		drm_mm_init(&ppgtt->base.mm, ppgtt->base.start,
 			    ppgtt->base.total);
 		i915_init_vm(dev_priv, &ppgtt->base);
-		if (INTEL_INFO(dev)->gen < 8) {
-			gen6_write_pdes(ppgtt);
-			DRM_DEBUG("Adding PPGTT at offset %x\n",
-				  ppgtt->pd_offset << 10);
+	}
+
+	return ret;
+}
+
+int i915_ppgtt_init_hw(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	int i, ret = 0;
+
+	/* In the case of execlists, PPGTT is enabled by the context descriptor
+	 * and the PDPs are contained within the context itself.  We don't
+	 * need to do anything here. */
+	if (i915.enable_execlists)
+		return 0;
+
+	if (!USES_PPGTT(dev))
+		return 0;
+
+	if (IS_GEN6(dev))
+		gen6_ppgtt_enable(dev);
+	else if (IS_GEN7(dev))
+		gen7_ppgtt_enable(dev);
+	else if (INTEL_INFO(dev)->gen >= 8)
+		gen8_ppgtt_enable(dev);
+	else
+		WARN_ON(1);
+
+	if (ppgtt) {
+		for_each_ring(ring, dev_priv, i) {
+			ret = ppgtt->switch_mm(ppgtt, ring);
+			if (ret != 0)
+				return ret;
 		}
 	}
 
 	return ret;
 }
+struct i915_hw_ppgtt *
+i915_ppgtt_create(struct drm_device *dev, struct drm_i915_file_private *fpriv)
+{
+	struct i915_hw_ppgtt *ppgtt;
+	int ret;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return ERR_PTR(-ENOMEM);
+
+	ret = i915_ppgtt_init(dev, ppgtt);
+	if (ret) {
+		kfree(ppgtt);
+		return ERR_PTR(ret);
+	}
+
+	ppgtt->file_priv = fpriv;
+
+	return ppgtt;
+}
+
+void  i915_ppgtt_release(struct kref *kref)
+{
+	struct i915_hw_ppgtt *ppgtt =
+		container_of(kref, struct i915_hw_ppgtt, ref);
+
+	/* vmas should already be unbound */
+	WARN_ON(!list_empty(&ppgtt->base.active_list));
+	WARN_ON(!list_empty(&ppgtt->base.inactive_list));
+
+	list_del(&ppgtt->base.global_link);
+	drm_mm_takedown(&ppgtt->base.mm);
+
+	ppgtt->base.cleanup(&ppgtt->base);
+	kfree(ppgtt);
+}
 
 static void
 ppgtt_bind_vma(struct i915_vma *vma,
 	       enum i915_cache_level cache_level,
 	       u32 flags)
 {
+	/* Currently applicable only to VLV */
+	if (vma->obj->gt_ro)
+		flags |= PTE_READ_ONLY;
+
 	vma->vm->insert_entries(vma->vm, vma->obj->pages, vma->node.start,
-				cache_level);
+				cache_level, flags);
 }
 
 static void ppgtt_unbind_vma(struct i915_vma *vma)
@@ -1297,6 +1273,16 @@ void i915_check_and_clear_faults(struct drm_device *dev)
 	POSTING_READ(RING_FAULT_REG(&dev_priv->ring[RCS]));
 }
 
+static void i915_ggtt_flush(struct drm_i915_private *dev_priv)
+{
+	if (INTEL_INFO(dev_priv->dev)->gen < 6) {
+		intel_gtt_chipset_flush();
+	} else {
+		I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
+		POSTING_READ(GFX_FLSH_CNTL_GEN6);
+	}
+}
+
 void i915_gem_suspend_gtt_mappings(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1313,6 +1299,8 @@ void i915_gem_suspend_gtt_mappings(struct drm_device *dev)
 				       dev_priv->gtt.base.start,
 				       dev_priv->gtt.base.total,
 				       true);
+
+	i915_ggtt_flush(dev_priv);
 }
 
 void i915_gem_restore_gtt_mappings(struct drm_device *dev)
@@ -1365,7 +1353,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 		gen6_write_pdes(container_of(vm, struct i915_hw_ppgtt, base));
 	}
 
-	i915_gem_chipset_flush(dev);
+	i915_ggtt_flush(dev_priv);
 }
 
 int i915_gem_gtt_prepare_object(struct drm_i915_gem_object *obj)
@@ -1394,7 +1382,7 @@ static inline void gen8_set_pte(void __iomem *addr, gen8_gtt_pte_t pte)
 static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 				     struct sg_table *st,
 				     uint64_t start,
-				     enum i915_cache_level level)
+				     enum i915_cache_level level, u32 unused)
 {
 	struct drm_i915_private *dev_priv = vm->dev->dev_private;
 	unsigned first_entry = start >> PAGE_SHIFT;
@@ -1402,7 +1390,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 		(gen8_gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
 	int i = 0;
 	struct sg_page_iter sg_iter;
-	dma_addr_t addr = 0;
+	dma_addr_t addr = 0; /* shut up gcc */
 
 	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
 		addr = sg_dma_address(sg_iter.sg) +
@@ -1440,7 +1428,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 				     struct sg_table *st,
 				     uint64_t start,
-				     enum i915_cache_level level)
+				     enum i915_cache_level level, u32 flags)
 {
 	struct drm_i915_private *dev_priv = vm->dev->dev_private;
 	unsigned first_entry = start >> PAGE_SHIFT;
@@ -1448,11 +1436,11 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 		(gen6_gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
 	int i = 0;
 	struct sg_page_iter sg_iter;
-	dma_addr_t addr;
+	dma_addr_t addr = 0;
 
 	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
 		addr = sg_page_iter_dma_address(&sg_iter);
-		iowrite32(vm->pte_encode(addr, level, true), &gtt_entries[i]);
+		iowrite32(vm->pte_encode(addr, level, true, flags), &gtt_entries[i]);
 		i++;
 	}
 
@@ -1462,9 +1450,10 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 	 * of NUMA access patterns. Therefore, even with the way we assume
 	 * hardware should work, we must keep this posting read for paranoia.
 	 */
-	if (i != 0)
-		WARN_ON(readl(&gtt_entries[i-1]) !=
-			vm->pte_encode(addr, level, true));
+	if (i != 0) {
+		unsigned long gtt = readl(&gtt_entries[i-1]);
+		WARN_ON(gtt != vm->pte_encode(addr, level, true, flags));
+	}
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -1518,7 +1507,7 @@ static void gen6_ggtt_clear_range(struct i915_address_space *vm,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, use_scratch);
+	scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, use_scratch, 0);
 
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
@@ -1567,6 +1556,10 @@ static void ggtt_bind_vma(struct i915_vma *vma,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_object *obj = vma->obj;
 
+	/* Currently applicable only to VLV */
+	if (obj->gt_ro)
+		flags |= PTE_READ_ONLY;
+
 	/* If there is no aliasing PPGTT, or the caller needs a global mapping,
 	 * or we have a global mapping already but the cacheability flags have
 	 * changed, set the global PTEs.
@@ -1583,7 +1576,7 @@ static void ggtt_bind_vma(struct i915_vma *vma,
 		    (cache_level != obj->cache_level)) {
 			vma->vm->insert_entries(vma->vm, obj->pages,
 						vma->node.start,
-						cache_level);
+						cache_level, flags);
 			obj->has_global_gtt_mapping = 1;
 		}
 	}
@@ -1595,7 +1588,7 @@ static void ggtt_bind_vma(struct i915_vma *vma,
 		appgtt->base.insert_entries(&appgtt->base,
 					    vma->obj->pages,
 					    vma->node.start,
-					    cache_level);
+					    cache_level, flags);
 		vma->obj->has_aliasing_ppgtt_mapping = 1;
 	}
 }
@@ -1657,10 +1650,10 @@ static void i915_gtt_color_adjust(struct drm_mm_node *node,
 	}
 }
 
-void i915_gem_setup_global_gtt(struct drm_device *dev,
-			       unsigned long start,
-			       unsigned long mappable_end,
-			       unsigned long end)
+int i915_gem_setup_global_gtt(struct drm_device *dev,
+			      unsigned long start,
+			      unsigned long mappable_end,
+			      unsigned long end)
 {
 	/* Let GEM Manage all of the aperture.
 	 *
@@ -1676,6 +1669,7 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	struct drm_mm_node *entry;
 	struct drm_i915_gem_object *obj;
 	unsigned long hole_start, hole_end;
+	int ret;
 
 	BUG_ON(mappable_end > end);
 
@@ -1687,14 +1681,16 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 	/* Mark any preallocated objects as occupied */
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		struct i915_vma *vma = i915_gem_obj_to_vma(obj, ggtt_vm);
-		int ret;
+
 		DRM_DEBUG_KMS("reserving preallocated space: %lx + %zx\n",
 			      i915_gem_obj_ggtt_offset(obj), obj->base.size);
 
 		WARN_ON(i915_gem_obj_ggtt_bound(obj));
 		ret = drm_mm_reserve_node(&ggtt_vm->mm, &vma->node);
-		if (ret)
-			DRM_DEBUG_KMS("Reservation failed\n");
+		if (ret) {
+			DRM_DEBUG_KMS("Reservation failed: %i\n", ret);
+			return ret;
+		}
 		obj->has_global_gtt_mapping = 1;
 	}
 
@@ -1711,6 +1707,22 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 
 	/* And finally clear the reserved guard page */
 	ggtt_vm->clear_range(ggtt_vm, end - PAGE_SIZE, PAGE_SIZE, true);
+
+	if (USES_PPGTT(dev) && !USES_FULL_PPGTT(dev)) {
+		struct i915_hw_ppgtt *ppgtt;
+
+		ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+		if (!ppgtt)
+			return -ENOMEM;
+
+		ret = __hw_ppgtt_init(dev, ppgtt);
+		if (ret != 0)
+			return ret;
+
+		dev_priv->mm.aliasing_ppgtt = ppgtt;
+	}
+
+	return 0;
 }
 
 void i915_gem_init_global_gtt(struct drm_device *dev)
@@ -1724,6 +1736,25 @@ void i915_gem_init_global_gtt(struct drm_device *dev)
 	i915_gem_setup_global_gtt(dev, 0, mappable_size, gtt_size);
 }
 
+void i915_global_gtt_cleanup(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_address_space *vm = &dev_priv->gtt.base;
+
+	if (dev_priv->mm.aliasing_ppgtt) {
+		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+
+		ppgtt->base.cleanup(&ppgtt->base);
+	}
+
+	if (drm_mm_initialized(&vm->mm)) {
+		drm_mm_takedown(&vm->mm);
+		list_del(&vm->global_link);
+	}
+
+	vm->cleanup(vm);
+}
+
 static int setup_scratch_page(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1871,6 +1902,22 @@ static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv)
 	      GEN8_PPAT(6, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(2)) |
 	      GEN8_PPAT(7, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(3));
 
+	if (!USES_PPGTT(dev_priv->dev))
+		/* Spec: "For GGTT, there is NO pat_sel[2:0] from the entry,
+		 * so RTL will always use the value corresponding to
+		 * pat_sel = 000".
+		 * So let's disable cache for GGTT to avoid screen corruptions.
+		 * MOCS still can be used though.
+		 * - System agent ggtt writes (i.e. cpu gtt mmaps) already work
+		 * before this patch, i.e. the same uncached + snooping access
+		 * like on gen6/7 seems to be in effect.
+		 * - So this just fixes blitter/render access. Again it looks
+		 * like it's not just uncached access, but uncached + snooping.
+		 * So we can still hold onto all our assumptions wrt cpu
+		 * clflushing on LLC machines.
+		 */
+		pat = GEN8_PPAT(0, GEN8_PPAT_UC);
+
 	/* XXX: spec defines this as 2 distinct registers. It's unclear if a 64b
 	 * write would work. */
 	I915_WRITE(GEN8_PRIVATE_PAT, pat);
@@ -1992,10 +2039,6 @@ static void gen6_gmch_remove(struct i915_address_space *vm)
 
 	struct i915_gtt *gtt = container_of(vm, struct i915_gtt, base);
 
-	if (drm_mm_initialized(&vm->mm)) {
-		drm_mm_takedown(&vm->mm);
-		list_del(&vm->global_link);
-	}
 	iounmap(gtt->gsm);
 	teardown_scratch_page(vm->dev);
 }
@@ -2028,10 +2071,6 @@ static int i915_gmch_probe(struct drm_device *dev,
 
 static void i915_gmch_remove(struct i915_address_space *vm)
 {
-	if (drm_mm_initialized(&vm->mm)) {
-		drm_mm_takedown(&vm->mm);
-		list_del(&vm->global_link);
-	}
 	intel_gmch_remove();
 }
 
@@ -2130,8 +2169,10 @@ static struct i915_vma *__i915_gem_vma_create(struct drm_i915_gem_object *obj,
 	/* Keep GGTT vmas first to make debug easier */
 	if (i915_is_ggtt(vm))
 		list_add(&vma->vma_link, &obj->vma_list);
-	else
+	else {
 		list_add_tail(&vma->vma_link, &obj->vma_list);
+		i915_ppgtt_get(i915_vm_to_ppgtt(vm));
+	}
 
 	return vma;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 1b96a06be3cb..d5c14af51e99 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -34,6 +34,8 @@
 #ifndef __I915_GEM_GTT_H__
 #define __I915_GEM_GTT_H__
 
+struct drm_i915_file_private;
+
 typedef uint32_t gen6_gtt_pte_t;
 typedef uint64_t gen8_gtt_pte_t;
 typedef gen8_gtt_pte_t gen8_ppgtt_pde_t;
@@ -154,6 +156,7 @@ struct i915_vma {
 	void (*unbind_vma)(struct i915_vma *vma);
 	/* Map an object into an address space with the given cache flags. */
 #define GLOBAL_BIND (1<<0)
+#define PTE_READ_ONLY (1<<1)
 	void (*bind_vma)(struct i915_vma *vma,
 			 enum i915_cache_level cache_level,
 			 u32 flags);
@@ -197,7 +200,7 @@ struct i915_address_space {
 	/* FIXME: Need a more generic return type */
 	gen6_gtt_pte_t (*pte_encode)(dma_addr_t addr,
 				     enum i915_cache_level level,
-				     bool valid); /* Create a valid PTE */
+				     bool valid, u32 flags); /* Create a valid PTE */
 	void (*clear_range)(struct i915_address_space *vm,
 			    uint64_t start,
 			    uint64_t length,
@@ -205,7 +208,7 @@ struct i915_address_space {
 	void (*insert_entries)(struct i915_address_space *vm,
 			       struct sg_table *st,
 			       uint64_t start,
-			       enum i915_cache_level cache_level);
+			       enum i915_cache_level cache_level, u32 flags);
 	void (*cleanup)(struct i915_address_space *vm);
 };
 
@@ -257,22 +260,36 @@ struct i915_hw_ppgtt {
 		dma_addr_t *gen8_pt_dma_addr[4];
 	};
 
-	struct intel_context *ctx;
+	struct drm_i915_file_private *file_priv;
 
 	int (*enable)(struct i915_hw_ppgtt *ppgtt);
 	int (*switch_mm)(struct i915_hw_ppgtt *ppgtt,
-			 struct intel_engine_cs *ring,
-			 bool synchronous);
+			 struct intel_engine_cs *ring);
 	void (*debug_dump)(struct i915_hw_ppgtt *ppgtt, struct seq_file *m);
 };
 
 int i915_gem_gtt_init(struct drm_device *dev);
 void i915_gem_init_global_gtt(struct drm_device *dev);
-void i915_gem_setup_global_gtt(struct drm_device *dev, unsigned long start,
-			       unsigned long mappable_end, unsigned long end);
-
-bool intel_enable_ppgtt(struct drm_device *dev, bool full);
-int i915_gem_init_ppgtt(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt);
+int i915_gem_setup_global_gtt(struct drm_device *dev, unsigned long start,
+			      unsigned long mappable_end, unsigned long end);
+void i915_global_gtt_cleanup(struct drm_device *dev);
+
+
+int i915_ppgtt_init(struct drm_device *dev, struct i915_hw_ppgtt *ppgtt);
+int i915_ppgtt_init_hw(struct drm_device *dev);
+void i915_ppgtt_release(struct kref *kref);
+struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_device *dev,
+					struct drm_i915_file_private *fpriv);
+static inline void i915_ppgtt_get(struct i915_hw_ppgtt *ppgtt)
+{
+	if (ppgtt)
+		kref_get(&ppgtt->ref);
+}
+static inline void i915_ppgtt_put(struct i915_hw_ppgtt *ppgtt)
+{
+	if (ppgtt)
+		kref_put(&ppgtt->ref, i915_ppgtt_release);
+}
 
 void i915_check_and_clear_faults(struct drm_device *dev);
 void i915_gem_suspend_gtt_mappings(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c
index 34894b573064..a9a62d75aa57 100644
--- a/drivers/gpu/drm/i915/i915_gem_render_state.c
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.c
@@ -28,64 +28,6 @@
 #include "i915_drv.h"
 #include "intel_renderstate.h"
 
-struct i915_render_state {
-	struct drm_i915_gem_object *obj;
-	unsigned long ggtt_offset;
-	u32 *batch;
-	u32 size;
-	u32 len;
-};
-
-static struct i915_render_state *render_state_alloc(struct drm_device *dev)
-{
-	struct i915_render_state *so;
-	struct page *page;
-	int ret;
-
-	so = kzalloc(sizeof(*so), GFP_KERNEL);
-	if (!so)
-		return ERR_PTR(-ENOMEM);
-
-	so->obj = i915_gem_alloc_object(dev, 4096);
-	if (so->obj == NULL) {
-		ret = -ENOMEM;
-		goto free;
-	}
-	so->size = 4096;
-
-	ret = i915_gem_obj_ggtt_pin(so->obj, 4096, 0);
-	if (ret)
-		goto free_gem;
-
-	BUG_ON(so->obj->pages->nents != 1);
-	page = sg_page(so->obj->pages->sgl);
-
-	so->batch = kmap(page);
-	if (!so->batch) {
-		ret = -ENOMEM;
-		goto unpin;
-	}
-
-	so->ggtt_offset = i915_gem_obj_ggtt_offset(so->obj);
-
-	return so;
-unpin:
-	i915_gem_object_ggtt_unpin(so->obj);
-free_gem:
-	drm_gem_object_unreference(&so->obj->base);
-free:
-	kfree(so);
-	return ERR_PTR(ret);
-}
-
-static void render_state_free(struct i915_render_state *so)
-{
-	kunmap(kmap_to_page(so->batch));
-	i915_gem_object_ggtt_unpin(so->obj);
-	drm_gem_object_unreference(&so->obj->base);
-	kfree(so);
-}
-
 static const struct intel_renderstate_rodata *
 render_state_get_rodata(struct drm_device *dev, const int gen)
 {
@@ -101,98 +43,137 @@ render_state_get_rodata(struct drm_device *dev, const int gen)
 	return NULL;
 }
 
-static int render_state_setup(const int gen,
-			      const struct intel_renderstate_rodata *rodata,
-			      struct i915_render_state *so)
+static int render_state_init(struct render_state *so, struct drm_device *dev)
 {
-	const u64 goffset = i915_gem_obj_ggtt_offset(so->obj);
-	u32 reloc_index = 0;
-	u32 * const d = so->batch;
-	unsigned int i = 0;
 	int ret;
 
-	if (!rodata || rodata->batch_items * 4 > so->size)
+	so->gen = INTEL_INFO(dev)->gen;
+	so->rodata = render_state_get_rodata(dev, so->gen);
+	if (so->rodata == NULL)
+		return 0;
+
+	if (so->rodata->batch_items * 4 > 4096)
 		return -EINVAL;
 
+	so->obj = i915_gem_alloc_object(dev, 4096);
+	if (so->obj == NULL)
+		return -ENOMEM;
+
+	ret = i915_gem_obj_ggtt_pin(so->obj, 4096, 0);
+	if (ret)
+		goto free_gem;
+
+	so->ggtt_offset = i915_gem_obj_ggtt_offset(so->obj);
+	return 0;
+
+free_gem:
+	drm_gem_object_unreference(&so->obj->base);
+	return ret;
+}
+
+static int render_state_setup(struct render_state *so)
+{
+	const struct intel_renderstate_rodata *rodata = so->rodata;
+	unsigned int i = 0, reloc_index = 0;
+	struct page *page;
+	u32 *d;
+	int ret;
+
 	ret = i915_gem_object_set_to_cpu_domain(so->obj, true);
 	if (ret)
 		return ret;
 
+	page = sg_page(so->obj->pages->sgl);
+	d = kmap(page);
+
 	while (i < rodata->batch_items) {
 		u32 s = rodata->batch[i];
 
-		if (reloc_index < rodata->reloc_items &&
-		    i * 4  == rodata->reloc[reloc_index]) {
-
-			s += goffset & 0xffffffff;
-
-			/* We keep batch offsets max 32bit */
-			if (gen >= 8) {
+		if (i * 4  == rodata->reloc[reloc_index]) {
+			u64 r = s + so->ggtt_offset;
+			s = lower_32_bits(r);
+			if (so->gen >= 8) {
 				if (i + 1 >= rodata->batch_items ||
 				    rodata->batch[i + 1] != 0)
 					return -EINVAL;
 
-				d[i] = s;
-				i++;
-				s = (goffset & 0xffffffff00000000ull) >> 32;
+				d[i++] = s;
+				s = upper_32_bits(r);
 			}
 
 			reloc_index++;
 		}
 
-		d[i] = s;
-		i++;
+		d[i++] = s;
 	}
+	kunmap(page);
 
 	ret = i915_gem_object_set_to_gtt_domain(so->obj, false);
 	if (ret)
 		return ret;
 
-	if (rodata->reloc_items != reloc_index) {
-		DRM_ERROR("not all relocs resolved, %d out of %d\n",
-			  reloc_index, rodata->reloc_items);
+	if (rodata->reloc[reloc_index] != -1) {
+		DRM_ERROR("only %d relocs resolved\n", reloc_index);
 		return -EINVAL;
 	}
 
-	so->len = rodata->batch_items * 4;
-
 	return 0;
 }
 
-int i915_gem_render_state_init(struct intel_engine_cs *ring)
+void i915_gem_render_state_fini(struct render_state *so)
+{
+	i915_gem_object_ggtt_unpin(so->obj);
+	drm_gem_object_unreference(&so->obj->base);
+}
+
+int i915_gem_render_state_prepare(struct intel_engine_cs *ring,
+				  struct render_state *so)
 {
-	const int gen = INTEL_INFO(ring->dev)->gen;
-	struct i915_render_state *so;
-	const struct intel_renderstate_rodata *rodata;
 	int ret;
 
 	if (WARN_ON(ring->id != RCS))
 		return -ENOENT;
 
-	rodata = render_state_get_rodata(ring->dev, gen);
-	if (rodata == NULL)
+	ret = render_state_init(so, ring->dev);
+	if (ret)
+		return ret;
+
+	if (so->rodata == NULL)
 		return 0;
 
-	so = render_state_alloc(ring->dev);
-	if (IS_ERR(so))
-		return PTR_ERR(so);
+	ret = render_state_setup(so);
+	if (ret) {
+		i915_gem_render_state_fini(so);
+		return ret;
+	}
+
+	return 0;
+}
+
+int i915_gem_render_state_init(struct intel_engine_cs *ring)
+{
+	struct render_state so;
+	int ret;
 
-	ret = render_state_setup(gen, rodata, so);
+	ret = i915_gem_render_state_prepare(ring, &so);
 	if (ret)
-		goto out;
+		return ret;
+
+	if (so.rodata == NULL)
+		return 0;
 
 	ret = ring->dispatch_execbuffer(ring,
-					i915_gem_obj_ggtt_offset(so->obj),
-					so->len,
+					so.ggtt_offset,
+					so.rodata->batch_items * 4,
 					I915_DISPATCH_SECURE);
 	if (ret)
 		goto out;
 
-	i915_vma_move_to_active(i915_gem_obj_to_ggtt(so->obj), ring);
+	i915_vma_move_to_active(i915_gem_obj_to_ggtt(so.obj), ring);
 
-	ret = __i915_add_request(ring, NULL, so->obj, NULL);
+	ret = __i915_add_request(ring, NULL, so.obj, NULL);
 	/* __i915_add_request moves object to inactive if it fails */
 out:
-	render_state_free(so);
+	i915_gem_render_state_fini(&so);
 	return ret;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.h b/drivers/gpu/drm/i915/i915_gem_render_state.h
new file mode 100644
index 000000000000..c44961ed3fad
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_render_state.h
@@ -0,0 +1,47 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _I915_GEM_RENDER_STATE_H_
+#define _I915_GEM_RENDER_STATE_H_
+
+#include <linux/types.h>
+
+struct intel_renderstate_rodata {
+	const u32 *reloc;
+	const u32 *batch;
+	const u32 batch_items;
+};
+
+struct render_state {
+	const struct intel_renderstate_rodata *rodata;
+	struct drm_i915_gem_object *obj;
+	u64 ggtt_offset;
+	int gen;
+};
+
+int i915_gem_render_state_init(struct intel_engine_cs *ring);
+void i915_gem_render_state_fini(struct render_state *so);
+int i915_gem_render_state_prepare(struct intel_engine_cs *ring,
+				  struct render_state *so);
+
+#endif /* _I915_GEM_RENDER_STATE_H_ */
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 7465ab0fd396..85fda6b803e4 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -147,30 +147,68 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 	return base;
 }
 
-static int i915_setup_compression(struct drm_device *dev, int size)
+static int find_compression_threshold(struct drm_device *dev,
+				      struct drm_mm_node *node,
+				      int size,
+				      int fb_cpp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mm_node *compressed_fb, *uninitialized_var(compressed_llb);
+	int compression_threshold = 1;
 	int ret;
 
-	compressed_fb = kzalloc(sizeof(*compressed_fb), GFP_KERNEL);
-	if (!compressed_fb)
-		goto err_llb;
+	/* HACK: This code depends on what we will do in *_enable_fbc. If that
+	 * code changes, this code needs to change as well.
+	 *
+	 * The enable_fbc code will attempt to use one of our 2 compression
+	 * thresholds, therefore, in that case, we only have 1 resort.
+	 */
 
-	/* Try to over-allocate to reduce reallocations and fragmentation */
-	ret = drm_mm_insert_node(&dev_priv->mm.stolen, compressed_fb,
+	/* Try to over-allocate to reduce reallocations and fragmentation. */
+	ret = drm_mm_insert_node(&dev_priv->mm.stolen, node,
 				 size <<= 1, 4096, DRM_MM_SEARCH_DEFAULT);
-	if (ret)
-		ret = drm_mm_insert_node(&dev_priv->mm.stolen, compressed_fb,
-					 size >>= 1, 4096,
-					 DRM_MM_SEARCH_DEFAULT);
-	if (ret)
+	if (ret == 0)
+		return compression_threshold;
+
+again:
+	/* HW's ability to limit the CFB is 1:4 */
+	if (compression_threshold > 4 ||
+	    (fb_cpp == 2 && compression_threshold == 2))
+		return 0;
+
+	ret = drm_mm_insert_node(&dev_priv->mm.stolen, node,
+				 size >>= 1, 4096,
+				 DRM_MM_SEARCH_DEFAULT);
+	if (ret && INTEL_INFO(dev)->gen <= 4) {
+		return 0;
+	} else if (ret) {
+		compression_threshold <<= 1;
+		goto again;
+	} else {
+		return compression_threshold;
+	}
+}
+
+static int i915_setup_compression(struct drm_device *dev, int size, int fb_cpp)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_mm_node *uninitialized_var(compressed_llb);
+	int ret;
+
+	ret = find_compression_threshold(dev, &dev_priv->fbc.compressed_fb,
+					 size, fb_cpp);
+	if (!ret)
 		goto err_llb;
+	else if (ret > 1) {
+		DRM_INFO("Reducing the compressed framebuffer size. This may lead to less power savings than a non-reduced-size. Try to increase stolen memory size if available in BIOS.\n");
+
+	}
+
+	dev_priv->fbc.threshold = ret;
 
 	if (HAS_PCH_SPLIT(dev))
-		I915_WRITE(ILK_DPFC_CB_BASE, compressed_fb->start);
+		I915_WRITE(ILK_DPFC_CB_BASE, dev_priv->fbc.compressed_fb.start);
 	else if (IS_GM45(dev)) {
-		I915_WRITE(DPFC_CB_BASE, compressed_fb->start);
+		I915_WRITE(DPFC_CB_BASE, dev_priv->fbc.compressed_fb.start);
 	} else {
 		compressed_llb = kzalloc(sizeof(*compressed_llb), GFP_KERNEL);
 		if (!compressed_llb)
@@ -184,13 +222,12 @@ static int i915_setup_compression(struct drm_device *dev, int size)
 		dev_priv->fbc.compressed_llb = compressed_llb;
 
 		I915_WRITE(FBC_CFB_BASE,
-			   dev_priv->mm.stolen_base + compressed_fb->start);
+			   dev_priv->mm.stolen_base + dev_priv->fbc.compressed_fb.start);
 		I915_WRITE(FBC_LL_BASE,
 			   dev_priv->mm.stolen_base + compressed_llb->start);
 	}
 
-	dev_priv->fbc.compressed_fb = compressed_fb;
-	dev_priv->fbc.size = size;
+	dev_priv->fbc.size = size / dev_priv->fbc.threshold;
 
 	DRM_DEBUG_KMS("reserved %d bytes of contiguous stolen space for FBC\n",
 		      size);
@@ -199,14 +236,13 @@ static int i915_setup_compression(struct drm_device *dev, int size)
 
 err_fb:
 	kfree(compressed_llb);
-	drm_mm_remove_node(compressed_fb);
+	drm_mm_remove_node(&dev_priv->fbc.compressed_fb);
 err_llb:
-	kfree(compressed_fb);
 	pr_info_once("drm: not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
 	return -ENOSPC;
 }
 
-int i915_gem_stolen_setup_compression(struct drm_device *dev, int size)
+int i915_gem_stolen_setup_compression(struct drm_device *dev, int size, int fb_cpp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -219,7 +255,7 @@ int i915_gem_stolen_setup_compression(struct drm_device *dev, int size)
 	/* Release any current block */
 	i915_gem_stolen_cleanup_compression(dev);
 
-	return i915_setup_compression(dev, size);
+	return i915_setup_compression(dev, size, fb_cpp);
 }
 
 void i915_gem_stolen_cleanup_compression(struct drm_device *dev)
@@ -229,10 +265,7 @@ void i915_gem_stolen_cleanup_compression(struct drm_device *dev)
 	if (dev_priv->fbc.size == 0)
 		return;
 
-	if (dev_priv->fbc.compressed_fb) {
-		drm_mm_remove_node(dev_priv->fbc.compressed_fb);
-		kfree(dev_priv->fbc.compressed_fb);
-	}
+	drm_mm_remove_node(&dev_priv->fbc.compressed_fb);
 
 	if (dev_priv->fbc.compressed_llb) {
 		drm_mm_remove_node(dev_priv->fbc.compressed_llb);
@@ -256,6 +289,7 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 int i915_gem_init_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 tmp;
 	int bios_reserved = 0;
 
 #ifdef CONFIG_INTEL_IOMMU
@@ -275,8 +309,16 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	DRM_DEBUG_KMS("found %zd bytes of stolen memory at %08lx\n",
 		      dev_priv->gtt.stolen_size, dev_priv->mm.stolen_base);
 
-	if (IS_VALLEYVIEW(dev))
-		bios_reserved = 1024*1024; /* top 1M on VLV/BYT */
+	if (INTEL_INFO(dev)->gen >= 8) {
+		tmp = I915_READ(GEN7_BIOS_RESERVED);
+		tmp >>= GEN8_BIOS_RESERVED_SHIFT;
+		tmp &= GEN8_BIOS_RESERVED_MASK;
+		bios_reserved = (1024*1024) << tmp;
+	} else if (IS_GEN7(dev)) {
+		tmp = I915_READ(GEN7_BIOS_RESERVED);
+		bios_reserved = tmp & GEN7_BIOS_RESERVED_256K ?
+			256*1024 : 1024*1024;
+	}
 
 	if (WARN_ON(bios_reserved > dev_priv->gtt.stolen_size))
 		return 0;
@@ -336,9 +378,20 @@ static void i915_gem_object_put_pages_stolen(struct drm_i915_gem_object *obj)
 	kfree(obj->pages);
 }
 
+
+static void
+i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
+{
+	if (obj->stolen) {
+		drm_mm_remove_node(obj->stolen);
+		kfree(obj->stolen);
+		obj->stolen = NULL;
+	}
+}
 static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
 	.get_pages = i915_gem_object_get_pages_stolen,
 	.put_pages = i915_gem_object_put_pages_stolen,
+	.release = i915_gem_object_release_stolen,
 };
 
 static struct drm_i915_gem_object *
@@ -496,13 +549,3 @@ err_out:
 	drm_gem_object_unreference(&obj->base);
 	return NULL;
 }
-
-void
-i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
-{
-	if (obj->stolen) {
-		drm_mm_remove_node(obj->stolen);
-		kfree(obj->stolen);
-		obj->stolen = NULL;
-	}
-}
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index cb150e8b4336..2b1eaa29ada4 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -91,7 +91,14 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 	uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
 	uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (INTEL_INFO(dev)->gen >= 8 || IS_VALLEYVIEW(dev)) {
+		/*
+		 * On BDW+, swizzling is not used. We leave the CPU memory
+		 * controller in charge of optimizing memory accesses without
+		 * the extra address manipulation GPU side.
+		 *
+		 * VLV and CHV don't have GPU swizzling.
+		 */
 		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
 		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
 	} else if (INTEL_INFO(dev)->gen >= 6) {
@@ -357,26 +364,13 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
 		 * has to also include the unfenced register the GPU uses
 		 * whilst executing a fenced command for an untiled object.
 		 */
-
-		obj->map_and_fenceable =
-			!i915_gem_obj_ggtt_bound(obj) ||
-			(i915_gem_obj_ggtt_offset(obj) +
-			 obj->base.size <= dev_priv->gtt.mappable_end &&
-			 i915_gem_object_fence_ok(obj, args->tiling_mode));
-
-		/* Rebind if we need a change of alignment */
-		if (!obj->map_and_fenceable) {
-			u32 unfenced_align =
-				i915_gem_get_gtt_alignment(dev, obj->base.size,
-							    args->tiling_mode,
-							    false);
-			if (i915_gem_obj_ggtt_offset(obj) & (unfenced_align - 1))
-				ret = i915_gem_object_ggtt_unbind(obj);
-		}
+		if (obj->map_and_fenceable &&
+		    !i915_gem_object_fence_ok(obj, args->tiling_mode))
+			ret = i915_gem_object_ggtt_unbind(obj);
 
 		if (ret == 0) {
 			obj->fence_dirty =
-				obj->fenced_gpu_access ||
+				obj->last_fenced_seqno ||
 				obj->fence_reg != I915_FENCE_REG_NONE;
 
 			obj->tiling_mode = args->tiling_mode;
diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c
index 21ea92886a56..d182058383a9 100644
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -32,6 +32,15 @@
 #include <linux/mempolicy.h>
 #include <linux/swap.h>
 
+struct i915_mm_struct {
+	struct mm_struct *mm;
+	struct drm_device *dev;
+	struct i915_mmu_notifier *mn;
+	struct hlist_node node;
+	struct kref kref;
+	struct work_struct work;
+};
+
 #if defined(CONFIG_MMU_NOTIFIER)
 #include <linux/interval_tree.h>
 
@@ -40,19 +49,83 @@ struct i915_mmu_notifier {
 	struct hlist_node node;
 	struct mmu_notifier mn;
 	struct rb_root objects;
-	struct drm_device *dev;
-	struct mm_struct *mm;
-	struct work_struct work;
-	unsigned long count;
+	struct list_head linear;
 	unsigned long serial;
+	bool has_linear;
 };
 
 struct i915_mmu_object {
-	struct i915_mmu_notifier *mmu;
+	struct i915_mmu_notifier *mn;
 	struct interval_tree_node it;
+	struct list_head link;
 	struct drm_i915_gem_object *obj;
+	bool is_linear;
 };
 
+static unsigned long cancel_userptr(struct drm_i915_gem_object *obj)
+{
+	struct drm_device *dev = obj->base.dev;
+	unsigned long end;
+
+	mutex_lock(&dev->struct_mutex);
+	/* Cancel any active worker and force us to re-evaluate gup */
+	obj->userptr.work = NULL;
+
+	if (obj->pages != NULL) {
+		struct drm_i915_private *dev_priv = to_i915(dev);
+		struct i915_vma *vma, *tmp;
+		bool was_interruptible;
+
+		was_interruptible = dev_priv->mm.interruptible;
+		dev_priv->mm.interruptible = false;
+
+		list_for_each_entry_safe(vma, tmp, &obj->vma_list, vma_link) {
+			int ret = i915_vma_unbind(vma);
+			WARN_ON(ret && ret != -EIO);
+		}
+		WARN_ON(i915_gem_object_put_pages(obj));
+
+		dev_priv->mm.interruptible = was_interruptible;
+	}
+
+	end = obj->userptr.ptr + obj->base.size;
+
+	drm_gem_object_unreference(&obj->base);
+	mutex_unlock(&dev->struct_mutex);
+
+	return end;
+}
+
+static void *invalidate_range__linear(struct i915_mmu_notifier *mn,
+				      struct mm_struct *mm,
+				      unsigned long start,
+				      unsigned long end)
+{
+	struct i915_mmu_object *mo;
+	unsigned long serial;
+
+restart:
+	serial = mn->serial;
+	list_for_each_entry(mo, &mn->linear, link) {
+		struct drm_i915_gem_object *obj;
+
+		if (mo->it.last < start || mo->it.start > end)
+			continue;
+
+		obj = mo->obj;
+		drm_gem_object_reference(&obj->base);
+		spin_unlock(&mn->lock);
+
+		cancel_userptr(obj);
+
+		spin_lock(&mn->lock);
+		if (serial != mn->serial)
+			goto restart;
+	}
+
+	return NULL;
+}
+
 static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn,
 						       struct mm_struct *mm,
 						       unsigned long start,
@@ -60,16 +133,18 @@ static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn,
 {
 	struct i915_mmu_notifier *mn = container_of(_mn, struct i915_mmu_notifier, mn);
 	struct interval_tree_node *it = NULL;
+	unsigned long next = start;
 	unsigned long serial = 0;
 
 	end--; /* interval ranges are inclusive, but invalidate range is exclusive */
-	while (start < end) {
-		struct drm_i915_gem_object *obj;
+	while (next < end) {
+		struct drm_i915_gem_object *obj = NULL;
 
-		obj = NULL;
 		spin_lock(&mn->lock);
-		if (serial == mn->serial)
-			it = interval_tree_iter_next(it, start, end);
+		if (mn->has_linear)
+			it = invalidate_range__linear(mn, mm, start, end);
+		else if (serial == mn->serial)
+			it = interval_tree_iter_next(it, next, end);
 		else
 			it = interval_tree_iter_first(&mn->objects, start, end);
 		if (it != NULL) {
@@ -81,31 +156,7 @@ static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn,
 		if (obj == NULL)
 			return;
 
-		mutex_lock(&mn->dev->struct_mutex);
-		/* Cancel any active worker and force us to re-evaluate gup */
-		obj->userptr.work = NULL;
-
-		if (obj->pages != NULL) {
-			struct drm_i915_private *dev_priv = to_i915(mn->dev);
-			struct i915_vma *vma, *tmp;
-			bool was_interruptible;
-
-			was_interruptible = dev_priv->mm.interruptible;
-			dev_priv->mm.interruptible = false;
-
-			list_for_each_entry_safe(vma, tmp, &obj->vma_list, vma_link) {
-				int ret = i915_vma_unbind(vma);
-				WARN_ON(ret && ret != -EIO);
-			}
-			WARN_ON(i915_gem_object_put_pages(obj));
-
-			dev_priv->mm.interruptible = was_interruptible;
-		}
-
-		start = obj->userptr.ptr + obj->base.size;
-
-		drm_gem_object_unreference(&obj->base);
-		mutex_unlock(&mn->dev->struct_mutex);
+		next = cancel_userptr(obj);
 	}
 }
 
@@ -114,113 +165,47 @@ static const struct mmu_notifier_ops i915_gem_userptr_notifier = {
 };
 
 static struct i915_mmu_notifier *
-__i915_mmu_notifier_lookup(struct drm_device *dev, struct mm_struct *mm)
+i915_mmu_notifier_create(struct mm_struct *mm)
 {
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct i915_mmu_notifier *mmu;
-
-	/* Protected by dev->struct_mutex */
-	hash_for_each_possible(dev_priv->mmu_notifiers, mmu, node, (unsigned long)mm)
-		if (mmu->mm == mm)
-			return mmu;
-
-	return NULL;
-}
-
-static struct i915_mmu_notifier *
-i915_mmu_notifier_get(struct drm_device *dev, struct mm_struct *mm)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct i915_mmu_notifier *mmu;
+	struct i915_mmu_notifier *mn;
 	int ret;
 
-	lockdep_assert_held(&dev->struct_mutex);
-
-	mmu = __i915_mmu_notifier_lookup(dev, mm);
-	if (mmu)
-		return mmu;
-
-	mmu = kmalloc(sizeof(*mmu), GFP_KERNEL);
-	if (mmu == NULL)
+	mn = kmalloc(sizeof(*mn), GFP_KERNEL);
+	if (mn == NULL)
 		return ERR_PTR(-ENOMEM);
 
-	spin_lock_init(&mmu->lock);
-	mmu->dev = dev;
-	mmu->mn.ops = &i915_gem_userptr_notifier;
-	mmu->mm = mm;
-	mmu->objects = RB_ROOT;
-	mmu->count = 0;
-	mmu->serial = 0;
+	spin_lock_init(&mn->lock);
+	mn->mn.ops = &i915_gem_userptr_notifier;
+	mn->objects = RB_ROOT;
+	mn->serial = 1;
+	INIT_LIST_HEAD(&mn->linear);
+	mn->has_linear = false;
 
-	/* Protected by mmap_sem (write-lock) */
-	ret = __mmu_notifier_register(&mmu->mn, mm);
+	 /* Protected by mmap_sem (write-lock) */
+	ret = __mmu_notifier_register(&mn->mn, mm);
 	if (ret) {
-		kfree(mmu);
+		kfree(mn);
 		return ERR_PTR(ret);
 	}
 
-	/* Protected by dev->struct_mutex */
-	hash_add(dev_priv->mmu_notifiers, &mmu->node, (unsigned long)mm);
-	return mmu;
-}
-
-static void
-__i915_mmu_notifier_destroy_worker(struct work_struct *work)
-{
-	struct i915_mmu_notifier *mmu = container_of(work, typeof(*mmu), work);
-	mmu_notifier_unregister(&mmu->mn, mmu->mm);
-	kfree(mmu);
+	return mn;
 }
 
-static void
-__i915_mmu_notifier_destroy(struct i915_mmu_notifier *mmu)
+static void __i915_mmu_notifier_update_serial(struct i915_mmu_notifier *mn)
 {
-	lockdep_assert_held(&mmu->dev->struct_mutex);
-
-	/* Protected by dev->struct_mutex */
-	hash_del(&mmu->node);
-
-	/* Our lock ordering is: mmap_sem, mmu_notifier_scru, struct_mutex.
-	 * We enter the function holding struct_mutex, therefore we need
-	 * to drop our mutex prior to calling mmu_notifier_unregister in
-	 * order to prevent lock inversion (and system-wide deadlock)
-	 * between the mmap_sem and struct-mutex. Hence we defer the
-	 * unregistration to a workqueue where we hold no locks.
-	 */
-	INIT_WORK(&mmu->work, __i915_mmu_notifier_destroy_worker);
-	schedule_work(&mmu->work);
-}
-
-static void __i915_mmu_notifier_update_serial(struct i915_mmu_notifier *mmu)
-{
-	if (++mmu->serial == 0)
-		mmu->serial = 1;
-}
-
-static void
-i915_mmu_notifier_del(struct i915_mmu_notifier *mmu,
-		      struct i915_mmu_object *mn)
-{
-	lockdep_assert_held(&mmu->dev->struct_mutex);
-
-	spin_lock(&mmu->lock);
-	interval_tree_remove(&mn->it, &mmu->objects);
-	__i915_mmu_notifier_update_serial(mmu);
-	spin_unlock(&mmu->lock);
-
-	/* Protected against _add() by dev->struct_mutex */
-	if (--mmu->count == 0)
-		__i915_mmu_notifier_destroy(mmu);
+	if (++mn->serial == 0)
+		mn->serial = 1;
 }
 
 static int
-i915_mmu_notifier_add(struct i915_mmu_notifier *mmu,
-		      struct i915_mmu_object *mn)
+i915_mmu_notifier_add(struct drm_device *dev,
+		      struct i915_mmu_notifier *mn,
+		      struct i915_mmu_object *mo)
 {
 	struct interval_tree_node *it;
 	int ret;
 
-	ret = i915_mutex_lock_interruptible(mmu->dev);
+	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
 
@@ -228,12 +213,11 @@ i915_mmu_notifier_add(struct i915_mmu_notifier *mmu,
 	 * remove the objects from the interval tree) before we do
 	 * the check for overlapping objects.
 	 */
-	i915_gem_retire_requests(mmu->dev);
+	i915_gem_retire_requests(dev);
 
-	/* Disallow overlapping userptr objects */
-	spin_lock(&mmu->lock);
-	it = interval_tree_iter_first(&mmu->objects,
-				      mn->it.start, mn->it.last);
+	spin_lock(&mn->lock);
+	it = interval_tree_iter_first(&mn->objects,
+				      mo->it.start, mo->it.last);
 	if (it) {
 		struct drm_i915_gem_object *obj;
 
@@ -243,85 +227,137 @@ i915_mmu_notifier_add(struct i915_mmu_notifier *mmu,
 		 * to flush their object references upon which the object will
 		 * be removed from the interval-tree, or the the range is
 		 * still in use by another client and the overlap is invalid.
+		 *
+		 * If we do have an overlap, we cannot use the interval tree
+		 * for fast range invalidation.
 		 */
 
 		obj = container_of(it, struct i915_mmu_object, it)->obj;
-		ret = obj->userptr.workers ? -EAGAIN : -EINVAL;
-	} else {
-		interval_tree_insert(&mn->it, &mmu->objects);
-		__i915_mmu_notifier_update_serial(mmu);
-		ret = 0;
+		if (!obj->userptr.workers)
+			mn->has_linear = mo->is_linear = true;
+		else
+			ret = -EAGAIN;
+	} else
+		interval_tree_insert(&mo->it, &mn->objects);
+
+	if (ret == 0) {
+		list_add(&mo->link, &mn->linear);
+		__i915_mmu_notifier_update_serial(mn);
 	}
-	spin_unlock(&mmu->lock);
-	mutex_unlock(&mmu->dev->struct_mutex);
+	spin_unlock(&mn->lock);
+	mutex_unlock(&dev->struct_mutex);
 
 	return ret;
 }
 
+static bool i915_mmu_notifier_has_linear(struct i915_mmu_notifier *mn)
+{
+	struct i915_mmu_object *mo;
+
+	list_for_each_entry(mo, &mn->linear, link)
+		if (mo->is_linear)
+			return true;
+
+	return false;
+}
+
+static void
+i915_mmu_notifier_del(struct i915_mmu_notifier *mn,
+		      struct i915_mmu_object *mo)
+{
+	spin_lock(&mn->lock);
+	list_del(&mo->link);
+	if (mo->is_linear)
+		mn->has_linear = i915_mmu_notifier_has_linear(mn);
+	else
+		interval_tree_remove(&mo->it, &mn->objects);
+	__i915_mmu_notifier_update_serial(mn);
+	spin_unlock(&mn->lock);
+}
+
 static void
 i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj)
 {
-	struct i915_mmu_object *mn;
+	struct i915_mmu_object *mo;
 
-	mn = obj->userptr.mn;
-	if (mn == NULL)
+	mo = obj->userptr.mmu_object;
+	if (mo == NULL)
 		return;
 
-	i915_mmu_notifier_del(mn->mmu, mn);
-	obj->userptr.mn = NULL;
+	i915_mmu_notifier_del(mo->mn, mo);
+	kfree(mo);
+
+	obj->userptr.mmu_object = NULL;
+}
+
+static struct i915_mmu_notifier *
+i915_mmu_notifier_find(struct i915_mm_struct *mm)
+{
+	struct i915_mmu_notifier *mn = mm->mn;
+
+	mn = mm->mn;
+	if (mn)
+		return mn;
+
+	down_write(&mm->mm->mmap_sem);
+	mutex_lock(&to_i915(mm->dev)->mm_lock);
+	if ((mn = mm->mn) == NULL) {
+		mn = i915_mmu_notifier_create(mm->mm);
+		if (!IS_ERR(mn))
+			mm->mn = mn;
+	}
+	mutex_unlock(&to_i915(mm->dev)->mm_lock);
+	up_write(&mm->mm->mmap_sem);
+
+	return mn;
 }
 
 static int
 i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj,
 				    unsigned flags)
 {
-	struct i915_mmu_notifier *mmu;
-	struct i915_mmu_object *mn;
+	struct i915_mmu_notifier *mn;
+	struct i915_mmu_object *mo;
 	int ret;
 
 	if (flags & I915_USERPTR_UNSYNCHRONIZED)
 		return capable(CAP_SYS_ADMIN) ? 0 : -EPERM;
 
-	down_write(&obj->userptr.mm->mmap_sem);
-	ret = i915_mutex_lock_interruptible(obj->base.dev);
-	if (ret == 0) {
-		mmu = i915_mmu_notifier_get(obj->base.dev, obj->userptr.mm);
-		if (!IS_ERR(mmu))
-			mmu->count++; /* preemptive add to act as a refcount */
-		else
-			ret = PTR_ERR(mmu);
-		mutex_unlock(&obj->base.dev->struct_mutex);
-	}
-	up_write(&obj->userptr.mm->mmap_sem);
-	if (ret)
-		return ret;
+	if (WARN_ON(obj->userptr.mm == NULL))
+		return -EINVAL;
 
-	mn = kzalloc(sizeof(*mn), GFP_KERNEL);
-	if (mn == NULL) {
-		ret = -ENOMEM;
-		goto destroy_mmu;
-	}
+	mn = i915_mmu_notifier_find(obj->userptr.mm);
+	if (IS_ERR(mn))
+		return PTR_ERR(mn);
 
-	mn->mmu = mmu;
-	mn->it.start = obj->userptr.ptr;
-	mn->it.last = mn->it.start + obj->base.size - 1;
-	mn->obj = obj;
+	mo = kzalloc(sizeof(*mo), GFP_KERNEL);
+	if (mo == NULL)
+		return -ENOMEM;
 
-	ret = i915_mmu_notifier_add(mmu, mn);
-	if (ret)
-		goto free_mn;
+	mo->mn = mn;
+	mo->it.start = obj->userptr.ptr;
+	mo->it.last = mo->it.start + obj->base.size - 1;
+	mo->obj = obj;
 
-	obj->userptr.mn = mn;
+	ret = i915_mmu_notifier_add(obj->base.dev, mn, mo);
+	if (ret) {
+		kfree(mo);
+		return ret;
+	}
+
+	obj->userptr.mmu_object = mo;
 	return 0;
+}
 
-free_mn:
+static void
+i915_mmu_notifier_free(struct i915_mmu_notifier *mn,
+		       struct mm_struct *mm)
+{
+	if (mn == NULL)
+		return;
+
+	mmu_notifier_unregister(&mn->mn, mm);
 	kfree(mn);
-destroy_mmu:
-	mutex_lock(&obj->base.dev->struct_mutex);
-	if (--mmu->count == 0)
-		__i915_mmu_notifier_destroy(mmu);
-	mutex_unlock(&obj->base.dev->struct_mutex);
-	return ret;
 }
 
 #else
@@ -343,15 +379,114 @@ i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj,
 
 	return 0;
 }
+
+static void
+i915_mmu_notifier_free(struct i915_mmu_notifier *mn,
+		       struct mm_struct *mm)
+{
+}
+
 #endif
 
+static struct i915_mm_struct *
+__i915_mm_struct_find(struct drm_i915_private *dev_priv, struct mm_struct *real)
+{
+	struct i915_mm_struct *mm;
+
+	/* Protected by dev_priv->mm_lock */
+	hash_for_each_possible(dev_priv->mm_structs, mm, node, (unsigned long)real)
+		if (mm->mm == real)
+			return mm;
+
+	return NULL;
+}
+
+static int
+i915_gem_userptr_init__mm_struct(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
+	struct i915_mm_struct *mm;
+	int ret = 0;
+
+	/* During release of the GEM object we hold the struct_mutex. This
+	 * precludes us from calling mmput() at that time as that may be
+	 * the last reference and so call exit_mmap(). exit_mmap() will
+	 * attempt to reap the vma, and if we were holding a GTT mmap
+	 * would then call drm_gem_vm_close() and attempt to reacquire
+	 * the struct mutex. So in order to avoid that recursion, we have
+	 * to defer releasing the mm reference until after we drop the
+	 * struct_mutex, i.e. we need to schedule a worker to do the clean
+	 * up.
+	 */
+	mutex_lock(&dev_priv->mm_lock);
+	mm = __i915_mm_struct_find(dev_priv, current->mm);
+	if (mm == NULL) {
+		mm = kmalloc(sizeof(*mm), GFP_KERNEL);
+		if (mm == NULL) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		kref_init(&mm->kref);
+		mm->dev = obj->base.dev;
+
+		mm->mm = current->mm;
+		atomic_inc(&current->mm->mm_count);
+
+		mm->mn = NULL;
+
+		/* Protected by dev_priv->mm_lock */
+		hash_add(dev_priv->mm_structs,
+			 &mm->node, (unsigned long)mm->mm);
+	} else
+		kref_get(&mm->kref);
+
+	obj->userptr.mm = mm;
+out:
+	mutex_unlock(&dev_priv->mm_lock);
+	return ret;
+}
+
+static void
+__i915_mm_struct_free__worker(struct work_struct *work)
+{
+	struct i915_mm_struct *mm = container_of(work, typeof(*mm), work);
+	i915_mmu_notifier_free(mm->mn, mm->mm);
+	mmdrop(mm->mm);
+	kfree(mm);
+}
+
+static void
+__i915_mm_struct_free(struct kref *kref)
+{
+	struct i915_mm_struct *mm = container_of(kref, typeof(*mm), kref);
+
+	/* Protected by dev_priv->mm_lock */
+	hash_del(&mm->node);
+	mutex_unlock(&to_i915(mm->dev)->mm_lock);
+
+	INIT_WORK(&mm->work, __i915_mm_struct_free__worker);
+	schedule_work(&mm->work);
+}
+
+static void
+i915_gem_userptr_release__mm_struct(struct drm_i915_gem_object *obj)
+{
+	if (obj->userptr.mm == NULL)
+		return;
+
+	kref_put_mutex(&obj->userptr.mm->kref,
+		       __i915_mm_struct_free,
+		       &to_i915(obj->base.dev)->mm_lock);
+	obj->userptr.mm = NULL;
+}
+
 struct get_pages_work {
 	struct work_struct work;
 	struct drm_i915_gem_object *obj;
 	struct task_struct *task;
 };
 
-
 #if IS_ENABLED(CONFIG_SWIOTLB)
 #define swiotlb_active() swiotlb_nr_tbl()
 #else
@@ -409,7 +544,7 @@ __i915_gem_userptr_get_pages_worker(struct work_struct *_work)
 	if (pvec == NULL)
 		pvec = drm_malloc_ab(num_pages, sizeof(struct page *));
 	if (pvec != NULL) {
-		struct mm_struct *mm = obj->userptr.mm;
+		struct mm_struct *mm = obj->userptr.mm->mm;
 
 		down_read(&mm->mmap_sem);
 		while (pinned < num_pages) {
@@ -475,7 +610,7 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
 
 	pvec = NULL;
 	pinned = 0;
-	if (obj->userptr.mm == current->mm) {
+	if (obj->userptr.mm->mm == current->mm) {
 		pvec = kmalloc(num_pages*sizeof(struct page *),
 			       GFP_TEMPORARY | __GFP_NOWARN | __GFP_NORETRY);
 		if (pvec == NULL) {
@@ -554,16 +689,15 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
 static void
 i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj)
 {
-	struct scatterlist *sg;
-	int i;
+	struct sg_page_iter sg_iter;
 
 	BUG_ON(obj->userptr.work != NULL);
 
 	if (obj->madv != I915_MADV_WILLNEED)
 		obj->dirty = 0;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
-		struct page *page = sg_page(sg);
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 
 		if (obj->dirty)
 			set_page_dirty(page);
@@ -581,17 +715,13 @@ static void
 i915_gem_userptr_release(struct drm_i915_gem_object *obj)
 {
 	i915_gem_userptr_release__mmu_notifier(obj);
-
-	if (obj->userptr.mm) {
-		mmput(obj->userptr.mm);
-		obj->userptr.mm = NULL;
-	}
+	i915_gem_userptr_release__mm_struct(obj);
 }
 
 static int
 i915_gem_userptr_dmabuf_export(struct drm_i915_gem_object *obj)
 {
-	if (obj->userptr.mn)
+	if (obj->userptr.mmu_object)
 		return 0;
 
 	return i915_gem_userptr_init__mmu_notifier(obj, 0);
@@ -611,12 +741,11 @@ static const struct drm_i915_gem_object_ops i915_gem_userptr_ops = {
  * We impose several restrictions upon the memory being mapped
  * into the GPU.
  * 1. It must be page aligned (both start/end addresses, i.e ptr and size).
- * 2. It cannot overlap any other userptr object in the same address space.
- * 3. It must be normal system memory, not a pointer into another map of IO
+ * 2. It must be normal system memory, not a pointer into another map of IO
  *    space (e.g. it must not be a GTT mmapping of another object).
- * 4. We only allow a bo as large as we could in theory map into the GTT,
+ * 3. We only allow a bo as large as we could in theory map into the GTT,
  *    that is we limit the size to the total size of the GTT.
- * 5. The bo is marked as being snoopable. The backing pages are left
+ * 4. The bo is marked as being snoopable. The backing pages are left
  *    accessible directly by the CPU, but reads and writes by the GPU may
  *    incur the cost of a snoop (unless you have an LLC architecture).
  *
@@ -667,7 +796,6 @@ i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file
 		return -ENODEV;
 	}
 
-	/* Allocate the new object */
 	obj = i915_gem_object_alloc(dev);
 	if (obj == NULL)
 		return -ENOMEM;
@@ -685,8 +813,8 @@ i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file
 	 * at binding. This means that we need to hook into the mmu_notifier
 	 * in order to detect if the mmu is destroyed.
 	 */
-	ret = -ENOMEM;
-	if ((obj->userptr.mm = get_task_mm(current)))
+	ret = i915_gem_userptr_init__mm_struct(obj);
+	if (ret == 0)
 		ret = i915_gem_userptr_init__mmu_notifier(obj, args->flags);
 	if (ret == 0)
 		ret = drm_gem_handle_create(file, &obj->base, &handle);
@@ -703,9 +831,8 @@ i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file
 int
 i915_gem_init_userptr(struct drm_device *dev)
 {
-#if defined(CONFIG_MMU_NOTIFIER)
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	hash_init(dev_priv->mmu_notifiers);
-#endif
+	mutex_init(&dev_priv->mm_lock);
+	hash_init(dev_priv->mm_structs);
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 66cf41765bf9..2c87a797213f 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -192,10 +192,10 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 				struct drm_i915_error_buffer *err,
 				int count)
 {
-	err_printf(m, "%s [%d]:\n", name, count);
+	err_printf(m, "  %s [%d]:\n", name, count);
 
 	while (count--) {
-		err_printf(m, "  %08x %8u %02x %02x %x %x",
+		err_printf(m, "    %08x %8u %02x %02x %x %x",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
@@ -208,7 +208,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 		err_puts(m, err->userptr ? " userptr" : "");
 		err_puts(m, err->ring != -1 ? " " : "");
 		err_puts(m, ring_str(err->ring));
-		err_puts(m, i915_cache_level_str(err->cache_level));
+		err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
 
 		if (err->name)
 			err_printf(m, " (name: %d)", err->name);
@@ -229,6 +229,8 @@ static const char *hangcheck_action_to_str(enum intel_ring_hangcheck_action a)
 		return "wait";
 	case HANGCHECK_ACTIVE:
 		return "active";
+	case HANGCHECK_ACTIVE_LOOP:
+		return "active (loop)";
 	case HANGCHECK_KICK:
 		return "kick";
 	case HANGCHECK_HUNG:
@@ -327,6 +329,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 	struct drm_device *dev = error_priv->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error = error_priv->error;
+	struct drm_i915_error_object *obj;
 	int i, j, offset, elt;
 	int max_hangcheck_score;
 
@@ -358,6 +361,12 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 	err_printf(m, "PCI ID: 0x%04x\n", dev->pdev->device);
 	err_printf(m, "EIR: 0x%08x\n", error->eir);
 	err_printf(m, "IER: 0x%08x\n", error->ier);
+	if (INTEL_INFO(dev)->gen >= 8) {
+		for (i = 0; i < 4; i++)
+			err_printf(m, "GTIER gt %d: 0x%08x\n", i,
+				   error->gtier[i]);
+	} else if (HAS_PCH_SPLIT(dev) || IS_VALLEYVIEW(dev))
+		err_printf(m, "GTIER: 0x%08x\n", error->gtier[0]);
 	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
 	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
 	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
@@ -384,19 +393,19 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		i915_ring_error_state(m, dev, &error->ring[i]);
 	}
 
-	if (error->active_bo)
+	for (i = 0; i < error->vm_count; i++) {
+		err_printf(m, "vm[%d]\n", i);
+
 		print_error_buffers(m, "Active",
-				    error->active_bo[0],
-				    error->active_bo_count[0]);
+				    error->active_bo[i],
+				    error->active_bo_count[i]);
 
-	if (error->pinned_bo)
 		print_error_buffers(m, "Pinned",
-				    error->pinned_bo[0],
-				    error->pinned_bo_count[0]);
+				    error->pinned_bo[i],
+				    error->pinned_bo_count[i]);
+	}
 
 	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
-		struct drm_i915_error_object *obj;
-
 		obj = error->ring[i].batchbuffer;
 		if (obj) {
 			err_puts(m, dev_priv->ring[i].name);
@@ -459,6 +468,18 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		}
 	}
 
+	if ((obj = error->semaphore_obj)) {
+		err_printf(m, "Semaphore page = 0x%08x\n", obj->gtt_offset);
+		for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
+			err_printf(m, "[%04x] %08x %08x %08x %08x\n",
+				   elt * 4,
+				   obj->pages[0][elt],
+				   obj->pages[0][elt+1],
+				   obj->pages[0][elt+2],
+				   obj->pages[0][elt+3]);
+		}
+	}
+
 	if (error->overlay)
 		intel_overlay_print_error_state(m, error->overlay);
 
@@ -473,9 +494,11 @@ out:
 }
 
 int i915_error_state_buf_init(struct drm_i915_error_state_buf *ebuf,
+			      struct drm_i915_private *i915,
 			      size_t count, loff_t pos)
 {
 	memset(ebuf, 0, sizeof(*ebuf));
+	ebuf->i915 = i915;
 
 	/* We need to have enough room to store any i915_error_state printf
 	 * so that we can move it to start position.
@@ -529,6 +552,7 @@ static void i915_error_state_free(struct kref *error_ref)
 		kfree(error->ring[i].requests);
 	}
 
+	i915_error_object_free(error->semaphore_obj);
 	kfree(error->active_bo);
 	kfree(error->overlay);
 	kfree(error->display);
@@ -536,24 +560,54 @@ static void i915_error_state_free(struct kref *error_ref)
 }
 
 static struct drm_i915_error_object *
-i915_error_object_create_sized(struct drm_i915_private *dev_priv,
-			       struct drm_i915_gem_object *src,
-			       struct i915_address_space *vm,
-			       const int num_pages)
+i915_error_object_create(struct drm_i915_private *dev_priv,
+			 struct drm_i915_gem_object *src,
+			 struct i915_address_space *vm)
 {
 	struct drm_i915_error_object *dst;
-	int i;
+	int num_pages;
+	bool use_ggtt;
+	int i = 0;
 	u32 reloc_offset;
 
 	if (src == NULL || src->pages == NULL)
 		return NULL;
 
+	num_pages = src->base.size >> PAGE_SHIFT;
+
 	dst = kmalloc(sizeof(*dst) + num_pages * sizeof(u32 *), GFP_ATOMIC);
 	if (dst == NULL)
 		return NULL;
 
-	reloc_offset = dst->gtt_offset = i915_gem_obj_offset(src, vm);
-	for (i = 0; i < num_pages; i++) {
+	if (i915_gem_obj_bound(src, vm))
+		dst->gtt_offset = i915_gem_obj_offset(src, vm);
+	else
+		dst->gtt_offset = -1;
+
+	reloc_offset = dst->gtt_offset;
+	use_ggtt = (src->cache_level == I915_CACHE_NONE &&
+		    i915_is_ggtt(vm) &&
+		    src->has_global_gtt_mapping &&
+		    reloc_offset + num_pages * PAGE_SIZE <= dev_priv->gtt.mappable_end);
+
+	/* Cannot access stolen address directly, try to use the aperture */
+	if (src->stolen) {
+		use_ggtt = true;
+
+		if (!src->has_global_gtt_mapping)
+			goto unwind;
+
+		reloc_offset = i915_gem_obj_ggtt_offset(src);
+		if (reloc_offset + num_pages * PAGE_SIZE > dev_priv->gtt.mappable_end)
+			goto unwind;
+	}
+
+	/* Cannot access snooped pages through the aperture */
+	if (use_ggtt && src->cache_level != I915_CACHE_NONE && !HAS_LLC(dev_priv->dev))
+		goto unwind;
+
+	dst->page_count = num_pages;
+	while (num_pages--) {
 		unsigned long flags;
 		void *d;
 
@@ -562,10 +616,7 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 			goto unwind;
 
 		local_irq_save(flags);
-		if (src->cache_level == I915_CACHE_NONE &&
-		    reloc_offset < dev_priv->gtt.mappable_end &&
-		    src->has_global_gtt_mapping &&
-		    i915_is_ggtt(vm)) {
+		if (use_ggtt) {
 			void __iomem *s;
 
 			/* Simply ignore tiling or any overlapping fence.
@@ -577,14 +628,6 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 						     reloc_offset);
 			memcpy_fromio(d, s, PAGE_SIZE);
 			io_mapping_unmap_atomic(s);
-		} else if (src->stolen) {
-			unsigned long offset;
-
-			offset = dev_priv->mm.stolen_base;
-			offset += src->stolen->start;
-			offset += i << PAGE_SHIFT;
-
-			memcpy_fromio(d, (void __iomem *) offset, PAGE_SIZE);
 		} else {
 			struct page *page;
 			void *s;
@@ -601,11 +644,9 @@ i915_error_object_create_sized(struct drm_i915_private *dev_priv,
 		}
 		local_irq_restore(flags);
 
-		dst->pages[i] = d;
-
+		dst->pages[i++] = d;
 		reloc_offset += PAGE_SIZE;
 	}
-	dst->page_count = num_pages;
 
 	return dst;
 
@@ -615,22 +656,19 @@ unwind:
 	kfree(dst);
 	return NULL;
 }
-#define i915_error_object_create(dev_priv, src, vm) \
-	i915_error_object_create_sized((dev_priv), (src), (vm), \
-				       (src)->base.size>>PAGE_SHIFT)
-
 #define i915_error_ggtt_object_create(dev_priv, src) \
-	i915_error_object_create_sized((dev_priv), (src), &(dev_priv)->gtt.base, \
-				       (src)->base.size>>PAGE_SHIFT)
+	i915_error_object_create((dev_priv), (src), &(dev_priv)->gtt.base)
 
 static void capture_bo(struct drm_i915_error_buffer *err,
-		       struct drm_i915_gem_object *obj)
+		       struct i915_vma *vma)
 {
+	struct drm_i915_gem_object *obj = vma->obj;
+
 	err->size = obj->base.size;
 	err->name = obj->base.name;
 	err->rseqno = obj->last_read_seqno;
 	err->wseqno = obj->last_write_seqno;
-	err->gtt_offset = i915_gem_obj_ggtt_offset(obj);
+	err->gtt_offset = vma->node.start;
 	err->read_domains = obj->base.read_domains;
 	err->write_domain = obj->base.write_domain;
 	err->fence_reg = obj->fence_reg;
@@ -654,7 +692,7 @@ static u32 capture_active_bo(struct drm_i915_error_buffer *err,
 	int i = 0;
 
 	list_for_each_entry(vma, head, mm_list) {
-		capture_bo(err++, vma->obj);
+		capture_bo(err++, vma);
 		if (++i == count)
 			break;
 	}
@@ -663,21 +701,27 @@ static u32 capture_active_bo(struct drm_i915_error_buffer *err,
 }
 
 static u32 capture_pinned_bo(struct drm_i915_error_buffer *err,
-			     int count, struct list_head *head)
+			     int count, struct list_head *head,
+			     struct i915_address_space *vm)
 {
 	struct drm_i915_gem_object *obj;
-	int i = 0;
+	struct drm_i915_error_buffer * const first = err;
+	struct drm_i915_error_buffer * const last = err + count;
 
 	list_for_each_entry(obj, head, global_list) {
-		if (!i915_gem_obj_is_pinned(obj))
-			continue;
+		struct i915_vma *vma;
 
-		capture_bo(err++, obj);
-		if (++i == count)
+		if (err == last)
 			break;
+
+		list_for_each_entry(vma, &obj->vma_list, vma_link)
+			if (vma->vm == vm && vma->pin_count > 0) {
+				capture_bo(err++, vma);
+				break;
+			}
 	}
 
-	return i;
+	return err - first;
 }
 
 /* Generate a semi-unique error code. The code is not meant to have meaning, The
@@ -746,7 +790,60 @@ static void i915_gem_record_fences(struct drm_device *dev,
 	}
 }
 
+
+static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv,
+					struct drm_i915_error_state *error,
+					struct intel_engine_cs *ring,
+					struct drm_i915_error_ring *ering)
+{
+	struct intel_engine_cs *to;
+	int i;
+
+	if (!i915_semaphore_is_enabled(dev_priv->dev))
+		return;
+
+	if (!error->semaphore_obj)
+		error->semaphore_obj =
+			i915_error_object_create(dev_priv,
+						 dev_priv->semaphore_obj,
+						 &dev_priv->gtt.base);
+
+	for_each_ring(to, dev_priv, i) {
+		int idx;
+		u16 signal_offset;
+		u32 *tmp;
+
+		if (ring == to)
+			continue;
+
+		signal_offset = (GEN8_SIGNAL_OFFSET(ring, i) & (PAGE_SIZE - 1))
+				/ 4;
+		tmp = error->semaphore_obj->pages[0];
+		idx = intel_ring_sync_index(ring, to);
+
+		ering->semaphore_mboxes[idx] = tmp[signal_offset];
+		ering->semaphore_seqno[idx] = ring->semaphore.sync_seqno[idx];
+	}
+}
+
+static void gen6_record_semaphore_state(struct drm_i915_private *dev_priv,
+					struct intel_engine_cs *ring,
+					struct drm_i915_error_ring *ering)
+{
+	ering->semaphore_mboxes[0] = I915_READ(RING_SYNC_0(ring->mmio_base));
+	ering->semaphore_mboxes[1] = I915_READ(RING_SYNC_1(ring->mmio_base));
+	ering->semaphore_seqno[0] = ring->semaphore.sync_seqno[0];
+	ering->semaphore_seqno[1] = ring->semaphore.sync_seqno[1];
+
+	if (HAS_VEBOX(dev_priv->dev)) {
+		ering->semaphore_mboxes[2] =
+			I915_READ(RING_SYNC_2(ring->mmio_base));
+		ering->semaphore_seqno[2] = ring->semaphore.sync_seqno[2];
+	}
+}
+
 static void i915_record_ring_state(struct drm_device *dev,
+				   struct drm_i915_error_state *error,
 				   struct intel_engine_cs *ring,
 				   struct drm_i915_error_ring *ering)
 {
@@ -755,18 +852,10 @@ static void i915_record_ring_state(struct drm_device *dev,
 	if (INTEL_INFO(dev)->gen >= 6) {
 		ering->rc_psmi = I915_READ(ring->mmio_base + 0x50);
 		ering->fault_reg = I915_READ(RING_FAULT_REG(ring));
-		ering->semaphore_mboxes[0]
-			= I915_READ(RING_SYNC_0(ring->mmio_base));
-		ering->semaphore_mboxes[1]
-			= I915_READ(RING_SYNC_1(ring->mmio_base));
-		ering->semaphore_seqno[0] = ring->semaphore.sync_seqno[0];
-		ering->semaphore_seqno[1] = ring->semaphore.sync_seqno[1];
-	}
-
-	if (HAS_VEBOX(dev)) {
-		ering->semaphore_mboxes[2] =
-			I915_READ(RING_SYNC_2(ring->mmio_base));
-		ering->semaphore_seqno[2] = ring->semaphore.sync_seqno[2];
+		if (INTEL_INFO(dev)->gen >= 8)
+			gen8_record_semaphore_state(dev_priv, error, ring, ering);
+		else
+			gen6_record_semaphore_state(dev_priv, ring, ering);
 	}
 
 	if (INTEL_INFO(dev)->gen >= 4) {
@@ -825,9 +914,6 @@ static void i915_record_ring_state(struct drm_device *dev,
 		ering->hws = I915_READ(mmio);
 	}
 
-	ering->cpu_ring_head = ring->buffer->head;
-	ering->cpu_ring_tail = ring->buffer->tail;
-
 	ering->hangcheck_score = ring->hangcheck.score;
 	ering->hangcheck_action = ring->hangcheck.action;
 
@@ -871,6 +957,9 @@ static void i915_gem_record_active_context(struct intel_engine_cs *ring,
 		return;
 
 	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
+		if (!i915_gem_obj_ggtt_bound(obj))
+			continue;
+
 		if ((error->ccid & PAGE_MASK) == i915_gem_obj_ggtt_offset(obj)) {
 			ering->ctx = i915_error_ggtt_object_create(dev_priv, obj);
 			break;
@@ -887,6 +976,7 @@ static void i915_gem_record_rings(struct drm_device *dev,
 
 	for (i = 0; i < I915_NUM_RINGS; i++) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
+		struct intel_ringbuffer *rbuf;
 
 		error->ring[i].pid = -1;
 
@@ -895,10 +985,16 @@ static void i915_gem_record_rings(struct drm_device *dev,
 
 		error->ring[i].valid = true;
 
-		i915_record_ring_state(dev, ring, &error->ring[i]);
+		i915_record_ring_state(dev, error, ring, &error->ring[i]);
 
 		request = i915_gem_find_active_request(ring);
 		if (request) {
+			struct i915_address_space *vm;
+
+			vm = request->ctx && request->ctx->ppgtt ?
+				&request->ctx->ppgtt->base :
+				&dev_priv->gtt.base;
+
 			/* We need to copy these to an anonymous buffer
 			 * as the simplest method to avoid being overwritten
 			 * by userspace.
@@ -906,12 +1002,9 @@ static void i915_gem_record_rings(struct drm_device *dev,
 			error->ring[i].batchbuffer =
 				i915_error_object_create(dev_priv,
 							 request->batch_obj,
-							 request->ctx ?
-							 request->ctx->vm :
-							 &dev_priv->gtt.base);
+							 vm);
 
-			if (HAS_BROKEN_CS_TLB(dev_priv->dev) &&
-			    ring->scratch.obj)
+			if (HAS_BROKEN_CS_TLB(dev_priv->dev))
 				error->ring[i].wa_batchbuffer =
 					i915_error_ggtt_object_create(dev_priv,
 							     ring->scratch.obj);
@@ -930,12 +1023,27 @@ static void i915_gem_record_rings(struct drm_device *dev,
 			}
 		}
 
+		if (i915.enable_execlists) {
+			/* TODO: This is only a small fix to keep basic error
+			 * capture working, but we need to add more information
+			 * for it to be useful (e.g. dump the context being
+			 * executed).
+			 */
+			if (request)
+				rbuf = request->ctx->engine[ring->id].ringbuf;
+			else
+				rbuf = ring->default_context->engine[ring->id].ringbuf;
+		} else
+			rbuf = ring->buffer;
+
+		error->ring[i].cpu_ring_head = rbuf->head;
+		error->ring[i].cpu_ring_tail = rbuf->tail;
+
 		error->ring[i].ringbuffer =
-			i915_error_ggtt_object_create(dev_priv, ring->buffer->obj);
+			i915_error_ggtt_object_create(dev_priv, rbuf->obj);
 
-		if (ring->status_page.obj)
-			error->ring[i].hws_page =
-				i915_error_ggtt_object_create(dev_priv, ring->status_page.obj);
+		error->ring[i].hws_page =
+			i915_error_ggtt_object_create(dev_priv, ring->status_page.obj);
 
 		i915_gem_record_active_context(ring, error, &error->ring[i]);
 
@@ -981,9 +1089,14 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 	list_for_each_entry(vma, &vm->active_list, mm_list)
 		i++;
 	error->active_bo_count[ndx] = i;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
-		if (i915_gem_obj_is_pinned(obj))
-			i++;
+
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
+		list_for_each_entry(vma, &obj->vma_list, vma_link)
+			if (vma->vm == vm && vma->pin_count > 0) {
+				i++;
+				break;
+			}
+	}
 	error->pinned_bo_count[ndx] = i - error->active_bo_count[ndx];
 
 	if (i) {
@@ -1002,7 +1115,7 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 		error->pinned_bo_count[ndx] =
 			capture_pinned_bo(pinned_bo,
 					  error->pinned_bo_count[ndx],
-					  &dev_priv->mm.bound_list);
+					  &dev_priv->mm.bound_list, vm);
 	error->active_bo[ndx] = active_bo;
 	error->pinned_bo[ndx] = pinned_bo;
 }
@@ -1023,8 +1136,25 @@ static void i915_gem_capture_buffers(struct drm_i915_private *dev_priv,
 	error->pinned_bo_count = kcalloc(cnt, sizeof(*error->pinned_bo_count),
 					 GFP_ATOMIC);
 
-	list_for_each_entry(vm, &dev_priv->vm_list, global_link)
-		i915_gem_capture_vm(dev_priv, error, vm, i++);
+	if (error->active_bo == NULL ||
+	    error->pinned_bo == NULL ||
+	    error->active_bo_count == NULL ||
+	    error->pinned_bo_count == NULL) {
+		kfree(error->active_bo);
+		kfree(error->active_bo_count);
+		kfree(error->pinned_bo);
+		kfree(error->pinned_bo_count);
+
+		error->active_bo = NULL;
+		error->active_bo_count = NULL;
+		error->pinned_bo = NULL;
+		error->pinned_bo_count = NULL;
+	} else {
+		list_for_each_entry(vm, &dev_priv->vm_list, global_link)
+			i915_gem_capture_vm(dev_priv, error, vm, i++);
+
+		error->vm_count = cnt;
+	}
 }
 
 /* Capture all registers which don't fit into another category. */
@@ -1032,6 +1162,7 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 				   struct drm_i915_error_state *error)
 {
 	struct drm_device *dev = dev_priv->dev;
+	int i;
 
 	/* General organization
 	 * 1. Registers specific to a single generation
@@ -1043,7 +1174,8 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 
 	/* 1: Registers specific to a single generation */
 	if (IS_VALLEYVIEW(dev)) {
-		error->ier = I915_READ(GTIER) | I915_READ(VLV_IER);
+		error->gtier[0] = I915_READ(GTIER);
+		error->ier = I915_READ(VLV_IER);
 		error->forcewake = I915_READ(FORCEWAKE_VLV);
 	}
 
@@ -1076,16 +1208,18 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 	if (HAS_HW_CONTEXTS(dev))
 		error->ccid = I915_READ(CCID);
 
-	if (HAS_PCH_SPLIT(dev))
-		error->ier = I915_READ(DEIER) | I915_READ(GTIER);
-	else {
-		if (IS_GEN2(dev))
-			error->ier = I915_READ16(IER);
-		else
-			error->ier = I915_READ(IER);
+	if (INTEL_INFO(dev)->gen >= 8) {
+		error->ier = I915_READ(GEN8_DE_MISC_IER);
+		for (i = 0; i < 4; i++)
+			error->gtier[i] = I915_READ(GEN8_GT_IER(i));
+	} else if (HAS_PCH_SPLIT(dev)) {
+		error->ier = I915_READ(DEIER);
+		error->gtier[0] = I915_READ(GTIER);
+	} else if (IS_GEN2(dev)) {
+		error->ier = I915_READ16(IER);
+	} else if (!IS_VALLEYVIEW(dev)) {
+		error->ier = I915_READ(IER);
 	}
-
-	/* 4: Everything else */
 	error->eir = I915_READ(EIR);
 	error->pgtbl_er = I915_READ(PGTBL_ER);
 
@@ -1223,11 +1357,11 @@ void i915_destroy_error_state(struct drm_device *dev)
 		kref_put(&error->ref, i915_error_state_free);
 }
 
-const char *i915_cache_level_str(int type)
+const char *i915_cache_level_str(struct drm_i915_private *i915, int type)
 {
 	switch (type) {
 	case I915_CACHE_NONE: return " uncached";
-	case I915_CACHE_LLC: return " snooped or LLC";
+	case I915_CACHE_LLC: return HAS_LLC(i915) ? " LLC" : " snooped";
 	case I915_CACHE_L3_LLC: return " L3+LLC";
 	case I915_CACHE_WT: return " WT";
 	default: return "";
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index c05c84f3f091..f66392b6e287 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -136,7 +136,7 @@ ironlake_enable_display_irq(struct drm_i915_private *dev_priv, u32 mask)
 {
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	if ((dev_priv->irq_mask & mask) != 0) {
@@ -151,7 +151,7 @@ ironlake_disable_display_irq(struct drm_i915_private *dev_priv, u32 mask)
 {
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	if ((dev_priv->irq_mask & mask) != mask) {
@@ -173,7 +173,7 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv,
 {
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	dev_priv->gt_irq_mask &= ~interrupt_mask;
@@ -182,12 +182,12 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv,
 	POSTING_READ(GTIMR);
 }
 
-void ilk_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	ilk_update_gt_irq(dev_priv, mask, mask);
 }
 
-void ilk_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	ilk_update_gt_irq(dev_priv, mask, 0);
 }
@@ -206,7 +206,7 @@ static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	new_val = dev_priv->pm_irq_mask;
@@ -220,12 +220,12 @@ static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
 	}
 }
 
-void snb_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	snb_update_pm_irq(dev_priv, mask, mask);
 }
 
-void snb_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	snb_update_pm_irq(dev_priv, mask, 0);
 }
@@ -238,7 +238,7 @@ static bool ivb_can_enable_err_int(struct drm_device *dev)
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 
 		if (crtc->cpu_fifo_underrun_disabled)
@@ -264,7 +264,7 @@ static void bdw_update_pm_irq(struct drm_i915_private *dev_priv,
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	new_val = dev_priv->pm_irq_mask;
@@ -278,12 +278,12 @@ static void bdw_update_pm_irq(struct drm_i915_private *dev_priv,
 	}
 }
 
-void bdw_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	bdw_update_pm_irq(dev_priv, mask, mask);
 }
 
-void bdw_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
+void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
 {
 	bdw_update_pm_irq(dev_priv, mask, 0);
 }
@@ -296,7 +296,7 @@ static bool cpt_can_enable_serr_int(struct drm_device *dev)
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 
 		if (crtc->pch_fifo_underrun_disabled)
@@ -420,7 +420,7 @@ static void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
 
 	assert_spin_locked(&dev_priv->irq_lock);
 
-	if (WARN_ON(dev_priv->pm.irqs_disabled))
+	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
 
 	I915_WRITE(SDEIMR, sdeimr);
@@ -497,7 +497,7 @@ static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
 	old = !intel_crtc->cpu_fifo_underrun_disabled;
 	intel_crtc->cpu_fifo_underrun_disabled = !enable;
 
-	if (INTEL_INFO(dev)->gen < 5 || IS_VALLEYVIEW(dev))
+	if (HAS_GMCH_DISPLAY(dev))
 		i9xx_set_fifo_underrun_reporting(dev, pipe, enable, old);
 	else if (IS_GEN5(dev) || IS_GEN6(dev))
 		ironlake_set_fifo_underrun_reporting(dev, pipe, enable);
@@ -1020,7 +1020,7 @@ static int i915_get_crtc_scanoutpos(struct drm_device *dev, int pipe,
 
 	/* In vblank? */
 	if (in_vbl)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 
 	return ret;
 }
@@ -1090,6 +1090,53 @@ static bool intel_hpd_irq_event(struct drm_device *dev,
 	return true;
 }
 
+static void i915_digport_work_func(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private, dig_port_work);
+	unsigned long irqflags;
+	u32 long_port_mask, short_port_mask;
+	struct intel_digital_port *intel_dig_port;
+	int i, ret;
+	u32 old_bits = 0;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	long_port_mask = dev_priv->long_hpd_port_mask;
+	dev_priv->long_hpd_port_mask = 0;
+	short_port_mask = dev_priv->short_hpd_port_mask;
+	dev_priv->short_hpd_port_mask = 0;
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	for (i = 0; i < I915_MAX_PORTS; i++) {
+		bool valid = false;
+		bool long_hpd = false;
+		intel_dig_port = dev_priv->hpd_irq_port[i];
+		if (!intel_dig_port || !intel_dig_port->hpd_pulse)
+			continue;
+
+		if (long_port_mask & (1 << i))  {
+			valid = true;
+			long_hpd = true;
+		} else if (short_port_mask & (1 << i))
+			valid = true;
+
+		if (valid) {
+			ret = intel_dig_port->hpd_pulse(intel_dig_port, long_hpd);
+			if (ret == true) {
+				/* if we get true fallback to old school hpd */
+				old_bits |= (1 << intel_dig_port->base.hpd_pin);
+			}
+		}
+	}
+
+	if (old_bits) {
+		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+		dev_priv->hpd_event_bits |= old_bits;
+		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+		schedule_work(&dev_priv->hotplug_work);
+	}
+}
+
 /*
  * Handle hotplug events outside the interrupt handler proper.
  */
@@ -1109,10 +1156,6 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	bool changed = false;
 	u32 hpd_event_bits;
 
-	/* HPD irq before everything is fully set up. */
-	if (!dev_priv->enable_hotplug_processing)
-		return;
-
 	mutex_lock(&mode_config->mutex);
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
@@ -1122,6 +1165,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	dev_priv->hpd_event_bits = 0;
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
+		if (!intel_connector->encoder)
+			continue;
 		intel_encoder = intel_connector->encoder;
 		if (intel_encoder->hpd_pin > HPD_NONE &&
 		    dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_MARK_DISABLED &&
@@ -1144,14 +1189,16 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	  * some connectors */
 	if (hpd_disabled) {
 		drm_kms_helper_poll_enable(dev);
-		mod_timer(&dev_priv->hotplug_reenable_timer,
-			  jiffies + msecs_to_jiffies(I915_REENABLE_HOTPLUG_DELAY));
+		mod_delayed_work(system_wq, &dev_priv->hotplug_reenable_work,
+				 msecs_to_jiffies(I915_REENABLE_HOTPLUG_DELAY));
 	}
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
+		if (!intel_connector->encoder)
+			continue;
 		intel_encoder = intel_connector->encoder;
 		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
 			if (intel_encoder->hot_plug)
@@ -1166,11 +1213,6 @@ static void i915_hotplug_work_func(struct work_struct *work)
 		drm_kms_helper_hotplug_event(dev);
 }
 
-static void intel_hpd_irq_uninstall(struct drm_i915_private *dev_priv)
-{
-	del_timer_sync(&dev_priv->hotplug_reenable_timer);
-}
-
 static void ironlake_rps_change_irq_handler(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1218,10 +1260,138 @@ static void notify_ring(struct drm_device *dev,
 
 	trace_i915_gem_request_complete(ring);
 
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		intel_notify_mmio_flip(ring);
+
 	wake_up_all(&ring->irq_queue);
 	i915_queue_hangcheck(dev);
 }
 
+static u32 vlv_c0_residency(struct drm_i915_private *dev_priv,
+			    struct intel_rps_ei *rps_ei)
+{
+	u32 cz_ts, cz_freq_khz;
+	u32 render_count, media_count;
+	u32 elapsed_render, elapsed_media, elapsed_time;
+	u32 residency = 0;
+
+	cz_ts = vlv_punit_read(dev_priv, PUNIT_REG_CZ_TIMESTAMP);
+	cz_freq_khz = DIV_ROUND_CLOSEST(dev_priv->mem_freq * 1000, 4);
+
+	render_count = I915_READ(VLV_RENDER_C0_COUNT_REG);
+	media_count = I915_READ(VLV_MEDIA_C0_COUNT_REG);
+
+	if (rps_ei->cz_clock == 0) {
+		rps_ei->cz_clock = cz_ts;
+		rps_ei->render_c0 = render_count;
+		rps_ei->media_c0 = media_count;
+
+		return dev_priv->rps.cur_freq;
+	}
+
+	elapsed_time = cz_ts - rps_ei->cz_clock;
+	rps_ei->cz_clock = cz_ts;
+
+	elapsed_render = render_count - rps_ei->render_c0;
+	rps_ei->render_c0 = render_count;
+
+	elapsed_media = media_count - rps_ei->media_c0;
+	rps_ei->media_c0 = media_count;
+
+	/* Convert all the counters into common unit of milli sec */
+	elapsed_time /= VLV_CZ_CLOCK_TO_MILLI_SEC;
+	elapsed_render /=  cz_freq_khz;
+	elapsed_media /= cz_freq_khz;
+
+	/*
+	 * Calculate overall C0 residency percentage
+	 * only if elapsed time is non zero
+	 */
+	if (elapsed_time) {
+		residency =
+			((max(elapsed_render, elapsed_media) * 100)
+				/ elapsed_time);
+	}
+
+	return residency;
+}
+
+/**
+ * vlv_calc_delay_from_C0_counters - Increase/Decrease freq based on GPU
+ * busy-ness calculated from C0 counters of render & media power wells
+ * @dev_priv: DRM device private
+ *
+ */
+static int vlv_calc_delay_from_C0_counters(struct drm_i915_private *dev_priv)
+{
+	u32 residency_C0_up = 0, residency_C0_down = 0;
+	int new_delay, adj;
+
+	dev_priv->rps.ei_interrupt_count++;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+
+	if (dev_priv->rps.up_ei.cz_clock == 0) {
+		vlv_c0_residency(dev_priv, &dev_priv->rps.up_ei);
+		vlv_c0_residency(dev_priv, &dev_priv->rps.down_ei);
+		return dev_priv->rps.cur_freq;
+	}
+
+
+	/*
+	 * To down throttle, C0 residency should be less than down threshold
+	 * for continous EI intervals. So calculate down EI counters
+	 * once in VLV_INT_COUNT_FOR_DOWN_EI
+	 */
+	if (dev_priv->rps.ei_interrupt_count == VLV_INT_COUNT_FOR_DOWN_EI) {
+
+		dev_priv->rps.ei_interrupt_count = 0;
+
+		residency_C0_down = vlv_c0_residency(dev_priv,
+						     &dev_priv->rps.down_ei);
+	} else {
+		residency_C0_up = vlv_c0_residency(dev_priv,
+						   &dev_priv->rps.up_ei);
+	}
+
+	new_delay = dev_priv->rps.cur_freq;
+
+	adj = dev_priv->rps.last_adj;
+	/* C0 residency is greater than UP threshold. Increase Frequency */
+	if (residency_C0_up >= VLV_RP_UP_EI_THRESHOLD) {
+		if (adj > 0)
+			adj *= 2;
+		else
+			adj = 1;
+
+		if (dev_priv->rps.cur_freq < dev_priv->rps.max_freq_softlimit)
+			new_delay = dev_priv->rps.cur_freq + adj;
+
+		/*
+		 * For better performance, jump directly
+		 * to RPe if we're below it.
+		 */
+		if (new_delay < dev_priv->rps.efficient_freq)
+			new_delay = dev_priv->rps.efficient_freq;
+
+	} else if (!dev_priv->rps.ei_interrupt_count &&
+			(residency_C0_down < VLV_RP_DOWN_EI_THRESHOLD)) {
+		if (adj < 0)
+			adj *= 2;
+		else
+			adj = -1;
+		/*
+		 * This means, C0 residency is less than down threshold over
+		 * a period of VLV_INT_COUNT_FOR_DOWN_EI. So, reduce the freq
+		 */
+		if (dev_priv->rps.cur_freq > dev_priv->rps.min_freq_softlimit)
+			new_delay = dev_priv->rps.cur_freq + adj;
+	}
+
+	return new_delay;
+}
+
 static void gen6_pm_rps_work(struct work_struct *work)
 {
 	struct drm_i915_private *dev_priv =
@@ -1232,11 +1402,11 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	spin_lock_irq(&dev_priv->irq_lock);
 	pm_iir = dev_priv->rps.pm_iir;
 	dev_priv->rps.pm_iir = 0;
-	if (IS_BROADWELL(dev_priv->dev))
-		bdw_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
+	if (INTEL_INFO(dev_priv->dev)->gen >= 8)
+		gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 	else {
 		/* Make sure not to corrupt PMIMR state used by ringbuffer */
-		snb_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
+		gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 	}
 	spin_unlock_irq(&dev_priv->irq_lock);
 
@@ -1252,8 +1422,10 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) {
 		if (adj > 0)
 			adj *= 2;
-		else
-			adj = 1;
+		else {
+			/* CHV needs even encode values */
+			adj = IS_CHERRYVIEW(dev_priv->dev) ? 2 : 1;
+		}
 		new_delay = dev_priv->rps.cur_freq + adj;
 
 		/*
@@ -1268,11 +1440,15 @@ static void gen6_pm_rps_work(struct work_struct *work)
 		else
 			new_delay = dev_priv->rps.min_freq_softlimit;
 		adj = 0;
+	} else if (pm_iir & GEN6_PM_RP_UP_EI_EXPIRED) {
+		new_delay = vlv_calc_delay_from_C0_counters(dev_priv);
 	} else if (pm_iir & GEN6_PM_RP_DOWN_THRESHOLD) {
 		if (adj < 0)
 			adj *= 2;
-		else
-			adj = -1;
+		else {
+			/* CHV needs even encode values */
+			adj = IS_CHERRYVIEW(dev_priv->dev) ? -2 : -1;
+		}
 		new_delay = dev_priv->rps.cur_freq + adj;
 	} else { /* unknown event */
 		new_delay = dev_priv->rps.cur_freq;
@@ -1372,7 +1548,7 @@ static void ivybridge_parity_work(struct work_struct *work)
 out:
 	WARN_ON(dev_priv->l3_parity.which_slice);
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	ilk_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv->dev));
+	gen5_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv->dev));
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
 	mutex_unlock(&dev_priv->dev->struct_mutex);
@@ -1386,7 +1562,7 @@ static void ivybridge_parity_error_irq_handler(struct drm_device *dev, u32 iir)
 		return;
 
 	spin_lock(&dev_priv->irq_lock);
-	ilk_disable_gt_irq(dev_priv, GT_PARITY_ERROR(dev));
+	gen5_disable_gt_irq(dev_priv, GT_PARITY_ERROR(dev));
 	spin_unlock(&dev_priv->irq_lock);
 
 	iir &= GT_PARITY_ERROR(dev);
@@ -1441,7 +1617,7 @@ static void gen8_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 
 	spin_lock(&dev_priv->irq_lock);
 	dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
-	bdw_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
+	gen8_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
 	spin_unlock(&dev_priv->irq_lock);
 
 	queue_work(dev_priv->wq, &dev_priv->rps.work);
@@ -1451,6 +1627,7 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 				       struct drm_i915_private *dev_priv,
 				       u32 master_ctl)
 {
+	struct intel_engine_cs *ring;
 	u32 rcs, bcs, vcs;
 	uint32_t tmp = 0;
 	irqreturn_t ret = IRQ_NONE;
@@ -1458,14 +1635,22 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 	if (master_ctl & (GEN8_GT_RCS_IRQ | GEN8_GT_BCS_IRQ)) {
 		tmp = I915_READ(GEN8_GT_IIR(0));
 		if (tmp) {
+			I915_WRITE(GEN8_GT_IIR(0), tmp);
 			ret = IRQ_HANDLED;
+
 			rcs = tmp >> GEN8_RCS_IRQ_SHIFT;
-			bcs = tmp >> GEN8_BCS_IRQ_SHIFT;
+			ring = &dev_priv->ring[RCS];
 			if (rcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[RCS]);
+				notify_ring(dev, ring);
+			if (rcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
+
+			bcs = tmp >> GEN8_BCS_IRQ_SHIFT;
+			ring = &dev_priv->ring[BCS];
 			if (bcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[BCS]);
-			I915_WRITE(GEN8_GT_IIR(0), tmp);
+				notify_ring(dev, ring);
+			if (bcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT0)!\n");
 	}
@@ -1473,14 +1658,22 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 	if (master_ctl & (GEN8_GT_VCS1_IRQ | GEN8_GT_VCS2_IRQ)) {
 		tmp = I915_READ(GEN8_GT_IIR(1));
 		if (tmp) {
+			I915_WRITE(GEN8_GT_IIR(1), tmp);
 			ret = IRQ_HANDLED;
+
 			vcs = tmp >> GEN8_VCS1_IRQ_SHIFT;
+			ring = &dev_priv->ring[VCS];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VCS]);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
+
 			vcs = tmp >> GEN8_VCS2_IRQ_SHIFT;
+			ring = &dev_priv->ring[VCS2];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VCS2]);
-			I915_WRITE(GEN8_GT_IIR(1), tmp);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT1)!\n");
 	}
@@ -1488,10 +1681,10 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 	if (master_ctl & GEN8_GT_PM_IRQ) {
 		tmp = I915_READ(GEN8_GT_IIR(2));
 		if (tmp & dev_priv->pm_rps_events) {
-			ret = IRQ_HANDLED;
-			gen8_rps_irq_handler(dev_priv, tmp);
 			I915_WRITE(GEN8_GT_IIR(2),
 				   tmp & dev_priv->pm_rps_events);
+			ret = IRQ_HANDLED;
+			gen8_rps_irq_handler(dev_priv, tmp);
 		} else
 			DRM_ERROR("The master control interrupt lied (PM)!\n");
 	}
@@ -1499,11 +1692,15 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 	if (master_ctl & GEN8_GT_VECS_IRQ) {
 		tmp = I915_READ(GEN8_GT_IIR(3));
 		if (tmp) {
+			I915_WRITE(GEN8_GT_IIR(3), tmp);
 			ret = IRQ_HANDLED;
+
 			vcs = tmp >> GEN8_VECS_IRQ_SHIFT;
+			ring = &dev_priv->ring[VECS];
 			if (vcs & GT_RENDER_USER_INTERRUPT)
-				notify_ring(dev, &dev_priv->ring[VECS]);
-			I915_WRITE(GEN8_GT_IIR(3), tmp);
+				notify_ring(dev, ring);
+			if (vcs & GT_CONTEXT_SWITCH_INTERRUPT)
+				intel_execlists_handle_ctx_events(ring);
 		} else
 			DRM_ERROR("The master control interrupt lied (GT3)!\n");
 	}
@@ -1514,23 +1711,106 @@ static irqreturn_t gen8_gt_irq_handler(struct drm_device *dev,
 #define HPD_STORM_DETECT_PERIOD 1000
 #define HPD_STORM_THRESHOLD 5
 
+static int pch_port_to_hotplug_shift(enum port port)
+{
+	switch (port) {
+	case PORT_A:
+	case PORT_E:
+	default:
+		return -1;
+	case PORT_B:
+		return 0;
+	case PORT_C:
+		return 8;
+	case PORT_D:
+		return 16;
+	}
+}
+
+static int i915_port_to_hotplug_shift(enum port port)
+{
+	switch (port) {
+	case PORT_A:
+	case PORT_E:
+	default:
+		return -1;
+	case PORT_B:
+		return 17;
+	case PORT_C:
+		return 19;
+	case PORT_D:
+		return 21;
+	}
+}
+
+static inline enum port get_port_from_pin(enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_B:
+		return PORT_B;
+	case HPD_PORT_C:
+		return PORT_C;
+	case HPD_PORT_D:
+		return PORT_D;
+	default:
+		return PORT_A; /* no hpd */
+	}
+}
+
 static inline void intel_hpd_irq_handler(struct drm_device *dev,
 					 u32 hotplug_trigger,
+					 u32 dig_hotplug_reg,
 					 const u32 *hpd)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
+	enum port port;
 	bool storm_detected = false;
+	bool queue_dig = false, queue_hp = false;
+	u32 dig_shift;
+	u32 dig_port_mask = 0;
 
 	if (!hotplug_trigger)
 		return;
 
-	DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
-			  hotplug_trigger);
+	DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x, dig 0x%08x\n",
+			 hotplug_trigger, dig_hotplug_reg);
 
 	spin_lock(&dev_priv->irq_lock);
 	for (i = 1; i < HPD_NUM_PINS; i++) {
+		if (!(hpd[i] & hotplug_trigger))
+			continue;
+
+		port = get_port_from_pin(i);
+		if (port && dev_priv->hpd_irq_port[port]) {
+			bool long_hpd;
 
+			if (HAS_PCH_SPLIT(dev)) {
+				dig_shift = pch_port_to_hotplug_shift(port);
+				long_hpd = (dig_hotplug_reg >> dig_shift) & PORTB_HOTPLUG_LONG_DETECT;
+			} else {
+				dig_shift = i915_port_to_hotplug_shift(port);
+				long_hpd = (hotplug_trigger >> dig_shift) & PORTB_HOTPLUG_LONG_DETECT;
+			}
+
+			DRM_DEBUG_DRIVER("digital hpd port %c - %s\n",
+					 port_name(port),
+					 long_hpd ? "long" : "short");
+			/* for long HPD pulses we want to have the digital queue happen,
+			   but we still want HPD storm detection to function. */
+			if (long_hpd) {
+				dev_priv->long_hpd_port_mask |= (1 << port);
+				dig_port_mask |= hpd[i];
+			} else {
+				/* for short HPD just trigger the digital queue */
+				dev_priv->short_hpd_port_mask |= (1 << port);
+				hotplug_trigger &= ~hpd[i];
+			}
+			queue_dig = true;
+		}
+	}
+
+	for (i = 1; i < HPD_NUM_PINS; i++) {
 		if (hpd[i] & hotplug_trigger &&
 		    dev_priv->hpd_stats[i].hpd_mark == HPD_DISABLED) {
 			/*
@@ -1550,7 +1830,11 @@ static inline void intel_hpd_irq_handler(struct drm_device *dev,
 		    dev_priv->hpd_stats[i].hpd_mark != HPD_ENABLED)
 			continue;
 
-		dev_priv->hpd_event_bits |= (1 << i);
+		if (!(dig_port_mask & hpd[i])) {
+			dev_priv->hpd_event_bits |= (1 << i);
+			queue_hp = true;
+		}
+
 		if (!time_in_range(jiffies, dev_priv->hpd_stats[i].hpd_last_jiffies,
 				   dev_priv->hpd_stats[i].hpd_last_jiffies
 				   + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD))) {
@@ -1579,7 +1863,10 @@ static inline void intel_hpd_irq_handler(struct drm_device *dev,
 	 * queue for otherwise the flush_work in the pageflip code will
 	 * deadlock.
 	 */
-	schedule_work(&dev_priv->hotplug_work);
+	if (queue_dig)
+		queue_work(dev_priv->dp_wq, &dev_priv->dig_port_work);
+	if (queue_hp)
+		schedule_work(&dev_priv->hotplug_work);
 }
 
 static void gmbus_irq_handler(struct drm_device *dev)
@@ -1700,7 +1987,7 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 	if (pm_iir & dev_priv->pm_rps_events) {
 		spin_lock(&dev_priv->irq_lock);
 		dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
-		snb_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
+		gen6_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
 		spin_unlock(&dev_priv->irq_lock);
 
 		queue_work(dev_priv->wq, &dev_priv->rps.work);
@@ -1720,14 +2007,9 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 
 static bool intel_pipe_handle_vblank(struct drm_device *dev, enum pipe pipe)
 {
-	struct intel_crtc *crtc;
-
 	if (!drm_handle_vblank(dev, pipe))
 		return false;
 
-	crtc = to_intel_crtc(intel_get_crtc_for_pipe(dev, pipe));
-	wake_up(&crtc->vbl_wait);
-
 	return true;
 }
 
@@ -1738,7 +2020,7 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 	int pipe;
 
 	spin_lock(&dev_priv->irq_lock);
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		int reg;
 		u32 mask, iir_bit = 0;
 
@@ -1783,9 +2065,10 @@ static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
 	}
 	spin_unlock(&dev_priv->irq_lock);
 
-	for_each_pipe(pipe) {
-		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV) {
 			intel_prepare_page_flip(dev, pipe);
@@ -1809,26 +2092,28 @@ static void i9xx_hpd_irq_handler(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
 
-	if (IS_G4X(dev)) {
-		u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_G4X;
+	if (hotplug_status) {
+		I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
+		/*
+		 * Make sure hotplug status is cleared before we clear IIR, or else we
+		 * may miss hotplug events.
+		 */
+		POSTING_READ(PORT_HOTPLUG_STAT);
 
-		intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_g4x);
-	} else {
-		u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
+		if (IS_G4X(dev)) {
+			u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_G4X;
 
-		intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_i915);
-	}
+			intel_hpd_irq_handler(dev, hotplug_trigger, 0, hpd_status_g4x);
+		} else {
+			u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
 
-	if ((IS_G4X(dev) || IS_VALLEYVIEW(dev)) &&
-	    hotplug_status & DP_AUX_CHANNEL_MASK_INT_STATUS_G4X)
-		dp_aux_irq_handler(dev);
+			intel_hpd_irq_handler(dev, hotplug_trigger, 0, hpd_status_i915);
+		}
 
-	I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
-	/*
-	 * Make sure hotplug status is cleared before we clear IIR, or else we
-	 * may miss hotplug events.
-	 */
-	POSTING_READ(PORT_HOTPLUG_STAT);
+		if ((IS_G4X(dev) || IS_VALLEYVIEW(dev)) &&
+		    hotplug_status & DP_AUX_CHANNEL_MASK_INT_STATUS_G4X)
+			dp_aux_irq_handler(dev);
+	}
 }
 
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
@@ -1839,29 +2124,36 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 	irqreturn_t ret = IRQ_NONE;
 
 	while (true) {
-		iir = I915_READ(VLV_IIR);
+		/* Find, clear, then process each source of interrupt */
+
 		gt_iir = I915_READ(GTIIR);
+		if (gt_iir)
+			I915_WRITE(GTIIR, gt_iir);
+
 		pm_iir = I915_READ(GEN6_PMIIR);
+		if (pm_iir)
+			I915_WRITE(GEN6_PMIIR, pm_iir);
+
+		iir = I915_READ(VLV_IIR);
+		if (iir) {
+			/* Consume port before clearing IIR or we'll miss events */
+			if (iir & I915_DISPLAY_PORT_INTERRUPT)
+				i9xx_hpd_irq_handler(dev);
+			I915_WRITE(VLV_IIR, iir);
+		}
 
 		if (gt_iir == 0 && pm_iir == 0 && iir == 0)
 			goto out;
 
 		ret = IRQ_HANDLED;
 
-		snb_gt_irq_handler(dev, dev_priv, gt_iir);
-
-		valleyview_pipestat_irq_handler(dev, iir);
-
-		/* Consume port.  Then clear IIR or we'll miss events */
-		if (iir & I915_DISPLAY_PORT_INTERRUPT)
-			i9xx_hpd_irq_handler(dev);
-
+		if (gt_iir)
+			snb_gt_irq_handler(dev, dev_priv, gt_iir);
 		if (pm_iir)
 			gen6_rps_irq_handler(dev_priv, pm_iir);
-
-		I915_WRITE(GTIIR, gt_iir);
-		I915_WRITE(GEN6_PMIIR, pm_iir);
-		I915_WRITE(VLV_IIR, iir);
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		valleyview_pipestat_irq_handler(dev, iir);
 	}
 
 out:
@@ -1882,21 +2174,27 @@ static irqreturn_t cherryview_irq_handler(int irq, void *arg)
 		if (master_ctl == 0 && iir == 0)
 			break;
 
+		ret = IRQ_HANDLED;
+
 		I915_WRITE(GEN8_MASTER_IRQ, 0);
 
-		gen8_gt_irq_handler(dev, dev_priv, master_ctl);
+		/* Find, clear, then process each source of interrupt */
 
-		valleyview_pipestat_irq_handler(dev, iir);
+		if (iir) {
+			/* Consume port before clearing IIR or we'll miss events */
+			if (iir & I915_DISPLAY_PORT_INTERRUPT)
+				i9xx_hpd_irq_handler(dev);
+			I915_WRITE(VLV_IIR, iir);
+		}
 
-		/* Consume port.  Then clear IIR or we'll miss events */
-		i9xx_hpd_irq_handler(dev);
+		gen8_gt_irq_handler(dev, dev_priv, master_ctl);
 
-		I915_WRITE(VLV_IIR, iir);
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		valleyview_pipestat_irq_handler(dev, iir);
 
 		I915_WRITE(GEN8_MASTER_IRQ, DE_MASTER_IRQ_CONTROL);
 		POSTING_READ(GEN8_MASTER_IRQ);
-
-		ret = IRQ_HANDLED;
 	}
 
 	return ret;
@@ -1907,8 +2205,12 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
+	u32 dig_hotplug_reg;
+
+	dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG);
+	I915_WRITE(PCH_PORT_HOTPLUG, dig_hotplug_reg);
 
-	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_ibx);
+	intel_hpd_irq_handler(dev, hotplug_trigger, dig_hotplug_reg, hpd_ibx);
 
 	if (pch_iir & SDE_AUDIO_POWER_MASK) {
 		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK) >>
@@ -1933,7 +2235,7 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_ERROR("PCH poison interrupt\n");
 
 	if (pch_iir & SDE_FDI_MASK)
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
@@ -1964,7 +2266,7 @@ static void ivb_err_int_handler(struct drm_device *dev)
 	if (err_int & ERR_INT_POISON)
 		DRM_ERROR("Poison interrupt\n");
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		if (err_int & ERR_INT_FIFO_UNDERRUN(pipe)) {
 			if (intel_set_cpu_fifo_underrun_reporting(dev, pipe,
 								  false))
@@ -2014,8 +2316,12 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
+	u32 dig_hotplug_reg;
+
+	dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG);
+	I915_WRITE(PCH_PORT_HOTPLUG, dig_hotplug_reg);
 
-	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_cpt);
+	intel_hpd_irq_handler(dev, hotplug_trigger, dig_hotplug_reg, hpd_cpt);
 
 	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT) {
 		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
@@ -2037,7 +2343,7 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 		DRM_DEBUG_DRIVER("Audio CP change interrupt\n");
 
 	if (pch_iir & SDE_FDI_MASK_CPT)
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
@@ -2060,9 +2366,10 @@ static void ilk_display_irq_handler(struct drm_device *dev, u32 de_iir)
 	if (de_iir & DE_POISON)
 		DRM_ERROR("Poison interrupt\n");
 
-	for_each_pipe(pipe) {
-		if (de_iir & DE_PIPE_VBLANK(pipe))
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & DE_PIPE_VBLANK(pipe) &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
 			if (intel_set_cpu_fifo_underrun_reporting(dev, pipe, false))
@@ -2110,9 +2417,10 @@ static void ivb_display_irq_handler(struct drm_device *dev, u32 de_iir)
 	if (de_iir & DE_GSE_IVB)
 		intel_opregion_asle_intr(dev);
 
-	for_each_pipe(pipe) {
-		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)))
-			intel_pipe_handle_vblank(dev, pipe);
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)) &&
+		    intel_pipe_handle_vblank(dev, pipe))
+			intel_check_page_flip(dev, pipe);
 
 		/* plane/pipes map 1:1 on ilk+ */
 		if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe)) {
@@ -2132,6 +2440,14 @@ static void ivb_display_irq_handler(struct drm_device *dev, u32 de_iir)
 	}
 }
 
+/*
+ * To handle irqs with the minimum potential races with fresh interrupts, we:
+ * 1 - Disable Master Interrupt Control.
+ * 2 - Find the source(s) of the interrupt.
+ * 3 - Clear the Interrupt Identity bits (IIR).
+ * 4 - Process the interrupt(s) that had bits set in the IIRs.
+ * 5 - Re-enable Master Interrupt Control.
+ */
 static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
@@ -2159,32 +2475,34 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 		POSTING_READ(SDEIER);
 	}
 
+	/* Find, clear, then process each source of interrupt */
+
 	gt_iir = I915_READ(GTIIR);
 	if (gt_iir) {
+		I915_WRITE(GTIIR, gt_iir);
+		ret = IRQ_HANDLED;
 		if (INTEL_INFO(dev)->gen >= 6)
 			snb_gt_irq_handler(dev, dev_priv, gt_iir);
 		else
 			ilk_gt_irq_handler(dev, dev_priv, gt_iir);
-		I915_WRITE(GTIIR, gt_iir);
-		ret = IRQ_HANDLED;
 	}
 
 	de_iir = I915_READ(DEIIR);
 	if (de_iir) {
+		I915_WRITE(DEIIR, de_iir);
+		ret = IRQ_HANDLED;
 		if (INTEL_INFO(dev)->gen >= 7)
 			ivb_display_irq_handler(dev, de_iir);
 		else
 			ilk_display_irq_handler(dev, de_iir);
-		I915_WRITE(DEIIR, de_iir);
-		ret = IRQ_HANDLED;
 	}
 
 	if (INTEL_INFO(dev)->gen >= 6) {
 		u32 pm_iir = I915_READ(GEN6_PMIIR);
 		if (pm_iir) {
-			gen6_rps_irq_handler(dev_priv, pm_iir);
 			I915_WRITE(GEN6_PMIIR, pm_iir);
 			ret = IRQ_HANDLED;
+			gen6_rps_irq_handler(dev_priv, pm_iir);
 		}
 	}
 
@@ -2215,72 +2533,72 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 	I915_WRITE(GEN8_MASTER_IRQ, 0);
 	POSTING_READ(GEN8_MASTER_IRQ);
 
+	/* Find, clear, then process each source of interrupt */
+
 	ret = gen8_gt_irq_handler(dev, dev_priv, master_ctl);
 
 	if (master_ctl & GEN8_DE_MISC_IRQ) {
 		tmp = I915_READ(GEN8_DE_MISC_IIR);
-		if (tmp & GEN8_DE_MISC_GSE)
-			intel_opregion_asle_intr(dev);
-		else if (tmp)
-			DRM_ERROR("Unexpected DE Misc interrupt\n");
-		else
-			DRM_ERROR("The master control interrupt lied (DE MISC)!\n");
-
 		if (tmp) {
 			I915_WRITE(GEN8_DE_MISC_IIR, tmp);
 			ret = IRQ_HANDLED;
+			if (tmp & GEN8_DE_MISC_GSE)
+				intel_opregion_asle_intr(dev);
+			else
+				DRM_ERROR("Unexpected DE Misc interrupt\n");
 		}
+		else
+			DRM_ERROR("The master control interrupt lied (DE MISC)!\n");
 	}
 
 	if (master_ctl & GEN8_DE_PORT_IRQ) {
 		tmp = I915_READ(GEN8_DE_PORT_IIR);
-		if (tmp & GEN8_AUX_CHANNEL_A)
-			dp_aux_irq_handler(dev);
-		else if (tmp)
-			DRM_ERROR("Unexpected DE Port interrupt\n");
-		else
-			DRM_ERROR("The master control interrupt lied (DE PORT)!\n");
-
 		if (tmp) {
 			I915_WRITE(GEN8_DE_PORT_IIR, tmp);
 			ret = IRQ_HANDLED;
+			if (tmp & GEN8_AUX_CHANNEL_A)
+				dp_aux_irq_handler(dev);
+			else
+				DRM_ERROR("Unexpected DE Port interrupt\n");
 		}
+		else
+			DRM_ERROR("The master control interrupt lied (DE PORT)!\n");
 	}
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		uint32_t pipe_iir;
 
 		if (!(master_ctl & GEN8_DE_PIPE_IRQ(pipe)))
 			continue;
 
 		pipe_iir = I915_READ(GEN8_DE_PIPE_IIR(pipe));
-		if (pipe_iir & GEN8_PIPE_VBLANK)
-			intel_pipe_handle_vblank(dev, pipe);
-
-		if (pipe_iir & GEN8_PIPE_PRIMARY_FLIP_DONE) {
-			intel_prepare_page_flip(dev, pipe);
-			intel_finish_page_flip_plane(dev, pipe);
-		}
+		if (pipe_iir) {
+			ret = IRQ_HANDLED;
+			I915_WRITE(GEN8_DE_PIPE_IIR(pipe), pipe_iir);
+			if (pipe_iir & GEN8_PIPE_VBLANK &&
+			    intel_pipe_handle_vblank(dev, pipe))
+				intel_check_page_flip(dev, pipe);
 
-		if (pipe_iir & GEN8_PIPE_CDCLK_CRC_DONE)
-			hsw_pipe_crc_irq_handler(dev, pipe);
+			if (pipe_iir & GEN8_PIPE_PRIMARY_FLIP_DONE) {
+				intel_prepare_page_flip(dev, pipe);
+				intel_finish_page_flip_plane(dev, pipe);
+			}
 
-		if (pipe_iir & GEN8_PIPE_FIFO_UNDERRUN) {
-			if (intel_set_cpu_fifo_underrun_reporting(dev, pipe,
-								  false))
-				DRM_ERROR("Pipe %c FIFO underrun\n",
-					  pipe_name(pipe));
-		}
+			if (pipe_iir & GEN8_PIPE_CDCLK_CRC_DONE)
+				hsw_pipe_crc_irq_handler(dev, pipe);
 
-		if (pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS) {
-			DRM_ERROR("Fault errors on pipe %c\n: 0x%08x",
-				  pipe_name(pipe),
-				  pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS);
-		}
+			if (pipe_iir & GEN8_PIPE_FIFO_UNDERRUN) {
+				if (intel_set_cpu_fifo_underrun_reporting(dev, pipe,
+									  false))
+					DRM_ERROR("Pipe %c FIFO underrun\n",
+						  pipe_name(pipe));
+			}
 
-		if (pipe_iir) {
-			ret = IRQ_HANDLED;
-			I915_WRITE(GEN8_DE_PIPE_IIR(pipe), pipe_iir);
+			if (pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS) {
+				DRM_ERROR("Fault errors on pipe %c\n: 0x%08x",
+					  pipe_name(pipe),
+					  pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS);
+			}
 		} else
 			DRM_ERROR("The master control interrupt lied (DE PIPE)!\n");
 	}
@@ -2292,13 +2610,13 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg)
 		 * on older pch-split platforms. But this needs testing.
 		 */
 		u32 pch_iir = I915_READ(SDEIIR);
-
-		cpt_irq_handler(dev, pch_iir);
-
 		if (pch_iir) {
 			I915_WRITE(SDEIIR, pch_iir);
 			ret = IRQ_HANDLED;
-		}
+			cpt_irq_handler(dev, pch_iir);
+		} else
+			DRM_ERROR("The master control interrupt lied (SDE)!\n");
+
 	}
 
 	I915_WRITE(GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL);
@@ -2467,7 +2785,7 @@ static void i915_report_and_clear_eir(struct drm_device *dev)
 
 	if (eir & I915_ERROR_MEMORY_REFRESH) {
 		pr_err("memory refresh error:\n");
-		for_each_pipe(pipe)
+		for_each_pipe(dev_priv, pipe)
 			pr_err("pipe %c stat: 0x%08x\n",
 			       pipe_name(pipe), I915_READ(PIPESTAT(pipe)));
 		/* pipestat has already been acked */
@@ -2564,52 +2882,6 @@ void i915_handle_error(struct drm_device *dev, bool wedged,
 	schedule_work(&dev_priv->gpu_error.work);
 }
 
-static void __always_unused i915_pageflip_stall_check(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj;
-	struct intel_unpin_work *work;
-	unsigned long flags;
-	bool stall_detected;
-
-	/* Ignore early vblank irqs */
-	if (intel_crtc == NULL)
-		return;
-
-	spin_lock_irqsave(&dev->event_lock, flags);
-	work = intel_crtc->unpin_work;
-
-	if (work == NULL ||
-	    atomic_read(&work->pending) >= INTEL_FLIP_COMPLETE ||
-	    !work->enable_stall_check) {
-		/* Either the pending flip IRQ arrived, or we're too early. Don't check */
-		spin_unlock_irqrestore(&dev->event_lock, flags);
-		return;
-	}
-
-	/* Potential stall - if we see that the flip has happened, assume a missed interrupt */
-	obj = work->pending_flip_obj;
-	if (INTEL_INFO(dev)->gen >= 4) {
-		int dspsurf = DSPSURF(intel_crtc->plane);
-		stall_detected = I915_HI_DISPBASE(I915_READ(dspsurf)) ==
-					i915_gem_obj_ggtt_offset(obj);
-	} else {
-		int dspaddr = DSPADDR(intel_crtc->plane);
-		stall_detected = I915_READ(dspaddr) == (i915_gem_obj_ggtt_offset(obj) +
-							crtc->y * crtc->primary->fb->pitches[0] +
-							crtc->x * crtc->primary->fb->bits_per_pixel/8);
-	}
-
-	spin_unlock_irqrestore(&dev->event_lock, flags);
-
-	if (stall_detected) {
-		DRM_DEBUG_DRIVER("Pageflip stall detected\n");
-		intel_prepare_page_flip(dev, intel_crtc->plane);
-	}
-}
-
 /* Called from drm generic code, passed 'crtc' which
  * we use as a pipe index
  */
@@ -2753,12 +3025,7 @@ static bool
 ipehr_is_semaphore_wait(struct drm_device *dev, u32 ipehr)
 {
 	if (INTEL_INFO(dev)->gen >= 8) {
-		/*
-		 * FIXME: gen8 semaphore support - currently we don't emit
-		 * semaphores on bdw anyway, but this needs to be addressed when
-		 * we merge that code.
-		 */
-		return false;
+		return (ipehr >> 23) == 0x1c;
 	} else {
 		ipehr &= ~MI_SEMAPHORE_SYNC_MASK;
 		return ipehr == (MI_SEMAPHORE_MBOX | MI_SEMAPHORE_COMPARE |
@@ -2767,19 +3034,20 @@ ipehr_is_semaphore_wait(struct drm_device *dev, u32 ipehr)
 }
 
 static struct intel_engine_cs *
-semaphore_wait_to_signaller_ring(struct intel_engine_cs *ring, u32 ipehr)
+semaphore_wait_to_signaller_ring(struct intel_engine_cs *ring, u32 ipehr, u64 offset)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct intel_engine_cs *signaller;
 	int i;
 
 	if (INTEL_INFO(dev_priv->dev)->gen >= 8) {
-		/*
-		 * FIXME: gen8 semaphore support - currently we don't emit
-		 * semaphores on bdw anyway, but this needs to be addressed when
-		 * we merge that code.
-		 */
-		return NULL;
+		for_each_ring(signaller, dev_priv, i) {
+			if (ring == signaller)
+				continue;
+
+			if (offset == signaller->semaphore.signal_ggtt[ring->id])
+				return signaller;
+		}
 	} else {
 		u32 sync_bits = ipehr & MI_SEMAPHORE_SYNC_MASK;
 
@@ -2792,8 +3060,8 @@ semaphore_wait_to_signaller_ring(struct intel_engine_cs *ring, u32 ipehr)
 		}
 	}
 
-	DRM_ERROR("No signaller ring found for ring %i, ipehr 0x%08x\n",
-		  ring->id, ipehr);
+	DRM_ERROR("No signaller ring found for ring %i, ipehr 0x%08x, offset 0x%016llx\n",
+		  ring->id, ipehr, offset);
 
 	return NULL;
 }
@@ -2803,7 +3071,8 @@ semaphore_waits_for(struct intel_engine_cs *ring, u32 *seqno)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	u32 cmd, ipehr, head;
-	int i;
+	u64 offset = 0;
+	int i, backwards;
 
 	ipehr = I915_READ(RING_IPEHR(ring->mmio_base));
 	if (!ipehr_is_semaphore_wait(ring->dev, ipehr))
@@ -2812,13 +3081,15 @@ semaphore_waits_for(struct intel_engine_cs *ring, u32 *seqno)
 	/*
 	 * HEAD is likely pointing to the dword after the actual command,
 	 * so scan backwards until we find the MBOX. But limit it to just 3
-	 * dwords. Note that we don't care about ACTHD here since that might
+	 * or 4 dwords depending on the semaphore wait command size.
+	 * Note that we don't care about ACTHD here since that might
 	 * point at at batch, and semaphores are always emitted into the
 	 * ringbuffer itself.
 	 */
 	head = I915_READ_HEAD(ring) & HEAD_ADDR;
+	backwards = (INTEL_INFO(ring->dev)->gen >= 8) ? 5 : 4;
 
-	for (i = 4; i; --i) {
+	for (i = backwards; i; --i) {
 		/*
 		 * Be paranoid and presume the hw has gone off into the wild -
 		 * our ring is smaller than what the hardware (and hence
@@ -2838,7 +3109,12 @@ semaphore_waits_for(struct intel_engine_cs *ring, u32 *seqno)
 		return NULL;
 
 	*seqno = ioread32(ring->buffer->virtual_start + head + 4) + 1;
-	return semaphore_wait_to_signaller_ring(ring, ipehr);
+	if (INTEL_INFO(ring->dev)->gen >= 8) {
+		offset = ioread32(ring->buffer->virtual_start + head + 12);
+		offset <<= 32;
+		offset = ioread32(ring->buffer->virtual_start + head + 8);
+	}
+	return semaphore_wait_to_signaller_ring(ring, ipehr, offset);
 }
 
 static int semaphore_passed(struct intel_engine_cs *ring)
@@ -2884,8 +3160,14 @@ ring_stuck(struct intel_engine_cs *ring, u64 acthd)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 tmp;
 
-	if (ring->hangcheck.acthd != acthd)
-		return HANGCHECK_ACTIVE;
+	if (acthd != ring->hangcheck.acthd) {
+		if (acthd > ring->hangcheck.max_acthd) {
+			ring->hangcheck.max_acthd = acthd;
+			return HANGCHECK_ACTIVE;
+		}
+
+		return HANGCHECK_ACTIVE_LOOP;
+	}
 
 	if (IS_GEN2(dev))
 		return HANGCHECK_HUNG;
@@ -2996,8 +3278,9 @@ static void i915_hangcheck_elapsed(unsigned long data)
 				switch (ring->hangcheck.action) {
 				case HANGCHECK_IDLE:
 				case HANGCHECK_WAIT:
-					break;
 				case HANGCHECK_ACTIVE:
+					break;
+				case HANGCHECK_ACTIVE_LOOP:
 					ring->hangcheck.score += BUSY;
 					break;
 				case HANGCHECK_KICK:
@@ -3017,6 +3300,8 @@ static void i915_hangcheck_elapsed(unsigned long data)
 			 */
 			if (ring->hangcheck.score > 0)
 				ring->hangcheck.score--;
+
+			ring->hangcheck.acthd = ring->hangcheck.max_acthd = 0;
 		}
 
 		ring->hangcheck.seqno = seqno;
@@ -3132,7 +3417,7 @@ static void valleyview_irq_preinstall(struct drm_device *dev)
 
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 	I915_WRITE(VLV_IIR, 0xffffffff);
 	I915_WRITE(VLV_IMR, 0xffffffff);
@@ -3158,8 +3443,10 @@ static void gen8_irq_reset(struct drm_device *dev)
 
 	gen8_gt_irq_reset(dev_priv);
 
-	for_each_pipe(pipe)
-		GEN8_IRQ_RESET_NDX(DE_PIPE, pipe);
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_enabled(dev_priv,
+						POWER_DOMAIN_PIPE(pipe)))
+			GEN8_IRQ_RESET_NDX(DE_PIPE, pipe);
 
 	GEN5_IRQ_RESET(GEN8_DE_PORT_);
 	GEN5_IRQ_RESET(GEN8_DE_MISC_);
@@ -3168,6 +3455,19 @@ static void gen8_irq_reset(struct drm_device *dev)
 	ibx_irq_reset(dev);
 }
 
+void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv)
+{
+	unsigned long irqflags;
+	uint32_t extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	GEN8_IRQ_INIT_NDX(DE_PIPE, PIPE_B, dev_priv->de_irq_mask[PIPE_B],
+			  ~dev_priv->de_irq_mask[PIPE_B] | extra_ier);
+	GEN8_IRQ_INIT_NDX(DE_PIPE, PIPE_C, dev_priv->de_irq_mask[PIPE_C],
+			  ~dev_priv->de_irq_mask[PIPE_C] | extra_ier);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
 static void cherryview_irq_preinstall(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3187,7 +3487,7 @@ static void cherryview_irq_preinstall(struct drm_device *dev)
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
 	I915_WRITE(VLV_IMR, 0xffffffff);
@@ -3199,18 +3499,17 @@ static void cherryview_irq_preinstall(struct drm_device *dev)
 static void ibx_hpd_irq_setup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_irqs, hotplug, enabled_irqs = 0;
 
 	if (HAS_PCH_IBX(dev)) {
 		hotplug_irqs = SDE_HOTPLUG_MASK;
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				enabled_irqs |= hpd_ibx[intel_encoder->hpd_pin];
 	} else {
 		hotplug_irqs = SDE_HOTPLUG_MASK_CPT;
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				enabled_irqs |= hpd_cpt[intel_encoder->hpd_pin];
 	}
@@ -3459,28 +3758,31 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 
 static void gen8_gt_irq_postinstall(struct drm_i915_private *dev_priv)
 {
-	int i;
-
 	/* These are interrupts we'll toggle with the ring mask register */
 	uint32_t gt_interrupts[] = {
 		GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
 			GT_RENDER_L3_PARITY_ERROR_INTERRUPT |
-			GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT,
+			GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT,
 		GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
-			GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT,
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
+			GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT,
 		0,
-		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT
+		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT |
+			GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT
 		};
 
-	for (i = 0; i < ARRAY_SIZE(gt_interrupts); i++)
-		GEN8_IRQ_INIT_NDX(GT, i, ~gt_interrupts[i], gt_interrupts[i]);
-
 	dev_priv->pm_irq_mask = 0xffffffff;
+	GEN8_IRQ_INIT_NDX(GT, 0, ~gt_interrupts[0], gt_interrupts[0]);
+	GEN8_IRQ_INIT_NDX(GT, 1, ~gt_interrupts[1], gt_interrupts[1]);
+	GEN8_IRQ_INIT_NDX(GT, 2, dev_priv->pm_irq_mask, dev_priv->pm_rps_events);
+	GEN8_IRQ_INIT_NDX(GT, 3, ~gt_interrupts[3], gt_interrupts[3]);
 }
 
 static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 {
-	struct drm_device *dev = dev_priv->dev;
 	uint32_t de_pipe_masked = GEN8_PIPE_PRIMARY_FLIP_DONE |
 		GEN8_PIPE_CDCLK_CRC_DONE |
 		GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
@@ -3491,9 +3793,12 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 	dev_priv->de_irq_mask[PIPE_B] = ~de_pipe_masked;
 	dev_priv->de_irq_mask[PIPE_C] = ~de_pipe_masked;
 
-	for_each_pipe(pipe)
-		GEN8_IRQ_INIT_NDX(DE_PIPE, pipe, dev_priv->de_irq_mask[pipe],
-				  de_pipe_enables);
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_enabled(dev_priv,
+				POWER_DOMAIN_PIPE(pipe)))
+			GEN8_IRQ_INIT_NDX(DE_PIPE, pipe,
+					  dev_priv->de_irq_mask[pipe],
+					  de_pipe_enables);
 
 	GEN5_IRQ_INIT(GEN8_DE_PORT_, ~GEN8_AUX_CHANNEL_A, GEN8_AUX_CHANNEL_A);
 }
@@ -3533,12 +3838,12 @@ static int cherryview_irq_postinstall(struct drm_device *dev)
 	 */
 	dev_priv->irq_mask = ~enable_mask;
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		i915_enable_pipestat(dev_priv, pipe, pipestat_enable);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
@@ -3561,8 +3866,6 @@ static void gen8_irq_uninstall(struct drm_device *dev)
 	if (!dev_priv)
 		return;
 
-	intel_hpd_irq_uninstall(dev_priv);
-
 	gen8_irq_reset(dev);
 }
 
@@ -3577,9 +3880,7 @@ static void valleyview_irq_uninstall(struct drm_device *dev)
 
 	I915_WRITE(VLV_MASTER_IER, 0);
 
-	intel_hpd_irq_uninstall(dev_priv);
-
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
 	I915_WRITE(HWSTAM, 0xffffffff);
@@ -3641,7 +3942,7 @@ do {							\
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0xffff);
 
 	I915_WRITE(VLV_IMR, 0xffffffff);
@@ -3657,8 +3958,6 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	if (!dev_priv)
 		return;
 
-	intel_hpd_irq_uninstall(dev_priv);
-
 	ironlake_irq_reset(dev);
 }
 
@@ -3667,7 +3966,7 @@ static void i8xx_irq_preinstall(struct drm_device * dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE16(IMR, 0xffff);
 	I915_WRITE16(IER, 0x0);
@@ -3721,7 +4020,7 @@ static bool i8xx_handle_vblank(struct drm_device *dev,
 		return false;
 
 	if ((iir & flip_pending) == 0)
-		return false;
+		goto check_page_flip;
 
 	intel_prepare_page_flip(dev, plane);
 
@@ -3732,11 +4031,14 @@ static bool i8xx_handle_vblank(struct drm_device *dev,
 	 * an interrupt per se, we watch for the change at vblank.
 	 */
 	if (I915_READ16(ISR) & flip_pending)
-		return false;
+		goto check_page_flip;
 
 	intel_finish_page_flip(dev, pipe);
-
 	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev, pipe);
+	return false;
 }
 
 static irqreturn_t i8xx_irq_handler(int irq, void *arg)
@@ -3767,7 +4069,7 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 					  "Command parser error, iir 0x%08x",
 					  iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -3787,7 +4089,7 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 		if (iir & I915_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[RCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int plane = pipe;
 			if (HAS_FBC(dev))
 				plane = !plane;
@@ -3815,7 +4117,7 @@ static void i8xx_irq_uninstall(struct drm_device * dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		/* Clear enable bits; then clear status bits */
 		I915_WRITE(PIPESTAT(pipe), 0);
 		I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
@@ -3836,7 +4138,7 @@ static void i915_irq_preinstall(struct drm_device * dev)
 	}
 
 	I915_WRITE16(HWSTAM, 0xeffe);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
@@ -3906,7 +4208,7 @@ static bool i915_handle_vblank(struct drm_device *dev,
 		return false;
 
 	if ((iir & flip_pending) == 0)
-		return false;
+		goto check_page_flip;
 
 	intel_prepare_page_flip(dev, plane);
 
@@ -3917,11 +4219,14 @@ static bool i915_handle_vblank(struct drm_device *dev,
 	 * an interrupt per se, we watch for the change at vblank.
 	 */
 	if (I915_READ(ISR) & flip_pending)
-		return false;
+		goto check_page_flip;
 
 	intel_finish_page_flip(dev, pipe);
-
 	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev, pipe);
+	return false;
 }
 
 static irqreturn_t i915_irq_handler(int irq, void *arg)
@@ -3951,7 +4256,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 					  "Command parser error, iir 0x%08x",
 					  iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -3977,7 +4282,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 		if (iir & I915_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[RCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int plane = pipe;
 			if (HAS_FBC(dev))
 				plane = !plane;
@@ -4029,15 +4334,13 @@ static void i915_irq_uninstall(struct drm_device * dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	intel_hpd_irq_uninstall(dev_priv);
-
 	if (I915_HAS_HOTPLUG(dev)) {
 		I915_WRITE(PORT_HOTPLUG_EN, 0);
 		I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 	}
 
 	I915_WRITE16(HWSTAM, 0xffff);
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		/* Clear enable bits; then clear status bits */
 		I915_WRITE(PIPESTAT(pipe), 0);
 		I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
@@ -4057,7 +4360,7 @@ static void i965_irq_preinstall(struct drm_device * dev)
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
 	I915_WRITE(HWSTAM, 0xeffe);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
@@ -4126,7 +4429,6 @@ static int i965_irq_postinstall(struct drm_device *dev)
 static void i915_hpd_irq_setup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_en;
 
@@ -4137,7 +4439,7 @@ static void i915_hpd_irq_setup(struct drm_device *dev)
 		hotplug_en &= ~HOTPLUG_INT_EN_MASK;
 		/* Note HDMI and DP share hotplug bits */
 		/* enable bits are the same for all generations */
-		list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
+		for_each_intel_encoder(dev, intel_encoder)
 			if (dev_priv->hpd_stats[intel_encoder->hpd_pin].hpd_mark == HPD_ENABLED)
 				hotplug_en |= hpd_mask_i915[intel_encoder->hpd_pin];
 		/* Programming the CRT detection parameters tends
@@ -4183,7 +4485,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 					  "Command parser error, iir 0x%08x",
 					  iir);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			int reg = PIPESTAT(pipe);
 			pipe_stats[pipe] = I915_READ(reg);
 
@@ -4214,7 +4516,7 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 		if (iir & I915_BSD_USER_INTERRUPT)
 			notify_ring(dev, &dev_priv->ring[VCS]);
 
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
 			    i915_handle_vblank(dev, pipe, pipe, iir))
 				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
@@ -4267,31 +4569,33 @@ static void i965_irq_uninstall(struct drm_device * dev)
 	if (!dev_priv)
 		return;
 
-	intel_hpd_irq_uninstall(dev_priv);
-
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
 
 	I915_WRITE(HWSTAM, 0xffffffff);
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe), 0);
 	I915_WRITE(IMR, 0xffffffff);
 	I915_WRITE(IER, 0x0);
 
-	for_each_pipe(pipe)
+	for_each_pipe(dev_priv, pipe)
 		I915_WRITE(PIPESTAT(pipe),
 			   I915_READ(PIPESTAT(pipe)) & 0x8000ffff);
 	I915_WRITE(IIR, I915_READ(IIR));
 }
 
-static void intel_hpd_irq_reenable(unsigned long data)
+static void intel_hpd_irq_reenable(struct work_struct *work)
 {
-	struct drm_i915_private *dev_priv = (struct drm_i915_private *)data;
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv),
+			     hotplug_reenable_work.work);
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	unsigned long irqflags;
 	int i;
 
+	intel_runtime_pm_get(dev_priv);
+
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	for (i = (HPD_NONE + 1); i < HPD_NUM_PINS; i++) {
 		struct drm_connector *connector;
@@ -4317,6 +4621,8 @@ static void intel_hpd_irq_reenable(unsigned long data)
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	intel_runtime_pm_put(dev_priv);
 }
 
 void intel_irq_init(struct drm_device *dev)
@@ -4324,21 +4630,29 @@ void intel_irq_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func);
+	INIT_WORK(&dev_priv->dig_port_work, i915_digport_work_func);
 	INIT_WORK(&dev_priv->gpu_error.work, i915_error_work_func);
 	INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work);
 	INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
 
 	/* Let's track the enabled rps events */
-	dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS;
+	if (IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev))
+		/* WaGsvRC0ResidencyMethod:vlv */
+		dev_priv->pm_rps_events = GEN6_PM_RP_UP_EI_EXPIRED;
+	else
+		dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS;
 
 	setup_timer(&dev_priv->gpu_error.hangcheck_timer,
 		    i915_hangcheck_elapsed,
 		    (unsigned long) dev);
-	setup_timer(&dev_priv->hotplug_reenable_timer, intel_hpd_irq_reenable,
-		    (unsigned long) dev_priv);
+	INIT_DELAYED_WORK(&dev_priv->hotplug_reenable_work,
+			  intel_hpd_irq_reenable);
 
 	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
 
+	/* Haven't installed the IRQ handler yet */
+	dev_priv->pm._irqs_disabled = true;
+
 	if (IS_GEN2(dev)) {
 		dev->max_vblank_count = 0;
 		dev->driver->get_vblank_counter = i8xx_get_vblank_counter;
@@ -4350,6 +4664,14 @@ void intel_irq_init(struct drm_device *dev)
 		dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
 	}
 
+	/*
+	 * Opt out of the vblank disable timer on everything except gen2.
+	 * Gen2 doesn't have a hardware frame counter and so depends on
+	 * vblank interrupts to produce sane vblank seuquence numbers.
+	 */
+	if (!IS_GEN2(dev))
+		dev->vblank_disable_immediate = true;
+
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
 		dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
@@ -4426,7 +4748,9 @@ void intel_hpd_init(struct drm_device *dev)
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		struct intel_connector *intel_connector = to_intel_connector(connector);
 		connector->polled = intel_connector->polled;
-		if (!connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE)
+		if (connector->encoder && !connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE)
+			connector->polled = DRM_CONNECTOR_POLL_HPD;
+		if (intel_connector->mst_port)
 			connector->polled = DRM_CONNECTOR_POLL_HPD;
 	}
 
@@ -4444,7 +4768,7 @@ void intel_runtime_pm_disable_interrupts(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	dev->driver->irq_uninstall(dev);
-	dev_priv->pm.irqs_disabled = true;
+	dev_priv->pm._irqs_disabled = true;
 }
 
 /* Restore interrupts so we can recover from runtime PM. */
@@ -4452,7 +4776,7 @@ void intel_runtime_pm_restore_interrupts(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->pm.irqs_disabled = false;
+	dev_priv->pm._irqs_disabled = false;
 	dev->driver->irq_preinstall(dev);
 	dev->driver->irq_postinstall(dev);
 }
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index d05a2afa17dc..c91cb2033cc5 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -35,6 +35,7 @@ struct i915_params i915 __read_mostly = {
 	.vbt_sdvo_panel_type = -1,
 	.enable_rc6 = -1,
 	.enable_fbc = -1,
+	.enable_execlists = 0,
 	.enable_hangcheck = true,
 	.enable_ppgtt = -1,
 	.enable_psr = 0,
@@ -48,6 +49,8 @@ struct i915_params i915 __read_mostly = {
 	.disable_display = 0,
 	.enable_cmd_parser = 1,
 	.disable_vtd_wa = 0,
+	.use_mmio_flip = 0,
+	.mmio_debug = 0,
 };
 
 module_param_named(modeset, i915.modeset, int, 0400);
@@ -64,12 +67,12 @@ module_param_named(powersave, i915.powersave, int, 0600);
 MODULE_PARM_DESC(powersave,
 	"Enable powersavings, fbc, downclocking, etc. (default: true)");
 
-module_param_named(semaphores, i915.semaphores, int, 0400);
+module_param_named_unsafe(semaphores, i915.semaphores, int, 0400);
 MODULE_PARM_DESC(semaphores,
 	"Use semaphores for inter-ring sync "
 	"(default: -1 (use per-chip defaults))");
 
-module_param_named(enable_rc6, i915.enable_rc6, int, 0400);
+module_param_named_unsafe(enable_rc6, i915.enable_rc6, int, 0400);
 MODULE_PARM_DESC(enable_rc6,
 	"Enable power-saving render C-state 6. "
 	"Different stages can be selected via bitmask values "
@@ -77,7 +80,7 @@ MODULE_PARM_DESC(enable_rc6,
 	"For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. "
 	"default: -1 (use per-chip default)");
 
-module_param_named(enable_fbc, i915.enable_fbc, int, 0600);
+module_param_named_unsafe(enable_fbc, i915.enable_fbc, int, 0600);
 MODULE_PARM_DESC(enable_fbc,
 	"Enable frame buffer compression for power savings "
 	"(default: -1 (use per-chip default))");
@@ -111,11 +114,16 @@ MODULE_PARM_DESC(enable_hangcheck,
 	"WARNING: Disabling this can cause system wide hangs. "
 	"(default: true)");
 
-module_param_named(enable_ppgtt, i915.enable_ppgtt, int, 0400);
+module_param_named_unsafe(enable_ppgtt, i915.enable_ppgtt, int, 0400);
 MODULE_PARM_DESC(enable_ppgtt,
 	"Override PPGTT usage. "
 	"(-1=auto [default], 0=disabled, 1=aliasing, 2=full)");
 
+module_param_named(enable_execlists, i915.enable_execlists, int, 0400);
+MODULE_PARM_DESC(enable_execlists,
+	"Override execlists usage. "
+	"(-1=auto, 0=disabled [default], 1=enabled)");
+
 module_param_named(enable_psr, i915.enable_psr, int, 0600);
 MODULE_PARM_DESC(enable_psr, "Enable PSR (default: false)");
 
@@ -156,3 +164,12 @@ MODULE_PARM_DESC(disable_vtd_wa, "Disable all VT-d workarounds (default: false)"
 module_param_named(enable_cmd_parser, i915.enable_cmd_parser, int, 0600);
 MODULE_PARM_DESC(enable_cmd_parser,
 		 "Enable command parsing (1=enabled [default], 0=disabled)");
+
+module_param_named(use_mmio_flip, i915.use_mmio_flip, int, 0600);
+MODULE_PARM_DESC(use_mmio_flip,
+		 "use MMIO flips (-1=never, 0=driver discretion [default], 1=always)");
+
+module_param_named(mmio_debug, i915.mmio_debug, bool, 0600);
+MODULE_PARM_DESC(mmio_debug,
+	"Enable the MMIO debug code (default: false). This may negatively "
+	"affect performance.");
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index a5bab61bfc00..c01e5f31430e 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -29,8 +29,8 @@
 #define _TRANSCODER(tran, a, b) ((a) + (tran)*((b)-(a)))
 
 #define _PORT(port, a, b) ((a) + (port)*((b)-(a)))
-#define _PIPE3(pipe, a, b, c) (pipe < 2 ? _PIPE(pipe, a, b) : c)
-#define _PORT3(port, a, b, c) (port < 2 ? _PORT(port, a, b) : c)
+#define _PIPE3(pipe, a, b, c) ((pipe) == PIPE_A ? (a) : \
+			       (pipe) == PIPE_B ? (b) : (c))
 
 #define _MASKED_BIT_ENABLE(a) (((a) << 16) | (a))
 #define _MASKED_BIT_DISABLE(a) ((a) << 16)
@@ -143,6 +143,14 @@
 #define GAB_CTL				0x24000
 #define   GAB_CTL_CONT_AFTER_PAGEFAULT	(1<<8)
 
+#define GEN7_BIOS_RESERVED		0x1082C0
+#define GEN7_BIOS_RESERVED_1M		(0 << 5)
+#define GEN7_BIOS_RESERVED_256K		(1 << 5)
+#define GEN8_BIOS_RESERVED_SHIFT       7
+#define GEN7_BIOS_RESERVED_MASK        0x1
+#define GEN8_BIOS_RESERVED_MASK        0x3
+
+
 /* VGA stuff */
 
 #define VGA_ST01_MDA 0x3ba
@@ -240,7 +248,7 @@
 #define   MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19)
 #define   MI_DISPLAY_FLIP_IVB_PLANE_C  (4 << 19)
 #define   MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19)
-#define MI_SEMAPHORE_MBOX	MI_INSTR(0x16, 1) /* gen6+ */
+#define MI_SEMAPHORE_MBOX	MI_INSTR(0x16, 1) /* gen6, gen7 */
 #define   MI_SEMAPHORE_GLOBAL_GTT    (1<<22)
 #define   MI_SEMAPHORE_UPDATE	    (1<<21)
 #define   MI_SEMAPHORE_COMPARE	    (1<<20)
@@ -266,7 +274,13 @@
 #define   MI_RESTORE_EXT_STATE_EN	(1<<2)
 #define   MI_FORCE_RESTORE		(1<<1)
 #define   MI_RESTORE_INHIBIT		(1<<0)
+#define MI_SEMAPHORE_SIGNAL	MI_INSTR(0x1b, 0) /* GEN8+ */
+#define   MI_SEMAPHORE_TARGET(engine)	((engine)<<15)
+#define MI_SEMAPHORE_WAIT	MI_INSTR(0x1c, 2) /* GEN8+ */
+#define   MI_SEMAPHORE_POLL		(1<<15)
+#define   MI_SEMAPHORE_SAD_GTE_SDD	(1<<12)
 #define MI_STORE_DWORD_IMM	MI_INSTR(0x20, 1)
+#define MI_STORE_DWORD_IMM_GEN8	MI_INSTR(0x20, 2)
 #define   MI_MEM_VIRTUAL	(1 << 22) /* 965+ only */
 #define MI_STORE_DWORD_INDEX	MI_INSTR(0x21, 1)
 #define   MI_STORE_DWORD_INDEX_SHIFT 2
@@ -277,6 +291,7 @@
  *   address/value pairs. Don't overdue it, though, x <= 2^4 must hold!
  */
 #define MI_LOAD_REGISTER_IMM(x)	MI_INSTR(0x22, 2*(x)-1)
+#define   MI_LRI_FORCE_POSTED		(1<<12)
 #define MI_STORE_REGISTER_MEM(x) MI_INSTR(0x24, 2*(x)-1)
 #define MI_STORE_REGISTER_MEM_GEN8(x) MI_INSTR(0x24, 3*(x)-1)
 #define   MI_SRM_LRM_GLOBAL_GTT		(1<<22)
@@ -329,16 +344,20 @@
 #define GFX_OP_DESTBUFFER_INFO	 ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1)
 #define GFX_OP_DRAWRECT_INFO     ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3))
 #define GFX_OP_DRAWRECT_INFO_I965  ((0x7900<<16)|0x2)
-#define SRC_COPY_BLT_CMD                ((2<<29)|(0x43<<22)|4)
+
+#define COLOR_BLT_CMD			(2<<29 | 0x40<<22 | (5-2))
+#define SRC_COPY_BLT_CMD		((2<<29)|(0x43<<22)|4)
 #define XY_SRC_COPY_BLT_CMD		((2<<29)|(0x53<<22)|6)
 #define XY_MONO_SRC_COPY_IMM_BLT	((2<<29)|(0x71<<22)|5)
-#define XY_SRC_COPY_BLT_WRITE_ALPHA	(1<<21)
-#define XY_SRC_COPY_BLT_WRITE_RGB	(1<<20)
+#define   BLT_WRITE_A			(2<<20)
+#define   BLT_WRITE_RGB			(1<<20)
+#define   BLT_WRITE_RGBA		(BLT_WRITE_RGB | BLT_WRITE_A)
 #define   BLT_DEPTH_8			(0<<24)
 #define   BLT_DEPTH_16_565		(1<<24)
 #define   BLT_DEPTH_16_1555		(2<<24)
 #define   BLT_DEPTH_32			(3<<24)
-#define   BLT_ROP_GXCOPY		(0xcc<<16)
+#define   BLT_ROP_SRC_COPY		(0xcc<<16)
+#define   BLT_ROP_COLOR_COPY		(0xf0<<16)
 #define XY_SRC_COPY_BLT_SRC_TILED	(1<<15) /* 965+ only */
 #define XY_SRC_COPY_BLT_DST_TILED	(1<<11) /* 965+ only */
 #define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2)
@@ -360,6 +379,7 @@
 #define   PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE		(1<<10) /* GM45+ only */
 #define   PIPE_CONTROL_INDIRECT_STATE_DISABLE		(1<<9)
 #define   PIPE_CONTROL_NOTIFY				(1<<8)
+#define   PIPE_CONTROL_FLUSH_ENABLE			(1<<7) /* gen7+ */
 #define   PIPE_CONTROL_VF_CACHE_INVALIDATE		(1<<4)
 #define   PIPE_CONTROL_CONST_CACHE_INVALIDATE		(1<<3)
 #define   PIPE_CONTROL_STATE_CACHE_INVALIDATE		(1<<2)
@@ -491,10 +511,26 @@
 #define BUNIT_REG_BISOC				0x11
 
 #define PUNIT_REG_DSPFREQ			0x36
+#define   DSPFREQSTAT_SHIFT_CHV			24
+#define   DSPFREQSTAT_MASK_CHV			(0x1f << DSPFREQSTAT_SHIFT_CHV)
+#define   DSPFREQGUAR_SHIFT_CHV			8
+#define   DSPFREQGUAR_MASK_CHV			(0x1f << DSPFREQGUAR_SHIFT_CHV)
 #define   DSPFREQSTAT_SHIFT			30
 #define   DSPFREQSTAT_MASK			(0x3 << DSPFREQSTAT_SHIFT)
 #define   DSPFREQGUAR_SHIFT			14
 #define   DSPFREQGUAR_MASK			(0x3 << DSPFREQGUAR_SHIFT)
+#define   _DP_SSC(val, pipe)			((val) << (2 * (pipe)))
+#define   DP_SSC_MASK(pipe)			_DP_SSC(0x3, (pipe))
+#define   DP_SSC_PWR_ON(pipe)			_DP_SSC(0x0, (pipe))
+#define   DP_SSC_CLK_GATE(pipe)			_DP_SSC(0x1, (pipe))
+#define   DP_SSC_RESET(pipe)			_DP_SSC(0x2, (pipe))
+#define   DP_SSC_PWR_GATE(pipe)			_DP_SSC(0x3, (pipe))
+#define   _DP_SSS(val, pipe)			((val) << (2 * (pipe) + 16))
+#define   DP_SSS_MASK(pipe)			_DP_SSS(0x3, (pipe))
+#define   DP_SSS_PWR_ON(pipe)			_DP_SSS(0x0, (pipe))
+#define   DP_SSS_CLK_GATE(pipe)			_DP_SSS(0x1, (pipe))
+#define   DP_SSS_RESET(pipe)			_DP_SSS(0x2, (pipe))
+#define   DP_SSS_PWR_GATE(pipe)			_DP_SSS(0x3, (pipe))
 
 /* See the PUNIT HAS v0.8 for the below bits */
 enum punit_power_well {
@@ -508,6 +544,11 @@ enum punit_power_well {
 	PUNIT_POWER_WELL_DPIO_TX_C_LANES_23	= 9,
 	PUNIT_POWER_WELL_DPIO_RX0		= 10,
 	PUNIT_POWER_WELL_DPIO_RX1		= 11,
+	PUNIT_POWER_WELL_DPIO_CMN_D		= 12,
+	/* FIXME: guesswork below */
+	PUNIT_POWER_WELL_DPIO_TX_D_LANES_01	= 13,
+	PUNIT_POWER_WELL_DPIO_TX_D_LANES_23	= 14,
+	PUNIT_POWER_WELL_DPIO_RX2		= 15,
 
 	PUNIT_POWER_WELL_NUM,
 };
@@ -525,10 +566,21 @@ enum punit_power_well {
 #define PUNIT_REG_GPU_FREQ_STS			0xd8
 #define   GENFREQSTATUS				(1<<0)
 #define PUNIT_REG_MEDIA_TURBO_FREQ_REQ		0xdc
+#define PUNIT_REG_CZ_TIMESTAMP			0xce
 
 #define PUNIT_FUSE_BUS2				0xf6 /* bits 47:40 */
 #define PUNIT_FUSE_BUS1				0xf5 /* bits 55:48 */
 
+#define PUNIT_GPU_STATUS_REG			0xdb
+#define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT	16
+#define PUNIT_GPU_STATUS_MAX_FREQ_MASK		0xff
+#define PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT	8
+#define PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK	0xff
+
+#define PUNIT_GPU_DUTYCYCLE_REG		0xdf
+#define PUNIT_GPU_DUTYCYCLE_RPE_FREQ_SHIFT	8
+#define PUNIT_GPU_DUTYCYCLE_RPE_FREQ_MASK	0xff
+
 #define IOSF_NC_FB_GFX_FREQ_FUSE		0x1c
 #define   FB_GFX_MAX_FREQ_FUSE_SHIFT		3
 #define   FB_GFX_MAX_FREQ_FUSE_MASK		0x000007f8
@@ -540,6 +592,11 @@ enum punit_power_well {
 #define   FB_FMAX_VMIN_FREQ_LO_SHIFT		27
 #define   FB_FMAX_VMIN_FREQ_LO_MASK		0xf8000000
 
+#define VLV_CZ_CLOCK_TO_MILLI_SEC		100000
+#define VLV_RP_UP_EI_THRESHOLD			90
+#define VLV_RP_DOWN_EI_THRESHOLD		70
+#define VLV_INT_COUNT_FOR_DOWN_EI		5
+
 /* vlv2 north clock has */
 #define CCK_FUSE_REG				0x8
 #define  CCK_FUSE_HPLL_FREQ_MASK		0x3
@@ -574,6 +631,11 @@ enum punit_power_well {
 #define  DSI_PLL_M1_DIV_SHIFT			0
 #define  DSI_PLL_M1_DIV_MASK			(0x1ff << 0)
 #define CCK_DISPLAY_CLOCK_CONTROL		0x6b
+#define  DISPLAY_TRUNK_FORCE_ON			(1 << 17)
+#define  DISPLAY_TRUNK_FORCE_OFF		(1 << 16)
+#define  DISPLAY_FREQUENCY_STATUS		(0x1f << 8)
+#define  DISPLAY_FREQUENCY_STATUS_SHIFT		8
+#define  DISPLAY_FREQUENCY_VALUES		(0x1f << 0)
 
 /**
  * DOC: DPIO
@@ -761,6 +823,8 @@ enum punit_power_well {
 
 #define _VLV_PCS_DW8_CH0		0x8220
 #define _VLV_PCS_DW8_CH1		0x8420
+#define   CHV_PCS_USEDCLKCHANNEL_OVRRIDE	(1 << 20)
+#define   CHV_PCS_USEDCLKCHANNEL		(1 << 21)
 #define VLV_PCS_DW8(ch) _PORT(ch, _VLV_PCS_DW8_CH0, _VLV_PCS_DW8_CH1)
 
 #define _VLV_PCS01_DW8_CH0		0x0220
@@ -805,8 +869,8 @@ enum punit_power_well {
 
 #define _VLV_TX_DW2_CH0			0x8288
 #define _VLV_TX_DW2_CH1			0x8488
-#define   DPIO_SWING_MARGIN_SHIFT	16
-#define   DPIO_SWING_MARGIN_MASK	(0xff << DPIO_SWING_MARGIN_SHIFT)
+#define   DPIO_SWING_MARGIN000_SHIFT	16
+#define   DPIO_SWING_MARGIN000_MASK	(0xff << DPIO_SWING_MARGIN000_SHIFT)
 #define   DPIO_UNIQ_TRANS_SCALE_SHIFT	8
 #define VLV_TX_DW2(ch) _PORT(ch, _VLV_TX_DW2_CH0, _VLV_TX_DW2_CH1)
 
@@ -814,12 +878,16 @@ enum punit_power_well {
 #define _VLV_TX_DW3_CH1			0x848c
 /* The following bit for CHV phy */
 #define   DPIO_TX_UNIQ_TRANS_SCALE_EN	(1<<27)
+#define   DPIO_SWING_MARGIN101_SHIFT	16
+#define   DPIO_SWING_MARGIN101_MASK	(0xff << DPIO_SWING_MARGIN101_SHIFT)
 #define VLV_TX_DW3(ch) _PORT(ch, _VLV_TX_DW3_CH0, _VLV_TX_DW3_CH1)
 
 #define _VLV_TX_DW4_CH0			0x8290
 #define _VLV_TX_DW4_CH1			0x8490
 #define   DPIO_SWING_DEEMPH9P5_SHIFT	24
 #define   DPIO_SWING_DEEMPH9P5_MASK	(0xff << DPIO_SWING_DEEMPH9P5_SHIFT)
+#define   DPIO_SWING_DEEMPH6P0_SHIFT	16
+#define   DPIO_SWING_DEEMPH6P0_MASK	(0xff << DPIO_SWING_DEEMPH6P0_SHIFT)
 #define VLV_TX_DW4(ch) _PORT(ch, _VLV_TX_DW4_CH0, _VLV_TX_DW4_CH1)
 
 #define _VLV_TX3_DW4_CH0		0x690
@@ -869,6 +937,16 @@ enum punit_power_well {
 #define   DPIO_CHV_PROP_COEFF_SHIFT	0
 #define CHV_PLL_DW6(ch) _PIPE(ch, _CHV_PLL_DW6_CH0, _CHV_PLL_DW6_CH1)
 
+#define _CHV_CMN_DW5_CH0               0x8114
+#define   CHV_BUFRIGHTENA1_DISABLE	(0 << 20)
+#define   CHV_BUFRIGHTENA1_NORMAL	(1 << 20)
+#define   CHV_BUFRIGHTENA1_FORCE	(3 << 20)
+#define   CHV_BUFRIGHTENA1_MASK		(3 << 20)
+#define   CHV_BUFLEFTENA1_DISABLE	(0 << 22)
+#define   CHV_BUFLEFTENA1_NORMAL	(1 << 22)
+#define   CHV_BUFLEFTENA1_FORCE		(3 << 22)
+#define   CHV_BUFLEFTENA1_MASK		(3 << 22)
+
 #define _CHV_CMN_DW13_CH0		0x8134
 #define _CHV_CMN_DW0_CH1		0x8080
 #define   DPIO_CHV_S1_DIV_SHIFT		21
@@ -883,8 +961,21 @@ enum punit_power_well {
 #define _CHV_CMN_DW1_CH1		0x8084
 #define   DPIO_AFC_RECAL		(1 << 14)
 #define   DPIO_DCLKP_EN			(1 << 13)
+#define   CHV_BUFLEFTENA2_DISABLE	(0 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_NORMAL	(1 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_FORCE		(3 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_MASK		(3 << 17) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_DISABLE	(0 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_NORMAL	(1 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_FORCE	(3 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_MASK		(3 << 19) /* CL2 DW1 only */
 #define CHV_CMN_DW14(ch) _PIPE(ch, _CHV_CMN_DW14_CH0, _CHV_CMN_DW1_CH1)
 
+#define _CHV_CMN_DW19_CH0		0x814c
+#define _CHV_CMN_DW6_CH1		0x8098
+#define   CHV_CMN_USEDCLKCHANNEL	(1 << 13)
+#define CHV_CMN_DW19(ch) _PIPE(ch, _CHV_CMN_DW19_CH0, _CHV_CMN_DW6_CH1)
+
 #define CHV_CMN_DW30			0x8178
 #define   DPIO_LRC_BYPASS		(1 << 3)
 
@@ -933,6 +1024,7 @@ enum punit_power_well {
 #define   SANDYBRIDGE_FENCE_PITCH_SHIFT	32
 #define   GEN7_FENCE_MAX_PITCH_VAL	0x0800
 
+
 /* control register for cpu gtt access */
 #define TILECTL				0x101000
 #define   TILECTL_SWZCTL			(1 << 0)
@@ -946,6 +1038,13 @@ enum punit_power_well {
 #define   PGTBL_ADDRESS_LO_MASK	0xfffff000 /* bits [31:12] */
 #define   PGTBL_ADDRESS_HI_MASK	0x000000f0 /* bits [35:32] (gen4) */
 #define PGTBL_ER	0x02024
+#define PRB0_BASE (0x2030-0x30)
+#define PRB1_BASE (0x2040-0x30) /* 830,gen3 */
+#define PRB2_BASE (0x2050-0x30) /* gen3 */
+#define SRB0_BASE (0x2100-0x30) /* gen2 */
+#define SRB1_BASE (0x2110-0x30) /* gen2 */
+#define SRB2_BASE (0x2120-0x30) /* 830 */
+#define SRB3_BASE (0x2130-0x30) /* 830 */
 #define RENDER_RING_BASE	0x02000
 #define BSD_RING_BASE		0x04000
 #define GEN6_BSD_RING_BASE	0x12000
@@ -1007,6 +1106,7 @@ enum punit_power_well {
 #define RING_ACTHD_UDW(base)	((base)+0x5c)
 #define RING_NOPID(base)	((base)+0x94)
 #define RING_IMR(base)		((base)+0xa8)
+#define RING_HWSTAM(base)	((base)+0x98)
 #define RING_TIMESTAMP(base)	((base)+0x358)
 #define   TAIL_ADDR		0x001FFFF8
 #define   HEAD_WRAP_COUNT	0xFFE00000
@@ -1170,6 +1270,8 @@ enum punit_power_well {
 #define VLV_IMR		(VLV_DISPLAY_BASE + 0x20a8)
 #define VLV_ISR		(VLV_DISPLAY_BASE + 0x20ac)
 #define VLV_PCBR	(VLV_DISPLAY_BASE + 0x2120)
+#define VLV_PCBR_ADDR_SHIFT	12
+
 #define   DISPLAY_PLANE_FLIP_PENDING(plane) (1<<(11-(plane))) /* A and B only */
 #define EIR		0x020b0
 #define EMR		0x020b4
@@ -1189,6 +1291,10 @@ enum punit_power_well {
 #define   INSTPM_TLB_INVALIDATE	(1<<9)
 #define   INSTPM_SYNC_FLUSH	(1<<5)
 #define ACTHD	        0x020c8
+#define MEM_MODE	0x020cc
+#define   MEM_DISPLAY_B_TRICKLE_FEED_DISABLE (1<<3) /* 830 only */
+#define   MEM_DISPLAY_A_TRICKLE_FEED_DISABLE (1<<2) /* 830/845 only */
+#define   MEM_DISPLAY_TRICKLE_FEED_DISABLE (1<<2) /* 85x only */
 #define FW_BLC		0x020d8
 #define FW_BLC2		0x020dc
 #define FW_BLC_SELF	0x020e0 /* 915+ only */
@@ -1321,6 +1427,7 @@ enum punit_power_well {
 #define GT_BSD_CS_ERROR_INTERRUPT		(1 << 15)
 #define GT_BSD_USER_INTERRUPT			(1 << 12)
 #define GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1	(1 << 11) /* hsw+; rsvd on snb, ivb, vlv */
+#define GT_CONTEXT_SWITCH_INTERRUPT		(1 <<  8)
 #define GT_RENDER_L3_PARITY_ERROR_INTERRUPT	(1 <<  5) /* !snb */
 #define GT_RENDER_PIPECTL_NOTIFY_INTERRUPT	(1 <<  4)
 #define GT_RENDER_CS_MASTER_ERROR_INTERRUPT	(1 <<  3)
@@ -1460,6 +1567,7 @@ enum punit_power_well {
 /* Framebuffer compression for Ironlake */
 #define ILK_DPFC_CB_BASE	0x43200
 #define ILK_DPFC_CONTROL	0x43208
+#define   FBC_CTL_FALSE_COLOR	(1<<10)
 /* The bit 28-8 is reserved */
 #define   DPFC_RESERVED		(0x1FFFFF00)
 #define ILK_DPFC_RECOMP_CTL	0x4320c
@@ -1570,11 +1678,10 @@ enum punit_power_well {
 /*
  * Clock control & power management
  */
-#define DPLL_A_OFFSET 0x6014
-#define DPLL_B_OFFSET 0x6018
-#define CHV_DPLL_C_OFFSET 0x6030
-#define DPLL(pipe) (dev_priv->info.dpll_offsets[pipe] + \
-		    dev_priv->info.display_mmio_offset)
+#define _DPLL_A (dev_priv->info.display_mmio_offset + 0x6014)
+#define _DPLL_B (dev_priv->info.display_mmio_offset + 0x6018)
+#define _CHV_DPLL_C (dev_priv->info.display_mmio_offset + 0x6030)
+#define DPLL(pipe) _PIPE3((pipe), _DPLL_A, _DPLL_B, _CHV_DPLL_C)
 
 #define VGA0	0x6000
 #define VGA1	0x6004
@@ -1617,12 +1724,9 @@ enum punit_power_well {
 #define DPIO_PHY_STATUS			(VLV_DISPLAY_BASE + 0x6240)
 #define   DPLL_PORTD_READY_MASK		(0xf)
 #define DISPLAY_PHY_CONTROL (VLV_DISPLAY_BASE + 0x60100)
-#define   PHY_COM_LANE_RESET_DEASSERT(phy, val) \
-				((phy == DPIO_PHY0) ? (val | 1) : (val | 2))
-#define   PHY_COM_LANE_RESET_ASSERT(phy, val) \
-				((phy == DPIO_PHY0) ? (val & ~1) : (val & ~2))
+#define   PHY_COM_LANE_RESET_DEASSERT(phy) (1 << (phy))
 #define DISPLAY_PHY_STATUS (VLV_DISPLAY_BASE + 0x60104)
-#define   PHY_POWERGOOD(phy)	((phy == DPIO_PHY0) ? (1<<31) : (1<<30))
+#define   PHY_POWERGOOD(phy)	(((phy) == DPIO_PHY0) ? (1<<31) : (1<<30))
 
 /*
  * The i830 generation, in LVDS mode, defines P1 as the bit number set within
@@ -1662,11 +1766,10 @@ enum punit_power_well {
 #define   SDVO_MULTIPLIER_SHIFT_HIRES		4
 #define   SDVO_MULTIPLIER_SHIFT_VGA		0
 
-#define DPLL_A_MD_OFFSET 0x601c /* 965+ only */
-#define DPLL_B_MD_OFFSET 0x6020 /* 965+ only */
-#define CHV_DPLL_C_MD_OFFSET 0x603c
-#define DPLL_MD(pipe) (dev_priv->info.dpll_md_offsets[pipe] + \
-		       dev_priv->info.display_mmio_offset)
+#define _DPLL_A_MD (dev_priv->info.display_mmio_offset + 0x601c)
+#define _DPLL_B_MD (dev_priv->info.display_mmio_offset + 0x6020)
+#define _CHV_DPLL_C_MD (dev_priv->info.display_mmio_offset + 0x603c)
+#define DPLL_MD(pipe) _PIPE3((pipe), _DPLL_A_MD, _DPLL_B_MD, _CHV_DPLL_C_MD)
 
 /*
  * UDI pixel divider, controlling how many pixels are stuffed into a packet.
@@ -2231,7 +2334,7 @@ enum punit_power_well {
 /* Same as Haswell, but 72064 bytes now. */
 #define GEN8_CXT_TOTAL_SIZE		(18 * PAGE_SIZE)
 
-
+#define CHV_CLK_CTL1			0x101100
 #define VLV_CLK_CTL2			0x101104
 #define   CLK_CTL2_CZCOUNT_30NS_SHIFT	28
 
@@ -2340,6 +2443,7 @@ enum punit_power_well {
 #define _PIPEASRC	0x6001c
 #define _BCLRPAT_A	0x60020
 #define _VSYNCSHIFT_A	0x60028
+#define _PIPE_MULT_A	0x6002c
 
 /* Pipe B timing regs */
 #define _HTOTAL_B	0x61000
@@ -2351,6 +2455,7 @@ enum punit_power_well {
 #define _PIPEBSRC	0x6101c
 #define _BCLRPAT_B	0x61020
 #define _VSYNCSHIFT_B	0x61028
+#define _PIPE_MULT_B	0x6102c
 
 #define TRANSCODER_A_OFFSET 0x60000
 #define TRANSCODER_B_OFFSET 0x61000
@@ -2371,11 +2476,13 @@ enum punit_power_well {
 #define BCLRPAT(trans) _TRANSCODER2(trans, _BCLRPAT_A)
 #define VSYNCSHIFT(trans) _TRANSCODER2(trans, _VSYNCSHIFT_A)
 #define PIPESRC(trans) _TRANSCODER2(trans, _PIPEASRC)
+#define PIPE_MULT(trans) _TRANSCODER2(trans, _PIPE_MULT_A)
 
 /* HSW+ eDP PSR registers */
 #define EDP_PSR_BASE(dev)                       (IS_HASWELL(dev) ? 0x64800 : 0x6f800)
 #define EDP_PSR_CTL(dev)			(EDP_PSR_BASE(dev) + 0)
 #define   EDP_PSR_ENABLE			(1<<31)
+#define   BDW_PSR_SINGLE_FRAME			(1<<30)
 #define   EDP_PSR_LINK_DISABLE			(0<<27)
 #define   EDP_PSR_LINK_STANDBY			(1<<27)
 #define   EDP_PSR_MIN_LINK_ENTRY_TIME_MASK	(3<<25)
@@ -2533,8 +2640,14 @@ enum punit_power_well {
 #define   PORTC_HOTPLUG_LIVE_STATUS_VLV		(1 << 28)
 #define   PORTB_HOTPLUG_LIVE_STATUS_VLV		(1 << 29)
 #define   PORTD_HOTPLUG_INT_STATUS		(3 << 21)
+#define   PORTD_HOTPLUG_INT_LONG_PULSE		(2 << 21)
+#define   PORTD_HOTPLUG_INT_SHORT_PULSE		(1 << 21)
 #define   PORTC_HOTPLUG_INT_STATUS		(3 << 19)
+#define   PORTC_HOTPLUG_INT_LONG_PULSE		(2 << 19)
+#define   PORTC_HOTPLUG_INT_SHORT_PULSE		(1 << 19)
 #define   PORTB_HOTPLUG_INT_STATUS		(3 << 17)
+#define   PORTB_HOTPLUG_INT_LONG_PULSE		(2 << 17)
+#define   PORTB_HOTPLUG_INT_SHORT_PLUSE		(1 << 17)
 /* CRT/TV common between gen3+ */
 #define   CRT_HOTPLUG_INT_STATUS		(1 << 11)
 #define   TV_HOTPLUG_INT_STATUS			(1 << 10)
@@ -2588,7 +2701,7 @@ enum punit_power_well {
 
 #define PORT_DFT_I9XX				0x61150
 #define   DC_BALANCE_RESET			(1 << 25)
-#define PORT_DFT2_G4X				0x61154
+#define PORT_DFT2_G4X		(dev_priv->info.display_mmio_offset + 0x61154)
 #define   DC_BALANCE_RESET_VLV			(1 << 31)
 #define   PIPE_SCRAMBLE_RESET_MASK		(0x3 << 0)
 #define   PIPE_B_SCRAMBLE_RESET			(1 << 1)
@@ -3412,6 +3525,8 @@ enum punit_power_well {
 #define   DP_LINK_TRAIN_OFF		(3 << 28)
 #define   DP_LINK_TRAIN_MASK		(3 << 28)
 #define   DP_LINK_TRAIN_SHIFT		28
+#define   DP_LINK_TRAIN_PAT_3_CHV	(1 << 14)
+#define   DP_LINK_TRAIN_MASK_CHV	((3 << 28)|(1<<14))
 
 /* CPT Link training mode */
 #define   DP_LINK_TRAIN_PAT_1_CPT	(0 << 8)
@@ -3668,7 +3783,6 @@ enum punit_power_well {
 #define   PIPE_VSYNC_INTERRUPT_STATUS		(1UL<<9)
 #define   PIPE_DISPLAY_LINE_COMPARE_STATUS	(1UL<<8)
 #define   PIPE_DPST_EVENT_STATUS		(1UL<<7)
-#define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL<<6)
 #define   PIPE_A_PSR_STATUS_VLV			(1UL<<6)
 #define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL<<6)
 #define   PIPE_ODD_FIELD_INTERRUPT_STATUS	(1UL<<5)
@@ -3778,73 +3892,151 @@ enum punit_power_well {
 #define   DSPARB_BEND_SHIFT	9 /* on 855 */
 #define   DSPARB_AEND_SHIFT	0
 
+/* pnv/gen4/g4x/vlv/chv */
 #define DSPFW1			(dev_priv->info.display_mmio_offset + 0x70034)
-#define   DSPFW_SR_SHIFT	23
-#define   DSPFW_SR_MASK		(0x1ff<<23)
-#define   DSPFW_CURSORB_SHIFT	16
-#define   DSPFW_CURSORB_MASK	(0x3f<<16)
-#define   DSPFW_PLANEB_SHIFT	8
-#define   DSPFW_PLANEB_MASK	(0x7f<<8)
-#define   DSPFW_PLANEA_MASK	(0x7f)
+#define   DSPFW_SR_SHIFT		23
+#define   DSPFW_SR_MASK			(0x1ff<<23)
+#define   DSPFW_CURSORB_SHIFT		16
+#define   DSPFW_CURSORB_MASK		(0x3f<<16)
+#define   DSPFW_PLANEB_SHIFT		8
+#define   DSPFW_PLANEB_MASK		(0x7f<<8)
+#define   DSPFW_PLANEB_MASK_VLV		(0xff<<8) /* vlv/chv */
+#define   DSPFW_PLANEA_SHIFT		0
+#define   DSPFW_PLANEA_MASK		(0x7f<<0)
+#define   DSPFW_PLANEA_MASK_VLV		(0xff<<0) /* vlv/chv */
 #define DSPFW2			(dev_priv->info.display_mmio_offset + 0x70038)
-#define   DSPFW_CURSORA_MASK	0x00003f00
-#define   DSPFW_CURSORA_SHIFT	8
-#define   DSPFW_PLANEC_MASK	(0x7f)
+#define   DSPFW_FBC_SR_EN		(1<<31)	  /* g4x */
+#define   DSPFW_FBC_SR_SHIFT		28
+#define   DSPFW_FBC_SR_MASK		(0x7<<28) /* g4x */
+#define   DSPFW_FBC_HPLL_SR_SHIFT	24
+#define   DSPFW_FBC_HPLL_SR_MASK	(0xf<<24) /* g4x */
+#define   DSPFW_SPRITEB_SHIFT		(16)
+#define   DSPFW_SPRITEB_MASK		(0x7f<<16) /* g4x */
+#define   DSPFW_SPRITEB_MASK_VLV	(0xff<<16) /* vlv/chv */
+#define   DSPFW_CURSORA_SHIFT		8
+#define   DSPFW_CURSORA_MASK		(0x3f<<8)
+#define   DSPFW_PLANEC_SHIFT_OLD	0
+#define   DSPFW_PLANEC_MASK_OLD		(0x7f<<0) /* pre-gen4 sprite C */
+#define   DSPFW_SPRITEA_SHIFT		0
+#define   DSPFW_SPRITEA_MASK		(0x7f<<0) /* g4x */
+#define   DSPFW_SPRITEA_MASK_VLV	(0xff<<0) /* vlv/chv */
 #define DSPFW3			(dev_priv->info.display_mmio_offset + 0x7003c)
-#define   DSPFW_HPLL_SR_EN	(1<<31)
-#define   DSPFW_CURSOR_SR_SHIFT	24
+#define   DSPFW_HPLL_SR_EN		(1<<31)
 #define   PINEVIEW_SELF_REFRESH_EN	(1<<30)
+#define   DSPFW_CURSOR_SR_SHIFT		24
 #define   DSPFW_CURSOR_SR_MASK		(0x3f<<24)
 #define   DSPFW_HPLL_CURSOR_SHIFT	16
 #define   DSPFW_HPLL_CURSOR_MASK	(0x3f<<16)
-#define   DSPFW_HPLL_SR_MASK		(0x1ff)
-#define DSPFW4			(dev_priv->info.display_mmio_offset + 0x70070)
-#define DSPFW7			(dev_priv->info.display_mmio_offset + 0x7007c)
+#define   DSPFW_HPLL_SR_SHIFT		0
+#define   DSPFW_HPLL_SR_MASK		(0x1ff<<0)
+
+/* vlv/chv */
+#define DSPFW4			(VLV_DISPLAY_BASE + 0x70070)
+#define   DSPFW_SPRITEB_WM1_SHIFT	16
+#define   DSPFW_SPRITEB_WM1_MASK	(0xff<<16)
+#define   DSPFW_CURSORA_WM1_SHIFT	8
+#define   DSPFW_CURSORA_WM1_MASK	(0x3f<<8)
+#define   DSPFW_SPRITEA_WM1_SHIFT	0
+#define   DSPFW_SPRITEA_WM1_MASK	(0xff<<0)
+#define DSPFW5			(VLV_DISPLAY_BASE + 0x70074)
+#define   DSPFW_PLANEB_WM1_SHIFT	24
+#define   DSPFW_PLANEB_WM1_MASK		(0xff<<24)
+#define   DSPFW_PLANEA_WM1_SHIFT	16
+#define   DSPFW_PLANEA_WM1_MASK		(0xff<<16)
+#define   DSPFW_CURSORB_WM1_SHIFT	8
+#define   DSPFW_CURSORB_WM1_MASK	(0x3f<<8)
+#define   DSPFW_CURSOR_SR_WM1_SHIFT	0
+#define   DSPFW_CURSOR_SR_WM1_MASK	(0x3f<<0)
+#define DSPFW6			(VLV_DISPLAY_BASE + 0x70078)
+#define   DSPFW_SR_WM1_SHIFT		0
+#define   DSPFW_SR_WM1_MASK		(0x1ff<<0)
+#define DSPFW7			(VLV_DISPLAY_BASE + 0x7007c)
+#define DSPFW7_CHV		(VLV_DISPLAY_BASE + 0x700b4) /* wtf #1? */
+#define   DSPFW_SPRITED_WM1_SHIFT	24
+#define   DSPFW_SPRITED_WM1_MASK	(0xff<<24)
+#define   DSPFW_SPRITED_SHIFT		16
+#define   DSPFW_SPRITED_MASK		(0xff<<16)
+#define   DSPFW_SPRITEC_WM1_SHIFT	8
+#define   DSPFW_SPRITEC_WM1_MASK	(0xff<<8)
+#define   DSPFW_SPRITEC_SHIFT		0
+#define   DSPFW_SPRITEC_MASK		(0xff<<0)
+#define DSPFW8_CHV		(VLV_DISPLAY_BASE + 0x700b8)
+#define   DSPFW_SPRITEF_WM1_SHIFT	24
+#define   DSPFW_SPRITEF_WM1_MASK	(0xff<<24)
+#define   DSPFW_SPRITEF_SHIFT		16
+#define   DSPFW_SPRITEF_MASK		(0xff<<16)
+#define   DSPFW_SPRITEE_WM1_SHIFT	8
+#define   DSPFW_SPRITEE_WM1_MASK	(0xff<<8)
+#define   DSPFW_SPRITEE_SHIFT		0
+#define   DSPFW_SPRITEE_MASK		(0xff<<0)
+#define DSPFW9_CHV		(VLV_DISPLAY_BASE + 0x7007c) /* wtf #2? */
+#define   DSPFW_PLANEC_WM1_SHIFT	24
+#define   DSPFW_PLANEC_WM1_MASK		(0xff<<24)
+#define   DSPFW_PLANEC_SHIFT		16
+#define   DSPFW_PLANEC_MASK		(0xff<<16)
+#define   DSPFW_CURSORC_WM1_SHIFT	8
+#define   DSPFW_CURSORC_WM1_MASK	(0x3f<<16)
+#define   DSPFW_CURSORC_SHIFT		0
+#define   DSPFW_CURSORC_MASK		(0x3f<<0)
+
+/* vlv/chv high order bits */
+#define DSPHOWM			(VLV_DISPLAY_BASE + 0x70064)
+#define   DSPFW_SR_HI_SHIFT		24
+#define   DSPFW_SR_HI_MASK		(1<<24)
+#define   DSPFW_SPRITEF_HI_SHIFT	23
+#define   DSPFW_SPRITEF_HI_MASK		(1<<23)
+#define   DSPFW_SPRITEE_HI_SHIFT	22
+#define   DSPFW_SPRITEE_HI_MASK		(1<<22)
+#define   DSPFW_PLANEC_HI_SHIFT		21
+#define   DSPFW_PLANEC_HI_MASK		(1<<21)
+#define   DSPFW_SPRITED_HI_SHIFT	20
+#define   DSPFW_SPRITED_HI_MASK		(1<<20)
+#define   DSPFW_SPRITEC_HI_SHIFT	16
+#define   DSPFW_SPRITEC_HI_MASK		(1<<16)
+#define   DSPFW_PLANEB_HI_SHIFT		12
+#define   DSPFW_PLANEB_HI_MASK		(1<<12)
+#define   DSPFW_SPRITEB_HI_SHIFT	8
+#define   DSPFW_SPRITEB_HI_MASK		(1<<8)
+#define   DSPFW_SPRITEA_HI_SHIFT	4
+#define   DSPFW_SPRITEA_HI_MASK		(1<<4)
+#define   DSPFW_PLANEA_HI_SHIFT		0
+#define   DSPFW_PLANEA_HI_MASK		(1<<0)
+#define DSPHOWM1		(VLV_DISPLAY_BASE + 0x70068)
+#define   DSPFW_SR_WM1_HI_SHIFT		24
+#define   DSPFW_SR_WM1_HI_MASK		(1<<24)
+#define   DSPFW_SPRITEF_WM1_HI_SHIFT	23
+#define   DSPFW_SPRITEF_WM1_HI_MASK	(1<<23)
+#define   DSPFW_SPRITEE_WM1_HI_SHIFT	22
+#define   DSPFW_SPRITEE_WM1_HI_MASK	(1<<22)
+#define   DSPFW_PLANEC_WM1_HI_SHIFT	21
+#define   DSPFW_PLANEC_WM1_HI_MASK	(1<<21)
+#define   DSPFW_SPRITED_WM1_HI_SHIFT	20
+#define   DSPFW_SPRITED_WM1_HI_MASK	(1<<20)
+#define   DSPFW_SPRITEC_WM1_HI_SHIFT	16
+#define   DSPFW_SPRITEC_WM1_HI_MASK	(1<<16)
+#define   DSPFW_PLANEB_WM1_HI_SHIFT	12
+#define   DSPFW_PLANEB_WM1_HI_MASK	(1<<12)
+#define   DSPFW_SPRITEB_WM1_HI_SHIFT	8
+#define   DSPFW_SPRITEB_WM1_HI_MASK	(1<<8)
+#define   DSPFW_SPRITEA_WM1_HI_SHIFT	4
+#define   DSPFW_SPRITEA_WM1_HI_MASK	(1<<4)
+#define   DSPFW_PLANEA_WM1_HI_SHIFT	0
+#define   DSPFW_PLANEA_WM1_HI_MASK	(1<<0)
 
 /* drain latency register values*/
 #define DRAIN_LATENCY_PRECISION_32	32
-#define DRAIN_LATENCY_PRECISION_16	16
-#define VLV_DDL1			(VLV_DISPLAY_BASE + 0x70050)
-#define DDL_CURSORA_PRECISION_32	(1<<31)
-#define DDL_CURSORA_PRECISION_16	(0<<31)
-#define DDL_CURSORA_SHIFT		24
-#define DDL_SPRITEB_PRECISION_32	(1<<23)
-#define DDL_SPRITEB_PRECISION_16	(0<<23)
-#define DDL_SPRITEB_SHIFT		16
-#define DDL_SPRITEA_PRECISION_32	(1<<15)
-#define DDL_SPRITEA_PRECISION_16	(0<<15)
-#define DDL_SPRITEA_SHIFT		8
-#define DDL_PLANEA_PRECISION_32		(1<<7)
-#define DDL_PLANEA_PRECISION_16		(0<<7)
-#define DDL_PLANEA_SHIFT		0
-
-#define VLV_DDL2			(VLV_DISPLAY_BASE + 0x70054)
-#define DDL_CURSORB_PRECISION_32	(1<<31)
-#define DDL_CURSORB_PRECISION_16	(0<<31)
-#define DDL_CURSORB_SHIFT		24
-#define DDL_SPRITED_PRECISION_32	(1<<23)
-#define DDL_SPRITED_PRECISION_16	(0<<23)
-#define DDL_SPRITED_SHIFT		16
-#define DDL_SPRITEC_PRECISION_32	(1<<15)
-#define DDL_SPRITEC_PRECISION_16	(0<<15)
-#define DDL_SPRITEC_SHIFT		8
-#define DDL_PLANEB_PRECISION_32		(1<<7)
-#define DDL_PLANEB_PRECISION_16		(0<<7)
-#define DDL_PLANEB_SHIFT		0
-
-#define VLV_DDL3			(VLV_DISPLAY_BASE + 0x70058)
-#define DDL_CURSORC_PRECISION_32	(1<<31)
-#define DDL_CURSORC_PRECISION_16	(0<<31)
-#define DDL_CURSORC_SHIFT		24
-#define DDL_SPRITEF_PRECISION_32	(1<<23)
-#define DDL_SPRITEF_PRECISION_16	(0<<23)
-#define DDL_SPRITEF_SHIFT		16
-#define DDL_SPRITEE_PRECISION_32	(1<<15)
-#define DDL_SPRITEE_PRECISION_16	(0<<15)
-#define DDL_SPRITEE_SHIFT		8
-#define DDL_PLANEC_PRECISION_32		(1<<7)
-#define DDL_PLANEC_PRECISION_16		(0<<7)
-#define DDL_PLANEC_SHIFT		0
+#define DRAIN_LATENCY_PRECISION_64	64
+#define VLV_DDL(pipe)			(VLV_DISPLAY_BASE + 0x70050 + 4 * (pipe))
+#define DDL_CURSOR_PRECISION_64		(1<<31)
+#define DDL_CURSOR_PRECISION_32		(0<<31)
+#define DDL_CURSOR_SHIFT		24
+#define DDL_SPRITE_PRECISION_64(sprite)	(1<<(15+8*(sprite)))
+#define DDL_SPRITE_PRECISION_32(sprite)	(0<<(15+8*(sprite)))
+#define DDL_SPRITE_SHIFT(sprite)	(8+8*(sprite))
+#define DDL_PLANE_PRECISION_64		(1<<7)
+#define DDL_PLANE_PRECISION_32		(0<<7)
+#define DDL_PLANE_SHIFT			0
+#define DRAIN_LATENCY_MASK		0x7f
 
 /* FIFO watermark sizes etc */
 #define G4X_FIFO_LINE_SIZE	64
@@ -3962,7 +4154,8 @@ enum punit_power_well {
 /* Old style CUR*CNTR flags (desktop 8xx) */
 #define   CURSOR_ENABLE		0x80000000
 #define   CURSOR_GAMMA_ENABLE	0x40000000
-#define   CURSOR_STRIDE_MASK	0x30000000
+#define   CURSOR_STRIDE_SHIFT	28
+#define   CURSOR_STRIDE(x)	((ffs(x)-9) << CURSOR_STRIDE_SHIFT) /* 256,512,1k,2k */
 #define   CURSOR_PIPE_CSC_ENABLE (1<<24)
 #define   CURSOR_FORMAT_SHIFT	24
 #define   CURSOR_FORMAT_MASK	(0x07 << CURSOR_FORMAT_SHIFT)
@@ -4047,6 +4240,7 @@ enum punit_power_well {
 #define   DISPPLANE_NO_LINE_DOUBLE		0
 #define   DISPPLANE_STEREO_POLARITY_FIRST	0
 #define   DISPPLANE_STEREO_POLARITY_SECOND	(1<<18)
+#define   DISPPLANE_ROTATE_180         (1<<15)
 #define   DISPPLANE_TRICKLE_FEED_DISABLE	(1<<14) /* Ironlake */
 #define   DISPPLANE_TILED			(1<<10)
 #define _DSPAADDR				0x70184
@@ -4131,6 +4325,7 @@ enum punit_power_well {
 #define   DVS_YUV_ORDER_UYVY	(1<<16)
 #define   DVS_YUV_ORDER_YVYU	(2<<16)
 #define   DVS_YUV_ORDER_VYUY	(3<<16)
+#define   DVS_ROTATE_180	(1<<15)
 #define   DVS_DEST_KEY		(1<<2)
 #define   DVS_TRICKLE_FEED_DISABLE (1<<14)
 #define   DVS_TILED		(1<<10)
@@ -4201,6 +4396,7 @@ enum punit_power_well {
 #define   SPRITE_YUV_ORDER_UYVY		(1<<16)
 #define   SPRITE_YUV_ORDER_YVYU		(2<<16)
 #define   SPRITE_YUV_ORDER_VYUY		(3<<16)
+#define   SPRITE_ROTATE_180		(1<<15)
 #define   SPRITE_TRICKLE_FEED_DISABLE	(1<<14)
 #define   SPRITE_INT_GAMMA_ENABLE	(1<<13)
 #define   SPRITE_TILED			(1<<10)
@@ -4274,6 +4470,7 @@ enum punit_power_well {
 #define   SP_YUV_ORDER_UYVY		(1<<16)
 #define   SP_YUV_ORDER_YVYU		(2<<16)
 #define   SP_YUV_ORDER_VYUY		(3<<16)
+#define   SP_ROTATE_180			(1<<15)
 #define   SP_TILED			(1<<10)
 #define _SPALINOFF		(VLV_DISPLAY_BASE + 0x72184)
 #define _SPASTRIDE		(VLV_DISPLAY_BASE + 0x72188)
@@ -4630,6 +4827,8 @@ enum punit_power_well {
 #define GEN7_L3CNTLREG1				0xB01C
 #define  GEN7_WA_FOR_GEN7_L3_CONTROL			0x3C47FF8C
 #define  GEN7_L3AGDIS				(1<<19)
+#define GEN7_L3CNTLREG2				0xB020
+#define GEN7_L3CNTLREG3				0xB024
 
 #define GEN7_L3_CHICKEN_MODE_REGISTER		0xB030
 #define  GEN7_WA_L3_CHICKEN_MODE				0x20000000
@@ -4876,8 +5075,7 @@ enum punit_power_well {
 #define _PCH_TRANSA_LINK_M2	0xe0048
 #define _PCH_TRANSA_LINK_N2	0xe004c
 
-/* Per-transcoder DIP controls */
-
+/* Per-transcoder DIP controls (PCH) */
 #define _VIDEO_DIP_CTL_A         0xe0200
 #define _VIDEO_DIP_DATA_A        0xe0208
 #define _VIDEO_DIP_GCP_A         0xe0210
@@ -4890,6 +5088,7 @@ enum punit_power_well {
 #define TVIDEO_DIP_DATA(pipe) _PIPE(pipe, _VIDEO_DIP_DATA_A, _VIDEO_DIP_DATA_B)
 #define TVIDEO_DIP_GCP(pipe) _PIPE(pipe, _VIDEO_DIP_GCP_A, _VIDEO_DIP_GCP_B)
 
+/* Per-transcoder DIP controls (VLV) */
 #define VLV_VIDEO_DIP_CTL_A		(VLV_DISPLAY_BASE + 0x60200)
 #define VLV_VIDEO_DIP_DATA_A		(VLV_DISPLAY_BASE + 0x60208)
 #define VLV_VIDEO_DIP_GDCP_PAYLOAD_A	(VLV_DISPLAY_BASE + 0x60210)
@@ -4898,12 +5097,19 @@ enum punit_power_well {
 #define VLV_VIDEO_DIP_DATA_B		(VLV_DISPLAY_BASE + 0x61174)
 #define VLV_VIDEO_DIP_GDCP_PAYLOAD_B	(VLV_DISPLAY_BASE + 0x61178)
 
+#define CHV_VIDEO_DIP_CTL_C		(VLV_DISPLAY_BASE + 0x611f0)
+#define CHV_VIDEO_DIP_DATA_C		(VLV_DISPLAY_BASE + 0x611f4)
+#define CHV_VIDEO_DIP_GDCP_PAYLOAD_C	(VLV_DISPLAY_BASE + 0x611f8)
+
 #define VLV_TVIDEO_DIP_CTL(pipe) \
-	 _PIPE(pipe, VLV_VIDEO_DIP_CTL_A, VLV_VIDEO_DIP_CTL_B)
+	_PIPE3((pipe), VLV_VIDEO_DIP_CTL_A, \
+	       VLV_VIDEO_DIP_CTL_B, CHV_VIDEO_DIP_CTL_C)
 #define VLV_TVIDEO_DIP_DATA(pipe) \
-	 _PIPE(pipe, VLV_VIDEO_DIP_DATA_A, VLV_VIDEO_DIP_DATA_B)
+	_PIPE3((pipe), VLV_VIDEO_DIP_DATA_A, \
+	       VLV_VIDEO_DIP_DATA_B, CHV_VIDEO_DIP_DATA_C)
 #define VLV_TVIDEO_DIP_GCP(pipe) \
-	_PIPE(pipe, VLV_VIDEO_DIP_GDCP_PAYLOAD_A, VLV_VIDEO_DIP_GDCP_PAYLOAD_B)
+	_PIPE3((pipe), VLV_VIDEO_DIP_GDCP_PAYLOAD_A, \
+		VLV_VIDEO_DIP_GDCP_PAYLOAD_B, CHV_VIDEO_DIP_GDCP_PAYLOAD_C)
 
 /* Haswell DIP controls */
 #define HSW_VIDEO_DIP_CTL_A		0x60200
@@ -5173,8 +5379,7 @@ enum punit_power_well {
 #define PIPEA_PP_STATUS         (VLV_DISPLAY_BASE + 0x61200)
 #define PIPEA_PP_CONTROL        (VLV_DISPLAY_BASE + 0x61204)
 #define PIPEA_PP_ON_DELAYS      (VLV_DISPLAY_BASE + 0x61208)
-#define  PANEL_PORT_SELECT_DPB_VLV	(1 << 30)
-#define  PANEL_PORT_SELECT_DPC_VLV	(2 << 30)
+#define  PANEL_PORT_SELECT_VLV(port)	((port) << 30)
 #define PIPEA_PP_OFF_DELAYS     (VLV_DISPLAY_BASE + 0x6120c)
 #define PIPEA_PP_DIVISOR        (VLV_DISPLAY_BASE + 0x61210)
 
@@ -5481,6 +5686,8 @@ enum punit_power_well {
 #define GEN6_GT_GFX_RC6_LOCKED			0x138104
 #define VLV_COUNTER_CONTROL			0x138104
 #define   VLV_COUNT_RANGE_HIGH			(1<<15)
+#define   VLV_MEDIA_RC0_COUNT_EN		(1<<5)
+#define   VLV_RENDER_RC0_COUNT_EN		(1<<4)
 #define   VLV_MEDIA_RC6_COUNT_EN		(1<<1)
 #define   VLV_RENDER_RC6_COUNT_EN		(1<<0)
 #define GEN6_GT_GFX_RC6				0x138108
@@ -5489,6 +5696,8 @@ enum punit_power_well {
 
 #define GEN6_GT_GFX_RC6p			0x13810C
 #define GEN6_GT_GFX_RC6pp			0x138110
+#define VLV_RENDER_C0_COUNT_REG		0x138118
+#define VLV_MEDIA_C0_COUNT_REG			0x13811C
 
 #define GEN6_PCODE_MAILBOX			0x138124
 #define   GEN6_PCODE_READY			(1<<31)
@@ -5723,6 +5932,7 @@ enum punit_power_well {
 #define  TRANS_DDI_FUNC_ENABLE		(1<<31)
 /* Those bits are ignored by pipe EDP since it can only connect to DDI A */
 #define  TRANS_DDI_PORT_MASK		(7<<28)
+#define  TRANS_DDI_PORT_SHIFT		28
 #define  TRANS_DDI_SELECT_PORT(x)	((x)<<28)
 #define  TRANS_DDI_PORT_NONE		(0<<28)
 #define  TRANS_DDI_MODE_SELECT_MASK	(7<<24)
@@ -5743,6 +5953,7 @@ enum punit_power_well {
 #define  TRANS_DDI_EDP_INPUT_A_ONOFF	(4<<12)
 #define  TRANS_DDI_EDP_INPUT_B_ONOFF	(5<<12)
 #define  TRANS_DDI_EDP_INPUT_C_ONOFF	(6<<12)
+#define  TRANS_DDI_DP_VC_PAYLOAD_ALLOC	(1<<8)
 #define  TRANS_DDI_BFI_ENABLE		(1<<4)
 
 /* DisplayPort Transport Control */
@@ -5752,6 +5963,7 @@ enum punit_power_well {
 #define  DP_TP_CTL_ENABLE			(1<<31)
 #define  DP_TP_CTL_MODE_SST			(0<<27)
 #define  DP_TP_CTL_MODE_MST			(1<<27)
+#define  DP_TP_CTL_FORCE_ACT			(1<<25)
 #define  DP_TP_CTL_ENHANCED_FRAME_ENABLE	(1<<18)
 #define  DP_TP_CTL_FDI_AUTOTRAIN		(1<<15)
 #define  DP_TP_CTL_LINK_TRAIN_MASK		(7<<8)
@@ -5766,34 +5978,20 @@ enum punit_power_well {
 #define DP_TP_STATUS_A			0x64044
 #define DP_TP_STATUS_B			0x64144
 #define DP_TP_STATUS(port) _PORT(port, DP_TP_STATUS_A, DP_TP_STATUS_B)
-#define  DP_TP_STATUS_IDLE_DONE		(1<<25)
-#define  DP_TP_STATUS_AUTOTRAIN_DONE	(1<<12)
+#define  DP_TP_STATUS_IDLE_DONE			(1<<25)
+#define  DP_TP_STATUS_ACT_SENT			(1<<24)
+#define  DP_TP_STATUS_MODE_STATUS_MST		(1<<23)
+#define  DP_TP_STATUS_AUTOTRAIN_DONE		(1<<12)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC2	(3 << 8)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC1	(3 << 4)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC0	(3 << 0)
 
 /* DDI Buffer Control */
 #define DDI_BUF_CTL_A				0x64000
 #define DDI_BUF_CTL_B				0x64100
 #define DDI_BUF_CTL(port) _PORT(port, DDI_BUF_CTL_A, DDI_BUF_CTL_B)
 #define  DDI_BUF_CTL_ENABLE			(1<<31)
-/* Haswell */
-#define  DDI_BUF_EMP_400MV_0DB_HSW		(0<<24)   /* Sel0 */
-#define  DDI_BUF_EMP_400MV_3_5DB_HSW		(1<<24)   /* Sel1 */
-#define  DDI_BUF_EMP_400MV_6DB_HSW		(2<<24)   /* Sel2 */
-#define  DDI_BUF_EMP_400MV_9_5DB_HSW		(3<<24)   /* Sel3 */
-#define  DDI_BUF_EMP_600MV_0DB_HSW		(4<<24)   /* Sel4 */
-#define  DDI_BUF_EMP_600MV_3_5DB_HSW		(5<<24)   /* Sel5 */
-#define  DDI_BUF_EMP_600MV_6DB_HSW		(6<<24)   /* Sel6 */
-#define  DDI_BUF_EMP_800MV_0DB_HSW		(7<<24)   /* Sel7 */
-#define  DDI_BUF_EMP_800MV_3_5DB_HSW		(8<<24)   /* Sel8 */
-/* Broadwell */
-#define  DDI_BUF_EMP_400MV_0DB_BDW		(0<<24)   /* Sel0 */
-#define  DDI_BUF_EMP_400MV_3_5DB_BDW		(1<<24)   /* Sel1 */
-#define  DDI_BUF_EMP_400MV_6DB_BDW		(2<<24)   /* Sel2 */
-#define  DDI_BUF_EMP_600MV_0DB_BDW		(3<<24)   /* Sel3 */
-#define  DDI_BUF_EMP_600MV_3_5DB_BDW		(4<<24)   /* Sel4 */
-#define  DDI_BUF_EMP_600MV_6DB_BDW		(5<<24)   /* Sel5 */
-#define  DDI_BUF_EMP_800MV_0DB_BDW		(6<<24)   /* Sel6 */
-#define  DDI_BUF_EMP_800MV_3_5DB_BDW		(7<<24)   /* Sel7 */
-#define  DDI_BUF_EMP_1200MV_0DB_BDW		(8<<24)   /* Sel8 */
+#define  DDI_BUF_TRANS_SELECT(n)	((n) << 24)
 #define  DDI_BUF_EMP_MASK			(0xf<<24)
 #define  DDI_BUF_PORT_REVERSAL			(1<<16)
 #define  DDI_BUF_IS_IDLE			(1<<7)
@@ -5861,10 +6059,12 @@ enum punit_power_well {
 /* WRPLL */
 #define WRPLL_CTL1			0x46040
 #define WRPLL_CTL2			0x46060
+#define WRPLL_CTL(pll)			(pll == 0 ? WRPLL_CTL1 : WRPLL_CTL2)
 #define  WRPLL_PLL_ENABLE		(1<<31)
-#define  WRPLL_PLL_SELECT_SSC		(0x01<<28)
-#define  WRPLL_PLL_SELECT_NON_SSC	(0x02<<28)
-#define  WRPLL_PLL_SELECT_LCPLL_2700	(0x03<<28)
+#define  WRPLL_PLL_SSC			(1<<28)
+#define  WRPLL_PLL_NON_SSC		(2<<28)
+#define  WRPLL_PLL_LCPLL		(3<<28)
+#define  WRPLL_PLL_REF_MASK		(3<<28)
 /* WRPLL divider programming */
 #define  WRPLL_DIVIDER_REFERENCE(x)	((x)<<0)
 #define  WRPLL_DIVIDER_REF_MASK		(0xff)
@@ -5883,6 +6083,7 @@ enum punit_power_well {
 #define  PORT_CLK_SEL_LCPLL_1350	(1<<29)
 #define  PORT_CLK_SEL_LCPLL_810		(2<<29)
 #define  PORT_CLK_SEL_SPLL		(3<<29)
+#define  PORT_CLK_SEL_WRPLL(pll)	(((pll)+4)<<29)
 #define  PORT_CLK_SEL_WRPLL1		(4<<29)
 #define  PORT_CLK_SEL_WRPLL2		(5<<29)
 #define  PORT_CLK_SEL_NONE		(7<<29)
@@ -5924,7 +6125,10 @@ enum punit_power_well {
 #define  LCPLL_CD_SOURCE_FCLK		(1<<21)
 #define  LCPLL_CD_SOURCE_FCLK_DONE	(1<<19)
 
-#define D_COMP				(MCHBAR_MIRROR_BASE_SNB + 0x5F0C)
+/* Please see hsw_read_dcomp() and hsw_write_dcomp() before using this register,
+ * since on HSW we can't write to it using I915_WRITE. */
+#define D_COMP_HSW			(MCHBAR_MIRROR_BASE_SNB + 0x5F0C)
+#define D_COMP_BDW			0x138144
 #define  D_COMP_RCOMP_IN_PROGRESS	(1<<9)
 #define  D_COMP_COMP_FORCE		(1<<8)
 #define  D_COMP_COMP_DISABLE		(1<<0)
@@ -6005,7 +6209,8 @@ enum punit_power_well {
 
 #define _MIPIA_PORT_CTRL			(VLV_DISPLAY_BASE + 0x61190)
 #define _MIPIB_PORT_CTRL			(VLV_DISPLAY_BASE + 0x61700)
-#define MIPI_PORT_CTRL(pipe)		_PIPE(pipe, _MIPIA_PORT_CTRL, _MIPIB_PORT_CTRL)
+#define MIPI_PORT_CTRL(tc)		_TRANSCODER(tc, _MIPIA_PORT_CTRL, \
+						_MIPIB_PORT_CTRL)
 #define  DPI_ENABLE					(1 << 31) /* A + B */
 #define  MIPIA_MIPI4DPHY_DELAY_COUNT_SHIFT		27
 #define  MIPIA_MIPI4DPHY_DELAY_COUNT_MASK		(0xf << 27)
@@ -6047,18 +6252,20 @@ enum punit_power_well {
 
 #define _MIPIA_TEARING_CTRL			(VLV_DISPLAY_BASE + 0x61194)
 #define _MIPIB_TEARING_CTRL			(VLV_DISPLAY_BASE + 0x61704)
-#define MIPI_TEARING_CTRL(pipe)		_PIPE(pipe, _MIPIA_TEARING_CTRL, _MIPIB_TEARING_CTRL)
+#define MIPI_TEARING_CTRL(tc)			_TRANSCODER(tc, \
+				_MIPIA_TEARING_CTRL, _MIPIB_TEARING_CTRL)
 #define  TEARING_EFFECT_DELAY_SHIFT			0
 #define  TEARING_EFFECT_DELAY_MASK			(0xffff << 0)
 
 /* XXX: all bits reserved */
-#define _MIPIA_AUTOPWG				(VLV_DISPLAY_BASE + 0x611a0)
+#define _MIPIA_AUTOPWG			(VLV_DISPLAY_BASE + 0x611a0)
 
 /* MIPI DSI Controller and D-PHY registers */
 
-#define _MIPIA_DEVICE_READY			(VLV_DISPLAY_BASE + 0xb000)
-#define _MIPIB_DEVICE_READY			(VLV_DISPLAY_BASE + 0xb800)
-#define MIPI_DEVICE_READY(pipe)		_PIPE(pipe, _MIPIA_DEVICE_READY, _MIPIB_DEVICE_READY)
+#define _MIPIA_DEVICE_READY		(dev_priv->mipi_mmio_base + 0xb000)
+#define _MIPIB_DEVICE_READY		(dev_priv->mipi_mmio_base + 0xb800)
+#define MIPI_DEVICE_READY(tc)		_TRANSCODER(tc, _MIPIA_DEVICE_READY, \
+						_MIPIB_DEVICE_READY)
 #define  BUS_POSSESSION					(1 << 3) /* set to give bus to receiver */
 #define  ULPS_STATE_MASK				(3 << 1)
 #define  ULPS_STATE_ENTER				(2 << 1)
@@ -6066,12 +6273,14 @@ enum punit_power_well {
 #define  ULPS_STATE_NORMAL_OPERATION			(0 << 1)
 #define  DEVICE_READY					(1 << 0)
 
-#define _MIPIA_INTR_STAT			(VLV_DISPLAY_BASE + 0xb004)
-#define _MIPIB_INTR_STAT			(VLV_DISPLAY_BASE + 0xb804)
-#define MIPI_INTR_STAT(pipe)		_PIPE(pipe, _MIPIA_INTR_STAT, _MIPIB_INTR_STAT)
-#define _MIPIA_INTR_EN				(VLV_DISPLAY_BASE + 0xb008)
-#define _MIPIB_INTR_EN				(VLV_DISPLAY_BASE + 0xb808)
-#define MIPI_INTR_EN(pipe)		_PIPE(pipe, _MIPIA_INTR_EN, _MIPIB_INTR_EN)
+#define _MIPIA_INTR_STAT		(dev_priv->mipi_mmio_base + 0xb004)
+#define _MIPIB_INTR_STAT		(dev_priv->mipi_mmio_base + 0xb804)
+#define MIPI_INTR_STAT(tc)		_TRANSCODER(tc, _MIPIA_INTR_STAT, \
+					_MIPIB_INTR_STAT)
+#define _MIPIA_INTR_EN			(dev_priv->mipi_mmio_base + 0xb008)
+#define _MIPIB_INTR_EN			(dev_priv->mipi_mmio_base + 0xb808)
+#define MIPI_INTR_EN(tc)		_TRANSCODER(tc, _MIPIA_INTR_EN, \
+					_MIPIB_INTR_EN)
 #define  TEARING_EFFECT					(1 << 31)
 #define  SPL_PKT_SENT_INTERRUPT				(1 << 30)
 #define  GEN_READ_DATA_AVAIL				(1 << 29)
@@ -6105,9 +6314,10 @@ enum punit_power_well {
 #define  RXSOT_SYNC_ERROR				(1 << 1)
 #define  RXSOT_ERROR					(1 << 0)
 
-#define _MIPIA_DSI_FUNC_PRG			(VLV_DISPLAY_BASE + 0xb00c)
-#define _MIPIB_DSI_FUNC_PRG			(VLV_DISPLAY_BASE + 0xb80c)
-#define MIPI_DSI_FUNC_PRG(pipe)		_PIPE(pipe, _MIPIA_DSI_FUNC_PRG, _MIPIB_DSI_FUNC_PRG)
+#define _MIPIA_DSI_FUNC_PRG		(dev_priv->mipi_mmio_base + 0xb00c)
+#define _MIPIB_DSI_FUNC_PRG		(dev_priv->mipi_mmio_base + 0xb80c)
+#define MIPI_DSI_FUNC_PRG(tc)		_TRANSCODER(tc, _MIPIA_DSI_FUNC_PRG, \
+						_MIPIB_DSI_FUNC_PRG)
 #define  CMD_MODE_DATA_WIDTH_MASK			(7 << 13)
 #define  CMD_MODE_NOT_SUPPORTED				(0 << 13)
 #define  CMD_MODE_DATA_WIDTH_16_BIT			(1 << 13)
@@ -6128,78 +6338,94 @@ enum punit_power_well {
 #define  DATA_LANES_PRG_REG_SHIFT			0
 #define  DATA_LANES_PRG_REG_MASK			(7 << 0)
 
-#define _MIPIA_HS_TX_TIMEOUT			(VLV_DISPLAY_BASE + 0xb010)
-#define _MIPIB_HS_TX_TIMEOUT			(VLV_DISPLAY_BASE + 0xb810)
-#define MIPI_HS_TX_TIMEOUT(pipe)	_PIPE(pipe, _MIPIA_HS_TX_TIMEOUT, _MIPIB_HS_TX_TIMEOUT)
+#define _MIPIA_HS_TX_TIMEOUT		(dev_priv->mipi_mmio_base + 0xb010)
+#define _MIPIB_HS_TX_TIMEOUT		(dev_priv->mipi_mmio_base + 0xb810)
+#define MIPI_HS_TX_TIMEOUT(tc)	_TRANSCODER(tc, _MIPIA_HS_TX_TIMEOUT, \
+					_MIPIB_HS_TX_TIMEOUT)
 #define  HIGH_SPEED_TX_TIMEOUT_COUNTER_MASK		0xffffff
 
-#define _MIPIA_LP_RX_TIMEOUT			(VLV_DISPLAY_BASE + 0xb014)
-#define _MIPIB_LP_RX_TIMEOUT			(VLV_DISPLAY_BASE + 0xb814)
-#define MIPI_LP_RX_TIMEOUT(pipe)	_PIPE(pipe, _MIPIA_LP_RX_TIMEOUT, _MIPIB_LP_RX_TIMEOUT)
+#define _MIPIA_LP_RX_TIMEOUT		(dev_priv->mipi_mmio_base + 0xb014)
+#define _MIPIB_LP_RX_TIMEOUT		(dev_priv->mipi_mmio_base + 0xb814)
+#define MIPI_LP_RX_TIMEOUT(tc)	_TRANSCODER(tc, _MIPIA_LP_RX_TIMEOUT, \
+					_MIPIB_LP_RX_TIMEOUT)
 #define  LOW_POWER_RX_TIMEOUT_COUNTER_MASK		0xffffff
 
-#define _MIPIA_TURN_AROUND_TIMEOUT		(VLV_DISPLAY_BASE + 0xb018)
-#define _MIPIB_TURN_AROUND_TIMEOUT		(VLV_DISPLAY_BASE + 0xb818)
-#define MIPI_TURN_AROUND_TIMEOUT(pipe)	_PIPE(pipe, _MIPIA_TURN_AROUND_TIMEOUT, _MIPIB_TURN_AROUND_TIMEOUT)
+#define _MIPIA_TURN_AROUND_TIMEOUT	(dev_priv->mipi_mmio_base + 0xb018)
+#define _MIPIB_TURN_AROUND_TIMEOUT	(dev_priv->mipi_mmio_base + 0xb818)
+#define MIPI_TURN_AROUND_TIMEOUT(tc)	_TRANSCODER(tc, \
+			_MIPIA_TURN_AROUND_TIMEOUT, _MIPIB_TURN_AROUND_TIMEOUT)
 #define  TURN_AROUND_TIMEOUT_MASK			0x3f
 
-#define _MIPIA_DEVICE_RESET_TIMER		(VLV_DISPLAY_BASE + 0xb01c)
-#define _MIPIB_DEVICE_RESET_TIMER		(VLV_DISPLAY_BASE + 0xb81c)
-#define MIPI_DEVICE_RESET_TIMER(pipe)	_PIPE(pipe, _MIPIA_DEVICE_RESET_TIMER, _MIPIB_DEVICE_RESET_TIMER)
+#define _MIPIA_DEVICE_RESET_TIMER	(dev_priv->mipi_mmio_base + 0xb01c)
+#define _MIPIB_DEVICE_RESET_TIMER	(dev_priv->mipi_mmio_base + 0xb81c)
+#define MIPI_DEVICE_RESET_TIMER(tc)	_TRANSCODER(tc, \
+			_MIPIA_DEVICE_RESET_TIMER, _MIPIB_DEVICE_RESET_TIMER)
 #define  DEVICE_RESET_TIMER_MASK			0xffff
 
-#define _MIPIA_DPI_RESOLUTION			(VLV_DISPLAY_BASE + 0xb020)
-#define _MIPIB_DPI_RESOLUTION			(VLV_DISPLAY_BASE + 0xb820)
-#define MIPI_DPI_RESOLUTION(pipe)	_PIPE(pipe, _MIPIA_DPI_RESOLUTION, _MIPIB_DPI_RESOLUTION)
+#define _MIPIA_DPI_RESOLUTION		(dev_priv->mipi_mmio_base + 0xb020)
+#define _MIPIB_DPI_RESOLUTION		(dev_priv->mipi_mmio_base + 0xb820)
+#define MIPI_DPI_RESOLUTION(tc)	_TRANSCODER(tc, _MIPIA_DPI_RESOLUTION, \
+					_MIPIB_DPI_RESOLUTION)
 #define  VERTICAL_ADDRESS_SHIFT				16
 #define  VERTICAL_ADDRESS_MASK				(0xffff << 16)
 #define  HORIZONTAL_ADDRESS_SHIFT			0
 #define  HORIZONTAL_ADDRESS_MASK			0xffff
 
-#define _MIPIA_DBI_FIFO_THROTTLE		(VLV_DISPLAY_BASE + 0xb024)
-#define _MIPIB_DBI_FIFO_THROTTLE		(VLV_DISPLAY_BASE + 0xb824)
-#define MIPI_DBI_FIFO_THROTTLE(pipe)	_PIPE(pipe, _MIPIA_DBI_FIFO_THROTTLE, _MIPIB_DBI_FIFO_THROTTLE)
+#define _MIPIA_DBI_FIFO_THROTTLE	(dev_priv->mipi_mmio_base + 0xb024)
+#define _MIPIB_DBI_FIFO_THROTTLE	(dev_priv->mipi_mmio_base + 0xb824)
+#define MIPI_DBI_FIFO_THROTTLE(tc)	_TRANSCODER(tc, \
+			_MIPIA_DBI_FIFO_THROTTLE, _MIPIB_DBI_FIFO_THROTTLE)
 #define  DBI_FIFO_EMPTY_HALF				(0 << 0)
 #define  DBI_FIFO_EMPTY_QUARTER				(1 << 0)
 #define  DBI_FIFO_EMPTY_7_LOCATIONS			(2 << 0)
 
 /* regs below are bits 15:0 */
-#define _MIPIA_HSYNC_PADDING_COUNT		(VLV_DISPLAY_BASE + 0xb028)
-#define _MIPIB_HSYNC_PADDING_COUNT		(VLV_DISPLAY_BASE + 0xb828)
-#define MIPI_HSYNC_PADDING_COUNT(pipe)	_PIPE(pipe, _MIPIA_HSYNC_PADDING_COUNT, _MIPIB_HSYNC_PADDING_COUNT)
-
-#define _MIPIA_HBP_COUNT			(VLV_DISPLAY_BASE + 0xb02c)
-#define _MIPIB_HBP_COUNT			(VLV_DISPLAY_BASE + 0xb82c)
-#define MIPI_HBP_COUNT(pipe)		_PIPE(pipe, _MIPIA_HBP_COUNT, _MIPIB_HBP_COUNT)
-
-#define _MIPIA_HFP_COUNT			(VLV_DISPLAY_BASE + 0xb030)
-#define _MIPIB_HFP_COUNT			(VLV_DISPLAY_BASE + 0xb830)
-#define MIPI_HFP_COUNT(pipe)		_PIPE(pipe, _MIPIA_HFP_COUNT, _MIPIB_HFP_COUNT)
-
-#define _MIPIA_HACTIVE_AREA_COUNT		(VLV_DISPLAY_BASE + 0xb034)
-#define _MIPIB_HACTIVE_AREA_COUNT		(VLV_DISPLAY_BASE + 0xb834)
-#define MIPI_HACTIVE_AREA_COUNT(pipe)	_PIPE(pipe, _MIPIA_HACTIVE_AREA_COUNT, _MIPIB_HACTIVE_AREA_COUNT)
-
-#define _MIPIA_VSYNC_PADDING_COUNT		(VLV_DISPLAY_BASE + 0xb038)
-#define _MIPIB_VSYNC_PADDING_COUNT		(VLV_DISPLAY_BASE + 0xb838)
-#define MIPI_VSYNC_PADDING_COUNT(pipe)	_PIPE(pipe, _MIPIA_VSYNC_PADDING_COUNT, _MIPIB_VSYNC_PADDING_COUNT)
+#define _MIPIA_HSYNC_PADDING_COUNT	(dev_priv->mipi_mmio_base + 0xb028)
+#define _MIPIB_HSYNC_PADDING_COUNT	(dev_priv->mipi_mmio_base + 0xb828)
+#define MIPI_HSYNC_PADDING_COUNT(tc)	_TRANSCODER(tc, \
+			_MIPIA_HSYNC_PADDING_COUNT, _MIPIB_HSYNC_PADDING_COUNT)
+
+#define _MIPIA_HBP_COUNT		(dev_priv->mipi_mmio_base + 0xb02c)
+#define _MIPIB_HBP_COUNT		(dev_priv->mipi_mmio_base + 0xb82c)
+#define MIPI_HBP_COUNT(tc)		_TRANSCODER(tc, _MIPIA_HBP_COUNT, \
+					_MIPIB_HBP_COUNT)
+
+#define _MIPIA_HFP_COUNT		(dev_priv->mipi_mmio_base + 0xb030)
+#define _MIPIB_HFP_COUNT		(dev_priv->mipi_mmio_base + 0xb830)
+#define MIPI_HFP_COUNT(tc)		_TRANSCODER(tc, _MIPIA_HFP_COUNT, \
+					_MIPIB_HFP_COUNT)
+
+#define _MIPIA_HACTIVE_AREA_COUNT	(dev_priv->mipi_mmio_base + 0xb034)
+#define _MIPIB_HACTIVE_AREA_COUNT	(dev_priv->mipi_mmio_base + 0xb834)
+#define MIPI_HACTIVE_AREA_COUNT(tc)	_TRANSCODER(tc, \
+			_MIPIA_HACTIVE_AREA_COUNT, _MIPIB_HACTIVE_AREA_COUNT)
+
+#define _MIPIA_VSYNC_PADDING_COUNT	(dev_priv->mipi_mmio_base + 0xb038)
+#define _MIPIB_VSYNC_PADDING_COUNT	(dev_priv->mipi_mmio_base + 0xb838)
+#define MIPI_VSYNC_PADDING_COUNT(tc)	_TRANSCODER(tc, \
+			_MIPIA_VSYNC_PADDING_COUNT, _MIPIB_VSYNC_PADDING_COUNT)
+
+#define _MIPIA_VBP_COUNT		(dev_priv->mipi_mmio_base + 0xb03c)
+#define _MIPIB_VBP_COUNT		(dev_priv->mipi_mmio_base + 0xb83c)
+#define MIPI_VBP_COUNT(tc)		_TRANSCODER(tc, _MIPIA_VBP_COUNT, \
+					_MIPIB_VBP_COUNT)
+
+#define _MIPIA_VFP_COUNT		(dev_priv->mipi_mmio_base + 0xb040)
+#define _MIPIB_VFP_COUNT		(dev_priv->mipi_mmio_base + 0xb840)
+#define MIPI_VFP_COUNT(tc)		_TRANSCODER(tc, _MIPIA_VFP_COUNT, \
+					_MIPIB_VFP_COUNT)
+
+#define _MIPIA_HIGH_LOW_SWITCH_COUNT	(dev_priv->mipi_mmio_base + 0xb044)
+#define _MIPIB_HIGH_LOW_SWITCH_COUNT	(dev_priv->mipi_mmio_base + 0xb844)
+#define MIPI_HIGH_LOW_SWITCH_COUNT(tc)	_TRANSCODER(tc,	\
+		_MIPIA_HIGH_LOW_SWITCH_COUNT, _MIPIB_HIGH_LOW_SWITCH_COUNT)
 
-#define _MIPIA_VBP_COUNT			(VLV_DISPLAY_BASE + 0xb03c)
-#define _MIPIB_VBP_COUNT			(VLV_DISPLAY_BASE + 0xb83c)
-#define MIPI_VBP_COUNT(pipe)		_PIPE(pipe, _MIPIA_VBP_COUNT, _MIPIB_VBP_COUNT)
-
-#define _MIPIA_VFP_COUNT			(VLV_DISPLAY_BASE + 0xb040)
-#define _MIPIB_VFP_COUNT			(VLV_DISPLAY_BASE + 0xb840)
-#define MIPI_VFP_COUNT(pipe)		_PIPE(pipe, _MIPIA_VFP_COUNT, _MIPIB_VFP_COUNT)
-
-#define _MIPIA_HIGH_LOW_SWITCH_COUNT		(VLV_DISPLAY_BASE + 0xb044)
-#define _MIPIB_HIGH_LOW_SWITCH_COUNT		(VLV_DISPLAY_BASE + 0xb844)
-#define MIPI_HIGH_LOW_SWITCH_COUNT(pipe)	_PIPE(pipe, _MIPIA_HIGH_LOW_SWITCH_COUNT, _MIPIB_HIGH_LOW_SWITCH_COUNT)
 /* regs above are bits 15:0 */
 
-#define _MIPIA_DPI_CONTROL			(VLV_DISPLAY_BASE + 0xb048)
-#define _MIPIB_DPI_CONTROL			(VLV_DISPLAY_BASE + 0xb848)
-#define MIPI_DPI_CONTROL(pipe)		_PIPE(pipe, _MIPIA_DPI_CONTROL, _MIPIB_DPI_CONTROL)
+#define _MIPIA_DPI_CONTROL		(dev_priv->mipi_mmio_base + 0xb048)
+#define _MIPIB_DPI_CONTROL		(dev_priv->mipi_mmio_base + 0xb848)
+#define MIPI_DPI_CONTROL(tc)		_TRANSCODER(tc, _MIPIA_DPI_CONTROL, \
+					_MIPIB_DPI_CONTROL)
 #define  DPI_LP_MODE					(1 << 6)
 #define  BACKLIGHT_OFF					(1 << 5)
 #define  BACKLIGHT_ON					(1 << 4)
@@ -6208,27 +6434,31 @@ enum punit_power_well {
 #define  TURN_ON					(1 << 1)
 #define  SHUTDOWN					(1 << 0)
 
-#define _MIPIA_DPI_DATA				(VLV_DISPLAY_BASE + 0xb04c)
-#define _MIPIB_DPI_DATA				(VLV_DISPLAY_BASE + 0xb84c)
-#define MIPI_DPI_DATA(pipe)		_PIPE(pipe, _MIPIA_DPI_DATA, _MIPIB_DPI_DATA)
+#define _MIPIA_DPI_DATA			(dev_priv->mipi_mmio_base + 0xb04c)
+#define _MIPIB_DPI_DATA			(dev_priv->mipi_mmio_base + 0xb84c)
+#define MIPI_DPI_DATA(tc)		_TRANSCODER(tc, _MIPIA_DPI_DATA, \
+					_MIPIB_DPI_DATA)
 #define  COMMAND_BYTE_SHIFT				0
 #define  COMMAND_BYTE_MASK				(0x3f << 0)
 
-#define _MIPIA_INIT_COUNT			(VLV_DISPLAY_BASE + 0xb050)
-#define _MIPIB_INIT_COUNT			(VLV_DISPLAY_BASE + 0xb850)
-#define MIPI_INIT_COUNT(pipe)		_PIPE(pipe, _MIPIA_INIT_COUNT, _MIPIB_INIT_COUNT)
+#define _MIPIA_INIT_COUNT		(dev_priv->mipi_mmio_base + 0xb050)
+#define _MIPIB_INIT_COUNT		(dev_priv->mipi_mmio_base + 0xb850)
+#define MIPI_INIT_COUNT(tc)		_TRANSCODER(tc, _MIPIA_INIT_COUNT, \
+					_MIPIB_INIT_COUNT)
 #define  MASTER_INIT_TIMER_SHIFT			0
 #define  MASTER_INIT_TIMER_MASK				(0xffff << 0)
 
-#define _MIPIA_MAX_RETURN_PKT_SIZE		(VLV_DISPLAY_BASE + 0xb054)
-#define _MIPIB_MAX_RETURN_PKT_SIZE		(VLV_DISPLAY_BASE + 0xb854)
-#define MIPI_MAX_RETURN_PKT_SIZE(pipe)	_PIPE(pipe, _MIPIA_MAX_RETURN_PKT_SIZE, _MIPIB_MAX_RETURN_PKT_SIZE)
+#define _MIPIA_MAX_RETURN_PKT_SIZE	(dev_priv->mipi_mmio_base + 0xb054)
+#define _MIPIB_MAX_RETURN_PKT_SIZE	(dev_priv->mipi_mmio_base + 0xb854)
+#define MIPI_MAX_RETURN_PKT_SIZE(tc)	_TRANSCODER(tc, \
+			_MIPIA_MAX_RETURN_PKT_SIZE, _MIPIB_MAX_RETURN_PKT_SIZE)
 #define  MAX_RETURN_PKT_SIZE_SHIFT			0
 #define  MAX_RETURN_PKT_SIZE_MASK			(0x3ff << 0)
 
-#define _MIPIA_VIDEO_MODE_FORMAT		(VLV_DISPLAY_BASE + 0xb058)
-#define _MIPIB_VIDEO_MODE_FORMAT		(VLV_DISPLAY_BASE + 0xb858)
-#define MIPI_VIDEO_MODE_FORMAT(pipe)	_PIPE(pipe, _MIPIA_VIDEO_MODE_FORMAT, _MIPIB_VIDEO_MODE_FORMAT)
+#define _MIPIA_VIDEO_MODE_FORMAT	(dev_priv->mipi_mmio_base + 0xb058)
+#define _MIPIB_VIDEO_MODE_FORMAT	(dev_priv->mipi_mmio_base + 0xb858)
+#define MIPI_VIDEO_MODE_FORMAT(tc)	_TRANSCODER(tc, \
+			_MIPIA_VIDEO_MODE_FORMAT, _MIPIB_VIDEO_MODE_FORMAT)
 #define  RANDOM_DPI_DISPLAY_RESOLUTION			(1 << 4)
 #define  DISABLE_VIDEO_BTA				(1 << 3)
 #define  IP_TG_CONFIG					(1 << 2)
@@ -6236,9 +6466,10 @@ enum punit_power_well {
 #define  VIDEO_MODE_NON_BURST_WITH_SYNC_EVENTS		(2 << 0)
 #define  VIDEO_MODE_BURST				(3 << 0)
 
-#define _MIPIA_EOT_DISABLE			(VLV_DISPLAY_BASE + 0xb05c)
-#define _MIPIB_EOT_DISABLE			(VLV_DISPLAY_BASE + 0xb85c)
-#define MIPI_EOT_DISABLE(pipe)		_PIPE(pipe, _MIPIA_EOT_DISABLE, _MIPIB_EOT_DISABLE)
+#define _MIPIA_EOT_DISABLE		(dev_priv->mipi_mmio_base + 0xb05c)
+#define _MIPIB_EOT_DISABLE		(dev_priv->mipi_mmio_base + 0xb85c)
+#define MIPI_EOT_DISABLE(tc)		_TRANSCODER(tc, _MIPIA_EOT_DISABLE, \
+					_MIPIB_EOT_DISABLE)
 #define  LP_RX_TIMEOUT_ERROR_RECOVERY_DISABLE		(1 << 7)
 #define  HS_RX_TIMEOUT_ERROR_RECOVERY_DISABLE		(1 << 6)
 #define  LOW_CONTENTION_RECOVERY_DISABLE		(1 << 5)
@@ -6248,28 +6479,33 @@ enum punit_power_well {
 #define  CLOCKSTOP					(1 << 1)
 #define  EOT_DISABLE					(1 << 0)
 
-#define _MIPIA_LP_BYTECLK			(VLV_DISPLAY_BASE + 0xb060)
-#define _MIPIB_LP_BYTECLK			(VLV_DISPLAY_BASE + 0xb860)
-#define MIPI_LP_BYTECLK(pipe)		_PIPE(pipe, _MIPIA_LP_BYTECLK, _MIPIB_LP_BYTECLK)
+#define _MIPIA_LP_BYTECLK		(dev_priv->mipi_mmio_base + 0xb060)
+#define _MIPIB_LP_BYTECLK		(dev_priv->mipi_mmio_base + 0xb860)
+#define MIPI_LP_BYTECLK(tc)		_TRANSCODER(tc, _MIPIA_LP_BYTECLK, \
+					_MIPIB_LP_BYTECLK)
 #define  LP_BYTECLK_SHIFT				0
 #define  LP_BYTECLK_MASK				(0xffff << 0)
 
 /* bits 31:0 */
-#define _MIPIA_LP_GEN_DATA			(VLV_DISPLAY_BASE + 0xb064)
-#define _MIPIB_LP_GEN_DATA			(VLV_DISPLAY_BASE + 0xb864)
-#define MIPI_LP_GEN_DATA(pipe)		_PIPE(pipe, _MIPIA_LP_GEN_DATA, _MIPIB_LP_GEN_DATA)
+#define _MIPIA_LP_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb064)
+#define _MIPIB_LP_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb864)
+#define MIPI_LP_GEN_DATA(tc)		_TRANSCODER(tc, _MIPIA_LP_GEN_DATA, \
+					_MIPIB_LP_GEN_DATA)
 
 /* bits 31:0 */
-#define _MIPIA_HS_GEN_DATA			(VLV_DISPLAY_BASE + 0xb068)
-#define _MIPIB_HS_GEN_DATA			(VLV_DISPLAY_BASE + 0xb868)
-#define MIPI_HS_GEN_DATA(pipe)		_PIPE(pipe, _MIPIA_HS_GEN_DATA, _MIPIB_HS_GEN_DATA)
-
-#define _MIPIA_LP_GEN_CTRL			(VLV_DISPLAY_BASE + 0xb06c)
-#define _MIPIB_LP_GEN_CTRL			(VLV_DISPLAY_BASE + 0xb86c)
-#define MIPI_LP_GEN_CTRL(pipe)		_PIPE(pipe, _MIPIA_LP_GEN_CTRL, _MIPIB_LP_GEN_CTRL)
-#define _MIPIA_HS_GEN_CTRL			(VLV_DISPLAY_BASE + 0xb070)
-#define _MIPIB_HS_GEN_CTRL			(VLV_DISPLAY_BASE + 0xb870)
-#define MIPI_HS_GEN_CTRL(pipe)		_PIPE(pipe, _MIPIA_HS_GEN_CTRL, _MIPIB_HS_GEN_CTRL)
+#define _MIPIA_HS_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb068)
+#define _MIPIB_HS_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb868)
+#define MIPI_HS_GEN_DATA(tc)		_TRANSCODER(tc, _MIPIA_HS_GEN_DATA, \
+					_MIPIB_HS_GEN_DATA)
+
+#define _MIPIA_LP_GEN_CTRL		(dev_priv->mipi_mmio_base + 0xb06c)
+#define _MIPIB_LP_GEN_CTRL		(dev_priv->mipi_mmio_base + 0xb86c)
+#define MIPI_LP_GEN_CTRL(tc)		_TRANSCODER(tc, _MIPIA_LP_GEN_CTRL, \
+					_MIPIB_LP_GEN_CTRL)
+#define _MIPIA_HS_GEN_CTRL		(dev_priv->mipi_mmio_base + 0xb070)
+#define _MIPIB_HS_GEN_CTRL		(dev_priv->mipi_mmio_base + 0xb870)
+#define MIPI_HS_GEN_CTRL(tc)		_TRANSCODER(tc, _MIPIA_HS_GEN_CTRL, \
+					_MIPIB_HS_GEN_CTRL)
 #define  LONG_PACKET_WORD_COUNT_SHIFT			8
 #define  LONG_PACKET_WORD_COUNT_MASK			(0xffff << 8)
 #define  SHORT_PACKET_PARAM_SHIFT			8
@@ -6280,9 +6516,10 @@ enum punit_power_well {
 #define  DATA_TYPE_MASK					(3f << 0)
 /* data type values, see include/video/mipi_display.h */
 
-#define _MIPIA_GEN_FIFO_STAT			(VLV_DISPLAY_BASE + 0xb074)
-#define _MIPIB_GEN_FIFO_STAT			(VLV_DISPLAY_BASE + 0xb874)
-#define MIPI_GEN_FIFO_STAT(pipe)	_PIPE(pipe, _MIPIA_GEN_FIFO_STAT, _MIPIB_GEN_FIFO_STAT)
+#define _MIPIA_GEN_FIFO_STAT		(dev_priv->mipi_mmio_base + 0xb074)
+#define _MIPIB_GEN_FIFO_STAT		(dev_priv->mipi_mmio_base + 0xb874)
+#define MIPI_GEN_FIFO_STAT(tc)	_TRANSCODER(tc, _MIPIA_GEN_FIFO_STAT, \
+					_MIPIB_GEN_FIFO_STAT)
 #define  DPI_FIFO_EMPTY					(1 << 28)
 #define  DBI_FIFO_EMPTY					(1 << 27)
 #define  LP_CTRL_FIFO_EMPTY				(1 << 26)
@@ -6298,16 +6535,18 @@ enum punit_power_well {
 #define  HS_DATA_FIFO_HALF_EMPTY			(1 << 1)
 #define  HS_DATA_FIFO_FULL				(1 << 0)
 
-#define _MIPIA_HS_LS_DBI_ENABLE			(VLV_DISPLAY_BASE + 0xb078)
-#define _MIPIB_HS_LS_DBI_ENABLE			(VLV_DISPLAY_BASE + 0xb878)
-#define MIPI_HS_LP_DBI_ENABLE(pipe)	_PIPE(pipe, _MIPIA_HS_LS_DBI_ENABLE, _MIPIB_HS_LS_DBI_ENABLE)
+#define _MIPIA_HS_LS_DBI_ENABLE		(dev_priv->mipi_mmio_base + 0xb078)
+#define _MIPIB_HS_LS_DBI_ENABLE		(dev_priv->mipi_mmio_base + 0xb878)
+#define MIPI_HS_LP_DBI_ENABLE(tc)	_TRANSCODER(tc, \
+			_MIPIA_HS_LS_DBI_ENABLE, _MIPIB_HS_LS_DBI_ENABLE)
 #define  DBI_HS_LP_MODE_MASK				(1 << 0)
 #define  DBI_LP_MODE					(1 << 0)
 #define  DBI_HS_MODE					(0 << 0)
 
-#define _MIPIA_DPHY_PARAM			(VLV_DISPLAY_BASE + 0xb080)
-#define _MIPIB_DPHY_PARAM			(VLV_DISPLAY_BASE + 0xb880)
-#define MIPI_DPHY_PARAM(pipe)		_PIPE(pipe, _MIPIA_DPHY_PARAM, _MIPIB_DPHY_PARAM)
+#define _MIPIA_DPHY_PARAM		(dev_priv->mipi_mmio_base + 0xb080)
+#define _MIPIB_DPHY_PARAM		(dev_priv->mipi_mmio_base + 0xb880)
+#define MIPI_DPHY_PARAM(tc)		_TRANSCODER(tc, _MIPIA_DPHY_PARAM, \
+					_MIPIB_DPHY_PARAM)
 #define  EXIT_ZERO_COUNT_SHIFT				24
 #define  EXIT_ZERO_COUNT_MASK				(0x3f << 24)
 #define  TRAIL_COUNT_SHIFT				16
@@ -6318,34 +6557,41 @@ enum punit_power_well {
 #define  PREPARE_COUNT_MASK				(0x3f << 0)
 
 /* bits 31:0 */
-#define _MIPIA_DBI_BW_CTRL			(VLV_DISPLAY_BASE + 0xb084)
-#define _MIPIB_DBI_BW_CTRL			(VLV_DISPLAY_BASE + 0xb884)
-#define MIPI_DBI_BW_CTRL(pipe)		_PIPE(pipe, _MIPIA_DBI_BW_CTRL, _MIPIB_DBI_BW_CTRL)
-
-#define _MIPIA_CLK_LANE_SWITCH_TIME_CNT		(VLV_DISPLAY_BASE + 0xb088)
-#define _MIPIB_CLK_LANE_SWITCH_TIME_CNT		(VLV_DISPLAY_BASE + 0xb888)
-#define MIPI_CLK_LANE_SWITCH_TIME_CNT(pipe)	_PIPE(pipe, _MIPIA_CLK_LANE_SWITCH_TIME_CNT, _MIPIB_CLK_LANE_SWITCH_TIME_CNT)
+#define _MIPIA_DBI_BW_CTRL		(dev_priv->mipi_mmio_base + 0xb084)
+#define _MIPIB_DBI_BW_CTRL		(dev_priv->mipi_mmio_base + 0xb884)
+#define MIPI_DBI_BW_CTRL(tc)		_TRANSCODER(tc, _MIPIA_DBI_BW_CTRL, \
+					_MIPIB_DBI_BW_CTRL)
+
+#define _MIPIA_CLK_LANE_SWITCH_TIME_CNT		(dev_priv->mipi_mmio_base \
+							+ 0xb088)
+#define _MIPIB_CLK_LANE_SWITCH_TIME_CNT		(dev_priv->mipi_mmio_base \
+							+ 0xb888)
+#define MIPI_CLK_LANE_SWITCH_TIME_CNT(tc)	_TRANSCODER(tc, \
+	_MIPIA_CLK_LANE_SWITCH_TIME_CNT, _MIPIB_CLK_LANE_SWITCH_TIME_CNT)
 #define  LP_HS_SSW_CNT_SHIFT				16
 #define  LP_HS_SSW_CNT_MASK				(0xffff << 16)
 #define  HS_LP_PWR_SW_CNT_SHIFT				0
 #define  HS_LP_PWR_SW_CNT_MASK				(0xffff << 0)
 
-#define _MIPIA_STOP_STATE_STALL			(VLV_DISPLAY_BASE + 0xb08c)
-#define _MIPIB_STOP_STATE_STALL			(VLV_DISPLAY_BASE + 0xb88c)
-#define MIPI_STOP_STATE_STALL(pipe)	_PIPE(pipe, _MIPIA_STOP_STATE_STALL, _MIPIB_STOP_STATE_STALL)
+#define _MIPIA_STOP_STATE_STALL		(dev_priv->mipi_mmio_base + 0xb08c)
+#define _MIPIB_STOP_STATE_STALL		(dev_priv->mipi_mmio_base + 0xb88c)
+#define MIPI_STOP_STATE_STALL(tc)	_TRANSCODER(tc, \
+			_MIPIA_STOP_STATE_STALL, _MIPIB_STOP_STATE_STALL)
 #define  STOP_STATE_STALL_COUNTER_SHIFT			0
 #define  STOP_STATE_STALL_COUNTER_MASK			(0xff << 0)
 
-#define _MIPIA_INTR_STAT_REG_1			(VLV_DISPLAY_BASE + 0xb090)
-#define _MIPIB_INTR_STAT_REG_1			(VLV_DISPLAY_BASE + 0xb890)
-#define MIPI_INTR_STAT_REG_1(pipe)	_PIPE(pipe, _MIPIA_INTR_STAT_REG_1, _MIPIB_INTR_STAT_REG_1)
-#define _MIPIA_INTR_EN_REG_1			(VLV_DISPLAY_BASE + 0xb094)
-#define _MIPIB_INTR_EN_REG_1			(VLV_DISPLAY_BASE + 0xb894)
-#define MIPI_INTR_EN_REG_1(pipe)	_PIPE(pipe, _MIPIA_INTR_EN_REG_1, _MIPIB_INTR_EN_REG_1)
+#define _MIPIA_INTR_STAT_REG_1		(dev_priv->mipi_mmio_base + 0xb090)
+#define _MIPIB_INTR_STAT_REG_1		(dev_priv->mipi_mmio_base + 0xb890)
+#define MIPI_INTR_STAT_REG_1(tc)	_TRANSCODER(tc, \
+				_MIPIA_INTR_STAT_REG_1, _MIPIB_INTR_STAT_REG_1)
+#define _MIPIA_INTR_EN_REG_1		(dev_priv->mipi_mmio_base + 0xb094)
+#define _MIPIB_INTR_EN_REG_1		(dev_priv->mipi_mmio_base + 0xb894)
+#define MIPI_INTR_EN_REG_1(tc)	_TRANSCODER(tc, _MIPIA_INTR_EN_REG_1, \
+					_MIPIB_INTR_EN_REG_1)
 #define  RX_CONTENTION_DETECTED				(1 << 0)
 
 /* XXX: only pipe A ?!? */
-#define MIPIA_DBI_TYPEC_CTRL			(VLV_DISPLAY_BASE + 0xb100)
+#define MIPIA_DBI_TYPEC_CTRL		(dev_priv->mipi_mmio_base + 0xb100)
 #define  DBI_TYPEC_ENABLE				(1 << 31)
 #define  DBI_TYPEC_WIP					(1 << 30)
 #define  DBI_TYPEC_OPTION_SHIFT				28
@@ -6359,9 +6605,10 @@ enum punit_power_well {
 
 /* MIPI adapter registers */
 
-#define _MIPIA_CTRL				(VLV_DISPLAY_BASE + 0xb104)
-#define _MIPIB_CTRL				(VLV_DISPLAY_BASE + 0xb904)
-#define MIPI_CTRL(pipe)			_PIPE(pipe, _MIPIA_CTRL, _MIPIB_CTRL)
+#define _MIPIA_CTRL			(dev_priv->mipi_mmio_base + 0xb104)
+#define _MIPIB_CTRL			(dev_priv->mipi_mmio_base + 0xb904)
+#define MIPI_CTRL(tc)			_TRANSCODER(tc, _MIPIA_CTRL, \
+					_MIPIB_CTRL)
 #define  ESCAPE_CLOCK_DIVIDER_SHIFT			5 /* A only */
 #define  ESCAPE_CLOCK_DIVIDER_MASK			(3 << 5)
 #define  ESCAPE_CLOCK_DIVIDER_1				(0 << 5)
@@ -6373,50 +6620,52 @@ enum punit_power_well {
 #define  READ_REQUEST_PRIORITY_HIGH			(3 << 3)
 #define  RGB_FLIP_TO_BGR				(1 << 2)
 
-#define _MIPIA_DATA_ADDRESS			(VLV_DISPLAY_BASE + 0xb108)
-#define _MIPIB_DATA_ADDRESS			(VLV_DISPLAY_BASE + 0xb908)
-#define MIPI_DATA_ADDRESS(pipe)		_PIPE(pipe, _MIPIA_DATA_ADDRESS, _MIPIB_DATA_ADDRESS)
+#define _MIPIA_DATA_ADDRESS		(dev_priv->mipi_mmio_base + 0xb108)
+#define _MIPIB_DATA_ADDRESS		(dev_priv->mipi_mmio_base + 0xb908)
+#define MIPI_DATA_ADDRESS(tc)		_TRANSCODER(tc, _MIPIA_DATA_ADDRESS, \
+					_MIPIB_DATA_ADDRESS)
 #define  DATA_MEM_ADDRESS_SHIFT				5
 #define  DATA_MEM_ADDRESS_MASK				(0x7ffffff << 5)
 #define  DATA_VALID					(1 << 0)
 
-#define _MIPIA_DATA_LENGTH			(VLV_DISPLAY_BASE + 0xb10c)
-#define _MIPIB_DATA_LENGTH			(VLV_DISPLAY_BASE + 0xb90c)
-#define MIPI_DATA_LENGTH(pipe)		_PIPE(pipe, _MIPIA_DATA_LENGTH, _MIPIB_DATA_LENGTH)
+#define _MIPIA_DATA_LENGTH		(dev_priv->mipi_mmio_base + 0xb10c)
+#define _MIPIB_DATA_LENGTH		(dev_priv->mipi_mmio_base + 0xb90c)
+#define MIPI_DATA_LENGTH(tc)		_TRANSCODER(tc, _MIPIA_DATA_LENGTH, \
+					_MIPIB_DATA_LENGTH)
 #define  DATA_LENGTH_SHIFT				0
 #define  DATA_LENGTH_MASK				(0xfffff << 0)
 
-#define _MIPIA_COMMAND_ADDRESS			(VLV_DISPLAY_BASE + 0xb110)
-#define _MIPIB_COMMAND_ADDRESS			(VLV_DISPLAY_BASE + 0xb910)
-#define MIPI_COMMAND_ADDRESS(pipe)	_PIPE(pipe, _MIPIA_COMMAND_ADDRESS, _MIPIB_COMMAND_ADDRESS)
+#define _MIPIA_COMMAND_ADDRESS		(dev_priv->mipi_mmio_base + 0xb110)
+#define _MIPIB_COMMAND_ADDRESS		(dev_priv->mipi_mmio_base + 0xb910)
+#define MIPI_COMMAND_ADDRESS(tc)	_TRANSCODER(tc, \
+				_MIPIA_COMMAND_ADDRESS, _MIPIB_COMMAND_ADDRESS)
 #define  COMMAND_MEM_ADDRESS_SHIFT			5
 #define  COMMAND_MEM_ADDRESS_MASK			(0x7ffffff << 5)
 #define  AUTO_PWG_ENABLE				(1 << 2)
 #define  MEMORY_WRITE_DATA_FROM_PIPE_RENDERING		(1 << 1)
 #define  COMMAND_VALID					(1 << 0)
 
-#define _MIPIA_COMMAND_LENGTH			(VLV_DISPLAY_BASE + 0xb114)
-#define _MIPIB_COMMAND_LENGTH			(VLV_DISPLAY_BASE + 0xb914)
-#define MIPI_COMMAND_LENGTH(pipe)	_PIPE(pipe, _MIPIA_COMMAND_LENGTH, _MIPIB_COMMAND_LENGTH)
+#define _MIPIA_COMMAND_LENGTH		(dev_priv->mipi_mmio_base + 0xb114)
+#define _MIPIB_COMMAND_LENGTH		(dev_priv->mipi_mmio_base + 0xb914)
+#define MIPI_COMMAND_LENGTH(tc)	_TRANSCODER(tc, _MIPIA_COMMAND_LENGTH, \
+					_MIPIB_COMMAND_LENGTH)
 #define  COMMAND_LENGTH_SHIFT(n)			(8 * (n)) /* n: 0...3 */
 #define  COMMAND_LENGTH_MASK(n)				(0xff << (8 * (n)))
 
-#define _MIPIA_READ_DATA_RETURN0		(VLV_DISPLAY_BASE + 0xb118)
-#define _MIPIB_READ_DATA_RETURN0		(VLV_DISPLAY_BASE + 0xb918)
-#define MIPI_READ_DATA_RETURN(pipe, n) \
-	(_PIPE(pipe, _MIPIA_READ_DATA_RETURN0, _MIPIB_READ_DATA_RETURN0) + 4 * (n)) /* n: 0...7 */
+#define _MIPIA_READ_DATA_RETURN0	(dev_priv->mipi_mmio_base + 0xb118)
+#define _MIPIB_READ_DATA_RETURN0	(dev_priv->mipi_mmio_base + 0xb918)
+#define MIPI_READ_DATA_RETURN(tc, n) \
+	(_TRANSCODER(tc, _MIPIA_READ_DATA_RETURN0, _MIPIB_READ_DATA_RETURN0) \
+					+ 4 * (n)) /* n: 0...7 */
 
-#define _MIPIA_READ_DATA_VALID			(VLV_DISPLAY_BASE + 0xb138)
-#define _MIPIB_READ_DATA_VALID			(VLV_DISPLAY_BASE + 0xb938)
-#define MIPI_READ_DATA_VALID(pipe)	_PIPE(pipe, _MIPIA_READ_DATA_VALID, _MIPIB_READ_DATA_VALID)
+#define _MIPIA_READ_DATA_VALID		(dev_priv->mipi_mmio_base + 0xb138)
+#define _MIPIB_READ_DATA_VALID		(dev_priv->mipi_mmio_base + 0xb938)
+#define MIPI_READ_DATA_VALID(tc)	_TRANSCODER(tc, \
+				_MIPIA_READ_DATA_VALID, _MIPIB_READ_DATA_VALID)
 #define  READ_DATA_VALID(n)				(1 << (n))
 
 /* For UMS only (deprecated): */
 #define _PALETTE_A (dev_priv->info.display_mmio_offset + 0xa000)
 #define _PALETTE_B (dev_priv->info.display_mmio_offset + 0xa800)
-#define _DPLL_A (dev_priv->info.display_mmio_offset + 0x6014)
-#define _DPLL_B (dev_priv->info.display_mmio_offset + 0x6018)
-#define _DPLL_A_MD (dev_priv->info.display_mmio_offset + 0x601c)
-#define _DPLL_B_MD (dev_priv->info.display_mmio_offset + 0x6020)
 
 #endif /* _I915_REG_H_ */
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index 86ce39aad0ff..503847f18fdd 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -47,22 +47,45 @@ static u32 calc_residency(struct drm_device *dev, const u32 reg)
 
 	intel_runtime_pm_get(dev_priv);
 
-	/* On VLV, residency time is in CZ units rather than 1.28us */
+	/* On VLV and CHV, residency time is in CZ units rather than 1.28us */
 	if (IS_VALLEYVIEW(dev)) {
-		u32 clkctl2;
+		u32 reg, czcount_30ns;
 
-		clkctl2 = I915_READ(VLV_CLK_CTL2) >>
-			CLK_CTL2_CZCOUNT_30NS_SHIFT;
-		if (!clkctl2) {
-			WARN(!clkctl2, "bogus CZ count value");
+		if (IS_CHERRYVIEW(dev))
+			reg = CHV_CLK_CTL1;
+		else
+			reg = VLV_CLK_CTL2;
+
+		czcount_30ns = I915_READ(reg) >> CLK_CTL2_CZCOUNT_30NS_SHIFT;
+
+		if (!czcount_30ns) {
+			WARN(!czcount_30ns, "bogus CZ count value");
 			ret = 0;
 			goto out;
 		}
-		units = DIV_ROUND_UP_ULL(30ULL * bias, (u64)clkctl2);
+
+		units = 0;
+		div = 1000000ULL;
+
+		if (IS_CHERRYVIEW(dev)) {
+			/* Special case for 320Mhz */
+			if (czcount_30ns == 1) {
+				div = 10000000ULL;
+				units = 3125ULL;
+			} else {
+				/* chv counts are one less */
+				czcount_30ns += 1;
+			}
+		}
+
+		if (units == 0)
+			units = DIV_ROUND_UP_ULL(30ULL * bias,
+						 (u64)czcount_30ns);
+
 		if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH)
 			units <<= 8;
 
-		div = 1000000ULL * bias;
+		div = div * bias;
 	}
 
 	raw_time = I915_READ(reg) * units;
@@ -461,11 +484,20 @@ static ssize_t gt_rp_mhz_show(struct device *kdev, struct device_attribute *attr
 	mutex_unlock(&dev->struct_mutex);
 
 	if (attr == &dev_attr_gt_RP0_freq_mhz) {
-		val = ((rp_state_cap & 0x0000ff) >> 0) * GT_FREQUENCY_MULTIPLIER;
+		if (IS_VALLEYVIEW(dev))
+			val = vlv_gpu_freq(dev_priv, dev_priv->rps.rp0_freq);
+		else
+			val = ((rp_state_cap & 0x0000ff) >> 0) * GT_FREQUENCY_MULTIPLIER;
 	} else if (attr == &dev_attr_gt_RP1_freq_mhz) {
-		val = ((rp_state_cap & 0x00ff00) >> 8) * GT_FREQUENCY_MULTIPLIER;
+		if (IS_VALLEYVIEW(dev))
+			val = vlv_gpu_freq(dev_priv, dev_priv->rps.rp1_freq);
+		else
+			val = ((rp_state_cap & 0x00ff00) >> 8) * GT_FREQUENCY_MULTIPLIER;
 	} else if (attr == &dev_attr_gt_RPn_freq_mhz) {
-		val = ((rp_state_cap & 0xff0000) >> 16) * GT_FREQUENCY_MULTIPLIER;
+		if (IS_VALLEYVIEW(dev))
+			val = vlv_gpu_freq(dev_priv, dev_priv->rps.min_freq);
+		else
+			val = ((rp_state_cap & 0xff0000) >> 16) * GT_FREQUENCY_MULTIPLIER;
 	} else {
 		BUG();
 	}
@@ -486,6 +518,9 @@ static const struct attribute *vlv_attrs[] = {
 	&dev_attr_gt_cur_freq_mhz.attr,
 	&dev_attr_gt_max_freq_mhz.attr,
 	&dev_attr_gt_min_freq_mhz.attr,
+	&dev_attr_gt_RP0_freq_mhz.attr,
+	&dev_attr_gt_RP1_freq_mhz.attr,
+	&dev_attr_gt_RPn_freq_mhz.attr,
 	&dev_attr_vlv_rpe_freq_mhz.attr,
 	NULL,
 };
@@ -505,7 +540,7 @@ static ssize_t error_state_read(struct file *filp, struct kobject *kobj,
 
 	memset(&error_priv, 0, sizeof(error_priv));
 
-	ret = i915_error_state_buf_init(&error_str, count, off);
+	ret = i915_error_state_buf_init(&error_str, to_i915(dev), count, off);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 21b347efcf1e..a4bd90f36a03 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -336,11 +336,12 @@ parse_lfp_backlight(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	dev_priv->vbt.backlight.pwm_freq_hz = entry->pwm_freq_hz;
 	dev_priv->vbt.backlight.active_low_pwm = entry->active_low_pwm;
+	dev_priv->vbt.backlight.min_brightness = entry->min_brightness;
 	DRM_DEBUG_KMS("VBT backlight PWM modulation frequency %u Hz, "
 		      "active %s, min brightness %u, level %u\n",
 		      dev_priv->vbt.backlight.pwm_freq_hz,
 		      dev_priv->vbt.backlight.active_low_pwm ? "low" : "high",
-		      entry->min_brightness,
+		      dev_priv->vbt.backlight.min_brightness,
 		      backlight_data->level[panel_type]);
 }
 
@@ -626,16 +627,16 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->preemphasis) {
 	case EDP_PREEMPHASIS_NONE:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_0;
 		break;
 	case EDP_PREEMPHASIS_3_5dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_1;
 		break;
 	case EDP_PREEMPHASIS_6dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_2;
 		break;
 	case EDP_PREEMPHASIS_9_5dB:
-		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_3;
 		break;
 	default:
 		DRM_DEBUG_KMS("VBT has unknown eDP pre-emphasis value %u\n",
@@ -645,16 +646,16 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	switch (edp_link_params->vswing) {
 	case EDP_VSWING_0_4V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 		break;
 	case EDP_VSWING_0_6V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
 		break;
 	case EDP_VSWING_0_8V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 		break;
 	case EDP_VSWING_1_2V:
-		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 		break;
 	default:
 		DRM_DEBUG_KMS("VBT has unknown eDP voltage swing value %u\n",
@@ -877,7 +878,7 @@ err:
 
 	/* error during parsing so set all pointers to null
 	 * because of partial parsing */
-	memset(dev_priv->vbt.dsi.sequence, 0, MIPI_SEQ_MAX);
+	memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence));
 }
 
 static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
@@ -975,12 +976,10 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
 	if (bdb->version >= 158) {
 		/* The VBT HDMI level shift values match the table we have. */
 		hdmi_level_shift = child->raw[7] & 0xF;
-		if (hdmi_level_shift < 0xC) {
-			DRM_DEBUG_KMS("VBT HDMI level shift for port %c: %d\n",
-				      port_name(port),
-				      hdmi_level_shift);
-			info->hdmi_level_shift = hdmi_level_shift;
-		}
+		DRM_DEBUG_KMS("VBT HDMI level shift for port %c: %d\n",
+			      port_name(port),
+			      hdmi_level_shift);
+		info->hdmi_level_shift = hdmi_level_shift;
 	}
 }
 
@@ -1113,8 +1112,7 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 		struct ddi_vbt_port_info *info =
 			&dev_priv->vbt.ddi_port_info[port];
 
-		/* Recommended BSpec default: 800mV 0dB. */
-		info->hdmi_level_shift = 6;
+		info->hdmi_level_shift = HDMI_LEVEL_SHIFT_UNKNOWN;
 
 		info->supports_dvi = (port != PORT_A && port != PORT_E);
 		info->supports_hdmi = info->supports_dvi;
@@ -1122,7 +1120,7 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 	}
 }
 
-static int __init intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
+static int intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
 {
 	DRM_DEBUG_KMS("Falling back to manually reading VBT from "
 		      "VBIOS ROM for %s\n",
diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
index b98667796337..905999bee2ac 100644
--- a/drivers/gpu/drm/i915/intel_bios.h
+++ b/drivers/gpu/drm/i915/intel_bios.h
@@ -802,7 +802,8 @@ struct mipi_config {
 
 	u16 rsvd4;
 
-	u8 rsvd5[5];
+	u8 rsvd5;
+	u32 target_burst_mode_freq;
 	u32 dsi_ddr_clk;
 	u32 bridge_ref_clk;
 
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 5a045d3bd77e..9212e6504e0f 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -137,6 +137,18 @@ static void hsw_crt_get_config(struct intel_encoder *encoder,
 	pipe_config->adjusted_mode.flags |= intel_crt_get_flags(encoder);
 }
 
+static void hsw_crt_pre_enable(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	WARN(I915_READ(SPLL_CTL) & SPLL_PLL_ENABLE, "SPLL already enabled\n");
+	I915_WRITE(SPLL_CTL,
+		   SPLL_PLL_ENABLE | SPLL_PLL_FREQ_1350MHz | SPLL_PLL_SSC);
+	POSTING_READ(SPLL_CTL);
+	udelay(20);
+}
+
 /* Note: The caller is required to filter out dpms modes not supported by the
  * platform. */
 static void intel_crt_set_dpms(struct intel_encoder *encoder, int mode)
@@ -194,6 +206,20 @@ static void intel_disable_crt(struct intel_encoder *encoder)
 	intel_crt_set_dpms(encoder, DRM_MODE_DPMS_OFF);
 }
 
+
+static void hsw_crt_post_disable(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t val;
+
+	DRM_DEBUG_KMS("Disabling SPLL\n");
+	val = I915_READ(SPLL_CTL);
+	WARN_ON(!(val & SPLL_PLL_ENABLE));
+	I915_WRITE(SPLL_CTL, val & ~SPLL_PLL_ENABLE);
+	POSTING_READ(SPLL_CTL);
+}
+
 static void intel_enable_crt(struct intel_encoder *encoder)
 {
 	struct intel_crt *crt = intel_encoder_to_crt(encoder);
@@ -289,8 +315,10 @@ static bool intel_crt_compute_config(struct intel_encoder *encoder,
 		pipe_config->pipe_bpp = 24;
 
 	/* FDI must always be 2.7 GHz */
-	if (HAS_DDI(dev))
+	if (HAS_DDI(dev)) {
+		pipe_config->ddi_pll_sel = PORT_CLK_SEL_SPLL;
 		pipe_config->port_clock = 135000 * 2;
+	}
 
 	return true;
 }
@@ -632,8 +660,6 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 	struct intel_load_detect_pipe tmp;
 	struct drm_modeset_acquire_ctx ctx;
 
-	intel_runtime_pm_get(dev_priv);
-
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] force=%d\n",
 		      connector->base.id, connector->name,
 		      force);
@@ -673,20 +699,23 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 		goto out;
 	}
 
+	drm_modeset_acquire_init(&ctx, 0);
+
 	/* for pre-945g platforms use load detect */
 	if (intel_get_load_detect_pipe(connector, NULL, &tmp, &ctx)) {
 		if (intel_crt_detect_ddc(connector))
 			status = connector_status_connected;
 		else
 			status = intel_crt_load_detect(crt);
-		intel_release_load_detect_pipe(connector, &tmp, &ctx);
+		intel_release_load_detect_pipe(connector, &tmp);
 	} else
 		status = connector_status_unknown;
 
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+
 out:
 	intel_display_power_put(dev_priv, power_domain);
-	intel_runtime_pm_put(dev_priv);
-
 	return status;
 }
 
@@ -775,7 +804,7 @@ static const struct drm_encoder_funcs intel_crt_enc_funcs = {
 	.destroy = intel_encoder_destroy,
 };
 
-static int __init intel_no_crt_dmi_callback(const struct dmi_system_id *id)
+static int intel_no_crt_dmi_callback(const struct dmi_system_id *id)
 {
 	DRM_INFO("Skipping CRT initialization for %s\n", id->ident);
 	return 1;
@@ -860,6 +889,8 @@ void intel_crt_init(struct drm_device *dev)
 	if (HAS_DDI(dev)) {
 		crt->base.get_config = hsw_crt_get_config;
 		crt->base.get_hw_state = intel_ddi_get_hw_state;
+		crt->base.pre_enable = hsw_crt_pre_enable;
+		crt->base.post_disable = hsw_crt_post_disable;
 	} else {
 		crt->base.get_config = intel_crt_get_config;
 		crt->base.get_hw_state = intel_crt_get_hw_state;
@@ -869,7 +900,7 @@ void intel_crt_init(struct drm_device *dev)
 
 	drm_connector_helper_add(connector, &intel_crt_connector_helper_funcs);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	if (!I915_HAS_HOTPLUG(dev))
 		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index b17b9c7c769f..b63d4fa204a3 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -28,87 +28,103 @@
 #include "i915_drv.h"
 #include "intel_drv.h"
 
+struct ddi_buf_trans {
+	u32 trans1;	/* balance leg enable, de-emph level */
+	u32 trans2;	/* vref sel, vswing */
+};
+
 /* HDMI/DVI modes ignore everything but the last 2 items. So we share
  * them for both DP and FDI transports, allowing those ports to
  * automatically adapt to HDMI connections as well
  */
-static const u32 hsw_ddi_translations_dp[] = {
-	0x00FFFFFF, 0x0006000E,		/* DP parameters */
-	0x00D75FFF, 0x0005000A,
-	0x00C30FFF, 0x00040006,
-	0x80AAAFFF, 0x000B0000,
-	0x00FFFFFF, 0x0005000A,
-	0x00D75FFF, 0x000C0004,
-	0x80C30FFF, 0x000B0000,
-	0x00FFFFFF, 0x00040006,
-	0x80D75FFF, 0x000B0000,
+static const struct ddi_buf_trans hsw_ddi_translations_dp[] = {
+	{ 0x00FFFFFF, 0x0006000E },
+	{ 0x00D75FFF, 0x0005000A },
+	{ 0x00C30FFF, 0x00040006 },
+	{ 0x80AAAFFF, 0x000B0000 },
+	{ 0x00FFFFFF, 0x0005000A },
+	{ 0x00D75FFF, 0x000C0004 },
+	{ 0x80C30FFF, 0x000B0000 },
+	{ 0x00FFFFFF, 0x00040006 },
+	{ 0x80D75FFF, 0x000B0000 },
+};
+
+static const struct ddi_buf_trans hsw_ddi_translations_fdi[] = {
+	{ 0x00FFFFFF, 0x0007000E },
+	{ 0x00D75FFF, 0x000F000A },
+	{ 0x00C30FFF, 0x00060006 },
+	{ 0x00AAAFFF, 0x001E0000 },
+	{ 0x00FFFFFF, 0x000F000A },
+	{ 0x00D75FFF, 0x00160004 },
+	{ 0x00C30FFF, 0x001E0000 },
+	{ 0x00FFFFFF, 0x00060006 },
+	{ 0x00D75FFF, 0x001E0000 },
 };
 
-static const u32 hsw_ddi_translations_fdi[] = {
-	0x00FFFFFF, 0x0007000E,		/* FDI parameters */
-	0x00D75FFF, 0x000F000A,
-	0x00C30FFF, 0x00060006,
-	0x00AAAFFF, 0x001E0000,
-	0x00FFFFFF, 0x000F000A,
-	0x00D75FFF, 0x00160004,
-	0x00C30FFF, 0x001E0000,
-	0x00FFFFFF, 0x00060006,
-	0x00D75FFF, 0x001E0000,
+static const struct ddi_buf_trans hsw_ddi_translations_hdmi[] = {
+					/* Idx	NT mV d	T mV d	db	*/
+	{ 0x00FFFFFF, 0x0006000E },	/* 0:	400	400	0	*/
+	{ 0x00E79FFF, 0x000E000C },	/* 1:	400	500	2	*/
+	{ 0x00D75FFF, 0x0005000A },	/* 2:	400	600	3.5	*/
+	{ 0x00FFFFFF, 0x0005000A },	/* 3:	600	600	0	*/
+	{ 0x00E79FFF, 0x001D0007 },	/* 4:	600	750	2	*/
+	{ 0x00D75FFF, 0x000C0004 },	/* 5:	600	900	3.5	*/
+	{ 0x00FFFFFF, 0x00040006 },	/* 6:	800	800	0	*/
+	{ 0x80E79FFF, 0x00030002 },	/* 7:	800	1000	2	*/
+	{ 0x00FFFFFF, 0x00140005 },	/* 8:	850	850	0	*/
+	{ 0x00FFFFFF, 0x000C0004 },	/* 9:	900	900	0	*/
+	{ 0x00FFFFFF, 0x001C0003 },	/* 10:	950	950	0	*/
+	{ 0x80FFFFFF, 0x00030002 },	/* 11:	1000	1000	0	*/
 };
 
-static const u32 hsw_ddi_translations_hdmi[] = {
-				/* Idx	NT mV diff	T mV diff	db  */
-	0x00FFFFFF, 0x0006000E, /* 0:	400		400		0   */
-	0x00E79FFF, 0x000E000C, /* 1:	400		500		2   */
-	0x00D75FFF, 0x0005000A, /* 2:	400		600		3.5 */
-	0x00FFFFFF, 0x0005000A, /* 3:	600		600		0   */
-	0x00E79FFF, 0x001D0007, /* 4:	600		750		2   */
-	0x00D75FFF, 0x000C0004, /* 5:	600		900		3.5 */
-	0x00FFFFFF, 0x00040006, /* 6:	800		800		0   */
-	0x80E79FFF, 0x00030002, /* 7:	800		1000		2   */
-	0x00FFFFFF, 0x00140005, /* 8:	850		850		0   */
-	0x00FFFFFF, 0x000C0004, /* 9:	900		900		0   */
-	0x00FFFFFF, 0x001C0003, /* 10:	950		950		0   */
-	0x80FFFFFF, 0x00030002, /* 11:	1000		1000		0   */
+static const struct ddi_buf_trans bdw_ddi_translations_edp[] = {
+	{ 0x00FFFFFF, 0x00000012 },
+	{ 0x00EBAFFF, 0x00020011 },
+	{ 0x00C71FFF, 0x0006000F },
+	{ 0x00AAAFFF, 0x000E000A },
+	{ 0x00FFFFFF, 0x00020011 },
+	{ 0x00DB6FFF, 0x0005000F },
+	{ 0x00BEEFFF, 0x000A000C },
+	{ 0x00FFFFFF, 0x0005000F },
+	{ 0x00DB6FFF, 0x000A000C },
 };
 
-static const u32 bdw_ddi_translations_edp[] = {
-	0x00FFFFFF, 0x00000012,		/* eDP parameters */
-	0x00EBAFFF, 0x00020011,
-	0x00C71FFF, 0x0006000F,
-	0x00FFFFFF, 0x00020011,
-	0x00DB6FFF, 0x0005000F,
-	0x00BEEFFF, 0x000A000C,
-	0x00FFFFFF, 0x0005000F,
-	0x00DB6FFF, 0x000A000C,
-	0x00FFFFFF, 0x000A000C,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans bdw_ddi_translations_dp[] = {
+	{ 0x00FFFFFF, 0x0007000E },
+	{ 0x00D75FFF, 0x000E000A },
+	{ 0x00BEFFFF, 0x00140006 },
+	{ 0x80B2CFFF, 0x001B0002 },
+	{ 0x00FFFFFF, 0x000E000A },
+	{ 0x00D75FFF, 0x00180004 },
+	{ 0x80CB2FFF, 0x001B0002 },
+	{ 0x00F7DFFF, 0x00180004 },
+	{ 0x80D75FFF, 0x001B0002 },
 };
 
-static const u32 bdw_ddi_translations_dp[] = {
-	0x00FFFFFF, 0x0007000E,		/* DP parameters */
-	0x00D75FFF, 0x000E000A,
-	0x00BEFFFF, 0x00140006,
-	0x00FFFFFF, 0x000E000A,
-	0x00D75FFF, 0x00180004,
-	0x80CB2FFF, 0x001B0002,
-	0x00F7DFFF, 0x00180004,
-	0x80D75FFF, 0x001B0002,
-	0x80FFFFFF, 0x001B0002,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans bdw_ddi_translations_fdi[] = {
+	{ 0x00FFFFFF, 0x0001000E },
+	{ 0x00D75FFF, 0x0004000A },
+	{ 0x00C30FFF, 0x00070006 },
+	{ 0x00AAAFFF, 0x000C0000 },
+	{ 0x00FFFFFF, 0x0004000A },
+	{ 0x00D75FFF, 0x00090004 },
+	{ 0x00C30FFF, 0x000C0000 },
+	{ 0x00FFFFFF, 0x00070006 },
+	{ 0x00D75FFF, 0x000C0000 },
 };
 
-static const u32 bdw_ddi_translations_fdi[] = {
-	0x00FFFFFF, 0x0001000E,		/* FDI parameters */
-	0x00D75FFF, 0x0004000A,
-	0x00C30FFF, 0x00070006,
-	0x00AAAFFF, 0x000C0000,
-	0x00FFFFFF, 0x0004000A,
-	0x00D75FFF, 0x00090004,
-	0x00C30FFF, 0x000C0000,
-	0x00FFFFFF, 0x00070006,
-	0x00D75FFF, 0x000C0000,
-	0x00FFFFFF, 0x00140006		/* HDMI parameters 800mV 0dB*/
+static const struct ddi_buf_trans bdw_ddi_translations_hdmi[] = {
+					/* Idx	NT mV d	T mV df	db	*/
+	{ 0x00FFFFFF, 0x0007000E },	/* 0:	400	400	0	*/
+	{ 0x00D75FFF, 0x000E000A },	/* 1:	400	600	3.5	*/
+	{ 0x00BEFFFF, 0x00140006 },	/* 2:	400	800	6	*/
+	{ 0x00FFFFFF, 0x0009000D },	/* 3:	450	450	0	*/
+	{ 0x00FFFFFF, 0x000E000A },	/* 4:	600	600	0	*/
+	{ 0x00D7FFFF, 0x00140006 },	/* 5:	600	800	2.5	*/
+	{ 0x80CB2FFF, 0x001B0002 },	/* 6:	600	1000	4.5	*/
+	{ 0x00FFFFFF, 0x00140006 },	/* 7:	800	800	0	*/
+	{ 0x80E79FFF, 0x001B0002 },	/* 8:	800	1000	2	*/
+	{ 0x80FFFFFF, 0x001B0002 },	/* 9:	1000	1000	0	*/
 };
 
 enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder)
@@ -116,7 +132,10 @@ enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder)
 	struct drm_encoder *encoder = &intel_encoder->base;
 	int type = intel_encoder->type;
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP ||
+	if (type == INTEL_OUTPUT_DP_MST) {
+		struct intel_digital_port *intel_dig_port = enc_to_mst(encoder)->primary;
+		return intel_dig_port->port;
+	} else if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP ||
 	    type == INTEL_OUTPUT_HDMI || type == INTEL_OUTPUT_UNKNOWN) {
 		struct intel_digital_port *intel_dig_port =
 			enc_to_dig_port(encoder);
@@ -142,26 +161,36 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 reg;
-	int i;
+	int i, n_hdmi_entries, hdmi_800mV_0dB;
 	int hdmi_level = dev_priv->vbt.ddi_port_info[port].hdmi_level_shift;
-	const u32 *ddi_translations_fdi;
-	const u32 *ddi_translations_dp;
-	const u32 *ddi_translations_edp;
-	const u32 *ddi_translations;
+	const struct ddi_buf_trans *ddi_translations_fdi;
+	const struct ddi_buf_trans *ddi_translations_dp;
+	const struct ddi_buf_trans *ddi_translations_edp;
+	const struct ddi_buf_trans *ddi_translations_hdmi;
+	const struct ddi_buf_trans *ddi_translations;
 
 	if (IS_BROADWELL(dev)) {
 		ddi_translations_fdi = bdw_ddi_translations_fdi;
 		ddi_translations_dp = bdw_ddi_translations_dp;
 		ddi_translations_edp = bdw_ddi_translations_edp;
+		ddi_translations_hdmi = bdw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(bdw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 7;
 	} else if (IS_HASWELL(dev)) {
 		ddi_translations_fdi = hsw_ddi_translations_fdi;
 		ddi_translations_dp = hsw_ddi_translations_dp;
 		ddi_translations_edp = hsw_ddi_translations_dp;
+		ddi_translations_hdmi = hsw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(hsw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 6;
 	} else {
 		WARN(1, "ddi translation table missing\n");
 		ddi_translations_edp = bdw_ddi_translations_dp;
 		ddi_translations_fdi = bdw_ddi_translations_fdi;
 		ddi_translations_dp = bdw_ddi_translations_dp;
+		ddi_translations_hdmi = bdw_ddi_translations_hdmi;
+		n_hdmi_entries = ARRAY_SIZE(bdw_ddi_translations_hdmi);
+		hdmi_800mV_0dB = 7;
 	}
 
 	switch (port) {
@@ -187,14 +216,22 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port)
 
 	for (i = 0, reg = DDI_BUF_TRANS(port);
 	     i < ARRAY_SIZE(hsw_ddi_translations_fdi); i++) {
-		I915_WRITE(reg, ddi_translations[i]);
+		I915_WRITE(reg, ddi_translations[i].trans1);
 		reg += 4;
-	}
-	/* Entry 9 is for HDMI: */
-	for (i = 0; i < 2; i++) {
-		I915_WRITE(reg, hsw_ddi_translations_hdmi[hdmi_level * 2 + i]);
+		I915_WRITE(reg, ddi_translations[i].trans2);
 		reg += 4;
 	}
+
+	/* Choose a good default if VBT is badly populated */
+	if (hdmi_level == HDMI_LEVEL_SHIFT_UNKNOWN ||
+	    hdmi_level >= n_hdmi_entries)
+		hdmi_level = hdmi_800mV_0dB;
+
+	/* Entry 9 is for HDMI: */
+	I915_WRITE(reg, ddi_translations_hdmi[hdmi_level].trans1);
+	reg += 4;
+	I915_WRITE(reg, ddi_translations_hdmi[hdmi_level].trans2);
+	reg += 4;
 }
 
 /* Program DDI buffers translations for DP. By default, program ports A-D in DP
@@ -211,18 +248,6 @@ void intel_prepare_ddi(struct drm_device *dev)
 		intel_prepare_ddi_buffers(dev, port);
 }
 
-static const long hsw_ddi_buf_ctl_values[] = {
-	DDI_BUF_EMP_400MV_0DB_HSW,
-	DDI_BUF_EMP_400MV_3_5DB_HSW,
-	DDI_BUF_EMP_400MV_6DB_HSW,
-	DDI_BUF_EMP_400MV_9_5DB_HSW,
-	DDI_BUF_EMP_600MV_0DB_HSW,
-	DDI_BUF_EMP_600MV_3_5DB_HSW,
-	DDI_BUF_EMP_600MV_6DB_HSW,
-	DDI_BUF_EMP_800MV_0DB_HSW,
-	DDI_BUF_EMP_800MV_3_5DB_HSW
-};
-
 static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv,
 				    enum port port)
 {
@@ -277,11 +302,12 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	I915_WRITE(_FDI_RXA_CTL, rx_ctl_val);
 
 	/* Configure Port Clock Select */
-	I915_WRITE(PORT_CLK_SEL(PORT_E), intel_crtc->ddi_pll_sel);
+	I915_WRITE(PORT_CLK_SEL(PORT_E), intel_crtc->config.ddi_pll_sel);
+	WARN_ON(intel_crtc->config.ddi_pll_sel != PORT_CLK_SEL_SPLL);
 
 	/* Start the training iterating through available voltages and emphasis,
 	 * testing each value twice. */
-	for (i = 0; i < ARRAY_SIZE(hsw_ddi_buf_ctl_values) * 2; i++) {
+	for (i = 0; i < ARRAY_SIZE(hsw_ddi_translations_fdi) * 2; i++) {
 		/* Configure DP_TP_CTL with auto-training */
 		I915_WRITE(DP_TP_CTL(PORT_E),
 					DP_TP_CTL_FDI_AUTOTRAIN |
@@ -296,7 +322,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
 			   ((intel_crtc->config.fdi_lanes - 1) << 1) |
-			   hsw_ddi_buf_ctl_values[i / 2]);
+			   DDI_BUF_TRANS_SELECT(i / 2));
 		POSTING_READ(DDI_BUF_CTL(PORT_E));
 
 		udelay(600);
@@ -364,6 +390,18 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	DRM_ERROR("FDI link training failed!\n");
 }
 
+void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	struct intel_digital_port *intel_dig_port =
+		enc_to_dig_port(&encoder->base);
+
+	intel_dp->DP = intel_dig_port->saved_port_bits |
+		DDI_BUF_CTL_ENABLE | DDI_BUF_TRANS_SELECT(0);
+	intel_dp->DP |= DDI_PORT_WIDTH(intel_dp->lane_count);
+
+}
+
 static struct intel_encoder *
 intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 {
@@ -385,55 +423,8 @@ intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 	return ret;
 }
 
-void intel_ddi_put_crtc_pll(struct drm_crtc *crtc)
-{
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
-	struct intel_ddi_plls *plls = &dev_priv->ddi_plls;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t val;
-
-	switch (intel_crtc->ddi_pll_sel) {
-	case PORT_CLK_SEL_SPLL:
-		plls->spll_refcount--;
-		if (plls->spll_refcount == 0) {
-			DRM_DEBUG_KMS("Disabling SPLL\n");
-			val = I915_READ(SPLL_CTL);
-			WARN_ON(!(val & SPLL_PLL_ENABLE));
-			I915_WRITE(SPLL_CTL, val & ~SPLL_PLL_ENABLE);
-			POSTING_READ(SPLL_CTL);
-		}
-		break;
-	case PORT_CLK_SEL_WRPLL1:
-		plls->wrpll1_refcount--;
-		if (plls->wrpll1_refcount == 0) {
-			DRM_DEBUG_KMS("Disabling WRPLL 1\n");
-			val = I915_READ(WRPLL_CTL1);
-			WARN_ON(!(val & WRPLL_PLL_ENABLE));
-			I915_WRITE(WRPLL_CTL1, val & ~WRPLL_PLL_ENABLE);
-			POSTING_READ(WRPLL_CTL1);
-		}
-		break;
-	case PORT_CLK_SEL_WRPLL2:
-		plls->wrpll2_refcount--;
-		if (plls->wrpll2_refcount == 0) {
-			DRM_DEBUG_KMS("Disabling WRPLL 2\n");
-			val = I915_READ(WRPLL_CTL2);
-			WARN_ON(!(val & WRPLL_PLL_ENABLE));
-			I915_WRITE(WRPLL_CTL2, val & ~WRPLL_PLL_ENABLE);
-			POSTING_READ(WRPLL_CTL2);
-		}
-		break;
-	}
-
-	WARN(plls->spll_refcount < 0, "Invalid SPLL refcount\n");
-	WARN(plls->wrpll1_refcount < 0, "Invalid WRPLL1 refcount\n");
-	WARN(plls->wrpll2_refcount < 0, "Invalid WRPLL2 refcount\n");
-
-	intel_crtc->ddi_pll_sel = PORT_CLK_SEL_NONE;
-}
-
 #define LC_FREQ 2700
-#define LC_FREQ_2K (LC_FREQ * 2000)
+#define LC_FREQ_2K U64_C(LC_FREQ * 2000)
 
 #define P_MIN 2
 #define P_MAX 64
@@ -445,7 +436,11 @@ void intel_ddi_put_crtc_pll(struct drm_crtc *crtc)
 #define VCO_MIN 2400
 #define VCO_MAX 4800
 
-#define ABS_DIFF(a, b) ((a > b) ? (a - b) : (b - a))
+#define abs_diff(a, b) ({			\
+	typeof(a) __a = (a);			\
+	typeof(b) __b = (b);			\
+	(void) (&__a == &__b);			\
+	__a > __b ? (__a - __b) : (__b - __a); })
 
 struct wrpll_rnp {
 	unsigned p, n2, r2;
@@ -555,9 +550,9 @@ static void wrpll_update_rnp(uint64_t freq2k, unsigned budget,
 	 */
 	a = freq2k * budget * p * r2;
 	b = freq2k * budget * best->p * best->r2;
-	diff = ABS_DIFF((freq2k * p * r2), (LC_FREQ_2K * n2));
-	diff_best = ABS_DIFF((freq2k * best->p * best->r2),
-			     (LC_FREQ_2K * best->n2));
+	diff = abs_diff(freq2k * p * r2, LC_FREQ_2K * n2);
+	diff_best = abs_diff(freq2k * best->p * best->r2,
+			     LC_FREQ_2K * best->n2);
 	c = 1000000 * diff;
 	d = 1000000 * diff_best;
 
@@ -592,9 +587,9 @@ static int intel_ddi_calc_wrpll_link(struct drm_i915_private *dev_priv,
 	u32 wrpll;
 
 	wrpll = I915_READ(reg);
-	switch (wrpll & SPLL_PLL_REF_MASK) {
-	case SPLL_PLL_SSC:
-	case SPLL_PLL_NON_SSC:
+	switch (wrpll & WRPLL_PLL_REF_MASK) {
+	case WRPLL_PLL_SSC:
+	case WRPLL_PLL_NON_SSC:
 		/*
 		 * We could calculate spread here, but our checking
 		 * code only cares about 5% accuracy, and spread is a max of
@@ -602,7 +597,7 @@ static int intel_ddi_calc_wrpll_link(struct drm_i915_private *dev_priv,
 		 */
 		refclk = 135;
 		break;
-	case SPLL_PLL_LCPLL:
+	case WRPLL_PLL_LCPLL:
 		refclk = LC_FREQ;
 		break;
 	default:
@@ -618,15 +613,14 @@ static int intel_ddi_calc_wrpll_link(struct drm_i915_private *dev_priv,
 	return (refclk * n * 100) / (p * r);
 }
 
-static void intel_ddi_clock_get(struct intel_encoder *encoder,
-				struct intel_crtc_config *pipe_config)
+static void hsw_ddi_clock_get(struct intel_encoder *encoder,
+			      struct intel_crtc_config *pipe_config)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
-	enum port port = intel_ddi_get_encoder_port(encoder);
 	int link_clock = 0;
 	u32 val, pll;
 
-	val = I915_READ(PORT_CLK_SEL(port));
+	val = pipe_config->ddi_pll_sel;
 	switch (val & PORT_CLK_SEL_MASK) {
 	case PORT_CLK_SEL_LCPLL_810:
 		link_clock = 81000;
@@ -675,9 +669,15 @@ static void intel_ddi_clock_get(struct intel_encoder *encoder,
 		pipe_config->adjusted_mode.crtc_clock = pipe_config->port_clock;
 }
 
+void intel_ddi_clock_get(struct intel_encoder *encoder,
+			 struct intel_crtc_config *pipe_config)
+{
+	hsw_ddi_clock_get(encoder, pipe_config);
+}
+
 static void
-intel_ddi_calculate_wrpll(int clock /* in Hz */,
-			  unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
+hsw_ddi_calculate_wrpll(int clock /* in Hz */,
+			unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
 {
 	uint64_t freq2k;
 	unsigned p, n2, r2;
@@ -740,181 +740,54 @@ intel_ddi_calculate_wrpll(int clock /* in Hz */,
 	*r2_out = best.r2;
 }
 
-/*
- * Tries to find a PLL for the CRTC. If it finds, it increases the refcount and
- * stores it in intel_crtc->ddi_pll_sel, so other mode sets won't be able to
- * steal the selected PLL. You need to call intel_ddi_pll_enable to actually
- * enable the PLL.
- */
-bool intel_ddi_pll_select(struct intel_crtc *intel_crtc)
+static bool
+hsw_ddi_pll_select(struct intel_crtc *intel_crtc,
+		   struct intel_encoder *intel_encoder,
+		   int clock)
 {
-	struct drm_crtc *crtc = &intel_crtc->base;
-	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
-	struct drm_encoder *encoder = &intel_encoder->base;
-	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
-	struct intel_ddi_plls *plls = &dev_priv->ddi_plls;
-	int type = intel_encoder->type;
-	enum pipe pipe = intel_crtc->pipe;
-	int clock = intel_crtc->config.port_clock;
-
-	intel_ddi_put_crtc_pll(crtc);
-
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
-		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-
-		switch (intel_dp->link_bw) {
-		case DP_LINK_BW_1_62:
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_LCPLL_810;
-			break;
-		case DP_LINK_BW_2_7:
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_LCPLL_1350;
-			break;
-		case DP_LINK_BW_5_4:
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_LCPLL_2700;
-			break;
-		default:
-			DRM_ERROR("Link bandwidth %d unsupported\n",
-				  intel_dp->link_bw);
-			return false;
-		}
-
-	} else if (type == INTEL_OUTPUT_HDMI) {
-		uint32_t reg, val;
+	if (intel_encoder->type == INTEL_OUTPUT_HDMI) {
+		struct intel_shared_dpll *pll;
+		uint32_t val;
 		unsigned p, n2, r2;
 
-		intel_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
+		hsw_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
 
-		val = WRPLL_PLL_ENABLE | WRPLL_PLL_SELECT_LCPLL_2700 |
+		val = WRPLL_PLL_ENABLE | WRPLL_PLL_LCPLL |
 		      WRPLL_DIVIDER_REFERENCE(r2) | WRPLL_DIVIDER_FEEDBACK(n2) |
 		      WRPLL_DIVIDER_POST(p);
 
-		if (val == I915_READ(WRPLL_CTL1)) {
-			DRM_DEBUG_KMS("Reusing WRPLL 1 on pipe %c\n",
-				      pipe_name(pipe));
-			reg = WRPLL_CTL1;
-		} else if (val == I915_READ(WRPLL_CTL2)) {
-			DRM_DEBUG_KMS("Reusing WRPLL 2 on pipe %c\n",
-				      pipe_name(pipe));
-			reg = WRPLL_CTL2;
-		} else if (plls->wrpll1_refcount == 0) {
-			DRM_DEBUG_KMS("Using WRPLL 1 on pipe %c\n",
-				      pipe_name(pipe));
-			reg = WRPLL_CTL1;
-		} else if (plls->wrpll2_refcount == 0) {
-			DRM_DEBUG_KMS("Using WRPLL 2 on pipe %c\n",
-				      pipe_name(pipe));
-			reg = WRPLL_CTL2;
-		} else {
-			DRM_ERROR("No WRPLLs available!\n");
-			return false;
-		}
-
-		DRM_DEBUG_KMS("WRPLL: %dKHz refresh rate with p=%d, n2=%d r2=%d\n",
-			      clock, p, n2, r2);
+		intel_crtc->config.dpll_hw_state.wrpll = val;
 
-		if (reg == WRPLL_CTL1) {
-			plls->wrpll1_refcount++;
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_WRPLL1;
-		} else {
-			plls->wrpll2_refcount++;
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_WRPLL2;
-		}
-
-	} else if (type == INTEL_OUTPUT_ANALOG) {
-		if (plls->spll_refcount == 0) {
-			DRM_DEBUG_KMS("Using SPLL on pipe %c\n",
-				      pipe_name(pipe));
-			plls->spll_refcount++;
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_SPLL;
-		} else {
-			DRM_ERROR("SPLL already in use\n");
+		pll = intel_get_shared_dpll(intel_crtc);
+		if (pll == NULL) {
+			DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
+					 pipe_name(intel_crtc->pipe));
 			return false;
 		}
 
-	} else {
-		WARN(1, "Invalid DDI encoder type %d\n", type);
-		return false;
+		intel_crtc->config.ddi_pll_sel = PORT_CLK_SEL_WRPLL(pll->id);
 	}
 
 	return true;
 }
 
+
 /*
- * To be called after intel_ddi_pll_select(). That one selects the PLL to be
- * used, this one actually enables the PLL.
+ * Tries to find a *shared* PLL for the CRTC and store it in
+ * intel_crtc->ddi_pll_sel.
+ *
+ * For private DPLLs, compute_config() should do the selection for us. This
+ * function should be folded into compute_config() eventually.
  */
-void intel_ddi_pll_enable(struct intel_crtc *crtc)
+bool intel_ddi_pll_select(struct intel_crtc *intel_crtc)
 {
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_ddi_plls *plls = &dev_priv->ddi_plls;
-	int clock = crtc->config.port_clock;
-	uint32_t reg, cur_val, new_val;
-	int refcount;
-	const char *pll_name;
-	uint32_t enable_bit = (1 << 31);
-	unsigned int p, n2, r2;
-
-	BUILD_BUG_ON(enable_bit != SPLL_PLL_ENABLE);
-	BUILD_BUG_ON(enable_bit != WRPLL_PLL_ENABLE);
-
-	switch (crtc->ddi_pll_sel) {
-	case PORT_CLK_SEL_LCPLL_2700:
-	case PORT_CLK_SEL_LCPLL_1350:
-	case PORT_CLK_SEL_LCPLL_810:
-		/*
-		 * LCPLL should always be enabled at this point of the mode set
-		 * sequence, so nothing to do.
-		 */
-		return;
-
-	case PORT_CLK_SEL_SPLL:
-		pll_name = "SPLL";
-		reg = SPLL_CTL;
-		refcount = plls->spll_refcount;
-		new_val = SPLL_PLL_ENABLE | SPLL_PLL_FREQ_1350MHz |
-			  SPLL_PLL_SSC;
-		break;
-
-	case PORT_CLK_SEL_WRPLL1:
-	case PORT_CLK_SEL_WRPLL2:
-		if (crtc->ddi_pll_sel == PORT_CLK_SEL_WRPLL1) {
-			pll_name = "WRPLL1";
-			reg = WRPLL_CTL1;
-			refcount = plls->wrpll1_refcount;
-		} else {
-			pll_name = "WRPLL2";
-			reg = WRPLL_CTL2;
-			refcount = plls->wrpll2_refcount;
-		}
-
-		intel_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
-
-		new_val = WRPLL_PLL_ENABLE | WRPLL_PLL_SELECT_LCPLL_2700 |
-			  WRPLL_DIVIDER_REFERENCE(r2) |
-			  WRPLL_DIVIDER_FEEDBACK(n2) | WRPLL_DIVIDER_POST(p);
-
-		break;
-
-	case PORT_CLK_SEL_NONE:
-		WARN(1, "Bad selected pll: PORT_CLK_SEL_NONE\n");
-		return;
-	default:
-		WARN(1, "Bad selected pll: 0x%08x\n", crtc->ddi_pll_sel);
-		return;
-	}
+	struct drm_crtc *crtc = &intel_crtc->base;
+	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
+	int clock = intel_crtc->config.port_clock;
 
-	cur_val = I915_READ(reg);
+	intel_put_shared_dpll(intel_crtc);
 
-	WARN(refcount < 1, "Bad %s refcount: %d\n", pll_name, refcount);
-	if (refcount == 1) {
-		WARN(cur_val & enable_bit, "%s already enabled\n", pll_name);
-		I915_WRITE(reg, new_val);
-		POSTING_READ(reg);
-		udelay(20);
-	} else {
-		WARN((cur_val & enable_bit) == 0, "%s disabled\n", pll_name);
-	}
+	return hsw_ddi_pll_select(intel_crtc, intel_encoder, clock);
 }
 
 void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
@@ -926,8 +799,7 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 	int type = intel_encoder->type;
 	uint32_t temp;
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
-
+	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP || type == INTEL_OUTPUT_DP_MST) {
 		temp = TRANS_MSA_SYNC_CLK;
 		switch (intel_crtc->config.pipe_bpp) {
 		case 18:
@@ -949,6 +821,21 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 	}
 }
 
+void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
+	uint32_t temp;
+	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (state == true)
+		temp |= TRANS_DDI_DP_VC_PAYLOAD_ALLOC;
+	else
+		temp &= ~TRANS_DDI_DP_VC_PAYLOAD_ALLOC;
+	I915_WRITE(TRANS_DDI_FUNC_CTL(cpu_transcoder), temp);
+}
+
 void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -995,7 +882,9 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 			 * eDP when not using the panel fitter, and when not
 			 * using motion blur mitigation (which we don't
 			 * support). */
-			if (IS_HASWELL(dev) && intel_crtc->config.pch_pfit.enabled)
+			if (IS_HASWELL(dev) &&
+			    (intel_crtc->config.pch_pfit.enabled ||
+			     intel_crtc->config.pch_pfit.force_thru))
 				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
 			else
 				temp |= TRANS_DDI_EDP_INPUT_A_ON;
@@ -1026,7 +915,19 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 		   type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
-		temp |= TRANS_DDI_MODE_SELECT_DP_SST;
+		if (intel_dp->is_mst) {
+			temp |= TRANS_DDI_MODE_SELECT_DP_MST;
+		} else
+			temp |= TRANS_DDI_MODE_SELECT_DP_SST;
+
+		temp |= DDI_PORT_WIDTH(intel_dp->lane_count);
+	} else if (type == INTEL_OUTPUT_DP_MST) {
+		struct intel_dp *intel_dp = &enc_to_mst(encoder)->primary->dp;
+
+		if (intel_dp->is_mst) {
+			temp |= TRANS_DDI_MODE_SELECT_DP_MST;
+		} else
+			temp |= TRANS_DDI_MODE_SELECT_DP_SST;
 
 		temp |= DDI_PORT_WIDTH(intel_dp->lane_count);
 	} else {
@@ -1043,7 +944,7 @@ void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 	uint32_t reg = TRANS_DDI_FUNC_CTL(cpu_transcoder);
 	uint32_t val = I915_READ(reg);
 
-	val &= ~(TRANS_DDI_FUNC_ENABLE | TRANS_DDI_PORT_MASK);
+	val &= ~(TRANS_DDI_FUNC_ENABLE | TRANS_DDI_PORT_MASK | TRANS_DDI_DP_VC_PAYLOAD_ALLOC);
 	val |= TRANS_DDI_PORT_NONE;
 	I915_WRITE(reg, val);
 }
@@ -1082,8 +983,11 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 	case TRANS_DDI_MODE_SELECT_DP_SST:
 		if (type == DRM_MODE_CONNECTOR_eDP)
 			return true;
-	case TRANS_DDI_MODE_SELECT_DP_MST:
 		return (type == DRM_MODE_CONNECTOR_DisplayPort);
+	case TRANS_DDI_MODE_SELECT_DP_MST:
+		/* if the transcoder is in MST state then
+		 * connector isn't connected */
+		return false;
 
 	case TRANS_DDI_MODE_SELECT_FDI:
 		return (type == DRM_MODE_CONNECTOR_VGA);
@@ -1135,6 +1039,9 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 
 			if ((tmp & TRANS_DDI_PORT_MASK)
 			    == TRANS_DDI_SELECT_PORT(port)) {
+				if ((tmp & TRANS_DDI_MODE_SELECT_MASK) == TRANS_DDI_MODE_SELECT_DP_MST)
+					return false;
+
 				*pipe = i;
 				return true;
 			}
@@ -1146,76 +1053,6 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 	return false;
 }
 
-static uint32_t intel_ddi_get_crtc_pll(struct drm_i915_private *dev_priv,
-				       enum pipe pipe)
-{
-	uint32_t temp, ret;
-	enum port port = I915_MAX_PORTS;
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
-	int i;
-
-	if (cpu_transcoder == TRANSCODER_EDP) {
-		port = PORT_A;
-	} else {
-		temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
-		temp &= TRANS_DDI_PORT_MASK;
-
-		for (i = PORT_B; i <= PORT_E; i++)
-			if (temp == TRANS_DDI_SELECT_PORT(i))
-				port = i;
-	}
-
-	if (port == I915_MAX_PORTS) {
-		WARN(1, "Pipe %c enabled on an unknown port\n",
-		     pipe_name(pipe));
-		ret = PORT_CLK_SEL_NONE;
-	} else {
-		ret = I915_READ(PORT_CLK_SEL(port));
-		DRM_DEBUG_KMS("Pipe %c connected to port %c using clock "
-			      "0x%08x\n", pipe_name(pipe), port_name(port),
-			      ret);
-	}
-
-	return ret;
-}
-
-void intel_ddi_setup_hw_pll_state(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum pipe pipe;
-	struct intel_crtc *intel_crtc;
-
-	dev_priv->ddi_plls.spll_refcount = 0;
-	dev_priv->ddi_plls.wrpll1_refcount = 0;
-	dev_priv->ddi_plls.wrpll2_refcount = 0;
-
-	for_each_pipe(pipe) {
-		intel_crtc =
-			to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-
-		if (!intel_crtc->active) {
-			intel_crtc->ddi_pll_sel = PORT_CLK_SEL_NONE;
-			continue;
-		}
-
-		intel_crtc->ddi_pll_sel = intel_ddi_get_crtc_pll(dev_priv,
-								 pipe);
-
-		switch (intel_crtc->ddi_pll_sel) {
-		case PORT_CLK_SEL_SPLL:
-			dev_priv->ddi_plls.spll_refcount++;
-			break;
-		case PORT_CLK_SEL_WRPLL1:
-			dev_priv->ddi_plls.wrpll1_refcount++;
-			break;
-		case PORT_CLK_SEL_WRPLL2:
-			dev_priv->ddi_plls.wrpll2_refcount++;
-			break;
-		}
-	}
-}
-
 void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
@@ -1261,17 +1098,13 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder)
 		intel_edp_panel_on(intel_dp);
 	}
 
-	WARN_ON(crtc->ddi_pll_sel == PORT_CLK_SEL_NONE);
-	I915_WRITE(PORT_CLK_SEL(port), crtc->ddi_pll_sel);
+	WARN_ON(crtc->config.ddi_pll_sel == PORT_CLK_SEL_NONE);
+	I915_WRITE(PORT_CLK_SEL(port), crtc->config.ddi_pll_sel);
 
 	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-		struct intel_digital_port *intel_dig_port =
-			enc_to_dig_port(encoder);
 
-		intel_dp->DP = intel_dig_port->saved_port_bits |
-			       DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
-		intel_dp->DP |= DDI_PORT_WIDTH(intel_dp->lane_count);
+		intel_ddi_init_dp_buf_reg(intel_encoder);
 
 		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
 		intel_dp_start_link_train(intel_dp);
@@ -1391,30 +1224,105 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	}
 }
 
-int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
+static int bdw_get_cdclk_freq(struct drm_i915_private *dev_priv)
+{
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		return 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
+		return 540000;
+	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
+		return 337500;
+	else
+		return 675000;
+}
+
+static int hsw_get_cdclk_freq(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 	uint32_t lcpll = I915_READ(LCPLL_CTL);
 	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
 
-	if (lcpll & LCPLL_CD_SOURCE_FCLK) {
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
 		return 800000;
-	} else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT) {
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
 		return 450000;
-	} else if (freq == LCPLL_CLK_FREQ_450) {
+	else if (freq == LCPLL_CLK_FREQ_450)
 		return 450000;
-	} else if (IS_HASWELL(dev)) {
-		if (IS_ULT(dev))
-			return 337500;
-		else
-			return 540000;
-	} else {
-		if (freq == LCPLL_CLK_FREQ_54O_BDW)
-			return 540000;
-		else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
-			return 337500;
-		else
-			return 675000;
+	else if (IS_ULT(dev))
+		return 337500;
+	else
+		return 540000;
+}
+
+int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+
+	if (IS_BROADWELL(dev))
+		return bdw_get_cdclk_freq(dev_priv);
+
+	/* Haswell */
+	return hsw_get_cdclk_freq(dev_priv);
+}
+
+static void hsw_ddi_pll_enable(struct drm_i915_private *dev_priv,
+			       struct intel_shared_dpll *pll)
+{
+	I915_WRITE(WRPLL_CTL(pll->id), pll->hw_state.wrpll);
+	POSTING_READ(WRPLL_CTL(pll->id));
+	udelay(20);
+}
+
+static void hsw_ddi_pll_disable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	uint32_t val;
+
+	val = I915_READ(WRPLL_CTL(pll->id));
+	I915_WRITE(WRPLL_CTL(pll->id), val & ~WRPLL_PLL_ENABLE);
+	POSTING_READ(WRPLL_CTL(pll->id));
+}
+
+static bool hsw_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				     struct intel_shared_dpll *pll,
+				     struct intel_dpll_hw_state *hw_state)
+{
+	uint32_t val;
+
+	if (!intel_display_power_enabled(dev_priv, POWER_DOMAIN_PLLS))
+		return false;
+
+	val = I915_READ(WRPLL_CTL(pll->id));
+	hw_state->wrpll = val;
+
+	return val & WRPLL_PLL_ENABLE;
+}
+
+static const char * const hsw_ddi_pll_names[] = {
+	"WRPLL 1",
+	"WRPLL 2",
+};
+
+static void hsw_shared_dplls_init(struct drm_i915_private *dev_priv)
+{
+	int i;
+
+	dev_priv->num_shared_dpll = 2;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		dev_priv->shared_dplls[i].id = i;
+		dev_priv->shared_dplls[i].name = hsw_ddi_pll_names[i];
+		dev_priv->shared_dplls[i].disable = hsw_ddi_pll_disable;
+		dev_priv->shared_dplls[i].enable = hsw_ddi_pll_enable;
+		dev_priv->shared_dplls[i].get_hw_state =
+			hsw_ddi_pll_get_hw_state;
 	}
 }
 
@@ -1423,6 +1331,8 @@ void intel_ddi_pll_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t val = I915_READ(LCPLL_CTL);
 
+	hsw_shared_dplls_init(dev_priv);
+
 	/* The LCPLL register should be turned on by the BIOS. For now let's
 	 * just check its state and print errors in case something is wrong.
 	 * Don't even try to turn it on.
@@ -1465,10 +1375,15 @@ void intel_ddi_prepare_link_retrain(struct drm_encoder *encoder)
 			intel_wait_ddi_buf_idle(dev_priv, port);
 	}
 
-	val = DP_TP_CTL_ENABLE | DP_TP_CTL_MODE_SST |
+	val = DP_TP_CTL_ENABLE |
 	      DP_TP_CTL_LINK_TRAIN_PAT1 | DP_TP_CTL_SCRAMBLE_DISABLE;
-	if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
-		val |= DP_TP_CTL_ENHANCED_FRAME_ENABLE;
+	if (intel_dp->is_mst)
+		val |= DP_TP_CTL_MODE_MST;
+	else {
+		val |= DP_TP_CTL_MODE_SST;
+		if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+			val |= DP_TP_CTL_ENHANCED_FRAME_ENABLE;
+	}
 	I915_WRITE(DP_TP_CTL(port), val);
 	POSTING_READ(DP_TP_CTL(port));
 
@@ -1507,11 +1422,16 @@ void intel_ddi_fdi_disable(struct drm_crtc *crtc)
 
 static void intel_ddi_hot_plug(struct intel_encoder *intel_encoder)
 {
-	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
-	int type = intel_encoder->type;
+	struct intel_digital_port *intel_dig_port = enc_to_dig_port(&intel_encoder->base);
+	int type = intel_dig_port->base.type;
 
-	if (type == INTEL_OUTPUT_DISPLAYPORT || type == INTEL_OUTPUT_EDP)
-		intel_dp_check_link_status(intel_dp);
+	if (type != INTEL_OUTPUT_DISPLAYPORT &&
+	    type != INTEL_OUTPUT_EDP &&
+	    type != INTEL_OUTPUT_UNKNOWN) {
+		return;
+	}
+
+	intel_dp_hot_plug(intel_encoder);
 }
 
 void intel_ddi_get_config(struct intel_encoder *encoder,
@@ -1592,7 +1512,7 @@ void intel_ddi_get_config(struct intel_encoder *encoder,
 		dev_priv->vbt.edp_bpp = pipe_config->pipe_bpp;
 	}
 
-	intel_ddi_clock_get(encoder, pipe_config);
+	hsw_ddi_clock_get(encoder, pipe_config);
 }
 
 static void intel_ddi_destroy(struct drm_encoder *encoder)
@@ -1663,15 +1583,13 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
 	struct drm_encoder *encoder;
-	struct intel_connector *hdmi_connector = NULL;
-	struct intel_connector *dp_connector = NULL;
 	bool init_hdmi, init_dp;
 
 	init_hdmi = (dev_priv->vbt.ddi_port_info[port].supports_dvi ||
 		     dev_priv->vbt.ddi_port_info[port].supports_hdmi);
 	init_dp = dev_priv->vbt.ddi_port_info[port].supports_dp;
 	if (!init_dp && !init_hdmi) {
-		DRM_DEBUG_KMS("VBT says port %c is not DVI/HDMI/DP compatible\n",
+		DRM_DEBUG_KMS("VBT says port %c is not DVI/HDMI/DP compatible, assuming it is\n",
 			      port_name(port));
 		init_hdmi = true;
 		init_dp = true;
@@ -1701,20 +1619,28 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 					   DDI_A_4_LANES);
 
 	intel_encoder->type = INTEL_OUTPUT_UNKNOWN;
-	intel_encoder->crtc_mask =  (1 << 0) | (1 << 1) | (1 << 2);
+	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
 	intel_encoder->cloneable = 0;
 	intel_encoder->hot_plug = intel_ddi_hot_plug;
 
-	if (init_dp)
-		dp_connector = intel_ddi_init_dp_connector(intel_dig_port);
+	if (init_dp) {
+		if (!intel_ddi_init_dp_connector(intel_dig_port))
+			goto err;
+
+		intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+		dev_priv->hpd_irq_port[port] = intel_dig_port;
+	}
 
 	/* In theory we don't need the encoder->type check, but leave it just in
 	 * case we have some really bad VBTs... */
-	if (intel_encoder->type != INTEL_OUTPUT_EDP && init_hdmi)
-		hdmi_connector = intel_ddi_init_hdmi_connector(intel_dig_port);
-
-	if (!dp_connector && !hdmi_connector) {
-		drm_encoder_cleanup(encoder);
-		kfree(intel_dig_port);
+	if (intel_encoder->type != INTEL_OUTPUT_EDP && init_hdmi) {
+		if (!intel_ddi_init_hdmi_connector(intel_dig_port))
+			goto err;
 	}
+
+	return;
+
+err:
+	drm_encoder_cleanup(encoder);
+	kfree(intel_dig_port);
 }
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index f0be855ddf45..f0a1a56406eb 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -39,12 +39,42 @@
 #include "i915_trace.h"
 #include <drm/drm_dp_helper.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_plane_helper.h>
+#include <drm/drm_rect.h>
 #include <linux/dma_remapping.h>
 
-#define DIV_ROUND_CLOSEST_ULL(ll, d)	\
-	({ unsigned long long _tmp = (ll)+(d)/2; do_div(_tmp, d); _tmp; })
+/* Primary plane formats supported by all gen */
+#define COMMON_PRIMARY_FORMATS \
+	DRM_FORMAT_C8, \
+	DRM_FORMAT_RGB565, \
+	DRM_FORMAT_XRGB8888, \
+	DRM_FORMAT_ARGB8888
+
+/* Primary plane formats for gen <= 3 */
+static const uint32_t intel_primary_formats_gen2[] = {
+	COMMON_PRIMARY_FORMATS,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_ARGB1555,
+};
+
+/* Primary plane formats for gen >= 4 */
+static const uint32_t intel_primary_formats_gen4[] = {
+	COMMON_PRIMARY_FORMATS, \
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ABGR2101010,
+};
+
+/* Cursor formats */
+static const uint32_t intel_cursor_formats[] = {
+	DRM_FORMAT_ARGB8888,
+};
 
-static void intel_increase_pllclock(struct drm_crtc *crtc);
+static void intel_increase_pllclock(struct drm_device *dev,
+				    enum pipe pipe);
 static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on);
 
 static void i9xx_crtc_clock_get(struct intel_crtc *crtc,
@@ -58,15 +88,24 @@ static int intel_framebuffer_init(struct drm_device *dev,
 				  struct intel_framebuffer *ifb,
 				  struct drm_mode_fb_cmd2 *mode_cmd,
 				  struct drm_i915_gem_object *obj);
-static void intel_dp_set_m_n(struct intel_crtc *crtc);
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc);
 static void intel_set_pipe_timings(struct intel_crtc *intel_crtc);
 static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2);
 static void ironlake_set_pipeconf(struct drm_crtc *crtc);
 static void haswell_set_pipeconf(struct drm_crtc *crtc);
 static void intel_set_pipe_csc(struct drm_crtc *crtc);
 static void vlv_prepare_pll(struct intel_crtc *crtc);
+static void chv_prepare_pll(struct intel_crtc *crtc);
+
+static struct intel_encoder *intel_find_encoder(struct intel_connector *connector, int pipe)
+{
+	if (!connector->mst_port)
+		return connector->encoder;
+	else
+		return &connector->mst_port->mst_encoders[pipe]->base;
+}
 
 typedef struct {
 	int	min, max;
@@ -858,7 +897,8 @@ static void g4x_wait_for_vblank(struct drm_device *dev, int pipe)
 	frame = I915_READ(frame_reg);
 
 	if (wait_for(I915_READ_NOTRACE(frame_reg) != frame, 50))
-		WARN(1, "vblank wait timed out\n");
+		WARN(1, "vblank wait on pipe %c timed out\n",
+		     pipe_name(pipe));
 }
 
 /**
@@ -899,7 +939,8 @@ void intel_wait_for_vblank(struct drm_device *dev, int pipe)
 	if (wait_for(I915_READ(pipestat_reg) &
 		     PIPE_VBLANK_INTERRUPT_STATUS,
 		     50))
-		DRM_DEBUG_KMS("vblank wait timed out\n");
+		DRM_DEBUG_KMS("vblank wait on pipe %c timed out\n",
+			      pipe_name(pipe));
 }
 
 static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
@@ -923,8 +964,7 @@ static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
 
 /*
  * intel_wait_for_pipe_off - wait for pipe to turn off
- * @dev: drm device
- * @pipe: pipe to wait for
+ * @crtc: crtc whose pipe to wait for
  *
  * After disabling a pipe, we can't wait for vblank in the usual way,
  * spinning on the vblank interrupt status bit, since we won't actually
@@ -938,11 +978,12 @@ static bool pipe_dsl_stopped(struct drm_device *dev, enum pipe pipe)
  *   ends up stopping at the start of the next frame).
  *
  */
-void intel_wait_for_pipe_off(struct drm_device *dev, int pipe)
+static void intel_wait_for_pipe_off(struct intel_crtc *crtc)
 {
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		int reg = PIPECONF(cpu_transcoder);
@@ -1061,11 +1102,6 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv,
 	bool cur_state;
 	struct intel_dpll_hw_state hw_state;
 
-	if (HAS_PCH_LPT(dev_priv->dev)) {
-		DRM_DEBUG_DRIVER("LPT detected: skipping PCH PLL test\n");
-		return;
-	}
-
 	if (WARN (!pll,
 		  "asserting DPLL %s with no DPLL\n", state_string(state)))
 		return;
@@ -1156,27 +1192,40 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv,
 static void assert_panel_unlocked(struct drm_i915_private *dev_priv,
 				  enum pipe pipe)
 {
-	int pp_reg, lvds_reg;
+	struct drm_device *dev = dev_priv->dev;
+	int pp_reg;
 	u32 val;
 	enum pipe panel_pipe = PIPE_A;
 	bool locked = true;
 
-	if (HAS_PCH_SPLIT(dev_priv->dev)) {
+	if (WARN_ON(HAS_DDI(dev)))
+		return;
+
+	if (HAS_PCH_SPLIT(dev)) {
+		u32 port_sel;
+
 		pp_reg = PCH_PP_CONTROL;
-		lvds_reg = PCH_LVDS;
+		port_sel = I915_READ(PCH_PP_ON_DELAYS) & PANEL_PORT_SELECT_MASK;
+
+		if (port_sel == PANEL_PORT_SELECT_LVDS &&
+		    I915_READ(PCH_LVDS) & LVDS_PIPEB_SELECT)
+			panel_pipe = PIPE_B;
+		/* XXX: else fix for eDP */
+	} else if (IS_VALLEYVIEW(dev)) {
+		/* presumably write lock depends on pipe, not port select */
+		pp_reg = VLV_PIPE_PP_CONTROL(pipe);
+		panel_pipe = pipe;
 	} else {
 		pp_reg = PP_CONTROL;
-		lvds_reg = LVDS;
+		if (I915_READ(LVDS) & LVDS_PIPEB_SELECT)
+			panel_pipe = PIPE_B;
 	}
 
 	val = I915_READ(pp_reg);
 	if (!(val & PANEL_POWER_ON) ||
-	    ((val & PANEL_UNLOCK_REGS) == PANEL_UNLOCK_REGS))
+	    ((val & PANEL_UNLOCK_MASK) == PANEL_UNLOCK_REGS))
 		locked = false;
 
-	if (I915_READ(lvds_reg) & LVDS_PIPEB_SELECT)
-		panel_pipe = PIPE_B;
-
 	WARN(panel_pipe == pipe && locked,
 	     "panel assertion failure, pipe %c regs locked\n",
 	     pipe_name(pipe));
@@ -1209,8 +1258,9 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 
-	/* if we need the pipe A quirk it must be always on */
-	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
+	/* if we need the pipe quirk it must be always on */
+	if ((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 		state = true;
 
 	if (!intel_display_power_enabled(dev_priv,
@@ -1264,7 +1314,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	}
 
 	/* Need to check both planes against the pipe */
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		reg = DSPCNTR(i);
 		val = I915_READ(reg);
 		cur_pipe = (val & DISPPLANE_SEL_PIPE_MASK) >>
@@ -1305,6 +1355,12 @@ static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 	}
 }
 
+static void assert_vblank_disabled(struct drm_crtc *crtc)
+{
+	if (WARN_ON(drm_crtc_vblank_get(crtc) == 0))
+		drm_crtc_vblank_put(crtc);
+}
+
 static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
 {
 	u32 val;
@@ -1477,57 +1533,6 @@ static void intel_init_dpio(struct drm_device *dev)
 	}
 }
 
-static void intel_reset_dpio(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (!IS_VALLEYVIEW(dev))
-		return;
-
-	if (IS_CHERRYVIEW(dev)) {
-		enum dpio_phy phy;
-		u32 val;
-
-		for (phy = DPIO_PHY0; phy < I915_NUM_PHYS_VLV; phy++) {
-			/* Poll for phypwrgood signal */
-			if (wait_for(I915_READ(DISPLAY_PHY_STATUS) &
-						PHY_POWERGOOD(phy), 1))
-				DRM_ERROR("Display PHY %d is not power up\n", phy);
-
-			/*
-			 * Deassert common lane reset for PHY.
-			 *
-			 * This should only be done on init and resume from S3
-			 * with both PLLs disabled, or we risk losing DPIO and
-			 * PLL synchronization.
-			 */
-			val = I915_READ(DISPLAY_PHY_CONTROL);
-			I915_WRITE(DISPLAY_PHY_CONTROL,
-				PHY_COM_LANE_RESET_DEASSERT(phy, val));
-		}
-
-	} else {
-		/*
-		 * If DPIO has already been reset, e.g. by BIOS, just skip all
-		 * this.
-		 */
-		if (I915_READ(DPIO_CTL) & DPIO_CMNRST)
-			return;
-
-		/*
-		 * From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
-		 * Need to assert and de-assert PHY SB reset by gating the
-		 * common lane power, then un-gating it.
-		 * Simply ungating isn't enough to reset the PHY enough to get
-		 * ports and lanes running.
-		 */
-		__vlv_set_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC,
-				     false);
-		__vlv_set_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC,
-				     true);
-	}
-}
-
 static void vlv_enable_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
@@ -1541,7 +1546,7 @@ static void vlv_enable_pll(struct intel_crtc *crtc)
 	BUG_ON(!IS_VALLEYVIEW(dev_priv->dev));
 
 	/* PLL is protected by panel, make sure we can write it */
-	if (IS_MOBILE(dev_priv->dev) && !IS_I830(dev_priv->dev))
+	if (IS_MOBILE(dev_priv->dev))
 		assert_panel_unlocked(dev_priv, crtc->pipe);
 
 	I915_WRITE(reg, dpll);
@@ -1604,6 +1609,18 @@ static void chv_enable_pll(struct intel_crtc *crtc)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static int intel_num_dvo_pipes(struct drm_device *dev)
+{
+	struct intel_crtc *crtc;
+	int count = 0;
+
+	for_each_intel_crtc(dev, crtc)
+		count += crtc->active &&
+			intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DVO);
+
+	return count;
+}
+
 static void i9xx_enable_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
@@ -1620,7 +1637,18 @@ static void i9xx_enable_pll(struct intel_crtc *crtc)
 	if (IS_MOBILE(dev) && !IS_I830(dev))
 		assert_panel_unlocked(dev_priv, crtc->pipe);
 
-	I915_WRITE(reg, dpll);
+	/* Enable DVO 2x clock on both PLLs if necessary */
+	if (IS_I830(dev) && intel_num_dvo_pipes(dev) > 0) {
+		/*
+		 * It appears to be important that we don't enable this
+		 * for the current pipe before otherwise configuring the
+		 * PLL. No idea how this should be handled if multiple
+		 * DVO outputs are enabled simultaneosly.
+		 */
+		dpll |= DPLL_DVO_2X_MODE;
+		I915_WRITE(DPLL(!crtc->pipe),
+			   I915_READ(DPLL(!crtc->pipe)) | DPLL_DVO_2X_MODE);
+	}
 
 	/* Wait for the clocks to stabilize. */
 	POSTING_READ(reg);
@@ -1659,10 +1687,25 @@ static void i9xx_enable_pll(struct intel_crtc *crtc)
  *
  * Note!  This is for pre-ILK only.
  */
-static void i9xx_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
+static void i9xx_disable_pll(struct intel_crtc *crtc)
 {
-	/* Don't disable pipe A or pipe A PLLs if needed */
-	if (pipe == PIPE_A && (dev_priv->quirks & QUIRK_PIPEA_FORCE))
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe = crtc->pipe;
+
+	/* Disable DVO 2x clock on both PLLs if necessary */
+	if (IS_I830(dev) &&
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DVO) &&
+	    intel_num_dvo_pipes(dev) == 1) {
+		I915_WRITE(DPLL(PIPE_B),
+			   I915_READ(DPLL(PIPE_B)) & ~DPLL_DVO_2X_MODE);
+		I915_WRITE(DPLL(PIPE_A),
+			   I915_READ(DPLL(PIPE_A)) & ~DPLL_DVO_2X_MODE);
+	}
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if ((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 		return;
 
 	/* Make sure the pipe isn't still relying on us */
@@ -1699,7 +1742,7 @@ static void chv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	assert_pipe_disabled(dev_priv, pipe);
 
 	/* Set PLL en = 0 */
-	val = DPLL_SSC_REF_CLOCK_CHV;
+	val = DPLL_SSC_REF_CLOCK_CHV | DPLL_REFA_CLK_ENABLE_VLV;
 	if (pipe != PIPE_A)
 		val |= DPLL_INTEGRATED_CRI_CLK_VLV;
 	I915_WRITE(DPLL(pipe), val);
@@ -1712,6 +1755,17 @@ static void chv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	val &= ~DPIO_DCLKP_EN;
 	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW14(port), val);
 
+	/* disable left/right clock distribution */
+	if (pipe != PIPE_B) {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW5_CH0);
+		val &= ~(CHV_BUFLEFTENA1_MASK | CHV_BUFRIGHTENA1_MASK);
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW5_CH0, val);
+	} else {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW1_CH1);
+		val &= ~(CHV_BUFLEFTENA2_MASK | CHV_BUFRIGHTENA2_MASK);
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW1_CH1, val);
+	}
+
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
@@ -1749,6 +1803,9 @@ static void intel_prepare_shared_dpll(struct intel_crtc *crtc)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
+	if (WARN_ON(pll == NULL))
+		return;
+
 	WARN_ON(!pll->refcount);
 	if (pll->active == 0) {
 		DRM_DEBUG_DRIVER("setting up %s\n", pll->name);
@@ -1779,7 +1836,7 @@ static void intel_enable_shared_dpll(struct intel_crtc *crtc)
 	if (WARN_ON(pll->refcount == 0))
 		return;
 
-	DRM_DEBUG_KMS("enable %s (active %d, on? %d)for crtc %d\n",
+	DRM_DEBUG_KMS("enable %s (active %d, on? %d) for crtc %d\n",
 		      pll->name, pll->active, pll->on,
 		      crtc->base.base.id);
 
@@ -1790,6 +1847,8 @@ static void intel_enable_shared_dpll(struct intel_crtc *crtc)
 	}
 	WARN_ON(pll->on);
 
+	intel_display_power_get(dev_priv, POWER_DOMAIN_PLLS);
+
 	DRM_DEBUG_KMS("enabling %s\n", pll->name);
 	pll->enable(dev_priv, pll);
 	pll->on = true;
@@ -1826,6 +1885,8 @@ static void intel_disable_shared_dpll(struct intel_crtc *crtc)
 	DRM_DEBUG_KMS("disabling %s\n", pll->name);
 	pll->disable(dev_priv, pll);
 	pll->on = false;
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS);
 }
 
 static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
@@ -1837,7 +1898,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	uint32_t reg, val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
-	BUG_ON(INTEL_INFO(dev)->gen < 5);
+	BUG_ON(!HAS_PCH_SPLIT(dev));
 
 	/* Make sure PCH DPLL is enabled */
 	assert_shared_dpll_enabled(dev_priv,
@@ -1890,7 +1951,7 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	u32 val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
-	BUG_ON(INTEL_INFO(dev_priv->dev)->gen < 5);
+	BUG_ON(!HAS_PCH_SPLIT(dev_priv->dev));
 
 	/* FDI must be feeding us bits for PCH ports */
 	assert_fdi_tx_enabled(dev_priv, (enum pipe) cpu_transcoder);
@@ -2012,8 +2073,8 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 	reg = PIPECONF(cpu_transcoder);
 	val = I915_READ(reg);
 	if (val & PIPECONF_ENABLE) {
-		WARN_ON(!(pipe == PIPE_A &&
-			  dev_priv->quirks & QUIRK_PIPEA_FORCE));
+		WARN_ON(!((pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+			  (pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE)));
 		return;
 	}
 
@@ -2023,21 +2084,19 @@ static void intel_enable_pipe(struct intel_crtc *crtc)
 
 /**
  * intel_disable_pipe - disable a pipe, asserting requirements
- * @dev_priv: i915 private structure
- * @pipe: pipe to disable
- *
- * Disable @pipe, making sure that various hardware specific requirements
- * are met, if applicable, e.g. plane disabled, panel fitter off, etc.
+ * @crtc: crtc whose pipes is to be disabled
  *
- * @pipe should be %PIPE_A or %PIPE_B.
+ * Disable the pipe of @crtc, making sure that various hardware
+ * specific requirements are met, if applicable, e.g. plane
+ * disabled, panel fitter off, etc.
  *
  * Will wait until the pipe has shut down before returning.
  */
-static void intel_disable_pipe(struct drm_i915_private *dev_priv,
-			       enum pipe pipe)
+static void intel_disable_pipe(struct intel_crtc *crtc)
 {
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
 	int reg;
 	u32 val;
 
@@ -2049,17 +2108,26 @@ static void intel_disable_pipe(struct drm_i915_private *dev_priv,
 	assert_cursor_disabled(dev_priv, pipe);
 	assert_sprites_disabled(dev_priv, pipe);
 
-	/* Don't disable pipe A or pipe A PLLs if needed */
-	if (pipe == PIPE_A && (dev_priv->quirks & QUIRK_PIPEA_FORCE))
-		return;
-
 	reg = PIPECONF(cpu_transcoder);
 	val = I915_READ(reg);
 	if ((val & PIPECONF_ENABLE) == 0)
 		return;
 
-	I915_WRITE(reg, val & ~PIPECONF_ENABLE);
-	intel_wait_for_pipe_off(dev_priv->dev, pipe);
+	/*
+	 * Double wide has implications for planes
+	 * so best keep it disabled when not needed.
+	 */
+	if (crtc->config.double_wide)
+		val &= ~PIPECONF_DOUBLE_WIDE;
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if (!(pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) &&
+	    !(pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
+		val &= ~PIPECONF_ENABLE;
+
+	I915_WRITE(reg, val);
+	if ((val & PIPECONF_ENABLE) == 0)
+		intel_wait_for_pipe_off(crtc);
 }
 
 /*
@@ -2078,35 +2146,28 @@ void intel_flush_primary_plane(struct drm_i915_private *dev_priv,
 
 /**
  * intel_enable_primary_hw_plane - enable the primary plane on a given pipe
- * @dev_priv: i915 private structure
- * @plane: plane to enable
- * @pipe: pipe being fed
+ * @plane:  plane to be enabled
+ * @crtc: crtc for the plane
  *
- * Enable @plane on @pipe, making sure that @pipe is running first.
+ * Enable @plane on @crtc, making sure that the pipe is running first.
  */
-static void intel_enable_primary_hw_plane(struct drm_i915_private *dev_priv,
-					  enum plane plane, enum pipe pipe)
+static void intel_enable_primary_hw_plane(struct drm_plane *plane,
+					  struct drm_crtc *crtc)
 {
-	struct drm_device *dev = dev_priv->dev;
-	struct intel_crtc *intel_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-	int reg;
-	u32 val;
+	struct drm_device *dev = plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
 	/* If the pipe isn't enabled, we can't pump pixels and may hang */
-	assert_pipe_enabled(dev_priv, pipe);
+	assert_pipe_enabled(dev_priv, intel_crtc->pipe);
 
 	if (intel_crtc->primary_enabled)
 		return;
 
 	intel_crtc->primary_enabled = true;
 
-	reg = DSPCNTR(plane);
-	val = I915_READ(reg);
-	WARN_ON(val & DISPLAY_PLANE_ENABLE);
-
-	I915_WRITE(reg, val | DISPLAY_PLANE_ENABLE);
-	intel_flush_primary_plane(dev_priv, plane);
+	dev_priv->display.update_primary_plane(crtc, plane->fb,
+					       crtc->x, crtc->y);
 
 	/*
 	 * BDW signals flip done immediately if the plane
@@ -2119,31 +2180,27 @@ static void intel_enable_primary_hw_plane(struct drm_i915_private *dev_priv,
 
 /**
  * intel_disable_primary_hw_plane - disable the primary hardware plane
- * @dev_priv: i915 private structure
- * @plane: plane to disable
- * @pipe: pipe consuming the data
+ * @plane: plane to be disabled
+ * @crtc: crtc for the plane
  *
- * Disable @plane; should be an independent operation.
+ * Disable @plane on @crtc, making sure that the pipe is running first.
  */
-static void intel_disable_primary_hw_plane(struct drm_i915_private *dev_priv,
-					   enum plane plane, enum pipe pipe)
+static void intel_disable_primary_hw_plane(struct drm_plane *plane,
+					   struct drm_crtc *crtc)
 {
-	struct intel_crtc *intel_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-	int reg;
-	u32 val;
+	struct drm_device *dev = plane->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	assert_pipe_enabled(dev_priv, intel_crtc->pipe);
 
 	if (!intel_crtc->primary_enabled)
 		return;
 
 	intel_crtc->primary_enabled = false;
 
-	reg = DSPCNTR(plane);
-	val = I915_READ(reg);
-	WARN_ON((val & DISPLAY_PLANE_ENABLE) == 0);
-
-	I915_WRITE(reg, val & ~DISPLAY_PLANE_ENABLE);
-	intel_flush_primary_plane(dev_priv, plane);
+	dev_priv->display.update_primary_plane(crtc, plane->fb,
+					       crtc->x, crtc->y);
 }
 
 static bool need_vtd_wa(struct drm_device *dev)
@@ -2172,6 +2229,8 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 	u32 alignment;
 	int ret;
 
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
 	switch (obj->tiling_mode) {
 	case I915_TILING_NONE:
 		if (IS_BROADWATER(dev) || IS_CRESTLINE(dev))
@@ -2200,6 +2259,15 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 	if (need_vtd_wa(dev) && alignment < 256 * 1024)
 		alignment = 256 * 1024;
 
+	/*
+	 * Global gtt pte registers are special registers which actually forward
+	 * writes to a chunk of system memory. Which means that there is no risk
+	 * that the register values disappear as soon as we call
+	 * intel_runtime_pm_put(), so it is correct to wrap only the
+	 * pin/unpin/fence and not more.
+	 */
+	intel_runtime_pm_get(dev_priv);
+
 	dev_priv->mm.interruptible = false;
 	ret = i915_gem_object_pin_to_display_plane(obj, alignment, pipelined);
 	if (ret)
@@ -2217,17 +2285,21 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 	i915_gem_object_pin_fence(obj);
 
 	dev_priv->mm.interruptible = true;
+	intel_runtime_pm_put(dev_priv);
 	return 0;
 
 err_unpin:
 	i915_gem_object_unpin_from_display_plane(obj);
 err_interruptible:
 	dev_priv->mm.interruptible = true;
+	intel_runtime_pm_put(dev_priv);
 	return ret;
 }
 
 void intel_unpin_fb_obj(struct drm_i915_gem_object *obj)
 {
+	WARN_ON(!mutex_is_locked(&obj->base.dev->struct_mutex));
+
 	i915_gem_object_unpin_fence(obj);
 	i915_gem_object_unpin_from_display_plane(obj);
 }
@@ -2314,6 +2386,7 @@ static bool intel_alloc_plane_obj(struct intel_crtc *crtc,
 		goto out_unref_obj;
 	}
 
+	obj->frontbuffer_bits = INTEL_FRONTBUFFER_PRIMARY(crtc->pipe);
 	mutex_unlock(&dev->struct_mutex);
 
 	DRM_DEBUG_KMS("plane fb obj %p\n", obj);
@@ -2331,7 +2404,7 @@ static void intel_find_plane_obj(struct intel_crtc *intel_crtc,
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_crtc *c;
 	struct intel_crtc *i;
-	struct intel_framebuffer *fb;
+	struct drm_i915_gem_object *obj;
 
 	if (!intel_crtc->base.primary->fb)
 		return;
@@ -2352,13 +2425,17 @@ static void intel_find_plane_obj(struct intel_crtc *intel_crtc,
 		if (c == &intel_crtc->base)
 			continue;
 
-		if (!i->active || !c->primary->fb)
+		if (!i->active)
+			continue;
+
+		obj = intel_fb_obj(c->primary->fb);
+		if (obj == NULL)
 			continue;
 
-		fb = to_intel_framebuffer(c->primary->fb);
-		if (i915_gem_obj_ggtt_offset(fb->obj) == plane_config->base) {
+		if (i915_gem_obj_ggtt_offset(obj) == plane_config->base) {
 			drm_framebuffer_reference(c->primary->fb);
 			intel_crtc->base.primary->fb = c->primary->fb;
+			obj->frontbuffer_bits |= INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe);
 			break;
 		}
 	}
@@ -2371,20 +2448,46 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
 	int plane = intel_crtc->plane;
 	unsigned long linear_offset;
 	u32 dspcntr;
-	u32 reg;
+	u32 reg = DSPCNTR(plane);
+	int pixel_size;
 
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
+	if (!intel_crtc->primary_enabled) {
+		I915_WRITE(reg, 0);
+		if (INTEL_INFO(dev)->gen >= 4)
+			I915_WRITE(DSPSURF(plane), 0);
+		else
+			I915_WRITE(DSPADDR(plane), 0);
+		POSTING_READ(reg);
+		return;
+	}
+
+	obj = intel_fb_obj(fb);
+	if (WARN_ON(obj == NULL))
+		return;
+
+	pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	dspcntr = DISPPLANE_GAMMA_ENABLE;
+
+	dspcntr |= DISPLAY_PLANE_ENABLE;
+
+	if (INTEL_INFO(dev)->gen < 4) {
+		if (intel_crtc->pipe == PIPE_B)
+			dspcntr |= DISPPLANE_SEL_PIPE_B;
+
+		/* pipesrc and dspsize control the size that is scaled from,
+		 * which should always be the user's requested size.
+		 */
+		I915_WRITE(DSPSIZE(plane),
+			   ((intel_crtc->config.pipe_src_h - 1) << 16) |
+			   (intel_crtc->config.pipe_src_w - 1));
+		I915_WRITE(DSPPOS(plane), 0);
+	}
 
-	reg = DSPCNTR(plane);
-	dspcntr = I915_READ(reg);
-	/* Mask out pixel format bits in case we change it */
-	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_C8:
 		dspcntr |= DISPPLANE_8BPP;
@@ -2416,30 +2519,40 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc,
 		BUG();
 	}
 
-	if (INTEL_INFO(dev)->gen >= 4) {
-		if (obj->tiling_mode != I915_TILING_NONE)
-			dspcntr |= DISPPLANE_TILED;
-		else
-			dspcntr &= ~DISPPLANE_TILED;
-	}
+	if (INTEL_INFO(dev)->gen >= 4 &&
+	    obj->tiling_mode != I915_TILING_NONE)
+		dspcntr |= DISPPLANE_TILED;
 
 	if (IS_G4X(dev))
 		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
 
-	I915_WRITE(reg, dspcntr);
-
-	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
+	linear_offset = y * fb->pitches[0] + x * pixel_size;
 
 	if (INTEL_INFO(dev)->gen >= 4) {
 		intel_crtc->dspaddr_offset =
 			intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
-						       fb->bits_per_pixel / 8,
+						       pixel_size,
 						       fb->pitches[0]);
 		linear_offset -= intel_crtc->dspaddr_offset;
 	} else {
 		intel_crtc->dspaddr_offset = linear_offset;
 	}
 
+	if (to_intel_plane(crtc->primary)->rotation == BIT(DRM_ROTATE_180)) {
+		dspcntr |= DISPPLANE_ROTATE_180;
+
+		x += (intel_crtc->config.pipe_src_w - 1);
+		y += (intel_crtc->config.pipe_src_h - 1);
+
+		/* Finding the last pixel of the last line of the display
+		data and adding to linear_offset*/
+		linear_offset +=
+			(intel_crtc->config.pipe_src_h - 1) * fb->pitches[0] +
+			(intel_crtc->config.pipe_src_w - 1) * pixel_size;
+	}
+
+	I915_WRITE(reg, dspcntr);
+
 	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
 		      i915_gem_obj_ggtt_offset(obj), linear_offset, x, y,
 		      fb->pitches[0]);
@@ -2461,20 +2574,33 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
 	int plane = intel_crtc->plane;
 	unsigned long linear_offset;
 	u32 dspcntr;
-	u32 reg;
+	u32 reg = DSPCNTR(plane);
+	int pixel_size;
 
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
+	if (!intel_crtc->primary_enabled) {
+		I915_WRITE(reg, 0);
+		I915_WRITE(DSPSURF(plane), 0);
+		POSTING_READ(reg);
+		return;
+	}
+
+	obj = intel_fb_obj(fb);
+	if (WARN_ON(obj == NULL))
+		return;
+
+	pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+
+	dspcntr = DISPPLANE_GAMMA_ENABLE;
+
+	dspcntr |= DISPLAY_PLANE_ENABLE;
+
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		dspcntr |= DISPPLANE_PIPE_CSC_ENABLE;
 
-	reg = DSPCNTR(plane);
-	dspcntr = I915_READ(reg);
-	/* Mask out pixel format bits in case we change it */
-	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_C8:
 		dspcntr |= DISPPLANE_8BPP;
@@ -2504,22 +2630,32 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc,
 
 	if (obj->tiling_mode != I915_TILING_NONE)
 		dspcntr |= DISPPLANE_TILED;
-	else
-		dspcntr &= ~DISPPLANE_TILED;
 
-	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
-		dspcntr &= ~DISPPLANE_TRICKLE_FEED_DISABLE;
-	else
+	if (!IS_HASWELL(dev) && !IS_BROADWELL(dev))
 		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
 
-	I915_WRITE(reg, dspcntr);
-
-	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
+	linear_offset = y * fb->pitches[0] + x * pixel_size;
 	intel_crtc->dspaddr_offset =
 		intel_gen4_compute_page_offset(&x, &y, obj->tiling_mode,
-					       fb->bits_per_pixel / 8,
+					       pixel_size,
 					       fb->pitches[0]);
 	linear_offset -= intel_crtc->dspaddr_offset;
+	if (to_intel_plane(crtc->primary)->rotation == BIT(DRM_ROTATE_180)) {
+		dspcntr |= DISPPLANE_ROTATE_180;
+
+		if (!IS_HASWELL(dev) && !IS_BROADWELL(dev)) {
+			x += (intel_crtc->config.pipe_src_w - 1);
+			y += (intel_crtc->config.pipe_src_h - 1);
+
+			/* Finding the last pixel of the last line of the display
+			data and adding to linear_offset*/
+			linear_offset +=
+				(intel_crtc->config.pipe_src_h - 1) * fb->pitches[0] +
+				(intel_crtc->config.pipe_src_w - 1) * pixel_size;
+		}
+	}
+
+	I915_WRITE(reg, dspcntr);
 
 	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
 		      i915_gem_obj_ggtt_offset(obj), linear_offset, x, y,
@@ -2546,7 +2682,7 @@ intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 
 	if (dev_priv->display.disable_fbc)
 		dev_priv->display.disable_fbc(dev);
-	intel_increase_pllclock(crtc);
+	intel_increase_pllclock(dev, to_intel_crtc(crtc)->pipe);
 
 	dev_priv->display.update_primary_plane(crtc, fb, x, y);
 
@@ -2601,7 +2737,7 @@ void intel_display_handle_reset(struct drm_device *dev)
 static int
 intel_finish_fb(struct drm_framebuffer *old_fb)
 {
-	struct drm_i915_gem_object *obj = to_intel_framebuffer(old_fb)->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(old_fb);
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
 	bool was_interruptible = dev_priv->mm.interruptible;
 	int ret;
@@ -2647,7 +2783,10 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_framebuffer *old_fb;
+	enum pipe pipe = intel_crtc->pipe;
+	struct drm_framebuffer *old_fb = crtc->primary->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *old_obj = intel_fb_obj(old_fb);
 	int ret;
 
 	if (intel_crtc_has_pending_flip(crtc)) {
@@ -2669,9 +2808,10 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	}
 
 	mutex_lock(&dev->struct_mutex);
-	ret = intel_pin_and_fence_fb_obj(dev,
-					 to_intel_framebuffer(fb)->obj,
-					 NULL);
+	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
+	if (ret == 0)
+		i915_gem_track_fb(old_obj, obj,
+				  INTEL_FRONTBUFFER_PRIMARY(pipe));
 	mutex_unlock(&dev->struct_mutex);
 	if (ret != 0) {
 		DRM_ERROR("pin & fence failed\n");
@@ -2711,7 +2851,9 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 
 	dev_priv->display.update_primary_plane(crtc, fb, x, y);
 
-	old_fb = crtc->primary->fb;
+	if (intel_crtc->active)
+		intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_PRIMARY(pipe));
+
 	crtc->primary->fb = fb;
 	crtc->x = x;
 	crtc->y = y;
@@ -2720,13 +2862,12 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		if (intel_crtc->active && old_fb != fb)
 			intel_wait_for_vblank(dev, intel_crtc->pipe);
 		mutex_lock(&dev->struct_mutex);
-		intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
+		intel_unpin_fb_obj(old_obj);
 		mutex_unlock(&dev->struct_mutex);
 	}
 
 	mutex_lock(&dev->struct_mutex);
 	intel_update_fbc(dev);
-	intel_edp_psr_update(dev);
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -3298,23 +3439,54 @@ bool intel_has_pending_fb_unpin(struct drm_device *dev)
 	return false;
 }
 
+static void page_flip_completed(struct intel_crtc *intel_crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
+	struct intel_unpin_work *work = intel_crtc->unpin_work;
+
+	/* ensure that the unpin work is consistent wrt ->pending. */
+	smp_rmb();
+	intel_crtc->unpin_work = NULL;
+
+	if (work->event)
+		drm_send_vblank_event(intel_crtc->base.dev,
+				      intel_crtc->pipe,
+				      work->event);
+
+	drm_crtc_vblank_put(&intel_crtc->base);
+
+	wake_up_all(&dev_priv->pending_flip_queue);
+	queue_work(dev_priv->wq, &work->work);
+
+	trace_i915_flip_complete(intel_crtc->plane,
+				 work->pending_flip_obj);
+}
+
 void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (crtc->primary->fb == NULL)
-		return;
-
 	WARN_ON(waitqueue_active(&dev_priv->pending_flip_queue));
+	if (WARN_ON(wait_event_timeout(dev_priv->pending_flip_queue,
+				       !intel_crtc_has_pending_flip(crtc),
+				       60*HZ) == 0)) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+		unsigned long flags;
 
-	WARN_ON(wait_event_timeout(dev_priv->pending_flip_queue,
-				   !intel_crtc_has_pending_flip(crtc),
-				   60*HZ) == 0);
+		spin_lock_irqsave(&dev->event_lock, flags);
+		if (intel_crtc->unpin_work) {
+			WARN_ONCE(1, "Removing stuck page flip\n");
+			page_flip_completed(intel_crtc);
+		}
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	}
 
-	mutex_lock(&dev->struct_mutex);
-	intel_finish_fb(crtc->primary->fb);
-	mutex_unlock(&dev->struct_mutex);
+	if (crtc->primary->fb) {
+		mutex_lock(&dev->struct_mutex);
+		intel_finish_fb(crtc->primary->fb);
+		mutex_unlock(&dev->struct_mutex);
+	}
 }
 
 /* Program iCLKIP clock to the desired frequency */
@@ -3587,7 +3759,7 @@ static void lpt_pch_enable(struct drm_crtc *crtc)
 	lpt_enable_pch_transcoder(dev_priv, cpu_transcoder);
 }
 
-static void intel_put_shared_dpll(struct intel_crtc *crtc)
+void intel_put_shared_dpll(struct intel_crtc *crtc)
 {
 	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
@@ -3607,7 +3779,7 @@ static void intel_put_shared_dpll(struct intel_crtc *crtc)
 	crtc->config.shared_dpll = DPLL_ID_PRIVATE;
 }
 
-static struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
+struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
 	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
@@ -3818,7 +3990,7 @@ static void intel_crtc_load_lut(struct drm_crtc *crtc)
 	}
 
 	/* use legacy palette for Ironlake */
-	if (HAS_PCH_SPLIT(dev))
+	if (!HAS_GMCH_DISPLAY(dev))
 		palreg = LGC_PALETTE(pipe);
 
 	/* Workaround : Do not read or write the pipe palette/gamma data while
@@ -3860,43 +4032,18 @@ static void intel_crtc_dpms_overlay(struct intel_crtc *intel_crtc, bool enable)
 	 */
 }
 
-/**
- * i9xx_fixup_plane - ugly workaround for G45 to fire up the hardware
- * cursor plane briefly if not already running after enabling the display
- * plane.
- * This workaround avoids occasional blank screens when self refresh is
- * enabled.
- */
-static void
-g4x_fixup_plane(struct drm_i915_private *dev_priv, enum pipe pipe)
-{
-	u32 cntl = I915_READ(CURCNTR(pipe));
-
-	if ((cntl & CURSOR_MODE) == 0) {
-		u32 fw_bcl_self = I915_READ(FW_BLC_SELF);
-
-		I915_WRITE(FW_BLC_SELF, fw_bcl_self & ~FW_BLC_SELF_EN);
-		I915_WRITE(CURCNTR(pipe), CURSOR_MODE_64_ARGB_AX);
-		intel_wait_for_vblank(dev_priv->dev, pipe);
-		I915_WRITE(CURCNTR(pipe), cntl);
-		I915_WRITE(CURBASE(pipe), I915_READ(CURBASE(pipe)));
-		I915_WRITE(FW_BLC_SELF, fw_bcl_self);
-	}
-}
-
 static void intel_crtc_enable_planes(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
 
-	intel_enable_primary_hw_plane(dev_priv, plane, pipe);
+	assert_vblank_disabled(crtc);
+
+	drm_vblank_on(dev, pipe);
+
+	intel_enable_primary_hw_plane(crtc->primary, crtc);
 	intel_enable_planes(crtc);
-	/* The fixup needs to happen before cursor is enabled */
-	if (IS_G4X(dev))
-		g4x_fixup_plane(dev_priv, pipe);
 	intel_crtc_update_cursor(crtc, true);
 	intel_crtc_dpms_overlay(intel_crtc, true);
 
@@ -3904,8 +4051,14 @@ static void intel_crtc_enable_planes(struct drm_crtc *crtc)
 
 	mutex_lock(&dev->struct_mutex);
 	intel_update_fbc(dev);
-	intel_edp_psr_update(dev);
 	mutex_unlock(&dev->struct_mutex);
+
+	/*
+	 * FIXME: Once we grow proper nuclear flip support out of this we need
+	 * to compute the mask of flip planes precisely. For the time being
+	 * consider this a flip from a NULL plane.
+	 */
+	intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_ALL_MASK(pipe));
 }
 
 static void intel_crtc_disable_planes(struct drm_crtc *crtc)
@@ -3917,7 +4070,6 @@ static void intel_crtc_disable_planes(struct drm_crtc *crtc)
 	int plane = intel_crtc->plane;
 
 	intel_crtc_wait_for_pending_flips(crtc);
-	drm_crtc_vblank_off(crtc);
 
 	if (dev_priv->fbc.plane == plane)
 		intel_disable_fbc(dev);
@@ -3927,7 +4079,18 @@ static void intel_crtc_disable_planes(struct drm_crtc *crtc)
 	intel_crtc_dpms_overlay(intel_crtc, false);
 	intel_crtc_update_cursor(crtc, false);
 	intel_disable_planes(crtc);
-	intel_disable_primary_hw_plane(dev_priv, plane, pipe);
+	intel_disable_primary_hw_plane(crtc->primary, crtc);
+
+	/*
+	 * FIXME: Once we grow proper nuclear flip support out of this we need
+	 * to compute the mask of flip planes precisely. For the time being
+	 * consider this a flip to a NULL plane.
+	 */
+	intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_ALL_MASK(pipe));
+
+	drm_vblank_off(dev, pipe);
+
+	assert_vblank_disabled(crtc);
 }
 
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
@@ -3937,7 +4100,6 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	enum plane plane = intel_crtc->plane;
 
 	WARN_ON(!crtc->enabled);
 
@@ -3954,18 +4116,11 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 
 	if (intel_crtc->config.has_pch_encoder) {
 		intel_cpu_transcoder_set_m_n(intel_crtc,
-					     &intel_crtc->config.fdi_m_n);
+				     &intel_crtc->config.fdi_m_n, NULL);
 	}
 
 	ironlake_set_pipeconf(crtc);
 
-	/* Set up the display plane register */
-	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
 	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
@@ -4006,8 +4161,6 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		cpt_verify_modeset(dev, intel_crtc->pipe);
 
 	intel_crtc_enable_planes(crtc);
-
-	drm_crtc_vblank_on(crtc);
 }
 
 /* IPS only exists on ULT machines and is tied to pipe A. */
@@ -4052,47 +4205,46 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	enum plane plane = intel_crtc->plane;
 
 	WARN_ON(!crtc->enabled);
 
 	if (intel_crtc->active)
 		return;
 
+	if (intel_crtc_to_shared_dpll(intel_crtc))
+		intel_enable_shared_dpll(intel_crtc);
+
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_set_pipe_timings(intel_crtc);
 
+	if (intel_crtc->config.cpu_transcoder != TRANSCODER_EDP) {
+		I915_WRITE(PIPE_MULT(intel_crtc->config.cpu_transcoder),
+			   intel_crtc->config.pixel_multiplier - 1);
+	}
+
 	if (intel_crtc->config.has_pch_encoder) {
 		intel_cpu_transcoder_set_m_n(intel_crtc,
-					     &intel_crtc->config.fdi_m_n);
+				     &intel_crtc->config.fdi_m_n, NULL);
 	}
 
 	haswell_set_pipeconf(crtc);
 
 	intel_set_pipe_csc(crtc);
 
-	/* Set up the display plane register */
-	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE | DISPPLANE_PIPE_CSC_ENABLE);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
 	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
-	if (intel_crtc->config.has_pch_encoder)
-		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
-
-	if (intel_crtc->config.has_pch_encoder)
-		dev_priv->display.fdi_link_train(crtc);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
+	if (intel_crtc->config.has_pch_encoder) {
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
+		dev_priv->display.fdi_link_train(crtc);
+	}
+
 	intel_ddi_enable_pipe_clock(intel_crtc);
 
 	ironlake_pfit_enable(intel_crtc);
@@ -4112,6 +4264,9 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	if (intel_crtc->config.has_pch_encoder)
 		lpt_pch_enable(crtc);
 
+	if (intel_crtc->config.dp_encoder_is_mst)
+		intel_ddi_set_vc_payload_alloc(crtc, true);
+
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		encoder->enable(encoder);
 		intel_opregion_notify_encoder(encoder, true);
@@ -4121,8 +4276,6 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	 * to change the workaround. */
 	haswell_mode_set_planes_workaround(intel_crtc);
 	intel_crtc_enable_planes(crtc);
-
-	drm_crtc_vblank_on(crtc);
 }
 
 static void ironlake_pfit_disable(struct intel_crtc *crtc)
@@ -4160,7 +4313,7 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 	if (intel_crtc->config.has_pch_encoder)
 		intel_set_pch_fifo_underrun_reporting(dev, pipe, false);
 
-	intel_disable_pipe(dev_priv, pipe);
+	intel_disable_pipe(intel_crtc);
 
 	ironlake_pfit_disable(intel_crtc);
 
@@ -4200,7 +4353,6 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 	mutex_lock(&dev->struct_mutex);
 	intel_update_fbc(dev);
-	intel_edp_psr_update(dev);
 	mutex_unlock(&dev->struct_mutex);
 }
 
@@ -4210,7 +4362,6 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
-	int pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
 	if (!intel_crtc->active)
@@ -4225,7 +4376,10 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	if (intel_crtc->config.has_pch_encoder)
 		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, false);
-	intel_disable_pipe(dev_priv, pipe);
+	intel_disable_pipe(intel_crtc);
+
+	if (intel_crtc->config.dp_encoder_is_mst)
+		intel_ddi_set_vc_payload_alloc(crtc, false);
 
 	intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
@@ -4233,23 +4387,25 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	intel_ddi_disable_pipe_clock(intel_crtc);
 
-	for_each_encoder_on_crtc(dev, crtc, encoder)
-		if (encoder->post_disable)
-			encoder->post_disable(encoder);
-
 	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
 		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
 		intel_ddi_fdi_disable(crtc);
 	}
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->post_disable)
+			encoder->post_disable(encoder);
+
 	intel_crtc->active = false;
 	intel_update_watermarks(crtc);
 
 	mutex_lock(&dev->struct_mutex);
 	intel_update_fbc(dev);
-	intel_edp_psr_update(dev);
 	mutex_unlock(&dev->struct_mutex);
+
+	if (intel_crtc_to_shared_dpll(intel_crtc))
+		intel_disable_shared_dpll(intel_crtc);
 }
 
 static void ironlake_crtc_off(struct drm_crtc *crtc)
@@ -4258,10 +4414,6 @@ static void ironlake_crtc_off(struct drm_crtc *crtc)
 	intel_put_shared_dpll(intel_crtc);
 }
 
-static void haswell_crtc_off(struct drm_crtc *crtc)
-{
-	intel_ddi_put_crtc_pll(crtc);
-}
 
 static void i9xx_pfit_enable(struct intel_crtc *crtc)
 {
@@ -4287,6 +4439,23 @@ static void i9xx_pfit_enable(struct intel_crtc *crtc)
 	I915_WRITE(BCLRPAT(crtc->pipe), 0);
 }
 
+static enum intel_display_power_domain port_to_power_domain(enum port port)
+{
+	switch (port) {
+	case PORT_A:
+		return POWER_DOMAIN_PORT_DDI_A_4_LANES;
+	case PORT_B:
+		return POWER_DOMAIN_PORT_DDI_B_4_LANES;
+	case PORT_C:
+		return POWER_DOMAIN_PORT_DDI_C_4_LANES;
+	case PORT_D:
+		return POWER_DOMAIN_PORT_DDI_D_4_LANES;
+	default:
+		WARN_ON_ONCE(1);
+		return POWER_DOMAIN_PORT_OTHER;
+	}
+}
+
 #define for_each_power_domain(domain, mask)				\
 	for ((domain) = 0; (domain) < POWER_DOMAIN_NUM; (domain)++)	\
 		if ((1 << (domain)) & (mask))
@@ -4305,19 +4474,10 @@ intel_display_port_power_domain(struct intel_encoder *intel_encoder)
 	case INTEL_OUTPUT_HDMI:
 	case INTEL_OUTPUT_EDP:
 		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
-		switch (intel_dig_port->port) {
-		case PORT_A:
-			return POWER_DOMAIN_PORT_DDI_A_4_LANES;
-		case PORT_B:
-			return POWER_DOMAIN_PORT_DDI_B_4_LANES;
-		case PORT_C:
-			return POWER_DOMAIN_PORT_DDI_C_4_LANES;
-		case PORT_D:
-			return POWER_DOMAIN_PORT_DDI_D_4_LANES;
-		default:
-			WARN_ON_ONCE(1);
-			return POWER_DOMAIN_PORT_OTHER;
-		}
+		return port_to_power_domain(intel_dig_port->port);
+	case INTEL_OUTPUT_DP_MST:
+		intel_dig_port = enc_to_mst(&intel_encoder->base)->primary;
+		return port_to_power_domain(intel_dig_port->port);
 	case INTEL_OUTPUT_ANALOG:
 		return POWER_DOMAIN_PORT_CRT;
 	case INTEL_OUTPUT_DSI:
@@ -4333,7 +4493,6 @@ static unsigned long get_crtc_power_domains(struct drm_crtc *crtc)
 	struct intel_encoder *intel_encoder;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
-	bool pfit_enabled = intel_crtc->config.pch_pfit.enabled;
 	unsigned long mask;
 	enum transcoder transcoder;
 
@@ -4341,7 +4500,8 @@ static unsigned long get_crtc_power_domains(struct drm_crtc *crtc)
 
 	mask = BIT(POWER_DOMAIN_PIPE(pipe));
 	mask |= BIT(POWER_DOMAIN_TRANSCODER(transcoder));
-	if (pfit_enabled)
+	if (intel_crtc->config.pch_pfit.enabled ||
+	    intel_crtc->config.pch_pfit.force_thru)
 		mask |= BIT(POWER_DOMAIN_PIPE_PANEL_FITTER(pipe));
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder)
@@ -4398,7 +4558,8 @@ static void modeset_update_crtc_power_domains(struct drm_device *dev)
 	intel_display_set_init_power(dev_priv, false);
 }
 
-int valleyview_get_vco(struct drm_i915_private *dev_priv)
+/* returns HPLL frequency in kHz */
+static int valleyview_get_vco(struct drm_i915_private *dev_priv)
 {
 	int hpll_freq, vco_freq[] = { 800, 1600, 2000, 2400 };
 
@@ -4408,7 +4569,23 @@ int valleyview_get_vco(struct drm_i915_private *dev_priv)
 		CCK_FUSE_HPLL_FREQ_MASK;
 	mutex_unlock(&dev_priv->dpio_lock);
 
-	return vco_freq[hpll_freq];
+	return vco_freq[hpll_freq] * 1000;
+}
+
+static void vlv_update_cdclk(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->vlv_cdclk_freq = dev_priv->display.get_display_clock_speed(dev);
+	DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz",
+			 dev_priv->vlv_cdclk_freq);
+
+	/*
+	 * Program the gmbus_freq based on the cdclk frequency.
+	 * BSpec erroneously claims we should aim for 4MHz, but
+	 * in fact 1MHz is the correct frequency.
+	 */
+	I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->vlv_cdclk_freq, 1000));
 }
 
 /* Adjust CDclk dividers to allow high res or save power if possible */
@@ -4417,12 +4594,11 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val, cmd;
 
-	WARN_ON(valleyview_cur_cdclk(dev_priv) != dev_priv->vlv_cdclk_freq);
-	dev_priv->vlv_cdclk_freq = cdclk;
+	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->vlv_cdclk_freq);
 
-	if (cdclk >= 320) /* jump to highest voltage for 400MHz too */
+	if (cdclk >= 320000) /* jump to highest voltage for 400MHz too */
 		cmd = 2;
-	else if (cdclk == 266)
+	else if (cdclk == 266667)
 		cmd = 1;
 	else
 		cmd = 0;
@@ -4439,18 +4615,23 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	if (cdclk == 400) {
+	if (cdclk == 400000) {
 		u32 divider, vco;
 
 		vco = valleyview_get_vco(dev_priv);
-		divider = ((vco << 1) / cdclk) - 1;
+		divider = DIV_ROUND_CLOSEST(vco << 1, cdclk) - 1;
 
 		mutex_lock(&dev_priv->dpio_lock);
 		/* adjust cdclk divider */
 		val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
-		val &= ~0xf;
+		val &= ~DISPLAY_FREQUENCY_VALUES;
 		val |= divider;
 		vlv_cck_write(dev_priv, CCK_DISPLAY_CLOCK_CONTROL, val);
+
+		if (wait_for((vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL) &
+			      DISPLAY_FREQUENCY_STATUS) == (divider << DISPLAY_FREQUENCY_STATUS_SHIFT),
+			     50))
+			DRM_ERROR("timed out waiting for CDclk change\n");
 		mutex_unlock(&dev_priv->dpio_lock);
 	}
 
@@ -4463,54 +4644,88 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	 * For high bandwidth configs, we set a higher latency in the bunit
 	 * so that the core display fetch happens in time to avoid underruns.
 	 */
-	if (cdclk == 400)
+	if (cdclk == 400000)
 		val |= 4500 / 250; /* 4.5 usec */
 	else
 		val |= 3000 / 250; /* 3.0 usec */
 	vlv_bunit_write(dev_priv, BUNIT_REG_BISOC, val);
 	mutex_unlock(&dev_priv->dpio_lock);
 
-	/* Since we changed the CDclk, we need to update the GMBUSFREQ too */
-	intel_i2c_reset(dev);
+	vlv_update_cdclk(dev);
 }
 
-int valleyview_cur_cdclk(struct drm_i915_private *dev_priv)
+static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
 {
-	int cur_cdclk, vco;
-	int divider;
-
-	vco = valleyview_get_vco(dev_priv);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val, cmd;
 
-	mutex_lock(&dev_priv->dpio_lock);
-	divider = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
-	mutex_unlock(&dev_priv->dpio_lock);
+	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->vlv_cdclk_freq);
 
-	divider &= 0xf;
+	switch (cdclk) {
+	case 400000:
+		cmd = 3;
+		break;
+	case 333333:
+	case 320000:
+		cmd = 2;
+		break;
+	case 266667:
+		cmd = 1;
+		break;
+	case 200000:
+		cmd = 0;
+		break;
+	default:
+		WARN_ON(1);
+		return;
+	}
 
-	cur_cdclk = (vco << 1) / (divider + 1);
+	mutex_lock(&dev_priv->rps.hw_lock);
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	val &= ~DSPFREQGUAR_MASK_CHV;
+	val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
+		      DSPFREQSTAT_MASK_CHV) == (cmd << DSPFREQSTAT_SHIFT_CHV),
+		     50)) {
+		DRM_ERROR("timed out waiting for CDclk change\n");
+	}
+	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	return cur_cdclk;
+	vlv_update_cdclk(dev);
 }
 
 static int valleyview_calc_cdclk(struct drm_i915_private *dev_priv,
 				 int max_pixclk)
 {
+	int vco = valleyview_get_vco(dev_priv);
+	int freq_320 = (vco <<  1) % 320000 != 0 ? 333333 : 320000;
+
+	/* FIXME: Punit isn't quite ready yet */
+	if (IS_CHERRYVIEW(dev_priv->dev))
+		return 400000;
+
 	/*
 	 * Really only a few cases to deal with, as only 4 CDclks are supported:
 	 *   200MHz
 	 *   267MHz
-	 *   320MHz
+	 *   320/333MHz (depends on HPLL freq)
 	 *   400MHz
 	 * So we check to see whether we're above 90% of the lower bin and
 	 * adjust if needed.
+	 *
+	 * We seem to get an unstable or solid color picture at 200MHz.
+	 * Not sure what's wrong. For now use 200MHz only when all pipes
+	 * are off.
 	 */
-	if (max_pixclk > 288000) {
-		return 400;
-	} else if (max_pixclk > 240000) {
-		return 320;
-	} else
-		return 266;
-	/* Looks like the 200MHz CDclk freq doesn't work on some configs */
+	if (max_pixclk > freq_320*9/10)
+		return 400000;
+	else if (max_pixclk > 266667*9/10)
+		return freq_320;
+	else if (max_pixclk > 0)
+		return 266667;
+	else
+		return 200000;
 }
 
 /* compute the max pixel clock for new configuration */
@@ -4552,21 +4767,23 @@ static void valleyview_modeset_global_resources(struct drm_device *dev)
 	int max_pixclk = intel_mode_max_pixclk(dev_priv);
 	int req_cdclk = valleyview_calc_cdclk(dev_priv, max_pixclk);
 
-	if (req_cdclk != dev_priv->vlv_cdclk_freq)
-		valleyview_set_cdclk(dev, req_cdclk);
+	if (req_cdclk != dev_priv->vlv_cdclk_freq) {
+		if (IS_CHERRYVIEW(dev))
+			cherryview_set_cdclk(dev, req_cdclk);
+		else
+			valleyview_set_cdclk(dev, req_cdclk);
+	}
+
 	modeset_update_crtc_power_domains(dev);
 }
 
 static void valleyview_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
 	bool is_dsi;
-	u32 dspcntr;
 
 	WARN_ON(!crtc->enabled);
 
@@ -4575,33 +4792,20 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 
 	is_dsi = intel_pipe_has_type(crtc, INTEL_OUTPUT_DSI);
 
-	if (!is_dsi && !IS_CHERRYVIEW(dev))
-		vlv_prepare_pll(intel_crtc);
-
-	/* Set up the display plane register */
-	dspcntr = DISPPLANE_GAMMA_ENABLE;
+	if (!is_dsi) {
+		if (IS_CHERRYVIEW(dev))
+			chv_prepare_pll(intel_crtc);
+		else
+			vlv_prepare_pll(intel_crtc);
+	}
 
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_set_pipe_timings(intel_crtc);
 
-	/* pipesrc and dspsize control the size that is scaled from,
-	 * which should always be the user's requested size.
-	 */
-	I915_WRITE(DSPSIZE(plane),
-		   ((intel_crtc->config.pipe_src_h - 1) << 16) |
-		   (intel_crtc->config.pipe_src_w - 1));
-	I915_WRITE(DSPPOS(plane), 0);
-
 	i9xx_set_pipeconf(intel_crtc);
 
-	I915_WRITE(DSPCNTR(plane), dspcntr);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
 	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
@@ -4633,8 +4837,6 @@ static void valleyview_crtc_enable(struct drm_crtc *crtc)
 
 	intel_crtc_enable_planes(crtc);
 
-	drm_crtc_vblank_on(crtc);
-
 	/* Underruns don't raise interrupts, so check manually. */
 	i9xx_check_fifo_underruns(dev);
 }
@@ -4651,12 +4853,9 @@ static void i9xx_set_pll_dividers(struct intel_crtc *crtc)
 static void i9xx_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	int pipe = intel_crtc->pipe;
-	int plane = intel_crtc->plane;
-	u32 dspcntr;
 
 	WARN_ON(!crtc->enabled);
 
@@ -4665,35 +4864,13 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 
 	i9xx_set_pll_dividers(intel_crtc);
 
-	/* Set up the display plane register */
-	dspcntr = DISPPLANE_GAMMA_ENABLE;
-
-	if (pipe == 0)
-		dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
-	else
-		dspcntr |= DISPPLANE_SEL_PIPE_B;
-
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_set_pipe_timings(intel_crtc);
 
-	/* pipesrc and dspsize control the size that is scaled from,
-	 * which should always be the user's requested size.
-	 */
-	I915_WRITE(DSPSIZE(plane),
-		   ((intel_crtc->config.pipe_src_h - 1) << 16) |
-		   (intel_crtc->config.pipe_src_w - 1));
-	I915_WRITE(DSPPOS(plane), 0);
-
 	i9xx_set_pipeconf(intel_crtc);
 
-	I915_WRITE(DSPCNTR(plane), dspcntr);
-	POSTING_READ(DSPCNTR(plane));
-
-	dev_priv->display.update_primary_plane(crtc, crtc->primary->fb,
-					       crtc->x, crtc->y);
-
 	intel_crtc->active = true;
 
 	if (!IS_GEN2(dev))
@@ -4727,8 +4904,6 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 	if (IS_GEN2(dev))
 		intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
 
-	drm_crtc_vblank_on(crtc);
-
 	/* Underruns don't raise interrupts, so check manually. */
 	i9xx_check_fifo_underruns(dev);
 }
@@ -4768,6 +4943,16 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	if (IS_GEN2(dev))
 		intel_set_cpu_fifo_underrun_reporting(dev, pipe, false);
 
+	/*
+	 * Vblank time updates from the shadow to live plane control register
+	 * are blocked if the memory self-refresh mode is active at that
+	 * moment. So to make sure the plane gets truly disabled, disable
+	 * first the self-refresh mode. The self-refresh enable bit in turn
+	 * will be checked/applied by the HW only at the next frame start
+	 * event which is after the vblank start event, so we need to have a
+	 * wait-for-vblank between disabling the plane and the pipe.
+	 */
+	intel_set_memory_cxsr(dev_priv, false);
 	intel_crtc_disable_planes(crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
@@ -4776,11 +4961,12 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	/*
 	 * On gen2 planes are double buffered but the pipe isn't, so we must
 	 * wait for planes to fully turn off before disabling the pipe.
+	 * We also need to wait on all gmch platforms because of the
+	 * self-refresh mode constraint explained above.
 	 */
-	if (IS_GEN2(dev))
-		intel_wait_for_vblank(dev, pipe);
+	intel_wait_for_vblank(dev, pipe);
 
-	intel_disable_pipe(dev_priv, pipe);
+	intel_disable_pipe(intel_crtc);
 
 	i9xx_pfit_disable(intel_crtc);
 
@@ -4794,7 +4980,7 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 		else if (IS_VALLEYVIEW(dev))
 			vlv_disable_pll(dev_priv, pipe);
 		else
-			i9xx_disable_pll(dev_priv, pipe);
+			i9xx_disable_pll(intel_crtc);
 	}
 
 	if (!IS_GEN2(dev))
@@ -4805,7 +4991,6 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 
 	mutex_lock(&dev->struct_mutex);
 	intel_update_fbc(dev);
-	intel_edp_psr_update(dev);
 	mutex_unlock(&dev->struct_mutex);
 }
 
@@ -4843,23 +5028,49 @@ static void intel_crtc_update_sarea(struct drm_crtc *crtc,
 	}
 }
 
+/* Master function to enable/disable CRTC and corresponding power wells */
+void intel_crtc_control(struct drm_crtc *crtc, bool enable)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum intel_display_power_domain domain;
+	unsigned long domains;
+
+	if (enable) {
+		if (!intel_crtc->active) {
+			domains = get_crtc_power_domains(crtc);
+			for_each_power_domain(domain, domains)
+				intel_display_power_get(dev_priv, domain);
+			intel_crtc->enabled_power_domains = domains;
+
+			dev_priv->display.crtc_enable(crtc);
+		}
+	} else {
+		if (intel_crtc->active) {
+			dev_priv->display.crtc_disable(crtc);
+
+			domains = intel_crtc->enabled_power_domains;
+			for_each_power_domain(domain, domains)
+				intel_display_power_put(dev_priv, domain);
+			intel_crtc->enabled_power_domains = 0;
+		}
+	}
+}
+
 /**
  * Sets the power management mode of the pipe and plane.
  */
 void intel_crtc_update_dpms(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *intel_encoder;
 	bool enable = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder)
 		enable |= intel_encoder->connectors_active;
 
-	if (enable)
-		dev_priv->display.crtc_enable(crtc);
-	else
-		dev_priv->display.crtc_disable(crtc);
+	intel_crtc_control(crtc, enable);
 
 	intel_crtc_update_sarea(crtc, enable);
 }
@@ -4869,6 +5080,8 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_connector *connector;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *old_obj = intel_fb_obj(crtc->primary->fb);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
 
 	/* crtc should still be enabled when we disable it. */
 	WARN_ON(!crtc->enabled);
@@ -4877,13 +5090,11 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	intel_crtc_update_sarea(crtc, false);
 	dev_priv->display.off(crtc);
 
-	assert_plane_disabled(dev->dev_private, to_intel_crtc(crtc)->plane);
-	assert_cursor_disabled(dev_priv, to_intel_crtc(crtc)->pipe);
-	assert_pipe_disabled(dev->dev_private, to_intel_crtc(crtc)->pipe);
-
 	if (crtc->primary->fb) {
 		mutex_lock(&dev->struct_mutex);
-		intel_unpin_fb_obj(to_intel_framebuffer(crtc->primary->fb)->obj);
+		intel_unpin_fb_obj(old_obj);
+		i915_gem_track_fb(old_obj, NULL,
+				  INTEL_FRONTBUFFER_PRIMARY(pipe));
 		mutex_unlock(&dev->struct_mutex);
 		crtc->primary->fb = NULL;
 	}
@@ -4939,24 +5150,31 @@ static void intel_connector_check_state(struct intel_connector *connector)
 			      connector->base.base.id,
 			      connector->base.name);
 
+		/* there is no real hw state for MST connectors */
+		if (connector->mst_port)
+			return;
+
 		WARN(connector->base.dpms == DRM_MODE_DPMS_OFF,
 		     "wrong connector dpms state\n");
 		WARN(connector->base.encoder != &encoder->base,
 		     "active connector not linked to encoder\n");
-		WARN(!encoder->connectors_active,
-		     "encoder->connectors_active not set\n");
 
-		encoder_enabled = encoder->get_hw_state(encoder, &pipe);
-		WARN(!encoder_enabled, "encoder not enabled\n");
-		if (WARN_ON(!encoder->base.crtc))
-			return;
+		if (encoder) {
+			WARN(!encoder->connectors_active,
+			     "encoder->connectors_active not set\n");
 
-		crtc = encoder->base.crtc;
+			encoder_enabled = encoder->get_hw_state(encoder, &pipe);
+			WARN(!encoder_enabled, "encoder not enabled\n");
+			if (WARN_ON(!encoder->base.crtc))
+				return;
 
-		WARN(!crtc->enabled, "crtc not enabled\n");
-		WARN(!to_intel_crtc(crtc)->active, "crtc not active\n");
-		WARN(pipe != to_intel_crtc(crtc)->pipe,
-		     "encoder active on the wrong pipe\n");
+			crtc = encoder->base.crtc;
+
+			WARN(!crtc->enabled, "crtc not enabled\n");
+			WARN(!to_intel_crtc(crtc)->active, "crtc not active\n");
+			WARN(pipe != to_intel_crtc(crtc)->pipe,
+			     "encoder active on the wrong pipe\n");
+		}
 	}
 }
 
@@ -5161,9 +5379,11 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	if (HAS_IPS(dev))
 		hsw_compute_ips_config(crtc, pipe_config);
 
-	/* XXX: PCH clock sharing is done in ->mode_set, so make sure the old
-	 * clock survives for now. */
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+	/*
+	 * XXX: PCH/WRPLL clock sharing is done in ->mode_set, so make sure the
+	 * old clock survives for now.
+	 */
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev) || HAS_DDI(dev))
 		pipe_config->shared_dpll = crtc->config.shared_dpll;
 
 	if (pipe_config->has_pch_encoder)
@@ -5174,7 +5394,26 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 
 static int valleyview_get_display_clock_speed(struct drm_device *dev)
 {
-	return 400000; /* FIXME */
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int vco = valleyview_get_vco(dev_priv);
+	u32 val;
+	int divider;
+
+	/* FIXME: Punit isn't quite ready yet */
+	if (IS_CHERRYVIEW(dev))
+		return 400000;
+
+	mutex_lock(&dev_priv->dpio_lock);
+	val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	divider = val & DISPLAY_FREQUENCY_VALUES;
+
+	WARN((val & DISPLAY_FREQUENCY_STATUS) !=
+	     (divider << DISPLAY_FREQUENCY_STATUS_SHIFT),
+	     "cdclk change in progress\n");
+
+	return DIV_ROUND_CLOSEST(vco << 1, divider + 1);
 }
 
 static int i945_get_display_clock_speed(struct drm_device *dev)
@@ -5408,7 +5647,8 @@ static void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
 }
 
 static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n)
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -5420,6 +5660,18 @@ static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
 		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
 		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
 		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
+		/* M2_N2 registers to be set only for gen < 8 (M2_N2 available
+		 * for gen < 8) and if DRRS is supported (to make sure the
+		 * registers are not unnecessarily accessed).
+		 */
+		if (m2_n2 && INTEL_INFO(dev)->gen < 8 &&
+			crtc->config.has_drrs) {
+			I915_WRITE(PIPE_DATA_M2(transcoder),
+					TU_SIZE(m2_n2->tu) | m2_n2->gmch_m);
+			I915_WRITE(PIPE_DATA_N2(transcoder), m2_n2->gmch_n);
+			I915_WRITE(PIPE_LINK_M2(transcoder), m2_n2->link_m);
+			I915_WRITE(PIPE_LINK_N2(transcoder), m2_n2->link_n);
+		}
 	} else {
 		I915_WRITE(PIPE_DATA_M_G4X(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
 		I915_WRITE(PIPE_DATA_N_G4X(pipe), m_n->gmch_n);
@@ -5428,12 +5680,13 @@ static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
 	}
 }
 
-static void intel_dp_set_m_n(struct intel_crtc *crtc)
+void intel_dp_set_m_n(struct intel_crtc *crtc)
 {
 	if (crtc->config.has_pch_encoder)
 		intel_pch_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
 	else
-		intel_cpu_transcoder_set_m_n(crtc, &crtc->config.dp_m_n);
+		intel_cpu_transcoder_set_m_n(crtc, &crtc->config.dp_m_n,
+						   &crtc->config.dp_m2_n2);
 }
 
 static void vlv_update_pll(struct intel_crtc *crtc)
@@ -5551,6 +5804,18 @@ static void vlv_prepare_pll(struct intel_crtc *crtc)
 
 static void chv_update_pll(struct intel_crtc *crtc)
 {
+	crtc->config.dpll_hw_state.dpll = DPLL_SSC_REF_CLOCK_CHV |
+		DPLL_REFA_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS |
+		DPLL_VCO_ENABLE;
+	if (crtc->pipe != PIPE_A)
+		crtc->config.dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	crtc->config.dpll_hw_state.dpll_md =
+		(crtc->config.pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+static void chv_prepare_pll(struct intel_crtc *crtc)
+{
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe = crtc->pipe;
@@ -5560,15 +5825,6 @@ static void chv_update_pll(struct intel_crtc *crtc)
 	u32 bestn, bestm1, bestm2, bestp1, bestp2, bestm2_frac;
 	int refclk;
 
-	crtc->config.dpll_hw_state.dpll = DPLL_SSC_REF_CLOCK_CHV |
-		DPLL_REFA_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS |
-		DPLL_VCO_ENABLE;
-	if (pipe != PIPE_A)
-		crtc->config.dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
-
-	crtc->config.dpll_hw_state.dpll_md =
-		(crtc->config.pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
-
 	bestn = crtc->config.dpll.n;
 	bestm2_frac = crtc->config.dpll.m2 & 0x3fffff;
 	bestm1 = crtc->config.dpll.m1;
@@ -5728,7 +5984,7 @@ static void i8xx_update_pll(struct intel_crtc *crtc,
 			dpll |= PLL_P2_DIVIDE_BY_4;
 	}
 
-	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DVO))
+	if (!IS_I830(dev) && intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DVO))
 		dpll |= DPLL_DVO_2X_MODE;
 
 	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
@@ -5879,9 +6135,9 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 
 	pipeconf = 0;
 
-	if (dev_priv->quirks & QUIRK_PIPEA_FORCE &&
-	    I915_READ(PIPECONF(intel_crtc->pipe)) & PIPECONF_ENABLE)
-		pipeconf |= PIPECONF_ENABLE;
+	if ((intel_crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+	    (intel_crtc->pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
+		pipeconf |= I915_READ(PIPECONF(intel_crtc->pipe)) & PIPECONF_ENABLE;
 
 	if (intel_crtc->config.double_wide)
 		pipeconf |= PIPECONF_DOUBLE_WIDE;
@@ -6060,6 +6316,10 @@ static void vlv_crtc_clock_get(struct intel_crtc *crtc,
 	u32 mdiv;
 	int refclk = 100000;
 
+	/* In case of MIPI DPLL will not even be used */
+	if (!(pipe_config->dpll_hw_state.dpll & DPLL_VCO_ENABLE))
+		return;
+
 	mutex_lock(&dev_priv->dpio_lock);
 	mdiv = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW3(pipe));
 	mutex_unlock(&dev_priv->dpio_lock);
@@ -6120,13 +6380,13 @@ static void i9xx_get_plane_config(struct intel_crtc *crtc,
 	crtc->base.primary->fb->height = ((val >> 0) & 0xfff) + 1;
 
 	val = I915_READ(DSPSTRIDE(pipe));
-	crtc->base.primary->fb->pitches[0] = val & 0xffffff80;
+	crtc->base.primary->fb->pitches[0] = val & 0xffffffc0;
 
 	aligned_height = intel_align_height(dev, crtc->base.primary->fb->height,
 					    plane_config->tiled);
 
-	plane_config->size = ALIGN(crtc->base.primary->fb->pitches[0] *
-				   aligned_height, PAGE_SIZE);
+	plane_config->size = PAGE_ALIGN(crtc->base.primary->fb->pitches[0] *
+					aligned_height);
 
 	DRM_DEBUG_KMS("pipe/plane %d/%d with fb: size=%dx%d@%d, offset=%x, pitch %d, size 0x%x\n",
 		      pipe, plane, crtc->base.primary->fb->width,
@@ -6230,6 +6490,14 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 	}
 	pipe_config->dpll_hw_state.dpll = I915_READ(DPLL(crtc->pipe));
 	if (!IS_VALLEYVIEW(dev)) {
+		/*
+		 * DPLL_DVO_2X_MODE must be enabled for both DPLLs
+		 * on 830. Filter it out here so that we don't
+		 * report errors due to that.
+		 */
+		if (IS_I830(dev))
+			pipe_config->dpll_hw_state.dpll &= ~DPLL_DVO_2X_MODE;
+
 		pipe_config->dpll_hw_state.fp0 = I915_READ(FP0(crtc->pipe));
 		pipe_config->dpll_hw_state.fp1 = I915_READ(FP1(crtc->pipe));
 	} else {
@@ -6252,7 +6520,6 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 static void ironlake_init_pch_refclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 	u32 val, final;
 	bool has_lvds = false;
@@ -6262,8 +6529,7 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 	bool can_ssc = false;
 
 	/* We need to take the global config into account */
-	list_for_each_entry(encoder, &mode_config->encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		switch (encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			has_panel = true;
@@ -6570,11 +6836,10 @@ static void lpt_disable_clkout_dp(struct drm_device *dev)
 
 static void lpt_init_pch_refclk(struct drm_device *dev)
 {
-	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct intel_encoder *encoder;
 	bool has_vga = false;
 
-	list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		switch (encoder->type) {
 		case INTEL_OUTPUT_ANALOG:
 			has_vga = true;
@@ -7030,7 +7295,8 @@ static void intel_pch_transcoder_get_m_n(struct intel_crtc *crtc,
 
 static void intel_cpu_transcoder_get_m_n(struct intel_crtc *crtc,
 					 enum transcoder transcoder,
-					 struct intel_link_m_n *m_n)
+					 struct intel_link_m_n *m_n,
+					 struct intel_link_m_n *m2_n2)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -7044,6 +7310,20 @@ static void intel_cpu_transcoder_get_m_n(struct intel_crtc *crtc,
 		m_n->gmch_n = I915_READ(PIPE_DATA_N1(transcoder));
 		m_n->tu = ((I915_READ(PIPE_DATA_M1(transcoder))
 			    & TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+		/* Read M2_N2 registers only for gen < 8 (M2_N2 available for
+		 * gen < 8) and if DRRS is supported (to make sure the
+		 * registers are not unnecessarily read).
+		 */
+		if (m2_n2 && INTEL_INFO(dev)->gen < 8 &&
+			crtc->config.has_drrs) {
+			m2_n2->link_m = I915_READ(PIPE_LINK_M2(transcoder));
+			m2_n2->link_n =	I915_READ(PIPE_LINK_N2(transcoder));
+			m2_n2->gmch_m =	I915_READ(PIPE_DATA_M2(transcoder))
+					& ~TU_SIZE_MASK;
+			m2_n2->gmch_n =	I915_READ(PIPE_DATA_N2(transcoder));
+			m2_n2->tu = ((I915_READ(PIPE_DATA_M2(transcoder))
+					& TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+		}
 	} else {
 		m_n->link_m = I915_READ(PIPE_LINK_M_G4X(pipe));
 		m_n->link_n = I915_READ(PIPE_LINK_N_G4X(pipe));
@@ -7062,14 +7342,15 @@ void intel_dp_get_m_n(struct intel_crtc *crtc,
 		intel_pch_transcoder_get_m_n(crtc, &pipe_config->dp_m_n);
 	else
 		intel_cpu_transcoder_get_m_n(crtc, pipe_config->cpu_transcoder,
-					     &pipe_config->dp_m_n);
+					     &pipe_config->dp_m_n,
+					     &pipe_config->dp_m2_n2);
 }
 
 static void ironlake_get_fdi_m_n_config(struct intel_crtc *crtc,
 					struct intel_crtc_config *pipe_config)
 {
 	intel_cpu_transcoder_get_m_n(crtc, pipe_config->cpu_transcoder,
-				     &pipe_config->fdi_m_n);
+				     &pipe_config->fdi_m_n, NULL);
 }
 
 static void ironlake_get_pfit_config(struct intel_crtc *crtc,
@@ -7140,13 +7421,13 @@ static void ironlake_get_plane_config(struct intel_crtc *crtc,
 	crtc->base.primary->fb->height = ((val >> 0) & 0xfff) + 1;
 
 	val = I915_READ(DSPSTRIDE(pipe));
-	crtc->base.primary->fb->pitches[0] = val & 0xffffff80;
+	crtc->base.primary->fb->pitches[0] = val & 0xffffffc0;
 
 	aligned_height = intel_align_height(dev, crtc->base.primary->fb->height,
 					    plane_config->tiled);
 
-	plane_config->size = ALIGN(crtc->base.primary->fb->pitches[0] *
-				   aligned_height, PAGE_SIZE);
+	plane_config->size = PAGE_ALIGN(crtc->base.primary->fb->pitches[0] *
+					aligned_height);
 
 	DRM_DEBUG_KMS("pipe/plane %d/%d with fb: size=%dx%d@%d, offset=%x, pitch %d, size 0x%x\n",
 		      pipe, plane, crtc->base.primary->fb->width,
@@ -7163,6 +7444,10 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	if (!intel_display_power_enabled(dev_priv,
+					 POWER_DOMAIN_PIPE(crtc->pipe)))
+		return false;
+
 	pipe_config->cpu_transcoder = (enum transcoder) crtc->pipe;
 	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
 
@@ -7237,7 +7522,6 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
-	struct intel_ddi_plls *plls = &dev_priv->ddi_plls;
 	struct intel_crtc *crtc;
 
 	for_each_intel_crtc(dev, crtc)
@@ -7245,14 +7529,15 @@ static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
 		     pipe_name(crtc->pipe));
 
 	WARN(I915_READ(HSW_PWR_WELL_DRIVER), "Power well on\n");
-	WARN(plls->spll_refcount, "SPLL enabled\n");
-	WARN(plls->wrpll1_refcount, "WRPLL1 enabled\n");
-	WARN(plls->wrpll2_refcount, "WRPLL2 enabled\n");
+	WARN(I915_READ(SPLL_CTL) & SPLL_PLL_ENABLE, "SPLL enabled\n");
+	WARN(I915_READ(WRPLL_CTL1) & WRPLL_PLL_ENABLE, "WRPLL1 enabled\n");
+	WARN(I915_READ(WRPLL_CTL2) & WRPLL_PLL_ENABLE, "WRPLL2 enabled\n");
 	WARN(I915_READ(PCH_PP_STATUS) & PP_ON, "Panel power on\n");
 	WARN(I915_READ(BLC_PWM_CPU_CTL2) & BLM_PWM_ENABLE,
 	     "CPU PWM1 enabled\n");
-	WARN(I915_READ(HSW_BLC_PWM2_CTL) & BLM_PWM_ENABLE,
-	     "CPU PWM2 enabled\n");
+	if (IS_HASWELL(dev))
+		WARN(I915_READ(HSW_BLC_PWM2_CTL) & BLM_PWM_ENABLE,
+		     "CPU PWM2 enabled\n");
 	WARN(I915_READ(BLC_PWM_PCH_CTL1) & BLM_PCH_PWM_ENABLE,
 	     "PCH PWM1 enabled\n");
 	WARN(I915_READ(UTIL_PIN_CTL) & UTIL_PIN_ENABLE,
@@ -7265,7 +7550,17 @@ static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
 	 * gen-specific and since we only disable LCPLL after we fully disable
 	 * the interrupts, the check below should be enough.
 	 */
-	WARN(!dev_priv->pm.irqs_disabled, "IRQs enabled\n");
+	WARN(intel_irqs_enabled(dev_priv), "IRQs enabled\n");
+}
+
+static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+
+	if (IS_HASWELL(dev))
+		return I915_READ(D_COMP_HSW);
+	else
+		return I915_READ(D_COMP_BDW);
 }
 
 static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val)
@@ -7276,12 +7571,12 @@ static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val)
 		mutex_lock(&dev_priv->rps.hw_lock);
 		if (sandybridge_pcode_write(dev_priv, GEN6_PCODE_WRITE_D_COMP,
 					    val))
-			DRM_ERROR("Failed to disable D_COMP\n");
+			DRM_ERROR("Failed to write to D_COMP\n");
 		mutex_unlock(&dev_priv->rps.hw_lock);
 	} else {
-		I915_WRITE(D_COMP, val);
+		I915_WRITE(D_COMP_BDW, val);
+		POSTING_READ(D_COMP_BDW);
 	}
-	POSTING_READ(D_COMP);
 }
 
 /*
@@ -7319,12 +7614,13 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
 	if (wait_for((I915_READ(LCPLL_CTL) & LCPLL_PLL_LOCK) == 0, 1))
 		DRM_ERROR("LCPLL still locked\n");
 
-	val = I915_READ(D_COMP);
+	val = hsw_read_dcomp(dev_priv);
 	val |= D_COMP_COMP_DISABLE;
 	hsw_write_dcomp(dev_priv, val);
 	ndelay(100);
 
-	if (wait_for((I915_READ(D_COMP) & D_COMP_RCOMP_IN_PROGRESS) == 0, 1))
+	if (wait_for((hsw_read_dcomp(dev_priv) & D_COMP_RCOMP_IN_PROGRESS) == 0,
+		     1))
 		DRM_ERROR("D_COMP RCOMP still in progress\n");
 
 	if (allow_power_down) {
@@ -7373,7 +7669,7 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 		POSTING_READ(LCPLL_CTL);
 	}
 
-	val = I915_READ(D_COMP);
+	val = hsw_read_dcomp(dev_priv);
 	val |= D_COMP_COMP_FORCE;
 	val &= ~D_COMP_COMP_DISABLE;
 	hsw_write_dcomp(dev_priv, val);
@@ -7479,13 +7775,66 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 
 	if (!intel_ddi_pll_select(intel_crtc))
 		return -EINVAL;
-	intel_ddi_pll_enable(intel_crtc);
 
 	intel_crtc->lowfreq_avail = false;
 
 	return 0;
 }
 
+static void haswell_get_ddi_pll(struct drm_i915_private *dev_priv,
+				enum port port,
+				struct intel_crtc_config *pipe_config)
+{
+	pipe_config->ddi_pll_sel = I915_READ(PORT_CLK_SEL(port));
+
+	switch (pipe_config->ddi_pll_sel) {
+	case PORT_CLK_SEL_WRPLL1:
+		pipe_config->shared_dpll = DPLL_ID_WRPLL1;
+		break;
+	case PORT_CLK_SEL_WRPLL2:
+		pipe_config->shared_dpll = DPLL_ID_WRPLL2;
+		break;
+	}
+}
+
+static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
+				       struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_shared_dpll *pll;
+	enum port port;
+	uint32_t tmp;
+
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder));
+
+	port = (tmp & TRANS_DDI_PORT_MASK) >> TRANS_DDI_PORT_SHIFT;
+
+	haswell_get_ddi_pll(dev_priv, port, pipe_config);
+
+	if (pipe_config->shared_dpll >= 0) {
+		pll = &dev_priv->shared_dplls[pipe_config->shared_dpll];
+
+		WARN_ON(!pll->get_hw_state(dev_priv, pll,
+					   &pipe_config->dpll_hw_state));
+	}
+
+	/*
+	 * Haswell has only FDI/PCH transcoder A. It is which is connected to
+	 * DDI E. So just check whether this pipe is wired to DDI E and whether
+	 * the PCH transcoder is on.
+	 */
+	if ((port == PORT_E) && I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
+		pipe_config->has_pch_encoder = true;
+
+		tmp = I915_READ(FDI_RX_CTL(PIPE_A));
+		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+		ironlake_get_fdi_m_n_config(crtc, pipe_config);
+	}
+}
+
 static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 				    struct intel_crtc_config *pipe_config)
 {
@@ -7531,22 +7880,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
-	/*
-	 * Haswell has only FDI/PCH transcoder A. It is which is connected to
-	 * DDI E. So just check whether this pipe is wired to DDI E and whether
-	 * the PCH transcoder is on.
-	 */
-	tmp = I915_READ(TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder));
-	if ((tmp & TRANS_DDI_PORT_MASK) == TRANS_DDI_SELECT_PORT(PORT_E) &&
-	    I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
-		pipe_config->has_pch_encoder = true;
-
-		tmp = I915_READ(FDI_RX_CTL(PIPE_A));
-		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
-					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
-
-		ironlake_get_fdi_m_n_config(crtc, pipe_config);
-	}
+	haswell_get_ddi_port_state(crtc, pipe_config);
 
 	intel_get_pipe_timings(crtc, pipe_config);
 
@@ -7558,7 +7892,12 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 		pipe_config->ips_enabled = hsw_crtc_supports_ips(crtc) &&
 			(I915_READ(IPS_CTL) & IPS_ENABLE);
 
-	pipe_config->pixel_multiplier = 1;
+	if (pipe_config->cpu_transcoder != TRANSCODER_EDP) {
+		pipe_config->pixel_multiplier =
+			I915_READ(PIPE_MULT(pipe_config->cpu_transcoder)) + 1;
+	} else {
+		pipe_config->pixel_multiplier = 1;
+	}
 
 	return true;
 }
@@ -7876,74 +8215,62 @@ static void i845_update_cursor(struct drm_crtc *crtc, u32 base)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t cntl;
+	uint32_t cntl = 0, size = 0;
 
-	if (base != intel_crtc->cursor_base) {
-		/* On these chipsets we can only modify the base whilst
-		 * the cursor is disabled.
-		 */
-		if (intel_crtc->cursor_cntl) {
-			I915_WRITE(_CURACNTR, 0);
-			POSTING_READ(_CURACNTR);
-			intel_crtc->cursor_cntl = 0;
+	if (base) {
+		unsigned int width = intel_crtc->cursor_width;
+		unsigned int height = intel_crtc->cursor_height;
+		unsigned int stride = roundup_pow_of_two(width) * 4;
+
+		switch (stride) {
+		default:
+			WARN_ONCE(1, "Invalid cursor width/stride, width=%u, stride=%u\n",
+				  width, stride);
+			stride = 256;
+			/* fallthrough */
+		case 256:
+		case 512:
+		case 1024:
+		case 2048:
+			break;
 		}
 
-		I915_WRITE(_CURABASE, base);
-		POSTING_READ(_CURABASE);
+		cntl |= CURSOR_ENABLE |
+			CURSOR_GAMMA_ENABLE |
+			CURSOR_FORMAT_ARGB |
+			CURSOR_STRIDE(stride);
+
+		size = (height << 12) | width;
 	}
 
-	/* XXX width must be 64, stride 256 => 0x00 << 28 */
-	cntl = 0;
-	if (base)
-		cntl = (CURSOR_ENABLE |
-			CURSOR_GAMMA_ENABLE |
-			CURSOR_FORMAT_ARGB);
-	if (intel_crtc->cursor_cntl != cntl) {
-		I915_WRITE(_CURACNTR, cntl);
+	if (intel_crtc->cursor_cntl != 0 &&
+	    (intel_crtc->cursor_base != base ||
+	     intel_crtc->cursor_size != size ||
+	     intel_crtc->cursor_cntl != cntl)) {
+		/* On these chipsets we can only modify the base/size/stride
+		 * whilst the cursor is disabled.
+		 */
+		I915_WRITE(_CURACNTR, 0);
 		POSTING_READ(_CURACNTR);
-		intel_crtc->cursor_cntl = cntl;
+		intel_crtc->cursor_cntl = 0;
 	}
-}
 
-static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
-	uint32_t cntl;
+	if (intel_crtc->cursor_base != base)
+		I915_WRITE(_CURABASE, base);
 
-	cntl = 0;
-	if (base) {
-		cntl = MCURSOR_GAMMA_ENABLE;
-		switch (intel_crtc->cursor_width) {
-			case 64:
-				cntl |= CURSOR_MODE_64_ARGB_AX;
-				break;
-			case 128:
-				cntl |= CURSOR_MODE_128_ARGB_AX;
-				break;
-			case 256:
-				cntl |= CURSOR_MODE_256_ARGB_AX;
-				break;
-			default:
-				WARN_ON(1);
-				return;
-		}
-		cntl |= pipe << 28; /* Connect to correct pipe */
+	if (intel_crtc->cursor_size != size) {
+		I915_WRITE(CURSIZE, size);
+		intel_crtc->cursor_size = size;
 	}
+
 	if (intel_crtc->cursor_cntl != cntl) {
-		I915_WRITE(CURCNTR(pipe), cntl);
-		POSTING_READ(CURCNTR(pipe));
+		I915_WRITE(_CURACNTR, cntl);
+		POSTING_READ(_CURACNTR);
 		intel_crtc->cursor_cntl = cntl;
 	}
-
-	/* and commit changes on next vblank */
-	I915_WRITE(CURBASE(pipe), base);
-	POSTING_READ(CURBASE(pipe));
 }
 
-static void ivb_update_cursor(struct drm_crtc *crtc, u32 base)
+static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -7968,6 +8295,7 @@ static void ivb_update_cursor(struct drm_crtc *crtc, u32 base)
 				WARN_ON(1);
 				return;
 		}
+		cntl |= pipe << 28; /* Connect to correct pipe */
 	}
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		cntl |= CURSOR_PIPE_CSC_ENABLE;
@@ -7991,8 +8319,8 @@ static void intel_crtc_update_cursor(struct drm_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
-	int x = intel_crtc->cursor_x;
-	int y = intel_crtc->cursor_y;
+	int x = crtc->cursor_x;
+	int y = crtc->cursor_y;
 	u32 base = 0, pos = 0;
 
 	if (on)
@@ -8027,51 +8355,86 @@ static void intel_crtc_update_cursor(struct drm_crtc *crtc,
 
 	I915_WRITE(CURPOS(pipe), pos);
 
-	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_BROADWELL(dev))
-		ivb_update_cursor(crtc, base);
-	else if (IS_845G(dev) || IS_I865G(dev))
+	if (IS_845G(dev) || IS_I865G(dev))
 		i845_update_cursor(crtc, base);
 	else
 		i9xx_update_cursor(crtc, base);
 	intel_crtc->cursor_base = base;
 }
 
-static int intel_crtc_cursor_set(struct drm_crtc *crtc,
-				 struct drm_file *file,
-				 uint32_t handle,
-				 uint32_t width, uint32_t height)
+static bool cursor_size_ok(struct drm_device *dev,
+			   uint32_t width, uint32_t height)
+{
+	if (width == 0 || height == 0)
+		return false;
+
+	/*
+	 * 845g/865g are special in that they are only limited by
+	 * the width of their cursors, the height is arbitrary up to
+	 * the precision of the register. Everything else requires
+	 * square cursors, limited to a few power-of-two sizes.
+	 */
+	if (IS_845G(dev) || IS_I865G(dev)) {
+		if ((width & 63) != 0)
+			return false;
+
+		if (width > (IS_845G(dev) ? 64 : 512))
+			return false;
+
+		if (height > 1023)
+			return false;
+	} else {
+		switch (width | height) {
+		case 256:
+		case 128:
+			if (IS_GEN2(dev))
+				return false;
+		case 64:
+			break;
+		default:
+			return false;
+		}
+	}
+
+	return true;
+}
+
+/*
+ * intel_crtc_cursor_set_obj - Set cursor to specified GEM object
+ *
+ * Note that the object's reference will be consumed if the update fails.  If
+ * the update succeeds, the reference of the old object (if any) will be
+ * consumed.
+ */
+static int intel_crtc_cursor_set_obj(struct drm_crtc *crtc,
+				     struct drm_i915_gem_object *obj,
+				     uint32_t width, uint32_t height)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj;
-	unsigned old_width;
+	enum pipe pipe = intel_crtc->pipe;
+	unsigned old_width, stride;
 	uint32_t addr;
 	int ret;
 
 	/* if we want to turn off the cursor ignore width and height */
-	if (!handle) {
+	if (!obj) {
 		DRM_DEBUG_KMS("cursor off\n");
 		addr = 0;
-		obj = NULL;
 		mutex_lock(&dev->struct_mutex);
 		goto finish;
 	}
 
 	/* Check for which cursor types we support */
-	if (!((width == 64 && height == 64) ||
-			(width == 128 && height == 128 && !IS_GEN2(dev)) ||
-			(width == 256 && height == 256 && !IS_GEN2(dev)))) {
+	if (!cursor_size_ok(dev, width, height)) {
 		DRM_DEBUG("Cursor dimension not supported\n");
 		return -EINVAL;
 	}
 
-	obj = to_intel_bo(drm_gem_object_lookup(dev, file, handle));
-	if (&obj->base == NULL)
-		return -ENOENT;
-
-	if (obj->base.size < width * height * 4) {
-		DRM_DEBUG_KMS("buffer is to small\n");
+	stride = roundup_pow_of_two(width) * 4;
+	if (obj->base.size < stride * height) {
+		DRM_DEBUG_KMS("buffer is too small\n");
 		ret = -ENOMEM;
 		goto fail;
 	}
@@ -8087,6 +8450,15 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 			goto fail_locked;
 		}
 
+		/*
+		 * Global gtt pte registers are special registers which actually
+		 * forward writes to a chunk of system memory. Which means that
+		 * there is no risk that the register values disappear as soon
+		 * as we call intel_runtime_pm_put(), so it is correct to wrap
+		 * only the pin/unpin/fence and not more.
+		 */
+		intel_runtime_pm_get(dev_priv);
+
 		/* Note that the w/a also requires 2 PTE of padding following
 		 * the bo. We currently fill all unused PTE with the shadow
 		 * page and so we should always have valid PTE following the
@@ -8099,16 +8471,20 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 		ret = i915_gem_object_pin_to_display_plane(obj, alignment, NULL);
 		if (ret) {
 			DRM_DEBUG_KMS("failed to move cursor bo into the GTT\n");
+			intel_runtime_pm_put(dev_priv);
 			goto fail_locked;
 		}
 
 		ret = i915_gem_object_put_fence(obj);
 		if (ret) {
 			DRM_DEBUG_KMS("failed to release fence for cursor");
+			intel_runtime_pm_put(dev_priv);
 			goto fail_unpin;
 		}
 
 		addr = i915_gem_obj_ggtt_offset(obj);
+
+		intel_runtime_pm_put(dev_priv);
 	} else {
 		int align = IS_I830(dev) ? 16 * 1024 : 256;
 		ret = i915_gem_object_attach_phys(obj, align);
@@ -8119,16 +8495,14 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 		addr = obj->phys_handle->busaddr;
 	}
 
-	if (IS_GEN2(dev))
-		I915_WRITE(CURSIZE, (height << 12) | width);
-
  finish:
 	if (intel_crtc->cursor_bo) {
 		if (!INTEL_INFO(dev)->cursor_needs_physical)
 			i915_gem_object_unpin_from_display_plane(intel_crtc->cursor_bo);
-		drm_gem_object_unreference(&intel_crtc->cursor_bo->base);
 	}
 
+	i915_gem_track_fb(intel_crtc->cursor_bo, obj,
+			  INTEL_FRONTBUFFER_CURSOR(pipe));
 	mutex_unlock(&dev->struct_mutex);
 
 	old_width = intel_crtc->cursor_width;
@@ -8144,6 +8518,8 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 		intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
 	}
 
+	intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_CURSOR(pipe));
+
 	return 0;
 fail_unpin:
 	i915_gem_object_unpin_from_display_plane(obj);
@@ -8154,19 +8530,6 @@ fail:
 	return ret;
 }
 
-static int intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
-{
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	intel_crtc->cursor_x = clamp_t(int, x, SHRT_MIN, SHRT_MAX);
-	intel_crtc->cursor_y = clamp_t(int, y, SHRT_MIN, SHRT_MAX);
-
-	if (intel_crtc->active)
-		intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
-
-	return 0;
-}
-
 static void intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
 				 u16 *blue, uint32_t start, uint32_t size)
 {
@@ -8242,7 +8605,7 @@ static u32
 intel_framebuffer_size_for_mode(struct drm_display_mode *mode, int bpp)
 {
 	u32 pitch = intel_framebuffer_pitch_for_width(mode->hdisplay, bpp);
-	return ALIGN(pitch * mode->vdisplay, PAGE_SIZE);
+	return PAGE_ALIGN(pitch * mode->vdisplay);
 }
 
 static struct drm_framebuffer *
@@ -8319,8 +8682,6 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 		      connector->base.id, connector->name,
 		      encoder->base.id, encoder->name);
 
-	drm_modeset_acquire_init(ctx, 0);
-
 retry:
 	ret = drm_modeset_lock(&config->connection_mutex, ctx);
 	if (ret)
@@ -8359,10 +8720,14 @@ retry:
 		i++;
 		if (!(encoder->possible_crtcs & (1 << i)))
 			continue;
-		if (!possible_crtc->enabled) {
-			crtc = possible_crtc;
-			break;
-		}
+		if (possible_crtc->enabled)
+			continue;
+		/* This can occur when applying the pipe A quirk on resume. */
+		if (to_intel_crtc(possible_crtc)->new_enabled)
+			continue;
+
+		crtc = possible_crtc;
+		break;
 	}
 
 	/*
@@ -8431,15 +8796,11 @@ fail_unlock:
 		goto retry;
 	}
 
-	drm_modeset_drop_locks(ctx);
-	drm_modeset_acquire_fini(ctx);
-
 	return false;
 }
 
 void intel_release_load_detect_pipe(struct drm_connector *connector,
-				    struct intel_load_detect_pipe *old,
-				    struct drm_modeset_acquire_ctx *ctx)
+				    struct intel_load_detect_pipe *old)
 {
 	struct intel_encoder *intel_encoder =
 		intel_attached_encoder(connector);
@@ -8463,17 +8824,12 @@ void intel_release_load_detect_pipe(struct drm_connector *connector,
 			drm_framebuffer_unreference(old->release_fb);
 		}
 
-		goto unlock;
 		return;
 	}
 
 	/* Switch crtc and encoder back off if necessary */
 	if (old->dpms_mode != DRM_MODE_DPMS_ON)
 		connector->funcs->dpms(connector, old->dpms_mode);
-
-unlock:
-	drm_modeset_drop_locks(ctx);
-	drm_modeset_acquire_fini(ctx);
 }
 
 static int i9xx_pll_refclk(struct drm_device *dev,
@@ -8667,16 +9023,14 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 	return mode;
 }
 
-static void intel_increase_pllclock(struct drm_crtc *crtc)
+static void intel_increase_pllclock(struct drm_device *dev,
+				    enum pipe pipe)
 {
-	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
 	int dpll_reg = DPLL(pipe);
 	int dpll;
 
-	if (HAS_PCH_SPLIT(dev))
+	if (!HAS_GMCH_DISPLAY(dev))
 		return;
 
 	if (!dev_priv->lvds_downclock_avail)
@@ -8704,7 +9058,7 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
-	if (HAS_PCH_SPLIT(dev))
+	if (!HAS_GMCH_DISPLAY(dev))
 		return;
 
 	if (!dev_priv->lvds_downclock_avail)
@@ -8773,28 +9127,188 @@ out:
 	intel_runtime_pm_put(dev_priv);
 }
 
-void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
-			struct intel_engine_cs *ring)
+
+/**
+ * intel_mark_fb_busy - mark given planes as busy
+ * @dev: DRM device
+ * @frontbuffer_bits: bits for the affected planes
+ * @ring: optional ring for asynchronous commands
+ *
+ * This function gets called every time the screen contents change. It can be
+ * used to keep e.g. the update rate at the nominal refresh rate with DRRS.
+ */
+static void intel_mark_fb_busy(struct drm_device *dev,
+			       unsigned frontbuffer_bits,
+			       struct intel_engine_cs *ring)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_crtc *crtc;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
 
 	if (!i915.powersave)
 		return;
 
-	for_each_crtc(dev, crtc) {
-		if (!crtc->primary->fb)
-			continue;
-
-		if (to_intel_framebuffer(crtc->primary->fb)->obj != obj)
+	for_each_pipe(dev_priv, pipe) {
+		if (!(frontbuffer_bits & INTEL_FRONTBUFFER_ALL_MASK(pipe)))
 			continue;
 
-		intel_increase_pllclock(crtc);
+		intel_increase_pllclock(dev, pipe);
 		if (ring && intel_fbc_enabled(dev))
 			ring->fbc_dirty = true;
 	}
 }
 
+/**
+ * intel_fb_obj_invalidate - invalidate frontbuffer object
+ * @obj: GEM object to invalidate
+ * @ring: set for asynchronous rendering
+ *
+ * This function gets called every time rendering on the given object starts and
+ * frontbuffer caching (fbc, low refresh rate for DRRS, panel self refresh) must
+ * be invalidated. If @ring is non-NULL any subsequent invalidation will be delayed
+ * until the rendering completes or a flip on this frontbuffer plane is
+ * scheduled.
+ */
+void intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
+			     struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (!obj->frontbuffer_bits)
+		return;
+
+	if (ring) {
+		mutex_lock(&dev_priv->fb_tracking.lock);
+		dev_priv->fb_tracking.busy_bits
+			|= obj->frontbuffer_bits;
+		dev_priv->fb_tracking.flip_bits
+			&= ~obj->frontbuffer_bits;
+		mutex_unlock(&dev_priv->fb_tracking.lock);
+	}
+
+	intel_mark_fb_busy(dev, obj->frontbuffer_bits, ring);
+
+	intel_edp_psr_invalidate(dev, obj->frontbuffer_bits);
+}
+
+/**
+ * intel_frontbuffer_flush - flush frontbuffer
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called every time rendering on the given planes has
+ * completed and frontbuffer caching can be started again. Flushes will get
+ * delayed if they're blocked by some oustanding asynchronous rendering.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flush(struct drm_device *dev,
+			     unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* Delay flushing when rings are still busy.*/
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	frontbuffer_bits &= ~dev_priv->fb_tracking.busy_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+
+	intel_mark_fb_busy(dev, frontbuffer_bits, NULL);
+
+	intel_edp_psr_flush(dev, frontbuffer_bits);
+
+	/*
+	 * FIXME: Unconditional fbc flushing here is a rather gross hack and
+	 * needs to be reworked into a proper frontbuffer tracking scheme like
+	 * psr employs.
+	 */
+	if (IS_BROADWELL(dev))
+		gen8_fbc_sw_flush(dev, FBC_REND_CACHE_CLEAN);
+}
+
+/**
+ * intel_fb_obj_flush - flush frontbuffer object
+ * @obj: GEM object to flush
+ * @retire: set when retiring asynchronous rendering
+ *
+ * This function gets called every time rendering on the given object has
+ * completed and frontbuffer caching can be started again. If @retire is true
+ * then any delayed flushes will be unblocked.
+ */
+void intel_fb_obj_flush(struct drm_i915_gem_object *obj,
+			bool retire)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned frontbuffer_bits;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	if (!obj->frontbuffer_bits)
+		return;
+
+	frontbuffer_bits = obj->frontbuffer_bits;
+
+	if (retire) {
+		mutex_lock(&dev_priv->fb_tracking.lock);
+		/* Filter out new bits since rendering started. */
+		frontbuffer_bits &= dev_priv->fb_tracking.busy_bits;
+
+		dev_priv->fb_tracking.busy_bits &= ~frontbuffer_bits;
+		mutex_unlock(&dev_priv->fb_tracking.lock);
+	}
+
+	intel_frontbuffer_flush(dev, frontbuffer_bits);
+}
+
+/**
+ * intel_frontbuffer_flip_prepare - prepare asnychronous frontbuffer flip
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after scheduling a flip on @obj. The actual
+ * frontbuffer flushing will be delayed until completion is signalled with
+ * intel_frontbuffer_flip_complete. If an invalidate happens in between this
+ * flush will be cancelled.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_prepare(struct drm_device *dev,
+				    unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	dev_priv->fb_tracking.flip_bits
+		|= frontbuffer_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+}
+
+/**
+ * intel_frontbuffer_flip_complete - complete asynchronous frontbuffer flush
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after the flip has been latched and will complete
+ * on the next vblank. It will execute the fush if it hasn't been cancalled yet.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_complete(struct drm_device *dev,
+				     unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	mutex_lock(&dev_priv->fb_tracking.lock);
+	/* Mask any cancelled flips. */
+	frontbuffer_bits &= dev_priv->fb_tracking.flip_bits;
+	dev_priv->fb_tracking.flip_bits &= ~frontbuffer_bits;
+	mutex_unlock(&dev_priv->fb_tracking.lock);
+
+	intel_frontbuffer_flush(dev, frontbuffer_bits);
+}
+
 static void intel_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -8812,8 +9326,6 @@ static void intel_crtc_destroy(struct drm_crtc *crtc)
 		kfree(work);
 	}
 
-	intel_crtc_cursor_set(crtc, NULL, 0, 0, 0);
-
 	drm_crtc_cleanup(crtc);
 
 	kfree(intel_crtc);
@@ -8824,6 +9336,7 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 	struct intel_unpin_work *work =
 		container_of(__work, struct intel_unpin_work, work);
 	struct drm_device *dev = work->crtc->dev;
+	enum pipe pipe = to_intel_crtc(work->crtc)->pipe;
 
 	mutex_lock(&dev->struct_mutex);
 	intel_unpin_fb_obj(work->old_fb_obj);
@@ -8833,6 +9346,8 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 	intel_update_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
+	intel_frontbuffer_flip_complete(dev, INTEL_FRONTBUFFER_PRIMARY(pipe));
+
 	BUG_ON(atomic_read(&to_intel_crtc(work->crtc)->unpin_work_count) == 0);
 	atomic_dec(&to_intel_crtc(work->crtc)->unpin_work_count);
 
@@ -8842,7 +9357,6 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 static void do_intel_finish_page_flip(struct drm_device *dev,
 				      struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_unpin_work *work;
 	unsigned long flags;
@@ -8862,23 +9376,9 @@ static void do_intel_finish_page_flip(struct drm_device *dev,
 		return;
 	}
 
-	/* and that the unpin work is consistent wrt ->pending. */
-	smp_rmb();
-
-	intel_crtc->unpin_work = NULL;
-
-	if (work->event)
-		drm_send_vblank_event(dev, intel_crtc->pipe, work->event);
-
-	drm_crtc_vblank_put(crtc);
+	page_flip_completed(intel_crtc);
 
 	spin_unlock_irqrestore(&dev->event_lock, flags);
-
-	wake_up_all(&dev_priv->pending_flip_queue);
-
-	queue_work(dev_priv->wq, &work->work);
-
-	trace_i915_flip_complete(intel_crtc->plane, work->pending_flip_obj);
 }
 
 void intel_finish_page_flip(struct drm_device *dev, int pipe)
@@ -9202,6 +9702,152 @@ static int intel_gen7_queue_flip(struct drm_device *dev,
 	return 0;
 }
 
+static bool use_mmio_flip(struct intel_engine_cs *ring,
+			  struct drm_i915_gem_object *obj)
+{
+	/*
+	 * This is not being used for older platforms, because
+	 * non-availability of flip done interrupt forces us to use
+	 * CS flips. Older platforms derive flip done using some clever
+	 * tricks involving the flip_pending status bits and vblank irqs.
+	 * So using MMIO flips there would disrupt this mechanism.
+	 */
+
+	if (ring == NULL)
+		return true;
+
+	if (INTEL_INFO(ring->dev)->gen < 5)
+		return false;
+
+	if (i915.use_mmio_flip < 0)
+		return false;
+	else if (i915.use_mmio_flip > 0)
+		return true;
+	else if (i915.enable_execlists)
+		return true;
+	else
+		return ring != obj->ring;
+}
+
+static void intel_do_mmio_flip(struct intel_crtc *intel_crtc)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_framebuffer *intel_fb =
+		to_intel_framebuffer(intel_crtc->base.primary->fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	u32 dspcntr;
+	u32 reg;
+
+	intel_mark_page_flip_active(intel_crtc);
+
+	reg = DSPCNTR(intel_crtc->plane);
+	dspcntr = I915_READ(reg);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		if (obj->tiling_mode != I915_TILING_NONE)
+			dspcntr |= DISPPLANE_TILED;
+		else
+			dspcntr &= ~DISPPLANE_TILED;
+	}
+	I915_WRITE(reg, dspcntr);
+
+	I915_WRITE(DSPSURF(intel_crtc->plane),
+		   intel_crtc->unpin_work->gtt_offset);
+	POSTING_READ(DSPSURF(intel_crtc->plane));
+}
+
+static int intel_postpone_flip(struct drm_i915_gem_object *obj)
+{
+	struct intel_engine_cs *ring;
+	int ret;
+
+	lockdep_assert_held(&obj->base.dev->struct_mutex);
+
+	if (!obj->last_write_seqno)
+		return 0;
+
+	ring = obj->ring;
+
+	if (i915_seqno_passed(ring->get_seqno(ring, true),
+			      obj->last_write_seqno))
+		return 0;
+
+	ret = i915_gem_check_olr(ring, obj->last_write_seqno);
+	if (ret)
+		return ret;
+
+	if (WARN_ON(!ring->irq_get(ring)))
+		return 0;
+
+	return 1;
+}
+
+void intel_notify_mmio_flip(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = to_i915(ring->dev);
+	struct intel_crtc *intel_crtc;
+	unsigned long irq_flags;
+	u32 seqno;
+
+	seqno = ring->get_seqno(ring, false);
+
+	spin_lock_irqsave(&dev_priv->mmio_flip_lock, irq_flags);
+	for_each_intel_crtc(ring->dev, intel_crtc) {
+		struct intel_mmio_flip *mmio_flip;
+
+		mmio_flip = &intel_crtc->mmio_flip;
+		if (mmio_flip->seqno == 0)
+			continue;
+
+		if (ring->id != mmio_flip->ring_id)
+			continue;
+
+		if (i915_seqno_passed(seqno, mmio_flip->seqno)) {
+			intel_do_mmio_flip(intel_crtc);
+			mmio_flip->seqno = 0;
+			ring->irq_put(ring);
+		}
+	}
+	spin_unlock_irqrestore(&dev_priv->mmio_flip_lock, irq_flags);
+}
+
+static int intel_queue_mmio_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct intel_engine_cs *ring,
+				 uint32_t flags)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned long irq_flags;
+	int ret;
+
+	if (WARN_ON(intel_crtc->mmio_flip.seqno))
+		return -EBUSY;
+
+	ret = intel_postpone_flip(obj);
+	if (ret < 0)
+		return ret;
+	if (ret == 0) {
+		intel_do_mmio_flip(intel_crtc);
+		return 0;
+	}
+
+	spin_lock_irqsave(&dev_priv->mmio_flip_lock, irq_flags);
+	intel_crtc->mmio_flip.seqno = obj->last_write_seqno;
+	intel_crtc->mmio_flip.ring_id = obj->ring->id;
+	spin_unlock_irqrestore(&dev_priv->mmio_flip_lock, irq_flags);
+
+	/*
+	 * Double check to catch cases where irq fired before
+	 * mmio flip data was ready
+	 */
+	intel_notify_mmio_flip(obj->ring);
+	return 0;
+}
+
 static int intel_default_queue_flip(struct drm_device *dev,
 				    struct drm_crtc *crtc,
 				    struct drm_framebuffer *fb,
@@ -9212,6 +9858,65 @@ static int intel_default_queue_flip(struct drm_device *dev,
 	return -ENODEV;
 }
 
+static bool __intel_pageflip_stall_check(struct drm_device *dev,
+					 struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_unpin_work *work = intel_crtc->unpin_work;
+	u32 addr;
+
+	if (atomic_read(&work->pending) >= INTEL_FLIP_COMPLETE)
+		return true;
+
+	if (!work->enable_stall_check)
+		return false;
+
+	if (work->flip_ready_vblank == 0) {
+		if (work->flip_queued_ring &&
+		    !i915_seqno_passed(work->flip_queued_ring->get_seqno(work->flip_queued_ring, true),
+				       work->flip_queued_seqno))
+			return false;
+
+		work->flip_ready_vblank = drm_vblank_count(dev, intel_crtc->pipe);
+	}
+
+	if (drm_vblank_count(dev, intel_crtc->pipe) - work->flip_ready_vblank < 3)
+		return false;
+
+	/* Potential stall - if we see that the flip has happened,
+	 * assume a missed interrupt. */
+	if (INTEL_INFO(dev)->gen >= 4)
+		addr = I915_HI_DISPBASE(I915_READ(DSPSURF(intel_crtc->plane)));
+	else
+		addr = I915_READ(DSPADDR(intel_crtc->plane));
+
+	/* There is a potential issue here with a false positive after a flip
+	 * to the same address. We could address this by checking for a
+	 * non-incrementing frame counter.
+	 */
+	return addr == work->gtt_offset;
+}
+
+void intel_check_page_flip(struct drm_device *dev, int pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned long flags;
+
+	if (crtc == NULL)
+		return;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (intel_crtc->unpin_work && __intel_pageflip_stall_check(dev, crtc)) {
+		WARN_ONCE(1, "Kicking stuck page flip: queued at %d, now %d\n",
+			 intel_crtc->unpin_work->flip_queued_vblank, drm_vblank_count(dev, pipe));
+		page_flip_completed(intel_crtc);
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
 static int intel_crtc_page_flip(struct drm_crtc *crtc,
 				struct drm_framebuffer *fb,
 				struct drm_pending_vblank_event *event,
@@ -9220,13 +9925,22 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_framebuffer *old_fb = crtc->primary->fb;
-	struct drm_i915_gem_object *obj = to_intel_framebuffer(fb)->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum pipe pipe = intel_crtc->pipe;
 	struct intel_unpin_work *work;
 	struct intel_engine_cs *ring;
 	unsigned long flags;
 	int ret;
 
+	/*
+	 * drm_mode_page_flip_ioctl() should already catch this, but double
+	 * check to be safe.  In the future we may enable pageflipping from
+	 * a disabled primary plane.
+	 */
+	if (WARN_ON(intel_fb_obj(old_fb) == NULL))
+		return -EBUSY;
+
 	/* Can't change pixel format via MI display flips. */
 	if (fb->pixel_format != crtc->primary->fb->pixel_format)
 		return -EINVAL;
@@ -9249,7 +9963,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 
 	work->event = event;
 	work->crtc = crtc;
-	work->old_fb_obj = to_intel_framebuffer(old_fb)->obj;
+	work->old_fb_obj = intel_fb_obj(old_fb);
 	INIT_WORK(&work->work, intel_unpin_work_fn);
 
 	ret = drm_crtc_vblank_get(crtc);
@@ -9259,12 +9973,20 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	/* We borrow the event spin lock for protecting unpin_work */
 	spin_lock_irqsave(&dev->event_lock, flags);
 	if (intel_crtc->unpin_work) {
-		spin_unlock_irqrestore(&dev->event_lock, flags);
-		kfree(work);
-		drm_crtc_vblank_put(crtc);
+		/* Before declaring the flip queue wedged, check if
+		 * the hardware completed the operation behind our backs.
+		 */
+		if (__intel_pageflip_stall_check(dev, crtc)) {
+			DRM_DEBUG_DRIVER("flip queue: previous flip completed, continuing\n");
+			page_flip_completed(intel_crtc);
+		} else {
+			DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
+			spin_unlock_irqrestore(&dev->event_lock, flags);
 
-		DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
-		return -EBUSY;
+			drm_crtc_vblank_put(crtc);
+			kfree(work);
+			return -EBUSY;
+		}
 	}
 	intel_crtc->unpin_work = work;
 	spin_unlock_irqrestore(&dev->event_lock, flags);
@@ -9284,16 +10006,19 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 
 	work->pending_flip_obj = obj;
 
-	work->enable_stall_check = true;
-
 	atomic_inc(&intel_crtc->unpin_work_count);
 	intel_crtc->reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 
 	if (INTEL_INFO(dev)->gen >= 5 || IS_G4X(dev))
-		work->flip_count = I915_READ(PIPE_FLIPCOUNT_GM45(intel_crtc->pipe)) + 1;
+		work->flip_count = I915_READ(PIPE_FLIPCOUNT_GM45(pipe)) + 1;
 
 	if (IS_VALLEYVIEW(dev)) {
 		ring = &dev_priv->ring[BCS];
+		if (obj->tiling_mode != work->old_fb_obj->tiling_mode)
+			/* vlv: DISPLAY_FLIP fails to change tiling */
+			ring = NULL;
+	} else if (IS_IVYBRIDGE(dev)) {
+		ring = &dev_priv->ring[BCS];
 	} else if (INTEL_INFO(dev)->gen >= 7) {
 		ring = obj->ring;
 		if (ring == NULL || ring->id != RCS)
@@ -9309,12 +10034,32 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	work->gtt_offset =
 		i915_gem_obj_ggtt_offset(obj) + intel_crtc->dspaddr_offset;
 
-	ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, ring, page_flip_flags);
-	if (ret)
-		goto cleanup_unpin;
+	if (use_mmio_flip(ring, obj)) {
+		ret = intel_queue_mmio_flip(dev, crtc, fb, obj, ring,
+					    page_flip_flags);
+		if (ret)
+			goto cleanup_unpin;
+
+		work->flip_queued_seqno = obj->last_write_seqno;
+		work->flip_queued_ring = obj->ring;
+	} else {
+		ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, ring,
+						   page_flip_flags);
+		if (ret)
+			goto cleanup_unpin;
+
+		work->flip_queued_seqno = intel_ring_get_seqno(ring);
+		work->flip_queued_ring = ring;
+	}
+
+	work->flip_queued_vblank = drm_vblank_count(dev, intel_crtc->pipe);
+	work->enable_stall_check = true;
+
+	i915_gem_track_fb(work->old_fb_obj, obj,
+			  INTEL_FRONTBUFFER_PRIMARY(pipe));
 
 	intel_disable_fbc(dev);
-	intel_mark_fb_busy(obj, NULL);
+	intel_frontbuffer_flip_prepare(dev, INTEL_FRONTBUFFER_PRIMARY(pipe));
 	mutex_unlock(&dev->struct_mutex);
 
 	trace_i915_flip_request(intel_crtc->plane, obj);
@@ -9343,8 +10088,11 @@ free_work:
 out_hang:
 		intel_crtc_wait_for_pending_flips(crtc);
 		ret = intel_pipe_set_base(crtc, crtc->x, crtc->y, fb);
-		if (ret == 0 && event)
-			drm_send_vblank_event(dev, intel_crtc->pipe, event);
+		if (ret == 0 && event) {
+			spin_lock_irqsave(&dev->event_lock, flags);
+			drm_send_vblank_event(dev, pipe, event);
+			spin_unlock_irqrestore(&dev->event_lock, flags);
+		}
 	}
 	return ret;
 }
@@ -9372,8 +10120,7 @@ static void intel_modeset_update_staged_output_state(struct drm_device *dev)
 			to_intel_encoder(connector->base.encoder);
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->new_crtc =
 			to_intel_crtc(encoder->base.crtc);
 	}
@@ -9404,8 +10151,7 @@ static void intel_modeset_commit_output_state(struct drm_device *dev)
 		connector->base.encoder = &connector->new_encoder->base;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->base.crtc = &encoder->new_crtc->base;
 	}
 
@@ -9532,6 +10278,15 @@ static void intel_dump_pipe_config(struct intel_crtc *crtc,
 		      pipe_config->dp_m_n.gmch_m, pipe_config->dp_m_n.gmch_n,
 		      pipe_config->dp_m_n.link_m, pipe_config->dp_m_n.link_n,
 		      pipe_config->dp_m_n.tu);
+
+	DRM_DEBUG_KMS("dp: %i, gmch_m2: %u, gmch_n2: %u, link_m2: %u, link_n2: %u, tu2: %u\n",
+		      pipe_config->has_dp_encoder,
+		      pipe_config->dp_m2_n2.gmch_m,
+		      pipe_config->dp_m2_n2.gmch_n,
+		      pipe_config->dp_m2_n2.link_m,
+		      pipe_config->dp_m2_n2.link_n,
+		      pipe_config->dp_m2_n2.tu);
+
 	DRM_DEBUG_KMS("requested mode:\n");
 	drm_mode_debug_printmodeline(&pipe_config->requested_mode);
 	DRM_DEBUG_KMS("adjusted mode:\n");
@@ -9566,8 +10321,7 @@ static bool check_single_encoder_cloning(struct intel_crtc *crtc,
 	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *source_encoder;
 
-	list_for_each_entry(source_encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, source_encoder) {
 		if (source_encoder->new_crtc != crtc)
 			continue;
 
@@ -9583,8 +10337,7 @@ static bool check_encoder_cloning(struct intel_crtc *crtc)
 	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *encoder;
 
-	list_for_each_entry(encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc != crtc)
 			continue;
 
@@ -9668,8 +10421,7 @@ encoder_retry:
 	 * adjust it according to limitations or connector properties, and also
 	 * a chance to reject the mode entirely.
 	 */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 
 		if (&encoder->new_crtc->base != crtc)
 			continue;
@@ -9747,8 +10499,7 @@ intel_modeset_affected_pipes(struct drm_crtc *crtc, unsigned *modeset_pipes,
 				1 << connector->new_encoder->new_crtc->pipe;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->base.crtc == &encoder->new_crtc->base)
 			continue;
 
@@ -9822,8 +10573,7 @@ intel_modeset_update_state(struct drm_device *dev, unsigned prepare_pipes)
 	struct intel_crtc *intel_crtc;
 	struct drm_connector *connector;
 
-	list_for_each_entry(intel_encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, intel_encoder) {
 		if (!intel_encoder->base.crtc)
 			continue;
 
@@ -9912,6 +10662,22 @@ intel_pipe_config_compare(struct drm_device *dev,
 		return false; \
 	}
 
+/* This is required for BDW+ where there is only one set of registers for
+ * switching between high and low RR.
+ * This macro can be used whenever a comparison has to be made between one
+ * hw state and multiple sw state variables.
+ */
+#define PIPE_CONF_CHECK_I_ALT(name, alt_name) \
+	if ((current_config->name != pipe_config->name) && \
+		(current_config->alt_name != pipe_config->name)) { \
+			DRM_ERROR("mismatch in " #name " " \
+				  "(expected %i or %i, found %i)\n", \
+				  current_config->name, \
+				  current_config->alt_name, \
+				  pipe_config->name); \
+			return false; \
+	}
+
 #define PIPE_CONF_CHECK_FLAGS(name, mask)	\
 	if ((current_config->name ^ pipe_config->name) & (mask)) { \
 		DRM_ERROR("mismatch in " #name "(" #mask ") "	   \
@@ -9944,11 +10710,28 @@ intel_pipe_config_compare(struct drm_device *dev,
 	PIPE_CONF_CHECK_I(fdi_m_n.tu);
 
 	PIPE_CONF_CHECK_I(has_dp_encoder);
-	PIPE_CONF_CHECK_I(dp_m_n.gmch_m);
-	PIPE_CONF_CHECK_I(dp_m_n.gmch_n);
-	PIPE_CONF_CHECK_I(dp_m_n.link_m);
-	PIPE_CONF_CHECK_I(dp_m_n.link_n);
-	PIPE_CONF_CHECK_I(dp_m_n.tu);
+
+	if (INTEL_INFO(dev)->gen < 8) {
+		PIPE_CONF_CHECK_I(dp_m_n.gmch_m);
+		PIPE_CONF_CHECK_I(dp_m_n.gmch_n);
+		PIPE_CONF_CHECK_I(dp_m_n.link_m);
+		PIPE_CONF_CHECK_I(dp_m_n.link_n);
+		PIPE_CONF_CHECK_I(dp_m_n.tu);
+
+		if (current_config->has_drrs) {
+			PIPE_CONF_CHECK_I(dp_m2_n2.gmch_m);
+			PIPE_CONF_CHECK_I(dp_m2_n2.gmch_n);
+			PIPE_CONF_CHECK_I(dp_m2_n2.link_m);
+			PIPE_CONF_CHECK_I(dp_m2_n2.link_n);
+			PIPE_CONF_CHECK_I(dp_m2_n2.tu);
+		}
+	} else {
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.gmch_m, dp_m2_n2.gmch_m);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.gmch_n, dp_m2_n2.gmch_n);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.link_m, dp_m2_n2.link_m);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.link_n, dp_m2_n2.link_n);
+		PIPE_CONF_CHECK_I_ALT(dp_m_n.tu, dp_m2_n2.tu);
+	}
 
 	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hdisplay);
 	PIPE_CONF_CHECK_I(adjusted_mode.crtc_htotal);
@@ -10017,11 +10800,14 @@ intel_pipe_config_compare(struct drm_device *dev,
 
 	PIPE_CONF_CHECK_I(double_wide);
 
+	PIPE_CONF_CHECK_X(ddi_pll_sel);
+
 	PIPE_CONF_CHECK_I(shared_dpll);
 	PIPE_CONF_CHECK_X(dpll_hw_state.dpll);
 	PIPE_CONF_CHECK_X(dpll_hw_state.dpll_md);
 	PIPE_CONF_CHECK_X(dpll_hw_state.fp0);
 	PIPE_CONF_CHECK_X(dpll_hw_state.fp1);
+	PIPE_CONF_CHECK_X(dpll_hw_state.wrpll);
 
 	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5)
 		PIPE_CONF_CHECK_I(pipe_bpp);
@@ -10031,6 +10817,7 @@ intel_pipe_config_compare(struct drm_device *dev,
 
 #undef PIPE_CONF_CHECK_X
 #undef PIPE_CONF_CHECK_I
+#undef PIPE_CONF_CHECK_I_ALT
 #undef PIPE_CONF_CHECK_FLAGS
 #undef PIPE_CONF_CHECK_CLOCK_FUZZY
 #undef PIPE_CONF_QUIRK
@@ -10060,8 +10847,7 @@ check_encoder_state(struct drm_device *dev)
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		bool enabled = false;
 		bool active = false;
 		enum pipe pipe, tracked_pipe;
@@ -10083,6 +10869,14 @@ check_encoder_state(struct drm_device *dev)
 			if (connector->base.dpms != DRM_MODE_DPMS_OFF)
 				active = true;
 		}
+		/*
+		 * for MST connectors if we unplug the connector is gone
+		 * away but the encoder is still connected to a crtc
+		 * until a modeset happens in response to the hotplug.
+		 */
+		if (!enabled && encoder->base.encoder_type == DRM_MODE_ENCODER_DPMST)
+			continue;
+
 		WARN(!!encoder->base.crtc != enabled,
 		     "encoder's enabled state mismatch "
 		     "(expected %i, found %i)\n",
@@ -10132,8 +10926,7 @@ check_crtc_state(struct drm_device *dev)
 		WARN(crtc->active && !crtc->base.enabled,
 		     "active crtc, but not enabled in sw tracking\n");
 
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			if (encoder->base.crtc != &crtc->base)
 				continue;
 			enabled = true;
@@ -10151,12 +10944,12 @@ check_crtc_state(struct drm_device *dev)
 		active = dev_priv->display.get_pipe_config(crtc,
 							   &pipe_config);
 
-		/* hw state is inconsistent with the pipe A quirk */
-		if (crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
+		/* hw state is inconsistent with the pipe quirk */
+		if ((crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) ||
+		    (crtc->pipe == PIPE_B && dev_priv->quirks & QUIRK_PIPEB_FORCE))
 			active = crtc->active;
 
-		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			enum pipe pipe;
 			if (encoder->base.crtc != &crtc->base)
 				continue;
@@ -10378,20 +11171,23 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * on the DPLL.
 	 */
 	for_each_intel_crtc_masked(dev, modeset_pipes, intel_crtc) {
-		struct drm_framebuffer *old_fb;
+		struct drm_framebuffer *old_fb = crtc->primary->fb;
+		struct drm_i915_gem_object *old_obj = intel_fb_obj(old_fb);
+		struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 
 		mutex_lock(&dev->struct_mutex);
 		ret = intel_pin_and_fence_fb_obj(dev,
-						 to_intel_framebuffer(fb)->obj,
+						 obj,
 						 NULL);
 		if (ret != 0) {
 			DRM_ERROR("pin & fence failed\n");
 			mutex_unlock(&dev->struct_mutex);
 			goto done;
 		}
-		old_fb = crtc->primary->fb;
 		if (old_fb)
-			intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
+			intel_unpin_fb_obj(old_obj);
+		i915_gem_track_fb(old_obj, obj,
+				  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
 		mutex_unlock(&dev->struct_mutex);
 
 		crtc->primary->fb = fb;
@@ -10521,7 +11317,7 @@ static void intel_set_config_restore_state(struct drm_device *dev,
 	}
 
 	count = 0;
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		encoder->new_crtc =
 			to_intel_crtc(config->save_encoder_crtcs[count++]);
 	}
@@ -10563,12 +11359,17 @@ intel_set_config_compute_mode_changes(struct drm_mode_set *set,
 	if (is_crtc_connector_off(set)) {
 		config->mode_changed = true;
 	} else if (set->crtc->primary->fb != set->fb) {
-		/* If we have no fb then treat it as a full mode set */
+		/*
+		 * If we have no fb, we can only flip as long as the crtc is
+		 * active, otherwise we need a full mode set.  The crtc may
+		 * be active if we've only disabled the primary plane, or
+		 * in fastboot situations.
+		 */
 		if (set->crtc->primary->fb == NULL) {
 			struct intel_crtc *intel_crtc =
 				to_intel_crtc(set->crtc);
 
-			if (intel_crtc->active && i915.fastboot) {
+			if (intel_crtc->active) {
 				DRM_DEBUG_KMS("crtc has no fb, will flip\n");
 				config->fb_changed = true;
 			} else {
@@ -10620,7 +11421,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 		 * for them. */
 		for (ro = 0; ro < set->num_connectors; ro++) {
 			if (set->connectors[ro] == &connector->base) {
-				connector->new_encoder = connector->encoder;
+				connector->new_encoder = intel_find_encoder(connector, to_intel_crtc(set->crtc)->pipe);
 				break;
 			}
 		}
@@ -10666,7 +11467,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 					 new_crtc)) {
 			return -EINVAL;
 		}
-		connector->encoder->new_crtc = to_intel_crtc(new_crtc);
+		connector->new_encoder->new_crtc = to_intel_crtc(new_crtc);
 
 		DRM_DEBUG_KMS("[CONNECTOR:%d:%s] to [CRTC:%d]\n",
 			connector->base.base.id,
@@ -10675,8 +11476,7 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 	}
 
 	/* Check for any encoders that needs to be disabled. */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		int num_connectors = 0;
 		list_for_each_entry(connector,
 				    &dev->mode_config.connector_list,
@@ -10700,13 +11500,16 @@ intel_modeset_stage_output_state(struct drm_device *dev,
 		}
 	}
 	/* Now we've also updated encoder->new_crtc for all encoders. */
-
+	list_for_each_entry(connector, &dev->mode_config.connector_list,
+			    base.head) {
+		if (connector->new_encoder)
+			if (connector->new_encoder != connector->encoder)
+				connector->encoder = connector->new_encoder;
+	}
 	for_each_intel_crtc(dev, crtc) {
 		crtc->new_enabled = false;
 
-		list_for_each_entry(encoder,
-				    &dev->mode_config.encoder_list,
-				    base.head) {
+		for_each_intel_encoder(dev, encoder) {
 			if (encoder->new_crtc == crtc) {
 				crtc->new_enabled = true;
 				break;
@@ -10743,7 +11546,7 @@ static void disable_crtc_nofb(struct intel_crtc *crtc)
 			connector->new_encoder = NULL;
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc == crtc)
 			encoder->new_crtc = NULL;
 	}
@@ -10806,10 +11609,22 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 		ret = intel_set_mode(set->crtc, set->mode,
 				     set->x, set->y, set->fb);
 	} else if (config->fb_changed) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(set->crtc);
+
 		intel_crtc_wait_for_pending_flips(set->crtc);
 
 		ret = intel_pipe_set_base(set->crtc,
 					  set->x, set->y, set->fb);
+
+		/*
+		 * We need to make sure the primary plane is re-enabled if it
+		 * has previously been turned off.
+		 */
+		if (!intel_crtc->primary_enabled && ret == 0) {
+			WARN_ON(!intel_crtc->active);
+			intel_enable_primary_hw_plane(set->crtc->primary, set->crtc);
+		}
+
 		/*
 		 * In the fastboot case this may be our only check of the
 		 * state after boot.  It would be better to only do it on
@@ -10850,26 +11665,21 @@ out_config:
 }
 
 static const struct drm_crtc_funcs intel_crtc_funcs = {
-	.cursor_set = intel_crtc_cursor_set,
-	.cursor_move = intel_crtc_cursor_move,
 	.gamma_set = intel_crtc_gamma_set,
 	.set_config = intel_crtc_set_config,
 	.destroy = intel_crtc_destroy,
 	.page_flip = intel_crtc_page_flip,
 };
 
-static void intel_cpu_pll_init(struct drm_device *dev)
-{
-	if (HAS_DDI(dev))
-		intel_ddi_pll_init(dev);
-}
-
 static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
 				      struct intel_shared_dpll *pll,
 				      struct intel_dpll_hw_state *hw_state)
 {
 	uint32_t val;
 
+	if (!intel_display_power_enabled(dev_priv, POWER_DOMAIN_PLLS))
+		return false;
+
 	val = I915_READ(PCH_DPLL(pll->id));
 	hw_state->dpll = val;
 	hw_state->fp0 = I915_READ(PCH_FP0(pll->id));
@@ -10951,7 +11761,9 @@ static void intel_shared_dpll_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+	if (HAS_DDI(dev))
+		intel_ddi_pll_init(dev);
+	else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
 		ibx_pch_dpll_init(dev);
 	else
 		dev_priv->num_shared_dpll = 0;
@@ -10959,17 +11771,383 @@ static void intel_shared_dpll_init(struct drm_device *dev)
 	BUG_ON(dev_priv->num_shared_dpll > I915_NUM_PLLS);
 }
 
+static int
+intel_primary_plane_disable(struct drm_plane *plane)
+{
+	struct drm_device *dev = plane->dev;
+	struct intel_crtc *intel_crtc;
+
+	if (!plane->fb)
+		return 0;
+
+	BUG_ON(!plane->crtc);
+
+	intel_crtc = to_intel_crtc(plane->crtc);
+
+	/*
+	 * Even though we checked plane->fb above, it's still possible that
+	 * the primary plane has been implicitly disabled because the crtc
+	 * coordinates given weren't visible, or because we detected
+	 * that it was 100% covered by a sprite plane.  Or, the CRTC may be
+	 * off and we've set a fb, but haven't actually turned on the CRTC yet.
+	 * In either case, we need to unpin the FB and let the fb pointer get
+	 * updated, but otherwise we don't need to touch the hardware.
+	 */
+	if (!intel_crtc->primary_enabled)
+		goto disable_unpin;
+
+	intel_crtc_wait_for_pending_flips(plane->crtc);
+	intel_disable_primary_hw_plane(plane, plane->crtc);
+
+disable_unpin:
+	mutex_lock(&dev->struct_mutex);
+	i915_gem_track_fb(intel_fb_obj(plane->fb), NULL,
+			  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
+	intel_unpin_fb_obj(intel_fb_obj(plane->fb));
+	mutex_unlock(&dev->struct_mutex);
+	plane->fb = NULL;
+
+	return 0;
+}
+
+static int
+intel_primary_plane_setplane(struct drm_plane *plane, struct drm_crtc *crtc,
+			     struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			     unsigned int crtc_w, unsigned int crtc_h,
+			     uint32_t src_x, uint32_t src_y,
+			     uint32_t src_w, uint32_t src_h)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_gem_object *old_obj = intel_fb_obj(plane->fb);
+	struct drm_rect dest = {
+		/* integer pixels */
+		.x1 = crtc_x,
+		.y1 = crtc_y,
+		.x2 = crtc_x + crtc_w,
+		.y2 = crtc_y + crtc_h,
+	};
+	struct drm_rect src = {
+		/* 16.16 fixed point */
+		.x1 = src_x,
+		.y1 = src_y,
+		.x2 = src_x + src_w,
+		.y2 = src_y + src_h,
+	};
+	const struct drm_rect clip = {
+		/* integer pixels */
+		.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0,
+		.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0,
+	};
+	const struct {
+		int crtc_x, crtc_y;
+		unsigned int crtc_w, crtc_h;
+		uint32_t src_x, src_y, src_w, src_h;
+	} orig = {
+		.crtc_x = crtc_x,
+		.crtc_y = crtc_y,
+		.crtc_w = crtc_w,
+		.crtc_h = crtc_h,
+		.src_x = src_x,
+		.src_y = src_y,
+		.src_w = src_w,
+		.src_h = src_h,
+	};
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	bool visible;
+	int ret;
+
+	ret = drm_plane_helper_check_update(plane, crtc, fb,
+					    &src, &dest, &clip,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    false, true, &visible);
+
+	if (ret)
+		return ret;
+
+	/*
+	 * If the CRTC isn't enabled, we're just pinning the framebuffer,
+	 * updating the fb pointer, and returning without touching the
+	 * hardware.  This allows us to later do a drmModeSetCrtc with fb=-1 to
+	 * turn on the display with all planes setup as desired.
+	 */
+	if (!crtc->enabled) {
+		mutex_lock(&dev->struct_mutex);
+
+		/*
+		 * If we already called setplane while the crtc was disabled,
+		 * we may have an fb pinned; unpin it.
+		 */
+		if (plane->fb)
+			intel_unpin_fb_obj(old_obj);
+
+		i915_gem_track_fb(old_obj, obj,
+				  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
+
+		/* Pin and return without programming hardware */
+		ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
+		mutex_unlock(&dev->struct_mutex);
+
+		return ret;
+	}
+
+	intel_crtc_wait_for_pending_flips(crtc);
+
+	/*
+	 * If clipping results in a non-visible primary plane, we'll disable
+	 * the primary plane.  Note that this is a bit different than what
+	 * happens if userspace explicitly disables the plane by passing fb=0
+	 * because plane->fb still gets set and pinned.
+	 */
+	if (!visible) {
+		mutex_lock(&dev->struct_mutex);
+
+		/*
+		 * Try to pin the new fb first so that we can bail out if we
+		 * fail.
+		 */
+		if (plane->fb != fb) {
+			ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
+			if (ret) {
+				mutex_unlock(&dev->struct_mutex);
+				return ret;
+			}
+		}
+
+		i915_gem_track_fb(old_obj, obj,
+				  INTEL_FRONTBUFFER_PRIMARY(intel_crtc->pipe));
+
+		if (intel_crtc->primary_enabled)
+			intel_disable_primary_hw_plane(plane, crtc);
+
+
+		if (plane->fb != fb)
+			if (plane->fb)
+				intel_unpin_fb_obj(old_obj);
+
+		mutex_unlock(&dev->struct_mutex);
+
+	} else {
+		if (intel_crtc && intel_crtc->active &&
+		    intel_crtc->primary_enabled) {
+			/*
+			 * FBC does not work on some platforms for rotated
+			 * planes, so disable it when rotation is not 0 and
+			 * update it when rotation is set back to 0.
+			 *
+			 * FIXME: This is redundant with the fbc update done in
+			 * the primary plane enable function except that that
+			 * one is done too late. We eventually need to unify
+			 * this.
+			 */
+			if (INTEL_INFO(dev)->gen <= 4 && !IS_G4X(dev) &&
+			    dev_priv->fbc.plane == intel_crtc->plane &&
+			    intel_plane->rotation != BIT(DRM_ROTATE_0)) {
+				intel_disable_fbc(dev);
+			}
+		}
+		ret = intel_pipe_set_base(crtc, src.x1, src.y1, fb);
+		if (ret)
+			return ret;
+
+		if (!intel_crtc->primary_enabled)
+			intel_enable_primary_hw_plane(plane, crtc);
+	}
+
+	intel_plane->crtc_x = orig.crtc_x;
+	intel_plane->crtc_y = orig.crtc_y;
+	intel_plane->crtc_w = orig.crtc_w;
+	intel_plane->crtc_h = orig.crtc_h;
+	intel_plane->src_x = orig.src_x;
+	intel_plane->src_y = orig.src_y;
+	intel_plane->src_w = orig.src_w;
+	intel_plane->src_h = orig.src_h;
+	intel_plane->obj = obj;
+
+	return 0;
+}
+
+/* Common destruction function for both primary and cursor planes */
+static void intel_plane_destroy(struct drm_plane *plane)
+{
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	drm_plane_cleanup(plane);
+	kfree(intel_plane);
+}
+
+static const struct drm_plane_funcs intel_primary_plane_funcs = {
+	.update_plane = intel_primary_plane_setplane,
+	.disable_plane = intel_primary_plane_disable,
+	.destroy = intel_plane_destroy,
+	.set_property = intel_plane_set_property
+};
+
+static struct drm_plane *intel_primary_plane_create(struct drm_device *dev,
+						    int pipe)
+{
+	struct intel_plane *primary;
+	const uint32_t *intel_primary_formats;
+	int num_formats;
+
+	primary = kzalloc(sizeof(*primary), GFP_KERNEL);
+	if (primary == NULL)
+		return NULL;
+
+	primary->can_scale = false;
+	primary->max_downscale = 1;
+	primary->pipe = pipe;
+	primary->plane = pipe;
+	primary->rotation = BIT(DRM_ROTATE_0);
+	if (HAS_FBC(dev) && INTEL_INFO(dev)->gen < 4)
+		primary->plane = !pipe;
+
+	if (INTEL_INFO(dev)->gen <= 3) {
+		intel_primary_formats = intel_primary_formats_gen2;
+		num_formats = ARRAY_SIZE(intel_primary_formats_gen2);
+	} else {
+		intel_primary_formats = intel_primary_formats_gen4;
+		num_formats = ARRAY_SIZE(intel_primary_formats_gen4);
+	}
+
+	drm_universal_plane_init(dev, &primary->base, 0,
+				 &intel_primary_plane_funcs,
+				 intel_primary_formats, num_formats,
+				 DRM_PLANE_TYPE_PRIMARY);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		if (!dev->mode_config.rotation_property)
+			dev->mode_config.rotation_property =
+				drm_mode_create_rotation_property(dev,
+							BIT(DRM_ROTATE_0) |
+							BIT(DRM_ROTATE_180));
+		if (dev->mode_config.rotation_property)
+			drm_object_attach_property(&primary->base.base,
+				dev->mode_config.rotation_property,
+				primary->rotation);
+	}
+
+	return &primary->base;
+}
+
+static int
+intel_cursor_plane_disable(struct drm_plane *plane)
+{
+	if (!plane->fb)
+		return 0;
+
+	BUG_ON(!plane->crtc);
+
+	return intel_crtc_cursor_set_obj(plane->crtc, NULL, 0, 0);
+}
+
+static int
+intel_cursor_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+			  struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+			  unsigned int crtc_w, unsigned int crtc_h,
+			  uint32_t src_x, uint32_t src_y,
+			  uint32_t src_w, uint32_t src_h)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_rect dest = {
+		/* integer pixels */
+		.x1 = crtc_x,
+		.y1 = crtc_y,
+		.x2 = crtc_x + crtc_w,
+		.y2 = crtc_y + crtc_h,
+	};
+	struct drm_rect src = {
+		/* 16.16 fixed point */
+		.x1 = src_x,
+		.y1 = src_y,
+		.x2 = src_x + src_w,
+		.y2 = src_y + src_h,
+	};
+	const struct drm_rect clip = {
+		/* integer pixels */
+		.x2 = intel_crtc->active ? intel_crtc->config.pipe_src_w : 0,
+		.y2 = intel_crtc->active ? intel_crtc->config.pipe_src_h : 0,
+	};
+	bool visible;
+	int ret;
+
+	ret = drm_plane_helper_check_update(plane, crtc, fb,
+					    &src, &dest, &clip,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    DRM_PLANE_HELPER_NO_SCALING,
+					    true, true, &visible);
+	if (ret)
+		return ret;
+
+	crtc->cursor_x = crtc_x;
+	crtc->cursor_y = crtc_y;
+	if (fb != crtc->cursor->fb) {
+		return intel_crtc_cursor_set_obj(crtc, obj, crtc_w, crtc_h);
+	} else {
+		intel_crtc_update_cursor(crtc, visible);
+
+		intel_frontbuffer_flip(crtc->dev,
+				       INTEL_FRONTBUFFER_CURSOR(intel_crtc->pipe));
+
+		return 0;
+	}
+}
+static const struct drm_plane_funcs intel_cursor_plane_funcs = {
+	.update_plane = intel_cursor_plane_update,
+	.disable_plane = intel_cursor_plane_disable,
+	.destroy = intel_plane_destroy,
+};
+
+static struct drm_plane *intel_cursor_plane_create(struct drm_device *dev,
+						   int pipe)
+{
+	struct intel_plane *cursor;
+
+	cursor = kzalloc(sizeof(*cursor), GFP_KERNEL);
+	if (cursor == NULL)
+		return NULL;
+
+	cursor->can_scale = false;
+	cursor->max_downscale = 1;
+	cursor->pipe = pipe;
+	cursor->plane = pipe;
+
+	drm_universal_plane_init(dev, &cursor->base, 0,
+				 &intel_cursor_plane_funcs,
+				 intel_cursor_formats,
+				 ARRAY_SIZE(intel_cursor_formats),
+				 DRM_PLANE_TYPE_CURSOR);
+	return &cursor->base;
+}
+
 static void intel_crtc_init(struct drm_device *dev, int pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc;
-	int i;
+	struct drm_plane *primary = NULL;
+	struct drm_plane *cursor = NULL;
+	int i, ret;
 
 	intel_crtc = kzalloc(sizeof(*intel_crtc), GFP_KERNEL);
 	if (intel_crtc == NULL)
 		return;
 
-	drm_crtc_init(dev, &intel_crtc->base, &intel_crtc_funcs);
+	primary = intel_primary_plane_create(dev, pipe);
+	if (!primary)
+		goto fail;
+
+	cursor = intel_cursor_plane_create(dev, pipe);
+	if (!cursor)
+		goto fail;
+
+	ret = drm_crtc_init_with_planes(dev, &intel_crtc->base, primary,
+					cursor, &intel_crtc_funcs);
+	if (ret)
+		goto fail;
 
 	drm_mode_crtc_set_gamma_size(&intel_crtc->base, 256);
 	for (i = 0; i < 256; i++) {
@@ -10980,7 +12158,7 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 
 	/*
 	 * On gen2/3 only plane A can do fbc, but the panel fitter and lvds port
-	 * is hooked to plane B. Hence we want plane A feeding pipe B.
+	 * is hooked to pipe B. Hence we want plane A feeding pipe B.
 	 */
 	intel_crtc->pipe = pipe;
 	intel_crtc->plane = pipe;
@@ -10991,8 +12169,7 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 
 	intel_crtc->cursor_base = ~0;
 	intel_crtc->cursor_cntl = ~0;
-
-	init_waitqueue_head(&intel_crtc->vbl_wait);
+	intel_crtc->cursor_size = ~0;
 
 	BUG_ON(pipe >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) ||
 	       dev_priv->plane_to_crtc_mapping[intel_crtc->plane] != NULL);
@@ -11002,6 +12179,14 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	drm_crtc_helper_add(&intel_crtc->base, &intel_helper_funcs);
 
 	WARN_ON(drm_crtc_index(&intel_crtc->base) != intel_crtc->pipe);
+	return;
+
+fail:
+	if (primary)
+		drm_plane_cleanup(primary);
+	if (cursor)
+		drm_plane_cleanup(cursor);
+	kfree(intel_crtc);
 }
 
 enum pipe intel_get_pipe_from_connector(struct intel_connector *connector)
@@ -11021,21 +12206,20 @@ int intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data,
 				struct drm_file *file)
 {
 	struct drm_i915_get_pipe_from_crtc_id *pipe_from_crtc_id = data;
-	struct drm_mode_object *drmmode_obj;
+	struct drm_crtc *drmmode_crtc;
 	struct intel_crtc *crtc;
 
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		return -ENODEV;
 
-	drmmode_obj = drm_mode_object_find(dev, pipe_from_crtc_id->crtc_id,
-			DRM_MODE_OBJECT_CRTC);
+	drmmode_crtc = drm_crtc_find(dev, pipe_from_crtc_id->crtc_id);
 
-	if (!drmmode_obj) {
+	if (!drmmode_crtc) {
 		DRM_ERROR("no such CRTC id\n");
 		return -ENOENT;
 	}
 
-	crtc = to_intel_crtc(obj_to_crtc(drmmode_obj));
+	crtc = to_intel_crtc(drmmode_crtc);
 	pipe_from_crtc_id->pipe = crtc->pipe;
 
 	return 0;
@@ -11048,8 +12232,7 @@ static int intel_encoder_clones(struct intel_encoder *encoder)
 	int index_mask = 0;
 	int entry = 0;
 
-	list_for_each_entry(source_encoder,
-			    &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, source_encoder) {
 		if (encoders_cloneable(encoder, source_encoder))
 			index_mask |= (1 << entry);
 
@@ -11171,27 +12354,36 @@ static void intel_setup_outputs(struct drm_device *dev)
 		if (I915_READ(PCH_DP_D) & DP_DETECTED)
 			intel_dp_init(dev, PCH_DP_D, PORT_D);
 	} else if (IS_VALLEYVIEW(dev)) {
-		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIB) & SDVO_DETECTED) {
+		/*
+		 * The DP_DETECTED bit is the latched state of the DDC
+		 * SDA pin at boot. However since eDP doesn't require DDC
+		 * (no way to plug in a DP->HDMI dongle) the DDC pins for
+		 * eDP ports may have been muxed to an alternate function.
+		 * Thus we can't rely on the DP_DETECTED bit alone to detect
+		 * eDP ports. Consult the VBT as well as DP_DETECTED to
+		 * detect eDP ports.
+		 */
+		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIB) & SDVO_DETECTED)
 			intel_hdmi_init(dev, VLV_DISPLAY_BASE + GEN4_HDMIB,
 					PORT_B);
-			if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED)
-				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
-		}
+		if (I915_READ(VLV_DISPLAY_BASE + DP_B) & DP_DETECTED ||
+		    intel_dp_is_edp(dev, PORT_B))
+			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_B, PORT_B);
 
-		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIC) & SDVO_DETECTED) {
+		if (I915_READ(VLV_DISPLAY_BASE + GEN4_HDMIC) & SDVO_DETECTED)
 			intel_hdmi_init(dev, VLV_DISPLAY_BASE + GEN4_HDMIC,
 					PORT_C);
-			if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED)
-				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
-		}
+		if (I915_READ(VLV_DISPLAY_BASE + DP_C) & DP_DETECTED ||
+		    intel_dp_is_edp(dev, PORT_C))
+			intel_dp_init(dev, VLV_DISPLAY_BASE + DP_C, PORT_C);
 
 		if (IS_CHERRYVIEW(dev)) {
-			if (I915_READ(VLV_DISPLAY_BASE + CHV_HDMID) & SDVO_DETECTED) {
+			if (I915_READ(VLV_DISPLAY_BASE + CHV_HDMID) & SDVO_DETECTED)
 				intel_hdmi_init(dev, VLV_DISPLAY_BASE + CHV_HDMID,
 						PORT_D);
-				if (I915_READ(VLV_DISPLAY_BASE + DP_D) & DP_DETECTED)
-					intel_dp_init(dev, VLV_DISPLAY_BASE + DP_D, PORT_D);
-			}
+			/* eDP not supported on port D, so don't check VBT */
+			if (I915_READ(VLV_DISPLAY_BASE + DP_D) & DP_DETECTED)
+				intel_dp_init(dev, VLV_DISPLAY_BASE + DP_D, PORT_D);
 		}
 
 		intel_dsi_init(dev);
@@ -11236,7 +12428,9 @@ static void intel_setup_outputs(struct drm_device *dev)
 	if (SUPPORTS_TV(dev))
 		intel_tv_init(dev);
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	intel_edp_psr_init(dev);
+
+	for_each_intel_encoder(dev, encoder) {
 		encoder->base.possible_crtcs = encoder->crtc_mask;
 		encoder->base.possible_clones =
 			intel_encoder_clones(encoder);
@@ -11249,11 +12443,14 @@ static void intel_setup_outputs(struct drm_device *dev)
 
 static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb)
 {
+	struct drm_device *dev = fb->dev;
 	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
 
 	drm_framebuffer_cleanup(fb);
+	mutex_lock(&dev->struct_mutex);
 	WARN_ON(!intel_fb->obj->framebuffer_references--);
-	drm_gem_object_unreference_unlocked(&intel_fb->obj->base);
+	drm_gem_object_unreference(&intel_fb->obj->base);
+	mutex_unlock(&dev->struct_mutex);
 	kfree(intel_fb);
 }
 
@@ -11438,7 +12635,7 @@ static void intel_init_display(struct drm_device *dev)
 		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
 		dev_priv->display.crtc_enable = haswell_crtc_enable;
 		dev_priv->display.crtc_disable = haswell_crtc_disable;
-		dev_priv->display.off = haswell_crtc_off;
+		dev_priv->display.off = ironlake_crtc_off;
 		dev_priv->display.update_primary_plane =
 			ironlake_update_primary_plane;
 	} else if (HAS_PCH_SPLIT(dev)) {
@@ -11499,29 +12696,27 @@ static void intel_init_display(struct drm_device *dev)
 		dev_priv->display.get_display_clock_speed =
 			i830_get_display_clock_speed;
 
-	if (HAS_PCH_SPLIT(dev)) {
-		if (IS_GEN5(dev)) {
-			dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-		} else if (IS_GEN6(dev)) {
-			dev_priv->display.fdi_link_train = gen6_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-			dev_priv->display.modeset_global_resources =
-				snb_modeset_global_resources;
-		} else if (IS_IVYBRIDGE(dev)) {
-			/* FIXME: detect B0+ stepping and use auto training */
-			dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
-			dev_priv->display.write_eld = ironlake_write_eld;
-			dev_priv->display.modeset_global_resources =
-				ivb_modeset_global_resources;
-		} else if (IS_HASWELL(dev) || IS_GEN8(dev)) {
-			dev_priv->display.fdi_link_train = hsw_fdi_link_train;
-			dev_priv->display.write_eld = haswell_write_eld;
-			dev_priv->display.modeset_global_resources =
-				haswell_modeset_global_resources;
-		}
-	} else if (IS_G4X(dev)) {
+	if (IS_G4X(dev)) {
 		dev_priv->display.write_eld = g4x_write_eld;
+	} else if (IS_GEN5(dev)) {
+		dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
+		dev_priv->display.write_eld = ironlake_write_eld;
+	} else if (IS_GEN6(dev)) {
+		dev_priv->display.fdi_link_train = gen6_fdi_link_train;
+		dev_priv->display.write_eld = ironlake_write_eld;
+		dev_priv->display.modeset_global_resources =
+			snb_modeset_global_resources;
+	} else if (IS_IVYBRIDGE(dev)) {
+		/* FIXME: detect B0+ stepping and use auto training */
+		dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
+		dev_priv->display.write_eld = ironlake_write_eld;
+		dev_priv->display.modeset_global_resources =
+			ivb_modeset_global_resources;
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+		dev_priv->display.fdi_link_train = hsw_fdi_link_train;
+		dev_priv->display.write_eld = haswell_write_eld;
+		dev_priv->display.modeset_global_resources =
+			haswell_modeset_global_resources;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.modeset_global_resources =
 			valleyview_modeset_global_resources;
@@ -11555,6 +12750,8 @@ static void intel_init_display(struct drm_device *dev)
 	}
 
 	intel_panel_init_backlight_funcs(dev);
+
+	mutex_init(&dev_priv->pps_mutex);
 }
 
 /*
@@ -11570,6 +12767,14 @@ static void quirk_pipea_force(struct drm_device *dev)
 	DRM_INFO("applying pipe a force quirk\n");
 }
 
+static void quirk_pipeb_force(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->quirks |= QUIRK_PIPEB_FORCE;
+	DRM_INFO("applying pipe b force quirk\n");
+}
+
 /*
  * Some machines (Lenovo U160) do not work with SSC on LVDS for some reason
  */
@@ -11644,6 +12849,12 @@ static struct intel_quirk intel_quirks[] = {
 	/* ThinkPad T60 needs pipe A force quirk (bug #16494) */
 	{ 0x2782, 0x17aa, 0x201a, quirk_pipea_force },
 
+	/* 830 needs to leave pipe A & dpll A up */
+	{ 0x3577, PCI_ANY_ID, PCI_ANY_ID, quirk_pipea_force },
+
+	/* 830 needs to leave pipe B & dpll B up */
+	{ 0x3577, PCI_ANY_ID, PCI_ANY_ID, quirk_pipeb_force },
+
 	/* Lenovo U160 cannot use SSC on LVDS */
 	{ 0x0046, 0x17aa, 0x3920, quirk_ssc_force_disable },
 
@@ -11671,6 +12882,12 @@ static struct intel_quirk intel_quirks[] = {
 	/* Acer C720 and C720P Chromebooks (Celeron 2955U) have backlights */
 	{ 0x0a06, 0x1025, 0x0a11, quirk_backlight_present },
 
+	/* Acer C720 Chromebook (Core i3 4005U) */
+	{ 0x0a16, 0x1025, 0x0a11, quirk_backlight_present },
+
+	/* Apple Macbook 2,1 (Core 2 T7400) */
+	{ 0x27a2, 0x8086, 0x7270, quirk_backlight_present },
+
 	/* Toshiba CB35 Chromebook (Celeron 2955U) */
 	{ 0x0a06, 0x1179, 0x0a88, quirk_backlight_present },
 
@@ -11714,7 +12931,11 @@ static void i915_disable_vga(struct drm_device *dev)
 	vga_put(dev->pdev, VGA_RSRC_LEGACY_IO);
 	udelay(300);
 
-	I915_WRITE(vga_reg, VGA_DISP_DISABLE);
+	/*
+	 * Fujitsu-Siemens Lifebook S6010 (830) has problems resuming
+	 * from S3 without preserving (some of?) the other bits.
+	 */
+	I915_WRITE(vga_reg, dev_priv->bios_vgacntr | VGA_DISP_DISABLE);
 	POSTING_READ(vga_reg);
 }
 
@@ -11722,9 +12943,10 @@ void intel_modeset_init_hw(struct drm_device *dev)
 {
 	intel_prepare_ddi(dev);
 
-	intel_init_clock_gating(dev);
+	if (IS_VALLEYVIEW(dev))
+		vlv_update_cdclk(dev);
 
-	intel_reset_dpio(dev);
+	intel_init_clock_gating(dev);
 
 	intel_enable_gt_powersave(dev);
 }
@@ -11771,7 +12993,10 @@ void intel_modeset_init(struct drm_device *dev)
 		dev->mode_config.max_height = 8192;
 	}
 
-	if (IS_GEN2(dev)) {
+	if (IS_845G(dev) || IS_I865G(dev)) {
+		dev->mode_config.cursor_width = IS_845G(dev) ? 64 : 512;
+		dev->mode_config.cursor_height = 1023;
+	} else if (IS_GEN2(dev)) {
 		dev->mode_config.cursor_width = GEN2_CURSOR_WIDTH;
 		dev->mode_config.cursor_height = GEN2_CURSOR_HEIGHT;
 	} else {
@@ -11785,7 +13010,7 @@ void intel_modeset_init(struct drm_device *dev)
 		      INTEL_INFO(dev)->num_pipes,
 		      INTEL_INFO(dev)->num_pipes > 1 ? "s" : "");
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		intel_crtc_init(dev, pipe);
 		for_each_sprite(pipe, sprite) {
 			ret = intel_plane_init(dev, pipe, sprite);
@@ -11796,11 +13021,11 @@ void intel_modeset_init(struct drm_device *dev)
 	}
 
 	intel_init_dpio(dev);
-	intel_reset_dpio(dev);
 
-	intel_cpu_pll_init(dev);
 	intel_shared_dpll_init(dev);
 
+	/* save the BIOS value before clobbering it */
+	dev_priv->bios_vgacntr = I915_READ(i915_vgacntrl_reg(dev));
 	/* Just disable it once at startup */
 	i915_disable_vga(dev);
 	intel_setup_outputs(dev);
@@ -11840,7 +13065,7 @@ static void intel_enable_pipe_a(struct drm_device *dev)
 	struct intel_connector *connector;
 	struct drm_connector *crt = NULL;
 	struct intel_load_detect_pipe load_detect_temp;
-	struct drm_modeset_acquire_ctx ctx;
+	struct drm_modeset_acquire_ctx *ctx = dev->mode_config.acquire_ctx;
 
 	/* We can't just switch on the pipe A, we need to set things up with a
 	 * proper mode and output configuration. As a gross hack, enable pipe A
@@ -11857,10 +13082,8 @@ static void intel_enable_pipe_a(struct drm_device *dev)
 	if (!crt)
 		return;
 
-	if (intel_get_load_detect_pipe(crt, NULL, &load_detect_temp, &ctx))
-		intel_release_load_detect_pipe(crt, &load_detect_temp, &ctx);
-
-
+	if (intel_get_load_detect_pipe(crt, NULL, &load_detect_temp, ctx))
+		intel_release_load_detect_pipe(crt, &load_detect_temp);
 }
 
 static bool
@@ -11894,9 +13117,10 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 	I915_WRITE(reg, I915_READ(reg) & ~PIPECONF_FRAME_START_DELAY_MASK);
 
 	/* restore vblank interrupts to correct state */
-	if (crtc->active)
+	if (crtc->active) {
+		update_scanline_offset(crtc);
 		drm_vblank_on(dev, crtc->pipe);
-	else
+	} else
 		drm_vblank_off(dev, crtc->pipe);
 
 	/* We need to sanitize the plane -> pipe mapping first because this will
@@ -11979,7 +13203,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 		}
 	}
 
-	if (crtc->active || IS_VALLEYVIEW(dev) || INTEL_INFO(dev)->gen < 5) {
+	if (crtc->active || HAS_GMCH_DISPLAY(dev)) {
 		/*
 		 * We start out with underrun reporting disabled to avoid races.
 		 * For correct bookkeeping mark this on active crtcs.
@@ -11995,8 +13219,6 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 		 */
 		crtc->cpu_fifo_underrun_disabled = true;
 		crtc->pch_fifo_underrun_disabled = true;
-
-		update_scanline_offset(crtc);
 	}
 }
 
@@ -12024,6 +13246,8 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder)
 				      encoder->base.base.id,
 				      encoder->base.name);
 			encoder->disable(encoder);
+			if (encoder->post_disable)
+				encoder->post_disable(encoder);
 		}
 		encoder->base.crtc = NULL;
 		encoder->connectors_active = false;
@@ -12108,10 +13332,6 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 			      crtc->active ? "enabled" : "disabled");
 	}
 
-	/* FIXME: Smash this into the new shared dpll infrastructure. */
-	if (HAS_DDI(dev))
-		intel_ddi_setup_hw_pll_state(dev);
-
 	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
 		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
 
@@ -12125,10 +13345,12 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 
 		DRM_DEBUG_KMS("%s hw state readout: refcount %i, on %i\n",
 			      pll->name, pll->refcount, pll->on);
+
+		if (pll->refcount)
+			intel_display_power_get(dev_priv, POWER_DOMAIN_PLLS);
 	}
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		pipe = 0;
 
 		if (encoder->get_hw_state(encoder, &pipe)) {
@@ -12192,12 +13414,11 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 	}
 
 	/* HW state is read out, now we need to sanitize this mess. */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		intel_sanitize_encoder(encoder);
 	}
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 		intel_sanitize_crtc(crtc);
 		intel_dump_pipe_config(crtc, &crtc->config, "[setup_hw_state]");
@@ -12225,7 +13446,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 		 * We need to use raw interfaces for restoring state to avoid
 		 * checking (bogus) intermediate states.
 		 */
-		for_each_pipe(pipe) {
+		for_each_pipe(dev_priv, pipe) {
 			struct drm_crtc *crtc =
 				dev_priv->pipe_to_crtc_mapping[pipe];
 
@@ -12242,7 +13463,7 @@ void intel_modeset_setup_hw_state(struct drm_device *dev,
 void intel_modeset_gem_init(struct drm_device *dev)
 {
 	struct drm_crtc *c;
-	struct intel_framebuffer *fb;
+	struct drm_i915_gem_object *obj;
 
 	mutex_lock(&dev->struct_mutex);
 	intel_init_gt_powersave(dev);
@@ -12259,11 +13480,11 @@ void intel_modeset_gem_init(struct drm_device *dev)
 	 */
 	mutex_lock(&dev->struct_mutex);
 	for_each_crtc(dev, c) {
-		if (!c->primary->fb)
+		obj = intel_fb_obj(c->primary->fb);
+		if (obj == NULL)
 			continue;
 
-		fb = to_intel_framebuffer(c->primary->fb);
-		if (intel_pin_and_fence_fb_obj(dev, fb->obj, NULL)) {
+		if (intel_pin_and_fence_fb_obj(dev, obj, NULL)) {
 			DRM_ERROR("failed to pin boot fb on pipe %d\n",
 				  to_intel_crtc(c)->pipe);
 			drm_framebuffer_unreference(c->primary->fb);
@@ -12278,13 +13499,12 @@ void intel_connector_unregister(struct intel_connector *intel_connector)
 	struct drm_connector *connector = &intel_connector->base;
 
 	intel_panel_destroy_backlight(connector);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 }
 
 void intel_modeset_cleanup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc;
 	struct drm_connector *connector;
 
 	/*
@@ -12293,7 +13513,9 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	 * experience fancy races otherwise.
 	 */
 	drm_irq_uninstall(dev);
-	cancel_work_sync(&dev_priv->hotplug_work);
+	intel_hpd_cancel_work(dev_priv);
+	dev_priv->pm._irqs_disabled = true;
+
 	/*
 	 * Due to the hpd irq storm handling the hotplug work can re-arm the
 	 * poll handlers. Hence disable polling after hpd handling is shut down.
@@ -12304,14 +13526,6 @@ void intel_modeset_cleanup(struct drm_device *dev)
 
 	intel_unregister_dsm_handler();
 
-	for_each_crtc(dev, crtc) {
-		/* Skip inactive CRTCs */
-		if (!crtc->primary->fb)
-			continue;
-
-		intel_increase_pllclock(crtc);
-	}
-
 	intel_disable_fbc(dev);
 
 	intel_disable_gt_powersave(dev);
@@ -12453,7 +13667,7 @@ intel_display_capture_error_state(struct drm_device *dev)
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		error->power_well_driver = I915_READ(HSW_PWR_WELL_DRIVER);
 
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		error->pipe[i].power_domain_on =
 			intel_display_power_enabled_unlocked(dev_priv,
 							   POWER_DOMAIN_PIPE(i));
@@ -12479,7 +13693,7 @@ intel_display_capture_error_state(struct drm_device *dev)
 
 		error->pipe[i].source = I915_READ(PIPESRC(i));
 
-		if (!HAS_PCH_SPLIT(dev))
+		if (HAS_GMCH_DISPLAY(dev))
 			error->pipe[i].stat = I915_READ(PIPESTAT(i));
 	}
 
@@ -12517,6 +13731,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
 	if (!error)
@@ -12526,7 +13741,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		err_printf(m, "PWR_WELL_CTL2: %08x\n",
 			   error->power_well_driver);
-	for_each_pipe(i) {
+	for_each_pipe(dev_priv, i) {
 		err_printf(m, "Pipe [%d]:\n", i);
 		err_printf(m, "  Power: %s\n",
 			   error->pipe[i].power_domain_on ? "on" : "off");
@@ -12567,3 +13782,25 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 		err_printf(m, "  VSYNC: %08x\n", error->transcoder[i].vsync);
 	}
 }
+
+void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_unpin_work *work;
+		unsigned long irqflags;
+
+		spin_lock_irqsave(&dev->event_lock, irqflags);
+
+		work = crtc->unpin_work;
+
+		if (work && work->event &&
+		    work->event->base.file_priv == file) {
+			kfree(work->event);
+			work->event = NULL;
+		}
+
+		spin_unlock_irqrestore(&dev->event_lock, irqflags);
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 8a1a4fbc06ac..5ad45bfff3fe 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -111,10 +111,10 @@ static struct intel_dp *intel_attached_dp(struct drm_connector *connector)
 }
 
 static void intel_dp_link_down(struct intel_dp *intel_dp);
-static bool _edp_panel_vdd_on(struct intel_dp *intel_dp);
+static bool edp_panel_vdd_on(struct intel_dp *intel_dp);
 static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
 
-static int
+int
 intel_dp_max_link_bw(struct intel_dp *intel_dp)
 {
 	int max_link_bw = intel_dp->dpcd[DP_MAX_LINK_RATE];
@@ -290,32 +290,201 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 					      struct intel_dp *intel_dp,
 					      struct edp_power_seq *out);
 
+static void pps_lock(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+
+	/*
+	 * See vlv_power_sequencer_reset() why we need
+	 * a power domain reference here.
+	 */
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_get(dev_priv, power_domain);
+
+	mutex_lock(&dev_priv->pps_mutex);
+}
+
+static void pps_unlock(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+
+	mutex_unlock(&dev_priv->pps_mutex);
+
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_put(dev_priv, power_domain);
+}
+
 static enum pipe
 vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct drm_crtc *crtc = intel_dig_port->base.base.crtc;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum port port = intel_dig_port->port;
-	enum pipe pipe;
+	struct intel_encoder *encoder;
+	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
+	struct edp_power_seq power_seq;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (intel_dp->pps_pipe != INVALID_PIPE)
+		return intel_dp->pps_pipe;
+
+	/*
+	 * We don't have power sequencer currently.
+	 * Pick one that's not used by other ports.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		struct intel_dp *tmp;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		tmp = enc_to_intel_dp(&encoder->base);
+
+		if (tmp->pps_pipe != INVALID_PIPE)
+			pipes &= ~(1 << tmp->pps_pipe);
+	}
+
+	/*
+	 * Didn't find one. This should not happen since there
+	 * are two power sequencers and up to two eDP ports.
+	 */
+	if (WARN_ON(pipes == 0))
+		return PIPE_A;
+
+	intel_dp->pps_pipe = ffs(pipes) - 1;
+
+	DRM_DEBUG_KMS("picked pipe %c power sequencer for port %c\n",
+		      pipe_name(intel_dp->pps_pipe),
+		      port_name(intel_dig_port->port));
+
+	/* init power sequencer on this pipe and port */
+	intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
+						      &power_seq);
+
+	return intel_dp->pps_pipe;
+}
+
+typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
+			       enum pipe pipe);
+
+static bool vlv_pipe_has_pp_on(struct drm_i915_private *dev_priv,
+			       enum pipe pipe)
+{
+	return I915_READ(VLV_PIPE_PP_STATUS(pipe)) & PP_ON;
+}
+
+static bool vlv_pipe_has_vdd_on(struct drm_i915_private *dev_priv,
+				enum pipe pipe)
+{
+	return I915_READ(VLV_PIPE_PP_CONTROL(pipe)) & EDP_FORCE_VDD;
+}
+
+static bool vlv_pipe_any(struct drm_i915_private *dev_priv,
+			 enum pipe pipe)
+{
+	return true;
+}
 
-	/* modeset should have pipe */
-	if (crtc)
-		return to_intel_crtc(crtc)->pipe;
+static enum pipe
+vlv_initial_pps_pipe(struct drm_i915_private *dev_priv,
+		     enum port port,
+		     vlv_pipe_check pipe_check)
+{
+	enum pipe pipe;
 
-	/* init time, try to find a pipe with this port selected */
 	for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) {
 		u32 port_sel = I915_READ(VLV_PIPE_PP_ON_DELAYS(pipe)) &
 			PANEL_PORT_SELECT_MASK;
-		if (port_sel == PANEL_PORT_SELECT_DPB_VLV && port == PORT_B)
-			return pipe;
-		if (port_sel == PANEL_PORT_SELECT_DPC_VLV && port == PORT_C)
-			return pipe;
+
+		if (port_sel != PANEL_PORT_SELECT_VLV(port))
+			continue;
+
+		if (!pipe_check(dev_priv, pipe))
+			continue;
+
+		return pipe;
 	}
 
-	/* shrug */
-	return PIPE_A;
+	return INVALID_PIPE;
+}
+
+static void
+vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct edp_power_seq power_seq;
+	enum port port = intel_dig_port->port;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	/* try to find a pipe with this port selected */
+	/* first pick one where the panel is on */
+	intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+						  vlv_pipe_has_pp_on);
+	/* didn't find one? pick one where vdd is on */
+	if (intel_dp->pps_pipe == INVALID_PIPE)
+		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							  vlv_pipe_has_vdd_on);
+	/* didn't find one? pick one with just the correct port */
+	if (intel_dp->pps_pipe == INVALID_PIPE)
+		intel_dp->pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							  vlv_pipe_any);
+
+	/* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */
+	if (intel_dp->pps_pipe == INVALID_PIPE) {
+		DRM_DEBUG_KMS("no initial power sequencer for port %c\n",
+			      port_name(port));
+		return;
+	}
+
+	DRM_DEBUG_KMS("initial power sequencer for port %c: pipe %c\n",
+		      port_name(port), pipe_name(intel_dp->pps_pipe));
+
+	intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
+						      &power_seq);
+}
+
+void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_encoder *encoder;
+
+	if (WARN_ON(!IS_VALLEYVIEW(dev)))
+		return;
+
+	/*
+	 * We can't grab pps_mutex here due to deadlock with power_domain
+	 * mutex when power_domain functions are called while holding pps_mutex.
+	 * That also means that in order to use pps_pipe the code needs to
+	 * hold both a power domain reference and pps_mutex, and the power domain
+	 * reference get/put must be done while _not_ holding pps_mutex.
+	 * pps_{lock,unlock}() do these steps in the correct order, so one
+	 * should use them always.
+	 */
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+		struct intel_dp *intel_dp;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		intel_dp = enc_to_intel_dp(&encoder->base);
+		intel_dp->pps_pipe = INVALID_PIPE;
+	}
 }
 
 static u32 _pp_ctrl_reg(struct intel_dp *intel_dp)
@@ -349,12 +518,15 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp_div;
 	u32 pp_ctrl_reg, pp_div_reg;
-	enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
 
 	if (!is_edp(intel_dp) || code != SYS_RESTART)
 		return 0;
 
+	pps_lock(intel_dp);
+
 	if (IS_VALLEYVIEW(dev)) {
+		enum pipe pipe = vlv_power_sequencer_pipe(intel_dp);
+
 		pp_ctrl_reg = VLV_PIPE_PP_CONTROL(pipe);
 		pp_div_reg  = VLV_PIPE_PP_DIVISOR(pipe);
 		pp_div = I915_READ(pp_div_reg);
@@ -366,6 +538,8 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code,
 		msleep(intel_dp->panel_power_cycle_delay);
 	}
 
+	pps_unlock(intel_dp);
+
 	return 0;
 }
 
@@ -374,6 +548,8 @@ static bool edp_have_panel_power(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	return (I915_READ(_pp_stat_reg(intel_dp)) & PP_ON) != 0;
 }
 
@@ -381,13 +557,10 @@ static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	enum intel_display_power_domain power_domain;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	return intel_display_power_enabled(dev_priv, power_domain) &&
-	       (I915_READ(_pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD) != 0;
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	return I915_READ(_pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD;
 }
 
 static void
@@ -535,7 +708,15 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	bool has_aux_irq = HAS_AUX_IRQ(dev);
 	bool vdd;
 
-	vdd = _edp_panel_vdd_on(intel_dp);
+	pps_lock(intel_dp);
+
+	/*
+	 * We will be called with VDD already enabled for dpcd/edid/oui reads.
+	 * In such cases we want to leave VDD enabled and it's up to upper layers
+	 * to turn it off. But for eg. i2c-dev access we need to turn it on/off
+	 * ourselves.
+	 */
+	vdd = edp_panel_vdd_on(intel_dp);
 
 	/* dp aux is extremely sensitive to irq latency, hence request the
 	 * lowest possible wakeup latency and so prevent the cpu from going into
@@ -644,6 +825,8 @@ out:
 	if (vdd)
 		edp_panel_vdd_off(intel_dp, false);
 
+	pps_unlock(intel_dp);
+
 	return ret;
 }
 
@@ -773,12 +956,29 @@ intel_dp_connector_unregister(struct intel_connector *intel_connector)
 {
 	struct intel_dp *intel_dp = intel_attached_dp(&intel_connector->base);
 
-	sysfs_remove_link(&intel_connector->base.kdev->kobj,
-			  intel_dp->aux.ddc.dev.kobj.name);
+	if (!intel_connector->mst_port)
+		sysfs_remove_link(&intel_connector->base.kdev->kobj,
+				  intel_dp->aux.ddc.dev.kobj.name);
 	intel_connector_unregister(intel_connector);
 }
 
 static void
+hsw_dp_set_ddi_pll_sel(struct intel_crtc_config *pipe_config, int link_bw)
+{
+	switch (link_bw) {
+	case DP_LINK_BW_1_62:
+		pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_810;
+		break;
+	case DP_LINK_BW_2_7:
+		pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_1350;
+		break;
+	case DP_LINK_BW_5_4:
+		pipe_config->ddi_pll_sel = PORT_CLK_SEL_LCPLL_2700;
+		break;
+	}
+}
+
+static void
 intel_dp_set_clock(struct intel_encoder *encoder,
 		   struct intel_crtc_config *pipe_config, int link_bw)
 {
@@ -789,8 +989,6 @@ intel_dp_set_clock(struct intel_encoder *encoder,
 	if (IS_G4X(dev)) {
 		divisor = gen4_dpll;
 		count = ARRAY_SIZE(gen4_dpll);
-	} else if (IS_HASWELL(dev)) {
-		/* Haswell has special-purpose DP DDI clocks. */
 	} else if (HAS_PCH_SPLIT(dev)) {
 		divisor = pch_dpll;
 		count = ARRAY_SIZE(pch_dpll);
@@ -813,20 +1011,6 @@ intel_dp_set_clock(struct intel_encoder *encoder,
 	}
 }
 
-static void
-intel_dp_set_m2_n2(struct intel_crtc *crtc, struct intel_link_m_n *m_n)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum transcoder transcoder = crtc->config.cpu_transcoder;
-
-	I915_WRITE(PIPE_DATA_M2(transcoder),
-		TU_SIZE(m_n->tu) | m_n->gmch_m);
-	I915_WRITE(PIPE_DATA_N2(transcoder), m_n->gmch_n);
-	I915_WRITE(PIPE_LINK_M2(transcoder), m_n->link_m);
-	I915_WRITE(PIPE_LINK_N2(transcoder), m_n->link_n);
-}
-
 bool
 intel_dp_compute_config(struct intel_encoder *encoder,
 			struct intel_crtc_config *pipe_config)
@@ -852,6 +1036,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 		pipe_config->has_pch_encoder = true;
 
 	pipe_config->has_dp_encoder = true;
+	pipe_config->has_drrs = false;
 	pipe_config->has_audio = intel_dp->has_audio;
 
 	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
@@ -883,23 +1068,15 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 			bpp = dev_priv->vbt.edp_bpp;
 		}
 
-		if (IS_BROADWELL(dev)) {
-			/* Yes, it's an ugly hack. */
-			min_lane_count = max_lane_count;
-			DRM_DEBUG_KMS("forcing lane count to max (%u) on BDW\n",
-				      min_lane_count);
-		} else if (dev_priv->vbt.edp_lanes) {
-			min_lane_count = min(dev_priv->vbt.edp_lanes,
-					     max_lane_count);
-			DRM_DEBUG_KMS("using min %u lanes per VBT\n",
-				      min_lane_count);
-		}
-
-		if (dev_priv->vbt.edp_rate) {
-			min_clock = min(dev_priv->vbt.edp_rate >> 3, max_clock);
-			DRM_DEBUG_KMS("using min %02x link bw per VBT\n",
-				      bws[min_clock]);
-		}
+		/*
+		 * Use the maximum clock and number of lanes the eDP panel
+		 * advertizes being capable of. The panels are generally
+		 * designed to support only a single clock and lane
+		 * configuration, and typically these values correspond to the
+		 * native resolution of the panel.
+		 */
+		min_lane_count = max_lane_count;
+		min_clock = max_clock;
 	}
 
 	for (; bpp >= 6*3; bpp -= 2*3) {
@@ -955,13 +1132,17 @@ found:
 
 	if (intel_connector->panel.downclock_mode != NULL &&
 		intel_dp->drrs_state.type == SEAMLESS_DRRS_SUPPORT) {
+			pipe_config->has_drrs = true;
 			intel_link_compute_m_n(bpp, lane_count,
 				intel_connector->panel.downclock_mode->clock,
 				pipe_config->port_clock,
 				&pipe_config->dp_m2_n2);
 	}
 
-	intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw);
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
+		hsw_dp_set_ddi_pll_sel(pipe_config, intel_dp->link_bw);
+	else
+		intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw);
 
 	return true;
 }
@@ -1092,6 +1273,8 @@ static void wait_panel_status(struct intel_dp *intel_dp,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp_stat_reg, pp_ctrl_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	pp_stat_reg = _pp_stat_reg(intel_dp);
 	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
 
@@ -1155,13 +1338,20 @@ static  u32 ironlake_get_pp_control(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 control;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	control = I915_READ(_pp_ctrl_reg(intel_dp));
 	control &= ~PANEL_UNLOCK_MASK;
 	control |= PANEL_UNLOCK_REGS;
 	return control;
 }
 
-static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
+/*
+ * Must be paired with edp_panel_vdd_off().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
+ */
+static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -1172,6 +1362,8 @@ static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
 	u32 pp_stat_reg, pp_ctrl_reg;
 	bool need_to_disable = !intel_dp->want_panel_vdd;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return false;
 
@@ -1209,66 +1401,103 @@ static bool _edp_panel_vdd_on(struct intel_dp *intel_dp)
 	return need_to_disable;
 }
 
+/*
+ * Must be paired with intel_edp_panel_vdd_off() or
+ * intel_edp_panel_off().
+ * Nested calls to these functions are not allowed since
+ * we drop the lock. Caller must use some higher level
+ * locking to prevent nested calls from other threads.
+ */
 void intel_edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
-	if (is_edp(intel_dp)) {
-		bool vdd = _edp_panel_vdd_on(intel_dp);
+	bool vdd;
 
-		WARN(!vdd, "eDP VDD already requested on\n");
-	}
+	if (!is_edp(intel_dp))
+		return;
+
+	pps_lock(intel_dp);
+	vdd = edp_panel_vdd_on(intel_dp);
+	pps_unlock(intel_dp);
+
+	WARN(!vdd, "eDP VDD already requested on\n");
 }
 
 static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *intel_dig_port =
+		dp_to_dig_port(intel_dp);
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	enum intel_display_power_domain power_domain;
 	u32 pp;
 	u32 pp_stat_reg, pp_ctrl_reg;
 
-	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	lockdep_assert_held(&dev_priv->pps_mutex);
 
-	if (!intel_dp->want_panel_vdd && edp_have_panel_vdd(intel_dp)) {
-		struct intel_digital_port *intel_dig_port =
-						dp_to_dig_port(intel_dp);
-		struct intel_encoder *intel_encoder = &intel_dig_port->base;
-		enum intel_display_power_domain power_domain;
+	WARN_ON(intel_dp->want_panel_vdd);
 
-		DRM_DEBUG_KMS("Turning eDP VDD off\n");
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
 
-		pp = ironlake_get_pp_control(intel_dp);
-		pp &= ~EDP_FORCE_VDD;
+	DRM_DEBUG_KMS("Turning eDP VDD off\n");
 
-		pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
-		pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp = ironlake_get_pp_control(intel_dp);
+	pp &= ~EDP_FORCE_VDD;
 
-		I915_WRITE(pp_ctrl_reg, pp);
-		POSTING_READ(pp_ctrl_reg);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp_stat_reg = _pp_stat_reg(intel_dp);
 
-		/* Make sure sequencer is idle before allowing subsequent activity */
-		DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
-		I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
+	I915_WRITE(pp_ctrl_reg, pp);
+	POSTING_READ(pp_ctrl_reg);
 
-		if ((pp & POWER_TARGET_ON) == 0)
-			intel_dp->last_power_cycle = jiffies;
+	/* Make sure sequencer is idle before allowing subsequent activity */
+	DRM_DEBUG_KMS("PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+	I915_READ(pp_stat_reg), I915_READ(pp_ctrl_reg));
 
-		power_domain = intel_display_port_power_domain(intel_encoder);
-		intel_display_power_put(dev_priv, power_domain);
-	}
+	if ((pp & POWER_TARGET_ON) == 0)
+		intel_dp->last_power_cycle = jiffies;
+
+	power_domain = intel_display_port_power_domain(intel_encoder);
+	intel_display_power_put(dev_priv, power_domain);
 }
 
 static void edp_panel_vdd_work(struct work_struct *__work)
 {
 	struct intel_dp *intel_dp = container_of(to_delayed_work(__work),
 						 struct intel_dp, panel_vdd_work);
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	edp_panel_vdd_off_sync(intel_dp);
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	pps_lock(intel_dp);
+	if (!intel_dp->want_panel_vdd)
+		edp_panel_vdd_off_sync(intel_dp);
+	pps_unlock(intel_dp);
+}
+
+static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
+{
+	unsigned long delay;
+
+	/*
+	 * Queue the timer to fire a long time from now (relative to the power
+	 * down delay) to keep the panel power up across a sequence of
+	 * operations.
+	 */
+	delay = msecs_to_jiffies(intel_dp->panel_power_cycle_delay * 5);
+	schedule_delayed_work(&intel_dp->panel_vdd_work, delay);
 }
 
+/*
+ * Must be paired with edp_panel_vdd_on().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_edp_panel_vdd_{on,off}() calls.
+ */
 static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 {
+	struct drm_i915_private *dev_priv =
+		intel_dp_to_dev(intel_dp)->dev_private;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (!is_edp(intel_dp))
 		return;
 
@@ -1276,17 +1505,26 @@ static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
 
 	intel_dp->want_panel_vdd = false;
 
-	if (sync) {
+	if (sync)
 		edp_panel_vdd_off_sync(intel_dp);
-	} else {
-		/*
-		 * Queue the timer to fire a long
-		 * time from now (relative to the power down delay)
-		 * to keep the panel power up across a sequence of operations
-		 */
-		schedule_delayed_work(&intel_dp->panel_vdd_work,
-				      msecs_to_jiffies(intel_dp->panel_power_cycle_delay * 5));
-	}
+	else
+		edp_panel_vdd_schedule_off(intel_dp);
+}
+
+/*
+ * Must be paired with intel_edp_panel_vdd_on().
+ * Nested calls to these functions are not allowed since
+ * we drop the lock. Caller must use some higher level
+ * locking to prevent nested calls from other threads.
+ */
+static void intel_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	pps_lock(intel_dp);
+	edp_panel_vdd_off(intel_dp, sync);
+	pps_unlock(intel_dp);
 }
 
 void intel_edp_panel_on(struct intel_dp *intel_dp)
@@ -1301,9 +1539,11 @@ void intel_edp_panel_on(struct intel_dp *intel_dp)
 
 	DRM_DEBUG_KMS("Turn eDP power on\n");
 
+	pps_lock(intel_dp);
+
 	if (edp_have_panel_power(intel_dp)) {
 		DRM_DEBUG_KMS("eDP power already on\n");
-		return;
+		goto out;
 	}
 
 	wait_panel_power_cycle(intel_dp);
@@ -1332,6 +1572,9 @@ void intel_edp_panel_on(struct intel_dp *intel_dp)
 		I915_WRITE(pp_ctrl_reg, pp);
 		POSTING_READ(pp_ctrl_reg);
 	}
+
+ out:
+	pps_unlock(intel_dp);
 }
 
 void intel_edp_panel_off(struct intel_dp *intel_dp)
@@ -1349,7 +1592,7 @@ void intel_edp_panel_off(struct intel_dp *intel_dp)
 
 	DRM_DEBUG_KMS("Turn eDP power off\n");
 
-	edp_wait_backlight_off(intel_dp);
+	pps_lock(intel_dp);
 
 	WARN(!intel_dp->want_panel_vdd, "Need VDD to turn off panel\n");
 
@@ -1372,9 +1615,12 @@ void intel_edp_panel_off(struct intel_dp *intel_dp)
 	/* We got a reference when we enabled the VDD. */
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_put(dev_priv, power_domain);
+
+	pps_unlock(intel_dp);
 }
 
-void intel_edp_backlight_on(struct intel_dp *intel_dp)
+/* Enable backlight in the panel power control. */
+static void _intel_edp_backlight_on(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
@@ -1382,10 +1628,6 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 	u32 pp;
 	u32 pp_ctrl_reg;
 
-	if (!is_edp(intel_dp))
-		return;
-
-	DRM_DEBUG_KMS("\n");
 	/*
 	 * If we enable the backlight right away following a panel power
 	 * on, we may see slight flicker as the panel syncs with the eDP
@@ -1393,6 +1635,9 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 	 * allowing it to appear.
 	 */
 	wait_backlight_on(intel_dp);
+
+	pps_lock(intel_dp);
+
 	pp = ironlake_get_pp_control(intel_dp);
 	pp |= EDP_BLC_ENABLE;
 
@@ -1401,10 +1646,23 @@ void intel_edp_backlight_on(struct intel_dp *intel_dp)
 	I915_WRITE(pp_ctrl_reg, pp);
 	POSTING_READ(pp_ctrl_reg);
 
+	pps_unlock(intel_dp);
+}
+
+/* Enable backlight PWM and backlight PP control. */
+void intel_edp_backlight_on(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	DRM_DEBUG_KMS("\n");
+
 	intel_panel_enable_backlight(intel_dp->attached_connector);
+	_intel_edp_backlight_on(intel_dp);
 }
 
-void intel_edp_backlight_off(struct intel_dp *intel_dp)
+/* Disable backlight in the panel power control. */
+static void _intel_edp_backlight_off(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1414,9 +1672,8 @@ void intel_edp_backlight_off(struct intel_dp *intel_dp)
 	if (!is_edp(intel_dp))
 		return;
 
-	intel_panel_disable_backlight(intel_dp->attached_connector);
+	pps_lock(intel_dp);
 
-	DRM_DEBUG_KMS("\n");
 	pp = ironlake_get_pp_control(intel_dp);
 	pp &= ~EDP_BLC_ENABLE;
 
@@ -1424,7 +1681,49 @@ void intel_edp_backlight_off(struct intel_dp *intel_dp)
 
 	I915_WRITE(pp_ctrl_reg, pp);
 	POSTING_READ(pp_ctrl_reg);
+
+	pps_unlock(intel_dp);
+
 	intel_dp->last_backlight_off = jiffies;
+	edp_wait_backlight_off(intel_dp);
+}
+
+/* Disable backlight PP control and backlight PWM. */
+void intel_edp_backlight_off(struct intel_dp *intel_dp)
+{
+	if (!is_edp(intel_dp))
+		return;
+
+	DRM_DEBUG_KMS("\n");
+
+	_intel_edp_backlight_off(intel_dp);
+	intel_panel_disable_backlight(intel_dp->attached_connector);
+}
+
+/*
+ * Hook for controlling the panel power control backlight through the bl_power
+ * sysfs attribute. Take care to handle multiple calls.
+ */
+static void intel_edp_backlight_power(struct intel_connector *connector,
+				      bool enable)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(&connector->base);
+	bool is_enabled;
+
+	pps_lock(intel_dp);
+	is_enabled = ironlake_get_pp_control(intel_dp) & EDP_BLC_ENABLE;
+	pps_unlock(intel_dp);
+
+	if (is_enabled == enable)
+		return;
+
+	DRM_DEBUG_KMS("panel power control backlight %s\n",
+		      enable ? "enable" : "disable");
+
+	if (enable)
+		_intel_edp_backlight_on(intel_dp);
+	else
+		_intel_edp_backlight_off(intel_dp);
 }
 
 static void ironlake_edp_pll_on(struct intel_dp *intel_dp)
@@ -1490,8 +1789,6 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
 	if (mode != DRM_MODE_DPMS_ON) {
 		ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER,
 					 DP_SET_POWER_D3);
-		if (ret != 1)
-			DRM_DEBUG_DRIVER("failed to write sink power state\n");
 	} else {
 		/*
 		 * When turning on, we need to retry for 1ms to give the sink
@@ -1505,6 +1802,10 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
 			msleep(1);
 		}
 	}
+
+	if (ret != 1)
+		DRM_DEBUG_KMS("failed to %s sink power state\n",
+			      mode == DRM_MODE_DPMS_ON ? "enable" : "disable");
 }
 
 static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
@@ -1551,7 +1852,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 			return true;
 		}
 
-		for_each_pipe(i) {
+		for_each_pipe(dev_priv, i) {
 			trans_dp = I915_READ(TRANS_DP_CTL(i));
 			if ((trans_dp & TRANS_DP_PORT_SEL_MASK) == trans_sel) {
 				*pipe = i;
@@ -1606,6 +1907,10 @@ static void intel_dp_get_config(struct intel_encoder *encoder,
 
 	pipe_config->adjusted_mode.flags |= flags;
 
+	if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev) &&
+	    tmp & DP_COLOR_RANGE_16_235)
+		pipe_config->limited_color_range = true;
+
 	pipe_config->has_dp_encoder = true;
 
 	intel_dp_get_m_n(crtc, pipe_config);
@@ -1646,11 +1951,9 @@ static void intel_dp_get_config(struct intel_encoder *encoder,
 	}
 }
 
-static bool is_edp_psr(struct drm_device *dev)
+static bool is_edp_psr(struct intel_dp *intel_dp)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	return dev_priv->psr.sink_support;
+	return intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED;
 }
 
 static bool intel_edp_is_psr_enabled(struct drm_device *dev)
@@ -1698,9 +2001,6 @@ static void intel_edp_psr_setup(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct edp_vsc_psr psr_vsc;
 
-	if (intel_dp->psr_setup_done)
-		return;
-
 	/* Prepare VSC packet as per EDP 1.3 spec, Table 3.10 */
 	memset(&psr_vsc, 0, sizeof(psr_vsc));
 	psr_vsc.sdp_header.HB0 = 0;
@@ -1712,22 +2012,25 @@ static void intel_edp_psr_setup(struct intel_dp *intel_dp)
 	/* Avoid continuous PSR exit by masking memup and hpd */
 	I915_WRITE(EDP_PSR_DEBUG_CTL(dev), EDP_PSR_DEBUG_MASK_MEMUP |
 		   EDP_PSR_DEBUG_MASK_HPD | EDP_PSR_DEBUG_MASK_LPSP);
-
-	intel_dp->psr_setup_done = true;
 }
 
 static void intel_edp_psr_enable_sink(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t aux_clock_divider;
 	int precharge = 0x3;
 	int msg_size = 5;       /* Header(4) + Message(1) */
+	bool only_standby = false;
 
 	aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0);
 
+	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
+		only_standby = true;
+
 	/* Enable PSR in sink */
-	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT)
+	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby)
 		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG,
 				   DP_PSR_ENABLE & ~DP_PSR_MAIN_LINK_ACTIVE);
 	else
@@ -1746,18 +2049,24 @@ static void intel_edp_psr_enable_sink(struct intel_dp *intel_dp)
 
 static void intel_edp_psr_enable_source(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t max_sleep_time = 0x1f;
 	uint32_t idle_frames = 1;
 	uint32_t val = 0x0;
 	const uint32_t link_entry_time = EDP_PSR_MIN_LINK_ENTRY_TIME_8_LINES;
+	bool only_standby = false;
 
-	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT) {
+	if (IS_BROADWELL(dev) && dig_port->port != PORT_A)
+		only_standby = true;
+
+	if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT || only_standby) {
 		val |= EDP_PSR_LINK_STANDBY;
 		val |= EDP_PSR_TP2_TP3_TIME_0us;
 		val |= EDP_PSR_TP1_TIME_0us;
 		val |= EDP_PSR_SKIP_AUX_EXIT;
+		val |= IS_BROADWELL(dev) ? BDW_PSR_SINGLE_FRAME : 0;
 	} else
 		val |= EDP_PSR_LINK_DISABLE;
 
@@ -1775,18 +2084,14 @@ static bool intel_edp_psr_match_conditions(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc *crtc = dig_port->base.base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_i915_gem_object *obj = to_intel_framebuffer(crtc->primary->fb)->obj;
-	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 
-	dev_priv->psr.source_ok = false;
+	lockdep_assert_held(&dev_priv->psr.lock);
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+	WARN_ON(!drm_modeset_is_locked(&crtc->mutex));
 
-	if (!HAS_PSR(dev)) {
-		DRM_DEBUG_KMS("PSR not supported on this platform\n");
-		return false;
-	}
+	dev_priv->psr.source_ok = false;
 
-	if ((intel_encoder->type != INTEL_OUTPUT_EDP) ||
-	    (dig_port->port != PORT_A)) {
+	if (IS_HASWELL(dev) && dig_port->port != PORT_A) {
 		DRM_DEBUG_KMS("HSW ties PSR to DDI A (eDP)\n");
 		return false;
 	}
@@ -1796,29 +2101,9 @@ static bool intel_edp_psr_match_conditions(struct intel_dp *intel_dp)
 		return false;
 	}
 
-	crtc = dig_port->base.base.crtc;
-	if (crtc == NULL) {
-		DRM_DEBUG_KMS("crtc not active for PSR\n");
-		return false;
-	}
-
-	intel_crtc = to_intel_crtc(crtc);
-	if (!intel_crtc_active(crtc)) {
-		DRM_DEBUG_KMS("crtc not active for PSR\n");
-		return false;
-	}
-
-	obj = to_intel_framebuffer(crtc->primary->fb)->obj;
-	if (obj->tiling_mode != I915_TILING_X ||
-	    obj->fence_reg == I915_FENCE_REG_NONE) {
-		DRM_DEBUG_KMS("PSR condition failed: fb not tiled or fenced\n");
-		return false;
-	}
-
-	if (I915_READ(SPRCTL(intel_crtc->pipe)) & SPRITE_ENABLE) {
-		DRM_DEBUG_KMS("PSR condition failed: Sprite is Enabled\n");
-		return false;
-	}
+	/* Below limitations aren't valid for Broadwell */
+	if (IS_BROADWELL(dev))
+		goto out;
 
 	if (I915_READ(HSW_STEREO_3D_CTL(intel_crtc->config.cpu_transcoder)) &
 	    S3D_ENABLE) {
@@ -1831,35 +2116,60 @@ static bool intel_edp_psr_match_conditions(struct intel_dp *intel_dp)
 		return false;
 	}
 
+ out:
 	dev_priv->psr.source_ok = true;
 	return true;
 }
 
 static void intel_edp_psr_do_enable(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
-
-	if (!intel_edp_psr_match_conditions(intel_dp) ||
-	    intel_edp_is_psr_enabled(dev))
-		return;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/* Setup PSR once */
-	intel_edp_psr_setup(intel_dp);
+	WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
+	WARN_ON(dev_priv->psr.active);
+	lockdep_assert_held(&dev_priv->psr.lock);
 
 	/* Enable PSR on the panel */
 	intel_edp_psr_enable_sink(intel_dp);
 
 	/* Enable PSR on the host */
 	intel_edp_psr_enable_source(intel_dp);
+
+	dev_priv->psr.active = true;
 }
 
 void intel_edp_psr_enable(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!HAS_PSR(dev)) {
+		DRM_DEBUG_KMS("PSR not supported on this platform\n");
+		return;
+	}
 
-	if (intel_edp_psr_match_conditions(intel_dp) &&
-	    !intel_edp_is_psr_enabled(dev))
-		intel_edp_psr_do_enable(intel_dp);
+	if (!is_edp_psr(intel_dp)) {
+		DRM_DEBUG_KMS("PSR not supported by this panel\n");
+		return;
+	}
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (dev_priv->psr.enabled) {
+		DRM_DEBUG_KMS("PSR already in use\n");
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	dev_priv->psr.busy_frontbuffer_bits = 0;
+
+	/* Setup PSR once */
+	intel_edp_psr_setup(intel_dp);
+
+	if (intel_edp_psr_match_conditions(intel_dp))
+		dev_priv->psr.enabled = intel_dp;
+	mutex_unlock(&dev_priv->psr.lock);
 }
 
 void intel_edp_psr_disable(struct intel_dp *intel_dp)
@@ -1867,42 +2177,141 @@ void intel_edp_psr_disable(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (!intel_edp_is_psr_enabled(dev))
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
 		return;
+	}
+
+	if (dev_priv->psr.active) {
+		I915_WRITE(EDP_PSR_CTL(dev),
+			   I915_READ(EDP_PSR_CTL(dev)) & ~EDP_PSR_ENABLE);
 
-	I915_WRITE(EDP_PSR_CTL(dev),
-		   I915_READ(EDP_PSR_CTL(dev)) & ~EDP_PSR_ENABLE);
+		/* Wait till PSR is idle */
+		if (_wait_for((I915_READ(EDP_PSR_STATUS_CTL(dev)) &
+			       EDP_PSR_STATUS_STATE_MASK) == 0, 2000, 10))
+			DRM_ERROR("Timed out waiting for PSR Idle State\n");
+
+		dev_priv->psr.active = false;
+	} else {
+		WARN_ON(I915_READ(EDP_PSR_CTL(dev)) & EDP_PSR_ENABLE);
+	}
 
-	/* Wait till PSR is idle */
-	if (_wait_for((I915_READ(EDP_PSR_STATUS_CTL(dev)) &
-		       EDP_PSR_STATUS_STATE_MASK) == 0, 2000, 10))
-		DRM_ERROR("Timed out waiting for PSR Idle State\n");
+	dev_priv->psr.enabled = NULL;
+	mutex_unlock(&dev_priv->psr.lock);
+
+	cancel_delayed_work_sync(&dev_priv->psr.work);
 }
 
-void intel_edp_psr_update(struct drm_device *dev)
+static void intel_edp_psr_work(struct work_struct *work)
 {
-	struct intel_encoder *encoder;
-	struct intel_dp *intel_dp = NULL;
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv), psr.work.work);
+	struct intel_dp *intel_dp = dev_priv->psr.enabled;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head)
-		if (encoder->type == INTEL_OUTPUT_EDP) {
-			intel_dp = enc_to_intel_dp(&encoder->base);
+	mutex_lock(&dev_priv->psr.lock);
+	intel_dp = dev_priv->psr.enabled;
 
-			if (!is_edp_psr(dev))
-				return;
+	if (!intel_dp)
+		goto unlock;
 
-			if (!intel_edp_psr_match_conditions(intel_dp))
-				intel_edp_psr_disable(intel_dp);
-			else
-				if (!intel_edp_is_psr_enabled(dev))
-					intel_edp_psr_do_enable(intel_dp);
-		}
+	/*
+	 * The delayed work can race with an invalidate hence we need to
+	 * recheck. Since psr_flush first clears this and then reschedules we
+	 * won't ever miss a flush when bailing out here.
+	 */
+	if (dev_priv->psr.busy_frontbuffer_bits)
+		goto unlock;
+
+	intel_edp_psr_do_enable(intel_dp);
+unlock:
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+static void intel_edp_psr_do_exit(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->psr.active) {
+		u32 val = I915_READ(EDP_PSR_CTL(dev));
+
+		WARN_ON(!(val & EDP_PSR_ENABLE));
+
+		I915_WRITE(EDP_PSR_CTL(dev), val & ~EDP_PSR_ENABLE);
+
+		dev_priv->psr.active = false;
+	}
+
+}
+
+void intel_edp_psr_invalidate(struct drm_device *dev,
+			      unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	enum pipe pipe;
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
+	pipe = to_intel_crtc(crtc)->pipe;
+
+	intel_edp_psr_do_exit(dev);
+
+	frontbuffer_bits &= INTEL_FRONTBUFFER_ALL_MASK(pipe);
+
+	dev_priv->psr.busy_frontbuffer_bits |= frontbuffer_bits;
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+void intel_edp_psr_flush(struct drm_device *dev,
+			 unsigned frontbuffer_bits)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	enum pipe pipe;
+
+	mutex_lock(&dev_priv->psr.lock);
+	if (!dev_priv->psr.enabled) {
+		mutex_unlock(&dev_priv->psr.lock);
+		return;
+	}
+
+	crtc = dp_to_dig_port(dev_priv->psr.enabled)->base.base.crtc;
+	pipe = to_intel_crtc(crtc)->pipe;
+	dev_priv->psr.busy_frontbuffer_bits &= ~frontbuffer_bits;
+
+	/*
+	 * On Haswell sprite plane updates don't result in a psr invalidating
+	 * signal in the hardware. Which means we need to manually fake this in
+	 * software for all flushes, not just when we've seen a preceding
+	 * invalidation through frontbuffer rendering.
+	 */
+	if (IS_HASWELL(dev) &&
+	    (frontbuffer_bits & INTEL_FRONTBUFFER_SPRITE(pipe)))
+		intel_edp_psr_do_exit(dev);
+
+	if (!dev_priv->psr.active && !dev_priv->psr.busy_frontbuffer_bits)
+		schedule_delayed_work(&dev_priv->psr.work,
+				      msecs_to_jiffies(100));
+	mutex_unlock(&dev_priv->psr.lock);
+}
+
+void intel_edp_psr_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	INIT_DELAYED_WORK(&dev_priv->psr.work, intel_edp_psr_work);
+	mutex_init(&dev_priv->psr.lock);
 }
 
 static void intel_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
-	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 
 	/* Make sure the panel is off before trying to change the mode. But also
@@ -1912,21 +2321,19 @@ static void intel_disable_dp(struct intel_encoder *encoder)
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF);
 	intel_edp_panel_off(intel_dp);
 
-	/* cpu edp my only be disable _after_ the cpu pipe/plane is disabled. */
-	if (!(port == PORT_A || IS_VALLEYVIEW(dev)))
+	/* disable the port before the pipe on g4x */
+	if (INTEL_INFO(dev)->gen < 5)
 		intel_dp_link_down(intel_dp);
 }
 
-static void g4x_post_disable_dp(struct intel_encoder *encoder)
+static void ilk_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (port != PORT_A)
-		return;
-
 	intel_dp_link_down(intel_dp);
-	ironlake_edp_pll_off(intel_dp);
+	if (port == PORT_A)
+		ironlake_edp_pll_off(intel_dp);
 }
 
 static void vlv_post_disable_dp(struct intel_encoder *encoder)
@@ -1972,6 +2379,104 @@ static void chv_post_disable_dp(struct intel_encoder *encoder)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static void
+_intel_dp_set_link_train(struct intel_dp *intel_dp,
+			 uint32_t *DP,
+			 uint8_t dp_train_pat)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = intel_dig_port->port;
+
+	if (HAS_DDI(dev)) {
+		uint32_t temp = I915_READ(DP_TP_CTL(port));
+
+		if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE)
+			temp |= DP_TP_CTL_SCRAMBLE_DISABLE;
+		else
+			temp &= ~DP_TP_CTL_SCRAMBLE_DISABLE;
+
+		temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
+
+			break;
+		case DP_TRAINING_PATTERN_1:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT2;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			temp |= DP_TP_CTL_LINK_TRAIN_PAT3;
+			break;
+		}
+		I915_WRITE(DP_TP_CTL(port), temp);
+
+	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
+		*DP &= ~DP_LINK_TRAIN_MASK_CPT;
+
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			*DP |= DP_LINK_TRAIN_OFF_CPT;
+			break;
+		case DP_TRAINING_PATTERN_1:
+			*DP |= DP_LINK_TRAIN_PAT_1_CPT;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			DRM_ERROR("DP training pattern 3 not supported\n");
+			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
+			break;
+		}
+
+	} else {
+		if (IS_CHERRYVIEW(dev))
+			*DP &= ~DP_LINK_TRAIN_MASK_CHV;
+		else
+			*DP &= ~DP_LINK_TRAIN_MASK;
+
+		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			*DP |= DP_LINK_TRAIN_OFF;
+			break;
+		case DP_TRAINING_PATTERN_1:
+			*DP |= DP_LINK_TRAIN_PAT_1;
+			break;
+		case DP_TRAINING_PATTERN_2:
+			*DP |= DP_LINK_TRAIN_PAT_2;
+			break;
+		case DP_TRAINING_PATTERN_3:
+			if (IS_CHERRYVIEW(dev)) {
+				*DP |= DP_LINK_TRAIN_PAT_3_CHV;
+			} else {
+				DRM_ERROR("DP training pattern 3 not supported\n");
+				*DP |= DP_LINK_TRAIN_PAT_2;
+			}
+			break;
+		}
+	}
+}
+
+static void intel_dp_enable_port(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	intel_dp->DP |= DP_PORT_EN;
+
+	/* enable with pattern 1 (as per spec) */
+	_intel_dp_set_link_train(intel_dp, &intel_dp->DP,
+				 DP_TRAINING_PATTERN_1);
+
+	I915_WRITE(intel_dp->output_reg, intel_dp->DP);
+	POSTING_READ(intel_dp->output_reg);
+}
+
 static void intel_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
@@ -1982,11 +2487,12 @@ static void intel_enable_dp(struct intel_encoder *encoder)
 	if (WARN_ON(dp_reg & DP_PORT_EN))
 		return;
 
+	intel_dp_enable_port(intel_dp);
 	intel_edp_panel_vdd_on(intel_dp);
+	intel_edp_panel_on(intel_dp);
+	intel_edp_panel_vdd_off(intel_dp, true);
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
 	intel_dp_start_link_train(intel_dp);
-	intel_edp_panel_on(intel_dp);
-	edp_panel_vdd_off(intel_dp, true);
 	intel_dp_complete_link_train(intel_dp);
 	intel_dp_stop_link_train(intel_dp);
 }
@@ -2020,6 +2526,78 @@ static void g4x_pre_enable_dp(struct intel_encoder *encoder)
 	}
 }
 
+static void vlv_steal_power_sequencer(struct drm_device *dev,
+				      enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *encoder;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+			    base.head) {
+		struct intel_dp *intel_dp;
+		enum port port;
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		intel_dp = enc_to_intel_dp(&encoder->base);
+		port = dp_to_dig_port(intel_dp)->port;
+
+		if (intel_dp->pps_pipe != pipe)
+			continue;
+
+		DRM_DEBUG_KMS("stealing pipe %c power sequencer from port %c\n",
+			      pipe_name(pipe), port_name(port));
+
+		/* make sure vdd is off before we steal it */
+		edp_panel_vdd_off_sync(intel_dp);
+
+		intel_dp->pps_pipe = INVALID_PIPE;
+	}
+}
+
+static void vlv_init_panel_power_sequencer(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+	struct edp_power_seq power_seq;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
+	if (intel_dp->pps_pipe == crtc->pipe)
+		return;
+
+	/*
+	 * If another power sequencer was being used on this
+	 * port previously make sure to turn off vdd there while
+	 * we still have control of it.
+	 */
+	if (intel_dp->pps_pipe != INVALID_PIPE)
+		edp_panel_vdd_off_sync(intel_dp);
+
+	/*
+	 * We may be stealing the power
+	 * sequencer from another port.
+	 */
+	vlv_steal_power_sequencer(dev, crtc->pipe);
+
+	/* now it's all ours */
+	intel_dp->pps_pipe = crtc->pipe;
+
+	DRM_DEBUG_KMS("initializing pipe %c power sequencer for port %c\n",
+		      pipe_name(intel_dp->pps_pipe), port_name(intel_dig_port->port));
+
+	/* init power sequencer on this pipe and port */
+	intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
+						      &power_seq);
+}
+
 static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
@@ -2029,7 +2607,6 @@ static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel port = vlv_dport_to_channel(dport);
 	int pipe = intel_crtc->pipe;
-	struct edp_power_seq power_seq;
 	u32 val;
 
 	mutex_lock(&dev_priv->dpio_lock);
@@ -2048,10 +2625,9 @@ static void vlv_pre_enable_dp(struct intel_encoder *encoder)
 	mutex_unlock(&dev_priv->dpio_lock);
 
 	if (is_edp(intel_dp)) {
-		/* init power sequencer on this pipe and port */
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
+		pps_lock(intel_dp);
+		vlv_init_panel_power_sequencer(intel_dp);
+		pps_unlock(intel_dp);
 	}
 
 	intel_enable_dp(encoder);
@@ -2095,7 +2671,6 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edp_power_seq power_seq;
 	struct intel_crtc *intel_crtc =
 		to_intel_crtc(encoder->base.crtc);
 	enum dpio_channel ch = vlv_dport_to_channel(dport);
@@ -2141,10 +2716,9 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 	mutex_unlock(&dev_priv->dpio_lock);
 
 	if (is_edp(intel_dp)) {
-		/* init power sequencer on this pipe and port */
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
+		pps_lock(intel_dp);
+		vlv_init_panel_power_sequencer(intel_dp);
+		pps_unlock(intel_dp);
 	}
 
 	intel_enable_dp(encoder);
@@ -2152,6 +2726,72 @@ static void chv_pre_enable_dp(struct intel_encoder *encoder)
 	vlv_wait_port_ready(dev_priv, dport);
 }
 
+static void chv_dp_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc =
+		to_intel_crtc(encoder->base.crtc);
+	enum dpio_channel ch = vlv_dport_to_channel(dport);
+	enum pipe pipe = intel_crtc->pipe;
+	u32 val;
+
+	intel_dp_prepare(encoder);
+
+	mutex_lock(&dev_priv->dpio_lock);
+
+	/* program left/right clock distribution */
+	if (pipe != PIPE_B) {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW5_CH0);
+		val &= ~(CHV_BUFLEFTENA1_MASK | CHV_BUFRIGHTENA1_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA1_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA1_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW5_CH0, val);
+	} else {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW1_CH1);
+		val &= ~(CHV_BUFLEFTENA2_MASK | CHV_BUFRIGHTENA2_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA2_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA2_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW1_CH1, val);
+	}
+
+	/* program clock channel usage */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW8(ch));
+	val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+	if (pipe != PIPE_B)
+		val &= ~CHV_PCS_USEDCLKCHANNEL;
+	else
+		val |= CHV_PCS_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW8(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW8(ch));
+	val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+	if (pipe != PIPE_B)
+		val &= ~CHV_PCS_USEDCLKCHANNEL;
+	else
+		val |= CHV_PCS_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW8(ch), val);
+
+	/*
+	 * This a a bit weird since generally CL
+	 * matches the pipe, but here we need to
+	 * pick the CL based on the port.
+	 */
+	val = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW19(ch));
+	if (pipe != PIPE_B)
+		val &= ~CHV_CMN_USEDCLKCHANNEL;
+	else
+		val |= CHV_CMN_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW19(ch), val);
+
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 /*
  * Native read with retry for link status and receiver capability reads for
  * cases where the sink may still be asleep.
@@ -2166,6 +2806,13 @@ intel_dp_dpcd_read_wake(struct drm_dp_aux *aux, unsigned int offset,
 	ssize_t ret;
 	int i;
 
+	/*
+	 * Sometime we just get the same incorrect byte repeated
+	 * over the entire buffer. Doing just one throw away read
+	 * initially seems to "solve" it.
+	 */
+	drm_dp_dpcd_read(aux, DP_DPCD_REV, buffer, 1);
+
 	for (i = 0; i < 3; i++) {
 		ret = drm_dp_dpcd_read(aux, offset, buffer, size);
 		if (ret == size)
@@ -2189,25 +2836,21 @@ intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_
 				       DP_LINK_STATUS_SIZE) == DP_LINK_STATUS_SIZE;
 }
 
-/*
- * These are source-specific values; current Intel hardware supports
- * a maximum voltage of 800mV and a maximum pre-emphasis of 6dB
- */
-
+/* These are source-specific values. */
 static uint8_t
 intel_dp_voltage_max(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_VALLEYVIEW(dev) || IS_BROADWELL(dev))
-		return DP_TRAIN_VOLTAGE_SWING_1200;
+	if (IS_VALLEYVIEW(dev))
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else if (IS_GEN7(dev) && port == PORT_A)
-		return DP_TRAIN_VOLTAGE_SWING_800;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 	else if (HAS_PCH_CPT(dev) && port != PORT_A)
-		return DP_TRAIN_VOLTAGE_SWING_1200;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else
-		return DP_TRAIN_VOLTAGE_SWING_800;
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
 }
 
 static uint8_t
@@ -2216,62 +2859,51 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_BROADWELL(dev)) {
-		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
-		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
-		}
-	} else if (IS_HASWELL(dev)) {
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_9_5;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_3;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else if (IS_VALLEYVIEW(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_9_5;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_3;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else if (IS_GEN7(dev) && port == PORT_A) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	} else {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			return DP_TRAIN_PRE_EMPH_LEVEL_2;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			return DP_TRAIN_PRE_EMPH_LEVEL_1;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
+			return DP_TRAIN_PRE_EMPH_LEVEL_0;
 		}
 	}
 }
@@ -2290,22 +2922,22 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 	int pipe = intel_crtc->pipe;
 
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 		preemph_reg_value = 0x0004000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B405555;
 			uniqtranscale_reg_value = 0x552AB83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x5548B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			demph_reg_value = 0x2B245555;
 			uniqtranscale_reg_value = 0x5560B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 			demph_reg_value = 0x2B405555;
 			uniqtranscale_reg_value = 0x5598DA3A;
 			break;
@@ -2313,18 +2945,18 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		preemph_reg_value = 0x0002000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x5552B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B404848;
 			uniqtranscale_reg_value = 0x5580B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			demph_reg_value = 0x2B404040;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2332,14 +2964,14 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		preemph_reg_value = 0x0000000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x2B305555;
 			uniqtranscale_reg_value = 0x5570B83A;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			demph_reg_value = 0x2B2B4040;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2347,10 +2979,10 @@ static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		preemph_reg_value = 0x0006000;
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			demph_reg_value = 0x1B405555;
 			uniqtranscale_reg_value = 0x55ADDA3A;
 			break;
@@ -2389,21 +3021,21 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	int i;
 
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 128;
 			margin_reg_value = 52;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 128;
 			margin_reg_value = 77;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			deemph_reg_value = 128;
 			margin_reg_value = 102;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 			deemph_reg_value = 128;
 			margin_reg_value = 154;
 			/* FIXME extra to set for 1200 */
@@ -2412,17 +3044,17 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 85;
 			margin_reg_value = 78;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 85;
 			margin_reg_value = 116;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_800:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 			deemph_reg_value = 85;
 			margin_reg_value = 154;
 			break;
@@ -2430,13 +3062,13 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 64;
 			margin_reg_value = 104;
 			break;
-		case DP_TRAIN_VOLTAGE_SWING_600:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 			deemph_reg_value = 64;
 			margin_reg_value = 154;
 			break;
@@ -2444,9 +3076,9 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 			return 0;
 		}
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 			deemph_reg_value = 43;
 			margin_reg_value = 154;
 			break;
@@ -2480,8 +3112,8 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	/* Program swing margin */
 	for (i = 0; i < 4; i++) {
 		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW2(ch, i));
-		val &= ~DPIO_SWING_MARGIN_MASK;
-		val |= margin_reg_value << DPIO_SWING_MARGIN_SHIFT;
+		val &= ~DPIO_SWING_MARGIN000_MASK;
+		val |= margin_reg_value << DPIO_SWING_MARGIN000_SHIFT;
 		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW2(ch, i), val);
 	}
 
@@ -2493,9 +3125,9 @@ static uint32_t intel_chv_signal_levels(struct intel_dp *intel_dp)
 	}
 
 	if (((train_set & DP_TRAIN_PRE_EMPHASIS_MASK)
-			== DP_TRAIN_PRE_EMPHASIS_0) &&
+			== DP_TRAIN_PRE_EMPH_LEVEL_0) &&
 		((train_set & DP_TRAIN_VOLTAGE_SWING_MASK)
-			== DP_TRAIN_VOLTAGE_SWING_1200)) {
+			== DP_TRAIN_VOLTAGE_SWING_LEVEL_3)) {
 
 		/*
 		 * The document said it needs to set bit 27 for ch0 and bit 26
@@ -2574,32 +3206,32 @@ intel_gen4_signal_levels(uint8_t train_set)
 	uint32_t	signal_levels = 0;
 
 	switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
-	case DP_TRAIN_VOLTAGE_SWING_400:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
 	default:
 		signal_levels |= DP_VOLTAGE_0_4;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_600:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
 		signal_levels |= DP_VOLTAGE_0_6;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_800:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
 		signal_levels |= DP_VOLTAGE_0_8;
 		break;
-	case DP_TRAIN_VOLTAGE_SWING_1200:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
 		signal_levels |= DP_VOLTAGE_1_2;
 		break;
 	}
 	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
-	case DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
 	default:
 		signal_levels |= DP_PRE_EMPHASIS_0;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
 		signal_levels |= DP_PRE_EMPHASIS_3_5;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
 		signal_levels |= DP_PRE_EMPHASIS_6;
 		break;
-	case DP_TRAIN_PRE_EMPHASIS_9_5:
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
 		signal_levels |= DP_PRE_EMPHASIS_9_5;
 		break;
 	}
@@ -2613,19 +3245,19 @@ intel_gen6_edp_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_400MV_3_5DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
 		return EDP_LINK_TRAIN_400_600MV_6DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-	case DP_TRAIN_VOLTAGE_SWING_1200 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B;
 	default:
 		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
@@ -2641,21 +3273,21 @@ intel_gen7_edp_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_400MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_400MV_3_5DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
 		return EDP_LINK_TRAIN_400MV_6DB_IVB;
 
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_600MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_600MV_3_5DB_IVB;
 
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
 		return EDP_LINK_TRAIN_800MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
 		return EDP_LINK_TRAIN_800MV_3_5DB_IVB;
 
 	default:
@@ -2672,65 +3304,30 @@ intel_hsw_signal_levels(uint8_t train_set)
 	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
 					 DP_TRAIN_PRE_EMPHASIS_MASK);
 	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_400MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_400MV_3_5DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_400MV_6DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_9_5:
-		return DDI_BUF_EMP_400MV_9_5DB_HSW;
-
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_600MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_600MV_3_5DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_600MV_6DB_HSW;
-
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_800MV_0DB_HSW;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_800MV_3_5DB_HSW;
-	default:
-		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
-			      "0x%x\n", signal_levels);
-		return DDI_BUF_EMP_400MV_0DB_HSW;
-	}
-}
-
-static uint32_t
-intel_bdw_signal_levels(uint8_t train_set)
-{
-	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
-					 DP_TRAIN_PRE_EMPHASIS_MASK);
-	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_400MV_0DB_BDW;	/* Sel0 */
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_400MV_3_5DB_BDW;	/* Sel1 */
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_400MV_6DB_BDW;	/* Sel2 */
-
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_600MV_0DB_BDW;	/* Sel3 */
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_600MV_3_5DB_BDW;	/* Sel4 */
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
-		return DDI_BUF_EMP_600MV_6DB_BDW;	/* Sel5 */
-
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_800MV_0DB_BDW;	/* Sel6 */
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return DDI_BUF_EMP_800MV_3_5DB_BDW;	/* Sel7 */
-
-	case DP_TRAIN_VOLTAGE_SWING_1200 | DP_TRAIN_PRE_EMPHASIS_0:
-		return DDI_BUF_EMP_1200MV_0DB_BDW;	/* Sel8 */
-
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(0);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(1);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return DDI_BUF_TRANS_SELECT(2);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_3:
+		return DDI_BUF_TRANS_SELECT(3);
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(4);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(5);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return DDI_BUF_TRANS_SELECT(6);
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DDI_BUF_TRANS_SELECT(7);
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DDI_BUF_TRANS_SELECT(8);
 	default:
 		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
 			      "0x%x\n", signal_levels);
-		return DDI_BUF_EMP_400MV_0DB_BDW;	/* Sel0 */
+		return DDI_BUF_TRANS_SELECT(0);
 	}
 }
 
@@ -2744,10 +3341,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 	uint32_t signal_levels, mask;
 	uint8_t train_set = intel_dp->train_set[0];
 
-	if (IS_BROADWELL(dev)) {
-		signal_levels = intel_bdw_signal_levels(train_set);
-		mask = DDI_BUF_EMP_MASK;
-	} else if (IS_HASWELL(dev)) {
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		signal_levels = intel_hsw_signal_levels(train_set);
 		mask = DDI_BUF_EMP_MASK;
 	} else if (IS_CHERRYVIEW(dev)) {
@@ -2780,74 +3374,10 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum port port = intel_dig_port->port;
 	uint8_t buf[sizeof(intel_dp->train_set) + 1];
 	int ret, len;
 
-	if (HAS_DDI(dev)) {
-		uint32_t temp = I915_READ(DP_TP_CTL(port));
-
-		if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE)
-			temp |= DP_TP_CTL_SCRAMBLE_DISABLE;
-		else
-			temp &= ~DP_TP_CTL_SCRAMBLE_DISABLE;
-
-		temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
-
-			break;
-		case DP_TRAINING_PATTERN_1:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT2;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			temp |= DP_TP_CTL_LINK_TRAIN_PAT3;
-			break;
-		}
-		I915_WRITE(DP_TP_CTL(port), temp);
-
-	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
-		*DP &= ~DP_LINK_TRAIN_MASK_CPT;
-
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			*DP |= DP_LINK_TRAIN_OFF_CPT;
-			break;
-		case DP_TRAINING_PATTERN_1:
-			*DP |= DP_LINK_TRAIN_PAT_1_CPT;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			DRM_ERROR("DP training pattern 3 not supported\n");
-			*DP |= DP_LINK_TRAIN_PAT_2_CPT;
-			break;
-		}
-
-	} else {
-		*DP &= ~DP_LINK_TRAIN_MASK;
-
-		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
-		case DP_TRAINING_PATTERN_DISABLE:
-			*DP |= DP_LINK_TRAIN_OFF;
-			break;
-		case DP_TRAINING_PATTERN_1:
-			*DP |= DP_LINK_TRAIN_PAT_1;
-			break;
-		case DP_TRAINING_PATTERN_2:
-			*DP |= DP_LINK_TRAIN_PAT_2;
-			break;
-		case DP_TRAINING_PATTERN_3:
-			DRM_ERROR("DP training pattern 3 not supported\n");
-			*DP |= DP_LINK_TRAIN_PAT_2;
-			break;
-		}
-	}
+	_intel_dp_set_link_train(intel_dp, DP, dp_train_pat);
 
 	I915_WRITE(intel_dp->output_reg, *DP);
 	POSTING_READ(intel_dp->output_reg);
@@ -3131,7 +3661,10 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE_CPT);
 	} else {
-		DP &= ~DP_LINK_TRAIN_MASK;
+		if (IS_CHERRYVIEW(dev))
+			DP &= ~DP_LINK_TRAIN_MASK_CHV;
+		else
+			DP &= ~DP_LINK_TRAIN_MASK;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE);
 	}
 	POSTING_READ(intel_dp->output_reg);
@@ -3177,15 +3710,11 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp)
 	struct drm_device *dev = dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	char dpcd_hex_dump[sizeof(intel_dp->dpcd) * 3];
-
 	if (intel_dp_dpcd_read_wake(&intel_dp->aux, 0x000, intel_dp->dpcd,
 				    sizeof(intel_dp->dpcd)) < 0)
 		return false; /* aux transfer failed */
 
-	hex_dump_to_buffer(intel_dp->dpcd, sizeof(intel_dp->dpcd),
-			   32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false);
-	DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump);
+	DRM_DEBUG_KMS("DPCD: %*ph\n", (int) sizeof(intel_dp->dpcd), intel_dp->dpcd);
 
 	if (intel_dp->dpcd[DP_DPCD_REV] == 0)
 		return false; /* DPCD not present */
@@ -3202,11 +3731,12 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp)
 		}
 	}
 
-	/* Training Pattern 3 support */
+	/* Training Pattern 3 support, both source and sink */
 	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x12 &&
-	    intel_dp->dpcd[DP_MAX_LANE_COUNT] & DP_TPS3_SUPPORTED) {
+	    intel_dp->dpcd[DP_MAX_LANE_COUNT] & DP_TPS3_SUPPORTED &&
+	    (IS_HASWELL(dev_priv) || INTEL_INFO(dev_priv)->gen >= 8)) {
 		intel_dp->use_tps3 = true;
-		DRM_DEBUG_KMS("Displayport TPS3 supported");
+		DRM_DEBUG_KMS("Displayport TPS3 supported\n");
 	} else
 		intel_dp->use_tps3 = false;
 
@@ -3243,7 +3773,34 @@ intel_dp_probe_oui(struct intel_dp *intel_dp)
 		DRM_DEBUG_KMS("Branch OUI: %02hx%02hx%02hx\n",
 			      buf[0], buf[1], buf[2]);
 
-	edp_panel_vdd_off(intel_dp, false);
+	intel_edp_panel_vdd_off(intel_dp, false);
+}
+
+static bool
+intel_dp_probe_mst(struct intel_dp *intel_dp)
+{
+	u8 buf[1];
+
+	if (!intel_dp->can_mst)
+		return false;
+
+	if (intel_dp->dpcd[DP_DPCD_REV] < 0x12)
+		return false;
+
+	intel_edp_panel_vdd_on(intel_dp);
+	if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_MSTM_CAP, buf, 1)) {
+		if (buf[0] & DP_MST_CAP) {
+			DRM_DEBUG_KMS("Sink is MST capable\n");
+			intel_dp->is_mst = true;
+		} else {
+			DRM_DEBUG_KMS("Sink is not MST capable\n");
+			intel_dp->is_mst = false;
+		}
+	}
+	intel_edp_panel_vdd_off(intel_dp, false);
+
+	drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, intel_dp->is_mst);
+	return intel_dp->is_mst;
 }
 
 int intel_dp_sink_crc(struct intel_dp *intel_dp, u8 *crc)
@@ -3255,21 +3812,21 @@ int intel_dp_sink_crc(struct intel_dp *intel_dp, u8 *crc)
 	u8 buf[1];
 
 	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK_MISC, buf) < 0)
-		return -EAGAIN;
+		return -EIO;
 
 	if (!(buf[0] & DP_TEST_CRC_SUPPORTED))
 		return -ENOTTY;
 
 	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK,
 			       DP_TEST_SINK_START) < 0)
-		return -EAGAIN;
+		return -EIO;
 
 	/* Wait 2 vblanks to be sure we will have the correct CRC value */
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 
 	if (drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_CRC_R_CR, crc, 6) < 0)
-		return -EAGAIN;
+		return -EIO;
 
 	drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK, 0);
 	return 0;
@@ -3283,6 +3840,20 @@ intel_dp_get_sink_irq(struct intel_dp *intel_dp, u8 *sink_irq_vector)
 				       sink_irq_vector, 1) == 1;
 }
 
+static bool
+intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *sink_irq_vector)
+{
+	int ret;
+
+	ret = intel_dp_dpcd_read_wake(&intel_dp->aux,
+					     DP_SINK_COUNT_ESI,
+					     sink_irq_vector, 14);
+	if (ret != 14)
+		return false;
+
+	return true;
+}
+
 static void
 intel_dp_handle_test_request(struct intel_dp *intel_dp)
 {
@@ -3290,6 +3861,63 @@ intel_dp_handle_test_request(struct intel_dp *intel_dp)
 	drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_RESPONSE, DP_TEST_NAK);
 }
 
+static int
+intel_dp_check_mst_status(struct intel_dp *intel_dp)
+{
+	bool bret;
+
+	if (intel_dp->is_mst) {
+		u8 esi[16] = { 0 };
+		int ret = 0;
+		int retry;
+		bool handled;
+		bret = intel_dp_get_sink_irq_esi(intel_dp, esi);
+go_again:
+		if (bret == true) {
+
+			/* check link status - esi[10] = 0x200c */
+			if (intel_dp->active_mst_links && !drm_dp_channel_eq_ok(&esi[10], intel_dp->lane_count)) {
+				DRM_DEBUG_KMS("channel EQ not ok, retraining\n");
+				intel_dp_start_link_train(intel_dp);
+				intel_dp_complete_link_train(intel_dp);
+				intel_dp_stop_link_train(intel_dp);
+			}
+
+			DRM_DEBUG_KMS("got esi %02x %02x %02x\n", esi[0], esi[1], esi[2]);
+			ret = drm_dp_mst_hpd_irq(&intel_dp->mst_mgr, esi, &handled);
+
+			if (handled) {
+				for (retry = 0; retry < 3; retry++) {
+					int wret;
+					wret = drm_dp_dpcd_write(&intel_dp->aux,
+								 DP_SINK_COUNT_ESI+1,
+								 &esi[1], 3);
+					if (wret == 3) {
+						break;
+					}
+				}
+
+				bret = intel_dp_get_sink_irq_esi(intel_dp, esi);
+				if (bret == true) {
+					DRM_DEBUG_KMS("got esi2 %02x %02x %02x\n", esi[0], esi[1], esi[2]);
+					goto go_again;
+				}
+			} else
+				ret = 0;
+
+			return ret;
+		} else {
+			struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+			DRM_DEBUG_KMS("failed to get ESI - device may have failed\n");
+			intel_dp->is_mst = false;
+			drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, intel_dp->is_mst);
+			/* send a hotplug event */
+			drm_kms_helper_hotplug_event(intel_dig_port->base.base.dev);
+		}
+	}
+	return -EINVAL;
+}
+
 /*
  * According to DP spec
  * 5.1.2:
@@ -3298,21 +3926,25 @@ intel_dp_handle_test_request(struct intel_dp *intel_dp)
  *  3. Use Link Training from 2.5.3.3 and 3.5.1.3
  *  4. Check link status on receipt of hot-plug interrupt
  */
-
 void
 intel_dp_check_link_status(struct intel_dp *intel_dp)
 {
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 	u8 sink_irq_vector;
 	u8 link_status[DP_LINK_STATUS_SIZE];
 
-	/* FIXME: This access isn't protected by any locks. */
+	WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+
 	if (!intel_encoder->connectors_active)
 		return;
 
 	if (WARN_ON(!intel_encoder->base.crtc))
 		return;
 
+	if (!to_intel_crtc(intel_encoder->base.crtc)->active)
+		return;
+
 	/* Try to read receiver status if the link appears to be up */
 	if (!intel_dp_get_link_status(intel_dp, link_status)) {
 		return;
@@ -3397,20 +4029,24 @@ intel_dp_detect_dpcd(struct intel_dp *intel_dp)
 }
 
 static enum drm_connector_status
+edp_detect(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum drm_connector_status status;
+
+	status = intel_panel_detect(dev);
+	if (status == connector_status_unknown)
+		status = connector_status_connected;
+
+	return status;
+}
+
+static enum drm_connector_status
 ironlake_dp_detect(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	enum drm_connector_status status;
-
-	/* Can't disconnect eDP, but you can close the lid... */
-	if (is_edp(intel_dp)) {
-		status = intel_panel_detect(dev);
-		if (status == connector_status_unknown)
-			status = connector_status_connected;
-		return status;
-	}
 
 	if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
 		return connector_status_disconnected;
@@ -3418,24 +4054,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
 	return intel_dp_detect_dpcd(intel_dp);
 }
 
-static enum drm_connector_status
-g4x_dp_detect(struct intel_dp *intel_dp)
+static int g4x_digital_port_connected(struct drm_device *dev,
+				       struct intel_digital_port *intel_dig_port)
 {
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	uint32_t bit;
 
-	/* Can't disconnect eDP, but you can close the lid... */
-	if (is_edp(intel_dp)) {
-		enum drm_connector_status status;
-
-		status = intel_panel_detect(dev);
-		if (status == connector_status_unknown)
-			status = connector_status_connected;
-		return status;
-	}
-
 	if (IS_VALLEYVIEW(dev)) {
 		switch (intel_dig_port->port) {
 		case PORT_B:
@@ -3448,7 +4072,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 			bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
 			break;
 		default:
-			return connector_status_unknown;
+			return -EINVAL;
 		}
 	} else {
 		switch (intel_dig_port->port) {
@@ -3462,20 +4086,45 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 			bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
 			break;
 		default:
-			return connector_status_unknown;
+			return -EINVAL;
 		}
 	}
 
 	if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
+		return 0;
+	return 1;
+}
+
+static enum drm_connector_status
+g4x_dp_detect(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	int ret;
+
+	/* Can't disconnect eDP, but you can close the lid... */
+	if (is_edp(intel_dp)) {
+		enum drm_connector_status status;
+
+		status = intel_panel_detect(dev);
+		if (status == connector_status_unknown)
+			status = connector_status_connected;
+		return status;
+	}
+
+	ret = g4x_digital_port_connected(dev, intel_dig_port);
+	if (ret == -EINVAL)
+		return connector_status_unknown;
+	else if (ret == 0)
 		return connector_status_disconnected;
 
 	return intel_dp_detect_dpcd(intel_dp);
 }
 
 static struct edid *
-intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
+intel_dp_get_edid(struct intel_dp *intel_dp)
 {
-	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
 
 	/* use cached edid if we have one */
 	if (intel_connector->edid) {
@@ -3484,27 +4133,55 @@ intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 			return NULL;
 
 		return drm_edid_duplicate(intel_connector->edid);
-	}
+	} else
+		return drm_get_edid(&intel_connector->base,
+				    &intel_dp->aux.ddc);
+}
 
-	return drm_get_edid(connector, adapter);
+static void
+intel_dp_set_edid(struct intel_dp *intel_dp)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct edid *edid;
+
+	edid = intel_dp_get_edid(intel_dp);
+	intel_connector->detect_edid = edid;
+
+	if (intel_dp->force_audio != HDMI_AUDIO_AUTO)
+		intel_dp->has_audio = intel_dp->force_audio == HDMI_AUDIO_ON;
+	else
+		intel_dp->has_audio = drm_detect_monitor_audio(edid);
 }
 
-static int
-intel_dp_get_edid_modes(struct drm_connector *connector, struct i2c_adapter *adapter)
+static void
+intel_dp_unset_edid(struct intel_dp *intel_dp)
 {
-	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
 
-	/* use cached edid if we have one */
-	if (intel_connector->edid) {
-		/* invalid edid */
-		if (IS_ERR(intel_connector->edid))
-			return 0;
+	kfree(intel_connector->detect_edid);
+	intel_connector->detect_edid = NULL;
 
-		return intel_connector_update_modes(connector,
-						    intel_connector->edid);
-	}
+	intel_dp->has_audio = false;
+}
+
+static enum intel_display_power_domain
+intel_dp_power_get(struct intel_dp *dp)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(dp)->base;
+	enum intel_display_power_domain power_domain;
+
+	power_domain = intel_display_port_power_domain(encoder);
+	intel_display_power_get(to_i915(encoder->base.dev), power_domain);
 
-	return intel_ddc_get_modes(connector, adapter);
+	return power_domain;
+}
+
+static void
+intel_dp_power_put(struct intel_dp *dp,
+		   enum intel_display_power_domain power_domain)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(dp)->base;
+	intel_display_power_put(to_i915(encoder->base.dev), power_domain);
 }
 
 static enum drm_connector_status
@@ -3514,110 +4191,117 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum drm_connector_status status;
 	enum intel_display_power_domain power_domain;
-	struct edid *edid = NULL;
-
-	intel_runtime_pm_get(dev_priv);
-
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	bool ret;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
+	intel_dp_unset_edid(intel_dp);
 
-	intel_dp->has_audio = false;
+	if (intel_dp->is_mst) {
+		/* MST devices are disconnected from a monitor POV */
+		if (intel_encoder->type != INTEL_OUTPUT_EDP)
+			intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+		return connector_status_disconnected;
+	}
 
-	if (HAS_PCH_SPLIT(dev))
+	power_domain = intel_dp_power_get(intel_dp);
+
+	/* Can't disconnect eDP, but you can close the lid... */
+	if (is_edp(intel_dp))
+		status = edp_detect(intel_dp);
+	else if (HAS_PCH_SPLIT(dev))
 		status = ironlake_dp_detect(intel_dp);
 	else
 		status = g4x_dp_detect(intel_dp);
-
 	if (status != connector_status_connected)
 		goto out;
 
 	intel_dp_probe_oui(intel_dp);
 
-	if (intel_dp->force_audio != HDMI_AUDIO_AUTO) {
-		intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON);
-	} else {
-		edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc);
-		if (edid) {
-			intel_dp->has_audio = drm_detect_monitor_audio(edid);
-			kfree(edid);
-		}
+	ret = intel_dp_probe_mst(intel_dp);
+	if (ret) {
+		/* if we are in MST mode then this connector
+		   won't appear connected or have anything with EDID on it */
+		if (intel_encoder->type != INTEL_OUTPUT_EDP)
+			intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+		status = connector_status_disconnected;
+		goto out;
 	}
 
+	intel_dp_set_edid(intel_dp);
+
 	if (intel_encoder->type != INTEL_OUTPUT_EDP)
 		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
 	status = connector_status_connected;
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
-
-	intel_runtime_pm_put(dev_priv);
-
+	intel_dp_power_put(intel_dp, power_domain);
 	return status;
 }
 
-static int intel_dp_get_modes(struct drm_connector *connector)
+static void
+intel_dp_force(struct drm_connector *connector)
 {
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct intel_connector *intel_connector = to_intel_connector(connector);
-	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 	enum intel_display_power_domain power_domain;
-	int ret;
 
-	/* We should parse the EDID data and find out if it has an audio sink
-	 */
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+	intel_dp_unset_edid(intel_dp);
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	if (connector->status != connector_status_connected)
+		return;
 
-	ret = intel_dp_get_edid_modes(connector, &intel_dp->aux.ddc);
-	intel_display_power_put(dev_priv, power_domain);
-	if (ret)
-		return ret;
+	power_domain = intel_dp_power_get(intel_dp);
+
+	intel_dp_set_edid(intel_dp);
+
+	intel_dp_power_put(intel_dp, power_domain);
+
+	if (intel_encoder->type != INTEL_OUTPUT_EDP)
+		intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT;
+}
+
+static int intel_dp_get_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct edid *edid;
+
+	edid = intel_connector->detect_edid;
+	if (edid) {
+		int ret = intel_connector_update_modes(connector, edid);
+		if (ret)
+			return ret;
+	}
 
 	/* if eDP has no EDID, fall back to fixed mode */
-	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
+	if (is_edp(intel_attached_dp(connector)) &&
+	    intel_connector->panel.fixed_mode) {
 		struct drm_display_mode *mode;
-		mode = drm_mode_duplicate(dev,
+
+		mode = drm_mode_duplicate(connector->dev,
 					  intel_connector->panel.fixed_mode);
 		if (mode) {
 			drm_mode_probed_add(connector, mode);
 			return 1;
 		}
 	}
+
 	return 0;
 }
 
 static bool
 intel_dp_detect_audio(struct drm_connector *connector)
 {
-	struct intel_dp *intel_dp = intel_attached_dp(connector);
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_device *dev = connector->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	struct edid *edid;
 	bool has_audio = false;
+	struct edid *edid;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
-	edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc);
-	if (edid) {
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid)
 		has_audio = drm_detect_monitor_audio(edid);
-		kfree(edid);
-	}
-
-	intel_display_power_put(dev_priv, power_domain);
 
 	return has_audio;
 }
@@ -3715,6 +4399,8 @@ intel_dp_connector_destroy(struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 
+	kfree(intel_connector->detect_edid);
+
 	if (!IS_ERR_OR_NULL(intel_connector->edid))
 		kfree(intel_connector->edid);
 
@@ -3731,15 +4417,20 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 {
 	struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder);
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
-	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 
 	drm_dp_aux_unregister(&intel_dp->aux);
+	intel_dp_mst_encoder_cleanup(intel_dig_port);
 	drm_encoder_cleanup(encoder);
 	if (is_edp(intel_dp)) {
 		cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-		drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+		/*
+		 * vdd might still be enabled do to the delayed vdd off.
+		 * Make sure vdd is actually turned off here.
+		 */
+		pps_lock(intel_dp);
 		edp_panel_vdd_off_sync(intel_dp);
-		drm_modeset_unlock(&dev->mode_config.connection_mutex);
+		pps_unlock(intel_dp);
+
 		if (intel_dp->edp_notifier.notifier_call) {
 			unregister_reboot_notifier(&intel_dp->edp_notifier);
 			intel_dp->edp_notifier.notifier_call = NULL;
@@ -3748,9 +4439,31 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 	kfree(intel_dig_port);
 }
 
+static void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
+
+	if (!is_edp(intel_dp))
+		return;
+
+	/*
+	 * vdd might still be enabled do to the delayed vdd off.
+	 * Make sure vdd is actually turned off here.
+	 */
+	pps_lock(intel_dp);
+	edp_panel_vdd_off_sync(intel_dp);
+	pps_unlock(intel_dp);
+}
+
+static void intel_dp_encoder_reset(struct drm_encoder *encoder)
+{
+	intel_edp_panel_vdd_sanitize(to_intel_encoder(encoder));
+}
+
 static const struct drm_connector_funcs intel_dp_connector_funcs = {
 	.dpms = intel_connector_dpms,
 	.detect = intel_dp_detect,
+	.force = intel_dp_force,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_set_property,
 	.destroy = intel_dp_connector_destroy,
@@ -3763,15 +4476,96 @@ static const struct drm_connector_helper_funcs intel_dp_connector_helper_funcs =
 };
 
 static const struct drm_encoder_funcs intel_dp_enc_funcs = {
+	.reset = intel_dp_encoder_reset,
 	.destroy = intel_dp_encoder_destroy,
 };
 
-static void
+void
 intel_dp_hot_plug(struct intel_encoder *intel_encoder)
 {
-	struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base);
+	return;
+}
+
+bool
+intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
+{
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	struct intel_encoder *intel_encoder = &intel_dig_port->base;
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain power_domain;
+	bool ret = true;
+
+	if (intel_dig_port->base.type != INTEL_OUTPUT_EDP)
+		intel_dig_port->base.type = INTEL_OUTPUT_DISPLAYPORT;
+
+	if (long_hpd && intel_dig_port->base.type == INTEL_OUTPUT_EDP) {
+		/*
+		 * vdd off can generate a long pulse on eDP which
+		 * would require vdd on to handle it, and thus we
+		 * would end up in an endless cycle of
+		 * "vdd off -> long hpd -> vdd on -> detect -> vdd off -> ..."
+		 */
+		DRM_DEBUG_KMS("ignoring long hpd on eDP port %c\n",
+			      port_name(intel_dig_port->port));
+		return false;
+	}
+
+	DRM_DEBUG_KMS("got hpd irq on port %c - %s\n",
+		      port_name(intel_dig_port->port),
+		      long_hpd ? "long" : "short");
+
+	power_domain = intel_display_port_power_domain(intel_encoder);
+	intel_display_power_get(dev_priv, power_domain);
+
+	if (long_hpd) {
+
+		if (HAS_PCH_SPLIT(dev)) {
+			if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
+				goto mst_fail;
+		} else {
+			if (g4x_digital_port_connected(dev, intel_dig_port) != 1)
+				goto mst_fail;
+		}
+
+		if (!intel_dp_get_dpcd(intel_dp)) {
+			goto mst_fail;
+		}
 
-	intel_dp_check_link_status(intel_dp);
+		intel_dp_probe_oui(intel_dp);
+
+		if (!intel_dp_probe_mst(intel_dp))
+			goto mst_fail;
+
+	} else {
+		if (intel_dp->is_mst) {
+			if (intel_dp_check_mst_status(intel_dp) == -EINVAL)
+				goto mst_fail;
+		}
+
+		if (!intel_dp->is_mst) {
+			/*
+			 * we'll check the link status via the normal hot plug path later -
+			 * but for short hpds we should check it now
+			 */
+			drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+			intel_dp_check_link_status(intel_dp);
+			drm_modeset_unlock(&dev->mode_config.connection_mutex);
+		}
+	}
+	ret = false;
+	goto put_power;
+mst_fail:
+	/* if we were in MST mode, and device is not there get out of MST mode */
+	if (intel_dp->is_mst) {
+		DRM_DEBUG_KMS("MST device may have disappeared %d vs %d\n", intel_dp->is_mst, intel_dp->mst_mgr.mst_state);
+		intel_dp->is_mst = false;
+		drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, intel_dp->is_mst);
+	}
+put_power:
+	intel_display_power_put(dev_priv, power_domain);
+
+	return ret;
 }
 
 /* Return which DP Port should be selected for Transcoder DP control */
@@ -3822,7 +4616,7 @@ bool intel_dp_is_edp(struct drm_device *dev, enum port port)
 	return false;
 }
 
-static void
+void
 intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
@@ -3858,6 +4652,8 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	u32 pp_on, pp_off, pp_div, pp;
 	int pp_ctrl_reg, pp_on_reg, pp_off_reg, pp_div_reg;
 
+	lockdep_assert_held(&dev_priv->pps_mutex);
+
 	if (HAS_PCH_SPLIT(dev)) {
 		pp_ctrl_reg = PCH_PP_CONTROL;
 		pp_on_reg = PCH_PP_ON_DELAYS;
@@ -3957,6 +4753,9 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 	u32 pp_on, pp_off, pp_div, port_sel = 0;
 	int div = HAS_PCH_SPLIT(dev) ? intel_pch_rawclk(dev) : intel_hrawclk(dev);
 	int pp_on_reg, pp_off_reg, pp_div_reg;
+	enum port port = dp_to_dig_port(intel_dp)->port;
+
+	lockdep_assert_held(&dev_priv->pps_mutex);
 
 	if (HAS_PCH_SPLIT(dev)) {
 		pp_on_reg = PCH_PP_ON_DELAYS;
@@ -3991,12 +4790,9 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 	/* Haswell doesn't have any port selection bits for the panel
 	 * power sequencer any more. */
 	if (IS_VALLEYVIEW(dev)) {
-		if (dp_to_dig_port(intel_dp)->port == PORT_B)
-			port_sel = PANEL_PORT_SELECT_DPB_VLV;
-		else
-			port_sel = PANEL_PORT_SELECT_DPC_VLV;
+		port_sel = PANEL_PORT_SELECT_VLV(port);
 	} else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		if (dp_to_dig_port(intel_dp)->port == PORT_A)
+		if (port == PORT_A)
 			port_sel = PANEL_PORT_SELECT_DPA;
 		else
 			port_sel = PANEL_PORT_SELECT_DPD;
@@ -4035,6 +4831,11 @@ void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 		return;
 	}
 
+	/*
+	 * FIXME: This needs proper synchronization with psr state. But really
+	 * hard to tell without seeing the user of this function of this code.
+	 * Check locking and ordering once that lands.
+	 */
 	if (INTEL_INFO(dev)->gen < 8 && intel_edp_is_psr_enabled(dev)) {
 		DRM_DEBUG_KMS("DRRS is disabled as PSR is enabled\n");
 		return;
@@ -4075,7 +4876,7 @@ void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate)
 		val = I915_READ(reg);
 		if (index > DRRS_HIGH_RR) {
 			val |= PIPECONF_EDP_RR_MODE_SWITCH;
-			intel_dp_set_m2_n2(intel_crtc, &config->dp_m2_n2);
+			intel_dp_set_m_n(intel_crtc);
 		} else {
 			val &= ~PIPECONF_EDP_RR_MODE_SWITCH;
 		}
@@ -4115,7 +4916,7 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 	}
 
 	if (dev_priv->vbt.drrs_type != SEAMLESS_DRRS_SUPPORT) {
-		DRM_INFO("VBT doesn't support DRRS\n");
+		DRM_DEBUG_KMS("VBT doesn't support DRRS\n");
 		return NULL;
 	}
 
@@ -4123,7 +4924,7 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 					(dev, fixed_mode, connector);
 
 	if (!downclock_mode) {
-		DRM_INFO("DRRS not supported\n");
+		DRM_DEBUG_KMS("DRRS not supported\n");
 		return NULL;
 	}
 
@@ -4134,10 +4935,41 @@ intel_dp_drrs_init(struct intel_digital_port *intel_dig_port,
 	intel_dp->drrs_state.type = dev_priv->vbt.drrs_type;
 
 	intel_dp->drrs_state.refresh_rate_type = DRRS_HIGH_RR;
-	DRM_INFO("seamless DRRS supported for eDP panel.\n");
+	DRM_DEBUG_KMS("seamless DRRS supported for eDP panel.\n");
 	return downclock_mode;
 }
 
+void intel_edp_panel_vdd_sanitize(struct intel_encoder *intel_encoder)
+{
+	struct drm_device *dev = intel_encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_dp *intel_dp;
+	enum intel_display_power_domain power_domain;
+
+	if (intel_encoder->type != INTEL_OUTPUT_EDP)
+		return;
+
+	intel_dp = enc_to_intel_dp(&intel_encoder->base);
+
+	pps_lock(intel_dp);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		goto out;
+	/*
+	 * The VDD bit needs a power domain reference, so if the bit is
+	 * already enabled when we boot or resume, grab this reference and
+	 * schedule a vdd off, so we don't hold on to the reference
+	 * indefinitely.
+	 */
+	DRM_DEBUG_KMS("VDD left on by BIOS, adjusting state tracking\n");
+	power_domain = intel_display_port_power_domain(intel_encoder);
+	intel_display_power_get(dev_priv, power_domain);
+
+	edp_panel_vdd_schedule_off(intel_dp);
+ out:
+	pps_unlock(intel_dp);
+}
+
 static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 				     struct intel_connector *intel_connector,
 				     struct edp_power_seq *power_seq)
@@ -4158,18 +4990,12 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	if (!is_edp(intel_dp))
 		return true;
 
-	/* The VDD bit needs a power domain reference, so if the bit is already
-	 * enabled when we boot, grab this reference. */
-	if (edp_have_panel_vdd(intel_dp)) {
-		enum intel_display_power_domain power_domain;
-		power_domain = intel_display_port_power_domain(intel_encoder);
-		intel_display_power_get(dev_priv, power_domain);
-	}
+	intel_edp_panel_vdd_sanitize(intel_encoder);
 
 	/* Cache DPCD and EDID for edp. */
 	intel_edp_panel_vdd_on(intel_dp);
 	has_dpcd = intel_dp_get_dpcd(intel_dp);
-	edp_panel_vdd_off(intel_dp, false);
+	intel_edp_panel_vdd_off(intel_dp, false);
 
 	if (has_dpcd) {
 		if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
@@ -4183,7 +5009,9 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	}
 
 	/* We now know it's not a ghost, init power sequence regs. */
+	pps_lock(intel_dp);
 	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp, power_seq);
+	pps_unlock(intel_dp);
 
 	mutex_lock(&dev->mode_config.mutex);
 	edid = drm_get_edid(connector, &intel_dp->aux.ddc);
@@ -4227,6 +5055,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp,
 	}
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
+	intel_connector->panel.backlight_power = intel_edp_backlight_power;
 	intel_panel_setup_backlight(connector);
 
 	return true;
@@ -4245,6 +5074,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	struct edp_power_seq power_seq = { 0 };
 	int type;
 
+	intel_dp->pps_pipe = INVALID_PIPE;
+
 	/* intel_dp vfuncs */
 	if (IS_VALLEYVIEW(dev))
 		intel_dp->get_aux_clock_divider = vlv_get_aux_clock_divider;
@@ -4288,7 +5119,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 			  edp_panel_vdd_work);
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	if (HAS_DDI(dev))
 		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
@@ -4315,23 +5146,40 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	}
 
 	if (is_edp(intel_dp)) {
-		intel_dp_init_panel_power_timestamps(intel_dp);
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+		pps_lock(intel_dp);
+		if (IS_VALLEYVIEW(dev)) {
+			vlv_initial_power_sequencer_setup(intel_dp);
+		} else {
+			intel_dp_init_panel_power_timestamps(intel_dp);
+			intel_dp_init_panel_power_sequencer(dev, intel_dp,
+							    &power_seq);
+		}
+		pps_unlock(intel_dp);
 	}
 
 	intel_dp_aux_init(intel_dp, intel_connector);
 
-	intel_dp->psr_setup_done = false;
+	/* init MST on ports that can support it */
+	if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
+		if (port == PORT_B || port == PORT_C || port == PORT_D) {
+			intel_dp_mst_encoder_init(intel_dig_port,
+						  intel_connector->base.base.id);
+		}
+	}
 
 	if (!intel_edp_init_connector(intel_dp, intel_connector, &power_seq)) {
 		drm_dp_aux_unregister(&intel_dp->aux);
 		if (is_edp(intel_dp)) {
 			cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-			drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+			/*
+			 * vdd might still be enabled do to the delayed vdd off.
+			 * Make sure vdd is actually turned off here.
+			 */
+			pps_lock(intel_dp);
 			edp_panel_vdd_off_sync(intel_dp);
-			drm_modeset_unlock(&dev->mode_config.connection_mutex);
+			pps_unlock(intel_dp);
 		}
-		drm_sysfs_connector_remove(connector);
+		drm_connector_unregister(connector);
 		drm_connector_cleanup(connector);
 		return false;
 	}
@@ -4353,6 +5201,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 void
 intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_digital_port *intel_dig_port;
 	struct intel_encoder *intel_encoder;
 	struct drm_encoder *encoder;
@@ -4378,7 +5227,9 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->disable = intel_disable_dp;
 	intel_encoder->get_hw_state = intel_dp_get_hw_state;
 	intel_encoder->get_config = intel_dp_get_config;
+	intel_encoder->suspend = intel_dp_encoder_suspend;
 	if (IS_CHERRYVIEW(dev)) {
+		intel_encoder->pre_pll_enable = chv_dp_pre_pll_enable;
 		intel_encoder->pre_enable = chv_pre_enable_dp;
 		intel_encoder->enable = vlv_enable_dp;
 		intel_encoder->post_disable = chv_post_disable_dp;
@@ -4390,7 +5241,8 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	} else {
 		intel_encoder->pre_enable = g4x_pre_enable_dp;
 		intel_encoder->enable = g4x_enable_dp;
-		intel_encoder->post_disable = g4x_post_disable_dp;
+		if (INTEL_INFO(dev)->gen >= 5)
+			intel_encoder->post_disable = ilk_post_disable_dp;
 	}
 
 	intel_dig_port->port = port;
@@ -4408,9 +5260,55 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->cloneable = 0;
 	intel_encoder->hot_plug = intel_dp_hot_plug;
 
+	intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
+	dev_priv->hpd_irq_port[port] = intel_dig_port;
+
 	if (!intel_dp_init_connector(intel_dig_port, intel_connector)) {
 		drm_encoder_cleanup(encoder);
 		kfree(intel_dig_port);
 		kfree(intel_connector);
 	}
 }
+
+void intel_dp_mst_suspend(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	/* disable MST */
+	for (i = 0; i < I915_MAX_PORTS; i++) {
+		struct intel_digital_port *intel_dig_port = dev_priv->hpd_irq_port[i];
+		if (!intel_dig_port)
+			continue;
+
+		if (intel_dig_port->base.type == INTEL_OUTPUT_DISPLAYPORT) {
+			if (!intel_dig_port->dp.can_mst)
+				continue;
+			if (intel_dig_port->dp.is_mst)
+				drm_dp_mst_topology_mgr_suspend(&intel_dig_port->dp.mst_mgr);
+		}
+	}
+}
+
+void intel_dp_mst_resume(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	for (i = 0; i < I915_MAX_PORTS; i++) {
+		struct intel_digital_port *intel_dig_port = dev_priv->hpd_irq_port[i];
+		if (!intel_dig_port)
+			continue;
+		if (intel_dig_port->base.type == INTEL_OUTPUT_DISPLAYPORT) {
+			int ret;
+
+			if (!intel_dig_port->dp.can_mst)
+				continue;
+
+			ret = drm_dp_mst_topology_mgr_resume(&intel_dig_port->dp.mst_mgr);
+			if (ret != 0) {
+				intel_dp_check_mst_status(&intel_dig_port->dp);
+			}
+		}
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
new file mode 100644
index 000000000000..d9a7a7865f66
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -0,0 +1,548 @@
+/*
+ * Copyright © 2008 Intel Corporation
+ *             2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <drm/drmP.h>
+#include "i915_drv.h"
+#include "intel_drv.h"
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_edid.h>
+
+static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
+					struct intel_crtc_config *pipe_config)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	struct drm_device *dev = encoder->base.dev;
+	int bpp;
+	int lane_count, slots;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	struct intel_connector *found = NULL, *intel_connector;
+	int mst_pbn;
+
+	pipe_config->dp_encoder_is_mst = true;
+	pipe_config->has_pch_encoder = false;
+	pipe_config->has_dp_encoder = true;
+	bpp = 24;
+	/*
+	 * for MST we always configure max link bw - the spec doesn't
+	 * seem to suggest we should do otherwise.
+	 */
+	lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
+	intel_dp->link_bw = intel_dp_max_link_bw(intel_dp);
+	intel_dp->lane_count = lane_count;
+
+	pipe_config->pipe_bpp = 24;
+	pipe_config->port_clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
+
+	list_for_each_entry(intel_connector, &dev->mode_config.connector_list, base.head) {
+		if (intel_connector->new_encoder == encoder) {
+			found = intel_connector;
+			break;
+		}
+	}
+
+	if (!found) {
+		DRM_ERROR("can't find connector\n");
+		return false;
+	}
+
+	mst_pbn = drm_dp_calc_pbn_mode(adjusted_mode->clock, bpp);
+
+	pipe_config->pbn = mst_pbn;
+	slots = drm_dp_find_vcpi_slots(&intel_dp->mst_mgr, mst_pbn);
+
+	intel_link_compute_m_n(bpp, lane_count,
+			       adjusted_mode->crtc_clock,
+			       pipe_config->port_clock,
+			       &pipe_config->dp_m_n);
+
+	pipe_config->dp_m_n.tu = slots;
+	return true;
+
+}
+
+static void intel_mst_disable_dp(struct intel_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	int ret;
+
+	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
+
+	drm_dp_mst_reset_vcpi_slots(&intel_dp->mst_mgr, intel_mst->port);
+
+	ret = drm_dp_update_payload_part1(&intel_dp->mst_mgr);
+	if (ret) {
+		DRM_ERROR("failed to update payload %d\n", ret);
+	}
+}
+
+static void intel_mst_post_disable_dp(struct intel_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+
+	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
+
+	/* this can fail */
+	drm_dp_check_act_status(&intel_dp->mst_mgr);
+	/* and this can also fail */
+	drm_dp_update_payload_part2(&intel_dp->mst_mgr);
+
+	drm_dp_mst_deallocate_vcpi(&intel_dp->mst_mgr, intel_mst->port);
+
+	intel_dp->active_mst_links--;
+	intel_mst->port = NULL;
+	if (intel_dp->active_mst_links == 0) {
+		intel_dig_port->base.post_disable(&intel_dig_port->base);
+		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF);
+	}
+}
+
+static void intel_mst_pre_enable_dp(struct intel_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = intel_dig_port->port;
+	int ret;
+	uint32_t temp;
+	struct intel_connector *found = NULL, *intel_connector;
+	int slots;
+	struct drm_crtc *crtc = encoder->base.crtc;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	list_for_each_entry(intel_connector, &dev->mode_config.connector_list, base.head) {
+		if (intel_connector->new_encoder == encoder) {
+			found = intel_connector;
+			break;
+		}
+	}
+
+	if (!found) {
+		DRM_ERROR("can't find connector\n");
+		return;
+	}
+
+	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
+	intel_mst->port = found->port;
+
+	if (intel_dp->active_mst_links == 0) {
+		enum port port = intel_ddi_get_encoder_port(encoder);
+
+		I915_WRITE(PORT_CLK_SEL(port), intel_crtc->config.ddi_pll_sel);
+
+		intel_ddi_init_dp_buf_reg(&intel_dig_port->base);
+
+		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
+
+
+		intel_dp_start_link_train(intel_dp);
+		intel_dp_complete_link_train(intel_dp);
+		intel_dp_stop_link_train(intel_dp);
+	}
+
+	ret = drm_dp_mst_allocate_vcpi(&intel_dp->mst_mgr,
+				       intel_mst->port, intel_crtc->config.pbn, &slots);
+	if (ret == false) {
+		DRM_ERROR("failed to allocate vcpi\n");
+		return;
+	}
+
+
+	intel_dp->active_mst_links++;
+	temp = I915_READ(DP_TP_STATUS(port));
+	I915_WRITE(DP_TP_STATUS(port), temp);
+
+	ret = drm_dp_update_payload_part1(&intel_dp->mst_mgr);
+}
+
+static void intel_mst_enable_dp(struct intel_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = intel_dig_port->port;
+	int ret;
+
+	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
+
+	if (wait_for((I915_READ(DP_TP_STATUS(port)) & DP_TP_STATUS_ACT_SENT),
+		     1))
+		DRM_ERROR("Timed out waiting for ACT sent\n");
+
+	ret = drm_dp_check_act_status(&intel_dp->mst_mgr);
+
+	ret = drm_dp_update_payload_part2(&intel_dp->mst_mgr);
+}
+
+static bool intel_dp_mst_enc_get_hw_state(struct intel_encoder *encoder,
+				      enum pipe *pipe)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	*pipe = intel_mst->pipe;
+	if (intel_mst->port)
+		return true;
+	return false;
+}
+
+static void intel_dp_mst_enc_get_config(struct intel_encoder *encoder,
+					struct intel_crtc_config *pipe_config)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(&encoder->base);
+	struct intel_digital_port *intel_dig_port = intel_mst->primary;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+	u32 temp, flags = 0;
+
+	pipe_config->has_dp_encoder = true;
+
+	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (temp & TRANS_DDI_PHSYNC)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (temp & TRANS_DDI_PVSYNC)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	switch (temp & TRANS_DDI_BPC_MASK) {
+	case TRANS_DDI_BPC_6:
+		pipe_config->pipe_bpp = 18;
+		break;
+	case TRANS_DDI_BPC_8:
+		pipe_config->pipe_bpp = 24;
+		break;
+	case TRANS_DDI_BPC_10:
+		pipe_config->pipe_bpp = 30;
+		break;
+	case TRANS_DDI_BPC_12:
+		pipe_config->pipe_bpp = 36;
+		break;
+	default:
+		break;
+	}
+	pipe_config->adjusted_mode.flags |= flags;
+	intel_dp_get_m_n(crtc, pipe_config);
+
+	intel_ddi_clock_get(&intel_dig_port->base, pipe_config);
+}
+
+static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+	struct edid *edid;
+	int ret;
+
+	edid = drm_dp_mst_get_edid(connector, &intel_dp->mst_mgr, intel_connector->port);
+	if (!edid)
+		return 0;
+
+	ret = intel_connector_update_modes(connector, edid);
+	kfree(edid);
+
+	return ret;
+}
+
+static enum drm_connector_status
+intel_mst_port_dp_detect(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+
+	return drm_dp_mst_detect_port(&intel_dp->mst_mgr, intel_connector->port);
+}
+
+static enum drm_connector_status
+intel_dp_mst_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+	status = intel_mst_port_dp_detect(connector);
+	return status;
+}
+
+static int
+intel_dp_mst_set_property(struct drm_connector *connector,
+			  struct drm_property *property,
+			  uint64_t val)
+{
+	return 0;
+}
+
+static void
+intel_dp_mst_connector_destroy(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+
+	if (!IS_ERR_OR_NULL(intel_connector->edid))
+		kfree(intel_connector->edid);
+
+	drm_connector_cleanup(connector);
+	kfree(connector);
+}
+
+static const struct drm_connector_funcs intel_dp_mst_connector_funcs = {
+	.dpms = intel_connector_dpms,
+	.detect = intel_dp_mst_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.set_property = intel_dp_mst_set_property,
+	.destroy = intel_dp_mst_connector_destroy,
+};
+
+static int intel_dp_mst_get_modes(struct drm_connector *connector)
+{
+	return intel_dp_mst_get_ddc_modes(connector);
+}
+
+static enum drm_mode_status
+intel_dp_mst_mode_valid(struct drm_connector *connector,
+			struct drm_display_mode *mode)
+{
+	/* TODO - validate mode against available PBN for link */
+	if (mode->clock < 10000)
+		return MODE_CLOCK_LOW;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK)
+		return MODE_H_ILLEGAL;
+
+	return MODE_OK;
+}
+
+static struct drm_encoder *intel_mst_best_encoder(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+	return &intel_dp->mst_encoders[0]->base.base;
+}
+
+static const struct drm_connector_helper_funcs intel_dp_mst_connector_helper_funcs = {
+	.get_modes = intel_dp_mst_get_modes,
+	.mode_valid = intel_dp_mst_mode_valid,
+	.best_encoder = intel_mst_best_encoder,
+};
+
+static void intel_dp_mst_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+
+	drm_encoder_cleanup(encoder);
+	kfree(intel_mst);
+}
+
+static const struct drm_encoder_funcs intel_dp_mst_enc_funcs = {
+	.destroy = intel_dp_mst_encoder_destroy,
+};
+
+static bool intel_dp_mst_get_hw_state(struct intel_connector *connector)
+{
+	if (connector->encoder) {
+		enum pipe pipe;
+		if (!connector->encoder->get_hw_state(connector->encoder, &pipe))
+			return false;
+		return true;
+	}
+	return false;
+}
+
+static void intel_connector_add_to_fbdev(struct intel_connector *connector)
+{
+#ifdef CONFIG_DRM_I915_FBDEV
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	drm_fb_helper_add_one_connector(&dev_priv->fbdev->helper, &connector->base);
+#endif
+}
+
+static void intel_connector_remove_from_fbdev(struct intel_connector *connector)
+{
+#ifdef CONFIG_DRM_I915_FBDEV
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	drm_fb_helper_remove_one_connector(&dev_priv->fbdev->helper, &connector->base);
+#endif
+}
+
+static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port, char *pathprop)
+{
+	struct intel_dp *intel_dp = container_of(mgr, struct intel_dp, mst_mgr);
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+	int i;
+
+	intel_connector = kzalloc(sizeof(*intel_connector), GFP_KERNEL);
+	if (!intel_connector)
+		return NULL;
+
+	connector = &intel_connector->base;
+	drm_connector_init(dev, connector, &intel_dp_mst_connector_funcs, DRM_MODE_CONNECTOR_DisplayPort);
+	drm_connector_helper_add(connector, &intel_dp_mst_connector_helper_funcs);
+
+	intel_connector->unregister = intel_connector_unregister;
+	intel_connector->get_hw_state = intel_dp_mst_get_hw_state;
+	intel_connector->mst_port = intel_dp;
+	intel_connector->port = port;
+
+	for (i = PIPE_A; i <= PIPE_C; i++) {
+		drm_mode_connector_attach_encoder(&intel_connector->base,
+						  &intel_dp->mst_encoders[i]->base.base);
+	}
+	intel_dp_add_properties(intel_dp, connector);
+
+	drm_object_attach_property(&connector->base, dev->mode_config.path_property, 0);
+	drm_mode_connector_set_path_property(connector, pathprop);
+	drm_reinit_primary_mode_group(dev);
+	mutex_lock(&dev->mode_config.mutex);
+	intel_connector_add_to_fbdev(intel_connector);
+	mutex_unlock(&dev->mode_config.mutex);
+	drm_connector_register(&intel_connector->base);
+	return connector;
+}
+
+static void intel_dp_destroy_mst_connector(struct drm_dp_mst_topology_mgr *mgr,
+					   struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct drm_device *dev = connector->dev;
+	/* need to nuke the connector */
+	mutex_lock(&dev->mode_config.mutex);
+	intel_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	intel_connector->unregister(intel_connector);
+
+	mutex_lock(&dev->mode_config.mutex);
+	intel_connector_remove_from_fbdev(intel_connector);
+	drm_connector_cleanup(connector);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	drm_reinit_primary_mode_group(dev);
+
+	kfree(intel_connector);
+	DRM_DEBUG_KMS("\n");
+}
+
+static void intel_dp_mst_hotplug(struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct intel_dp *intel_dp = container_of(mgr, struct intel_dp, mst_mgr);
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+
+	drm_kms_helper_hotplug_event(dev);
+}
+
+static struct drm_dp_mst_topology_cbs mst_cbs = {
+	.add_connector = intel_dp_add_mst_connector,
+	.destroy_connector = intel_dp_destroy_mst_connector,
+	.hotplug = intel_dp_mst_hotplug,
+};
+
+static struct intel_dp_mst_encoder *
+intel_dp_create_fake_mst_encoder(struct intel_digital_port *intel_dig_port, enum pipe pipe)
+{
+	struct intel_dp_mst_encoder *intel_mst;
+	struct intel_encoder *intel_encoder;
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+
+	intel_mst = kzalloc(sizeof(*intel_mst), GFP_KERNEL);
+
+	if (!intel_mst)
+		return NULL;
+
+	intel_mst->pipe = pipe;
+	intel_encoder = &intel_mst->base;
+	intel_mst->primary = intel_dig_port;
+
+	drm_encoder_init(dev, &intel_encoder->base, &intel_dp_mst_enc_funcs,
+			 DRM_MODE_ENCODER_DPMST);
+
+	intel_encoder->type = INTEL_OUTPUT_DP_MST;
+	intel_encoder->crtc_mask = 0x7;
+	intel_encoder->cloneable = 0;
+
+	intel_encoder->compute_config = intel_dp_mst_compute_config;
+	intel_encoder->disable = intel_mst_disable_dp;
+	intel_encoder->post_disable = intel_mst_post_disable_dp;
+	intel_encoder->pre_enable = intel_mst_pre_enable_dp;
+	intel_encoder->enable = intel_mst_enable_dp;
+	intel_encoder->get_hw_state = intel_dp_mst_enc_get_hw_state;
+	intel_encoder->get_config = intel_dp_mst_enc_get_config;
+
+	return intel_mst;
+
+}
+
+static bool
+intel_dp_create_fake_mst_encoders(struct intel_digital_port *intel_dig_port)
+{
+	int i;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+
+	for (i = PIPE_A; i <= PIPE_C; i++)
+		intel_dp->mst_encoders[i] = intel_dp_create_fake_mst_encoder(intel_dig_port, i);
+	return true;
+}
+
+int
+intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_base_id)
+{
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	int ret;
+
+	intel_dp->can_mst = true;
+	intel_dp->mst_mgr.cbs = &mst_cbs;
+
+	/* create encoders */
+	intel_dp_create_fake_mst_encoders(intel_dig_port);
+	ret = drm_dp_mst_topology_mgr_init(&intel_dp->mst_mgr, dev->dev, &intel_dp->aux, 16, 3, conn_base_id);
+	if (ret) {
+		intel_dp->can_mst = false;
+		return ret;
+	}
+	return 0;
+}
+
+void
+intel_dp_mst_encoder_cleanup(struct intel_digital_port *intel_dig_port)
+{
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+
+	if (!intel_dp->can_mst)
+		return;
+
+	drm_dp_mst_topology_mgr_destroy(&intel_dp->mst_mgr);
+	/* encoders will get killed by normal cleanup */
+}
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index f67340ed2c12..ba715229a540 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -25,6 +25,7 @@
 #ifndef __INTEL_DRV_H__
 #define __INTEL_DRV_H__
 
+#include <linux/async.h>
 #include <linux/i2c.h>
 #include <linux/hdmi.h>
 #include <drm/i915_drm.h>
@@ -32,7 +33,10 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
-#include <drm/drm_dp_helper.h>
+#include <drm/drm_dp_mst_helper.h>
+
+#define DIV_ROUND_CLOSEST_ULL(ll, d)	\
+({ unsigned long long _tmp = (ll)+(d)/2; do_div(_tmp, d); _tmp; })
 
 /**
  * _wait_for - magic (register) wait macro
@@ -100,6 +104,7 @@
 #define INTEL_OUTPUT_EDP 8
 #define INTEL_OUTPUT_DSI 9
 #define INTEL_OUTPUT_UNKNOWN 10
+#define INTEL_OUTPUT_DP_MST 11
 
 #define INTEL_DVO_CHIP_NONE 0
 #define INTEL_DVO_CHIP_LVDS 1
@@ -152,6 +157,12 @@ struct intel_encoder {
 	 * be set correctly before calling this function. */
 	void (*get_config)(struct intel_encoder *,
 			   struct intel_crtc_config *pipe_config);
+	/*
+	 * Called during system suspend after all pending requests for the
+	 * encoder are flushed (for example for DP AUX transactions) and
+	 * device interrupts are disabled.
+	 */
+	void (*suspend)(struct intel_encoder *);
 	int crtc_mask;
 	enum hpd_pin hpd_pin;
 };
@@ -165,12 +176,15 @@ struct intel_panel {
 	struct {
 		bool present;
 		u32 level;
+		u32 min;
 		u32 max;
 		bool enabled;
 		bool combination_mode;	/* gen 2/4 only */
 		bool active_low_pwm;
 		struct backlight_device *device;
 	} backlight;
+
+	void (*backlight_power)(struct intel_connector *, bool enable);
 };
 
 struct intel_connector {
@@ -203,10 +217,15 @@ struct intel_connector {
 
 	/* Cached EDID for eDP and LVDS. May hold ERR_PTR for invalid EDID. */
 	struct edid *edid;
+	struct edid *detect_edid;
 
 	/* since POLL and HPD connectors may use the same HPD line keep the native
 	   state of connector->polled in case hotplug storm detection changes it */
 	u8 polled;
+
+	void *port; /* store this opaque as its illegal to dereference it */
+
+	struct intel_dp *mst_port;
 };
 
 typedef struct dpll {
@@ -307,6 +326,9 @@ struct intel_crtc_config {
 	/* Selected dpll when shared or DPLL_ID_PRIVATE. */
 	enum intel_dpll_id shared_dpll;
 
+	/* PORT_CLK_SEL for DDI ports. */
+	uint32_t ddi_pll_sel;
+
 	/* Actual register state of the dpll, for shared dpll cross-checking. */
 	struct intel_dpll_hw_state dpll_hw_state;
 
@@ -315,6 +337,7 @@ struct intel_crtc_config {
 
 	/* m2_n2 for eDP downclock */
 	struct intel_link_m_n dp_m2_n2;
+	bool has_drrs;
 
 	/*
 	 * Frequence the dpll for the port should run at. Differs from the
@@ -338,6 +361,7 @@ struct intel_crtc_config {
 		u32 pos;
 		u32 size;
 		bool enabled;
+		bool force_thru;
 	} pch_pfit;
 
 	/* FDI configuration, only valid if has_pch_encoder is set. */
@@ -347,6 +371,9 @@ struct intel_crtc_config {
 	bool ips_enabled;
 
 	bool double_wide;
+
+	bool dp_encoder_is_mst;
+	int pbn;
 };
 
 struct intel_pipe_wm {
@@ -358,6 +385,11 @@ struct intel_pipe_wm {
 	bool sprites_scaled;
 };
 
+struct intel_mmio_flip {
+	u32 seqno;
+	u32 ring_id;
+};
+
 struct intel_crtc {
 	struct drm_crtc base;
 	enum pipe pipe;
@@ -384,9 +416,9 @@ struct intel_crtc {
 
 	struct drm_i915_gem_object *cursor_bo;
 	uint32_t cursor_addr;
-	int16_t cursor_x, cursor_y;
 	int16_t cursor_width, cursor_height;
 	uint32_t cursor_cntl;
+	uint32_t cursor_size;
 	uint32_t cursor_base;
 
 	struct intel_plane_config plane_config;
@@ -394,8 +426,6 @@ struct intel_crtc {
 	struct intel_crtc_config *new_config;
 	bool new_enabled;
 
-	uint32_t ddi_pll_sel;
-
 	/* reset counter value when the last flip was submitted */
 	unsigned int reset_counter;
 
@@ -409,13 +439,13 @@ struct intel_crtc {
 		struct intel_pipe_wm active;
 	} wm;
 
-	wait_queue_head_t vbl_wait;
-
 	int scanline_offset;
+	struct intel_mmio_flip mmio_flip;
 };
 
 struct intel_plane_wm_parameters {
 	uint32_t horiz_pixels;
+	uint32_t vert_pixels;
 	uint8_t bytes_per_pixel;
 	bool enabled;
 	bool scaled;
@@ -428,11 +458,11 @@ struct intel_plane {
 	struct drm_i915_gem_object *obj;
 	bool can_scale;
 	int max_downscale;
-	u32 lut_r[1024], lut_g[1024], lut_b[1024];
 	int crtc_x, crtc_y;
 	unsigned int crtc_w, crtc_h;
 	uint32_t src_x, src_y;
 	uint32_t src_w, src_h;
+	unsigned int rotation;
 
 	/* Since we need to change the watermarks before/after
 	 * enabling/disabling the planes, we need to store the parameters here
@@ -481,6 +511,7 @@ struct cxsr_latency {
 #define to_intel_encoder(x) container_of(x, struct intel_encoder, base)
 #define to_intel_framebuffer(x) container_of(x, struct intel_framebuffer, base)
 #define to_intel_plane(x) container_of(x, struct intel_plane, base)
+#define intel_fb_obj(x) (x ? to_intel_framebuffer(x)->obj : NULL)
 
 struct intel_hdmi {
 	u32 hdmi_reg;
@@ -491,6 +522,7 @@ struct intel_hdmi {
 	bool has_audio;
 	enum hdmi_force_audio force_audio;
 	bool rgb_quant_range_selectable;
+	enum hdmi_picture_aspect aspect_ratio;
 	void (*write_infoframe)(struct drm_encoder *encoder,
 				enum hdmi_infoframe_type type,
 				const void *frame, ssize_t len);
@@ -499,6 +531,7 @@ struct intel_hdmi {
 			       struct drm_display_mode *adjusted_mode);
 };
 
+struct intel_dp_mst_encoder;
 #define DP_MAX_DOWNSTREAM_PORTS		0x10
 
 /**
@@ -537,12 +570,26 @@ struct intel_dp {
 	unsigned long last_power_cycle;
 	unsigned long last_power_on;
 	unsigned long last_backlight_off;
-	bool psr_setup_done;
+
 	struct notifier_block edp_notifier;
 
+	/*
+	 * Pipe whose power sequencer is currently locked into
+	 * this port. Only relevant on VLV/CHV.
+	 */
+	enum pipe pps_pipe;
+
 	bool use_tps3;
+	bool can_mst; /* this port supports mst */
+	bool is_mst;
+	int active_mst_links;
+	/* connector directly attached - won't be use for modeset in mst world */
 	struct intel_connector *attached_connector;
 
+	/* mst connector list */
+	struct intel_dp_mst_encoder *mst_encoders[I915_MAX_PIPES];
+	struct drm_dp_mst_topology_mgr mst_mgr;
+
 	uint32_t (*get_aux_clock_divider)(struct intel_dp *dp, int index);
 	/*
 	 * This function returns the value we have to program the AUX_CTL
@@ -566,6 +613,14 @@ struct intel_digital_port {
 	u32 saved_port_bits;
 	struct intel_dp dp;
 	struct intel_hdmi hdmi;
+	bool (*hpd_pulse)(struct intel_digital_port *, bool);
+};
+
+struct intel_dp_mst_encoder {
+	struct intel_encoder base;
+	enum pipe pipe;
+	struct intel_digital_port *primary;
+	void *port; /* store this opaque as its illegal to dereference it */
 };
 
 static inline int
@@ -622,6 +677,10 @@ struct intel_unpin_work {
 #define INTEL_FLIP_COMPLETE	2
 	u32 flip_count;
 	u32 gtt_offset;
+	struct intel_engine_cs *flip_queued_ring;
+	u32 flip_queued_seqno;
+	int flip_queued_vblank;
+	int flip_ready_vblank;
 	bool enable_stall_check;
 };
 
@@ -652,6 +711,12 @@ enc_to_dig_port(struct drm_encoder *encoder)
 	return container_of(encoder, struct intel_digital_port, base.base);
 }
 
+static inline struct intel_dp_mst_encoder *
+enc_to_mst(struct drm_encoder *encoder)
+{
+	return container_of(encoder, struct intel_dp_mst_encoder, base.base);
+}
+
 static inline struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
 {
 	return &enc_to_dig_port(encoder)->dp;
@@ -676,17 +741,26 @@ bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
 bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
 					   enum transcoder pch_transcoder,
 					   bool enable);
-void ilk_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void ilk_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void snb_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void snb_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void bdw_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
-void bdw_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
+void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask);
 void intel_runtime_pm_disable_interrupts(struct drm_device *dev);
 void intel_runtime_pm_restore_interrupts(struct drm_device *dev);
+static inline bool intel_irqs_enabled(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * We only use drm_irq_uninstall() at unload and VT switch, so
+	 * this is the only thing we need to check.
+	 */
+	return !dev_priv->pm._irqs_disabled;
+}
+
 int intel_get_crtc_scanline(struct intel_crtc *crtc);
 void i9xx_check_fifo_underruns(struct drm_device *dev);
-
+void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv);
 
 /* intel_crt.c */
 void intel_crt_init(struct drm_device *dev);
@@ -705,10 +779,7 @@ void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 				       enum transcoder cpu_transcoder);
 void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc);
 void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc);
-void intel_ddi_setup_hw_pll_state(struct drm_device *dev);
 bool intel_ddi_pll_select(struct intel_crtc *crtc);
-void intel_ddi_pll_enable(struct intel_crtc *crtc);
-void intel_ddi_put_crtc_pll(struct drm_crtc *crtc);
 void intel_ddi_set_pipe_settings(struct drm_crtc *crtc);
 void intel_ddi_prepare_link_retrain(struct drm_encoder *encoder);
 bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
@@ -716,17 +787,46 @@ void intel_ddi_fdi_disable(struct drm_crtc *crtc);
 void intel_ddi_get_config(struct intel_encoder *encoder,
 			  struct intel_crtc_config *pipe_config);
 
+void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder);
+void intel_ddi_clock_get(struct intel_encoder *encoder,
+			 struct intel_crtc_config *pipe_config);
+void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state);
 
 /* intel_display.c */
 const char *intel_output_name(int output);
 bool intel_has_pending_fb_unpin(struct drm_device *dev);
 int intel_pch_rawclk(struct drm_device *dev);
-int valleyview_cur_cdclk(struct drm_i915_private *dev_priv);
 void intel_mark_busy(struct drm_device *dev);
-void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
-			struct intel_engine_cs *ring);
+void intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
+			     struct intel_engine_cs *ring);
+void intel_frontbuffer_flip_prepare(struct drm_device *dev,
+				    unsigned frontbuffer_bits);
+void intel_frontbuffer_flip_complete(struct drm_device *dev,
+				     unsigned frontbuffer_bits);
+void intel_frontbuffer_flush(struct drm_device *dev,
+			     unsigned frontbuffer_bits);
+/**
+ * intel_frontbuffer_flip - prepare frontbuffer flip
+ * @dev: DRM device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after scheduling a flip on @obj. This is for
+ * synchronous plane updates which will happen on the next vblank and which will
+ * not get delayed by pending gpu rendering.
+ *
+ * Can be called without any locks held.
+ */
+static inline
+void intel_frontbuffer_flip(struct drm_device *dev,
+			    unsigned frontbuffer_bits)
+{
+	intel_frontbuffer_flush(dev, frontbuffer_bits);
+}
+
+void intel_fb_obj_flush(struct drm_i915_gem_object *obj, bool retire);
 void intel_mark_idle(struct drm_device *dev);
 void intel_crtc_restore_mode(struct drm_crtc *crtc);
+void intel_crtc_control(struct drm_crtc *crtc, bool enable);
 void intel_crtc_update_dpms(struct drm_crtc *crtc);
 void intel_encoder_destroy(struct drm_encoder *encoder);
 void intel_connector_dpms(struct drm_connector *, int mode);
@@ -745,7 +845,6 @@ int intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data,
 enum transcoder intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 					     enum pipe pipe);
 void intel_wait_for_vblank(struct drm_device *dev, int pipe);
-void intel_wait_for_pipe_off(struct drm_device *dev, int pipe);
 int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp);
 void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
 			 struct intel_digital_port *dport);
@@ -754,8 +853,7 @@ bool intel_get_load_detect_pipe(struct drm_connector *connector,
 				struct intel_load_detect_pipe *old,
 				struct drm_modeset_acquire_ctx *ctx);
 void intel_release_load_detect_pipe(struct drm_connector *connector,
-				    struct intel_load_detect_pipe *old,
-				    struct drm_modeset_acquire_ctx *ctx);
+				    struct intel_load_detect_pipe *old);
 int intel_pin_and_fence_fb_obj(struct drm_device *dev,
 			       struct drm_i915_gem_object *obj,
 			       struct intel_engine_cs *pipelined);
@@ -767,12 +865,19 @@ __intel_framebuffer_create(struct drm_device *dev,
 void intel_prepare_page_flip(struct drm_device *dev, int plane);
 void intel_finish_page_flip(struct drm_device *dev, int pipe);
 void intel_finish_page_flip_plane(struct drm_device *dev, int plane);
+void intel_check_page_flip(struct drm_device *dev, int pipe);
+
+/* shared dpll functions */
 struct intel_shared_dpll *intel_crtc_to_shared_dpll(struct intel_crtc *crtc);
 void assert_shared_dpll(struct drm_i915_private *dev_priv,
 			struct intel_shared_dpll *pll,
 			bool state);
 #define assert_shared_dpll_enabled(d, p) assert_shared_dpll(d, p, true)
 #define assert_shared_dpll_disabled(d, p) assert_shared_dpll(d, p, false)
+struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc);
+void intel_put_shared_dpll(struct intel_crtc *crtc);
+
+/* modesetting asserts */
 void assert_pll(struct drm_i915_private *dev_priv,
 		enum pipe pipe, bool state);
 #define assert_pll_enabled(d, p) assert_pll(d, p, true)
@@ -795,6 +900,7 @@ void hsw_enable_pc8(struct drm_i915_private *dev_priv);
 void hsw_disable_pc8(struct drm_i915_private *dev_priv);
 void intel_dp_get_m_n(struct intel_crtc *crtc,
 		      struct intel_crtc_config *pipe_config);
+void intel_dp_set_m_n(struct intel_crtc *crtc);
 int intel_dotclock_calculate(int link_freq, const struct intel_link_m_n *m_n);
 void
 ironlake_check_encoder_dotclock(const struct intel_crtc_config *pipe_config,
@@ -805,12 +911,11 @@ void hsw_disable_ips(struct intel_crtc *crtc);
 void intel_display_set_init_power(struct drm_i915_private *dev, bool enable);
 enum intel_display_power_domain
 intel_display_port_power_domain(struct intel_encoder *intel_encoder);
-int valleyview_get_vco(struct drm_i915_private *dev_priv);
 void intel_mode_from_pipe_config(struct drm_display_mode *mode,
 				 struct intel_crtc_config *pipe_config);
 int intel_format_to_fourcc(int format);
 void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc);
-
+void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file);
 
 /* intel_dp.c */
 void intel_dp_init(struct drm_device *dev, int output_reg, enum port port);
@@ -826,18 +931,35 @@ int intel_dp_sink_crc(struct intel_dp *intel_dp, u8 *crc);
 bool intel_dp_compute_config(struct intel_encoder *encoder,
 			     struct intel_crtc_config *pipe_config);
 bool intel_dp_is_edp(struct drm_device *dev, enum port port);
+bool intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port,
+			bool long_hpd);
 void intel_edp_backlight_on(struct intel_dp *intel_dp);
 void intel_edp_backlight_off(struct intel_dp *intel_dp);
 void intel_edp_panel_vdd_on(struct intel_dp *intel_dp);
+void intel_edp_panel_vdd_sanitize(struct intel_encoder *intel_encoder);
 void intel_edp_panel_on(struct intel_dp *intel_dp);
 void intel_edp_panel_off(struct intel_dp *intel_dp);
 void intel_edp_psr_enable(struct intel_dp *intel_dp);
 void intel_edp_psr_disable(struct intel_dp *intel_dp);
-void intel_edp_psr_update(struct drm_device *dev);
 void intel_dp_set_drrs_state(struct drm_device *dev, int refresh_rate);
-
+void intel_edp_psr_invalidate(struct drm_device *dev,
+			      unsigned frontbuffer_bits);
+void intel_edp_psr_flush(struct drm_device *dev,
+			 unsigned frontbuffer_bits);
+void intel_edp_psr_init(struct drm_device *dev);
+
+int intel_dp_handle_hpd_irq(struct intel_digital_port *digport, bool long_hpd);
+void intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connector);
+void intel_dp_mst_suspend(struct drm_device *dev);
+void intel_dp_mst_resume(struct drm_device *dev);
+int intel_dp_max_link_bw(struct intel_dp *intel_dp);
+void intel_dp_hot_plug(struct intel_encoder *intel_encoder);
+void vlv_power_sequencer_reset(struct drm_i915_private *dev_priv);
+/* intel_dp_mst.c */
+int intel_dp_mst_encoder_init(struct intel_digital_port *intel_dig_port, int conn_id);
+void intel_dp_mst_encoder_cleanup(struct intel_digital_port *intel_dig_port);
 /* intel_dsi.c */
-bool intel_dsi_init(struct drm_device *dev);
+void intel_dsi_init(struct drm_device *dev);
 
 
 /* intel_dvo.c */
@@ -847,9 +969,9 @@ void intel_dvo_init(struct drm_device *dev);
 /* legacy fbdev emulation in intel_fbdev.c */
 #ifdef CONFIG_DRM_I915_FBDEV
 extern int intel_fbdev_init(struct drm_device *dev);
-extern void intel_fbdev_initial_config(struct drm_device *dev);
+extern void intel_fbdev_initial_config(void *data, async_cookie_t cookie);
 extern void intel_fbdev_fini(struct drm_device *dev);
-extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
+extern void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous);
 extern void intel_fbdev_output_poll_changed(struct drm_device *dev);
 extern void intel_fbdev_restore_mode(struct drm_device *dev);
 #else
@@ -858,7 +980,7 @@ static inline int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-static inline void intel_fbdev_initial_config(struct drm_device *dev)
+static inline void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
 }
 
@@ -866,7 +988,7 @@ static inline void intel_fbdev_fini(struct drm_device *dev)
 {
 }
 
-static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state)
+static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
 {
 }
 
@@ -920,8 +1042,8 @@ void intel_pch_panel_fitting(struct intel_crtc *crtc,
 void intel_gmch_panel_fitting(struct intel_crtc *crtc,
 			      struct intel_crtc_config *pipe_config,
 			      int fitting_mode);
-void intel_panel_set_backlight(struct intel_connector *connector, u32 level,
-			       u32 max);
+void intel_panel_set_backlight_acpi(struct intel_connector *connector,
+				    u32 level, u32 max);
 int intel_panel_setup_backlight(struct drm_connector *connector);
 void intel_panel_enable_backlight(struct intel_connector *connector);
 void intel_panel_disable_backlight(struct intel_connector *connector);
@@ -940,7 +1062,9 @@ int ilk_wm_max_level(const struct drm_device *dev);
 void intel_update_watermarks(struct drm_crtc *crtc);
 void intel_update_sprite_watermarks(struct drm_plane *plane,
 				    struct drm_crtc *crtc,
-				    uint32_t sprite_width, int pixel_size,
+				    uint32_t sprite_width,
+				    uint32_t sprite_height,
+				    int pixel_size,
 				    bool enabled, bool scaled);
 void intel_init_pm(struct drm_device *dev);
 void intel_pm_setup(struct drm_device *dev);
@@ -963,6 +1087,7 @@ void intel_init_gt_powersave(struct drm_device *dev);
 void intel_cleanup_gt_powersave(struct drm_device *dev);
 void intel_enable_gt_powersave(struct drm_device *dev);
 void intel_disable_gt_powersave(struct drm_device *dev);
+void intel_suspend_gt_powersave(struct drm_device *dev);
 void intel_reset_gt_powersave(struct drm_device *dev);
 void ironlake_teardown_rc6(struct drm_device *dev);
 void gen6_update_ring_freq(struct drm_device *dev);
@@ -976,8 +1101,7 @@ void intel_runtime_pm_put(struct drm_i915_private *dev_priv);
 void intel_init_runtime_pm(struct drm_i915_private *dev_priv);
 void intel_fini_runtime_pm(struct drm_i915_private *dev_priv);
 void ilk_wm_get_hw_state(struct drm_device *dev);
-void __vlv_set_power_well(struct drm_i915_private *dev_priv,
-			  enum punit_power_well power_well_id, bool enable);
+
 
 /* intel_sdvo.c */
 bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob);
@@ -987,7 +1111,10 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob);
 int intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane);
 void intel_flush_primary_plane(struct drm_i915_private *dev_priv,
 			       enum plane plane);
-void intel_plane_restore(struct drm_plane *plane);
+int intel_plane_set_property(struct drm_plane *plane,
+			     struct drm_property *prop,
+			     uint64_t val);
+int intel_plane_restore(struct drm_plane *plane);
 void intel_plane_disable(struct drm_plane *plane);
 int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c
index 3fd082933c87..5bd9e09ad3c5 100644
--- a/drivers/gpu/drm/i915/intel_dsi.c
+++ b/drivers/gpu/drm/i915/intel_dsi.c
@@ -92,6 +92,9 @@ static bool intel_dsi_compute_config(struct intel_encoder *encoder,
 	if (fixed_mode)
 		intel_fixed_panel_mode(fixed_mode, adjusted_mode);
 
+	/* DSI uses short packets for sync events, so clear mode flags for DSI */
+	adjusted_mode->flags = 0;
+
 	if (intel_dsi->dev.dev_ops->mode_fixup)
 		return intel_dsi->dev.dev_ops->mode_fixup(&intel_dsi->dev,
 							  mode, adjusted_mode);
@@ -152,6 +155,8 @@ static void intel_dsi_enable(struct intel_encoder *encoder)
 		if (intel_dsi->dev.dev_ops->enable)
 			intel_dsi->dev.dev_ops->enable(&intel_dsi->dev);
 
+		wait_for_dsi_fifo_empty(intel_dsi);
+
 		/* assert ip_tg_enable signal */
 		temp = I915_READ(MIPI_PORT_CTRL(pipe)) & ~LANE_CONFIGURATION_MASK;
 		temp = temp | intel_dsi->port_bits;
@@ -177,6 +182,10 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder)
 	tmp |= DPLL_REFA_CLK_ENABLE_VLV;
 	I915_WRITE(DPLL(pipe), tmp);
 
+	/* update the hw state for DPLL */
+	intel_crtc->config.dpll_hw_state.dpll = DPLL_INTEGRATED_CLOCK_VLV |
+		DPLL_REFA_CLK_ENABLE_VLV;
+
 	tmp = I915_READ(DSPCLK_GATE_D);
 	tmp |= DPOUNIT_CLOCK_GATE_DISABLE;
 	I915_WRITE(DSPCLK_GATE_D, tmp);
@@ -192,6 +201,8 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder)
 	if (intel_dsi->dev.dev_ops->send_otp_cmds)
 		intel_dsi->dev.dev_ops->send_otp_cmds(&intel_dsi->dev);
 
+	wait_for_dsi_fifo_empty(intel_dsi);
+
 	/* Enable port in pre-enable phase itself because as per hw team
 	 * recommendation, port should be enabled befor plane & pipe */
 	intel_dsi_enable(encoder);
@@ -232,6 +243,8 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 	DRM_DEBUG_KMS("\n");
 
 	if (is_vid_mode(intel_dsi)) {
+		wait_for_dsi_fifo_empty(intel_dsi);
+
 		/* de-assert ip_tg_enable signal */
 		temp = I915_READ(MIPI_PORT_CTRL(pipe));
 		I915_WRITE(MIPI_PORT_CTRL(pipe), temp & ~DPI_ENABLE);
@@ -246,8 +259,8 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 	temp = I915_READ(MIPI_CTRL(pipe));
 	temp &= ~ESCAPE_CLOCK_DIVIDER_MASK;
 	I915_WRITE(MIPI_CTRL(pipe), temp |
-			intel_dsi->escape_clk_div <<
-			ESCAPE_CLOCK_DIVIDER_SHIFT);
+		   intel_dsi->escape_clk_div <<
+		   ESCAPE_CLOCK_DIVIDER_SHIFT);
 
 	I915_WRITE(MIPI_EOT_DISABLE(pipe), CLOCKSTOP);
 
@@ -261,6 +274,8 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 	 * some next enable sequence send turn on packet error is observed */
 	if (intel_dsi->dev.dev_ops->disable)
 		intel_dsi->dev.dev_ops->disable(&intel_dsi->dev);
+
+	wait_for_dsi_fifo_empty(intel_dsi);
 }
 
 static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
@@ -282,7 +297,7 @@ static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 	usleep_range(2000, 2500);
 
 	if (wait_for(((I915_READ(MIPI_PORT_CTRL(pipe)) & AFE_LATCHOUT)
-					== 0x00000), 30))
+		      == 0x00000), 30))
 		DRM_ERROR("DSI LP not going Low\n");
 
 	val = I915_READ(MIPI_PORT_CTRL(pipe));
@@ -351,9 +366,21 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
 static void intel_dsi_get_config(struct intel_encoder *encoder,
 				 struct intel_crtc_config *pipe_config)
 {
+	u32 pclk;
 	DRM_DEBUG_KMS("\n");
 
-	/* XXX: read flags, set to adjusted_mode */
+	/*
+	 * DPLL_MD is not used in case of DSI, reading will get some default value
+	 * set dpll_md = 0
+	 */
+	pipe_config->dpll_hw_state.dpll_md = 0;
+
+	pclk = vlv_get_dsi_pclk(encoder, pipe_config->pipe_bpp);
+	if (!pclk)
+		return;
+
+	pipe_config->adjusted_mode.crtc_clock = pclk;
+	pipe_config->port_clock = pclk;
 }
 
 static enum drm_mode_status
@@ -396,9 +423,11 @@ static u16 txclkesc(u32 divider, unsigned int us)
 }
 
 /* return pixels in terms of txbyteclkhs */
-static u16 txbyteclkhs(u16 pixels, int bpp, int lane_count)
+static u16 txbyteclkhs(u16 pixels, int bpp, int lane_count,
+		       u16 burst_mode_ratio)
 {
-	return DIV_ROUND_UP(DIV_ROUND_UP(pixels * bpp, 8), lane_count);
+	return DIV_ROUND_UP(DIV_ROUND_UP(pixels * bpp * burst_mode_ratio,
+					 8 * 100), lane_count);
 }
 
 static void set_dsi_timings(struct drm_encoder *encoder,
@@ -424,10 +453,12 @@ static void set_dsi_timings(struct drm_encoder *encoder,
 	vbp = mode->vtotal - mode->vsync_end;
 
 	/* horizontal values are in terms of high speed byte clock */
-	hactive = txbyteclkhs(hactive, bpp, lane_count);
-	hfp = txbyteclkhs(hfp, bpp, lane_count);
-	hsync = txbyteclkhs(hsync, bpp, lane_count);
-	hbp = txbyteclkhs(hbp, bpp, lane_count);
+	hactive = txbyteclkhs(hactive, bpp, lane_count,
+			      intel_dsi->burst_mode_ratio);
+	hfp = txbyteclkhs(hfp, bpp, lane_count, intel_dsi->burst_mode_ratio);
+	hsync = txbyteclkhs(hsync, bpp, lane_count,
+			    intel_dsi->burst_mode_ratio);
+	hbp = txbyteclkhs(hbp, bpp, lane_count, intel_dsi->burst_mode_ratio);
 
 	I915_WRITE(MIPI_HACTIVE_AREA_COUNT(pipe), hactive);
 	I915_WRITE(MIPI_HFP_COUNT(pipe), hfp);
@@ -514,12 +545,14 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 	    intel_dsi->video_mode_format == VIDEO_MODE_BURST) {
 		I915_WRITE(MIPI_HS_TX_TIMEOUT(pipe),
 			   txbyteclkhs(adjusted_mode->htotal, bpp,
-				       intel_dsi->lane_count) + 1);
+				       intel_dsi->lane_count,
+				       intel_dsi->burst_mode_ratio) + 1);
 	} else {
 		I915_WRITE(MIPI_HS_TX_TIMEOUT(pipe),
 			   txbyteclkhs(adjusted_mode->vtotal *
 				       adjusted_mode->htotal,
-				       bpp, intel_dsi->lane_count) + 1);
+				       bpp, intel_dsi->lane_count,
+				       intel_dsi->burst_mode_ratio) + 1);
 	}
 	I915_WRITE(MIPI_LP_RX_TIMEOUT(pipe), intel_dsi->lp_rx_timeout);
 	I915_WRITE(MIPI_TURN_AROUND_TIMEOUT(pipe), intel_dsi->turn_arnd_val);
@@ -549,7 +582,7 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 	 * XXX: write MIPI_STOP_STATE_STALL?
 	 */
 	I915_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT(pipe),
-						intel_dsi->hs_to_lp_count);
+		   intel_dsi->hs_to_lp_count);
 
 	/* XXX: low power clock equivalence in terms of byte clock. the number
 	 * of byte clocks occupied in one low power clock. based on txbyteclkhs
@@ -574,10 +607,10 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder)
 		 * 64 like 1366 x 768. Enable RANDOM resolution support for such
 		 * panels by default */
 		I915_WRITE(MIPI_VIDEO_MODE_FORMAT(pipe),
-				intel_dsi->video_frmt_cfg_bits |
-				intel_dsi->video_mode_format |
-				IP_TG_CONFIG |
-				RANDOM_DPI_DISPLAY_RESOLUTION);
+			   intel_dsi->video_frmt_cfg_bits |
+			   intel_dsi->video_mode_format |
+			   IP_TG_CONFIG |
+			   RANDOM_DPI_DISPLAY_RESOLUTION);
 }
 
 static void intel_dsi_pre_pll_enable(struct intel_encoder *encoder)
@@ -658,7 +691,7 @@ static const struct drm_connector_funcs intel_dsi_connector_funcs = {
 	.fill_modes = drm_helper_probe_single_connector_modes,
 };
 
-bool intel_dsi_init(struct drm_device *dev)
+void intel_dsi_init(struct drm_device *dev)
 {
 	struct intel_dsi *intel_dsi;
 	struct intel_encoder *intel_encoder;
@@ -674,29 +707,29 @@ bool intel_dsi_init(struct drm_device *dev)
 
 	/* There is no detection method for MIPI so rely on VBT */
 	if (!dev_priv->vbt.has_mipi)
-		return false;
+		return;
+
+	if (IS_VALLEYVIEW(dev)) {
+		dev_priv->mipi_mmio_base = VLV_MIPI_BASE;
+	} else {
+		DRM_ERROR("Unsupported Mipi device to reg base");
+		return;
+	}
 
 	intel_dsi = kzalloc(sizeof(*intel_dsi), GFP_KERNEL);
 	if (!intel_dsi)
-		return false;
+		return;
 
 	intel_connector = kzalloc(sizeof(*intel_connector), GFP_KERNEL);
 	if (!intel_connector) {
 		kfree(intel_dsi);
-		return false;
+		return;
 	}
 
 	intel_encoder = &intel_dsi->base;
 	encoder = &intel_encoder->base;
 	intel_dsi->attached_connector = intel_connector;
 
-	if (IS_VALLEYVIEW(dev)) {
-		dev_priv->mipi_mmio_base = VLV_MIPI_BASE;
-	} else {
-		DRM_ERROR("Unsupported Mipi device to reg base");
-		return false;
-	}
-
 	connector = &intel_connector->base;
 
 	drm_encoder_init(dev, encoder, &intel_dsi_funcs, DRM_MODE_ENCODER_DSI);
@@ -743,7 +776,7 @@ bool intel_dsi_init(struct drm_device *dev)
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	fixed_mode = dsi->dev_ops->get_modes(&intel_dsi->dev);
 	if (!fixed_mode) {
@@ -754,12 +787,10 @@ bool intel_dsi_init(struct drm_device *dev)
 	fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
 	intel_panel_init(&intel_connector->panel, fixed_mode, NULL);
 
-	return true;
+	return;
 
 err:
 	drm_encoder_cleanup(&intel_encoder->base);
 	kfree(intel_dsi);
 	kfree(intel_connector);
-
-	return false;
 }
diff --git a/drivers/gpu/drm/i915/intel_dsi.h b/drivers/gpu/drm/i915/intel_dsi.h
index 31db33d3e5cc..657eb5c1b9d8 100644
--- a/drivers/gpu/drm/i915/intel_dsi.h
+++ b/drivers/gpu/drm/i915/intel_dsi.h
@@ -116,6 +116,8 @@ struct intel_dsi {
 	u16 clk_hs_to_lp_count;
 
 	u16 init_count;
+	u32 pclk;
+	u16 burst_mode_ratio;
 
 	/* all delays in ms */
 	u16 backlight_off_delay;
@@ -132,6 +134,7 @@ static inline struct intel_dsi *enc_to_intel_dsi(struct drm_encoder *encoder)
 
 extern void vlv_enable_dsi_pll(struct intel_encoder *encoder);
 extern void vlv_disable_dsi_pll(struct intel_encoder *encoder);
+extern u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp);
 
 extern struct intel_dsi_dev_ops vbt_generic_dsi_display_ops;
 
diff --git a/drivers/gpu/drm/i915/intel_dsi_cmd.c b/drivers/gpu/drm/i915/intel_dsi_cmd.c
index 933c86305237..f4767fd2ebeb 100644
--- a/drivers/gpu/drm/i915/intel_dsi_cmd.c
+++ b/drivers/gpu/drm/i915/intel_dsi_cmd.c
@@ -419,3 +419,19 @@ int dpi_send_cmd(struct intel_dsi *intel_dsi, u32 cmd, bool hs)
 
 	return 0;
 }
+
+void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi)
+{
+	struct drm_encoder *encoder = &intel_dsi->base.base;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->crtc);
+	enum pipe pipe = intel_crtc->pipe;
+	u32 mask;
+
+	mask = LP_CTRL_FIFO_EMPTY | HS_CTRL_FIFO_EMPTY |
+		LP_DATA_FIFO_EMPTY | HS_DATA_FIFO_EMPTY;
+
+	if (wait_for((I915_READ(MIPI_GEN_FIFO_STAT(pipe)) & mask) == mask, 100))
+		DRM_ERROR("DPI FIFOs are not empty\n");
+}
diff --git a/drivers/gpu/drm/i915/intel_dsi_cmd.h b/drivers/gpu/drm/i915/intel_dsi_cmd.h
index 9a18cbfa5460..46aa1acc00eb 100644
--- a/drivers/gpu/drm/i915/intel_dsi_cmd.h
+++ b/drivers/gpu/drm/i915/intel_dsi_cmd.h
@@ -51,6 +51,7 @@ int dsi_vc_generic_read(struct intel_dsi *intel_dsi, int channel,
 			u8 *reqdata, int reqlen, u8 *buf, int buflen);
 
 int dpi_send_cmd(struct intel_dsi *intel_dsi, u32 cmd, bool hs);
+void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi);
 
 /* XXX: questionable write helpers */
 static inline int dsi_vc_dcs_write_0(struct intel_dsi *intel_dsi,
diff --git a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
index 21a0d348cedc..f6bdd44069ce 100644
--- a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
+++ b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
@@ -143,7 +143,7 @@ static u8 *mipi_exec_send_packet(struct intel_dsi *intel_dsi, u8 *data)
 	case MIPI_DSI_DCS_LONG_WRITE:
 		dsi_vc_dcs_write(intel_dsi, vc, data, len);
 		break;
-	};
+	}
 
 	data += len;
 
@@ -271,6 +271,8 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	u32 ths_prepare_ns, tclk_trail_ns;
 	u32 tclk_prepare_clkzero, ths_prepare_hszero;
 	u32 lp_to_hs_switch, hs_to_lp_switch;
+	u32 pclk, computed_ddr;
+	u16 burst_mode_ratio;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -284,8 +286,6 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	else if (intel_dsi->pixel_format == VID_MODE_FORMAT_RGB565)
 		bits_per_pixel = 16;
 
-	bitrate = (mode->clock * bits_per_pixel) / intel_dsi->lane_count;
-
 	intel_dsi->operation_mode = mipi_config->is_cmd_mode;
 	intel_dsi->video_mode_format = mipi_config->video_transfer_mode;
 	intel_dsi->escape_clk_div = mipi_config->byte_clk_sel;
@@ -294,7 +294,42 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	intel_dsi->rst_timer_val = mipi_config->device_reset_timer;
 	intel_dsi->init_count = mipi_config->master_init_timer;
 	intel_dsi->bw_timer = mipi_config->dbi_bw_timer;
-	intel_dsi->video_frmt_cfg_bits = mipi_config->bta_enabled ? DISABLE_VIDEO_BTA : 0;
+	intel_dsi->video_frmt_cfg_bits =
+		mipi_config->bta_enabled ? DISABLE_VIDEO_BTA : 0;
+
+	pclk = mode->clock;
+
+	/* Burst Mode Ratio
+	 * Target ddr frequency from VBT / non burst ddr freq
+	 * multiply by 100 to preserve remainder
+	 */
+	if (intel_dsi->video_mode_format == VIDEO_MODE_BURST) {
+		if (mipi_config->target_burst_mode_freq) {
+			computed_ddr =
+				(pclk * bits_per_pixel) / intel_dsi->lane_count;
+
+			if (mipi_config->target_burst_mode_freq <
+								computed_ddr) {
+				DRM_ERROR("Burst mode freq is less than computed\n");
+				return false;
+			}
+
+			burst_mode_ratio = DIV_ROUND_UP(
+				mipi_config->target_burst_mode_freq * 100,
+				computed_ddr);
+
+			pclk = DIV_ROUND_UP(pclk * burst_mode_ratio, 100);
+		} else {
+			DRM_ERROR("Burst mode target is not set\n");
+			return false;
+		}
+	} else
+		burst_mode_ratio = 100;
+
+	intel_dsi->burst_mode_ratio = burst_mode_ratio;
+	intel_dsi->pclk = pclk;
+
+	bitrate = (pclk * bits_per_pixel) / intel_dsi->lane_count;
 
 	switch (intel_dsi->escape_clk_div) {
 	case 0:
@@ -351,7 +386,8 @@ static bool generic_init(struct intel_dsi_device *dsi)
 	 *
 	 * prepare count
 	 */
-	ths_prepare_ns = max(mipi_config->ths_prepare, mipi_config->tclk_prepare);
+	ths_prepare_ns = max(mipi_config->ths_prepare,
+			     mipi_config->tclk_prepare);
 	prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * ui_den, ui_num * 2);
 
 	/* exit zero count */
diff --git a/drivers/gpu/drm/i915/intel_dsi_pll.c b/drivers/gpu/drm/i915/intel_dsi_pll.c
index ba79ec19da3b..fa7a6ca34cd6 100644
--- a/drivers/gpu/drm/i915/intel_dsi_pll.c
+++ b/drivers/gpu/drm/i915/intel_dsi_pll.c
@@ -134,8 +134,7 @@ static u32 dsi_rr_formula(const struct drm_display_mode *mode,
 #else
 
 /* Get DSI clock from pixel clock */
-static u32 dsi_clk_from_pclk(const struct drm_display_mode *mode,
-			  int pixel_format, int lane_count)
+static u32 dsi_clk_from_pclk(u32 pclk, int pixel_format, int lane_count)
 {
 	u32 dsi_clk_khz;
 	u32 bpp;
@@ -156,7 +155,7 @@ static u32 dsi_clk_from_pclk(const struct drm_display_mode *mode,
 
 	/* DSI data rate = pixel clock * bits per pixel / lane count
 	   pixel clock is converted from KHz to Hz */
-	dsi_clk_khz = DIV_ROUND_CLOSEST(mode->clock * bpp, lane_count);
+	dsi_clk_khz = DIV_ROUND_CLOSEST(pclk * bpp, lane_count);
 
 	return dsi_clk_khz;
 }
@@ -191,7 +190,7 @@ static int dsi_calc_mnp(u32 dsi_clk, struct dsi_mnp *dsi_mnp)
 	for (m = 62; m <= 92; m++) {
 		for (p = 2; p <= 6; p++) {
 			/* Find the optimal m and p divisors
-			with minimal error +/- the required clock */
+			   with minimal error +/- the required clock */
 			calc_dsi_clk = (m * ref_clk) / p;
 			if (calc_dsi_clk == target_dsi_clk) {
 				calc_m = m;
@@ -228,15 +227,13 @@ static int dsi_calc_mnp(u32 dsi_clk, struct dsi_mnp *dsi_mnp)
 static void vlv_configure_dsi_pll(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
-	const struct drm_display_mode *mode = &intel_crtc->config.adjusted_mode;
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	int ret;
 	struct dsi_mnp dsi_mnp;
 	u32 dsi_clk;
 
-	dsi_clk = dsi_clk_from_pclk(mode, intel_dsi->pixel_format,
-						intel_dsi->lane_count);
+	dsi_clk = dsi_clk_from_pclk(intel_dsi->pclk, intel_dsi->pixel_format,
+				    intel_dsi->lane_count);
 
 	ret = dsi_calc_mnp(dsi_clk, &dsi_mnp);
 	if (ret) {
@@ -298,3 +295,84 @@ void vlv_disable_dsi_pll(struct intel_encoder *encoder)
 
 	mutex_unlock(&dev_priv->dpio_lock);
 }
+
+static void assert_bpp_mismatch(int pixel_format, int pipe_bpp)
+{
+	int bpp;
+
+	switch (pixel_format) {
+	default:
+	case VID_MODE_FORMAT_RGB888:
+	case VID_MODE_FORMAT_RGB666_LOOSE:
+		bpp = 24;
+		break;
+	case VID_MODE_FORMAT_RGB666:
+		bpp = 18;
+		break;
+	case VID_MODE_FORMAT_RGB565:
+		bpp = 16;
+		break;
+	}
+
+	WARN(bpp != pipe_bpp,
+	     "bpp match assertion failure (expected %d, current %d)\n",
+	     bpp, pipe_bpp);
+}
+
+u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	u32 dsi_clock, pclk;
+	u32 pll_ctl, pll_div;
+	u32 m = 0, p = 0;
+	int refclk = 25000;
+	int i;
+
+	DRM_DEBUG_KMS("\n");
+
+	mutex_lock(&dev_priv->dpio_lock);
+	pll_ctl = vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_CONTROL);
+	pll_div = vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_DIVIDER);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	/* mask out other bits and extract the P1 divisor */
+	pll_ctl &= DSI_PLL_P1_POST_DIV_MASK;
+	pll_ctl = pll_ctl >> (DSI_PLL_P1_POST_DIV_SHIFT - 2);
+
+	/* mask out the other bits and extract the M1 divisor */
+	pll_div &= DSI_PLL_M1_DIV_MASK;
+	pll_div = pll_div >> DSI_PLL_M1_DIV_SHIFT;
+
+	while (pll_ctl) {
+		pll_ctl = pll_ctl >> 1;
+		p++;
+	}
+	p--;
+
+	if (!p) {
+		DRM_ERROR("wrong P1 divisor\n");
+		return 0;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(lfsr_converts); i++) {
+		if (lfsr_converts[i] == pll_div)
+			break;
+	}
+
+	if (i == ARRAY_SIZE(lfsr_converts)) {
+		DRM_ERROR("wrong m_seed programmed\n");
+		return 0;
+	}
+
+	m = i + 62;
+
+	dsi_clock = (m * refclk) / p;
+
+	/* pixel_format and pipe_bpp should agree */
+	assert_bpp_mismatch(intel_dsi->pixel_format, pipe_bpp);
+
+	pclk = DIV_ROUND_CLOSEST(dsi_clock * intel_dsi->lane_count, pipe_bpp);
+
+	return pclk;
+}
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index a3631c0a5c28..e40e3df33517 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -85,7 +85,7 @@ static const struct intel_dvo_device intel_dvo_devices[] = {
 	{
 	        .type = INTEL_DVO_CHIP_TMDS,
 		.name = "ns2501",
-		.dvo_reg = DVOC,
+		.dvo_reg = DVOB,
 		.slave_addr = NS2501_ADDR,
 		.dev_ops = &ns2501_ops,
        }
@@ -112,7 +112,15 @@ static struct intel_dvo *intel_attached_dvo(struct drm_connector *connector)
 
 static bool intel_dvo_connector_get_hw_state(struct intel_connector *connector)
 {
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dvo *intel_dvo = intel_attached_dvo(&connector->base);
+	u32 tmp;
+
+	tmp = I915_READ(intel_dvo->dev.dvo_reg);
+
+	if (!(tmp & DVO_ENABLE))
+		return false;
 
 	return intel_dvo->dev.dev_ops->get_hw_state(&intel_dvo->dev);
 }
@@ -177,12 +185,13 @@ static void intel_enable_dvo(struct intel_encoder *encoder)
 	u32 dvo_reg = intel_dvo->dev.dvo_reg;
 	u32 temp = I915_READ(dvo_reg);
 
-	I915_WRITE(dvo_reg, temp | DVO_ENABLE);
-	I915_READ(dvo_reg);
 	intel_dvo->dev.dev_ops->mode_set(&intel_dvo->dev,
 					 &crtc->config.requested_mode,
 					 &crtc->config.adjusted_mode);
 
+	I915_WRITE(dvo_reg, temp | DVO_ENABLE);
+	I915_READ(dvo_reg);
+
 	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 }
 
@@ -218,10 +227,6 @@ static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 
 		intel_crtc_update_dpms(crtc);
 
-		intel_dvo->dev.dev_ops->mode_set(&intel_dvo->dev,
-						 &config->requested_mode,
-						 &config->adjusted_mode);
-
 		intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 	} else {
 		intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, false);
@@ -558,7 +563,7 @@ void intel_dvo_init(struct drm_device *dev)
 			intel_dvo->panel_wants_dither = true;
 		}
 
-		drm_sysfs_connector_add(connector);
+		drm_connector_register(connector);
 		return;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c
index 088fe9378a4c..9b584f3fbb99 100644
--- a/drivers/gpu/drm/i915/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/intel_fbdev.c
@@ -24,8 +24,10 @@
  *     David Airlie
  */
 
+#include <linux/async.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
+#include <linux/console.h>
 #include <linux/errno.h>
 #include <linux/string.h>
 #include <linux/mm.h>
@@ -43,10 +45,36 @@
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
 
+static int intel_fbdev_set_par(struct fb_info *info)
+{
+	struct drm_fb_helper *fb_helper = info->par;
+	struct intel_fbdev *ifbdev =
+		container_of(fb_helper, struct intel_fbdev, helper);
+	int ret;
+
+	ret = drm_fb_helper_set_par(info);
+
+	if (ret == 0) {
+		/*
+		 * FIXME: fbdev presumes that all callbacks also work from
+		 * atomic contexts and relies on that for emergency oops
+		 * printing. KMS totally doesn't do that and the locking here is
+		 * by far not the only place this goes wrong.  Ignore this for
+		 * now until we solve this for real.
+		 */
+		mutex_lock(&fb_helper->dev->struct_mutex);
+		ret = i915_gem_object_set_to_gtt_domain(ifbdev->fb->obj,
+							true);
+		mutex_unlock(&fb_helper->dev->struct_mutex);
+	}
+
+	return ret;
+}
+
 static struct fb_ops intelfb_ops = {
 	.owner = THIS_MODULE,
 	.fb_check_var = drm_fb_helper_check_var,
-	.fb_set_par = drm_fb_helper_set_par,
+	.fb_set_par = intel_fbdev_set_par,
 	.fb_fillrect = cfb_fillrect,
 	.fb_copyarea = cfb_copyarea,
 	.fb_imageblit = cfb_imageblit,
@@ -81,7 +109,7 @@ static int intelfb_alloc(struct drm_fb_helper *helper,
 							  sizes->surface_depth);
 
 	size = mode_cmd.pitches[0] * mode_cmd.height;
-	size = ALIGN(size, PAGE_SIZE);
+	size = PAGE_ALIGN(size);
 	obj = i915_gem_object_create_stolen(dev, size);
 	if (obj == NULL)
 		obj = i915_gem_alloc_object(dev, size);
@@ -305,24 +333,6 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 	int num_connectors_enabled = 0;
 	int num_connectors_detected = 0;
 
-	/*
-	 * If the user specified any force options, just bail here
-	 * and use that config.
-	 */
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		struct drm_fb_helper_connector *fb_conn;
-		struct drm_connector *connector;
-
-		fb_conn = fb_helper->connector_info[i];
-		connector = fb_conn->connector;
-
-		if (!enabled[i])
-			continue;
-
-		if (connector->force != DRM_FORCE_UNSPECIFIED)
-			return false;
-	}
-
 	save_enabled = kcalloc(dev->mode_config.num_connector, sizeof(bool),
 			       GFP_KERNEL);
 	if (!save_enabled)
@@ -348,8 +358,18 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 			continue;
 		}
 
+		if (connector->force == DRM_FORCE_OFF) {
+			DRM_DEBUG_KMS("connector %s is disabled by user, skipping\n",
+				      connector->name);
+			enabled[i] = false;
+			continue;
+		}
+
 		encoder = connector->encoder;
 		if (!encoder || WARN_ON(!encoder->crtc)) {
+			if (connector->force > DRM_FORCE_OFF)
+				goto bail;
+
 			DRM_DEBUG_KMS("connector %s has no encoder or crtc, skipping\n",
 				      connector->name);
 			enabled[i] = false;
@@ -368,8 +388,7 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		for (j = 0; j < fb_helper->connector_count; j++) {
 			if (crtcs[j] == new_crtc) {
 				DRM_DEBUG_KMS("fallback: cloned configuration\n");
-				fallback = true;
-				goto out;
+				goto bail;
 			}
 		}
 
@@ -417,7 +436,7 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		}
 		crtcs[i] = new_crtc;
 
-		DRM_DEBUG_KMS("connector %s on pipe %d [CRTC:%d]: %dx%d%s\n",
+		DRM_DEBUG_KMS("connector %s on pipe %c [CRTC:%d]: %dx%d%s\n",
 			      connector->name,
 			      pipe_name(to_intel_crtc(encoder->crtc)->pipe),
 			      encoder->crtc->base.id,
@@ -440,8 +459,8 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 		fallback = true;
 	}
 
-out:
 	if (fallback) {
+bail:
 		DRM_DEBUG_KMS("Not using firmware configuration\n");
 		memcpy(enabled, save_enabled, dev->mode_config.num_connector);
 		kfree(save_enabled);
@@ -452,7 +471,7 @@ out:
 	return true;
 }
 
-static struct drm_fb_helper_funcs intel_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs intel_fb_helper_funcs = {
 	.initial_config = intel_fb_initial_config,
 	.gamma_set = intel_crtc_fb_gamma_set,
 	.gamma_get = intel_crtc_fb_gamma_get,
@@ -610,6 +629,15 @@ out:
 	return false;
 }
 
+static void intel_fbdev_suspend_worker(struct work_struct *work)
+{
+	intel_fbdev_set_suspend(container_of(work,
+					     struct drm_i915_private,
+					     fbdev_suspend_work)->dev,
+				FBINFO_STATE_RUNNING,
+				true);
+}
+
 int intel_fbdev_init(struct drm_device *dev)
 {
 	struct intel_fbdev *ifbdev;
@@ -623,7 +651,8 @@ int intel_fbdev_init(struct drm_device *dev)
 	if (ifbdev == NULL)
 		return -ENOMEM;
 
-	ifbdev->helper.funcs = &intel_fb_helper_funcs;
+	drm_fb_helper_prepare(dev, &ifbdev->helper, &intel_fb_helper_funcs);
+
 	if (!intel_fbdev_init_bios(dev, ifbdev))
 		ifbdev->preferred_bpp = 32;
 
@@ -635,14 +664,16 @@ int intel_fbdev_init(struct drm_device *dev)
 	}
 
 	dev_priv->fbdev = ifbdev;
+	INIT_WORK(&dev_priv->fbdev_suspend_work, intel_fbdev_suspend_worker);
+
 	drm_fb_helper_single_add_all_connectors(&ifbdev->helper);
 
 	return 0;
 }
 
-void intel_fbdev_initial_config(struct drm_device *dev)
+void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = data;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
@@ -655,12 +686,15 @@ void intel_fbdev_fini(struct drm_device *dev)
 	if (!dev_priv->fbdev)
 		return;
 
+	flush_work(&dev_priv->fbdev_suspend_work);
+
+	async_synchronize_full();
 	intel_fbdev_destroy(dev, dev_priv->fbdev);
 	kfree(dev_priv->fbdev);
 	dev_priv->fbdev = NULL;
 }
 
-void intel_fbdev_set_suspend(struct drm_device *dev, int state)
+void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
@@ -671,6 +705,33 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 
 	info = ifbdev->helper.fbdev;
 
+	if (synchronous) {
+		/* Flush any pending work to turn the console on, and then
+		 * wait to turn it off. It must be synchronous as we are
+		 * about to suspend or unload the driver.
+		 *
+		 * Note that from within the work-handler, we cannot flush
+		 * ourselves, so only flush outstanding work upon suspend!
+		 */
+		if (state != FBINFO_STATE_RUNNING)
+			flush_work(&dev_priv->fbdev_suspend_work);
+		console_lock();
+	} else {
+		/*
+		 * The console lock can be pretty contented on resume due
+		 * to all the printk activity.  Try to keep it out of the hot
+		 * path of resume if possible.
+		 */
+		WARN_ON(state != FBINFO_STATE_RUNNING);
+		if (!console_trylock()) {
+			/* Don't block our own workqueue as this can
+			 * be run in parallel with other i915.ko tasks.
+			 */
+			schedule_work(&dev_priv->fbdev_suspend_work);
+			return;
+		}
+	}
+
 	/* On resume from hibernation: If the object is shmemfs backed, it has
 	 * been restored from swap. If the object is stolen however, it will be
 	 * full of whatever garbage was left in there.
@@ -679,6 +740,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	fb_set_suspend(info, state);
+	console_unlock();
 }
 
 void intel_fbdev_output_poll_changed(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index eee2bbec2958..29ec1535992d 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -367,6 +367,9 @@ static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder,
 	union hdmi_infoframe frame;
 	int ret;
 
+	/* Set user selected PAR to incoming mode's member */
+	adjusted_mode->picture_aspect_ratio = intel_hdmi->aspect_ratio;
+
 	ret = drm_hdmi_avi_infoframe_from_display_mode(&frame.avi,
 						       adjusted_mode);
 	if (ret < 0) {
@@ -709,7 +712,8 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder,
 				  struct intel_crtc_config *pipe_config)
 {
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
-	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 tmp, flags = 0;
 	int dotclock;
 
@@ -728,9 +732,13 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder,
 	if (tmp & HDMI_MODE_SELECT_HDMI)
 		pipe_config->has_hdmi_sink = true;
 
-	if (tmp & HDMI_MODE_SELECT_HDMI)
+	if (tmp & SDVO_AUDIO_ENABLE)
 		pipe_config->has_audio = true;
 
+	if (!HAS_PCH_SPLIT(dev) &&
+	    tmp & HDMI_COLOR_RANGE_16_235)
+		pipe_config->limited_color_range = true;
+
 	pipe_config->adjusted_mode.flags |= flags;
 
 	if ((tmp & SDVO_COLOR_FORMAT_MASK) == HDMI_COLOR_FORMAT_12bpc)
@@ -861,10 +869,15 @@ static enum drm_mode_status
 intel_hdmi_mode_valid(struct drm_connector *connector,
 		      struct drm_display_mode *mode)
 {
-	if (mode->clock > hdmi_portclock_limit(intel_attached_hdmi(connector),
-					       true))
+	int clock = mode->clock;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK)
+		clock *= 2;
+
+	if (clock > hdmi_portclock_limit(intel_attached_hdmi(connector),
+					 true))
 		return MODE_CLOCK_HIGH;
-	if (mode->clock < 20000)
+	if (clock < 20000)
 		return MODE_CLOCK_LOW;
 
 	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
@@ -879,10 +892,10 @@ static bool hdmi_12bpc_possible(struct intel_crtc *crtc)
 	struct intel_encoder *encoder;
 	int count = 0, count_hdmi = 0;
 
-	if (!HAS_PCH_SPLIT(dev))
+	if (HAS_GMCH_DISPLAY(dev))
 		return false;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->new_crtc != crtc)
 			continue;
 
@@ -918,6 +931,10 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 			intel_hdmi->color_range = 0;
 	}
 
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK) {
+		pipe_config->pixel_multiplier = 2;
+	}
+
 	if (intel_hdmi->color_range)
 		pipe_config->limited_color_range = true;
 
@@ -959,104 +976,117 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	return true;
 }
 
-static enum drm_connector_status
-intel_hdmi_detect(struct drm_connector *connector, bool force)
+static void
+intel_hdmi_unset_edid(struct drm_connector *connector)
 {
-	struct drm_device *dev = connector->dev;
 	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
-	struct intel_digital_port *intel_dig_port =
-		hdmi_to_dig_port(intel_hdmi);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct edid *edid;
-	enum intel_display_power_domain power_domain;
-	enum drm_connector_status status = connector_status_disconnected;
 
-	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
-		      connector->base.id, connector->name);
+	intel_hdmi->has_hdmi_sink = false;
+	intel_hdmi->has_audio = false;
+	intel_hdmi->rgb_quant_range_selectable = false;
+
+	kfree(to_intel_connector(connector)->detect_edid);
+	to_intel_connector(connector)->detect_edid = NULL;
+}
+
+static bool
+intel_hdmi_set_edid(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+	struct intel_encoder *intel_encoder =
+		&hdmi_to_dig_port(intel_hdmi)->base;
+	enum intel_display_power_domain power_domain;
+	struct edid *edid;
+	bool connected = false;
 
 	power_domain = intel_display_port_power_domain(intel_encoder);
 	intel_display_power_get(dev_priv, power_domain);
 
-	intel_hdmi->has_hdmi_sink = false;
-	intel_hdmi->has_audio = false;
-	intel_hdmi->rgb_quant_range_selectable = false;
 	edid = drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
 						    intel_hdmi->ddc_bus));
 
-	if (edid) {
-		if (edid->input & DRM_EDID_INPUT_DIGITAL) {
-			status = connector_status_connected;
-			if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
-				intel_hdmi->has_hdmi_sink =
-						drm_detect_hdmi_monitor(edid);
-			intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
-			intel_hdmi->rgb_quant_range_selectable =
-				drm_rgb_quant_range_selectable(edid);
-		}
-		kfree(edid);
-	}
+	intel_display_power_put(dev_priv, power_domain);
+
+	to_intel_connector(connector)->detect_edid = edid;
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL) {
+		intel_hdmi->rgb_quant_range_selectable =
+			drm_rgb_quant_range_selectable(edid);
 
-	if (status == connector_status_connected) {
+		intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
 		if (intel_hdmi->force_audio != HDMI_AUDIO_AUTO)
 			intel_hdmi->has_audio =
-				(intel_hdmi->force_audio == HDMI_AUDIO_ON);
-		intel_encoder->type = INTEL_OUTPUT_HDMI;
+				intel_hdmi->force_audio == HDMI_AUDIO_ON;
+
+		if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
+			intel_hdmi->has_hdmi_sink =
+				drm_detect_hdmi_monitor(edid);
+
+		connected = true;
 	}
 
-	intel_display_power_put(dev_priv, power_domain);
+	return connected;
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	intel_hdmi_unset_edid(connector);
+
+	if (intel_hdmi_set_edid(connector)) {
+		struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
+
+		hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+		status = connector_status_connected;
+	} else
+		status = connector_status_disconnected;
 
 	return status;
 }
 
-static int intel_hdmi_get_modes(struct drm_connector *connector)
+static void
+intel_hdmi_force(struct drm_connector *connector)
 {
-	struct intel_encoder *intel_encoder = intel_attached_encoder(connector);
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&intel_encoder->base);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	int ret;
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector);
 
-	/* We should parse the EDID data and find out if it's an HDMI sink so
-	 * we can send audio to it.
-	 */
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_hdmi_unset_edid(connector);
 
-	ret = intel_ddc_get_modes(connector,
-				   intel_gmbus_get_adapter(dev_priv,
-							   intel_hdmi->ddc_bus));
+	if (connector->status != connector_status_connected)
+		return;
 
-	intel_display_power_put(dev_priv, power_domain);
+	intel_hdmi_set_edid(connector);
+	hdmi_to_dig_port(intel_hdmi)->base.type = INTEL_OUTPUT_HDMI;
+}
 
-	return ret;
+static int intel_hdmi_get_modes(struct drm_connector *connector)
+{
+	struct edid *edid;
+
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid == NULL)
+		return 0;
+
+	return intel_connector_update_modes(connector, edid);
 }
 
 static bool
 intel_hdmi_detect_audio(struct drm_connector *connector)
 {
-	struct intel_encoder *intel_encoder = intel_attached_encoder(connector);
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&intel_encoder->base);
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	enum intel_display_power_domain power_domain;
-	struct edid *edid;
 	bool has_audio = false;
+	struct edid *edid;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
-
-	edid = drm_get_edid(connector,
-			    intel_gmbus_get_adapter(dev_priv,
-						    intel_hdmi->ddc_bus));
-	if (edid) {
-		if (edid->input & DRM_EDID_INPUT_DIGITAL)
-			has_audio = drm_detect_monitor_audio(edid);
-		kfree(edid);
-	}
-
-	intel_display_power_put(dev_priv, power_domain);
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL)
+		has_audio = drm_detect_monitor_audio(edid);
 
 	return has_audio;
 }
@@ -1124,6 +1154,23 @@ intel_hdmi_set_property(struct drm_connector *connector,
 		goto done;
 	}
 
+	if (property == connector->dev->mode_config.aspect_ratio_property) {
+		switch (val) {
+		case DRM_MODE_PICTURE_ASPECT_NONE:
+			intel_hdmi->aspect_ratio = HDMI_PICTURE_ASPECT_NONE;
+			break;
+		case DRM_MODE_PICTURE_ASPECT_4_3:
+			intel_hdmi->aspect_ratio = HDMI_PICTURE_ASPECT_4_3;
+			break;
+		case DRM_MODE_PICTURE_ASPECT_16_9:
+			intel_hdmi->aspect_ratio = HDMI_PICTURE_ASPECT_16_9;
+			break;
+		default:
+			return -EINVAL;
+		}
+		goto done;
+	}
+
 	return -EINVAL;
 
 done:
@@ -1229,6 +1276,72 @@ static void vlv_hdmi_pre_pll_enable(struct intel_encoder *encoder)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static void chv_hdmi_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc =
+		to_intel_crtc(encoder->base.crtc);
+	enum dpio_channel ch = vlv_dport_to_channel(dport);
+	enum pipe pipe = intel_crtc->pipe;
+	u32 val;
+
+	intel_hdmi_prepare(encoder);
+
+	mutex_lock(&dev_priv->dpio_lock);
+
+	/* program left/right clock distribution */
+	if (pipe != PIPE_B) {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW5_CH0);
+		val &= ~(CHV_BUFLEFTENA1_MASK | CHV_BUFRIGHTENA1_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA1_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA1_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW5_CH0, val);
+	} else {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW1_CH1);
+		val &= ~(CHV_BUFLEFTENA2_MASK | CHV_BUFRIGHTENA2_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA2_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA2_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW1_CH1, val);
+	}
+
+	/* program clock channel usage */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW8(ch));
+	val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+	if (pipe != PIPE_B)
+		val &= ~CHV_PCS_USEDCLKCHANNEL;
+	else
+		val |= CHV_PCS_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW8(ch), val);
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW8(ch));
+	val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+	if (pipe != PIPE_B)
+		val &= ~CHV_PCS_USEDCLKCHANNEL;
+	else
+		val |= CHV_PCS_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW8(ch), val);
+
+	/*
+	 * This a a bit weird since generally CL
+	 * matches the pipe, but here we need to
+	 * pick the CL based on the port.
+	 */
+	val = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW19(ch));
+	if (pipe != PIPE_B)
+		val &= ~CHV_CMN_USEDCLKCHANNEL;
+	else
+		val |= CHV_CMN_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW19(ch), val);
+
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 static void vlv_hdmi_post_disable(struct intel_encoder *encoder)
 {
 	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
@@ -1345,8 +1458,8 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 	for (i = 0; i < 4; i++) {
 		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW2(ch, i));
-		val &= ~DPIO_SWING_MARGIN_MASK;
-		val |= 102 << DPIO_SWING_MARGIN_SHIFT;
+		val &= ~DPIO_SWING_MARGIN000_MASK;
+		val |= 102 << DPIO_SWING_MARGIN000_SHIFT;
 		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW2(ch, i), val);
 	}
 
@@ -1393,6 +1506,7 @@ static void chv_hdmi_pre_enable(struct intel_encoder *encoder)
 
 static void intel_hdmi_destroy(struct drm_connector *connector)
 {
+	kfree(to_intel_connector(connector)->detect_edid);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -1400,6 +1514,7 @@ static void intel_hdmi_destroy(struct drm_connector *connector)
 static const struct drm_connector_funcs intel_hdmi_connector_funcs = {
 	.dpms = intel_connector_dpms,
 	.detect = intel_hdmi_detect,
+	.force = intel_hdmi_force,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_hdmi_set_property,
 	.destroy = intel_hdmi_destroy,
@@ -1416,11 +1531,22 @@ static const struct drm_encoder_funcs intel_hdmi_enc_funcs = {
 };
 
 static void
+intel_attach_aspect_ratio_property(struct drm_connector *connector)
+{
+	if (!drm_mode_create_aspect_ratio_property(connector->dev))
+		drm_object_attach_property(&connector->base,
+			connector->dev->mode_config.aspect_ratio_property,
+			DRM_MODE_PICTURE_ASPECT_NONE);
+}
+
+static void
 intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *connector)
 {
 	intel_attach_force_audio_property(connector);
 	intel_attach_broadcast_rgb_property(connector);
 	intel_hdmi->color_range_auto = true;
+	intel_attach_aspect_ratio_property(connector);
+	intel_hdmi->aspect_ratio = HDMI_PICTURE_ASPECT_NONE;
 }
 
 void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
@@ -1467,7 +1593,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	if (IS_VALLEYVIEW(dev)) {
 		intel_hdmi->write_infoframe = vlv_write_infoframe;
 		intel_hdmi->set_infoframes = vlv_set_infoframes;
-	} else if (!HAS_PCH_SPLIT(dev)) {
+	} else if (IS_G4X(dev)) {
 		intel_hdmi->write_infoframe = g4x_write_infoframe;
 		intel_hdmi->set_infoframes = g4x_set_infoframes;
 	} else if (HAS_DDI(dev)) {
@@ -1490,7 +1616,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 	intel_hdmi_add_properties(intel_hdmi, connector);
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
 	 * 0xd.  Failure to do so will result in spurious interrupts being
@@ -1528,6 +1654,7 @@ void intel_hdmi_init(struct drm_device *dev, int hdmi_reg, enum port port)
 	intel_encoder->get_hw_state = intel_hdmi_get_hw_state;
 	intel_encoder->get_config = intel_hdmi_get_config;
 	if (IS_CHERRYVIEW(dev)) {
+		intel_encoder->pre_pll_enable = chv_hdmi_pre_pll_enable;
 		intel_encoder->pre_enable = chv_hdmi_pre_enable;
 		intel_encoder->enable = vlv_enable_hdmi;
 		intel_encoder->post_disable = chv_hdmi_post_disable;
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index d33b61d0dd33..b31088a551f2 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -34,11 +34,6 @@
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
 
-enum disp_clk {
-	CDCLK,
-	CZCLK
-};
-
 struct gmbus_port {
 	const char *name;
 	int reg;
@@ -63,60 +58,11 @@ to_intel_gmbus(struct i2c_adapter *i2c)
 	return container_of(i2c, struct intel_gmbus, adapter);
 }
 
-static int get_disp_clk_div(struct drm_i915_private *dev_priv,
-			    enum disp_clk clk)
-{
-	u32 reg_val;
-	int clk_ratio;
-
-	reg_val = I915_READ(CZCLK_CDCLK_FREQ_RATIO);
-
-	if (clk == CDCLK)
-		clk_ratio =
-			((reg_val & CDCLK_FREQ_MASK) >> CDCLK_FREQ_SHIFT) + 1;
-	else
-		clk_ratio = (reg_val & CZCLK_FREQ_MASK) + 1;
-
-	return clk_ratio;
-}
-
-static void gmbus_set_freq(struct drm_i915_private *dev_priv)
-{
-	int vco, gmbus_freq = 0, cdclk_div;
-
-	BUG_ON(!IS_VALLEYVIEW(dev_priv->dev));
-
-	vco = valleyview_get_vco(dev_priv);
-
-	/* Get the CDCLK divide ratio */
-	cdclk_div = get_disp_clk_div(dev_priv, CDCLK);
-
-	/*
-	 * Program the gmbus_freq based on the cdclk frequency.
-	 * BSpec erroneously claims we should aim for 4MHz, but
-	 * in fact 1MHz is the correct frequency.
-	 */
-	if (cdclk_div)
-		gmbus_freq = (vco << 1) / cdclk_div;
-
-	if (WARN_ON(gmbus_freq == 0))
-		return;
-
-	I915_WRITE(GMBUSFREQ_VLV, gmbus_freq);
-}
-
 void
 intel_i2c_reset(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/*
-	 * In BIOS-less system, program the correct gmbus frequency
-	 * before reading edid.
-	 */
-	if (IS_VALLEYVIEW(dev))
-		gmbus_set_freq(dev_priv);
-
 	I915_WRITE(dev_priv->gpio_mmio_base + GMBUS0, 0);
 	I915_WRITE(dev_priv->gpio_mmio_base + GMBUS4, 0);
 }
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
new file mode 100644
index 000000000000..bafd38b5703e
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -0,0 +1,1766 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Ben Widawsky <ben@bwidawsk.net>
+ *    Michel Thierry <michel.thierry@intel.com>
+ *    Thomas Daniel <thomas.daniel@intel.com>
+ *    Oscar Mateo <oscar.mateo@intel.com>
+ *
+ */
+
+/**
+ * DOC: Logical Rings, Logical Ring Contexts and Execlists
+ *
+ * Motivation:
+ * GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
+ * These expanded contexts enable a number of new abilities, especially
+ * "Execlists" (also implemented in this file).
+ *
+ * One of the main differences with the legacy HW contexts is that logical
+ * ring contexts incorporate many more things to the context's state, like
+ * PDPs or ringbuffer control registers:
+ *
+ * The reason why PDPs are included in the context is straightforward: as
+ * PPGTTs (per-process GTTs) are actually per-context, having the PDPs
+ * contained there mean you don't need to do a ppgtt->switch_mm yourself,
+ * instead, the GPU will do it for you on the context switch.
+ *
+ * But, what about the ringbuffer control registers (head, tail, etc..)?
+ * shouldn't we just need a set of those per engine command streamer? This is
+ * where the name "Logical Rings" starts to make sense: by virtualizing the
+ * rings, the engine cs shifts to a new "ring buffer" with every context
+ * switch. When you want to submit a workload to the GPU you: A) choose your
+ * context, B) find its appropriate virtualized ring, C) write commands to it
+ * and then, finally, D) tell the GPU to switch to that context.
+ *
+ * Instead of the legacy MI_SET_CONTEXT, the way you tell the GPU to switch
+ * to a contexts is via a context execution list, ergo "Execlists".
+ *
+ * LRC implementation:
+ * Regarding the creation of contexts, we have:
+ *
+ * - One global default context.
+ * - One local default context for each opened fd.
+ * - One local extra context for each context create ioctl call.
+ *
+ * Now that ringbuffers belong per-context (and not per-engine, like before)
+ * and that contexts are uniquely tied to a given engine (and not reusable,
+ * like before) we need:
+ *
+ * - One ringbuffer per-engine inside each context.
+ * - One backing object per-engine inside each context.
+ *
+ * The global default context starts its life with these new objects fully
+ * allocated and populated. The local default context for each opened fd is
+ * more complex, because we don't know at creation time which engine is going
+ * to use them. To handle this, we have implemented a deferred creation of LR
+ * contexts:
+ *
+ * The local context starts its life as a hollow or blank holder, that only
+ * gets populated for a given engine once we receive an execbuffer. If later
+ * on we receive another execbuffer ioctl for the same context but a different
+ * engine, we allocate/populate a new ringbuffer and context backing object and
+ * so on.
+ *
+ * Finally, regarding local contexts created using the ioctl call: as they are
+ * only allowed with the render ring, we can allocate & populate them right
+ * away (no need to defer anything, at least for now).
+ *
+ * Execlists implementation:
+ * Execlists are the new method by which, on gen8+ hardware, workloads are
+ * submitted for execution (as opposed to the legacy, ringbuffer-based, method).
+ * This method works as follows:
+ *
+ * When a request is committed, its commands (the BB start and any leading or
+ * trailing commands, like the seqno breadcrumbs) are placed in the ringbuffer
+ * for the appropriate context. The tail pointer in the hardware context is not
+ * updated at this time, but instead, kept by the driver in the ringbuffer
+ * structure. A structure representing this request is added to a request queue
+ * for the appropriate engine: this structure contains a copy of the context's
+ * tail after the request was written to the ring buffer and a pointer to the
+ * context itself.
+ *
+ * If the engine's request queue was empty before the request was added, the
+ * queue is processed immediately. Otherwise the queue will be processed during
+ * a context switch interrupt. In any case, elements on the queue will get sent
+ * (in pairs) to the GPU's ExecLists Submit Port (ELSP, for short) with a
+ * globally unique 20-bits submission ID.
+ *
+ * When execution of a request completes, the GPU updates the context status
+ * buffer with a context complete event and generates a context switch interrupt.
+ * During the interrupt handling, the driver examines the events in the buffer:
+ * for each context complete event, if the announced ID matches that on the head
+ * of the request queue, then that request is retired and removed from the queue.
+ *
+ * After processing, if any requests were retired and the queue is not empty
+ * then a new execution list can be submitted. The two requests at the front of
+ * the queue are next to be submitted but since a context may not occur twice in
+ * an execution list, if subsequent requests have the same ID as the first then
+ * the two requests must be combined. This is done simply by discarding requests
+ * at the head of the queue until either only one requests is left (in which case
+ * we use a NULL second context) or the first two requests have unique IDs.
+ *
+ * By always executing the first two requests in the queue the driver ensures
+ * that the GPU is kept as busy as possible. In the case where a single context
+ * completes but a second context is still executing, the request for this second
+ * context will be at the head of the queue when we remove the first one. This
+ * request will then be resubmitted along with a new request for a different context,
+ * which will cause the hardware to continue executing the second request and queue
+ * the new request (the GPU detects the condition of a context getting preempted
+ * with the same context and optimizes the context switch flow by not doing
+ * preemption, but just sampling the new tail pointer).
+ *
+ */
+
+#include <drm/drmP.h>
+#include <drm/i915_drm.h>
+#include "i915_drv.h"
+
+#define GEN8_LR_CONTEXT_RENDER_SIZE (20 * PAGE_SIZE)
+#define GEN8_LR_CONTEXT_OTHER_SIZE (2 * PAGE_SIZE)
+
+#define GEN8_LR_CONTEXT_ALIGN 4096
+
+#define RING_EXECLIST_QFULL		(1 << 0x2)
+#define RING_EXECLIST1_VALID		(1 << 0x3)
+#define RING_EXECLIST0_VALID		(1 << 0x4)
+#define RING_EXECLIST_ACTIVE_STATUS	(3 << 0xE)
+#define RING_EXECLIST1_ACTIVE		(1 << 0x11)
+#define RING_EXECLIST0_ACTIVE		(1 << 0x12)
+
+#define GEN8_CTX_STATUS_IDLE_ACTIVE	(1 << 0)
+#define GEN8_CTX_STATUS_PREEMPTED	(1 << 1)
+#define GEN8_CTX_STATUS_ELEMENT_SWITCH	(1 << 2)
+#define GEN8_CTX_STATUS_ACTIVE_IDLE	(1 << 3)
+#define GEN8_CTX_STATUS_COMPLETE	(1 << 4)
+#define GEN8_CTX_STATUS_LITE_RESTORE	(1 << 15)
+
+#define CTX_LRI_HEADER_0		0x01
+#define CTX_CONTEXT_CONTROL		0x02
+#define CTX_RING_HEAD			0x04
+#define CTX_RING_TAIL			0x06
+#define CTX_RING_BUFFER_START		0x08
+#define CTX_RING_BUFFER_CONTROL		0x0a
+#define CTX_BB_HEAD_U			0x0c
+#define CTX_BB_HEAD_L			0x0e
+#define CTX_BB_STATE			0x10
+#define CTX_SECOND_BB_HEAD_U		0x12
+#define CTX_SECOND_BB_HEAD_L		0x14
+#define CTX_SECOND_BB_STATE		0x16
+#define CTX_BB_PER_CTX_PTR		0x18
+#define CTX_RCS_INDIRECT_CTX		0x1a
+#define CTX_RCS_INDIRECT_CTX_OFFSET	0x1c
+#define CTX_LRI_HEADER_1		0x21
+#define CTX_CTX_TIMESTAMP		0x22
+#define CTX_PDP3_UDW			0x24
+#define CTX_PDP3_LDW			0x26
+#define CTX_PDP2_UDW			0x28
+#define CTX_PDP2_LDW			0x2a
+#define CTX_PDP1_UDW			0x2c
+#define CTX_PDP1_LDW			0x2e
+#define CTX_PDP0_UDW			0x30
+#define CTX_PDP0_LDW			0x32
+#define CTX_LRI_HEADER_2		0x41
+#define CTX_R_PWR_CLK_STATE		0x42
+#define CTX_GPGPU_CSR_BASE_ADDRESS	0x44
+
+#define GEN8_CTX_VALID (1<<0)
+#define GEN8_CTX_FORCE_PD_RESTORE (1<<1)
+#define GEN8_CTX_FORCE_RESTORE (1<<2)
+#define GEN8_CTX_L3LLC_COHERENT (1<<5)
+#define GEN8_CTX_PRIVILEGE (1<<8)
+enum {
+	ADVANCED_CONTEXT = 0,
+	LEGACY_CONTEXT,
+	ADVANCED_AD_CONTEXT,
+	LEGACY_64B_CONTEXT
+};
+#define GEN8_CTX_MODE_SHIFT 3
+enum {
+	FAULT_AND_HANG = 0,
+	FAULT_AND_HALT, /* Debug only */
+	FAULT_AND_STREAM,
+	FAULT_AND_CONTINUE /* Unsupported */
+};
+#define GEN8_CTX_ID_SHIFT 32
+
+/**
+ * intel_sanitize_enable_execlists() - sanitize i915.enable_execlists
+ * @dev: DRM device.
+ * @enable_execlists: value of i915.enable_execlists module parameter.
+ *
+ * Only certain platforms support Execlists (the prerequisites being
+ * support for Logical Ring Contexts and Aliasing PPGTT or better),
+ * and only when enabled via module parameter.
+ *
+ * Return: 1 if Execlists is supported and has to be enabled.
+ */
+int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists)
+{
+	WARN_ON(i915.enable_ppgtt == -1);
+
+	if (enable_execlists == 0)
+		return 0;
+
+	if (HAS_LOGICAL_RING_CONTEXTS(dev) && USES_PPGTT(dev) &&
+	    i915.use_mmio_flip >= 0)
+		return 1;
+
+	return 0;
+}
+
+/**
+ * intel_execlists_ctx_id() - get the Execlists Context ID
+ * @ctx_obj: Logical Ring Context backing object.
+ *
+ * Do not confuse with ctx->id! Unfortunately we have a name overload
+ * here: the old context ID we pass to userspace as a handler so that
+ * they can refer to a context, and the new context ID we pass to the
+ * ELSP so that the GPU can inform us of the context status via
+ * interrupts.
+ *
+ * Return: 20-bits globally unique context ID.
+ */
+u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj)
+{
+	u32 lrca = i915_gem_obj_ggtt_offset(ctx_obj);
+
+	/* LRCA is required to be 4K aligned so the more significant 20 bits
+	 * are globally unique */
+	return lrca >> 12;
+}
+
+static uint64_t execlists_ctx_descriptor(struct drm_i915_gem_object *ctx_obj)
+{
+	uint64_t desc;
+	uint64_t lrca = i915_gem_obj_ggtt_offset(ctx_obj);
+
+	WARN_ON(lrca & 0xFFFFFFFF00000FFFULL);
+
+	desc = GEN8_CTX_VALID;
+	desc |= LEGACY_CONTEXT << GEN8_CTX_MODE_SHIFT;
+	desc |= GEN8_CTX_L3LLC_COHERENT;
+	desc |= GEN8_CTX_PRIVILEGE;
+	desc |= lrca;
+	desc |= (u64)intel_execlists_ctx_id(ctx_obj) << GEN8_CTX_ID_SHIFT;
+
+	/* TODO: WaDisableLiteRestore when we start using semaphore
+	 * signalling between Command Streamers */
+	/* desc |= GEN8_CTX_FORCE_RESTORE; */
+
+	return desc;
+}
+
+static void execlists_elsp_write(struct intel_engine_cs *ring,
+				 struct drm_i915_gem_object *ctx_obj0,
+				 struct drm_i915_gem_object *ctx_obj1)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	uint64_t temp = 0;
+	uint32_t desc[4];
+	unsigned long flags;
+
+	/* XXX: You must always write both descriptors in the order below. */
+	if (ctx_obj1)
+		temp = execlists_ctx_descriptor(ctx_obj1);
+	else
+		temp = 0;
+	desc[1] = (u32)(temp >> 32);
+	desc[0] = (u32)temp;
+
+	temp = execlists_ctx_descriptor(ctx_obj0);
+	desc[3] = (u32)(temp >> 32);
+	desc[2] = (u32)temp;
+
+	/* Set Force Wakeup bit to prevent GT from entering C6 while ELSP writes
+	 * are in progress.
+	 *
+	 * The other problem is that we can't just call gen6_gt_force_wake_get()
+	 * because that function calls intel_runtime_pm_get(), which might sleep.
+	 * Instead, we do the runtime_pm_get/put when creating/destroying requests.
+	 */
+	spin_lock_irqsave(&dev_priv->uncore.lock, flags);
+	if (IS_CHERRYVIEW(dev_priv->dev)) {
+		if (dev_priv->uncore.fw_rendercount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_RENDER);
+		if (dev_priv->uncore.fw_mediacount++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_MEDIA);
+	} else {
+		if (dev_priv->uncore.forcewake_count++ == 0)
+			dev_priv->uncore.funcs.force_wake_get(dev_priv,
+							      FORCEWAKE_ALL);
+	}
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, flags);
+
+	I915_WRITE(RING_ELSP(ring), desc[1]);
+	I915_WRITE(RING_ELSP(ring), desc[0]);
+	I915_WRITE(RING_ELSP(ring), desc[3]);
+	/* The context is automatically loaded after the following */
+	I915_WRITE(RING_ELSP(ring), desc[2]);
+
+	/* ELSP is a wo register, so use another nearby reg for posting instead */
+	POSTING_READ(RING_EXECLIST_STATUS(ring));
+
+	/* Release Force Wakeup (see the big comment above). */
+	spin_lock_irqsave(&dev_priv->uncore.lock, flags);
+	if (IS_CHERRYVIEW(dev_priv->dev)) {
+		if (--dev_priv->uncore.fw_rendercount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_RENDER);
+		if (--dev_priv->uncore.fw_mediacount == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_MEDIA);
+	} else {
+		if (--dev_priv->uncore.forcewake_count == 0)
+			dev_priv->uncore.funcs.force_wake_put(dev_priv,
+							      FORCEWAKE_ALL);
+	}
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, flags);
+}
+
+static int execlists_ctx_write_tail(struct drm_i915_gem_object *ctx_obj, u32 tail)
+{
+	struct page *page;
+	uint32_t *reg_state;
+
+	page = i915_gem_object_get_page(ctx_obj, 1);
+	reg_state = kmap_atomic(page);
+
+	reg_state[CTX_RING_TAIL+1] = tail;
+
+	kunmap_atomic(reg_state);
+
+	return 0;
+}
+
+static int execlists_submit_context(struct intel_engine_cs *ring,
+				    struct intel_context *to0, u32 tail0,
+				    struct intel_context *to1, u32 tail1)
+{
+	struct drm_i915_gem_object *ctx_obj0;
+	struct drm_i915_gem_object *ctx_obj1 = NULL;
+
+	ctx_obj0 = to0->engine[ring->id].state;
+	BUG_ON(!ctx_obj0);
+	WARN_ON(!i915_gem_obj_is_pinned(ctx_obj0));
+
+	execlists_ctx_write_tail(ctx_obj0, tail0);
+
+	if (to1) {
+		ctx_obj1 = to1->engine[ring->id].state;
+		BUG_ON(!ctx_obj1);
+		WARN_ON(!i915_gem_obj_is_pinned(ctx_obj1));
+
+		execlists_ctx_write_tail(ctx_obj1, tail1);
+	}
+
+	execlists_elsp_write(ring, ctx_obj0, ctx_obj1);
+
+	return 0;
+}
+
+static void execlists_context_unqueue(struct intel_engine_cs *ring)
+{
+	struct intel_ctx_submit_request *req0 = NULL, *req1 = NULL;
+	struct intel_ctx_submit_request *cursor = NULL, *tmp = NULL;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+
+	assert_spin_locked(&ring->execlist_lock);
+
+	if (list_empty(&ring->execlist_queue))
+		return;
+
+	/* Try to read in pairs */
+	list_for_each_entry_safe(cursor, tmp, &ring->execlist_queue,
+				 execlist_link) {
+		if (!req0) {
+			req0 = cursor;
+		} else if (req0->ctx == cursor->ctx) {
+			/* Same ctx: ignore first request, as second request
+			 * will update tail past first request's workload */
+			cursor->elsp_submitted = req0->elsp_submitted;
+			list_del(&req0->execlist_link);
+			queue_work(dev_priv->wq, &req0->work);
+			req0 = cursor;
+		} else {
+			req1 = cursor;
+			break;
+		}
+	}
+
+	WARN_ON(req1 && req1->elsp_submitted);
+
+	WARN_ON(execlists_submit_context(ring, req0->ctx, req0->tail,
+					 req1 ? req1->ctx : NULL,
+					 req1 ? req1->tail : 0));
+
+	req0->elsp_submitted++;
+	if (req1)
+		req1->elsp_submitted++;
+}
+
+static bool execlists_check_remove_request(struct intel_engine_cs *ring,
+					   u32 request_id)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct intel_ctx_submit_request *head_req;
+
+	assert_spin_locked(&ring->execlist_lock);
+
+	head_req = list_first_entry_or_null(&ring->execlist_queue,
+					    struct intel_ctx_submit_request,
+					    execlist_link);
+
+	if (head_req != NULL) {
+		struct drm_i915_gem_object *ctx_obj =
+				head_req->ctx->engine[ring->id].state;
+		if (intel_execlists_ctx_id(ctx_obj) == request_id) {
+			WARN(head_req->elsp_submitted == 0,
+			     "Never submitted head request\n");
+
+			if (--head_req->elsp_submitted <= 0) {
+				list_del(&head_req->execlist_link);
+				queue_work(dev_priv->wq, &head_req->work);
+				return true;
+			}
+		}
+	}
+
+	return false;
+}
+
+/**
+ * intel_execlists_handle_ctx_events() - handle Context Switch interrupts
+ * @ring: Engine Command Streamer to handle.
+ *
+ * Check the unread Context Status Buffers and manage the submission of new
+ * contexts to the ELSP accordingly.
+ */
+void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	u32 status_pointer;
+	u8 read_pointer;
+	u8 write_pointer;
+	u32 status;
+	u32 status_id;
+	u32 submit_contexts = 0;
+
+	status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
+
+	read_pointer = ring->next_context_status_buffer;
+	write_pointer = status_pointer & 0x07;
+	if (read_pointer > write_pointer)
+		write_pointer += 6;
+
+	spin_lock(&ring->execlist_lock);
+
+	while (read_pointer < write_pointer) {
+		read_pointer++;
+		status = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
+				(read_pointer % 6) * 8);
+		status_id = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
+				(read_pointer % 6) * 8 + 4);
+
+		if (status & GEN8_CTX_STATUS_PREEMPTED) {
+			if (status & GEN8_CTX_STATUS_LITE_RESTORE) {
+				if (execlists_check_remove_request(ring, status_id))
+					WARN(1, "Lite Restored request removed from queue\n");
+			} else
+				WARN(1, "Preemption without Lite Restore\n");
+		}
+
+		 if ((status & GEN8_CTX_STATUS_ACTIVE_IDLE) ||
+		     (status & GEN8_CTX_STATUS_ELEMENT_SWITCH)) {
+			if (execlists_check_remove_request(ring, status_id))
+				submit_contexts++;
+		}
+	}
+
+	if (submit_contexts != 0)
+		execlists_context_unqueue(ring);
+
+	spin_unlock(&ring->execlist_lock);
+
+	WARN(submit_contexts > 2, "More than two context complete events?\n");
+	ring->next_context_status_buffer = write_pointer % 6;
+
+	I915_WRITE(RING_CONTEXT_STATUS_PTR(ring),
+		   ((u32)ring->next_context_status_buffer & 0x07) << 8);
+}
+
+static void execlists_free_request_task(struct work_struct *work)
+{
+	struct intel_ctx_submit_request *req =
+		container_of(work, struct intel_ctx_submit_request, work);
+	struct drm_device *dev = req->ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	intel_runtime_pm_put(dev_priv);
+
+	mutex_lock(&dev->struct_mutex);
+	i915_gem_context_unreference(req->ctx);
+	mutex_unlock(&dev->struct_mutex);
+
+	kfree(req);
+}
+
+static int execlists_context_queue(struct intel_engine_cs *ring,
+				   struct intel_context *to,
+				   u32 tail)
+{
+	struct intel_ctx_submit_request *req = NULL, *cursor;
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	unsigned long flags;
+	int num_elements = 0;
+
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (req == NULL)
+		return -ENOMEM;
+	req->ctx = to;
+	i915_gem_context_reference(req->ctx);
+	req->ring = ring;
+	req->tail = tail;
+	INIT_WORK(&req->work, execlists_free_request_task);
+
+	intel_runtime_pm_get(dev_priv);
+
+	spin_lock_irqsave(&ring->execlist_lock, flags);
+
+	list_for_each_entry(cursor, &ring->execlist_queue, execlist_link)
+		if (++num_elements > 2)
+			break;
+
+	if (num_elements > 2) {
+		struct intel_ctx_submit_request *tail_req;
+
+		tail_req = list_last_entry(&ring->execlist_queue,
+					   struct intel_ctx_submit_request,
+					   execlist_link);
+
+		if (to == tail_req->ctx) {
+			WARN(tail_req->elsp_submitted != 0,
+			     "More than 2 already-submitted reqs queued\n");
+			list_del(&tail_req->execlist_link);
+			queue_work(dev_priv->wq, &tail_req->work);
+		}
+	}
+
+	list_add_tail(&req->execlist_link, &ring->execlist_queue);
+	if (num_elements == 0)
+		execlists_context_unqueue(ring);
+
+	spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+	return 0;
+}
+
+static int logical_ring_invalidate_all_caches(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	uint32_t flush_domains;
+	int ret;
+
+	flush_domains = 0;
+	if (ring->gpu_caches_dirty)
+		flush_domains = I915_GEM_GPU_DOMAINS;
+
+	ret = ring->emit_flush(ringbuf, I915_GEM_GPU_DOMAINS, flush_domains);
+	if (ret)
+		return ret;
+
+	ring->gpu_caches_dirty = false;
+	return 0;
+}
+
+static int execlists_move_to_gpu(struct intel_ringbuffer *ringbuf,
+				 struct list_head *vmas)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct i915_vma *vma;
+	uint32_t flush_domains = 0;
+	bool flush_chipset = false;
+	int ret;
+
+	list_for_each_entry(vma, vmas, exec_list) {
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		ret = i915_gem_object_sync(obj, ring);
+		if (ret)
+			return ret;
+
+		if (obj->base.write_domain & I915_GEM_DOMAIN_CPU)
+			flush_chipset |= i915_gem_clflush_object(obj, false);
+
+		flush_domains |= obj->base.write_domain;
+	}
+
+	if (flush_domains & I915_GEM_DOMAIN_GTT)
+		wmb();
+
+	/* Unconditionally invalidate gpu caches and ensure that we do flush
+	 * any residual writes from the previous batch.
+	 */
+	return logical_ring_invalidate_all_caches(ringbuf);
+}
+
+/**
+ * execlists_submission() - submit a batchbuffer for execution, Execlists style
+ * @dev: DRM device.
+ * @file: DRM file.
+ * @ring: Engine Command Streamer to submit to.
+ * @ctx: Context to employ for this submission.
+ * @args: execbuffer call arguments.
+ * @vmas: list of vmas.
+ * @batch_obj: the batchbuffer to submit.
+ * @exec_start: batchbuffer start virtual address pointer.
+ * @flags: translated execbuffer call flags.
+ *
+ * This is the evil twin version of i915_gem_ringbuffer_submission. It abstracts
+ * away the submission details of the execbuffer ioctl call.
+ *
+ * Return: non-zero if the submission fails.
+ */
+int intel_execlists_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	int instp_mode;
+	u32 instp_mask;
+	int ret;
+
+	instp_mode = args->flags & I915_EXEC_CONSTANTS_MASK;
+	instp_mask = I915_EXEC_CONSTANTS_MASK;
+	switch (instp_mode) {
+	case I915_EXEC_CONSTANTS_REL_GENERAL:
+	case I915_EXEC_CONSTANTS_ABSOLUTE:
+	case I915_EXEC_CONSTANTS_REL_SURFACE:
+		if (instp_mode != 0 && ring != &dev_priv->ring[RCS]) {
+			DRM_DEBUG("non-0 rel constants mode on non-RCS\n");
+			return -EINVAL;
+		}
+
+		if (instp_mode != dev_priv->relative_constants_mode) {
+			if (instp_mode == I915_EXEC_CONSTANTS_REL_SURFACE) {
+				DRM_DEBUG("rel surface constants mode invalid on gen5+\n");
+				return -EINVAL;
+			}
+
+			/* The HW changed the meaning on this bit on gen6 */
+			instp_mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
+		}
+		break;
+	default:
+		DRM_DEBUG("execbuf with unknown constants: %d\n", instp_mode);
+		return -EINVAL;
+	}
+
+	if (args->num_cliprects != 0) {
+		DRM_DEBUG("clip rectangles are only valid on pre-gen5\n");
+		return -EINVAL;
+	} else {
+		if (args->DR4 == 0xffffffff) {
+			DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
+			args->DR4 = 0;
+		}
+
+		if (args->DR1 || args->DR4 || args->cliprects_ptr) {
+			DRM_DEBUG("0 cliprects but dirt in cliprects fields\n");
+			return -EINVAL;
+		}
+	}
+
+	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
+		DRM_DEBUG("sol reset is gen7 only\n");
+		return -EINVAL;
+	}
+
+	ret = execlists_move_to_gpu(ringbuf, vmas);
+	if (ret)
+		return ret;
+
+	if (ring == &dev_priv->ring[RCS] &&
+	    instp_mode != dev_priv->relative_constants_mode) {
+		ret = intel_logical_ring_begin(ringbuf, 4);
+		if (ret)
+			return ret;
+
+		intel_logical_ring_emit(ringbuf, MI_NOOP);
+		intel_logical_ring_emit(ringbuf, MI_LOAD_REGISTER_IMM(1));
+		intel_logical_ring_emit(ringbuf, INSTPM);
+		intel_logical_ring_emit(ringbuf, instp_mask << 16 | instp_mode);
+		intel_logical_ring_advance(ringbuf);
+
+		dev_priv->relative_constants_mode = instp_mode;
+	}
+
+	ret = ring->emit_bb_start(ringbuf, exec_start, flags);
+	if (ret)
+		return ret;
+
+	i915_gem_execbuffer_move_to_active(vmas, ring);
+	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
+
+	return 0;
+}
+
+void intel_logical_ring_stop(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	int ret;
+
+	if (!intel_ring_initialized(ring))
+		return;
+
+	ret = intel_ring_idle(ring);
+	if (ret && !i915_reset_in_progress(&to_i915(ring->dev)->gpu_error))
+		DRM_ERROR("failed to quiesce %s whilst cleaning up: %d\n",
+			  ring->name, ret);
+
+	/* TODO: Is this correct with Execlists enabled? */
+	I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING));
+	if (wait_for_atomic((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
+		DRM_ERROR("%s :timed out trying to stop ring\n", ring->name);
+		return;
+	}
+	I915_WRITE_MODE(ring, _MASKED_BIT_DISABLE(STOP_RING));
+}
+
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	int ret;
+
+	if (!ring->gpu_caches_dirty)
+		return 0;
+
+	ret = ring->emit_flush(ringbuf, 0, I915_GEM_GPU_DOMAINS);
+	if (ret)
+		return ret;
+
+	ring->gpu_caches_dirty = false;
+	return 0;
+}
+
+/**
+ * intel_logical_ring_advance_and_submit() - advance the tail and submit the workload
+ * @ringbuf: Logical Ringbuffer to advance.
+ *
+ * The tail is updated in our logical ringbuffer struct, not in the actual context. What
+ * really happens during submission is that the context and current tail will be placed
+ * on a queue waiting for the ELSP to be ready to accept a new context submission. At that
+ * point, the tail *inside* the context is updated and the ELSP written to.
+ */
+void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct intel_context *ctx = ringbuf->FIXME_lrc_ctx;
+
+	intel_logical_ring_advance(ringbuf);
+
+	if (intel_ring_stopped(ring))
+		return;
+
+	execlists_context_queue(ring, ctx, ringbuf->tail);
+}
+
+static int logical_ring_alloc_seqno(struct intel_engine_cs *ring,
+				    struct intel_context *ctx)
+{
+	if (ring->outstanding_lazy_seqno)
+		return 0;
+
+	if (ring->preallocated_lazy_request == NULL) {
+		struct drm_i915_gem_request *request;
+
+		request = kmalloc(sizeof(*request), GFP_KERNEL);
+		if (request == NULL)
+			return -ENOMEM;
+
+		/* Hold a reference to the context this request belongs to
+		 * (we will need it when the time comes to emit/retire the
+		 * request).
+		 */
+		request->ctx = ctx;
+		i915_gem_context_reference(request->ctx);
+
+		ring->preallocated_lazy_request = request;
+	}
+
+	return i915_gem_get_seqno(ring->dev, &ring->outstanding_lazy_seqno);
+}
+
+static int logical_ring_wait_request(struct intel_ringbuffer *ringbuf,
+				     int bytes)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_i915_gem_request *request;
+	u32 seqno = 0;
+	int ret;
+
+	if (ringbuf->last_retired_head != -1) {
+		ringbuf->head = ringbuf->last_retired_head;
+		ringbuf->last_retired_head = -1;
+
+		ringbuf->space = intel_ring_space(ringbuf);
+		if (ringbuf->space >= bytes)
+			return 0;
+	}
+
+	list_for_each_entry(request, &ring->request_list, list) {
+		if (__intel_ring_space(request->tail, ringbuf->tail,
+				       ringbuf->size) >= bytes) {
+			seqno = request->seqno;
+			break;
+		}
+	}
+
+	if (seqno == 0)
+		return -ENOSPC;
+
+	ret = i915_wait_seqno(ring, seqno);
+	if (ret)
+		return ret;
+
+	i915_gem_retire_requests_ring(ring);
+	ringbuf->head = ringbuf->last_retired_head;
+	ringbuf->last_retired_head = -1;
+
+	ringbuf->space = intel_ring_space(ringbuf);
+	return 0;
+}
+
+static int logical_ring_wait_for_space(struct intel_ringbuffer *ringbuf,
+				       int bytes)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long end;
+	int ret;
+
+	ret = logical_ring_wait_request(ringbuf, bytes);
+	if (ret != -ENOSPC)
+		return ret;
+
+	/* Force the context submission in case we have been skipping it */
+	intel_logical_ring_advance_and_submit(ringbuf);
+
+	/* With GEM the hangcheck timer should kick us out of the loop,
+	 * leaving it early runs the risk of corrupting GEM state (due
+	 * to running on almost untested codepaths). But on resume
+	 * timers don't work yet, so prevent a complete hang in that
+	 * case by choosing an insanely large timeout. */
+	end = jiffies + 60 * HZ;
+
+	do {
+		ringbuf->head = I915_READ_HEAD(ring);
+		ringbuf->space = intel_ring_space(ringbuf);
+		if (ringbuf->space >= bytes) {
+			ret = 0;
+			break;
+		}
+
+		msleep(1);
+
+		if (dev_priv->mm.interruptible && signal_pending(current)) {
+			ret = -ERESTARTSYS;
+			break;
+		}
+
+		ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+					   dev_priv->mm.interruptible);
+		if (ret)
+			break;
+
+		if (time_after(jiffies, end)) {
+			ret = -EBUSY;
+			break;
+		}
+	} while (1);
+
+	return ret;
+}
+
+static int logical_ring_wrap_buffer(struct intel_ringbuffer *ringbuf)
+{
+	uint32_t __iomem *virt;
+	int rem = ringbuf->size - ringbuf->tail;
+
+	if (ringbuf->space < rem) {
+		int ret = logical_ring_wait_for_space(ringbuf, rem);
+
+		if (ret)
+			return ret;
+	}
+
+	virt = ringbuf->virtual_start + ringbuf->tail;
+	rem /= 4;
+	while (rem--)
+		iowrite32(MI_NOOP, virt++);
+
+	ringbuf->tail = 0;
+	ringbuf->space = intel_ring_space(ringbuf);
+
+	return 0;
+}
+
+static int logical_ring_prepare(struct intel_ringbuffer *ringbuf, int bytes)
+{
+	int ret;
+
+	if (unlikely(ringbuf->tail + bytes > ringbuf->effective_size)) {
+		ret = logical_ring_wrap_buffer(ringbuf);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	if (unlikely(ringbuf->space < bytes)) {
+		ret = logical_ring_wait_for_space(ringbuf, bytes);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_logical_ring_begin() - prepare the logical ringbuffer to accept some commands
+ *
+ * @ringbuf: Logical ringbuffer.
+ * @num_dwords: number of DWORDs that we plan to write to the ringbuffer.
+ *
+ * The ringbuffer might not be ready to accept the commands right away (maybe it needs to
+ * be wrapped, or wait a bit for the tail to be updated). This function takes care of that
+ * and also preallocates a request (every workload submission is still mediated through
+ * requests, same as it did with legacy ringbuffer submission).
+ *
+ * Return: non-zero if the ringbuffer is not ready to be written to.
+ */
+int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = i915_gem_check_wedge(&dev_priv->gpu_error,
+				   dev_priv->mm.interruptible);
+	if (ret)
+		return ret;
+
+	ret = logical_ring_prepare(ringbuf, num_dwords * sizeof(uint32_t));
+	if (ret)
+		return ret;
+
+	/* Preallocate the olr before touching the ring */
+	ret = logical_ring_alloc_seqno(ring, ringbuf->FIXME_lrc_ctx);
+	if (ret)
+		return ret;
+
+	ringbuf->space -= num_dwords * sizeof(uint32_t);
+	return 0;
+}
+
+static int gen8_init_common_ring(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE_IMR(ring, ~(ring->irq_enable_mask | ring->irq_keep_mask));
+	I915_WRITE(RING_HWSTAM(ring->mmio_base), 0xffffffff);
+
+	I915_WRITE(RING_MODE_GEN7(ring),
+		   _MASKED_BIT_DISABLE(GFX_REPLAY_MODE) |
+		   _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE));
+	POSTING_READ(RING_MODE_GEN7(ring));
+	DRM_DEBUG_DRIVER("Execlists enabled for %s\n", ring->name);
+
+	memset(&ring->hangcheck, 0, sizeof(ring->hangcheck));
+
+	return 0;
+}
+
+static int gen8_init_render_ring(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = gen8_init_common_ring(ring);
+	if (ret)
+		return ret;
+
+	/* We need to disable the AsyncFlip performance optimisations in order
+	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
+	 * programmed to '1' on all products.
+	 *
+	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv,bdw,chv
+	 */
+	I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));
+
+	ret = intel_init_pipe_control(ring);
+	if (ret)
+		return ret;
+
+	I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
+
+	return ret;
+}
+
+static int gen8_emit_bb_start(struct intel_ringbuffer *ringbuf,
+			      u64 offset, unsigned flags)
+{
+	bool ppgtt = !(flags & I915_DISPATCH_SECURE);
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 4);
+	if (ret)
+		return ret;
+
+	/* FIXME(BDW): Address space and security selectors. */
+	intel_logical_ring_emit(ringbuf, MI_BATCH_BUFFER_START_GEN8 | (ppgtt<<8));
+	intel_logical_ring_emit(ringbuf, lower_32_bits(offset));
+	intel_logical_ring_emit(ringbuf, upper_32_bits(offset));
+	intel_logical_ring_emit(ringbuf, MI_NOOP);
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static bool gen8_logical_ring_get_irq(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (!dev->irq_enabled)
+		return false;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	if (ring->irq_refcount++ == 0) {
+		I915_WRITE_IMR(ring, ~(ring->irq_enable_mask | ring->irq_keep_mask));
+		POSTING_READ(RING_IMR(ring->mmio_base));
+	}
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+
+	return true;
+}
+
+static void gen8_logical_ring_put_irq(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	if (--ring->irq_refcount == 0) {
+		I915_WRITE_IMR(ring, ~ring->irq_keep_mask);
+		POSTING_READ(RING_IMR(ring->mmio_base));
+	}
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+}
+
+static int gen8_emit_flush(struct intel_ringbuffer *ringbuf,
+			   u32 invalidate_domains,
+			   u32 unused)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t cmd;
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 4);
+	if (ret)
+		return ret;
+
+	cmd = MI_FLUSH_DW + 1;
+
+	if (ring == &dev_priv->ring[VCS]) {
+		if (invalidate_domains & I915_GEM_GPU_DOMAINS)
+			cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD |
+				MI_FLUSH_DW_STORE_INDEX |
+				MI_FLUSH_DW_OP_STOREDW;
+	} else {
+		if (invalidate_domains & I915_GEM_DOMAIN_RENDER)
+			cmd |= MI_INVALIDATE_TLB | MI_FLUSH_DW_STORE_INDEX |
+				MI_FLUSH_DW_OP_STOREDW;
+	}
+
+	intel_logical_ring_emit(ringbuf, cmd);
+	intel_logical_ring_emit(ringbuf,
+				I915_GEM_HWS_SCRATCH_ADDR |
+				MI_FLUSH_DW_USE_GTT);
+	intel_logical_ring_emit(ringbuf, 0); /* upper addr */
+	intel_logical_ring_emit(ringbuf, 0); /* value */
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static int gen8_emit_flush_render(struct intel_ringbuffer *ringbuf,
+				  u32 invalidate_domains,
+				  u32 flush_domains)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES;
+	u32 flags = 0;
+	int ret;
+
+	flags |= PIPE_CONTROL_CS_STALL;
+
+	if (flush_domains) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+	}
+
+	if (invalidate_domains) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
+	}
+
+	ret = intel_logical_ring_begin(ringbuf, 6);
+	if (ret)
+		return ret;
+
+	intel_logical_ring_emit(ringbuf, GFX_OP_PIPE_CONTROL(6));
+	intel_logical_ring_emit(ringbuf, flags);
+	intel_logical_ring_emit(ringbuf, scratch_addr);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_advance(ringbuf);
+
+	return 0;
+}
+
+static u32 gen8_get_seqno(struct intel_engine_cs *ring, bool lazy_coherency)
+{
+	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
+}
+
+static void gen8_set_seqno(struct intel_engine_cs *ring, u32 seqno)
+{
+	intel_write_status_page(ring, I915_GEM_HWS_INDEX, seqno);
+}
+
+static int gen8_emit_request(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	u32 cmd;
+	int ret;
+
+	ret = intel_logical_ring_begin(ringbuf, 6);
+	if (ret)
+		return ret;
+
+	cmd = MI_STORE_DWORD_IMM_GEN8;
+	cmd |= MI_GLOBAL_GTT;
+
+	intel_logical_ring_emit(ringbuf, cmd);
+	intel_logical_ring_emit(ringbuf,
+				(ring->status_page.gfx_addr +
+				(I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT)));
+	intel_logical_ring_emit(ringbuf, 0);
+	intel_logical_ring_emit(ringbuf, ring->outstanding_lazy_seqno);
+	intel_logical_ring_emit(ringbuf, MI_USER_INTERRUPT);
+	intel_logical_ring_emit(ringbuf, MI_NOOP);
+	intel_logical_ring_advance_and_submit(ringbuf);
+
+	return 0;
+}
+
+/**
+ * intel_logical_ring_cleanup() - deallocate the Engine Command Streamer
+ *
+ * @ring: Engine Command Streamer.
+ *
+ */
+void intel_logical_ring_cleanup(struct intel_engine_cs *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+
+	if (!intel_ring_initialized(ring))
+		return;
+
+	intel_logical_ring_stop(ring);
+	WARN_ON((I915_READ_MODE(ring) & MODE_IDLE) == 0);
+	ring->preallocated_lazy_request = NULL;
+	ring->outstanding_lazy_seqno = 0;
+
+	if (ring->cleanup)
+		ring->cleanup(ring);
+
+	i915_cmd_parser_fini_ring(ring);
+
+	if (ring->status_page.obj) {
+		kunmap(sg_page(ring->status_page.obj->pages->sgl));
+		ring->status_page.obj = NULL;
+	}
+}
+
+static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *ring)
+{
+	int ret;
+
+	/* Intentionally left blank. */
+	ring->buffer = NULL;
+
+	ring->dev = dev;
+	INIT_LIST_HEAD(&ring->active_list);
+	INIT_LIST_HEAD(&ring->request_list);
+	init_waitqueue_head(&ring->irq_queue);
+
+	INIT_LIST_HEAD(&ring->execlist_queue);
+	spin_lock_init(&ring->execlist_lock);
+	ring->next_context_status_buffer = 0;
+
+	ret = i915_cmd_parser_init_ring(ring);
+	if (ret)
+		return ret;
+
+	if (ring->init) {
+		ret = ring->init(ring);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_lr_context_deferred_create(ring->default_context, ring);
+
+	return ret;
+}
+
+static int logical_render_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[RCS];
+
+	ring->name = "render ring";
+	ring->id = RCS;
+	ring->mmio_base = RENDER_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT;
+	if (HAS_L3_DPF(dev))
+		ring->irq_keep_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+
+	ring->init = gen8_init_render_ring;
+	ring->cleanup = intel_fini_pipe_control;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush_render;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_bsd_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VCS];
+
+	ring->name = "bsd ring";
+	ring->id = VCS;
+	ring->mmio_base = GEN6_BSD_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_bsd2_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VCS2];
+
+	ring->name = "bds2 ring";
+	ring->id = VCS2;
+	ring->mmio_base = GEN8_BSD2_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_blt_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[BCS];
+
+	ring->name = "blitter ring";
+	ring->id = BCS;
+	ring->mmio_base = BLT_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+static int logical_vebox_ring_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring = &dev_priv->ring[VECS];
+
+	ring->name = "video enhancement ring";
+	ring->id = VECS;
+	ring->mmio_base = VEBOX_RING_BASE;
+	ring->irq_enable_mask =
+		GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT;
+	ring->irq_keep_mask =
+		GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT;
+
+	ring->init = gen8_init_common_ring;
+	ring->get_seqno = gen8_get_seqno;
+	ring->set_seqno = gen8_set_seqno;
+	ring->emit_request = gen8_emit_request;
+	ring->emit_flush = gen8_emit_flush;
+	ring->irq_get = gen8_logical_ring_get_irq;
+	ring->irq_put = gen8_logical_ring_put_irq;
+	ring->emit_bb_start = gen8_emit_bb_start;
+
+	return logical_ring_init(dev, ring);
+}
+
+/**
+ * intel_logical_rings_init() - allocate, populate and init the Engine Command Streamers
+ * @dev: DRM device.
+ *
+ * This function inits the engines for an Execlists submission style (the equivalent in the
+ * legacy ringbuffer submission world would be i915_gem_init_rings). It does it only for
+ * those engines that are present in the hardware.
+ *
+ * Return: non-zero if the initialization failed.
+ */
+int intel_logical_rings_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = logical_render_ring_init(dev);
+	if (ret)
+		return ret;
+
+	if (HAS_BSD(dev)) {
+		ret = logical_bsd_ring_init(dev);
+		if (ret)
+			goto cleanup_render_ring;
+	}
+
+	if (HAS_BLT(dev)) {
+		ret = logical_blt_ring_init(dev);
+		if (ret)
+			goto cleanup_bsd_ring;
+	}
+
+	if (HAS_VEBOX(dev)) {
+		ret = logical_vebox_ring_init(dev);
+		if (ret)
+			goto cleanup_blt_ring;
+	}
+
+	if (HAS_BSD2(dev)) {
+		ret = logical_bsd2_ring_init(dev);
+		if (ret)
+			goto cleanup_vebox_ring;
+	}
+
+	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
+	if (ret)
+		goto cleanup_bsd2_ring;
+
+	return 0;
+
+cleanup_bsd2_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VCS2]);
+cleanup_vebox_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VECS]);
+cleanup_blt_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[BCS]);
+cleanup_bsd_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[VCS]);
+cleanup_render_ring:
+	intel_logical_ring_cleanup(&dev_priv->ring[RCS]);
+
+	return ret;
+}
+
+int intel_lr_context_render_state_init(struct intel_engine_cs *ring,
+				       struct intel_context *ctx)
+{
+	struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf;
+	struct render_state so;
+	struct drm_i915_file_private *file_priv = ctx->file_priv;
+	struct drm_file *file = file_priv ? file_priv->file : NULL;
+	int ret;
+
+	ret = i915_gem_render_state_prepare(ring, &so);
+	if (ret)
+		return ret;
+
+	if (so.rodata == NULL)
+		return 0;
+
+	ret = ring->emit_bb_start(ringbuf,
+			so.ggtt_offset,
+			I915_DISPATCH_SECURE);
+	if (ret)
+		goto out;
+
+	i915_vma_move_to_active(i915_gem_obj_to_ggtt(so.obj), ring);
+
+	ret = __i915_add_request(ring, file, so.obj, NULL);
+	/* intel_logical_ring_add_request moves object to inactive if it
+	 * fails */
+out:
+	i915_gem_render_state_fini(&so);
+	return ret;
+}
+
+static int
+populate_lr_context(struct intel_context *ctx, struct drm_i915_gem_object *ctx_obj,
+		    struct intel_engine_cs *ring, struct intel_ringbuffer *ringbuf)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *ring_obj = ringbuf->obj;
+	struct i915_hw_ppgtt *ppgtt = ctx->ppgtt;
+	struct page *page;
+	uint32_t *reg_state;
+	int ret;
+
+	if (!ppgtt)
+		ppgtt = dev_priv->mm.aliasing_ppgtt;
+
+	ret = i915_gem_object_set_to_cpu_domain(ctx_obj, true);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Could not set to CPU domain\n");
+		return ret;
+	}
+
+	ret = i915_gem_object_get_pages(ctx_obj);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Could not get object pages\n");
+		return ret;
+	}
+
+	i915_gem_object_pin_pages(ctx_obj);
+
+	/* The second page of the context object contains some fields which must
+	 * be set up prior to the first execution. */
+	page = i915_gem_object_get_page(ctx_obj, 1);
+	reg_state = kmap_atomic(page);
+
+	/* A context is actually a big batch buffer with several MI_LOAD_REGISTER_IMM
+	 * commands followed by (reg, value) pairs. The values we are setting here are
+	 * only for the first context restore: on a subsequent save, the GPU will
+	 * recreate this batchbuffer with new values (including all the missing
+	 * MI_LOAD_REGISTER_IMM commands that we are not initializing here). */
+	if (ring->id == RCS)
+		reg_state[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(14);
+	else
+		reg_state[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(11);
+	reg_state[CTX_LRI_HEADER_0] |= MI_LRI_FORCE_POSTED;
+	reg_state[CTX_CONTEXT_CONTROL] = RING_CONTEXT_CONTROL(ring);
+	reg_state[CTX_CONTEXT_CONTROL+1] =
+			_MASKED_BIT_ENABLE((1<<3) | MI_RESTORE_INHIBIT);
+	reg_state[CTX_RING_HEAD] = RING_HEAD(ring->mmio_base);
+	reg_state[CTX_RING_HEAD+1] = 0;
+	reg_state[CTX_RING_TAIL] = RING_TAIL(ring->mmio_base);
+	reg_state[CTX_RING_TAIL+1] = 0;
+	reg_state[CTX_RING_BUFFER_START] = RING_START(ring->mmio_base);
+	reg_state[CTX_RING_BUFFER_START+1] = i915_gem_obj_ggtt_offset(ring_obj);
+	reg_state[CTX_RING_BUFFER_CONTROL] = RING_CTL(ring->mmio_base);
+	reg_state[CTX_RING_BUFFER_CONTROL+1] =
+			((ringbuf->size - PAGE_SIZE) & RING_NR_PAGES) | RING_VALID;
+	reg_state[CTX_BB_HEAD_U] = ring->mmio_base + 0x168;
+	reg_state[CTX_BB_HEAD_U+1] = 0;
+	reg_state[CTX_BB_HEAD_L] = ring->mmio_base + 0x140;
+	reg_state[CTX_BB_HEAD_L+1] = 0;
+	reg_state[CTX_BB_STATE] = ring->mmio_base + 0x110;
+	reg_state[CTX_BB_STATE+1] = (1<<5);
+	reg_state[CTX_SECOND_BB_HEAD_U] = ring->mmio_base + 0x11c;
+	reg_state[CTX_SECOND_BB_HEAD_U+1] = 0;
+	reg_state[CTX_SECOND_BB_HEAD_L] = ring->mmio_base + 0x114;
+	reg_state[CTX_SECOND_BB_HEAD_L+1] = 0;
+	reg_state[CTX_SECOND_BB_STATE] = ring->mmio_base + 0x118;
+	reg_state[CTX_SECOND_BB_STATE+1] = 0;
+	if (ring->id == RCS) {
+		/* TODO: according to BSpec, the register state context
+		 * for CHV does not have these. OTOH, these registers do
+		 * exist in CHV. I'm waiting for a clarification */
+		reg_state[CTX_BB_PER_CTX_PTR] = ring->mmio_base + 0x1c0;
+		reg_state[CTX_BB_PER_CTX_PTR+1] = 0;
+		reg_state[CTX_RCS_INDIRECT_CTX] = ring->mmio_base + 0x1c4;
+		reg_state[CTX_RCS_INDIRECT_CTX+1] = 0;
+		reg_state[CTX_RCS_INDIRECT_CTX_OFFSET] = ring->mmio_base + 0x1c8;
+		reg_state[CTX_RCS_INDIRECT_CTX_OFFSET+1] = 0;
+	}
+	reg_state[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9);
+	reg_state[CTX_LRI_HEADER_1] |= MI_LRI_FORCE_POSTED;
+	reg_state[CTX_CTX_TIMESTAMP] = ring->mmio_base + 0x3a8;
+	reg_state[CTX_CTX_TIMESTAMP+1] = 0;
+	reg_state[CTX_PDP3_UDW] = GEN8_RING_PDP_UDW(ring, 3);
+	reg_state[CTX_PDP3_LDW] = GEN8_RING_PDP_LDW(ring, 3);
+	reg_state[CTX_PDP2_UDW] = GEN8_RING_PDP_UDW(ring, 2);
+	reg_state[CTX_PDP2_LDW] = GEN8_RING_PDP_LDW(ring, 2);
+	reg_state[CTX_PDP1_UDW] = GEN8_RING_PDP_UDW(ring, 1);
+	reg_state[CTX_PDP1_LDW] = GEN8_RING_PDP_LDW(ring, 1);
+	reg_state[CTX_PDP0_UDW] = GEN8_RING_PDP_UDW(ring, 0);
+	reg_state[CTX_PDP0_LDW] = GEN8_RING_PDP_LDW(ring, 0);
+	reg_state[CTX_PDP3_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[3]);
+	reg_state[CTX_PDP3_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[3]);
+	reg_state[CTX_PDP2_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[2]);
+	reg_state[CTX_PDP2_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[2]);
+	reg_state[CTX_PDP1_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[1]);
+	reg_state[CTX_PDP1_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[1]);
+	reg_state[CTX_PDP0_UDW+1] = upper_32_bits(ppgtt->pd_dma_addr[0]);
+	reg_state[CTX_PDP0_LDW+1] = lower_32_bits(ppgtt->pd_dma_addr[0]);
+	if (ring->id == RCS) {
+		reg_state[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
+		reg_state[CTX_R_PWR_CLK_STATE] = 0x20c8;
+		reg_state[CTX_R_PWR_CLK_STATE+1] = 0;
+	}
+
+	kunmap_atomic(reg_state);
+
+	ctx_obj->dirty = 1;
+	set_page_dirty(page);
+	i915_gem_object_unpin_pages(ctx_obj);
+
+	return 0;
+}
+
+/**
+ * intel_lr_context_free() - free the LRC specific bits of a context
+ * @ctx: the LR context to free.
+ *
+ * The real context freeing is done in i915_gem_context_free: this only
+ * takes care of the bits that are LRC related: the per-engine backing
+ * objects and the logical ringbuffer.
+ */
+void intel_lr_context_free(struct intel_context *ctx)
+{
+	int i;
+
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct drm_i915_gem_object *ctx_obj = ctx->engine[i].state;
+		struct intel_ringbuffer *ringbuf = ctx->engine[i].ringbuf;
+
+		if (ctx_obj) {
+			intel_destroy_ringbuffer_obj(ringbuf);
+			kfree(ringbuf);
+			i915_gem_object_ggtt_unpin(ctx_obj);
+			drm_gem_object_unreference(&ctx_obj->base);
+		}
+	}
+}
+
+static uint32_t get_lr_context_size(struct intel_engine_cs *ring)
+{
+	int ret = 0;
+
+	WARN_ON(INTEL_INFO(ring->dev)->gen != 8);
+
+	switch (ring->id) {
+	case RCS:
+		ret = GEN8_LR_CONTEXT_RENDER_SIZE;
+		break;
+	case VCS:
+	case BCS:
+	case VECS:
+	case VCS2:
+		ret = GEN8_LR_CONTEXT_OTHER_SIZE;
+		break;
+	}
+
+	return ret;
+}
+
+/**
+ * intel_lr_context_deferred_create() - create the LRC specific bits of a context
+ * @ctx: LR context to create.
+ * @ring: engine to be used with the context.
+ *
+ * This function can be called more than once, with different engines, if we plan
+ * to use the context with them. The context backing objects and the ringbuffers
+ * (specially the ringbuffer backing objects) suck a lot of memory up, and that's why
+ * the creation is a deferred call: it's better to make sure first that we need to use
+ * a given ring with the context.
+ *
+ * Return: non-zero on eror.
+ */
+int intel_lr_context_deferred_create(struct intel_context *ctx,
+				     struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_gem_object *ctx_obj;
+	uint32_t context_size;
+	struct intel_ringbuffer *ringbuf;
+	int ret;
+
+	WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL);
+	if (ctx->engine[ring->id].state)
+		return 0;
+
+	context_size = round_up(get_lr_context_size(ring), 4096);
+
+	ctx_obj = i915_gem_alloc_context_obj(dev, context_size);
+	if (IS_ERR(ctx_obj)) {
+		ret = PTR_ERR(ctx_obj);
+		DRM_DEBUG_DRIVER("Alloc LRC backing obj failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = i915_gem_obj_ggtt_pin(ctx_obj, GEN8_LR_CONTEXT_ALIGN, 0);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Pin LRC backing obj failed: %d\n", ret);
+		drm_gem_object_unreference(&ctx_obj->base);
+		return ret;
+	}
+
+	ringbuf = kzalloc(sizeof(*ringbuf), GFP_KERNEL);
+	if (!ringbuf) {
+		DRM_DEBUG_DRIVER("Failed to allocate ringbuffer %s\n",
+				ring->name);
+		i915_gem_object_ggtt_unpin(ctx_obj);
+		drm_gem_object_unreference(&ctx_obj->base);
+		ret = -ENOMEM;
+		return ret;
+	}
+
+	ringbuf->ring = ring;
+	ringbuf->FIXME_lrc_ctx = ctx;
+
+	ringbuf->size = 32 * PAGE_SIZE;
+	ringbuf->effective_size = ringbuf->size;
+	ringbuf->head = 0;
+	ringbuf->tail = 0;
+	ringbuf->space = ringbuf->size;
+	ringbuf->last_retired_head = -1;
+
+	/* TODO: For now we put this in the mappable region so that we can reuse
+	 * the existing ringbuffer code which ioremaps it. When we start
+	 * creating many contexts, this will no longer work and we must switch
+	 * to a kmapish interface.
+	 */
+	ret = intel_alloc_ringbuffer_obj(dev, ringbuf);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Failed to allocate ringbuffer obj %s: %d\n",
+				ring->name, ret);
+		goto error;
+	}
+
+	ret = populate_lr_context(ctx, ctx_obj, ring, ringbuf);
+	if (ret) {
+		DRM_DEBUG_DRIVER("Failed to populate LRC: %d\n", ret);
+		intel_destroy_ringbuffer_obj(ringbuf);
+		goto error;
+	}
+
+	ctx->engine[ring->id].ringbuf = ringbuf;
+	ctx->engine[ring->id].state = ctx_obj;
+
+	if (ctx == ring->default_context) {
+		/* The status page is offset 0 from the default context object
+		 * in LRC mode. */
+		ring->status_page.gfx_addr = i915_gem_obj_ggtt_offset(ctx_obj);
+		ring->status_page.page_addr =
+				kmap(sg_page(ctx_obj->pages->sgl));
+		if (ring->status_page.page_addr == NULL)
+			return -ENOMEM;
+		ring->status_page.obj = ctx_obj;
+	}
+
+	if (ring->id == RCS && !ctx->rcs_initialized) {
+		ret = intel_lr_context_render_state_init(ring, ctx);
+		if (ret) {
+			DRM_ERROR("Init render state failed: %d\n", ret);
+			ctx->engine[ring->id].ringbuf = NULL;
+			ctx->engine[ring->id].state = NULL;
+			intel_destroy_ringbuffer_obj(ringbuf);
+			goto error;
+		}
+		ctx->rcs_initialized = true;
+	}
+
+	return 0;
+
+error:
+	kfree(ringbuf);
+	i915_gem_object_ggtt_unpin(ctx_obj);
+	drm_gem_object_unreference(&ctx_obj->base);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
new file mode 100644
index 000000000000..33c3b4bf28c5
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_lrc.h
@@ -0,0 +1,114 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _INTEL_LRC_H_
+#define _INTEL_LRC_H_
+
+/* Execlists regs */
+#define RING_ELSP(ring)			((ring)->mmio_base+0x230)
+#define RING_EXECLIST_STATUS(ring)	((ring)->mmio_base+0x234)
+#define RING_CONTEXT_CONTROL(ring)	((ring)->mmio_base+0x244)
+#define RING_CONTEXT_STATUS_BUF(ring)	((ring)->mmio_base+0x370)
+#define RING_CONTEXT_STATUS_PTR(ring)	((ring)->mmio_base+0x3a0)
+
+/* Logical Rings */
+void intel_logical_ring_stop(struct intel_engine_cs *ring);
+void intel_logical_ring_cleanup(struct intel_engine_cs *ring);
+int intel_logical_rings_init(struct drm_device *dev);
+
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf);
+void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf);
+/**
+ * intel_logical_ring_advance() - advance the ringbuffer tail
+ * @ringbuf: Ringbuffer to advance.
+ *
+ * The tail is only updated in our logical ringbuffer struct.
+ */
+static inline void intel_logical_ring_advance(struct intel_ringbuffer *ringbuf)
+{
+	ringbuf->tail &= ringbuf->size - 1;
+}
+/**
+ * intel_logical_ring_emit() - write a DWORD to the ringbuffer.
+ * @ringbuf: Ringbuffer to write to.
+ * @data: DWORD to write.
+ */
+static inline void intel_logical_ring_emit(struct intel_ringbuffer *ringbuf,
+					   u32 data)
+{
+	iowrite32(data, ringbuf->virtual_start + ringbuf->tail);
+	ringbuf->tail += 4;
+}
+int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords);
+
+/* Logical Ring Contexts */
+int intel_lr_context_render_state_init(struct intel_engine_cs *ring,
+				       struct intel_context *ctx);
+void intel_lr_context_free(struct intel_context *ctx);
+int intel_lr_context_deferred_create(struct intel_context *ctx,
+				     struct intel_engine_cs *ring);
+
+/* Execlists */
+int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists);
+int intel_execlists_submission(struct drm_device *dev, struct drm_file *file,
+			       struct intel_engine_cs *ring,
+			       struct intel_context *ctx,
+			       struct drm_i915_gem_execbuffer2 *args,
+			       struct list_head *vmas,
+			       struct drm_i915_gem_object *batch_obj,
+			       u64 exec_start, u32 flags);
+u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj);
+
+/**
+ * struct intel_ctx_submit_request - queued context submission request
+ * @ctx: Context to submit to the ELSP.
+ * @ring: Engine to submit it to.
+ * @tail: how far in the context's ringbuffer this request goes to.
+ * @execlist_link: link in the submission queue.
+ * @work: workqueue for processing this request in a bottom half.
+ * @elsp_submitted: no. of times this request has been sent to the ELSP.
+ *
+ * The ELSP only accepts two elements at a time, so we queue context/tail
+ * pairs on a given queue (ring->execlist_queue) until the hardware is
+ * available. The queue serves a double purpose: we also use it to keep track
+ * of the up to 2 contexts currently in the hardware (usually one in execution
+ * and the other queued up by the GPU): We only remove elements from the head
+ * of the queue when the hardware informs us that an element has been
+ * completed.
+ *
+ * All accesses to the queue are mediated by a spinlock (ring->execlist_lock).
+ */
+struct intel_ctx_submit_request {
+	struct intel_context *ctx;
+	struct intel_engine_cs *ring;
+	u32 tail;
+
+	struct list_head execlist_link;
+	struct work_struct work;
+
+	int elsp_submitted;
+};
+
+void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring);
+
+#endif /* _INTEL_LRC_H_ */
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 5e5a72fca5fb..a6bd1422e38f 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -51,6 +51,7 @@ struct intel_lvds_encoder {
 
 	bool is_dual_link;
 	u32 reg;
+	u32 a3_power;
 
 	struct intel_lvds_connector *attached_connector;
 };
@@ -71,8 +72,13 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	enum intel_display_power_domain power_domain;
 	u32 tmp;
 
+	power_domain = intel_display_port_power_domain(encoder);
+	if (!intel_display_power_enabled(dev_priv, power_domain))
+		return false;
+
 	tmp = I915_READ(lvds_encoder->reg);
 
 	if (!(tmp & LVDS_PORT_EN))
@@ -172,8 +178,11 @@ static void intel_pre_enable_lvds(struct intel_encoder *encoder)
 
 	/* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP)
 	 * appropriately here, but we need to look more thoroughly into how
-	 * panels behave in the two modes.
+	 * panels behave in the two modes. For now, let's just maintain the
+	 * value we got from the BIOS.
 	 */
+	 temp &= ~LVDS_A3_POWER_MASK;
+	 temp |= lvds_encoder->a3_power;
 
 	/* Set the dithering flag on LVDS as needed, note that there is no
 	 * special lvds dither control bit on pch-split platforms, dithering is
@@ -271,7 +280,6 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 				      struct intel_crtc_config *pipe_config)
 {
 	struct drm_device *dev = intel_encoder->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds_encoder *lvds_encoder =
 		to_lvds_encoder(&intel_encoder->base);
 	struct intel_connector *intel_connector =
@@ -286,8 +294,7 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 		return false;
 	}
 
-	if ((I915_READ(lvds_encoder->reg) & LVDS_A3_POWER_MASK) ==
-	    LVDS_A3_POWER_UP)
+	if (lvds_encoder->a3_power == LVDS_A3_POWER_UP)
 		lvds_bpp = 8*3;
 	else
 		lvds_bpp = 6*3;
@@ -531,7 +538,7 @@ static const struct drm_encoder_funcs intel_lvds_enc_funcs = {
 	.destroy = intel_encoder_destroy,
 };
 
-static int __init intel_no_lvds_dmi_callback(const struct dmi_system_id *id)
+static int intel_no_lvds_dmi_callback(const struct dmi_system_id *id)
 {
 	DRM_INFO("Skipping LVDS initialization for %s\n", id->ident);
 	return 1;
@@ -816,8 +823,7 @@ bool intel_is_dual_link_lvds(struct drm_device *dev)
 	struct intel_encoder *encoder;
 	struct intel_lvds_encoder *lvds_encoder;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
+	for_each_intel_encoder(dev, encoder) {
 		if (encoder->type == INTEL_OUTPUT_LVDS) {
 			lvds_encoder = to_lvds_encoder(&encoder->base);
 
@@ -1088,6 +1094,9 @@ out:
 	DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
 		      lvds_encoder->is_dual_link ? "dual" : "single");
 
+	lvds_encoder->a3_power = I915_READ(lvds_encoder->reg) &
+				 LVDS_A3_POWER_MASK;
+
 	/*
 	 * Unlock registers and just
 	 * leave them unlocked
@@ -1104,7 +1113,7 @@ out:
 		DRM_DEBUG_KMS("lid notifier registration failed\n");
 		lvds_connector->lid_notifier.notifier_call = NULL;
 	}
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	intel_panel_init(&intel_connector->panel, fixed_mode, downclock_mode);
 	intel_panel_setup_backlight(connector);
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index 4f6b53998d79..d8de1d5140a7 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -352,6 +352,7 @@ int intel_opregion_notify_encoder(struct intel_encoder *intel_encoder,
 	case INTEL_OUTPUT_UNKNOWN:
 	case INTEL_OUTPUT_DISPLAYPORT:
 	case INTEL_OUTPUT_HDMI:
+	case INTEL_OUTPUT_DP_MST:
 		type = DISPLAY_TYPE_EXTERNAL_FLAT_PANEL;
 		break;
 	case INTEL_OUTPUT_EDP:
@@ -395,6 +396,16 @@ int intel_opregion_notify_adapter(struct drm_device *dev, pci_power_t state)
 	return -EINVAL;
 }
 
+/*
+ * If the vendor backlight interface is not in use and ACPI backlight interface
+ * is broken, do not bother processing backlight change requests from firmware.
+ */
+static bool should_ignore_backlight_request(void)
+{
+	return acpi_video_backlight_support() &&
+	       !acpi_video_verify_backlight_support();
+}
+
 static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -403,11 +414,7 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 
 	DRM_DEBUG_DRIVER("bclp = 0x%08x\n", bclp);
 
-	/*
-	 * If the acpi_video interface is not supposed to be used, don't
-	 * bother processing backlight level change requests from firmware.
-	 */
-	if (!acpi_video_verify_backlight_support()) {
+	if (should_ignore_backlight_request()) {
 		DRM_DEBUG_KMS("opregion backlight request ignored\n");
 		return 0;
 	}
@@ -427,7 +434,7 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 	 */
 	DRM_DEBUG_KMS("updating opregion backlight %d/255\n", bclp);
 	list_for_each_entry(intel_connector, &dev->mode_config.connector_list, base.head)
-		intel_panel_set_backlight(intel_connector, bclp, 255);
+		intel_panel_set_backlight_acpi(intel_connector, bclp, 255);
 	iowrite32(DIV_ROUND_UP(bclp * 100, 255) | ASLE_CBLV_VALID, &asle->cblv);
 
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index daa118978eec..dc2f4f26c961 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -415,6 +415,10 @@ static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
 	}
 
 	intel_overlay_release_old_vid_tail(overlay);
+
+
+	i915_gem_track_fb(overlay->old_vid_bo, NULL,
+			  INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe));
 	return 0;
 }
 
@@ -686,6 +690,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	bool scale_changed = false;
 	struct drm_device *dev = overlay->dev;
 	u32 swidth, swidthsw, sheight, ostride;
+	enum pipe pipe = overlay->crtc->pipe;
 
 	BUG_ON(!mutex_is_locked(&dev->struct_mutex));
 	BUG_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
@@ -713,7 +718,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 		oconfig = OCONF_CC_OUT_8BIT;
 		if (IS_GEN4(overlay->dev))
 			oconfig |= OCONF_CSC_MODE_BT709;
-		oconfig |= overlay->crtc->pipe == 0 ?
+		oconfig |= pipe == 0 ?
 			OCONF_PIPE_A : OCONF_PIPE_B;
 		iowrite32(oconfig, &regs->OCONFIG);
 		intel_overlay_unmap_regs(overlay, regs);
@@ -776,9 +781,15 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	if (ret)
 		goto out_unpin;
 
+	i915_gem_track_fb(overlay->vid_bo, new_bo,
+			  INTEL_FRONTBUFFER_OVERLAY(pipe));
+
 	overlay->old_vid_bo = overlay->vid_bo;
 	overlay->vid_bo = new_bo;
 
+	intel_frontbuffer_flip(dev,
+			       INTEL_FRONTBUFFER_OVERLAY(pipe));
+
 	return 0;
 
 out_unpin:
@@ -1028,7 +1039,7 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 	struct drm_intel_overlay_put_image *put_image_rec = data;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay;
-	struct drm_mode_object *drmmode_obj;
+	struct drm_crtc *drmmode_crtc;
 	struct intel_crtc *crtc;
 	struct drm_i915_gem_object *new_bo;
 	struct put_image_params *params;
@@ -1057,13 +1068,12 @@ int intel_overlay_put_image(struct drm_device *dev, void *data,
 	if (!params)
 		return -ENOMEM;
 
-	drmmode_obj = drm_mode_object_find(dev, put_image_rec->crtc_id,
-					   DRM_MODE_OBJECT_CRTC);
-	if (!drmmode_obj) {
+	drmmode_crtc = drm_crtc_find(dev, put_image_rec->crtc_id);
+	if (!drmmode_crtc) {
 		ret = -ENOENT;
 		goto out_free;
 	}
-	crtc = to_intel_crtc(obj_to_crtc(drmmode_obj));
+	crtc = to_intel_crtc(drmmode_crtc);
 
 	new_bo = to_intel_bo(drm_gem_object_lookup(dev, file_priv,
 						   put_image_rec->bo_handle));
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 12b02fe1d0ae..41b3be217493 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -398,6 +398,68 @@ intel_panel_detect(struct drm_device *dev)
 	}
 }
 
+/**
+ * scale - scale values from one range to another
+ *
+ * @source_val: value in range [@source_min..@source_max]
+ *
+ * Return @source_val in range [@source_min..@source_max] scaled to range
+ * [@target_min..@target_max].
+ */
+static uint32_t scale(uint32_t source_val,
+		      uint32_t source_min, uint32_t source_max,
+		      uint32_t target_min, uint32_t target_max)
+{
+	uint64_t target_val;
+
+	WARN_ON(source_min > source_max);
+	WARN_ON(target_min > target_max);
+
+	/* defensive */
+	source_val = clamp(source_val, source_min, source_max);
+
+	/* avoid overflows */
+	target_val = DIV_ROUND_CLOSEST_ULL((uint64_t)(source_val - source_min) *
+			(target_max - target_min), source_max - source_min);
+	target_val += target_min;
+
+	return target_val;
+}
+
+/* Scale user_level in range [0..user_max] to [hw_min..hw_max]. */
+static inline u32 scale_user_to_hw(struct intel_connector *connector,
+				   u32 user_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return scale(user_level, 0, user_max,
+		     panel->backlight.min, panel->backlight.max);
+}
+
+/* Scale user_level in range [0..user_max] to [0..hw_max], clamping the result
+ * to [hw_min..hw_max]. */
+static inline u32 clamp_user_to_hw(struct intel_connector *connector,
+				   u32 user_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+	u32 hw_level;
+
+	hw_level = scale(user_level, 0, user_max, 0, panel->backlight.max);
+	hw_level = clamp(hw_level, panel->backlight.min, panel->backlight.max);
+
+	return hw_level;
+}
+
+/* Scale hw_level in range [hw_min..hw_max] to [0..user_max]. */
+static inline u32 scale_hw_to_user(struct intel_connector *connector,
+				   u32 hw_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return scale(hw_level, panel->backlight.min, panel->backlight.max,
+		     0, user_max);
+}
+
 static u32 intel_panel_compute_brightness(struct intel_connector *connector,
 					  u32 val)
 {
@@ -557,17 +619,16 @@ intel_panel_actually_set_backlight(struct intel_connector *connector, u32 level)
 	dev_priv->display.set_backlight(connector, level);
 }
 
-/* set backlight brightness to level in range [0..max] */
-void intel_panel_set_backlight(struct intel_connector *connector, u32 level,
-			       u32 max)
+/* set backlight brightness to level in range [0..max], scaling wrt hw min */
+static void intel_panel_set_backlight(struct intel_connector *connector,
+				      u32 user_level, u32 user_max)
 {
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_panel *panel = &connector->panel;
 	enum pipe pipe = intel_get_pipe_from_connector(connector);
-	u32 freq;
+	u32 hw_level;
 	unsigned long flags;
-	u64 n;
 
 	if (!panel->backlight.present || pipe == INVALID_PIPE)
 		return;
@@ -576,18 +637,46 @@ void intel_panel_set_backlight(struct intel_connector *connector, u32 level,
 
 	WARN_ON(panel->backlight.max == 0);
 
-	/* scale to hardware max, but be careful to not overflow */
-	freq = panel->backlight.max;
-	n = (u64)level * freq;
-	do_div(n, max);
-	level = n;
+	hw_level = scale_user_to_hw(connector, user_level, user_max);
+	panel->backlight.level = hw_level;
+
+	if (panel->backlight.enabled)
+		intel_panel_actually_set_backlight(connector, hw_level);
+
+	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
+}
+
+/* set backlight brightness to level in range [0..max], assuming hw min is
+ * respected.
+ */
+void intel_panel_set_backlight_acpi(struct intel_connector *connector,
+				    u32 user_level, u32 user_max)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_panel *panel = &connector->panel;
+	enum pipe pipe = intel_get_pipe_from_connector(connector);
+	u32 hw_level;
+	unsigned long flags;
+
+	if (!panel->backlight.present || pipe == INVALID_PIPE)
+		return;
+
+	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
+
+	WARN_ON(panel->backlight.max == 0);
+
+	hw_level = clamp_user_to_hw(connector, user_level, user_max);
+	panel->backlight.level = hw_level;
 
-	panel->backlight.level = level;
 	if (panel->backlight.device)
-		panel->backlight.device->props.brightness = level;
+		panel->backlight.device->props.brightness =
+			scale_hw_to_user(connector,
+					 panel->backlight.level,
+					 panel->backlight.device->props.max_brightness);
 
 	if (panel->backlight.enabled)
-		intel_panel_actually_set_backlight(connector, level);
+		intel_panel_actually_set_backlight(connector, hw_level);
 
 	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
 }
@@ -661,6 +750,8 @@ void intel_panel_disable_backlight(struct intel_connector *connector)
 
 	spin_lock_irqsave(&dev_priv->backlight_lock, flags);
 
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_POWERDOWN;
 	panel->backlight.enabled = false;
 	dev_priv->display.disable_backlight(connector);
 
@@ -711,7 +802,7 @@ static void pch_enable_backlight(struct intel_connector *connector)
 
 	cpu_ctl2 = I915_READ(BLC_PWM_CPU_CTL2);
 	if (cpu_ctl2 & BLM_PWM_ENABLE) {
-		WARN(1, "cpu backlight already enabled\n");
+		DRM_DEBUG_KMS("cpu backlight already enabled\n");
 		cpu_ctl2 &= ~BLM_PWM_ENABLE;
 		I915_WRITE(BLC_PWM_CPU_CTL2, cpu_ctl2);
 	}
@@ -755,7 +846,7 @@ static void i9xx_enable_backlight(struct intel_connector *connector)
 
 	ctl = I915_READ(BLC_PWM_CTL);
 	if (ctl & BACKLIGHT_DUTY_CYCLE_MASK_PNV) {
-		WARN(1, "backlight already enabled\n");
+		DRM_DEBUG_KMS("backlight already enabled\n");
 		I915_WRITE(BLC_PWM_CTL, 0);
 	}
 
@@ -786,7 +877,7 @@ static void i965_enable_backlight(struct intel_connector *connector)
 
 	ctl2 = I915_READ(BLC_PWM_CTL2);
 	if (ctl2 & BLM_PWM_ENABLE) {
-		WARN(1, "backlight already enabled\n");
+		DRM_DEBUG_KMS("backlight already enabled\n");
 		ctl2 &= ~BLM_PWM_ENABLE;
 		I915_WRITE(BLC_PWM_CTL2, ctl2);
 	}
@@ -820,7 +911,7 @@ static void vlv_enable_backlight(struct intel_connector *connector)
 
 	ctl2 = I915_READ(VLV_BLC_PWM_CTL2(pipe));
 	if (ctl2 & BLM_PWM_ENABLE) {
-		WARN(1, "backlight already enabled\n");
+		DRM_DEBUG_KMS("backlight already enabled\n");
 		ctl2 &= ~BLM_PWM_ENABLE;
 		I915_WRITE(VLV_BLC_PWM_CTL2(pipe), ctl2);
 	}
@@ -860,11 +951,15 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 		panel->backlight.level = panel->backlight.max;
 		if (panel->backlight.device)
 			panel->backlight.device->props.brightness =
-				panel->backlight.level;
+				scale_hw_to_user(connector,
+						 panel->backlight.level,
+						 panel->backlight.device->props.max_brightness);
 	}
 
 	dev_priv->display.enable_backlight(connector);
 	panel->backlight.enabled = true;
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_UNBLANK;
 
 	spin_unlock_irqrestore(&dev_priv->backlight_lock, flags);
 }
@@ -873,6 +968,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 static int intel_backlight_device_update_status(struct backlight_device *bd)
 {
 	struct intel_connector *connector = bl_get_data(bd);
+	struct intel_panel *panel = &connector->panel;
 	struct drm_device *dev = connector->base.dev;
 
 	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
@@ -880,6 +976,23 @@ static int intel_backlight_device_update_status(struct backlight_device *bd)
 		      bd->props.brightness, bd->props.max_brightness);
 	intel_panel_set_backlight(connector, bd->props.brightness,
 				  bd->props.max_brightness);
+
+	/*
+	 * Allow flipping bl_power as a sub-state of enabled. Sadly the
+	 * backlight class device does not make it easy to to differentiate
+	 * between callbacks for brightness and bl_power, so our backlight_power
+	 * callback needs to take this into account.
+	 */
+	if (panel->backlight.enabled) {
+		if (panel->backlight_power) {
+			bool enable = bd->props.power == FB_BLANK_UNBLANK &&
+				bd->props.brightness != 0;
+			panel->backlight_power(connector, enable);
+		}
+	} else {
+		bd->props.power = FB_BLANK_POWERDOWN;
+	}
+
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
 	return 0;
 }
@@ -889,11 +1002,15 @@ static int intel_backlight_device_get_brightness(struct backlight_device *bd)
 	struct intel_connector *connector = bl_get_data(bd);
 	struct drm_device *dev = connector->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 hw_level;
 	int ret;
 
 	intel_runtime_pm_get(dev_priv);
 	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	ret = intel_panel_get_backlight(connector);
+
+	hw_level = intel_panel_get_backlight(connector);
+	ret = scale_hw_to_user(connector, hw_level, bd->props.max_brightness);
+
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
 	intel_runtime_pm_put(dev_priv);
 
@@ -913,12 +1030,24 @@ static int intel_backlight_device_register(struct intel_connector *connector)
 	if (WARN_ON(panel->backlight.device))
 		return -ENODEV;
 
-	BUG_ON(panel->backlight.max == 0);
+	WARN_ON(panel->backlight.max == 0);
 
 	memset(&props, 0, sizeof(props));
 	props.type = BACKLIGHT_RAW;
-	props.brightness = panel->backlight.level;
+
+	/*
+	 * Note: Everything should work even if the backlight device max
+	 * presented to the userspace is arbitrarily chosen.
+	 */
 	props.max_brightness = panel->backlight.max;
+	props.brightness = scale_hw_to_user(connector,
+					    panel->backlight.level,
+					    props.max_brightness);
+
+	if (panel->backlight.enabled)
+		props.power = FB_BLANK_UNBLANK;
+	else
+		props.power = FB_BLANK_POWERDOWN;
 
 	/*
 	 * Note: using the same name independent of the connector prevents
@@ -964,6 +1093,32 @@ static void intel_backlight_device_unregister(struct intel_connector *connector)
  * XXX: Query mode clock or hardware clock and program PWM modulation frequency
  * appropriately when it's 0. Use VBT and/or sane defaults.
  */
+static u32 get_backlight_min_vbt(struct intel_connector *connector)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_panel *panel = &connector->panel;
+	int min;
+
+	WARN_ON(panel->backlight.max == 0);
+
+	/*
+	 * XXX: If the vbt value is 255, it makes min equal to max, which leads
+	 * to problems. There are such machines out there. Either our
+	 * interpretation is wrong or the vbt has bogus data. Or both. Safeguard
+	 * against this by letting the minimum be at most (arbitrarily chosen)
+	 * 25% of the max.
+	 */
+	min = clamp_t(int, dev_priv->vbt.backlight.min_brightness, 0, 64);
+	if (min != dev_priv->vbt.backlight.min_brightness) {
+		DRM_DEBUG_KMS("clamping VBT min backlight %d/255 to %d/255\n",
+			      dev_priv->vbt.backlight.min_brightness, min);
+	}
+
+	/* vbt value is a coefficient in range [0..255] */
+	return scale(min, 0, 255, 0, panel->backlight.max);
+}
+
 static int bdw_setup_backlight(struct intel_connector *connector)
 {
 	struct drm_device *dev = connector->base.dev;
@@ -979,6 +1134,8 @@ static int bdw_setup_backlight(struct intel_connector *connector)
 	if (!panel->backlight.max)
 		return -ENODEV;
 
+	panel->backlight.min = get_backlight_min_vbt(connector);
+
 	val = bdw_get_backlight(connector);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
@@ -1003,6 +1160,8 @@ static int pch_setup_backlight(struct intel_connector *connector)
 	if (!panel->backlight.max)
 		return -ENODEV;
 
+	panel->backlight.min = get_backlight_min_vbt(connector);
+
 	val = pch_get_backlight(connector);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
@@ -1035,6 +1194,8 @@ static int i9xx_setup_backlight(struct intel_connector *connector)
 	if (!panel->backlight.max)
 		return -ENODEV;
 
+	panel->backlight.min = get_backlight_min_vbt(connector);
+
 	val = i9xx_get_backlight(connector);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
@@ -1062,6 +1223,8 @@ static int i965_setup_backlight(struct intel_connector *connector)
 	if (!panel->backlight.max)
 		return -ENODEV;
 
+	panel->backlight.min = get_backlight_min_vbt(connector);
+
 	val = i9xx_get_backlight(connector);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
@@ -1079,7 +1242,7 @@ static int vlv_setup_backlight(struct intel_connector *connector)
 	enum pipe pipe;
 	u32 ctl, ctl2, val;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		u32 cur_val = I915_READ(VLV_BLC_PWM_CTL(pipe));
 
 		/* Skip if the modulation freq is already set */
@@ -1099,6 +1262,8 @@ static int vlv_setup_backlight(struct intel_connector *connector)
 	if (!panel->backlight.max)
 		return -ENODEV;
 
+	panel->backlight.min = get_backlight_min_vbt(connector);
+
 	val = _vlv_get_backlight(dev, PIPE_A);
 	panel->backlight.level = intel_panel_compute_brightness(connector, val);
 
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index f1233f544f3e..c27b6140bfd1 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -93,8 +93,7 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_framebuffer *fb = crtc->primary->fb;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int cfb_pitch;
 	int i;
@@ -150,8 +149,7 @@ static void g4x_enable_fbc(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_framebuffer *fb = crtc->primary->fb;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
@@ -222,16 +220,26 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_framebuffer *fb = crtc->primary->fb;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
 	dpfc_ctl = DPFC_CTL_PLANE(intel_crtc->plane);
 	if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
+		dev_priv->fbc.threshold++;
+
+	switch (dev_priv->fbc.threshold) {
+	case 4:
+	case 3:
+		dpfc_ctl |= DPFC_CTL_LIMIT_4X;
+		break;
+	case 2:
 		dpfc_ctl |= DPFC_CTL_LIMIT_2X;
-	else
+		break;
+	case 1:
 		dpfc_ctl |= DPFC_CTL_LIMIT_1X;
+		break;
+	}
 	dpfc_ctl |= DPFC_CTL_FENCE_EN;
 	if (IS_GEN5(dev))
 		dpfc_ctl |= obj->fence_reg;
@@ -278,18 +286,32 @@ static void gen7_enable_fbc(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_framebuffer *fb = crtc->primary->fb;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	u32 dpfc_ctl;
 
 	dpfc_ctl = IVB_DPFC_CTL_PLANE(intel_crtc->plane);
 	if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
+		dev_priv->fbc.threshold++;
+
+	switch (dev_priv->fbc.threshold) {
+	case 4:
+	case 3:
+		dpfc_ctl |= DPFC_CTL_LIMIT_4X;
+		break;
+	case 2:
 		dpfc_ctl |= DPFC_CTL_LIMIT_2X;
-	else
+		break;
+	case 1:
 		dpfc_ctl |= DPFC_CTL_LIMIT_1X;
+		break;
+	}
+
 	dpfc_ctl |= IVB_DPFC_CTL_FENCE_EN;
 
+	if (dev_priv->fbc.false_color)
+		dpfc_ctl |= FBC_CTL_FALSE_COLOR;
+
 	I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN);
 
 	if (IS_IVYBRIDGE(dev)) {
@@ -323,6 +345,16 @@ bool intel_fbc_enabled(struct drm_device *dev)
 	return dev_priv->display.fbc_enabled(dev);
 }
 
+void gen8_fbc_sw_flush(struct drm_device *dev, u32 value)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!IS_GEN8(dev))
+		return;
+
+	I915_WRITE(MSG_FBC_REND_STATE, value);
+}
+
 static void intel_fbc_work_fn(struct work_struct *__work)
 {
 	struct intel_fbc_work *work =
@@ -462,7 +494,6 @@ void intel_update_fbc(struct drm_device *dev)
 	struct drm_crtc *crtc = NULL, *tmp_crtc;
 	struct intel_crtc *intel_crtc;
 	struct drm_framebuffer *fb;
-	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
 	const struct drm_display_mode *adjusted_mode;
 	unsigned int max_width, max_height;
@@ -507,8 +538,7 @@ void intel_update_fbc(struct drm_device *dev)
 
 	intel_crtc = to_intel_crtc(crtc);
 	fb = crtc->primary->fb;
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
+	obj = intel_fb_obj(fb);
 	adjusted_mode = &intel_crtc->config.adjusted_mode;
 
 	if (i915.enable_fbc < 0) {
@@ -529,7 +559,10 @@ void intel_update_fbc(struct drm_device *dev)
 		goto out_disable;
 	}
 
-	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
+	if (INTEL_INFO(dev)->gen >= 8 || IS_HASWELL(dev)) {
+		max_width = 4096;
+		max_height = 4096;
+	} else if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
 		max_width = 4096;
 		max_height = 2048;
 	} else {
@@ -558,12 +591,19 @@ void intel_update_fbc(struct drm_device *dev)
 			DRM_DEBUG_KMS("framebuffer not tiled or fenced, disabling compression\n");
 		goto out_disable;
 	}
+	if (INTEL_INFO(dev)->gen <= 4 && !IS_G4X(dev) &&
+	    to_intel_plane(crtc->primary)->rotation != BIT(DRM_ROTATE_0)) {
+		if (set_no_fbc_reason(dev_priv, FBC_UNSUPPORTED_MODE))
+			DRM_DEBUG_KMS("Rotation unsupported, disabling\n");
+		goto out_disable;
+	}
 
 	/* If the kernel debugger is active, always disable compression */
 	if (in_dbg_master())
 		goto out_disable;
 
-	if (i915_gem_stolen_setup_compression(dev, intel_fb->obj->base.size)) {
+	if (i915_gem_stolen_setup_compression(dev, obj->base.size,
+					      drm_format_plane_cpp(fb->pixel_format, 0))) {
 		if (set_no_fbc_reason(dev_priv, FBC_STOLEN_TOO_SMALL))
 			DRM_DEBUG_KMS("framebuffer too large, disabling compression\n");
 		goto out_disable;
@@ -789,12 +829,33 @@ static const struct cxsr_latency *intel_get_cxsr_latency(int is_desktop,
 	return NULL;
 }
 
-static void pineview_disable_cxsr(struct drm_device *dev)
+void intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_device *dev = dev_priv->dev;
+	u32 val;
 
-	/* deactivate cxsr */
-	I915_WRITE(DSPFW3, I915_READ(DSPFW3) & ~PINEVIEW_SELF_REFRESH_EN);
+	if (IS_VALLEYVIEW(dev)) {
+		I915_WRITE(FW_BLC_SELF_VLV, enable ? FW_CSPWRDWNEN : 0);
+	} else if (IS_G4X(dev) || IS_CRESTLINE(dev)) {
+		I915_WRITE(FW_BLC_SELF, enable ? FW_BLC_SELF_EN : 0);
+	} else if (IS_PINEVIEW(dev)) {
+		val = I915_READ(DSPFW3) & ~PINEVIEW_SELF_REFRESH_EN;
+		val |= enable ? PINEVIEW_SELF_REFRESH_EN : 0;
+		I915_WRITE(DSPFW3, val);
+	} else if (IS_I945G(dev) || IS_I945GM(dev)) {
+		val = enable ? _MASKED_BIT_ENABLE(FW_BLC_SELF_EN) :
+			       _MASKED_BIT_DISABLE(FW_BLC_SELF_EN);
+		I915_WRITE(FW_BLC_SELF, val);
+	} else if (IS_I915GM(dev)) {
+		val = enable ? _MASKED_BIT_ENABLE(INSTPM_SELF_EN) :
+			       _MASKED_BIT_DISABLE(INSTPM_SELF_EN);
+		I915_WRITE(INSTPM, val);
+	} else {
+		return;
+	}
+
+	DRM_DEBUG_KMS("memory self-refresh is %s\n",
+		      enable ? "enabled" : "disabled");
 }
 
 /*
@@ -811,7 +872,7 @@ static void pineview_disable_cxsr(struct drm_device *dev)
  * A value of 5us seems to be a good balance; safe for very low end
  * platforms but not overly aggressive on lower latency configs.
  */
-static const int latency_ns = 5000;
+static const int pessimal_latency_ns = 5000;
 
 static int i9xx_get_fifo_size(struct drm_device *dev, int plane)
 {
@@ -864,95 +925,102 @@ static int i845_get_fifo_size(struct drm_device *dev, int plane)
 
 /* Pineview has different values for various configs */
 static const struct intel_watermark_params pineview_display_wm = {
-	PINEVIEW_DISPLAY_FIFO,
-	PINEVIEW_MAX_WM,
-	PINEVIEW_DFT_WM,
-	PINEVIEW_GUARD_WM,
-	PINEVIEW_FIFO_LINE_SIZE
+	.fifo_size = PINEVIEW_DISPLAY_FIFO,
+	.max_wm = PINEVIEW_MAX_WM,
+	.default_wm = PINEVIEW_DFT_WM,
+	.guard_size = PINEVIEW_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params pineview_display_hplloff_wm = {
-	PINEVIEW_DISPLAY_FIFO,
-	PINEVIEW_MAX_WM,
-	PINEVIEW_DFT_HPLLOFF_WM,
-	PINEVIEW_GUARD_WM,
-	PINEVIEW_FIFO_LINE_SIZE
+	.fifo_size = PINEVIEW_DISPLAY_FIFO,
+	.max_wm = PINEVIEW_MAX_WM,
+	.default_wm = PINEVIEW_DFT_HPLLOFF_WM,
+	.guard_size = PINEVIEW_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params pineview_cursor_wm = {
-	PINEVIEW_CURSOR_FIFO,
-	PINEVIEW_CURSOR_MAX_WM,
-	PINEVIEW_CURSOR_DFT_WM,
-	PINEVIEW_CURSOR_GUARD_WM,
-	PINEVIEW_FIFO_LINE_SIZE,
+	.fifo_size = PINEVIEW_CURSOR_FIFO,
+	.max_wm = PINEVIEW_CURSOR_MAX_WM,
+	.default_wm = PINEVIEW_CURSOR_DFT_WM,
+	.guard_size = PINEVIEW_CURSOR_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params pineview_cursor_hplloff_wm = {
-	PINEVIEW_CURSOR_FIFO,
-	PINEVIEW_CURSOR_MAX_WM,
-	PINEVIEW_CURSOR_DFT_WM,
-	PINEVIEW_CURSOR_GUARD_WM,
-	PINEVIEW_FIFO_LINE_SIZE
+	.fifo_size = PINEVIEW_CURSOR_FIFO,
+	.max_wm = PINEVIEW_CURSOR_MAX_WM,
+	.default_wm = PINEVIEW_CURSOR_DFT_WM,
+	.guard_size = PINEVIEW_CURSOR_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params g4x_wm_info = {
-	G4X_FIFO_SIZE,
-	G4X_MAX_WM,
-	G4X_MAX_WM,
-	2,
-	G4X_FIFO_LINE_SIZE,
+	.fifo_size = G4X_FIFO_SIZE,
+	.max_wm = G4X_MAX_WM,
+	.default_wm = G4X_MAX_WM,
+	.guard_size = 2,
+	.cacheline_size = G4X_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params g4x_cursor_wm_info = {
-	I965_CURSOR_FIFO,
-	I965_CURSOR_MAX_WM,
-	I965_CURSOR_DFT_WM,
-	2,
-	G4X_FIFO_LINE_SIZE,
+	.fifo_size = I965_CURSOR_FIFO,
+	.max_wm = I965_CURSOR_MAX_WM,
+	.default_wm = I965_CURSOR_DFT_WM,
+	.guard_size = 2,
+	.cacheline_size = G4X_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params valleyview_wm_info = {
-	VALLEYVIEW_FIFO_SIZE,
-	VALLEYVIEW_MAX_WM,
-	VALLEYVIEW_MAX_WM,
-	2,
-	G4X_FIFO_LINE_SIZE,
+	.fifo_size = VALLEYVIEW_FIFO_SIZE,
+	.max_wm = VALLEYVIEW_MAX_WM,
+	.default_wm = VALLEYVIEW_MAX_WM,
+	.guard_size = 2,
+	.cacheline_size = G4X_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params valleyview_cursor_wm_info = {
-	I965_CURSOR_FIFO,
-	VALLEYVIEW_CURSOR_MAX_WM,
-	I965_CURSOR_DFT_WM,
-	2,
-	G4X_FIFO_LINE_SIZE,
+	.fifo_size = I965_CURSOR_FIFO,
+	.max_wm = VALLEYVIEW_CURSOR_MAX_WM,
+	.default_wm = I965_CURSOR_DFT_WM,
+	.guard_size = 2,
+	.cacheline_size = G4X_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params i965_cursor_wm_info = {
-	I965_CURSOR_FIFO,
-	I965_CURSOR_MAX_WM,
-	I965_CURSOR_DFT_WM,
-	2,
-	I915_FIFO_LINE_SIZE,
+	.fifo_size = I965_CURSOR_FIFO,
+	.max_wm = I965_CURSOR_MAX_WM,
+	.default_wm = I965_CURSOR_DFT_WM,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params i945_wm_info = {
-	I945_FIFO_SIZE,
-	I915_MAX_WM,
-	1,
-	2,
-	I915_FIFO_LINE_SIZE
+	.fifo_size = I945_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params i915_wm_info = {
-	I915_FIFO_SIZE,
-	I915_MAX_WM,
-	1,
-	2,
-	I915_FIFO_LINE_SIZE
+	.fifo_size = I915_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
 };
-static const struct intel_watermark_params i830_wm_info = {
-	I855GM_FIFO_SIZE,
-	I915_MAX_WM,
-	1,
-	2,
-	I830_FIFO_LINE_SIZE
+static const struct intel_watermark_params i830_a_wm_info = {
+	.fifo_size = I855GM_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
+};
+static const struct intel_watermark_params i830_bc_wm_info = {
+	.fifo_size = I855GM_FIFO_SIZE,
+	.max_wm = I915_MAX_WM/2,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
 };
 static const struct intel_watermark_params i845_wm_info = {
-	I830_FIFO_SIZE,
-	I915_MAX_WM,
-	1,
-	2,
-	I830_FIFO_LINE_SIZE
+	.fifo_size = I830_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
 };
 
 /**
@@ -1002,6 +1070,17 @@ static unsigned long intel_calculate_wm(unsigned long clock_in_khz,
 		wm_size = wm->max_wm;
 	if (wm_size <= 0)
 		wm_size = wm->default_wm;
+
+	/*
+	 * Bspec seems to indicate that the value shouldn't be lower than
+	 * 'burst size + 1'. Certainly 830 is quite unhappy with low values.
+	 * Lets go for 8 which is the burst size since certain platforms
+	 * already use a hardcoded 8 (which is what the spec says should be
+	 * done).
+	 */
+	if (wm_size <= 8)
+		wm_size = 8;
+
 	return wm_size;
 }
 
@@ -1033,7 +1112,7 @@ static void pineview_update_wm(struct drm_crtc *unused_crtc)
 					 dev_priv->fsb_freq, dev_priv->mem_freq);
 	if (!latency) {
 		DRM_DEBUG_KMS("Unknown FSB/MEM found, disable CxSR\n");
-		pineview_disable_cxsr(dev);
+		intel_set_memory_cxsr(dev_priv, false);
 		return;
 	}
 
@@ -1084,13 +1163,9 @@ static void pineview_update_wm(struct drm_crtc *unused_crtc)
 		I915_WRITE(DSPFW3, reg);
 		DRM_DEBUG_KMS("DSPFW3 register is %x\n", reg);
 
-		/* activate cxsr */
-		I915_WRITE(DSPFW3,
-			   I915_READ(DSPFW3) | PINEVIEW_SELF_REFRESH_EN);
-		DRM_DEBUG_KMS("Self-refresh is enabled\n");
+		intel_set_memory_cxsr(dev_priv, true);
 	} else {
-		pineview_disable_cxsr(dev);
-		DRM_DEBUG_KMS("Self-refresh is disabled\n");
+		intel_set_memory_cxsr(dev_priv, false);
 	}
 }
 
@@ -1230,34 +1305,27 @@ static bool g4x_compute_srwm(struct drm_device *dev,
 			      display, cursor);
 }
 
-static bool vlv_compute_drain_latency(struct drm_device *dev,
-				     int plane,
-				     int *plane_prec_mult,
-				     int *plane_dl,
-				     int *cursor_prec_mult,
-				     int *cursor_dl)
+static bool vlv_compute_drain_latency(struct drm_crtc *crtc,
+				      int pixel_size,
+				      int *prec_mult,
+				      int *drain_latency)
 {
-	struct drm_crtc *crtc;
-	int clock, pixel_size;
 	int entries;
+	int clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock;
 
-	crtc = intel_get_crtc_for_plane(dev, plane);
-	if (!intel_crtc_active(crtc))
+	if (WARN(clock == 0, "Pixel clock is zero!\n"))
 		return false;
 
-	clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock;
-	pixel_size = crtc->primary->fb->bits_per_pixel / 8;	/* BPP */
+	if (WARN(pixel_size == 0, "Pixel size is zero!\n"))
+		return false;
 
-	entries = (clock / 1000) * pixel_size;
-	*plane_prec_mult = (entries > 256) ?
-		DRAIN_LATENCY_PRECISION_32 : DRAIN_LATENCY_PRECISION_16;
-	*plane_dl = (64 * (*plane_prec_mult) * 4) / ((clock / 1000) *
-						     pixel_size);
+	entries = DIV_ROUND_UP(clock, 1000) * pixel_size;
+	*prec_mult = (entries > 128) ? DRAIN_LATENCY_PRECISION_64 :
+				       DRAIN_LATENCY_PRECISION_32;
+	*drain_latency = (64 * (*prec_mult) * 4) / entries;
 
-	entries = (clock / 1000) * 4;	/* BPP is always 4 for cursor */
-	*cursor_prec_mult = (entries > 256) ?
-		DRAIN_LATENCY_PRECISION_32 : DRAIN_LATENCY_PRECISION_16;
-	*cursor_dl = (64 * (*cursor_prec_mult) * 4) / ((clock / 1000) * 4);
+	if (*drain_latency > DRAIN_LATENCY_MASK)
+		*drain_latency = DRAIN_LATENCY_MASK;
 
 	return true;
 }
@@ -1270,39 +1338,48 @@ static bool vlv_compute_drain_latency(struct drm_device *dev,
  * latency value.
  */
 
-static void vlv_update_drain_latency(struct drm_device *dev)
+static void vlv_update_drain_latency(struct drm_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int planea_prec, planea_dl, planeb_prec, planeb_dl;
-	int cursora_prec, cursora_dl, cursorb_prec, cursorb_dl;
-	int plane_prec_mult, cursor_prec_mult; /* Precision multiplier is
-							either 16 or 32 */
+	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pixel_size;
+	int drain_latency;
+	enum pipe pipe = intel_crtc->pipe;
+	int plane_prec, prec_mult, plane_dl;
 
-	/* For plane A, Cursor A */
-	if (vlv_compute_drain_latency(dev, 0, &plane_prec_mult, &planea_dl,
-				      &cursor_prec_mult, &cursora_dl)) {
-		cursora_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_CURSORA_PRECISION_32 : DDL_CURSORA_PRECISION_16;
-		planea_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_PLANEA_PRECISION_32 : DDL_PLANEA_PRECISION_16;
+	plane_dl = I915_READ(VLV_DDL(pipe)) & ~(DDL_PLANE_PRECISION_64 |
+		   DRAIN_LATENCY_MASK | DDL_CURSOR_PRECISION_64 |
+		   (DRAIN_LATENCY_MASK << DDL_CURSOR_SHIFT));
+
+	if (!intel_crtc_active(crtc)) {
+		I915_WRITE(VLV_DDL(pipe), plane_dl);
+		return;
+	}
 
-		I915_WRITE(VLV_DDL1, cursora_prec |
-				(cursora_dl << DDL_CURSORA_SHIFT) |
-				planea_prec | planea_dl);
+	/* Primary plane Drain Latency */
+	pixel_size = crtc->primary->fb->bits_per_pixel / 8;	/* BPP */
+	if (vlv_compute_drain_latency(crtc, pixel_size, &prec_mult, &drain_latency)) {
+		plane_prec = (prec_mult == DRAIN_LATENCY_PRECISION_64) ?
+					   DDL_PLANE_PRECISION_64 :
+					   DDL_PLANE_PRECISION_32;
+		plane_dl |= plane_prec | drain_latency;
 	}
 
-	/* For plane B, Cursor B */
-	if (vlv_compute_drain_latency(dev, 1, &plane_prec_mult, &planeb_dl,
-				      &cursor_prec_mult, &cursorb_dl)) {
-		cursorb_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_CURSORB_PRECISION_32 : DDL_CURSORB_PRECISION_16;
-		planeb_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
-			DDL_PLANEB_PRECISION_32 : DDL_PLANEB_PRECISION_16;
+	/* Cursor Drain Latency
+	 * BPP is always 4 for cursor
+	 */
+	pixel_size = 4;
 
-		I915_WRITE(VLV_DDL2, cursorb_prec |
-				(cursorb_dl << DDL_CURSORB_SHIFT) |
-				planeb_prec | planeb_dl);
+	/* Program cursor DL only if it is enabled */
+	if (intel_crtc->cursor_base &&
+	    vlv_compute_drain_latency(crtc, pixel_size, &prec_mult, &drain_latency)) {
+		plane_prec = (prec_mult == DRAIN_LATENCY_PRECISION_64) ?
+					   DDL_CURSOR_PRECISION_64 :
+					   DDL_CURSOR_PRECISION_32;
+		plane_dl |= plane_prec | (drain_latency << DDL_CURSOR_SHIFT);
 	}
+
+	I915_WRITE(VLV_DDL(pipe), plane_dl);
 }
 
 #define single_plane_enabled(mask) is_power_of_2(mask)
@@ -1316,21 +1393,94 @@ static void valleyview_update_wm(struct drm_crtc *crtc)
 	int plane_sr, cursor_sr;
 	int ignore_plane_sr, ignore_cursor_sr;
 	unsigned int enabled = 0;
+	bool cxsr_enabled;
+
+	vlv_update_drain_latency(crtc);
+
+	if (g4x_compute_wm0(dev, PIPE_A,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planea_wm, &cursora_wm))
+		enabled |= 1 << PIPE_A;
+
+	if (g4x_compute_wm0(dev, PIPE_B,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planeb_wm, &cursorb_wm))
+		enabled |= 1 << PIPE_B;
+
+	if (single_plane_enabled(enabled) &&
+	    g4x_compute_srwm(dev, ffs(enabled) - 1,
+			     sr_latency_ns,
+			     &valleyview_wm_info,
+			     &valleyview_cursor_wm_info,
+			     &plane_sr, &ignore_cursor_sr) &&
+	    g4x_compute_srwm(dev, ffs(enabled) - 1,
+			     2*sr_latency_ns,
+			     &valleyview_wm_info,
+			     &valleyview_cursor_wm_info,
+			     &ignore_plane_sr, &cursor_sr)) {
+		cxsr_enabled = true;
+	} else {
+		cxsr_enabled = false;
+		intel_set_memory_cxsr(dev_priv, false);
+		plane_sr = cursor_sr = 0;
+	}
+
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+		      planea_wm, cursora_wm,
+		      planeb_wm, cursorb_wm,
+		      plane_sr, cursor_sr);
+
+	I915_WRITE(DSPFW1,
+		   (plane_sr << DSPFW_SR_SHIFT) |
+		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
+		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
+	I915_WRITE(DSPFW2,
+		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
+		   (cursora_wm << DSPFW_CURSORA_SHIFT));
+	I915_WRITE(DSPFW3,
+		   (I915_READ(DSPFW3) & ~DSPFW_CURSOR_SR_MASK) |
+		   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
+}
+
+static void cherryview_update_wm(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	static const int sr_latency_ns = 12000;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int planea_wm, planeb_wm, planec_wm;
+	int cursora_wm, cursorb_wm, cursorc_wm;
+	int plane_sr, cursor_sr;
+	int ignore_plane_sr, ignore_cursor_sr;
+	unsigned int enabled = 0;
+	bool cxsr_enabled;
 
-	vlv_update_drain_latency(dev);
+	vlv_update_drain_latency(crtc);
 
 	if (g4x_compute_wm0(dev, PIPE_A,
-			    &valleyview_wm_info, latency_ns,
-			    &valleyview_cursor_wm_info, latency_ns,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
 			    &planea_wm, &cursora_wm))
 		enabled |= 1 << PIPE_A;
 
 	if (g4x_compute_wm0(dev, PIPE_B,
-			    &valleyview_wm_info, latency_ns,
-			    &valleyview_cursor_wm_info, latency_ns,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
 			    &planeb_wm, &cursorb_wm))
 		enabled |= 1 << PIPE_B;
 
+	if (g4x_compute_wm0(dev, PIPE_C,
+			    &valleyview_wm_info, pessimal_latency_ns,
+			    &valleyview_cursor_wm_info, pessimal_latency_ns,
+			    &planec_wm, &cursorc_wm))
+		enabled |= 1 << PIPE_C;
+
 	if (single_plane_enabled(enabled) &&
 	    g4x_compute_srwm(dev, ffs(enabled) - 1,
 			     sr_latency_ns,
@@ -1342,29 +1492,71 @@ static void valleyview_update_wm(struct drm_crtc *crtc)
 			     &valleyview_wm_info,
 			     &valleyview_cursor_wm_info,
 			     &ignore_plane_sr, &cursor_sr)) {
-		I915_WRITE(FW_BLC_SELF_VLV, FW_CSPWRDWNEN);
+		cxsr_enabled = true;
 	} else {
-		I915_WRITE(FW_BLC_SELF_VLV,
-			   I915_READ(FW_BLC_SELF_VLV) & ~FW_CSPWRDWNEN);
+		cxsr_enabled = false;
+		intel_set_memory_cxsr(dev_priv, false);
 		plane_sr = cursor_sr = 0;
 	}
 
-	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, C: plane=%d, cursor=%d, "
+		      "SR: plane=%d, cursor=%d\n",
 		      planea_wm, cursora_wm,
 		      planeb_wm, cursorb_wm,
+		      planec_wm, cursorc_wm,
 		      plane_sr, cursor_sr);
 
 	I915_WRITE(DSPFW1,
 		   (plane_sr << DSPFW_SR_SHIFT) |
 		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
 		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
-		   planea_wm);
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
 	I915_WRITE(DSPFW2,
 		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
 		   (cursora_wm << DSPFW_CURSORA_SHIFT));
 	I915_WRITE(DSPFW3,
 		   (I915_READ(DSPFW3) & ~DSPFW_CURSOR_SR_MASK) |
 		   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+	I915_WRITE(DSPFW9_CHV,
+		   (I915_READ(DSPFW9_CHV) & ~(DSPFW_PLANEC_MASK |
+					      DSPFW_CURSORC_MASK)) |
+		   (planec_wm << DSPFW_PLANEC_SHIFT) |
+		   (cursorc_wm << DSPFW_CURSORC_SHIFT));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
+}
+
+static void valleyview_update_sprite_wm(struct drm_plane *plane,
+					struct drm_crtc *crtc,
+					uint32_t sprite_width,
+					uint32_t sprite_height,
+					int pixel_size,
+					bool enabled, bool scaled)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = to_intel_plane(plane)->pipe;
+	int sprite = to_intel_plane(plane)->plane;
+	int drain_latency;
+	int plane_prec;
+	int sprite_dl;
+	int prec_mult;
+
+	sprite_dl = I915_READ(VLV_DDL(pipe)) & ~(DDL_SPRITE_PRECISION_64(sprite) |
+		    (DRAIN_LATENCY_MASK << DDL_SPRITE_SHIFT(sprite)));
+
+	if (enabled && vlv_compute_drain_latency(crtc, pixel_size, &prec_mult,
+						 &drain_latency)) {
+		plane_prec = (prec_mult == DRAIN_LATENCY_PRECISION_64) ?
+					   DDL_SPRITE_PRECISION_64(sprite) :
+					   DDL_SPRITE_PRECISION_32(sprite);
+		sprite_dl |= plane_prec |
+			     (drain_latency << DDL_SPRITE_SHIFT(sprite));
+	}
+
+	I915_WRITE(VLV_DDL(pipe), sprite_dl);
 }
 
 static void g4x_update_wm(struct drm_crtc *crtc)
@@ -1375,16 +1567,17 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 	int planea_wm, planeb_wm, cursora_wm, cursorb_wm;
 	int plane_sr, cursor_sr;
 	unsigned int enabled = 0;
+	bool cxsr_enabled;
 
 	if (g4x_compute_wm0(dev, PIPE_A,
-			    &g4x_wm_info, latency_ns,
-			    &g4x_cursor_wm_info, latency_ns,
+			    &g4x_wm_info, pessimal_latency_ns,
+			    &g4x_cursor_wm_info, pessimal_latency_ns,
 			    &planea_wm, &cursora_wm))
 		enabled |= 1 << PIPE_A;
 
 	if (g4x_compute_wm0(dev, PIPE_B,
-			    &g4x_wm_info, latency_ns,
-			    &g4x_cursor_wm_info, latency_ns,
+			    &g4x_wm_info, pessimal_latency_ns,
+			    &g4x_cursor_wm_info, pessimal_latency_ns,
 			    &planeb_wm, &cursorb_wm))
 		enabled |= 1 << PIPE_B;
 
@@ -1394,14 +1587,15 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 			     &g4x_wm_info,
 			     &g4x_cursor_wm_info,
 			     &plane_sr, &cursor_sr)) {
-		I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN);
+		cxsr_enabled = true;
 	} else {
-		I915_WRITE(FW_BLC_SELF,
-			   I915_READ(FW_BLC_SELF) & ~FW_BLC_SELF_EN);
+		cxsr_enabled = false;
+		intel_set_memory_cxsr(dev_priv, false);
 		plane_sr = cursor_sr = 0;
 	}
 
-	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
+	DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, "
+		      "B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
 		      planea_wm, cursora_wm,
 		      planeb_wm, cursorb_wm,
 		      plane_sr, cursor_sr);
@@ -1410,7 +1604,7 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 		   (plane_sr << DSPFW_SR_SHIFT) |
 		   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
 		   (planeb_wm << DSPFW_PLANEB_SHIFT) |
-		   planea_wm);
+		   (planea_wm << DSPFW_PLANEA_SHIFT));
 	I915_WRITE(DSPFW2,
 		   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
 		   (cursora_wm << DSPFW_CURSORA_SHIFT));
@@ -1418,6 +1612,9 @@ static void g4x_update_wm(struct drm_crtc *crtc)
 	I915_WRITE(DSPFW3,
 		   (I915_READ(DSPFW3) & ~(DSPFW_HPLL_SR_EN | DSPFW_CURSOR_SR_MASK)) |
 		   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
 }
 
 static void i965_update_wm(struct drm_crtc *unused_crtc)
@@ -1427,6 +1624,7 @@ static void i965_update_wm(struct drm_crtc *unused_crtc)
 	struct drm_crtc *crtc;
 	int srwm = 1;
 	int cursor_sr = 16;
+	bool cxsr_enabled;
 
 	/* Calc sr entries for one plane configs */
 	crtc = single_enabled_crtc(dev);
@@ -1468,13 +1666,11 @@ static void i965_update_wm(struct drm_crtc *unused_crtc)
 		DRM_DEBUG_KMS("self-refresh watermark: display plane %d "
 			      "cursor %d\n", srwm, cursor_sr);
 
-		if (IS_CRESTLINE(dev))
-			I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN);
+		cxsr_enabled = true;
 	} else {
+		cxsr_enabled = false;
 		/* Turn off self refresh if both pipes are enabled */
-		if (IS_CRESTLINE(dev))
-			I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF)
-				   & ~FW_BLC_SELF_EN);
+		intel_set_memory_cxsr(dev_priv, false);
 	}
 
 	DRM_DEBUG_KMS("Setting FIFO watermarks - A: 8, B: 8, C: 8, SR %d\n",
@@ -1482,10 +1678,16 @@ static void i965_update_wm(struct drm_crtc *unused_crtc)
 
 	/* 965 has limitations... */
 	I915_WRITE(DSPFW1, (srwm << DSPFW_SR_SHIFT) |
-		   (8 << 16) | (8 << 8) | (8 << 0));
-	I915_WRITE(DSPFW2, (8 << 8) | (8 << 0));
+		   (8 << DSPFW_CURSORB_SHIFT) |
+		   (8 << DSPFW_PLANEB_SHIFT) |
+		   (8 << DSPFW_PLANEA_SHIFT));
+	I915_WRITE(DSPFW2, (8 << DSPFW_CURSORA_SHIFT) |
+		   (8 << DSPFW_PLANEC_SHIFT_OLD));
 	/* update cursor SR watermark */
 	I915_WRITE(DSPFW3, (cursor_sr << DSPFW_CURSOR_SR_SHIFT));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
 }
 
 static void i9xx_update_wm(struct drm_crtc *unused_crtc)
@@ -1505,7 +1707,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 	else if (!IS_GEN2(dev))
 		wm_info = &i915_wm_info;
 	else
-		wm_info = &i830_wm_info;
+		wm_info = &i830_a_wm_info;
 
 	fifo_size = dev_priv->display.get_fifo_size(dev, 0);
 	crtc = intel_get_crtc_for_plane(dev, 0);
@@ -1518,10 +1720,16 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 		adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
 		planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 					       wm_info, fifo_size, cpp,
-					       latency_ns);
+					       pessimal_latency_ns);
 		enabled = crtc;
-	} else
+	} else {
 		planea_wm = fifo_size - wm_info->guard_size;
+		if (planea_wm > (long)wm_info->max_wm)
+			planea_wm = wm_info->max_wm;
+	}
+
+	if (IS_GEN2(dev))
+		wm_info = &i830_bc_wm_info;
 
 	fifo_size = dev_priv->display.get_fifo_size(dev, 1);
 	crtc = intel_get_crtc_for_plane(dev, 1);
@@ -1534,23 +1742,26 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 		adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
 		planeb_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 					       wm_info, fifo_size, cpp,
-					       latency_ns);
+					       pessimal_latency_ns);
 		if (enabled == NULL)
 			enabled = crtc;
 		else
 			enabled = NULL;
-	} else
+	} else {
 		planeb_wm = fifo_size - wm_info->guard_size;
+		if (planeb_wm > (long)wm_info->max_wm)
+			planeb_wm = wm_info->max_wm;
+	}
 
 	DRM_DEBUG_KMS("FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm);
 
 	if (IS_I915GM(dev) && enabled) {
-		struct intel_framebuffer *fb;
+		struct drm_i915_gem_object *obj;
 
-		fb = to_intel_framebuffer(enabled->primary->fb);
+		obj = intel_fb_obj(enabled->primary->fb);
 
 		/* self-refresh seems busted with untiled */
-		if (fb->obj->tiling_mode == I915_TILING_NONE)
+		if (obj->tiling_mode == I915_TILING_NONE)
 			enabled = NULL;
 	}
 
@@ -1560,10 +1771,7 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 	cwm = 2;
 
 	/* Play safe and disable self-refresh before adjusting watermarks. */
-	if (IS_I945G(dev) || IS_I945GM(dev))
-		I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN_MASK | 0);
-	else if (IS_I915GM(dev))
-		I915_WRITE(INSTPM, _MASKED_BIT_DISABLE(INSTPM_SELF_EN));
+	intel_set_memory_cxsr(dev_priv, false);
 
 	/* Calc sr entries for one plane configs */
 	if (HAS_FW_BLC(dev) && enabled) {
@@ -1609,17 +1817,8 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc)
 	I915_WRITE(FW_BLC, fwater_lo);
 	I915_WRITE(FW_BLC2, fwater_hi);
 
-	if (HAS_FW_BLC(dev)) {
-		if (enabled) {
-			if (IS_I945G(dev) || IS_I945GM(dev))
-				I915_WRITE(FW_BLC_SELF,
-					   FW_BLC_SELF_EN_MASK | FW_BLC_SELF_EN);
-			else if (IS_I915GM(dev))
-				I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_SELF_EN));
-			DRM_DEBUG_KMS("memory self refresh enabled\n");
-		} else
-			DRM_DEBUG_KMS("memory self refresh disabled\n");
-	}
+	if (enabled)
+		intel_set_memory_cxsr(dev_priv, true);
 }
 
 static void i845_update_wm(struct drm_crtc *unused_crtc)
@@ -1639,7 +1838,7 @@ static void i845_update_wm(struct drm_crtc *unused_crtc)
 	planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
 				       &i845_wm_info,
 				       dev_priv->display.get_fifo_size(dev, 0),
-				       4, latency_ns);
+				       4, pessimal_latency_ns);
 	fwater_lo = I915_READ(FW_BLC) & ~0xfff;
 	fwater_lo |= (3<<8) | planea_wm;
 
@@ -2492,7 +2691,7 @@ static struct intel_pipe_wm *ilk_find_best_result(struct drm_device *dev,
 #define WM_DIRTY_FBC (1 << 24)
 #define WM_DIRTY_DDB (1 << 25)
 
-static unsigned int ilk_compute_wm_dirty(struct drm_device *dev,
+static unsigned int ilk_compute_wm_dirty(struct drm_i915_private *dev_priv,
 					 const struct ilk_wm_values *old,
 					 const struct ilk_wm_values *new)
 {
@@ -2500,7 +2699,7 @@ static unsigned int ilk_compute_wm_dirty(struct drm_device *dev,
 	enum pipe pipe;
 	int wm_lp;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		if (old->wm_linetime[pipe] != new->wm_linetime[pipe]) {
 			dirty |= WM_DIRTY_LINETIME(pipe);
 			/* Must disable LP1+ watermarks too */
@@ -2586,7 +2785,7 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 	unsigned int dirty;
 	uint32_t val;
 
-	dirty = ilk_compute_wm_dirty(dev, previous, results);
+	dirty = ilk_compute_wm_dirty(dev_priv, previous, results);
 	if (!dirty)
 		return;
 
@@ -2707,10 +2906,11 @@ static void ilk_update_wm(struct drm_crtc *crtc)
 	ilk_write_wm_values(dev_priv, &results);
 }
 
-static void ilk_update_sprite_wm(struct drm_plane *plane,
-				     struct drm_crtc *crtc,
-				     uint32_t sprite_width, int pixel_size,
-				     bool enabled, bool scaled)
+static void
+ilk_update_sprite_wm(struct drm_plane *plane,
+		     struct drm_crtc *crtc,
+		     uint32_t sprite_width, uint32_t sprite_height,
+		     int pixel_size, bool enabled, bool scaled)
 {
 	struct drm_device *dev = plane->dev;
 	struct intel_plane *intel_plane = to_intel_plane(plane);
@@ -2718,6 +2918,7 @@ static void ilk_update_sprite_wm(struct drm_plane *plane,
 	intel_plane->wm.enabled = enabled;
 	intel_plane->wm.scaled = scaled;
 	intel_plane->wm.horiz_pixels = sprite_width;
+	intel_plane->wm.vert_pixels = sprite_width;
 	intel_plane->wm.bytes_per_pixel = pixel_size;
 
 	/*
@@ -2852,13 +3053,16 @@ void intel_update_watermarks(struct drm_crtc *crtc)
 
 void intel_update_sprite_watermarks(struct drm_plane *plane,
 				    struct drm_crtc *crtc,
-				    uint32_t sprite_width, int pixel_size,
+				    uint32_t sprite_width,
+				    uint32_t sprite_height,
+				    int pixel_size,
 				    bool enabled, bool scaled)
 {
 	struct drm_i915_private *dev_priv = plane->dev->dev_private;
 
 	if (dev_priv->display.update_sprite_wm)
-		dev_priv->display.update_sprite_wm(plane, crtc, sprite_width,
+		dev_priv->display.update_sprite_wm(plane, crtc,
+						   sprite_width, sprite_height,
 						   pixel_size, enabled, scaled);
 }
 
@@ -3147,6 +3351,9 @@ static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val)
 	if (val < dev_priv->rps.max_freq_softlimit)
 		mask |= GEN6_PM_RP_UP_THRESHOLD;
 
+	mask |= dev_priv->pm_rps_events & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED);
+	mask &= dev_priv->pm_rps_events;
+
 	/* IVB and SNB hard hangs on looping batchbuffer
 	 * if GEN6_PM_UP_EI_EXPIRED is masked.
 	 */
@@ -3250,7 +3457,9 @@ void gen6_rps_idle(struct drm_i915_private *dev_priv)
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 	if (dev_priv->rps.enabled) {
-		if (IS_VALLEYVIEW(dev))
+		if (IS_CHERRYVIEW(dev))
+			valleyview_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
+		else if (IS_VALLEYVIEW(dev))
 			vlv_set_rps_idle(dev_priv);
 		else
 			gen6_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
@@ -3282,13 +3491,18 @@ void valleyview_set_rps(struct drm_device *dev, u8 val)
 	WARN_ON(val > dev_priv->rps.max_freq_softlimit);
 	WARN_ON(val < dev_priv->rps.min_freq_softlimit);
 
-	DRM_DEBUG_DRIVER("GPU freq request from %d MHz (%u) to %d MHz (%u)\n",
-			 vlv_gpu_freq(dev_priv, dev_priv->rps.cur_freq),
-			 dev_priv->rps.cur_freq,
-			 vlv_gpu_freq(dev_priv, val), val);
+	if (WARN_ONCE(IS_CHERRYVIEW(dev) && (val & 1),
+		      "Odd GPU freq value\n"))
+		val &= ~1;
+
+	if (val != dev_priv->rps.cur_freq) {
+		DRM_DEBUG_DRIVER("GPU freq request from %d MHz (%u) to %d MHz (%u)\n",
+				 vlv_gpu_freq(dev_priv, dev_priv->rps.cur_freq),
+				 dev_priv->rps.cur_freq,
+				 vlv_gpu_freq(dev_priv, val), val);
 
-	if (val != dev_priv->rps.cur_freq)
 		vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
+	}
 
 	I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val));
 
@@ -3348,12 +3562,27 @@ static void gen6_disable_rps(struct drm_device *dev)
 		gen6_disable_rps_interrupts(dev);
 }
 
+static void cherryview_disable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE(GEN6_RC_CONTROL, 0);
+
+	gen8_disable_rps_interrupts(dev);
+}
+
 static void valleyview_disable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* we're doing forcewake before Disabling RC6,
+	 * This what the BIOS expects when going into suspend */
+	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
+
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 
+	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
+
 	gen6_disable_rps_interrupts(dev);
 }
 
@@ -3365,10 +3594,10 @@ static void intel_print_rc6_info(struct drm_device *dev, u32 mode)
 		else
 			mode = 0;
 	}
-	DRM_INFO("Enabling RC6 states: RC6 %s, RC6p %s, RC6pp %s\n",
-		 (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off",
-		 (mode & GEN6_RC_CTL_RC6p_ENABLE) ? "on" : "off",
-		 (mode & GEN6_RC_CTL_RC6pp_ENABLE) ? "on" : "off");
+	DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s, RC6p %s, RC6pp %s\n",
+		      (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off",
+		      (mode & GEN6_RC_CTL_RC6p_ENABLE) ? "on" : "off",
+		      (mode & GEN6_RC_CTL_RC6pp_ENABLE) ? "on" : "off");
 }
 
 static int sanitize_rc6_option(const struct drm_device *dev, int enable_rc6)
@@ -3392,8 +3621,8 @@ static int sanitize_rc6_option(const struct drm_device *dev, int enable_rc6)
 			mask = INTEL_RC6_ENABLE;
 
 		if ((enable_rc6 & mask) != enable_rc6)
-			DRM_INFO("Adjusting RC6 mask to %d (requested %d, valid %d)\n",
-				 enable_rc6 & mask, enable_rc6, mask);
+			DRM_DEBUG_KMS("Adjusting RC6 mask to %d (requested %d, valid %d)\n",
+				      enable_rc6 & mask, enable_rc6, mask);
 
 		return enable_rc6 & mask;
 	}
@@ -3419,7 +3648,7 @@ static void gen8_enable_rps_interrupts(struct drm_device *dev)
 
 	spin_lock_irq(&dev_priv->irq_lock);
 	WARN_ON(dev_priv->rps.pm_iir);
-	bdw_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
+	gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 	I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events);
 	spin_unlock_irq(&dev_priv->irq_lock);
 }
@@ -3430,7 +3659,7 @@ static void gen6_enable_rps_interrupts(struct drm_device *dev)
 
 	spin_lock_irq(&dev_priv->irq_lock);
 	WARN_ON(dev_priv->rps.pm_iir);
-	snb_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
+	gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
 	I915_WRITE(GEN6_PMIIR, dev_priv->pm_rps_events);
 	spin_unlock_irq(&dev_priv->irq_lock);
 }
@@ -3483,15 +3712,23 @@ static void gen8_enable_rps(struct drm_device *dev)
 	for_each_ring(ring, dev_priv, unused)
 		I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
 	I915_WRITE(GEN6_RC_SLEEP, 0);
-	I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */
+	if (IS_BROADWELL(dev))
+		I915_WRITE(GEN6_RC6_THRESHOLD, 625); /* 800us/1.28 for TO */
+	else
+		I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */
 
 	/* 3: Enable RC6 */
 	if (intel_enable_rc6(dev) & INTEL_RC6_ENABLE)
 		rc6_mask = GEN6_RC_CTL_RC6_ENABLE;
 	intel_print_rc6_info(dev, rc6_mask);
-	I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
-				    GEN6_RC_CTL_EI_MODE(1) |
-				    rc6_mask);
+	if (IS_BROADWELL(dev))
+		I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
+				GEN7_RC_CTL_TO_MODE |
+				rc6_mask);
+	else
+		I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
+				GEN6_RC_CTL_EI_MODE(1) |
+				rc6_mask);
 
 	/* 4 Program defaults and thresholds for RPS*/
 	I915_WRITE(GEN6_RPNSWREQ,
@@ -3536,7 +3773,6 @@ static void gen6_enable_rps(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
 	u32 rp_state_cap;
-	u32 gt_perf_status;
 	u32 rc6vids, pcu_mbox = 0, rc6_mask = 0;
 	u32 gtfifodbg;
 	int rc6_mode;
@@ -3561,7 +3797,6 @@ static void gen6_enable_rps(struct drm_device *dev)
 	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
 
 	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 
 	parse_rp_state_cap(dev_priv, rp_state_cap);
 
@@ -3727,7 +3962,57 @@ void gen6_update_ring_freq(struct drm_device *dev)
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
-int valleyview_rps_max_freq(struct drm_i915_private *dev_priv)
+static int cherryview_rps_max_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rp0;
+
+	val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG);
+	rp0 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK;
+
+	return rp0;
+}
+
+static int cherryview_rps_rpe_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rpe;
+
+	val = vlv_punit_read(dev_priv, PUNIT_GPU_DUTYCYCLE_REG);
+	rpe = (val >> PUNIT_GPU_DUTYCYCLE_RPE_FREQ_SHIFT) & PUNIT_GPU_DUTYCYCLE_RPE_FREQ_MASK;
+
+	return rpe;
+}
+
+static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rp1;
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+	rp1 = (val >> PUNIT_GPU_STATUS_MAX_FREQ_SHIFT) & PUNIT_GPU_STATUS_MAX_FREQ_MASK;
+
+	return rp1;
+}
+
+static int cherryview_rps_min_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rpn;
+
+	val = vlv_punit_read(dev_priv, PUNIT_GPU_STATUS_REG);
+	rpn = (val >> PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT) & PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK;
+	return rpn;
+}
+
+static int valleyview_rps_guar_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rp1;
+
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FREQ_FUSE);
+
+	rp1 = (val & FB_GFX_FGUARANTEED_FREQ_FUSE_MASK) >> FB_GFX_FGUARANTEED_FREQ_FUSE_SHIFT;
+
+	return rp1;
+}
+
+static int valleyview_rps_max_freq(struct drm_i915_private *dev_priv)
 {
 	u32 val, rp0;
 
@@ -3752,7 +4037,7 @@ static int valleyview_rps_rpe_freq(struct drm_i915_private *dev_priv)
 	return rpe;
 }
 
-int valleyview_rps_min_freq(struct drm_i915_private *dev_priv)
+static int valleyview_rps_min_freq(struct drm_i915_private *dev_priv)
 {
 	return vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM) & 0xff;
 }
@@ -3766,6 +4051,35 @@ static void valleyview_check_pctx(struct drm_i915_private *dev_priv)
 			     dev_priv->vlv_pctx->stolen->start);
 }
 
+
+/* Check that the pcbr address is not empty. */
+static void cherryview_check_pctx(struct drm_i915_private *dev_priv)
+{
+	unsigned long pctx_addr = I915_READ(VLV_PCBR) & ~4095;
+
+	WARN_ON((pctx_addr >> VLV_PCBR_ADDR_SHIFT) == 0);
+}
+
+static void cherryview_setup_pctx(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long pctx_paddr, paddr;
+	struct i915_gtt *gtt = &dev_priv->gtt;
+	u32 pcbr;
+	int pctx_size = 32*1024;
+
+	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+
+	pcbr = I915_READ(VLV_PCBR);
+	if ((pcbr >> VLV_PCBR_ADDR_SHIFT) == 0) {
+		paddr = (dev_priv->mm.stolen_base +
+			 (gtt->stolen_size - pctx_size));
+
+		pctx_paddr = (paddr & (~4095));
+		I915_WRITE(VLV_PCBR, pctx_paddr);
+	}
+}
+
 static void valleyview_setup_pctx(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3824,11 +4138,27 @@ static void valleyview_cleanup_pctx(struct drm_device *dev)
 static void valleyview_init_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val;
 
 	valleyview_setup_pctx(dev);
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+	switch ((val >> 6) & 3) {
+	case 0:
+	case 1:
+		dev_priv->mem_freq = 800;
+		break;
+	case 2:
+		dev_priv->mem_freq = 1066;
+		break;
+	case 3:
+		dev_priv->mem_freq = 1333;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
+
 	dev_priv->rps.max_freq = valleyview_rps_max_freq(dev_priv);
 	dev_priv->rps.rp0_freq = dev_priv->rps.max_freq;
 	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
@@ -3840,6 +4170,11 @@ static void valleyview_init_gt_powersave(struct drm_device *dev)
 			 vlv_gpu_freq(dev_priv, dev_priv->rps.efficient_freq),
 			 dev_priv->rps.efficient_freq);
 
+	dev_priv->rps.rp1_freq = valleyview_rps_guar_freq(dev_priv);
+	DRM_DEBUG_DRIVER("RP1(Guar Freq) GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.rp1_freq),
+			 dev_priv->rps.rp1_freq);
+
 	dev_priv->rps.min_freq = valleyview_rps_min_freq(dev_priv);
 	DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n",
 			 vlv_gpu_freq(dev_priv, dev_priv->rps.min_freq),
@@ -3855,11 +4190,175 @@ static void valleyview_init_gt_powersave(struct drm_device *dev)
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
+static void cherryview_init_gt_powersave(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 val;
+
+	cherryview_setup_pctx(dev);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+	val = vlv_punit_read(dev_priv, CCK_FUSE_REG);
+	switch ((val >> 2) & 0x7) {
+	case 0:
+	case 1:
+		dev_priv->rps.cz_freq = 200;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 2:
+		dev_priv->rps.cz_freq = 267;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 3:
+		dev_priv->rps.cz_freq = 333;
+		dev_priv->mem_freq = 2000;
+		break;
+	case 4:
+		dev_priv->rps.cz_freq = 320;
+		dev_priv->mem_freq = 1600;
+		break;
+	case 5:
+		dev_priv->rps.cz_freq = 400;
+		dev_priv->mem_freq = 1600;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
+
+	dev_priv->rps.max_freq = cherryview_rps_max_freq(dev_priv);
+	dev_priv->rps.rp0_freq = dev_priv->rps.max_freq;
+	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.max_freq),
+			 dev_priv->rps.max_freq);
+
+	dev_priv->rps.efficient_freq = cherryview_rps_rpe_freq(dev_priv);
+	DRM_DEBUG_DRIVER("RPe GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.efficient_freq),
+			 dev_priv->rps.efficient_freq);
+
+	dev_priv->rps.rp1_freq = cherryview_rps_guar_freq(dev_priv);
+	DRM_DEBUG_DRIVER("RP1(Guar) GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.rp1_freq),
+			 dev_priv->rps.rp1_freq);
+
+	dev_priv->rps.min_freq = cherryview_rps_min_freq(dev_priv);
+	DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.min_freq),
+			 dev_priv->rps.min_freq);
+
+	WARN_ONCE((dev_priv->rps.max_freq |
+		   dev_priv->rps.efficient_freq |
+		   dev_priv->rps.rp1_freq |
+		   dev_priv->rps.min_freq) & 1,
+		  "Odd GPU freq values\n");
+
+	/* Preserve min/max settings in case of re-init */
+	if (dev_priv->rps.max_freq_softlimit == 0)
+		dev_priv->rps.max_freq_softlimit = dev_priv->rps.max_freq;
+
+	if (dev_priv->rps.min_freq_softlimit == 0)
+		dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq;
+
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
 static void valleyview_cleanup_gt_powersave(struct drm_device *dev)
 {
 	valleyview_cleanup_pctx(dev);
 }
 
+static void cherryview_enable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *ring;
+	u32 gtfifodbg, val, rc6_mode = 0, pcbr;
+	int i;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	gtfifodbg = I915_READ(GTFIFODBG);
+	if (gtfifodbg) {
+		DRM_DEBUG_DRIVER("GT fifo had a previous error %x\n",
+				 gtfifodbg);
+		I915_WRITE(GTFIFODBG, gtfifodbg);
+	}
+
+	cherryview_check_pctx(dev_priv);
+
+	/* 1a & 1b: Get forcewake during program sequence. Although the driver
+	 * hasn't enabled a state yet where we need forcewake, BIOS may have.*/
+	gen6_gt_force_wake_get(dev_priv, FORCEWAKE_ALL);
+
+	/* 2a: Program RC6 thresholds.*/
+	I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16);
+	I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+	I915_WRITE(GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+
+	for_each_ring(ring, dev_priv, i)
+		I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
+	I915_WRITE(GEN6_RC_SLEEP, 0);
+
+	I915_WRITE(GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */
+
+	/* allows RC6 residency counter to work */
+	I915_WRITE(VLV_COUNTER_CONTROL,
+		   _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH |
+				      VLV_MEDIA_RC6_COUNT_EN |
+				      VLV_RENDER_RC6_COUNT_EN));
+
+	/* For now we assume BIOS is allocating and populating the PCBR  */
+	pcbr = I915_READ(VLV_PCBR);
+
+	DRM_DEBUG_DRIVER("PCBR offset : 0x%x\n", pcbr);
+
+	/* 3: Enable RC6 */
+	if ((intel_enable_rc6(dev) & INTEL_RC6_ENABLE) &&
+						(pcbr >> VLV_PCBR_ADDR_SHIFT))
+		rc6_mode = GEN6_RC_CTL_EI_MODE(1);
+
+	I915_WRITE(GEN6_RC_CONTROL, rc6_mode);
+
+	/* 4 Program defaults and thresholds for RPS*/
+	I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400);
+	I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000);
+	I915_WRITE(GEN6_RP_UP_EI, 66000);
+	I915_WRITE(GEN6_RP_DOWN_EI, 350000);
+
+	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	/* WaDisablePwrmtrEvent:chv (pre-production hw) */
+	I915_WRITE(0xA80C, I915_READ(0xA80C) & 0x00ffffff);
+	I915_WRITE(0xA810, I915_READ(0xA810) & 0xffffff00);
+
+	/* 5: Enable RPS */
+	I915_WRITE(GEN6_RP_CONTROL,
+		   GEN6_RP_MEDIA_HW_NORMAL_MODE |
+		   GEN6_RP_MEDIA_IS_GFX | /* WaSetMaskForGfxBusyness:chv (pre-production hw ?) */
+		   GEN6_RP_ENABLE |
+		   GEN6_RP_UP_BUSY_AVG |
+		   GEN6_RP_DOWN_IDLE_AVG);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+
+	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & 0x10 ? "yes" : "no");
+	DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val);
+
+	dev_priv->rps.cur_freq = (val >> 8) & 0xff;
+	DRM_DEBUG_DRIVER("current GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.cur_freq),
+			 dev_priv->rps.cur_freq);
+
+	DRM_DEBUG_DRIVER("setting GPU freq to %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv, dev_priv->rps.efficient_freq),
+			 dev_priv->rps.efficient_freq);
+
+	valleyview_set_rps(dev_priv->dev, dev_priv->rps.efficient_freq);
+
+	gen8_enable_rps_interrupts(dev);
+
+	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
+}
+
 static void valleyview_enable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3886,6 +4385,7 @@ static void valleyview_enable_rps(struct drm_device *dev)
 	I915_WRITE(GEN6_RP_DOWN_EI, 350000);
 
 	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+	I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 0xf4240);
 
 	I915_WRITE(GEN6_RP_CONTROL,
 		   GEN6_RP_MEDIA_TURBO |
@@ -3906,9 +4406,11 @@ static void valleyview_enable_rps(struct drm_device *dev)
 
 	/* allows RC6 residency counter to work */
 	I915_WRITE(VLV_COUNTER_CONTROL,
-		   _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH |
+		   _MASKED_BIT_ENABLE(VLV_MEDIA_RC0_COUNT_EN |
+				      VLV_RENDER_RC0_COUNT_EN |
 				      VLV_MEDIA_RC6_COUNT_EN |
 				      VLV_RENDER_RC6_COUNT_EN));
+
 	if (intel_enable_rc6(dev) & INTEL_RC6_ENABLE)
 		rc6_mode = GEN7_RC_CTL_TO_MODE | VLV_RC_CTL_CTX_RST_PARALLEL;
 
@@ -4666,33 +5168,60 @@ void intel_init_gt_powersave(struct drm_device *dev)
 {
 	i915.enable_rc6 = sanitize_rc6_option(dev, i915.enable_rc6);
 
-	if (IS_VALLEYVIEW(dev))
+	if (IS_CHERRYVIEW(dev))
+		cherryview_init_gt_powersave(dev);
+	else if (IS_VALLEYVIEW(dev))
 		valleyview_init_gt_powersave(dev);
 }
 
 void intel_cleanup_gt_powersave(struct drm_device *dev)
 {
-	if (IS_VALLEYVIEW(dev))
+	if (IS_CHERRYVIEW(dev))
+		return;
+	else if (IS_VALLEYVIEW(dev))
 		valleyview_cleanup_gt_powersave(dev);
 }
 
+/**
+ * intel_suspend_gt_powersave - suspend PM work and helper threads
+ * @dev: drm device
+ *
+ * We don't want to disable RC6 or other features here, we just want
+ * to make sure any work we've queued has finished and won't bother
+ * us while we're suspended.
+ */
+void intel_suspend_gt_powersave(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/* Interrupts should be disabled already to avoid re-arming. */
+	WARN_ON(intel_irqs_enabled(dev_priv));
+
+	flush_delayed_work(&dev_priv->rps.delayed_resume_work);
+
+	cancel_work_sync(&dev_priv->rps.work);
+
+	/* Force GPU to min freq during suspend */
+	gen6_rps_idle(dev_priv);
+}
+
 void intel_disable_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* Interrupts should be disabled already to avoid re-arming. */
-	WARN_ON(dev->irq_enabled);
+	WARN_ON(intel_irqs_enabled(dev_priv));
 
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_disable_drps(dev);
 		ironlake_disable_rc6(dev);
-	} else if (IS_GEN6(dev) || IS_GEN7(dev) || IS_BROADWELL(dev)) {
-		if (cancel_delayed_work_sync(&dev_priv->rps.delayed_resume_work))
-			intel_runtime_pm_put(dev_priv);
+	} else if (INTEL_INFO(dev)->gen >= 6) {
+		intel_suspend_gt_powersave(dev);
 
-		cancel_work_sync(&dev_priv->rps.work);
 		mutex_lock(&dev_priv->rps.hw_lock);
-		if (IS_VALLEYVIEW(dev))
+		if (IS_CHERRYVIEW(dev))
+			cherryview_disable_rps(dev);
+		else if (IS_VALLEYVIEW(dev))
 			valleyview_disable_rps(dev);
 		else
 			gen6_disable_rps(dev);
@@ -4710,7 +5239,9 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (IS_CHERRYVIEW(dev)) {
+		cherryview_enable_rps(dev);
+	} else if (IS_VALLEYVIEW(dev)) {
 		valleyview_enable_rps(dev);
 	} else if (IS_BROADWELL(dev)) {
 		gen8_enable_rps(dev);
@@ -4735,7 +5266,7 @@ void intel_enable_gt_powersave(struct drm_device *dev)
 		ironlake_enable_rc6(dev);
 		intel_init_emon(dev);
 		mutex_unlock(&dev->struct_mutex);
-	} else if (IS_GEN6(dev) || IS_GEN7(dev) || IS_BROADWELL(dev)) {
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		/*
 		 * PCU communication is slow and this doesn't need to be
 		 * done at any specific time, so do this out of our fast path
@@ -4779,7 +5310,7 @@ static void g4x_disable_trickle_feed(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe;
 
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(DSPCNTR(pipe),
 			   I915_READ(DSPCNTR(pipe)) |
 			   DISPPLANE_TRICKLE_FEED_DISABLE);
@@ -4894,7 +5425,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 	/* The below fixes the weird display corruption, a few pixels shifted
 	 * downward, on (only) LVDS of some HP laptops with IVY.
 	 */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		val = I915_READ(TRANS_CHICKEN2(pipe));
 		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
@@ -4906,7 +5437,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(TRANS_CHICKEN2(pipe), val);
 	}
 	/* WADP0ClockGatingDisable */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(TRANS_CHICKEN1(pipe),
 			   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
 	}
@@ -4918,11 +5449,9 @@ static void gen6_check_mch_setup(struct drm_device *dev)
 	uint32_t tmp;
 
 	tmp = I915_READ(MCH_SSKPD);
-	if ((tmp & MCH_SSKPD_WM0_MASK) != MCH_SSKPD_WM0_VAL) {
-		DRM_INFO("Wrong MCH_SSKPD value: 0x%08x\n", tmp);
-		DRM_INFO("This can cause pipe underruns and display issues.\n");
-		DRM_INFO("Please upgrade your BIOS to fix this.\n");
-	}
+	if ((tmp & MCH_SSKPD_WM0_MASK) != MCH_SSKPD_WM0_VAL)
+		DRM_DEBUG_KMS("Wrong MCH_SSKPD value: 0x%08x This can cause underruns.\n",
+			      tmp);
 }
 
 static void gen6_init_clock_gating(struct drm_device *dev)
@@ -5076,7 +5605,7 @@ static void lpt_suspend_hw(struct drm_device *dev)
 	}
 }
 
-static void gen8_init_clock_gating(struct drm_device *dev)
+static void broadwell_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
@@ -5088,37 +5617,12 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 	/* FIXME(BDW): Check all the w/a, some might only apply to
 	 * pre-production hw. */
 
-	/* WaDisablePartialInstShootdown:bdw */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE));
-
-	/* WaDisableThreadStallDopClockGating:bdw */
-	/* FIXME: Unclear whether we really need this on production bdw. */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE));
 
-	/*
-	 * This GEN8_CENTROID_PIXEL_OPT_DIS W/A is only needed for
-	 * pre-production hardware
-	 */
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_CENTROID_PIXEL_OPT_DIS));
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
 	I915_WRITE(GAMTARBMODE, _MASKED_BIT_ENABLE(ARB_MODE_BWGTLB_DISABLE));
 
 	I915_WRITE(_3D_CHICKEN3,
-		   _3D_CHICKEN_SDE_LIMIT_FIFO_POLY_DEPTH(2));
-
-	I915_WRITE(COMMON_SLICE_CHICKEN2,
-		   _MASKED_BIT_ENABLE(GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE));
+		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SDE_LIMIT_FIFO_POLY_DEPTH(2)));
 
-	I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
-		   _MASKED_BIT_ENABLE(GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE));
-
-	/* WaDisableDopClockGating:bdw May not be needed for production */
-	I915_WRITE(GEN7_ROW_CHICKEN2,
-		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
 	/* WaSwitchSolVfFArbitrationPriority:bdw */
 	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
@@ -5128,37 +5632,18 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 		   I915_READ(CHICKEN_PAR1_1) | DPA_MASK_VBLANK_SRD);
 
 	/* WaPsrDPRSUnmaskVBlankInSRD:bdw */
-	for_each_pipe(pipe) {
+	for_each_pipe(dev_priv, pipe) {
 		I915_WRITE(CHICKEN_PIPESL_1(pipe),
 			   I915_READ(CHICKEN_PIPESL_1(pipe)) |
 			   BDW_DPRS_MASK_VBLANK_SRD);
 	}
 
-	/* Use Force Non-Coherent whenever executing a 3D context. This is a
-	 * workaround for for a possible hang in the unlikely event a TLB
-	 * invalidation occurs during a PSD flush.
-	 */
-	I915_WRITE(HDC_CHICKEN0,
-		   I915_READ(HDC_CHICKEN0) |
-		   _MASKED_BIT_ENABLE(HDC_FORCE_NON_COHERENT));
-
 	/* WaVSRefCountFullforceMissDisable:bdw */
 	/* WaDSRefCountFullforceMissDisable:bdw */
 	I915_WRITE(GEN7_FF_THREAD_MODE,
 		   I915_READ(GEN7_FF_THREAD_MODE) &
 		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
 
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	I915_WRITE(GEN7_GT_MODE,
-		   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
-
 	I915_WRITE(GEN6_RC_SLEEP_PSMI_CONTROL,
 		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
 
@@ -5166,9 +5651,7 @@ static void gen8_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) |
 		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
-	/* Wa4x4STCOptimizationDisable:bdw */
-	I915_WRITE(CACHE_MODE_1,
-		   _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE));
+	lpt_init_clock_gating(dev);
 }
 
 static void haswell_init_clock_gating(struct drm_device *dev)
@@ -5324,28 +5807,6 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 static void valleyview_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 val;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-	switch ((val >> 6) & 3) {
-	case 0:
-	case 1:
-		dev_priv->mem_freq = 800;
-		break;
-	case 2:
-		dev_priv->mem_freq = 1066;
-		break;
-	case 3:
-		dev_priv->mem_freq = 1333;
-		break;
-	}
-	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
-
-	dev_priv->vlv_cdclk_freq = valleyview_cur_cdclk(dev_priv);
-	DRM_DEBUG_DRIVER("Current CD clock rate: %d MHz",
-			 dev_priv->vlv_cdclk_freq);
 
 	I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE);
 
@@ -5426,14 +5887,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev)
 
 	I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
 
-	/* WaDisablePartialInstShootdown:chv */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE));
-
-	/* WaDisableThreadStallDopClockGating:chv */
-	I915_WRITE(GEN8_ROW_CHICKEN,
-		   _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE));
-
 	/* WaVSRefCountFullforceMissDisable:chv */
 	/* WaDSRefCountFullforceMissDisable:chv */
 	I915_WRITE(GEN7_FF_THREAD_MODE,
@@ -5452,10 +5905,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) |
 		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
-	/* WaDisableSamplerPowerBypass:chv (pre-production hw) */
-	I915_WRITE(HALF_SLICE_CHICKEN3,
-		   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
-
 	/* WaDisableGunitClockGating:chv (pre-production hw) */
 	I915_WRITE(VLV_GUNIT_CLOCK_GATE, I915_READ(VLV_GUNIT_CLOCK_GATE) |
 		   GINT_DIS);
@@ -5465,8 +5914,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev)
 		   _MASKED_BIT_ENABLE(GEN8_FF_DOP_CLOCK_GATE_DISABLE));
 
 	/* WaDisableDopClockGating:chv (pre-production hw) */
-	I915_WRITE(GEN7_ROW_CHICKEN2,
-		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 	I915_WRITE(GEN6_UCGCTL1, I915_READ(GEN6_UCGCTL1) |
 		   GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
 }
@@ -5551,6 +5998,9 @@ static void gen3_init_clock_gating(struct drm_device *dev)
 
 	/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
 	I915_WRITE(MI_ARB_STATE, _MASKED_BIT_ENABLE(MI_ARB_C3_LP_WRITE_ENABLE));
+
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void i85x_init_clock_gating(struct drm_device *dev)
@@ -5562,6 +6012,9 @@ static void i85x_init_clock_gating(struct drm_device *dev)
 	/* interrupts should cause a wake up from C3 */
 	I915_WRITE(MI_STATE, _MASKED_BIT_ENABLE(MI_AGPBUSY_INT_EN) |
 		   _MASKED_BIT_DISABLE(MI_AGPBUSY_830_MODE));
+
+	I915_WRITE(MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void i830_init_clock_gating(struct drm_device *dev)
@@ -5569,6 +6022,10 @@ static void i830_init_clock_gating(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	I915_WRITE(DSPCLK_GATE_D, OVRUNIT_CLOCK_GATE_DISABLE);
+
+	I915_WRITE(MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_A_TRICKLE_FEED_DISABLE) |
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_B_TRICKLE_FEED_DISABLE));
 }
 
 void intel_init_clock_gating(struct drm_device *dev)
@@ -5661,7 +6118,6 @@ bool intel_display_power_enabled(struct drm_i915_private *dev_priv,
 static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
-	unsigned long irqflags;
 
 	/*
 	 * After we re-enable the power well, if we touch VGA register 0x3d5
@@ -5677,21 +6133,8 @@ static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv)
 	outb(inb(VGA_MSR_READ), VGA_MSR_WRITE);
 	vga_put(dev->pdev, VGA_RSRC_LEGACY_IO);
 
-	if (IS_BROADWELL(dev)) {
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-		I915_WRITE(GEN8_DE_PIPE_IMR(PIPE_B),
-			   dev_priv->de_irq_mask[PIPE_B]);
-		I915_WRITE(GEN8_DE_PIPE_IER(PIPE_B),
-			   ~dev_priv->de_irq_mask[PIPE_B] |
-			   GEN8_PIPE_VBLANK);
-		I915_WRITE(GEN8_DE_PIPE_IMR(PIPE_C),
-			   dev_priv->de_irq_mask[PIPE_C]);
-		I915_WRITE(GEN8_DE_PIPE_IER(PIPE_C),
-			   ~dev_priv->de_irq_mask[PIPE_C] |
-			   GEN8_PIPE_VBLANK);
-		POSTING_READ(GEN8_DE_PIPE_IER(PIPE_C));
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
-	}
+	if (IS_BROADWELL(dev))
+		gen8_irq_power_well_post_enable(dev_priv);
 }
 
 static void hsw_set_power_well(struct drm_i915_private *dev_priv,
@@ -5762,34 +6205,13 @@ static bool i9xx_always_on_power_well_enabled(struct drm_i915_private *dev_priv,
 	return true;
 }
 
-void __vlv_set_power_well(struct drm_i915_private *dev_priv,
-			  enum punit_power_well power_well_id, bool enable)
+static void vlv_set_power_well(struct drm_i915_private *dev_priv,
+			       struct i915_power_well *power_well, bool enable)
 {
-	struct drm_device *dev = dev_priv->dev;
+	enum punit_power_well power_well_id = power_well->data;
 	u32 mask;
 	u32 state;
 	u32 ctrl;
-	enum pipe pipe;
-
-	if (power_well_id == PUNIT_POWER_WELL_DPIO_CMN_BC) {
-		if (enable) {
-			/*
-			 * Enable the CRI clock source so we can get at the
-			 * display and the reference clock for VGA
-			 * hotplug / manual detection.
-			 */
-			I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
-				   DPLL_REFA_CLK_ENABLE_VLV |
-				   DPLL_INTEGRATED_CRI_CLK_VLV);
-			udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
-		} else {
-			for_each_pipe(pipe)
-				assert_pll_disabled(dev_priv, pipe);
-			/* Assert common reset */
-			I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) &
-				   ~DPIO_CMNRST);
-		}
-	}
 
 	mask = PUNIT_PWRGT_MASK(power_well_id);
 	state = enable ? PUNIT_PWRGT_PWR_ON(power_well_id) :
@@ -5817,28 +6239,6 @@ void __vlv_set_power_well(struct drm_i915_private *dev_priv,
 
 out:
 	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	/*
-	 * From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
-	 *  6.	De-assert cmn_reset/side_reset. Same as VLV X0.
-	 *   a.	GUnit 0x2110 bit[0] set to 1 (def 0)
-	 *   b.	The other bits such as sfr settings / modesel may all
-	 *	be set to 0.
-	 *
-	 * This should only be done on init and resume from S3 with
-	 * both PLLs disabled, or we risk losing DPIO and PLL
-	 * synchronization.
-	 */
-	if (power_well_id == PUNIT_POWER_WELL_DPIO_CMN_BC && enable)
-		I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) | DPIO_CMNRST);
-}
-
-static void vlv_set_power_well(struct drm_i915_private *dev_priv,
-			       struct i915_power_well *power_well, bool enable)
-{
-	enum punit_power_well power_well_id = power_well->data;
-
-	__vlv_set_power_well(dev_priv, power_well_id, enable);
 }
 
 static void vlv_power_well_sync_hw(struct drm_i915_private *dev_priv,
@@ -5928,6 +6328,201 @@ static void vlv_display_power_well_disable(struct drm_i915_private *dev_priv,
 	spin_unlock_irq(&dev_priv->irq_lock);
 
 	vlv_set_power_well(dev_priv, power_well, false);
+
+	vlv_power_sequencer_reset(dev_priv);
+}
+
+static void vlv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
+
+	/*
+	 * Enable the CRI clock source so we can get at the
+	 * display and the reference clock for VGA
+	 * hotplug / manual detection.
+	 */
+	I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+		   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/*
+	 * From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
+	 *  6.	De-assert cmn_reset/side_reset. Same as VLV X0.
+	 *   a.	GUnit 0x2110 bit[0] set to 1 (def 0)
+	 *   b.	The other bits such as sfr settings / modesel may all
+	 *	be set to 0.
+	 *
+	 * This should only be done on init and resume from S3 with
+	 * both PLLs disabled, or we risk losing DPIO and PLL
+	 * synchronization.
+	 */
+	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) | DPIO_CMNRST);
+}
+
+static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum pipe pipe;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC);
+
+	for_each_pipe(dev_priv, pipe)
+		assert_pll_disabled(dev_priv, pipe);
+
+	/* Assert common reset */
+	I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) & ~DPIO_CMNRST);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	enum dpio_phy phy;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC &&
+		     power_well->data != PUNIT_POWER_WELL_DPIO_CMN_D);
+
+	/*
+	 * Enable the CRI clock source so we can get at the
+	 * display and the reference clock for VGA
+	 * hotplug / manual detection.
+	 */
+	if (power_well->data == PUNIT_POWER_WELL_DPIO_CMN_BC) {
+		phy = DPIO_PHY0;
+		I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+			   DPLL_REFA_CLK_ENABLE_VLV);
+		I915_WRITE(DPLL(PIPE_B), I915_READ(DPLL(PIPE_B)) |
+			   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	} else {
+		phy = DPIO_PHY1;
+		I915_WRITE(DPLL(PIPE_C), I915_READ(DPLL(PIPE_C)) |
+			   DPLL_REFA_CLK_ENABLE_VLV | DPLL_INTEGRATED_CRI_CLK_VLV);
+	}
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/* Poll for phypwrgood signal */
+	if (wait_for(I915_READ(DISPLAY_PHY_STATUS) & PHY_POWERGOOD(phy), 1))
+		DRM_ERROR("Display PHY %d is not power up\n", phy);
+
+	I915_WRITE(DISPLAY_PHY_CONTROL, I915_READ(DISPLAY_PHY_CONTROL) |
+		   PHY_COM_LANE_RESET_DEASSERT(phy));
+}
+
+static void chv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum dpio_phy phy;
+
+	WARN_ON_ONCE(power_well->data != PUNIT_POWER_WELL_DPIO_CMN_BC &&
+		     power_well->data != PUNIT_POWER_WELL_DPIO_CMN_D);
+
+	if (power_well->data == PUNIT_POWER_WELL_DPIO_CMN_BC) {
+		phy = DPIO_PHY0;
+		assert_pll_disabled(dev_priv, PIPE_A);
+		assert_pll_disabled(dev_priv, PIPE_B);
+	} else {
+		phy = DPIO_PHY1;
+		assert_pll_disabled(dev_priv, PIPE_C);
+	}
+
+	I915_WRITE(DISPLAY_PHY_CONTROL, I915_READ(DISPLAY_PHY_CONTROL) &
+		   ~PHY_COM_LANE_RESET_DEASSERT(phy));
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	enum pipe pipe = power_well->data;
+	bool enabled;
+	u32 state, ctrl;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+	state = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe);
+	/*
+	 * We only ever set the power-on and power-gate states, anything
+	 * else is unexpected.
+	 */
+	WARN_ON(state != DP_SSS_PWR_ON(pipe) && state != DP_SSS_PWR_GATE(pipe));
+	enabled = state == DP_SSS_PWR_ON(pipe);
+
+	/*
+	 * A transient state at this point would mean some unexpected party
+	 * is poking at the power controls too.
+	 */
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSC_MASK(pipe);
+	WARN_ON(ctrl << 16 != state);
+
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	return enabled;
+}
+
+static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
+				    struct i915_power_well *power_well,
+				    bool enable)
+{
+	enum pipe pipe = power_well->data;
+	u32 state;
+	u32 ctrl;
+
+	state = enable ? DP_SSS_PWR_ON(pipe) : DP_SSS_PWR_GATE(pipe);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+#define COND \
+	((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe)) == state)
+
+	if (COND)
+		goto out;
+
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	ctrl &= ~DP_SSC_MASK(pipe);
+	ctrl |= enable ? DP_SSC_PWR_ON(pipe) : DP_SSC_PWR_GATE(pipe);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, ctrl);
+
+	if (wait_for(COND, 100))
+		DRM_ERROR("timout setting power well state %08x (%08x)\n",
+			  state,
+			  vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ));
+
+#undef COND
+
+out:
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void chv_pipe_power_well_sync_hw(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	chv_set_pipe_power_well(dev_priv, power_well, power_well->count > 0);
+}
+
+static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
+				       struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PIPE_A &&
+		     power_well->data != PIPE_B &&
+		     power_well->data != PIPE_C);
+
+	chv_set_pipe_power_well(dev_priv, power_well, true);
+}
+
+static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	WARN_ON_ONCE(power_well->data != PIPE_A &&
+		     power_well->data != PIPE_B &&
+		     power_well->data != PIPE_C);
+
+	chv_set_pipe_power_well(dev_priv, power_well, false);
 }
 
 static void check_power_well_state(struct drm_i915_private *dev_priv,
@@ -6079,6 +6674,7 @@ EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
 	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |		\
 	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |		\
 	BIT(POWER_DOMAIN_PORT_CRT) |			\
+	BIT(POWER_DOMAIN_PLLS) |			\
 	BIT(POWER_DOMAIN_INIT))
 #define HSW_DISPLAY_POWER_DOMAINS (				\
 	(POWER_DOMAIN_MASK & ~HSW_ALWAYS_ON_POWER_DOMAINS) |	\
@@ -6120,6 +6716,39 @@ EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
 	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
 	BIT(POWER_DOMAIN_INIT))
 
+#define CHV_PIPE_A_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_A) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_PIPE_B_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_B) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_PIPE_C_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PIPE_C) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_CMN_BC_POWER_DOMAINS (		\
+	BIT(POWER_DOMAIN_PORT_DDI_B_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_B_4_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_C_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_CMN_D_POWER_DOMAINS (		\
+	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_2_LANES) |	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
+#define CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS (	\
+	BIT(POWER_DOMAIN_PORT_DDI_D_4_LANES) |	\
+	BIT(POWER_DOMAIN_INIT))
+
 static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
 	.sync_hw = i9xx_always_on_power_well_noop,
 	.enable = i9xx_always_on_power_well_noop,
@@ -6127,6 +6756,20 @@ static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
 	.is_enabled = i9xx_always_on_power_well_enabled,
 };
 
+static const struct i915_power_well_ops chv_pipe_power_well_ops = {
+	.sync_hw = chv_pipe_power_well_sync_hw,
+	.enable = chv_pipe_power_well_enable,
+	.disable = chv_pipe_power_well_disable,
+	.is_enabled = chv_pipe_power_well_enabled,
+};
+
+static const struct i915_power_well_ops chv_dpio_cmn_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = chv_dpio_cmn_power_well_enable,
+	.disable = chv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
 static struct i915_power_well i9xx_always_on_power_well[] = {
 	{
 		.name = "always-on",
@@ -6178,6 +6821,13 @@ static const struct i915_power_well_ops vlv_display_power_well_ops = {
 	.is_enabled = vlv_power_well_enabled,
 };
 
+static const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
+	.sync_hw = vlv_power_well_sync_hw,
+	.enable = vlv_dpio_cmn_power_well_enable,
+	.disable = vlv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
 static const struct i915_power_well_ops vlv_dpio_power_well_ops = {
 	.sync_hw = vlv_power_well_sync_hw,
 	.enable = vlv_power_well_enable,
@@ -6238,10 +6888,126 @@ static struct i915_power_well vlv_power_wells[] = {
 		.name = "dpio-common",
 		.domains = VLV_DPIO_CMN_BC_POWER_DOMAINS,
 		.data = PUNIT_POWER_WELL_DPIO_CMN_BC,
+		.ops = &vlv_dpio_cmn_power_well_ops,
+	},
+};
+
+static struct i915_power_well chv_power_wells[] = {
+	{
+		.name = "always-on",
+		.always_on = 1,
+		.domains = VLV_ALWAYS_ON_POWER_DOMAINS,
+		.ops = &i9xx_always_on_power_well_ops,
+	},
+#if 0
+	{
+		.name = "display",
+		.domains = VLV_DISPLAY_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DISP2D,
+		.ops = &vlv_display_power_well_ops,
+	},
+	{
+		.name = "pipe-a",
+		.domains = CHV_PIPE_A_POWER_DOMAINS,
+		.data = PIPE_A,
+		.ops = &chv_pipe_power_well_ops,
+	},
+	{
+		.name = "pipe-b",
+		.domains = CHV_PIPE_B_POWER_DOMAINS,
+		.data = PIPE_B,
+		.ops = &chv_pipe_power_well_ops,
+	},
+	{
+		.name = "pipe-c",
+		.domains = CHV_PIPE_C_POWER_DOMAINS,
+		.data = PIPE_C,
+		.ops = &chv_pipe_power_well_ops,
+	},
+#endif
+	{
+		.name = "dpio-common-bc",
+		/*
+		 * XXX: cmnreset for one PHY seems to disturb the other.
+		 * As a workaround keep both powered on at the same
+		 * time for now.
+		 */
+		.domains = CHV_DPIO_CMN_BC_POWER_DOMAINS | CHV_DPIO_CMN_D_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DPIO_CMN_BC,
+		.ops = &chv_dpio_cmn_power_well_ops,
+	},
+	{
+		.name = "dpio-common-d",
+		/*
+		 * XXX: cmnreset for one PHY seems to disturb the other.
+		 * As a workaround keep both powered on at the same
+		 * time for now.
+		 */
+		.domains = CHV_DPIO_CMN_BC_POWER_DOMAINS | CHV_DPIO_CMN_D_POWER_DOMAINS,
+		.data = PUNIT_POWER_WELL_DPIO_CMN_D,
+		.ops = &chv_dpio_cmn_power_well_ops,
+	},
+#if 0
+	{
+		.name = "dpio-tx-b-01",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_01,
+	},
+	{
+		.name = "dpio-tx-b-23",
+		.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_B_LANES_23,
+	},
+	{
+		.name = "dpio-tx-c-01",
+		.domains = VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
 		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_01,
 	},
+	{
+		.name = "dpio-tx-c-23",
+		.domains = VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
+			   VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_C_LANES_23,
+	},
+	{
+		.name = "dpio-tx-d-01",
+		.domains = CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS |
+			   CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_D_LANES_01,
+	},
+	{
+		.name = "dpio-tx-d-23",
+		.domains = CHV_DPIO_TX_D_LANES_01_POWER_DOMAINS |
+			   CHV_DPIO_TX_D_LANES_23_POWER_DOMAINS,
+		.ops = &vlv_dpio_power_well_ops,
+		.data = PUNIT_POWER_WELL_DPIO_TX_D_LANES_23,
+	},
+#endif
 };
 
+static struct i915_power_well *lookup_power_well(struct drm_i915_private *dev_priv,
+						 enum punit_power_well power_well_id)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+	struct i915_power_well *power_well;
+	int i;
+
+	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
+		if (power_well->data == power_well_id)
+			return power_well;
+	}
+
+	return NULL;
+}
+
 #define set_power_wells(power_domains, __power_wells) ({		\
 	(power_domains)->power_wells = (__power_wells);			\
 	(power_domains)->power_well_count = ARRAY_SIZE(__power_wells);	\
@@ -6263,6 +7029,8 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
 	} else if (IS_BROADWELL(dev_priv->dev)) {
 		set_power_wells(power_domains, bdw_power_wells);
 		hsw_pwr = power_domains;
+	} else if (IS_CHERRYVIEW(dev_priv->dev)) {
+		set_power_wells(power_domains, chv_power_wells);
 	} else if (IS_VALLEYVIEW(dev_priv->dev)) {
 		set_power_wells(power_domains, vlv_power_wells);
 	} else {
@@ -6292,11 +7060,50 @@ static void intel_power_domains_resume(struct drm_i915_private *dev_priv)
 	mutex_unlock(&power_domains->lock);
 }
 
+static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *cmn =
+		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
+	struct i915_power_well *disp2d =
+		lookup_power_well(dev_priv, PUNIT_POWER_WELL_DISP2D);
+
+	/* nothing to do if common lane is already off */
+	if (!cmn->ops->is_enabled(dev_priv, cmn))
+		return;
+
+	/* If the display might be already active skip this */
+	if (disp2d->ops->is_enabled(dev_priv, disp2d) &&
+	    I915_READ(DPIO_CTL) & DPIO_CMNRST)
+		return;
+
+	DRM_DEBUG_KMS("toggling display PHY side reset\n");
+
+	/* cmnlane needs DPLL registers */
+	disp2d->ops->enable(dev_priv, disp2d);
+
+	/*
+	 * From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
+	 * Need to assert and de-assert PHY SB reset by gating the
+	 * common lane power, then un-gating it.
+	 * Simply ungating isn't enough to reset the PHY enough to get
+	 * ports and lanes running.
+	 */
+	cmn->ops->disable(dev_priv, cmn);
+}
+
 void intel_power_domains_init_hw(struct drm_i915_private *dev_priv)
 {
+	struct drm_device *dev = dev_priv->dev;
 	struct i915_power_domains *power_domains = &dev_priv->power_domains;
 
 	power_domains->initializing = true;
+
+	if (IS_VALLEYVIEW(dev) && !IS_CHERRYVIEW(dev)) {
+		mutex_lock(&power_domains->lock);
+		vlv_cmnlane_wa(dev_priv);
+		mutex_unlock(&power_domains->lock);
+	}
+
 	/* For now, we need the power well to be always enabled. */
 	intel_display_set_init_power(dev_priv, true);
 	intel_power_domains_resume(dev_priv);
@@ -6449,13 +7256,15 @@ void intel_init_pm(struct drm_device *dev)
 		else if (IS_HASWELL(dev))
 			dev_priv->display.init_clock_gating = haswell_init_clock_gating;
 		else if (INTEL_INFO(dev)->gen == 8)
-			dev_priv->display.init_clock_gating = gen8_init_clock_gating;
+			dev_priv->display.init_clock_gating = broadwell_init_clock_gating;
 	} else if (IS_CHERRYVIEW(dev)) {
-		dev_priv->display.update_wm = valleyview_update_wm;
+		dev_priv->display.update_wm = cherryview_update_wm;
+		dev_priv->display.update_sprite_wm = valleyview_update_sprite_wm;
 		dev_priv->display.init_clock_gating =
 			cherryview_init_clock_gating;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.update_wm = valleyview_update_wm;
+		dev_priv->display.update_sprite_wm = valleyview_update_sprite_wm;
 		dev_priv->display.init_clock_gating =
 			valleyview_init_clock_gating;
 	} else if (IS_PINEVIEW(dev)) {
@@ -6469,7 +7278,7 @@ void intel_init_pm(struct drm_device *dev)
 				 (dev_priv->is_ddr3 == 1) ? "3" : "2",
 				 dev_priv->fsb_freq, dev_priv->mem_freq);
 			/* Disable CxSR and never update its watermark again */
-			pineview_disable_cxsr(dev);
+			intel_set_memory_cxsr(dev_priv, false);
 			dev_priv->display.update_wm = NULL;
 		} else
 			dev_priv->display.update_wm = pineview_update_wm;
@@ -6552,7 +7361,7 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
 	return 0;
 }
 
-int vlv_gpu_freq(struct drm_i915_private *dev_priv, int val)
+static int byt_gpu_freq(struct drm_i915_private *dev_priv, int val)
 {
 	int div;
 
@@ -6574,7 +7383,7 @@ int vlv_gpu_freq(struct drm_i915_private *dev_priv, int val)
 	return DIV_ROUND_CLOSEST(dev_priv->mem_freq * (val + 6 - 0xbd), 4 * div);
 }
 
-int vlv_freq_opcode(struct drm_i915_private *dev_priv, int val)
+static int byt_freq_opcode(struct drm_i915_private *dev_priv, int val)
 {
 	int mul;
 
@@ -6596,6 +7405,81 @@ int vlv_freq_opcode(struct drm_i915_private *dev_priv, int val)
 	return DIV_ROUND_CLOSEST(4 * mul * val, dev_priv->mem_freq) + 0xbd - 6;
 }
 
+static int chv_gpu_freq(struct drm_i915_private *dev_priv, int val)
+{
+	int div, freq;
+
+	switch (dev_priv->rps.cz_freq) {
+	case 200:
+		div = 5;
+		break;
+	case 267:
+		div = 6;
+		break;
+	case 320:
+	case 333:
+	case 400:
+		div = 8;
+		break;
+	default:
+		return -1;
+	}
+
+	freq = (DIV_ROUND_CLOSEST((dev_priv->rps.cz_freq * val), 2 * div) / 2);
+
+	return freq;
+}
+
+static int chv_freq_opcode(struct drm_i915_private *dev_priv, int val)
+{
+	int mul, opcode;
+
+	switch (dev_priv->rps.cz_freq) {
+	case 200:
+		mul = 5;
+		break;
+	case 267:
+		mul = 6;
+		break;
+	case 320:
+	case 333:
+	case 400:
+		mul = 8;
+		break;
+	default:
+		return -1;
+	}
+
+	/* CHV needs even values */
+	opcode = (DIV_ROUND_CLOSEST((val * 2 * mul), dev_priv->rps.cz_freq) * 2);
+
+	return opcode;
+}
+
+int vlv_gpu_freq(struct drm_i915_private *dev_priv, int val)
+{
+	int ret = -1;
+
+	if (IS_CHERRYVIEW(dev_priv->dev))
+		ret = chv_gpu_freq(dev_priv, val);
+	else if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = byt_gpu_freq(dev_priv, val);
+
+	return ret;
+}
+
+int vlv_freq_opcode(struct drm_i915_private *dev_priv, int val)
+{
+	int ret = -1;
+
+	if (IS_CHERRYVIEW(dev_priv->dev))
+		ret = chv_freq_opcode(dev_priv, val);
+	else if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = byt_freq_opcode(dev_priv, val);
+
+	return ret;
+}
+
 void intel_pm_setup(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -6606,5 +7490,5 @@ void intel_pm_setup(struct drm_device *dev)
 			  intel_gen6_powersave_work);
 
 	dev_priv->pm.suspended = false;
-	dev_priv->pm.irqs_disabled = false;
+	dev_priv->pm._irqs_disabled = false;
 }
diff --git a/drivers/gpu/drm/i915/intel_renderstate.h b/drivers/gpu/drm/i915/intel_renderstate.h
index a5e783a9928a..6c792d3a9c9c 100644
--- a/drivers/gpu/drm/i915/intel_renderstate.h
+++ b/drivers/gpu/drm/i915/intel_renderstate.h
@@ -24,14 +24,7 @@
 #ifndef _INTEL_RENDERSTATE_H
 #define _INTEL_RENDERSTATE_H
 
-#include <linux/types.h>
-
-struct intel_renderstate_rodata {
-	const u32 *reloc;
-	const u32 reloc_items;
-	const u32 *batch;
-	const u32 batch_items;
-};
+#include "i915_drv.h"
 
 extern const struct intel_renderstate_rodata gen6_null_state;
 extern const struct intel_renderstate_rodata gen7_null_state;
@@ -40,7 +33,6 @@ extern const struct intel_renderstate_rodata gen8_null_state;
 #define RO_RENDERSTATE(_g)						\
 	const struct intel_renderstate_rodata gen ## _g ## _null_state = { \
 		.reloc = gen ## _g ## _null_state_relocs,		\
-		.reloc_items = sizeof(gen ## _g ## _null_state_relocs)/4, \
 		.batch = gen ## _g ## _null_state_batch,		\
 		.batch_items = sizeof(gen ## _g ## _null_state_batch)/4, \
 	}
diff --git a/drivers/gpu/drm/i915/intel_renderstate_gen6.c b/drivers/gpu/drm/i915/intel_renderstate_gen6.c
index 740538ad0977..56c1429d8a60 100644
--- a/drivers/gpu/drm/i915/intel_renderstate_gen6.c
+++ b/drivers/gpu/drm/i915/intel_renderstate_gen6.c
@@ -6,6 +6,7 @@ static const u32 gen6_null_state_relocs[] = {
 	0x0000002c,
 	0x000001e0,
 	0x000001e4,
+	-1,
 };
 
 static const u32 gen6_null_state_batch[] = {
diff --git a/drivers/gpu/drm/i915/intel_renderstate_gen7.c b/drivers/gpu/drm/i915/intel_renderstate_gen7.c
index 6fa7ff2a1298..419e35a7b0ff 100644
--- a/drivers/gpu/drm/i915/intel_renderstate_gen7.c
+++ b/drivers/gpu/drm/i915/intel_renderstate_gen7.c
@@ -5,6 +5,7 @@ static const u32 gen7_null_state_relocs[] = {
 	0x00000010,
 	0x00000018,
 	0x000001ec,
+	-1,
 };
 
 static const u32 gen7_null_state_batch[] = {
diff --git a/drivers/gpu/drm/i915/intel_renderstate_gen8.c b/drivers/gpu/drm/i915/intel_renderstate_gen8.c
index 5c875615d42a..75ef1b5de45c 100644
--- a/drivers/gpu/drm/i915/intel_renderstate_gen8.c
+++ b/drivers/gpu/drm/i915/intel_renderstate_gen8.c
@@ -5,6 +5,7 @@ static const u32 gen8_null_state_relocs[] = {
 	0x00000050,
 	0x00000060,
 	0x000003ec,
+	-1,
 };
 
 static const u32 gen8_null_state_batch[] = {
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 279488addf3f..0a80e419b589 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -33,14 +33,24 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
- * but keeps the logic simple. Indeed, the whole purpose of this macro is just
- * to give some inclination as to some of the magic values used in the various
- * workarounds!
- */
-#define CACHELINE_BYTES 64
+bool
+intel_ring_initialized(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+
+	if (!dev)
+		return false;
 
-static inline int __ring_space(int head, int tail, int size)
+	if (i915.enable_execlists) {
+		struct intel_context *dctx = ring->default_context;
+		struct intel_ringbuffer *ringbuf = dctx->engine[ring->id].ringbuf;
+
+		return ringbuf->obj;
+	} else
+		return ring->buffer && ring->buffer->obj;
+}
+
+int __intel_ring_space(int head, int tail, int size)
 {
 	int space = head - (tail + I915_RING_FREE_SPACE);
 	if (space < 0)
@@ -48,13 +58,13 @@ static inline int __ring_space(int head, int tail, int size)
 	return space;
 }
 
-static inline int ring_space(struct intel_engine_cs *ring)
+int intel_ring_space(struct intel_ringbuffer *ringbuf)
 {
-	struct intel_ringbuffer *ringbuf = ring->buffer;
-	return __ring_space(ringbuf->head & HEAD_ADDR, ringbuf->tail, ringbuf->size);
+	return __intel_ring_space(ringbuf->head & HEAD_ADDR,
+				  ringbuf->tail, ringbuf->size);
 }
 
-static bool intel_ring_stopped(struct intel_engine_cs *ring)
+bool intel_ring_stopped(struct intel_engine_cs *ring)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	return dev_priv->gpu_error.stop_rings & intel_ring_flag(ring);
@@ -381,6 +391,27 @@ gen7_render_ring_flush(struct intel_engine_cs *ring,
 }
 
 static int
+gen8_emit_pipe_control(struct intel_engine_cs *ring,
+		       u32 flags, u32 scratch_addr)
+{
+	int ret;
+
+	ret = intel_ring_begin(ring, 6);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
+	intel_ring_emit(ring, flags);
+	intel_ring_emit(ring, scratch_addr);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_advance(ring);
+
+	return 0;
+}
+
+static int
 gen8_render_ring_flush(struct intel_engine_cs *ring,
 		       u32 invalidate_domains, u32 flush_domains)
 {
@@ -403,22 +434,24 @@ gen8_render_ring_flush(struct intel_engine_cs *ring,
 		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
 		flags |= PIPE_CONTROL_QW_WRITE;
 		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
+
+		/* WaCsStallBeforeStateCacheInvalidate:bdw,chv */
+		ret = gen8_emit_pipe_control(ring,
+					     PIPE_CONTROL_CS_STALL |
+					     PIPE_CONTROL_STALL_AT_SCOREBOARD,
+					     0);
+		if (ret)
+			return ret;
 	}
 
-	ret = intel_ring_begin(ring, 6);
+	ret = gen8_emit_pipe_control(ring, flags, scratch_addr);
 	if (ret)
 		return ret;
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_advance(ring);
+	if (!invalidate_domains && flush_domains)
+		return gen7_ring_fbc_flush(ring, FBC_REND_NUKE);
 
 	return 0;
-
 }
 
 static void ring_write_tail(struct intel_engine_cs *ring,
@@ -461,9 +494,14 @@ static bool stop_ring(struct intel_engine_cs *ring)
 
 	if (!IS_GEN2(ring->dev)) {
 		I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING));
-		if (wait_for_atomic((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
-			DRM_ERROR("%s :timed out trying to stop ring\n", ring->name);
-			return false;
+		if (wait_for((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) {
+			DRM_ERROR("%s : timed out trying to stop ring\n", ring->name);
+			/* Sometimes we observe that the idle flag is not
+			 * set even though the ring is empty. So double
+			 * check before giving up.
+			 */
+			if (I915_READ_HEAD(ring) != I915_READ_TAIL(ring))
+				return false;
 		}
 	}
 
@@ -517,11 +555,22 @@ static int init_ring_common(struct intel_engine_cs *ring)
 	else
 		ring_setup_phys_status_page(ring);
 
+	/* Enforce ordering by reading HEAD register back */
+	I915_READ_HEAD(ring);
+
 	/* Initialize the ring. This must happen _after_ we've cleared the ring
 	 * registers with the above sequence (the readback of the HEAD registers
 	 * also enforces ordering), otherwise the hw might lose the new ring
 	 * register values. */
 	I915_WRITE_START(ring, i915_gem_obj_ggtt_offset(obj));
+
+	/* WaClearRingBufHeadRegAtInit:ctg,elk */
+	if (I915_READ_HEAD(ring))
+		DRM_DEBUG("%s initialization failed [head=%08x], fudging\n",
+			  ring->name, I915_READ_HEAD(ring));
+	I915_WRITE_HEAD(ring, 0);
+	(void)I915_READ_HEAD(ring);
+
 	I915_WRITE_CTL(ring,
 			((ringbuf->size - PAGE_SIZE) & RING_NR_PAGES)
 			| RING_VALID);
@@ -545,7 +594,7 @@ static int init_ring_common(struct intel_engine_cs *ring)
 	else {
 		ringbuf->head = I915_READ_HEAD(ring);
 		ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR;
-		ringbuf->space = ring_space(ring);
+		ringbuf->space = intel_ring_space(ringbuf);
 		ringbuf->last_retired_head = -1;
 	}
 
@@ -557,8 +606,25 @@ out:
 	return ret;
 }
 
-static int
-init_pipe_control(struct intel_engine_cs *ring)
+void
+intel_fini_pipe_control(struct intel_engine_cs *ring)
+{
+	struct drm_device *dev = ring->dev;
+
+	if (ring->scratch.obj == NULL)
+		return;
+
+	if (INTEL_INFO(dev)->gen >= 5) {
+		kunmap(sg_page(ring->scratch.obj->pages->sgl));
+		i915_gem_object_ggtt_unpin(ring->scratch.obj);
+	}
+
+	drm_gem_object_unreference(&ring->scratch.obj->base);
+	ring->scratch.obj = NULL;
+}
+
+int
+intel_init_pipe_control(struct intel_engine_cs *ring)
 {
 	int ret;
 
@@ -599,11 +665,142 @@ err:
 	return ret;
 }
 
+static inline void intel_ring_emit_wa(struct intel_engine_cs *ring,
+				       u32 addr, u32 value)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (WARN_ON(dev_priv->num_wa_regs >= I915_MAX_WA_REGS))
+		return;
+
+	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+	intel_ring_emit(ring, addr);
+	intel_ring_emit(ring, value);
+
+	dev_priv->intel_wa_regs[dev_priv->num_wa_regs].addr = addr;
+	dev_priv->intel_wa_regs[dev_priv->num_wa_regs].mask = value & 0xFFFF;
+	/* value is updated with the status of remaining bits of this
+	 * register when it is read from debugfs file
+	 */
+	dev_priv->intel_wa_regs[dev_priv->num_wa_regs].value = value;
+	dev_priv->num_wa_regs++;
+
+	return;
+}
+
+static int bdw_init_workarounds(struct intel_engine_cs *ring)
+{
+	int ret;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/*
+	 * workarounds applied in this fn are part of register state context,
+	 * they need to be re-initialized followed by gpu reset, suspend/resume,
+	 * module reload.
+	 */
+	dev_priv->num_wa_regs = 0;
+	memset(dev_priv->intel_wa_regs, 0, sizeof(dev_priv->intel_wa_regs));
+
+	/*
+	 * update the number of dwords required based on the
+	 * actual number of workarounds applied
+	 */
+	ret = intel_ring_begin(ring, 18);
+	if (ret)
+		return ret;
+
+	/* WaDisablePartialInstShootdown:bdw */
+	/* WaDisableThreadStallDopClockGating:bdw */
+	/* FIXME: Unclear whether we really need this on production bdw. */
+	intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN,
+			   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE
+					     | STALL_DOP_GATING_DISABLE));
+
+	/* WaDisableDopClockGating:bdw May not be needed for production */
+	intel_ring_emit_wa(ring, GEN7_ROW_CHICKEN2,
+			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+
+	intel_ring_emit_wa(ring, HALF_SLICE_CHICKEN3,
+			   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
+
+	/* Use Force Non-Coherent whenever executing a 3D context. This is a
+	 * workaround for for a possible hang in the unlikely event a TLB
+	 * invalidation occurs during a PSD flush.
+	 */
+	intel_ring_emit_wa(ring, HDC_CHICKEN0,
+			   _MASKED_BIT_ENABLE(HDC_FORCE_NON_COHERENT));
+
+	/* Wa4x4STCOptimizationDisable:bdw */
+	intel_ring_emit_wa(ring, CACHE_MODE_1,
+			   _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE));
+
+	/*
+	 * BSpec recommends 8x4 when MSAA is used,
+	 * however in practice 16x4 seems fastest.
+	 *
+	 * Note that PS/WM thread counts depend on the WIZ hashing
+	 * disable bit, which we don't touch here, but it's good
+	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+	 */
+	intel_ring_emit_wa(ring, GEN7_GT_MODE,
+			   GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4);
+
+	intel_ring_advance(ring);
+
+	DRM_DEBUG_DRIVER("Number of Workarounds applied: %d\n",
+			 dev_priv->num_wa_regs);
+
+	return 0;
+}
+
+static int chv_init_workarounds(struct intel_engine_cs *ring)
+{
+	int ret;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	/*
+	 * workarounds applied in this fn are part of register state context,
+	 * they need to be re-initialized followed by gpu reset, suspend/resume,
+	 * module reload.
+	 */
+	dev_priv->num_wa_regs = 0;
+	memset(dev_priv->intel_wa_regs, 0, sizeof(dev_priv->intel_wa_regs));
+
+	ret = intel_ring_begin(ring, 12);
+	if (ret)
+		return ret;
+
+	/* WaDisablePartialInstShootdown:chv */
+	intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN,
+			   _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE));
+
+	/* WaDisableThreadStallDopClockGating:chv */
+	intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN,
+			   _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE));
+
+	/* WaDisableDopClockGating:chv (pre-production hw) */
+	intel_ring_emit_wa(ring, GEN7_ROW_CHICKEN2,
+			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+
+	/* WaDisableSamplerPowerBypass:chv (pre-production hw) */
+	intel_ring_emit_wa(ring, HALF_SLICE_CHICKEN3,
+			   _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS));
+
+	intel_ring_advance(ring);
+
+	return 0;
+}
+
 static int init_render_ring(struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ring->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret = init_ring_common(ring);
+	if (ret)
+		return ret;
 
 	/* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */
 	if (INTEL_INFO(dev)->gen >= 4 && INTEL_INFO(dev)->gen < 7)
@@ -631,7 +828,7 @@ static int init_render_ring(struct intel_engine_cs *ring)
 			   _MASKED_BIT_ENABLE(GFX_REPLAY_MODE));
 
 	if (INTEL_INFO(dev)->gen >= 5) {
-		ret = init_pipe_control(ring);
+		ret = intel_init_pipe_control(ring);
 		if (ret)
 			return ret;
 	}
@@ -658,17 +855,89 @@ static int init_render_ring(struct intel_engine_cs *ring)
 static void render_ring_cleanup(struct intel_engine_cs *ring)
 {
 	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (ring->scratch.obj == NULL)
-		return;
+	if (dev_priv->semaphore_obj) {
+		i915_gem_object_ggtt_unpin(dev_priv->semaphore_obj);
+		drm_gem_object_unreference(&dev_priv->semaphore_obj->base);
+		dev_priv->semaphore_obj = NULL;
+	}
 
-	if (INTEL_INFO(dev)->gen >= 5) {
-		kunmap(sg_page(ring->scratch.obj->pages->sgl));
-		i915_gem_object_ggtt_unpin(ring->scratch.obj);
+	intel_fini_pipe_control(ring);
+}
+
+static int gen8_rcs_signal(struct intel_engine_cs *signaller,
+			   unsigned int num_dwords)
+{
+#define MBOX_UPDATE_DWORDS 8
+	struct drm_device *dev = signaller->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *waiter;
+	int i, ret, num_rings;
+
+	num_rings = hweight32(INTEL_INFO(dev)->ring_mask);
+	num_dwords += (num_rings-1) * MBOX_UPDATE_DWORDS;
+#undef MBOX_UPDATE_DWORDS
+
+	ret = intel_ring_begin(signaller, num_dwords);
+	if (ret)
+		return ret;
+
+	for_each_ring(waiter, dev_priv, i) {
+		u64 gtt_offset = signaller->semaphore.signal_ggtt[i];
+		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
+			continue;
+
+		intel_ring_emit(signaller, GFX_OP_PIPE_CONTROL(6));
+		intel_ring_emit(signaller, PIPE_CONTROL_GLOBAL_GTT_IVB |
+					   PIPE_CONTROL_QW_WRITE |
+					   PIPE_CONTROL_FLUSH_ENABLE);
+		intel_ring_emit(signaller, lower_32_bits(gtt_offset));
+		intel_ring_emit(signaller, upper_32_bits(gtt_offset));
+		intel_ring_emit(signaller, signaller->outstanding_lazy_seqno);
+		intel_ring_emit(signaller, 0);
+		intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL |
+					   MI_SEMAPHORE_TARGET(waiter->id));
+		intel_ring_emit(signaller, 0);
 	}
 
-	drm_gem_object_unreference(&ring->scratch.obj->base);
-	ring->scratch.obj = NULL;
+	return 0;
+}
+
+static int gen8_xcs_signal(struct intel_engine_cs *signaller,
+			   unsigned int num_dwords)
+{
+#define MBOX_UPDATE_DWORDS 6
+	struct drm_device *dev = signaller->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_engine_cs *waiter;
+	int i, ret, num_rings;
+
+	num_rings = hweight32(INTEL_INFO(dev)->ring_mask);
+	num_dwords += (num_rings-1) * MBOX_UPDATE_DWORDS;
+#undef MBOX_UPDATE_DWORDS
+
+	ret = intel_ring_begin(signaller, num_dwords);
+	if (ret)
+		return ret;
+
+	for_each_ring(waiter, dev_priv, i) {
+		u64 gtt_offset = signaller->semaphore.signal_ggtt[i];
+		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
+			continue;
+
+		intel_ring_emit(signaller, (MI_FLUSH_DW + 1) |
+					   MI_FLUSH_DW_OP_STOREDW);
+		intel_ring_emit(signaller, lower_32_bits(gtt_offset) |
+					   MI_FLUSH_DW_USE_GTT);
+		intel_ring_emit(signaller, upper_32_bits(gtt_offset));
+		intel_ring_emit(signaller, signaller->outstanding_lazy_seqno);
+		intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL |
+					   MI_SEMAPHORE_TARGET(waiter->id));
+		intel_ring_emit(signaller, 0);
+	}
+
+	return 0;
 }
 
 static int gen6_signal(struct intel_engine_cs *signaller,
@@ -677,23 +946,16 @@ static int gen6_signal(struct intel_engine_cs *signaller,
 	struct drm_device *dev = signaller->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *useless;
-	int i, ret;
+	int i, ret, num_rings;
 
-	/* NB: In order to be able to do semaphore MBOX updates for varying
-	 * number of rings, it's easiest if we round up each individual update
-	 * to a multiple of 2 (since ring updates must always be a multiple of
-	 * 2) even though the actual update only requires 3 dwords.
-	 */
-#define MBOX_UPDATE_DWORDS 4
-	if (i915_semaphore_is_enabled(dev))
-		num_dwords += ((I915_NUM_RINGS-1) * MBOX_UPDATE_DWORDS);
-	else
-		return intel_ring_begin(signaller, num_dwords);
+#define MBOX_UPDATE_DWORDS 3
+	num_rings = hweight32(INTEL_INFO(dev)->ring_mask);
+	num_dwords += round_up((num_rings-1) * MBOX_UPDATE_DWORDS, 2);
+#undef MBOX_UPDATE_DWORDS
 
 	ret = intel_ring_begin(signaller, num_dwords);
 	if (ret)
 		return ret;
-#undef MBOX_UPDATE_DWORDS
 
 	for_each_ring(useless, dev_priv, i) {
 		u32 mbox_reg = signaller->semaphore.mbox.signal[i];
@@ -701,15 +963,13 @@ static int gen6_signal(struct intel_engine_cs *signaller,
 			intel_ring_emit(signaller, MI_LOAD_REGISTER_IMM(1));
 			intel_ring_emit(signaller, mbox_reg);
 			intel_ring_emit(signaller, signaller->outstanding_lazy_seqno);
-			intel_ring_emit(signaller, MI_NOOP);
-		} else {
-			intel_ring_emit(signaller, MI_NOOP);
-			intel_ring_emit(signaller, MI_NOOP);
-			intel_ring_emit(signaller, MI_NOOP);
-			intel_ring_emit(signaller, MI_NOOP);
 		}
 	}
 
+	/* If num_dwords was rounded, make sure the tail pointer is correct */
+	if (num_rings % 2 == 0)
+		intel_ring_emit(signaller, MI_NOOP);
+
 	return 0;
 }
 
@@ -727,7 +987,11 @@ gen6_add_request(struct intel_engine_cs *ring)
 {
 	int ret;
 
-	ret = ring->semaphore.signal(ring, 4);
+	if (ring->semaphore.signal)
+		ret = ring->semaphore.signal(ring, 4);
+	else
+		ret = intel_ring_begin(ring, 4);
+
 	if (ret)
 		return ret;
 
@@ -754,6 +1018,32 @@ static inline bool i915_gem_has_seqno_wrapped(struct drm_device *dev,
  * @signaller - ring which has, or will signal
  * @seqno - seqno which the waiter will block on
  */
+
+static int
+gen8_ring_sync(struct intel_engine_cs *waiter,
+	       struct intel_engine_cs *signaller,
+	       u32 seqno)
+{
+	struct drm_i915_private *dev_priv = waiter->dev->dev_private;
+	int ret;
+
+	ret = intel_ring_begin(waiter, 4);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(waiter, MI_SEMAPHORE_WAIT |
+				MI_SEMAPHORE_GLOBAL_GTT |
+				MI_SEMAPHORE_POLL |
+				MI_SEMAPHORE_SAD_GTE_SDD);
+	intel_ring_emit(waiter, seqno);
+	intel_ring_emit(waiter,
+			lower_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id)));
+	intel_ring_emit(waiter,
+			upper_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id)));
+	intel_ring_advance(waiter);
+	return 0;
+}
+
 static int
 gen6_ring_sync(struct intel_engine_cs *waiter,
 	       struct intel_engine_cs *signaller,
@@ -901,7 +1191,7 @@ gen5_ring_get_irq(struct intel_engine_cs *ring)
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (ring->irq_refcount++ == 0)
-		ilk_enable_gt_irq(dev_priv, ring->irq_enable_mask);
+		gen5_enable_gt_irq(dev_priv, ring->irq_enable_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
 	return true;
@@ -916,7 +1206,7 @@ gen5_ring_put_irq(struct intel_engine_cs *ring)
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (--ring->irq_refcount == 0)
-		ilk_disable_gt_irq(dev_priv, ring->irq_enable_mask);
+		gen5_disable_gt_irq(dev_priv, ring->irq_enable_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 }
 
@@ -1109,7 +1399,7 @@ gen6_ring_get_irq(struct intel_engine_cs *ring)
 					 GT_PARITY_ERROR(dev)));
 		else
 			I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
-		ilk_enable_gt_irq(dev_priv, ring->irq_enable_mask);
+		gen5_enable_gt_irq(dev_priv, ring->irq_enable_mask);
 	}
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -1129,7 +1419,7 @@ gen6_ring_put_irq(struct intel_engine_cs *ring)
 			I915_WRITE_IMR(ring, ~GT_PARITY_ERROR(dev));
 		else
 			I915_WRITE_IMR(ring, ~0);
-		ilk_disable_gt_irq(dev_priv, ring->irq_enable_mask);
+		gen5_disable_gt_irq(dev_priv, ring->irq_enable_mask);
 	}
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 }
@@ -1147,7 +1437,7 @@ hsw_vebox_get_irq(struct intel_engine_cs *ring)
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (ring->irq_refcount++ == 0) {
 		I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
-		snb_enable_pm_irq(dev_priv, ring->irq_enable_mask);
+		gen6_enable_pm_irq(dev_priv, ring->irq_enable_mask);
 	}
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -1167,7 +1457,7 @@ hsw_vebox_put_irq(struct intel_engine_cs *ring)
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
 	if (--ring->irq_refcount == 0) {
 		I915_WRITE_IMR(ring, ~0);
-		snb_disable_pm_irq(dev_priv, ring->irq_enable_mask);
+		gen6_disable_pm_irq(dev_priv, ring->irq_enable_mask);
 	}
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 }
@@ -1241,54 +1531,66 @@ i965_dispatch_execbuffer(struct intel_engine_cs *ring,
 
 /* Just userspace ABI convention to limit the wa batch bo to a resonable size */
 #define I830_BATCH_LIMIT (256*1024)
+#define I830_TLB_ENTRIES (2)
+#define I830_WA_SIZE max(I830_TLB_ENTRIES*4096, I830_BATCH_LIMIT)
 static int
 i830_dispatch_execbuffer(struct intel_engine_cs *ring,
 				u64 offset, u32 len,
 				unsigned flags)
 {
+	u32 cs_offset = ring->scratch.gtt_offset;
 	int ret;
 
-	if (flags & I915_DISPATCH_PINNED) {
-		ret = intel_ring_begin(ring, 4);
-		if (ret)
-			return ret;
+	ret = intel_ring_begin(ring, 6);
+	if (ret)
+		return ret;
 
-		intel_ring_emit(ring, MI_BATCH_BUFFER);
-		intel_ring_emit(ring, offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE));
-		intel_ring_emit(ring, offset + len - 8);
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_advance(ring);
-	} else {
-		u32 cs_offset = ring->scratch.gtt_offset;
+	/* Evict the invalid PTE TLBs */
+	intel_ring_emit(ring, COLOR_BLT_CMD | BLT_WRITE_RGBA);
+	intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096);
+	intel_ring_emit(ring, I830_TLB_ENTRIES << 16 | 4); /* load each page */
+	intel_ring_emit(ring, cs_offset);
+	intel_ring_emit(ring, 0xdeadbeef);
+	intel_ring_emit(ring, MI_NOOP);
+	intel_ring_advance(ring);
 
+	if ((flags & I915_DISPATCH_PINNED) == 0) {
 		if (len > I830_BATCH_LIMIT)
 			return -ENOSPC;
 
-		ret = intel_ring_begin(ring, 9+3);
+		ret = intel_ring_begin(ring, 6 + 2);
 		if (ret)
 			return ret;
-		/* Blit the batch (which has now all relocs applied) to the stable batch
-		 * scratch bo area (so that the CS never stumbles over its tlb
-		 * invalidation bug) ... */
-		intel_ring_emit(ring, XY_SRC_COPY_BLT_CMD |
-				XY_SRC_COPY_BLT_WRITE_ALPHA |
-				XY_SRC_COPY_BLT_WRITE_RGB);
-		intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_GXCOPY | 4096);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, (DIV_ROUND_UP(len, 4096) << 16) | 1024);
+
+		/* Blit the batch (which has now all relocs applied) to the
+		 * stable batch scratch bo area (so that the CS never
+		 * stumbles over its tlb invalidation bug) ...
+		 */
+		intel_ring_emit(ring, SRC_COPY_BLT_CMD | BLT_WRITE_RGBA);
+		intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096);
+		intel_ring_emit(ring, DIV_ROUND_UP(len, 4096) << 16 | 4096);
 		intel_ring_emit(ring, cs_offset);
-		intel_ring_emit(ring, 0);
 		intel_ring_emit(ring, 4096);
 		intel_ring_emit(ring, offset);
+
 		intel_ring_emit(ring, MI_FLUSH);
+		intel_ring_emit(ring, MI_NOOP);
+		intel_ring_advance(ring);
 
 		/* ... and execute it. */
-		intel_ring_emit(ring, MI_BATCH_BUFFER);
-		intel_ring_emit(ring, cs_offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE));
-		intel_ring_emit(ring, cs_offset + len - 8);
-		intel_ring_advance(ring);
+		offset = cs_offset;
 	}
 
+	ret = intel_ring_begin(ring, 4);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_BATCH_BUFFER);
+	intel_ring_emit(ring, offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE));
+	intel_ring_emit(ring, offset + len - 8);
+	intel_ring_emit(ring, MI_NOOP);
+	intel_ring_advance(ring);
+
 	return 0;
 }
 
@@ -1329,6 +1631,7 @@ static int init_status_page(struct intel_engine_cs *ring)
 	struct drm_i915_gem_object *obj;
 
 	if ((obj = ring->status_page.obj) == NULL) {
+		unsigned flags;
 		int ret;
 
 		obj = i915_gem_alloc_object(ring->dev, 4096);
@@ -1341,7 +1644,20 @@ static int init_status_page(struct intel_engine_cs *ring)
 		if (ret)
 			goto err_unref;
 
-		ret = i915_gem_obj_ggtt_pin(obj, 4096, 0);
+		flags = 0;
+		if (!HAS_LLC(ring->dev))
+			/* On g33, we cannot place HWS above 256MiB, so
+			 * restrict its pinning to the low mappable arena.
+			 * Though this restriction is not documented for
+			 * gen4, gen5, or byt, they also behave similarly
+			 * and hang if the HWS is placed at the top of the
+			 * GTT. To generalise, it appears that all !llc
+			 * platforms have issues with us placing the HWS
+			 * above the mappable region (even though we never
+			 * actualy map it).
+			 */
+			flags |= PIN_MAPPABLE;
+		ret = i915_gem_obj_ggtt_pin(obj, 4096, flags);
 		if (ret) {
 err_unref:
 			drm_gem_object_unreference(&obj->base);
@@ -1378,15 +1694,25 @@ static int init_phys_status_page(struct intel_engine_cs *ring)
 	return 0;
 }
 
-static int allocate_ring_buffer(struct intel_engine_cs *ring)
+void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf)
+{
+	if (!ringbuf->obj)
+		return;
+
+	iounmap(ringbuf->virtual_start);
+	i915_gem_object_ggtt_unpin(ringbuf->obj);
+	drm_gem_object_unreference(&ringbuf->obj->base);
+	ringbuf->obj = NULL;
+}
+
+int intel_alloc_ringbuffer_obj(struct drm_device *dev,
+			       struct intel_ringbuffer *ringbuf)
 {
-	struct drm_device *dev = ring->dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_ringbuffer *ringbuf = ring->buffer;
 	struct drm_i915_gem_object *obj;
 	int ret;
 
-	if (intel_ring_initialized(ring))
+	if (ringbuf->obj)
 		return 0;
 
 	obj = NULL;
@@ -1397,6 +1723,9 @@ static int allocate_ring_buffer(struct intel_engine_cs *ring)
 	if (obj == NULL)
 		return -ENOMEM;
 
+	/* mark ring buffers as read-only from GPU side by default */
+	obj->gt_ro = 1;
+
 	ret = i915_gem_obj_ggtt_pin(obj, PAGE_SIZE, PIN_MAPPABLE);
 	if (ret)
 		goto err_unref;
@@ -1439,7 +1768,9 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 	ring->dev = dev;
 	INIT_LIST_HEAD(&ring->active_list);
 	INIT_LIST_HEAD(&ring->request_list);
+	INIT_LIST_HEAD(&ring->execlist_queue);
 	ringbuf->size = 32 * PAGE_SIZE;
+	ringbuf->ring = ring;
 	memset(ring->semaphore.sync_seqno, 0, sizeof(ring->semaphore.sync_seqno));
 
 	init_waitqueue_head(&ring->irq_queue);
@@ -1455,7 +1786,7 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 			goto error;
 	}
 
-	ret = allocate_ring_buffer(ring);
+	ret = intel_alloc_ringbuffer_obj(dev, ringbuf);
 	if (ret) {
 		DRM_ERROR("Failed to allocate ringbuffer %s: %d\n", ring->name, ret);
 		goto error;
@@ -1496,11 +1827,7 @@ void intel_cleanup_ring_buffer(struct intel_engine_cs *ring)
 	intel_stop_ring_buffer(ring);
 	WARN_ON(!IS_GEN2(ring->dev) && (I915_READ_MODE(ring) & MODE_IDLE) == 0);
 
-	iounmap(ringbuf->virtual_start);
-
-	i915_gem_object_ggtt_unpin(ringbuf->obj);
-	drm_gem_object_unreference(&ringbuf->obj->base);
-	ringbuf->obj = NULL;
+	intel_destroy_ringbuffer_obj(ringbuf);
 	ring->preallocated_lazy_request = NULL;
 	ring->outstanding_lazy_seqno = 0;
 
@@ -1526,13 +1853,14 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n)
 		ringbuf->head = ringbuf->last_retired_head;
 		ringbuf->last_retired_head = -1;
 
-		ringbuf->space = ring_space(ring);
+		ringbuf->space = intel_ring_space(ringbuf);
 		if (ringbuf->space >= n)
 			return 0;
 	}
 
 	list_for_each_entry(request, &ring->request_list, list) {
-		if (__ring_space(request->tail, ringbuf->tail, ringbuf->size) >= n) {
+		if (__intel_ring_space(request->tail, ringbuf->tail,
+				       ringbuf->size) >= n) {
 			seqno = request->seqno;
 			break;
 		}
@@ -1549,7 +1877,7 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n)
 	ringbuf->head = ringbuf->last_retired_head;
 	ringbuf->last_retired_head = -1;
 
-	ringbuf->space = ring_space(ring);
+	ringbuf->space = intel_ring_space(ringbuf);
 	return 0;
 }
 
@@ -1578,7 +1906,7 @@ static int ring_wait_for_space(struct intel_engine_cs *ring, int n)
 	trace_i915_ring_wait_begin(ring);
 	do {
 		ringbuf->head = I915_READ_HEAD(ring);
-		ringbuf->space = ring_space(ring);
+		ringbuf->space = intel_ring_space(ringbuf);
 		if (ringbuf->space >= n) {
 			ret = 0;
 			break;
@@ -1630,7 +1958,7 @@ static int intel_wrap_ring_buffer(struct intel_engine_cs *ring)
 		iowrite32(MI_NOOP, virt++);
 
 	ringbuf->tail = 0;
-	ringbuf->space = ring_space(ring);
+	ringbuf->space = intel_ring_space(ringbuf);
 
 	return 0;
 }
@@ -1746,14 +2074,15 @@ int intel_ring_cacheline_align(struct intel_engine_cs *ring)
 
 void intel_ring_init_seqno(struct intel_engine_cs *ring, u32 seqno)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	BUG_ON(ring->outstanding_lazy_seqno);
 
-	if (INTEL_INFO(ring->dev)->gen >= 6) {
+	if (INTEL_INFO(dev)->gen == 6 || INTEL_INFO(dev)->gen == 7) {
 		I915_WRITE(RING_SYNC_0(ring->mmio_base), 0);
 		I915_WRITE(RING_SYNC_1(ring->mmio_base), 0);
-		if (HAS_VEBOX(ring->dev))
+		if (HAS_VEBOX(dev))
 			I915_WRITE(RING_SYNC_2(ring->mmio_base), 0);
 	}
 
@@ -1834,9 +2163,7 @@ gen8_ring_dispatch_execbuffer(struct intel_engine_cs *ring,
 			      u64 offset, u32 len,
 			      unsigned flags)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	bool ppgtt = dev_priv->mm.aliasing_ppgtt != NULL &&
-		!(flags & I915_DISPATCH_SECURE);
+	bool ppgtt = USES_PPGTT(ring->dev) && !(flags & I915_DISPATCH_SECURE);
 	int ret;
 
 	ret = intel_ring_begin(ring, 4);
@@ -1865,8 +2192,9 @@ hsw_ring_dispatch_execbuffer(struct intel_engine_cs *ring,
 		return ret;
 
 	intel_ring_emit(ring,
-			MI_BATCH_BUFFER_START | MI_BATCH_PPGTT_HSW |
-			(flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE_HSW));
+			MI_BATCH_BUFFER_START |
+			(flags & I915_DISPATCH_SECURE ?
+			 0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW));
 	/* bit0-7 is the length on GEN6+ */
 	intel_ring_emit(ring, offset);
 	intel_ring_advance(ring);
@@ -1941,45 +2269,78 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring = &dev_priv->ring[RCS];
+	struct drm_i915_gem_object *obj;
+	int ret;
 
 	ring->name = "render ring";
 	ring->id = RCS;
 	ring->mmio_base = RENDER_RING_BASE;
 
-	if (INTEL_INFO(dev)->gen >= 6) {
+	if (INTEL_INFO(dev)->gen >= 8) {
+		if (i915_semaphore_is_enabled(dev)) {
+			obj = i915_gem_alloc_object(dev, 4096);
+			if (obj == NULL) {
+				DRM_ERROR("Failed to allocate semaphore bo. Disabling semaphores\n");
+				i915.semaphores = 0;
+			} else {
+				i915_gem_object_set_cache_level(obj, I915_CACHE_LLC);
+				ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_NONBLOCK);
+				if (ret != 0) {
+					drm_gem_object_unreference(&obj->base);
+					DRM_ERROR("Failed to pin semaphore bo. Disabling semaphores\n");
+					i915.semaphores = 0;
+				} else
+					dev_priv->semaphore_obj = obj;
+			}
+		}
+		if (IS_CHERRYVIEW(dev))
+			ring->init_context = chv_init_workarounds;
+		else
+			ring->init_context = bdw_init_workarounds;
+		ring->add_request = gen6_add_request;
+		ring->flush = gen8_render_ring_flush;
+		ring->irq_get = gen8_ring_get_irq;
+		ring->irq_put = gen8_ring_put_irq;
+		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
+		ring->get_seqno = gen6_ring_get_seqno;
+		ring->set_seqno = ring_set_seqno;
+		if (i915_semaphore_is_enabled(dev)) {
+			WARN_ON(!dev_priv->semaphore_obj);
+			ring->semaphore.sync_to = gen8_ring_sync;
+			ring->semaphore.signal = gen8_rcs_signal;
+			GEN8_RING_SEMAPHORE_INIT;
+		}
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		ring->add_request = gen6_add_request;
 		ring->flush = gen7_render_ring_flush;
 		if (INTEL_INFO(dev)->gen == 6)
 			ring->flush = gen6_render_ring_flush;
-		if (INTEL_INFO(dev)->gen >= 8) {
-			ring->flush = gen8_render_ring_flush;
-			ring->irq_get = gen8_ring_get_irq;
-			ring->irq_put = gen8_ring_put_irq;
-		} else {
-			ring->irq_get = gen6_ring_get_irq;
-			ring->irq_put = gen6_ring_put_irq;
-		}
+		ring->irq_get = gen6_ring_get_irq;
+		ring->irq_put = gen6_ring_put_irq;
 		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
 		ring->get_seqno = gen6_ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
-		ring->semaphore.sync_to = gen6_ring_sync;
-		ring->semaphore.signal = gen6_signal;
-		/*
-		 * The current semaphore is only applied on pre-gen8 platform.
-		 * And there is no VCS2 ring on the pre-gen8 platform. So the
-		 * semaphore between RCS and VCS2 is initialized as INVALID.
-		 * Gen8 will initialize the sema between VCS2 and RCS later.
-		 */
-		ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_RV;
-		ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_RB;
-		ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_RVE;
-		ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC;
-		ring->semaphore.mbox.signal[VCS] = GEN6_VRSYNC;
-		ring->semaphore.mbox.signal[BCS] = GEN6_BRSYNC;
-		ring->semaphore.mbox.signal[VECS] = GEN6_VERSYNC;
-		ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
+		if (i915_semaphore_is_enabled(dev)) {
+			ring->semaphore.sync_to = gen6_ring_sync;
+			ring->semaphore.signal = gen6_signal;
+			/*
+			 * The current semaphore is only applied on pre-gen8
+			 * platform.  And there is no VCS2 ring on the pre-gen8
+			 * platform. So the semaphore between RCS and VCS2 is
+			 * initialized as INVALID.  Gen8 will initialize the
+			 * sema between VCS2 and RCS later.
+			 */
+			ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_RV;
+			ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_RB;
+			ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_RVE;
+			ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC;
+			ring->semaphore.mbox.signal[VCS] = GEN6_VRSYNC;
+			ring->semaphore.mbox.signal[BCS] = GEN6_BRSYNC;
+			ring->semaphore.mbox.signal[VECS] = GEN6_VERSYNC;
+			ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
+		}
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->flush = gen4_render_ring_flush;
@@ -2007,6 +2368,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		ring->irq_enable_mask = I915_USER_INTERRUPT;
 	}
 	ring->write_tail = ring_write_tail;
+
 	if (IS_HASWELL(dev))
 		ring->dispatch_execbuffer = hsw_ring_dispatch_execbuffer;
 	else if (IS_GEN8(dev))
@@ -2024,10 +2386,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 
 	/* Workaround batchbuffer to combat CS tlb bug. */
 	if (HAS_BROKEN_CS_TLB(dev)) {
-		struct drm_i915_gem_object *obj;
-		int ret;
-
-		obj = i915_gem_alloc_object(dev, I830_BATCH_LIMIT);
+		obj = i915_gem_alloc_object(dev, I830_WA_SIZE);
 		if (obj == NULL) {
 			DRM_ERROR("Failed to allocate batch bo\n");
 			return -ENOMEM;
@@ -2157,31 +2516,32 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 			ring->irq_put = gen8_ring_put_irq;
 			ring->dispatch_execbuffer =
 				gen8_ring_dispatch_execbuffer;
+			if (i915_semaphore_is_enabled(dev)) {
+				ring->semaphore.sync_to = gen8_ring_sync;
+				ring->semaphore.signal = gen8_xcs_signal;
+				GEN8_RING_SEMAPHORE_INIT;
+			}
 		} else {
 			ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
 			ring->irq_get = gen6_ring_get_irq;
 			ring->irq_put = gen6_ring_put_irq;
 			ring->dispatch_execbuffer =
 				gen6_ring_dispatch_execbuffer;
+			if (i915_semaphore_is_enabled(dev)) {
+				ring->semaphore.sync_to = gen6_ring_sync;
+				ring->semaphore.signal = gen6_signal;
+				ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VR;
+				ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID;
+				ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VB;
+				ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_VVE;
+				ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
+				ring->semaphore.mbox.signal[RCS] = GEN6_RVSYNC;
+				ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC;
+				ring->semaphore.mbox.signal[BCS] = GEN6_BVSYNC;
+				ring->semaphore.mbox.signal[VECS] = GEN6_VEVSYNC;
+				ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
+			}
 		}
-		ring->semaphore.sync_to = gen6_ring_sync;
-		ring->semaphore.signal = gen6_signal;
-		/*
-		 * The current semaphore is only applied on pre-gen8 platform.
-		 * And there is no VCS2 ring on the pre-gen8 platform. So the
-		 * semaphore between VCS and VCS2 is initialized as INVALID.
-		 * Gen8 will initialize the sema between VCS2 and VCS later.
-		 */
-		ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VR;
-		ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VB;
-		ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_VVE;
-		ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore.mbox.signal[RCS] = GEN6_RVSYNC;
-		ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC;
-		ring->semaphore.mbox.signal[BCS] = GEN6_BVSYNC;
-		ring->semaphore.mbox.signal[VECS] = GEN6_VEVSYNC;
-		ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
 	} else {
 		ring->mmio_base = BSD_RING_BASE;
 		ring->flush = bsd_ring_flush;
@@ -2218,7 +2578,7 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev)
 		return -EINVAL;
 	}
 
-	ring->name = "bds2_ring";
+	ring->name = "bsd2 ring";
 	ring->id = VCS2;
 
 	ring->write_tail = ring_write_tail;
@@ -2233,25 +2593,11 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev)
 	ring->irq_put = gen8_ring_put_irq;
 	ring->dispatch_execbuffer =
 			gen8_ring_dispatch_execbuffer;
-	ring->semaphore.sync_to = gen6_ring_sync;
-	ring->semaphore.signal = gen6_signal;
-	/*
-	 * The current semaphore is only applied on the pre-gen8. And there
-	 * is no bsd2 ring on the pre-gen8. So now the semaphore_register
-	 * between VCS2 and other ring is initialized as invalid.
-	 * Gen8 will initialize the sema between VCS2 and other ring later.
-	 */
-	ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
-
+	if (i915_semaphore_is_enabled(dev)) {
+		ring->semaphore.sync_to = gen8_ring_sync;
+		ring->semaphore.signal = gen8_xcs_signal;
+		GEN8_RING_SEMAPHORE_INIT;
+	}
 	ring->init = init_ring_common;
 
 	return intel_init_ring_buffer(dev, ring);
@@ -2277,30 +2623,38 @@ int intel_init_blt_ring_buffer(struct drm_device *dev)
 		ring->irq_get = gen8_ring_get_irq;
 		ring->irq_put = gen8_ring_put_irq;
 		ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
+		if (i915_semaphore_is_enabled(dev)) {
+			ring->semaphore.sync_to = gen8_ring_sync;
+			ring->semaphore.signal = gen8_xcs_signal;
+			GEN8_RING_SEMAPHORE_INIT;
+		}
 	} else {
 		ring->irq_enable_mask = GT_BLT_USER_INTERRUPT;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
 		ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
+		if (i915_semaphore_is_enabled(dev)) {
+			ring->semaphore.signal = gen6_signal;
+			ring->semaphore.sync_to = gen6_ring_sync;
+			/*
+			 * The current semaphore is only applied on pre-gen8
+			 * platform.  And there is no VCS2 ring on the pre-gen8
+			 * platform. So the semaphore between BCS and VCS2 is
+			 * initialized as INVALID.  Gen8 will initialize the
+			 * sema between BCS and VCS2 later.
+			 */
+			ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_BR;
+			ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_BV;
+			ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_BVE;
+			ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.signal[RCS] = GEN6_RBSYNC;
+			ring->semaphore.mbox.signal[VCS] = GEN6_VBSYNC;
+			ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC;
+			ring->semaphore.mbox.signal[VECS] = GEN6_VEBSYNC;
+			ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
+		}
 	}
-	ring->semaphore.sync_to = gen6_ring_sync;
-	ring->semaphore.signal = gen6_signal;
-	/*
-	 * The current semaphore is only applied on pre-gen8 platform. And
-	 * there is no VCS2 ring on the pre-gen8 platform. So the semaphore
-	 * between BCS and VCS2 is initialized as INVALID.
-	 * Gen8 will initialize the sema between BCS and VCS2 later.
-	 */
-	ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_BR;
-	ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_BV;
-	ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_BVE;
-	ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.signal[RCS] = GEN6_RBSYNC;
-	ring->semaphore.mbox.signal[VCS] = GEN6_VBSYNC;
-	ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[VECS] = GEN6_VEBSYNC;
-	ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
 	ring->init = init_ring_common;
 
 	return intel_init_ring_buffer(dev, ring);
@@ -2327,24 +2681,31 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev)
 		ring->irq_get = gen8_ring_get_irq;
 		ring->irq_put = gen8_ring_put_irq;
 		ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer;
+		if (i915_semaphore_is_enabled(dev)) {
+			ring->semaphore.sync_to = gen8_ring_sync;
+			ring->semaphore.signal = gen8_xcs_signal;
+			GEN8_RING_SEMAPHORE_INIT;
+		}
 	} else {
 		ring->irq_enable_mask = PM_VEBOX_USER_INTERRUPT;
 		ring->irq_get = hsw_vebox_get_irq;
 		ring->irq_put = hsw_vebox_put_irq;
 		ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
+		if (i915_semaphore_is_enabled(dev)) {
+			ring->semaphore.sync_to = gen6_ring_sync;
+			ring->semaphore.signal = gen6_signal;
+			ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VER;
+			ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_VEV;
+			ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VEB;
+			ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
+			ring->semaphore.mbox.signal[RCS] = GEN6_RVESYNC;
+			ring->semaphore.mbox.signal[VCS] = GEN6_VVESYNC;
+			ring->semaphore.mbox.signal[BCS] = GEN6_BVESYNC;
+			ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC;
+			ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
+		}
 	}
-	ring->semaphore.sync_to = gen6_ring_sync;
-	ring->semaphore.signal = gen6_signal;
-	ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VER;
-	ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_VEV;
-	ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VEB;
-	ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->semaphore.mbox.signal[RCS] = GEN6_RVESYNC;
-	ring->semaphore.mbox.signal[VCS] = GEN6_VVESYNC;
-	ring->semaphore.mbox.signal[BCS] = GEN6_BVESYNC;
-	ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC;
-	ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC;
 	ring->init = init_ring_common;
 
 	return intel_init_ring_buffer(dev, ring);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index e72017bdcd7f..96479c89f4bd 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -5,6 +5,13 @@
 
 #define I915_CMD_HASH_ORDER 9
 
+/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
+ * but keeps the logic simple. Indeed, the whole purpose of this macro is just
+ * to give some inclination as to some of the magic values used in the various
+ * workarounds!
+ */
+#define CACHELINE_BYTES 64
+
 /*
  * Gen2 BSpec "1. Programming Environment" / 1.4.4.6 "Ring Buffer Use"
  * Gen3 BSpec "vol1c Memory Interface Functions" / 2.3.4.5 "Ring Buffer Use"
@@ -40,10 +47,37 @@ struct  intel_hw_status_page {
 #define I915_READ_MODE(ring) I915_READ(RING_MI_MODE((ring)->mmio_base))
 #define I915_WRITE_MODE(ring, val) I915_WRITE(RING_MI_MODE((ring)->mmio_base), val)
 
+/* seqno size is actually only a uint32, but since we plan to use MI_FLUSH_DW to
+ * do the writes, and that must have qw aligned offsets, simply pretend it's 8b.
+ */
+#define i915_semaphore_seqno_size sizeof(uint64_t)
+#define GEN8_SIGNAL_OFFSET(__ring, to)			     \
+	(i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj) + \
+	((__ring)->id * I915_NUM_RINGS * i915_semaphore_seqno_size) +	\
+	(i915_semaphore_seqno_size * (to)))
+
+#define GEN8_WAIT_OFFSET(__ring, from)			     \
+	(i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj) + \
+	((from) * I915_NUM_RINGS * i915_semaphore_seqno_size) + \
+	(i915_semaphore_seqno_size * (__ring)->id))
+
+#define GEN8_RING_SEMAPHORE_INIT do { \
+	if (!dev_priv->semaphore_obj) { \
+		break; \
+	} \
+	ring->semaphore.signal_ggtt[RCS] = GEN8_SIGNAL_OFFSET(ring, RCS); \
+	ring->semaphore.signal_ggtt[VCS] = GEN8_SIGNAL_OFFSET(ring, VCS); \
+	ring->semaphore.signal_ggtt[BCS] = GEN8_SIGNAL_OFFSET(ring, BCS); \
+	ring->semaphore.signal_ggtt[VECS] = GEN8_SIGNAL_OFFSET(ring, VECS); \
+	ring->semaphore.signal_ggtt[VCS2] = GEN8_SIGNAL_OFFSET(ring, VCS2); \
+	ring->semaphore.signal_ggtt[ring->id] = MI_SEMAPHORE_SYNC_INVALID; \
+	} while(0)
+
 enum intel_ring_hangcheck_action {
 	HANGCHECK_IDLE = 0,
 	HANGCHECK_WAIT,
 	HANGCHECK_ACTIVE,
+	HANGCHECK_ACTIVE_LOOP,
 	HANGCHECK_KICK,
 	HANGCHECK_HUNG,
 };
@@ -52,6 +86,7 @@ enum intel_ring_hangcheck_action {
 
 struct intel_ring_hangcheck {
 	u64 acthd;
+	u64 max_acthd;
 	u32 seqno;
 	int score;
 	enum intel_ring_hangcheck_action action;
@@ -62,6 +97,15 @@ struct intel_ringbuffer {
 	struct drm_i915_gem_object *obj;
 	void __iomem *virtual_start;
 
+	struct intel_engine_cs *ring;
+
+	/*
+	 * FIXME: This backpointer is an artifact of the history of how the
+	 * execlist patches came into being. It will get removed once the basic
+	 * code has landed.
+	 */
+	struct intel_context *FIXME_lrc_ctx;
+
 	u32 head;
 	u32 tail;
 	int space;
@@ -104,6 +148,8 @@ struct  intel_engine_cs {
 
 	int		(*init)(struct intel_engine_cs *ring);
 
+	int		(*init_context)(struct intel_engine_cs *ring);
+
 	void		(*write_tail)(struct intel_engine_cs *ring,
 				      u32 value);
 	int __must_check (*flush)(struct intel_engine_cs *ring,
@@ -127,15 +173,55 @@ struct  intel_engine_cs {
 #define I915_DISPATCH_PINNED 0x2
 	void		(*cleanup)(struct intel_engine_cs *ring);
 
+	/* GEN8 signal/wait table - never trust comments!
+	 *	  signal to	signal to    signal to   signal to      signal to
+	 *	    RCS		   VCS          BCS        VECS		 VCS2
+	 *      --------------------------------------------------------------------
+	 *  RCS | NOP (0x00) | VCS (0x08) | BCS (0x10) | VECS (0x18) | VCS2 (0x20) |
+	 *	|-------------------------------------------------------------------
+	 *  VCS | RCS (0x28) | NOP (0x30) | BCS (0x38) | VECS (0x40) | VCS2 (0x48) |
+	 *	|-------------------------------------------------------------------
+	 *  BCS | RCS (0x50) | VCS (0x58) | NOP (0x60) | VECS (0x68) | VCS2 (0x70) |
+	 *	|-------------------------------------------------------------------
+	 * VECS | RCS (0x78) | VCS (0x80) | BCS (0x88) |  NOP (0x90) | VCS2 (0x98) |
+	 *	|-------------------------------------------------------------------
+	 * VCS2 | RCS (0xa0) | VCS (0xa8) | BCS (0xb0) | VECS (0xb8) | NOP  (0xc0) |
+	 *	|-------------------------------------------------------------------
+	 *
+	 * Generalization:
+	 *  f(x, y) := (x->id * NUM_RINGS * seqno_size) + (seqno_size * y->id)
+	 *  ie. transpose of g(x, y)
+	 *
+	 *	 sync from	sync from    sync from    sync from	sync from
+	 *	    RCS		   VCS          BCS        VECS		 VCS2
+	 *      --------------------------------------------------------------------
+	 *  RCS | NOP (0x00) | VCS (0x28) | BCS (0x50) | VECS (0x78) | VCS2 (0xa0) |
+	 *	|-------------------------------------------------------------------
+	 *  VCS | RCS (0x08) | NOP (0x30) | BCS (0x58) | VECS (0x80) | VCS2 (0xa8) |
+	 *	|-------------------------------------------------------------------
+	 *  BCS | RCS (0x10) | VCS (0x38) | NOP (0x60) | VECS (0x88) | VCS2 (0xb0) |
+	 *	|-------------------------------------------------------------------
+	 * VECS | RCS (0x18) | VCS (0x40) | BCS (0x68) |  NOP (0x90) | VCS2 (0xb8) |
+	 *	|-------------------------------------------------------------------
+	 * VCS2 | RCS (0x20) | VCS (0x48) | BCS (0x70) | VECS (0x98) |  NOP (0xc0) |
+	 *	|-------------------------------------------------------------------
+	 *
+	 * Generalization:
+	 *  g(x, y) := (y->id * NUM_RINGS * seqno_size) + (seqno_size * x->id)
+	 *  ie. transpose of f(x, y)
+	 */
 	struct {
 		u32	sync_seqno[I915_NUM_RINGS-1];
 
-		struct {
-			/* our mbox written by others */
-			u32		wait[I915_NUM_RINGS];
-			/* mboxes this ring signals to */
-			u32		signal[I915_NUM_RINGS];
-		} mbox;
+		union {
+			struct {
+				/* our mbox written by others */
+				u32		wait[I915_NUM_RINGS];
+				/* mboxes this ring signals to */
+				u32		signal[I915_NUM_RINGS];
+			} mbox;
+			u64		signal_ggtt[I915_NUM_RINGS];
+		};
 
 		/* AKA wait() */
 		int	(*sync_to)(struct intel_engine_cs *ring,
@@ -146,6 +232,18 @@ struct  intel_engine_cs {
 				  unsigned int num_dwords);
 	} semaphore;
 
+	/* Execlists */
+	spinlock_t execlist_lock;
+	struct list_head execlist_queue;
+	u8 next_context_status_buffer;
+	u32             irq_keep_mask; /* bitmask for interrupts that should not be masked */
+	int		(*emit_request)(struct intel_ringbuffer *ringbuf);
+	int		(*emit_flush)(struct intel_ringbuffer *ringbuf,
+				      u32 invalidate_domains,
+				      u32 flush_domains);
+	int		(*emit_bb_start)(struct intel_ringbuffer *ringbuf,
+					 u64 offset, unsigned flags);
+
 	/**
 	 * List of objects currently involved in rendering from the
 	 * ringbuffer.
@@ -219,11 +317,7 @@ struct  intel_engine_cs {
 	u32 (*get_cmd_length_mask)(u32 cmd_header);
 };
 
-static inline bool
-intel_ring_initialized(struct intel_engine_cs *ring)
-{
-	return ring->buffer && ring->buffer->obj;
-}
+bool intel_ring_initialized(struct intel_engine_cs *ring);
 
 static inline unsigned
 intel_ring_flag(struct intel_engine_cs *ring)
@@ -238,9 +332,11 @@ intel_ring_sync_index(struct intel_engine_cs *ring,
 	int idx;
 
 	/*
-	 * cs -> 0 = vcs, 1 = bcs
-	 * vcs -> 0 = bcs, 1 = cs,
-	 * bcs -> 0 = cs, 1 = vcs.
+	 * rcs -> 0 = vcs, 1 = bcs, 2 = vecs, 3 = vcs2;
+	 * vcs -> 0 = bcs, 1 = vecs, 2 = vcs2, 3 = rcs;
+	 * bcs -> 0 = vecs, 1 = vcs2. 2 = rcs, 3 = vcs;
+	 * vecs -> 0 = vcs2, 1 = rcs, 2 = vcs, 3 = bcs;
+	 * vcs2 -> 0 = rcs, 1 = vcs, 2 = bcs, 3 = vecs;
 	 */
 
 	idx = (other - ring) - 1;
@@ -285,6 +381,10 @@ intel_write_status_page(struct intel_engine_cs *ring,
 #define I915_GEM_HWS_SCRATCH_INDEX	0x30
 #define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT)
 
+void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf);
+int intel_alloc_ringbuffer_obj(struct drm_device *dev,
+			       struct intel_ringbuffer *ringbuf);
+
 void intel_stop_ring_buffer(struct intel_engine_cs *ring);
 void intel_cleanup_ring_buffer(struct intel_engine_cs *ring);
 
@@ -302,6 +402,9 @@ static inline void intel_ring_advance(struct intel_engine_cs *ring)
 	struct intel_ringbuffer *ringbuf = ring->buffer;
 	ringbuf->tail &= ringbuf->size - 1;
 }
+int __intel_ring_space(int head, int tail, int size);
+int intel_ring_space(struct intel_ringbuffer *ringbuf);
+bool intel_ring_stopped(struct intel_engine_cs *ring);
 void __intel_ring_advance(struct intel_engine_cs *ring);
 
 int __must_check intel_ring_idle(struct intel_engine_cs *ring);
@@ -309,6 +412,9 @@ void intel_ring_init_seqno(struct intel_engine_cs *ring, u32 seqno);
 int intel_ring_flush_all_caches(struct intel_engine_cs *ring);
 int intel_ring_invalidate_all_caches(struct intel_engine_cs *ring);
 
+void intel_fini_pipe_control(struct intel_engine_cs *ring);
+int intel_init_pipe_control(struct intel_engine_cs *ring);
+
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
 int intel_init_bsd2_ring_buffer(struct drm_device *dev);
@@ -318,9 +424,9 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev);
 u64 intel_ring_get_active_head(struct intel_engine_cs *ring);
 void intel_ring_setup_status_page(struct intel_engine_cs *ring);
 
-static inline u32 intel_ring_get_tail(struct intel_engine_cs *ring)
+static inline u32 intel_ring_get_tail(struct intel_ringbuffer *ringbuf)
 {
-	return ring->buffer->tail;
+	return ringbuf->tail;
 }
 
 static inline u32 intel_ring_get_seqno(struct intel_engine_cs *ring)
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 20375cc7f82d..9350edd6728d 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -2433,7 +2433,7 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 	connector->base.unregister = intel_sdvo_connector_unregister;
 
 	intel_connector_attach_encoder(&connector->base, &encoder->base);
-	ret = drm_sysfs_connector_add(drm_connector);
+	ret = drm_connector_register(drm_connector);
 	if (ret < 0)
 		goto err1;
 
@@ -2446,7 +2446,7 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 	return 0;
 
 err2:
-	drm_sysfs_connector_remove(drm_connector);
+	drm_connector_unregister(drm_connector);
 err1:
 	drm_connector_cleanup(drm_connector);
 
@@ -2559,7 +2559,7 @@ intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type)
 	return true;
 
 err:
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	intel_sdvo_destroy(connector);
 	return false;
 }
@@ -2638,7 +2638,7 @@ intel_sdvo_lvds_init(struct intel_sdvo *intel_sdvo, int device)
 	return true;
 
 err:
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	intel_sdvo_destroy(connector);
 	return false;
 }
@@ -2711,7 +2711,7 @@ static void intel_sdvo_output_cleanup(struct intel_sdvo *intel_sdvo)
 	list_for_each_entry_safe(connector, tmp,
 				 &dev->mode_config.connector_list, head) {
 		if (intel_attached_encoder(connector) == &intel_sdvo->base) {
-			drm_sysfs_connector_remove(connector);
+			drm_connector_unregister(connector);
 			intel_sdvo_destroy(connector);
 		}
 	}
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 9a17b4e92ef4..07a74ef589bd 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -53,6 +53,7 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 	enum pipe pipe = crtc->pipe;
 	long timeout = msecs_to_jiffies_timeout(1);
 	int scanline, min, max, vblank_start;
+	wait_queue_head_t *wq = drm_crtc_vblank_waitqueue(&crtc->base);
 	DEFINE_WAIT(wait);
 
 	WARN_ON(!drm_modeset_is_locked(&crtc->base.mutex));
@@ -81,7 +82,7 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 		 * other CPUs can see the task state update by the time we
 		 * read the scanline.
 		 */
-		prepare_to_wait(&crtc->vbl_wait, &wait, TASK_UNINTERRUPTIBLE);
+		prepare_to_wait(wq, &wait, TASK_UNINTERRUPTIBLE);
 
 		scanline = intel_get_crtc_scanline(crtc);
 		if (scanline < min || scanline > max)
@@ -100,7 +101,7 @@ static bool intel_pipe_update_start(struct intel_crtc *crtc, uint32_t *start_vbl
 		local_irq_disable();
 	}
 
-	finish_wait(&crtc->vbl_wait, &wait);
+	finish_wait(wq, &wait);
 
 	drm_vblank_put(dev, pipe);
 
@@ -163,6 +164,7 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 	sprctl &= ~SP_PIXFORMAT_MASK;
 	sprctl &= ~SP_YUV_BYTE_ORDER_MASK;
 	sprctl &= ~SP_TILED;
+	sprctl &= ~SP_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_YUYV:
@@ -218,7 +220,8 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 
 	sprctl |= SP_ENABLE;
 
-	intel_update_sprite_watermarks(dplane, crtc, src_w, pixel_size, true,
+	intel_update_sprite_watermarks(dplane, crtc, src_w, src_h,
+				       pixel_size, true,
 				       src_w != crtc_w || src_h != crtc_h);
 
 	/* Sizes are 0 based */
@@ -234,6 +237,14 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc,
 							fb->pitches[0]);
 	linear_offset -= sprsurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		sprctl |= SP_ROTATE_180;
+
+		x += src_w;
+		y += src_h;
+		linear_offset += src_h * fb->pitches[0] + src_w * pixel_size;
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
@@ -283,7 +294,7 @@ vlv_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc)
 	if (atomic_update)
 		intel_pipe_update_end(intel_crtc, start_vbl_count);
 
-	intel_update_sprite_watermarks(dplane, crtc, 0, 0, false, false);
+	intel_update_sprite_watermarks(dplane, crtc, 0, 0, 0, false, false);
 }
 
 static int
@@ -363,6 +374,7 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	sprctl &= ~SPRITE_RGB_ORDER_RGBX;
 	sprctl &= ~SPRITE_YUV_BYTE_ORDER_MASK;
 	sprctl &= ~SPRITE_TILED;
+	sprctl &= ~SPRITE_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_XBGR8888:
@@ -406,7 +418,8 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		sprctl |= SPRITE_PIPE_CSC_ENABLE;
 
-	intel_update_sprite_watermarks(plane, crtc, src_w, pixel_size, true,
+	intel_update_sprite_watermarks(plane, crtc, src_w, src_h, pixel_size,
+				       true,
 				       src_w != crtc_w || src_h != crtc_h);
 
 	/* Sizes are 0 based */
@@ -424,6 +437,18 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 					       pixel_size, fb->pitches[0]);
 	linear_offset -= sprsurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		sprctl |= SPRITE_ROTATE_180;
+
+		/* HSW and BDW does this automagically in hardware */
+		if (!IS_HASWELL(dev) && !IS_BROADWELL(dev)) {
+			x += src_w;
+			y += src_h;
+			linear_offset += src_h * fb->pitches[0] +
+				src_w * pixel_size;
+		}
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
@@ -486,7 +511,7 @@ ivb_disable_plane(struct drm_plane *plane, struct drm_crtc *crtc)
 	 */
 	intel_wait_for_vblank(dev, pipe);
 
-	intel_update_sprite_watermarks(plane, crtc, 0, 0, false, false);
+	intel_update_sprite_watermarks(plane, crtc, 0, 0, 0, false, false);
 }
 
 static int
@@ -569,6 +594,7 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	dvscntr &= ~DVS_RGB_ORDER_XBGR;
 	dvscntr &= ~DVS_YUV_BYTE_ORDER_MASK;
 	dvscntr &= ~DVS_TILED;
+	dvscntr &= ~DVS_ROTATE_180;
 
 	switch (fb->pixel_format) {
 	case DRM_FORMAT_XBGR8888:
@@ -606,7 +632,8 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		dvscntr |= DVS_TRICKLE_FEED_DISABLE; /* must disable */
 	dvscntr |= DVS_ENABLE;
 
-	intel_update_sprite_watermarks(plane, crtc, src_w, pixel_size, true,
+	intel_update_sprite_watermarks(plane, crtc, src_w, src_h,
+				       pixel_size, true,
 				       src_w != crtc_w || src_h != crtc_h);
 
 	/* Sizes are 0 based */
@@ -625,6 +652,14 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 					       pixel_size, fb->pitches[0]);
 	linear_offset -= dvssurf_offset;
 
+	if (intel_plane->rotation == BIT(DRM_ROTATE_180)) {
+		dvscntr |= DVS_ROTATE_180;
+
+		x += src_w;
+		y += src_h;
+		linear_offset += src_h * fb->pitches[0] + src_w * pixel_size;
+	}
+
 	atomic_update = intel_pipe_update_start(intel_crtc, &start_vbl_count);
 
 	intel_update_primary_plane(intel_crtc);
@@ -681,7 +716,7 @@ ilk_disable_plane(struct drm_plane *plane, struct drm_crtc *crtc)
 	 */
 	intel_wait_for_vblank(dev, pipe);
 
-	intel_update_sprite_watermarks(plane, crtc, 0, 0, false, false);
+	intel_update_sprite_watermarks(plane, crtc, 0, 0, 0, false, false);
 }
 
 static void
@@ -819,6 +854,7 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	struct drm_device *dev = plane->dev;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_plane *intel_plane = to_intel_plane(plane);
+	enum pipe pipe = intel_crtc->pipe;
 	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
 	struct drm_i915_gem_object *obj = intel_fb->obj;
 	struct drm_i915_gem_object *old_obj = intel_plane->obj;
@@ -891,6 +927,9 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	max_scale = intel_plane->max_downscale << 16;
 	min_scale = intel_plane->can_scale ? 1 : (1 << 16);
 
+	drm_rect_rotate(&src, fb->width << 16, fb->height << 16,
+			intel_plane->rotation);
+
 	hscale = drm_rect_calc_hscale_relaxed(&src, &dst, min_scale, max_scale);
 	BUG_ON(hscale < 0);
 
@@ -929,6 +968,9 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 				     drm_rect_width(&dst) * hscale - drm_rect_width(&src),
 				     drm_rect_height(&dst) * vscale - drm_rect_height(&src));
 
+		drm_rect_rotate_inv(&src, fb->width << 16, fb->height << 16,
+				    intel_plane->rotation);
+
 		/* sanity check to make sure the src viewport wasn't enlarged */
 		WARN_ON(src.x1 < (int) src_x ||
 			src.y1 < (int) src_y ||
@@ -1006,6 +1048,8 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	 */
 	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
 
+	i915_gem_track_fb(old_obj, obj,
+			  INTEL_FRONTBUFFER_SPRITE(pipe));
 	mutex_unlock(&dev->struct_mutex);
 
 	if (ret)
@@ -1039,6 +1083,8 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		else
 			intel_plane->disable_plane(plane, crtc);
 
+		intel_frontbuffer_flip(dev, INTEL_FRONTBUFFER_SPRITE(pipe));
+
 		if (!primary_was_enabled && primary_enabled)
 			intel_post_enable_primary(crtc);
 	}
@@ -1068,6 +1114,7 @@ intel_disable_plane(struct drm_plane *plane)
 	struct drm_device *dev = plane->dev;
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	struct intel_crtc *intel_crtc;
+	enum pipe pipe;
 
 	if (!plane->fb)
 		return 0;
@@ -1076,6 +1123,7 @@ intel_disable_plane(struct drm_plane *plane)
 		return -EINVAL;
 
 	intel_crtc = to_intel_crtc(plane->crtc);
+	pipe = intel_crtc->pipe;
 
 	if (intel_crtc->active) {
 		bool primary_was_enabled = intel_crtc->primary_enabled;
@@ -1094,6 +1142,8 @@ intel_disable_plane(struct drm_plane *plane)
 
 		mutex_lock(&dev->struct_mutex);
 		intel_unpin_fb_obj(intel_plane->obj);
+		i915_gem_track_fb(intel_plane->obj, NULL,
+				  INTEL_FRONTBUFFER_SPRITE(pipe));
 		mutex_unlock(&dev->struct_mutex);
 
 		intel_plane->obj = NULL;
@@ -1114,7 +1164,6 @@ int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv)
 {
 	struct drm_intel_sprite_colorkey *set = data;
-	struct drm_mode_object *obj;
 	struct drm_plane *plane;
 	struct intel_plane *intel_plane;
 	int ret = 0;
@@ -1128,13 +1177,12 @@ int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
-	obj = drm_mode_object_find(dev, set->plane_id, DRM_MODE_OBJECT_PLANE);
-	if (!obj) {
+	plane = drm_plane_find(dev, set->plane_id);
+	if (!plane) {
 		ret = -ENOENT;
 		goto out_unlock;
 	}
 
-	plane = obj_to_plane(obj);
 	intel_plane = to_intel_plane(plane);
 	ret = intel_plane->update_colorkey(plane, set);
 
@@ -1147,7 +1195,6 @@ int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv)
 {
 	struct drm_intel_sprite_colorkey *get = data;
-	struct drm_mode_object *obj;
 	struct drm_plane *plane;
 	struct intel_plane *intel_plane;
 	int ret = 0;
@@ -1157,13 +1204,12 @@ int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 
 	drm_modeset_lock_all(dev);
 
-	obj = drm_mode_object_find(dev, get->plane_id, DRM_MODE_OBJECT_PLANE);
-	if (!obj) {
+	plane = drm_plane_find(dev, get->plane_id);
+	if (!plane) {
 		ret = -ENOENT;
 		goto out_unlock;
 	}
 
-	plane = obj_to_plane(obj);
 	intel_plane = to_intel_plane(plane);
 	intel_plane->get_colorkey(plane, get);
 
@@ -1172,18 +1218,45 @@ out_unlock:
 	return ret;
 }
 
-void intel_plane_restore(struct drm_plane *plane)
+int intel_plane_set_property(struct drm_plane *plane,
+			     struct drm_property *prop,
+			     uint64_t val)
+{
+	struct drm_device *dev = plane->dev;
+	struct intel_plane *intel_plane = to_intel_plane(plane);
+	uint64_t old_val;
+	int ret = -ENOENT;
+
+	if (prop == dev->mode_config.rotation_property) {
+		/* exactly one rotation angle please */
+		if (hweight32(val & 0xf) != 1)
+			return -EINVAL;
+
+		if (intel_plane->rotation == val)
+			return 0;
+
+		old_val = intel_plane->rotation;
+		intel_plane->rotation = val;
+		ret = intel_plane_restore(plane);
+		if (ret)
+			intel_plane->rotation = old_val;
+	}
+
+	return ret;
+}
+
+int intel_plane_restore(struct drm_plane *plane)
 {
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 
 	if (!plane->crtc || !plane->fb)
-		return;
+		return 0;
 
-	intel_update_plane(plane, plane->crtc, plane->fb,
-			   intel_plane->crtc_x, intel_plane->crtc_y,
-			   intel_plane->crtc_w, intel_plane->crtc_h,
-			   intel_plane->src_x, intel_plane->src_y,
-			   intel_plane->src_w, intel_plane->src_h);
+	return plane->funcs->update_plane(plane, plane->crtc, plane->fb,
+				  intel_plane->crtc_x, intel_plane->crtc_y,
+				  intel_plane->crtc_w, intel_plane->crtc_h,
+				  intel_plane->src_x, intel_plane->src_y,
+				  intel_plane->src_w, intel_plane->src_h);
 }
 
 void intel_plane_disable(struct drm_plane *plane)
@@ -1198,6 +1271,7 @@ static const struct drm_plane_funcs intel_plane_funcs = {
 	.update_plane = intel_update_plane,
 	.disable_plane = intel_disable_plane,
 	.destroy = intel_destroy_plane,
+	.set_property = intel_plane_set_property,
 };
 
 static uint32_t ilk_plane_formats[] = {
@@ -1302,13 +1376,28 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 
 	intel_plane->pipe = pipe;
 	intel_plane->plane = plane;
+	intel_plane->rotation = BIT(DRM_ROTATE_0);
 	possible_crtcs = (1 << pipe);
-	ret = drm_plane_init(dev, &intel_plane->base, possible_crtcs,
-			     &intel_plane_funcs,
-			     plane_formats, num_plane_formats,
-			     false);
-	if (ret)
+	ret = drm_universal_plane_init(dev, &intel_plane->base, possible_crtcs,
+				       &intel_plane_funcs,
+				       plane_formats, num_plane_formats,
+				       DRM_PLANE_TYPE_OVERLAY);
+	if (ret) {
 		kfree(intel_plane);
+		goto out;
+	}
+
+	if (!dev->mode_config.rotation_property)
+		dev->mode_config.rotation_property =
+			drm_mode_create_rotation_property(dev,
+							  BIT(DRM_ROTATE_0) |
+							  BIT(DRM_ROTATE_180));
+
+	if (dev->mode_config.rotation_property)
+		drm_object_attach_property(&intel_plane->base.base,
+					   dev->mode_config.rotation_property,
+					   intel_plane->rotation);
 
+ out:
 	return ret;
 }
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index 67c6c9a2eb1c..c14341ca3ef9 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -854,6 +854,10 @@ intel_enable_tv(struct intel_encoder *encoder)
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* Prevents vblank waits from timing out in intel_tv_detect_type() */
+	intel_wait_for_vblank(encoder->base.dev,
+			      to_intel_crtc(encoder->base.crtc)->pipe);
+
 	I915_WRITE(TV_CTL, I915_READ(TV_CTL) | TV_ENC_ENABLE);
 }
 
@@ -1311,6 +1315,7 @@ intel_tv_detect(struct drm_connector *connector, bool force)
 {
 	struct drm_display_mode mode;
 	struct intel_tv *intel_tv = intel_attached_tv(connector);
+	enum drm_connector_status status;
 	int type;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] force=%d\n",
@@ -1323,16 +1328,24 @@ intel_tv_detect(struct drm_connector *connector, bool force)
 		struct intel_load_detect_pipe tmp;
 		struct drm_modeset_acquire_ctx ctx;
 
+		drm_modeset_acquire_init(&ctx, 0);
+
 		if (intel_get_load_detect_pipe(connector, &mode, &tmp, &ctx)) {
 			type = intel_tv_detect_type(intel_tv, connector);
-			intel_release_load_detect_pipe(connector, &tmp, &ctx);
+			intel_release_load_detect_pipe(connector, &tmp);
+			status = type < 0 ?
+				connector_status_disconnected :
+				connector_status_connected;
 		} else
-			return connector_status_unknown;
+			status = connector_status_unknown;
+
+		drm_modeset_drop_locks(&ctx);
+		drm_modeset_acquire_fini(&ctx);
 	} else
 		return connector->status;
 
-	if (type < 0)
-		return connector_status_disconnected;
+	if (status != connector_status_connected)
+		return status;
 
 	intel_tv->type = type;
 	intel_tv_find_better_format(connector);
@@ -1680,5 +1693,5 @@ intel_tv_init(struct drm_device *dev)
 	drm_object_attach_property(&connector->base,
 				   dev->mode_config.tv_bottom_margin_property,
 				   intel_tv->margin[TV_MARGIN_BOTTOM]);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 }
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index 4f6fef7ac069..918b76163965 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -101,7 +101,7 @@ static void __gen7_gt_force_wake_mt_get(struct drm_i915_private *dev_priv,
 {
 	u32 forcewake_ack;
 
-	if (IS_HASWELL(dev_priv->dev) || IS_GEN8(dev_priv->dev))
+	if (IS_HASWELL(dev_priv->dev) || IS_BROADWELL(dev_priv->dev))
 		forcewake_ack = FORCEWAKE_ACK_HSW;
 	else
 		forcewake_ack = FORCEWAKE_MT_ACK;
@@ -231,8 +231,8 @@ static void __vlv_force_wake_get(struct drm_i915_private *dev_priv,
 	}
 
 	/* WaRsForcewakeWaitTC0:vlv */
-	__gen6_gt_wait_for_thread_c0(dev_priv);
-
+	if (!IS_CHERRYVIEW(dev_priv->dev))
+		__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
 static void __vlv_force_wake_put(struct drm_i915_private *dev_priv,
@@ -250,9 +250,10 @@ static void __vlv_force_wake_put(struct drm_i915_private *dev_priv,
 		__raw_i915_write32(dev_priv, FORCEWAKE_MEDIA_VLV,
 				_MASKED_BIT_DISABLE(FORCEWAKE_KERNEL));
 
-	/* The below doubles as a POSTING_READ */
-	gen6_gt_check_fifodbg(dev_priv);
-
+	/* something from same cacheline, but !FORCEWAKE_VLV */
+	__raw_posting_read(dev_priv, FORCEWAKE_ACK_VLV);
+	if (!IS_CHERRYVIEW(dev_priv->dev))
+		gen6_gt_check_fifodbg(dev_priv);
 }
 
 static void vlv_force_wake_get(struct drm_i915_private *dev_priv, int fw_engine)
@@ -315,7 +316,7 @@ static void gen6_force_wake_timer(unsigned long arg)
 	intel_runtime_pm_put(dev_priv);
 }
 
-static void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
+void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long irqflags;
@@ -333,7 +334,7 @@ static void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 	else if (IS_GEN6(dev) || IS_GEN7(dev))
 		__gen6_gt_force_wake_reset(dev_priv);
 
-	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_GEN8(dev))
+	if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev) || IS_BROADWELL(dev))
 		__gen7_gt_force_wake_mt_reset(dev_priv);
 
 	if (restore) { /* If reset with a user forcewake, try to restore */
@@ -357,16 +358,12 @@ static void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore)
 			dev_priv->uncore.fifo_count =
 				__raw_i915_read32(dev_priv, GTFIFOCTL) &
 				GT_FIFO_FREE_ENTRIES_MASK;
-	} else {
-		dev_priv->uncore.forcewake_count = 0;
-		dev_priv->uncore.fw_rendercount = 0;
-		dev_priv->uncore.fw_mediacount = 0;
 	}
 
 	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
 }
 
-void intel_uncore_early_sanitize(struct drm_device *dev)
+void intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -389,7 +386,7 @@ void intel_uncore_early_sanitize(struct drm_device *dev)
 		__raw_i915_write32(dev_priv, GTFIFODBG,
 				   __raw_i915_read32(dev_priv, GTFIFODBG));
 
-	intel_uncore_forcewake_reset(dev, false);
+	intel_uncore_forcewake_reset(dev, restore_forcewake);
 }
 
 void intel_uncore_sanitize(struct drm_device *dev)
@@ -469,16 +466,43 @@ void assert_force_wake_inactive(struct drm_i915_private *dev_priv)
 #define NEEDS_FORCE_WAKE(dev_priv, reg) \
 	 ((reg) < 0x40000 && (reg) != FORCEWAKE)
 
-#define FORCEWAKE_VLV_RENDER_RANGE_OFFSET(reg) \
-	(((reg) >= 0x2000 && (reg) < 0x4000) ||\
-	((reg) >= 0x5000 && (reg) < 0x8000) ||\
-	((reg) >= 0xB000 && (reg) < 0x12000) ||\
-	((reg) >= 0x2E000 && (reg) < 0x30000))
+#define REG_RANGE(reg, start, end) ((reg) >= (start) && (reg) < (end))
 
-#define FORCEWAKE_VLV_MEDIA_RANGE_OFFSET(reg)\
-	(((reg) >= 0x12000 && (reg) < 0x14000) ||\
-	((reg) >= 0x22000 && (reg) < 0x24000) ||\
-	((reg) >= 0x30000 && (reg) < 0x40000))
+#define FORCEWAKE_VLV_RENDER_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x2000, 0x4000) || \
+	 REG_RANGE((reg), 0x5000, 0x8000) || \
+	 REG_RANGE((reg), 0xB000, 0x12000) || \
+	 REG_RANGE((reg), 0x2E000, 0x30000))
+
+#define FORCEWAKE_VLV_MEDIA_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x12000, 0x14000) || \
+	 REG_RANGE((reg), 0x22000, 0x24000) || \
+	 REG_RANGE((reg), 0x30000, 0x40000))
+
+#define FORCEWAKE_CHV_RENDER_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x2000, 0x4000) || \
+	 REG_RANGE((reg), 0x5000, 0x8000) || \
+	 REG_RANGE((reg), 0x8300, 0x8500) || \
+	 REG_RANGE((reg), 0xB000, 0xC000) || \
+	 REG_RANGE((reg), 0xE000, 0xE800))
+
+#define FORCEWAKE_CHV_MEDIA_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x8800, 0x8900) || \
+	 REG_RANGE((reg), 0xD000, 0xD800) || \
+	 REG_RANGE((reg), 0x12000, 0x14000) || \
+	 REG_RANGE((reg), 0x1A000, 0x1C000) || \
+	 REG_RANGE((reg), 0x1E800, 0x1EA00) || \
+	 REG_RANGE((reg), 0x30000, 0x40000))
+
+#define FORCEWAKE_CHV_COMMON_RANGE_OFFSET(reg) \
+	(REG_RANGE((reg), 0x4000, 0x5000) || \
+	 REG_RANGE((reg), 0x8000, 0x8300) || \
+	 REG_RANGE((reg), 0x8500, 0x8600) || \
+	 REG_RANGE((reg), 0x9000, 0xB000) || \
+	 REG_RANGE((reg), 0xC000, 0xC800) || \
+	 REG_RANGE((reg), 0xF000, 0x10000) || \
+	 REG_RANGE((reg), 0x14000, 0x14400) || \
+	 REG_RANGE((reg), 0x22000, 0x24000))
 
 static void
 ilk_dummy_write(struct drm_i915_private *dev_priv)
@@ -490,20 +514,30 @@ ilk_dummy_write(struct drm_i915_private *dev_priv)
 }
 
 static void
-hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
+hsw_unclaimed_reg_debug(struct drm_i915_private *dev_priv, u32 reg, bool read,
+			bool before)
 {
+	const char *op = read ? "reading" : "writing to";
+	const char *when = before ? "before" : "after";
+
+	if (!i915.mmio_debug)
+		return;
+
 	if (__raw_i915_read32(dev_priv, FPGA_DBG) & FPGA_DBG_RM_NOCLAIM) {
-		DRM_ERROR("Unknown unclaimed register before writing to %x\n",
-			  reg);
+		WARN(1, "Unclaimed register detected %s %s register 0x%x\n",
+		     when, op, reg);
 		__raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
 	}
 }
 
 static void
-hsw_unclaimed_reg_check(struct drm_i915_private *dev_priv, u32 reg)
+hsw_unclaimed_reg_detect(struct drm_i915_private *dev_priv)
 {
+	if (i915.mmio_debug)
+		return;
+
 	if (__raw_i915_read32(dev_priv, FPGA_DBG) & FPGA_DBG_RM_NOCLAIM) {
-		DRM_ERROR("Unclaimed write to %x\n", reg);
+		DRM_ERROR("Unclaimed register detected. Please use the i915.mmio_debug=1 to debug this problem.");
 		__raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
 	}
 }
@@ -540,6 +574,7 @@ gen5_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
 static u##x \
 gen6_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
 	REG_READ_HEADER(x); \
+	hsw_unclaimed_reg_debug(dev_priv, reg, true, true); \
 	if (dev_priv->uncore.forcewake_count == 0 && \
 	    NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
 		dev_priv->uncore.funcs.force_wake_get(dev_priv, \
@@ -550,6 +585,7 @@ gen6_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
 	} else { \
 		val = __raw_i915_read##x(dev_priv, reg); \
 	} \
+	hsw_unclaimed_reg_debug(dev_priv, reg, true, false); \
 	REG_READ_FOOTER; \
 }
 
@@ -573,7 +609,35 @@ vlv_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
 	REG_READ_FOOTER; \
 }
 
+#define __chv_read(x) \
+static u##x \
+chv_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \
+	unsigned fwengine = 0; \
+	REG_READ_HEADER(x); \
+	if (FORCEWAKE_CHV_RENDER_RANGE_OFFSET(reg)) { \
+		if (dev_priv->uncore.fw_rendercount == 0) \
+			fwengine = FORCEWAKE_RENDER; \
+	} else if (FORCEWAKE_CHV_MEDIA_RANGE_OFFSET(reg)) { \
+		if (dev_priv->uncore.fw_mediacount == 0) \
+			fwengine = FORCEWAKE_MEDIA; \
+	} else if (FORCEWAKE_CHV_COMMON_RANGE_OFFSET(reg)) { \
+		if (dev_priv->uncore.fw_rendercount == 0) \
+			fwengine |= FORCEWAKE_RENDER; \
+		if (dev_priv->uncore.fw_mediacount == 0) \
+			fwengine |= FORCEWAKE_MEDIA; \
+	} \
+	if (fwengine) \
+		dev_priv->uncore.funcs.force_wake_get(dev_priv, fwengine); \
+	val = __raw_i915_read##x(dev_priv, reg); \
+	if (fwengine) \
+		dev_priv->uncore.funcs.force_wake_put(dev_priv, fwengine); \
+	REG_READ_FOOTER; \
+}
 
+__chv_read(8)
+__chv_read(16)
+__chv_read(32)
+__chv_read(64)
 __vlv_read(8)
 __vlv_read(16)
 __vlv_read(32)
@@ -591,6 +655,7 @@ __gen4_read(16)
 __gen4_read(32)
 __gen4_read(64)
 
+#undef __chv_read
 #undef __vlv_read
 #undef __gen6_read
 #undef __gen5_read
@@ -647,12 +712,13 @@ hsw_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace)
 	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
 		__fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \
 	} \
-	hsw_unclaimed_reg_clear(dev_priv, reg); \
+	hsw_unclaimed_reg_debug(dev_priv, reg, false, true); \
 	__raw_i915_write##x(dev_priv, reg, val); \
 	if (unlikely(__fifo_ret)) { \
 		gen6_gt_check_fifodbg(dev_priv); \
 	} \
-	hsw_unclaimed_reg_check(dev_priv, reg); \
+	hsw_unclaimed_reg_debug(dev_priv, reg, false, false); \
+	hsw_unclaimed_reg_detect(dev_priv); \
 	REG_WRITE_FOOTER; \
 }
 
@@ -681,6 +747,7 @@ static bool is_gen8_shadowed(struct drm_i915_private *dev_priv, u32 reg)
 static void \
 gen8_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace) { \
 	REG_WRITE_HEADER; \
+	hsw_unclaimed_reg_debug(dev_priv, reg, false, true); \
 	if (reg < 0x40000 && !is_gen8_shadowed(dev_priv, reg)) { \
 		if (dev_priv->uncore.forcewake_count == 0) \
 			dev_priv->uncore.funcs.force_wake_get(dev_priv,	\
@@ -692,9 +759,43 @@ gen8_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace
 	} else { \
 		__raw_i915_write##x(dev_priv, reg, val); \
 	} \
+	hsw_unclaimed_reg_debug(dev_priv, reg, false, false); \
+	hsw_unclaimed_reg_detect(dev_priv); \
 	REG_WRITE_FOOTER; \
 }
 
+#define __chv_write(x) \
+static void \
+chv_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace) { \
+	unsigned fwengine = 0; \
+	bool shadowed = is_gen8_shadowed(dev_priv, reg); \
+	REG_WRITE_HEADER; \
+	if (!shadowed) { \
+		if (FORCEWAKE_CHV_RENDER_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine = FORCEWAKE_RENDER; \
+		} else if (FORCEWAKE_CHV_MEDIA_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine = FORCEWAKE_MEDIA; \
+		} else if (FORCEWAKE_CHV_COMMON_RANGE_OFFSET(reg)) { \
+			if (dev_priv->uncore.fw_rendercount == 0) \
+				fwengine |= FORCEWAKE_RENDER; \
+			if (dev_priv->uncore.fw_mediacount == 0) \
+				fwengine |= FORCEWAKE_MEDIA; \
+		} \
+	} \
+	if (fwengine) \
+		dev_priv->uncore.funcs.force_wake_get(dev_priv, fwengine); \
+	__raw_i915_write##x(dev_priv, reg, val); \
+	if (fwengine) \
+		dev_priv->uncore.funcs.force_wake_put(dev_priv, fwengine); \
+	REG_WRITE_FOOTER; \
+}
+
+__chv_write(8)
+__chv_write(16)
+__chv_write(32)
+__chv_write(64)
 __gen8_write(8)
 __gen8_write(16)
 __gen8_write(32)
@@ -716,6 +817,7 @@ __gen4_write(16)
 __gen4_write(32)
 __gen4_write(64)
 
+#undef __chv_write
 #undef __gen8_write
 #undef __hsw_write
 #undef __gen6_write
@@ -731,12 +833,12 @@ void intel_uncore_init(struct drm_device *dev)
 	setup_timer(&dev_priv->uncore.force_wake_timer,
 		    gen6_force_wake_timer, (unsigned long)dev_priv);
 
-	intel_uncore_early_sanitize(dev);
+	intel_uncore_early_sanitize(dev, false);
 
 	if (IS_VALLEYVIEW(dev)) {
 		dev_priv->uncore.funcs.force_wake_get = __vlv_force_wake_get;
 		dev_priv->uncore.funcs.force_wake_put = __vlv_force_wake_put;
-	} else if (IS_HASWELL(dev) || IS_GEN8(dev)) {
+	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		dev_priv->uncore.funcs.force_wake_get = __gen7_gt_force_wake_mt_get;
 		dev_priv->uncore.funcs.force_wake_put = __gen7_gt_force_wake_mt_put;
 	} else if (IS_IVYBRIDGE(dev)) {
@@ -779,14 +881,26 @@ void intel_uncore_init(struct drm_device *dev)
 
 	switch (INTEL_INFO(dev)->gen) {
 	default:
-		dev_priv->uncore.funcs.mmio_writeb  = gen8_write8;
-		dev_priv->uncore.funcs.mmio_writew  = gen8_write16;
-		dev_priv->uncore.funcs.mmio_writel  = gen8_write32;
-		dev_priv->uncore.funcs.mmio_writeq  = gen8_write64;
-		dev_priv->uncore.funcs.mmio_readb  = gen6_read8;
-		dev_priv->uncore.funcs.mmio_readw  = gen6_read16;
-		dev_priv->uncore.funcs.mmio_readl  = gen6_read32;
-		dev_priv->uncore.funcs.mmio_readq  = gen6_read64;
+		if (IS_CHERRYVIEW(dev)) {
+			dev_priv->uncore.funcs.mmio_writeb  = chv_write8;
+			dev_priv->uncore.funcs.mmio_writew  = chv_write16;
+			dev_priv->uncore.funcs.mmio_writel  = chv_write32;
+			dev_priv->uncore.funcs.mmio_writeq  = chv_write64;
+			dev_priv->uncore.funcs.mmio_readb  = chv_read8;
+			dev_priv->uncore.funcs.mmio_readw  = chv_read16;
+			dev_priv->uncore.funcs.mmio_readl  = chv_read32;
+			dev_priv->uncore.funcs.mmio_readq  = chv_read64;
+
+		} else {
+			dev_priv->uncore.funcs.mmio_writeb  = gen8_write8;
+			dev_priv->uncore.funcs.mmio_writew  = gen8_write16;
+			dev_priv->uncore.funcs.mmio_writel  = gen8_write32;
+			dev_priv->uncore.funcs.mmio_writeq  = gen8_write64;
+			dev_priv->uncore.funcs.mmio_readb  = gen6_read8;
+			dev_priv->uncore.funcs.mmio_readw  = gen6_read16;
+			dev_priv->uncore.funcs.mmio_readl  = gen6_read32;
+			dev_priv->uncore.funcs.mmio_readq  = gen6_read64;
+		}
 		break;
 	case 7:
 	case 6:
@@ -912,7 +1026,7 @@ int i915_get_reset_stats_ioctl(struct drm_device *dev,
 	if (args->flags || args->pad)
 		return -EINVAL;
 
-	if (args->ctx_id == DEFAULT_CONTEXT_ID && !capable(CAP_SYS_ADMIN))
+	if (args->ctx_id == DEFAULT_CONTEXT_HANDLE && !capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -1053,18 +1167,16 @@ static int gen6_do_reset(struct drm_device *dev)
 
 int intel_gpu_reset(struct drm_device *dev)
 {
-	switch (INTEL_INFO(dev)->gen) {
-	case 8:
-	case 7:
-	case 6: return gen6_do_reset(dev);
-	case 5: return ironlake_do_reset(dev);
-	case 4:
-		if (IS_G4X(dev))
-			return g4x_do_reset(dev);
-		else
-			return i965_do_reset(dev);
-	default: return -ENODEV;
-	}
+	if (INTEL_INFO(dev)->gen >= 6)
+		return gen6_do_reset(dev);
+	else if (IS_GEN5(dev))
+		return ironlake_do_reset(dev);
+	else if (IS_G4X(dev))
+		return g4x_do_reset(dev);
+	else if (IS_GEN4(dev))
+		return i965_do_reset(dev);
+	else
+		return -ENODEV;
 }
 
 void intel_uncore_check_errors(struct drm_device *dev)
diff --git a/drivers/gpu/drm/mga/mga_dma.c b/drivers/gpu/drm/mga/mga_dma.c
index c3bf059ba720..8cfa9cb74c86 100644
--- a/drivers/gpu/drm/mga/mga_dma.c
+++ b/drivers/gpu/drm/mga/mga_dma.c
@@ -502,31 +502,31 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 		return err;
 	}
 
-	/* Make drm_addbufs happy by not trying to create a mapping for less
-	 * than a page.
+	/* Make drm_legacy_addbufs happy by not trying to create a mapping for
+	 * less than a page.
 	 */
 	if (warp_size < PAGE_SIZE)
 		warp_size = PAGE_SIZE;
 
 	offset = 0;
-	err = drm_addmap(dev, offset, warp_size,
-			 _DRM_AGP, _DRM_READ_ONLY, &dev_priv->warp);
+	err = drm_legacy_addmap(dev, offset, warp_size,
+				_DRM_AGP, _DRM_READ_ONLY, &dev_priv->warp);
 	if (err) {
 		DRM_ERROR("Unable to map WARP microcode: %d\n", err);
 		return err;
 	}
 
 	offset += warp_size;
-	err = drm_addmap(dev, offset, dma_bs->primary_size,
-			 _DRM_AGP, _DRM_READ_ONLY, &dev_priv->primary);
+	err = drm_legacy_addmap(dev, offset, dma_bs->primary_size,
+				_DRM_AGP, _DRM_READ_ONLY, &dev_priv->primary);
 	if (err) {
 		DRM_ERROR("Unable to map primary DMA region: %d\n", err);
 		return err;
 	}
 
 	offset += dma_bs->primary_size;
-	err = drm_addmap(dev, offset, secondary_size,
-			 _DRM_AGP, 0, &dev->agp_buffer_map);
+	err = drm_legacy_addmap(dev, offset, secondary_size,
+				_DRM_AGP, 0, &dev->agp_buffer_map);
 	if (err) {
 		DRM_ERROR("Unable to map secondary DMA region: %d\n", err);
 		return err;
@@ -538,7 +538,7 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 	req.flags = _DRM_AGP_BUFFER;
 	req.agp_start = offset;
 
-	err = drm_addbufs_agp(dev, &req);
+	err = drm_legacy_addbufs_agp(dev, &req);
 	if (err) {
 		DRM_ERROR("Unable to add secondary DMA buffers: %d\n", err);
 		return err;
@@ -559,16 +559,16 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 	}
 
 	offset += secondary_size;
-	err = drm_addmap(dev, offset, agp_size - offset,
-			 _DRM_AGP, 0, &dev_priv->agp_textures);
+	err = drm_legacy_addmap(dev, offset, agp_size - offset,
+				_DRM_AGP, 0, &dev_priv->agp_textures);
 	if (err) {
 		DRM_ERROR("Unable to map AGP texture region %d\n", err);
 		return err;
 	}
 
-	drm_core_ioremap(dev_priv->warp, dev);
-	drm_core_ioremap(dev_priv->primary, dev);
-	drm_core_ioremap(dev->agp_buffer_map, dev);
+	drm_legacy_ioremap(dev_priv->warp, dev);
+	drm_legacy_ioremap(dev_priv->primary, dev);
+	drm_legacy_ioremap(dev->agp_buffer_map, dev);
 
 	if (!dev_priv->warp->handle ||
 	    !dev_priv->primary->handle || !dev->agp_buffer_map->handle) {
@@ -602,7 +602,7 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
  *
  * \todo
  * Determine whether the maximum address passed to drm_pci_alloc is correct.
- * The same goes for drm_addbufs_pci.
+ * The same goes for drm_legacy_addbufs_pci.
  *
  * \sa mga_do_dma_bootstrap, mga_do_agp_dma_bootstrap
  */
@@ -622,15 +622,15 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 		return -EFAULT;
 	}
 
-	/* Make drm_addbufs happy by not trying to create a mapping for less
-	 * than a page.
+	/* Make drm_legacy_addbufs happy by not trying to create a mapping for
+	 * less than a page.
 	 */
 	if (warp_size < PAGE_SIZE)
 		warp_size = PAGE_SIZE;
 
 	/* The proper alignment is 0x100 for this mapping */
-	err = drm_addmap(dev, 0, warp_size, _DRM_CONSISTENT,
-			 _DRM_READ_ONLY, &dev_priv->warp);
+	err = drm_legacy_addmap(dev, 0, warp_size, _DRM_CONSISTENT,
+				_DRM_READ_ONLY, &dev_priv->warp);
 	if (err != 0) {
 		DRM_ERROR("Unable to create mapping for WARP microcode: %d\n",
 			  err);
@@ -645,8 +645,8 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 	for (primary_size = dma_bs->primary_size; primary_size != 0;
 	     primary_size >>= 1) {
 		/* The proper alignment for this mapping is 0x04 */
-		err = drm_addmap(dev, 0, primary_size, _DRM_CONSISTENT,
-				 _DRM_READ_ONLY, &dev_priv->primary);
+		err = drm_legacy_addmap(dev, 0, primary_size, _DRM_CONSISTENT,
+					_DRM_READ_ONLY, &dev_priv->primary);
 		if (!err)
 			break;
 	}
@@ -669,7 +669,7 @@ static int mga_do_pci_dma_bootstrap(struct drm_device *dev,
 		req.count = bin_count;
 		req.size = dma_bs->secondary_bin_size;
 
-		err = drm_addbufs_pci(dev, &req);
+		err = drm_legacy_addbufs_pci(dev, &req);
 		if (!err)
 			break;
 	}
@@ -708,15 +708,16 @@ static int mga_do_dma_bootstrap(struct drm_device *dev,
 	/* The first steps are the same for both PCI and AGP based DMA.  Map
 	 * the cards MMIO registers and map a status page.
 	 */
-	err = drm_addmap(dev, dev_priv->mmio_base, dev_priv->mmio_size,
-			 _DRM_REGISTERS, _DRM_READ_ONLY, &dev_priv->mmio);
+	err = drm_legacy_addmap(dev, dev_priv->mmio_base, dev_priv->mmio_size,
+				_DRM_REGISTERS, _DRM_READ_ONLY,
+				&dev_priv->mmio);
 	if (err) {
 		DRM_ERROR("Unable to map MMIO region: %d\n", err);
 		return err;
 	}
 
-	err = drm_addmap(dev, 0, SAREA_MAX, _DRM_SHM,
-			 _DRM_READ_ONLY | _DRM_LOCKED | _DRM_KERNEL,
+	err = drm_legacy_addmap(dev, 0, SAREA_MAX, _DRM_SHM,
+				_DRM_READ_ONLY | _DRM_LOCKED | _DRM_KERNEL,
 			 &dev_priv->status);
 	if (err) {
 		DRM_ERROR("Unable to map status region: %d\n", err);
@@ -809,7 +810,7 @@ static int mga_do_init_dma(struct drm_device *dev, drm_mga_init_t *init)
 	dev_priv->texture_offset = init->texture_offset[0];
 	dev_priv->texture_size = init->texture_size[0];
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("failed to find sarea!\n");
 		return -EINVAL;
@@ -820,37 +821,37 @@ static int mga_do_init_dma(struct drm_device *dev, drm_mga_init_t *init)
 		dev_priv->dma_access = MGA_PAGPXFER;
 		dev_priv->wagp_enable = MGA_WAGP_ENABLE;
 
-		dev_priv->status = drm_core_findmap(dev, init->status_offset);
+		dev_priv->status = drm_legacy_findmap(dev, init->status_offset);
 		if (!dev_priv->status) {
 			DRM_ERROR("failed to find status page!\n");
 			return -EINVAL;
 		}
-		dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+		dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 		if (!dev_priv->mmio) {
 			DRM_ERROR("failed to find mmio region!\n");
 			return -EINVAL;
 		}
-		dev_priv->warp = drm_core_findmap(dev, init->warp_offset);
+		dev_priv->warp = drm_legacy_findmap(dev, init->warp_offset);
 		if (!dev_priv->warp) {
 			DRM_ERROR("failed to find warp microcode region!\n");
 			return -EINVAL;
 		}
-		dev_priv->primary = drm_core_findmap(dev, init->primary_offset);
+		dev_priv->primary = drm_legacy_findmap(dev, init->primary_offset);
 		if (!dev_priv->primary) {
 			DRM_ERROR("failed to find primary dma region!\n");
 			return -EINVAL;
 		}
 		dev->agp_buffer_token = init->buffers_offset;
 		dev->agp_buffer_map =
-		    drm_core_findmap(dev, init->buffers_offset);
+		    drm_legacy_findmap(dev, init->buffers_offset);
 		if (!dev->agp_buffer_map) {
 			DRM_ERROR("failed to find dma buffer region!\n");
 			return -EINVAL;
 		}
 
-		drm_core_ioremap(dev_priv->warp, dev);
-		drm_core_ioremap(dev_priv->primary, dev);
-		drm_core_ioremap(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap(dev_priv->warp, dev);
+		drm_legacy_ioremap(dev_priv->primary, dev);
+		drm_legacy_ioremap(dev->agp_buffer_map, dev);
 	}
 
 	dev_priv->sarea_priv =
@@ -936,14 +937,14 @@ static int mga_do_cleanup_dma(struct drm_device *dev, int full_cleanup)
 
 		if ((dev_priv->warp != NULL)
 		    && (dev_priv->warp->type != _DRM_CONSISTENT))
-			drm_core_ioremapfree(dev_priv->warp, dev);
+			drm_legacy_ioremapfree(dev_priv->warp, dev);
 
 		if ((dev_priv->primary != NULL)
 		    && (dev_priv->primary->type != _DRM_CONSISTENT))
-			drm_core_ioremapfree(dev_priv->primary, dev);
+			drm_legacy_ioremapfree(dev_priv->primary, dev);
 
 		if (dev->agp_buffer_map != NULL)
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 
 		if (dev_priv->used_new_dma_init) {
 #if __OS_HAS_AGP
diff --git a/drivers/gpu/drm/mga/mga_drv.c b/drivers/gpu/drm/mga/mga_drv.c
index 6b1a87c8aac5..5e2f131a6a72 100644
--- a/drivers/gpu/drm/mga/mga_drv.c
+++ b/drivers/gpu/drm/mga/mga_drv.c
@@ -48,7 +48,7 @@ static const struct file_operations mga_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = mga_compat_ioctl,
@@ -64,6 +64,7 @@ static struct drm_driver driver = {
 	.load = mga_driver_load,
 	.unload = mga_driver_unload,
 	.lastclose = mga_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.dma_quiescent = mga_driver_dma_quiescent,
 	.device_is_agp = mga_driver_device_is_agp,
 	.get_vblank_counter = mga_get_vblank_counter,
diff --git a/drivers/gpu/drm/mga/mga_drv.h b/drivers/gpu/drm/mga/mga_drv.h
index fe453213600a..b4a2014917e5 100644
--- a/drivers/gpu/drm/mga/mga_drv.h
+++ b/drivers/gpu/drm/mga/mga_drv.h
@@ -31,6 +31,8 @@
 #ifndef __MGA_DRV_H__
 #define __MGA_DRV_H__
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.c b/drivers/gpu/drm/mgag200/mgag200_drv.c
index f15ea3c4a90a..97745991544d 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.c
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.c
@@ -28,7 +28,7 @@ module_param_named(modeset, mgag200_modeset, int, 0400);
 
 static struct drm_driver driver;
 
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+static const struct pci_device_id pciidlist[] = {
 	{ PCI_VENDOR_ID_MATROX, 0x522, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_SE_A },
 	{ PCI_VENDOR_ID_MATROX, 0x524, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_SE_B },
 	{ PCI_VENDOR_ID_MATROX, 0x530, PCI_ANY_ID, PCI_ANY_ID, 0, 0, G200_EV },
@@ -91,6 +91,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_GEM | DRIVER_MODESET,
 	.load = mgag200_driver_load,
 	.unload = mgag200_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.fops = &mgag200_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
index cf11ee68a6d9..e9eea1d4e7c3 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.h
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
@@ -22,6 +22,8 @@
 #include <drm/ttm/ttm_memory.h>
 #include <drm/ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
 
@@ -190,8 +192,6 @@ struct mga_device {
 	resource_size_t			rmmio_size;
 	void __iomem			*rmmio;
 
-	drm_local_map_t			*framebuffer;
-
 	struct mga_mc			mc;
 	struct mga_mode_info		mode_info;
 
@@ -224,7 +224,7 @@ struct mgag200_bo {
 	struct ttm_placement placement;
 	struct ttm_bo_kmap_obj kmap;
 	struct drm_gem_object gem;
-	u32 placements[3];
+	struct ttm_place placements[3];
 	int pin_count;
 };
 #define gem_to_mga_bo(gobj) container_of((gobj), struct mgag200_bo, gem)
@@ -280,7 +280,7 @@ static inline int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
 {
 	int ret;
 
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, NULL);
 	if (ret) {
 		if (ret != -ERESTARTSYS && ret != -EBUSY)
 			DRM_ERROR("reserve failed %p\n", bo);
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 13b7dd83faa9..4415af3666ab 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -158,7 +158,8 @@ static int mgag200fb_create_object(struct mga_fbdev *afbdev,
 static int mgag200fb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct mga_fbdev *mfbdev = (struct mga_fbdev *)helper;
+	struct mga_fbdev *mfbdev =
+		container_of(helper, struct mga_fbdev, helper);
 	struct drm_device *dev = mfbdev->helper.dev;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct mga_device *mdev = dev->dev_private;
@@ -272,7 +273,7 @@ static int mga_fbdev_destroy(struct drm_device *dev,
 	return 0;
 }
 
-static struct drm_fb_helper_funcs mga_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs mga_fb_helper_funcs = {
 	.gamma_set = mga_crtc_fb_gamma_set,
 	.gamma_get = mga_crtc_fb_gamma_get,
 	.fb_probe = mgag200fb_create,
@@ -293,9 +294,10 @@ int mgag200_fbdev_init(struct mga_device *mdev)
 		return -ENOMEM;
 
 	mdev->mfbdev = mfbdev;
-	mfbdev->helper.funcs = &mga_fb_helper_funcs;
 	spin_lock_init(&mfbdev->dirty_lock);
 
+	drm_fb_helper_prepare(mdev->dev, &mfbdev->helper, &mga_fb_helper_funcs);
+
 	ret = drm_fb_helper_init(mdev->dev, &mfbdev->helper,
 				 mdev->num_crtc, MGAG200FB_CONN_LIMIT);
 	if (ret)
diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
index a034ed408252..83485ab81ce8 100644
--- a/drivers/gpu/drm/mgag200/mgag200_mode.c
+++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
@@ -1483,11 +1483,7 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 {
 	struct drm_device *dev = connector->dev;
 	struct mga_device *mdev = (struct mga_device*)dev->dev_private;
-	struct mga_fbdev *mfbdev = mdev->mfbdev;
-	struct drm_fb_helper *fb_helper = &mfbdev->helper;
-	struct drm_fb_helper_connector *fb_helper_conn = NULL;
 	int bpp = 32;
-	int i = 0;
 
 	if (IS_G200_SE(mdev)) {
 		if (mdev->unique_rev_id == 0x01) {
@@ -1537,21 +1533,14 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 	}
 
 	/* Validate the mode input by the user */
-	for (i = 0; i < fb_helper->connector_count; i++) {
-		if (fb_helper->connector_info[i]->connector == connector) {
-			/* Found the helper for this connector */
-			fb_helper_conn = fb_helper->connector_info[i];
-			if (fb_helper_conn->cmdline_mode.specified) {
-				if (fb_helper_conn->cmdline_mode.bpp_specified) {
-					bpp = fb_helper_conn->cmdline_mode.bpp;
-				}
-			}
-		}
+	if (connector->cmdline_mode.specified) {
+		if (connector->cmdline_mode.bpp_specified)
+			bpp = connector->cmdline_mode.bpp;
 	}
 
 	if ((mode->hdisplay * mode->vdisplay * (bpp/8)) > mdev->mc.vram_size) {
-		if (fb_helper_conn)
-			fb_helper_conn->cmdline_mode.specified = false;
+		if (connector->cmdline_mode.specified)
+			connector->cmdline_mode.specified = false;
 		return MODE_BAD;
 	}
 
@@ -1562,19 +1551,9 @@ static struct drm_encoder *mga_connector_best_encoder(struct drm_connector
 						  *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj =
-		    drm_mode_object_find(connector->dev, enc_id,
-					 DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -1621,7 +1600,7 @@ static struct drm_connector *mga_vga_init(struct drm_device *dev)
 
 	drm_connector_helper_add(connector, &mga_vga_connector_helper_funcs);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	mga_connector->i2c = mgag200_i2c_create(dev);
 	if (!mga_connector->i2c)
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 5a00e90696de..d16964ea0ed4 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -293,18 +293,22 @@ void mgag200_mm_fini(struct mga_device *mdev)
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
 {
 	u32 c = 0;
-	bo->placement.fpfn = 0;
-	bo->placement.lpfn = 0;
+	unsigned i;
+
 	bo->placement.placement = bo->placements;
 	bo->placement.busy_placement = bo->placements;
 	if (domain & TTM_PL_FLAG_VRAM)
-		bo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
+		bo->placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM;
 	if (domain & TTM_PL_FLAG_SYSTEM)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	if (!c)
-		bo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		bo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	bo->placement.num_placement = c;
 	bo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		bo->placements[i].fpfn = 0;
+		bo->placements[i].lpfn = 0;
+	}
 }
 
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
@@ -335,7 +339,7 @@ int mgag200_bo_create(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&mdev->ttm.bdev, &mgabo->bo, size,
 			  ttm_bo_type_device, &mgabo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size,
-			  NULL, mgag200_bo_ttm_destroy);
+			  NULL, NULL, mgag200_bo_ttm_destroy);
 	if (ret)
 		return ret;
 
@@ -361,7 +365,7 @@ int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr)
 
 	mgag200_ttm_placement(bo, pl_flag);
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -384,7 +388,7 @@ int mgag200_bo_unpin(struct mgag200_bo *bo)
 		return 0;
 
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret)
 		return ret;
@@ -408,7 +412,7 @@ int mgag200_bo_push_sysram(struct mgag200_bo *bo)
 
 	mgag200_ttm_placement(bo, TTM_PL_FLAG_SYSTEM);
 	for (i = 0; i < bo->placement.num_placement ; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 
 	ret = ttm_bo_validate(&bo->bo, &bo->placement, false, false);
 	if (ret) {
@@ -424,7 +428,7 @@ int mgag200_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct mga_device *mdev;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	file_priv = filp->private_data;
 	mdev = file_priv->minor->dev->dev_private;
diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig
index f12388967856..9d907c526c94 100644
--- a/drivers/gpu/drm/msm/Kconfig
+++ b/drivers/gpu/drm/msm/Kconfig
@@ -2,9 +2,9 @@
 config DRM_MSM
 	tristate "MSM DRM"
 	depends on DRM
-	depends on MSM_IOMMU
 	depends on ARCH_QCOM || (ARM && COMPILE_TEST)
 	select DRM_KMS_HELPER
+	select DRM_PANEL
 	select SHMEM
 	select TMPFS
 	default y
diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile
index 93ca49c8df44..6283dcb96af5 100644
--- a/drivers/gpu/drm/msm/Makefile
+++ b/drivers/gpu/drm/msm/Makefile
@@ -4,6 +4,7 @@ ifeq (, $(findstring -W,$(EXTRA_CFLAGS)))
 endif
 
 msm-y := \
+	adreno/adreno_device.o \
 	adreno/adreno_gpu.o \
 	adreno/a3xx_gpu.o \
 	hdmi/hdmi.o \
@@ -18,6 +19,8 @@ msm-y := \
 	mdp/mdp_kms.o \
 	mdp/mdp4/mdp4_crtc.o \
 	mdp/mdp4/mdp4_dtv_encoder.o \
+	mdp/mdp4/mdp4_lcdc_encoder.o \
+	mdp/mdp4/mdp4_lvds_connector.o \
 	mdp/mdp4/mdp4_irq.o \
 	mdp/mdp4/mdp4_kms.o \
 	mdp/mdp4/mdp4_plane.o \
@@ -39,5 +42,6 @@ msm-y := \
 	msm_ringbuffer.o
 
 msm-$(CONFIG_DRM_MSM_FBDEV) += msm_fbdev.o
+msm-$(CONFIG_COMMON_CLK) += mdp/mdp4/mdp4_lvds_pll.o
 
 obj-$(CONFIG_DRM_MSM)	+= msm.o
diff --git a/drivers/gpu/drm/msm/adreno/a2xx.xml.h b/drivers/gpu/drm/msm/adreno/a2xx.xml.h
index 85d615e7d62f..a3104598c27f 100644
--- a/drivers/gpu/drm/msm/adreno/a2xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a2xx.xml.h
@@ -10,13 +10,13 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32814 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   8900 bytes, from 2013-10-22 23:57:49)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  10574 bytes, from 2013-11-13 05:44:45)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  53644 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (   8344 bytes, from 2013-11-30 14:49:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14960 bytes, from 2014-07-27 17:22:13)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  41068 bytes, from 2014-08-01 12:22:48)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -203,6 +203,15 @@ enum a2xx_rb_copy_sample_select {
 	SAMPLE_0123 = 6,
 };
 
+enum a2xx_rb_blend_opcode {
+	BLEND_DST_PLUS_SRC = 0,
+	BLEND_SRC_MINUS_DST = 1,
+	BLEND_MIN_DST_SRC = 2,
+	BLEND_MAX_DST_SRC = 3,
+	BLEND_DST_MINUS_SRC = 4,
+	BLEND_DST_PLUS_SRC_BIAS = 5,
+};
+
 enum adreno_mmu_clnt_beh {
 	BEH_NEVR = 0,
 	BEH_TRAN_RNG = 1,
@@ -890,6 +899,39 @@ static inline uint32_t A2XX_SQ_CONTEXT_MISC_PARAM_GEN_POS(uint32_t val)
 #define REG_A2XX_VGT_EVENT_INITIATOR				0x000021f9
 
 #define REG_A2XX_VGT_DRAW_INITIATOR				0x000021fc
+#define A2XX_VGT_DRAW_INITIATOR_PRIM_TYPE__MASK			0x0000003f
+#define A2XX_VGT_DRAW_INITIATOR_PRIM_TYPE__SHIFT		0
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_PRIM_TYPE(enum pc_di_primtype val)
+{
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_PRIM_TYPE__SHIFT) & A2XX_VGT_DRAW_INITIATOR_PRIM_TYPE__MASK;
+}
+#define A2XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__MASK		0x000000c0
+#define A2XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__SHIFT		6
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_SOURCE_SELECT(enum pc_di_src_sel val)
+{
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__SHIFT) & A2XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__MASK;
+}
+#define A2XX_VGT_DRAW_INITIATOR_VIS_CULL__MASK			0x00000600
+#define A2XX_VGT_DRAW_INITIATOR_VIS_CULL__SHIFT			9
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_VIS_CULL(enum pc_di_vis_cull_mode val)
+{
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_VIS_CULL__SHIFT) & A2XX_VGT_DRAW_INITIATOR_VIS_CULL__MASK;
+}
+#define A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE__MASK		0x00000800
+#define A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE__SHIFT		11
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE(enum pc_di_index_size val)
+{
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE__SHIFT) & A2XX_VGT_DRAW_INITIATOR_INDEX_SIZE__MASK;
+}
+#define A2XX_VGT_DRAW_INITIATOR_NOT_EOP				0x00001000
+#define A2XX_VGT_DRAW_INITIATOR_SMALL_INDEX			0x00002000
+#define A2XX_VGT_DRAW_INITIATOR_PRE_DRAW_INITIATOR_ENABLE	0x00004000
+#define A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK		0xffff0000
+#define A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT		16
+static inline uint32_t A2XX_VGT_DRAW_INITIATOR_NUM_INDICES(uint32_t val)
+{
+	return ((val) << A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT) & A2XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK;
+}
 
 #define REG_A2XX_VGT_IMMED_DATA					0x000021fd
 
@@ -963,7 +1005,7 @@ static inline uint32_t A2XX_RB_BLEND_CONTROL_COLOR_SRCBLEND(enum adreno_rb_blend
 }
 #define A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN__MASK		0x000000e0
 #define A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN__SHIFT		5
-static inline uint32_t A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN(enum adreno_rb_blend_opcode val)
+static inline uint32_t A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN(enum a2xx_rb_blend_opcode val)
 {
 	return ((val) << A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN__SHIFT) & A2XX_RB_BLEND_CONTROL_COLOR_COMB_FCN__MASK;
 }
@@ -981,7 +1023,7 @@ static inline uint32_t A2XX_RB_BLEND_CONTROL_ALPHA_SRCBLEND(enum adreno_rb_blend
 }
 #define A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN__MASK		0x00e00000
 #define A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN__SHIFT		21
-static inline uint32_t A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN(enum adreno_rb_blend_opcode val)
+static inline uint32_t A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN(enum a2xx_rb_blend_opcode val)
 {
 	return ((val) << A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN__SHIFT) & A2XX_RB_BLEND_CONTROL_ALPHA_COMB_FCN__MASK;
 }
diff --git a/drivers/gpu/drm/msm/adreno/a3xx.xml.h b/drivers/gpu/drm/msm/adreno/a3xx.xml.h
index a7be56163d23..82d015279b47 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx.xml.h
+++ b/drivers/gpu/drm/msm/adreno/a3xx.xml.h
@@ -10,13 +10,13 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32814 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   8900 bytes, from 2013-10-22 23:57:49)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  10574 bytes, from 2013-11-13 05:44:45)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  53644 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (   8344 bytes, from 2013-11-30 14:49:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14960 bytes, from 2014-07-27 17:22:13)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  41068 bytes, from 2014-08-01 12:22:48)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -41,31 +41,11 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 */
 
 
-enum a3xx_render_mode {
-	RB_RENDERING_PASS = 0,
-	RB_TILING_PASS = 1,
-	RB_RESOLVE_PASS = 2,
-};
-
 enum a3xx_tile_mode {
 	LINEAR = 0,
 	TILE_32X32 = 2,
 };
 
-enum a3xx_threadmode {
-	MULTI = 0,
-	SINGLE = 1,
-};
-
-enum a3xx_instrbuffermode {
-	BUFFER = 1,
-};
-
-enum a3xx_threadsize {
-	TWO_QUADS = 0,
-	FOUR_QUADS = 1,
-};
-
 enum a3xx_state_block_id {
 	HLSQ_BLOCK_ID_TP_TEX = 2,
 	HLSQ_BLOCK_ID_TP_MIPMAP = 3,
@@ -169,6 +149,8 @@ enum a3xx_color_fmt {
 	RB_R8G8B8A8_UNORM = 8,
 	RB_Z16_UNORM = 12,
 	RB_A8_UNORM = 20,
+	RB_R16G16B16A16_FLOAT = 27,
+	RB_R32G32B32A32_FLOAT = 51,
 };
 
 enum a3xx_color_swap {
@@ -178,12 +160,6 @@ enum a3xx_color_swap {
 	XYZW = 3,
 };
 
-enum a3xx_msaa_samples {
-	MSAA_ONE = 0,
-	MSAA_TWO = 1,
-	MSAA_FOUR = 2,
-};
-
 enum a3xx_sp_perfcounter_select {
 	SP_FS_CFLOW_INSTRUCTIONS = 12,
 	SP_FS_FULL_ALU_INSTRUCTIONS = 14,
@@ -191,21 +167,45 @@ enum a3xx_sp_perfcounter_select {
 	SP_ALU_ACTIVE_CYCLES = 29,
 };
 
-enum adreno_rb_copy_control_mode {
-	RB_COPY_RESOLVE = 1,
-	RB_COPY_DEPTH_STENCIL = 5,
+enum a3xx_rop_code {
+	ROP_CLEAR = 0,
+	ROP_NOR = 1,
+	ROP_AND_INVERTED = 2,
+	ROP_COPY_INVERTED = 3,
+	ROP_AND_REVERSE = 4,
+	ROP_INVERT = 5,
+	ROP_XOR = 6,
+	ROP_NAND = 7,
+	ROP_AND = 8,
+	ROP_EQUIV = 9,
+	ROP_NOOP = 10,
+	ROP_OR_INVERTED = 11,
+	ROP_COPY = 12,
+	ROP_OR_REVERSE = 13,
+	ROP_OR = 14,
+	ROP_SET = 15,
+};
+
+enum a3xx_rb_blend_opcode {
+	BLEND_DST_PLUS_SRC = 0,
+	BLEND_SRC_MINUS_DST = 1,
+	BLEND_DST_MINUS_SRC = 2,
+	BLEND_MIN_DST_SRC = 3,
+	BLEND_MAX_DST_SRC = 4,
 };
 
 enum a3xx_tex_filter {
 	A3XX_TEX_NEAREST = 0,
 	A3XX_TEX_LINEAR = 1,
+	A3XX_TEX_ANISO = 2,
 };
 
 enum a3xx_tex_clamp {
 	A3XX_TEX_REPEAT = 0,
 	A3XX_TEX_CLAMP_TO_EDGE = 1,
 	A3XX_TEX_MIRROR_REPEAT = 2,
-	A3XX_TEX_CLAMP_NONE = 3,
+	A3XX_TEX_CLAMP_TO_BORDER = 3,
+	A3XX_TEX_MIRROR_CLAMP = 4,
 };
 
 enum a3xx_tex_swiz {
@@ -316,6 +316,7 @@ enum a3xx_tex_type {
 #define REG_A3XX_RBBM_INT_0_STATUS				0x00000064
 
 #define REG_A3XX_RBBM_PERFCTR_CTL				0x00000080
+#define A3XX_RBBM_PERFCTR_CTL_ENABLE				0x00000001
 
 #define REG_A3XX_RBBM_PERFCTR_LOAD_CMD0				0x00000081
 
@@ -549,6 +550,10 @@ static inline uint32_t REG_A3XX_CP_PROTECT_REG(uint32_t i0) { return 0x00000460
 
 #define REG_A3XX_CP_AHB_FAULT					0x0000054d
 
+#define REG_A3XX_SP_GLOBAL_MEM_SIZE				0x00000e22
+
+#define REG_A3XX_SP_GLOBAL_MEM_ADDR				0x00000e23
+
 #define REG_A3XX_GRAS_CL_CLIP_CNTL				0x00002040
 #define A3XX_GRAS_CL_CLIP_CNTL_IJ_PERSP_CENTER			0x00001000
 #define A3XX_GRAS_CL_CLIP_CNTL_CLIP_DISABLE			0x00010000
@@ -556,6 +561,9 @@ static inline uint32_t REG_A3XX_CP_PROTECT_REG(uint32_t i0) { return 0x00000460
 #define A3XX_GRAS_CL_CLIP_CNTL_VP_CLIP_CODE_IGNORE		0x00080000
 #define A3XX_GRAS_CL_CLIP_CNTL_VP_XFORM_DISABLE			0x00100000
 #define A3XX_GRAS_CL_CLIP_CNTL_PERSP_DIVISION_DISABLE		0x00200000
+#define A3XX_GRAS_CL_CLIP_CNTL_ZCOORD				0x00800000
+#define A3XX_GRAS_CL_CLIP_CNTL_WCOORD				0x01000000
+#define A3XX_GRAS_CL_CLIP_CNTL_ZCLIP_DISABLE			0x02000000
 
 #define REG_A3XX_GRAS_CL_GB_CLIP_ADJ				0x00002044
 #define A3XX_GRAS_CL_GB_CLIP_ADJ_HORZ__MASK			0x000003ff
@@ -620,15 +628,33 @@ static inline uint32_t A3XX_GRAS_CL_VPORT_ZSCALE(float val)
 }
 
 #define REG_A3XX_GRAS_SU_POINT_MINMAX				0x00002068
+#define A3XX_GRAS_SU_POINT_MINMAX_MIN__MASK			0x0000ffff
+#define A3XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT			0
+static inline uint32_t A3XX_GRAS_SU_POINT_MINMAX_MIN(float val)
+{
+	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_MINMAX_MIN__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MIN__MASK;
+}
+#define A3XX_GRAS_SU_POINT_MINMAX_MAX__MASK			0xffff0000
+#define A3XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT			16
+static inline uint32_t A3XX_GRAS_SU_POINT_MINMAX_MAX(float val)
+{
+	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_MINMAX_MAX__SHIFT) & A3XX_GRAS_SU_POINT_MINMAX_MAX__MASK;
+}
 
 #define REG_A3XX_GRAS_SU_POINT_SIZE				0x00002069
+#define A3XX_GRAS_SU_POINT_SIZE__MASK				0xffffffff
+#define A3XX_GRAS_SU_POINT_SIZE__SHIFT				0
+static inline uint32_t A3XX_GRAS_SU_POINT_SIZE(float val)
+{
+	return ((((uint32_t)(val * 8.0))) << A3XX_GRAS_SU_POINT_SIZE__SHIFT) & A3XX_GRAS_SU_POINT_SIZE__MASK;
+}
 
 #define REG_A3XX_GRAS_SU_POLY_OFFSET_SCALE			0x0000206c
 #define A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__MASK		0x00ffffff
 #define A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT		0
 static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL(float val)
 {
-	return ((((uint32_t)(val * 40.0))) << A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__MASK;
+	return ((((uint32_t)(val * 28.0))) << A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_POLY_OFFSET_OFFSET			0x0000206d
@@ -636,7 +662,7 @@ static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_SCALE_VAL(float val)
 #define A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT			0
 static inline uint32_t A3XX_GRAS_SU_POLY_OFFSET_OFFSET(float val)
 {
-	return ((((uint32_t)(val * 44.0))) << A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK;
+	return ((((uint32_t)(val * 28.0))) << A3XX_GRAS_SU_POLY_OFFSET_OFFSET__SHIFT) & A3XX_GRAS_SU_POLY_OFFSET_OFFSET__MASK;
 }
 
 #define REG_A3XX_GRAS_SU_MODE_CONTROL				0x00002070
@@ -743,6 +769,7 @@ static inline uint32_t A3XX_RB_MODE_CONTROL_RENDER_MODE(enum a3xx_render_mode va
 #define A3XX_RB_MODE_CONTROL_PACKER_TIMER_ENABLE		0x00010000
 
 #define REG_A3XX_RB_RENDER_CONTROL				0x000020c1
+#define A3XX_RB_RENDER_CONTROL_FACENESS				0x00000008
 #define A3XX_RB_RENDER_CONTROL_BIN_WIDTH__MASK			0x00000ff0
 #define A3XX_RB_RENDER_CONTROL_BIN_WIDTH__SHIFT			4
 static inline uint32_t A3XX_RB_RENDER_CONTROL_BIN_WIDTH(uint32_t val)
@@ -751,6 +778,10 @@ static inline uint32_t A3XX_RB_RENDER_CONTROL_BIN_WIDTH(uint32_t val)
 }
 #define A3XX_RB_RENDER_CONTROL_DISABLE_COLOR_PIPE		0x00001000
 #define A3XX_RB_RENDER_CONTROL_ENABLE_GMEM			0x00002000
+#define A3XX_RB_RENDER_CONTROL_XCOORD				0x00004000
+#define A3XX_RB_RENDER_CONTROL_YCOORD				0x00008000
+#define A3XX_RB_RENDER_CONTROL_ZCOORD				0x00010000
+#define A3XX_RB_RENDER_CONTROL_WCOORD				0x00020000
 #define A3XX_RB_RENDER_CONTROL_ALPHA_TEST			0x00400000
 #define A3XX_RB_RENDER_CONTROL_ALPHA_TEST_FUNC__MASK		0x07000000
 #define A3XX_RB_RENDER_CONTROL_ALPHA_TEST_FUNC__SHIFT		24
@@ -796,7 +827,7 @@ static inline uint32_t REG_A3XX_RB_MRT_CONTROL(uint32_t i0) { return 0x000020c4
 #define A3XX_RB_MRT_CONTROL_BLEND2				0x00000020
 #define A3XX_RB_MRT_CONTROL_ROP_CODE__MASK			0x00000f00
 #define A3XX_RB_MRT_CONTROL_ROP_CODE__SHIFT			8
-static inline uint32_t A3XX_RB_MRT_CONTROL_ROP_CODE(uint32_t val)
+static inline uint32_t A3XX_RB_MRT_CONTROL_ROP_CODE(enum a3xx_rop_code val)
 {
 	return ((val) << A3XX_RB_MRT_CONTROL_ROP_CODE__SHIFT) & A3XX_RB_MRT_CONTROL_ROP_CODE__MASK;
 }
@@ -856,7 +887,7 @@ static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_RGB_SRC_FACTOR(enum adreno_rb_b
 }
 #define A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__MASK	0x000000e0
 #define A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__SHIFT	5
-static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE(enum adreno_rb_blend_opcode val)
+static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE(enum a3xx_rb_blend_opcode val)
 {
 	return ((val) << A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__SHIFT) & A3XX_RB_MRT_BLEND_CONTROL_RGB_BLEND_OPCODE__MASK;
 }
@@ -874,7 +905,7 @@ static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_ALPHA_SRC_FACTOR(enum adreno_rb
 }
 #define A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__MASK	0x00e00000
 #define A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__SHIFT	21
-static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE(enum adreno_rb_blend_opcode val)
+static inline uint32_t A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE(enum a3xx_rb_blend_opcode val)
 {
 	return ((val) << A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__SHIFT) & A3XX_RB_MRT_BLEND_CONTROL_ALPHA_BLEND_OPCODE__MASK;
 }
@@ -957,17 +988,24 @@ static inline uint32_t A3XX_RB_COPY_CONTROL_MSAA_RESOLVE(enum a3xx_msaa_samples
 {
 	return ((val) << A3XX_RB_COPY_CONTROL_MSAA_RESOLVE__SHIFT) & A3XX_RB_COPY_CONTROL_MSAA_RESOLVE__MASK;
 }
+#define A3XX_RB_COPY_CONTROL_DEPTHCLEAR				0x00000008
 #define A3XX_RB_COPY_CONTROL_MODE__MASK				0x00000070
 #define A3XX_RB_COPY_CONTROL_MODE__SHIFT			4
 static inline uint32_t A3XX_RB_COPY_CONTROL_MODE(enum adreno_rb_copy_control_mode val)
 {
 	return ((val) << A3XX_RB_COPY_CONTROL_MODE__SHIFT) & A3XX_RB_COPY_CONTROL_MODE__MASK;
 }
-#define A3XX_RB_COPY_CONTROL_GMEM_BASE__MASK			0xfffffc00
-#define A3XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT			10
+#define A3XX_RB_COPY_CONTROL_FASTCLEAR__MASK			0x00000f00
+#define A3XX_RB_COPY_CONTROL_FASTCLEAR__SHIFT			8
+static inline uint32_t A3XX_RB_COPY_CONTROL_FASTCLEAR(uint32_t val)
+{
+	return ((val) << A3XX_RB_COPY_CONTROL_FASTCLEAR__SHIFT) & A3XX_RB_COPY_CONTROL_FASTCLEAR__MASK;
+}
+#define A3XX_RB_COPY_CONTROL_GMEM_BASE__MASK			0xffffc000
+#define A3XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT			14
 static inline uint32_t A3XX_RB_COPY_CONTROL_GMEM_BASE(uint32_t val)
 {
-	return ((val >> 10) << A3XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT) & A3XX_RB_COPY_CONTROL_GMEM_BASE__MASK;
+	return ((val >> 14) << A3XX_RB_COPY_CONTROL_GMEM_BASE__SHIFT) & A3XX_RB_COPY_CONTROL_GMEM_BASE__MASK;
 }
 
 #define REG_A3XX_RB_COPY_DEST_BASE				0x000020ed
@@ -1005,6 +1043,12 @@ static inline uint32_t A3XX_RB_COPY_DEST_INFO_SWAP(enum a3xx_color_swap val)
 {
 	return ((val) << A3XX_RB_COPY_DEST_INFO_SWAP__SHIFT) & A3XX_RB_COPY_DEST_INFO_SWAP__MASK;
 }
+#define A3XX_RB_COPY_DEST_INFO_DITHER_MODE__MASK		0x00000c00
+#define A3XX_RB_COPY_DEST_INFO_DITHER_MODE__SHIFT		10
+static inline uint32_t A3XX_RB_COPY_DEST_INFO_DITHER_MODE(enum adreno_rb_dither_mode val)
+{
+	return ((val) << A3XX_RB_COPY_DEST_INFO_DITHER_MODE__SHIFT) & A3XX_RB_COPY_DEST_INFO_DITHER_MODE__MASK;
+}
 #define A3XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__MASK		0x0003c000
 #define A3XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE__SHIFT		14
 static inline uint32_t A3XX_RB_COPY_DEST_INFO_COMPONENT_ENABLE(uint32_t val)
@@ -1019,6 +1063,7 @@ static inline uint32_t A3XX_RB_COPY_DEST_INFO_ENDIAN(enum adreno_rb_surface_endi
 }
 
 #define REG_A3XX_RB_DEPTH_CONTROL				0x00002100
+#define A3XX_RB_DEPTH_CONTROL_FRAG_WRITES_Z			0x00000001
 #define A3XX_RB_DEPTH_CONTROL_Z_ENABLE				0x00000002
 #define A3XX_RB_DEPTH_CONTROL_Z_WRITE_ENABLE			0x00000004
 #define A3XX_RB_DEPTH_CONTROL_EARLY_Z_DISABLE			0x00000008
@@ -1044,7 +1089,7 @@ static inline uint32_t A3XX_RB_DEPTH_INFO_DEPTH_FORMAT(enum adreno_rb_depth_form
 #define A3XX_RB_DEPTH_INFO_DEPTH_BASE__SHIFT			11
 static inline uint32_t A3XX_RB_DEPTH_INFO_DEPTH_BASE(uint32_t val)
 {
-	return ((val >> 10) << A3XX_RB_DEPTH_INFO_DEPTH_BASE__SHIFT) & A3XX_RB_DEPTH_INFO_DEPTH_BASE__MASK;
+	return ((val >> 12) << A3XX_RB_DEPTH_INFO_DEPTH_BASE__SHIFT) & A3XX_RB_DEPTH_INFO_DEPTH_BASE__MASK;
 }
 
 #define REG_A3XX_RB_DEPTH_PITCH					0x00002103
@@ -1172,6 +1217,8 @@ static inline uint32_t A3XX_RB_WINDOW_OFFSET_Y(uint32_t val)
 }
 
 #define REG_A3XX_RB_SAMPLE_COUNT_CONTROL			0x00002110
+#define A3XX_RB_SAMPLE_COUNT_CONTROL_RESET			0x00000001
+#define A3XX_RB_SAMPLE_COUNT_CONTROL_COPY			0x00000002
 
 #define REG_A3XX_RB_SAMPLE_COUNT_ADDR				0x00002111
 
@@ -1179,7 +1226,23 @@ static inline uint32_t A3XX_RB_WINDOW_OFFSET_Y(uint32_t val)
 
 #define REG_A3XX_RB_Z_CLAMP_MAX					0x00002115
 
+#define REG_A3XX_VGT_BIN_BASE					0x000021e1
+
+#define REG_A3XX_VGT_BIN_SIZE					0x000021e2
+
 #define REG_A3XX_PC_VSTREAM_CONTROL				0x000021e4
+#define A3XX_PC_VSTREAM_CONTROL_SIZE__MASK			0x003f0000
+#define A3XX_PC_VSTREAM_CONTROL_SIZE__SHIFT			16
+static inline uint32_t A3XX_PC_VSTREAM_CONTROL_SIZE(uint32_t val)
+{
+	return ((val) << A3XX_PC_VSTREAM_CONTROL_SIZE__SHIFT) & A3XX_PC_VSTREAM_CONTROL_SIZE__MASK;
+}
+#define A3XX_PC_VSTREAM_CONTROL_N__MASK				0x07c00000
+#define A3XX_PC_VSTREAM_CONTROL_N__SHIFT			22
+static inline uint32_t A3XX_PC_VSTREAM_CONTROL_N(uint32_t val)
+{
+	return ((val) << A3XX_PC_VSTREAM_CONTROL_N__SHIFT) & A3XX_PC_VSTREAM_CONTROL_N__MASK;
+}
 
 #define REG_A3XX_PC_VERTEX_REUSE_BLOCK_CNTL			0x000021ea
 
@@ -1203,6 +1266,7 @@ static inline uint32_t A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE(enum adreno_pa_
 	return ((val) << A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE__SHIFT) & A3XX_PC_PRIM_VTX_CNTL_POLYMODE_BACK_PTYPE__MASK;
 }
 #define A3XX_PC_PRIM_VTX_CNTL_PROVOKING_VTX_LAST		0x02000000
+#define A3XX_PC_PRIM_VTX_CNTL_PSIZE				0x04000000
 
 #define REG_A3XX_PC_RESTART_INDEX				0x000021ed
 
@@ -1232,6 +1296,7 @@ static inline uint32_t A3XX_HLSQ_CONTROL_1_REG_VSTHREADSIZE(enum a3xx_threadsize
 }
 #define A3XX_HLSQ_CONTROL_1_REG_VSSUPERTHREADENABLE		0x00000100
 #define A3XX_HLSQ_CONTROL_1_REG_RESERVED1			0x00000200
+#define A3XX_HLSQ_CONTROL_1_REG_ZWCOORD				0x02000000
 
 #define REG_A3XX_HLSQ_CONTROL_2_REG				0x00002202
 #define A3XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD__MASK	0xfc000000
@@ -1242,6 +1307,12 @@ static inline uint32_t A3XX_HLSQ_CONTROL_2_REG_PRIMALLOCTHRESHOLD(uint32_t val)
 }
 
 #define REG_A3XX_HLSQ_CONTROL_3_REG				0x00002203
+#define A3XX_HLSQ_CONTROL_3_REG_REGID__MASK			0x000000ff
+#define A3XX_HLSQ_CONTROL_3_REG_REGID__SHIFT			0
+static inline uint32_t A3XX_HLSQ_CONTROL_3_REG_REGID(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CONTROL_3_REG_REGID__SHIFT) & A3XX_HLSQ_CONTROL_3_REG_REGID__MASK;
+}
 
 #define REG_A3XX_HLSQ_VS_CONTROL_REG				0x00002204
 #define A3XX_HLSQ_VS_CONTROL_REG_CONSTLENGTH__MASK		0x00000fff
@@ -1312,10 +1383,36 @@ static inline uint32_t A3XX_HLSQ_CONST_FSPRESV_RANGE_REG_ENDENTRY(uint32_t val)
 }
 
 #define REG_A3XX_HLSQ_CL_NDRANGE_0_REG				0x0000220a
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_WORKDIM__MASK		0x00000003
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_WORKDIM__SHIFT		0
+static inline uint32_t A3XX_HLSQ_CL_NDRANGE_0_REG_WORKDIM(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CL_NDRANGE_0_REG_WORKDIM__SHIFT) & A3XX_HLSQ_CL_NDRANGE_0_REG_WORKDIM__MASK;
+}
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE0__MASK		0x00000ffc
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE0__SHIFT		2
+static inline uint32_t A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE0(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE0__SHIFT) & A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE0__MASK;
+}
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE1__MASK		0x003ff000
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE1__SHIFT		12
+static inline uint32_t A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE1(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE1__SHIFT) & A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE1__MASK;
+}
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE2__MASK		0xffc00000
+#define A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE2__SHIFT		22
+static inline uint32_t A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE2(uint32_t val)
+{
+	return ((val) << A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE2__SHIFT) & A3XX_HLSQ_CL_NDRANGE_0_REG_LOCALSIZE2__MASK;
+}
+
+static inline uint32_t REG_A3XX_HLSQ_CL_GLOBAL_WORK(uint32_t i0) { return 0x0000220b + 0x2*i0; }
 
-#define REG_A3XX_HLSQ_CL_NDRANGE_1_REG				0x0000220b
+static inline uint32_t REG_A3XX_HLSQ_CL_GLOBAL_WORK_SIZE(uint32_t i0) { return 0x0000220b + 0x2*i0; }
 
-#define REG_A3XX_HLSQ_CL_NDRANGE_2_REG				0x0000220c
+static inline uint32_t REG_A3XX_HLSQ_CL_GLOBAL_WORK_OFFSET(uint32_t i0) { return 0x0000220c + 0x2*i0; }
 
 #define REG_A3XX_HLSQ_CL_CONTROL_0_REG				0x00002211
 
@@ -1323,7 +1420,9 @@ static inline uint32_t A3XX_HLSQ_CONST_FSPRESV_RANGE_REG_ENDENTRY(uint32_t val)
 
 #define REG_A3XX_HLSQ_CL_KERNEL_CONST_REG			0x00002214
 
-#define REG_A3XX_HLSQ_CL_KERNEL_GROUP_X_REG			0x00002215
+static inline uint32_t REG_A3XX_HLSQ_CL_KERNEL_GROUP(uint32_t i0) { return 0x00002215 + 0x1*i0; }
+
+static inline uint32_t REG_A3XX_HLSQ_CL_KERNEL_GROUP_RATIO(uint32_t i0) { return 0x00002215 + 0x1*i0; }
 
 #define REG_A3XX_HLSQ_CL_KERNEL_GROUP_Y_REG			0x00002216
 
@@ -1438,6 +1537,12 @@ static inline uint32_t A3XX_VFD_DECODE_INSTR_REGID(uint32_t val)
 {
 	return ((val) << A3XX_VFD_DECODE_INSTR_REGID__SHIFT) & A3XX_VFD_DECODE_INSTR_REGID__MASK;
 }
+#define A3XX_VFD_DECODE_INSTR_SWAP__MASK			0x00c00000
+#define A3XX_VFD_DECODE_INSTR_SWAP__SHIFT			22
+static inline uint32_t A3XX_VFD_DECODE_INSTR_SWAP(enum a3xx_color_swap val)
+{
+	return ((val) << A3XX_VFD_DECODE_INSTR_SWAP__SHIFT) & A3XX_VFD_DECODE_INSTR_SWAP__MASK;
+}
 #define A3XX_VFD_DECODE_INSTR_SHIFTCNT__MASK			0x1f000000
 #define A3XX_VFD_DECODE_INSTR_SHIFTCNT__SHIFT			24
 static inline uint32_t A3XX_VFD_DECODE_INSTR_SHIFTCNT(uint32_t val)
@@ -1462,12 +1567,13 @@ static inline uint32_t A3XX_VFD_VS_THREADING_THRESHOLD_REGID_VTXCNT(uint32_t val
 }
 
 #define REG_A3XX_VPC_ATTR					0x00002280
-#define A3XX_VPC_ATTR_TOTALATTR__MASK				0x00000fff
+#define A3XX_VPC_ATTR_TOTALATTR__MASK				0x000001ff
 #define A3XX_VPC_ATTR_TOTALATTR__SHIFT				0
 static inline uint32_t A3XX_VPC_ATTR_TOTALATTR(uint32_t val)
 {
 	return ((val) << A3XX_VPC_ATTR_TOTALATTR__SHIFT) & A3XX_VPC_ATTR_TOTALATTR__MASK;
 }
+#define A3XX_VPC_ATTR_PSIZE					0x00000200
 #define A3XX_VPC_ATTR_THRDASSIGN__MASK				0x0ffff000
 #define A3XX_VPC_ATTR_THRDASSIGN__SHIFT				12
 static inline uint32_t A3XX_VPC_ATTR_THRDASSIGN(uint32_t val)
@@ -1522,11 +1628,11 @@ static inline uint32_t A3XX_SP_SP_CTRL_REG_SLEEPMODE(uint32_t val)
 {
 	return ((val) << A3XX_SP_SP_CTRL_REG_SLEEPMODE__SHIFT) & A3XX_SP_SP_CTRL_REG_SLEEPMODE__MASK;
 }
-#define A3XX_SP_SP_CTRL_REG_LOMODE__MASK			0x00c00000
-#define A3XX_SP_SP_CTRL_REG_LOMODE__SHIFT			22
-static inline uint32_t A3XX_SP_SP_CTRL_REG_LOMODE(uint32_t val)
+#define A3XX_SP_SP_CTRL_REG_L0MODE__MASK			0x00c00000
+#define A3XX_SP_SP_CTRL_REG_L0MODE__SHIFT			22
+static inline uint32_t A3XX_SP_SP_CTRL_REG_L0MODE(uint32_t val)
 {
-	return ((val) << A3XX_SP_SP_CTRL_REG_LOMODE__SHIFT) & A3XX_SP_SP_CTRL_REG_LOMODE__MASK;
+	return ((val) << A3XX_SP_SP_CTRL_REG_L0MODE__SHIFT) & A3XX_SP_SP_CTRL_REG_L0MODE__MASK;
 }
 
 #define REG_A3XX_SP_VS_CTRL_REG0				0x000022c4
@@ -1569,6 +1675,7 @@ static inline uint32_t A3XX_SP_VS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val)
 }
 #define A3XX_SP_VS_CTRL_REG0_SUPERTHREADMODE			0x00200000
 #define A3XX_SP_VS_CTRL_REG0_PIXLODENABLE			0x00400000
+#define A3XX_SP_VS_CTRL_REG0_COMPUTEMODE			0x00800000
 #define A3XX_SP_VS_CTRL_REG0_LENGTH__MASK			0xff000000
 #define A3XX_SP_VS_CTRL_REG0_LENGTH__SHIFT			24
 static inline uint32_t A3XX_SP_VS_CTRL_REG0_LENGTH(uint32_t val)
@@ -1589,7 +1696,7 @@ static inline uint32_t A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT(uint32_t val)
 {
 	return ((val) << A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT__SHIFT) & A3XX_SP_VS_CTRL_REG1_CONSTFOOTPRINT__MASK;
 }
-#define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK		0x3f000000
+#define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__MASK		0x7f000000
 #define A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING__SHIFT		24
 static inline uint32_t A3XX_SP_VS_CTRL_REG1_INITIALOUTSTANDING(uint32_t val)
 {
@@ -1742,6 +1849,7 @@ static inline uint32_t A3XX_SP_FS_CTRL_REG0_THREADSIZE(enum a3xx_threadsize val)
 }
 #define A3XX_SP_FS_CTRL_REG0_SUPERTHREADMODE			0x00200000
 #define A3XX_SP_FS_CTRL_REG0_PIXLODENABLE			0x00400000
+#define A3XX_SP_FS_CTRL_REG0_COMPUTEMODE			0x00800000
 #define A3XX_SP_FS_CTRL_REG0_LENGTH__MASK			0xff000000
 #define A3XX_SP_FS_CTRL_REG0_LENGTH__SHIFT			24
 static inline uint32_t A3XX_SP_FS_CTRL_REG0_LENGTH(uint32_t val)
@@ -1802,6 +1910,13 @@ static inline uint32_t A3XX_SP_FS_OBJ_OFFSET_REG_SHADEROBJOFFSET(uint32_t val)
 #define REG_A3XX_SP_FS_FLAT_SHAD_MODE_REG_1			0x000022e9
 
 #define REG_A3XX_SP_FS_OUTPUT_REG				0x000022ec
+#define A3XX_SP_FS_OUTPUT_REG_DEPTH_ENABLE			0x00000080
+#define A3XX_SP_FS_OUTPUT_REG_DEPTH_REGID__MASK			0x0000ff00
+#define A3XX_SP_FS_OUTPUT_REG_DEPTH_REGID__SHIFT		8
+static inline uint32_t A3XX_SP_FS_OUTPUT_REG_DEPTH_REGID(uint32_t val)
+{
+	return ((val) << A3XX_SP_FS_OUTPUT_REG_DEPTH_REGID__SHIFT) & A3XX_SP_FS_OUTPUT_REG_DEPTH_REGID__MASK;
+}
 
 static inline uint32_t REG_A3XX_SP_FS_MRT(uint32_t i0) { return 0x000022f0 + 0x1*i0; }
 
@@ -1914,6 +2029,42 @@ static inline uint32_t A3XX_TPL1_TP_FS_TEX_OFFSET_BASETABLEPTR(uint32_t val)
 
 #define REG_A3XX_VBIF_OUT_AXI_AOOO				0x0000305f
 
+#define REG_A3XX_VBIF_PERF_CNT_EN				0x00003070
+#define A3XX_VBIF_PERF_CNT_EN_CNT0				0x00000001
+#define A3XX_VBIF_PERF_CNT_EN_CNT1				0x00000002
+#define A3XX_VBIF_PERF_CNT_EN_PWRCNT0				0x00000004
+#define A3XX_VBIF_PERF_CNT_EN_PWRCNT1				0x00000008
+#define A3XX_VBIF_PERF_CNT_EN_PWRCNT2				0x00000010
+
+#define REG_A3XX_VBIF_PERF_CNT_CLR				0x00003071
+#define A3XX_VBIF_PERF_CNT_CLR_CNT0				0x00000001
+#define A3XX_VBIF_PERF_CNT_CLR_CNT1				0x00000002
+#define A3XX_VBIF_PERF_CNT_CLR_PWRCNT0				0x00000004
+#define A3XX_VBIF_PERF_CNT_CLR_PWRCNT1				0x00000008
+#define A3XX_VBIF_PERF_CNT_CLR_PWRCNT2				0x00000010
+
+#define REG_A3XX_VBIF_PERF_CNT_SEL				0x00003072
+
+#define REG_A3XX_VBIF_PERF_CNT0_LO				0x00003073
+
+#define REG_A3XX_VBIF_PERF_CNT0_HI				0x00003074
+
+#define REG_A3XX_VBIF_PERF_CNT1_LO				0x00003075
+
+#define REG_A3XX_VBIF_PERF_CNT1_HI				0x00003076
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT0_LO				0x00003077
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT0_HI				0x00003078
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT1_LO				0x00003079
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT1_HI				0x0000307a
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT2_LO				0x0000307b
+
+#define REG_A3XX_VBIF_PERF_PWR_CNT2_HI				0x0000307c
+
 #define REG_A3XX_VSC_BIN_SIZE					0x00000c01
 #define A3XX_VSC_BIN_SIZE_WIDTH__MASK				0x0000001f
 #define A3XX_VSC_BIN_SIZE_WIDTH__SHIFT				0
@@ -2080,6 +2231,8 @@ static inline uint32_t A3XX_UCHE_CACHE_INVALIDATE1_REG_OPCODE(enum a3xx_cache_op
 }
 #define A3XX_UCHE_CACHE_INVALIDATE1_REG_ENTIRE_CACHE		0x80000000
 
+#define REG_A3XX_UNKNOWN_0EA6					0x00000ea6
+
 #define REG_A3XX_SP_PERFCOUNTER0_SELECT				0x00000ec4
 
 #define REG_A3XX_SP_PERFCOUNTER1_SELECT				0x00000ec5
@@ -2117,6 +2270,39 @@ static inline uint32_t A3XX_UCHE_CACHE_INVALIDATE1_REG_OPCODE(enum a3xx_cache_op
 #define REG_A3XX_VGT_EVENT_INITIATOR				0x000021f9
 
 #define REG_A3XX_VGT_DRAW_INITIATOR				0x000021fc
+#define A3XX_VGT_DRAW_INITIATOR_PRIM_TYPE__MASK			0x0000003f
+#define A3XX_VGT_DRAW_INITIATOR_PRIM_TYPE__SHIFT		0
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_PRIM_TYPE(enum pc_di_primtype val)
+{
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_PRIM_TYPE__SHIFT) & A3XX_VGT_DRAW_INITIATOR_PRIM_TYPE__MASK;
+}
+#define A3XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__MASK		0x000000c0
+#define A3XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__SHIFT		6
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_SOURCE_SELECT(enum pc_di_src_sel val)
+{
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__SHIFT) & A3XX_VGT_DRAW_INITIATOR_SOURCE_SELECT__MASK;
+}
+#define A3XX_VGT_DRAW_INITIATOR_VIS_CULL__MASK			0x00000600
+#define A3XX_VGT_DRAW_INITIATOR_VIS_CULL__SHIFT			9
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_VIS_CULL(enum pc_di_vis_cull_mode val)
+{
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_VIS_CULL__SHIFT) & A3XX_VGT_DRAW_INITIATOR_VIS_CULL__MASK;
+}
+#define A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE__MASK		0x00000800
+#define A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE__SHIFT		11
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE(enum pc_di_index_size val)
+{
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE__SHIFT) & A3XX_VGT_DRAW_INITIATOR_INDEX_SIZE__MASK;
+}
+#define A3XX_VGT_DRAW_INITIATOR_NOT_EOP				0x00001000
+#define A3XX_VGT_DRAW_INITIATOR_SMALL_INDEX			0x00002000
+#define A3XX_VGT_DRAW_INITIATOR_PRE_DRAW_INITIATOR_ENABLE	0x00004000
+#define A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK		0xffff0000
+#define A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT		16
+static inline uint32_t A3XX_VGT_DRAW_INITIATOR_NUM_INDICES(uint32_t val)
+{
+	return ((val) << A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__SHIFT) & A3XX_VGT_DRAW_INITIATOR_NUM_INDICES__MASK;
+}
 
 #define REG_A3XX_VGT_IMMED_DATA					0x000021fd
 
@@ -2152,6 +2338,12 @@ static inline uint32_t A3XX_TEX_SAMP_0_WRAP_R(enum a3xx_tex_clamp val)
 {
 	return ((val) << A3XX_TEX_SAMP_0_WRAP_R__SHIFT) & A3XX_TEX_SAMP_0_WRAP_R__MASK;
 }
+#define A3XX_TEX_SAMP_0_COMPARE_FUNC__MASK			0x00700000
+#define A3XX_TEX_SAMP_0_COMPARE_FUNC__SHIFT			20
+static inline uint32_t A3XX_TEX_SAMP_0_COMPARE_FUNC(enum adreno_compare_func val)
+{
+	return ((val) << A3XX_TEX_SAMP_0_COMPARE_FUNC__SHIFT) & A3XX_TEX_SAMP_0_COMPARE_FUNC__MASK;
+}
 #define A3XX_TEX_SAMP_0_UNNORM_COORDS				0x80000000
 
 #define REG_A3XX_TEX_SAMP_1					0x00000001
@@ -2170,6 +2362,7 @@ static inline uint32_t A3XX_TEX_SAMP_1_MIN_LOD(float val)
 
 #define REG_A3XX_TEX_CONST_0					0x00000000
 #define A3XX_TEX_CONST_0_TILED					0x00000001
+#define A3XX_TEX_CONST_0_SRGB					0x00000004
 #define A3XX_TEX_CONST_0_SWIZ_X__MASK				0x00000070
 #define A3XX_TEX_CONST_0_SWIZ_X__SHIFT				4
 static inline uint32_t A3XX_TEX_CONST_0_SWIZ_X(enum a3xx_tex_swiz val)
@@ -2206,6 +2399,7 @@ static inline uint32_t A3XX_TEX_CONST_0_FMT(enum a3xx_tex_fmt val)
 {
 	return ((val) << A3XX_TEX_CONST_0_FMT__SHIFT) & A3XX_TEX_CONST_0_FMT__MASK;
 }
+#define A3XX_TEX_CONST_0_NOCONVERT				0x20000000
 #define A3XX_TEX_CONST_0_TYPE__MASK				0xc0000000
 #define A3XX_TEX_CONST_0_TYPE__SHIFT				30
 static inline uint32_t A3XX_TEX_CONST_0_TYPE(enum a3xx_tex_type val)
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index 942e09d898a8..218c5b060398 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -35,10 +35,8 @@
 	 A3XX_INT0_CP_AHB_ERROR_HALT |     \
 	 A3XX_INT0_UCHE_OOB_ACCESS)
 
+extern bool hang_debug;
 
-static bool hang_debug = false;
-MODULE_PARM_DESC(hang_debug, "Dump registers when hang is detected (can be slow!)");
-module_param_named(hang_debug, hang_debug, bool, 0600);
 static void a3xx_dump(struct msm_gpu *gpu);
 
 static void a3xx_me_init(struct msm_gpu *gpu)
@@ -387,63 +385,26 @@ static const unsigned int a3xx_registers[] = {
 	0x2750, 0x2756, 0x2760, 0x2760, 0x300c, 0x300e, 0x301c, 0x301d,
 	0x302a, 0x302a, 0x302c, 0x302d, 0x3030, 0x3031, 0x3034, 0x3036,
 	0x303c, 0x303c, 0x305e, 0x305f,
+	~0   /* sentinel */
 };
 
 #ifdef CONFIG_DEBUG_FS
 static void a3xx_show(struct msm_gpu *gpu, struct seq_file *m)
 {
-	struct drm_device *dev = gpu->dev;
-	int i;
-
-	adreno_show(gpu, m);
-
-	mutex_lock(&dev->struct_mutex);
-
 	gpu->funcs->pm_resume(gpu);
-
 	seq_printf(m, "status:   %08x\n",
 			gpu_read(gpu, REG_A3XX_RBBM_STATUS));
-
-	/* dump these out in a form that can be parsed by demsm: */
-	seq_printf(m, "IO:region %s 00000000 00020000\n", gpu->name);
-	for (i = 0; i < ARRAY_SIZE(a3xx_registers); i += 2) {
-		uint32_t start = a3xx_registers[i];
-		uint32_t end   = a3xx_registers[i+1];
-		uint32_t addr;
-
-		for (addr = start; addr <= end; addr++) {
-			uint32_t val = gpu_read(gpu, addr);
-			seq_printf(m, "IO:R %08x %08x\n", addr<<2, val);
-		}
-	}
-
 	gpu->funcs->pm_suspend(gpu);
-
-	mutex_unlock(&dev->struct_mutex);
+	adreno_show(gpu, m);
 }
 #endif
 
 /* would be nice to not have to duplicate the _show() stuff with printk(): */
 static void a3xx_dump(struct msm_gpu *gpu)
 {
-	int i;
-
-	adreno_dump(gpu);
 	printk("status:   %08x\n",
 			gpu_read(gpu, REG_A3XX_RBBM_STATUS));
-
-	/* dump these out in a form that can be parsed by demsm: */
-	printk("IO:region %s 00000000 00020000\n", gpu->name);
-	for (i = 0; i < ARRAY_SIZE(a3xx_registers); i += 2) {
-		uint32_t start = a3xx_registers[i];
-		uint32_t end   = a3xx_registers[i+1];
-		uint32_t addr;
-
-		for (addr = start; addr <= end; addr++) {
-			uint32_t val = gpu_read(gpu, addr);
-			printk("IO:R %08x %08x\n", addr<<2, val);
-		}
-	}
+	adreno_dump(gpu);
 }
 
 static const struct adreno_gpu_funcs funcs = {
@@ -479,7 +440,6 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 	struct msm_gpu *gpu;
 	struct msm_drm_private *priv = dev->dev_private;
 	struct platform_device *pdev = priv->gpu_pdev;
-	struct adreno_platform_config *config;
 	int ret;
 
 	if (!pdev) {
@@ -488,8 +448,6 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 		goto fail;
 	}
 
-	config = pdev->dev.platform_data;
-
 	a3xx_gpu = kzalloc(sizeof(*a3xx_gpu), GFP_KERNEL);
 	if (!a3xx_gpu) {
 		ret = -ENOMEM;
@@ -501,20 +459,12 @@ struct msm_gpu *a3xx_gpu_init(struct drm_device *dev)
 
 	a3xx_gpu->pdev = pdev;
 
-	gpu->fast_rate = config->fast_rate;
-	gpu->slow_rate = config->slow_rate;
-	gpu->bus_freq  = config->bus_freq;
-#ifdef CONFIG_MSM_BUS_SCALING
-	gpu->bus_scale_table = config->bus_scale_table;
-#endif
-
-	DBG("fast_rate=%u, slow_rate=%u, bus_freq=%u",
-			gpu->fast_rate, gpu->slow_rate, gpu->bus_freq);
-
 	gpu->perfcntrs = perfcntrs;
 	gpu->num_perfcntrs = ARRAY_SIZE(perfcntrs);
 
-	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, config->rev);
+	adreno_gpu->registers = a3xx_registers;
+
+	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs);
 	if (ret)
 		goto fail;
 
@@ -554,156 +504,3 @@ fail:
 
 	return ERR_PTR(ret);
 }
-
-/*
- * The a3xx device:
- */
-
-#if defined(CONFIG_MSM_BUS_SCALING) && !defined(CONFIG_OF)
-#  include <mach/kgsl.h>
-#endif
-
-static void set_gpu_pdev(struct drm_device *dev,
-		struct platform_device *pdev)
-{
-	struct msm_drm_private *priv = dev->dev_private;
-	priv->gpu_pdev = pdev;
-}
-
-static int a3xx_bind(struct device *dev, struct device *master, void *data)
-{
-	static struct adreno_platform_config config = {};
-#ifdef CONFIG_OF
-	struct device_node *child, *node = dev->of_node;
-	u32 val;
-	int ret;
-
-	ret = of_property_read_u32(node, "qcom,chipid", &val);
-	if (ret) {
-		dev_err(dev, "could not find chipid: %d\n", ret);
-		return ret;
-	}
-
-	config.rev = ADRENO_REV((val >> 24) & 0xff,
-			(val >> 16) & 0xff, (val >> 8) & 0xff, val & 0xff);
-
-	/* find clock rates: */
-	config.fast_rate = 0;
-	config.slow_rate = ~0;
-	for_each_child_of_node(node, child) {
-		if (of_device_is_compatible(child, "qcom,gpu-pwrlevels")) {
-			struct device_node *pwrlvl;
-			for_each_child_of_node(child, pwrlvl) {
-				ret = of_property_read_u32(pwrlvl, "qcom,gpu-freq", &val);
-				if (ret) {
-					dev_err(dev, "could not find gpu-freq: %d\n", ret);
-					return ret;
-				}
-				config.fast_rate = max(config.fast_rate, val);
-				config.slow_rate = min(config.slow_rate, val);
-			}
-		}
-	}
-
-	if (!config.fast_rate) {
-		dev_err(dev, "could not find clk rates\n");
-		return -ENXIO;
-	}
-
-#else
-	struct kgsl_device_platform_data *pdata = dev->platform_data;
-	uint32_t version = socinfo_get_version();
-	if (cpu_is_apq8064ab()) {
-		config.fast_rate = 450000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 4;
-		config.rev = ADRENO_REV(3, 2, 1, 0);
-	} else if (cpu_is_apq8064()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 4;
-
-		if (SOCINFO_VERSION_MAJOR(version) == 2)
-			config.rev = ADRENO_REV(3, 2, 0, 2);
-		else if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
-				(SOCINFO_VERSION_MINOR(version) == 1))
-			config.rev = ADRENO_REV(3, 2, 0, 1);
-		else
-			config.rev = ADRENO_REV(3, 2, 0, 0);
-
-	} else if (cpu_is_msm8960ab()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 320000000;
-		config.bus_freq  = 4;
-
-		if (SOCINFO_VERSION_MINOR(version) == 0)
-			config.rev = ADRENO_REV(3, 2, 1, 0);
-		else
-			config.rev = ADRENO_REV(3, 2, 1, 1);
-
-	} else if (cpu_is_msm8930()) {
-		config.fast_rate = 400000000;
-		config.slow_rate = 27000000;
-		config.bus_freq  = 3;
-
-		if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
-			(SOCINFO_VERSION_MINOR(version) == 2))
-			config.rev = ADRENO_REV(3, 0, 5, 2);
-		else
-			config.rev = ADRENO_REV(3, 0, 5, 0);
-
-	}
-#  ifdef CONFIG_MSM_BUS_SCALING
-	config.bus_scale_table = pdata->bus_scale_table;
-#  endif
-#endif
-	dev->platform_data = &config;
-	set_gpu_pdev(dev_get_drvdata(master), to_platform_device(dev));
-	return 0;
-}
-
-static void a3xx_unbind(struct device *dev, struct device *master,
-		void *data)
-{
-	set_gpu_pdev(dev_get_drvdata(master), NULL);
-}
-
-static const struct component_ops a3xx_ops = {
-		.bind   = a3xx_bind,
-		.unbind = a3xx_unbind,
-};
-
-static int a3xx_probe(struct platform_device *pdev)
-{
-	return component_add(&pdev->dev, &a3xx_ops);
-}
-
-static int a3xx_remove(struct platform_device *pdev)
-{
-	component_del(&pdev->dev, &a3xx_ops);
-	return 0;
-}
-
-static const struct of_device_id dt_match[] = {
-	{ .compatible = "qcom,kgsl-3d0" },
-	{}
-};
-
-static struct platform_driver a3xx_driver = {
-	.probe = a3xx_probe,
-	.remove = a3xx_remove,
-	.driver = {
-		.name = "kgsl-3d0",
-		.of_match_table = dt_match,
-	},
-};
-
-void __init a3xx_register(void)
-{
-	platform_driver_register(&a3xx_driver);
-}
-
-void __exit a3xx_unregister(void)
-{
-	platform_driver_unregister(&a3xx_driver);
-}
diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.h b/drivers/gpu/drm/msm/adreno/a3xx_gpu.h
index bb9a8ca0507b..85ff66cbddd6 100644
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.h
@@ -19,6 +19,11 @@
 #define __A3XX_GPU_H__
 
 #include "adreno_gpu.h"
+
+/* arrg, somehow fb.h is getting pulled in: */
+#undef ROP_COPY
+#undef ROP_XOR
+
 #include "a3xx.xml.h"
 
 struct a3xx_gpu {
diff --git a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
index d6e6ce2d1abd..cc341bc62b51 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_common.xml.h
@@ -10,13 +10,13 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32814 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   8900 bytes, from 2013-10-22 23:57:49)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  10574 bytes, from 2013-11-13 05:44:45)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  53644 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (   8344 bytes, from 2013-11-30 14:49:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14960 bytes, from 2014-07-27 17:22:13)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  41068 bytes, from 2014-08-01 12:22:48)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -87,15 +87,6 @@ enum adreno_rb_blend_factor {
 	FACTOR_SRC_ALPHA_SATURATE = 16,
 };
 
-enum adreno_rb_blend_opcode {
-	BLEND_DST_PLUS_SRC = 0,
-	BLEND_SRC_MINUS_DST = 1,
-	BLEND_MIN_DST_SRC = 2,
-	BLEND_MAX_DST_SRC = 3,
-	BLEND_DST_MINUS_SRC = 4,
-	BLEND_DST_PLUS_SRC_BIAS = 5,
-};
-
 enum adreno_rb_surface_endian {
 	ENDIAN_NONE = 0,
 	ENDIAN_8IN16 = 1,
@@ -116,6 +107,39 @@ enum adreno_rb_depth_format {
 	DEPTHX_24_8 = 1,
 };
 
+enum adreno_rb_copy_control_mode {
+	RB_COPY_RESOLVE = 1,
+	RB_COPY_CLEAR = 2,
+	RB_COPY_DEPTH_STENCIL = 5,
+};
+
+enum a3xx_render_mode {
+	RB_RENDERING_PASS = 0,
+	RB_TILING_PASS = 1,
+	RB_RESOLVE_PASS = 2,
+	RB_COMPUTE_PASS = 3,
+};
+
+enum a3xx_msaa_samples {
+	MSAA_ONE = 0,
+	MSAA_TWO = 1,
+	MSAA_FOUR = 2,
+};
+
+enum a3xx_threadmode {
+	MULTI = 0,
+	SINGLE = 1,
+};
+
+enum a3xx_instrbuffermode {
+	BUFFER = 1,
+};
+
+enum a3xx_threadsize {
+	TWO_QUADS = 0,
+	FOUR_QUADS = 1,
+};
+
 #define REG_AXXX_CP_RB_BASE					0x000001c0
 
 #define REG_AXXX_CP_RB_CNTL					0x000001c1
@@ -264,6 +288,8 @@ static inline uint32_t AXXX_SCRATCH_UMSK_SWAP(uint32_t val)
 #define REG_AXXX_CP_INT_ACK					0x000001f4
 
 #define REG_AXXX_CP_ME_CNTL					0x000001f6
+#define AXXX_CP_ME_CNTL_BUSY					0x20000000
+#define AXXX_CP_ME_CNTL_HALT					0x10000000
 
 #define REG_AXXX_CP_ME_STATUS					0x000001f7
 
diff --git a/drivers/gpu/drm/msm/adreno/adreno_device.c b/drivers/gpu/drm/msm/adreno/adreno_device.c
new file mode 100644
index 000000000000..7ab85af3a7db
--- /dev/null
+++ b/drivers/gpu/drm/msm/adreno/adreno_device.c
@@ -0,0 +1,285 @@
+/*
+ * Copyright (C) 2013-2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "adreno_gpu.h"
+
+#if defined(CONFIG_MSM_BUS_SCALING) && !defined(CONFIG_OF)
+#  include <mach/kgsl.h>
+#endif
+
+#define ANY_ID 0xff
+
+bool hang_debug = false;
+MODULE_PARM_DESC(hang_debug, "Dump registers when hang is detected (can be slow!)");
+module_param_named(hang_debug, hang_debug, bool, 0600);
+
+struct msm_gpu *a3xx_gpu_init(struct drm_device *dev);
+
+static const struct adreno_info gpulist[] = {
+	{
+		.rev   = ADRENO_REV(3, 0, 5, ANY_ID),
+		.revn  = 305,
+		.name  = "A305",
+		.pm4fw = "a300_pm4.fw",
+		.pfpfw = "a300_pfp.fw",
+		.gmem  = SZ_256K,
+		.init  = a3xx_gpu_init,
+	}, {
+		.rev   = ADRENO_REV(3, 2, ANY_ID, ANY_ID),
+		.revn  = 320,
+		.name  = "A320",
+		.pm4fw = "a300_pm4.fw",
+		.pfpfw = "a300_pfp.fw",
+		.gmem  = SZ_512K,
+		.init  = a3xx_gpu_init,
+	}, {
+		.rev   = ADRENO_REV(3, 3, 0, ANY_ID),
+		.revn  = 330,
+		.name  = "A330",
+		.pm4fw = "a330_pm4.fw",
+		.pfpfw = "a330_pfp.fw",
+		.gmem  = SZ_1M,
+		.init  = a3xx_gpu_init,
+	},
+};
+
+MODULE_FIRMWARE("a300_pm4.fw");
+MODULE_FIRMWARE("a300_pfp.fw");
+MODULE_FIRMWARE("a330_pm4.fw");
+MODULE_FIRMWARE("a330_pfp.fw");
+
+static inline bool _rev_match(uint8_t entry, uint8_t id)
+{
+	return (entry == ANY_ID) || (entry == id);
+}
+
+const struct adreno_info *adreno_info(struct adreno_rev rev)
+{
+	int i;
+
+	/* identify gpu: */
+	for (i = 0; i < ARRAY_SIZE(gpulist); i++) {
+		const struct adreno_info *info = &gpulist[i];
+		if (_rev_match(info->rev.core, rev.core) &&
+				_rev_match(info->rev.major, rev.major) &&
+				_rev_match(info->rev.minor, rev.minor) &&
+				_rev_match(info->rev.patchid, rev.patchid))
+			return info;
+	}
+
+	return NULL;
+}
+
+struct msm_gpu *adreno_load_gpu(struct drm_device *dev)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	struct platform_device *pdev = priv->gpu_pdev;
+	struct adreno_platform_config *config;
+	struct adreno_rev rev;
+	const struct adreno_info *info;
+	struct msm_gpu *gpu = NULL;
+
+	if (!pdev) {
+		dev_err(dev->dev, "no adreno device\n");
+		return NULL;
+	}
+
+	config = pdev->dev.platform_data;
+	rev = config->rev;
+	info = adreno_info(config->rev);
+
+	if (!info) {
+		dev_warn(dev->dev, "Unknown GPU revision: %u.%u.%u.%u\n",
+				rev.core, rev.major, rev.minor, rev.patchid);
+		return NULL;
+	}
+
+	DBG("Found GPU: %u.%u.%u.%u",  rev.core, rev.major,
+			rev.minor, rev.patchid);
+
+	gpu = info->init(dev);
+	if (IS_ERR(gpu)) {
+		dev_warn(dev->dev, "failed to load adreno gpu\n");
+		gpu = NULL;
+		/* not fatal */
+	}
+
+	if (gpu) {
+		int ret;
+		mutex_lock(&dev->struct_mutex);
+		gpu->funcs->pm_resume(gpu);
+		mutex_unlock(&dev->struct_mutex);
+		ret = gpu->funcs->hw_init(gpu);
+		if (ret) {
+			dev_err(dev->dev, "gpu hw init failed: %d\n", ret);
+			gpu->funcs->destroy(gpu);
+			gpu = NULL;
+		} else {
+			/* give inactive pm a chance to kick in: */
+			msm_gpu_retire(gpu);
+		}
+	}
+
+	return gpu;
+}
+
+static void set_gpu_pdev(struct drm_device *dev,
+		struct platform_device *pdev)
+{
+	struct msm_drm_private *priv = dev->dev_private;
+	priv->gpu_pdev = pdev;
+}
+
+static int adreno_bind(struct device *dev, struct device *master, void *data)
+{
+	static struct adreno_platform_config config = {};
+#ifdef CONFIG_OF
+	struct device_node *child, *node = dev->of_node;
+	u32 val;
+	int ret;
+
+	ret = of_property_read_u32(node, "qcom,chipid", &val);
+	if (ret) {
+		dev_err(dev, "could not find chipid: %d\n", ret);
+		return ret;
+	}
+
+	config.rev = ADRENO_REV((val >> 24) & 0xff,
+			(val >> 16) & 0xff, (val >> 8) & 0xff, val & 0xff);
+
+	/* find clock rates: */
+	config.fast_rate = 0;
+	config.slow_rate = ~0;
+	for_each_child_of_node(node, child) {
+		if (of_device_is_compatible(child, "qcom,gpu-pwrlevels")) {
+			struct device_node *pwrlvl;
+			for_each_child_of_node(child, pwrlvl) {
+				ret = of_property_read_u32(pwrlvl, "qcom,gpu-freq", &val);
+				if (ret) {
+					dev_err(dev, "could not find gpu-freq: %d\n", ret);
+					return ret;
+				}
+				config.fast_rate = max(config.fast_rate, val);
+				config.slow_rate = min(config.slow_rate, val);
+			}
+		}
+	}
+
+	if (!config.fast_rate) {
+		dev_err(dev, "could not find clk rates\n");
+		return -ENXIO;
+	}
+
+#else
+	struct kgsl_device_platform_data *pdata = dev->platform_data;
+	uint32_t version = socinfo_get_version();
+	if (cpu_is_apq8064ab()) {
+		config.fast_rate = 450000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 4;
+		config.rev = ADRENO_REV(3, 2, 1, 0);
+	} else if (cpu_is_apq8064()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 4;
+
+		if (SOCINFO_VERSION_MAJOR(version) == 2)
+			config.rev = ADRENO_REV(3, 2, 0, 2);
+		else if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
+				(SOCINFO_VERSION_MINOR(version) == 1))
+			config.rev = ADRENO_REV(3, 2, 0, 1);
+		else
+			config.rev = ADRENO_REV(3, 2, 0, 0);
+
+	} else if (cpu_is_msm8960ab()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 320000000;
+		config.bus_freq  = 4;
+
+		if (SOCINFO_VERSION_MINOR(version) == 0)
+			config.rev = ADRENO_REV(3, 2, 1, 0);
+		else
+			config.rev = ADRENO_REV(3, 2, 1, 1);
+
+	} else if (cpu_is_msm8930()) {
+		config.fast_rate = 400000000;
+		config.slow_rate = 27000000;
+		config.bus_freq  = 3;
+
+		if ((SOCINFO_VERSION_MAJOR(version) == 1) &&
+			(SOCINFO_VERSION_MINOR(version) == 2))
+			config.rev = ADRENO_REV(3, 0, 5, 2);
+		else
+			config.rev = ADRENO_REV(3, 0, 5, 0);
+
+	}
+#  ifdef CONFIG_MSM_BUS_SCALING
+	config.bus_scale_table = pdata->bus_scale_table;
+#  endif
+#endif
+	dev->platform_data = &config;
+	set_gpu_pdev(dev_get_drvdata(master), to_platform_device(dev));
+	return 0;
+}
+
+static void adreno_unbind(struct device *dev, struct device *master,
+		void *data)
+{
+	set_gpu_pdev(dev_get_drvdata(master), NULL);
+}
+
+static const struct component_ops a3xx_ops = {
+		.bind   = adreno_bind,
+		.unbind = adreno_unbind,
+};
+
+static int adreno_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &a3xx_ops);
+}
+
+static int adreno_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &a3xx_ops);
+	return 0;
+}
+
+static const struct of_device_id dt_match[] = {
+	{ .compatible = "qcom,adreno-3xx" },
+	/* for backwards compat w/ downstream kgsl DT files: */
+	{ .compatible = "qcom,kgsl-3d0" },
+	{}
+};
+
+static struct platform_driver adreno_driver = {
+	.probe = adreno_probe,
+	.remove = adreno_remove,
+	.driver = {
+		.name = "adreno",
+		.of_match_table = dt_match,
+	},
+};
+
+void __init adreno_register(void)
+{
+	platform_driver_register(&adreno_driver);
+}
+
+void __exit adreno_unregister(void)
+{
+	platform_driver_unregister(&adreno_driver);
+}
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
index 28ca8cd8b09e..6afa29167fee 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
@@ -19,46 +19,6 @@
 #include "msm_gem.h"
 #include "msm_mmu.h"
 
-struct adreno_info {
-	struct adreno_rev rev;
-	uint32_t revn;
-	const char *name;
-	const char *pm4fw, *pfpfw;
-	uint32_t gmem;
-};
-
-#define ANY_ID 0xff
-
-static const struct adreno_info gpulist[] = {
-	{
-		.rev   = ADRENO_REV(3, 0, 5, ANY_ID),
-		.revn  = 305,
-		.name  = "A305",
-		.pm4fw = "a300_pm4.fw",
-		.pfpfw = "a300_pfp.fw",
-		.gmem  = SZ_256K,
-	}, {
-		.rev   = ADRENO_REV(3, 2, ANY_ID, ANY_ID),
-		.revn  = 320,
-		.name  = "A320",
-		.pm4fw = "a300_pm4.fw",
-		.pfpfw = "a300_pfp.fw",
-		.gmem  = SZ_512K,
-	}, {
-		.rev   = ADRENO_REV(3, 3, 0, ANY_ID),
-		.revn  = 330,
-		.name  = "A330",
-		.pm4fw = "a330_pm4.fw",
-		.pfpfw = "a330_pfp.fw",
-		.gmem  = SZ_1M,
-	},
-};
-
-MODULE_FIRMWARE("a300_pm4.fw");
-MODULE_FIRMWARE("a300_pfp.fw");
-MODULE_FIRMWARE("a330_pm4.fw");
-MODULE_FIRMWARE("a330_pfp.fw");
-
 #define RB_SIZE    SZ_32K
 #define RB_BLKSIZE 16
 
@@ -91,9 +51,17 @@ int adreno_get_param(struct msm_gpu *gpu, uint32_t param, uint64_t *value)
 int adreno_hw_init(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	int ret;
 
 	DBG("%s", gpu->name);
 
+	ret = msm_gem_get_iova(gpu->rb->bo, gpu->id, &gpu->rb_iova);
+	if (ret) {
+		gpu->rb_iova = 0;
+		dev_err(gpu->dev->dev, "could not map ringbuffer: %d\n", ret);
+		return ret;
+	}
+
 	/* Setup REG_CP_RB_CNTL: */
 	gpu_write(gpu, REG_AXXX_CP_RB_CNTL,
 			/* size is log2(quad-words): */
@@ -244,6 +212,7 @@ void adreno_idle(struct msm_gpu *gpu)
 void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	int i;
 
 	seq_printf(m, "revision: %d (%d.%d.%d.%d)\n",
 			adreno_gpu->info->revn, adreno_gpu->rev.core,
@@ -255,6 +224,23 @@ void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 	seq_printf(m, "rptr:     %d\n", adreno_gpu->memptrs->rptr);
 	seq_printf(m, "wptr:     %d\n", adreno_gpu->memptrs->wptr);
 	seq_printf(m, "rb wptr:  %d\n", get_wptr(gpu->rb));
+
+	gpu->funcs->pm_resume(gpu);
+
+	/* dump these out in a form that can be parsed by demsm: */
+	seq_printf(m, "IO:region %s 00000000 00020000\n", gpu->name);
+	for (i = 0; adreno_gpu->registers[i] != ~0; i += 2) {
+		uint32_t start = adreno_gpu->registers[i];
+		uint32_t end   = adreno_gpu->registers[i+1];
+		uint32_t addr;
+
+		for (addr = start; addr <= end; addr++) {
+			uint32_t val = gpu_read(gpu, addr);
+			seq_printf(m, "IO:R %08x %08x\n", addr<<2, val);
+		}
+	}
+
+	gpu->funcs->pm_suspend(gpu);
 }
 #endif
 
@@ -262,6 +248,7 @@ void adreno_show(struct msm_gpu *gpu, struct seq_file *m)
 void adreno_dump(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
+	int i;
 
 	printk("revision: %d (%d.%d.%d.%d)\n",
 			adreno_gpu->info->revn, adreno_gpu->rev.core,
@@ -274,6 +261,18 @@ void adreno_dump(struct msm_gpu *gpu)
 	printk("wptr:     %d\n", adreno_gpu->memptrs->wptr);
 	printk("rb wptr:  %d\n", get_wptr(gpu->rb));
 
+	/* dump these out in a form that can be parsed by demsm: */
+	printk("IO:region %s 00000000 00020000\n", gpu->name);
+	for (i = 0; adreno_gpu->registers[i] != ~0; i += 2) {
+		uint32_t start = adreno_gpu->registers[i];
+		uint32_t end   = adreno_gpu->registers[i+1];
+		uint32_t addr;
+
+		for (addr = start; addr <= end; addr++) {
+			uint32_t val = gpu_read(gpu, addr);
+			printk("IO:R %08x %08x\n", addr<<2, val);
+		}
+	}
 }
 
 static uint32_t ring_freewords(struct msm_gpu *gpu)
@@ -296,65 +295,51 @@ static const char *iommu_ports[] = {
 		"gfx3d1_user", "gfx3d1_priv",
 };
 
-static inline bool _rev_match(uint8_t entry, uint8_t id)
-{
-	return (entry == ANY_ID) || (entry == id);
-}
-
 int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
-		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs,
-		struct adreno_rev rev)
+		struct adreno_gpu *adreno_gpu, const struct adreno_gpu_funcs *funcs)
 {
+	struct adreno_platform_config *config = pdev->dev.platform_data;
+	struct msm_gpu *gpu = &adreno_gpu->base;
 	struct msm_mmu *mmu;
-	int i, ret;
-
-	/* identify gpu: */
-	for (i = 0; i < ARRAY_SIZE(gpulist); i++) {
-		const struct adreno_info *info = &gpulist[i];
-		if (_rev_match(info->rev.core, rev.core) &&
-				_rev_match(info->rev.major, rev.major) &&
-				_rev_match(info->rev.minor, rev.minor) &&
-				_rev_match(info->rev.patchid, rev.patchid)) {
-			gpu->info = info;
-			gpu->revn = info->revn;
-			break;
-		}
-	}
-
-	if (i == ARRAY_SIZE(gpulist)) {
-		dev_err(drm->dev, "Unknown GPU revision: %u.%u.%u.%u\n",
-				rev.core, rev.major, rev.minor, rev.patchid);
-		return -ENXIO;
-	}
+	int ret;
 
-	DBG("Found GPU: %s (%u.%u.%u.%u)", gpu->info->name,
-			rev.core, rev.major, rev.minor, rev.patchid);
+	adreno_gpu->funcs = funcs;
+	adreno_gpu->info = adreno_info(config->rev);
+	adreno_gpu->gmem = adreno_gpu->info->gmem;
+	adreno_gpu->revn = adreno_gpu->info->revn;
+	adreno_gpu->rev = config->rev;
+
+	gpu->fast_rate = config->fast_rate;
+	gpu->slow_rate = config->slow_rate;
+	gpu->bus_freq  = config->bus_freq;
+#ifdef CONFIG_MSM_BUS_SCALING
+	gpu->bus_scale_table = config->bus_scale_table;
+#endif
 
-	gpu->funcs = funcs;
-	gpu->gmem = gpu->info->gmem;
-	gpu->rev = rev;
+	DBG("fast_rate=%u, slow_rate=%u, bus_freq=%u",
+			gpu->fast_rate, gpu->slow_rate, gpu->bus_freq);
 
-	ret = request_firmware(&gpu->pm4, gpu->info->pm4fw, drm->dev);
+	ret = request_firmware(&adreno_gpu->pm4, adreno_gpu->info->pm4fw, drm->dev);
 	if (ret) {
 		dev_err(drm->dev, "failed to load %s PM4 firmware: %d\n",
-				gpu->info->pm4fw, ret);
+				adreno_gpu->info->pm4fw, ret);
 		return ret;
 	}
 
-	ret = request_firmware(&gpu->pfp, gpu->info->pfpfw, drm->dev);
+	ret = request_firmware(&adreno_gpu->pfp, adreno_gpu->info->pfpfw, drm->dev);
 	if (ret) {
 		dev_err(drm->dev, "failed to load %s PFP firmware: %d\n",
-				gpu->info->pfpfw, ret);
+				adreno_gpu->info->pfpfw, ret);
 		return ret;
 	}
 
-	ret = msm_gpu_init(drm, pdev, &gpu->base, &funcs->base,
-			gpu->info->name, "kgsl_3d0_reg_memory", "kgsl_3d0_irq",
+	ret = msm_gpu_init(drm, pdev, &adreno_gpu->base, &funcs->base,
+			adreno_gpu->info->name, "kgsl_3d0_reg_memory", "kgsl_3d0_irq",
 			RB_SIZE);
 	if (ret)
 		return ret;
 
-	mmu = gpu->base.mmu;
+	mmu = gpu->mmu;
 	if (mmu) {
 		ret = mmu->funcs->attach(mmu, iommu_ports,
 				ARRAY_SIZE(iommu_ports));
@@ -362,23 +347,25 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 			return ret;
 	}
 
-	gpu->memptrs_bo = msm_gem_new(drm, sizeof(*gpu->memptrs),
+	mutex_lock(&drm->struct_mutex);
+	adreno_gpu->memptrs_bo = msm_gem_new(drm, sizeof(*adreno_gpu->memptrs),
 			MSM_BO_UNCACHED);
-	if (IS_ERR(gpu->memptrs_bo)) {
-		ret = PTR_ERR(gpu->memptrs_bo);
-		gpu->memptrs_bo = NULL;
+	mutex_unlock(&drm->struct_mutex);
+	if (IS_ERR(adreno_gpu->memptrs_bo)) {
+		ret = PTR_ERR(adreno_gpu->memptrs_bo);
+		adreno_gpu->memptrs_bo = NULL;
 		dev_err(drm->dev, "could not allocate memptrs: %d\n", ret);
 		return ret;
 	}
 
-	gpu->memptrs = msm_gem_vaddr_locked(gpu->memptrs_bo);
-	if (!gpu->memptrs) {
+	adreno_gpu->memptrs = msm_gem_vaddr(adreno_gpu->memptrs_bo);
+	if (!adreno_gpu->memptrs) {
 		dev_err(drm->dev, "could not vmap memptrs\n");
 		return -ENOMEM;
 	}
 
-	ret = msm_gem_get_iova_locked(gpu->memptrs_bo, gpu->base.id,
-			&gpu->memptrs_iova);
+	ret = msm_gem_get_iova(adreno_gpu->memptrs_bo, gpu->id,
+			&adreno_gpu->memptrs_iova);
 	if (ret) {
 		dev_err(drm->dev, "could not map memptrs: %d\n", ret);
 		return ret;
diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
index 63c36ce33020..52f051579753 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
@@ -39,7 +39,16 @@ struct adreno_gpu_funcs {
 	struct msm_gpu_funcs base;
 };
 
-struct adreno_info;
+struct adreno_info {
+	struct adreno_rev rev;
+	uint32_t revn;
+	const char *name;
+	const char *pm4fw, *pfpfw;
+	uint32_t gmem;
+	struct msm_gpu *(*init)(struct drm_device *dev);
+};
+
+const struct adreno_info *adreno_info(struct adreno_rev rev);
 
 struct adreno_rbmemptrs {
 	volatile uint32_t rptr;
@@ -55,6 +64,9 @@ struct adreno_gpu {
 	uint32_t revn;  /* numeric revision name */
 	const struct adreno_gpu_funcs *funcs;
 
+	/* interesting register offsets to dump: */
+	const unsigned int *registers;
+
 	/* firmware: */
 	const struct firmware *pm4, *pfp;
 
@@ -131,8 +143,7 @@ void adreno_dump(struct msm_gpu *gpu);
 void adreno_wait_ring(struct msm_gpu *gpu, uint32_t ndwords);
 
 int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev,
-		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs,
-		struct adreno_rev rev);
+		struct adreno_gpu *gpu, const struct adreno_gpu_funcs *funcs);
 void adreno_gpu_cleanup(struct adreno_gpu *gpu);
 
 
diff --git a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
index ae992c71703f..6ef43f66c30a 100644
--- a/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h
@@ -10,13 +10,13 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/adreno.xml               (    364 bytes, from 2013-11-30 14:47:15)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml  (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32814 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   8900 bytes, from 2013-10-22 23:57:49)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  10574 bytes, from 2013-11-13 05:44:45)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  53644 bytes, from 2013-11-30 15:07:33)
-- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (   8344 bytes, from 2013-11-30 14:49:47)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a2xx.xml          (  32901 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_common.xml (   9859 bytes, from 2014-06-02 15:21:30)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/adreno_pm4.xml    (  14960 bytes, from 2014-07-27 17:22:13)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a3xx.xml          (  58020 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/adreno/a4xx.xml          (  41068 bytes, from 2014-08-01 12:22:48)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -105,6 +105,7 @@ enum pc_di_index_size {
 
 enum pc_di_vis_cull_mode {
 	IGNORE_VISIBILITY = 0,
+	USE_VISIBILITY = 1,
 };
 
 enum adreno_pm4_packet_type {
@@ -162,7 +163,16 @@ enum adreno_pm4_type3_packets {
 	CP_INDIRECT_BUFFER_PFE = 63,
 	CP_SET_BIN = 76,
 	CP_TEST_TWO_MEMS = 113,
+	CP_REG_WR_NO_CTXT = 120,
+	CP_RECORD_PFP_TIMESTAMP = 17,
 	CP_WAIT_FOR_ME = 19,
+	CP_SET_DRAW_STATE = 67,
+	CP_DRAW_INDX_OFFSET = 56,
+	CP_DRAW_INDIRECT = 40,
+	CP_DRAW_INDX_INDIRECT = 41,
+	CP_DRAW_AUTO = 36,
+	CP_UNKNOWN_1A = 26,
+	CP_WIDE_REG_WRITE = 116,
 	IN_IB_PREFETCH_END = 23,
 	IN_SUBBLK_PREFETCH = 31,
 	IN_INSTR_PREFETCH = 32,
@@ -232,6 +242,211 @@ static inline uint32_t CP_LOAD_STATE_1_EXT_SRC_ADDR(uint32_t val)
 	return ((val >> 2) << CP_LOAD_STATE_1_EXT_SRC_ADDR__SHIFT) & CP_LOAD_STATE_1_EXT_SRC_ADDR__MASK;
 }
 
+#define REG_CP_DRAW_INDX_0					0x00000000
+#define CP_DRAW_INDX_0_VIZ_QUERY__MASK				0xffffffff
+#define CP_DRAW_INDX_0_VIZ_QUERY__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_0_VIZ_QUERY(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_0_VIZ_QUERY__SHIFT) & CP_DRAW_INDX_0_VIZ_QUERY__MASK;
+}
+
+#define REG_CP_DRAW_INDX_1					0x00000001
+#define CP_DRAW_INDX_1_PRIM_TYPE__MASK				0x0000003f
+#define CP_DRAW_INDX_1_PRIM_TYPE__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_1_PRIM_TYPE(enum pc_di_primtype val)
+{
+	return ((val) << CP_DRAW_INDX_1_PRIM_TYPE__SHIFT) & CP_DRAW_INDX_1_PRIM_TYPE__MASK;
+}
+#define CP_DRAW_INDX_1_SOURCE_SELECT__MASK			0x000000c0
+#define CP_DRAW_INDX_1_SOURCE_SELECT__SHIFT			6
+static inline uint32_t CP_DRAW_INDX_1_SOURCE_SELECT(enum pc_di_src_sel val)
+{
+	return ((val) << CP_DRAW_INDX_1_SOURCE_SELECT__SHIFT) & CP_DRAW_INDX_1_SOURCE_SELECT__MASK;
+}
+#define CP_DRAW_INDX_1_VIS_CULL__MASK				0x00000600
+#define CP_DRAW_INDX_1_VIS_CULL__SHIFT				9
+static inline uint32_t CP_DRAW_INDX_1_VIS_CULL(enum pc_di_vis_cull_mode val)
+{
+	return ((val) << CP_DRAW_INDX_1_VIS_CULL__SHIFT) & CP_DRAW_INDX_1_VIS_CULL__MASK;
+}
+#define CP_DRAW_INDX_1_INDEX_SIZE__MASK				0x00000800
+#define CP_DRAW_INDX_1_INDEX_SIZE__SHIFT			11
+static inline uint32_t CP_DRAW_INDX_1_INDEX_SIZE(enum pc_di_index_size val)
+{
+	return ((val) << CP_DRAW_INDX_1_INDEX_SIZE__SHIFT) & CP_DRAW_INDX_1_INDEX_SIZE__MASK;
+}
+#define CP_DRAW_INDX_1_NOT_EOP					0x00001000
+#define CP_DRAW_INDX_1_SMALL_INDEX				0x00002000
+#define CP_DRAW_INDX_1_PRE_DRAW_INITIATOR_ENABLE		0x00004000
+#define CP_DRAW_INDX_1_NUM_INDICES__MASK			0xffff0000
+#define CP_DRAW_INDX_1_NUM_INDICES__SHIFT			16
+static inline uint32_t CP_DRAW_INDX_1_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_1_NUM_INDICES__SHIFT) & CP_DRAW_INDX_1_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2					0x00000002
+#define CP_DRAW_INDX_2_NUM_INDICES__MASK			0xffffffff
+#define CP_DRAW_INDX_2_NUM_INDICES__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_2_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_NUM_INDICES__SHIFT) & CP_DRAW_INDX_2_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2					0x00000002
+#define CP_DRAW_INDX_2_INDX_BASE__MASK				0xffffffff
+#define CP_DRAW_INDX_2_INDX_BASE__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_2_INDX_BASE(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_INDX_BASE__SHIFT) & CP_DRAW_INDX_2_INDX_BASE__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2					0x00000002
+#define CP_DRAW_INDX_2_INDX_SIZE__MASK				0xffffffff
+#define CP_DRAW_INDX_2_INDX_SIZE__SHIFT				0
+static inline uint32_t CP_DRAW_INDX_2_INDX_SIZE(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_INDX_SIZE__SHIFT) & CP_DRAW_INDX_2_INDX_SIZE__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2_0					0x00000000
+#define CP_DRAW_INDX_2_0_VIZ_QUERY__MASK			0xffffffff
+#define CP_DRAW_INDX_2_0_VIZ_QUERY__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_2_0_VIZ_QUERY(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_0_VIZ_QUERY__SHIFT) & CP_DRAW_INDX_2_0_VIZ_QUERY__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2_1					0x00000001
+#define CP_DRAW_INDX_2_1_PRIM_TYPE__MASK			0x0000003f
+#define CP_DRAW_INDX_2_1_PRIM_TYPE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_2_1_PRIM_TYPE(enum pc_di_primtype val)
+{
+	return ((val) << CP_DRAW_INDX_2_1_PRIM_TYPE__SHIFT) & CP_DRAW_INDX_2_1_PRIM_TYPE__MASK;
+}
+#define CP_DRAW_INDX_2_1_SOURCE_SELECT__MASK			0x000000c0
+#define CP_DRAW_INDX_2_1_SOURCE_SELECT__SHIFT			6
+static inline uint32_t CP_DRAW_INDX_2_1_SOURCE_SELECT(enum pc_di_src_sel val)
+{
+	return ((val) << CP_DRAW_INDX_2_1_SOURCE_SELECT__SHIFT) & CP_DRAW_INDX_2_1_SOURCE_SELECT__MASK;
+}
+#define CP_DRAW_INDX_2_1_VIS_CULL__MASK				0x00000600
+#define CP_DRAW_INDX_2_1_VIS_CULL__SHIFT			9
+static inline uint32_t CP_DRAW_INDX_2_1_VIS_CULL(enum pc_di_vis_cull_mode val)
+{
+	return ((val) << CP_DRAW_INDX_2_1_VIS_CULL__SHIFT) & CP_DRAW_INDX_2_1_VIS_CULL__MASK;
+}
+#define CP_DRAW_INDX_2_1_INDEX_SIZE__MASK			0x00000800
+#define CP_DRAW_INDX_2_1_INDEX_SIZE__SHIFT			11
+static inline uint32_t CP_DRAW_INDX_2_1_INDEX_SIZE(enum pc_di_index_size val)
+{
+	return ((val) << CP_DRAW_INDX_2_1_INDEX_SIZE__SHIFT) & CP_DRAW_INDX_2_1_INDEX_SIZE__MASK;
+}
+#define CP_DRAW_INDX_2_1_NOT_EOP				0x00001000
+#define CP_DRAW_INDX_2_1_SMALL_INDEX				0x00002000
+#define CP_DRAW_INDX_2_1_PRE_DRAW_INITIATOR_ENABLE		0x00004000
+#define CP_DRAW_INDX_2_1_NUM_INDICES__MASK			0xffff0000
+#define CP_DRAW_INDX_2_1_NUM_INDICES__SHIFT			16
+static inline uint32_t CP_DRAW_INDX_2_1_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_1_NUM_INDICES__SHIFT) & CP_DRAW_INDX_2_1_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_2_2					0x00000002
+#define CP_DRAW_INDX_2_2_NUM_INDICES__MASK			0xffffffff
+#define CP_DRAW_INDX_2_2_NUM_INDICES__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_2_2_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_2_2_NUM_INDICES__SHIFT) & CP_DRAW_INDX_2_2_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_OFFSET_0				0x00000000
+#define CP_DRAW_INDX_OFFSET_0_PRIM_TYPE__MASK			0x0000003f
+#define CP_DRAW_INDX_OFFSET_0_PRIM_TYPE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_PRIM_TYPE(enum pc_di_primtype val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_0_PRIM_TYPE__SHIFT) & CP_DRAW_INDX_OFFSET_0_PRIM_TYPE__MASK;
+}
+#define CP_DRAW_INDX_OFFSET_0_SOURCE_SELECT__MASK		0x000000c0
+#define CP_DRAW_INDX_OFFSET_0_SOURCE_SELECT__SHIFT		6
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_SOURCE_SELECT(enum pc_di_src_sel val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_0_SOURCE_SELECT__SHIFT) & CP_DRAW_INDX_OFFSET_0_SOURCE_SELECT__MASK;
+}
+#define CP_DRAW_INDX_OFFSET_0_VIS_CULL__MASK			0x00000700
+#define CP_DRAW_INDX_OFFSET_0_VIS_CULL__SHIFT			8
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_VIS_CULL(enum pc_di_vis_cull_mode val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_0_VIS_CULL__SHIFT) & CP_DRAW_INDX_OFFSET_0_VIS_CULL__MASK;
+}
+#define CP_DRAW_INDX_OFFSET_0_INDEX_SIZE__MASK			0x00000800
+#define CP_DRAW_INDX_OFFSET_0_INDEX_SIZE__SHIFT			11
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_INDEX_SIZE(enum pc_di_index_size val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_0_INDEX_SIZE__SHIFT) & CP_DRAW_INDX_OFFSET_0_INDEX_SIZE__MASK;
+}
+#define CP_DRAW_INDX_OFFSET_0_NOT_EOP				0x00001000
+#define CP_DRAW_INDX_OFFSET_0_SMALL_INDEX			0x00002000
+#define CP_DRAW_INDX_OFFSET_0_PRE_DRAW_INITIATOR_ENABLE		0x00004000
+#define CP_DRAW_INDX_OFFSET_0_NUM_INDICES__MASK			0xffff0000
+#define CP_DRAW_INDX_OFFSET_0_NUM_INDICES__SHIFT		16
+static inline uint32_t CP_DRAW_INDX_OFFSET_0_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_0_NUM_INDICES__SHIFT) & CP_DRAW_INDX_OFFSET_0_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_OFFSET_1				0x00000001
+
+#define REG_CP_DRAW_INDX_OFFSET_2				0x00000002
+#define CP_DRAW_INDX_OFFSET_2_NUM_INDICES__MASK			0xffffffff
+#define CP_DRAW_INDX_OFFSET_2_NUM_INDICES__SHIFT		0
+static inline uint32_t CP_DRAW_INDX_OFFSET_2_NUM_INDICES(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_2_NUM_INDICES__SHIFT) & CP_DRAW_INDX_OFFSET_2_NUM_INDICES__MASK;
+}
+
+#define REG_CP_DRAW_INDX_OFFSET_2				0x00000002
+#define CP_DRAW_INDX_OFFSET_2_INDX_BASE__MASK			0xffffffff
+#define CP_DRAW_INDX_OFFSET_2_INDX_BASE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_OFFSET_2_INDX_BASE(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_2_INDX_BASE__SHIFT) & CP_DRAW_INDX_OFFSET_2_INDX_BASE__MASK;
+}
+
+#define REG_CP_DRAW_INDX_OFFSET_2				0x00000002
+#define CP_DRAW_INDX_OFFSET_2_INDX_SIZE__MASK			0xffffffff
+#define CP_DRAW_INDX_OFFSET_2_INDX_SIZE__SHIFT			0
+static inline uint32_t CP_DRAW_INDX_OFFSET_2_INDX_SIZE(uint32_t val)
+{
+	return ((val) << CP_DRAW_INDX_OFFSET_2_INDX_SIZE__SHIFT) & CP_DRAW_INDX_OFFSET_2_INDX_SIZE__MASK;
+}
+
+#define REG_CP_SET_DRAW_STATE_0					0x00000000
+#define CP_SET_DRAW_STATE_0_COUNT__MASK				0x0000ffff
+#define CP_SET_DRAW_STATE_0_COUNT__SHIFT			0
+static inline uint32_t CP_SET_DRAW_STATE_0_COUNT(uint32_t val)
+{
+	return ((val) << CP_SET_DRAW_STATE_0_COUNT__SHIFT) & CP_SET_DRAW_STATE_0_COUNT__MASK;
+}
+#define CP_SET_DRAW_STATE_0_DIRTY				0x00010000
+#define CP_SET_DRAW_STATE_0_DISABLE				0x00020000
+#define CP_SET_DRAW_STATE_0_DISABLE_ALL_GROUPS			0x00040000
+#define CP_SET_DRAW_STATE_0_LOAD_IMMED				0x00080000
+#define CP_SET_DRAW_STATE_0_GROUP_ID__MASK			0x1f000000
+#define CP_SET_DRAW_STATE_0_GROUP_ID__SHIFT			24
+static inline uint32_t CP_SET_DRAW_STATE_0_GROUP_ID(uint32_t val)
+{
+	return ((val) << CP_SET_DRAW_STATE_0_GROUP_ID__SHIFT) & CP_SET_DRAW_STATE_0_GROUP_ID__MASK;
+}
+
+#define REG_CP_SET_DRAW_STATE_1					0x00000001
+#define CP_SET_DRAW_STATE_1_ADDR__MASK				0xffffffff
+#define CP_SET_DRAW_STATE_1_ADDR__SHIFT				0
+static inline uint32_t CP_SET_DRAW_STATE_1_ADDR(uint32_t val)
+{
+	return ((val) << CP_SET_DRAW_STATE_1_ADDR__SHIFT) & CP_SET_DRAW_STATE_1_ADDR__MASK;
+}
+
 #define REG_CP_SET_BIN_0					0x00000000
 
 #define REG_CP_SET_BIN_1					0x00000001
@@ -262,5 +477,21 @@ static inline uint32_t CP_SET_BIN_2_Y2(uint32_t val)
 	return ((val) << CP_SET_BIN_2_Y2__SHIFT) & CP_SET_BIN_2_Y2__MASK;
 }
 
+#define REG_CP_SET_BIN_DATA_0					0x00000000
+#define CP_SET_BIN_DATA_0_BIN_DATA_ADDR__MASK			0xffffffff
+#define CP_SET_BIN_DATA_0_BIN_DATA_ADDR__SHIFT			0
+static inline uint32_t CP_SET_BIN_DATA_0_BIN_DATA_ADDR(uint32_t val)
+{
+	return ((val) << CP_SET_BIN_DATA_0_BIN_DATA_ADDR__SHIFT) & CP_SET_BIN_DATA_0_BIN_DATA_ADDR__MASK;
+}
+
+#define REG_CP_SET_BIN_DATA_1					0x00000001
+#define CP_SET_BIN_DATA_1_BIN_SIZE_ADDRESS__MASK		0xffffffff
+#define CP_SET_BIN_DATA_1_BIN_SIZE_ADDRESS__SHIFT		0
+static inline uint32_t CP_SET_BIN_DATA_1_BIN_SIZE_ADDRESS(uint32_t val)
+{
+	return ((val) << CP_SET_BIN_DATA_1_BIN_SIZE_ADDRESS__SHIFT) & CP_SET_BIN_DATA_1_BIN_SIZE_ADDRESS__MASK;
+}
+
 
 #endif /* ADRENO_PM4_XML */
diff --git a/drivers/gpu/drm/msm/dsi/dsi.xml.h b/drivers/gpu/drm/msm/dsi/dsi.xml.h
index 87be647e3825..e965898dfda6 100644
--- a/drivers/gpu/drm/msm/dsi/dsi.xml.h
+++ b/drivers/gpu/drm/msm/dsi/dsi.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h b/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
index 747a6ef4211f..f2bdda957205 100644
--- a/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
+++ b/drivers/gpu/drm/msm/dsi/mmss_cc.xml.h
@@ -10,16 +10,16 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -112,5 +112,11 @@ static inline uint32_t MMSS_CC_CLK_NS_VAL(uint32_t val)
 	return ((val) << MMSS_CC_CLK_NS_VAL__SHIFT) & MMSS_CC_CLK_NS_VAL__MASK;
 }
 
+#define REG_MMSS_CC_DSI2_PIXEL_CC				0x00000094
+
+#define REG_MMSS_CC_DSI2_PIXEL_NS				0x000000e4
+
+#define REG_MMSS_CC_DSI2_PIXEL_CC2				0x00000264
+
 
 #endif /* MMSS_CC_XML */
diff --git a/drivers/gpu/drm/msm/dsi/sfpb.xml.h b/drivers/gpu/drm/msm/dsi/sfpb.xml.h
index 48e03acf19bf..e5b071ffd865 100644
--- a/drivers/gpu/drm/msm/dsi/sfpb.xml.h
+++ b/drivers/gpu/drm/msm/dsi/sfpb.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c
index 7f7aadef8a82..9d00dcba6959 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.c
@@ -123,7 +123,8 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 	for (i = 0; i < config->hpd_reg_cnt; i++) {
 		struct regulator *reg;
 
-		reg = devm_regulator_get(&pdev->dev, config->hpd_reg_names[i]);
+		reg = devm_regulator_get(&pdev->dev,
+				config->hpd_reg_names[i]);
 		if (IS_ERR(reg)) {
 			ret = PTR_ERR(reg);
 			dev_err(dev->dev, "failed to get hpd regulator: %s (%d)\n",
@@ -138,7 +139,8 @@ struct hdmi *hdmi_init(struct drm_device *dev, struct drm_encoder *encoder)
 	for (i = 0; i < config->pwr_reg_cnt; i++) {
 		struct regulator *reg;
 
-		reg = devm_regulator_get(&pdev->dev, config->pwr_reg_names[i]);
+		reg = devm_regulator_get(&pdev->dev,
+				config->pwr_reg_names[i]);
 		if (IS_ERR(reg)) {
 			ret = PTR_ERR(reg);
 			dev_err(dev->dev, "failed to get pwr regulator: %s (%d)\n",
@@ -256,47 +258,68 @@ static void set_hdmi_pdev(struct drm_device *dev,
 	priv->hdmi_pdev = pdev;
 }
 
-static int hdmi_bind(struct device *dev, struct device *master, void *data)
-{
-	static struct hdmi_platform_config config = {};
 #ifdef CONFIG_OF
-	struct device_node *of_node = dev->of_node;
-
-	int get_gpio(const char *name)
-	{
-		int gpio = of_get_named_gpio(of_node, name, 0);
+static int get_gpio(struct device *dev, struct device_node *of_node, const char *name)
+{
+	int gpio = of_get_named_gpio(of_node, name, 0);
+	if (gpio < 0) {
+		char name2[32];
+		snprintf(name2, sizeof(name2), "%s-gpio", name);
+		gpio = of_get_named_gpio(of_node, name2, 0);
 		if (gpio < 0) {
 			dev_err(dev, "failed to get gpio: %s (%d)\n",
 					name, gpio);
 			gpio = -1;
 		}
-		return gpio;
 	}
+	return gpio;
+}
+#endif
+
+static int hdmi_bind(struct device *dev, struct device *master, void *data)
+{
+	static struct hdmi_platform_config config = {};
+#ifdef CONFIG_OF
+	struct device_node *of_node = dev->of_node;
 
-	/* TODO actually use DT.. */
-	static const char *hpd_reg_names[] = {"hpd-gdsc", "hpd-5v"};
-	static const char *pwr_reg_names[] = {"core-vdda", "core-vcc"};
-	static const char *hpd_clk_names[] = {"iface_clk", "core_clk", "mdp_core_clk"};
-	static unsigned long hpd_clk_freq[] = {0, 19200000, 0};
-	static const char *pwr_clk_names[] = {"extp_clk", "alt_iface_clk"};
+	if (of_device_is_compatible(of_node, "qcom,hdmi-tx-8074")) {
+		static const char *hpd_reg_names[] = {"hpd-gdsc", "hpd-5v"};
+		static const char *pwr_reg_names[] = {"core-vdda", "core-vcc"};
+		static const char *hpd_clk_names[] = {"iface_clk", "core_clk", "mdp_core_clk"};
+		static unsigned long hpd_clk_freq[] = {0, 19200000, 0};
+		static const char *pwr_clk_names[] = {"extp_clk", "alt_iface_clk"};
+		config.phy_init      = hdmi_phy_8x74_init;
+		config.hpd_reg_names = hpd_reg_names;
+		config.hpd_reg_cnt   = ARRAY_SIZE(hpd_reg_names);
+		config.pwr_reg_names = pwr_reg_names;
+		config.pwr_reg_cnt   = ARRAY_SIZE(pwr_reg_names);
+		config.hpd_clk_names = hpd_clk_names;
+		config.hpd_freq      = hpd_clk_freq;
+		config.hpd_clk_cnt   = ARRAY_SIZE(hpd_clk_names);
+		config.pwr_clk_names = pwr_clk_names;
+		config.pwr_clk_cnt   = ARRAY_SIZE(pwr_clk_names);
+		config.shared_irq    = true;
+	} else if (of_device_is_compatible(of_node, "qcom,hdmi-tx-8960")) {
+		static const char *hpd_clk_names[] = {"core_clk", "master_iface_clk", "slave_iface_clk"};
+		static const char *hpd_reg_names[] = {"core-vdda", "hdmi-mux"};
+		config.phy_init      = hdmi_phy_8960_init;
+		config.hpd_reg_names = hpd_reg_names;
+		config.hpd_reg_cnt   = ARRAY_SIZE(hpd_reg_names);
+		config.hpd_clk_names = hpd_clk_names;
+		config.hpd_clk_cnt   = ARRAY_SIZE(hpd_clk_names);
+	} else if (of_device_is_compatible(of_node, "qcom,hdmi-tx-8660")) {
+		config.phy_init      = hdmi_phy_8x60_init;
+	} else {
+		dev_err(dev, "unknown phy: %s\n", of_node->name);
+	}
 
-	config.phy_init      = hdmi_phy_8x74_init;
 	config.mmio_name     = "core_physical";
-	config.hpd_reg_names = hpd_reg_names;
-	config.hpd_reg_cnt   = ARRAY_SIZE(hpd_reg_names);
-	config.pwr_reg_names = pwr_reg_names;
-	config.pwr_reg_cnt   = ARRAY_SIZE(pwr_reg_names);
-	config.hpd_clk_names = hpd_clk_names;
-	config.hpd_freq      = hpd_clk_freq;
-	config.hpd_clk_cnt   = ARRAY_SIZE(hpd_clk_names);
-	config.pwr_clk_names = pwr_clk_names;
-	config.pwr_clk_cnt   = ARRAY_SIZE(pwr_clk_names);
-	config.ddc_clk_gpio  = get_gpio("qcom,hdmi-tx-ddc-clk");
-	config.ddc_data_gpio = get_gpio("qcom,hdmi-tx-ddc-data");
-	config.hpd_gpio      = get_gpio("qcom,hdmi-tx-hpd");
-	config.mux_en_gpio   = get_gpio("qcom,hdmi-tx-mux-en");
-	config.mux_sel_gpio  = get_gpio("qcom,hdmi-tx-mux-sel");
-	config.shared_irq    = true;
+	config.ddc_clk_gpio  = get_gpio(dev, of_node, "qcom,hdmi-tx-ddc-clk");
+	config.ddc_data_gpio = get_gpio(dev, of_node, "qcom,hdmi-tx-ddc-data");
+	config.hpd_gpio      = get_gpio(dev, of_node, "qcom,hdmi-tx-hpd");
+	config.mux_en_gpio   = get_gpio(dev, of_node, "qcom,hdmi-tx-mux-en");
+	config.mux_sel_gpio  = get_gpio(dev, of_node, "qcom,hdmi-tx-mux-sel");
+	config.mux_lpm_gpio  = get_gpio(dev, of_node, "qcom,hdmi-tx-mux-lpm");
 
 #else
 	static const char *hpd_clk_names[] = {
@@ -373,7 +396,9 @@ static int hdmi_dev_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id dt_match[] = {
-	{ .compatible = "qcom,hdmi-tx" },
+	{ .compatible = "qcom,hdmi-tx-8074" },
+	{ .compatible = "qcom,hdmi-tx-8960" },
+	{ .compatible = "qcom,hdmi-tx-8660" },
 	{}
 };
 
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h
index 9d7723c6528a..b981995410b5 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.h
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.h
@@ -96,6 +96,7 @@ struct hdmi_platform_config {
 
 	/* gpio's: */
 	int ddc_clk_gpio, ddc_data_gpio, hpd_gpio, mux_en_gpio, mux_sel_gpio;
+	int mux_lpm_gpio;
 
 	/* older devices had their own irq, mdp5+ it is shared w/ mdp: */
 	bool shared_irq;
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.xml.h b/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
index e2636582cfd7..76fd0cfc6558 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
+++ b/drivers/gpu/drm/msm/hdmi/hdmi.xml.h
@@ -10,16 +10,16 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -148,9 +148,9 @@ static inline uint32_t REG_HDMI_GENERIC0(uint32_t i0) { return 0x00000088 + 0x4*
 
 static inline uint32_t REG_HDMI_GENERIC1(uint32_t i0) { return 0x000000a8 + 0x4*i0; }
 
-static inline uint32_t REG_HDMI_ACR(uint32_t i0) { return 0x000000c4 + 0x8*i0; }
+static inline uint32_t REG_HDMI_ACR(enum hdmi_acr_cts i0) { return 0x000000c4 + 0x8*i0; }
 
-static inline uint32_t REG_HDMI_ACR_0(uint32_t i0) { return 0x000000c4 + 0x8*i0; }
+static inline uint32_t REG_HDMI_ACR_0(enum hdmi_acr_cts i0) { return 0x000000c4 + 0x8*i0; }
 #define HDMI_ACR_0_CTS__MASK					0xfffff000
 #define HDMI_ACR_0_CTS__SHIFT					12
 static inline uint32_t HDMI_ACR_0_CTS(uint32_t val)
@@ -158,7 +158,7 @@ static inline uint32_t HDMI_ACR_0_CTS(uint32_t val)
 	return ((val) << HDMI_ACR_0_CTS__SHIFT) & HDMI_ACR_0_CTS__MASK;
 }
 
-static inline uint32_t REG_HDMI_ACR_1(uint32_t i0) { return 0x000000c8 + 0x8*i0; }
+static inline uint32_t REG_HDMI_ACR_1(enum hdmi_acr_cts i0) { return 0x000000c8 + 0x8*i0; }
 #define HDMI_ACR_1_N__MASK					0xffffffff
 #define HDMI_ACR_1_N__SHIFT					0
 static inline uint32_t HDMI_ACR_1_N(uint32_t val)
@@ -552,6 +552,103 @@ static inline uint32_t HDMI_8x60_PHY_REG1_OUTVOL_SWING_CTRL(uint32_t val)
 #define REG_HDMI_8960_PHY_REG11					0x0000042c
 
 #define REG_HDMI_8960_PHY_REG12					0x00000430
+#define HDMI_8960_PHY_REG12_SW_RESET				0x00000020
+#define HDMI_8960_PHY_REG12_PWRDN_B				0x00000080
+
+#define REG_HDMI_8960_PHY_REG_BIST_CFG				0x00000434
+
+#define REG_HDMI_8960_PHY_DEBUG_BUS_SEL				0x00000438
+
+#define REG_HDMI_8960_PHY_REG_MISC0				0x0000043c
+
+#define REG_HDMI_8960_PHY_REG13					0x00000440
+
+#define REG_HDMI_8960_PHY_REG14					0x00000444
+
+#define REG_HDMI_8960_PHY_REG15					0x00000448
+
+#define REG_HDMI_8960_PHY_PLL_REFCLK_CFG			0x00000500
+
+#define REG_HDMI_8960_PHY_PLL_CHRG_PUMP_CFG			0x00000504
+
+#define REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0			0x00000508
+
+#define REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1			0x0000050c
+
+#define REG_HDMI_8960_PHY_PLL_IDAC_ADJ_CFG			0x00000510
+
+#define REG_HDMI_8960_PHY_PLL_I_VI_KVCO_CFG			0x00000514
+
+#define REG_HDMI_8960_PHY_PLL_PWRDN_B				0x00000518
+#define HDMI_8960_PHY_PLL_PWRDN_B_PD_PLL			0x00000002
+#define HDMI_8960_PHY_PLL_PWRDN_B_PLL_PWRDN_B			0x00000008
+
+#define REG_HDMI_8960_PHY_PLL_SDM_CFG0				0x0000051c
+
+#define REG_HDMI_8960_PHY_PLL_SDM_CFG1				0x00000520
+
+#define REG_HDMI_8960_PHY_PLL_SDM_CFG2				0x00000524
+
+#define REG_HDMI_8960_PHY_PLL_SDM_CFG3				0x00000528
+
+#define REG_HDMI_8960_PHY_PLL_SDM_CFG4				0x0000052c
+
+#define REG_HDMI_8960_PHY_PLL_SSC_CFG0				0x00000530
+
+#define REG_HDMI_8960_PHY_PLL_SSC_CFG1				0x00000534
+
+#define REG_HDMI_8960_PHY_PLL_SSC_CFG2				0x00000538
+
+#define REG_HDMI_8960_PHY_PLL_SSC_CFG3				0x0000053c
+
+#define REG_HDMI_8960_PHY_PLL_LOCKDET_CFG0			0x00000540
+
+#define REG_HDMI_8960_PHY_PLL_LOCKDET_CFG1			0x00000544
+
+#define REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2			0x00000548
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0			0x0000054c
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1			0x00000550
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2			0x00000554
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3			0x00000558
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4			0x0000055c
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5			0x00000560
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG6			0x00000564
+
+#define REG_HDMI_8960_PHY_PLL_VCOCAL_CFG7			0x00000568
+
+#define REG_HDMI_8960_PHY_PLL_DEBUG_SEL				0x0000056c
+
+#define REG_HDMI_8960_PHY_PLL_MISC0				0x00000570
+
+#define REG_HDMI_8960_PHY_PLL_MISC1				0x00000574
+
+#define REG_HDMI_8960_PHY_PLL_MISC2				0x00000578
+
+#define REG_HDMI_8960_PHY_PLL_MISC3				0x0000057c
+
+#define REG_HDMI_8960_PHY_PLL_MISC4				0x00000580
+
+#define REG_HDMI_8960_PHY_PLL_MISC5				0x00000584
+
+#define REG_HDMI_8960_PHY_PLL_MISC6				0x00000588
+
+#define REG_HDMI_8960_PHY_PLL_DEBUG_BUS0			0x0000058c
+
+#define REG_HDMI_8960_PHY_PLL_DEBUG_BUS1			0x00000590
+
+#define REG_HDMI_8960_PHY_PLL_DEBUG_BUS2			0x00000594
+
+#define REG_HDMI_8960_PHY_PLL_STATUS0				0x00000598
+#define HDMI_8960_PHY_PLL_STATUS0_PLL_LOCK			0x00000001
+
+#define REG_HDMI_8960_PHY_PLL_STATUS1				0x0000059c
 
 #define REG_HDMI_8x74_ANA_CFG0					0x00000000
 
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
index 28f7e3ec6c28..4aca2a3c667c 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi_connector.c
@@ -63,7 +63,7 @@ static int gpio_config(struct hdmi *hdmi, bool on)
 			ret = gpio_request(config->mux_en_gpio, "HDMI_MUX_EN");
 			if (ret) {
 				dev_err(dev->dev, "'%s'(%d) gpio_request failed: %d\n",
-					"HDMI_MUX_SEL", config->mux_en_gpio, ret);
+					"HDMI_MUX_EN", config->mux_en_gpio, ret);
 				goto error4;
 			}
 			gpio_set_value_cansleep(config->mux_en_gpio, 1);
@@ -78,6 +78,19 @@ static int gpio_config(struct hdmi *hdmi, bool on)
 			}
 			gpio_set_value_cansleep(config->mux_sel_gpio, 0);
 		}
+
+		if (config->mux_lpm_gpio != -1) {
+			ret = gpio_request(config->mux_lpm_gpio,
+					"HDMI_MUX_LPM");
+			if (ret) {
+				dev_err(dev->dev,
+					"'%s'(%d) gpio_request failed: %d\n",
+					"HDMI_MUX_LPM",
+					config->mux_lpm_gpio, ret);
+				goto error6;
+			}
+			gpio_set_value_cansleep(config->mux_lpm_gpio, 1);
+		}
 		DBG("gpio on");
 	} else {
 		gpio_free(config->ddc_clk_gpio);
@@ -93,11 +106,19 @@ static int gpio_config(struct hdmi *hdmi, bool on)
 			gpio_set_value_cansleep(config->mux_sel_gpio, 1);
 			gpio_free(config->mux_sel_gpio);
 		}
+
+		if (config->mux_lpm_gpio != -1) {
+			gpio_set_value_cansleep(config->mux_lpm_gpio, 0);
+			gpio_free(config->mux_lpm_gpio);
+		}
 		DBG("gpio off");
 	}
 
 	return 0;
 
+error6:
+	if (config->mux_sel_gpio != -1)
+		gpio_free(config->mux_sel_gpio);
 error5:
 	if (config->mux_en_gpio != -1)
 		gpio_free(config->mux_en_gpio);
@@ -306,7 +327,7 @@ static void hdmi_connector_destroy(struct drm_connector *connector)
 
 	hdp_disable(hdmi_connector);
 
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 
 	hdmi_unreference(hdmi_connector->hdmi);
@@ -416,7 +437,7 @@ struct drm_connector *hdmi_connector_init(struct hdmi *hdmi)
 	connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	ret = hpd_enable(hdmi_connector);
 	if (ret) {
diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
index e5b7ed5b8f01..f408b69486a8 100644
--- a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
+++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8960.c
@@ -15,14 +15,377 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#ifdef CONFIG_COMMON_CLK
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#endif
+
 #include "hdmi.h"
 
 struct hdmi_phy_8960 {
 	struct hdmi_phy base;
 	struct hdmi *hdmi;
+#ifdef CONFIG_COMMON_CLK
+	struct clk_hw pll_hw;
+	struct clk *pll;
+	unsigned long pixclk;
+#endif
 };
 #define to_hdmi_phy_8960(x) container_of(x, struct hdmi_phy_8960, base)
 
+#ifdef CONFIG_COMMON_CLK
+#define clk_to_phy(x) container_of(x, struct hdmi_phy_8960, pll_hw)
+
+/*
+ * HDMI PLL:
+ *
+ * To get the parent clock setup properly, we need to plug in hdmi pll
+ * configuration into common-clock-framework.
+ */
+
+struct pll_rate {
+	unsigned long rate;
+	struct {
+		uint32_t val;
+		uint32_t reg;
+	} conf[32];
+};
+
+/* NOTE: keep sorted highest freq to lowest: */
+static const struct pll_rate freqtbl[] = {
+	/* 1080p60/1080p50 case */
+	{ 148500000, {
+		{ 0x02, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_CHRG_PUMP_CFG },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x2c, REG_HDMI_8960_PHY_PLL_IDAC_ADJ_CFG  },
+		{ 0x06, REG_HDMI_8960_PHY_PLL_I_VI_KVCO_CFG },
+		{ 0x0a, REG_HDMI_8960_PHY_PLL_PWRDN_B       },
+		{ 0x76, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0x4c, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0xc0, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0x9a, REG_HDMI_8960_PHY_PLL_SSC_CFG0      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG1      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG2      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG3      },
+		{ 0x10, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG0  },
+		{ 0x1a, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG1  },
+		{ 0x0d, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2  },
+		{ 0xe6, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG6   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG7   },
+		{ 0, 0 } }
+	},
+	{ 108000000, {
+		{ 0x08, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x21, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0xf9, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x1c, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x49, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x49, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0, 0 } }
+	},
+	/* 720p60/720p50/1080i60/1080i50/1080p24/1080p30/1080p25 */
+	{ 74250000, {
+		{ 0x0a, REG_HDMI_8960_PHY_PLL_PWRDN_B       },
+		{ 0x12, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x76, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0xe6, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0, 0 } }
+	},
+	{ 65000000, {
+		{ 0x18, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x20, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0xf9, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x8a, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x0b, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x4b, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0x7b, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0x09, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0, 0 } }
+	},
+	/* 480p60/480i60 */
+	{ 27030000, {
+		{ 0x0a, REG_HDMI_8960_PHY_PLL_PWRDN_B       },
+		{ 0x38, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_CHRG_PUMP_CFG },
+		{ 0x20, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0xff, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x4e, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0xd7, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0x03, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0x2a, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x03, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG6   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG7   },
+		{ 0, 0 } }
+	},
+	/* 576p50/576i50 */
+	{ 27000000, {
+		{ 0x32, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_CHRG_PUMP_CFG },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x2c, REG_HDMI_8960_PHY_PLL_IDAC_ADJ_CFG  },
+		{ 0x06, REG_HDMI_8960_PHY_PLL_I_VI_KVCO_CFG },
+		{ 0x0a, REG_HDMI_8960_PHY_PLL_PWRDN_B       },
+		{ 0x7b, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0x4c, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0xc0, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0x9a, REG_HDMI_8960_PHY_PLL_SSC_CFG0      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG1      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG2      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG3      },
+		{ 0x10, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG0  },
+		{ 0x1a, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG1  },
+		{ 0x0d, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2  },
+		{ 0x2a, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x03, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG6   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG7   },
+		{ 0, 0 } }
+	},
+	/* 640x480p60 */
+	{ 25200000, {
+		{ 0x32, REG_HDMI_8960_PHY_PLL_REFCLK_CFG    },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_CHRG_PUMP_CFG },
+		{ 0x01, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG0 },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_LOOP_FLT_CFG1 },
+		{ 0x2c, REG_HDMI_8960_PHY_PLL_IDAC_ADJ_CFG  },
+		{ 0x06, REG_HDMI_8960_PHY_PLL_I_VI_KVCO_CFG },
+		{ 0x0a, REG_HDMI_8960_PHY_PLL_PWRDN_B       },
+		{ 0x77, REG_HDMI_8960_PHY_PLL_SDM_CFG0      },
+		{ 0x4c, REG_HDMI_8960_PHY_PLL_SDM_CFG1      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG2      },
+		{ 0xc0, REG_HDMI_8960_PHY_PLL_SDM_CFG3      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SDM_CFG4      },
+		{ 0x9a, REG_HDMI_8960_PHY_PLL_SSC_CFG0      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG1      },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_SSC_CFG2      },
+		{ 0x20, REG_HDMI_8960_PHY_PLL_SSC_CFG3      },
+		{ 0x10, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG0  },
+		{ 0x1a, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG1  },
+		{ 0x0d, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2  },
+		{ 0xf4, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG0   },
+		{ 0x02, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG1   },
+		{ 0x3b, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG2   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG3   },
+		{ 0x86, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG4   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG5   },
+		{ 0x33, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG6   },
+		{ 0x00, REG_HDMI_8960_PHY_PLL_VCOCAL_CFG7   },
+		{ 0, 0 } }
+	},
+};
+
+static int hdmi_pll_enable(struct clk_hw *hw)
+{
+	struct hdmi_phy_8960 *phy_8960 = clk_to_phy(hw);
+	struct hdmi *hdmi = phy_8960->hdmi;
+	int timeout_count, pll_lock_retry = 10;
+	unsigned int val;
+
+	DBG("");
+
+	/* Assert PLL S/W reset */
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2, 0x8d);
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG0, 0x10);
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG1, 0x1a);
+
+	/* Wait for a short time before de-asserting
+	 * to allow the hardware to complete its job.
+	 * This much of delay should be fine for hardware
+	 * to assert and de-assert.
+	 */
+	udelay(10);
+
+	/* De-assert PLL S/W reset */
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2, 0x0d);
+
+	val = hdmi_read(hdmi, REG_HDMI_8960_PHY_REG12);
+	val |= HDMI_8960_PHY_REG12_SW_RESET;
+	/* Assert PHY S/W reset */
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG12, val);
+	val &= ~HDMI_8960_PHY_REG12_SW_RESET;
+	/* Wait for a short time before de-asserting
+	   to allow the hardware to complete its job.
+	   This much of delay should be fine for hardware
+	   to assert and de-assert. */
+	udelay(10);
+	/* De-assert PHY S/W reset */
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG12, val);
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG2,  0x3f);
+
+	val = hdmi_read(hdmi, REG_HDMI_8960_PHY_REG12);
+	val |= HDMI_8960_PHY_REG12_PWRDN_B;
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG12, val);
+	/* Wait 10 us for enabling global power for PHY */
+	mb();
+	udelay(10);
+
+	val = hdmi_read(hdmi, REG_HDMI_8960_PHY_PLL_PWRDN_B);
+	val |= HDMI_8960_PHY_PLL_PWRDN_B_PLL_PWRDN_B;
+	val &= ~HDMI_8960_PHY_PLL_PWRDN_B_PD_PLL;
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_PWRDN_B, val);
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG2, 0x80);
+
+	timeout_count = 1000;
+	while (--pll_lock_retry > 0) {
+
+		/* are we there yet? */
+		val = hdmi_read(hdmi, REG_HDMI_8960_PHY_PLL_STATUS0);
+		if (val & HDMI_8960_PHY_PLL_STATUS0_PLL_LOCK)
+			break;
+
+		udelay(1);
+
+		if (--timeout_count > 0)
+			continue;
+
+		/*
+		 * PLL has still not locked.
+		 * Do a software reset and try again
+		 * Assert PLL S/W reset first
+		 */
+		hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2, 0x8d);
+		udelay(10);
+		hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_LOCKDET_CFG2, 0x0d);
+
+		/*
+		 * Wait for a short duration for the PLL calibration
+		 * before checking if the PLL gets locked
+		 */
+		udelay(350);
+
+		timeout_count = 1000;
+	}
+
+	return 0;
+}
+
+static void hdmi_pll_disable(struct clk_hw *hw)
+{
+	struct hdmi_phy_8960 *phy_8960 = clk_to_phy(hw);
+	struct hdmi *hdmi = phy_8960->hdmi;
+	unsigned int val;
+
+	DBG("");
+
+	val = hdmi_read(hdmi, REG_HDMI_8960_PHY_REG12);
+	val &= ~HDMI_8960_PHY_REG12_PWRDN_B;
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG12, val);
+
+	val = hdmi_read(hdmi, REG_HDMI_8960_PHY_PLL_PWRDN_B);
+	val |= HDMI_8960_PHY_REG12_SW_RESET;
+	val &= ~HDMI_8960_PHY_REG12_PWRDN_B;
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_PLL_PWRDN_B, val);
+	/* Make sure HDMI PHY/PLL are powered down */
+	mb();
+}
+
+static const struct pll_rate *find_rate(unsigned long rate)
+{
+	int i;
+	for (i = 1; i < ARRAY_SIZE(freqtbl); i++)
+		if (rate > freqtbl[i].rate)
+			return &freqtbl[i-1];
+	return &freqtbl[i-1];
+}
+
+static unsigned long hdmi_pll_recalc_rate(struct clk_hw *hw,
+				unsigned long parent_rate)
+{
+	struct hdmi_phy_8960 *phy_8960 = clk_to_phy(hw);
+	return phy_8960->pixclk;
+}
+
+static long hdmi_pll_round_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long *parent_rate)
+{
+	const struct pll_rate *pll_rate = find_rate(rate);
+	return pll_rate->rate;
+}
+
+static int hdmi_pll_set_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long parent_rate)
+{
+	struct hdmi_phy_8960 *phy_8960 = clk_to_phy(hw);
+	struct hdmi *hdmi = phy_8960->hdmi;
+	const struct pll_rate *pll_rate = find_rate(rate);
+	int i;
+
+	DBG("rate=%lu", rate);
+
+	for (i = 0; pll_rate->conf[i].reg; i++)
+		hdmi_write(hdmi, pll_rate->conf[i].reg, pll_rate->conf[i].val);
+
+	phy_8960->pixclk = rate;
+
+	return 0;
+}
+
+
+static const struct clk_ops hdmi_pll_ops = {
+	.enable = hdmi_pll_enable,
+	.disable = hdmi_pll_disable,
+	.recalc_rate = hdmi_pll_recalc_rate,
+	.round_rate = hdmi_pll_round_rate,
+	.set_rate = hdmi_pll_set_rate,
+};
+
+static const char *hdmi_pll_parents[] = {
+	"pxo",
+};
+
+static struct clk_init_data pll_init = {
+	.name = "hdmi_pll",
+	.ops = &hdmi_pll_ops,
+	.parent_names = hdmi_pll_parents,
+	.num_parents = ARRAY_SIZE(hdmi_pll_parents),
+};
+#endif
+
+/*
+ * HDMI Phy:
+ */
+
 static void hdmi_phy_8960_destroy(struct hdmi_phy *phy)
 {
 	struct hdmi_phy_8960 *phy_8960 = to_hdmi_phy_8960(phy);
@@ -86,6 +449,9 @@ static void hdmi_phy_8960_powerup(struct hdmi_phy *phy,
 	struct hdmi_phy_8960 *phy_8960 = to_hdmi_phy_8960(phy);
 	struct hdmi *hdmi = phy_8960->hdmi;
 
+	DBG("pixclock: %lu", pixclock);
+
+	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG2, 0x00);
 	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG0, 0x1b);
 	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG1, 0xf2);
 	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG4, 0x00);
@@ -104,6 +470,8 @@ static void hdmi_phy_8960_powerdown(struct hdmi_phy *phy)
 	struct hdmi_phy_8960 *phy_8960 = to_hdmi_phy_8960(phy);
 	struct hdmi *hdmi = phy_8960->hdmi;
 
+	DBG("");
+
 	hdmi_write(hdmi, REG_HDMI_8960_PHY_REG2, 0x7f);
 }
 
@@ -119,6 +487,14 @@ struct hdmi_phy *hdmi_phy_8960_init(struct hdmi *hdmi)
 	struct hdmi_phy_8960 *phy_8960;
 	struct hdmi_phy *phy = NULL;
 	int ret;
+#ifdef CONFIG_COMMON_CLK
+	int i;
+
+	/* sanity check: */
+	for (i = 0; i < (ARRAY_SIZE(freqtbl) - 1); i++)
+		if (WARN_ON(freqtbl[i].rate < freqtbl[i+1].rate))
+			return ERR_PTR(-EINVAL);
+#endif
 
 	phy_8960 = kzalloc(sizeof(*phy_8960), GFP_KERNEL);
 	if (!phy_8960) {
@@ -132,6 +508,16 @@ struct hdmi_phy *hdmi_phy_8960_init(struct hdmi *hdmi)
 
 	phy_8960->hdmi = hdmi;
 
+#ifdef CONFIG_COMMON_CLK
+	phy_8960->pll_hw.init = &pll_init;
+	phy_8960->pll = devm_clk_register(hdmi->dev->dev, &phy_8960->pll_hw);
+	if (IS_ERR(phy_8960->pll)) {
+		ret = PTR_ERR(phy_8960->pll);
+		phy_8960->pll = NULL;
+		goto fail;
+	}
+#endif
+
 	return phy;
 
 fail:
diff --git a/drivers/gpu/drm/msm/hdmi/qfprom.xml.h b/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
index d591567173c4..d53c29327df9 100644
--- a/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
+++ b/drivers/gpu/drm/msm/hdmi/qfprom.xml.h
@@ -10,14 +10,14 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
index 416a26e1e58d..03c0bd9cd5b9 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4.xml.h
@@ -10,16 +10,16 @@ git clone https://github.com/freedreno/envytools.git
 The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/msm.xml                 (    647 bytes, from 2013-11-30 14:45:35)
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  20457 bytes, from 2014-08-01 12:22:48)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2014-07-17 15:34:33)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-07-17 15:34:33)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
-- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
+- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1686 bytes, from 2014-08-01 12:23:53)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-07-17 15:33:30)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -871,6 +871,101 @@ static inline uint32_t MDP4_LCDC_UNDERFLOW_CLR_COLOR(uint32_t val)
 #define MDP4_LCDC_CTRL_POLARITY_VSYNC_LOW			0x00000002
 #define MDP4_LCDC_CTRL_POLARITY_DATA_EN_LOW			0x00000004
 
+#define REG_MDP4_LCDC_LVDS_INTF_CTL				0x000c2000
+#define MDP4_LCDC_LVDS_INTF_CTL_MODE_SEL			0x00000004
+#define MDP4_LCDC_LVDS_INTF_CTL_RGB_OUT				0x00000008
+#define MDP4_LCDC_LVDS_INTF_CTL_CH_SWAP				0x00000010
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_RES_BIT			0x00000020
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_RES_BIT			0x00000040
+#define MDP4_LCDC_LVDS_INTF_CTL_ENABLE				0x00000080
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN		0x00000100
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN		0x00000200
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN		0x00000400
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN		0x00000800
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN		0x00001000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN		0x00002000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN		0x00004000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE3_EN		0x00008000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN			0x00010000
+#define MDP4_LCDC_LVDS_INTF_CTL_CH2_CLK_LANE_EN			0x00020000
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL(uint32_t i0) { return 0x000c2014 + 0x8*i0; }
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(uint32_t i0) { return 0x000c2014 + 0x8*i0; }
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__MASK		0x000000ff
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__SHIFT		0
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__MASK		0x0000ff00
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__SHIFT		8
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__MASK		0x00ff0000
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__SHIFT		16
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__MASK		0xff000000
+#define MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__SHIFT		24
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3__MASK;
+}
+
+static inline uint32_t REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(uint32_t i0) { return 0x000c2018 + 0x8*i0; }
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__MASK		0x000000ff
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__SHIFT		0
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__MASK		0x0000ff00
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__SHIFT		8
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5__MASK;
+}
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__MASK		0x00ff0000
+#define MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__SHIFT		16
+static inline uint32_t MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(uint32_t val)
+{
+	return ((val) << MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__SHIFT) & MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6__MASK;
+}
+
+#define REG_MDP4_LCDC_LVDS_PHY_RESET				0x000c2034
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_0				0x000c3000
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_1				0x000c3004
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_2				0x000c3008
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_3				0x000c300c
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_5				0x000c3014
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_6				0x000c3018
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_7				0x000c301c
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_8				0x000c3020
+
+#define REG_MDP4_LVDS_PHY_PLL_CTRL_9				0x000c3024
+
+#define REG_MDP4_LVDS_PHY_PLL_LOCKED				0x000c3080
+
+#define REG_MDP4_LVDS_PHY_CFG2					0x000c3108
+
+#define REG_MDP4_LVDS_PHY_CFG0					0x000c3100
+#define MDP4_LVDS_PHY_CFG0_SERIALIZATION_ENBLE			0x00000010
+#define MDP4_LVDS_PHY_CFG0_CHANNEL0				0x00000040
+#define MDP4_LVDS_PHY_CFG0_CHANNEL1				0x00000080
+
 #define REG_MDP4_DTV						0x000d0000
 
 #define REG_MDP4_DTV_ENABLE					0x000d0000
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
index 74cebb51e8c2..7d00f7fb5773 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c
@@ -273,14 +273,17 @@ static void blend_setup(struct drm_crtc *crtc)
 	};
 	bool alpha[4]= { false, false, false, false };
 
+	/* Don't rely on value read back from hw, but instead use our
+	 * own shadowed value.  Possibly disable/reenable looses the
+	 * previous value and goes back to power-on default?
+	 */
+	mixer_cfg = mdp4_kms->mixer_cfg;
+
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_LOW0(ovlp), 0);
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_LOW1(ovlp), 0);
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_HIGH0(ovlp), 0);
 	mdp4_write(mdp4_kms, REG_MDP4_OVLP_TRANSP_HIGH1(ovlp), 0);
 
-	/* TODO single register for all CRTCs, so this won't work properly
-	 * when multiple CRTCs are active..
-	 */
 	for (i = 0; i < ARRAY_SIZE(mdp4_crtc->planes); i++) {
 		struct drm_plane *plane = mdp4_crtc->planes[i];
 		if (plane) {
@@ -291,7 +294,8 @@ static void blend_setup(struct drm_crtc *crtc)
 					to_mdp_format(msm_framebuffer_format(plane->fb));
 				alpha[idx-1] = format->alpha_enable;
 			}
-			mixer_cfg |= mixercfg(mdp4_crtc->mixer, pipe_id, stages[idx]);
+			mixer_cfg = mixercfg(mixer_cfg, mdp4_crtc->mixer,
+					pipe_id, stages[idx]);
 		}
 	}
 
@@ -320,6 +324,7 @@ static void blend_setup(struct drm_crtc *crtc)
 		mdp4_write(mdp4_kms, REG_MDP4_OVLP_STAGE_TRANSP_HIGH1(ovlp, i), 0);
 	}
 
+	mdp4_kms->mixer_cfg = mixer_cfg;
 	mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG, mixer_cfg);
 }
 
@@ -397,6 +402,7 @@ static void mdp4_crtc_prepare(struct drm_crtc *crtc)
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	DBG("%s", mdp4_crtc->name);
 	/* make sure we hold a ref to mdp clks while setting up mode: */
+	drm_crtc_vblank_get(crtc);
 	mdp4_enable(get_kms(crtc));
 	mdp4_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
 }
@@ -407,6 +413,7 @@ static void mdp4_crtc_commit(struct drm_crtc *crtc)
 	crtc_flush(crtc);
 	/* drop the ref to mdp clk's that we got in prepare: */
 	mdp4_disable(get_kms(crtc));
+	drm_crtc_vblank_put(crtc);
 }
 
 static int mdp4_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
@@ -670,7 +677,7 @@ void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config)
 }
 
 /* set interface for routing crtc->encoder: */
-void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf)
+void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf, int mixer)
 {
 	struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc);
 	struct mdp4_kms *mdp4_kms = get_kms(crtc);
@@ -696,15 +703,13 @@ void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf)
 	if (intf == INTF_DSI_VIDEO) {
 		intf_sel &= ~MDP4_DISP_INTF_SEL_DSI_CMD;
 		intf_sel |= MDP4_DISP_INTF_SEL_DSI_VIDEO;
-		mdp4_crtc->mixer = 0;
 	} else if (intf == INTF_DSI_CMD) {
 		intf_sel &= ~MDP4_DISP_INTF_SEL_DSI_VIDEO;
 		intf_sel |= MDP4_DISP_INTF_SEL_DSI_CMD;
-		mdp4_crtc->mixer = 0;
-	} else if (intf == INTF_LCDC_DTV){
-		mdp4_crtc->mixer = 1;
 	}
 
+	mdp4_crtc->mixer = mixer;
+
 	blend_setup(crtc);
 
 	DBG("%s: intf_sel=%08x", mdp4_crtc->name, intf_sel);
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
index 067ed03b35fe..c3878420180b 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c
@@ -233,7 +233,7 @@ static void mdp4_dtv_encoder_commit(struct drm_encoder *encoder)
 			MDP4_DMA_CONFIG_G_BPC(BPC8) |
 			MDP4_DMA_CONFIG_B_BPC(BPC8) |
 			MDP4_DMA_CONFIG_PACK(0x21));
-	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV);
+	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV, 1);
 	mdp4_dtv_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
 }
 
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
index 0bb4faa17523..79d804e61cc4 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c
@@ -106,6 +106,7 @@ static int mdp4_hw_init(struct msm_kms *kms)
 
 	if (mdp4_kms->rev >= 2)
 		mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG_UPDATE_METHOD, 1);
+	mdp4_write(mdp4_kms, REG_MDP4_LAYERMIXER_IN_CFG, 0);
 
 	/* disable CSC matrix / YUV by default: */
 	mdp4_write(mdp4_kms, REG_MDP4_PIPE_OP_MODE(VG1), 0);
@@ -147,7 +148,7 @@ static void mdp4_destroy(struct msm_kms *kms)
 	if (mdp4_kms->blank_cursor_iova)
 		msm_gem_put_iova(mdp4_kms->blank_cursor_bo, mdp4_kms->id);
 	if (mdp4_kms->blank_cursor_bo)
-		drm_gem_object_unreference(mdp4_kms->blank_cursor_bo);
+		drm_gem_object_unreference_unlocked(mdp4_kms->blank_cursor_bo);
 	kfree(mdp4_kms);
 }
 
@@ -176,6 +177,8 @@ int mdp4_disable(struct mdp4_kms *mdp4_kms)
 	if (mdp4_kms->pclk)
 		clk_disable_unprepare(mdp4_kms->pclk);
 	clk_disable_unprepare(mdp4_kms->lut_clk);
+	if (mdp4_kms->axi_clk)
+		clk_disable_unprepare(mdp4_kms->axi_clk);
 
 	return 0;
 }
@@ -188,10 +191,34 @@ int mdp4_enable(struct mdp4_kms *mdp4_kms)
 	if (mdp4_kms->pclk)
 		clk_prepare_enable(mdp4_kms->pclk);
 	clk_prepare_enable(mdp4_kms->lut_clk);
+	if (mdp4_kms->axi_clk)
+		clk_prepare_enable(mdp4_kms->axi_clk);
 
 	return 0;
 }
 
+#ifdef CONFIG_OF
+static struct drm_panel *detect_panel(struct drm_device *dev, const char *name)
+{
+	struct device_node *n;
+	struct drm_panel *panel = NULL;
+
+	n = of_parse_phandle(dev->dev->of_node, name, 0);
+	if (n) {
+		panel = of_drm_find_panel(n);
+		if (!panel)
+			panel = ERR_PTR(-EPROBE_DEFER);
+	}
+
+	return panel;
+}
+#else
+static struct drm_panel *detect_panel(struct drm_device *dev, const char *name)
+{
+	// ??? maybe use a module param to specify which panel is attached?
+}
+#endif
+
 static int modeset_init(struct mdp4_kms *mdp4_kms)
 {
 	struct drm_device *dev = mdp4_kms->dev;
@@ -199,14 +226,11 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 	struct drm_plane *plane;
 	struct drm_crtc *crtc;
 	struct drm_encoder *encoder;
+	struct drm_connector *connector;
+	struct drm_panel *panel;
 	struct hdmi *hdmi;
 	int ret;
 
-	/*
-	 *  NOTE: this is a bit simplistic until we add support
-	 * for more than just RGB1->DMA_E->DTV->HDMI
-	 */
-
 	/* construct non-private planes: */
 	plane = mdp4_plane_init(dev, VG1, false);
 	if (IS_ERR(plane)) {
@@ -224,7 +248,57 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 	}
 	priv->planes[priv->num_planes++] = plane;
 
-	/* the CRTCs get constructed with a private plane: */
+	/*
+	 * Setup the LCDC/LVDS path: RGB2 -> DMA_P -> LCDC -> LVDS:
+	 */
+
+	panel = detect_panel(dev, "qcom,lvds-panel");
+	if (IS_ERR(panel)) {
+		ret = PTR_ERR(panel);
+		dev_err(dev->dev, "failed to detect LVDS panel: %d\n", ret);
+		goto fail;
+	}
+
+	plane = mdp4_plane_init(dev, RGB2, true);
+	if (IS_ERR(plane)) {
+		dev_err(dev->dev, "failed to construct plane for RGB2\n");
+		ret = PTR_ERR(plane);
+		goto fail;
+	}
+
+	crtc  = mdp4_crtc_init(dev, plane, priv->num_crtcs, 0, DMA_P);
+	if (IS_ERR(crtc)) {
+		dev_err(dev->dev, "failed to construct crtc for DMA_P\n");
+		ret = PTR_ERR(crtc);
+		goto fail;
+	}
+
+	encoder = mdp4_lcdc_encoder_init(dev, panel);
+	if (IS_ERR(encoder)) {
+		dev_err(dev->dev, "failed to construct LCDC encoder\n");
+		ret = PTR_ERR(encoder);
+		goto fail;
+	}
+
+	/* LCDC can be hooked to DMA_P: */
+	encoder->possible_crtcs = 1 << priv->num_crtcs;
+
+	priv->crtcs[priv->num_crtcs++] = crtc;
+	priv->encoders[priv->num_encoders++] = encoder;
+
+	connector = mdp4_lvds_connector_init(dev, panel, encoder);
+	if (IS_ERR(connector)) {
+		ret = PTR_ERR(connector);
+		dev_err(dev->dev, "failed to initialize LVDS connector: %d\n", ret);
+		goto fail;
+	}
+
+	priv->connectors[priv->num_connectors++] = connector;
+
+	/*
+	 * Setup DTV/HDMI path: RGB1 -> DMA_E -> DTV -> HDMI:
+	 */
+
 	plane = mdp4_plane_init(dev, RGB1, true);
 	if (IS_ERR(plane)) {
 		dev_err(dev->dev, "failed to construct plane for RGB1\n");
@@ -238,7 +312,6 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 		ret = PTR_ERR(crtc);
 		goto fail;
 	}
-	priv->crtcs[priv->num_crtcs++] = crtc;
 
 	encoder = mdp4_dtv_encoder_init(dev);
 	if (IS_ERR(encoder)) {
@@ -246,7 +319,11 @@ static int modeset_init(struct mdp4_kms *mdp4_kms)
 		ret = PTR_ERR(encoder);
 		goto fail;
 	}
-	encoder->possible_crtcs = 0x1;     /* DTV can be hooked to DMA_E */
+
+	/* DTV can be hooked to DMA_E: */
+	encoder->possible_crtcs = 1 << priv->num_crtcs;
+
+	priv->crtcs[priv->num_crtcs++] = crtc;
 	priv->encoders[priv->num_encoders++] = encoder;
 
 	hdmi = hdmi_init(dev, encoder);
@@ -294,15 +371,17 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev)
 		goto fail;
 	}
 
-	mdp4_kms->dsi_pll_vdda = devm_regulator_get(&pdev->dev, "dsi_pll_vdda");
+	mdp4_kms->dsi_pll_vdda =
+			devm_regulator_get_optional(&pdev->dev, "dsi_pll_vdda");
 	if (IS_ERR(mdp4_kms->dsi_pll_vdda))
 		mdp4_kms->dsi_pll_vdda = NULL;
 
-	mdp4_kms->dsi_pll_vddio = devm_regulator_get(&pdev->dev, "dsi_pll_vddio");
+	mdp4_kms->dsi_pll_vddio =
+			devm_regulator_get_optional(&pdev->dev, "dsi_pll_vddio");
 	if (IS_ERR(mdp4_kms->dsi_pll_vddio))
 		mdp4_kms->dsi_pll_vddio = NULL;
 
-	mdp4_kms->vdd = devm_regulator_get(&pdev->dev, "vdd");
+	mdp4_kms->vdd = devm_regulator_get_exclusive(&pdev->dev, "vdd");
 	if (IS_ERR(mdp4_kms->vdd))
 		mdp4_kms->vdd = NULL;
 
@@ -333,6 +412,13 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev)
 		goto fail;
 	}
 
+	mdp4_kms->axi_clk = devm_clk_get(&pdev->dev, "mdp_axi_clk");
+	if (IS_ERR(mdp4_kms->axi_clk)) {
+		dev_err(dev->dev, "failed to get axi_clk\n");
+		ret = PTR_ERR(mdp4_kms->axi_clk);
+		goto fail;
+	}
+
 	clk_set_rate(mdp4_kms->clk, config->max_clk);
 	clk_set_rate(mdp4_kms->lut_clk, config->max_clk);
 
@@ -348,7 +434,7 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev)
 	mdelay(16);
 
 	if (config->iommu) {
-		mmu = msm_iommu_new(dev, config->iommu);
+		mmu = msm_iommu_new(&pdev->dev, config->iommu);
 		if (IS_ERR(mmu)) {
 			ret = PTR_ERR(mmu);
 			goto fail;
@@ -406,6 +492,8 @@ static struct mdp4_platform_config *mdp4_get_config(struct platform_device *dev)
 	static struct mdp4_platform_config config = {};
 #ifdef CONFIG_OF
 	/* TODO */
+	config.max_clk = 266667000;
+	config.iommu = iommu_domain_alloc(&platform_bus_type);
 #else
 	if (cpu_is_apq8064())
 		config.max_clk = 266667000;
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
index 715520c54cde..9ff6e7ccfe90 100644
--- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h
@@ -23,6 +23,8 @@
 #include "mdp/mdp_kms.h"
 #include "mdp4.xml.h"
 
+#include "drm_panel.h"
+
 struct mdp4_kms {
 	struct mdp_kms base;
 
@@ -30,6 +32,13 @@ struct mdp4_kms {
 
 	int rev;
 
+	/* Shadow value for MDP4_LAYERMIXER_IN_CFG.. since setup for all
+	 * crtcs/encoders is in one shared register, we need to update it
+	 * via read/modify/write.  But to avoid getting confused by power-
+	 * on-default values after resume, use this shadow value instead:
+	 */
+	uint32_t mixer_cfg;
+
 	/* mapper-id used to request GEM buffer mapped for scanout: */
 	int id;
 
@@ -42,6 +51,7 @@ struct mdp4_kms {
 	struct clk *clk;
 	struct clk *pclk;
 	struct clk *lut_clk;
+	struct clk *axi_clk;
 
 	struct mdp_irq error_handler;
 
@@ -73,7 +83,7 @@ static inline uint32_t pipe2flush(enum mdp4_pipe pipe)
 	case VG1:      return MDP4_OVERLAY_FLUSH_VG1;
 	case VG2:      return MDP4_OVERLAY_FLUSH_VG2;
 	case RGB1:     return MDP4_OVERLAY_FLUSH_RGB1;
-	case RGB2:     return MDP4_OVERLAY_FLUSH_RGB1;
+	case RGB2:     return MDP4_OVERLAY_FLUSH_RGB2;
 	default:       return 0;
 	}
 }
@@ -107,38 +117,50 @@ static inline uint32_t dma2err(enum mdp4_dma dma)
 	}
 }
 
-static inline uint32_t mixercfg(int mixer, enum mdp4_pipe pipe,
-		enum mdp_mixer_stage_id stage)
+static inline uint32_t mixercfg(uint32_t mixer_cfg, int mixer,
+		enum mdp4_pipe pipe, enum mdp_mixer_stage_id stage)
 {
-	uint32_t mixer_cfg = 0;
-
 	switch (pipe) {
 	case VG1:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE0(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE0__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE0_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE0(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE0_MIXER1);
 		break;
 	case VG2:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE1(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE1__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE1_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE1(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE1_MIXER1);
 		break;
 	case RGB1:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE2(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE2__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE2_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE2(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE2_MIXER1);
 		break;
 	case RGB2:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE3(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE3__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE3_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE3(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE3_MIXER1);
 		break;
 	case RGB3:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE4(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE4__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE4_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE4(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE4_MIXER1);
 		break;
 	case VG3:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE5(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE5__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE5_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE5(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE5_MIXER1);
 		break;
 	case VG4:
-		mixer_cfg = MDP4_LAYERMIXER_IN_CFG_PIPE6(stage) |
+		mixer_cfg &= ~(MDP4_LAYERMIXER_IN_CFG_PIPE6__MASK |
+				MDP4_LAYERMIXER_IN_CFG_PIPE6_MIXER1);
+		mixer_cfg |= MDP4_LAYERMIXER_IN_CFG_PIPE6(stage) |
 			COND(mixer == 1, MDP4_LAYERMIXER_IN_CFG_PIPE6_MIXER1);
 		break;
 	default:
@@ -187,7 +209,7 @@ struct drm_plane *mdp4_plane_init(struct drm_device *dev,
 uint32_t mdp4_crtc_vblank(struct drm_crtc *crtc);
 void mdp4_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file);
 void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config);
-void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf);
+void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf, int mixer);
 void mdp4_crtc_attach(struct drm_crtc *crtc, struct drm_plane *plane);
 void mdp4_crtc_detach(struct drm_crtc *crtc, struct drm_plane *plane);
 struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
@@ -197,6 +219,22 @@ struct drm_crtc *mdp4_crtc_init(struct drm_device *dev,
 long mdp4_dtv_round_pixclk(struct drm_encoder *encoder, unsigned long rate);
 struct drm_encoder *mdp4_dtv_encoder_init(struct drm_device *dev);
 
+long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate);
+struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev,
+		struct drm_panel *panel);
+
+struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev,
+		struct drm_panel *panel, struct drm_encoder *encoder);
+
+#ifdef CONFIG_COMMON_CLK
+struct clk *mpd4_lvds_pll_init(struct drm_device *dev);
+#else
+static inline struct clk *mpd4_lvds_pll_init(struct drm_device *dev)
+{
+	return ERR_PTR(-ENODEV);
+}
+#endif
+
 #ifdef CONFIG_MSM_BUS_SCALING
 static inline int match_dev_name(struct device *dev, void *data)
 {
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c
new file mode 100644
index 000000000000..41f6436754fc
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c
@@ -0,0 +1,506 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ * Author: Vinay Simha <vinaysimha@inforcecomputing.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "mdp4_kms.h"
+
+#include "drm_crtc.h"
+#include "drm_crtc_helper.h"
+
+struct mdp4_lcdc_encoder {
+	struct drm_encoder base;
+	struct drm_panel *panel;
+	struct clk *lcdc_clk;
+	unsigned long int pixclock;
+	struct regulator *regs[3];
+	bool enabled;
+	uint32_t bsc;
+};
+#define to_mdp4_lcdc_encoder(x) container_of(x, struct mdp4_lcdc_encoder, base)
+
+static struct mdp4_kms *get_kms(struct drm_encoder *encoder)
+{
+	struct msm_drm_private *priv = encoder->dev->dev_private;
+	return to_mdp4_kms(to_mdp_kms(priv->kms));
+}
+
+#ifdef CONFIG_MSM_BUS_SCALING
+#include <mach/board.h>
+static void bs_init(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder)
+{
+	struct drm_device *dev = mdp4_lcdc_encoder->base.dev;
+	struct lcdc_platform_data *lcdc_pdata = mdp4_find_pdata("lvds.0");
+
+	if (!lcdc_pdata) {
+		dev_err(dev->dev, "could not find lvds pdata\n");
+		return;
+	}
+
+	if (lcdc_pdata->bus_scale_table) {
+		mdp4_lcdc_encoder->bsc = msm_bus_scale_register_client(
+				lcdc_pdata->bus_scale_table);
+		DBG("lvds : bus scale client: %08x", mdp4_lcdc_encoder->bsc);
+	}
+}
+
+static void bs_fini(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder)
+{
+	if (mdp4_lcdc_encoder->bsc) {
+		msm_bus_scale_unregister_client(mdp4_lcdc_encoder->bsc);
+		mdp4_lcdc_encoder->bsc = 0;
+	}
+}
+
+static void bs_set(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder, int idx)
+{
+	if (mdp4_lcdc_encoder->bsc) {
+		DBG("set bus scaling: %d", idx);
+		msm_bus_scale_client_update_request(mdp4_lcdc_encoder->bsc, idx);
+	}
+}
+#else
+static void bs_init(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder) {}
+static void bs_fini(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder) {}
+static void bs_set(struct mdp4_lcdc_encoder *mdp4_lcdc_encoder, int idx) {}
+#endif
+
+static void mdp4_lcdc_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	bs_fini(mdp4_lcdc_encoder);
+	drm_encoder_cleanup(encoder);
+	kfree(mdp4_lcdc_encoder);
+}
+
+static const struct drm_encoder_funcs mdp4_lcdc_encoder_funcs = {
+	.destroy = mdp4_lcdc_encoder_destroy,
+};
+
+/* this should probably be a helper: */
+struct drm_connector *get_connector(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head)
+		if (connector->encoder == encoder)
+			return connector;
+
+	return NULL;
+}
+
+static void setup_phy(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector = get_connector(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	uint32_t lvds_intf = 0, lvds_phy_cfg0 = 0;
+	int bpp, nchan, swap;
+
+	if (!connector)
+		return;
+
+	bpp = 3 * connector->display_info.bpc;
+
+	if (!bpp)
+		bpp = 18;
+
+	/* TODO, these should come from panel somehow: */
+	nchan = 1;
+	swap = 0;
+
+	switch (bpp) {
+	case 24:
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(0),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x08) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x05) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x04) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x03));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(0),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x02) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x01) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x00));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(1),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x11) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x10) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x0d) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0c));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(1),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0b) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x0a) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x09));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(2),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x19) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x18) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x15));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(2),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x14) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x13) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x12));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(3),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1b) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x17) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x16) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0f));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(3),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0e) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x07) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x06));
+		if (nchan == 2) {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		} else {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE3_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		}
+		break;
+
+	case 18:
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(0),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x0a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x07) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x06) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x05));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(0),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x04) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x03) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x02));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(1),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x13) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x12) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x0f) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x0e));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(1),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x0d) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x0c) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x0b));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_3_TO_0(2),
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT0(0x1a) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT1(0x19) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT2(0x18) |
+				MDP4_LCDC_LVDS_MUX_CTL_3_TO_0_BIT3(0x17));
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_MUX_CTL_6_TO_4(2),
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT4(0x16) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT5(0x15) |
+				MDP4_LCDC_LVDS_MUX_CTL_6_TO_4_BIT6(0x14));
+		if (nchan == 2) {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH2_DATA_LANE0_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		} else {
+			lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE2_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE1_EN |
+					MDP4_LCDC_LVDS_INTF_CTL_CH1_DATA_LANE0_EN;
+		}
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_RGB_OUT;
+		break;
+
+	default:
+		dev_err(dev->dev, "unknown bpp: %d\n", bpp);
+		return;
+	}
+
+	switch (nchan) {
+	case 1:
+		lvds_phy_cfg0 = MDP4_LVDS_PHY_CFG0_CHANNEL0;
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN |
+				MDP4_LCDC_LVDS_INTF_CTL_MODE_SEL;
+		break;
+	case 2:
+		lvds_phy_cfg0 = MDP4_LVDS_PHY_CFG0_CHANNEL0 |
+				MDP4_LVDS_PHY_CFG0_CHANNEL1;
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH2_CLK_LANE_EN |
+				MDP4_LCDC_LVDS_INTF_CTL_CH1_CLK_LANE_EN;
+		break;
+	default:
+		dev_err(dev->dev, "unknown # of channels: %d\n", nchan);
+		return;
+	}
+
+	if (swap)
+		lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_CH_SWAP;
+
+	lvds_intf |= MDP4_LCDC_LVDS_INTF_CTL_ENABLE;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, lvds_phy_cfg0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_INTF_CTL, lvds_intf);
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG2, 0x30);
+
+	mb();
+	udelay(1);
+	lvds_phy_cfg0 |= MDP4_LVDS_PHY_CFG0_SERIALIZATION_ENBLE;
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, lvds_phy_cfg0);
+}
+
+static void mdp4_lcdc_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	struct drm_panel *panel = mdp4_lcdc_encoder->panel;
+	bool enabled = (mode == DRM_MODE_DPMS_ON);
+	int i, ret;
+
+	DBG("mode=%d", mode);
+
+	if (enabled == mdp4_lcdc_encoder->enabled)
+		return;
+
+	if (enabled) {
+		unsigned long pc = mdp4_lcdc_encoder->pixclock;
+		int ret;
+
+		bs_set(mdp4_lcdc_encoder, 1);
+
+		for (i = 0; i < ARRAY_SIZE(mdp4_lcdc_encoder->regs); i++) {
+			ret = regulator_enable(mdp4_lcdc_encoder->regs[i]);
+			if (ret)
+				dev_err(dev->dev, "failed to enable regulator: %d\n", ret);
+		}
+
+		DBG("setting lcdc_clk=%lu", pc);
+		ret = clk_set_rate(mdp4_lcdc_encoder->lcdc_clk, pc);
+		if (ret)
+			dev_err(dev->dev, "failed to configure lcdc_clk: %d\n", ret);
+		ret = clk_prepare_enable(mdp4_lcdc_encoder->lcdc_clk);
+		if (ret)
+			dev_err(dev->dev, "failed to enable lcdc_clk: %d\n", ret);
+
+		if (panel)
+			drm_panel_enable(panel);
+
+		setup_phy(encoder);
+
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 1);
+	} else {
+		mdp4_write(mdp4_kms, REG_MDP4_LCDC_ENABLE, 0);
+
+		if (panel)
+			drm_panel_disable(panel);
+
+		/*
+		 * Wait for a vsync so we know the ENABLE=0 latched before
+		 * the (connector) source of the vsync's gets disabled,
+		 * otherwise we end up in a funny state if we re-enable
+		 * before the disable latches, which results that some of
+		 * the settings changes for the new modeset (like new
+		 * scanout buffer) don't latch properly..
+		 */
+		mdp_irq_wait(&mdp4_kms->base, MDP4_IRQ_PRIMARY_VSYNC);
+
+		clk_disable_unprepare(mdp4_lcdc_encoder->lcdc_clk);
+
+		for (i = 0; i < ARRAY_SIZE(mdp4_lcdc_encoder->regs); i++) {
+			ret = regulator_disable(mdp4_lcdc_encoder->regs[i]);
+			if (ret)
+				dev_err(dev->dev, "failed to disable regulator: %d\n", ret);
+		}
+
+		bs_set(mdp4_lcdc_encoder, 0);
+	}
+
+	mdp4_lcdc_encoder->enabled = enabled;
+}
+
+static bool mdp4_lcdc_encoder_mode_fixup(struct drm_encoder *encoder,
+		const struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void mdp4_lcdc_encoder_mode_set(struct drm_encoder *encoder,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	struct mdp4_kms *mdp4_kms = get_kms(encoder);
+	uint32_t lcdc_hsync_skew, vsync_period, vsync_len, ctrl_pol;
+	uint32_t display_v_start, display_v_end;
+	uint32_t hsync_start_x, hsync_end_x;
+
+	mode = adjusted_mode;
+
+	DBG("set mode: %d:\"%s\" %d %d %d %d %d %d %d %d %d %d 0x%x 0x%x",
+			mode->base.id, mode->name,
+			mode->vrefresh, mode->clock,
+			mode->hdisplay, mode->hsync_start,
+			mode->hsync_end, mode->htotal,
+			mode->vdisplay, mode->vsync_start,
+			mode->vsync_end, mode->vtotal,
+			mode->type, mode->flags);
+
+	mdp4_lcdc_encoder->pixclock = mode->clock * 1000;
+
+	DBG("pixclock=%lu", mdp4_lcdc_encoder->pixclock);
+
+	ctrl_pol = 0;
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		ctrl_pol |= MDP4_LCDC_CTRL_POLARITY_HSYNC_LOW;
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		ctrl_pol |= MDP4_LCDC_CTRL_POLARITY_VSYNC_LOW;
+	/* probably need to get DATA_EN polarity from panel.. */
+
+	lcdc_hsync_skew = 0;  /* get this from panel? */
+
+	hsync_start_x = (mode->htotal - mode->hsync_start);
+	hsync_end_x = mode->htotal - (mode->hsync_start - mode->hdisplay) - 1;
+
+	vsync_period = mode->vtotal * mode->htotal;
+	vsync_len = (mode->vsync_end - mode->vsync_start) * mode->htotal;
+	display_v_start = (mode->vtotal - mode->vsync_start) * mode->htotal + lcdc_hsync_skew;
+	display_v_end = vsync_period - ((mode->vsync_start - mode->vdisplay) * mode->htotal) + lcdc_hsync_skew - 1;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_HSYNC_CTRL,
+			MDP4_LCDC_HSYNC_CTRL_PULSEW(mode->hsync_end - mode->hsync_start) |
+			MDP4_LCDC_HSYNC_CTRL_PERIOD(mode->htotal));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_VSYNC_PERIOD, vsync_period);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_VSYNC_LEN, vsync_len);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_HCTRL,
+			MDP4_LCDC_DISPLAY_HCTRL_START(hsync_start_x) |
+			MDP4_LCDC_DISPLAY_HCTRL_END(hsync_end_x));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_VSTART, display_v_start);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_DISPLAY_VEND, display_v_end);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_BORDER_CLR, 0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_UNDERFLOW_CLR,
+			MDP4_LCDC_UNDERFLOW_CLR_ENABLE_RECOVERY |
+			MDP4_LCDC_UNDERFLOW_CLR_COLOR(0xff));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_HSYNC_SKEW, lcdc_hsync_skew);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_CTRL_POLARITY, ctrl_pol);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_HCTL,
+			MDP4_LCDC_ACTIVE_HCTL_START(0) |
+			MDP4_LCDC_ACTIVE_HCTL_END(0));
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_VSTART, 0);
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_ACTIVE_VEND, 0);
+}
+
+static void mdp4_lcdc_encoder_prepare(struct drm_encoder *encoder)
+{
+	mdp4_lcdc_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+}
+
+static void mdp4_lcdc_encoder_commit(struct drm_encoder *encoder)
+{
+	/* TODO: hard-coded for 18bpp: */
+	mdp4_crtc_set_config(encoder->crtc,
+			MDP4_DMA_CONFIG_R_BPC(BPC6) |
+			MDP4_DMA_CONFIG_G_BPC(BPC6) |
+			MDP4_DMA_CONFIG_B_BPC(BPC6) |
+			MDP4_DMA_CONFIG_PACK_ALIGN_MSB |
+			MDP4_DMA_CONFIG_PACK(0x21) |
+			MDP4_DMA_CONFIG_DEFLKR_EN |
+			MDP4_DMA_CONFIG_DITHER_EN);
+	mdp4_crtc_set_intf(encoder->crtc, INTF_LCDC_DTV, 0);
+	mdp4_lcdc_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+}
+
+static const struct drm_encoder_helper_funcs mdp4_lcdc_encoder_helper_funcs = {
+	.dpms = mdp4_lcdc_encoder_dpms,
+	.mode_fixup = mdp4_lcdc_encoder_mode_fixup,
+	.mode_set = mdp4_lcdc_encoder_mode_set,
+	.prepare = mdp4_lcdc_encoder_prepare,
+	.commit = mdp4_lcdc_encoder_commit,
+};
+
+long mdp4_lcdc_round_pixclk(struct drm_encoder *encoder, unsigned long rate)
+{
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder =
+			to_mdp4_lcdc_encoder(encoder);
+	return clk_round_rate(mdp4_lcdc_encoder->lcdc_clk, rate);
+}
+
+/* initialize encoder */
+struct drm_encoder *mdp4_lcdc_encoder_init(struct drm_device *dev,
+		struct drm_panel *panel)
+{
+	struct drm_encoder *encoder = NULL;
+	struct mdp4_lcdc_encoder *mdp4_lcdc_encoder;
+	struct regulator *reg;
+	int ret;
+
+	mdp4_lcdc_encoder = kzalloc(sizeof(*mdp4_lcdc_encoder), GFP_KERNEL);
+	if (!mdp4_lcdc_encoder) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	mdp4_lcdc_encoder->panel = panel;
+
+	encoder = &mdp4_lcdc_encoder->base;
+
+	drm_encoder_init(dev, encoder, &mdp4_lcdc_encoder_funcs,
+			 DRM_MODE_ENCODER_LVDS);
+	drm_encoder_helper_add(encoder, &mdp4_lcdc_encoder_helper_funcs);
+
+	/* TODO: do we need different pll in other cases? */
+	mdp4_lcdc_encoder->lcdc_clk = mpd4_lvds_pll_init(dev);
+	if (IS_ERR(mdp4_lcdc_encoder->lcdc_clk)) {
+		dev_err(dev->dev, "failed to get lvds_clk\n");
+		ret = PTR_ERR(mdp4_lcdc_encoder->lcdc_clk);
+		goto fail;
+	}
+
+	/* TODO: different regulators in other cases? */
+	reg = devm_regulator_get(dev->dev, "lvds-vccs-3p3v");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-vccs-3p3v: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[0] = reg;
+
+	reg = devm_regulator_get(dev->dev, "lvds-pll-vdda");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-pll-vdda: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[1] = reg;
+
+	reg = devm_regulator_get(dev->dev, "lvds-vdda");
+	if (IS_ERR(reg)) {
+		ret = PTR_ERR(reg);
+		dev_err(dev->dev, "failed to get lvds-vdda: %d\n", ret);
+		goto fail;
+	}
+	mdp4_lcdc_encoder->regs[2] = reg;
+
+	bs_init(mdp4_lcdc_encoder);
+
+	return encoder;
+
+fail:
+	if (encoder)
+		mdp4_lcdc_encoder_destroy(encoder);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c
new file mode 100644
index 000000000000..310034688c15
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_connector.c
@@ -0,0 +1,151 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ * Author: Vinay Simha <vinaysimha@inforcecomputing.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/gpio.h>
+
+#include "mdp4_kms.h"
+
+struct mdp4_lvds_connector {
+	struct drm_connector base;
+	struct drm_encoder *encoder;
+	struct drm_panel *panel;
+};
+#define to_mdp4_lvds_connector(x) container_of(x, struct mdp4_lvds_connector, base)
+
+static enum drm_connector_status mdp4_lvds_connector_detect(
+		struct drm_connector *connector, bool force)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+
+	return mdp4_lvds_connector->panel ?
+			connector_status_connected :
+			connector_status_disconnected;
+}
+
+static void mdp4_lvds_connector_destroy(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_panel *panel = mdp4_lvds_connector->panel;
+
+	if (panel)
+		drm_panel_detach(panel);
+
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+
+	kfree(mdp4_lvds_connector);
+}
+
+static int mdp4_lvds_connector_get_modes(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_panel *panel = mdp4_lvds_connector->panel;
+	int ret = 0;
+
+	if (panel)
+		ret = panel->funcs->get_modes(panel);
+
+	return ret;
+}
+
+static int mdp4_lvds_connector_mode_valid(struct drm_connector *connector,
+				 struct drm_display_mode *mode)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	struct drm_encoder *encoder = mdp4_lvds_connector->encoder;
+	long actual, requested;
+
+	requested = 1000 * mode->clock;
+	actual = mdp4_lcdc_round_pixclk(encoder, requested);
+
+	DBG("requested=%ld, actual=%ld", requested, actual);
+
+	if (actual != requested)
+		return MODE_CLOCK_RANGE;
+
+	return MODE_OK;
+}
+
+static struct drm_encoder *
+mdp4_lvds_connector_best_encoder(struct drm_connector *connector)
+{
+	struct mdp4_lvds_connector *mdp4_lvds_connector =
+			to_mdp4_lvds_connector(connector);
+	return mdp4_lvds_connector->encoder;
+}
+
+static const struct drm_connector_funcs mdp4_lvds_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = mdp4_lvds_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = mdp4_lvds_connector_destroy,
+};
+
+static const struct drm_connector_helper_funcs mdp4_lvds_connector_helper_funcs = {
+	.get_modes = mdp4_lvds_connector_get_modes,
+	.mode_valid = mdp4_lvds_connector_mode_valid,
+	.best_encoder = mdp4_lvds_connector_best_encoder,
+};
+
+/* initialize connector */
+struct drm_connector *mdp4_lvds_connector_init(struct drm_device *dev,
+		struct drm_panel *panel, struct drm_encoder *encoder)
+{
+	struct drm_connector *connector = NULL;
+	struct mdp4_lvds_connector *mdp4_lvds_connector;
+	int ret;
+
+	mdp4_lvds_connector = kzalloc(sizeof(*mdp4_lvds_connector), GFP_KERNEL);
+	if (!mdp4_lvds_connector) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	mdp4_lvds_connector->encoder = encoder;
+	mdp4_lvds_connector->panel = panel;
+
+	connector = &mdp4_lvds_connector->base;
+
+	drm_connector_init(dev, connector, &mdp4_lvds_connector_funcs,
+			DRM_MODE_CONNECTOR_LVDS);
+	drm_connector_helper_add(connector, &mdp4_lvds_connector_helper_funcs);
+
+	connector->polled = 0;
+
+	connector->interlace_allowed = 0;
+	connector->doublescan_allowed = 0;
+
+	drm_connector_register(connector);
+
+	drm_mode_connector_attach_encoder(connector, encoder);
+
+	if (panel)
+		drm_panel_attach(panel, connector);
+
+	return connector;
+
+fail:
+	if (connector)
+		mdp4_lvds_connector_destroy(connector);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c
new file mode 100644
index 000000000000..ce4245971673
--- /dev/null
+++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lvds_pll.c
@@ -0,0 +1,172 @@
+/*
+ * Copyright (C) 2014 Red Hat
+ * Author: Rob Clark <robdclark@gmail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+
+#include "mdp4_kms.h"
+
+struct mdp4_lvds_pll {
+	struct clk_hw pll_hw;
+	struct drm_device *dev;
+	unsigned long pixclk;
+};
+#define to_mdp4_lvds_pll(x) container_of(x, struct mdp4_lvds_pll, pll_hw)
+
+static struct mdp4_kms *get_kms(struct mdp4_lvds_pll *lvds_pll)
+{
+	struct msm_drm_private *priv = lvds_pll->dev->dev_private;
+	return to_mdp4_kms(to_mdp_kms(priv->kms));
+}
+
+struct pll_rate {
+	unsigned long rate;
+	struct {
+		uint32_t val;
+		uint32_t reg;
+	} conf[32];
+};
+
+/* NOTE: keep sorted highest freq to lowest: */
+static const struct pll_rate freqtbl[] = {
+	{ 72000000, {
+		{ 0x8f, REG_MDP4_LVDS_PHY_PLL_CTRL_1 },
+		{ 0x30, REG_MDP4_LVDS_PHY_PLL_CTRL_2 },
+		{ 0xc6, REG_MDP4_LVDS_PHY_PLL_CTRL_3 },
+		{ 0x10, REG_MDP4_LVDS_PHY_PLL_CTRL_5 },
+		{ 0x07, REG_MDP4_LVDS_PHY_PLL_CTRL_6 },
+		{ 0x62, REG_MDP4_LVDS_PHY_PLL_CTRL_7 },
+		{ 0x41, REG_MDP4_LVDS_PHY_PLL_CTRL_8 },
+		{ 0x0d, REG_MDP4_LVDS_PHY_PLL_CTRL_9 },
+		{ 0, 0 } }
+	},
+};
+
+static const struct pll_rate *find_rate(unsigned long rate)
+{
+	int i;
+	for (i = 1; i < ARRAY_SIZE(freqtbl); i++)
+		if (rate > freqtbl[i].rate)
+			return &freqtbl[i-1];
+	return &freqtbl[i-1];
+}
+
+static int mpd4_lvds_pll_enable(struct clk_hw *hw)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	struct mdp4_kms *mdp4_kms = get_kms(lvds_pll);
+	const struct pll_rate *pll_rate = find_rate(lvds_pll->pixclk);
+	int i;
+
+	DBG("pixclk=%lu (%lu)", lvds_pll->pixclk, pll_rate->rate);
+
+	if (WARN_ON(!pll_rate))
+		return -EINVAL;
+
+	mdp4_write(mdp4_kms, REG_MDP4_LCDC_LVDS_PHY_RESET, 0x33);
+
+	for (i = 0; pll_rate->conf[i].reg; i++)
+		mdp4_write(mdp4_kms, pll_rate->conf[i].reg, pll_rate->conf[i].val);
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_CTRL_0, 0x01);
+
+	/* Wait until LVDS PLL is locked and ready */
+	while (!mdp4_read(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_LOCKED))
+		cpu_relax();
+
+	return 0;
+}
+
+static void mpd4_lvds_pll_disable(struct clk_hw *hw)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	struct mdp4_kms *mdp4_kms = get_kms(lvds_pll);
+
+	DBG("");
+
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, 0x0);
+	mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_PLL_CTRL_0, 0x0);
+}
+
+static unsigned long mpd4_lvds_pll_recalc_rate(struct clk_hw *hw,
+				unsigned long parent_rate)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	return lvds_pll->pixclk;
+}
+
+static long mpd4_lvds_pll_round_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long *parent_rate)
+{
+	const struct pll_rate *pll_rate = find_rate(rate);
+	return pll_rate->rate;
+}
+
+static int mpd4_lvds_pll_set_rate(struct clk_hw *hw, unsigned long rate,
+		unsigned long parent_rate)
+{
+	struct mdp4_lvds_pll *lvds_pll = to_mdp4_lvds_pll(hw);
+	lvds_pll->pixclk = rate;
+	return 0;
+}
+
+
+static const struct clk_ops mpd4_lvds_pll_ops = {
+	.enable = mpd4_lvds_pll_enable,
+	.disable = mpd4_lvds_pll_disable,
+	.recalc_rate = mpd4_lvds_pll_recalc_rate,
+	.round_rate = mpd4_lvds_pll_round_rate,
+	.set_rate = mpd4_lvds_pll_set_rate,
+};
+
+static const char *mpd4_lvds_pll_parents[] = {
+	"pxo",
+};
+
+static struct clk_init_data pll_init = {
+	.name = "mpd4_lvds_pll",
+	.ops = &mpd4_lvds_pll_ops,
+	.parent_names = mpd4_lvds_pll_parents,
+	.num_parents = ARRAY_SIZE(mpd4_lvds_pll_parents),
+};
+
+struct clk *mpd4_lvds_pll_init(struct drm_device *dev)
+{
+	struct mdp4_lvds_pll *lvds_pll;
+	struct clk *clk;
+	int ret;
+
+	lvds_pll = devm_kzalloc(dev->dev, sizeof(*lvds_pll), GFP_KERNEL);
+	if (!lvds_pll) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	lvds_pll->dev = dev;
+
+	lvds_pll->pll_hw.init = &pll_init;
+	clk = devm_clk_register(dev->dev, &lvds_pll->pll_hw);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto fail;
+	}
+
+	return clk;
+
+fail:
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
index 0aa51517f826..67f4f896ba8c 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h
@@ -12,14 +12,14 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
 - /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
 
-Copyright (C) 2013 by the following authors:
+Copyright (C) 2013-2014 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
 
 Permission is hereby granted, free of charge, to any person obtaining
@@ -68,6 +68,8 @@ enum mdp5_pipe {
 	SSPP_RGB2 = 5,
 	SSPP_DMA0 = 6,
 	SSPP_DMA1 = 7,
+	SSPP_VIG3 = 8,
+	SSPP_RGB3 = 9,
 };
 
 enum mdp5_ctl_mode {
@@ -126,7 +128,11 @@ enum mdp5_client_id {
 	CID_RGB0 = 16,
 	CID_RGB1 = 17,
 	CID_RGB2 = 18,
-	CID_MAX = 19,
+	CID_VIG3_Y = 19,
+	CID_VIG3_CR = 20,
+	CID_VIG3_CB = 21,
+	CID_RGB3 = 22,
+	CID_MAX = 23,
 };
 
 enum mdp5_igc_type {
@@ -299,11 +305,34 @@ static inline uint32_t MDP5_IGC_LUT_REG_VAL(uint32_t val)
 #define MDP5_IGC_LUT_REG_DISABLE_PIPE_1				0x20000000
 #define MDP5_IGC_LUT_REG_DISABLE_PIPE_2				0x40000000
 
-static inline uint32_t REG_MDP5_CTL(uint32_t i0) { return 0x00000600 + 0x100*i0; }
+static inline uint32_t __offset_CTL(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return (mdp5_cfg->ctl.base[0]);
+		case 1: return (mdp5_cfg->ctl.base[1]);
+		case 2: return (mdp5_cfg->ctl.base[2]);
+		case 3: return (mdp5_cfg->ctl.base[3]);
+		case 4: return (mdp5_cfg->ctl.base[4]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_CTL(uint32_t i0) { return 0x00000000 + __offset_CTL(i0); }
 
-static inline uint32_t REG_MDP5_CTL_LAYER(uint32_t i0, uint32_t i1) { return 0x00000600 + 0x100*i0 + 0x4*i1; }
+static inline uint32_t __offset_LAYER(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return 0x00000000;
+		case 1: return 0x00000004;
+		case 2: return 0x00000008;
+		case 3: return 0x0000000c;
+		case 4: return 0x00000010;
+		case 5: return 0x00000024;
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_CTL_LAYER(uint32_t i0, uint32_t i1) { return 0x00000000 + __offset_CTL(i0) + __offset_LAYER(i1); }
 
-static inline uint32_t REG_MDP5_CTL_LAYER_REG(uint32_t i0, uint32_t i1) { return 0x00000600 + 0x100*i0 + 0x4*i1; }
+static inline uint32_t REG_MDP5_CTL_LAYER_REG(uint32_t i0, uint32_t i1) { return 0x00000000 + __offset_CTL(i0) + __offset_LAYER(i1); }
 #define MDP5_CTL_LAYER_REG_VIG0__MASK				0x00000007
 #define MDP5_CTL_LAYER_REG_VIG0__SHIFT				0
 static inline uint32_t MDP5_CTL_LAYER_REG_VIG0(enum mdp_mixer_stage_id val)
@@ -354,8 +383,20 @@ static inline uint32_t MDP5_CTL_LAYER_REG_DMA1(enum mdp_mixer_stage_id val)
 }
 #define MDP5_CTL_LAYER_REG_BORDER_COLOR				0x01000000
 #define MDP5_CTL_LAYER_REG_CURSOR_OUT				0x02000000
+#define MDP5_CTL_LAYER_REG_VIG3__MASK				0x1c000000
+#define MDP5_CTL_LAYER_REG_VIG3__SHIFT				26
+static inline uint32_t MDP5_CTL_LAYER_REG_VIG3(enum mdp_mixer_stage_id val)
+{
+	return ((val) << MDP5_CTL_LAYER_REG_VIG3__SHIFT) & MDP5_CTL_LAYER_REG_VIG3__MASK;
+}
+#define MDP5_CTL_LAYER_REG_RGB3__MASK				0xe0000000
+#define MDP5_CTL_LAYER_REG_RGB3__SHIFT				29
+static inline uint32_t MDP5_CTL_LAYER_REG_RGB3(enum mdp_mixer_stage_id val)
+{
+	return ((val) << MDP5_CTL_LAYER_REG_RGB3__SHIFT) & MDP5_CTL_LAYER_REG_RGB3__MASK;
+}
 
-static inline uint32_t REG_MDP5_CTL_OP(uint32_t i0) { return 0x00000614 + 0x100*i0; }
+static inline uint32_t REG_MDP5_CTL_OP(uint32_t i0) { return 0x00000014 + __offset_CTL(i0); }
 #define MDP5_CTL_OP_MODE__MASK					0x0000000f
 #define MDP5_CTL_OP_MODE__SHIFT					0
 static inline uint32_t MDP5_CTL_OP_MODE(enum mdp5_ctl_mode val)
@@ -377,7 +418,7 @@ static inline uint32_t MDP5_CTL_OP_PACK_3D(enum mdp5_pack_3d val)
 	return ((val) << MDP5_CTL_OP_PACK_3D__SHIFT) & MDP5_CTL_OP_PACK_3D__MASK;
 }
 
-static inline uint32_t REG_MDP5_CTL_FLUSH(uint32_t i0) { return 0x00000618 + 0x100*i0; }
+static inline uint32_t REG_MDP5_CTL_FLUSH(uint32_t i0) { return 0x00000018 + __offset_CTL(i0); }
 #define MDP5_CTL_FLUSH_VIG0					0x00000001
 #define MDP5_CTL_FLUSH_VIG1					0x00000002
 #define MDP5_CTL_FLUSH_VIG2					0x00000004
@@ -387,26 +428,48 @@ static inline uint32_t REG_MDP5_CTL_FLUSH(uint32_t i0) { return 0x00000618 + 0x1
 #define MDP5_CTL_FLUSH_LM0					0x00000040
 #define MDP5_CTL_FLUSH_LM1					0x00000080
 #define MDP5_CTL_FLUSH_LM2					0x00000100
+#define MDP5_CTL_FLUSH_LM3					0x00000200
+#define MDP5_CTL_FLUSH_LM4					0x00000400
 #define MDP5_CTL_FLUSH_DMA0					0x00000800
 #define MDP5_CTL_FLUSH_DMA1					0x00001000
 #define MDP5_CTL_FLUSH_DSPP0					0x00002000
 #define MDP5_CTL_FLUSH_DSPP1					0x00004000
 #define MDP5_CTL_FLUSH_DSPP2					0x00008000
 #define MDP5_CTL_FLUSH_CTL					0x00020000
+#define MDP5_CTL_FLUSH_VIG3					0x00040000
+#define MDP5_CTL_FLUSH_RGB3					0x00080000
+#define MDP5_CTL_FLUSH_LM5					0x00100000
+#define MDP5_CTL_FLUSH_DSPP3					0x00200000
 
-static inline uint32_t REG_MDP5_CTL_START(uint32_t i0) { return 0x0000061c + 0x100*i0; }
+static inline uint32_t REG_MDP5_CTL_START(uint32_t i0) { return 0x0000001c + __offset_CTL(i0); }
 
-static inline uint32_t REG_MDP5_CTL_PACK_3D(uint32_t i0) { return 0x00000620 + 0x100*i0; }
+static inline uint32_t REG_MDP5_CTL_PACK_3D(uint32_t i0) { return 0x00000020 + __offset_CTL(i0); }
 
-static inline uint32_t REG_MDP5_PIPE(enum mdp5_pipe i0) { return 0x00001200 + 0x400*i0; }
+static inline uint32_t __offset_PIPE(enum mdp5_pipe idx)
+{
+	switch (idx) {
+		case SSPP_VIG0: return (mdp5_cfg->pipe_vig.base[0]);
+		case SSPP_VIG1: return (mdp5_cfg->pipe_vig.base[1]);
+		case SSPP_VIG2: return (mdp5_cfg->pipe_vig.base[2]);
+		case SSPP_RGB0: return (mdp5_cfg->pipe_rgb.base[0]);
+		case SSPP_RGB1: return (mdp5_cfg->pipe_rgb.base[1]);
+		case SSPP_RGB2: return (mdp5_cfg->pipe_rgb.base[2]);
+		case SSPP_DMA0: return (mdp5_cfg->pipe_dma.base[0]);
+		case SSPP_DMA1: return (mdp5_cfg->pipe_dma.base[1]);
+		case SSPP_VIG3: return (mdp5_cfg->pipe_vig.base[3]);
+		case SSPP_RGB3: return (mdp5_cfg->pipe_rgb.base[3]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_PIPE(enum mdp5_pipe i0) { return 0x00000000 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_HIST_CTL_BASE(enum mdp5_pipe i0) { return 0x000014c4 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_HIST_CTL_BASE(enum mdp5_pipe i0) { return 0x000002c4 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_HIST_LUT_BASE(enum mdp5_pipe i0) { return 0x000014f0 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_HIST_LUT_BASE(enum mdp5_pipe i0) { return 0x000002f0 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_HIST_LUT_SWAP(enum mdp5_pipe i0) { return 0x00001500 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_HIST_LUT_SWAP(enum mdp5_pipe i0) { return 0x00000300 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_SIZE(enum mdp5_pipe i0) { return 0x00001200 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_SIZE(enum mdp5_pipe i0) { return 0x00000000 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_SIZE_HEIGHT__MASK				0xffff0000
 #define MDP5_PIPE_SRC_SIZE_HEIGHT__SHIFT			16
 static inline uint32_t MDP5_PIPE_SRC_SIZE_HEIGHT(uint32_t val)
@@ -420,7 +483,7 @@ static inline uint32_t MDP5_PIPE_SRC_SIZE_WIDTH(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_SIZE_WIDTH__SHIFT) & MDP5_PIPE_SRC_SIZE_WIDTH__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_IMG_SIZE(enum mdp5_pipe i0) { return 0x00001204 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_IMG_SIZE(enum mdp5_pipe i0) { return 0x00000004 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_IMG_SIZE_HEIGHT__MASK			0xffff0000
 #define MDP5_PIPE_SRC_IMG_SIZE_HEIGHT__SHIFT			16
 static inline uint32_t MDP5_PIPE_SRC_IMG_SIZE_HEIGHT(uint32_t val)
@@ -434,7 +497,7 @@ static inline uint32_t MDP5_PIPE_SRC_IMG_SIZE_WIDTH(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_IMG_SIZE_WIDTH__SHIFT) & MDP5_PIPE_SRC_IMG_SIZE_WIDTH__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_XY(enum mdp5_pipe i0) { return 0x00001208 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_XY(enum mdp5_pipe i0) { return 0x00000008 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_XY_Y__MASK				0xffff0000
 #define MDP5_PIPE_SRC_XY_Y__SHIFT				16
 static inline uint32_t MDP5_PIPE_SRC_XY_Y(uint32_t val)
@@ -448,7 +511,7 @@ static inline uint32_t MDP5_PIPE_SRC_XY_X(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_XY_X__SHIFT) & MDP5_PIPE_SRC_XY_X__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_OUT_SIZE(enum mdp5_pipe i0) { return 0x0000120c + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_OUT_SIZE(enum mdp5_pipe i0) { return 0x0000000c + __offset_PIPE(i0); }
 #define MDP5_PIPE_OUT_SIZE_HEIGHT__MASK				0xffff0000
 #define MDP5_PIPE_OUT_SIZE_HEIGHT__SHIFT			16
 static inline uint32_t MDP5_PIPE_OUT_SIZE_HEIGHT(uint32_t val)
@@ -462,7 +525,7 @@ static inline uint32_t MDP5_PIPE_OUT_SIZE_WIDTH(uint32_t val)
 	return ((val) << MDP5_PIPE_OUT_SIZE_WIDTH__SHIFT) & MDP5_PIPE_OUT_SIZE_WIDTH__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_OUT_XY(enum mdp5_pipe i0) { return 0x00001210 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_OUT_XY(enum mdp5_pipe i0) { return 0x00000010 + __offset_PIPE(i0); }
 #define MDP5_PIPE_OUT_XY_Y__MASK				0xffff0000
 #define MDP5_PIPE_OUT_XY_Y__SHIFT				16
 static inline uint32_t MDP5_PIPE_OUT_XY_Y(uint32_t val)
@@ -476,15 +539,15 @@ static inline uint32_t MDP5_PIPE_OUT_XY_X(uint32_t val)
 	return ((val) << MDP5_PIPE_OUT_XY_X__SHIFT) & MDP5_PIPE_OUT_XY_X__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC0_ADDR(enum mdp5_pipe i0) { return 0x00001214 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC0_ADDR(enum mdp5_pipe i0) { return 0x00000014 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC1_ADDR(enum mdp5_pipe i0) { return 0x00001218 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC1_ADDR(enum mdp5_pipe i0) { return 0x00000018 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC2_ADDR(enum mdp5_pipe i0) { return 0x0000121c + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC2_ADDR(enum mdp5_pipe i0) { return 0x0000001c + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC3_ADDR(enum mdp5_pipe i0) { return 0x00001220 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC3_ADDR(enum mdp5_pipe i0) { return 0x00000020 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_STRIDE_A(enum mdp5_pipe i0) { return 0x00001224 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_STRIDE_A(enum mdp5_pipe i0) { return 0x00000024 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_STRIDE_A_P0__MASK				0x0000ffff
 #define MDP5_PIPE_SRC_STRIDE_A_P0__SHIFT			0
 static inline uint32_t MDP5_PIPE_SRC_STRIDE_A_P0(uint32_t val)
@@ -498,7 +561,7 @@ static inline uint32_t MDP5_PIPE_SRC_STRIDE_A_P1(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_STRIDE_A_P1__SHIFT) & MDP5_PIPE_SRC_STRIDE_A_P1__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_STRIDE_B(enum mdp5_pipe i0) { return 0x00001228 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_STRIDE_B(enum mdp5_pipe i0) { return 0x00000028 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_STRIDE_B_P2__MASK				0x0000ffff
 #define MDP5_PIPE_SRC_STRIDE_B_P2__SHIFT			0
 static inline uint32_t MDP5_PIPE_SRC_STRIDE_B_P2(uint32_t val)
@@ -512,9 +575,9 @@ static inline uint32_t MDP5_PIPE_SRC_STRIDE_B_P3(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_STRIDE_B_P3__SHIFT) & MDP5_PIPE_SRC_STRIDE_B_P3__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_STILE_FRAME_SIZE(enum mdp5_pipe i0) { return 0x0000122c + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_STILE_FRAME_SIZE(enum mdp5_pipe i0) { return 0x0000002c + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_FORMAT(enum mdp5_pipe i0) { return 0x00001230 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_FORMAT(enum mdp5_pipe i0) { return 0x00000030 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_FORMAT_G_BPC__MASK			0x00000003
 #define MDP5_PIPE_SRC_FORMAT_G_BPC__SHIFT			0
 static inline uint32_t MDP5_PIPE_SRC_FORMAT_G_BPC(enum mdp_bpc val)
@@ -568,7 +631,7 @@ static inline uint32_t MDP5_PIPE_SRC_FORMAT_CHROMA_SAMP(enum mdp5_chroma_samp_ty
 	return ((val) << MDP5_PIPE_SRC_FORMAT_CHROMA_SAMP__SHIFT) & MDP5_PIPE_SRC_FORMAT_CHROMA_SAMP__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_UNPACK(enum mdp5_pipe i0) { return 0x00001234 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_UNPACK(enum mdp5_pipe i0) { return 0x00000034 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_UNPACK_ELEM0__MASK			0x000000ff
 #define MDP5_PIPE_SRC_UNPACK_ELEM0__SHIFT			0
 static inline uint32_t MDP5_PIPE_SRC_UNPACK_ELEM0(uint32_t val)
@@ -594,7 +657,7 @@ static inline uint32_t MDP5_PIPE_SRC_UNPACK_ELEM3(uint32_t val)
 	return ((val) << MDP5_PIPE_SRC_UNPACK_ELEM3__SHIFT) & MDP5_PIPE_SRC_UNPACK_ELEM3__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_OP_MODE(enum mdp5_pipe i0) { return 0x00001238 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_OP_MODE(enum mdp5_pipe i0) { return 0x00000038 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SRC_OP_MODE_BWC_EN				0x00000001
 #define MDP5_PIPE_SRC_OP_MODE_BWC__MASK				0x00000006
 #define MDP5_PIPE_SRC_OP_MODE_BWC__SHIFT			1
@@ -610,29 +673,29 @@ static inline uint32_t MDP5_PIPE_SRC_OP_MODE_BWC(enum mdp5_pipe_bwc val)
 #define MDP5_PIPE_SRC_OP_MODE_DEINTERLACE			0x00400000
 #define MDP5_PIPE_SRC_OP_MODE_DEINTERLACE_ODD			0x00800000
 
-static inline uint32_t REG_MDP5_PIPE_SRC_CONSTANT_COLOR(enum mdp5_pipe i0) { return 0x0000123c + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_CONSTANT_COLOR(enum mdp5_pipe i0) { return 0x0000003c + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_FETCH_CONFIG(enum mdp5_pipe i0) { return 0x00001248 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_FETCH_CONFIG(enum mdp5_pipe i0) { return 0x00000048 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_VC1_RANGE(enum mdp5_pipe i0) { return 0x0000124c + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_VC1_RANGE(enum mdp5_pipe i0) { return 0x0000004c + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_0(enum mdp5_pipe i0) { return 0x00001250 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_0(enum mdp5_pipe i0) { return 0x00000050 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_1(enum mdp5_pipe i0) { return 0x00001254 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_1(enum mdp5_pipe i0) { return 0x00000054 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_2(enum mdp5_pipe i0) { return 0x00001258 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_REQPRIO_FIFO_WM_2(enum mdp5_pipe i0) { return 0x00000058 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SRC_ADDR_SW_STATUS(enum mdp5_pipe i0) { return 0x00001270 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SRC_ADDR_SW_STATUS(enum mdp5_pipe i0) { return 0x00000070 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC0_ADDR(enum mdp5_pipe i0) { return 0x000012a4 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC0_ADDR(enum mdp5_pipe i0) { return 0x000000a4 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC1_ADDR(enum mdp5_pipe i0) { return 0x000012a8 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC1_ADDR(enum mdp5_pipe i0) { return 0x000000a8 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC2_ADDR(enum mdp5_pipe i0) { return 0x000012ac + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC2_ADDR(enum mdp5_pipe i0) { return 0x000000ac + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC3_ADDR(enum mdp5_pipe i0) { return 0x000012b0 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_CURRENT_SRC3_ADDR(enum mdp5_pipe i0) { return 0x000000b0 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_DECIMATION(enum mdp5_pipe i0) { return 0x000012b4 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_DECIMATION(enum mdp5_pipe i0) { return 0x000000b4 + __offset_PIPE(i0); }
 #define MDP5_PIPE_DECIMATION_VERT__MASK				0x000000ff
 #define MDP5_PIPE_DECIMATION_VERT__SHIFT			0
 static inline uint32_t MDP5_PIPE_DECIMATION_VERT(uint32_t val)
@@ -646,7 +709,7 @@ static inline uint32_t MDP5_PIPE_DECIMATION_HORZ(uint32_t val)
 	return ((val) << MDP5_PIPE_DECIMATION_HORZ__SHIFT) & MDP5_PIPE_DECIMATION_HORZ__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SCALE_CONFIG(enum mdp5_pipe i0) { return 0x00001404 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SCALE_CONFIG(enum mdp5_pipe i0) { return 0x00000204 + __offset_PIPE(i0); }
 #define MDP5_PIPE_SCALE_CONFIG_SCALEX_EN			0x00000001
 #define MDP5_PIPE_SCALE_CONFIG_SCALEY_EN			0x00000002
 #define MDP5_PIPE_SCALE_CONFIG_SCALEX_MIN_FILTER__MASK		0x00000300
@@ -686,23 +749,34 @@ static inline uint32_t MDP5_PIPE_SCALE_CONFIG_SCALEY_MAX_FILTER(enum mdp5_scale_
 	return ((val) << MDP5_PIPE_SCALE_CONFIG_SCALEY_MAX_FILTER__SHIFT) & MDP5_PIPE_SCALE_CONFIG_SCALEY_MAX_FILTER__MASK;
 }
 
-static inline uint32_t REG_MDP5_PIPE_SCALE_PHASE_STEP_X(enum mdp5_pipe i0) { return 0x00001410 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SCALE_PHASE_STEP_X(enum mdp5_pipe i0) { return 0x00000210 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SCALE_PHASE_STEP_Y(enum mdp5_pipe i0) { return 0x00001414 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SCALE_PHASE_STEP_Y(enum mdp5_pipe i0) { return 0x00000214 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SCALE_INIT_PHASE_X(enum mdp5_pipe i0) { return 0x00001420 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SCALE_INIT_PHASE_X(enum mdp5_pipe i0) { return 0x00000220 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_PIPE_SCALE_INIT_PHASE_Y(enum mdp5_pipe i0) { return 0x00001424 + 0x400*i0; }
+static inline uint32_t REG_MDP5_PIPE_SCALE_INIT_PHASE_Y(enum mdp5_pipe i0) { return 0x00000224 + __offset_PIPE(i0); }
 
-static inline uint32_t REG_MDP5_LM(uint32_t i0) { return 0x00003200 + 0x400*i0; }
+static inline uint32_t __offset_LM(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return (mdp5_cfg->lm.base[0]);
+		case 1: return (mdp5_cfg->lm.base[1]);
+		case 2: return (mdp5_cfg->lm.base[2]);
+		case 3: return (mdp5_cfg->lm.base[3]);
+		case 4: return (mdp5_cfg->lm.base[4]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_LM(uint32_t i0) { return 0x00000000 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_BLEND_COLOR_OUT(uint32_t i0) { return 0x00003200 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_BLEND_COLOR_OUT(uint32_t i0) { return 0x00000000 + __offset_LM(i0); }
 #define MDP5_LM_BLEND_COLOR_OUT_STAGE0_FG_ALPHA			0x00000002
 #define MDP5_LM_BLEND_COLOR_OUT_STAGE1_FG_ALPHA			0x00000004
 #define MDP5_LM_BLEND_COLOR_OUT_STAGE2_FG_ALPHA			0x00000008
 #define MDP5_LM_BLEND_COLOR_OUT_STAGE3_FG_ALPHA			0x00000010
 
-static inline uint32_t REG_MDP5_LM_OUT_SIZE(uint32_t i0) { return 0x00003204 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_OUT_SIZE(uint32_t i0) { return 0x00000004 + __offset_LM(i0); }
 #define MDP5_LM_OUT_SIZE_HEIGHT__MASK				0xffff0000
 #define MDP5_LM_OUT_SIZE_HEIGHT__SHIFT				16
 static inline uint32_t MDP5_LM_OUT_SIZE_HEIGHT(uint32_t val)
@@ -716,13 +790,13 @@ static inline uint32_t MDP5_LM_OUT_SIZE_WIDTH(uint32_t val)
 	return ((val) << MDP5_LM_OUT_SIZE_WIDTH__SHIFT) & MDP5_LM_OUT_SIZE_WIDTH__MASK;
 }
 
-static inline uint32_t REG_MDP5_LM_BORDER_COLOR_0(uint32_t i0) { return 0x00003208 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_BORDER_COLOR_0(uint32_t i0) { return 0x00000008 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_BORDER_COLOR_1(uint32_t i0) { return 0x00003210 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_BORDER_COLOR_1(uint32_t i0) { return 0x00000010 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_BLEND(uint32_t i0, uint32_t i1) { return 0x00003220 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND(uint32_t i0, uint32_t i1) { return 0x00000020 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_OP_MODE(uint32_t i0, uint32_t i1) { return 0x00003220 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_OP_MODE(uint32_t i0, uint32_t i1) { return 0x00000020 + __offset_LM(i0) + 0x30*i1; }
 #define MDP5_LM_BLEND_OP_MODE_FG_ALPHA__MASK			0x00000003
 #define MDP5_LM_BLEND_OP_MODE_FG_ALPHA__SHIFT			0
 static inline uint32_t MDP5_LM_BLEND_OP_MODE_FG_ALPHA(enum mdp_alpha_type val)
@@ -744,57 +818,67 @@ static inline uint32_t MDP5_LM_BLEND_OP_MODE_BG_ALPHA(enum mdp_alpha_type val)
 #define MDP5_LM_BLEND_OP_MODE_BG_INV_MOD_ALPHA			0x00001000
 #define MDP5_LM_BLEND_OP_MODE_BG_TRANSP_EN			0x00002000
 
-static inline uint32_t REG_MDP5_LM_BLEND_FG_ALPHA(uint32_t i0, uint32_t i1) { return 0x00003224 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_FG_ALPHA(uint32_t i0, uint32_t i1) { return 0x00000024 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_BG_ALPHA(uint32_t i0, uint32_t i1) { return 0x00003228 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_BG_ALPHA(uint32_t i0, uint32_t i1) { return 0x00000028 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_LOW0(uint32_t i0, uint32_t i1) { return 0x0000322c + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_LOW0(uint32_t i0, uint32_t i1) { return 0x0000002c + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_LOW1(uint32_t i0, uint32_t i1) { return 0x00003230 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_LOW1(uint32_t i0, uint32_t i1) { return 0x00000030 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_HIGH0(uint32_t i0, uint32_t i1) { return 0x00003234 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_HIGH0(uint32_t i0, uint32_t i1) { return 0x00000034 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_HIGH1(uint32_t i0, uint32_t i1) { return 0x00003238 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_FG_TRANSP_HIGH1(uint32_t i0, uint32_t i1) { return 0x00000038 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_LOW0(uint32_t i0, uint32_t i1) { return 0x0000323c + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_LOW0(uint32_t i0, uint32_t i1) { return 0x0000003c + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_LOW1(uint32_t i0, uint32_t i1) { return 0x00003240 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_LOW1(uint32_t i0, uint32_t i1) { return 0x00000040 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_HIGH0(uint32_t i0, uint32_t i1) { return 0x00003244 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_HIGH0(uint32_t i0, uint32_t i1) { return 0x00000044 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_HIGH1(uint32_t i0, uint32_t i1) { return 0x00003248 + 0x400*i0 + 0x30*i1; }
+static inline uint32_t REG_MDP5_LM_BLEND_BG_TRANSP_HIGH1(uint32_t i0, uint32_t i1) { return 0x00000048 + __offset_LM(i0) + 0x30*i1; }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_IMG_SIZE(uint32_t i0) { return 0x000032e0 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_IMG_SIZE(uint32_t i0) { return 0x000000e0 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_SIZE(uint32_t i0) { return 0x000032e4 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_SIZE(uint32_t i0) { return 0x000000e4 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_XY(uint32_t i0) { return 0x000032e8 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_XY(uint32_t i0) { return 0x000000e8 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_STRIDE(uint32_t i0) { return 0x000032dc + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_STRIDE(uint32_t i0) { return 0x000000dc + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_FORMAT(uint32_t i0) { return 0x000032ec + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_FORMAT(uint32_t i0) { return 0x000000ec + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BASE_ADDR(uint32_t i0) { return 0x000032f0 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BASE_ADDR(uint32_t i0) { return 0x000000f0 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_START_XY(uint32_t i0) { return 0x000032f4 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_START_XY(uint32_t i0) { return 0x000000f4 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_CONFIG(uint32_t i0) { return 0x000032f8 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_CONFIG(uint32_t i0) { return 0x000000f8 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_PARAM(uint32_t i0) { return 0x000032fc + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_PARAM(uint32_t i0) { return 0x000000fc + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_LOW0(uint32_t i0) { return 0x00003300 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_LOW0(uint32_t i0) { return 0x00000100 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_LOW1(uint32_t i0) { return 0x00003304 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_LOW1(uint32_t i0) { return 0x00000104 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_HIGH0(uint32_t i0) { return 0x00003308 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_HIGH0(uint32_t i0) { return 0x00000108 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_HIGH1(uint32_t i0) { return 0x0000330c + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_CURSOR_BLEND_TRANSP_HIGH1(uint32_t i0) { return 0x0000010c + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_LM_GC_LUT_BASE(uint32_t i0) { return 0x00003310 + 0x400*i0; }
+static inline uint32_t REG_MDP5_LM_GC_LUT_BASE(uint32_t i0) { return 0x00000110 + __offset_LM(i0); }
 
-static inline uint32_t REG_MDP5_DSPP(uint32_t i0) { return 0x00004600 + 0x400*i0; }
+static inline uint32_t __offset_DSPP(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return (mdp5_cfg->dspp.base[0]);
+		case 1: return (mdp5_cfg->dspp.base[1]);
+		case 2: return (mdp5_cfg->dspp.base[2]);
+		case 3: return (mdp5_cfg->dspp.base[3]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_DSPP(uint32_t i0) { return 0x00000000 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_OP_MODE(uint32_t i0) { return 0x00004600 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_OP_MODE(uint32_t i0) { return 0x00000000 + __offset_DSPP(i0); }
 #define MDP5_DSPP_OP_MODE_IGC_LUT_EN				0x00000001
 #define MDP5_DSPP_OP_MODE_IGC_TBL_IDX__MASK			0x0000000e
 #define MDP5_DSPP_OP_MODE_IGC_TBL_IDX__SHIFT			1
@@ -811,29 +895,40 @@ static inline uint32_t MDP5_DSPP_OP_MODE_IGC_TBL_IDX(uint32_t val)
 #define MDP5_DSPP_OP_MODE_GAMUT_EN				0x00800000
 #define MDP5_DSPP_OP_MODE_GAMUT_ORDER				0x01000000
 
-static inline uint32_t REG_MDP5_DSPP_PCC_BASE(uint32_t i0) { return 0x00004630 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_PCC_BASE(uint32_t i0) { return 0x00000030 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_DITHER_DEPTH(uint32_t i0) { return 0x00004750 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_DITHER_DEPTH(uint32_t i0) { return 0x00000150 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_HIST_CTL_BASE(uint32_t i0) { return 0x00004810 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_HIST_CTL_BASE(uint32_t i0) { return 0x00000210 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_HIST_LUT_BASE(uint32_t i0) { return 0x00004830 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_HIST_LUT_BASE(uint32_t i0) { return 0x00000230 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_HIST_LUT_SWAP(uint32_t i0) { return 0x00004834 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_HIST_LUT_SWAP(uint32_t i0) { return 0x00000234 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_PA_BASE(uint32_t i0) { return 0x00004838 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_PA_BASE(uint32_t i0) { return 0x00000238 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_GAMUT_BASE(uint32_t i0) { return 0x000048dc + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_GAMUT_BASE(uint32_t i0) { return 0x000002dc + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_DSPP_GC_BASE(uint32_t i0) { return 0x000048b0 + 0x400*i0; }
+static inline uint32_t REG_MDP5_DSPP_GC_BASE(uint32_t i0) { return 0x000002b0 + __offset_DSPP(i0); }
 
-static inline uint32_t REG_MDP5_INTF(uint32_t i0) { return 0x00012500 + 0x200*i0; }
+static inline uint32_t __offset_INTF(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return (mdp5_cfg->intf.base[0]);
+		case 1: return (mdp5_cfg->intf.base[1]);
+		case 2: return (mdp5_cfg->intf.base[2]);
+		case 3: return (mdp5_cfg->intf.base[3]);
+		case 4: return (mdp5_cfg->intf.base[4]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_INTF(uint32_t i0) { return 0x00000000 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TIMING_ENGINE_EN(uint32_t i0) { return 0x00012500 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TIMING_ENGINE_EN(uint32_t i0) { return 0x00000000 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_CONFIG(uint32_t i0) { return 0x00012504 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_CONFIG(uint32_t i0) { return 0x00000004 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_HSYNC_CTL(uint32_t i0) { return 0x00012508 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_HSYNC_CTL(uint32_t i0) { return 0x00000008 + __offset_INTF(i0); }
 #define MDP5_INTF_HSYNC_CTL_PULSEW__MASK			0x0000ffff
 #define MDP5_INTF_HSYNC_CTL_PULSEW__SHIFT			0
 static inline uint32_t MDP5_INTF_HSYNC_CTL_PULSEW(uint32_t val)
@@ -847,23 +942,23 @@ static inline uint32_t MDP5_INTF_HSYNC_CTL_PERIOD(uint32_t val)
 	return ((val) << MDP5_INTF_HSYNC_CTL_PERIOD__SHIFT) & MDP5_INTF_HSYNC_CTL_PERIOD__MASK;
 }
 
-static inline uint32_t REG_MDP5_INTF_VSYNC_PERIOD_F0(uint32_t i0) { return 0x0001250c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_VSYNC_PERIOD_F0(uint32_t i0) { return 0x0000000c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_VSYNC_PERIOD_F1(uint32_t i0) { return 0x00012510 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_VSYNC_PERIOD_F1(uint32_t i0) { return 0x00000010 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_VSYNC_LEN_F0(uint32_t i0) { return 0x00012514 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_VSYNC_LEN_F0(uint32_t i0) { return 0x00000014 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_VSYNC_LEN_F1(uint32_t i0) { return 0x00012518 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_VSYNC_LEN_F1(uint32_t i0) { return 0x00000018 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DISPLAY_VSTART_F0(uint32_t i0) { return 0x0001251c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DISPLAY_VSTART_F0(uint32_t i0) { return 0x0000001c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DISPLAY_VSTART_F1(uint32_t i0) { return 0x00012520 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DISPLAY_VSTART_F1(uint32_t i0) { return 0x00000020 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DISPLAY_VEND_F0(uint32_t i0) { return 0x00012524 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DISPLAY_VEND_F0(uint32_t i0) { return 0x00000024 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DISPLAY_VEND_F1(uint32_t i0) { return 0x00012528 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DISPLAY_VEND_F1(uint32_t i0) { return 0x00000028 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_ACTIVE_VSTART_F0(uint32_t i0) { return 0x0001252c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_ACTIVE_VSTART_F0(uint32_t i0) { return 0x0000002c + __offset_INTF(i0); }
 #define MDP5_INTF_ACTIVE_VSTART_F0_VAL__MASK			0x7fffffff
 #define MDP5_INTF_ACTIVE_VSTART_F0_VAL__SHIFT			0
 static inline uint32_t MDP5_INTF_ACTIVE_VSTART_F0_VAL(uint32_t val)
@@ -872,7 +967,7 @@ static inline uint32_t MDP5_INTF_ACTIVE_VSTART_F0_VAL(uint32_t val)
 }
 #define MDP5_INTF_ACTIVE_VSTART_F0_ACTIVE_V_ENABLE		0x80000000
 
-static inline uint32_t REG_MDP5_INTF_ACTIVE_VSTART_F1(uint32_t i0) { return 0x00012530 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_ACTIVE_VSTART_F1(uint32_t i0) { return 0x00000030 + __offset_INTF(i0); }
 #define MDP5_INTF_ACTIVE_VSTART_F1_VAL__MASK			0x7fffffff
 #define MDP5_INTF_ACTIVE_VSTART_F1_VAL__SHIFT			0
 static inline uint32_t MDP5_INTF_ACTIVE_VSTART_F1_VAL(uint32_t val)
@@ -880,11 +975,11 @@ static inline uint32_t MDP5_INTF_ACTIVE_VSTART_F1_VAL(uint32_t val)
 	return ((val) << MDP5_INTF_ACTIVE_VSTART_F1_VAL__SHIFT) & MDP5_INTF_ACTIVE_VSTART_F1_VAL__MASK;
 }
 
-static inline uint32_t REG_MDP5_INTF_ACTIVE_VEND_F0(uint32_t i0) { return 0x00012534 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_ACTIVE_VEND_F0(uint32_t i0) { return 0x00000034 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_ACTIVE_VEND_F1(uint32_t i0) { return 0x00012538 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_ACTIVE_VEND_F1(uint32_t i0) { return 0x00000038 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DISPLAY_HCTL(uint32_t i0) { return 0x0001253c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DISPLAY_HCTL(uint32_t i0) { return 0x0000003c + __offset_INTF(i0); }
 #define MDP5_INTF_DISPLAY_HCTL_START__MASK			0x0000ffff
 #define MDP5_INTF_DISPLAY_HCTL_START__SHIFT			0
 static inline uint32_t MDP5_INTF_DISPLAY_HCTL_START(uint32_t val)
@@ -898,7 +993,7 @@ static inline uint32_t MDP5_INTF_DISPLAY_HCTL_END(uint32_t val)
 	return ((val) << MDP5_INTF_DISPLAY_HCTL_END__SHIFT) & MDP5_INTF_DISPLAY_HCTL_END__MASK;
 }
 
-static inline uint32_t REG_MDP5_INTF_ACTIVE_HCTL(uint32_t i0) { return 0x00012540 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_ACTIVE_HCTL(uint32_t i0) { return 0x00000040 + __offset_INTF(i0); }
 #define MDP5_INTF_ACTIVE_HCTL_START__MASK			0x00007fff
 #define MDP5_INTF_ACTIVE_HCTL_START__SHIFT			0
 static inline uint32_t MDP5_INTF_ACTIVE_HCTL_START(uint32_t val)
@@ -913,124 +1008,132 @@ static inline uint32_t MDP5_INTF_ACTIVE_HCTL_END(uint32_t val)
 }
 #define MDP5_INTF_ACTIVE_HCTL_ACTIVE_H_ENABLE			0x80000000
 
-static inline uint32_t REG_MDP5_INTF_BORDER_COLOR(uint32_t i0) { return 0x00012544 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_BORDER_COLOR(uint32_t i0) { return 0x00000044 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_UNDERFLOW_COLOR(uint32_t i0) { return 0x00012548 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_UNDERFLOW_COLOR(uint32_t i0) { return 0x00000048 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_HSYNC_SKEW(uint32_t i0) { return 0x0001254c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_HSYNC_SKEW(uint32_t i0) { return 0x0000004c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_POLARITY_CTL(uint32_t i0) { return 0x00012550 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_POLARITY_CTL(uint32_t i0) { return 0x00000050 + __offset_INTF(i0); }
 #define MDP5_INTF_POLARITY_CTL_HSYNC_LOW			0x00000001
 #define MDP5_INTF_POLARITY_CTL_VSYNC_LOW			0x00000002
 #define MDP5_INTF_POLARITY_CTL_DATA_EN_LOW			0x00000004
 
-static inline uint32_t REG_MDP5_INTF_TEST_CTL(uint32_t i0) { return 0x00012554 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TEST_CTL(uint32_t i0) { return 0x00000054 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TP_COLOR0(uint32_t i0) { return 0x00012558 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TP_COLOR0(uint32_t i0) { return 0x00000058 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TP_COLOR1(uint32_t i0) { return 0x0001255c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TP_COLOR1(uint32_t i0) { return 0x0000005c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DSI_CMD_MODE_TRIGGER_EN(uint32_t i0) { return 0x00012584 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DSI_CMD_MODE_TRIGGER_EN(uint32_t i0) { return 0x00000084 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_PANEL_FORMAT(uint32_t i0) { return 0x00012590 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_PANEL_FORMAT(uint32_t i0) { return 0x00000090 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_FRAME_LINE_COUNT_EN(uint32_t i0) { return 0x000125a8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_FRAME_LINE_COUNT_EN(uint32_t i0) { return 0x000000a8 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_FRAME_COUNT(uint32_t i0) { return 0x000125ac + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_FRAME_COUNT(uint32_t i0) { return 0x000000ac + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_LINE_COUNT(uint32_t i0) { return 0x000125b0 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_LINE_COUNT(uint32_t i0) { return 0x000000b0 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DEFLICKER_CONFIG(uint32_t i0) { return 0x000125f0 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DEFLICKER_CONFIG(uint32_t i0) { return 0x000000f0 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DEFLICKER_STRNG_COEFF(uint32_t i0) { return 0x000125f4 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DEFLICKER_STRNG_COEFF(uint32_t i0) { return 0x000000f4 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_DEFLICKER_WEAK_COEFF(uint32_t i0) { return 0x000125f8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_DEFLICKER_WEAK_COEFF(uint32_t i0) { return 0x000000f8 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_ENABLE(uint32_t i0) { return 0x00012600 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_ENABLE(uint32_t i0) { return 0x00000100 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_MAIN_CONTROL(uint32_t i0) { return 0x00012604 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_MAIN_CONTROL(uint32_t i0) { return 0x00000104 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_VIDEO_CONFIG(uint32_t i0) { return 0x00012608 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_VIDEO_CONFIG(uint32_t i0) { return 0x00000108 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_COMPONENT_LIMITS(uint32_t i0) { return 0x0001260c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_COMPONENT_LIMITS(uint32_t i0) { return 0x0000010c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_RECTANGLE(uint32_t i0) { return 0x00012610 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_RECTANGLE(uint32_t i0) { return 0x00000110 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_INITIAL_VALUE(uint32_t i0) { return 0x00012614 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_INITIAL_VALUE(uint32_t i0) { return 0x00000114 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_BLK_WHITE_PATTERN_FRAME(uint32_t i0) { return 0x00012618 + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_BLK_WHITE_PATTERN_FRAME(uint32_t i0) { return 0x00000118 + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_INTF_TPG_RGB_MAPPING(uint32_t i0) { return 0x0001261c + 0x200*i0; }
+static inline uint32_t REG_MDP5_INTF_TPG_RGB_MAPPING(uint32_t i0) { return 0x0000011c + __offset_INTF(i0); }
 
-static inline uint32_t REG_MDP5_AD(uint32_t i0) { return 0x00013100 + 0x200*i0; }
+static inline uint32_t __offset_AD(uint32_t idx)
+{
+	switch (idx) {
+		case 0: return (mdp5_cfg->ad.base[0]);
+		case 1: return (mdp5_cfg->ad.base[1]);
+		default: return INVALID_IDX(idx);
+	}
+}
+static inline uint32_t REG_MDP5_AD(uint32_t i0) { return 0x00000000 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BYPASS(uint32_t i0) { return 0x00013100 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BYPASS(uint32_t i0) { return 0x00000000 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CTRL_0(uint32_t i0) { return 0x00013104 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CTRL_0(uint32_t i0) { return 0x00000004 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CTRL_1(uint32_t i0) { return 0x00013108 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CTRL_1(uint32_t i0) { return 0x00000008 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_FRAME_SIZE(uint32_t i0) { return 0x0001310c + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_FRAME_SIZE(uint32_t i0) { return 0x0000000c + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CON_CTRL_0(uint32_t i0) { return 0x00013110 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CON_CTRL_0(uint32_t i0) { return 0x00000010 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CON_CTRL_1(uint32_t i0) { return 0x00013114 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CON_CTRL_1(uint32_t i0) { return 0x00000014 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_STR_MAN(uint32_t i0) { return 0x00013118 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_STR_MAN(uint32_t i0) { return 0x00000018 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_VAR(uint32_t i0) { return 0x0001311c + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_VAR(uint32_t i0) { return 0x0000001c + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_DITH(uint32_t i0) { return 0x00013120 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_DITH(uint32_t i0) { return 0x00000020 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_DITH_CTRL(uint32_t i0) { return 0x00013124 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_DITH_CTRL(uint32_t i0) { return 0x00000024 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_AMP_LIM(uint32_t i0) { return 0x00013128 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_AMP_LIM(uint32_t i0) { return 0x00000028 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_SLOPE(uint32_t i0) { return 0x0001312c + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_SLOPE(uint32_t i0) { return 0x0000002c + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BW_LVL(uint32_t i0) { return 0x00013130 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BW_LVL(uint32_t i0) { return 0x00000030 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_LOGO_POS(uint32_t i0) { return 0x00013134 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_LOGO_POS(uint32_t i0) { return 0x00000034 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_LUT_FI(uint32_t i0) { return 0x00013138 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_LUT_FI(uint32_t i0) { return 0x00000038 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_LUT_CC(uint32_t i0) { return 0x0001317c + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_LUT_CC(uint32_t i0) { return 0x0000007c + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_STR_LIM(uint32_t i0) { return 0x000131c8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_STR_LIM(uint32_t i0) { return 0x000000c8 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CALIB_AB(uint32_t i0) { return 0x000131cc + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CALIB_AB(uint32_t i0) { return 0x000000cc + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CALIB_CD(uint32_t i0) { return 0x000131d0 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CALIB_CD(uint32_t i0) { return 0x000000d0 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_MODE_SEL(uint32_t i0) { return 0x000131d4 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_MODE_SEL(uint32_t i0) { return 0x000000d4 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_TFILT_CTRL(uint32_t i0) { return 0x000131d8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_TFILT_CTRL(uint32_t i0) { return 0x000000d8 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BL_MINMAX(uint32_t i0) { return 0x000131dc + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BL_MINMAX(uint32_t i0) { return 0x000000dc + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BL(uint32_t i0) { return 0x000131e0 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BL(uint32_t i0) { return 0x000000e0 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BL_MAX(uint32_t i0) { return 0x000131e8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BL_MAX(uint32_t i0) { return 0x000000e8 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_AL(uint32_t i0) { return 0x000131ec + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_AL(uint32_t i0) { return 0x000000ec + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_AL_MIN(uint32_t i0) { return 0x000131f0 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_AL_MIN(uint32_t i0) { return 0x000000f0 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_AL_FILT(uint32_t i0) { return 0x000131f4 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_AL_FILT(uint32_t i0) { return 0x000000f4 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CFG_BUF(uint32_t i0) { return 0x000131f8 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CFG_BUF(uint32_t i0) { return 0x000000f8 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_LUT_AL(uint32_t i0) { return 0x00013200 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_LUT_AL(uint32_t i0) { return 0x00000100 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_TARG_STR(uint32_t i0) { return 0x00013244 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_TARG_STR(uint32_t i0) { return 0x00000144 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_START_CALC(uint32_t i0) { return 0x00013248 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_START_CALC(uint32_t i0) { return 0x00000148 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_STR_OUT(uint32_t i0) { return 0x0001324c + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_STR_OUT(uint32_t i0) { return 0x0000014c + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_BL_OUT(uint32_t i0) { return 0x00013254 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_BL_OUT(uint32_t i0) { return 0x00000154 + __offset_AD(i0); }
 
-static inline uint32_t REG_MDP5_AD_CALC_DONE(uint32_t i0) { return 0x00013258 + 0x200*i0; }
+static inline uint32_t REG_MDP5_AD_CALC_DONE(uint32_t i0) { return 0x00000158 + __offset_AD(i0); }
 
 
 #endif /* MDP5_XML */
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
index 71510ee26e96..31a2c6331a1d 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c
@@ -26,14 +26,98 @@ static const char *iommu_ports[] = {
 
 static struct mdp5_platform_config *mdp5_get_config(struct platform_device *dev);
 
-static int mdp5_hw_init(struct msm_kms *kms)
+const struct mdp5_config *mdp5_cfg;
+
+static const struct mdp5_config msm8x74_config = {
+	.name = "msm8x74",
+	.ctl = {
+		.count = 5,
+		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
+	},
+	.pipe_vig = {
+		.count = 3,
+		.base = { 0x01200, 0x01600, 0x01a00 },
+	},
+	.pipe_rgb = {
+		.count = 3,
+		.base = { 0x01e00, 0x02200, 0x02600 },
+	},
+	.pipe_dma = {
+		.count = 2,
+		.base = { 0x02a00, 0x02e00 },
+	},
+	.lm = {
+		.count = 5,
+		.base = { 0x03200, 0x03600, 0x03a00, 0x03e00, 0x04200 },
+	},
+	.dspp = {
+		.count = 3,
+		.base = { 0x04600, 0x04a00, 0x04e00 },
+	},
+	.ad = {
+		.count = 2,
+		.base = { 0x13100, 0x13300 }, /* NOTE: no ad in v1.0 */
+	},
+	.intf = {
+		.count = 4,
+		.base = { 0x12500, 0x12700, 0x12900, 0x12b00 },
+	},
+};
+
+static const struct mdp5_config apq8084_config = {
+	.name = "apq8084",
+	.ctl = {
+		.count = 5,
+		.base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 },
+	},
+	.pipe_vig = {
+		.count = 4,
+		.base = { 0x01200, 0x01600, 0x01a00, 0x01e00 },
+	},
+	.pipe_rgb = {
+		.count = 4,
+		.base = { 0x02200, 0x02600, 0x02a00, 0x02e00 },
+	},
+	.pipe_dma = {
+		.count = 2,
+		.base = { 0x03200, 0x03600 },
+	},
+	.lm = {
+		.count = 6,
+		.base = { 0x03a00, 0x03e00, 0x04200, 0x04600, 0x04a00, 0x04e00 },
+	},
+	.dspp = {
+		.count = 4,
+		.base = { 0x05200, 0x05600, 0x05a00, 0x05e00 },
+
+	},
+	.ad = {
+		.count = 3,
+		.base = { 0x13500, 0x13700, 0x13900 },
+	},
+	.intf = {
+		.count = 5,
+		.base = { 0x12500, 0x12700, 0x12900, 0x12b00, 0x12d00 },
+	},
+};
+
+struct mdp5_config_entry {
+	int revision;
+	const struct mdp5_config *config;
+};
+
+static const struct mdp5_config_entry mdp5_configs[] = {
+	{ .revision = 0, .config = &msm8x74_config },
+	{ .revision = 2, .config = &msm8x74_config },
+	{ .revision = 3, .config = &apq8084_config },
+};
+
+static int mdp5_select_hw_cfg(struct msm_kms *kms)
 {
 	struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms));
 	struct drm_device *dev = mdp5_kms->dev;
 	uint32_t version, major, minor;
-	int ret = 0;
-
-	pm_runtime_get_sync(dev->dev);
+	int i, ret = 0;
 
 	mdp5_enable(mdp5_kms);
 	version = mdp5_read(mdp5_kms, REG_MDP5_MDP_VERSION);
@@ -44,8 +128,8 @@ static int mdp5_hw_init(struct msm_kms *kms)
 
 	DBG("found MDP5 version v%d.%d", major, minor);
 
-	if ((major != 1) || ((minor != 0) && (minor != 2))) {
-		dev_err(dev->dev, "unexpected MDP version: v%d.%d\n",
+	if (major != 1) {
+		dev_err(dev->dev, "unexpected MDP major version: v%d.%d\n",
 				major, minor);
 		ret = -ENXIO;
 		goto out;
@@ -53,6 +137,35 @@ static int mdp5_hw_init(struct msm_kms *kms)
 
 	mdp5_kms->rev = minor;
 
+	/* only after mdp5_cfg global pointer's init can we access the hw */
+	for (i = 0; i < ARRAY_SIZE(mdp5_configs); i++) {
+		if (mdp5_configs[i].revision != minor)
+			continue;
+		mdp5_kms->hw_cfg = mdp5_cfg = mdp5_configs[i].config;
+		break;
+	}
+	if (unlikely(!mdp5_kms->hw_cfg)) {
+		dev_err(dev->dev, "unexpected MDP minor revision: v%d.%d\n",
+				major, minor);
+		ret = -ENXIO;
+		goto out;
+	}
+
+	DBG("MDP5: %s config selected", mdp5_kms->hw_cfg->name);
+
+	return 0;
+out:
+	return ret;
+}
+
+static int mdp5_hw_init(struct msm_kms *kms)
+{
+	struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms));
+	struct drm_device *dev = mdp5_kms->dev;
+	int i;
+
+	pm_runtime_get_sync(dev->dev);
+
 	/* Magic unknown register writes:
 	 *
 	 *    W VBIF:0x004 00000001      (mdss_mdp.c:839)
@@ -78,15 +191,13 @@ static int mdp5_hw_init(struct msm_kms *kms)
 	 */
 
 	mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(0), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(1), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(2), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(3), 0);
 
-out:
+	for (i = 0; i < mdp5_kms->hw_cfg->ctl.count; i++)
+		mdp5_write(mdp5_kms, REG_MDP5_CTL_OP(i), 0);
+
 	pm_runtime_put_sync(dev->dev);
 
-	return ret;
+	return 0;
 }
 
 static long mdp5_round_pixclk(struct msm_kms *kms, unsigned long rate,
@@ -161,7 +272,7 @@ int mdp5_enable(struct mdp5_kms *mdp5_kms)
 static int modeset_init(struct mdp5_kms *mdp5_kms)
 {
 	static const enum mdp5_pipe crtcs[] = {
-			SSPP_RGB0, SSPP_RGB1, SSPP_RGB2,
+			SSPP_RGB0, SSPP_RGB1, SSPP_RGB2, SSPP_RGB3,
 	};
 	struct drm_device *dev = mdp5_kms->dev;
 	struct msm_drm_private *priv = dev->dev_private;
@@ -169,7 +280,7 @@ static int modeset_init(struct mdp5_kms *mdp5_kms)
 	int i, ret;
 
 	/* construct CRTCs: */
-	for (i = 0; i < ARRAY_SIZE(crtcs); i++) {
+	for (i = 0; i < mdp5_kms->hw_cfg->pipe_rgb.count; i++) {
 		struct drm_plane *plane;
 		struct drm_crtc *crtc;
 
@@ -246,7 +357,7 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev)
 	struct mdp5_kms *mdp5_kms;
 	struct msm_kms *kms = NULL;
 	struct msm_mmu *mmu;
-	int ret;
+	int i, ret;
 
 	mdp5_kms = kzalloc(sizeof(*mdp5_kms), GFP_KERNEL);
 	if (!mdp5_kms) {
@@ -307,20 +418,22 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev)
 
 	ret = clk_set_rate(mdp5_kms->src_clk, config->max_clk);
 
+	ret = mdp5_select_hw_cfg(kms);
+	if (ret)
+		goto fail;
+
 	/* make sure things are off before attaching iommu (bootloader could
 	 * have left things on, in which case we'll start getting faults if
 	 * we don't disable):
 	 */
 	mdp5_enable(mdp5_kms);
-	mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(0), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(1), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(2), 0);
-	mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(3), 0);
+	for (i = 0; i < mdp5_kms->hw_cfg->intf.count; i++)
+		mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(i), 0);
 	mdp5_disable(mdp5_kms);
 	mdelay(16);
 
 	if (config->iommu) {
-		mmu = msm_iommu_new(dev, config->iommu);
+		mmu = msm_iommu_new(&pdev->dev, config->iommu);
 		if (IS_ERR(mmu)) {
 			ret = PTR_ERR(mmu);
 			dev_err(dev->dev, "failed to init iommu: %d\n", ret);
@@ -368,5 +481,11 @@ static struct mdp5_platform_config *mdp5_get_config(struct platform_device *dev)
 #ifdef CONFIG_OF
 	/* TODO */
 #endif
+	config.iommu = iommu_domain_alloc(&platform_bus_type);
+	/* TODO hard-coded in downstream mdss, but should it be? */
+	config.max_clk = 200000000;
+	/* TODO get from DT: */
+	config.smp_blk_cnt = 22;
+
 	return &config;
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
index 6e981b692d1d..5bf340dd0f00 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h
@@ -21,6 +21,24 @@
 #include "msm_drv.h"
 #include "msm_kms.h"
 #include "mdp/mdp_kms.h"
+/* dynamic offsets used by mdp5.xml.h (initialized in mdp5_kms.c) */
+#define MDP5_MAX_BASES		8
+struct mdp5_sub_block {
+	int	count;
+	uint32_t base[MDP5_MAX_BASES];
+};
+struct mdp5_config {
+	char  *name;
+	struct mdp5_sub_block ctl;
+	struct mdp5_sub_block pipe_vig;
+	struct mdp5_sub_block pipe_rgb;
+	struct mdp5_sub_block pipe_dma;
+	struct mdp5_sub_block lm;
+	struct mdp5_sub_block dspp;
+	struct mdp5_sub_block ad;
+	struct mdp5_sub_block intf;
+};
+extern const struct mdp5_config *mdp5_cfg;
 #include "mdp5.xml.h"
 #include "mdp5_smp.h"
 
@@ -30,6 +48,7 @@ struct mdp5_kms {
 	struct drm_device *dev;
 
 	int rev;
+	const struct mdp5_config *hw_cfg;
 
 	/* mapper-id used to request GEM buffer mapped for scanout: */
 	int id;
@@ -82,6 +101,7 @@ static inline const char *pipe2name(enum mdp5_pipe pipe)
 		NAME(VIG0), NAME(VIG1), NAME(VIG2),
 		NAME(RGB0), NAME(RGB1), NAME(RGB2),
 		NAME(DMA0), NAME(DMA1),
+		NAME(VIG3), NAME(RGB3),
 #undef NAME
 	};
 	return names[pipe];
@@ -98,6 +118,8 @@ static inline uint32_t pipe2flush(enum mdp5_pipe pipe)
 	case SSPP_RGB2: return MDP5_CTL_FLUSH_RGB2;
 	case SSPP_DMA0: return MDP5_CTL_FLUSH_DMA0;
 	case SSPP_DMA1: return MDP5_CTL_FLUSH_DMA1;
+	case SSPP_VIG3: return MDP5_CTL_FLUSH_VIG3;
+	case SSPP_RGB3: return MDP5_CTL_FLUSH_RGB3;
 	default:        return 0;
 	}
 }
@@ -108,6 +130,7 @@ static inline int pipe2nclients(enum mdp5_pipe pipe)
 	case SSPP_RGB0:
 	case SSPP_RGB1:
 	case SSPP_RGB2:
+	case SSPP_RGB3:
 		return 1;
 	default:
 		return 3;
@@ -126,6 +149,8 @@ static inline enum mdp5_client_id pipe2client(enum mdp5_pipe pipe, int plane)
 	case SSPP_RGB2: return CID_RGB2;
 	case SSPP_DMA0: return CID_DMA0_Y + plane;
 	case SSPP_DMA1: return CID_DMA1_Y + plane;
+	case SSPP_VIG3: return CID_VIG3_Y + plane;
+	case SSPP_RGB3: return CID_RGB3;
 	default:        return CID_UNUSED;
 	}
 }
diff --git a/drivers/gpu/drm/msm/mdp/mdp_common.xml.h b/drivers/gpu/drm/msm/mdp/mdp_common.xml.h
index a9629b85b983..64c1afd6030a 100644
--- a/drivers/gpu/drm/msm/mdp/mdp_common.xml.h
+++ b/drivers/gpu/drm/msm/mdp/mdp_common.xml.h
@@ -12,12 +12,12 @@ The rules-ng-ng source files this header was generated from are:
 - /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml (   1453 bytes, from 2013-03-31 16:51:27)
 - /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml            (  17996 bytes, from 2013-12-01 19:10:31)
 - /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml      (   1615 bytes, from 2013-11-30 15:00:52)
-- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2013-12-03 20:59:13)
+- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml            (  22517 bytes, from 2014-06-25 12:55:02)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml             (  11712 bytes, from 2013-08-17 17:13:43)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml            (    344 bytes, from 2013-08-11 19:26:32)
 - /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml         (   1544 bytes, from 2013-08-16 19:17:05)
 - /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml         (    600 bytes, from 2013-07-05 19:21:12)
-- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  20932 bytes, from 2013-12-01 15:13:04)
+- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml           (  23613 bytes, from 2014-06-25 12:53:44)
 
 Copyright (C) 2013 by the following authors:
 - Rob Clark <robdclark@gmail.com> (robclark)
diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c
index 9a5d87db5c23..b67ef5985125 100644
--- a/drivers/gpu/drm/msm/msm_drv.c
+++ b/drivers/gpu/drm/msm/msm_drv.c
@@ -52,7 +52,7 @@ module_param(reglog, bool, 0600);
 #define reglog 0
 #endif
 
-static char *vram;
+static char *vram = "16m";
 MODULE_PARM_DESC(vram, "Configure VRAM size (for devices without IOMMU/GPUMMU");
 module_param(vram, charp, 0);
 
@@ -181,7 +181,6 @@ static int msm_load(struct drm_device *dev, unsigned long flags)
 	struct msm_kms *kms;
 	int ret;
 
-
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv) {
 		dev_err(dev->dev, "failed to allocate private data\n");
@@ -281,7 +280,7 @@ static int msm_load(struct drm_device *dev, unsigned long flags)
 	dev->mode_config.max_height = 2048;
 	dev->mode_config.funcs = &mode_config_funcs;
 
-	ret = drm_vblank_init(dev, 1);
+	ret = drm_vblank_init(dev, priv->num_crtcs);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize vblank\n");
 		goto fail;
@@ -314,38 +313,15 @@ fail:
 
 static void load_gpu(struct drm_device *dev)
 {
+	static DEFINE_MUTEX(init_lock);
 	struct msm_drm_private *priv = dev->dev_private;
-	struct msm_gpu *gpu;
-
-	if (priv->gpu)
-		return;
-
-	mutex_lock(&dev->struct_mutex);
-	gpu = a3xx_gpu_init(dev);
-	if (IS_ERR(gpu)) {
-		dev_warn(dev->dev, "failed to load a3xx gpu\n");
-		gpu = NULL;
-		/* not fatal */
-	}
-
-	if (gpu) {
-		int ret;
-		gpu->funcs->pm_resume(gpu);
-		ret = gpu->funcs->hw_init(gpu);
-		if (ret) {
-			dev_err(dev->dev, "gpu hw init failed: %d\n", ret);
-			gpu->funcs->destroy(gpu);
-			gpu = NULL;
-		} else {
-			/* give inactive pm a chance to kick in: */
-			msm_gpu_retire(gpu);
-		}
 
-	}
+	mutex_lock(&init_lock);
 
-	priv->gpu = gpu;
+	if (!priv->gpu)
+		priv->gpu = adreno_load_gpu(dev);
 
-	mutex_unlock(&dev->struct_mutex);
+	mutex_unlock(&init_lock);
 }
 
 static int msm_open(struct drm_device *dev, struct drm_file *file)
@@ -833,6 +809,7 @@ static struct drm_driver msm_driver = {
 	.open               = msm_open,
 	.preclose           = msm_preclose,
 	.lastclose          = msm_lastclose,
+	.set_busid          = drm_platform_set_busid,
 	.irq_handler        = msm_irq,
 	.irq_preinstall     = msm_irq_preinstall,
 	.irq_postinstall    = msm_irq_postinstall,
@@ -906,25 +883,22 @@ static int compare_of(struct device *dev, void *data)
 	return dev->of_node == data;
 }
 
-static int msm_drm_add_components(struct device *master, struct master *m)
+static int add_components(struct device *dev, struct component_match **matchptr,
+		const char *name)
 {
-	struct device_node *np = master->of_node;
+	struct device_node *np = dev->of_node;
 	unsigned i;
-	int ret;
 
 	for (i = 0; ; i++) {
 		struct device_node *node;
 
-		node = of_parse_phandle(np, "connectors", i);
+		node = of_parse_phandle(np, name, i);
 		if (!node)
 			break;
 
-		ret = component_master_add_child(m, compare_of, node);
-		of_node_put(node);
-
-		if (ret)
-			return ret;
+		component_match_add(dev, matchptr, compare_of, node);
 	}
+
 	return 0;
 }
 #else
@@ -932,9 +906,34 @@ static int compare_dev(struct device *dev, void *data)
 {
 	return dev == data;
 }
+#endif
+
+static int msm_drm_bind(struct device *dev)
+{
+	return drm_platform_init(&msm_driver, to_platform_device(dev));
+}
+
+static void msm_drm_unbind(struct device *dev)
+{
+	drm_put_dev(platform_get_drvdata(to_platform_device(dev)));
+}
+
+static const struct component_master_ops msm_drm_ops = {
+	.bind = msm_drm_bind,
+	.unbind = msm_drm_unbind,
+};
 
-static int msm_drm_add_components(struct device *master, struct master *m)
+/*
+ * Platform driver:
+ */
+
+static int msm_pdev_probe(struct platform_device *pdev)
 {
+	struct component_match *match = NULL;
+#ifdef CONFIG_OF
+	add_components(&pdev->dev, &match, "connectors");
+	add_components(&pdev->dev, &match, "gpus");
+#else
 	/* For non-DT case, it kinda sucks.  We don't actually have a way
 	 * to know whether or not we are waiting for certain devices (or if
 	 * they are simply not present).  But for non-DT we only need to
@@ -949,50 +948,20 @@ static int msm_drm_add_components(struct device *master, struct master *m)
 
 	for (i = 0; i < ARRAY_SIZE(devnames); i++) {
 		struct device *dev;
-		int ret;
 
 		dev = bus_find_device_by_name(&platform_bus_type,
 				NULL, devnames[i]);
 		if (!dev) {
-			dev_info(master, "still waiting for %s\n", devnames[i]);
+			dev_info(&pdev->dev, "still waiting for %s\n", devnames[i]);
 			return -EPROBE_DEFER;
 		}
 
-		ret = component_master_add_child(m, compare_dev, dev);
-		if (ret) {
-			DBG("could not add child: %d", ret);
-			return ret;
-		}
+		component_match_add(&pdev->dev, &match, compare_dev, dev);
 	}
-
-	return 0;
-}
 #endif
 
-static int msm_drm_bind(struct device *dev)
-{
-	return drm_platform_init(&msm_driver, to_platform_device(dev));
-}
-
-static void msm_drm_unbind(struct device *dev)
-{
-	drm_put_dev(platform_get_drvdata(to_platform_device(dev)));
-}
-
-static const struct component_master_ops msm_drm_ops = {
-		.add_components = msm_drm_add_components,
-		.bind = msm_drm_bind,
-		.unbind = msm_drm_unbind,
-};
-
-/*
- * Platform driver:
- */
-
-static int msm_pdev_probe(struct platform_device *pdev)
-{
 	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
-	return component_master_add(&pdev->dev, &msm_drm_ops);
+	return component_master_add_with_match(&pdev->dev, &msm_drm_ops, match);
 }
 
 static int msm_pdev_remove(struct platform_device *pdev)
@@ -1008,7 +977,8 @@ static const struct platform_device_id msm_id[] = {
 };
 
 static const struct of_device_id dt_match[] = {
-	{ .compatible = "qcom,mdss_mdp" },
+	{ .compatible = "qcom,mdp" },      /* mdp4 */
+	{ .compatible = "qcom,mdss_mdp" }, /* mdp5 */
 	{}
 };
 MODULE_DEVICE_TABLE(of, dt_match);
@@ -1029,7 +999,7 @@ static int __init msm_drm_register(void)
 {
 	DBG("init");
 	hdmi_register();
-	a3xx_register();
+	adreno_register();
 	return platform_driver_register(&msm_platform_driver);
 }
 
@@ -1038,7 +1008,7 @@ static void __exit msm_drm_unregister(void)
 	DBG("fini");
 	platform_driver_unregister(&msm_platform_driver);
 	hdmi_unregister();
-	a3xx_unregister();
+	adreno_unregister();
 }
 
 module_init(msm_drm_register);
diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h
index 8a2c5fd0893e..67f9d0a2332c 100644
--- a/drivers/gpu/drm/msm/msm_drv.h
+++ b/drivers/gpu/drm/msm/msm_drv.h
@@ -51,6 +51,7 @@ static inline struct device *msm_iommu_get_ctx(const char *ctx_name)
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
 #include <drm/msm_drm.h>
+#include <drm/drm_gem.h>
 
 struct msm_kms;
 struct msm_gpu;
@@ -170,7 +171,7 @@ struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj);
 void *msm_gem_prime_vmap(struct drm_gem_object *obj);
 void msm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
 struct drm_gem_object *msm_gem_prime_import_sg_table(struct drm_device *dev,
-		size_t size, struct sg_table *sg);
+		struct dma_buf_attachment *attach, struct sg_table *sg);
 int msm_gem_prime_pin(struct drm_gem_object *obj);
 void msm_gem_prime_unpin(struct drm_gem_object *obj);
 void *msm_gem_vaddr_locked(struct drm_gem_object *obj);
diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c
index 5107fc4826bc..ab5bfd2d0ebf 100644
--- a/drivers/gpu/drm/msm/msm_fbdev.c
+++ b/drivers/gpu/drm/msm/msm_fbdev.c
@@ -19,6 +19,11 @@
 
 #include "drm_crtc.h"
 #include "drm_fb_helper.h"
+#include "msm_gem.h"
+
+extern int msm_gem_mmap_obj(struct drm_gem_object *obj,
+					struct vm_area_struct *vma);
+static int msm_fbdev_mmap(struct fb_info *info, struct vm_area_struct *vma);
 
 /*
  * fbdev funcs, to implement legacy fbdev interface on top of drm driver
@@ -43,6 +48,7 @@ static struct fb_ops msm_fb_ops = {
 	.fb_fillrect = sys_fillrect,
 	.fb_copyarea = sys_copyarea,
 	.fb_imageblit = sys_imageblit,
+	.fb_mmap = msm_fbdev_mmap,
 
 	.fb_check_var = drm_fb_helper_check_var,
 	.fb_set_par = drm_fb_helper_set_par,
@@ -51,6 +57,31 @@ static struct fb_ops msm_fb_ops = {
 	.fb_setcmap = drm_fb_helper_setcmap,
 };
 
+static int msm_fbdev_mmap(struct fb_info *info, struct vm_area_struct *vma)
+{
+	struct drm_fb_helper *helper = (struct drm_fb_helper *)info->par;
+	struct msm_fbdev *fbdev = to_msm_fbdev(helper);
+	struct drm_gem_object *drm_obj = fbdev->bo;
+	struct drm_device *dev = helper->dev;
+	int ret = 0;
+
+	if (drm_device_is_unplugged(dev))
+		return -ENODEV;
+
+	mutex_lock(&dev->struct_mutex);
+
+	ret = drm_gem_mmap_obj(drm_obj, drm_obj->size, vma);
+
+	mutex_unlock(&dev->struct_mutex);
+
+	if (ret) {
+		pr_err("%s:drm_gem_mmap_obj fail\n", __func__);
+		return ret;
+	}
+
+	return msm_gem_mmap_obj(drm_obj, vma);
+}
+
 static int msm_fbdev_create(struct drm_fb_helper *helper,
 		struct drm_fb_helper_surface_size *sizes)
 {
@@ -104,8 +135,16 @@ static int msm_fbdev_create(struct drm_fb_helper *helper,
 
 	mutex_lock(&dev->struct_mutex);
 
-	/* TODO implement our own fb_mmap so we don't need this: */
-	msm_gem_get_iova_locked(fbdev->bo, 0, &paddr);
+	/*
+	 * NOTE: if we can be guaranteed to be able to map buffer
+	 * in panic (ie. lock-safe, etc) we could avoid pinning the
+	 * buffer now:
+	 */
+	ret = msm_gem_get_iova_locked(fbdev->bo, 0, &paddr);
+	if (ret) {
+		dev_err(dev->dev, "failed to get buffer obj iova: %d\n", ret);
+		goto fail_unlock;
+	}
 
 	fbi = framebuffer_alloc(0, dev->dev);
 	if (!fbi) {
@@ -177,7 +216,7 @@ static void msm_crtc_fb_gamma_get(struct drm_crtc *crtc,
 	DBG("fbdev: get gamma");
 }
 
-static struct drm_fb_helper_funcs msm_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs msm_fb_helper_funcs = {
 	.gamma_set = msm_crtc_fb_gamma_set,
 	.gamma_get = msm_crtc_fb_gamma_get,
 	.fb_probe = msm_fbdev_create,
@@ -189,7 +228,7 @@ struct drm_fb_helper *msm_fbdev_init(struct drm_device *dev)
 	struct msm_drm_private *priv = dev->dev_private;
 	struct msm_fbdev *fbdev = NULL;
 	struct drm_fb_helper *helper;
-	int ret = 0;
+	int ret;
 
 	fbdev = kzalloc(sizeof(*fbdev), GFP_KERNEL);
 	if (!fbdev)
@@ -197,7 +236,7 @@ struct drm_fb_helper *msm_fbdev_init(struct drm_device *dev)
 
 	helper = &fbdev->base;
 
-	helper->funcs = &msm_fb_helper_funcs;
+	drm_fb_helper_prepare(dev, helper, &msm_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(dev, helper,
 			priv->num_crtcs, priv->num_connectors);
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index 690d7e7b6d1e..4b1b82adabde 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -73,7 +73,7 @@ static struct page **get_pages(struct drm_gem_object *obj)
 		int npages = obj->size >> PAGE_SHIFT;
 
 		if (iommu_present(&platform_bus_type))
-			p = drm_gem_get_pages(obj, 0);
+			p = drm_gem_get_pages(obj);
 		else
 			p = get_pages_vram(obj, npages);
 
@@ -278,24 +278,23 @@ int msm_gem_get_iova_locked(struct drm_gem_object *obj, int id,
 		uint32_t *iova)
 {
 	struct msm_gem_object *msm_obj = to_msm_bo(obj);
-	struct drm_device *dev = obj->dev;
 	int ret = 0;
 
 	if (!msm_obj->domain[id].iova) {
 		struct msm_drm_private *priv = obj->dev->dev_private;
-		struct msm_mmu *mmu = priv->mmus[id];
 		struct page **pages = get_pages(obj);
 
-		if (!mmu) {
-			dev_err(dev->dev, "null MMU pointer\n");
-			return -EINVAL;
-		}
-
 		if (IS_ERR(pages))
 			return PTR_ERR(pages);
 
 		if (iommu_present(&platform_bus_type)) {
-			uint32_t offset = (uint32_t)mmap_offset(obj);
+			struct msm_mmu *mmu = priv->mmus[id];
+			uint32_t offset;
+
+			if (WARN_ON(!mmu))
+				return -EINVAL;
+
+			offset = (uint32_t)mmap_offset(obj);
 			ret = mmu->funcs->map(mmu, offset, msm_obj->sgt,
 					obj->size, IOMMU_READ | IOMMU_WRITE);
 			msm_obj->domain[id].iova = offset;
diff --git a/drivers/gpu/drm/msm/msm_gem_prime.c b/drivers/gpu/drm/msm/msm_gem_prime.c
index d48f9fc5129b..ad772fe36115 100644
--- a/drivers/gpu/drm/msm/msm_gem_prime.c
+++ b/drivers/gpu/drm/msm/msm_gem_prime.c
@@ -18,6 +18,7 @@
 #include "msm_drv.h"
 #include "msm_gem.h"
 
+#include <linux/dma-buf.h>
 
 struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
@@ -37,9 +38,9 @@ void msm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 }
 
 struct drm_gem_object *msm_gem_prime_import_sg_table(struct drm_device *dev,
-		size_t size, struct sg_table *sg)
+		struct dma_buf_attachment *attach, struct sg_table *sg)
 {
-	return msm_gem_import(dev, size, sg);
+	return msm_gem_import(dev, attach->dmabuf->size, sg);
 }
 
 int msm_gem_prime_pin(struct drm_gem_object *obj)
diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c
index c6322197db8c..4a0dce587745 100644
--- a/drivers/gpu/drm/msm/msm_gpu.c
+++ b/drivers/gpu/drm/msm/msm_gpu.c
@@ -606,14 +606,17 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 	iommu = iommu_domain_alloc(&platform_bus_type);
 	if (iommu) {
 		dev_info(drm->dev, "%s: using IOMMU\n", name);
-		gpu->mmu = msm_iommu_new(drm, iommu);
+		gpu->mmu = msm_iommu_new(&pdev->dev, iommu);
 	} else {
 		dev_info(drm->dev, "%s: no IOMMU, fallback to VRAM carveout!\n", name);
 	}
 	gpu->id = msm_register_mmu(drm, gpu->mmu);
 
+
 	/* Create ringbuffer: */
+	mutex_lock(&drm->struct_mutex);
 	gpu->rb = msm_ringbuffer_new(gpu, ringsz);
+	mutex_unlock(&drm->struct_mutex);
 	if (IS_ERR(gpu->rb)) {
 		ret = PTR_ERR(gpu->rb);
 		gpu->rb = NULL;
@@ -621,13 +624,6 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 		goto fail;
 	}
 
-	ret = msm_gem_get_iova_locked(gpu->rb->bo, gpu->id, &gpu->rb_iova);
-	if (ret) {
-		gpu->rb_iova = 0;
-		dev_err(drm->dev, "could not map ringbuffer: %d\n", ret);
-		goto fail;
-	}
-
 	bs_init(gpu);
 
 	return 0;
diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h
index 9b579b792840..fd1e4b4a6d40 100644
--- a/drivers/gpu/drm/msm/msm_gpu.h
+++ b/drivers/gpu/drm/msm/msm_gpu.h
@@ -166,8 +166,8 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev,
 		const char *name, const char *ioname, const char *irqname, int ringsz);
 void msm_gpu_cleanup(struct msm_gpu *gpu);
 
-struct msm_gpu *a3xx_gpu_init(struct drm_device *dev);
-void __init a3xx_register(void);
-void __exit a3xx_unregister(void);
+struct msm_gpu *adreno_load_gpu(struct drm_device *dev);
+void __init adreno_register(void);
+void __exit adreno_unregister(void);
 
 #endif /* __MSM_GPU_H__ */
diff --git a/drivers/gpu/drm/msm/msm_iommu.c b/drivers/gpu/drm/msm/msm_iommu.c
index 4b2ad9181edf..7acdaa5688b7 100644
--- a/drivers/gpu/drm/msm/msm_iommu.c
+++ b/drivers/gpu/drm/msm/msm_iommu.c
@@ -27,45 +27,20 @@ struct msm_iommu {
 static int msm_fault_handler(struct iommu_domain *iommu, struct device *dev,
 		unsigned long iova, int flags, void *arg)
 {
-	DBG("*** fault: iova=%08lx, flags=%d", iova, flags);
-	return -ENOSYS;
+	pr_warn_ratelimited("*** fault: iova=%08lx, flags=%d\n", iova, flags);
+	return 0;
 }
 
 static int msm_iommu_attach(struct msm_mmu *mmu, const char **names, int cnt)
 {
-	struct drm_device *dev = mmu->dev;
 	struct msm_iommu *iommu = to_msm_iommu(mmu);
-	int i, ret;
-
-	for (i = 0; i < cnt; i++) {
-		struct device *msm_iommu_get_ctx(const char *ctx_name);
-		struct device *ctx = msm_iommu_get_ctx(names[i]);
-		if (IS_ERR_OR_NULL(ctx)) {
-			dev_warn(dev->dev, "couldn't get %s context", names[i]);
-			continue;
-		}
-		ret = iommu_attach_device(iommu->domain, ctx);
-		if (ret) {
-			dev_warn(dev->dev, "could not attach iommu to %s", names[i]);
-			return ret;
-		}
-	}
-
-	return 0;
+	return iommu_attach_device(iommu->domain, mmu->dev);
 }
 
 static void msm_iommu_detach(struct msm_mmu *mmu, const char **names, int cnt)
 {
 	struct msm_iommu *iommu = to_msm_iommu(mmu);
-	int i;
-
-	for (i = 0; i < cnt; i++) {
-		struct device *msm_iommu_get_ctx(const char *ctx_name);
-		struct device *ctx = msm_iommu_get_ctx(names[i]);
-		if (IS_ERR_OR_NULL(ctx))
-			continue;
-		iommu_detach_device(iommu->domain, ctx);
-	}
+	iommu_detach_device(iommu->domain, mmu->dev);
 }
 
 static int msm_iommu_map(struct msm_mmu *mmu, uint32_t iova,
@@ -149,7 +124,7 @@ static const struct msm_mmu_funcs funcs = {
 		.destroy = msm_iommu_destroy,
 };
 
-struct msm_mmu *msm_iommu_new(struct drm_device *dev, struct iommu_domain *domain)
+struct msm_mmu *msm_iommu_new(struct device *dev, struct iommu_domain *domain)
 {
 	struct msm_iommu *iommu;
 
diff --git a/drivers/gpu/drm/msm/msm_mmu.h b/drivers/gpu/drm/msm/msm_mmu.h
index 21da6d154f71..7cd88d9dc155 100644
--- a/drivers/gpu/drm/msm/msm_mmu.h
+++ b/drivers/gpu/drm/msm/msm_mmu.h
@@ -32,17 +32,17 @@ struct msm_mmu_funcs {
 
 struct msm_mmu {
 	const struct msm_mmu_funcs *funcs;
-	struct drm_device *dev;
+	struct device *dev;
 };
 
-static inline void msm_mmu_init(struct msm_mmu *mmu, struct drm_device *dev,
+static inline void msm_mmu_init(struct msm_mmu *mmu, struct device *dev,
 		const struct msm_mmu_funcs *funcs)
 {
 	mmu->dev = dev;
 	mmu->funcs = funcs;
 }
 
-struct msm_mmu *msm_iommu_new(struct drm_device *dev, struct iommu_domain *domain);
-struct msm_mmu *msm_gpummu_new(struct drm_device *dev, struct msm_gpu *gpu);
+struct msm_mmu *msm_iommu_new(struct device *dev, struct iommu_domain *domain);
+struct msm_mmu *msm_gpummu_new(struct device *dev, struct msm_gpu *gpu);
 
 #endif /* __MSM_MMU_H__ */
diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index 637c29a33127..40afc69a3778 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -1,5 +1,5 @@
 config DRM_NOUVEAU
-	tristate "Nouveau (nVidia) cards"
+	tristate "Nouveau (NVIDIA) cards"
 	depends on DRM && PCI
         select FW_LOADER
 	select DRM_KMS_HELPER
@@ -23,7 +23,15 @@ config DRM_NOUVEAU
 	select THERMAL if ACPI && X86
 	select ACPI_VIDEO if ACPI && X86
 	help
-	  Choose this option for open-source nVidia support.
+	  Choose this option for open-source NVIDIA support.
+
+config NOUVEAU_PLATFORM_DRIVER
+	tristate "Nouveau (NVIDIA) SoC GPUs"
+	depends on DRM_NOUVEAU && ARCH_TEGRA
+	default y
+	help
+	  Support for Nouveau platform driver, used for SoC GPUs as found
+	  on NVIDIA Tegra K1.
 
 config NOUVEAU_DEBUG
 	int "Maximum debug level"
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 8b307e143632..12c24c8abf7f 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -14,8 +14,10 @@ nouveau-y += core/core/enum.o
 nouveau-y += core/core/event.o
 nouveau-y += core/core/gpuobj.o
 nouveau-y += core/core/handle.o
+nouveau-y += core/core/ioctl.o
 nouveau-y += core/core/mm.o
 nouveau-y += core/core/namedb.o
+nouveau-y += core/core/notify.o
 nouveau-y += core/core/object.o
 nouveau-y += core/core/option.o
 nouveau-y += core/core/parent.o
@@ -26,6 +28,7 @@ nouveau-y += core/core/subdev.o
 nouveau-y += core/subdev/bar/base.o
 nouveau-y += core/subdev/bar/nv50.o
 nouveau-y += core/subdev/bar/nvc0.o
+nouveau-y += core/subdev/bar/gk20a.o
 nouveau-y += core/subdev/bios/base.o
 nouveau-y += core/subdev/bios/bit.o
 nouveau-y += core/subdev/bios/boost.o
@@ -35,6 +38,7 @@ nouveau-y += core/subdev/bios/dcb.o
 nouveau-y += core/subdev/bios/disp.o
 nouveau-y += core/subdev/bios/dp.o
 nouveau-y += core/subdev/bios/extdev.o
+nouveau-y += core/subdev/bios/fan.o
 nouveau-y += core/subdev/bios/gpio.o
 nouveau-y += core/subdev/bios/i2c.o
 nouveau-y += core/subdev/bios/init.o
@@ -48,6 +52,8 @@ nouveau-y += core/subdev/bios/therm.o
 nouveau-y += core/subdev/bios/vmap.o
 nouveau-y += core/subdev/bios/volt.o
 nouveau-y += core/subdev/bios/xpio.o
+nouveau-y += core/subdev/bios/M0205.o
+nouveau-y += core/subdev/bios/M0209.o
 nouveau-y += core/subdev/bios/P0260.o
 nouveau-y += core/subdev/bus/hwsq.o
 nouveau-y += core/subdev/bus/nv04.o
@@ -64,6 +70,7 @@ nouveau-y += core/subdev/clock/nva3.o
 nouveau-y += core/subdev/clock/nvaa.o
 nouveau-y += core/subdev/clock/nvc0.o
 nouveau-y += core/subdev/clock/nve0.o
+nouveau-y += core/subdev/clock/gk20a.o
 nouveau-y += core/subdev/clock/pllnv04.o
 nouveau-y += core/subdev/clock/pllnva3.o
 nouveau-y += core/subdev/devinit/base.o
@@ -120,12 +127,17 @@ nouveau-y += core/subdev/fb/ramnvc0.o
 nouveau-y += core/subdev/fb/ramnve0.o
 nouveau-y += core/subdev/fb/ramgk20a.o
 nouveau-y += core/subdev/fb/ramgm107.o
+nouveau-y += core/subdev/fb/sddr2.o
 nouveau-y += core/subdev/fb/sddr3.o
 nouveau-y += core/subdev/fb/gddr5.o
+nouveau-y += core/subdev/fuse/base.o
+nouveau-y += core/subdev/fuse/g80.o
+nouveau-y += core/subdev/fuse/gf100.o
+nouveau-y += core/subdev/fuse/gm107.o
 nouveau-y += core/subdev/gpio/base.o
 nouveau-y += core/subdev/gpio/nv10.o
 nouveau-y += core/subdev/gpio/nv50.o
-nouveau-y += core/subdev/gpio/nv92.o
+nouveau-y += core/subdev/gpio/nv94.o
 nouveau-y += core/subdev/gpio/nvd0.o
 nouveau-y += core/subdev/gpio/nve0.o
 nouveau-y += core/subdev/i2c/base.o
@@ -149,8 +161,10 @@ nouveau-y += core/subdev/instmem/base.o
 nouveau-y += core/subdev/instmem/nv04.o
 nouveau-y += core/subdev/instmem/nv40.o
 nouveau-y += core/subdev/instmem/nv50.o
-nouveau-y += core/subdev/ltcg/gf100.o
-nouveau-y += core/subdev/ltcg/gm107.o
+nouveau-y += core/subdev/ltc/base.o
+nouveau-y += core/subdev/ltc/gf100.o
+nouveau-y += core/subdev/ltc/gk104.o
+nouveau-y += core/subdev/ltc/gm107.o
 nouveau-y += core/subdev/mc/base.o
 nouveau-y += core/subdev/mc/nv04.o
 nouveau-y += core/subdev/mc/nv40.o
@@ -161,6 +175,7 @@ nouveau-y += core/subdev/mc/nv94.o
 nouveau-y += core/subdev/mc/nv98.o
 nouveau-y += core/subdev/mc/nvc0.o
 nouveau-y += core/subdev/mc/nvc3.o
+nouveau-y += core/subdev/mc/gk20a.o
 nouveau-y += core/subdev/mxm/base.o
 nouveau-y += core/subdev/mxm/mxms.o
 nouveau-y += core/subdev/mxm/nv50.o
@@ -169,6 +184,7 @@ nouveau-y += core/subdev/pwr/memx.o
 nouveau-y += core/subdev/pwr/nva3.o
 nouveau-y += core/subdev/pwr/nvc0.o
 nouveau-y += core/subdev/pwr/nvd0.o
+nouveau-y += core/subdev/pwr/gk104.o
 nouveau-y += core/subdev/pwr/nv108.o
 nouveau-y += core/subdev/therm/base.o
 nouveau-y += core/subdev/therm/fan.o
@@ -182,6 +198,7 @@ nouveau-y += core/subdev/therm/nv50.o
 nouveau-y += core/subdev/therm/nv84.o
 nouveau-y += core/subdev/therm/nva3.o
 nouveau-y += core/subdev/therm/nvd0.o
+nouveau-y += core/subdev/therm/gm107.o
 nouveau-y += core/subdev/timer/base.o
 nouveau-y += core/subdev/timer/nv04.o
 nouveau-y += core/subdev/timer/gk20a.o
@@ -211,6 +228,7 @@ nouveau-y += core/engine/copy/nvc0.o
 nouveau-y += core/engine/copy/nve0.o
 nouveau-y += core/engine/crypt/nv84.o
 nouveau-y += core/engine/crypt/nv98.o
+nouveau-y += core/engine/device/acpi.o
 nouveau-y += core/engine/device/base.o
 nouveau-y += core/engine/device/ctrl.o
 nouveau-y += core/engine/device/nv04.o
@@ -243,6 +261,7 @@ nouveau-y += core/engine/disp/hdanvd0.o
 nouveau-y += core/engine/disp/hdminv84.o
 nouveau-y += core/engine/disp/hdminva3.o
 nouveau-y += core/engine/disp/hdminvd0.o
+nouveau-y += core/engine/disp/hdminve0.o
 nouveau-y += core/engine/disp/piornv50.o
 nouveau-y += core/engine/disp/sornv50.o
 nouveau-y += core/engine/disp/sornv94.o
@@ -270,6 +289,7 @@ nouveau-y += core/engine/graph/ctxnvd9.o
 nouveau-y += core/engine/graph/ctxnve4.o
 nouveau-y += core/engine/graph/ctxgk20a.o
 nouveau-y += core/engine/graph/ctxnvf0.o
+nouveau-y += core/engine/graph/ctxgk110b.o
 nouveau-y += core/engine/graph/ctxnv108.o
 nouveau-y += core/engine/graph/ctxgm107.o
 nouveau-y += core/engine/graph/nv04.o
@@ -291,6 +311,7 @@ nouveau-y += core/engine/graph/nvd9.o
 nouveau-y += core/engine/graph/nve4.o
 nouveau-y += core/engine/graph/gk20a.o
 nouveau-y += core/engine/graph/nvf0.o
+nouveau-y += core/engine/graph/gk110b.o
 nouveau-y += core/engine/graph/nv108.o
 nouveau-y += core/engine/graph/gm107.o
 nouveau-y += core/engine/mpeg/nv31.o
@@ -318,11 +339,18 @@ nouveau-y += core/engine/vp/nv98.o
 nouveau-y += core/engine/vp/nvc0.o
 nouveau-y += core/engine/vp/nve0.o
 
+# nvif
+nouveau-y += nvif/object.o
+nouveau-y += nvif/client.o
+nouveau-y += nvif/device.o
+nouveau-y += nvif/notify.o
+
 # drm/core
 nouveau-y += nouveau_drm.o nouveau_chan.o nouveau_dma.o nouveau_fence.o
 nouveau-y += nouveau_vga.o nouveau_agp.o
 nouveau-y += nouveau_ttm.o nouveau_sgdma.o nouveau_bo.o nouveau_gem.o
 nouveau-y += nouveau_prime.o nouveau_abi16.o
+nouveau-y += nouveau_nvif.o nouveau_usif.o
 nouveau-y += nv04_fence.o nv10_fence.o nv17_fence.o
 nouveau-y += nv50_fence.o nv84_fence.o nvc0_fence.o
 
@@ -349,3 +377,6 @@ nouveau-$(CONFIG_DRM_NOUVEAU_BACKLIGHT) += nouveau_backlight.o
 nouveau-$(CONFIG_DEBUG_FS) += nouveau_debugfs.o
 
 obj-$(CONFIG_DRM_NOUVEAU)+= nouveau.o
+
+# platform driver
+obj-$(CONFIG_NOUVEAU_PLATFORM_DRIVER) += nouveau_platform.o
diff --git a/drivers/gpu/drm/nouveau/core/core/client.c b/drivers/gpu/drm/nouveau/core/core/client.c
index 9079c0ac58e6..e962433294c3 100644
--- a/drivers/gpu/drm/nouveau/core/core/client.c
+++ b/drivers/gpu/drm/nouveau/core/core/client.c
@@ -26,13 +26,168 @@
 #include <core/client.h>
 #include <core/handle.h>
 #include <core/option.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+
+#include <nvif/unpack.h>
+#include <nvif/event.h>
 
 #include <engine/device.h>
 
+struct nvkm_client_notify {
+	struct nouveau_client *client;
+	struct nvkm_notify n;
+	u8 version;
+	u8 size;
+	union {
+		struct nvif_notify_rep_v0 v0;
+	} rep;
+};
+
+static int
+nvkm_client_notify(struct nvkm_notify *n)
+{
+	struct nvkm_client_notify *notify = container_of(n, typeof(*notify), n);
+	struct nouveau_client *client = notify->client;
+	return client->ntfy(&notify->rep, notify->size, n->data, n->size);
+}
+
+int
+nvkm_client_notify_put(struct nouveau_client *client, int index)
+{
+	if (index < ARRAY_SIZE(client->notify)) {
+		if (client->notify[index]) {
+			nvkm_notify_put(&client->notify[index]->n);
+			return 0;
+		}
+	}
+	return -ENOENT;
+}
+
+int
+nvkm_client_notify_get(struct nouveau_client *client, int index)
+{
+	if (index < ARRAY_SIZE(client->notify)) {
+		if (client->notify[index]) {
+			nvkm_notify_get(&client->notify[index]->n);
+			return 0;
+		}
+	}
+	return -ENOENT;
+}
+
+int
+nvkm_client_notify_del(struct nouveau_client *client, int index)
+{
+	if (index < ARRAY_SIZE(client->notify)) {
+		if (client->notify[index]) {
+			nvkm_notify_fini(&client->notify[index]->n);
+			kfree(client->notify[index]);
+			client->notify[index] = NULL;
+			return 0;
+		}
+	}
+	return -ENOENT;
+}
+
+int
+nvkm_client_notify_new(struct nouveau_object *object,
+		       struct nvkm_event *event, void *data, u32 size)
+{
+	struct nouveau_client *client = nouveau_client(object);
+	struct nvkm_client_notify *notify;
+	union {
+		struct nvif_notify_req_v0 v0;
+	} *req = data;
+	u8  index, reply;
+	int ret;
+
+	for (index = 0; index < ARRAY_SIZE(client->notify); index++) {
+		if (!client->notify[index])
+			break;
+	}
+
+	if (index == ARRAY_SIZE(client->notify))
+		return -ENOSPC;
+
+	notify = kzalloc(sizeof(*notify), GFP_KERNEL);
+	if (!notify)
+		return -ENOMEM;
+
+	nv_ioctl(client, "notify new size %d\n", size);
+	if (nvif_unpack(req->v0, 0, 0, true)) {
+		nv_ioctl(client, "notify new vers %d reply %d route %02x "
+				 "token %llx\n", req->v0.version,
+			 req->v0.reply, req->v0.route, req->v0.token);
+		notify->version = req->v0.version;
+		notify->size = sizeof(notify->rep.v0);
+		notify->rep.v0.version = req->v0.version;
+		notify->rep.v0.route = req->v0.route;
+		notify->rep.v0.token = req->v0.token;
+		reply = req->v0.reply;
+	}
+
+	if (ret == 0) {
+		ret = nvkm_notify_init(object, event, nvkm_client_notify,
+				       false, data, size, reply, &notify->n);
+		if (ret == 0) {
+			client->notify[index] = notify;
+			notify->client = client;
+			return index;
+		}
+	}
+
+	kfree(notify);
+	return ret;
+}
+
+static int
+nouveau_client_devlist(struct nouveau_object *object, void *data, u32 size)
+{
+	union {
+		struct nv_client_devlist_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "client devlist size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "client devlist vers %d count %d\n",
+			 args->v0.version, args->v0.count);
+		if (size == sizeof(args->v0.device[0]) * args->v0.count) {
+			ret = nouveau_device_list(args->v0.device,
+						  args->v0.count);
+			if (ret >= 0) {
+				args->v0.count = ret;
+				ret = 0;
+			}
+		} else {
+			ret = -EINVAL;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nouveau_client_mthd(struct nouveau_object *object, u32 mthd,
+		    void *data, u32 size)
+{
+	switch (mthd) {
+	case NV_CLIENT_DEVLIST:
+		return nouveau_client_devlist(object, data, size);
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
 static void
 nouveau_client_dtor(struct nouveau_object *object)
 {
 	struct nouveau_client *client = (void *)object;
+	int i;
+	for (i = 0; i < ARRAY_SIZE(client->notify); i++)
+		nvkm_client_notify_del(client, i);
 	nouveau_object_ref(NULL, &client->device);
 	nouveau_handle_destroy(client->root);
 	nouveau_namedb_destroy(&client->base);
@@ -42,6 +197,7 @@ static struct nouveau_oclass
 nouveau_client_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.dtor = nouveau_client_dtor,
+		.mthd = nouveau_client_mthd,
 	},
 };
 
@@ -93,9 +249,12 @@ int
 nouveau_client_fini(struct nouveau_client *client, bool suspend)
 {
 	const char *name[2] = { "fini", "suspend" };
-	int ret;
-
+	int ret, i;
 	nv_debug(client, "%s running\n", name[suspend]);
+	nv_debug(client, "%s notify\n", name[suspend]);
+	for (i = 0; i < ARRAY_SIZE(client->notify); i++)
+		nvkm_client_notify_put(client, i);
+	nv_debug(client, "%s object\n", name[suspend]);
 	ret = nouveau_handle_fini(client->root, suspend);
 	nv_debug(client, "%s completed with %d\n", name[suspend], ret);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/core/core/event.c b/drivers/gpu/drm/nouveau/core/core/event.c
index ae81d3b5d8b7..ff2b434b3db4 100644
--- a/drivers/gpu/drm/nouveau/core/core/event.c
+++ b/drivers/gpu/drm/nouveau/core/core/event.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2013 Red Hat Inc.
+ * Copyright 2013-2014 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,177 +20,81 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <core/os.h>
+#include <core/object.h>
 #include <core/event.h>
 
 void
-nouveau_event_put(struct nouveau_eventh *handler)
+nvkm_event_put(struct nvkm_event *event, u32 types, int index)
 {
-	struct nouveau_event *event = handler->event;
-	unsigned long flags;
-	u32 m, t;
-
-	if (!__test_and_clear_bit(NVKM_EVENT_ENABLE, &handler->flags))
-		return;
-
-	spin_lock_irqsave(&event->refs_lock, flags);
-	for (m = handler->types; t = __ffs(m), m; m &= ~(1 << t)) {
-		if (!--event->refs[handler->index * event->types_nr + t]) {
-			if (event->disable)
-				event->disable(event, 1 << t, handler->index);
+	BUG_ON(!spin_is_locked(&event->refs_lock));
+	while (types) {
+		int type = __ffs(types); types &= ~(1 << type);
+		if (--event->refs[index * event->types_nr + type] == 0) {
+			if (event->func->fini)
+				event->func->fini(event, 1 << type, index);
 		}
-
 	}
-	spin_unlock_irqrestore(&event->refs_lock, flags);
 }
 
 void
-nouveau_event_get(struct nouveau_eventh *handler)
+nvkm_event_get(struct nvkm_event *event, u32 types, int index)
 {
-	struct nouveau_event *event = handler->event;
-	unsigned long flags;
-	u32 m, t;
-
-	if (__test_and_set_bit(NVKM_EVENT_ENABLE, &handler->flags))
-		return;
-
-	spin_lock_irqsave(&event->refs_lock, flags);
-	for (m = handler->types; t = __ffs(m), m; m &= ~(1 << t)) {
-		if (!event->refs[handler->index * event->types_nr + t]++) {
-			if (event->enable)
-				event->enable(event, 1 << t, handler->index);
+	BUG_ON(!spin_is_locked(&event->refs_lock));
+	while (types) {
+		int type = __ffs(types); types &= ~(1 << type);
+		if (++event->refs[index * event->types_nr + type] == 1) {
+			if (event->func->init)
+				event->func->init(event, 1 << type, index);
 		}
-
 	}
-	spin_unlock_irqrestore(&event->refs_lock, flags);
-}
-
-static void
-nouveau_event_fini(struct nouveau_eventh *handler)
-{
-	struct nouveau_event *event = handler->event;
-	unsigned long flags;
-	nouveau_event_put(handler);
-	spin_lock_irqsave(&event->list_lock, flags);
-	list_del(&handler->head);
-	spin_unlock_irqrestore(&event->list_lock, flags);
-}
-
-static int
-nouveau_event_init(struct nouveau_event *event, u32 types, int index,
-		   int (*func)(void *, u32, int), void *priv,
-		   struct nouveau_eventh *handler)
-{
-	unsigned long flags;
-
-	if (types & ~((1 << event->types_nr) - 1))
-		return -EINVAL;
-	if (index >= event->index_nr)
-		return -EINVAL;
-
-	handler->event = event;
-	handler->flags = 0;
-	handler->types = types;
-	handler->index = index;
-	handler->func = func;
-	handler->priv = priv;
-
-	spin_lock_irqsave(&event->list_lock, flags);
-	list_add_tail(&handler->head, &event->list[index]);
-	spin_unlock_irqrestore(&event->list_lock, flags);
-	return 0;
-}
-
-int
-nouveau_event_new(struct nouveau_event *event, u32 types, int index,
-		  int (*func)(void *, u32, int), void *priv,
-		  struct nouveau_eventh **phandler)
-{
-	struct nouveau_eventh *handler;
-	int ret = -ENOMEM;
-
-	if (event->check) {
-		ret = event->check(event, types, index);
-		if (ret)
-			return ret;
-	}
-
-	handler = *phandler = kmalloc(sizeof(*handler), GFP_KERNEL);
-	if (handler) {
-		ret = nouveau_event_init(event, types, index, func, priv, handler);
-		if (ret)
-			kfree(handler);
-	}
-
-	return ret;
-}
-
-void
-nouveau_event_ref(struct nouveau_eventh *handler, struct nouveau_eventh **ref)
-{
-	BUG_ON(handler != NULL);
-	if (*ref) {
-		nouveau_event_fini(*ref);
-		kfree(*ref);
-	}
-	*ref = handler;
 }
 
 void
-nouveau_event_trigger(struct nouveau_event *event, u32 types, int index)
+nvkm_event_send(struct nvkm_event *event, u32 types, int index,
+		void *data, u32 size)
 {
-	struct nouveau_eventh *handler;
+	struct nvkm_notify *notify;
 	unsigned long flags;
 
-	if (WARN_ON(index >= event->index_nr))
+	if (!event->refs || WARN_ON(index >= event->index_nr))
 		return;
 
 	spin_lock_irqsave(&event->list_lock, flags);
-	list_for_each_entry(handler, &event->list[index], head) {
-		if (!test_bit(NVKM_EVENT_ENABLE, &handler->flags))
-			continue;
-		if (!(handler->types & types))
-			continue;
-		if (handler->func(handler->priv, handler->types & types, index)
-				!= NVKM_EVENT_DROP)
-			continue;
-		nouveau_event_put(handler);
+	list_for_each_entry(notify, &event->list, head) {
+		if (notify->index == index && (notify->types & types)) {
+			if (event->func->send) {
+				event->func->send(data, size, notify);
+				continue;
+			}
+			nvkm_notify_send(notify, data, size);
+		}
 	}
 	spin_unlock_irqrestore(&event->list_lock, flags);
 }
 
 void
-nouveau_event_destroy(struct nouveau_event **pevent)
+nvkm_event_fini(struct nvkm_event *event)
 {
-	struct nouveau_event *event = *pevent;
-	if (event) {
-		kfree(event);
-		*pevent = NULL;
+	if (event->refs) {
+		kfree(event->refs);
+		event->refs = NULL;
 	}
 }
 
 int
-nouveau_event_create(int types_nr, int index_nr, struct nouveau_event **pevent)
+nvkm_event_init(const struct nvkm_event_func *func, int types_nr, int index_nr,
+		struct nvkm_event *event)
 {
-	struct nouveau_event *event;
-	int i;
-
-	event = *pevent = kzalloc(sizeof(*event) + (index_nr * types_nr) *
-				  sizeof(event->refs[0]), GFP_KERNEL);
-	if (!event)
-		return -ENOMEM;
-
-	event->list = kmalloc(sizeof(*event->list) * index_nr, GFP_KERNEL);
-	if (!event->list) {
-		kfree(event);
+	event->refs = kzalloc(sizeof(*event->refs) * index_nr * types_nr,
+			      GFP_KERNEL);
+	if (!event->refs)
 		return -ENOMEM;
-	}
 
-	spin_lock_init(&event->list_lock);
-	spin_lock_init(&event->refs_lock);
-	for (i = 0; i < index_nr; i++)
-		INIT_LIST_HEAD(&event->list[i]);
+	event->func = func;
 	event->types_nr = types_nr;
 	event->index_nr = index_nr;
+	spin_lock_init(&event->refs_lock);
+	spin_lock_init(&event->list_lock);
+	INIT_LIST_HEAD(&event->list);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/core/gpuobj.c b/drivers/gpu/drm/nouveau/core/core/gpuobj.c
index 560b2214cf1c..daee87702502 100644
--- a/drivers/gpu/drm/nouveau/core/core/gpuobj.c
+++ b/drivers/gpu/drm/nouveau/core/core/gpuobj.c
@@ -115,7 +115,7 @@ nouveau_gpuobj_create_(struct nouveau_object *parent,
 	gpuobj->size = size;
 
 	if (heap) {
-		ret = nouveau_mm_head(heap, 1, size, size,
+		ret = nouveau_mm_head(heap, 0, 1, size, size,
 				      max(align, (u32)1), &gpuobj->node);
 		if (ret)
 			return ret;
diff --git a/drivers/gpu/drm/nouveau/core/core/handle.c b/drivers/gpu/drm/nouveau/core/core/handle.c
index 264c2b338ac3..a490b805d7e3 100644
--- a/drivers/gpu/drm/nouveau/core/core/handle.c
+++ b/drivers/gpu/drm/nouveau/core/core/handle.c
@@ -146,9 +146,7 @@ nouveau_handle_create(struct nouveau_object *parent, u32 _parent, u32 _handle,
 	}
 
 	hprintk(handle, TRACE, "created\n");
-
 	*phandle = handle;
-
 	return 0;
 }
 
@@ -224,3 +222,116 @@ nouveau_handle_put(struct nouveau_handle *handle)
 	if (handle)
 		nouveau_namedb_put(handle);
 }
+
+int
+nouveau_handle_new(struct nouveau_object *client, u32 _parent, u32 _handle,
+		   u16 _oclass, void *data, u32 size,
+		   struct nouveau_object **pobject)
+{
+	struct nouveau_object *parent = NULL;
+	struct nouveau_object *engctx = NULL;
+	struct nouveau_object *object = NULL;
+	struct nouveau_object *engine;
+	struct nouveau_oclass *oclass;
+	struct nouveau_handle *handle;
+	int ret;
+
+	/* lookup parent object and ensure it *is* a parent */
+	parent = nouveau_handle_ref(client, _parent);
+	if (!parent) {
+		nv_error(client, "parent 0x%08x not found\n", _parent);
+		return -ENOENT;
+	}
+
+	if (!nv_iclass(parent, NV_PARENT_CLASS)) {
+		nv_error(parent, "cannot have children\n");
+		ret = -EINVAL;
+		goto fail_class;
+	}
+
+	/* check that parent supports the requested subclass */
+	ret = nouveau_parent_sclass(parent, _oclass, &engine, &oclass);
+	if (ret) {
+		nv_debug(parent, "illegal class 0x%04x\n", _oclass);
+		goto fail_class;
+	}
+
+	/* make sure engine init has been completed *before* any objects
+	 * it controls are created - the constructors may depend on
+	 * state calculated at init (ie. default context construction)
+	 */
+	if (engine) {
+		ret = nouveau_object_inc(engine);
+		if (ret)
+			goto fail_class;
+	}
+
+	/* if engine requires it, create a context object to insert
+	 * between the parent and its children (eg. PGRAPH context)
+	 */
+	if (engine && nv_engine(engine)->cclass) {
+		ret = nouveau_object_ctor(parent, engine,
+					  nv_engine(engine)->cclass,
+					  data, size, &engctx);
+		if (ret)
+			goto fail_engctx;
+	} else {
+		nouveau_object_ref(parent, &engctx);
+	}
+
+	/* finally, create new object and bind it to its handle */
+	ret = nouveau_object_ctor(engctx, engine, oclass, data, size, &object);
+	*pobject = object;
+	if (ret)
+		goto fail_ctor;
+
+	ret = nouveau_object_inc(object);
+	if (ret)
+		goto fail_init;
+
+	ret = nouveau_handle_create(parent, _parent, _handle, object, &handle);
+	if (ret)
+		goto fail_handle;
+
+	ret = nouveau_handle_init(handle);
+	if (ret)
+		nouveau_handle_destroy(handle);
+
+fail_handle:
+	nouveau_object_dec(object, false);
+fail_init:
+	nouveau_object_ref(NULL, &object);
+fail_ctor:
+	nouveau_object_ref(NULL, &engctx);
+fail_engctx:
+	if (engine)
+		nouveau_object_dec(engine, false);
+fail_class:
+	nouveau_object_ref(NULL, &parent);
+	return ret;
+}
+
+int
+nouveau_handle_del(struct nouveau_object *client, u32 _parent, u32 _handle)
+{
+	struct nouveau_object *parent = NULL;
+	struct nouveau_object *namedb = NULL;
+	struct nouveau_handle *handle = NULL;
+
+	parent = nouveau_handle_ref(client, _parent);
+	if (!parent)
+		return -ENOENT;
+
+	namedb = nv_pclass(parent, NV_NAMEDB_CLASS);
+	if (namedb) {
+		handle = nouveau_namedb_get(nv_namedb(namedb), _handle);
+		if (handle) {
+			nouveau_namedb_put(handle);
+			nouveau_handle_fini(handle, false);
+			nouveau_handle_destroy(handle);
+		}
+	}
+
+	nouveau_object_ref(NULL, &parent);
+	return handle ? 0 : -EINVAL;
+}
diff --git a/drivers/gpu/drm/nouveau/core/core/ioctl.c b/drivers/gpu/drm/nouveau/core/core/ioctl.c
new file mode 100644
index 000000000000..692aa92dd850
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/core/ioctl.c
@@ -0,0 +1,530 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <core/object.h>
+#include <core/parent.h>
+#include <core/handle.h>
+#include <core/namedb.h>
+#include <core/client.h>
+#include <core/device.h>
+#include <core/ioctl.h>
+#include <core/event.h>
+
+#include <nvif/unpack.h>
+#include <nvif/ioctl.h>
+
+static int
+nvkm_ioctl_nop(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_nop none;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "nop size %d\n", size);
+	if (nvif_unvers(args->none)) {
+		nv_ioctl(object, "nop\n");
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_sclass(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_sclass_v0 v0;
+	} *args = data;
+	int ret;
+
+	if (!nv_iclass(object, NV_PARENT_CLASS)) {
+		nv_debug(object, "cannot have children (sclass)\n");
+		return -ENODEV;
+	}
+
+	nv_ioctl(object, "sclass size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "sclass vers %d count %d\n",
+			 args->v0.version, args->v0.count);
+		if (size == args->v0.count * sizeof(args->v0.oclass[0])) {
+			ret = nouveau_parent_lclass(object, args->v0.oclass,
+							    args->v0.count);
+			if (ret >= 0) {
+				args->v0.count = ret;
+				ret = 0;
+			}
+		} else {
+			ret = -EINVAL;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_new(struct nouveau_handle *parent, void *data, u32 size)
+{
+	union {
+		struct nvif_ioctl_new_v0 v0;
+	} *args = data;
+	struct nouveau_client *client = nouveau_client(parent->object);
+	struct nouveau_object *engctx = NULL;
+	struct nouveau_object *object = NULL;
+	struct nouveau_object *engine;
+	struct nouveau_oclass *oclass;
+	struct nouveau_handle *handle;
+	u32 _handle, _oclass;
+	int ret;
+
+	nv_ioctl(client, "new size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		_handle = args->v0.handle;
+		_oclass = args->v0.oclass;
+	} else
+		return ret;
+
+	nv_ioctl(client, "new vers %d handle %08x class %08x "
+			 "route %02x token %llx\n",
+		args->v0.version, _handle, _oclass,
+		args->v0.route, args->v0.token);
+
+	if (!nv_iclass(parent->object, NV_PARENT_CLASS)) {
+		nv_debug(parent->object, "cannot have children (ctor)\n");
+		ret = -ENODEV;
+		goto fail_class;
+	}
+
+	/* check that parent supports the requested subclass */
+	ret = nouveau_parent_sclass(parent->object, _oclass, &engine, &oclass);
+	if (ret) {
+		nv_debug(parent->object, "illegal class 0x%04x\n", _oclass);
+		goto fail_class;
+	}
+
+	/* make sure engine init has been completed *before* any objects
+	 * it controls are created - the constructors may depend on
+	 * state calculated at init (ie. default context construction)
+	 */
+	if (engine) {
+		ret = nouveau_object_inc(engine);
+		if (ret)
+			goto fail_class;
+	}
+
+	/* if engine requires it, create a context object to insert
+	 * between the parent and its children (eg. PGRAPH context)
+	 */
+	if (engine && nv_engine(engine)->cclass) {
+		ret = nouveau_object_ctor(parent->object, engine,
+					  nv_engine(engine)->cclass,
+					  data, size, &engctx);
+		if (ret)
+			goto fail_engctx;
+	} else {
+		nouveau_object_ref(parent->object, &engctx);
+	}
+
+	/* finally, create new object and bind it to its handle */
+	ret = nouveau_object_ctor(engctx, engine, oclass, data, size, &object);
+	client->data = object;
+	if (ret)
+		goto fail_ctor;
+
+	ret = nouveau_object_inc(object);
+	if (ret)
+		goto fail_init;
+
+	ret = nouveau_handle_create(parent->object, parent->name,
+				    _handle, object, &handle);
+	if (ret)
+		goto fail_handle;
+
+	ret = nouveau_handle_init(handle);
+	handle->route = args->v0.route;
+	handle->token = args->v0.token;
+	if (ret)
+		nouveau_handle_destroy(handle);
+
+fail_handle:
+	nouveau_object_dec(object, false);
+fail_init:
+	nouveau_object_ref(NULL, &object);
+fail_ctor:
+	nouveau_object_ref(NULL, &engctx);
+fail_engctx:
+	if (engine)
+		nouveau_object_dec(engine, false);
+fail_class:
+	return ret;
+}
+
+static int
+nvkm_ioctl_del(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_del none;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "delete size %d\n", size);
+	if (nvif_unvers(args->none)) {
+		nv_ioctl(object, "delete\n");
+		nouveau_handle_fini(handle, false);
+		nouveau_handle_destroy(handle);
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_mthd(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
+	union {
+		struct nvif_ioctl_mthd_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "mthd size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "mthd vers %d mthd %02x\n",
+			 args->v0.version, args->v0.method);
+		if (ret = -ENODEV, ofuncs->mthd)
+			ret = ofuncs->mthd(object, args->v0.method, data, size);
+	}
+
+	return ret;
+}
+
+
+static int
+nvkm_ioctl_rd(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
+	union {
+		struct nvif_ioctl_rd_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "rd size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "rd vers %d size %d addr %016llx\n",
+			args->v0.version, args->v0.size, args->v0.addr);
+		switch (args->v0.size) {
+		case 1:
+			if (ret = -ENODEV, ofuncs->rd08) {
+				args->v0.data = nv_ro08(object, args->v0.addr);
+				ret = 0;
+			}
+			break;
+		case 2:
+			if (ret = -ENODEV, ofuncs->rd16) {
+				args->v0.data = nv_ro16(object, args->v0.addr);
+				ret = 0;
+			}
+			break;
+		case 4:
+			if (ret = -ENODEV, ofuncs->rd32) {
+				args->v0.data = nv_ro32(object, args->v0.addr);
+				ret = 0;
+			}
+			break;
+		default:
+			ret = -EINVAL;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_wr(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
+	union {
+		struct nvif_ioctl_wr_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "wr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "wr vers %d size %d addr %016llx data %08x\n",
+			 args->v0.version, args->v0.size, args->v0.addr,
+			 args->v0.data);
+		switch (args->v0.size) {
+		case 1:
+			if (ret = -ENODEV, ofuncs->wr08) {
+				nv_wo08(object, args->v0.addr, args->v0.data);
+				ret = 0;
+			}
+			break;
+		case 2:
+			if (ret = -ENODEV, ofuncs->wr16) {
+				nv_wo16(object, args->v0.addr, args->v0.data);
+				ret = 0;
+			}
+			break;
+		case 4:
+			if (ret = -ENODEV, ofuncs->wr32) {
+				nv_wo32(object, args->v0.addr, args->v0.data);
+				ret = 0;
+			}
+			break;
+		default:
+			ret = -EINVAL;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_map(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
+	union {
+		struct nvif_ioctl_map_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "map size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "map vers %d\n", args->v0.version);
+		if (ret = -ENODEV, ofuncs->map) {
+			ret = ofuncs->map(object, &args->v0.handle,
+						  &args->v0.length);
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_unmap(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_unmap none;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "unmap size %d\n", size);
+	if (nvif_unvers(args->none)) {
+		nv_ioctl(object, "unmap\n");
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_ntfy_new(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_object *object = handle->object;
+	struct nouveau_ofuncs *ofuncs = object->oclass->ofuncs;
+	union {
+		struct nvif_ioctl_ntfy_new_v0 v0;
+	} *args = data;
+	struct nvkm_event *event;
+	int ret;
+
+	nv_ioctl(object, "ntfy new size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "ntfy new vers %d event %02x\n",
+			 args->v0.version, args->v0.event);
+		if (ret = -ENODEV, ofuncs->ntfy)
+			ret = ofuncs->ntfy(object, args->v0.event, &event);
+		if (ret == 0) {
+			ret = nvkm_client_notify_new(object, event, data, size);
+			if (ret >= 0) {
+				args->v0.index = ret;
+				ret = 0;
+			}
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_ntfy_del(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_client *client = nouveau_client(handle->object);
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_ntfy_del_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "ntfy del size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "ntfy del vers %d index %d\n",
+			 args->v0.version, args->v0.index);
+		ret = nvkm_client_notify_del(client, args->v0.index);
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_ntfy_get(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_client *client = nouveau_client(handle->object);
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_ntfy_get_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "ntfy get size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "ntfy get vers %d index %d\n",
+			 args->v0.version, args->v0.index);
+		ret = nvkm_client_notify_get(client, args->v0.index);
+	}
+
+	return ret;
+}
+
+static int
+nvkm_ioctl_ntfy_put(struct nouveau_handle *handle, void *data, u32 size)
+{
+	struct nouveau_client *client = nouveau_client(handle->object);
+	struct nouveau_object *object = handle->object;
+	union {
+		struct nvif_ioctl_ntfy_put_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "ntfy put size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "ntfy put vers %d index %d\n",
+			 args->v0.version, args->v0.index);
+		ret = nvkm_client_notify_put(client, args->v0.index);
+	}
+
+	return ret;
+}
+
+static struct {
+	int version;
+	int (*func)(struct nouveau_handle *, void *, u32);
+}
+nvkm_ioctl_v0[] = {
+	{ 0x00, nvkm_ioctl_nop },
+	{ 0x00, nvkm_ioctl_sclass },
+	{ 0x00, nvkm_ioctl_new },
+	{ 0x00, nvkm_ioctl_del },
+	{ 0x00, nvkm_ioctl_mthd },
+	{ 0x00, nvkm_ioctl_rd },
+	{ 0x00, nvkm_ioctl_wr },
+	{ 0x00, nvkm_ioctl_map },
+	{ 0x00, nvkm_ioctl_unmap },
+	{ 0x00, nvkm_ioctl_ntfy_new },
+	{ 0x00, nvkm_ioctl_ntfy_del },
+	{ 0x00, nvkm_ioctl_ntfy_get },
+	{ 0x00, nvkm_ioctl_ntfy_put },
+};
+
+static int
+nvkm_ioctl_path(struct nouveau_handle *parent, u32 type, u32 nr,
+		  u32 *path, void *data, u32 size,
+		  u8 owner, u8 *route, u64 *token)
+{
+	struct nouveau_handle *handle = parent;
+	struct nouveau_namedb *namedb;
+	struct nouveau_object *object;
+	int ret;
+
+	while ((object = parent->object), nr--) {
+		nv_ioctl(object, "path 0x%08x\n", path[nr]);
+		if (!nv_iclass(object, NV_PARENT_CLASS)) {
+			nv_debug(object, "cannot have children (path)\n");
+			return -EINVAL;
+		}
+
+		if (!(namedb = (void *)nv_pclass(object, NV_NAMEDB_CLASS)) ||
+		    !(handle = nouveau_namedb_get(namedb, path[nr]))) {
+			nv_debug(object, "handle 0x%08x not found\n", path[nr]);
+			return -ENOENT;
+		}
+		nouveau_namedb_put(handle);
+		parent = handle;
+	}
+
+	if (owner != NVIF_IOCTL_V0_OWNER_ANY &&
+	    owner != handle->route) {
+		nv_ioctl(object, "object route != owner\n");
+		return -EACCES;
+	}
+	*route = handle->route;
+	*token = handle->token;
+
+	if (ret = -EINVAL, type < ARRAY_SIZE(nvkm_ioctl_v0)) {
+		if (nvkm_ioctl_v0[type].version == 0) {
+			ret = nvkm_ioctl_v0[type].func(handle, data, size);
+		}
+	}
+
+	return ret;
+}
+
+int
+nvkm_ioctl(struct nouveau_client *client, bool supervisor,
+	   void *data, u32 size, void **hack)
+{
+	union {
+		struct nvif_ioctl_v0 v0;
+	} *args = data;
+	int ret;
+
+	client->super = supervisor;
+	nv_ioctl(client, "size %d\n", size);
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(client, "vers %d type %02x path %d owner %02x\n",
+			 args->v0.version, args->v0.type, args->v0.path_nr,
+			 args->v0.owner);
+		ret = nvkm_ioctl_path(client->root, args->v0.type,
+				      args->v0.path_nr, args->v0.path,
+				      data, size, args->v0.owner,
+				     &args->v0.route, &args->v0.token);
+	}
+
+	nv_ioctl(client, "return %d\n", ret);
+	if (hack) {
+		*hack = client->data;
+		client->data = NULL;
+	}
+	client->super = false;
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/core/mm.c b/drivers/gpu/drm/nouveau/core/core/mm.c
index 7a4e0891c5f8..b4f5db66d5b5 100644
--- a/drivers/gpu/drm/nouveau/core/core/mm.c
+++ b/drivers/gpu/drm/nouveau/core/core/mm.c
@@ -28,6 +28,24 @@
 #define node(root, dir) ((root)->nl_entry.dir == &mm->nodes) ? NULL : \
 	list_entry((root)->nl_entry.dir, struct nouveau_mm_node, nl_entry)
 
+static void
+nouveau_mm_dump(struct nouveau_mm *mm, const char *header)
+{
+	struct nouveau_mm_node *node;
+
+	printk(KERN_ERR "nouveau: %s\n", header);
+	printk(KERN_ERR "nouveau: node list:\n");
+	list_for_each_entry(node, &mm->nodes, nl_entry) {
+		printk(KERN_ERR "nouveau: \t%08x %08x %d\n",
+		       node->offset, node->length, node->type);
+	}
+	printk(KERN_ERR "nouveau: free list:\n");
+	list_for_each_entry(node, &mm->free, fl_entry) {
+		printk(KERN_ERR "nouveau: \t%08x %08x %d\n",
+		       node->offset, node->length, node->type);
+	}
+}
+
 void
 nouveau_mm_free(struct nouveau_mm *mm, struct nouveau_mm_node **pthis)
 {
@@ -37,29 +55,29 @@ nouveau_mm_free(struct nouveau_mm *mm, struct nouveau_mm_node **pthis)
 		struct nouveau_mm_node *prev = node(this, prev);
 		struct nouveau_mm_node *next = node(this, next);
 
-		if (prev && prev->type == 0) {
+		if (prev && prev->type == NVKM_MM_TYPE_NONE) {
 			prev->length += this->length;
 			list_del(&this->nl_entry);
 			kfree(this); this = prev;
 		}
 
-		if (next && next->type == 0) {
+		if (next && next->type == NVKM_MM_TYPE_NONE) {
 			next->offset  = this->offset;
 			next->length += this->length;
-			if (this->type == 0)
+			if (this->type == NVKM_MM_TYPE_NONE)
 				list_del(&this->fl_entry);
 			list_del(&this->nl_entry);
 			kfree(this); this = NULL;
 		}
 
-		if (this && this->type != 0) {
+		if (this && this->type != NVKM_MM_TYPE_NONE) {
 			list_for_each_entry(prev, &mm->free, fl_entry) {
 				if (this->offset < prev->offset)
 					break;
 			}
 
 			list_add_tail(&this->fl_entry, &prev->fl_entry);
-			this->type = 0;
+			this->type = NVKM_MM_TYPE_NONE;
 		}
 	}
 
@@ -80,27 +98,32 @@ region_head(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
 
 	b->offset = a->offset;
 	b->length = size;
+	b->heap   = a->heap;
 	b->type   = a->type;
 	a->offset += size;
 	a->length -= size;
 	list_add_tail(&b->nl_entry, &a->nl_entry);
-	if (b->type == 0)
+	if (b->type == NVKM_MM_TYPE_NONE)
 		list_add_tail(&b->fl_entry, &a->fl_entry);
 	return b;
 }
 
 int
-nouveau_mm_head(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
-		u32 align, struct nouveau_mm_node **pnode)
+nouveau_mm_head(struct nouveau_mm *mm, u8 heap, u8 type, u32 size_max,
+		u32 size_min, u32 align, struct nouveau_mm_node **pnode)
 {
 	struct nouveau_mm_node *prev, *this, *next;
 	u32 mask = align - 1;
 	u32 splitoff;
 	u32 s, e;
 
-	BUG_ON(!type);
+	BUG_ON(type == NVKM_MM_TYPE_NONE || type == NVKM_MM_TYPE_HOLE);
 
 	list_for_each_entry(this, &mm->free, fl_entry) {
+		if (unlikely(heap != NVKM_MM_HEAP_ANY)) {
+			if (this->heap != heap)
+				continue;
+		}
 		e = this->offset + this->length;
 		s = this->offset;
 
@@ -149,27 +172,32 @@ region_tail(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
 	a->length -= size;
 	b->offset  = a->offset + a->length;
 	b->length  = size;
+	b->heap    = a->heap;
 	b->type    = a->type;
 
 	list_add(&b->nl_entry, &a->nl_entry);
-	if (b->type == 0)
+	if (b->type == NVKM_MM_TYPE_NONE)
 		list_add(&b->fl_entry, &a->fl_entry);
 	return b;
 }
 
 int
-nouveau_mm_tail(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
-		u32 align, struct nouveau_mm_node **pnode)
+nouveau_mm_tail(struct nouveau_mm *mm, u8 heap, u8 type, u32 size_max,
+		u32 size_min, u32 align, struct nouveau_mm_node **pnode)
 {
 	struct nouveau_mm_node *prev, *this, *next;
 	u32 mask = align - 1;
 
-	BUG_ON(!type);
+	BUG_ON(type == NVKM_MM_TYPE_NONE || type == NVKM_MM_TYPE_HOLE);
 
 	list_for_each_entry_reverse(this, &mm->free, fl_entry) {
 		u32 e = this->offset + this->length;
 		u32 s = this->offset;
 		u32 c = 0, a;
+		if (unlikely(heap != NVKM_MM_HEAP_ANY)) {
+			if (this->heap != heap)
+				continue;
+		}
 
 		prev = node(this, prev);
 		if (prev && prev->type != type)
@@ -209,9 +237,23 @@ nouveau_mm_tail(struct nouveau_mm *mm, u8 type, u32 size_max, u32 size_min,
 int
 nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 {
-	struct nouveau_mm_node *node;
+	struct nouveau_mm_node *node, *prev;
+	u32 next;
 
-	if (block) {
+	if (nouveau_mm_initialised(mm)) {
+		prev = list_last_entry(&mm->nodes, typeof(*node), nl_entry);
+		next = prev->offset + prev->length;
+		if (next != offset) {
+			BUG_ON(next > offset);
+			if (!(node = kzalloc(sizeof(*node), GFP_KERNEL)))
+				return -ENOMEM;
+			node->type   = NVKM_MM_TYPE_HOLE;
+			node->offset = next;
+			node->length = offset - next;
+			list_add_tail(&node->nl_entry, &mm->nodes);
+		}
+		BUG_ON(block != mm->block_size);
+	} else {
 		INIT_LIST_HEAD(&mm->nodes);
 		INIT_LIST_HEAD(&mm->free);
 		mm->block_size = block;
@@ -230,25 +272,32 @@ nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 
 	list_add_tail(&node->nl_entry, &mm->nodes);
 	list_add_tail(&node->fl_entry, &mm->free);
-	mm->heap_nodes++;
+	node->heap = ++mm->heap_nodes;
 	return 0;
 }
 
 int
 nouveau_mm_fini(struct nouveau_mm *mm)
 {
-	if (nouveau_mm_initialised(mm)) {
-		struct nouveau_mm_node *node, *heap =
-			list_first_entry(&mm->nodes, typeof(*heap), nl_entry);
-		int nodes = 0;
+	struct nouveau_mm_node *node, *temp;
+	int nodes = 0;
 
-		list_for_each_entry(node, &mm->nodes, nl_entry) {
-			if (WARN_ON(nodes++ == mm->heap_nodes))
+	if (!nouveau_mm_initialised(mm))
+		return 0;
+
+	list_for_each_entry(node, &mm->nodes, nl_entry) {
+		if (node->type != NVKM_MM_TYPE_HOLE) {
+			if (++nodes > mm->heap_nodes) {
+				nouveau_mm_dump(mm, "mm not clean!");
 				return -EBUSY;
+			}
 		}
-
-		kfree(heap);
 	}
 
+	list_for_each_entry_safe(node, temp, &mm->nodes, nl_entry) {
+		list_del(&node->nl_entry);
+		kfree(node);
+	}
+	mm->heap_nodes = 0;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/core/notify.c b/drivers/gpu/drm/nouveau/core/core/notify.c
new file mode 100644
index 000000000000..d1bcde55e9d7
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/core/notify.c
@@ -0,0 +1,168 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <core/client.h>
+#include <core/event.h>
+#include <core/notify.h>
+
+#include <nvif/unpack.h>
+#include <nvif/event.h>
+
+static inline void
+nvkm_notify_put_locked(struct nvkm_notify *notify)
+{
+	if (notify->block++ == 0)
+		nvkm_event_put(notify->event, notify->types, notify->index);
+}
+
+void
+nvkm_notify_put(struct nvkm_notify *notify)
+{
+	struct nvkm_event *event = notify->event;
+	unsigned long flags;
+	if (likely(event) &&
+	    test_and_clear_bit(NVKM_NOTIFY_USER, &notify->flags)) {
+		spin_lock_irqsave(&event->refs_lock, flags);
+		nvkm_notify_put_locked(notify);
+		spin_unlock_irqrestore(&event->refs_lock, flags);
+		if (test_bit(NVKM_NOTIFY_WORK, &notify->flags))
+			flush_work(&notify->work);
+	}
+}
+
+static inline void
+nvkm_notify_get_locked(struct nvkm_notify *notify)
+{
+	if (--notify->block == 0)
+		nvkm_event_get(notify->event, notify->types, notify->index);
+}
+
+void
+nvkm_notify_get(struct nvkm_notify *notify)
+{
+	struct nvkm_event *event = notify->event;
+	unsigned long flags;
+	if (likely(event) &&
+	    !test_and_set_bit(NVKM_NOTIFY_USER, &notify->flags)) {
+		spin_lock_irqsave(&event->refs_lock, flags);
+		nvkm_notify_get_locked(notify);
+		spin_unlock_irqrestore(&event->refs_lock, flags);
+	}
+}
+
+static inline void
+nvkm_notify_func(struct nvkm_notify *notify)
+{
+	struct nvkm_event *event = notify->event;
+	int ret = notify->func(notify);
+	unsigned long flags;
+	if ((ret == NVKM_NOTIFY_KEEP) ||
+	    !test_and_clear_bit(NVKM_NOTIFY_USER, &notify->flags)) {
+		spin_lock_irqsave(&event->refs_lock, flags);
+		nvkm_notify_get_locked(notify);
+		spin_unlock_irqrestore(&event->refs_lock, flags);
+	}
+}
+
+static void
+nvkm_notify_work(struct work_struct *work)
+{
+	struct nvkm_notify *notify = container_of(work, typeof(*notify), work);
+	nvkm_notify_func(notify);
+}
+
+void
+nvkm_notify_send(struct nvkm_notify *notify, void *data, u32 size)
+{
+	struct nvkm_event *event = notify->event;
+	unsigned long flags;
+
+	BUG_ON(!spin_is_locked(&event->list_lock));
+	BUG_ON(size != notify->size);
+
+	spin_lock_irqsave(&event->refs_lock, flags);
+	if (notify->block) {
+		spin_unlock_irqrestore(&event->refs_lock, flags);
+		return;
+	}
+	nvkm_notify_put_locked(notify);
+	spin_unlock_irqrestore(&event->refs_lock, flags);
+
+	if (test_bit(NVKM_NOTIFY_WORK, &notify->flags)) {
+		memcpy((void *)notify->data, data, size);
+		schedule_work(&notify->work);
+	} else {
+		notify->data = data;
+		nvkm_notify_func(notify);
+		notify->data = NULL;
+	}
+}
+
+void
+nvkm_notify_fini(struct nvkm_notify *notify)
+{
+	unsigned long flags;
+	if (notify->event) {
+		nvkm_notify_put(notify);
+		spin_lock_irqsave(&notify->event->list_lock, flags);
+		list_del(&notify->head);
+		spin_unlock_irqrestore(&notify->event->list_lock, flags);
+		kfree((void *)notify->data);
+		notify->event = NULL;
+	}
+}
+
+int
+nvkm_notify_init(struct nouveau_object *object, struct nvkm_event *event,
+		 int (*func)(struct nvkm_notify *), bool work,
+		 void *data, u32 size, u32 reply,
+		 struct nvkm_notify *notify)
+{
+	unsigned long flags;
+	int ret = -ENODEV;
+	if ((notify->event = event), event->refs) {
+		ret = event->func->ctor(object, data, size, notify);
+		if (ret == 0 && (ret = -EINVAL, notify->size == reply)) {
+			notify->flags = 0;
+			notify->block = 1;
+			notify->func = func;
+			notify->data = NULL;
+			if (ret = 0, work) {
+				INIT_WORK(&notify->work, nvkm_notify_work);
+				set_bit(NVKM_NOTIFY_WORK, &notify->flags);
+				notify->data = kmalloc(reply, GFP_KERNEL);
+				if (!notify->data)
+					ret = -ENOMEM;
+			}
+		}
+		if (ret == 0) {
+			spin_lock_irqsave(&event->list_lock, flags);
+			list_add_tail(&notify->head, &event->list);
+			spin_unlock_irqrestore(&event->list_lock, flags);
+		}
+	}
+	if (ret)
+		notify->event = NULL;
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/core/object.c b/drivers/gpu/drm/nouveau/core/core/object.c
index 124538555904..b08630577c82 100644
--- a/drivers/gpu/drm/nouveau/core/core/object.c
+++ b/drivers/gpu/drm/nouveau/core/core/object.c
@@ -23,9 +23,6 @@
  */
 
 #include <core/object.h>
-#include <core/parent.h>
-#include <core/namedb.h>
-#include <core/handle.h>
 #include <core/engine.h>
 
 #ifdef NOUVEAU_OBJECT_MAGIC
@@ -61,21 +58,15 @@ nouveau_object_create_(struct nouveau_object *parent,
 	return 0;
 }
 
-static int
+int
 _nouveau_object_ctor(struct nouveau_object *parent,
 		     struct nouveau_object *engine,
 		     struct nouveau_oclass *oclass, void *data, u32 size,
 		     struct nouveau_object **pobject)
 {
-	struct nouveau_object *object;
-	int ret;
-
-	ret = nouveau_object_create(parent, engine, oclass, 0, &object);
-	*pobject = nv_object(object);
-	if (ret)
-		return ret;
-
-	return 0;
+	if (size != 0)
+		return -ENOSYS;
+	return nouveau_object_create(parent, engine, oclass, 0, pobject);
 }
 
 void
@@ -91,42 +82,24 @@ nouveau_object_destroy(struct nouveau_object *object)
 	kfree(object);
 }
 
-static void
-_nouveau_object_dtor(struct nouveau_object *object)
-{
-	nouveau_object_destroy(object);
-}
-
 int
 nouveau_object_init(struct nouveau_object *object)
 {
 	return 0;
 }
 
-static int
-_nouveau_object_init(struct nouveau_object *object)
-{
-	return nouveau_object_init(object);
-}
-
 int
 nouveau_object_fini(struct nouveau_object *object, bool suspend)
 {
 	return 0;
 }
 
-static int
-_nouveau_object_fini(struct nouveau_object *object, bool suspend)
-{
-	return nouveau_object_fini(object, suspend);
-}
-
 struct nouveau_ofuncs
 nouveau_object_ofuncs = {
 	.ctor = _nouveau_object_ctor,
-	.dtor = _nouveau_object_dtor,
-	.init = _nouveau_object_init,
-	.fini = _nouveau_object_fini,
+	.dtor = nouveau_object_destroy,
+	.init = nouveau_object_init,
+	.fini = nouveau_object_fini,
 };
 
 int
@@ -189,119 +162,6 @@ nouveau_object_ref(struct nouveau_object *obj, struct nouveau_object **ref)
 }
 
 int
-nouveau_object_new(struct nouveau_object *client, u32 _parent, u32 _handle,
-		   u16 _oclass, void *data, u32 size,
-		   struct nouveau_object **pobject)
-{
-	struct nouveau_object *parent = NULL;
-	struct nouveau_object *engctx = NULL;
-	struct nouveau_object *object = NULL;
-	struct nouveau_object *engine;
-	struct nouveau_oclass *oclass;
-	struct nouveau_handle *handle;
-	int ret;
-
-	/* lookup parent object and ensure it *is* a parent */
-	parent = nouveau_handle_ref(client, _parent);
-	if (!parent) {
-		nv_error(client, "parent 0x%08x not found\n", _parent);
-		return -ENOENT;
-	}
-
-	if (!nv_iclass(parent, NV_PARENT_CLASS)) {
-		nv_error(parent, "cannot have children\n");
-		ret = -EINVAL;
-		goto fail_class;
-	}
-
-	/* check that parent supports the requested subclass */
-	ret = nouveau_parent_sclass(parent, _oclass, &engine, &oclass);
-	if (ret) {
-		nv_debug(parent, "illegal class 0x%04x\n", _oclass);
-		goto fail_class;
-	}
-
-	/* make sure engine init has been completed *before* any objects
-	 * it controls are created - the constructors may depend on
-	 * state calculated at init (ie. default context construction)
-	 */
-	if (engine) {
-		ret = nouveau_object_inc(engine);
-		if (ret)
-			goto fail_class;
-	}
-
-	/* if engine requires it, create a context object to insert
-	 * between the parent and its children (eg. PGRAPH context)
-	 */
-	if (engine && nv_engine(engine)->cclass) {
-		ret = nouveau_object_ctor(parent, engine,
-					  nv_engine(engine)->cclass,
-					  data, size, &engctx);
-		if (ret)
-			goto fail_engctx;
-	} else {
-		nouveau_object_ref(parent, &engctx);
-	}
-
-	/* finally, create new object and bind it to its handle */
-	ret = nouveau_object_ctor(engctx, engine, oclass, data, size, &object);
-	*pobject = object;
-	if (ret)
-		goto fail_ctor;
-
-	ret = nouveau_object_inc(object);
-	if (ret)
-		goto fail_init;
-
-	ret = nouveau_handle_create(parent, _parent, _handle, object, &handle);
-	if (ret)
-		goto fail_handle;
-
-	ret = nouveau_handle_init(handle);
-	if (ret)
-		nouveau_handle_destroy(handle);
-
-fail_handle:
-	nouveau_object_dec(object, false);
-fail_init:
-	nouveau_object_ref(NULL, &object);
-fail_ctor:
-	nouveau_object_ref(NULL, &engctx);
-fail_engctx:
-	if (engine)
-		nouveau_object_dec(engine, false);
-fail_class:
-	nouveau_object_ref(NULL, &parent);
-	return ret;
-}
-
-int
-nouveau_object_del(struct nouveau_object *client, u32 _parent, u32 _handle)
-{
-	struct nouveau_object *parent = NULL;
-	struct nouveau_object *namedb = NULL;
-	struct nouveau_handle *handle = NULL;
-
-	parent = nouveau_handle_ref(client, _parent);
-	if (!parent)
-		return -ENOENT;
-
-	namedb = nv_pclass(parent, NV_NAMEDB_CLASS);
-	if (namedb) {
-		handle = nouveau_namedb_get(nv_namedb(namedb), _handle);
-		if (handle) {
-			nouveau_namedb_put(handle);
-			nouveau_handle_fini(handle, false);
-			nouveau_handle_destroy(handle);
-		}
-	}
-
-	nouveau_object_ref(NULL, &parent);
-	return handle ? 0 : -EINVAL;
-}
-
-int
 nouveau_object_inc(struct nouveau_object *object)
 {
 	int ref = atomic_add_return(1, &object->usecount);
diff --git a/drivers/gpu/drm/nouveau/core/core/parent.c b/drivers/gpu/drm/nouveau/core/core/parent.c
index dee5d1235e9b..30a2911878f8 100644
--- a/drivers/gpu/drm/nouveau/core/core/parent.c
+++ b/drivers/gpu/drm/nouveau/core/core/parent.c
@@ -75,6 +75,39 @@ nouveau_parent_sclass(struct nouveau_object *parent, u16 handle,
 }
 
 int
+nouveau_parent_lclass(struct nouveau_object *parent, u32 *lclass, int size)
+{
+	struct nouveau_sclass *sclass;
+	struct nouveau_engine *engine;
+	struct nouveau_oclass *oclass;
+	int nr = -1, i;
+	u64 mask;
+
+	sclass = nv_parent(parent)->sclass;
+	while (sclass) {
+		if (++nr < size)
+			lclass[nr] = sclass->oclass->handle & 0xffff;
+		sclass = sclass->sclass;
+	}
+
+	mask = nv_parent(parent)->engine;
+	while (i = __ffs64(mask), mask) {
+		engine = nouveau_engine(parent, i);
+		if (engine && (oclass = engine->sclass)) {
+			while (oclass->ofuncs) {
+				if (++nr < size)
+					lclass[nr] = oclass->handle & 0xffff;
+				oclass++;
+			}
+		}
+
+		mask &= ~(1ULL << i);
+	}
+
+	return nr + 1;
+}
+
+int
 nouveau_parent_create_(struct nouveau_object *parent,
 		       struct nouveau_object *engine,
 		       struct nouveau_oclass *oclass, u32 pclass,
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
index f31527733e00..abb410ef09ea 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
@@ -30,7 +30,6 @@
 #include <subdev/vm.h>
 
 #include <core/client.h>
-#include <core/class.h>
 #include <core/enum.h>
 
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
index ac3291f781f6..9261694d0d35 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
@@ -26,9 +26,7 @@
 #include <engine/fifo.h>
 #include <engine/copy.h>
 
-#include <core/class.h>
 #include <core/enum.h>
-#include <core/class.h>
 #include <core/enum.h>
 
 #include "fuc/nvc0.fuc.h"
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
index 748a61eb3c6f..c7194b354605 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
@@ -24,7 +24,6 @@
 
 #include <core/os.h>
 #include <core/enum.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <engine/copy.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index 2551dafbec73..ea5c42f31791 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -25,7 +25,6 @@
 #include <core/client.h>
 #include <core/os.h>
 #include <core/enum.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/gpuobj.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
index c7082377ec76..5571c09534cb 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
@@ -25,7 +25,6 @@
 #include <core/client.h>
 #include <core/os.h>
 #include <core/enum.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <subdev/timer.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/acpi.c b/drivers/gpu/drm/nouveau/core/engine/device/acpi.c
new file mode 100644
index 000000000000..4dbf0ba89e5c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/device/acpi.c
@@ -0,0 +1,59 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "acpi.h"
+
+#ifdef CONFIG_ACPI
+static int
+nvkm_acpi_ntfy(struct notifier_block *nb, unsigned long val, void *data)
+{
+	struct nouveau_device *device =
+		container_of(nb, typeof(*device), acpi.nb);
+	struct acpi_bus_event *info = data;
+
+	if (!strcmp(info->device_class, "ac_adapter"))
+		nvkm_event_send(&device->event, 1, 0, NULL, 0);
+
+	return NOTIFY_DONE;
+}
+#endif
+
+int
+nvkm_acpi_fini(struct nouveau_device *device, bool suspend)
+{
+#ifdef CONFIG_ACPI
+	unregister_acpi_notifier(&device->acpi.nb);
+#endif
+	return 0;
+}
+
+int
+nvkm_acpi_init(struct nouveau_device *device)
+{
+#ifdef CONFIG_ACPI
+	device->acpi.nb.notifier_call = nvkm_acpi_ntfy;
+	register_acpi_notifier(&device->acpi.nb);
+#endif
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/acpi.h b/drivers/gpu/drm/nouveau/core/engine/device/acpi.h
new file mode 100644
index 000000000000..cc49f4f568cd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/device/acpi.h
@@ -0,0 +1,9 @@
+#ifndef __NVKM_DEVICE_ACPI_H__
+#define __NVKM_DEVICE_ACPI_H__
+
+#include <engine/device.h>
+
+int nvkm_acpi_init(struct nouveau_device *);
+int nvkm_acpi_fini(struct nouveau_device *, bool);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/base.c b/drivers/gpu/drm/nouveau/core/engine/device/base.c
index 18c8c7245b73..0ef5a5713182 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/base.c
@@ -26,10 +26,14 @@
 #include <core/device.h>
 #include <core/client.h>
 #include <core/option.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
-#include <core/class.h>
+#include <subdev/fb.h>
+#include <subdev/instmem.h>
 
 #include "priv.h"
+#include "acpi.h"
 
 static DEFINE_MUTEX(nv_devices_mutex);
 static LIST_HEAD(nv_devices);
@@ -49,74 +53,258 @@ nouveau_device_find(u64 name)
 	return match;
 }
 
+int
+nouveau_device_list(u64 *name, int size)
+{
+	struct nouveau_device *device;
+	int nr = 0;
+	mutex_lock(&nv_devices_mutex);
+	list_for_each_entry(device, &nv_devices, head) {
+		if (nr++ < size)
+			name[nr - 1] = device->handle;
+	}
+	mutex_unlock(&nv_devices_mutex);
+	return nr;
+}
+
 /******************************************************************************
  * nouveau_devobj (0x0080): class implementation
  *****************************************************************************/
+
 struct nouveau_devobj {
 	struct nouveau_parent base;
 	struct nouveau_object *subdev[NVDEV_SUBDEV_NR];
 };
 
+static int
+nouveau_devobj_info(struct nouveau_object *object, void *data, u32 size)
+{
+	struct nouveau_device *device = nv_device(object);
+	struct nouveau_fb *pfb = nouveau_fb(device);
+	struct nouveau_instmem *imem = nouveau_instmem(device);
+	union {
+		struct nv_device_info_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "device info size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "device info vers %d\n", args->v0.version);
+	} else
+		return ret;
+
+	switch (device->chipset) {
+	case 0x01a:
+	case 0x01f:
+	case 0x04c:
+	case 0x04e:
+	case 0x063:
+	case 0x067:
+	case 0x068:
+	case 0x0aa:
+	case 0x0ac:
+	case 0x0af:
+		args->v0.platform = NV_DEVICE_INFO_V0_IGP;
+		break;
+	default:
+		if (device->pdev) {
+			if (pci_find_capability(device->pdev, PCI_CAP_ID_AGP))
+				args->v0.platform = NV_DEVICE_INFO_V0_AGP;
+			else
+			if (pci_is_pcie(device->pdev))
+				args->v0.platform = NV_DEVICE_INFO_V0_PCIE;
+			else
+				args->v0.platform = NV_DEVICE_INFO_V0_PCI;
+		} else {
+			args->v0.platform = NV_DEVICE_INFO_V0_SOC;
+		}
+		break;
+	}
+
+	switch (device->card_type) {
+	case NV_04: args->v0.family = NV_DEVICE_INFO_V0_TNT; break;
+	case NV_10:
+	case NV_11: args->v0.family = NV_DEVICE_INFO_V0_CELSIUS; break;
+	case NV_20: args->v0.family = NV_DEVICE_INFO_V0_KELVIN; break;
+	case NV_30: args->v0.family = NV_DEVICE_INFO_V0_RANKINE; break;
+	case NV_40: args->v0.family = NV_DEVICE_INFO_V0_CURIE; break;
+	case NV_50: args->v0.family = NV_DEVICE_INFO_V0_TESLA; break;
+	case NV_C0: args->v0.family = NV_DEVICE_INFO_V0_FERMI; break;
+	case NV_E0: args->v0.family = NV_DEVICE_INFO_V0_KEPLER; break;
+	case GM100: args->v0.family = NV_DEVICE_INFO_V0_MAXWELL; break;
+	default:
+		args->v0.family = 0;
+		break;
+	}
+
+	args->v0.chipset  = device->chipset;
+	args->v0.revision = device->chipset >= 0x10 ? nv_rd32(device, 0) : 0x00;
+	if (pfb)  args->v0.ram_size = args->v0.ram_user = pfb->ram->size;
+	else      args->v0.ram_size = args->v0.ram_user = 0;
+	if (imem) args->v0.ram_user = args->v0.ram_user - imem->reserved;
+	return 0;
+}
+
+static int
+nouveau_devobj_mthd(struct nouveau_object *object, u32 mthd,
+		    void *data, u32 size)
+{
+	switch (mthd) {
+	case NV_DEVICE_V0_INFO:
+		return nouveau_devobj_info(object, data, size);
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
+static u8
+nouveau_devobj_rd08(struct nouveau_object *object, u64 addr)
+{
+	return nv_rd08(object->engine, addr);
+}
+
+static u16
+nouveau_devobj_rd16(struct nouveau_object *object, u64 addr)
+{
+	return nv_rd16(object->engine, addr);
+}
+
+static u32
+nouveau_devobj_rd32(struct nouveau_object *object, u64 addr)
+{
+	return nv_rd32(object->engine, addr);
+}
+
+static void
+nouveau_devobj_wr08(struct nouveau_object *object, u64 addr, u8 data)
+{
+	nv_wr08(object->engine, addr, data);
+}
+
+static void
+nouveau_devobj_wr16(struct nouveau_object *object, u64 addr, u16 data)
+{
+	nv_wr16(object->engine, addr, data);
+}
+
+static void
+nouveau_devobj_wr32(struct nouveau_object *object, u64 addr, u32 data)
+{
+	nv_wr32(object->engine, addr, data);
+}
+
+static int
+nouveau_devobj_map(struct nouveau_object *object, u64 *addr, u32 *size)
+{
+	struct nouveau_device *device = nv_device(object);
+	*addr = nv_device_resource_start(device, 0);
+	*size = nv_device_resource_len(device, 0);
+	return 0;
+}
+
 static const u64 disable_map[] = {
-	[NVDEV_SUBDEV_VBIOS]	= NV_DEVICE_DISABLE_VBIOS,
-	[NVDEV_SUBDEV_DEVINIT]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_GPIO]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_I2C]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_CLOCK]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_MXM]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_MC]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_BUS]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_TIMER]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_FB]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_LTCG]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_IBUS]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_INSTMEM]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_VM]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_BAR]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_VOLT]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_THERM]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_SUBDEV_PWR]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_ENGINE_DMAOBJ]	= NV_DEVICE_DISABLE_CORE,
-	[NVDEV_ENGINE_PERFMON]  = NV_DEVICE_DISABLE_CORE,
-	[NVDEV_ENGINE_FIFO]	= NV_DEVICE_DISABLE_FIFO,
-	[NVDEV_ENGINE_SW]	= NV_DEVICE_DISABLE_FIFO,
-	[NVDEV_ENGINE_GR]	= NV_DEVICE_DISABLE_GRAPH,
-	[NVDEV_ENGINE_MPEG]	= NV_DEVICE_DISABLE_MPEG,
-	[NVDEV_ENGINE_ME]	= NV_DEVICE_DISABLE_ME,
-	[NVDEV_ENGINE_VP]	= NV_DEVICE_DISABLE_VP,
-	[NVDEV_ENGINE_CRYPT]	= NV_DEVICE_DISABLE_CRYPT,
-	[NVDEV_ENGINE_BSP]	= NV_DEVICE_DISABLE_BSP,
-	[NVDEV_ENGINE_PPP]	= NV_DEVICE_DISABLE_PPP,
-	[NVDEV_ENGINE_COPY0]	= NV_DEVICE_DISABLE_COPY0,
-	[NVDEV_ENGINE_COPY1]	= NV_DEVICE_DISABLE_COPY1,
-	[NVDEV_ENGINE_VIC]	= NV_DEVICE_DISABLE_VIC,
-	[NVDEV_ENGINE_VENC]	= NV_DEVICE_DISABLE_VENC,
-	[NVDEV_ENGINE_DISP]	= NV_DEVICE_DISABLE_DISP,
+	[NVDEV_SUBDEV_VBIOS]	= NV_DEVICE_V0_DISABLE_VBIOS,
+	[NVDEV_SUBDEV_DEVINIT]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_GPIO]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_I2C]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_CLOCK]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_MXM]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_MC]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_BUS]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_TIMER]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_FB]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_LTC]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_IBUS]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_INSTMEM]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_VM]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_BAR]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_VOLT]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_THERM]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_SUBDEV_PWR]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_ENGINE_DMAOBJ]	= NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_ENGINE_PERFMON]  = NV_DEVICE_V0_DISABLE_CORE,
+	[NVDEV_ENGINE_FIFO]	= NV_DEVICE_V0_DISABLE_FIFO,
+	[NVDEV_ENGINE_SW]	= NV_DEVICE_V0_DISABLE_FIFO,
+	[NVDEV_ENGINE_GR]	= NV_DEVICE_V0_DISABLE_GRAPH,
+	[NVDEV_ENGINE_MPEG]	= NV_DEVICE_V0_DISABLE_MPEG,
+	[NVDEV_ENGINE_ME]	= NV_DEVICE_V0_DISABLE_ME,
+	[NVDEV_ENGINE_VP]	= NV_DEVICE_V0_DISABLE_VP,
+	[NVDEV_ENGINE_CRYPT]	= NV_DEVICE_V0_DISABLE_CRYPT,
+	[NVDEV_ENGINE_BSP]	= NV_DEVICE_V0_DISABLE_BSP,
+	[NVDEV_ENGINE_PPP]	= NV_DEVICE_V0_DISABLE_PPP,
+	[NVDEV_ENGINE_COPY0]	= NV_DEVICE_V0_DISABLE_COPY0,
+	[NVDEV_ENGINE_COPY1]	= NV_DEVICE_V0_DISABLE_COPY1,
+	[NVDEV_ENGINE_VIC]	= NV_DEVICE_V0_DISABLE_VIC,
+	[NVDEV_ENGINE_VENC]	= NV_DEVICE_V0_DISABLE_VENC,
+	[NVDEV_ENGINE_DISP]	= NV_DEVICE_V0_DISABLE_DISP,
 	[NVDEV_SUBDEV_NR]	= 0,
 };
 
+static void
+nouveau_devobj_dtor(struct nouveau_object *object)
+{
+	struct nouveau_devobj *devobj = (void *)object;
+	int i;
+
+	for (i = NVDEV_SUBDEV_NR - 1; i >= 0; i--)
+		nouveau_object_ref(NULL, &devobj->subdev[i]);
+
+	nouveau_parent_destroy(&devobj->base);
+}
+
+static struct nouveau_oclass
+nouveau_devobj_oclass_super = {
+	.handle = NV_DEVICE,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.dtor = nouveau_devobj_dtor,
+		.init = _nouveau_parent_init,
+		.fini = _nouveau_parent_fini,
+		.mthd = nouveau_devobj_mthd,
+		.map  = nouveau_devobj_map,
+		.rd08 = nouveau_devobj_rd08,
+		.rd16 = nouveau_devobj_rd16,
+		.rd32 = nouveau_devobj_rd32,
+		.wr08 = nouveau_devobj_wr08,
+		.wr16 = nouveau_devobj_wr16,
+		.wr32 = nouveau_devobj_wr32,
+	}
+};
+
 static int
 nouveau_devobj_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv_device_v0 v0;
+	} *args = data;
 	struct nouveau_client *client = nv_client(parent);
 	struct nouveau_device *device;
 	struct nouveau_devobj *devobj;
-	struct nv_device_class *args = data;
 	u32 boot0, strap;
 	u64 disable, mmio_base, mmio_size;
 	void __iomem *map;
 	int ret, i, c;
 
-	if (size < sizeof(struct nv_device_class))
-		return -EINVAL;
+	nv_ioctl(parent, "create device size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create device v%d device %016llx "
+				 "disable %016llx debug0 %016llx\n",
+			 args->v0.version, args->v0.device,
+			 args->v0.disable, args->v0.debug0);
+	} else
+		return ret;
+
+	/* give priviledged clients register access */
+	if (client->super)
+		oclass = &nouveau_devobj_oclass_super;
 
 	/* find the device subdev that matches what the client requested */
 	device = nv_device(client->device);
-	if (args->device != ~0) {
-		device = nouveau_device_find(args->device);
+	if (args->v0.device != ~0) {
+		device = nouveau_device_find(args->v0.device);
 		if (!device)
 			return -ENODEV;
 	}
@@ -135,14 +323,14 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 	mmio_size = nv_device_resource_len(device, 0);
 
 	/* translate api disable mask into internal mapping */
-	disable = args->debug0;
+	disable = args->v0.debug0;
 	for (i = 0; i < NVDEV_SUBDEV_NR; i++) {
-		if (args->disable & disable_map[i])
+		if (args->v0.disable & disable_map[i])
 			disable |= (1ULL << i);
 	}
 
 	/* identify the chipset, and determine classes of subdev/engines */
-	if (!(args->disable & NV_DEVICE_DISABLE_IDENTIFY) &&
+	if (!(args->v0.disable & NV_DEVICE_V0_DISABLE_IDENTIFY) &&
 	    !device->card_type) {
 		map = ioremap(mmio_base, 0x102000);
 		if (map == NULL)
@@ -180,8 +368,8 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 			case 0x080:
 			case 0x090:
 			case 0x0a0: device->card_type = NV_50; break;
-			case 0x0c0: device->card_type = NV_C0; break;
-			case 0x0d0: device->card_type = NV_D0; break;
+			case 0x0c0:
+			case 0x0d0: device->card_type = NV_C0; break;
 			case 0x0e0:
 			case 0x0f0:
 			case 0x100: device->card_type = NV_E0; break;
@@ -206,8 +394,7 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 		case NV_30: ret = nv30_identify(device); break;
 		case NV_40: ret = nv40_identify(device); break;
 		case NV_50: ret = nv50_identify(device); break;
-		case NV_C0:
-		case NV_D0: ret = nvc0_identify(device); break;
+		case NV_C0: ret = nvc0_identify(device); break;
 		case NV_E0: ret = nve0_identify(device); break;
 		case GM100: ret = gm100_identify(device); break;
 		default:
@@ -242,7 +429,7 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 		nv_debug(device, "crystal freq: %dKHz\n", device->crystal);
 	}
 
-	if (!(args->disable & NV_DEVICE_DISABLE_MMIO) &&
+	if (!(args->v0.disable & NV_DEVICE_V0_DISABLE_MMIO) &&
 	    !nv_subdev(device)->mmio) {
 		nv_subdev(device)->mmio  = ioremap(mmio_base, mmio_size);
 		if (!nv_subdev(device)->mmio) {
@@ -298,71 +485,19 @@ nouveau_devobj_ctor(struct nouveau_object *parent,
 	return 0;
 }
 
-static void
-nouveau_devobj_dtor(struct nouveau_object *object)
-{
-	struct nouveau_devobj *devobj = (void *)object;
-	int i;
-
-	for (i = NVDEV_SUBDEV_NR - 1; i >= 0; i--)
-		nouveau_object_ref(NULL, &devobj->subdev[i]);
-
-	nouveau_parent_destroy(&devobj->base);
-}
-
-static u8
-nouveau_devobj_rd08(struct nouveau_object *object, u64 addr)
-{
-	return nv_rd08(object->engine, addr);
-}
-
-static u16
-nouveau_devobj_rd16(struct nouveau_object *object, u64 addr)
-{
-	return nv_rd16(object->engine, addr);
-}
-
-static u32
-nouveau_devobj_rd32(struct nouveau_object *object, u64 addr)
-{
-	return nv_rd32(object->engine, addr);
-}
-
-static void
-nouveau_devobj_wr08(struct nouveau_object *object, u64 addr, u8 data)
-{
-	nv_wr08(object->engine, addr, data);
-}
-
-static void
-nouveau_devobj_wr16(struct nouveau_object *object, u64 addr, u16 data)
-{
-	nv_wr16(object->engine, addr, data);
-}
-
-static void
-nouveau_devobj_wr32(struct nouveau_object *object, u64 addr, u32 data)
-{
-	nv_wr32(object->engine, addr, data);
-}
-
 static struct nouveau_ofuncs
 nouveau_devobj_ofuncs = {
 	.ctor = nouveau_devobj_ctor,
 	.dtor = nouveau_devobj_dtor,
 	.init = _nouveau_parent_init,
 	.fini = _nouveau_parent_fini,
-	.rd08 = nouveau_devobj_rd08,
-	.rd16 = nouveau_devobj_rd16,
-	.rd32 = nouveau_devobj_rd32,
-	.wr08 = nouveau_devobj_wr08,
-	.wr16 = nouveau_devobj_wr16,
-	.wr32 = nouveau_devobj_wr32,
+	.mthd = nouveau_devobj_mthd,
 };
 
 /******************************************************************************
  * nouveau_device: engine functions
  *****************************************************************************/
+
 static struct nouveau_oclass
 nouveau_device_sclass[] = {
 	{ 0x0080, &nouveau_devobj_ofuncs },
@@ -370,6 +505,24 @@ nouveau_device_sclass[] = {
 };
 
 static int
+nouveau_device_event_ctor(struct nouveau_object *object, void *data, u32 size,
+			  struct nvkm_notify *notify)
+{
+	if (!WARN_ON(size != 0)) {
+		notify->size  = 0;
+		notify->types = 1;
+		notify->index = 0;
+		return 0;
+	}
+	return -EINVAL;
+}
+
+static const struct nvkm_event_func
+nouveau_device_event_func = {
+	.ctor = nouveau_device_event_ctor,
+};
+
+static int
 nouveau_device_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nouveau_device *device = (void *)object;
@@ -386,7 +539,7 @@ nouveau_device_fini(struct nouveau_object *object, bool suspend)
 		}
 	}
 
-	ret = 0;
+	ret = nvkm_acpi_fini(device, suspend);
 fail:
 	for (; ret && i < NVDEV_SUBDEV_NR; i++) {
 		if ((subdev = device->subdev[i])) {
@@ -407,7 +560,11 @@ nouveau_device_init(struct nouveau_object *object)
 {
 	struct nouveau_device *device = (void *)object;
 	struct nouveau_object *subdev;
-	int ret, i;
+	int ret, i = 0;
+
+	ret = nvkm_acpi_init(device);
+	if (ret)
+		goto fail;
 
 	for (i = 0; i < NVDEV_SUBDEV_NR; i++) {
 		if ((subdev = device->subdev[i])) {
@@ -430,6 +587,8 @@ fail:
 		}
 	}
 
+	if (ret)
+		nvkm_acpi_fini(device, false);
 	return ret;
 }
 
@@ -438,6 +597,8 @@ nouveau_device_dtor(struct nouveau_object *object)
 {
 	struct nouveau_device *device = (void *)object;
 
+	nvkm_event_fini(&device->event);
+
 	mutex_lock(&nv_devices_mutex);
 	list_del(&device->head);
 	mutex_unlock(&nv_devices_mutex);
@@ -478,31 +639,6 @@ nv_device_resource_len(struct nouveau_device *device, unsigned int bar)
 	}
 }
 
-dma_addr_t
-nv_device_map_page(struct nouveau_device *device, struct page *page)
-{
-	dma_addr_t ret;
-
-	if (nv_device_is_pci(device)) {
-		ret = pci_map_page(device->pdev, page, 0, PAGE_SIZE,
-				   PCI_DMA_BIDIRECTIONAL);
-		if (pci_dma_mapping_error(device->pdev, ret))
-			ret = 0;
-	} else {
-		ret = page_to_phys(page);
-	}
-
-	return ret;
-}
-
-void
-nv_device_unmap_page(struct nouveau_device *device, dma_addr_t addr)
-{
-	if (nv_device_is_pci(device))
-		pci_unmap_page(device->pdev, addr, PAGE_SIZE,
-			       PCI_DMA_BIDIRECTIONAL);
-}
-
 int
 nv_device_get_irq(struct nouveau_device *device, bool stall)
 {
@@ -560,6 +696,9 @@ nouveau_device_create_(void *dev, enum nv_bus_type type, u64 name,
 	nv_subdev(device)->debug = nouveau_dbgopt(device->dbgopt, "DEVICE");
 	nv_engine(device)->sclass = nouveau_device_sclass;
 	list_add(&device->head, &nv_devices);
+
+	ret = nvkm_event_init(&nouveau_device_event_func, 1, 1,
+			      &device->event);
 done:
 	mutex_unlock(&nv_devices_mutex);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/ctrl.c b/drivers/gpu/drm/nouveau/core/engine/device/ctrl.c
index 4b69bf56ed01..e34101a3490e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/ctrl.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/ctrl.c
@@ -22,55 +22,82 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
+#include <core/client.h>
 #include <core/object.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+#include <nvif/ioctl.h>
 
 #include <subdev/clock.h>
 
 #include "priv.h"
 
 static int
-nouveau_control_mthd_pstate_info(struct nouveau_object *object, u32 mthd,
-				void *data, u32 size)
+nouveau_control_mthd_pstate_info(struct nouveau_object *object,
+				 void *data, u32 size)
 {
+	union {
+		struct nvif_control_pstate_info_v0 v0;
+	} *args = data;
 	struct nouveau_clock *clk = nouveau_clock(object);
-	struct nv_control_pstate_info *args = data;
+	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(object, "control pstate info size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "control pstate info vers %d\n",
+			 args->v0.version);
+	} else
+		return ret;
 
 	if (clk) {
-		args->count  = clk->state_nr;
-		args->ustate = clk->ustate;
-		args->pstate = clk->pstate;
+		args->v0.count = clk->state_nr;
+		args->v0.ustate_ac = clk->ustate_ac;
+		args->v0.ustate_dc = clk->ustate_dc;
+		args->v0.pwrsrc = clk->pwrsrc;
+		args->v0.pstate = clk->pstate;
 	} else {
-		args->count  = 0;
-		args->ustate = NV_CONTROL_PSTATE_INFO_USTATE_DISABLE;
-		args->pstate = NV_CONTROL_PSTATE_INFO_PSTATE_UNKNOWN;
+		args->v0.count = 0;
+		args->v0.ustate_ac = NVIF_CONTROL_PSTATE_INFO_V0_USTATE_DISABLE;
+		args->v0.ustate_dc = NVIF_CONTROL_PSTATE_INFO_V0_USTATE_DISABLE;
+		args->v0.pwrsrc = -ENOSYS;
+		args->v0.pstate = NVIF_CONTROL_PSTATE_INFO_V0_PSTATE_UNKNOWN;
 	}
 
 	return 0;
 }
 
 static int
-nouveau_control_mthd_pstate_attr(struct nouveau_object *object, u32 mthd,
-				void *data, u32 size)
+nouveau_control_mthd_pstate_attr(struct nouveau_object *object,
+				 void *data, u32 size)
 {
+	union {
+		struct nvif_control_pstate_attr_v0 v0;
+	} *args = data;
 	struct nouveau_clock *clk = nouveau_clock(object);
-	struct nv_control_pstate_attr *args = data;
 	struct nouveau_clocks *domain;
 	struct nouveau_pstate *pstate;
 	struct nouveau_cstate *cstate;
 	int i = 0, j = -1;
 	u32 lo, hi;
-
-	if ((size < sizeof(*args)) || !clk ||
-	    (args->state >= 0 && args->state >= clk->state_nr))
-		return -EINVAL;
+	int ret;
+
+	nv_ioctl(object, "control pstate attr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "control pstate attr vers %d state %d "
+				 "index %d\n",
+			 args->v0.version, args->v0.state, args->v0.index);
+		if (!clk)
+			return -ENODEV;
+		if (args->v0.state < NVIF_CONTROL_PSTATE_ATTR_V0_STATE_CURRENT)
+			return -EINVAL;
+		if (args->v0.state >= clk->state_nr)
+			return -EINVAL;
+	} else
+		return ret;
 	domain = clk->domains;
 
 	while (domain->name != nv_clk_src_max) {
-		if (domain->mname && ++j == args->index)
+		if (domain->mname && ++j == args->v0.index)
 			break;
 		domain++;
 	}
@@ -78,9 +105,9 @@ nouveau_control_mthd_pstate_attr(struct nouveau_object *object, u32 mthd,
 	if (domain->name == nv_clk_src_max)
 		return -EINVAL;
 
-	if (args->state != NV_CONTROL_PSTATE_ATTR_STATE_CURRENT) {
+	if (args->v0.state != NVIF_CONTROL_PSTATE_ATTR_V0_STATE_CURRENT) {
 		list_for_each_entry(pstate, &clk->states, head) {
-			if (i++ == args->state)
+			if (i++ == args->v0.state)
 				break;
 		}
 
@@ -91,21 +118,21 @@ nouveau_control_mthd_pstate_attr(struct nouveau_object *object, u32 mthd,
 			hi = max(hi, cstate->domain[domain->name]);
 		}
 
-		args->state = pstate->pstate;
+		args->v0.state = pstate->pstate;
 	} else {
 		lo = max(clk->read(clk, domain->name), 0);
 		hi = lo;
 	}
 
-	snprintf(args->name, sizeof(args->name), "%s", domain->mname);
-	snprintf(args->unit, sizeof(args->unit), "MHz");
-	args->min = lo / domain->mdiv;
-	args->max = hi / domain->mdiv;
+	snprintf(args->v0.name, sizeof(args->v0.name), "%s", domain->mname);
+	snprintf(args->v0.unit, sizeof(args->v0.unit), "MHz");
+	args->v0.min = lo / domain->mdiv;
+	args->v0.max = hi / domain->mdiv;
 
-	args->index = 0;
+	args->v0.index = 0;
 	while ((++domain)->name != nv_clk_src_max) {
 		if (domain->mname) {
-			args->index = ++j;
+			args->v0.index = ++j;
 			break;
 		}
 	}
@@ -114,31 +141,65 @@ nouveau_control_mthd_pstate_attr(struct nouveau_object *object, u32 mthd,
 }
 
 static int
-nouveau_control_mthd_pstate_user(struct nouveau_object *object, u32 mthd,
-				void *data, u32 size)
+nouveau_control_mthd_pstate_user(struct nouveau_object *object,
+				 void *data, u32 size)
 {
+	union {
+		struct nvif_control_pstate_user_v0 v0;
+	} *args = data;
 	struct nouveau_clock *clk = nouveau_clock(object);
-	struct nv_control_pstate_user *args = data;
+	int ret;
+
+	nv_ioctl(object, "control pstate user size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "control pstate user vers %d ustate %d "
+				 "pwrsrc %d\n", args->v0.version,
+			 args->v0.ustate, args->v0.pwrsrc);
+		if (!clk)
+			return -ENODEV;
+	} else
+		return ret;
+
+	if (args->v0.pwrsrc >= 0) {
+		ret |= nouveau_clock_ustate(clk, args->v0.ustate, args->v0.pwrsrc);
+	} else {
+		ret |= nouveau_clock_ustate(clk, args->v0.ustate, 0);
+		ret |= nouveau_clock_ustate(clk, args->v0.ustate, 1);
+	}
 
-	if (size < sizeof(*args) || !clk)
-		return -EINVAL;
+	return ret;
+}
 
-	return nouveau_clock_ustate(clk, args->state);
+static int
+nouveau_control_mthd(struct nouveau_object *object, u32 mthd,
+		     void *data, u32 size)
+{
+	switch (mthd) {
+	case NVIF_CONTROL_PSTATE_INFO:
+		return nouveau_control_mthd_pstate_info(object, data, size);
+	case NVIF_CONTROL_PSTATE_ATTR:
+		return nouveau_control_mthd_pstate_attr(object, data, size);
+	case NVIF_CONTROL_PSTATE_USER:
+		return nouveau_control_mthd_pstate_user(object, data, size);
+	default:
+		break;
+	}
+	return -EINVAL;
 }
 
+static struct nouveau_ofuncs
+nouveau_control_ofuncs = {
+	.ctor = _nouveau_object_ctor,
+	.dtor = nouveau_object_destroy,
+	.init = nouveau_object_init,
+	.fini = nouveau_object_fini,
+	.mthd = nouveau_control_mthd,
+};
+
 struct nouveau_oclass
 nouveau_control_oclass[] = {
-	{ .handle = NV_CONTROL_CLASS,
-	  .ofuncs = &nouveau_object_ofuncs,
-	  .omthds = (struct nouveau_omthds[]) {
-		  { NV_CONTROL_PSTATE_INFO,
-		    NV_CONTROL_PSTATE_INFO, nouveau_control_mthd_pstate_info },
-		  { NV_CONTROL_PSTATE_ATTR,
-		    NV_CONTROL_PSTATE_ATTR, nouveau_control_mthd_pstate_attr },
-		  { NV_CONTROL_PSTATE_USER,
-		    NV_CONTROL_PSTATE_USER, nouveau_control_mthd_pstate_user },
-		  {},
-	  },
+	{ .handle = NVIF_IOCTL_NEW_V0_CONTROL,
+	  .ofuncs = &nouveau_control_ofuncs
 	},
 	{}
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/gm100.c b/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
index a520029e25d9..6295668e29a5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/gm100.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -33,7 +34,7 @@
 #include <subdev/mc.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/ltcg.h>
+#include <subdev/ltc.h>
 #include <subdev/ibus.h>
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
@@ -62,26 +63,26 @@ gm100_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gm107_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
-#if 0
-		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
-#endif
+		device->oclass[NVDEV_SUBDEV_THERM  ] = &gm107_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] =  gm107_devinit_oclass;
-		device->oclass[NVDEV_SUBDEV_MC     ] =  nvc3_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &gk20a_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  gm107_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gm107_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gm107_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nv108_pwr_oclass;
+
 #if 0
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nv108_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
 #endif
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv108_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  gm107_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv04.c b/drivers/gpu/drm/nouveau/core/engine/device/nv04.c
index 40b29d0214cb..573b55f5c2f9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv04.c
@@ -56,7 +56,7 @@ nv04_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv04_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv04_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv04_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv04_graph_oclass;
@@ -74,7 +74,7 @@ nv04_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv04_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv04_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv04_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv04_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv10.c b/drivers/gpu/drm/nouveau/core/engine/device/nv10.c
index 5f7c25ff523d..183a85a6204e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv10.c
@@ -58,7 +58,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] =  nv04_disp_oclass;
 		break;
@@ -75,7 +75,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv10_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -94,7 +94,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv10_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -113,7 +113,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv1a_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv10_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -132,7 +132,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv10_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -151,7 +151,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -170,7 +170,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv1a_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
@@ -189,7 +189,7 @@ nv10_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv10_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv20.c b/drivers/gpu/drm/nouveau/core/engine/device/nv20.c
index 75fed11bba0a..aa564c68a920 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv20.c
@@ -59,7 +59,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv20_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv20_graph_oclass;
@@ -78,7 +78,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv25_graph_oclass;
@@ -97,7 +97,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv25_graph_oclass;
@@ -116,7 +116,7 @@ nv20_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv25_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv2a_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv30.c b/drivers/gpu/drm/nouveau/core/engine/device/nv30.c
index 36919d7db7cc..11bd31da82ab 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv30.c
@@ -59,7 +59,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv30_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv30_graph_oclass;
@@ -78,7 +78,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv35_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv35_graph_oclass;
@@ -97,7 +97,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv30_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv30_graph_oclass;
@@ -117,7 +117,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv36_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv35_graph_oclass;
@@ -137,7 +137,7 @@ nv30_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nv10_fb_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv04_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv17_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv34_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv40.c b/drivers/gpu/drm/nouveau/core/engine/device/nv40.c
index 1130a62be2c7..e96c223cb797 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv40.c
@@ -65,7 +65,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -88,7 +88,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -111,7 +111,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -134,7 +134,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -157,7 +157,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv04_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -180,7 +180,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -203,7 +203,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -226,7 +226,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv41_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -249,7 +249,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -272,7 +272,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -295,7 +295,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -318,7 +318,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -341,7 +341,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -364,7 +364,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -387,7 +387,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
@@ -410,7 +410,7 @@ nv40_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv40_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv44_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv04_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv04_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv40_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv10_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv40_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
index ef0b0bde1a91..96f568d1321b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -62,6 +63,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv50_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv50_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -74,7 +76,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv50_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -87,6 +89,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -99,7 +102,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -115,6 +118,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -127,7 +131,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -141,8 +145,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x92:
 		device->cname = "G92";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv50_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -155,7 +160,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -169,8 +174,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x94:
 		device->cname = "G94";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -183,7 +189,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -197,8 +203,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x96:
 		device->cname = "G96";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -211,7 +218,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -225,8 +232,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0x98:
 		device->cname = "G98";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -239,7 +247,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -253,8 +261,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa0:
 		device->cname = "G200";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv50_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nv84_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -267,7 +276,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -281,8 +290,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xaa:
 		device->cname = "MCP77/MCP78";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nvaa_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -295,7 +305,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -309,8 +319,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xac:
 		device->cname = "MCP79/MCP7A";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] =  nvaa_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nv84_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -323,7 +334,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -337,8 +348,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa3:
 		device->cname = "GT215";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -350,9 +362,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nva3_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nva3_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -367,8 +379,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa5:
 		device->cname = "GT216";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -380,9 +393,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nva3_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nva3_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -396,8 +409,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xa8:
 		device->cname = "GT218";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -409,9 +423,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nva3_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nva3_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
@@ -425,8 +439,9 @@ nv50_identify(struct nouveau_device *device)
 	case 0xaf:
 		device->cname = "MCP89";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] =  &g80_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -438,9 +453,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nv50_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nv50_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nva3_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nva3_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nv50_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nv50_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
index 8d55ed633b19..cd05677ad4b7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -33,7 +34,7 @@
 #include <subdev/mc.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/ltcg.h>
+#include <subdev/ltc.h>
 #include <subdev/ibus.h>
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
@@ -60,8 +61,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc0:
 		device->cname = "GF100";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -70,14 +72,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc0_graph_oclass;
@@ -92,8 +94,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc4:
 		device->cname = "GF104";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -102,14 +105,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc4_graph_oclass;
@@ -124,8 +127,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc3:
 		device->cname = "GF106";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -134,14 +138,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc4_graph_oclass;
@@ -155,8 +159,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xce:
 		device->cname = "GF114";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -165,14 +170,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc4_graph_oclass;
@@ -187,8 +192,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xcf:
 		device->cname = "GF116";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -197,14 +203,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc4_graph_oclass;
@@ -219,8 +225,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc1:
 		device->cname = "GF108";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -229,14 +236,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc1_graph_oclass;
@@ -250,8 +257,9 @@ nvc0_identify(struct nouveau_device *device)
 	case 0xc8:
 		device->cname = "GF110";
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
-		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv92_gpio_oclass;
+		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nv94_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nv94_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -260,14 +268,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvc0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvc0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvc8_graph_oclass;
@@ -284,6 +292,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -292,14 +301,14 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvd0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvd9_graph_oclass;
@@ -315,6 +324,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nvd0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  gf117_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -323,12 +333,12 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nvc0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gf100_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nvc0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvd7_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
index 2d1e97d4264f..b1b2e484ecfa 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
@@ -26,6 +26,7 @@
 #include <subdev/bus.h>
 #include <subdev/gpio.h>
 #include <subdev/i2c.h>
+#include <subdev/fuse.h>
 #include <subdev/clock.h>
 #include <subdev/therm.h>
 #include <subdev/mxm.h>
@@ -33,7 +34,7 @@
 #include <subdev/mc.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/ltcg.h>
+#include <subdev/ltc.h>
 #include <subdev/ibus.h>
 #include <subdev/instmem.h>
 #include <subdev/vm.h>
@@ -62,6 +63,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -70,14 +72,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  gk104_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
@@ -95,6 +97,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -103,14 +106,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvd0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
@@ -128,6 +131,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -136,14 +140,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  gk104_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
@@ -158,15 +162,18 @@ nve0_identify(struct nouveau_device *device)
 		break;
 	case 0xea:
 		device->cname = "GK20A";
-		device->oclass[NVDEV_SUBDEV_MC     ] =  nvc3_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &gk20a_clock_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &gk20a_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  gk20a_fb_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &gk20a_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] = nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
-		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_SUBDEV_BAR    ] = &gk20a_bar_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  gk20a_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  gk20a_graph_oclass;
@@ -178,6 +185,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -186,14 +194,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvd0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nvf0_graph_oclass;
@@ -211,6 +219,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nvd0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
@@ -219,17 +228,17 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nvd0_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nvd0_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] =  nvf0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  gk110b_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] =  nvf0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
@@ -244,22 +253,23 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VBIOS  ] = &nouveau_bios_oclass;
 		device->oclass[NVDEV_SUBDEV_GPIO   ] =  nve0_gpio_oclass;
 		device->oclass[NVDEV_SUBDEV_I2C    ] =  nve0_i2c_oclass;
+		device->oclass[NVDEV_SUBDEV_FUSE   ] = &gf100_fuse_oclass;
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nve0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
 		device->oclass[NVDEV_SUBDEV_DEVINIT] =  nvc0_devinit_oclass;
-		device->oclass[NVDEV_SUBDEV_MC     ] =  nvc3_mc_oclass;
+		device->oclass[NVDEV_SUBDEV_MC     ] =  gk20a_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] =  nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
 		device->oclass[NVDEV_SUBDEV_FB     ] =  nve0_fb_oclass;
-		device->oclass[NVDEV_SUBDEV_LTCG   ] =  gf100_ltcg_oclass;
+		device->oclass[NVDEV_SUBDEV_LTC    ] =  gk104_ltc_oclass;
 		device->oclass[NVDEV_SUBDEV_IBUS   ] = &nve0_ibus_oclass;
 		device->oclass[NVDEV_SUBDEV_INSTMEM] =  nv50_instmem_oclass;
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
-		device->oclass[NVDEV_SUBDEV_PWR    ] = &nv108_pwr_oclass;
+		device->oclass[NVDEV_SUBDEV_PWR    ] =  nv108_pwr_oclass;
 		device->oclass[NVDEV_SUBDEV_VOLT   ] = &nv40_volt_oclass;
-		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
+		device->oclass[NVDEV_ENGINE_DMAOBJ ] =  nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] =  nv108_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] =  nvc0_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] =  nv108_graph_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/base.c b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
index 9c38c5e40500..64b84667f3a5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/base.c
@@ -22,23 +22,95 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/os.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+#include <nvif/event.h>
+
 #include "priv.h"
 #include "outp.h"
 #include "conn.h"
 
+int
+nouveau_disp_vblank_ctor(struct nouveau_object *object, void *data, u32 size,
+			 struct nvkm_notify *notify)
+{
+	struct nouveau_disp *disp =
+		container_of(notify->event, typeof(*disp), vblank);
+	union {
+		struct nvif_notify_head_req_v0 v0;
+	} *req = data;
+	int ret;
+
+	if (nvif_unpack(req->v0, 0, 0, false)) {
+		notify->size = sizeof(struct nvif_notify_head_rep_v0);
+		if (ret = -ENXIO, req->v0.head <= disp->vblank.index_nr) {
+			notify->types = 1;
+			notify->index = req->v0.head;
+			return 0;
+		}
+	}
+
+	return ret;
+}
+
+void
+nouveau_disp_vblank(struct nouveau_disp *disp, int head)
+{
+	struct nvif_notify_head_rep_v0 rep = {};
+	nvkm_event_send(&disp->vblank, 1, head, &rep, sizeof(rep));
+}
+
 static int
-nouveau_disp_hpd_check(struct nouveau_event *event, u32 types, int index)
+nouveau_disp_hpd_ctor(struct nouveau_object *object, void *data, u32 size,
+		      struct nvkm_notify *notify)
 {
-	struct nouveau_disp *disp = event->priv;
+	struct nouveau_disp *disp =
+		container_of(notify->event, typeof(*disp), hpd);
+	union {
+		struct nvif_notify_conn_req_v0 v0;
+	} *req = data;
 	struct nvkm_output *outp;
-	list_for_each_entry(outp, &disp->outp, head) {
-		if (outp->conn->index == index) {
-			if (outp->conn->hpd.event)
-				return 0;
-			break;
+	int ret;
+
+	if (nvif_unpack(req->v0, 0, 0, false)) {
+		notify->size = sizeof(struct nvif_notify_conn_rep_v0);
+		list_for_each_entry(outp, &disp->outp, head) {
+			if (ret = -ENXIO, outp->conn->index == req->v0.conn) {
+				if (ret = -ENODEV, outp->conn->hpd.event) {
+					notify->types = req->v0.mask;
+					notify->index = req->v0.conn;
+					ret = 0;
+				}
+				break;
+			}
 		}
 	}
-	return -ENOSYS;
+
+	return ret;
+}
+
+static const struct nvkm_event_func
+nouveau_disp_hpd_func = {
+	.ctor = nouveau_disp_hpd_ctor
+};
+
+int
+nouveau_disp_ntfy(struct nouveau_object *object, u32 type,
+		  struct nvkm_event **event)
+{
+	struct nouveau_disp *disp = (void *)object->engine;
+	switch (type) {
+	case NV04_DISP_NTFY_VBLANK:
+		*event = &disp->vblank;
+		return 0;
+	case NV04_DISP_NTFY_CONN:
+		*event = &disp->hpd;
+		return 0;
+	default:
+		break;
+	}
+	return -EINVAL;
 }
 
 int
@@ -97,7 +169,8 @@ _nouveau_disp_dtor(struct nouveau_object *object)
 	struct nouveau_disp *disp = (void *)object;
 	struct nvkm_output *outp, *outt;
 
-	nouveau_event_destroy(&disp->vblank);
+	nvkm_event_fini(&disp->vblank);
+	nvkm_event_fini(&disp->hpd);
 
 	if (disp->outp.next) {
 		list_for_each_entry_safe(outp, outt, &disp->outp, head) {
@@ -157,14 +230,11 @@ nouveau_disp_create_(struct nouveau_object *parent,
 		hpd = max(hpd, (u8)(dcbE.connector + 1));
 	}
 
-	ret = nouveau_event_create(3, hpd, &disp->hpd);
+	ret = nvkm_event_init(&nouveau_disp_hpd_func, 3, hpd, &disp->hpd);
 	if (ret)
 		return ret;
 
-	disp->hpd->priv = disp;
-	disp->hpd->check = nouveau_disp_hpd_check;
-
-	ret = nouveau_event_create(1, heads, &disp->vblank);
+	ret = nvkm_event_init(impl->vblank, 1, heads, &disp->vblank);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/conn.c b/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
index 4ffbc70ecf5a..1496b567dd4a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/conn.c
@@ -22,39 +22,41 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/os.h>
+#include <nvif/event.h>
+
 #include <subdev/gpio.h>
 
 #include "conn.h"
 #include "outp.h"
 
-static void
-nvkm_connector_hpd_work(struct work_struct *w)
+static int
+nvkm_connector_hpd(struct nvkm_notify *notify)
 {
-	struct nvkm_connector *conn = container_of(w, typeof(*conn), hpd.work);
+	struct nvkm_connector *conn = container_of(notify, typeof(*conn), hpd);
 	struct nouveau_disp *disp = nouveau_disp(conn);
 	struct nouveau_gpio *gpio = nouveau_gpio(conn);
-	u32 send = NVKM_HPD_UNPLUG;
-	if (gpio->get(gpio, 0, DCB_GPIO_UNUSED, conn->hpd.event->index))
-		send = NVKM_HPD_PLUG;
-	nouveau_event_trigger(disp->hpd, send, conn->index);
-	nouveau_event_get(conn->hpd.event);
-}
+	const struct nvkm_gpio_ntfy_rep *line = notify->data;
+	struct nvif_notify_conn_rep_v0 rep;
+	int index = conn->index;
 
-static int
-nvkm_connector_hpd(void *data, u32 type, int index)
-{
-	struct nvkm_connector *conn = data;
-	DBG("HPD: %d\n", type);
-	schedule_work(&conn->hpd.work);
-	return NVKM_EVENT_DROP;
+	DBG("HPD: %d\n", line->mask);
+
+	if (!gpio->get(gpio, 0, DCB_GPIO_UNUSED, conn->hpd.index))
+		rep.mask = NVIF_NOTIFY_CONN_V0_UNPLUG;
+	else
+		rep.mask = NVIF_NOTIFY_CONN_V0_PLUG;
+	rep.version = 0;
+
+	nvkm_event_send(&disp->hpd, rep.mask, index, &rep, sizeof(rep));
+	return NVKM_NOTIFY_KEEP;
 }
 
 int
 _nvkm_connector_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nvkm_connector *conn = (void *)object;
-	if (conn->hpd.event)
-		nouveau_event_put(conn->hpd.event);
+	nvkm_notify_put(&conn->hpd);
 	return nouveau_object_fini(&conn->base, suspend);
 }
 
@@ -63,10 +65,8 @@ _nvkm_connector_init(struct nouveau_object *object)
 {
 	struct nvkm_connector *conn = (void *)object;
 	int ret = nouveau_object_init(&conn->base);
-	if (ret == 0) {
-		if (conn->hpd.event)
-			nouveau_event_get(conn->hpd.event);
-	}
+	if (ret == 0)
+		nvkm_notify_get(&conn->hpd);
 	return ret;
 }
 
@@ -74,7 +74,7 @@ void
 _nvkm_connector_dtor(struct nouveau_object *object)
 {
 	struct nvkm_connector *conn = (void *)object;
-	nouveau_event_ref(NULL, &conn->hpd.event);
+	nvkm_notify_fini(&conn->hpd);
 	nouveau_object_destroy(&conn->base);
 }
 
@@ -116,19 +116,24 @@ nvkm_connector_create_(struct nouveau_object *parent,
 	if ((info->hpd = ffs(info->hpd))) {
 		if (--info->hpd >= ARRAY_SIZE(hpd)) {
 			ERR("hpd %02x unknown\n", info->hpd);
-			goto done;
+			return 0;
 		}
 		info->hpd = hpd[info->hpd];
 
 		ret = gpio->find(gpio, 0, info->hpd, DCB_GPIO_UNUSED, &func);
 		if (ret) {
 			ERR("func %02x lookup failed, %d\n", info->hpd, ret);
-			goto done;
+			return 0;
 		}
 
-		ret = nouveau_event_new(gpio->events, NVKM_GPIO_TOGGLED,
-					func.line, nvkm_connector_hpd,
-					conn, &conn->hpd.event);
+		ret = nvkm_notify_init(NULL, &gpio->event, nvkm_connector_hpd,
+				       true, &(struct nvkm_gpio_ntfy_req) {
+					.mask = NVKM_GPIO_TOGGLED,
+					.line = func.line,
+				       },
+				       sizeof(struct nvkm_gpio_ntfy_req),
+				       sizeof(struct nvkm_gpio_ntfy_rep),
+				       &conn->hpd);
 		if (ret) {
 			ERR("func %02x failed, %d\n", info->hpd, ret);
 		} else {
@@ -136,8 +141,6 @@ nvkm_connector_create_(struct nouveau_object *parent,
 		}
 	}
 
-done:
-	INIT_WORK(&conn->hpd.work, nvkm_connector_hpd_work);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/conn.h b/drivers/gpu/drm/nouveau/core/engine/disp/conn.h
index 035ebeacbb1c..55e5f5c82c14 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/conn.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/conn.h
@@ -10,10 +10,7 @@ struct nvkm_connector {
 	struct nvbios_connE info;
 	int index;
 
-	struct {
-		struct nouveau_eventh *event;
-		struct work_struct work;
-	} hpd;
+	struct nvkm_notify hpd;
 };
 
 #define nvkm_connector_create(p,e,c,b,i,d)                                     \
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dacnv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/dacnv50.c
index a66b27c0fcab..b36addff06a9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/dacnv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dacnv50.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
@@ -32,13 +33,28 @@
 #include "nv50.h"
 
 int
-nv50_dac_power(struct nv50_disp_priv *priv, int or, u32 data)
+nv50_dac_power(NV50_DISP_MTHD_V1)
 {
-	const u32 stat = (data & NV50_DISP_DAC_PWR_HSYNC) |
-		         (data & NV50_DISP_DAC_PWR_VSYNC) |
-		         (data & NV50_DISP_DAC_PWR_DATA) |
-		         (data & NV50_DISP_DAC_PWR_STATE);
-	const u32 doff = (or * 0x800);
+	const u32 doff = outp->or * 0x800;
+	union {
+		struct nv50_disp_dac_pwr_v0 v0;
+	} *args = data;
+	u32 stat;
+	int ret;
+
+	nv_ioctl(object, "disp dac pwr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp dac pwr vers %d state %d data %d "
+				 "vsync %d hsync %d\n",
+			 args->v0.version, args->v0.state, args->v0.data,
+			 args->v0.vsync, args->v0.hsync);
+		stat  = 0x00000040 * !args->v0.state;
+		stat |= 0x00000010 * !args->v0.data;
+		stat |= 0x00000004 * !args->v0.vsync;
+		stat |= 0x00000001 * !args->v0.hsync;
+	} else
+		return ret;
+
 	nv_wait(priv, 0x61a004 + doff, 0x80000000, 0x00000000);
 	nv_mask(priv, 0x61a004 + doff, 0xc000007f, 0x80000000 | stat);
 	nv_wait(priv, 0x61a004 + doff, 0x80000000, 0x00000000);
@@ -46,9 +62,24 @@ nv50_dac_power(struct nv50_disp_priv *priv, int or, u32 data)
 }
 
 int
-nv50_dac_sense(struct nv50_disp_priv *priv, int or, u32 loadval)
+nv50_dac_sense(NV50_DISP_MTHD_V1)
 {
-	const u32 doff = (or * 0x800);
+	union {
+		struct nv50_disp_dac_load_v0 v0;
+	} *args = data;
+	const u32 doff = outp->or * 0x800;
+	u32 loadval;
+	int ret;
+
+	nv_ioctl(object, "disp dac load size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp dac load vers %d data %08x\n",
+			 args->v0.version, args->v0.data);
+		if (args->v0.data & 0xfff00000)
+			return -EINVAL;
+		loadval = args->v0.data;
+	} else
+		return ret;
 
 	nv_mask(priv, 0x61a004 + doff, 0x807f0000, 0x80150000);
 	nv_wait(priv, 0x61a004 + doff, 0x80000000, 0x00000000);
@@ -61,38 +92,10 @@ nv50_dac_sense(struct nv50_disp_priv *priv, int or, u32 loadval)
 	nv_mask(priv, 0x61a004 + doff, 0x807f0000, 0x80550000);
 	nv_wait(priv, 0x61a004 + doff, 0x80000000, 0x00000000);
 
-	nv_debug(priv, "DAC%d sense: 0x%08x\n", or, loadval);
+	nv_debug(priv, "DAC%d sense: 0x%08x\n", outp->or, loadval);
 	if (!(loadval & 0x80000000))
 		return -ETIMEDOUT;
 
-	return (loadval & 0x38000000) >> 27;
-}
-
-int
-nv50_dac_mthd(struct nouveau_object *object, u32 mthd, void *args, u32 size)
-{
-	struct nv50_disp_priv *priv = (void *)object->engine;
-	const u8 or = (mthd & NV50_DISP_DAC_MTHD_OR);
-	u32 *data = args;
-	int ret;
-
-	if (size < sizeof(u32))
-		return -EINVAL;
-
-	switch (mthd & ~0x3f) {
-	case NV50_DISP_DAC_PWR:
-		ret = priv->dac.power(priv, or, data[0]);
-		break;
-	case NV50_DISP_DAC_LOAD:
-		ret = priv->dac.sense(priv, or, data[0]);
-		if (ret >= 0) {
-			data[0] = ret;
-			ret = 0;
-		}
-		break;
-	default:
-		BUG_ON(1);
-	}
-
-	return ret;
+	args->v0.load = (loadval & 0x38000000) >> 27;
+	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
index 5a5b59b21130..39890221b91c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/dport.c
@@ -30,7 +30,7 @@
 
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "dport.h"
 #include "outpdp.h"
@@ -335,7 +335,7 @@ nouveau_dp_train(struct work_struct *w)
 	int ret;
 
 	/* bring capabilities within encoder limits */
-	if (nv_mclass(disp) < NVD0_DISP_CLASS)
+	if (nv_mclass(disp) < GF110_DISP)
 		outp->dpcd[2] &= ~DPCD_RC02_TPS3_SUPPORTED;
 	if ((outp->dpcd[2] & 0x1f) > outp->base.info.dpconf.link_nr) {
 		outp->dpcd[2] &= ~DPCD_RC02_MAX_LANE_COUNT;
@@ -354,7 +354,7 @@ nouveau_dp_train(struct work_struct *w)
 	cfg--;
 
 	/* disable link interrupt handling during link training */
-	nouveau_event_put(outp->irq);
+	nvkm_notify_put(&outp->irq);
 
 	/* enable down-spreading and execute pre-train script from vbios */
 	dp_link_train_init(dp, outp->dpcd[3] & 0x01);
@@ -395,5 +395,5 @@ nouveau_dp_train(struct work_struct *w)
 	DBG("training complete\n");
 	atomic_set(&outp->lt.done, 1);
 	wake_up(&outp->lt.wait);
-	nouveau_event_get(outp->irq);
+	nvkm_notify_get(&outp->irq);
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c b/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
index 9fc7447fec90..b3df3fe2dc09 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/gm107.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -35,17 +35,17 @@
 
 static struct nouveau_oclass
 gm107_disp_sclass[] = {
-	{ GM107_DISP_MAST_CLASS, &nvd0_disp_mast_ofuncs },
-	{ GM107_DISP_SYNC_CLASS, &nvd0_disp_sync_ofuncs },
-	{ GM107_DISP_OVLY_CLASS, &nvd0_disp_ovly_ofuncs },
-	{ GM107_DISP_OIMM_CLASS, &nvd0_disp_oimm_ofuncs },
-	{ GM107_DISP_CURS_CLASS, &nvd0_disp_curs_ofuncs },
+	{ GM107_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
+	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
+	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
+	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
 	{}
 };
 
 static struct nouveau_oclass
 gm107_disp_base_oclass[] = {
-	{ GM107_DISP_CLASS, &nvd0_disp_base_ofuncs, nvd0_disp_base_omthds },
+	{ GM107_DISP, &nvd0_disp_base_ofuncs },
 	{}
 };
 
@@ -68,6 +68,10 @@ gm107_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = gm107_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
@@ -80,7 +84,7 @@ gm107_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -93,9 +97,11 @@ gm107_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
 	.mthd.core = &nve0_disp_mast_mthd_chan,
 	.mthd.base = &nvd0_disp_sync_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
+	.head.scanoutpos = nvd0_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
index a19e7d79b847..fe9ef5894dd4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdanva3.c
@@ -22,28 +22,47 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+
+#include <subdev/timer.h>
 
 #include "nv50.h"
 
 int
-nva3_hda_eld(struct nv50_disp_priv *priv, int or, u8 *data, u32 size)
+nva3_hda_eld(NV50_DISP_MTHD_V1)
 {
-	const u32 soff = (or * 0x800);
-	int i;
+	union {
+		struct nv50_disp_sor_hda_eld_v0 v0;
+	} *args = data;
+	const u32 soff = outp->or * 0x800;
+	int ret, i;
+
+	nv_ioctl(object, "disp sor hda eld size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "disp sor hda eld vers %d\n", args->v0.version);
+		if (size > 0x60)
+			return -E2BIG;
+	} else
+		return ret;
 
-	if (data && data[0]) {
+	if (size && args->v0.data[0]) {
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x61c1e0 + soff, 0x8000000d, 0x80000001);
+			nv_wait(priv, 0x61c1e0 + soff, 0x80000000, 0x00000000);
+		}
 		for (i = 0; i < size; i++)
-			nv_wr32(priv, 0x61c440 + soff, (i << 8) | data[i]);
+			nv_wr32(priv, 0x61c440 + soff, (i << 8) | args->v0.data[0]);
 		for (; i < 0x60; i++)
 			nv_wr32(priv, 0x61c440 + soff, (i << 8));
 		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000003);
-	} else
-	if (data) {
-		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000001);
 	} else {
-		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000000);
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x61c1e0 + soff, 0x80000001, 0x80000000);
+			nv_wait(priv, 0x61c1e0 + soff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x61c448 + soff, 0x80000003, 0x80000000 | !!size);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
index 717639386ced..1d4e8432d857 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdanvd0.c
@@ -22,33 +22,49 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
-#include <subdev/bios.h>
-#include <subdev/bios/dcb.h>
-#include <subdev/bios/dp.h>
-#include <subdev/bios/init.h>
+#include <subdev/timer.h>
 
 #include "nv50.h"
 
 int
-nvd0_hda_eld(struct nv50_disp_priv *priv, int or, u8 *data, u32 size)
+nvd0_hda_eld(NV50_DISP_MTHD_V1)
 {
-	const u32 soff = (or * 0x030);
-	int i;
+	union {
+		struct nv50_disp_sor_hda_eld_v0 v0;
+	} *args = data;
+	const u32 soff = outp->or * 0x030;
+	const u32 hoff = head * 0x800;
+	int ret, i;
 
-	if (data && data[0]) {
+	nv_ioctl(object, "disp sor hda eld size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "disp sor hda eld vers %d\n", args->v0.version);
+		if (size > 0x60)
+			return -E2BIG;
+	} else
+		return ret;
+
+	if (size && args->v0.data[0]) {
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x616618 + hoff, 0x8000000c, 0x80000001);
+			nv_wait(priv, 0x616618 + hoff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x616548 + hoff, 0x00000070, 0x00000000);
 		for (i = 0; i < size; i++)
-			nv_wr32(priv, 0x10ec00 + soff, (i << 8) | data[i]);
+			nv_wr32(priv, 0x10ec00 + soff, (i << 8) | args->v0.data[i]);
 		for (; i < 0x60; i++)
 			nv_wr32(priv, 0x10ec00 + soff, (i << 8));
 		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000003);
-	} else
-	if (data) {
-		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000001);
 	} else {
-		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000000);
+		if (outp->info.type == DCB_OUTPUT_DP) {
+			nv_mask(priv, 0x616618 + hoff, 0x80000001, 0x80000000);
+			nv_wait(priv, 0x616618 + hoff, 0x80000000, 0x00000000);
+		}
+		nv_mask(priv, 0x10ec10 + soff, 0x80000003, 0x80000000 | !!size);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminv84.c
index 7fdade6e604d..fa276dede9cd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminv84.c
@@ -22,17 +22,38 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
 int
-nv84_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
+nv84_hdmi_ctrl(NV50_DISP_MTHD_V1)
 {
 	const u32 hoff = (head * 0x800);
+	union {
+		struct nv50_disp_sor_hdmi_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl;
+	int ret;
 
-	if (!(data & NV84_DISP_SOR_HDMI_PWR_STATE_ON)) {
+	nv_ioctl(object, "disp sor hdmi ctrl size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor hdmi ctrl vers %d state %d "
+				 "max_ac_packet %d rekey %d\n",
+			 args->v0.version, args->v0.state,
+			 args->v0.max_ac_packet, args->v0.rekey);
+		if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f)
+			return -EINVAL;
+		ctrl  = 0x40000000 * !!args->v0.state;
+		ctrl |= args->v0.max_ac_packet << 16;
+		ctrl |= args->v0.rekey;
+		ctrl |= 0x1f000000; /* ??? */
+	} else
+		return ret;
+
+	if (!(ctrl & 0x40000000)) {
 		nv_mask(priv, 0x6165a4 + hoff, 0x40000000, 0x00000000);
 		nv_mask(priv, 0x616520 + hoff, 0x00000001, 0x00000000);
 		nv_mask(priv, 0x616500 + hoff, 0x00000001, 0x00000000);
@@ -65,6 +86,6 @@ nv84_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
 	nv_mask(priv, 0x61733c, 0x00100000, 0x00000000); /* !RESETF */
 
 	/* HDMI_CTRL */
-	nv_mask(priv, 0x6165a4 + hoff, 0x5f1f007f, data | 0x1f000000 /* ??? */);
+	nv_mask(priv, 0x6165a4 + hoff, 0x5f1f007f, ctrl);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
index db8c6fd46278..57eeed1d1942 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
@@ -22,17 +22,38 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
 int
-nva3_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
+nva3_hdmi_ctrl(NV50_DISP_MTHD_V1)
 {
-	const u32 soff = (or * 0x800);
+	const u32 soff = outp->or * 0x800;
+	union {
+		struct nv50_disp_sor_hdmi_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl;
+	int ret;
 
-	if (!(data & NV84_DISP_SOR_HDMI_PWR_STATE_ON)) {
+	nv_ioctl(object, "disp sor hdmi ctrl size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor hdmi ctrl vers %d state %d "
+				 "max_ac_packet %d rekey %d\n",
+			 args->v0.version, args->v0.state,
+			 args->v0.max_ac_packet, args->v0.rekey);
+		if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f)
+			return -EINVAL;
+		ctrl  = 0x40000000 * !!args->v0.state;
+		ctrl |= args->v0.max_ac_packet << 16;
+		ctrl |= args->v0.rekey;
+		ctrl |= 0x1f000000; /* ??? */
+	} else
+		return ret;
+
+	if (!(ctrl & 0x40000000)) {
 		nv_mask(priv, 0x61c5a4 + soff, 0x40000000, 0x00000000);
 		nv_mask(priv, 0x61c520 + soff, 0x00000001, 0x00000000);
 		nv_mask(priv, 0x61c500 + soff, 0x00000001, 0x00000000);
@@ -65,6 +86,6 @@ nva3_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
 	nv_mask(priv, 0x61733c, 0x00100000, 0x00000000); /* !RESETF */
 
 	/* HDMI_CTRL */
-	nv_mask(priv, 0x61c5a4 + soff, 0x5f1f007f, data | 0x1f000000 /* ??? */);
+	nv_mask(priv, 0x61c5a4 + soff, 0x5f1f007f, ctrl);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
index 5151bb261832..bac4fc4570f0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminvd0.c
@@ -22,17 +22,37 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
 int
-nvd0_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
+nvd0_hdmi_ctrl(NV50_DISP_MTHD_V1)
 {
 	const u32 hoff = (head * 0x800);
+	union {
+		struct nv50_disp_sor_hdmi_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl;
+	int ret;
 
-	if (!(data & NV84_DISP_SOR_HDMI_PWR_STATE_ON)) {
+	nv_ioctl(object, "disp sor hdmi ctrl size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor hdmi ctrl vers %d state %d "
+				 "max_ac_packet %d rekey %d\n",
+			 args->v0.version, args->v0.state,
+			 args->v0.max_ac_packet, args->v0.rekey);
+		if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f)
+			return -EINVAL;
+		ctrl  = 0x40000000 * !!args->v0.state;
+		ctrl |= args->v0.max_ac_packet << 16;
+		ctrl |= args->v0.rekey;
+	} else
+		return ret;
+
+	if (!(ctrl & 0x40000000)) {
 		nv_mask(priv, 0x616798 + hoff, 0x40000000, 0x00000000);
 		nv_mask(priv, 0x6167a4 + hoff, 0x00000001, 0x00000000);
 		nv_mask(priv, 0x616714 + hoff, 0x00000001, 0x00000000);
@@ -54,9 +74,6 @@ nvd0_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
 	nv_mask(priv, 0x6167a4 + hoff, 0x00000001, 0x00000001);
 
 	/* HDMI_CTRL */
-	nv_mask(priv, 0x616798 + hoff, 0x401f007f, data);
-
-	/* NFI, audio doesn't work without it though.. */
-	nv_mask(priv, 0x616548 + hoff, 0x00000070, 0x00000000);
+	nv_mask(priv, 0x616798 + hoff, 0x401f007f, ctrl);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c
new file mode 100644
index 000000000000..528d14ec2f7f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminve0.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+
+#include "nv50.h"
+
+int
+nve0_hdmi_ctrl(NV50_DISP_MTHD_V1)
+{
+	const u32 hoff = (head * 0x800);
+	const u32 hdmi = (head * 0x400);
+	union {
+		struct nv50_disp_sor_hdmi_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl;
+	int ret;
+
+	nv_ioctl(object, "disp sor hdmi ctrl size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor hdmi ctrl vers %d state %d "
+				 "max_ac_packet %d rekey %d\n",
+			 args->v0.version, args->v0.state,
+			 args->v0.max_ac_packet, args->v0.rekey);
+		if (args->v0.max_ac_packet > 0x1f || args->v0.rekey > 0x7f)
+			return -EINVAL;
+		ctrl  = 0x40000000 * !!args->v0.state;
+		ctrl |= args->v0.max_ac_packet << 16;
+		ctrl |= args->v0.rekey;
+	} else
+		return ret;
+
+	if (!(ctrl & 0x40000000)) {
+		nv_mask(priv, 0x616798 + hoff, 0x40000000, 0x00000000);
+		nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000000);
+		nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000000);
+		return 0;
+	}
+
+	/* AVI InfoFrame */
+	nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000000);
+	nv_wr32(priv, 0x690008 + hdmi, 0x000d0282);
+	nv_wr32(priv, 0x69000c + hdmi, 0x0000006f);
+	nv_wr32(priv, 0x690010 + hdmi, 0x00000000);
+	nv_wr32(priv, 0x690014 + hdmi, 0x00000000);
+	nv_wr32(priv, 0x690018 + hdmi, 0x00000000);
+	nv_mask(priv, 0x690000 + hdmi, 0x00000001, 0x00000001);
+
+	/* ??? InfoFrame? */
+	nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000000);
+	nv_wr32(priv, 0x6900cc + hdmi, 0x00000010);
+	nv_mask(priv, 0x6900c0 + hdmi, 0x00000001, 0x00000001);
+
+	/* ??? */
+	nv_wr32(priv, 0x690080 + hdmi, 0x82000000);
+
+	/* HDMI_CTRL */
+	nv_mask(priv, 0x616798 + hoff, 0x401f007f, ctrl);
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
index a32666ed0c47..366f315fc9a5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv04.c
@@ -24,60 +24,100 @@
 
 #include "priv.h"
 
+#include <core/client.h>
 #include <core/event.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 struct nv04_disp_priv {
 	struct nouveau_disp base;
 };
 
 static int
-nv04_disp_scanoutpos(struct nouveau_object *object, u32 mthd,
-		     void *data, u32 size)
+nv04_disp_scanoutpos(struct nouveau_object *object, struct nv04_disp_priv *priv,
+		     void *data, u32 size, int head)
 {
-	struct nv04_disp_priv *priv = (void *)object->engine;
-	struct nv04_display_scanoutpos *args = data;
-	const int head = (mthd & NV04_DISP_MTHD_HEAD);
+	const u32 hoff = head * 0x2000;
+	union {
+		struct nv04_disp_scanoutpos_v0 v0;
+	} *args = data;
 	u32 line;
+	int ret;
+
+	nv_ioctl(object, "disp scanoutpos size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp scanoutpos vers %d\n", args->v0.version);
+		args->v0.vblanks = nv_rd32(priv, 0x680800 + hoff) & 0xffff;
+		args->v0.vtotal  = nv_rd32(priv, 0x680804 + hoff) & 0xffff;
+		args->v0.vblanke = args->v0.vtotal - 1;
+
+		args->v0.hblanks = nv_rd32(priv, 0x680820 + hoff) & 0xffff;
+		args->v0.htotal  = nv_rd32(priv, 0x680824 + hoff) & 0xffff;
+		args->v0.hblanke = args->v0.htotal - 1;
+
+		/*
+		 * If output is vga instead of digital then vtotal/htotal is
+		 * invalid so we have to give up and trigger the timestamping
+		 * fallback in the drm core.
+		 */
+		if (!args->v0.vtotal || !args->v0.htotal)
+			return -ENOTSUPP;
+
+		args->v0.time[0] = ktime_to_ns(ktime_get());
+		line = nv_rd32(priv, 0x600868 + hoff);
+		args->v0.time[1] = ktime_to_ns(ktime_get());
+		args->v0.hline = (line & 0xffff0000) >> 16;
+		args->v0.vline = (line & 0x0000ffff);
+	} else
+		return ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
-
-	args->vblanks = nv_rd32(priv, 0x680800 + (head * 0x2000)) & 0xffff;
-	args->vtotal  = nv_rd32(priv, 0x680804 + (head * 0x2000)) & 0xffff;
-	args->vblanke = args->vtotal - 1;
-
-	args->hblanks = nv_rd32(priv, 0x680820 + (head * 0x2000)) & 0xffff;
-	args->htotal  = nv_rd32(priv, 0x680824 + (head * 0x2000)) & 0xffff;
-	args->hblanke = args->htotal - 1;
-
-	/*
-	 * If output is vga instead of digital then vtotal/htotal is invalid
-	 * so we have to give up and trigger the timestamping fallback in the
-	 * drm core.
-	 */
-	if (!args->vtotal || !args->htotal)
-		return -ENOTSUPP;
-
-	args->time[0] = ktime_to_ns(ktime_get());
-	line = nv_rd32(priv, 0x600868 + (head * 0x2000));
-	args->time[1] = ktime_to_ns(ktime_get());
-	args->hline = (line & 0xffff0000) >> 16;
-	args->vline = (line & 0x0000ffff);
 	return 0;
 }
 
-#define HEAD_MTHD(n) (n), (n) + 0x01
+static int
+nv04_disp_mthd(struct nouveau_object *object, u32 mthd, void *data, u32 size)
+{
+	union {
+		struct nv04_disp_mthd_v0 v0;
+	} *args = data;
+	struct nv04_disp_priv *priv = (void *)object->engine;
+	int head, ret;
+
+	nv_ioctl(object, "disp mthd size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "disp mthd vers %d mthd %02x head %d\n",
+			 args->v0.version, args->v0.method, args->v0.head);
+		mthd = args->v0.method;
+		head = args->v0.head;
+	} else
+		return ret;
 
-static struct nouveau_omthds
-nv04_disp_omthds[] = {
-	{ HEAD_MTHD(NV04_DISP_SCANOUTPOS), nv04_disp_scanoutpos },
-	{}
+	if (head < 0 || head >= 2)
+		return -ENXIO;
+
+	switch (mthd) {
+	case NV04_DISP_SCANOUTPOS:
+		return nv04_disp_scanoutpos(object, priv, data, size, head);
+	default:
+		break;
+	}
+
+	return -EINVAL;
+}
+
+static struct nouveau_ofuncs
+nv04_disp_ofuncs = {
+	.ctor = _nouveau_object_ctor,
+	.dtor = nouveau_object_destroy,
+	.init = nouveau_object_init,
+	.fini = nouveau_object_fini,
+	.mthd = nv04_disp_mthd,
+	.ntfy = nouveau_disp_ntfy,
 };
 
 static struct nouveau_oclass
 nv04_disp_sclass[] = {
-	{ NV04_DISP_CLASS, &nouveau_object_ofuncs, nv04_disp_omthds },
+	{ NV04_DISP, &nv04_disp_ofuncs },
 	{},
 };
 
@@ -86,17 +126,26 @@ nv04_disp_sclass[] = {
  ******************************************************************************/
 
 static void
-nv04_disp_vblank_enable(struct nouveau_event *event, int type, int head)
+nv04_disp_vblank_init(struct nvkm_event *event, int type, int head)
 {
-	nv_wr32(event->priv, 0x600140 + (head * 0x2000) , 0x00000001);
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_wr32(disp, 0x600140 + (head * 0x2000) , 0x00000001);
 }
 
 static void
-nv04_disp_vblank_disable(struct nouveau_event *event, int type, int head)
+nv04_disp_vblank_fini(struct nvkm_event *event, int type, int head)
 {
-	nv_wr32(event->priv, 0x600140 + (head * 0x2000) , 0x00000000);
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_wr32(disp, 0x600140 + (head * 0x2000) , 0x00000000);
 }
 
+static const struct nvkm_event_func
+nv04_disp_vblank_func = {
+	.ctor = nouveau_disp_vblank_ctor,
+	.init = nv04_disp_vblank_init,
+	.fini = nv04_disp_vblank_fini,
+};
+
 static void
 nv04_disp_intr(struct nouveau_subdev *subdev)
 {
@@ -106,12 +155,12 @@ nv04_disp_intr(struct nouveau_subdev *subdev)
 	u32 pvideo;
 
 	if (crtc0 & 0x00000001) {
-		nouveau_event_trigger(priv->base.vblank, 1, 0);
+		nouveau_disp_vblank(&priv->base, 0);
 		nv_wr32(priv, 0x600100, 0x00000001);
 	}
 
 	if (crtc1 & 0x00000001) {
-		nouveau_event_trigger(priv->base.vblank, 1, 1);
+		nouveau_disp_vblank(&priv->base, 1);
 		nv_wr32(priv, 0x602100, 0x00000001);
 	}
 
@@ -140,9 +189,6 @@ nv04_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	nv_engine(priv)->sclass = nv04_disp_sclass;
 	nv_subdev(priv)->intr = nv04_disp_intr;
-	priv->base.vblank->priv = priv;
-	priv->base.vblank->enable = nv04_disp_vblank_enable;
-	priv->base.vblank->disable = nv04_disp_vblank_disable;
 	return 0;
 }
 
@@ -155,4 +201,5 @@ nv04_disp_oclass = &(struct nouveau_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.vblank = &nv04_disp_vblank_func,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 2283c442a10d..2df3a937037d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -23,10 +23,13 @@
  */
 
 #include <core/object.h>
+#include <core/client.h>
 #include <core/parent.h>
 #include <core/handle.h>
-#include <core/class.h>
 #include <core/enum.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+#include <nvif/event.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
@@ -43,14 +46,16 @@
  * EVO channel base class
  ******************************************************************************/
 
-int
+static int
 nv50_disp_chan_create_(struct nouveau_object *parent,
 		       struct nouveau_object *engine,
-		       struct nouveau_oclass *oclass, int chid,
+		       struct nouveau_oclass *oclass, int head,
 		       int length, void **pobject)
 {
+	const struct nv50_disp_chan_impl *impl = (void *)oclass->ofuncs;
 	struct nv50_disp_base *base = (void *)parent;
 	struct nv50_disp_chan *chan;
+	int chid = impl->chid + head;
 	int ret;
 
 	if (base->chan & (1 << chid))
@@ -63,12 +68,14 @@ nv50_disp_chan_create_(struct nouveau_object *parent,
 	chan = *pobject;
 	if (ret)
 		return ret;
-
 	chan->chid = chid;
+
+	nv_parent(chan)->object_attach = impl->attach;
+	nv_parent(chan)->object_detach = impl->detach;
 	return 0;
 }
 
-void
+static void
 nv50_disp_chan_destroy(struct nv50_disp_chan *chan)
 {
 	struct nv50_disp_base *base = (void *)nv_object(chan)->parent;
@@ -76,6 +83,81 @@ nv50_disp_chan_destroy(struct nv50_disp_chan *chan)
 	nouveau_namedb_destroy(&chan->base);
 }
 
+static void
+nv50_disp_chan_uevent_fini(struct nvkm_event *event, int type, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610028, 0x00000001 << index, 0x00000000 << index);
+}
+
+static void
+nv50_disp_chan_uevent_init(struct nvkm_event *event, int types, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610028, 0x00000001 << index, 0x00000001 << index);
+}
+
+void
+nv50_disp_chan_uevent_send(struct nv50_disp_priv *priv, int chid)
+{
+	struct nvif_notify_uevent_rep {
+	} rep;
+
+	nvkm_event_send(&priv->uevent, 1, chid, &rep, sizeof(rep));
+}
+
+int
+nv50_disp_chan_uevent_ctor(struct nouveau_object *object, void *data, u32 size,
+			   struct nvkm_notify *notify)
+{
+	struct nv50_disp_dmac *dmac = (void *)object;
+	union {
+		struct nvif_notify_uevent_req none;
+	} *args = data;
+	int ret;
+
+	if (nvif_unvers(args->none)) {
+		notify->size  = sizeof(struct nvif_notify_uevent_rep);
+		notify->types = 1;
+		notify->index = dmac->base.chid;
+		return 0;
+	}
+
+	return ret;
+}
+
+const struct nvkm_event_func
+nv50_disp_chan_uevent = {
+	.ctor = nv50_disp_chan_uevent_ctor,
+	.init = nv50_disp_chan_uevent_init,
+	.fini = nv50_disp_chan_uevent_fini,
+};
+
+int
+nv50_disp_chan_ntfy(struct nouveau_object *object, u32 type,
+		    struct nvkm_event **pevent)
+{
+	struct nv50_disp_priv *priv = (void *)object->engine;
+	switch (type) {
+	case NV50_DISP_CORE_CHANNEL_DMA_V0_NTFY_UEVENT:
+		*pevent = &priv->uevent;
+		return 0;
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
+int
+nv50_disp_chan_map(struct nouveau_object *object, u64 *addr, u32 *size)
+{
+	struct nv50_disp_chan *chan = (void *)object;
+	*addr = nv_device_resource_start(nv_device(object), 0) +
+		0x640000 + (chan->chid * 0x1000);
+	*size = 0x001000;
+	return 0;
+}
+
 u32
 nv50_disp_chan_rd32(struct nouveau_object *object, u64 addr)
 {
@@ -115,16 +197,16 @@ nv50_disp_dmac_object_detach(struct nouveau_object *parent, int cookie)
 	nouveau_ramht_remove(base->ramht, cookie);
 }
 
-int
+static int
 nv50_disp_dmac_create_(struct nouveau_object *parent,
 		       struct nouveau_object *engine,
-		       struct nouveau_oclass *oclass, u32 pushbuf, int chid,
+		       struct nouveau_oclass *oclass, u32 pushbuf, int head,
 		       int length, void **pobject)
 {
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	ret = nv50_disp_chan_create_(parent, engine, oclass, chid,
+	ret = nv50_disp_chan_create_(parent, engine, oclass, head,
 				     length, pobject);
 	dmac = *pobject;
 	if (ret)
@@ -179,7 +261,7 @@ nv50_disp_dmac_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610028, 0x00010001 << chid, 0x00010001 << chid);
+	nv_mask(priv, 0x610028, 0x00010000 << chid, 0x00010000 << chid);
 
 	/* initialise channel for dma command submission */
 	nv_wr32(priv, 0x610204 + (chid * 0x0010), dmac->push);
@@ -216,7 +298,7 @@ nv50_disp_dmac_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notifications */
 	nv_mask(priv, 0x610028, 0x00010001 << chid, 0x00000000 << chid);
 
 	return nv50_disp_chan_fini(&dmac->base, suspend);
@@ -397,27 +479,32 @@ nv50_disp_mast_mthd_chan = {
 	}
 };
 
-static int
+int
 nv50_disp_mast_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
-	struct nv50_display_mast_class *args = data;
+	union {
+		struct nv50_disp_core_channel_dma_v0 v0;
+	} *args = data;
 	struct nv50_disp_dmac *mast;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create disp core channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create disp core channel dma vers %d "
+				 "pushbuf %08x\n",
+			 args->v0.version, args->v0.pushbuf);
+	} else
+		return ret;
 
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
+	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->v0.pushbuf,
 				     0, sizeof(*mast), (void **)&mast);
 	*pobject = nv_object(mast);
 	if (ret)
 		return ret;
 
-	nv_parent(mast)->object_attach = nv50_disp_dmac_object_attach;
-	nv_parent(mast)->object_detach = nv50_disp_dmac_object_detach;
 	return 0;
 }
 
@@ -433,7 +520,7 @@ nv50_disp_mast_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610028, 0x00010001, 0x00010001);
+	nv_mask(priv, 0x610028, 0x00010000, 0x00010000);
 
 	/* attempt to unstick channel from some unknown state */
 	if ((nv_rd32(priv, 0x610200) & 0x009f0000) == 0x00020000)
@@ -473,20 +560,25 @@ nv50_disp_mast_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notifications */
 	nv_mask(priv, 0x610028, 0x00010001, 0x00000000);
 
 	return nv50_disp_chan_fini(&mast->base, suspend);
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nv50_disp_mast_ofuncs = {
-	.ctor = nv50_disp_mast_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nv50_disp_mast_init,
-	.fini = nv50_disp_mast_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_mast_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nv50_disp_mast_init,
+	.base.fini = nv50_disp_mast_fini,
+	.base.map  = nv50_disp_chan_map,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 0,
+	.attach = nv50_disp_dmac_object_attach,
+	.detach = nv50_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -543,39 +635,52 @@ nv50_disp_sync_mthd_chan = {
 	}
 };
 
-static int
+int
 nv50_disp_sync_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
-	struct nv50_display_sync_class *args = data;
+	union {
+		struct nv50_disp_base_channel_dma_v0 v0;
+	} *args = data;
+	struct nv50_disp_priv *priv = (void *)engine;
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*args) || args->head > 1)
-		return -EINVAL;
+	nv_ioctl(parent, "create disp base channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create disp base channel dma vers %d "
+				 "pushbuf %08x head %d\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.head);
+		if (args->v0.head > priv->head.nr)
+			return -EINVAL;
+	} else
+		return ret;
 
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-				     1 + args->head, sizeof(*dmac),
+	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->v0.pushbuf,
+				     args->v0.head, sizeof(*dmac),
 				     (void **)&dmac);
 	*pobject = nv_object(dmac);
 	if (ret)
 		return ret;
 
-	nv_parent(dmac)->object_attach = nv50_disp_dmac_object_attach;
-	nv_parent(dmac)->object_detach = nv50_disp_dmac_object_detach;
 	return 0;
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nv50_disp_sync_ofuncs = {
-	.ctor = nv50_disp_sync_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nv50_disp_dmac_init,
-	.fini = nv50_disp_dmac_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_sync_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nv50_disp_dmac_init,
+	.base.fini = nv50_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 1,
+	.attach = nv50_disp_dmac_object_attach,
+	.detach = nv50_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -620,39 +725,52 @@ nv50_disp_ovly_mthd_chan = {
 	}
 };
 
-static int
+int
 nv50_disp_ovly_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
-	struct nv50_display_ovly_class *args = data;
+	union {
+		struct nv50_disp_overlay_channel_dma_v0 v0;
+	} *args = data;
+	struct nv50_disp_priv *priv = (void *)engine;
 	struct nv50_disp_dmac *dmac;
 	int ret;
 
-	if (size < sizeof(*args) || args->head > 1)
-		return -EINVAL;
+	nv_ioctl(parent, "create disp overlay channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create disp overlay channel dma vers %d "
+				 "pushbuf %08x head %d\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.head);
+		if (args->v0.head > priv->head.nr)
+			return -EINVAL;
+	} else
+		return ret;
 
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-				     3 + args->head, sizeof(*dmac),
+	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->v0.pushbuf,
+				     args->v0.head, sizeof(*dmac),
 				     (void **)&dmac);
 	*pobject = nv_object(dmac);
 	if (ret)
 		return ret;
 
-	nv_parent(dmac)->object_attach = nv50_disp_dmac_object_attach;
-	nv_parent(dmac)->object_detach = nv50_disp_dmac_object_detach;
 	return 0;
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nv50_disp_ovly_ofuncs = {
-	.ctor = nv50_disp_ovly_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nv50_disp_dmac_init,
-	.fini = nv50_disp_dmac_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_ovly_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nv50_disp_dmac_init,
+	.base.fini = nv50_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 3,
+	.attach = nv50_disp_dmac_object_attach,
+	.detach = nv50_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -662,14 +780,14 @@ nv50_disp_ovly_ofuncs = {
 static int
 nv50_disp_pioc_create_(struct nouveau_object *parent,
 		       struct nouveau_object *engine,
-		       struct nouveau_oclass *oclass, int chid,
+		       struct nouveau_oclass *oclass, int head,
 		       int length, void **pobject)
 {
-	return nv50_disp_chan_create_(parent, engine, oclass, chid,
+	return nv50_disp_chan_create_(parent, engine, oclass, head,
 				      length, pobject);
 }
 
-static void
+void
 nv50_disp_pioc_dtor(struct nouveau_object *object)
 {
 	struct nv50_disp_pioc *pioc = (void *)object;
@@ -727,20 +845,29 @@ nv50_disp_pioc_fini(struct nouveau_object *object, bool suspend)
  * EVO immediate overlay channel objects
  ******************************************************************************/
 
-static int
+int
 nv50_disp_oimm_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
-	struct nv50_display_oimm_class *args = data;
+	union {
+		struct nv50_disp_overlay_v0 v0;
+	} *args = data;
+	struct nv50_disp_priv *priv = (void *)engine;
 	struct nv50_disp_pioc *pioc;
 	int ret;
 
-	if (size < sizeof(*args) || args->head > 1)
-		return -EINVAL;
+	nv_ioctl(parent, "create disp overlay size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create disp overlay vers %d head %d\n",
+			 args->v0.version, args->v0.head);
+		if (args->v0.head > priv->head.nr)
+			return -EINVAL;
+	} else
+		return ret;
 
-	ret = nv50_disp_pioc_create_(parent, engine, oclass, 5 + args->head,
+	ret = nv50_disp_pioc_create_(parent, engine, oclass, args->v0.head,
 				     sizeof(*pioc), (void **)&pioc);
 	*pobject = nv_object(pioc);
 	if (ret)
@@ -749,34 +876,46 @@ nv50_disp_oimm_ctor(struct nouveau_object *parent,
 	return 0;
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nv50_disp_oimm_ofuncs = {
-	.ctor = nv50_disp_oimm_ctor,
-	.dtor = nv50_disp_pioc_dtor,
-	.init = nv50_disp_pioc_init,
-	.fini = nv50_disp_pioc_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_oimm_ctor,
+	.base.dtor = nv50_disp_pioc_dtor,
+	.base.init = nv50_disp_pioc_init,
+	.base.fini = nv50_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 5,
 };
 
 /*******************************************************************************
  * EVO cursor channel objects
  ******************************************************************************/
 
-static int
+int
 nv50_disp_curs_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
-	struct nv50_display_curs_class *args = data;
+	union {
+		struct nv50_disp_cursor_v0 v0;
+	} *args = data;
+	struct nv50_disp_priv *priv = (void *)engine;
 	struct nv50_disp_pioc *pioc;
 	int ret;
 
-	if (size < sizeof(*args) || args->head > 1)
-		return -EINVAL;
+	nv_ioctl(parent, "create disp cursor size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create disp cursor vers %d head %d\n",
+			 args->v0.version, args->v0.head);
+		if (args->v0.head > priv->head.nr)
+			return -EINVAL;
+	} else
+		return ret;
 
-	ret = nv50_disp_pioc_create_(parent, engine, oclass, 7 + args->head,
+	ret = nv50_disp_pioc_create_(parent, engine, oclass, args->v0.head,
 				     sizeof(*pioc), (void **)&pioc);
 	*pobject = nv_object(pioc);
 	if (ret)
@@ -785,14 +924,17 @@ nv50_disp_curs_ctor(struct nouveau_object *parent,
 	return 0;
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nv50_disp_curs_ofuncs = {
-	.ctor = nv50_disp_curs_ctor,
-	.dtor = nv50_disp_pioc_dtor,
-	.init = nv50_disp_pioc_init,
-	.fini = nv50_disp_pioc_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_curs_ctor,
+	.base.dtor = nv50_disp_pioc_dtor,
+	.base.init = nv50_disp_pioc_init,
+	.base.fini = nv50_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 7,
 };
 
 /*******************************************************************************
@@ -800,47 +942,162 @@ nv50_disp_curs_ofuncs = {
  ******************************************************************************/
 
 int
-nv50_disp_base_scanoutpos(struct nouveau_object *object, u32 mthd,
-			  void *data, u32 size)
+nv50_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
 {
-	struct nv50_disp_priv *priv = (void *)object->engine;
-	struct nv04_display_scanoutpos *args = data;
-	const int head = (mthd & NV50_DISP_MTHD_HEAD);
-	u32 blanke, blanks, total;
+	const u32 blanke = nv_rd32(priv, 0x610aec + (head * 0x540));
+	const u32 blanks = nv_rd32(priv, 0x610af4 + (head * 0x540));
+	const u32 total  = nv_rd32(priv, 0x610afc + (head * 0x540));
+	union {
+		struct nv04_disp_scanoutpos_v0 v0;
+	} *args = data;
+	int ret;
+
+	nv_ioctl(object, "disp scanoutpos size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp scanoutpos vers %d\n", args->v0.version);
+		args->v0.vblanke = (blanke & 0xffff0000) >> 16;
+		args->v0.hblanke = (blanke & 0x0000ffff);
+		args->v0.vblanks = (blanks & 0xffff0000) >> 16;
+		args->v0.hblanks = (blanks & 0x0000ffff);
+		args->v0.vtotal  = ( total & 0xffff0000) >> 16;
+		args->v0.htotal  = ( total & 0x0000ffff);
+		args->v0.time[0] = ktime_to_ns(ktime_get());
+		args->v0.vline = /* vline read locks hline */
+			nv_rd32(priv, 0x616340 + (head * 0x800)) & 0xffff;
+		args->v0.time[1] = ktime_to_ns(ktime_get());
+		args->v0.hline =
+			nv_rd32(priv, 0x616344 + (head * 0x800)) & 0xffff;
+	} else
+		return ret;
 
-	if (size < sizeof(*args) || head >= priv->head.nr)
-		return -EINVAL;
-	blanke = nv_rd32(priv, 0x610aec + (head * 0x540));
-	blanks = nv_rd32(priv, 0x610af4 + (head * 0x540));
-	total  = nv_rd32(priv, 0x610afc + (head * 0x540));
-
-	args->vblanke = (blanke & 0xffff0000) >> 16;
-	args->hblanke = (blanke & 0x0000ffff);
-	args->vblanks = (blanks & 0xffff0000) >> 16;
-	args->hblanks = (blanks & 0x0000ffff);
-	args->vtotal  = ( total & 0xffff0000) >> 16;
-	args->htotal  = ( total & 0x0000ffff);
-
-	args->time[0] = ktime_to_ns(ktime_get());
-	args->vline   = nv_rd32(priv, 0x616340 + (head * 0x800)) & 0xffff;
-	args->time[1] = ktime_to_ns(ktime_get()); /* vline read locks hline */
-	args->hline   = nv_rd32(priv, 0x616344 + (head * 0x800)) & 0xffff;
 	return 0;
 }
 
-static void
-nv50_disp_base_vblank_enable(struct nouveau_event *event, int type, int head)
+int
+nv50_disp_base_mthd(struct nouveau_object *object, u32 mthd,
+		    void *data, u32 size)
 {
-	nv_mask(event->priv, 0x61002c, (4 << head), (4 << head));
-}
+	const struct nv50_disp_impl *impl = (void *)nv_oclass(object->engine);
+	union {
+		struct nv50_disp_mthd_v0 v0;
+		struct nv50_disp_mthd_v1 v1;
+	} *args = data;
+	struct nv50_disp_priv *priv = (void *)object->engine;
+	struct nvkm_output *outp = NULL;
+	struct nvkm_output *temp;
+	u16 type, mask = 0;
+	int head, ret;
 
-static void
-nv50_disp_base_vblank_disable(struct nouveau_event *event, int type, int head)
-{
-	nv_mask(event->priv, 0x61002c, (4 << head), 0);
+	if (mthd != NV50_DISP_MTHD)
+		return -EINVAL;
+
+	nv_ioctl(object, "disp mthd size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(object, "disp mthd vers %d mthd %02x head %d\n",
+			 args->v0.version, args->v0.method, args->v0.head);
+		mthd = args->v0.method;
+		head = args->v0.head;
+	} else
+	if (nvif_unpack(args->v1, 1, 1, true)) {
+		nv_ioctl(object, "disp mthd vers %d mthd %02x "
+				 "type %04x mask %04x\n",
+			 args->v1.version, args->v1.method,
+			 args->v1.hasht, args->v1.hashm);
+		mthd = args->v1.method;
+		type = args->v1.hasht;
+		mask = args->v1.hashm;
+		head = ffs((mask >> 8) & 0x0f) - 1;
+	} else
+		return ret;
+
+	if (head < 0 || head >= priv->head.nr)
+		return -ENXIO;
+
+	if (mask) {
+		list_for_each_entry(temp, &priv->base.outp, head) {
+			if ((temp->info.hasht         == type) &&
+			    (temp->info.hashm & mask) == mask) {
+				outp = temp;
+				break;
+			}
+		}
+		if (outp == NULL)
+			return -ENXIO;
+	}
+
+	switch (mthd) {
+	case NV50_DISP_SCANOUTPOS:
+		return impl->head.scanoutpos(object, priv, data, size, head);
+	default:
+		break;
+	}
+
+	switch (mthd * !!outp) {
+	case NV50_DISP_MTHD_V1_DAC_PWR:
+		return priv->dac.power(object, priv, data, size, head, outp);
+	case NV50_DISP_MTHD_V1_DAC_LOAD:
+		return priv->dac.sense(object, priv, data, size, head, outp);
+	case NV50_DISP_MTHD_V1_SOR_PWR:
+		return priv->sor.power(object, priv, data, size, head, outp);
+	case NV50_DISP_MTHD_V1_SOR_HDA_ELD:
+		if (!priv->sor.hda_eld)
+			return -ENODEV;
+		return priv->sor.hda_eld(object, priv, data, size, head, outp);
+	case NV50_DISP_MTHD_V1_SOR_HDMI_PWR:
+		if (!priv->sor.hdmi)
+			return -ENODEV;
+		return priv->sor.hdmi(object, priv, data, size, head, outp);
+	case NV50_DISP_MTHD_V1_SOR_LVDS_SCRIPT: {
+		union {
+			struct nv50_disp_sor_lvds_script_v0 v0;
+		} *args = data;
+		nv_ioctl(object, "disp sor lvds script size %d\n", size);
+		if (nvif_unpack(args->v0, 0, 0, false)) {
+			nv_ioctl(object, "disp sor lvds script "
+					 "vers %d name %04x\n",
+				 args->v0.version, args->v0.script);
+			priv->sor.lvdsconf = args->v0.script;
+			return 0;
+		} else
+			return ret;
+	}
+		break;
+	case NV50_DISP_MTHD_V1_SOR_DP_PWR: {
+		struct nvkm_output_dp *outpdp = (void *)outp;
+		union {
+			struct nv50_disp_sor_dp_pwr_v0 v0;
+		} *args = data;
+		nv_ioctl(object, "disp sor dp pwr size %d\n", size);
+		if (nvif_unpack(args->v0, 0, 0, false)) {
+			nv_ioctl(object, "disp sor dp pwr vers %d state %d\n",
+				 args->v0.version, args->v0.state);
+			if (args->v0.state == 0) {
+				nvkm_notify_put(&outpdp->irq);
+				((struct nvkm_output_dp_impl *)nv_oclass(outp))
+					->lnk_pwr(outpdp, 0);
+				atomic_set(&outpdp->lt.done, 0);
+				return 0;
+			} else
+			if (args->v0.state != 0) {
+				nvkm_output_dp_train(&outpdp->base, 0, true);
+				return 0;
+			}
+		} else
+			return ret;
+	}
+		break;
+	case NV50_DISP_MTHD_V1_PIOR_PWR:
+		if (!priv->pior.power)
+			return -ENODEV;
+		return priv->pior.power(object, priv, data, size, head, outp);
+	default:
+		break;
+	}
+
+	return -EINVAL;
 }
 
-static int
+int
 nv50_disp_base_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
@@ -856,14 +1113,11 @@ nv50_disp_base_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	priv->base.vblank->priv = priv;
-	priv->base.vblank->enable = nv50_disp_base_vblank_enable;
-	priv->base.vblank->disable = nv50_disp_base_vblank_disable;
 	return nouveau_ramht_new(nv_object(base), nv_object(base), 0x1000, 0,
 				&base->ramht);
 }
 
-static void
+void
 nv50_disp_base_dtor(struct nouveau_object *object)
 {
 	struct nv50_disp_base *base = (void *)object;
@@ -958,34 +1212,23 @@ nv50_disp_base_ofuncs = {
 	.dtor = nv50_disp_base_dtor,
 	.init = nv50_disp_base_init,
 	.fini = nv50_disp_base_fini,
-};
-
-static struct nouveau_omthds
-nv50_disp_base_omthds[] = {
-	{ HEAD_MTHD(NV50_DISP_SCANOUTPOS)     , nv50_disp_base_scanoutpos },
-	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
-	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
-	{},
+	.mthd = nv50_disp_base_mthd,
+	.ntfy = nouveau_disp_ntfy,
 };
 
 static struct nouveau_oclass
 nv50_disp_base_oclass[] = {
-	{ NV50_DISP_CLASS, &nv50_disp_base_ofuncs, nv50_disp_base_omthds },
+	{ NV50_DISP, &nv50_disp_base_ofuncs },
 	{}
 };
 
 static struct nouveau_oclass
 nv50_disp_sclass[] = {
-	{ NV50_DISP_MAST_CLASS, &nv50_disp_mast_ofuncs },
-	{ NV50_DISP_SYNC_CLASS, &nv50_disp_sync_ofuncs },
-	{ NV50_DISP_OVLY_CLASS, &nv50_disp_ovly_ofuncs },
-	{ NV50_DISP_OIMM_CLASS, &nv50_disp_oimm_ofuncs },
-	{ NV50_DISP_CURS_CLASS, &nv50_disp_curs_ofuncs },
+	{ NV50_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
+	{ NV50_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ NV50_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
+	{ NV50_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
+	{ NV50_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
 	{}
 };
 
@@ -1005,7 +1248,7 @@ nv50_disp_data_ctor(struct nouveau_object *parent,
 	int ret = -EBUSY;
 
 	/* no context needed for channel objects... */
-	if (nv_mclass(parent) != NV_DEVICE_CLASS) {
+	if (nv_mclass(parent) != NV_DEVICE) {
 		atomic_inc(&parent->refcount);
 		*pobject = parent;
 		return 1;
@@ -1040,6 +1283,27 @@ nv50_disp_cclass = {
  * Display engine implementation
  ******************************************************************************/
 
+static void
+nv50_disp_vblank_fini(struct nvkm_event *event, int type, int head)
+{
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_mask(disp, 0x61002c, (4 << head), 0);
+}
+
+static void
+nv50_disp_vblank_init(struct nvkm_event *event, int type, int head)
+{
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_mask(disp, 0x61002c, (4 << head), (4 << head));
+}
+
+const struct nvkm_event_func
+nv50_disp_vblank_func = {
+	.ctor = nouveau_disp_vblank_ctor,
+	.init = nv50_disp_vblank_init,
+	.fini = nv50_disp_vblank_fini,
+};
+
 static const struct nouveau_enum
 nv50_disp_intr_error_type[] = {
 	{ 3, "ILLEGAL_MTHD" },
@@ -1366,7 +1630,7 @@ nv50_disp_intr_unk20_1(struct nv50_disp_priv *priv, int head)
 }
 
 static void
-nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
+nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv, int head,
 			  struct dcb_output *outp, u32 pclk)
 {
 	const int link = !(outp->sorconf.link & 1);
@@ -1375,24 +1639,36 @@ nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
 	const u32 loff = (link * 0x080) + soff;
 	const u32 ctrl = nv_rd32(priv, 0x610794 + (or * 8));
 	const u32 symbol = 100000;
-	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff) & 0x0000f0000;
+	const s32 vactive = nv_rd32(priv, 0x610af8 + (head * 0x540)) & 0xffff;
+	const s32 vblanke = nv_rd32(priv, 0x610ae8 + (head * 0x540)) & 0xffff;
+	const s32 vblanks = nv_rd32(priv, 0x610af0 + (head * 0x540)) & 0xffff;
+	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff);
 	u32 clksor = nv_rd32(priv, 0x614300 + soff);
 	int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0;
 	int TU, VTUi, VTUf, VTUa;
 	u64 link_data_rate, link_ratio, unk;
 	u32 best_diff = 64 * symbol;
-	u32 link_nr, link_bw, bits, r;
-
-	/* calculate packed data rate for each lane */
-	if      (dpctrl > 0x00030000) link_nr = 4;
-	else if (dpctrl > 0x00010000) link_nr = 2;
-	else			      link_nr = 1;
-
-	if (clksor & 0x000c0000)
-		link_bw = 270000;
-	else
-		link_bw = 162000;
-
+	u32 link_nr, link_bw, bits;
+	u64 value;
+
+	link_bw = (clksor & 0x000c0000) ? 270000 : 162000;
+	link_nr = hweight32(dpctrl & 0x000f0000);
+
+	/* symbols/hblank - algorithm taken from comments in tegra driver */
+	value = vblanke + vactive - vblanks - 7;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - (3 * !!(dpctrl & 0x00004000)) - (12 / link_nr);
+	nv_mask(priv, 0x61c1e8 + soff, 0x0000ffff, value);
+
+	/* symbols/vblank - algorithm taken from comments in tegra driver */
+	value = vblanks - vblanke - 25;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - ((36 / link_nr) + 3) - 1;
+	nv_mask(priv, 0x61c1ec + soff, 0x00ffffff, value);
+
+	/* watermark / activesym */
 	if      ((ctrl & 0xf0000) == 0x60000) bits = 30;
 	else if ((ctrl & 0xf0000) == 0x50000) bits = 24;
 	else                                  bits = 18;
@@ -1401,7 +1677,7 @@ nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
 
 	/* calculate ratio of packed data rate to link symbol rate */
 	link_ratio = link_data_rate * symbol;
-	r = do_div(link_ratio, link_bw);
+	do_div(link_ratio, link_bw);
 
 	for (TU = 64; TU >= 32; TU--) {
 		/* calculate average number of valid symbols in each TU */
@@ -1462,8 +1738,8 @@ nv50_disp_intr_unk20_2_dp(struct nv50_disp_priv *priv,
 	/* XXX close to vbios numbers, but not right */
 	unk  = (symbol - link_ratio) * bestTU;
 	unk *= link_ratio;
-	r = do_div(unk, symbol);
-	r = do_div(unk, symbol);
+	do_div(unk, symbol);
+	do_div(unk, symbol);
 	unk += 6;
 
 	nv_mask(priv, 0x61c10c + loff, 0x000001fc, bestTU << 2);
@@ -1538,7 +1814,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 	} else
 	if (!outp->info.location) {
 		if (outp->info.type == DCB_OUTPUT_DP)
-			nv50_disp_intr_unk20_2_dp(priv, &outp->info, pclk);
+			nv50_disp_intr_unk20_2_dp(priv, head, &outp->info, pclk);
 		oreg = 0x614300 + (ffs(outp->info.or) - 1) * 0x800;
 		oval = (conf & 0x0100) ? 0x00000101 : 0x00000000;
 		hval = 0x00000000;
@@ -1570,9 +1846,10 @@ nv50_disp_intr_unk40_0_tmds(struct nv50_disp_priv *priv, struct dcb_output *outp
 	const int   or = ffs(outp->or) - 1;
 	const u32 loff = (or * 0x800) + (link * 0x80);
 	const u16 mask = (outp->sorconf.link << 6) | outp->or;
+	struct dcb_output match;
 	u8  ver, hdr;
 
-	if (dcb_outp_match(bios, DCB_OUTPUT_DP, mask, &ver, &hdr, outp))
+	if (dcb_outp_match(bios, DCB_OUTPUT_DP, mask, &ver, &hdr, &match))
 		nv_mask(priv, 0x61c10c + loff, 0x00000001, 0x00000000);
 }
 
@@ -1653,14 +1930,20 @@ nv50_disp_intr(struct nouveau_subdev *subdev)
 		intr0 &= ~(0x00010000 << chid);
 	}
 
+	while (intr0 & 0x0000001f) {
+		u32 chid = __ffs(intr0 & 0x0000001f);
+		nv50_disp_chan_uevent_send(priv, chid);
+		intr0 &= ~(0x00000001 << chid);
+	}
+
 	if (intr1 & 0x00000004) {
-		nouveau_event_trigger(priv->base.vblank, 1, 0);
+		nouveau_disp_vblank(&priv->base, 0);
 		nv_wr32(priv, 0x610024, 0x00000004);
 		intr1 &= ~0x00000004;
 	}
 
 	if (intr1 & 0x00000008) {
-		nouveau_event_trigger(priv->base.vblank, 1, 1);
+		nouveau_disp_vblank(&priv->base, 1);
 		nv_wr32(priv, 0x610024, 0x00000008);
 		intr1 &= ~0x00000008;
 	}
@@ -1687,6 +1970,10 @@ nv50_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nv50_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
@@ -1718,9 +2005,11 @@ nv50_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
 	.mthd.core = &nv50_disp_mast_mthd_chan,
 	.mthd.base = &nv50_disp_sync_mthd_chan,
 	.mthd.ovly = &nv50_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
+	.head.scanoutpos = nv50_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
index 1a886472b6f5..5279feefec06 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.h
@@ -14,15 +14,10 @@
 #include "outp.h"
 #include "outpdp.h"
 
-struct nv50_disp_impl {
-	struct nouveau_disp_impl base;
-	struct {
-		const struct nv50_disp_mthd_chan *core;
-		const struct nv50_disp_mthd_chan *base;
-		const struct nv50_disp_mthd_chan *ovly;
-		int prev;
-	} mthd;
-};
+#define NV50_DISP_MTHD_ struct nouveau_object *object,                         \
+	struct nv50_disp_priv *priv, void *data, u32 size
+#define NV50_DISP_MTHD_V0 NV50_DISP_MTHD_, int head
+#define NV50_DISP_MTHD_V1 NV50_DISP_MTHD_, int head, struct nvkm_output *outp
 
 struct nv50_disp_priv {
 	struct nouveau_disp base;
@@ -31,49 +26,60 @@ struct nv50_disp_priv {
 	struct work_struct supervisor;
 	u32 super;
 
+	struct nvkm_event uevent;
+
 	struct {
 		int nr;
 	} head;
 	struct {
 		int nr;
-		int (*power)(struct nv50_disp_priv *, int dac, u32 data);
-		int (*sense)(struct nv50_disp_priv *, int dac, u32 load);
+		int (*power)(NV50_DISP_MTHD_V1);
+		int (*sense)(NV50_DISP_MTHD_V1);
 	} dac;
 	struct {
 		int nr;
-		int (*power)(struct nv50_disp_priv *, int sor, u32 data);
-		int (*hda_eld)(struct nv50_disp_priv *, int sor, u8 *, u32);
-		int (*hdmi)(struct nv50_disp_priv *, int head, int sor, u32);
+		int (*power)(NV50_DISP_MTHD_V1);
+		int (*hda_eld)(NV50_DISP_MTHD_V1);
+		int (*hdmi)(NV50_DISP_MTHD_V1);
 		u32 lvdsconf;
 	} sor;
 	struct {
 		int nr;
-		int (*power)(struct nv50_disp_priv *, int ext, u32 data);
+		int (*power)(NV50_DISP_MTHD_V1);
 		u8 type[3];
 	} pior;
 };
 
-#define HEAD_MTHD(n) (n), (n) + 0x03
-
-int nv50_disp_base_scanoutpos(struct nouveau_object *, u32, void *, u32);
+struct nv50_disp_impl {
+	struct nouveau_disp_impl base;
+	struct {
+		const struct nv50_disp_mthd_chan *core;
+		const struct nv50_disp_mthd_chan *base;
+		const struct nv50_disp_mthd_chan *ovly;
+		int prev;
+	} mthd;
+	struct {
+		int (*scanoutpos)(NV50_DISP_MTHD_V0);
+	} head;
+};
 
-#define DAC_MTHD(n) (n), (n) + 0x03
+int nv50_disp_base_scanoutpos(NV50_DISP_MTHD_V0);
+int nv50_disp_base_mthd(struct nouveau_object *, u32, void *, u32);
 
-int nv50_dac_mthd(struct nouveau_object *, u32, void *, u32);
-int nv50_dac_power(struct nv50_disp_priv *, int, u32);
-int nv50_dac_sense(struct nv50_disp_priv *, int, u32);
+int nvd0_disp_base_scanoutpos(NV50_DISP_MTHD_V0);
 
-#define SOR_MTHD(n) (n), (n) + 0x3f
+int nv50_dac_power(NV50_DISP_MTHD_V1);
+int nv50_dac_sense(NV50_DISP_MTHD_V1);
 
-int nva3_hda_eld(struct nv50_disp_priv *, int, u8 *, u32);
-int nvd0_hda_eld(struct nv50_disp_priv *, int, u8 *, u32);
+int nva3_hda_eld(NV50_DISP_MTHD_V1);
+int nvd0_hda_eld(NV50_DISP_MTHD_V1);
 
-int nv84_hdmi_ctrl(struct nv50_disp_priv *, int, int, u32);
-int nva3_hdmi_ctrl(struct nv50_disp_priv *, int, int, u32);
-int nvd0_hdmi_ctrl(struct nv50_disp_priv *, int, int, u32);
+int nv84_hdmi_ctrl(NV50_DISP_MTHD_V1);
+int nva3_hdmi_ctrl(NV50_DISP_MTHD_V1);
+int nvd0_hdmi_ctrl(NV50_DISP_MTHD_V1);
+int nve0_hdmi_ctrl(NV50_DISP_MTHD_V1);
 
-int nv50_sor_mthd(struct nouveau_object *, u32, void *, u32);
-int nv50_sor_power(struct nv50_disp_priv *, int, u32);
+int nv50_sor_power(NV50_DISP_MTHD_V1);
 
 int nv94_sor_dp_train_init(struct nv50_disp_priv *, int, int, int, u16, u16,
 		           u32, struct dcb_output *);
@@ -93,10 +99,7 @@ int nvd0_sor_dp_lnkctl(struct nv50_disp_priv *, int, int, int, u16, u16, u32,
 int nvd0_sor_dp_drvctl(struct nv50_disp_priv *, int, int, int, u16, u16, u32,
 		       struct dcb_output *);
 
-#define PIOR_MTHD(n) (n), (n) + 0x03
-
-int nv50_pior_mthd(struct nouveau_object *, u32, void *, u32);
-int nv50_pior_power(struct nv50_disp_priv *, int, u32);
+int nv50_pior_power(NV50_DISP_MTHD_V1);
 
 struct nv50_disp_base {
 	struct nouveau_parent base;
@@ -104,36 +107,48 @@ struct nv50_disp_base {
 	u32 chan;
 };
 
+struct nv50_disp_chan_impl {
+	struct nouveau_ofuncs base;
+	int chid;
+	int  (*attach)(struct nouveau_object *, struct nouveau_object *, u32);
+	void (*detach)(struct nouveau_object *, int);
+};
+
 struct nv50_disp_chan {
 	struct nouveau_namedb base;
 	int chid;
 };
 
-int  nv50_disp_chan_create_(struct nouveau_object *, struct nouveau_object *,
-			    struct nouveau_oclass *, int, int, void **);
-void nv50_disp_chan_destroy(struct nv50_disp_chan *);
+int  nv50_disp_chan_ntfy(struct nouveau_object *, u32, struct nvkm_event **);
+int  nv50_disp_chan_map(struct nouveau_object *, u64 *, u32 *);
 u32  nv50_disp_chan_rd32(struct nouveau_object *, u64);
 void nv50_disp_chan_wr32(struct nouveau_object *, u64, u32);
+extern const struct nvkm_event_func nv50_disp_chan_uevent;
+int  nv50_disp_chan_uevent_ctor(struct nouveau_object *, void *, u32,
+				struct nvkm_notify *);
+void nv50_disp_chan_uevent_send(struct nv50_disp_priv *, int);
+
+extern const struct nvkm_event_func nvd0_disp_chan_uevent;
 
 #define nv50_disp_chan_init(a)                                                 \
 	nouveau_namedb_init(&(a)->base)
 #define nv50_disp_chan_fini(a,b)                                               \
 	nouveau_namedb_fini(&(a)->base, (b))
 
-int  nv50_disp_dmac_create_(struct nouveau_object *, struct nouveau_object *,
-			    struct nouveau_oclass *, u32, int, int, void **);
-void nv50_disp_dmac_dtor(struct nouveau_object *);
-
 struct nv50_disp_dmac {
 	struct nv50_disp_chan base;
 	struct nouveau_dmaobj *pushdma;
 	u32 push;
 };
 
+void nv50_disp_dmac_dtor(struct nouveau_object *);
+
 struct nv50_disp_pioc {
 	struct nv50_disp_chan base;
 };
 
+void nv50_disp_pioc_dtor(struct nouveau_object *);
+
 struct nv50_disp_mthd_list {
 	u32 mthd;
 	u32 addr;
@@ -154,47 +169,67 @@ struct nv50_disp_mthd_chan {
 	} data[];
 };
 
-extern struct nouveau_ofuncs nv50_disp_mast_ofuncs;
+extern struct nv50_disp_chan_impl nv50_disp_mast_ofuncs;
+int nv50_disp_mast_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
 extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_base;
 extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_sor;
 extern const struct nv50_disp_mthd_list nv50_disp_mast_mthd_pior;
-extern struct nouveau_ofuncs nv50_disp_sync_ofuncs;
+extern struct nv50_disp_chan_impl nv50_disp_sync_ofuncs;
+int nv50_disp_sync_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
 extern const struct nv50_disp_mthd_list nv50_disp_sync_mthd_image;
-extern struct nouveau_ofuncs nv50_disp_ovly_ofuncs;
+extern struct nv50_disp_chan_impl nv50_disp_ovly_ofuncs;
+int nv50_disp_ovly_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
 extern const struct nv50_disp_mthd_list nv50_disp_ovly_mthd_base;
-extern struct nouveau_ofuncs nv50_disp_oimm_ofuncs;
-extern struct nouveau_ofuncs nv50_disp_curs_ofuncs;
+extern struct nv50_disp_chan_impl nv50_disp_oimm_ofuncs;
+int nv50_disp_oimm_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
+extern struct nv50_disp_chan_impl nv50_disp_curs_ofuncs;
+int nv50_disp_curs_ctor(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void *, u32,
+			struct nouveau_object **);
 extern struct nouveau_ofuncs nv50_disp_base_ofuncs;
+int  nv50_disp_base_ctor(struct nouveau_object *, struct nouveau_object *,
+			 struct nouveau_oclass *, void *, u32,
+			 struct nouveau_object **);
+void nv50_disp_base_dtor(struct nouveau_object *);
+extern struct nouveau_omthds nv50_disp_base_omthds[];
 extern struct nouveau_oclass nv50_disp_cclass;
 void nv50_disp_mthd_chan(struct nv50_disp_priv *, int debug, int head,
 			 const struct nv50_disp_mthd_chan *);
 void nv50_disp_intr_supervisor(struct work_struct *);
 void nv50_disp_intr(struct nouveau_subdev *);
+extern const struct nvkm_event_func nv50_disp_vblank_func;
 
 extern const struct nv50_disp_mthd_chan nv84_disp_mast_mthd_chan;
 extern const struct nv50_disp_mthd_list nv84_disp_mast_mthd_dac;
 extern const struct nv50_disp_mthd_list nv84_disp_mast_mthd_head;
 extern const struct nv50_disp_mthd_chan nv84_disp_sync_mthd_chan;
 extern const struct nv50_disp_mthd_chan nv84_disp_ovly_mthd_chan;
-extern struct nouveau_omthds nv84_disp_base_omthds[];
 
 extern const struct nv50_disp_mthd_chan nv94_disp_mast_mthd_chan;
 
-extern struct nouveau_ofuncs nvd0_disp_mast_ofuncs;
+extern struct nv50_disp_chan_impl nvd0_disp_mast_ofuncs;
 extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_base;
 extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_dac;
 extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_sor;
 extern const struct nv50_disp_mthd_list nvd0_disp_mast_mthd_pior;
-extern struct nouveau_ofuncs nvd0_disp_sync_ofuncs;
-extern struct nouveau_ofuncs nvd0_disp_ovly_ofuncs;
+extern struct nv50_disp_chan_impl nvd0_disp_sync_ofuncs;
+extern struct nv50_disp_chan_impl nvd0_disp_ovly_ofuncs;
 extern const struct nv50_disp_mthd_chan nvd0_disp_sync_mthd_chan;
-extern struct nouveau_ofuncs nvd0_disp_oimm_ofuncs;
-extern struct nouveau_ofuncs nvd0_disp_curs_ofuncs;
-extern struct nouveau_omthds nvd0_disp_base_omthds[];
+extern struct nv50_disp_chan_impl nvd0_disp_oimm_ofuncs;
+extern struct nv50_disp_chan_impl nvd0_disp_curs_ofuncs;
 extern struct nouveau_ofuncs nvd0_disp_base_ofuncs;
 extern struct nouveau_oclass nvd0_disp_cclass;
 void nvd0_disp_intr_supervisor(struct work_struct *);
 void nvd0_disp_intr(struct nouveau_subdev *);
+extern const struct nvkm_event_func nvd0_disp_vblank_func;
 
 extern const struct nv50_disp_mthd_chan nve0_disp_mast_mthd_chan;
 extern const struct nv50_disp_mthd_chan nve0_disp_ovly_mthd_chan;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
index 1cc62e434683..d36284715b2a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv84.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -204,31 +204,17 @@ nv84_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nv84_disp_sclass[] = {
-	{ NV84_DISP_MAST_CLASS, &nv50_disp_mast_ofuncs },
-	{ NV84_DISP_SYNC_CLASS, &nv50_disp_sync_ofuncs },
-	{ NV84_DISP_OVLY_CLASS, &nv50_disp_ovly_ofuncs },
-	{ NV84_DISP_OIMM_CLASS, &nv50_disp_oimm_ofuncs },
-	{ NV84_DISP_CURS_CLASS, &nv50_disp_curs_ofuncs },
+	{ G82_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
+	{ G82_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ G82_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
+	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
+	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
 	{}
 };
 
-struct nouveau_omthds
-nv84_disp_base_omthds[] = {
-	{ HEAD_MTHD(NV50_DISP_SCANOUTPOS)     , nv50_disp_base_scanoutpos },
-	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
-	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
-	{},
-};
-
 static struct nouveau_oclass
 nv84_disp_base_oclass[] = {
-	{ NV84_DISP_CLASS, &nv50_disp_base_ofuncs, nv84_disp_base_omthds },
+	{ G82_DISP, &nv50_disp_base_ofuncs },
 	{}
 };
 
@@ -250,6 +236,10 @@ nv84_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nv84_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
@@ -276,9 +266,11 @@ nv84_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
 	.mthd.core = &nv84_disp_mast_mthd_chan,
 	.mthd.base = &nv84_disp_sync_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
+	.head.scanoutpos = nv50_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
index 4f718a9f5aef..a117064002b1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv94.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -63,32 +63,17 @@ nv94_disp_mast_mthd_chan = {
 
 static struct nouveau_oclass
 nv94_disp_sclass[] = {
-	{ NV94_DISP_MAST_CLASS, &nv50_disp_mast_ofuncs },
-	{ NV94_DISP_SYNC_CLASS, &nv50_disp_sync_ofuncs },
-	{ NV94_DISP_OVLY_CLASS, &nv50_disp_ovly_ofuncs },
-	{ NV94_DISP_OIMM_CLASS, &nv50_disp_oimm_ofuncs },
-	{ NV94_DISP_CURS_CLASS, &nv50_disp_curs_ofuncs },
+	{ GT206_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
+	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT200_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
+	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
+	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
 	{}
 };
 
-static struct nouveau_omthds
-nv94_disp_base_omthds[] = {
-	{ HEAD_MTHD(NV50_DISP_SCANOUTPOS)     , nv50_disp_base_scanoutpos },
-	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
-	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_PWR)      , nv50_sor_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
-	{},
-};
-
 static struct nouveau_oclass
 nv94_disp_base_oclass[] = {
-	{ NV94_DISP_CLASS, &nv50_disp_base_ofuncs, nv94_disp_base_omthds },
+	{ GT206_DISP, &nv50_disp_base_ofuncs },
 	{}
 };
 
@@ -110,6 +95,10 @@ nv94_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nv94_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
@@ -143,9 +132,11 @@ nv94_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv94_disp_outp_sclass,
 	.mthd.core = &nv94_disp_mast_mthd_chan,
 	.mthd.base = &nv84_disp_sync_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
+	.head.scanoutpos = nv50_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
index 6237a9a36f70..c67e68aadd45 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva0.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -80,17 +80,17 @@ nva0_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nva0_disp_sclass[] = {
-	{ NVA0_DISP_MAST_CLASS, &nv50_disp_mast_ofuncs },
-	{ NVA0_DISP_SYNC_CLASS, &nv50_disp_sync_ofuncs },
-	{ NVA0_DISP_OVLY_CLASS, &nv50_disp_ovly_ofuncs },
-	{ NVA0_DISP_OIMM_CLASS, &nv50_disp_oimm_ofuncs },
-	{ NVA0_DISP_CURS_CLASS, &nv50_disp_curs_ofuncs },
+	{ GT200_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
+	{ GT200_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT200_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
+	{ G82_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
+	{ G82_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
 	{}
 };
 
 static struct nouveau_oclass
 nva0_disp_base_oclass[] = {
-	{ NVA0_DISP_CLASS, &nv50_disp_base_ofuncs, nv84_disp_base_omthds },
+	{ GT200_DISP, &nv50_disp_base_ofuncs },
 	{}
 };
 
@@ -112,6 +112,10 @@ nva0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nva0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
@@ -138,9 +142,11 @@ nva0_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv50_disp_outp_sclass,
 	.mthd.core = &nv84_disp_mast_mthd_chan,
 	.mthd.base = &nv84_disp_sync_mthd_chan,
 	.mthd.ovly = &nva0_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
+	.head.scanoutpos = nv50_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
index 019124d4782b..22969f355aae 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nva3.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -35,33 +35,17 @@
 
 static struct nouveau_oclass
 nva3_disp_sclass[] = {
-	{ NVA3_DISP_MAST_CLASS, &nv50_disp_mast_ofuncs },
-	{ NVA3_DISP_SYNC_CLASS, &nv50_disp_sync_ofuncs },
-	{ NVA3_DISP_OVLY_CLASS, &nv50_disp_ovly_ofuncs },
-	{ NVA3_DISP_OIMM_CLASS, &nv50_disp_oimm_ofuncs },
-	{ NVA3_DISP_CURS_CLASS, &nv50_disp_curs_ofuncs },
+	{ GT214_DISP_CORE_CHANNEL_DMA, &nv50_disp_mast_ofuncs.base },
+	{ GT214_DISP_BASE_CHANNEL_DMA, &nv50_disp_sync_ofuncs.base },
+	{ GT214_DISP_OVERLAY_CHANNEL_DMA, &nv50_disp_ovly_ofuncs.base },
+	{ GT214_DISP_OVERLAY, &nv50_disp_oimm_ofuncs.base },
+	{ GT214_DISP_CURSOR, &nv50_disp_curs_ofuncs.base },
 	{}
 };
 
-static struct nouveau_omthds
-nva3_disp_base_omthds[] = {
-	{ HEAD_MTHD(NV50_DISP_SCANOUTPOS)     , nv50_disp_base_scanoutpos },
-	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
-	{ SOR_MTHD(NVA3_DISP_SOR_HDA_ELD)     , nv50_sor_mthd },
-	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_PWR)      , nv50_sor_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
-	{},
-};
-
 static struct nouveau_oclass
 nva3_disp_base_oclass[] = {
-	{ NVA3_DISP_CLASS, &nv50_disp_base_ofuncs, nva3_disp_base_omthds },
+	{ GT214_DISP, &nv50_disp_base_ofuncs },
 	{}
 };
 
@@ -83,6 +67,10 @@ nva3_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nv50_disp_chan_uevent, 1, 9, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nva3_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nv50_disp_intr;
@@ -110,9 +98,11 @@ nva3_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nv50_disp_vblank_func,
 	.base.outp =  nv94_disp_outp_sclass,
 	.mthd.core = &nv94_disp_mast_mthd_chan,
 	.mthd.base = &nv84_disp_sync_mthd_chan,
 	.mthd.ovly = &nv84_disp_ovly_mthd_chan,
 	.mthd.prev = 0x000004,
+	.head.scanoutpos = nv50_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index fa30d8196f35..747e64bb9c06 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -23,9 +23,11 @@
  */
 
 #include <core/object.h>
+#include <core/client.h>
 #include <core/parent.h>
 #include <core/handle.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <engine/disp.h>
 
@@ -41,6 +43,31 @@
 #include "nv50.h"
 
 /*******************************************************************************
+ * EVO channel base class
+ ******************************************************************************/
+
+static void
+nvd0_disp_chan_uevent_fini(struct nvkm_event *event, int type, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610090, 0x00000001 << index, 0x00000000 << index);
+}
+
+static void
+nvd0_disp_chan_uevent_init(struct nvkm_event *event, int types, int index)
+{
+	struct nv50_disp_priv *priv = container_of(event, typeof(*priv), uevent);
+	nv_mask(priv, 0x610090, 0x00000001 << index, 0x00000001 << index);
+}
+
+const struct nvkm_event_func
+nvd0_disp_chan_uevent = {
+	.ctor = nv50_disp_chan_uevent_ctor,
+	.init = nvd0_disp_chan_uevent_init,
+	.fini = nvd0_disp_chan_uevent_fini,
+};
+
+/*******************************************************************************
  * EVO DMA channel base class
  ******************************************************************************/
 
@@ -75,7 +102,6 @@ nvd0_disp_dmac_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000001 << chid);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid);
 
 	/* initialise channel for dma command submission */
@@ -113,7 +139,7 @@ nvd0_disp_dmac_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000000);
 
@@ -265,30 +291,6 @@ nvd0_disp_mast_mthd_chan = {
 };
 
 static int
-nvd0_disp_mast_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
-{
-	struct nv50_display_mast_class *args = data;
-	struct nv50_disp_dmac *mast;
-	int ret;
-
-	if (size < sizeof(*args))
-		return -EINVAL;
-
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-				     0, sizeof(*mast), (void **)&mast);
-	*pobject = nv_object(mast);
-	if (ret)
-		return ret;
-
-	nv_parent(mast)->object_attach = nvd0_disp_dmac_object_attach;
-	nv_parent(mast)->object_detach = nvd0_disp_dmac_object_detach;
-	return 0;
-}
-
-static int
 nvd0_disp_mast_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
@@ -300,7 +302,6 @@ nvd0_disp_mast_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001, 0x00000001);
 	nv_mask(priv, 0x6100a0, 0x00000001, 0x00000001);
 
 	/* initialise channel for dma command submission */
@@ -335,21 +336,26 @@ nvd0_disp_mast_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001, 0x00000000);
 
 	return nv50_disp_chan_fini(&mast->base, suspend);
 }
 
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nvd0_disp_mast_ofuncs = {
-	.ctor = nvd0_disp_mast_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nvd0_disp_mast_init,
-	.fini = nvd0_disp_mast_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_mast_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nvd0_disp_mast_init,
+	.base.fini = nvd0_disp_mast_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 0,
+	.attach = nvd0_disp_dmac_object_attach,
+	.detach = nvd0_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -431,40 +437,19 @@ nvd0_disp_sync_mthd_chan = {
 	}
 };
 
-static int
-nvd0_disp_sync_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
-{
-	struct nv50_display_sync_class *args = data;
-	struct nv50_disp_priv *priv = (void *)engine;
-	struct nv50_disp_dmac *dmac;
-	int ret;
-
-	if (size < sizeof(*args) || args->head >= priv->head.nr)
-		return -EINVAL;
-
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-				     1 + args->head, sizeof(*dmac),
-				     (void **)&dmac);
-	*pobject = nv_object(dmac);
-	if (ret)
-		return ret;
-
-	nv_parent(dmac)->object_attach = nvd0_disp_dmac_object_attach;
-	nv_parent(dmac)->object_detach = nvd0_disp_dmac_object_detach;
-	return 0;
-}
-
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nvd0_disp_sync_ofuncs = {
-	.ctor = nvd0_disp_sync_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nvd0_disp_dmac_init,
-	.fini = nvd0_disp_dmac_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_sync_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nvd0_disp_dmac_init,
+	.base.fini = nvd0_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 1,
+	.attach = nvd0_disp_dmac_object_attach,
+	.detach = nvd0_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -533,40 +518,19 @@ nvd0_disp_ovly_mthd_chan = {
 	}
 };
 
-static int
-nvd0_disp_ovly_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
-{
-	struct nv50_display_ovly_class *args = data;
-	struct nv50_disp_priv *priv = (void *)engine;
-	struct nv50_disp_dmac *dmac;
-	int ret;
-
-	if (size < sizeof(*args) || args->head >= priv->head.nr)
-		return -EINVAL;
-
-	ret = nv50_disp_dmac_create_(parent, engine, oclass, args->pushbuf,
-				     5 + args->head, sizeof(*dmac),
-				     (void **)&dmac);
-	*pobject = nv_object(dmac);
-	if (ret)
-		return ret;
-
-	nv_parent(dmac)->object_attach = nvd0_disp_dmac_object_attach;
-	nv_parent(dmac)->object_detach = nvd0_disp_dmac_object_detach;
-	return 0;
-}
-
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nvd0_disp_ovly_ofuncs = {
-	.ctor = nvd0_disp_ovly_ctor,
-	.dtor = nv50_disp_dmac_dtor,
-	.init = nvd0_disp_dmac_init,
-	.fini = nvd0_disp_dmac_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_ovly_ctor,
+	.base.dtor = nv50_disp_dmac_dtor,
+	.base.init = nvd0_disp_dmac_init,
+	.base.fini = nvd0_disp_dmac_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 5,
+	.attach = nvd0_disp_dmac_object_attach,
+	.detach = nvd0_disp_dmac_object_detach,
 };
 
 /*******************************************************************************
@@ -574,23 +538,6 @@ nvd0_disp_ovly_ofuncs = {
  ******************************************************************************/
 
 static int
-nvd0_disp_pioc_create_(struct nouveau_object *parent,
-		       struct nouveau_object *engine,
-		       struct nouveau_oclass *oclass, int chid,
-		       int length, void **pobject)
-{
-	return nv50_disp_chan_create_(parent, engine, oclass, chid,
-				      length, pobject);
-}
-
-static void
-nvd0_disp_pioc_dtor(struct nouveau_object *object)
-{
-	struct nv50_disp_pioc *pioc = (void *)object;
-	nv50_disp_chan_destroy(&pioc->base);
-}
-
-static int
 nvd0_disp_pioc_init(struct nouveau_object *object)
 {
 	struct nv50_disp_priv *priv = (void *)object->engine;
@@ -603,7 +550,6 @@ nvd0_disp_pioc_init(struct nouveau_object *object)
 		return ret;
 
 	/* enable error reporting */
-	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000001 << chid);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid);
 
 	/* activate channel */
@@ -632,7 +578,7 @@ nvd0_disp_pioc_fini(struct nouveau_object *object, bool suspend)
 			return -EBUSY;
 	}
 
-	/* disable error reporting */
+	/* disable error reporting and completion notification */
 	nv_mask(priv, 0x610090, 0x00000001 << chid, 0x00000000);
 	nv_mask(priv, 0x6100a0, 0x00000001 << chid, 0x00000000);
 
@@ -643,152 +589,70 @@ nvd0_disp_pioc_fini(struct nouveau_object *object, bool suspend)
  * EVO immediate overlay channel objects
  ******************************************************************************/
 
-static int
-nvd0_disp_oimm_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
-{
-	struct nv50_display_oimm_class *args = data;
-	struct nv50_disp_priv *priv = (void *)engine;
-	struct nv50_disp_pioc *pioc;
-	int ret;
-
-	if (size < sizeof(*args) || args->head >= priv->head.nr)
-		return -EINVAL;
-
-	ret = nvd0_disp_pioc_create_(parent, engine, oclass, 9 + args->head,
-				     sizeof(*pioc), (void **)&pioc);
-	*pobject = nv_object(pioc);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nvd0_disp_oimm_ofuncs = {
-	.ctor = nvd0_disp_oimm_ctor,
-	.dtor = nvd0_disp_pioc_dtor,
-	.init = nvd0_disp_pioc_init,
-	.fini = nvd0_disp_pioc_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_oimm_ctor,
+	.base.dtor = nv50_disp_pioc_dtor,
+	.base.init = nvd0_disp_pioc_init,
+	.base.fini = nvd0_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 9,
 };
 
 /*******************************************************************************
  * EVO cursor channel objects
  ******************************************************************************/
 
-static int
-nvd0_disp_curs_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
-{
-	struct nv50_display_curs_class *args = data;
-	struct nv50_disp_priv *priv = (void *)engine;
-	struct nv50_disp_pioc *pioc;
-	int ret;
-
-	if (size < sizeof(*args) || args->head >= priv->head.nr)
-		return -EINVAL;
-
-	ret = nvd0_disp_pioc_create_(parent, engine, oclass, 13 + args->head,
-				     sizeof(*pioc), (void **)&pioc);
-	*pobject = nv_object(pioc);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-struct nouveau_ofuncs
+struct nv50_disp_chan_impl
 nvd0_disp_curs_ofuncs = {
-	.ctor = nvd0_disp_curs_ctor,
-	.dtor = nvd0_disp_pioc_dtor,
-	.init = nvd0_disp_pioc_init,
-	.fini = nvd0_disp_pioc_fini,
-	.rd32 = nv50_disp_chan_rd32,
-	.wr32 = nv50_disp_chan_wr32,
+	.base.ctor = nv50_disp_curs_ctor,
+	.base.dtor = nv50_disp_pioc_dtor,
+	.base.init = nvd0_disp_pioc_init,
+	.base.fini = nvd0_disp_pioc_fini,
+	.base.ntfy = nv50_disp_chan_ntfy,
+	.base.map  = nv50_disp_chan_map,
+	.base.rd32 = nv50_disp_chan_rd32,
+	.base.wr32 = nv50_disp_chan_wr32,
+	.chid = 13,
 };
 
 /*******************************************************************************
  * Base display object
  ******************************************************************************/
 
-static int
-nvd0_disp_base_scanoutpos(struct nouveau_object *object, u32 mthd,
-			  void *data, u32 size)
-{
-	struct nv50_disp_priv *priv = (void *)object->engine;
-	struct nv04_display_scanoutpos *args = data;
-	const int head = (mthd & NV50_DISP_MTHD_HEAD);
-	u32 blanke, blanks, total;
-
-	if (size < sizeof(*args) || head >= priv->head.nr)
-		return -EINVAL;
-
-	total  = nv_rd32(priv, 0x640414 + (head * 0x300));
-	blanke = nv_rd32(priv, 0x64041c + (head * 0x300));
-	blanks = nv_rd32(priv, 0x640420 + (head * 0x300));
-
-	args->vblanke = (blanke & 0xffff0000) >> 16;
-	args->hblanke = (blanke & 0x0000ffff);
-	args->vblanks = (blanks & 0xffff0000) >> 16;
-	args->hblanks = (blanks & 0x0000ffff);
-	args->vtotal  = ( total & 0xffff0000) >> 16;
-	args->htotal  = ( total & 0x0000ffff);
-
-	args->time[0] = ktime_to_ns(ktime_get());
-	args->vline   = nv_rd32(priv, 0x616340 + (head * 0x800)) & 0xffff;
-	args->time[1] = ktime_to_ns(ktime_get()); /* vline read locks hline */
-	args->hline   = nv_rd32(priv, 0x616344 + (head * 0x800)) & 0xffff;
-	return 0;
-}
-
-static void
-nvd0_disp_base_vblank_enable(struct nouveau_event *event, int type, int head)
-{
-	nv_mask(event->priv, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000001);
-}
-
-static void
-nvd0_disp_base_vblank_disable(struct nouveau_event *event, int type, int head)
-{
-	nv_mask(event->priv, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000000);
-}
-
-static int
-nvd0_disp_base_ctor(struct nouveau_object *parent,
-		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
+int
+nvd0_disp_base_scanoutpos(NV50_DISP_MTHD_V0)
 {
-	struct nv50_disp_priv *priv = (void *)engine;
-	struct nv50_disp_base *base;
+	const u32 total  = nv_rd32(priv, 0x640414 + (head * 0x300));
+	const u32 blanke = nv_rd32(priv, 0x64041c + (head * 0x300));
+	const u32 blanks = nv_rd32(priv, 0x640420 + (head * 0x300));
+	union {
+		struct nv04_disp_scanoutpos_v0 v0;
+	} *args = data;
 	int ret;
 
-	ret = nouveau_parent_create(parent, engine, oclass, 0,
-				    priv->sclass, 0, &base);
-	*pobject = nv_object(base);
-	if (ret)
+	nv_ioctl(object, "disp scanoutpos size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp scanoutpos vers %d\n", args->v0.version);
+		args->v0.vblanke = (blanke & 0xffff0000) >> 16;
+		args->v0.hblanke = (blanke & 0x0000ffff);
+		args->v0.vblanks = (blanks & 0xffff0000) >> 16;
+		args->v0.hblanks = (blanks & 0x0000ffff);
+		args->v0.vtotal  = ( total & 0xffff0000) >> 16;
+		args->v0.htotal  = ( total & 0x0000ffff);
+		args->v0.time[0] = ktime_to_ns(ktime_get());
+		args->v0.vline = /* vline read locks hline */
+			nv_rd32(priv, 0x616340 + (head * 0x800)) & 0xffff;
+		args->v0.time[1] = ktime_to_ns(ktime_get());
+		args->v0.hline =
+			nv_rd32(priv, 0x616344 + (head * 0x800)) & 0xffff;
+	} else
 		return ret;
 
-	priv->base.vblank->priv = priv;
-	priv->base.vblank->enable = nvd0_disp_base_vblank_enable;
-	priv->base.vblank->disable = nvd0_disp_base_vblank_disable;
-
-	return nouveau_ramht_new(nv_object(base), nv_object(base), 0x1000, 0,
-				&base->ramht);
-}
-
-static void
-nvd0_disp_base_dtor(struct nouveau_object *object)
-{
-	struct nv50_disp_base *base = (void *)object;
-	nouveau_ramht_ref(NULL, &base->ramht);
-	nouveau_parent_destroy(&base->base);
+	return 0;
 }
 
 static int
@@ -874,41 +738,27 @@ nvd0_disp_base_fini(struct nouveau_object *object, bool suspend)
 
 struct nouveau_ofuncs
 nvd0_disp_base_ofuncs = {
-	.ctor = nvd0_disp_base_ctor,
-	.dtor = nvd0_disp_base_dtor,
+	.ctor = nv50_disp_base_ctor,
+	.dtor = nv50_disp_base_dtor,
 	.init = nvd0_disp_base_init,
 	.fini = nvd0_disp_base_fini,
-};
-
-struct nouveau_omthds
-nvd0_disp_base_omthds[] = {
-	{ HEAD_MTHD(NV50_DISP_SCANOUTPOS)     , nvd0_disp_base_scanoutpos },
-	{ SOR_MTHD(NV50_DISP_SOR_PWR)         , nv50_sor_mthd },
-	{ SOR_MTHD(NVA3_DISP_SOR_HDA_ELD)     , nv50_sor_mthd },
-	{ SOR_MTHD(NV84_DISP_SOR_HDMI_PWR)    , nv50_sor_mthd },
-	{ SOR_MTHD(NV50_DISP_SOR_LVDS_SCRIPT) , nv50_sor_mthd },
-	{ SOR_MTHD(NV94_DISP_SOR_DP_PWR)      , nv50_sor_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_PWR)         , nv50_dac_mthd },
-	{ DAC_MTHD(NV50_DISP_DAC_LOAD)        , nv50_dac_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_PWR)       , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_TMDS_PWR)  , nv50_pior_mthd },
-	{ PIOR_MTHD(NV50_DISP_PIOR_DP_PWR)    , nv50_pior_mthd },
-	{},
+	.mthd = nv50_disp_base_mthd,
+	.ntfy = nouveau_disp_ntfy,
 };
 
 static struct nouveau_oclass
 nvd0_disp_base_oclass[] = {
-	{ NVD0_DISP_CLASS, &nvd0_disp_base_ofuncs, nvd0_disp_base_omthds },
+	{ GF110_DISP, &nvd0_disp_base_ofuncs },
 	{}
 };
 
 static struct nouveau_oclass
 nvd0_disp_sclass[] = {
-	{ NVD0_DISP_MAST_CLASS, &nvd0_disp_mast_ofuncs },
-	{ NVD0_DISP_SYNC_CLASS, &nvd0_disp_sync_ofuncs },
-	{ NVD0_DISP_OVLY_CLASS, &nvd0_disp_ovly_ofuncs },
-	{ NVD0_DISP_OIMM_CLASS, &nvd0_disp_oimm_ofuncs },
-	{ NVD0_DISP_CURS_CLASS, &nvd0_disp_curs_ofuncs },
+	{ GF110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
+	{ GF110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GF110_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
+	{ GF110_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
+	{ GF110_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
 	{}
 };
 
@@ -916,6 +766,27 @@ nvd0_disp_sclass[] = {
  * Display engine implementation
  ******************************************************************************/
 
+static void
+nvd0_disp_vblank_init(struct nvkm_event *event, int type, int head)
+{
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_mask(disp, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000001);
+}
+
+static void
+nvd0_disp_vblank_fini(struct nvkm_event *event, int type, int head)
+{
+	struct nouveau_disp *disp = container_of(event, typeof(*disp), vblank);
+	nv_mask(disp, 0x6100c0 + (head * 0x800), 0x00000001, 0x00000000);
+}
+
+const struct nvkm_event_func
+nvd0_disp_vblank_func = {
+	.ctor = nouveau_disp_vblank_ctor,
+	.init = nvd0_disp_vblank_init,
+	.fini = nvd0_disp_vblank_fini,
+};
+
 static struct nvkm_output *
 exec_lookup(struct nv50_disp_priv *priv, int head, int or, u32 ctrl,
 	    u32 *data, u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
@@ -1105,6 +976,9 @@ nvd0_disp_intr_unk2_2_tu(struct nv50_disp_priv *priv, int head,
 	const int or = ffs(outp->or) - 1;
 	const u32 ctrl = nv_rd32(priv, 0x660200 + (or   * 0x020));
 	const u32 conf = nv_rd32(priv, 0x660404 + (head * 0x300));
+	const s32 vactive = nv_rd32(priv, 0x660414 + (head * 0x300)) & 0xffff;
+	const s32 vblanke = nv_rd32(priv, 0x66041c + (head * 0x300)) & 0xffff;
+	const s32 vblanks = nv_rd32(priv, 0x660420 + (head * 0x300)) & 0xffff;
 	const u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
 	const u32 link = ((ctrl & 0xf00) == 0x800) ? 0 : 1;
 	const u32 hoff = (head * 0x800);
@@ -1112,23 +986,35 @@ nvd0_disp_intr_unk2_2_tu(struct nv50_disp_priv *priv, int head,
 	const u32 loff = (link * 0x080) + soff;
 	const u32 symbol = 100000;
 	const u32 TU = 64;
-	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff) & 0x000f0000;
+	u32 dpctrl = nv_rd32(priv, 0x61c10c + loff);
 	u32 clksor = nv_rd32(priv, 0x612300 + soff);
 	u32 datarate, link_nr, link_bw, bits;
 	u64 ratio, value;
 
+	link_nr  = hweight32(dpctrl & 0x000f0000);
+	link_bw  = (clksor & 0x007c0000) >> 18;
+	link_bw *= 27000;
+
+	/* symbols/hblank - algorithm taken from comments in tegra driver */
+	value = vblanke + vactive - vblanks - 7;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - (3 * !!(dpctrl & 0x00004000)) - (12 / link_nr);
+	nv_mask(priv, 0x616620 + hoff, 0x0000ffff, value);
+
+	/* symbols/vblank - algorithm taken from comments in tegra driver */
+	value = vblanks - vblanke - 25;
+	value = value * link_bw;
+	do_div(value, pclk);
+	value = value - ((36 / link_nr) + 3) - 1;
+	nv_mask(priv, 0x616624 + hoff, 0x00ffffff, value);
+
+	/* watermark */
 	if      ((conf & 0x3c0) == 0x180) bits = 30;
 	else if ((conf & 0x3c0) == 0x140) bits = 24;
 	else                              bits = 18;
 	datarate = (pclk * bits) / 8;
 
-	if      (dpctrl > 0x00030000) link_nr = 4;
-	else if (dpctrl > 0x00010000) link_nr = 2;
-	else			      link_nr = 1;
-
-	link_bw  = (clksor & 0x007c0000) >> 18;
-	link_bw *= 27000;
-
 	ratio  = datarate;
 	ratio *= symbol;
 	do_div(ratio, link_nr * link_bw);
@@ -1309,7 +1195,11 @@ nvd0_disp_intr(struct nouveau_subdev *subdev)
 
 	if (intr & 0x00000001) {
 		u32 stat = nv_rd32(priv, 0x61008c);
-		nv_wr32(priv, 0x61008c, stat);
+		while (stat) {
+			int chid = __ffs(stat); stat &= ~(1 << chid);
+			nv50_disp_chan_uevent_send(priv, chid);
+			nv_wr32(priv, 0x61008c, 1 << chid);
+		}
 		intr &= ~0x00000001;
 	}
 
@@ -1343,7 +1233,7 @@ nvd0_disp_intr(struct nouveau_subdev *subdev)
 		if (mask & intr) {
 			u32 stat = nv_rd32(priv, 0x6100bc + (i * 0x800));
 			if (stat & 0x00000001)
-				nouveau_event_trigger(priv->base.vblank, 1, i);
+				nouveau_disp_vblank(&priv->base, i);
 			nv_mask(priv, 0x6100bc + (i * 0x800), 0, 0);
 			nv_rd32(priv, 0x6100c0 + (i * 0x800));
 		}
@@ -1365,6 +1255,10 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nvd0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
@@ -1396,9 +1290,11 @@ nvd0_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
 	.mthd.core = &nvd0_disp_mast_mthd_chan,
 	.mthd.base = &nvd0_disp_sync_mthd_chan,
 	.mthd.ovly = &nvd0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
+	.head.scanoutpos = nvd0_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 11328e3f5df1..db144b2cf06b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -200,17 +200,17 @@ nve0_disp_ovly_mthd_chan = {
 
 static struct nouveau_oclass
 nve0_disp_sclass[] = {
-	{ NVE0_DISP_MAST_CLASS, &nvd0_disp_mast_ofuncs },
-	{ NVE0_DISP_SYNC_CLASS, &nvd0_disp_sync_ofuncs },
-	{ NVE0_DISP_OVLY_CLASS, &nvd0_disp_ovly_ofuncs },
-	{ NVE0_DISP_OIMM_CLASS, &nvd0_disp_oimm_ofuncs },
-	{ NVE0_DISP_CURS_CLASS, &nvd0_disp_curs_ofuncs },
+	{ GK104_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
+	{ GK104_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
+	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
+	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
 	{}
 };
 
 static struct nouveau_oclass
 nve0_disp_base_oclass[] = {
-	{ NVE0_DISP_CLASS, &nvd0_disp_base_ofuncs, nvd0_disp_base_omthds },
+	{ GK104_DISP, &nvd0_disp_base_ofuncs },
 	{}
 };
 
@@ -233,6 +233,10 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nve0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
@@ -245,7 +249,7 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -258,9 +262,11 @@ nve0_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
 	.mthd.core = &nve0_disp_mast_mthd_chan,
 	.mthd.base = &nvd0_disp_sync_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
+	.head.scanoutpos = nvd0_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
index 104388081d73..402d7d67d806 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
@@ -25,7 +25,7 @@
 #include <engine/software.h>
 #include <engine/disp.h>
 
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include "nv50.h"
 
@@ -35,17 +35,17 @@
 
 static struct nouveau_oclass
 nvf0_disp_sclass[] = {
-	{ NVF0_DISP_MAST_CLASS, &nvd0_disp_mast_ofuncs },
-	{ NVF0_DISP_SYNC_CLASS, &nvd0_disp_sync_ofuncs },
-	{ NVF0_DISP_OVLY_CLASS, &nvd0_disp_ovly_ofuncs },
-	{ NVF0_DISP_OIMM_CLASS, &nvd0_disp_oimm_ofuncs },
-	{ NVF0_DISP_CURS_CLASS, &nvd0_disp_curs_ofuncs },
+	{ GK110_DISP_CORE_CHANNEL_DMA, &nvd0_disp_mast_ofuncs.base },
+	{ GK110_DISP_BASE_CHANNEL_DMA, &nvd0_disp_sync_ofuncs.base },
+	{ GK104_DISP_OVERLAY_CONTROL_DMA, &nvd0_disp_ovly_ofuncs.base },
+	{ GK104_DISP_OVERLAY, &nvd0_disp_oimm_ofuncs.base },
+	{ GK104_DISP_CURSOR, &nvd0_disp_curs_ofuncs.base },
 	{}
 };
 
 static struct nouveau_oclass
 nvf0_disp_base_oclass[] = {
-	{ NVF0_DISP_CLASS, &nvd0_disp_base_ofuncs, nvd0_disp_base_omthds },
+	{ GK110_DISP, &nvd0_disp_base_ofuncs },
 	{}
 };
 
@@ -68,6 +68,10 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	ret = nvkm_event_init(&nvd0_disp_chan_uevent, 1, 17, &priv->uevent);
+	if (ret)
+		return ret;
+
 	nv_engine(priv)->sclass = nvf0_disp_base_oclass;
 	nv_engine(priv)->cclass = &nv50_disp_cclass;
 	nv_subdev(priv)->intr = nvd0_disp_intr;
@@ -80,7 +84,7 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->dac.sense = nv50_dac_sense;
 	priv->sor.power = nv50_sor_power;
 	priv->sor.hda_eld = nvd0_hda_eld;
-	priv->sor.hdmi = nvd0_hdmi_ctrl;
+	priv->sor.hdmi = nve0_hdmi_ctrl;
 	return 0;
 }
 
@@ -93,9 +97,11 @@ nvf0_disp_oclass = &(struct nv50_disp_impl) {
 		.init = _nouveau_disp_init,
 		.fini = _nouveau_disp_fini,
 	},
+	.base.vblank = &nvd0_disp_vblank_func,
 	.base.outp =  nvd0_disp_outp_sclass,
 	.mthd.core = &nve0_disp_mast_mthd_chan,
 	.mthd.base = &nvd0_disp_sync_mthd_chan,
 	.mthd.ovly = &nve0_disp_ovly_mthd_chan,
 	.mthd.prev = -0x020000,
+	.head.scanoutpos = nvd0_disp_base_scanoutpos,
 }.base.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outp.c b/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
index ad9ba7ccec7f..a5ff00a9cedc 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outp.c
@@ -78,6 +78,7 @@ nvkm_output_create_(struct nouveau_object *parent,
 
 	outp->info = *dcbE;
 	outp->index = index;
+	outp->or = ffs(outp->info.or) - 1;
 
 	DBG("type %02x loc %d or %d link %d con %x edid %x bus %d head %x\n",
 	    dcbE->type, dcbE->location, dcbE->or, dcbE->type >= 2 ?
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outp.h b/drivers/gpu/drm/nouveau/core/engine/disp/outp.h
index bc76fbf85710..187f435ad0e2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outp.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outp.h
@@ -9,6 +9,7 @@ struct nvkm_output {
 
 	struct dcb_output info;
 	int index;
+	int or;
 
 	struct nouveau_i2c_port *port;
 	struct nouveau_i2c_port *edid;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
index eb2d7789555d..667a9070e006 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.c
@@ -22,6 +22,9 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/os.h>
+#include <nvif/event.h>
+
 #include <subdev/i2c.h>
 
 #include "outpdp.h"
@@ -86,7 +89,7 @@ done:
 		atomic_set(&outp->lt.done, 0);
 		schedule_work(&outp->lt.work);
 	} else {
-		nouveau_event_get(outp->irq);
+		nvkm_notify_get(&outp->irq);
 	}
 
 	if (wait) {
@@ -133,46 +136,59 @@ nvkm_output_dp_detect(struct nvkm_output_dp *outp)
 	}
 }
 
-static void
-nvkm_output_dp_service_work(struct work_struct *work)
+static int
+nvkm_output_dp_hpd(struct nvkm_notify *notify)
 {
-	struct nvkm_output_dp *outp = container_of(work, typeof(*outp), work);
-	struct nouveau_disp *disp = nouveau_disp(outp);
-	int type = atomic_xchg(&outp->pending, 0);
-	u32 send = 0;
-
-	if (type & (NVKM_I2C_PLUG | NVKM_I2C_UNPLUG)) {
-		nvkm_output_dp_detect(outp);
-		if (type & NVKM_I2C_UNPLUG)
-			send |= NVKM_HPD_UNPLUG;
-		if (type & NVKM_I2C_PLUG)
-			send |= NVKM_HPD_PLUG;
-		nouveau_event_get(outp->base.conn->hpd.event);
-	}
-
-	if (type & NVKM_I2C_IRQ) {
-		nvkm_output_dp_train(&outp->base, 0, true);
-		send |= NVKM_HPD_IRQ;
+	struct nvkm_connector *conn = container_of(notify, typeof(*conn), hpd);
+	struct nvkm_output_dp *outp;
+	struct nouveau_disp *disp = nouveau_disp(conn);
+	const struct nvkm_i2c_ntfy_rep *line = notify->data;
+	struct nvif_notify_conn_rep_v0 rep = {};
+
+	list_for_each_entry(outp, &disp->outp, base.head) {
+		if (outp->base.conn == conn &&
+		    outp->info.type == DCB_OUTPUT_DP) {
+			DBG("HPD: %d\n", line->mask);
+			nvkm_output_dp_detect(outp);
+
+			if (line->mask & NVKM_I2C_UNPLUG)
+				rep.mask |= NVIF_NOTIFY_CONN_V0_UNPLUG;
+			if (line->mask & NVKM_I2C_PLUG)
+				rep.mask |= NVIF_NOTIFY_CONN_V0_PLUG;
+
+			nvkm_event_send(&disp->hpd, rep.mask, conn->index,
+					&rep, sizeof(rep));
+			return NVKM_NOTIFY_KEEP;
+		}
 	}
 
-	nouveau_event_trigger(disp->hpd, send, outp->base.info.connector);
+	WARN_ON(1);
+	return NVKM_NOTIFY_DROP;
 }
 
 static int
-nvkm_output_dp_service(void *data, u32 type, int index)
+nvkm_output_dp_irq(struct nvkm_notify *notify)
 {
-	struct nvkm_output_dp *outp = data;
-	DBG("HPD: %d\n", type);
-	atomic_or(type, &outp->pending);
-	schedule_work(&outp->work);
-	return NVKM_EVENT_DROP;
+	struct nvkm_output_dp *outp = container_of(notify, typeof(*outp), irq);
+	struct nouveau_disp *disp = nouveau_disp(outp);
+	const struct nvkm_i2c_ntfy_rep *line = notify->data;
+	struct nvif_notify_conn_rep_v0 rep = {
+		.mask = NVIF_NOTIFY_CONN_V0_IRQ,
+	};
+	int index = outp->base.info.connector;
+
+	DBG("IRQ: %d\n", line->mask);
+	nvkm_output_dp_train(&outp->base, 0, true);
+
+	nvkm_event_send(&disp->hpd, rep.mask, index, &rep, sizeof(rep));
+	return NVKM_NOTIFY_DROP;
 }
 
 int
 _nvkm_output_dp_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nvkm_output_dp *outp = (void *)object;
-	nouveau_event_put(outp->irq);
+	nvkm_notify_put(&outp->irq);
 	nvkm_output_dp_enable(outp, false);
 	return nvkm_output_fini(&outp->base, suspend);
 }
@@ -189,7 +205,7 @@ void
 _nvkm_output_dp_dtor(struct nouveau_object *object)
 {
 	struct nvkm_output_dp *outp = (void *)object;
-	nouveau_event_ref(NULL, &outp->irq);
+	nvkm_notify_fini(&outp->irq);
 	nvkm_output_destroy(&outp->base);
 }
 
@@ -213,7 +229,7 @@ nvkm_output_dp_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	nouveau_event_ref(NULL, &outp->base.conn->hpd.event);
+	nvkm_notify_fini(&outp->base.conn->hpd);
 
 	/* access to the aux channel is not optional... */
 	if (!outp->base.edid) {
@@ -238,20 +254,28 @@ nvkm_output_dp_create_(struct nouveau_object *parent,
 	atomic_set(&outp->lt.done, 0);
 
 	/* link maintenance */
-	ret = nouveau_event_new(i2c->ntfy, NVKM_I2C_IRQ, outp->base.edid->index,
-				nvkm_output_dp_service, outp, &outp->irq);
+	ret = nvkm_notify_init(NULL, &i2c->event, nvkm_output_dp_irq, true,
+			       &(struct nvkm_i2c_ntfy_req) {
+				.mask = NVKM_I2C_IRQ,
+				.port = outp->base.edid->index,
+			       },
+			       sizeof(struct nvkm_i2c_ntfy_req),
+			       sizeof(struct nvkm_i2c_ntfy_rep),
+			       &outp->irq);
 	if (ret) {
 		ERR("error monitoring aux irq event: %d\n", ret);
 		return ret;
 	}
 
-	INIT_WORK(&outp->work, nvkm_output_dp_service_work);
-
 	/* hotplug detect, replaces gpio-based mechanism with aux events */
-	ret = nouveau_event_new(i2c->ntfy, NVKM_I2C_PLUG | NVKM_I2C_UNPLUG,
-				outp->base.edid->index,
-				nvkm_output_dp_service, outp,
-			       &outp->base.conn->hpd.event);
+	ret = nvkm_notify_init(NULL, &i2c->event, nvkm_output_dp_hpd, true,
+			       &(struct nvkm_i2c_ntfy_req) {
+				.mask = NVKM_I2C_PLUG | NVKM_I2C_UNPLUG,
+				.port = outp->base.edid->index,
+			       },
+			       sizeof(struct nvkm_i2c_ntfy_req),
+			       sizeof(struct nvkm_i2c_ntfy_rep),
+			       &outp->base.conn->hpd);
 	if (ret) {
 		ERR("error monitoring aux hpd events: %d\n", ret);
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.h b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.h
index ff33ba12cb67..1fac367cc867 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/outpdp.h
@@ -12,10 +12,7 @@ struct nvkm_output_dp {
 	struct nvbios_dpout info;
 	u8 version;
 
-	struct nouveau_eventh *irq;
-	struct nouveau_eventh *hpd;
-	struct work_struct work;
-	atomic_t pending;
+	struct nvkm_notify irq;
 	bool present;
 	u8 dpcd[16];
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c
index fe0f256f11bf..d00f89a468a7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/piornv50.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
@@ -143,38 +144,29 @@ nv50_pior_dp_impl = {
  *****************************************************************************/
 
 int
-nv50_pior_power(struct nv50_disp_priv *priv, int or, u32 data)
+nv50_pior_power(NV50_DISP_MTHD_V1)
 {
-	const u32 stat = data & NV50_DISP_PIOR_PWR_STATE;
-	const u32 soff = (or * 0x800);
+	const u32 soff = outp->or * 0x800;
+	union {
+		struct nv50_disp_pior_pwr_v0 v0;
+	} *args = data;
+	u32 ctrl, type;
+	int ret;
+
+	nv_ioctl(object, "disp pior pwr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp pior pwr vers %d state %d type %x\n",
+			 args->v0.version, args->v0.state, args->v0.type);
+		if (args->v0.type > 0x0f)
+			return -EINVAL;
+		ctrl = !!args->v0.state;
+		type = args->v0.type;
+	} else
+		return ret;
+
 	nv_wait(priv, 0x61e004 + soff, 0x80000000, 0x00000000);
-	nv_mask(priv, 0x61e004 + soff, 0x80000101, 0x80000000 | stat);
+	nv_mask(priv, 0x61e004 + soff, 0x80000101, 0x80000000 | ctrl);
 	nv_wait(priv, 0x61e004 + soff, 0x80000000, 0x00000000);
+	priv->pior.type[outp->or] = type;
 	return 0;
 }
-
-int
-nv50_pior_mthd(struct nouveau_object *object, u32 mthd, void *args, u32 size)
-{
-	struct nv50_disp_priv *priv = (void *)object->engine;
-	const u8 type = (mthd & NV50_DISP_PIOR_MTHD_TYPE) >> 12;
-	const u8 or   = (mthd & NV50_DISP_PIOR_MTHD_OR);
-	u32 *data = args;
-	int ret;
-
-	if (size < sizeof(u32))
-		return -EINVAL;
-
-	mthd &= ~NV50_DISP_PIOR_MTHD_TYPE;
-	mthd &= ~NV50_DISP_PIOR_MTHD_OR;
-	switch (mthd) {
-	case NV50_DISP_PIOR_PWR:
-		ret = priv->pior.power(priv, or, data[0]);
-		priv->pior.type[or] = type;
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	return ret;
-}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/priv.h b/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
index 26e9a42569c7..6a0511d54ce6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/priv.h
@@ -11,6 +11,7 @@ struct nouveau_disp_impl {
 	struct nouveau_oclass base;
 	struct nouveau_oclass **outp;
 	struct nouveau_oclass **conn;
+	const struct nvkm_event_func *vblank;
 };
 
 #define nouveau_disp_create(p,e,c,h,i,x,d)                                     \
@@ -39,4 +40,9 @@ int  _nouveau_disp_fini(struct nouveau_object *, bool);
 extern struct nouveau_oclass *nvkm_output_oclass;
 extern struct nouveau_oclass *nvkm_connector_oclass;
 
+int  nouveau_disp_vblank_ctor(struct nouveau_object *, void *data, u32 size,
+			      struct nvkm_notify *);
+void nouveau_disp_vblank(struct nouveau_disp *, int head);
+int  nouveau_disp_ntfy(struct nouveau_object *, u32, struct nvkm_event **);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
index 7a1ebdfa9e1b..ddf1760c4400 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornv50.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
@@ -32,77 +33,26 @@
 #include "nv50.h"
 
 int
-nv50_sor_power(struct nv50_disp_priv *priv, int or, u32 data)
+nv50_sor_power(NV50_DISP_MTHD_V1)
 {
-	const u32 stat = data & NV50_DISP_SOR_PWR_STATE;
-	const u32 soff = (or * 0x800);
+	union {
+		struct nv50_disp_sor_pwr_v0 v0;
+	} *args = data;
+	const u32 soff = outp->or * 0x800;
+	u32 stat;
+	int ret;
+
+	nv_ioctl(object, "disp sor pwr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "disp sor pwr vers %d state %d\n",
+			 args->v0.version, args->v0.state);
+		stat = !!args->v0.state;
+	} else
+		return ret;
+
 	nv_wait(priv, 0x61c004 + soff, 0x80000000, 0x00000000);
 	nv_mask(priv, 0x61c004 + soff, 0x80000001, 0x80000000 | stat);
 	nv_wait(priv, 0x61c004 + soff, 0x80000000, 0x00000000);
 	nv_wait(priv, 0x61c030 + soff, 0x10000000, 0x00000000);
 	return 0;
 }
-
-int
-nv50_sor_mthd(struct nouveau_object *object, u32 mthd, void *args, u32 size)
-{
-	struct nv50_disp_priv *priv = (void *)object->engine;
-	const u8  type = (mthd & NV50_DISP_SOR_MTHD_TYPE) >> 12;
-	const u8  head = (mthd & NV50_DISP_SOR_MTHD_HEAD) >> 3;
-	const u8  link = (mthd & NV50_DISP_SOR_MTHD_LINK) >> 2;
-	const u8    or = (mthd & NV50_DISP_SOR_MTHD_OR);
-	const u16 mask = (0x0100 << head) | (0x0040 << link) | (0x0001 << or);
-	struct nvkm_output *outp = NULL, *temp;
-	u32 data;
-	int ret = -EINVAL;
-
-	if (size < sizeof(u32))
-		return -EINVAL;
-	data = *(u32 *)args;
-
-	list_for_each_entry(temp, &priv->base.outp, head) {
-		if ((temp->info.hasht & 0xff) == type &&
-		    (temp->info.hashm & mask) == mask) {
-			outp = temp;
-			break;
-		}
-	}
-
-	switch (mthd & ~0x3f) {
-	case NV50_DISP_SOR_PWR:
-		ret = priv->sor.power(priv, or, data);
-		break;
-	case NVA3_DISP_SOR_HDA_ELD:
-		ret = priv->sor.hda_eld(priv, or, args, size);
-		break;
-	case NV84_DISP_SOR_HDMI_PWR:
-		ret = priv->sor.hdmi(priv, head, or, data);
-		break;
-	case NV50_DISP_SOR_LVDS_SCRIPT:
-		priv->sor.lvdsconf = data & NV50_DISP_SOR_LVDS_SCRIPT_ID;
-		ret = 0;
-		break;
-	case NV94_DISP_SOR_DP_PWR:
-		if (outp) {
-			struct nvkm_output_dp *outpdp = (void *)outp;
-			switch (data) {
-			case NV94_DISP_SOR_DP_PWR_STATE_OFF:
-				nouveau_event_put(outpdp->irq);
-				((struct nvkm_output_dp_impl *)nv_oclass(outp))
-					->lnk_pwr(outpdp, 0);
-				atomic_set(&outpdp->lt.done, 0);
-				break;
-			case NV94_DISP_SOR_DP_PWR_STATE_ON:
-				nvkm_output_dp_train(&outpdp->base, 0, true);
-				break;
-			default:
-				return -EINVAL;
-			}
-		}
-		break;
-	default:
-		BUG_ON(1);
-	}
-
-	return ret;
-}
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
index 05487cda84a8..39f85d627336 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornv94.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
index 97f0e9cd3d40..7b7bbc3e459e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/sornvd0.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/base.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/base.c
index 5103e88d1877..e1500f77a56a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/base.c
@@ -23,98 +23,143 @@
  */
 
 #include <core/object.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/fb.h>
-#include <engine/dmaobj.h>
+#include <subdev/instmem.h>
+
+#include "priv.h"
 
 static int
-nouveau_dmaobj_ctor(struct nouveau_object *parent,
+nvkm_dmaobj_bind(struct nouveau_dmaobj *dmaobj, struct nouveau_object *parent,
+		 struct nouveau_gpuobj **pgpuobj)
+{
+	const struct nvkm_dmaeng_impl *impl = (void *)
+		nv_oclass(nv_object(dmaobj)->engine);
+	int ret = 0;
+
+	if (nv_object(dmaobj) == parent) { /* ctor bind */
+		if (nv_mclass(parent->parent) == NV_DEVICE) {
+			/* delayed, or no, binding */
+			return 0;
+		}
+		ret = impl->bind(dmaobj, parent, pgpuobj);
+		if (ret == 0)
+			nouveau_object_ref(NULL, &parent);
+		return ret;
+	}
+
+	return impl->bind(dmaobj, parent, pgpuobj);
+}
+
+int
+nvkm_dmaobj_create_(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *data, u32 size,
-		    struct nouveau_object **pobject)
+		    struct nouveau_oclass *oclass, void **pdata, u32 *psize,
+		    int length, void **pobject)
 {
-	struct nouveau_dmaeng *dmaeng = (void *)engine;
+	union {
+		struct nv_dma_v0 v0;
+	} *args = *pdata;
+	struct nouveau_instmem *instmem = nouveau_instmem(parent);
+	struct nouveau_client *client = nouveau_client(parent);
+	struct nouveau_device *device = nv_device(parent);
+	struct nouveau_fb *pfb = nouveau_fb(parent);
 	struct nouveau_dmaobj *dmaobj;
-	struct nouveau_gpuobj *gpuobj;
-	struct nv_dma_class *args = data;
+	void *data = *pdata;
+	u32 size = *psize;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
-
-	ret = nouveau_object_create(parent, engine, oclass, 0, &dmaobj);
-	*pobject = nv_object(dmaobj);
+	ret = nouveau_object_create_(parent, engine, oclass, 0, length, pobject);
+	dmaobj = *pobject;
 	if (ret)
 		return ret;
 
-	switch (args->flags & NV_DMA_TARGET_MASK) {
-	case NV_DMA_TARGET_VM:
+	nv_ioctl(parent, "create dma size %d\n", *psize);
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		nv_ioctl(parent, "create dma vers %d target %d access %d "
+				 "start %016llx limit %016llx\n",
+			 args->v0.version, args->v0.target, args->v0.access,
+			 args->v0.start, args->v0.limit);
+		dmaobj->target = args->v0.target;
+		dmaobj->access = args->v0.access;
+		dmaobj->start  = args->v0.start;
+		dmaobj->limit  = args->v0.limit;
+	} else
+		return ret;
+
+	*pdata = data;
+	*psize = size;
+
+	if (dmaobj->start > dmaobj->limit)
+		return -EINVAL;
+
+	switch (dmaobj->target) {
+	case NV_DMA_V0_TARGET_VM:
 		dmaobj->target = NV_MEM_TARGET_VM;
 		break;
-	case NV_DMA_TARGET_VRAM:
+	case NV_DMA_V0_TARGET_VRAM:
+		if (!client->super) {
+			if (dmaobj->limit >= pfb->ram->size - instmem->reserved)
+				return -EACCES;
+			if (device->card_type >= NV_50)
+				return -EACCES;
+		}
 		dmaobj->target = NV_MEM_TARGET_VRAM;
 		break;
-	case NV_DMA_TARGET_PCI:
+	case NV_DMA_V0_TARGET_PCI:
+		if (!client->super)
+			return -EACCES;
 		dmaobj->target = NV_MEM_TARGET_PCI;
 		break;
-	case NV_DMA_TARGET_PCI_US:
-	case NV_DMA_TARGET_AGP:
+	case NV_DMA_V0_TARGET_PCI_US:
+	case NV_DMA_V0_TARGET_AGP:
+		if (!client->super)
+			return -EACCES;
 		dmaobj->target = NV_MEM_TARGET_PCI_NOSNOOP;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	switch (args->flags & NV_DMA_ACCESS_MASK) {
-	case NV_DMA_ACCESS_VM:
+	switch (dmaobj->access) {
+	case NV_DMA_V0_ACCESS_VM:
 		dmaobj->access = NV_MEM_ACCESS_VM;
 		break;
-	case NV_DMA_ACCESS_RD:
+	case NV_DMA_V0_ACCESS_RD:
 		dmaobj->access = NV_MEM_ACCESS_RO;
 		break;
-	case NV_DMA_ACCESS_WR:
+	case NV_DMA_V0_ACCESS_WR:
 		dmaobj->access = NV_MEM_ACCESS_WO;
 		break;
-	case NV_DMA_ACCESS_RDWR:
+	case NV_DMA_V0_ACCESS_RDWR:
 		dmaobj->access = NV_MEM_ACCESS_RW;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	dmaobj->start = args->start;
-	dmaobj->limit = args->limit;
-	dmaobj->conf0 = args->conf0;
-
-	switch (nv_mclass(parent)) {
-	case NV_DEVICE_CLASS:
-		/* delayed, or no, binding */
-		break;
-	default:
-		ret = dmaeng->bind(dmaeng, *pobject, dmaobj, &gpuobj);
-		if (ret == 0) {
-			nouveau_object_ref(NULL, pobject);
-			*pobject = nv_object(gpuobj);
-		}
-		break;
-	}
-
 	return ret;
 }
 
-static struct nouveau_ofuncs
-nouveau_dmaobj_ofuncs = {
-	.ctor = nouveau_dmaobj_ctor,
-	.dtor = nouveau_object_destroy,
-	.init = nouveau_object_init,
-	.fini = nouveau_object_fini,
-};
-
-struct nouveau_oclass
-nouveau_dmaobj_sclass[] = {
-	{ NV_DMA_FROM_MEMORY_CLASS, &nouveau_dmaobj_ofuncs },
-	{ NV_DMA_TO_MEMORY_CLASS, &nouveau_dmaobj_ofuncs },
-	{ NV_DMA_IN_MEMORY_CLASS, &nouveau_dmaobj_ofuncs },
-	{}
-};
+int
+_nvkm_dmaeng_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	const struct nvkm_dmaeng_impl *impl = (void *)oclass;
+	struct nouveau_dmaeng *dmaeng;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true, "DMAOBJ",
+				    "dmaobj", &dmaeng);
+	*pobject = nv_object(dmaeng);
+	if (ret)
+		return ret;
+
+	nv_engine(dmaeng)->sclass = impl->sclass;
+	dmaeng->bind = nvkm_dmaobj_bind;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv04.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv04.c
index 027d8217c0fa..20c9dbfe3b2e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv04.c
@@ -23,121 +23,143 @@
  */
 
 #include <core/gpuobj.h>
-#include <core/class.h>
+#include <nvif/class.h>
 
 #include <subdev/fb.h>
 #include <subdev/vm/nv04.h>
 
-#include <engine/dmaobj.h>
+#include "priv.h"
 
-struct nv04_dmaeng_priv {
-	struct nouveau_dmaeng base;
+struct nv04_dmaobj_priv {
+	struct nouveau_dmaobj base;
+	bool clone;
+	u32 flags0;
+	u32 flags2;
 };
 
 static int
-nv04_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
+nv04_dmaobj_bind(struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_object *parent,
-		 struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_gpuobj **pgpuobj)
 {
-	struct nv04_vmmgr_priv *vmm = nv04_vmmgr(dmaeng);
+	struct nv04_dmaobj_priv *priv = (void *)dmaobj;
 	struct nouveau_gpuobj *gpuobj;
-	u32 flags0 = nv_mclass(dmaobj);
-	u32 flags2 = 0x00000000;
-	u64 offset = dmaobj->start & 0xfffff000;
-	u64 adjust = dmaobj->start & 0x00000fff;
-	u32 length = dmaobj->limit - dmaobj->start;
+	u64 offset = priv->base.start & 0xfffff000;
+	u64 adjust = priv->base.start & 0x00000fff;
+	u32 length = priv->base.limit - priv->base.start;
 	int ret;
 
 	if (!nv_iclass(parent, NV_ENGCTX_CLASS)) {
 		switch (nv_mclass(parent->parent)) {
-		case NV03_CHANNEL_DMA_CLASS:
-		case NV10_CHANNEL_DMA_CLASS:
-		case NV17_CHANNEL_DMA_CLASS:
-		case NV40_CHANNEL_DMA_CLASS:
+		case NV03_CHANNEL_DMA:
+		case NV10_CHANNEL_DMA:
+		case NV17_CHANNEL_DMA:
+		case NV40_CHANNEL_DMA:
 			break;
 		default:
 			return -EINVAL;
 		}
 	}
 
-	if (dmaobj->target == NV_MEM_TARGET_VM) {
-		if (nv_object(vmm)->oclass == &nv04_vmmgr_oclass) {
-			struct nouveau_gpuobj *pgt = vmm->vm->pgt[0].obj[0];
-			if (!dmaobj->start)
-				return nouveau_gpuobj_dup(parent, pgt, pgpuobj);
-			offset  = nv_ro32(pgt, 8 + (offset >> 10));
-			offset &= 0xfffff000;
-		}
+	if (priv->clone) {
+		struct nv04_vmmgr_priv *vmm = nv04_vmmgr(dmaobj);
+		struct nouveau_gpuobj *pgt = vmm->vm->pgt[0].obj[0];
+		if (!dmaobj->start)
+			return nouveau_gpuobj_dup(parent, pgt, pgpuobj);
+		offset  = nv_ro32(pgt, 8 + (offset >> 10));
+		offset &= 0xfffff000;
+	}
+
+	ret = nouveau_gpuobj_new(parent, parent, 16, 16, 0, &gpuobj);
+	*pgpuobj = gpuobj;
+	if (ret == 0) {
+		nv_wo32(*pgpuobj, 0x00, priv->flags0 | (adjust << 20));
+		nv_wo32(*pgpuobj, 0x04, length);
+		nv_wo32(*pgpuobj, 0x08, priv->flags2 | offset);
+		nv_wo32(*pgpuobj, 0x0c, priv->flags2 | offset);
+	}
+
+	return ret;
+}
+
+static int
+nv04_dmaobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		 struct nouveau_oclass *oclass, void *data, u32 size,
+		 struct nouveau_object **pobject)
+{
+	struct nouveau_dmaeng *dmaeng = (void *)engine;
+	struct nv04_vmmgr_priv *vmm = nv04_vmmgr(engine);
+	struct nv04_dmaobj_priv *priv;
+	int ret;
+
+	ret = nvkm_dmaobj_create(parent, engine, oclass, &data, &size, &priv);
+	*pobject = nv_object(priv);
+	if (ret || (ret = -ENOSYS, size))
+		return ret;
 
-		dmaobj->target = NV_MEM_TARGET_PCI;
-		dmaobj->access = NV_MEM_ACCESS_RW;
+	if (priv->base.target == NV_MEM_TARGET_VM) {
+		if (nv_object(vmm)->oclass == &nv04_vmmgr_oclass)
+			priv->clone = true;
+		priv->base.target = NV_MEM_TARGET_PCI;
+		priv->base.access = NV_MEM_ACCESS_RW;
 	}
 
-	switch (dmaobj->target) {
+	priv->flags0 = nv_mclass(priv);
+	switch (priv->base.target) {
 	case NV_MEM_TARGET_VRAM:
-		flags0 |= 0x00003000;
+		priv->flags0 |= 0x00003000;
 		break;
 	case NV_MEM_TARGET_PCI:
-		flags0 |= 0x00023000;
+		priv->flags0 |= 0x00023000;
 		break;
 	case NV_MEM_TARGET_PCI_NOSNOOP:
-		flags0 |= 0x00033000;
+		priv->flags0 |= 0x00033000;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	switch (dmaobj->access) {
+	switch (priv->base.access) {
 	case NV_MEM_ACCESS_RO:
-		flags0 |= 0x00004000;
+		priv->flags0 |= 0x00004000;
 		break;
 	case NV_MEM_ACCESS_WO:
-		flags0 |= 0x00008000;
+		priv->flags0 |= 0x00008000;
 	case NV_MEM_ACCESS_RW:
-		flags2 |= 0x00000002;
+		priv->flags2 |= 0x00000002;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	ret = nouveau_gpuobj_new(parent, parent, 16, 16, 0, &gpuobj);
-	*pgpuobj = gpuobj;
-	if (ret == 0) {
-		nv_wo32(*pgpuobj, 0x00, flags0 | (adjust << 20));
-		nv_wo32(*pgpuobj, 0x04, length);
-		nv_wo32(*pgpuobj, 0x08, flags2 | offset);
-		nv_wo32(*pgpuobj, 0x0c, flags2 | offset);
-	}
-
-	return ret;
+	return dmaeng->bind(&priv->base, nv_object(priv), (void *)pobject);
 }
 
-static int
-nv04_dmaeng_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		 struct nouveau_oclass *oclass, void *data, u32 size,
-		 struct nouveau_object **pobject)
-{
-	struct nv04_dmaeng_priv *priv;
-	int ret;
-
-	ret = nouveau_dmaeng_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
+static struct nouveau_ofuncs
+nv04_dmaobj_ofuncs = {
+	.ctor =  nv04_dmaobj_ctor,
+	.dtor = _nvkm_dmaobj_dtor,
+	.init = _nvkm_dmaobj_init,
+	.fini = _nvkm_dmaobj_fini,
+};
 
-	nv_engine(priv)->sclass = nouveau_dmaobj_sclass;
-	priv->base.bind = nv04_dmaobj_bind;
-	return 0;
-}
+static struct nouveau_oclass
+nv04_dmaeng_sclass[] = {
+	{ NV_DMA_FROM_MEMORY, &nv04_dmaobj_ofuncs },
+	{ NV_DMA_TO_MEMORY, &nv04_dmaobj_ofuncs },
+	{ NV_DMA_IN_MEMORY, &nv04_dmaobj_ofuncs },
+	{}
+};
 
-struct nouveau_oclass
-nv04_dmaeng_oclass = {
-	.handle = NV_ENGINE(DMAOBJ, 0x04),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nv04_dmaeng_ctor,
-		.dtor = _nouveau_dmaeng_dtor,
-		.init = _nouveau_dmaeng_init,
-		.fini = _nouveau_dmaeng_fini,
+struct nouveau_oclass *
+nv04_dmaeng_oclass = &(struct nvkm_dmaeng_impl) {
+	.base.handle = NV_ENGINE(DMAOBJ, 0x04),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nvkm_dmaeng_ctor,
+		.dtor = _nvkm_dmaeng_dtor,
+		.init = _nvkm_dmaeng_init,
+		.fini = _nvkm_dmaeng_fini,
 	},
-};
+	.sclass = nv04_dmaeng_sclass,
+	.bind = nv04_dmaobj_bind,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv50.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv50.c
index 750183f7c057..a740ddba2ee2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nv50.c
@@ -22,140 +22,176 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/gpuobj.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/fb.h>
-#include <engine/dmaobj.h>
 
-struct nv50_dmaeng_priv {
-	struct nouveau_dmaeng base;
+#include "priv.h"
+
+struct nv50_dmaobj_priv {
+	struct nouveau_dmaobj base;
+	u32 flags0;
+	u32 flags5;
 };
 
 static int
-nv50_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
+nv50_dmaobj_bind(struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_object *parent,
-		 struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_gpuobj **pgpuobj)
 {
-	u32 flags0 = nv_mclass(dmaobj);
-	u32 flags5 = 0x00000000;
+	struct nv50_dmaobj_priv *priv = (void *)dmaobj;
 	int ret;
 
 	if (!nv_iclass(parent, NV_ENGCTX_CLASS)) {
 		switch (nv_mclass(parent->parent)) {
-		case NV50_CHANNEL_DMA_CLASS:
-		case NV84_CHANNEL_DMA_CLASS:
-		case NV50_CHANNEL_IND_CLASS:
-		case NV84_CHANNEL_IND_CLASS:
-		case NV50_DISP_MAST_CLASS:
-		case NV84_DISP_MAST_CLASS:
-		case NV94_DISP_MAST_CLASS:
-		case NVA0_DISP_MAST_CLASS:
-		case NVA3_DISP_MAST_CLASS:
-		case NV50_DISP_SYNC_CLASS:
-		case NV84_DISP_SYNC_CLASS:
-		case NV94_DISP_SYNC_CLASS:
-		case NVA0_DISP_SYNC_CLASS:
-		case NVA3_DISP_SYNC_CLASS:
-		case NV50_DISP_OVLY_CLASS:
-		case NV84_DISP_OVLY_CLASS:
-		case NV94_DISP_OVLY_CLASS:
-		case NVA0_DISP_OVLY_CLASS:
-		case NVA3_DISP_OVLY_CLASS:
+		case NV40_CHANNEL_DMA:
+		case NV50_CHANNEL_GPFIFO:
+		case G82_CHANNEL_GPFIFO:
+		case NV50_DISP_CORE_CHANNEL_DMA:
+		case G82_DISP_CORE_CHANNEL_DMA:
+		case GT206_DISP_CORE_CHANNEL_DMA:
+		case GT200_DISP_CORE_CHANNEL_DMA:
+		case GT214_DISP_CORE_CHANNEL_DMA:
+		case NV50_DISP_BASE_CHANNEL_DMA:
+		case G82_DISP_BASE_CHANNEL_DMA:
+		case GT200_DISP_BASE_CHANNEL_DMA:
+		case GT214_DISP_BASE_CHANNEL_DMA:
+		case NV50_DISP_OVERLAY_CHANNEL_DMA:
+		case G82_DISP_OVERLAY_CHANNEL_DMA:
+		case GT200_DISP_OVERLAY_CHANNEL_DMA:
+		case GT214_DISP_OVERLAY_CHANNEL_DMA:
 			break;
 		default:
 			return -EINVAL;
 		}
 	}
 
-	if (!(dmaobj->conf0 & NV50_DMA_CONF0_ENABLE)) {
-		if (dmaobj->target == NV_MEM_TARGET_VM) {
-			dmaobj->conf0  = NV50_DMA_CONF0_PRIV_VM;
-			dmaobj->conf0 |= NV50_DMA_CONF0_PART_VM;
-			dmaobj->conf0 |= NV50_DMA_CONF0_COMP_VM;
-			dmaobj->conf0 |= NV50_DMA_CONF0_TYPE_VM;
+	ret = nouveau_gpuobj_new(parent, parent, 24, 32, 0, pgpuobj);
+	if (ret == 0) {
+		nv_wo32(*pgpuobj, 0x00, priv->flags0 | nv_mclass(dmaobj));
+		nv_wo32(*pgpuobj, 0x04, lower_32_bits(priv->base.limit));
+		nv_wo32(*pgpuobj, 0x08, lower_32_bits(priv->base.start));
+		nv_wo32(*pgpuobj, 0x0c, upper_32_bits(priv->base.limit) << 24 |
+					upper_32_bits(priv->base.start));
+		nv_wo32(*pgpuobj, 0x10, 0x00000000);
+		nv_wo32(*pgpuobj, 0x14, priv->flags5);
+	}
+
+	return ret;
+}
+
+static int
+nv50_dmaobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		 struct nouveau_oclass *oclass, void *data, u32 size,
+		 struct nouveau_object **pobject)
+{
+	struct nouveau_dmaeng *dmaeng = (void *)engine;
+	union {
+		struct nv50_dma_v0 v0;
+	} *args;
+	struct nv50_dmaobj_priv *priv;
+	u32 user, part, comp, kind;
+	int ret;
+
+	ret = nvkm_dmaobj_create(parent, engine, oclass, &data, &size, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+	args = data;
+
+	nv_ioctl(parent, "create nv50 dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create nv50 dma vers %d priv %d part %d "
+				 "comp %d kind %02x\n", args->v0.version,
+			 args->v0.priv, args->v0.part, args->v0.comp,
+			 args->v0.kind);
+		user = args->v0.priv;
+		part = args->v0.part;
+		comp = args->v0.comp;
+		kind = args->v0.kind;
+	} else
+	if (size == 0) {
+		if (priv->base.target != NV_MEM_TARGET_VM) {
+			user = NV50_DMA_V0_PRIV_US;
+			part = NV50_DMA_V0_PART_256;
+			comp = NV50_DMA_V0_COMP_NONE;
+			kind = NV50_DMA_V0_KIND_PITCH;
 		} else {
-			dmaobj->conf0  = NV50_DMA_CONF0_PRIV_US;
-			dmaobj->conf0 |= NV50_DMA_CONF0_PART_256;
-			dmaobj->conf0 |= NV50_DMA_CONF0_COMP_NONE;
-			dmaobj->conf0 |= NV50_DMA_CONF0_TYPE_LINEAR;
+			user = NV50_DMA_V0_PRIV_VM;
+			part = NV50_DMA_V0_PART_VM;
+			comp = NV50_DMA_V0_COMP_VM;
+			kind = NV50_DMA_V0_KIND_VM;
 		}
-	}
+	} else
+		return ret;
 
-	flags0 |= (dmaobj->conf0 & NV50_DMA_CONF0_COMP) << 22;
-	flags0 |= (dmaobj->conf0 & NV50_DMA_CONF0_TYPE) << 22;
-	flags0 |= (dmaobj->conf0 & NV50_DMA_CONF0_PRIV);
-	flags5 |= (dmaobj->conf0 & NV50_DMA_CONF0_PART);
+	if (user > 2 || part > 2 || comp > 3 || kind > 0x7f)
+		return -EINVAL;
+	priv->flags0 = (comp << 29) | (kind << 22) | (user << 20);
+	priv->flags5 = (part << 16);
 
-	switch (dmaobj->target) {
+	switch (priv->base.target) {
 	case NV_MEM_TARGET_VM:
-		flags0 |= 0x00000000;
+		priv->flags0 |= 0x00000000;
 		break;
 	case NV_MEM_TARGET_VRAM:
-		flags0 |= 0x00010000;
+		priv->flags0 |= 0x00010000;
 		break;
 	case NV_MEM_TARGET_PCI:
-		flags0 |= 0x00020000;
+		priv->flags0 |= 0x00020000;
 		break;
 	case NV_MEM_TARGET_PCI_NOSNOOP:
-		flags0 |= 0x00030000;
+		priv->flags0 |= 0x00030000;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	switch (dmaobj->access) {
+	switch (priv->base.access) {
 	case NV_MEM_ACCESS_VM:
 		break;
 	case NV_MEM_ACCESS_RO:
-		flags0 |= 0x00040000;
+		priv->flags0 |= 0x00040000;
 		break;
 	case NV_MEM_ACCESS_WO:
 	case NV_MEM_ACCESS_RW:
-		flags0 |= 0x00080000;
+		priv->flags0 |= 0x00080000;
 		break;
+	default:
+		return -EINVAL;
 	}
 
-	ret = nouveau_gpuobj_new(parent, parent, 24, 32, 0, pgpuobj);
-	if (ret == 0) {
-		nv_wo32(*pgpuobj, 0x00, flags0);
-		nv_wo32(*pgpuobj, 0x04, lower_32_bits(dmaobj->limit));
-		nv_wo32(*pgpuobj, 0x08, lower_32_bits(dmaobj->start));
-		nv_wo32(*pgpuobj, 0x0c, upper_32_bits(dmaobj->limit) << 24 |
-					upper_32_bits(dmaobj->start));
-		nv_wo32(*pgpuobj, 0x10, 0x00000000);
-		nv_wo32(*pgpuobj, 0x14, flags5);
-	}
-
-	return ret;
+	return dmaeng->bind(&priv->base, nv_object(priv), (void *)pobject);
 }
 
-static int
-nv50_dmaeng_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		 struct nouveau_oclass *oclass, void *data, u32 size,
-		 struct nouveau_object **pobject)
-{
-	struct nv50_dmaeng_priv *priv;
-	int ret;
-
-	ret = nouveau_dmaeng_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
+static struct nouveau_ofuncs
+nv50_dmaobj_ofuncs = {
+	.ctor =  nv50_dmaobj_ctor,
+	.dtor = _nvkm_dmaobj_dtor,
+	.init = _nvkm_dmaobj_init,
+	.fini = _nvkm_dmaobj_fini,
+};
 
-	nv_engine(priv)->sclass = nouveau_dmaobj_sclass;
-	priv->base.bind = nv50_dmaobj_bind;
-	return 0;
-}
+static struct nouveau_oclass
+nv50_dmaeng_sclass[] = {
+	{ NV_DMA_FROM_MEMORY, &nv50_dmaobj_ofuncs },
+	{ NV_DMA_TO_MEMORY, &nv50_dmaobj_ofuncs },
+	{ NV_DMA_IN_MEMORY, &nv50_dmaobj_ofuncs },
+	{}
+};
 
-struct nouveau_oclass
-nv50_dmaeng_oclass = {
-	.handle = NV_ENGINE(DMAOBJ, 0x50),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nv50_dmaeng_ctor,
-		.dtor = _nouveau_dmaeng_dtor,
-		.init = _nouveau_dmaeng_init,
-		.fini = _nouveau_dmaeng_fini,
+struct nouveau_oclass *
+nv50_dmaeng_oclass = &(struct nvkm_dmaeng_impl) {
+	.base.handle = NV_ENGINE(DMAOBJ, 0x50),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nvkm_dmaeng_ctor,
+		.dtor = _nvkm_dmaeng_dtor,
+		.init = _nvkm_dmaeng_init,
+		.fini = _nvkm_dmaeng_fini,
 	},
-};
+	.sclass = nv50_dmaeng_sclass,
+	.bind = nv50_dmaobj_bind,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvc0.c
index cd3970d03b80..88ec33b20048 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvc0.c
@@ -22,32 +22,35 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/device.h>
 #include <core/gpuobj.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/fb.h>
-#include <engine/dmaobj.h>
 
-struct nvc0_dmaeng_priv {
-	struct nouveau_dmaeng base;
+#include "priv.h"
+
+struct nvc0_dmaobj_priv {
+	struct nouveau_dmaobj base;
+	u32 flags0;
+	u32 flags5;
 };
 
 static int
-nvc0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
+nvc0_dmaobj_bind(struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_object *parent,
-		 struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_gpuobj **pgpuobj)
 {
-	u32 flags0 = nv_mclass(dmaobj);
-	u32 flags5 = 0x00000000;
+	struct nvc0_dmaobj_priv *priv = (void *)dmaobj;
 	int ret;
 
 	if (!nv_iclass(parent, NV_ENGCTX_CLASS)) {
 		switch (nv_mclass(parent->parent)) {
-		case NVA3_DISP_MAST_CLASS:
-		case NVA3_DISP_SYNC_CLASS:
-		case NVA3_DISP_OVLY_CLASS:
+		case GT214_DISP_CORE_CHANNEL_DMA:
+		case GT214_DISP_BASE_CHANNEL_DMA:
+		case GT214_DISP_OVERLAY_CHANNEL_DMA:
 			break;
 		default:
 			return -EINVAL;
@@ -55,89 +58,122 @@ nvc0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
 	} else
 		return 0;
 
-	if (!(dmaobj->conf0 & NVC0_DMA_CONF0_ENABLE)) {
-		if (dmaobj->target == NV_MEM_TARGET_VM) {
-			dmaobj->conf0  = NVC0_DMA_CONF0_PRIV_VM;
-			dmaobj->conf0 |= NVC0_DMA_CONF0_TYPE_VM;
+	ret = nouveau_gpuobj_new(parent, parent, 24, 32, 0, pgpuobj);
+	if (ret == 0) {
+		nv_wo32(*pgpuobj, 0x00, priv->flags0 | nv_mclass(dmaobj));
+		nv_wo32(*pgpuobj, 0x04, lower_32_bits(priv->base.limit));
+		nv_wo32(*pgpuobj, 0x08, lower_32_bits(priv->base.start));
+		nv_wo32(*pgpuobj, 0x0c, upper_32_bits(priv->base.limit) << 24 |
+					upper_32_bits(priv->base.start));
+		nv_wo32(*pgpuobj, 0x10, 0x00000000);
+		nv_wo32(*pgpuobj, 0x14, priv->flags5);
+	}
+
+	return ret;
+}
+
+static int
+nvc0_dmaobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		 struct nouveau_oclass *oclass, void *data, u32 size,
+		 struct nouveau_object **pobject)
+{
+	struct nouveau_dmaeng *dmaeng = (void *)engine;
+	union {
+		struct gf100_dma_v0 v0;
+	} *args;
+	struct nvc0_dmaobj_priv *priv;
+	u32 kind, user, unkn;
+	int ret;
+
+	ret = nvkm_dmaobj_create(parent, engine, oclass, &data, &size, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+	args = data;
+
+	nv_ioctl(parent, "create gf100 dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create gf100 dma vers %d priv %d kind %02x\n",
+			 args->v0.version, args->v0.priv, args->v0.kind);
+		kind = args->v0.kind;
+		user = args->v0.priv;
+		unkn = 0;
+	} else
+	if (size == 0) {
+		if (priv->base.target != NV_MEM_TARGET_VM) {
+			kind = GF100_DMA_V0_KIND_PITCH;
+			user = GF100_DMA_V0_PRIV_US;
+			unkn = 2;
 		} else {
-			dmaobj->conf0  = NVC0_DMA_CONF0_PRIV_US;
-			dmaobj->conf0 |= NVC0_DMA_CONF0_TYPE_LINEAR;
-			dmaobj->conf0 |= 0x00020000;
+			kind = GF100_DMA_V0_KIND_VM;
+			user = GF100_DMA_V0_PRIV_VM;
+			unkn = 0;
 		}
-	}
+	} else
+		return ret;
 
-	flags0 |= (dmaobj->conf0 & NVC0_DMA_CONF0_TYPE) << 22;
-	flags0 |= (dmaobj->conf0 & NVC0_DMA_CONF0_PRIV);
-	flags5 |= (dmaobj->conf0 & NVC0_DMA_CONF0_UNKN);
+	if (user > 2)
+		return -EINVAL;
+	priv->flags0 |= (kind << 22) | (user << 20);
+	priv->flags5 |= (unkn << 16);
 
-	switch (dmaobj->target) {
+	switch (priv->base.target) {
 	case NV_MEM_TARGET_VM:
-		flags0 |= 0x00000000;
+		priv->flags0 |= 0x00000000;
 		break;
 	case NV_MEM_TARGET_VRAM:
-		flags0 |= 0x00010000;
+		priv->flags0 |= 0x00010000;
 		break;
 	case NV_MEM_TARGET_PCI:
-		flags0 |= 0x00020000;
+		priv->flags0 |= 0x00020000;
 		break;
 	case NV_MEM_TARGET_PCI_NOSNOOP:
-		flags0 |= 0x00030000;
+		priv->flags0 |= 0x00030000;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	switch (dmaobj->access) {
+	switch (priv->base.access) {
 	case NV_MEM_ACCESS_VM:
 		break;
 	case NV_MEM_ACCESS_RO:
-		flags0 |= 0x00040000;
+		priv->flags0 |= 0x00040000;
 		break;
 	case NV_MEM_ACCESS_WO:
 	case NV_MEM_ACCESS_RW:
-		flags0 |= 0x00080000;
+		priv->flags0 |= 0x00080000;
 		break;
 	}
 
-	ret = nouveau_gpuobj_new(parent, parent, 24, 32, 0, pgpuobj);
-	if (ret == 0) {
-		nv_wo32(*pgpuobj, 0x00, flags0);
-		nv_wo32(*pgpuobj, 0x04, lower_32_bits(dmaobj->limit));
-		nv_wo32(*pgpuobj, 0x08, lower_32_bits(dmaobj->start));
-		nv_wo32(*pgpuobj, 0x0c, upper_32_bits(dmaobj->limit) << 24 |
-					upper_32_bits(dmaobj->start));
-		nv_wo32(*pgpuobj, 0x10, 0x00000000);
-		nv_wo32(*pgpuobj, 0x14, flags5);
-	}
-
-	return ret;
+	return dmaeng->bind(&priv->base, nv_object(priv), (void *)pobject);
 }
 
-static int
-nvc0_dmaeng_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		 struct nouveau_oclass *oclass, void *data, u32 size,
-		 struct nouveau_object **pobject)
-{
-	struct nvc0_dmaeng_priv *priv;
-	int ret;
-
-	ret = nouveau_dmaeng_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
+static struct nouveau_ofuncs
+nvc0_dmaobj_ofuncs = {
+	.ctor =  nvc0_dmaobj_ctor,
+	.dtor = _nvkm_dmaobj_dtor,
+	.init = _nvkm_dmaobj_init,
+	.fini = _nvkm_dmaobj_fini,
+};
 
-	nv_engine(priv)->sclass = nouveau_dmaobj_sclass;
-	priv->base.bind = nvc0_dmaobj_bind;
-	return 0;
-}
+static struct nouveau_oclass
+nvc0_dmaeng_sclass[] = {
+	{ NV_DMA_FROM_MEMORY, &nvc0_dmaobj_ofuncs },
+	{ NV_DMA_TO_MEMORY, &nvc0_dmaobj_ofuncs },
+	{ NV_DMA_IN_MEMORY, &nvc0_dmaobj_ofuncs },
+	{}
+};
 
-struct nouveau_oclass
-nvc0_dmaeng_oclass = {
-	.handle = NV_ENGINE(DMAOBJ, 0xc0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_dmaeng_ctor,
-		.dtor = _nouveau_dmaeng_dtor,
-		.init = _nouveau_dmaeng_init,
-		.fini = _nouveau_dmaeng_fini,
+struct nouveau_oclass *
+nvc0_dmaeng_oclass = &(struct nvkm_dmaeng_impl) {
+	.base.handle = NV_ENGINE(DMAOBJ, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nvkm_dmaeng_ctor,
+		.dtor = _nvkm_dmaeng_dtor,
+		.init = _nvkm_dmaeng_init,
+		.fini = _nvkm_dmaeng_fini,
 	},
-};
+	.sclass = nvc0_dmaeng_sclass,
+	.bind = nvc0_dmaobj_bind,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
index 1cfb3bb90131..3fc4f0b0eaca 100644
--- a/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/nvd0.c
@@ -22,40 +22,40 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/device.h>
 #include <core/gpuobj.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/fb.h>
-#include <engine/dmaobj.h>
 
-struct nvd0_dmaeng_priv {
-	struct nouveau_dmaeng base;
+#include "priv.h"
+
+struct nvd0_dmaobj_priv {
+	struct nouveau_dmaobj base;
+	u32 flags0;
 };
 
 static int
-nvd0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
+nvd0_dmaobj_bind(struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_object *parent,
-		 struct nouveau_dmaobj *dmaobj,
 		 struct nouveau_gpuobj **pgpuobj)
 {
-	u32 flags0 = 0x00000000;
+	struct nvd0_dmaobj_priv *priv = (void *)dmaobj;
 	int ret;
 
 	if (!nv_iclass(parent, NV_ENGCTX_CLASS)) {
 		switch (nv_mclass(parent->parent)) {
-		case NVD0_DISP_MAST_CLASS:
-		case NVD0_DISP_SYNC_CLASS:
-		case NVD0_DISP_OVLY_CLASS:
-		case NVE0_DISP_MAST_CLASS:
-		case NVE0_DISP_SYNC_CLASS:
-		case NVE0_DISP_OVLY_CLASS:
-		case NVF0_DISP_MAST_CLASS:
-		case NVF0_DISP_SYNC_CLASS:
-		case NVF0_DISP_OVLY_CLASS:
-		case GM107_DISP_MAST_CLASS:
-		case GM107_DISP_SYNC_CLASS:
-		case GM107_DISP_OVLY_CLASS:
+		case GF110_DISP_CORE_CHANNEL_DMA:
+		case GK104_DISP_CORE_CHANNEL_DMA:
+		case GK110_DISP_CORE_CHANNEL_DMA:
+		case GM107_DISP_CORE_CHANNEL_DMA:
+		case GF110_DISP_BASE_CHANNEL_DMA:
+		case GK104_DISP_BASE_CHANNEL_DMA:
+		case GK110_DISP_BASE_CHANNEL_DMA:
+		case GF110_DISP_OVERLAY_CONTROL_DMA:
+		case GK104_DISP_OVERLAY_CONTROL_DMA:
 			break;
 		default:
 			return -EINVAL;
@@ -63,33 +63,11 @@ nvd0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
 	} else
 		return 0;
 
-	if (!(dmaobj->conf0 & NVD0_DMA_CONF0_ENABLE)) {
-		if (dmaobj->target == NV_MEM_TARGET_VM) {
-			dmaobj->conf0 |= NVD0_DMA_CONF0_TYPE_VM;
-			dmaobj->conf0 |= NVD0_DMA_CONF0_PAGE_LP;
-		} else {
-			dmaobj->conf0 |= NVD0_DMA_CONF0_TYPE_LINEAR;
-			dmaobj->conf0 |= NVD0_DMA_CONF0_PAGE_SP;
-		}
-	}
-
-	flags0 |= (dmaobj->conf0 & NVD0_DMA_CONF0_TYPE) << 20;
-	flags0 |= (dmaobj->conf0 & NVD0_DMA_CONF0_PAGE) >> 4;
-
-	switch (dmaobj->target) {
-	case NV_MEM_TARGET_VRAM:
-		flags0 |= 0x00000009;
-		break;
-	default:
-		return -EINVAL;
-		break;
-	}
-
 	ret = nouveau_gpuobj_new(parent, parent, 24, 32, 0, pgpuobj);
 	if (ret == 0) {
-		nv_wo32(*pgpuobj, 0x00, flags0);
-		nv_wo32(*pgpuobj, 0x04, dmaobj->start >> 8);
-		nv_wo32(*pgpuobj, 0x08, dmaobj->limit >> 8);
+		nv_wo32(*pgpuobj, 0x00, priv->flags0);
+		nv_wo32(*pgpuobj, 0x04, priv->base.start >> 8);
+		nv_wo32(*pgpuobj, 0x08, priv->base.limit >> 8);
 		nv_wo32(*pgpuobj, 0x0c, 0x00000000);
 		nv_wo32(*pgpuobj, 0x10, 0x00000000);
 		nv_wo32(*pgpuobj, 0x14, 0x00000000);
@@ -99,30 +77,91 @@ nvd0_dmaobj_bind(struct nouveau_dmaeng *dmaeng,
 }
 
 static int
-nvd0_dmaeng_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+nvd0_dmaobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		 struct nouveau_oclass *oclass, void *data, u32 size,
 		 struct nouveau_object **pobject)
 {
-	struct nvd0_dmaeng_priv *priv;
+	struct nouveau_dmaeng *dmaeng = (void *)engine;
+	union {
+		struct gf110_dma_v0 v0;
+	} *args;
+	struct nvd0_dmaobj_priv *priv;
+	u32 kind, page;
 	int ret;
 
-	ret = nouveau_dmaeng_create(parent, engine, oclass, &priv);
+	ret = nvkm_dmaobj_create(parent, engine, oclass, &data, &size, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
+	args = data;
 
-	nv_engine(priv)->sclass = nouveau_dmaobj_sclass;
-	priv->base.bind = nvd0_dmaobj_bind;
-	return 0;
+	nv_ioctl(parent, "create gf110 dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create gf100 dma vers %d page %d kind %02x\n",
+			 args->v0.version, args->v0.page, args->v0.kind);
+		kind = args->v0.kind;
+		page = args->v0.page;
+	} else
+	if (size == 0) {
+		if (priv->base.target != NV_MEM_TARGET_VM) {
+			kind = GF110_DMA_V0_KIND_PITCH;
+			page = GF110_DMA_V0_PAGE_SP;
+		} else {
+			kind = GF110_DMA_V0_KIND_VM;
+			page = GF110_DMA_V0_PAGE_LP;
+		}
+	} else
+		return ret;
+
+	if (page > 1)
+		return -EINVAL;
+	priv->flags0 = (kind << 20) | (page << 6);
+
+	switch (priv->base.target) {
+	case NV_MEM_TARGET_VRAM:
+		priv->flags0 |= 0x00000009;
+		break;
+	case NV_MEM_TARGET_VM:
+	case NV_MEM_TARGET_PCI:
+	case NV_MEM_TARGET_PCI_NOSNOOP:
+		/* XXX: don't currently know how to construct a real one
+		 *      of these.  we only use them to represent pushbufs
+		 *      on these chipsets, and the classes that use them
+		 *      deal with the target themselves.
+		 */
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return dmaeng->bind(&priv->base, nv_object(priv), (void *)pobject);
 }
 
-struct nouveau_oclass
-nvd0_dmaeng_oclass = {
-	.handle = NV_ENGINE(DMAOBJ, 0xd0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvd0_dmaeng_ctor,
-		.dtor = _nouveau_dmaeng_dtor,
-		.init = _nouveau_dmaeng_init,
-		.fini = _nouveau_dmaeng_fini,
-	},
+static struct nouveau_ofuncs
+nvd0_dmaobj_ofuncs = {
+	.ctor =  nvd0_dmaobj_ctor,
+	.dtor = _nvkm_dmaobj_dtor,
+	.init = _nvkm_dmaobj_init,
+	.fini = _nvkm_dmaobj_fini,
 };
+
+static struct nouveau_oclass
+nvd0_dmaeng_sclass[] = {
+	{ NV_DMA_FROM_MEMORY, &nvd0_dmaobj_ofuncs },
+	{ NV_DMA_TO_MEMORY, &nvd0_dmaobj_ofuncs },
+	{ NV_DMA_IN_MEMORY, &nvd0_dmaobj_ofuncs },
+	{}
+};
+
+struct nouveau_oclass *
+nvd0_dmaeng_oclass = &(struct nvkm_dmaeng_impl) {
+	.base.handle = NV_ENGINE(DMAOBJ, 0xd0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nvkm_dmaeng_ctor,
+		.dtor = _nvkm_dmaeng_dtor,
+		.init = _nvkm_dmaeng_init,
+		.fini = _nvkm_dmaeng_fini,
+	},
+	.sclass = nvd0_dmaeng_sclass,
+	.bind = nvd0_dmaobj_bind,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/dmaobj/priv.h b/drivers/gpu/drm/nouveau/core/engine/dmaobj/priv.h
new file mode 100644
index 000000000000..36f743866937
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/dmaobj/priv.h
@@ -0,0 +1,30 @@
+#ifndef __NVKM_DMAOBJ_PRIV_H__
+#define __NVKM_DMAOBJ_PRIV_H__
+
+#include <engine/dmaobj.h>
+
+#define nvkm_dmaobj_create(p,e,c,pa,sa,d)                                      \
+	nvkm_dmaobj_create_((p), (e), (c), (pa), (sa), sizeof(**d), (void **)d)
+
+int nvkm_dmaobj_create_(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, void **, u32 *,
+			int, void **);
+#define _nvkm_dmaobj_dtor nouveau_object_destroy
+#define _nvkm_dmaobj_init nouveau_object_init
+#define _nvkm_dmaobj_fini nouveau_object_fini
+
+int _nvkm_dmaeng_ctor(struct nouveau_object *, struct nouveau_object *,
+		      struct nouveau_oclass *, void *, u32,
+		      struct nouveau_object **);
+#define _nvkm_dmaeng_dtor _nouveau_engine_dtor
+#define _nvkm_dmaeng_init _nouveau_engine_init
+#define _nvkm_dmaeng_fini _nouveau_engine_fini
+
+struct nvkm_dmaeng_impl {
+	struct nouveau_oclass base;
+	struct nouveau_oclass *sclass;
+	int (*bind)(struct nouveau_dmaobj *, struct nouveau_object *,
+		    struct nouveau_gpuobj **);
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
index 56ed3d73bf8e..ac8375cf4eef 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/base.c
@@ -26,11 +26,31 @@
 #include <core/object.h>
 #include <core/handle.h>
 #include <core/event.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+#include <nvif/event.h>
 
 #include <engine/dmaobj.h>
 #include <engine/fifo.h>
 
+static int
+nouveau_fifo_event_ctor(struct nouveau_object *object, void *data, u32 size,
+			struct nvkm_notify *notify)
+{
+	if (size == 0) {
+		notify->size  = 0;
+		notify->types = 1;
+		notify->index = 0;
+		return 0;
+	}
+	return -ENOSYS;
+}
+
+static const struct nvkm_event_func
+nouveau_fifo_event_func = {
+	.ctor = nouveau_fifo_event_ctor,
+};
+
 int
 nouveau_fifo_channel_create_(struct nouveau_object *parent,
 			     struct nouveau_object *engine,
@@ -59,14 +79,14 @@ nouveau_fifo_channel_create_(struct nouveau_object *parent,
 
 	dmaeng = (void *)chan->pushdma->base.engine;
 	switch (chan->pushdma->base.oclass->handle) {
-	case NV_DMA_FROM_MEMORY_CLASS:
-	case NV_DMA_IN_MEMORY_CLASS:
+	case NV_DMA_FROM_MEMORY:
+	case NV_DMA_IN_MEMORY:
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	ret = dmaeng->bind(dmaeng, parent, chan->pushdma, &chan->pushgpu);
+	ret = dmaeng->bind(chan->pushdma, parent, &chan->pushgpu);
 	if (ret)
 		return ret;
 
@@ -85,15 +105,10 @@ nouveau_fifo_channel_create_(struct nouveau_object *parent,
 		return -ENOSPC;
 	}
 
-	/* map fifo control registers */
-	chan->user = ioremap(nv_device_resource_start(device, bar) + addr +
-			     (chan->chid * size), size);
-	if (!chan->user)
-		return -EFAULT;
-
-	nouveau_event_trigger(priv->cevent, 1, 0);
-
+	chan->addr = nv_device_resource_start(device, bar) +
+		     addr + size * chan->chid;
 	chan->size = size;
+	nvkm_event_send(&priv->cevent, 1, 0, NULL, 0);
 	return 0;
 }
 
@@ -103,7 +118,8 @@ nouveau_fifo_channel_destroy(struct nouveau_fifo_chan *chan)
 	struct nouveau_fifo *priv = (void *)nv_object(chan)->engine;
 	unsigned long flags;
 
-	iounmap(chan->user);
+	if (chan->user)
+		iounmap(chan->user);
 
 	spin_lock_irqsave(&priv->lock, flags);
 	priv->channel[chan->chid] = NULL;
@@ -121,10 +137,24 @@ _nouveau_fifo_channel_dtor(struct nouveau_object *object)
 	nouveau_fifo_channel_destroy(chan);
 }
 
+int
+_nouveau_fifo_channel_map(struct nouveau_object *object, u64 *addr, u32 *size)
+{
+	struct nouveau_fifo_chan *chan = (void *)object;
+	*addr = chan->addr;
+	*size = chan->size;
+	return 0;
+}
+
 u32
 _nouveau_fifo_channel_rd32(struct nouveau_object *object, u64 addr)
 {
 	struct nouveau_fifo_chan *chan = (void *)object;
+	if (unlikely(!chan->user)) {
+		chan->user = ioremap(chan->addr, chan->size);
+		if (WARN_ON_ONCE(chan->user == NULL))
+			return 0;
+	}
 	return ioread32_native(chan->user + addr);
 }
 
@@ -132,9 +162,58 @@ void
 _nouveau_fifo_channel_wr32(struct nouveau_object *object, u64 addr, u32 data)
 {
 	struct nouveau_fifo_chan *chan = (void *)object;
+	if (unlikely(!chan->user)) {
+		chan->user = ioremap(chan->addr, chan->size);
+		if (WARN_ON_ONCE(chan->user == NULL))
+			return;
+	}
 	iowrite32_native(data, chan->user + addr);
 }
 
+int
+nouveau_fifo_uevent_ctor(struct nouveau_object *object, void *data, u32 size,
+			 struct nvkm_notify *notify)
+{
+	union {
+		struct nvif_notify_uevent_req none;
+	} *req = data;
+	int ret;
+
+	if (nvif_unvers(req->none)) {
+		notify->size  = sizeof(struct nvif_notify_uevent_rep);
+		notify->types = 1;
+		notify->index = 0;
+	}
+
+	return ret;
+}
+
+void
+nouveau_fifo_uevent(struct nouveau_fifo *fifo)
+{
+	struct nvif_notify_uevent_rep rep = {
+	};
+	nvkm_event_send(&fifo->uevent, 1, 0, &rep, sizeof(rep));
+}
+
+int
+_nouveau_fifo_channel_ntfy(struct nouveau_object *object, u32 type,
+			   struct nvkm_event **event)
+{
+	struct nouveau_fifo *fifo = (void *)object->engine;
+	switch (type) {
+	case G82_CHANNEL_DMA_V0_NTFY_UEVENT:
+		if (nv_mclass(object) >= G82_CHANNEL_DMA) {
+			*event = &fifo->uevent;
+			return 0;
+		}
+		break;
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
 static int
 nouveau_fifo_chid(struct nouveau_fifo *priv, struct nouveau_object *object)
 {
@@ -168,8 +247,8 @@ void
 nouveau_fifo_destroy(struct nouveau_fifo *priv)
 {
 	kfree(priv->channel);
-	nouveau_event_destroy(&priv->uevent);
-	nouveau_event_destroy(&priv->cevent);
+	nvkm_event_fini(&priv->uevent);
+	nvkm_event_fini(&priv->cevent);
 	nouveau_engine_destroy(&priv->base);
 }
 
@@ -194,11 +273,7 @@ nouveau_fifo_create_(struct nouveau_object *parent,
 	if (!priv->channel)
 		return -ENOMEM;
 
-	ret = nouveau_event_create(1, 1, &priv->cevent);
-	if (ret)
-		return ret;
-
-	ret = nouveau_event_create(1, 1, &priv->uevent);
+	ret = nvkm_event_init(&nouveau_fifo_event_func, 1, 1, &priv->cevent);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
index c61b16a63884..5ae6a43893b5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv04.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/engctx.h>
 #include <core/namedb.h>
 #include <core/handle.h>
@@ -117,16 +118,23 @@ nv04_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nv04_fifo_priv *priv = (void *)engine;
 	struct nv04_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0x800000,
-					  0x10000, args->pushbuf,
+					  0x10000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR), &chan);
@@ -134,13 +142,15 @@ nv04_fifo_chan_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->object_attach = nv04_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv04_fifo_object_detach;
 	nv_parent(chan)->context_attach = nv04_fifo_context_attach;
 	chan->ramfc = chan->base.chid * 32;
 
-	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->offset);
-	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->v0.offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->v0.offset);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x08, chan->base.pushgpu->addr >> 4);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x10,
 			     NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
@@ -242,13 +252,15 @@ nv04_fifo_ofuncs = {
 	.dtor = nv04_fifo_chan_dtor,
 	.init = nv04_fifo_chan_init,
 	.fini = nv04_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv04_fifo_sclass[] = {
-	{ NV03_CHANNEL_DMA_CLASS, &nv04_fifo_ofuncs },
+	{ NV03_CHANNEL_DMA, &nv04_fifo_ofuncs },
 	{}
 };
 
@@ -539,7 +551,7 @@ nv04_fifo_intr(struct nouveau_subdev *subdev)
 			}
 
 			if (status & 0x40000000) {
-				nouveau_event_trigger(priv->base.uevent, 1, 0);
+				nouveau_fifo_uevent(&priv->base);
 				nv_wr32(priv, 0x002100, 0x40000000);
 				status &= ~0x40000000;
 			}
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv10.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv10.c
index 571a22aa1ae5..2a32add51c81 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv10.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/engctx.h>
 #include <core/ramht.h>
 
@@ -59,16 +60,23 @@ nv10_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nv04_fifo_priv *priv = (void *)engine;
 	struct nv04_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0x800000,
-					  0x10000, args->pushbuf,
+					  0x10000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR), &chan);
@@ -76,13 +84,15 @@ nv10_fifo_chan_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->object_attach = nv04_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv04_fifo_object_detach;
 	nv_parent(chan)->context_attach = nv04_fifo_context_attach;
 	chan->ramfc = chan->base.chid * 32;
 
-	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->offset);
-	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->v0.offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->v0.offset);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x0c, chan->base.pushgpu->addr >> 4);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x14,
 			     NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
@@ -100,13 +110,15 @@ nv10_fifo_ofuncs = {
 	.dtor = nv04_fifo_chan_dtor,
 	.init = nv04_fifo_chan_init,
 	.fini = nv04_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv10_fifo_sclass[] = {
-	{ NV10_CHANNEL_DMA_CLASS, &nv10_fifo_ofuncs },
+	{ NV10_CHANNEL_DMA, &nv10_fifo_ofuncs },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv17.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv17.c
index f25760209316..12d76c8adb23 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv17.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv17.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/engctx.h>
 #include <core/ramht.h>
 
@@ -64,16 +65,23 @@ nv17_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nv04_fifo_priv *priv = (void *)engine;
 	struct nv04_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0x800000,
-					  0x10000, args->pushbuf,
+					  0x10000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -83,13 +91,15 @@ nv17_fifo_chan_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->object_attach = nv04_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv04_fifo_object_detach;
 	nv_parent(chan)->context_attach = nv04_fifo_context_attach;
 	chan->ramfc = chan->base.chid * 64;
 
-	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->offset);
-	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->v0.offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->v0.offset);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x0c, chan->base.pushgpu->addr >> 4);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x14,
 			     NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
@@ -107,13 +117,15 @@ nv17_fifo_ofuncs = {
 	.dtor = nv04_fifo_chan_dtor,
 	.init = nv04_fifo_chan_init,
 	.fini = nv04_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv17_fifo_sclass[] = {
-	{ NV17_CHANNEL_DMA_CLASS, &nv17_fifo_ofuncs },
+	{ NV17_CHANNEL_DMA, &nv17_fifo_ofuncs },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
index 343487ed2238..9f49c3a24dc6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
@@ -22,8 +22,9 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/os.h>
-#include <core/class.h>
+#include <core/client.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/engctx.h>
 #include <core/ramht.h>
 
@@ -182,16 +183,23 @@ nv40_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nv04_fifo_priv *priv = (void *)engine;
 	struct nv04_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0xc00000,
-					  0x1000, args->pushbuf,
+					  0x1000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -200,14 +208,16 @@ nv40_fifo_chan_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->context_attach = nv40_fifo_context_attach;
 	nv_parent(chan)->context_detach = nv40_fifo_context_detach;
 	nv_parent(chan)->object_attach = nv40_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv04_fifo_object_detach;
 	chan->ramfc = chan->base.chid * 128;
 
-	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->offset);
-	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x00, args->v0.offset);
+	nv_wo32(priv->ramfc, chan->ramfc + 0x04, args->v0.offset);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x0c, chan->base.pushgpu->addr >> 4);
 	nv_wo32(priv->ramfc, chan->ramfc + 0x18, 0x30000000 |
 			     NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
@@ -226,13 +236,15 @@ nv40_fifo_ofuncs = {
 	.dtor = nv04_fifo_chan_dtor,
 	.init = nv04_fifo_chan_init,
 	.fini = nv04_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv40_fifo_sclass[] = {
-	{ NV40_CHANNEL_DMA_CLASS, &nv40_fifo_ofuncs },
+	{ NV40_CHANNEL_DMA, &nv40_fifo_ofuncs },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
index e6352bd5b4ff..5d1e86bc244c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv50.c
@@ -25,7 +25,8 @@
 #include <core/client.h>
 #include <core/engctx.h>
 #include <core/ramht.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/timer.h>
 #include <subdev/bar.h>
@@ -194,17 +195,24 @@ nv50_fifo_chan_ctor_dma(struct nouveau_object *parent,
 			struct nouveau_oclass *oclass, void *data, u32 size,
 			struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nv50_fifo_base *base = (void *)parent;
 	struct nv50_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0xc00000,
-					  0x2000, args->pushbuf,
+					  0x2000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -213,6 +221,8 @@ nv50_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->context_attach = nv50_fifo_context_attach;
 	nv_parent(chan)->context_detach = nv50_fifo_context_detach;
 	nv_parent(chan)->object_attach = nv50_fifo_object_attach;
@@ -223,10 +233,10 @@ nv50_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	nv_wo32(base->ramfc, 0x08, lower_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x0c, upper_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x10, lower_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x14, upper_32_bits(args->offset));
+	nv_wo32(base->ramfc, 0x08, lower_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x0c, upper_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x10, lower_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x14, upper_32_bits(args->v0.offset));
 	nv_wo32(base->ramfc, 0x3c, 0x003f6078);
 	nv_wo32(base->ramfc, 0x44, 0x01003fff);
 	nv_wo32(base->ramfc, 0x48, chan->base.pushgpu->node->offset >> 4);
@@ -247,18 +257,26 @@ nv50_fifo_chan_ctor_ind(struct nouveau_object *parent,
 			struct nouveau_oclass *oclass, void *data, u32 size,
 			struct nouveau_object **pobject)
 {
-	struct nv50_channel_ind_class *args = data;
+	union {
+		struct nv50_channel_gpfifo_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nv50_fifo_base *base = (void *)parent;
 	struct nv50_fifo_chan *chan;
 	u64 ioffset, ilength;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel gpfifo size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel gpfifo vers %d pushbuf %08x "
+				 "ioffset %016llx ilength %08x\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.ioffset,
+			 args->v0.ilength);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0xc00000,
-					  0x2000, args->pushbuf,
+					  0x2000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -267,6 +285,8 @@ nv50_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->context_attach = nv50_fifo_context_attach;
 	nv_parent(chan)->context_detach = nv50_fifo_context_detach;
 	nv_parent(chan)->object_attach = nv50_fifo_object_attach;
@@ -277,8 +297,8 @@ nv50_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	ioffset = args->ioffset;
-	ilength = order_base_2(args->ilength / 8);
+	ioffset = args->v0.ioffset;
+	ilength = order_base_2(args->v0.ilength / 8);
 
 	nv_wo32(base->ramfc, 0x3c, 0x403f6078);
 	nv_wo32(base->ramfc, 0x44, 0x01003fff);
@@ -343,8 +363,10 @@ nv50_fifo_ofuncs_dma = {
 	.dtor = nv50_fifo_chan_dtor,
 	.init = nv50_fifo_chan_init,
 	.fini = nv50_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_ofuncs
@@ -353,14 +375,16 @@ nv50_fifo_ofuncs_ind = {
 	.dtor = nv50_fifo_chan_dtor,
 	.init = nv50_fifo_chan_init,
 	.fini = nv50_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv50_fifo_sclass[] = {
-	{ NV50_CHANNEL_DMA_CLASS, &nv50_fifo_ofuncs_dma },
-	{ NV50_CHANNEL_IND_CLASS, &nv50_fifo_ofuncs_ind },
+	{ NV50_CHANNEL_DMA, &nv50_fifo_ofuncs_dma },
+	{ NV50_CHANNEL_GPFIFO, &nv50_fifo_ofuncs_ind },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 6e5ac16e5460..1f42996b354a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -27,7 +27,8 @@
 #include <core/engctx.h>
 #include <core/ramht.h>
 #include <core/event.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 
 #include <subdev/timer.h>
 #include <subdev/bar.h>
@@ -160,17 +161,24 @@ nv84_fifo_chan_ctor_dma(struct nouveau_object *parent,
 			struct nouveau_oclass *oclass, void *data, u32 size,
 			struct nouveau_object **pobject)
 {
+	union {
+		struct nv03_channel_dma_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nv50_fifo_base *base = (void *)parent;
 	struct nv50_fifo_chan *chan;
-	struct nv03_channel_dma_class *args = data;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel dma size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel dma vers %d pushbuf %08x "
+				 "offset %016llx\n", args->v0.version,
+			 args->v0.pushbuf, args->v0.offset);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0xc00000,
-					  0x2000, args->pushbuf,
+					  0x2000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -186,6 +194,8 @@ nv84_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
 			       &chan->ramht);
 	if (ret)
@@ -196,10 +206,10 @@ nv84_fifo_chan_ctor_dma(struct nouveau_object *parent,
 	nv_parent(chan)->object_attach = nv84_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv50_fifo_object_detach;
 
-	nv_wo32(base->ramfc, 0x08, lower_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x0c, upper_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x10, lower_32_bits(args->offset));
-	nv_wo32(base->ramfc, 0x14, upper_32_bits(args->offset));
+	nv_wo32(base->ramfc, 0x08, lower_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x0c, upper_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x10, lower_32_bits(args->v0.offset));
+	nv_wo32(base->ramfc, 0x14, upper_32_bits(args->v0.offset));
 	nv_wo32(base->ramfc, 0x3c, 0x003f6078);
 	nv_wo32(base->ramfc, 0x44, 0x01003fff);
 	nv_wo32(base->ramfc, 0x48, chan->base.pushgpu->node->offset >> 4);
@@ -222,18 +232,26 @@ nv84_fifo_chan_ctor_ind(struct nouveau_object *parent,
 			struct nouveau_oclass *oclass, void *data, u32 size,
 			struct nouveau_object **pobject)
 {
+	union {
+		struct nv50_channel_gpfifo_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nv50_fifo_base *base = (void *)parent;
 	struct nv50_fifo_chan *chan;
-	struct nv50_channel_ind_class *args = data;
 	u64 ioffset, ilength;
 	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel gpfifo size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel gpfifo vers %d pushbuf %08x "
+				 "ioffset %016llx ilength %08x\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.ioffset,
+			 args->v0.ilength);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 0, 0xc00000,
-					  0x2000, args->pushbuf,
+					  0x2000, args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_DMAOBJ) |
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
@@ -249,6 +267,8 @@ nv84_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	ret = nouveau_ramht_new(nv_object(chan), nv_object(chan), 0x8000, 16,
 			       &chan->ramht);
 	if (ret)
@@ -259,8 +279,8 @@ nv84_fifo_chan_ctor_ind(struct nouveau_object *parent,
 	nv_parent(chan)->object_attach = nv84_fifo_object_attach;
 	nv_parent(chan)->object_detach = nv50_fifo_object_detach;
 
-	ioffset = args->ioffset;
-	ilength = order_base_2(args->ilength / 8);
+	ioffset = args->v0.ioffset;
+	ilength = order_base_2(args->v0.ilength / 8);
 
 	nv_wo32(base->ramfc, 0x3c, 0x403f6078);
 	nv_wo32(base->ramfc, 0x44, 0x01003fff);
@@ -304,8 +324,10 @@ nv84_fifo_ofuncs_dma = {
 	.dtor = nv50_fifo_chan_dtor,
 	.init = nv84_fifo_chan_init,
 	.fini = nv50_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_ofuncs
@@ -314,14 +336,16 @@ nv84_fifo_ofuncs_ind = {
 	.dtor = nv50_fifo_chan_dtor,
 	.init = nv84_fifo_chan_init,
 	.fini = nv50_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nv84_fifo_sclass[] = {
-	{ NV84_CHANNEL_DMA_CLASS, &nv84_fifo_ofuncs_dma },
-	{ NV84_CHANNEL_IND_CLASS, &nv84_fifo_ofuncs_ind },
+	{ G82_CHANNEL_DMA, &nv84_fifo_ofuncs_dma },
+	{ G82_CHANNEL_GPFIFO, &nv84_fifo_ofuncs_ind },
 	{}
 };
 
@@ -389,19 +413,26 @@ nv84_fifo_cclass = {
  ******************************************************************************/
 
 static void
-nv84_fifo_uevent_enable(struct nouveau_event *event, int type, int index)
+nv84_fifo_uevent_init(struct nvkm_event *event, int type, int index)
 {
-	struct nv84_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x40000000, 0x40000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x40000000, 0x40000000);
 }
 
 static void
-nv84_fifo_uevent_disable(struct nouveau_event *event, int type, int index)
+nv84_fifo_uevent_fini(struct nvkm_event *event, int type, int index)
 {
-	struct nv84_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x40000000, 0x00000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x40000000, 0x00000000);
 }
 
+static const struct nvkm_event_func
+nv84_fifo_uevent_func = {
+	.ctor = nouveau_fifo_uevent_ctor,
+	.init = nv84_fifo_uevent_init,
+	.fini = nv84_fifo_uevent_fini,
+};
+
 static int
 nv84_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -425,9 +456,9 @@ nv84_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.uevent->enable = nv84_fifo_uevent_enable;
-	priv->base.uevent->disable = nv84_fifo_uevent_disable;
-	priv->base.uevent->priv = priv;
+	ret = nvkm_event_init(&nv84_fifo_uevent_func, 1, 1, &priv->base.uevent);
+	if (ret)
+		return ret;
 
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nv04_fifo_intr;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
index ae4a4dc5642a..1fe1f8fbda0c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nvc0.c
@@ -28,7 +28,8 @@
 #include <core/gpuobj.h>
 #include <core/engctx.h>
 #include <core/event.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/enum.h>
 
 #include <subdev/timer.h>
@@ -187,20 +188,28 @@ nvc0_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct nv50_channel_gpfifo_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nvc0_fifo_priv *priv = (void *)engine;
 	struct nvc0_fifo_base *base = (void *)parent;
 	struct nvc0_fifo_chan *chan;
-	struct nv50_channel_ind_class *args = data;
 	u64 usermem, ioffset, ilength;
 	int ret, i;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel gpfifo size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel gpfifo vers %d pushbuf %08x "
+				 "ioffset %016llx ilength %08x\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.ioffset,
+			 args->v0.ilength);
+	} else
+		return ret;
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 1,
 					  priv->user.bar.offset, 0x1000,
-					  args->pushbuf,
+					  args->v0.pushbuf,
 					  (1ULL << NVDEV_ENGINE_SW) |
 					  (1ULL << NVDEV_ENGINE_GR) |
 					  (1ULL << NVDEV_ENGINE_COPY0) |
@@ -212,12 +221,14 @@ nvc0_fifo_chan_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->context_attach = nvc0_fifo_context_attach;
 	nv_parent(chan)->context_detach = nvc0_fifo_context_detach;
 
 	usermem = chan->base.chid * 0x1000;
-	ioffset = args->ioffset;
-	ilength = order_base_2(args->ilength / 8);
+	ioffset = args->v0.ioffset;
+	ilength = order_base_2(args->v0.ilength / 8);
 
 	for (i = 0; i < 0x1000; i += 4)
 		nv_wo32(priv->user.mem, usermem + i, 0x00000000);
@@ -291,13 +302,15 @@ nvc0_fifo_ofuncs = {
 	.dtor = _nouveau_fifo_channel_dtor,
 	.init = nvc0_fifo_chan_init,
 	.fini = nvc0_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nvc0_fifo_sclass[] = {
-	{ NVC0_CHANNEL_IND_CLASS, &nvc0_fifo_ofuncs },
+	{ FERMI_CHANNEL_GPFIFO, &nvc0_fifo_ofuncs },
 	{}
 };
 
@@ -654,7 +667,7 @@ nvc0_fifo_intr_fault(struct nvc0_fifo_priv *priv, int unit)
 	object = engctx;
 	while (object) {
 		switch (nv_mclass(object)) {
-		case NVC0_CHANNEL_IND_CLASS:
+		case FERMI_CHANNEL_GPFIFO:
 			nvc0_fifo_recover(priv, engine, (void *)object);
 			break;
 		}
@@ -730,7 +743,7 @@ nvc0_fifo_intr_engine_unit(struct nvc0_fifo_priv *priv, int engn)
 	for (unkn = 0; unkn < 8; unkn++) {
 		u32 ints = (intr >> (unkn * 0x04)) & inte;
 		if (ints & 0x1) {
-			nouveau_event_trigger(priv->base.uevent, 1, 0);
+			nouveau_fifo_uevent(&priv->base);
 			ints &= ~1;
 		}
 		if (ints) {
@@ -827,19 +840,26 @@ nvc0_fifo_intr(struct nouveau_subdev *subdev)
 }
 
 static void
-nvc0_fifo_uevent_enable(struct nouveau_event *event, int type, int index)
+nvc0_fifo_uevent_init(struct nvkm_event *event, int type, int index)
 {
-	struct nvc0_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x80000000, 0x80000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x80000000, 0x80000000);
 }
 
 static void
-nvc0_fifo_uevent_disable(struct nouveau_event *event, int type, int index)
+nvc0_fifo_uevent_fini(struct nvkm_event *event, int type, int index)
 {
-	struct nvc0_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x80000000, 0x00000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x80000000, 0x00000000);
 }
 
+static const struct nvkm_event_func
+nvc0_fifo_uevent_func = {
+	.ctor = nouveau_fifo_uevent_ctor,
+	.init = nvc0_fifo_uevent_init,
+	.fini = nvc0_fifo_uevent_fini,
+};
+
 static int
 nvc0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
@@ -877,9 +897,9 @@ nvc0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.uevent->enable = nvc0_fifo_uevent_enable;
-	priv->base.uevent->disable = nvc0_fifo_uevent_disable;
-	priv->base.uevent->priv = priv;
+	ret = nvkm_event_init(&nvc0_fifo_uevent_func, 1, 1, &priv->base.uevent);
+	if (ret)
+		return ret;
 
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nvc0_fifo_intr;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 298063edb92d..d2f0fd39c145 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -28,7 +28,8 @@
 #include <core/gpuobj.h>
 #include <core/engctx.h>
 #include <core/event.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
 #include <core/enum.h>
 
 #include <subdev/timer.h>
@@ -216,46 +217,56 @@ nve0_fifo_chan_ctor(struct nouveau_object *parent,
 		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	union {
+		struct kepler_channel_gpfifo_a_v0 v0;
+	} *args = data;
 	struct nouveau_bar *bar = nouveau_bar(parent);
 	struct nve0_fifo_priv *priv = (void *)engine;
 	struct nve0_fifo_base *base = (void *)parent;
 	struct nve0_fifo_chan *chan;
-	struct nve0_channel_ind_class *args = data;
 	u64 usermem, ioffset, ilength;
 	int ret, i;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create channel gpfifo size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create channel gpfifo vers %d pushbuf %08x "
+				 "ioffset %016llx ilength %08x engine %08x\n",
+			 args->v0.version, args->v0.pushbuf, args->v0.ioffset,
+			 args->v0.ilength, args->v0.engine);
+	} else
+		return ret;
 
 	for (i = 0; i < FIFO_ENGINE_NR; i++) {
-		if (args->engine & (1 << i)) {
+		if (args->v0.engine & (1 << i)) {
 			if (nouveau_engine(parent, fifo_engine[i].subdev)) {
-				args->engine = (1 << i);
+				args->v0.engine = (1 << i);
 				break;
 			}
 		}
 	}
 
 	if (i == FIFO_ENGINE_NR) {
-		nv_error(priv, "unsupported engines 0x%08x\n", args->engine);
+		nv_error(priv, "unsupported engines 0x%08x\n", args->v0.engine);
 		return -ENODEV;
 	}
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 1,
 					  priv->user.bar.offset, 0x200,
-					  args->pushbuf,
+					  args->v0.pushbuf,
 					  fifo_engine[i].mask, &chan);
 	*pobject = nv_object(chan);
 	if (ret)
 		return ret;
 
+	args->v0.chid = chan->base.chid;
+
 	nv_parent(chan)->context_attach = nve0_fifo_context_attach;
 	nv_parent(chan)->context_detach = nve0_fifo_context_detach;
 	chan->engine = i;
 
 	usermem = chan->base.chid * 0x200;
-	ioffset = args->ioffset;
-	ilength = order_base_2(args->ilength / 8);
+	ioffset = args->v0.ioffset;
+	ilength = order_base_2(args->v0.ilength / 8);
 
 	for (i = 0; i < 0x200; i += 4)
 		nv_wo32(priv->user.mem, usermem + i, 0x00000000);
@@ -325,13 +336,15 @@ nve0_fifo_ofuncs = {
 	.dtor = _nouveau_fifo_channel_dtor,
 	.init = nve0_fifo_chan_init,
 	.fini = nve0_fifo_chan_fini,
+	.map  = _nouveau_fifo_channel_map,
 	.rd32 = _nouveau_fifo_channel_rd32,
 	.wr32 = _nouveau_fifo_channel_wr32,
+	.ntfy = _nouveau_fifo_channel_ntfy
 };
 
 static struct nouveau_oclass
 nve0_fifo_sclass[] = {
-	{ NVE0_CHANNEL_IND_CLASS, &nve0_fifo_ofuncs },
+	{ KEPLER_CHANNEL_GPFIFO_A, &nve0_fifo_ofuncs },
 	{}
 };
 
@@ -769,7 +782,7 @@ nve0_fifo_intr_fault(struct nve0_fifo_priv *priv, int unit)
 	object = engctx;
 	while (object) {
 		switch (nv_mclass(object)) {
-		case NVE0_CHANNEL_IND_CLASS:
+		case KEPLER_CHANNEL_GPFIFO_A:
 			nve0_fifo_recover(priv, engine, (void *)object);
 			break;
 		}
@@ -859,7 +872,7 @@ nve0_fifo_intr_runlist(struct nve0_fifo_priv *priv)
 static void
 nve0_fifo_intr_engine(struct nve0_fifo_priv *priv)
 {
-	nouveau_event_trigger(priv->base.uevent, 1, 0);
+	nouveau_fifo_uevent(&priv->base);
 }
 
 static void
@@ -952,19 +965,26 @@ nve0_fifo_intr(struct nouveau_subdev *subdev)
 }
 
 static void
-nve0_fifo_uevent_enable(struct nouveau_event *event, int type, int index)
+nve0_fifo_uevent_init(struct nvkm_event *event, int type, int index)
 {
-	struct nve0_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x80000000, 0x80000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x80000000, 0x80000000);
 }
 
 static void
-nve0_fifo_uevent_disable(struct nouveau_event *event, int type, int index)
+nve0_fifo_uevent_fini(struct nvkm_event *event, int type, int index)
 {
-	struct nve0_fifo_priv *priv = event->priv;
-	nv_mask(priv, 0x002140, 0x80000000, 0x00000000);
+	struct nouveau_fifo *fifo = container_of(event, typeof(*fifo), uevent);
+	nv_mask(fifo, 0x002140, 0x80000000, 0x00000000);
 }
 
+static const struct nvkm_event_func
+nve0_fifo_uevent_func = {
+	.ctor = nouveau_fifo_uevent_ctor,
+	.init = nve0_fifo_uevent_init,
+	.fini = nve0_fifo_uevent_fini,
+};
+
 int
 nve0_fifo_fini(struct nouveau_object *object, bool suspend)
 {
@@ -1067,9 +1087,9 @@ nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.uevent->enable = nve0_fifo_uevent_enable;
-	priv->base.uevent->disable = nve0_fifo_uevent_disable;
-	priv->base.uevent->priv = priv;
+	ret = nvkm_event_init(&nve0_fifo_uevent_func, 1, 1, &priv->base.uevent);
+	if (ret)
+		return ret;
 
 	nv_subdev(priv)->unit = 0x00000100;
 	nv_subdev(priv)->intr = nve0_fifo_intr;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk110b.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk110b.c
new file mode 100644
index 000000000000..3adb7fe91772
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk110b.c
@@ -0,0 +1,104 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "ctxnvc0.h"
+
+/*******************************************************************************
+ * PGRAPH context register lists
+ ******************************************************************************/
+
+static const struct nvc0_graph_init
+gk110b_grctx_init_sm_0[] = {
+	{ 0x419e04,   1, 0x04, 0x00000000 },
+	{ 0x419e08,   1, 0x04, 0x0000001d },
+	{ 0x419e0c,   1, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00001c02 },
+	{ 0x419e44,   1, 0x04, 0x0013eff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000007f },
+	{ 0x419e50,   2, 0x04, 0x00000000 },
+	{ 0x419e58,   1, 0x04, 0x00000001 },
+	{ 0x419e5c,   3, 0x04, 0x00000000 },
+	{ 0x419e68,   1, 0x04, 0x00000002 },
+	{ 0x419e6c,  12, 0x04, 0x00000000 },
+	{ 0x419eac,   1, 0x04, 0x00001f8f },
+	{ 0x419eb0,   1, 0x04, 0x0db00d2f },
+	{ 0x419eb8,   1, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0001304f },
+	{ 0x419f30,   4, 0x04, 0x00000000 },
+	{ 0x419f40,   1, 0x04, 0x00000018 },
+	{ 0x419f44,   3, 0x04, 0x00000000 },
+	{ 0x419f58,   1, 0x04, 0x00000000 },
+	{ 0x419f70,   1, 0x04, 0x00006300 },
+	{ 0x419f78,   1, 0x04, 0x000000eb },
+	{ 0x419f7c,   1, 0x04, 0x00000404 },
+	{}
+};
+
+static const struct nvc0_graph_pack
+gk110b_grctx_pack_tpc[] = {
+	{ nvd7_grctx_init_pe_0 },
+	{ nvf0_grctx_init_tex_0 },
+	{ nvf0_grctx_init_mpc_0 },
+	{ nvf0_grctx_init_l1c_0 },
+	{ gk110b_grctx_init_sm_0 },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH context implementation
+ ******************************************************************************/
+
+struct nouveau_oclass *
+gk110b_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xf1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main  = nve4_grctx_generate_main,
+	.unkn  = nve4_grctx_generate_unkn,
+	.hub   = nvf0_grctx_pack_hub,
+	.gpc   = nvf0_grctx_pack_gpc,
+	.zcull = nvc0_grctx_pack_zcull,
+	.tpc   = gk110b_grctx_pack_tpc,
+	.ppc   = nvf0_grctx_pack_ppc,
+	.icmd  = nvf0_grctx_pack_icmd,
+	.mthd  = nvf0_grctx_pack_mthd,
+	.bundle = nve4_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0x180,
+	.bundle_token_limit = 0x600,
+	.pagepool = nve4_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x7ff,
+	.alpha_nr = 0x648,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk20a.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk20a.c
index 224ee0287ab7..36fc9831cc93 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk20a.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgk20a.c
@@ -41,7 +41,6 @@ gk20a_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nve4_grctx_generate_main,
-	.mods  = nve4_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = nve4_grctx_pack_hub,
 	.gpc   = nve4_grctx_pack_gpc,
@@ -50,4 +49,15 @@ gk20a_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = nve4_grctx_pack_ppc,
 	.icmd  = nve4_grctx_pack_icmd,
 	.mthd  = gk20a_grctx_pack_mthd,
+	.bundle = nve4_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.bundle_min_gpm_fifo_depth = 0x62,
+	.bundle_token_limit = 0x100,
+	.pagepool = nve4_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x240,
+	.attrib_nr = 0x240,
+	.alpha_nr_max = 0x648 + (0x648 / 2),
+	.alpha_nr = 0x648,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxgm107.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgm107.c
index b0d0fb2f4d08..62e918b9fa81 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxgm107.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxgm107.c
@@ -859,45 +859,74 @@ gm107_grctx_pack_ppc[] = {
  ******************************************************************************/
 
 static void
-gm107_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+gm107_grctx_generate_bundle(struct nvc0_grctx *info)
 {
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x200000, 0x1000, NV_MEM_ACCESS_RW);
-
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x418e30, 0x80000000,  0, 0);
-
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418e24, 0x00000000,  8, 0);
-	mmio_list(0x418e28, 0x80000030,  0, 0);
-
-	mmio_list(0x4064c8, 0x018002c0,  0, 0);
-
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x419c2c, 0x10000000, 12, 2);
-
-	mmio_list(0x405830, 0x0aa01000,  0, 0);
-	mmio_list(0x4064c4, 0x0400ffff,  0, 0);
-
-	/*XXX*/
-	mmio_list(0x5030c0, 0x00001540,  0, 0);
-	mmio_list(0x5030f4, 0x00000000,  0, 0);
-	mmio_list(0x5030e4, 0x00002000,  0, 0);
-	mmio_list(0x5030f8, 0x00003fc0,  0, 0);
-	mmio_list(0x418ea0, 0x07151540,  0, 0);
-
-	mmio_list(0x5032c0, 0x00001540,  0, 0);
-	mmio_list(0x5032f4, 0x00001fe0,  0, 0);
-	mmio_list(0x5032e4, 0x00002000,  0, 0);
-	mmio_list(0x5032f8, 0x00006fc0,  0, 0);
-	mmio_list(0x418ea4, 0x07151540,  0, 0);
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 state_limit = min(impl->bundle_min_gpm_fifo_depth,
+				    impl->bundle_size / 0x20);
+	const u32 token_limit = impl->bundle_token_limit;
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->bundle_size, (1 << s), access);
+	mmio_refn(info, 0x408004, 0x00000000, s, b);
+	mmio_refn(info, 0x408008, 0x80000000 | (impl->bundle_size >> s), 0, b);
+	mmio_refn(info, 0x418e24, 0x00000000, s, b);
+	mmio_refn(info, 0x418e28, 0x80000000 | (impl->bundle_size >> s), 0, b);
+	mmio_wr32(info, 0x4064c8, (state_limit << 16) | token_limit);
+}
+
+static void
+gm107_grctx_generate_pagepool(struct nvc0_grctx *info)
+{
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->pagepool_size, (1 << s), access);
+	mmio_refn(info, 0x40800c, 0x00000000, s, b);
+	mmio_wr32(info, 0x408010, 0x80000000);
+	mmio_refn(info, 0x419004, 0x00000000, s, b);
+	mmio_wr32(info, 0x419008, 0x00000000);
+	mmio_wr32(info, 0x4064cc, 0x80000000);
+	mmio_wr32(info, 0x418e30, 0x80000000); /* guess at it being related */
+}
+
+static void
+gm107_grctx_generate_attrib(struct nvc0_grctx *info)
+{
+	struct nvc0_graph_priv *priv = info->priv;
+	const struct nvc0_grctx_oclass *impl = (void *)nvc0_grctx_impl(priv);
+	const u32  alpha = impl->alpha_nr;
+	const u32 attrib = impl->attrib_nr;
+	const u32   size = 0x20 * (impl->attrib_nr_max + impl->alpha_nr_max);
+	const u32 access = NV_MEM_ACCESS_RW;
+	const int s = 12;
+	const int b = mmio_vram(info, size * priv->tpc_total, (1 << s), access);
+	const int max_batches = 0xffff;
+	u32 bo = 0;
+	u32 ao = bo + impl->attrib_nr_max * priv->tpc_total;
+	int gpc, ppc, n = 0;
+
+	mmio_refn(info, 0x418810, 0x80000000, s, b);
+	mmio_refn(info, 0x419848, 0x10000000, s, b);
+	mmio_refn(info, 0x419c2c, 0x10000000, s, b);
+	mmio_wr32(info, 0x405830, (attrib << 16) | alpha);
+	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		for (ppc = 0; ppc < priv->ppc_nr[gpc]; ppc++, n++) {
+			const u32 as =  alpha * priv->ppc_tpc_nr[gpc][ppc];
+			const u32 bs = attrib * priv->ppc_tpc_nr[gpc][ppc];
+			const u32 u = 0x418ea0 + (n * 0x04);
+			const u32 o = PPC_UNIT(gpc, ppc, 0);
+			mmio_wr32(info, o + 0xc0, bs);
+			mmio_wr32(info, o + 0xf4, bo);
+			bo += impl->attrib_nr_max * priv->ppc_tpc_nr[gpc][ppc];
+			mmio_wr32(info, o + 0xe4, as);
+			mmio_wr32(info, o + 0xf8, ao);
+			ao += impl->alpha_nr_max * priv->ppc_tpc_nr[gpc][ppc];
+			mmio_wr32(info, u, (0x715 /*XXX*/ << 16) | bs);
+		}
+	}
 }
 
 static void
@@ -934,7 +963,9 @@ gm107_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 
 	nv_wr32(priv, 0x404154, 0x00000000);
 
-	oclass->mods(priv, info);
+	oclass->bundle(info);
+	oclass->pagepool(info);
+	oclass->attrib(info);
 	oclass->unkn(priv);
 
 	gm107_grctx_generate_tpcid(priv);
@@ -979,7 +1010,6 @@ gm107_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = gm107_grctx_generate_main,
-	.mods  = gm107_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = gm107_grctx_pack_hub,
 	.gpc   = gm107_grctx_pack_gpc,
@@ -988,4 +1018,15 @@ gm107_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = gm107_grctx_pack_ppc,
 	.icmd  = gm107_grctx_pack_icmd,
 	.mthd  = gm107_grctx_pack_mthd,
+	.bundle = gm107_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0x180,
+	.bundle_token_limit = 0x2c0,
+	.pagepool = gm107_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = gm107_grctx_generate_attrib,
+	.attrib_nr_max = 0xff0,
+	.attrib_nr = 0xaa0,
+	.alpha_nr_max = 0x1800,
+	.alpha_nr = 0x1000,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv108.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv108.c
index 8de4a4291548..ce252adbef81 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv108.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv108.c
@@ -531,50 +531,6 @@ nv108_grctx_pack_ppc[] = {
  * PGRAPH context implementation
  ******************************************************************************/
 
-static void
-nv108_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
-{
-	u32 magic[GPC_MAX][2];
-	u32 offset;
-	int gpc;
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000030,  0, 0);
-	mmio_list(0x4064c8, 0x00c20200,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-
-	mmio_list(0x405830, 0x02180648,  0, 0);
-	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
-		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * priv->tpc_nr[gpc];
-	}
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * priv->tpc_nr[gpc];
-	}
-
-	mmio_list(0x17e91c, 0x0b040a0b, 0, 0);
-	mmio_list(0x17e920, 0x00090d08, 0, 0);
-}
-
 struct nouveau_oclass *
 nv108_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.base.handle = NV_ENGCTX(GR, 0x08),
@@ -587,7 +543,6 @@ nv108_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nve4_grctx_generate_main,
-	.mods  = nv108_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = nv108_grctx_pack_hub,
 	.gpc   = nv108_grctx_pack_gpc,
@@ -596,4 +551,15 @@ nv108_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = nv108_grctx_pack_ppc,
 	.icmd  = nv108_grctx_pack_icmd,
 	.mthd  = nvf0_grctx_pack_mthd,
+	.bundle = nve4_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0xc2,
+	.bundle_token_limit = 0x200,
+	.pagepool = nve4_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x7ff,
+	.alpha_nr = 0x648,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
index 552fdbd45ebe..1d0e33fb5f61 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnv50.c
@@ -113,6 +113,8 @@
 #define IS_NVA3F(x) (((x) > 0xa0 && (x) < 0xaa) || (x) == 0xaf)
 #define IS_NVAAF(x) ((x) >= 0xaa && (x) <= 0xac)
 
+#include <subdev/fb.h>
+
 /*
  * This code deals with PGRAPH contexts on NV50 family cards. Like NV40, it's
  * the GPU itself that does context-switching, but it needs a special
@@ -569,8 +571,12 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 			gr_def(ctx, 0x407d08, 0x00010040);
 		else if (device->chipset < 0xa0)
 			gr_def(ctx, 0x407d08, 0x00390040);
-		else
-			gr_def(ctx, 0x407d08, 0x003d0040);
+		else {
+			if (nouveau_fb(device)->ram->type != NV_MEM_TYPE_GDDR5)
+				gr_def(ctx, 0x407d08, 0x003d0040);
+			else
+				gr_def(ctx, 0x407d08, 0x003c0040);
+		}
 		gr_def(ctx, 0x407d0c, 0x00000022);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
index 833a96508c4e..b8e5fe60a1eb 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
@@ -982,34 +982,93 @@ nvc0_grctx_pack_tpc[] = {
  * PGRAPH context implementation
  ******************************************************************************/
 
+int
+nvc0_grctx_mmio_data(struct nvc0_grctx *info, u32 size, u32 align, u32 access)
+{
+	if (info->data) {
+		info->buffer[info->buffer_nr] = round_up(info->addr, align);
+		info->addr = info->buffer[info->buffer_nr] + size;
+		info->data->size = size;
+		info->data->align = align;
+		info->data->access = access;
+		info->data++;
+		return info->buffer_nr++;
+	}
+	return -1;
+}
+
+void
+nvc0_grctx_mmio_item(struct nvc0_grctx *info, u32 addr, u32 data,
+		     int shift, int buffer)
+{
+	if (info->data) {
+		if (shift >= 0) {
+			info->mmio->addr = addr;
+			info->mmio->data = data;
+			info->mmio->shift = shift;
+			info->mmio->buffer = buffer;
+			if (buffer >= 0)
+				data |= info->buffer[buffer] >> shift;
+			info->mmio++;
+		} else
+			return;
+	} else {
+		if (buffer >= 0)
+			return;
+	}
+
+	nv_wr32(info->priv, addr, data);
+}
+
+void
+nvc0_grctx_generate_bundle(struct nvc0_grctx *info)
+{
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->bundle_size, (1 << s), access);
+	mmio_refn(info, 0x408004, 0x00000000, s, b);
+	mmio_refn(info, 0x408008, 0x80000000 | (impl->bundle_size >> s), 0, b);
+	mmio_refn(info, 0x418808, 0x00000000, s, b);
+	mmio_refn(info, 0x41880c, 0x80000000 | (impl->bundle_size >> s), 0, b);
+}
+
 void
-nvc0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+nvc0_grctx_generate_pagepool(struct nvc0_grctx *info)
 {
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->pagepool_size, (1 << s), access);
+	mmio_refn(info, 0x40800c, 0x00000000, s, b);
+	mmio_wr32(info, 0x408010, 0x80000000);
+	mmio_refn(info, 0x419004, 0x00000000, s, b);
+	mmio_wr32(info, 0x419008, 0x00000000);
+}
+
+void
+nvc0_grctx_generate_attrib(struct nvc0_grctx *info)
+{
+	struct nvc0_graph_priv *priv = info->priv;
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(priv);
+	const u32 attrib = impl->attrib_nr;
+	const u32   size = 0x20 * (impl->attrib_nr_max + impl->alpha_nr_max);
+	const u32 access = NV_MEM_ACCESS_RW;
+	const int s = 12;
+	const int b = mmio_vram(info, size * priv->tpc_total, (1 << s), access);
 	int gpc, tpc;
-	u32 offset;
-
-	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000018,  0, 0);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000018,  0, 0);
-
-	mmio_list(0x405830, 0x02180000, 0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+	u32 bo = 0;
+
+	mmio_refn(info, 0x418810, 0x80000000, s, b);
+	mmio_refn(info, 0x419848, 0x10000000, s, b);
+	mmio_wr32(info, 0x405830, (attrib << 16));
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
 		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
-			mmio_list(addr, 0x02180000 | offset, 0, 0);
-			offset += 0x0324;
+			const u32 o = TPC_UNIT(gpc, tpc, 0x0520);
+			mmio_skip(info, o, (attrib << 16) | ++bo);
+			mmio_wr32(info, o, (attrib << 16) | --bo);
+			bo += impl->attrib_nr_max;
 		}
 	}
 }
@@ -1170,7 +1229,7 @@ nvc0_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
 	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
 
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 0);
 
 	nvc0_graph_mmio(priv, oclass->hub);
 	nvc0_graph_mmio(priv, oclass->gpc);
@@ -1180,7 +1239,9 @@ nvc0_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 
 	nv_wr32(priv, 0x404154, 0x00000000);
 
-	oclass->mods(priv, info);
+	oclass->bundle(info);
+	oclass->pagepool(info);
+	oclass->attrib(info);
 	oclass->unkn(priv);
 
 	nvc0_grctx_generate_tpcid(priv);
@@ -1192,7 +1253,7 @@ nvc0_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	nvc0_graph_icmd(priv, oclass->icmd);
 	nv_wr32(priv, 0x404154, 0x00000400);
 	nvc0_graph_mthd(priv, oclass->mthd);
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 1);
 }
 
 int
@@ -1308,7 +1369,6 @@ nvc0_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvc0_grctx_generate_main,
-	.mods  = nvc0_grctx_generate_mods,
 	.unkn  = nvc0_grctx_generate_unkn,
 	.hub   = nvc0_grctx_pack_hub,
 	.gpc   = nvc0_grctx_pack_gpc,
@@ -1316,4 +1376,11 @@ nvc0_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.tpc   = nvc0_grctx_pack_tpc,
 	.icmd  = nvc0_grctx_pack_icmd,
 	.mthd  = nvc0_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvc0_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.h b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.h
index 8da8b627b9d0..c776cd715e33 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.h
@@ -12,12 +12,19 @@ struct nvc0_grctx {
 	u64 addr;
 };
 
+int  nvc0_grctx_mmio_data(struct nvc0_grctx *, u32 size, u32 align, u32 access);
+void nvc0_grctx_mmio_item(struct nvc0_grctx *, u32 addr, u32 data, int s, int);
+
+#define mmio_vram(a,b,c,d) nvc0_grctx_mmio_data((a), (b), (c), (d))
+#define mmio_refn(a,b,c,d,e) nvc0_grctx_mmio_item((a), (b), (c), (d), (e))
+#define mmio_skip(a,b,c) mmio_refn((a), (b), (c), -1, -1)
+#define mmio_wr32(a,b,c) mmio_refn((a), (b), (c),  0, -1)
+
 struct nvc0_grctx_oclass {
 	struct nouveau_oclass base;
 	/* main context generation function */
 	void  (*main)(struct nvc0_graph_priv *, struct nvc0_grctx *);
 	/* context-specific modify-on-first-load list generation function */
-	void  (*mods)(struct nvc0_graph_priv *, struct nvc0_grctx *);
 	void  (*unkn)(struct nvc0_graph_priv *);
 	/* mmio context data */
 	const struct nvc0_graph_pack *hub;
@@ -28,30 +35,34 @@ struct nvc0_grctx_oclass {
 	/* indirect context data, generated with icmds/mthds */
 	const struct nvc0_graph_pack *icmd;
 	const struct nvc0_graph_pack *mthd;
+	/* bundle circular buffer */
+	void (*bundle)(struct nvc0_grctx *);
+	u32 bundle_size;
+	u32 bundle_min_gpm_fifo_depth;
+	u32 bundle_token_limit;
+	/* pagepool */
+	void (*pagepool)(struct nvc0_grctx *);
+	u32 pagepool_size;
+	/* attribute(/alpha) circular buffer */
+	void (*attrib)(struct nvc0_grctx *);
+	u32 attrib_nr_max;
+	u32 attrib_nr;
+	u32 alpha_nr_max;
+	u32 alpha_nr;
 };
 
-#define mmio_data(s,a,p) do {                                                  \
-	info->buffer[info->buffer_nr] = round_up(info->addr, (a));             \
-	info->addr = info->buffer[info->buffer_nr++] + (s);                    \
-	info->data->size = (s);                                                \
-	info->data->align = (a);                                               \
-	info->data->access = (p);                                              \
-	info->data++;                                                          \
-} while(0)
-
-#define mmio_list(r,d,s,b) do {                                                \
-	info->mmio->addr = (r);                                                \
-	info->mmio->data = (d);                                                \
-	info->mmio->shift = (s);                                               \
-	info->mmio->buffer = (b);                                              \
-	info->mmio++;                                                          \
-	nv_wr32(priv, (r), (d) | ((s) ? (info->buffer[(b)] >> (s)) : 0));      \
-} while(0)
+static inline const struct nvc0_grctx_oclass *
+nvc0_grctx_impl(struct nvc0_graph_priv *priv)
+{
+	return (void *)nv_engine(priv)->cclass;
+}
 
 extern struct nouveau_oclass *nvc0_grctx_oclass;
 int  nvc0_grctx_generate(struct nvc0_graph_priv *);
 void nvc0_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
-void nvc0_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc0_grctx_generate_bundle(struct nvc0_grctx *);
+void nvc0_grctx_generate_pagepool(struct nvc0_grctx *);
+void nvc0_grctx_generate_attrib(struct nvc0_grctx *);
 void nvc0_grctx_generate_unkn(struct nvc0_graph_priv *);
 void nvc0_grctx_generate_tpcid(struct nvc0_graph_priv *);
 void nvc0_grctx_generate_r406028(struct nvc0_graph_priv *);
@@ -60,22 +71,27 @@ void nvc0_grctx_generate_r418bb8(struct nvc0_graph_priv *);
 void nvc0_grctx_generate_r406800(struct nvc0_graph_priv *);
 
 extern struct nouveau_oclass *nvc1_grctx_oclass;
-void nvc1_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc1_grctx_generate_attrib(struct nvc0_grctx *);
 void nvc1_grctx_generate_unkn(struct nvc0_graph_priv *);
 
 extern struct nouveau_oclass *nvc4_grctx_oclass;
 extern struct nouveau_oclass *nvc8_grctx_oclass;
+
 extern struct nouveau_oclass *nvd7_grctx_oclass;
+void nvd7_grctx_generate_attrib(struct nvc0_grctx *);
+
 extern struct nouveau_oclass *nvd9_grctx_oclass;
 
 extern struct nouveau_oclass *nve4_grctx_oclass;
 extern struct nouveau_oclass *gk20a_grctx_oclass;
 void nve4_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
-void nve4_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nve4_grctx_generate_bundle(struct nvc0_grctx *);
+void nve4_grctx_generate_pagepool(struct nvc0_grctx *);
 void nve4_grctx_generate_unkn(struct nvc0_graph_priv *);
 void nve4_grctx_generate_r418bb8(struct nvc0_graph_priv *);
 
 extern struct nouveau_oclass *nvf0_grctx_oclass;
+extern struct nouveau_oclass *gk110b_grctx_oclass;
 extern struct nouveau_oclass *nv108_grctx_oclass;
 extern struct nouveau_oclass *gm107_grctx_oclass;
 
@@ -160,16 +176,23 @@ extern const struct nvc0_graph_pack nve4_grctx_pack_ppc[];
 extern const struct nvc0_graph_pack nve4_grctx_pack_icmd[];
 extern const struct nvc0_graph_init nve4_grctx_init_a097_0[];
 
+extern const struct nvc0_graph_pack nvf0_grctx_pack_icmd[];
+
 extern const struct nvc0_graph_pack nvf0_grctx_pack_mthd[];
 
+extern const struct nvc0_graph_pack nvf0_grctx_pack_hub[];
 extern const struct nvc0_graph_init nvf0_grctx_init_pri_0[];
 extern const struct nvc0_graph_init nvf0_grctx_init_cwd_0[];
 
+extern const struct nvc0_graph_pack nvf0_grctx_pack_gpc[];
 extern const struct nvc0_graph_init nvf0_grctx_init_gpc_unk_2[];
 
+extern const struct nvc0_graph_init nvf0_grctx_init_tex_0[];
 extern const struct nvc0_graph_init nvf0_grctx_init_mpc_0[];
 extern const struct nvc0_graph_init nvf0_grctx_init_l1c_0[];
 
+extern const struct nvc0_graph_pack nvf0_grctx_pack_ppc[];
+
 extern const struct nvc0_graph_init nv108_grctx_init_rstr2d_0[];
 
 extern const struct nvc0_graph_init nv108_grctx_init_prop_0[];
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
index 24a92c569c0a..c6ba8fed18f1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
@@ -727,38 +727,38 @@ nvc1_grctx_pack_tpc[] = {
  ******************************************************************************/
 
 void
-nvc1_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+nvc1_grctx_generate_attrib(struct nvc0_grctx *info)
 {
+	struct nvc0_graph_priv *priv = info->priv;
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(priv);
+	const u32  alpha = impl->alpha_nr;
+	const u32   beta = impl->attrib_nr;
+	const u32   size = 0x20 * (impl->attrib_nr_max + impl->alpha_nr_max);
+	const u32 access = NV_MEM_ACCESS_RW;
+	const int s = 12;
+	const int b = mmio_vram(info, size * priv->tpc_total, (1 << s), access);
+	const int timeslice_mode = 1;
+	const int max_batches = 0xffff;
+	u32 bo = 0;
+	u32 ao = bo + impl->attrib_nr_max * priv->tpc_total;
 	int gpc, tpc;
-	u32 offset;
 
-	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000018,  0, 0);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000018,  0, 0);
+	mmio_refn(info, 0x418810, 0x80000000, s, b);
+	mmio_refn(info, 0x419848, 0x10000000, s, b);
+	mmio_wr32(info, 0x405830, (beta << 16) | alpha);
+	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
 
-	mmio_list(0x405830, 0x02180218, 0, 0);
-	mmio_list(0x4064c4, 0x0086ffff, 0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
-			mmio_list(addr, 0x12180000 | offset, 0, 0);
-			offset += 0x0324;
-		}
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
 		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			u32 addr = TPC_UNIT(gpc, tpc, 0x0544);
-			mmio_list(addr, 0x02180000 | offset, 0, 0);
-			offset += 0x0324;
+			const u32 a = alpha;
+			const u32 b =  beta;
+			const u32 t = timeslice_mode;
+			const u32 o = TPC_UNIT(gpc, tpc, 0x500);
+			mmio_skip(info, o + 0x20, (t << 28) | (b << 16) | ++bo);
+			mmio_wr32(info, o + 0x20, (t << 28) | (b << 16) | --bo);
+			bo += impl->attrib_nr_max;
+			mmio_wr32(info, o + 0x44, (a << 16) | ao);
+			ao += impl->alpha_nr_max;
 		}
 	}
 }
@@ -786,7 +786,6 @@ nvc1_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvc0_grctx_generate_main,
-	.mods  = nvc1_grctx_generate_mods,
 	.unkn  = nvc1_grctx_generate_unkn,
 	.hub   = nvc1_grctx_pack_hub,
 	.gpc   = nvc1_grctx_pack_gpc,
@@ -794,4 +793,13 @@ nvc1_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.tpc   = nvc1_grctx_pack_tpc,
 	.icmd  = nvc1_grctx_pack_icmd,
 	.mthd  = nvc1_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvc1_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x324,
+	.alpha_nr = 0x218,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc4.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc4.c
index e11ed5538193..41705c60cc47 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc4.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc4.c
@@ -92,7 +92,6 @@ nvc4_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvc0_grctx_generate_main,
-	.mods  = nvc0_grctx_generate_mods,
 	.unkn  = nvc0_grctx_generate_unkn,
 	.hub   = nvc0_grctx_pack_hub,
 	.gpc   = nvc0_grctx_pack_gpc,
@@ -100,4 +99,11 @@ nvc4_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.tpc   = nvc4_grctx_pack_tpc,
 	.icmd  = nvc0_grctx_pack_icmd,
 	.mthd  = nvc0_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvc0_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
index feebd58dfe8d..8f804cd8f9c7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
@@ -343,7 +343,6 @@ nvc8_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvc0_grctx_generate_main,
-	.mods  = nvc0_grctx_generate_mods,
 	.unkn  = nvc0_grctx_generate_unkn,
 	.hub   = nvc0_grctx_pack_hub,
 	.gpc   = nvc8_grctx_pack_gpc,
@@ -351,4 +350,11 @@ nvc8_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.tpc   = nvc0_grctx_pack_tpc,
 	.icmd  = nvc8_grctx_pack_icmd,
 	.mthd  = nvc8_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvc0_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
index 1dbc8d7f2e86..fcf534fd9e65 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
@@ -177,44 +177,41 @@ nvd7_grctx_pack_ppc[] = {
  * PGRAPH context implementation
  ******************************************************************************/
 
-static void
-nvd7_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+void
+nvd7_grctx_generate_attrib(struct nvc0_grctx *info)
 {
-	u32 magic[GPC_MAX][2];
-	u32 offset;
-	int gpc;
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000018,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000018,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
+	struct nvc0_graph_priv *priv = info->priv;
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(priv);
+	const u32  alpha = impl->alpha_nr;
+	const u32   beta = impl->attrib_nr;
+	const u32   size = 0x20 * (impl->attrib_nr_max + impl->alpha_nr_max);
+	const u32 access = NV_MEM_ACCESS_RW;
+	const int s = 12;
+	const int b = mmio_vram(info, size * priv->tpc_total, (1 << s), access);
+	const int timeslice_mode = 1;
+	const int max_batches = 0xffff;
+	u32 bo = 0;
+	u32 ao = bo + impl->attrib_nr_max * priv->tpc_total;
+	int gpc, ppc;
 
-	mmio_list(0x405830, 0x02180324,  0, 0);
-	mmio_list(0x4064c4, 0x00c9ffff,  0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
-		u16 magic1 = 0x0324 * priv->tpc_nr[gpc];
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * priv->tpc_nr[gpc];
-	}
+	mmio_refn(info, 0x418810, 0x80000000, s, b);
+	mmio_refn(info, 0x419848, 0x10000000, s, b);
+	mmio_wr32(info, 0x405830, (beta << 16) | alpha);
+	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
 
 	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * priv->tpc_nr[gpc];
+		for (ppc = 0; ppc < priv->ppc_nr[gpc]; ppc++) {
+			const u32 a = alpha * priv->ppc_tpc_nr[gpc][ppc];
+			const u32 b =  beta * priv->ppc_tpc_nr[gpc][ppc];
+			const u32 t = timeslice_mode;
+			const u32 o = PPC_UNIT(gpc, ppc, 0);
+			mmio_skip(info, o + 0xc0, (t << 28) | (b << 16) | ++bo);
+			mmio_wr32(info, o + 0xc0, (t << 28) | (b << 16) | --bo);
+			bo += impl->attrib_nr_max * priv->ppc_tpc_nr[gpc][ppc];
+			mmio_wr32(info, o + 0xe4, (a << 16) | ao);
+			ao += impl->alpha_nr_max * priv->ppc_tpc_nr[gpc][ppc];
+		}
 	}
-	mmio_list(0x17e91c, 0x03060609, 0, 0); /* different from kepler */
 }
 
 void
@@ -223,7 +220,7 @@ nvd7_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
 	int i;
 
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 0);
 
 	nvc0_graph_mmio(priv, oclass->hub);
 	nvc0_graph_mmio(priv, oclass->gpc);
@@ -233,7 +230,9 @@ nvd7_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 
 	nv_wr32(priv, 0x404154, 0x00000000);
 
-	oclass->mods(priv, info);
+	oclass->bundle(info);
+	oclass->pagepool(info);
+	oclass->attrib(info);
 	oclass->unkn(priv);
 
 	nvc0_grctx_generate_tpcid(priv);
@@ -248,7 +247,7 @@ nvd7_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	nvc0_graph_icmd(priv, oclass->icmd);
 	nv_wr32(priv, 0x404154, 0x00000400);
 	nvc0_graph_mthd(priv, oclass->mthd);
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 1);
 }
 
 struct nouveau_oclass *
@@ -263,7 +262,6 @@ nvd7_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvd7_grctx_generate_main,
-	.mods  = nvd7_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = nvd7_grctx_pack_hub,
 	.gpc   = nvd7_grctx_pack_gpc,
@@ -272,4 +270,13 @@ nvd7_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = nvd7_grctx_pack_ppc,
 	.icmd  = nvd9_grctx_pack_icmd,
 	.mthd  = nvd9_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x7ff,
+	.alpha_nr = 0x324,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
index c665fb7e4660..b9a301b6fd9f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
@@ -511,7 +511,6 @@ nvd9_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nvc0_grctx_generate_main,
-	.mods  = nvc1_grctx_generate_mods,
 	.unkn  = nvc1_grctx_generate_unkn,
 	.hub   = nvd9_grctx_pack_hub,
 	.gpc   = nvd9_grctx_pack_gpc,
@@ -519,4 +518,13 @@ nvd9_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.tpc   = nvd9_grctx_pack_tpc,
 	.icmd  = nvd9_grctx_pack_icmd,
 	.mthd  = nvd9_grctx_pack_mthd,
+	.bundle = nvc0_grctx_generate_bundle,
+	.bundle_size = 0x1800,
+	.pagepool = nvc0_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvc1_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x324,
+	.alpha_nr = 0x218,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
index c5b249238587..ccac2ee1a1cb 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
@@ -839,47 +839,34 @@ nve4_grctx_pack_ppc[] = {
  ******************************************************************************/
 
 void
-nve4_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+nve4_grctx_generate_bundle(struct nvc0_grctx *info)
 {
-	u32 magic[GPC_MAX][2];
-	u32 offset;
-	int gpc;
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000030,  0, 0);
-	mmio_list(0x4064c8, 0x01800600,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-
-	mmio_list(0x405830, 0x02180648,  0, 0);
-	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
-		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * priv->tpc_nr[gpc];
-	}
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * priv->tpc_nr[gpc];
-	}
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 state_limit = min(impl->bundle_min_gpm_fifo_depth,
+				    impl->bundle_size / 0x20);
+	const u32 token_limit = impl->bundle_token_limit;
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->bundle_size, (1 << s), access);
+	mmio_refn(info, 0x408004, 0x00000000, s, b);
+	mmio_refn(info, 0x408008, 0x80000000 | (impl->bundle_size >> s), 0, b);
+	mmio_refn(info, 0x418808, 0x00000000, s, b);
+	mmio_refn(info, 0x41880c, 0x80000000 | (impl->bundle_size >> s), 0, b);
+	mmio_wr32(info, 0x4064c8, (state_limit << 16) | token_limit);
+}
 
-	mmio_list(0x17e91c, 0x06060609, 0, 0);
-	mmio_list(0x17e920, 0x00090a05, 0, 0);
+void
+nve4_grctx_generate_pagepool(struct nvc0_grctx *info)
+{
+	const struct nvc0_grctx_oclass *impl = nvc0_grctx_impl(info->priv);
+	const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS;
+	const int s = 8;
+	const int b = mmio_vram(info, impl->pagepool_size, (1 << s), access);
+	mmio_refn(info, 0x40800c, 0x00000000, s, b);
+	mmio_wr32(info, 0x408010, 0x80000000);
+	mmio_refn(info, 0x419004, 0x00000000, s, b);
+	mmio_wr32(info, 0x419008, 0x00000000);
+	mmio_wr32(info, 0x4064cc, 0x80000000);
 }
 
 void
@@ -957,7 +944,7 @@ nve4_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
 	int i;
 
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 0);
 
 	nvc0_graph_mmio(priv, oclass->hub);
 	nvc0_graph_mmio(priv, oclass->gpc);
@@ -967,7 +954,9 @@ nve4_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 
 	nv_wr32(priv, 0x404154, 0x00000000);
 
-	oclass->mods(priv, info);
+	oclass->bundle(info);
+	oclass->pagepool(info);
+	oclass->attrib(info);
 	oclass->unkn(priv);
 
 	nvc0_grctx_generate_tpcid(priv);
@@ -991,7 +980,7 @@ nve4_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 	nvc0_graph_icmd(priv, oclass->icmd);
 	nv_wr32(priv, 0x404154, 0x00000400);
 	nvc0_graph_mthd(priv, oclass->mthd);
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 1);
 
 	nv_mask(priv, 0x418800, 0x00200000, 0x00200000);
 	nv_mask(priv, 0x41be10, 0x00800000, 0x00800000);
@@ -1009,7 +998,6 @@ nve4_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nve4_grctx_generate_main,
-	.mods  = nve4_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = nve4_grctx_pack_hub,
 	.gpc   = nve4_grctx_pack_gpc,
@@ -1018,4 +1006,15 @@ nve4_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = nve4_grctx_pack_ppc,
 	.icmd  = nve4_grctx_pack_icmd,
 	.mthd  = nve4_grctx_pack_mthd,
+	.bundle = nve4_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0x180,
+	.bundle_token_limit = 0x600,
+	.pagepool = nve4_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x7ff,
+	.alpha_nr = 0x648,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
index dec03f04114d..e9b0dcf95a49 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
@@ -279,7 +279,7 @@ nvf0_grctx_init_icmd_0[] = {
 	{}
 };
 
-static const struct nvc0_graph_pack
+const struct nvc0_graph_pack
 nvf0_grctx_pack_icmd[] = {
 	{ nvf0_grctx_init_icmd_0 },
 	{}
@@ -668,7 +668,7 @@ nvf0_grctx_init_be_0[] = {
 	{}
 };
 
-static const struct nvc0_graph_pack
+const struct nvc0_graph_pack
 nvf0_grctx_pack_hub[] = {
 	{ nvc0_grctx_init_main_0 },
 	{ nvf0_grctx_init_fe_0 },
@@ -704,7 +704,7 @@ nvf0_grctx_init_gpc_unk_2[] = {
 	{}
 };
 
-static const struct nvc0_graph_pack
+const struct nvc0_graph_pack
 nvf0_grctx_pack_gpc[] = {
 	{ nvc0_grctx_init_gpc_unk_0 },
 	{ nvd9_grctx_init_prop_0 },
@@ -718,7 +718,7 @@ nvf0_grctx_pack_gpc[] = {
 	{}
 };
 
-static const struct nvc0_graph_init
+const struct nvc0_graph_init
 nvf0_grctx_init_tex_0[] = {
 	{ 0x419a00,   1, 0x04, 0x000000f0 },
 	{ 0x419a04,   1, 0x04, 0x00000001 },
@@ -797,7 +797,7 @@ nvf0_grctx_init_cbm_0[] = {
 	{}
 };
 
-static const struct nvc0_graph_pack
+const struct nvc0_graph_pack
 nvf0_grctx_pack_ppc[] = {
 	{ nve4_grctx_init_pes_0 },
 	{ nvf0_grctx_init_cbm_0 },
@@ -809,58 +809,6 @@ nvf0_grctx_pack_ppc[] = {
  * PGRAPH context implementation
  ******************************************************************************/
 
-static void
-nvf0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
-{
-	u32 magic[GPC_MAX][4];
-	u32 offset;
-	int gpc;
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000030,  0, 0);
-	mmio_list(0x4064c8, 0x01800600,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-
-	mmio_list(0x405830, 0x02180648,  0, 0);
-	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
-
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * (priv->tpc_nr[gpc] - 1);
-		u16 magic1 = 0x0648 * (priv->tpc_nr[gpc] - 1);
-		u16 magic2 = 0x0218;
-		u16 magic3 = 0x0648;
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * (priv->tpc_nr[gpc] - 1);
-		magic[gpc][2]  = 0x10000000 | (magic2 << 16) | offset;
-		magic[gpc][3]  = 0x00000000 | (magic3 << 16);
-		offset += 0x0324;
-	}
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * (priv->tpc_nr[gpc] - 1);
-		mmio_list(GPC_UNIT(gpc, 0x32c0), magic[gpc][2], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x32e4), magic[gpc][3] | offset, 0, 0);
-		offset += 0x07ff;
-	}
-
-	mmio_list(0x17e91c, 0x06060609, 0, 0);
-	mmio_list(0x17e920, 0x00090a05, 0, 0);
-}
-
 struct nouveau_oclass *
 nvf0_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.base.handle = NV_ENGCTX(GR, 0xf0),
@@ -873,7 +821,6 @@ nvf0_grctx_oclass = &(struct nvc0_grctx_oclass) {
 		.wr32 = _nouveau_graph_context_wr32,
 	},
 	.main  = nve4_grctx_generate_main,
-	.mods  = nvf0_grctx_generate_mods,
 	.unkn  = nve4_grctx_generate_unkn,
 	.hub   = nvf0_grctx_pack_hub,
 	.gpc   = nvf0_grctx_pack_gpc,
@@ -882,4 +829,15 @@ nvf0_grctx_oclass = &(struct nvc0_grctx_oclass) {
 	.ppc   = nvf0_grctx_pack_ppc,
 	.icmd  = nvf0_grctx_pack_icmd,
 	.mthd  = nvf0_grctx_pack_mthd,
+	.bundle = nve4_grctx_generate_bundle,
+	.bundle_size = 0x3000,
+	.bundle_min_gpm_fifo_depth = 0x180,
+	.bundle_token_limit = 0x7c0,
+	.pagepool = nve4_grctx_generate_pagepool,
+	.pagepool_size = 0x8000,
+	.attrib = nvd7_grctx_generate_attrib,
+	.attrib_nr_max = 0x324,
+	.attrib_nr = 0x218,
+	.alpha_nr_max = 0x7ff,
+	.alpha_nr = 0x648,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/gk110b.c b/drivers/gpu/drm/nouveau/core/engine/graph/gk110b.c
new file mode 100644
index 000000000000..d07b19dc168d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/gk110b.c
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+#include "ctxnvc0.h"
+
+/*******************************************************************************
+ * PGRAPH register lists
+ ******************************************************************************/
+
+static const struct nvc0_graph_init
+gk110b_graph_init_l1c_0[] = {
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x09000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00b08bea },
+	{ 0x419c84,   1, 0x04, 0x00010384 },
+	{ 0x419cbc,   1, 0x04, 0x281b3646 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c80,   1, 0x04, 0x00020230 },
+	{ 0x419ccc,   2, 0x04, 0x00000000 },
+	{}
+};
+
+static const struct nvc0_graph_init
+gk110b_graph_init_sm_0[] = {
+	{ 0x419e00,   1, 0x04, 0x00000080 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ee4,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00000000 },
+	{ 0x419eb4,   1, 0x04, 0x00000000 },
+	{ 0x419ebc,   2, 0x04, 0x00000000 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419ed0,   1, 0x04, 0x00002616 },
+	{ 0x419f74,   1, 0x04, 0x00015555 },
+	{ 0x419f80,   4, 0x04, 0x00000000 },
+	{}
+};
+
+static const struct nvc0_graph_pack
+gk110b_graph_pack_mmio[] = {
+	{ nve4_graph_init_main_0 },
+	{ nvf0_graph_init_fe_0 },
+	{ nvc0_graph_init_pri_0 },
+	{ nvc0_graph_init_rstr2d_0 },
+	{ nvd9_graph_init_pd_0 },
+	{ nvf0_graph_init_ds_0 },
+	{ nvc0_graph_init_scc_0 },
+	{ nvf0_graph_init_sked_0 },
+	{ nvf0_graph_init_cwd_0 },
+	{ nvd9_graph_init_prop_0 },
+	{ nvc1_graph_init_gpc_unk_0 },
+	{ nvc0_graph_init_setup_0 },
+	{ nvc0_graph_init_crstr_0 },
+	{ nvc1_graph_init_setup_1 },
+	{ nvc0_graph_init_zcull_0 },
+	{ nvd9_graph_init_gpm_0 },
+	{ nvf0_graph_init_gpc_unk_1 },
+	{ nvc0_graph_init_gcc_0 },
+	{ nve4_graph_init_tpccs_0 },
+	{ nvf0_graph_init_tex_0 },
+	{ nve4_graph_init_pe_0 },
+	{ gk110b_graph_init_l1c_0 },
+	{ nvc0_graph_init_mpc_0 },
+	{ gk110b_graph_init_sm_0 },
+	{ nvd7_graph_init_pes_0 },
+	{ nvd7_graph_init_wwdx_0 },
+	{ nvd7_graph_init_cbm_0 },
+	{ nve4_graph_init_be_0 },
+	{ nvc0_graph_init_fe_1 },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nouveau_oclass *
+gk110b_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xf1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nve4_graph_init,
+		.fini = nvf0_graph_fini,
+	},
+	.cclass = &gk110b_grctx_oclass,
+	.sclass =  nvf0_graph_sclass,
+	.mmio = gk110b_graph_pack_mmio,
+	.fecs.ucode = &nvf0_graph_fecs_ucode,
+	.gpccs.ucode = &nvf0_graph_gpccs_ucode,
+	.ppc_nr = 2,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/gk20a.c b/drivers/gpu/drm/nouveau/core/engine/graph/gk20a.c
index 83048a56430d..7d0abe9f3fe7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/gk20a.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/gk20a.c
@@ -27,8 +27,8 @@ static struct nouveau_oclass
 gk20a_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0xa040, &nouveau_object_ofuncs },
-	{ 0xa297, &nouveau_object_ofuncs },
-	{ 0xa0c0, &nouveau_object_ofuncs },
+	{ KEPLER_C, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ KEPLER_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
@@ -39,9 +39,10 @@ gk20a_graph_oclass = &(struct nvc0_graph_oclass) {
 		.ctor = nvc0_graph_ctor,
 		.dtor = nvc0_graph_dtor,
 		.init = nve4_graph_init,
-		.fini = nve4_graph_fini,
+		.fini = _nouveau_graph_fini,
 	},
 	.cclass = &gk20a_grctx_oclass,
 	.sclass = gk20a_graph_sclass,
 	.mmio = nve4_graph_pack_mmio,
+	.ppc_nr = 1,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/gm107.c b/drivers/gpu/drm/nouveau/core/engine/graph/gm107.c
index 21c5f31d607f..4bdbdab2fd9a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/gm107.c
@@ -36,8 +36,8 @@ static struct nouveau_oclass
 gm107_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0xa140, &nouveau_object_ofuncs },
-	{ 0xb097, &nouveau_object_ofuncs },
-	{ 0xb0c0, &nouveau_object_ofuncs },
+	{ MAXWELL_A, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ MAXWELL_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
@@ -425,6 +425,9 @@ gm107_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400134, 0xffffffff);
 
 	nv_wr32(priv, 0x400054, 0x2c350f63);
+
+	nvc0_graph_zbc_init(priv);
+
 	return nvc0_graph_init_ctxctl(priv);
 }
 
@@ -462,4 +465,5 @@ gm107_graph_oclass = &(struct nvc0_graph_oclass) {
 	.mmio = gm107_graph_pack_mmio,
 	.fecs.ucode = 0 ? &gm107_graph_fecs_ucode : NULL,
 	.gpccs.ucode = &gm107_graph_gpccs_ucode,
+	.ppc_nr = 2,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
index ad13dcdd15f9..f70e2f67a4dd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv04.c
@@ -24,7 +24,6 @@
 
 #include <core/client.h>
 #include <core/os.h>
-#include <core/class.h>
 #include <core/handle.h>
 #include <core/namedb.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
index 4532f7e5618c..2b12b09683c8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv10.c
@@ -24,7 +24,6 @@
 
 #include <core/client.h>
 #include <core/os.h>
-#include <core/class.h>
 #include <core/handle.h>
 
 #include <subdev/fb.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv108.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv108.c
index 00ea1a089822..2b0e8f48c029 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv108.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv108.c
@@ -33,7 +33,7 @@ static struct nouveau_oclass
 nv108_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0xa140, &nouveau_object_ofuncs },
-	{ 0xa197, &nouveau_object_ofuncs },
+	{ KEPLER_B, &nvc0_fermi_ofuncs },
 	{ 0xa1c0, &nouveau_object_ofuncs },
 	{}
 };
@@ -220,4 +220,5 @@ nv108_graph_oclass = &(struct nvc0_graph_oclass) {
 	.mmio = nv108_graph_pack_mmio,
 	.fecs.ucode = &nv108_graph_fecs_ucode,
 	.gpccs.ucode = &nv108_graph_gpccs_ucode,
+	.ppc_nr = 1,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
index d145e080899a..ceb9c746d94e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv20.c
@@ -1,6 +1,5 @@
 #include <core/client.h>
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/handle.h>
 #include <core/enum.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
index 7a80d005a974..f8a6fdd7d5e8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv25.c
@@ -1,5 +1,4 @@
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/enum.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
index 3e1f32ee43d4..5de9caa2ef67 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv2a.c
@@ -1,5 +1,4 @@
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/enum.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
index e451db32e92a..2f9dbc709389 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv30.c
@@ -1,5 +1,4 @@
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/enum.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
index 9385ac7b44a4..34dd26c70b64 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv34.c
@@ -1,5 +1,4 @@
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/enum.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
index 9ce84b73f86a..2fb5756d9f66 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv35.c
@@ -1,5 +1,4 @@
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/enum.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
index 6477fbf6a550..4f401174868d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv40.c
@@ -24,7 +24,6 @@
 
 #include <core/client.h>
 #include <core/os.h>
-#include <core/class.h>
 #include <core/handle.h>
 #include <core/engctx.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index 20665c21d80e..38e0aa26f1cd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/client.h>
 #include <core/handle.h>
 #include <core/engctx.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index aa0838916354..30fd1dc64f93 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -26,15 +26,232 @@
 #include "ctxnvc0.h"
 
 /*******************************************************************************
+ * Zero Bandwidth Clear
+ ******************************************************************************/
+
+static void
+nvc0_graph_zbc_clear_color(struct nvc0_graph_priv *priv, int zbc)
+{
+	if (priv->zbc_color[zbc].format) {
+		nv_wr32(priv, 0x405804, priv->zbc_color[zbc].ds[0]);
+		nv_wr32(priv, 0x405808, priv->zbc_color[zbc].ds[1]);
+		nv_wr32(priv, 0x40580c, priv->zbc_color[zbc].ds[2]);
+		nv_wr32(priv, 0x405810, priv->zbc_color[zbc].ds[3]);
+	}
+	nv_wr32(priv, 0x405814, priv->zbc_color[zbc].format);
+	nv_wr32(priv, 0x405820, zbc);
+	nv_wr32(priv, 0x405824, 0x00000004); /* TRIGGER | WRITE | COLOR */
+}
+
+static int
+nvc0_graph_zbc_color_get(struct nvc0_graph_priv *priv, int format,
+			 const u32 ds[4], const u32 l2[4])
+{
+	struct nouveau_ltc *ltc = nouveau_ltc(priv);
+	int zbc = -ENOSPC, i;
+
+	for (i = ltc->zbc_min; i <= ltc->zbc_max; i++) {
+		if (priv->zbc_color[i].format) {
+			if (priv->zbc_color[i].format != format)
+				continue;
+			if (memcmp(priv->zbc_color[i].ds, ds, sizeof(
+				   priv->zbc_color[i].ds)))
+				continue;
+			if (memcmp(priv->zbc_color[i].l2, l2, sizeof(
+				   priv->zbc_color[i].l2))) {
+				WARN_ON(1);
+				return -EINVAL;
+			}
+			return i;
+		} else {
+			zbc = (zbc < 0) ? i : zbc;
+		}
+	}
+
+	if (zbc < 0)
+		return zbc;
+
+	memcpy(priv->zbc_color[zbc].ds, ds, sizeof(priv->zbc_color[zbc].ds));
+	memcpy(priv->zbc_color[zbc].l2, l2, sizeof(priv->zbc_color[zbc].l2));
+	priv->zbc_color[zbc].format = format;
+	ltc->zbc_color_get(ltc, zbc, l2);
+	nvc0_graph_zbc_clear_color(priv, zbc);
+	return zbc;
+}
+
+static void
+nvc0_graph_zbc_clear_depth(struct nvc0_graph_priv *priv, int zbc)
+{
+	if (priv->zbc_depth[zbc].format)
+		nv_wr32(priv, 0x405818, priv->zbc_depth[zbc].ds);
+	nv_wr32(priv, 0x40581c, priv->zbc_depth[zbc].format);
+	nv_wr32(priv, 0x405820, zbc);
+	nv_wr32(priv, 0x405824, 0x00000005); /* TRIGGER | WRITE | DEPTH */
+}
+
+static int
+nvc0_graph_zbc_depth_get(struct nvc0_graph_priv *priv, int format,
+			 const u32 ds, const u32 l2)
+{
+	struct nouveau_ltc *ltc = nouveau_ltc(priv);
+	int zbc = -ENOSPC, i;
+
+	for (i = ltc->zbc_min; i <= ltc->zbc_max; i++) {
+		if (priv->zbc_depth[i].format) {
+			if (priv->zbc_depth[i].format != format)
+				continue;
+			if (priv->zbc_depth[i].ds != ds)
+				continue;
+			if (priv->zbc_depth[i].l2 != l2) {
+				WARN_ON(1);
+				return -EINVAL;
+			}
+			return i;
+		} else {
+			zbc = (zbc < 0) ? i : zbc;
+		}
+	}
+
+	if (zbc < 0)
+		return zbc;
+
+	priv->zbc_depth[zbc].format = format;
+	priv->zbc_depth[zbc].ds = ds;
+	priv->zbc_depth[zbc].l2 = l2;
+	ltc->zbc_depth_get(ltc, zbc, l2);
+	nvc0_graph_zbc_clear_depth(priv, zbc);
+	return zbc;
+}
+
+/*******************************************************************************
  * Graphics object classes
  ******************************************************************************/
 
+static int
+nvc0_fermi_mthd_zbc_color(struct nouveau_object *object, void *data, u32 size)
+{
+	struct nvc0_graph_priv *priv = (void *)object->engine;
+	union {
+		struct fermi_a_zbc_color_v0 v0;
+	} *args = data;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		switch (args->v0.format) {
+		case FERMI_A_ZBC_COLOR_V0_FMT_ZERO:
+		case FERMI_A_ZBC_COLOR_V0_FMT_UNORM_ONE:
+		case FERMI_A_ZBC_COLOR_V0_FMT_RF32_GF32_BF32_AF32:
+		case FERMI_A_ZBC_COLOR_V0_FMT_R16_G16_B16_A16:
+		case FERMI_A_ZBC_COLOR_V0_FMT_RN16_GN16_BN16_AN16:
+		case FERMI_A_ZBC_COLOR_V0_FMT_RS16_GS16_BS16_AS16:
+		case FERMI_A_ZBC_COLOR_V0_FMT_RU16_GU16_BU16_AU16:
+		case FERMI_A_ZBC_COLOR_V0_FMT_RF16_GF16_BF16_AF16:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A8R8G8B8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A8RL8GL8BL8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A2B10G10R10:
+		case FERMI_A_ZBC_COLOR_V0_FMT_AU2BU10GU10RU10:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A8B8G8R8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A8BL8GL8RL8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_AN8BN8GN8RN8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_AS8BS8GS8RS8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_AU8BU8GU8RU8:
+		case FERMI_A_ZBC_COLOR_V0_FMT_A2R10G10B10:
+		case FERMI_A_ZBC_COLOR_V0_FMT_BF10GF11RF11:
+			ret = nvc0_graph_zbc_color_get(priv, args->v0.format,
+							     args->v0.ds,
+							     args->v0.l2);
+			if (ret >= 0) {
+				args->v0.index = ret;
+				return 0;
+			}
+			break;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvc0_fermi_mthd_zbc_depth(struct nouveau_object *object, void *data, u32 size)
+{
+	struct nvc0_graph_priv *priv = (void *)object->engine;
+	union {
+		struct fermi_a_zbc_depth_v0 v0;
+	} *args = data;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		switch (args->v0.format) {
+		case FERMI_A_ZBC_DEPTH_V0_FMT_FP32:
+			ret = nvc0_graph_zbc_depth_get(priv, args->v0.format,
+							     args->v0.ds,
+							     args->v0.l2);
+			return (ret >= 0) ? 0 : -ENOSPC;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	return ret;
+}
+
+static int
+nvc0_fermi_mthd(struct nouveau_object *object, u32 mthd, void *data, u32 size)
+{
+	switch (mthd) {
+	case FERMI_A_ZBC_COLOR:
+		return nvc0_fermi_mthd_zbc_color(object, data, size);
+	case FERMI_A_ZBC_DEPTH:
+		return nvc0_fermi_mthd_zbc_depth(object, data, size);
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
+struct nouveau_ofuncs
+nvc0_fermi_ofuncs = {
+	.ctor = _nouveau_object_ctor,
+	.dtor = nouveau_object_destroy,
+	.init = nouveau_object_init,
+	.fini = nouveau_object_fini,
+	.mthd = nvc0_fermi_mthd,
+};
+
+static int
+nvc0_graph_set_shader_exceptions(struct nouveau_object *object, u32 mthd,
+				 void *pdata, u32 size)
+{
+	struct nvc0_graph_priv *priv = (void *)nv_engine(object);
+	if (size >= sizeof(u32)) {
+		u32 data = *(u32 *)pdata ? 0xffffffff : 0x00000000;
+		nv_wr32(priv, 0x419e44, data);
+		nv_wr32(priv, 0x419e4c, data);
+		return 0;
+	}
+	return -EINVAL;
+}
+
+struct nouveau_omthds
+nvc0_graph_9097_omthds[] = {
+	{ 0x1528, 0x1528, nvc0_graph_set_shader_exceptions },
+	{}
+};
+
+struct nouveau_omthds
+nvc0_graph_90c0_omthds[] = {
+	{ 0x1528, 0x1528, nvc0_graph_set_shader_exceptions },
+	{}
+};
+
 struct nouveau_oclass
 nvc0_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
+	{ FERMI_A, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
@@ -98,7 +315,7 @@ nvc0_graph_context_ctor(struct nouveau_object *parent,
 		u32 addr = mmio->addr;
 		u32 data = mmio->data;
 
-		if (mmio->shift) {
+		if (mmio->buffer >= 0) {
 			u64 info = chan->data[mmio->buffer].vma.offset;
 			data |= info >> mmio->shift;
 		}
@@ -407,6 +624,35 @@ nvc0_graph_pack_mmio[] = {
  ******************************************************************************/
 
 void
+nvc0_graph_zbc_init(struct nvc0_graph_priv *priv)
+{
+	const u32  zero[] = { 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+			      0x00000000, 0x00000000, 0x00000000, 0x00000000 };
+	const u32   one[] = { 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000,
+			      0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff };
+	const u32 f32_0[] = { 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+			      0x00000000, 0x00000000, 0x00000000, 0x00000000 };
+	const u32 f32_1[] = { 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000,
+			      0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000 };
+	struct nouveau_ltc *ltc = nouveau_ltc(priv);
+	int index;
+
+	if (!priv->zbc_color[0].format) {
+		nvc0_graph_zbc_color_get(priv, 1,  & zero[0],   &zero[4]);
+		nvc0_graph_zbc_color_get(priv, 2,  &  one[0],    &one[4]);
+		nvc0_graph_zbc_color_get(priv, 4,  &f32_0[0],  &f32_0[4]);
+		nvc0_graph_zbc_color_get(priv, 4,  &f32_1[0],  &f32_1[4]);
+		nvc0_graph_zbc_depth_get(priv, 1, 0x00000000, 0x00000000);
+		nvc0_graph_zbc_depth_get(priv, 1, 0x3f800000, 0x3f800000);
+	}
+
+	for (index = ltc->zbc_min; index <= ltc->zbc_max; index++)
+		nvc0_graph_zbc_clear_color(priv, index);
+	for (index = ltc->zbc_min; index <= ltc->zbc_max; index++)
+		nvc0_graph_zbc_clear_depth(priv, index);
+}
+
+void
 nvc0_graph_mmio(struct nvc0_graph_priv *priv, const struct nvc0_graph_pack *p)
 {
 	const struct nvc0_graph_pack *pack;
@@ -969,17 +1215,16 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 {
 	struct nvc0_graph_oclass *oclass = (void *)nv_object(priv)->oclass;
 	struct nvc0_grctx_oclass *cclass = (void *)nv_engine(priv)->cclass;
-	u32 r000260;
 	int i;
 
 	if (priv->firmware) {
 		/* load fuc microcode */
-		r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+		nouveau_mc(priv)->unk260(nouveau_mc(priv), 0);
 		nvc0_graph_init_fw(priv, 0x409000, &priv->fuc409c,
 						   &priv->fuc409d);
 		nvc0_graph_init_fw(priv, 0x41a000, &priv->fuc41ac,
 						   &priv->fuc41ad);
-		nv_wr32(priv, 0x000260, r000260);
+		nouveau_mc(priv)->unk260(nouveau_mc(priv), 1);
 
 		/* start both of them running */
 		nv_wr32(priv, 0x409840, 0xffffffff);
@@ -1066,7 +1311,7 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 	}
 
 	/* load HUB microcode */
-	r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 0);
 	nv_wr32(priv, 0x4091c0, 0x01000000);
 	for (i = 0; i < oclass->fecs.ucode->data.size / 4; i++)
 		nv_wr32(priv, 0x4091c4, oclass->fecs.ucode->data.data[i]);
@@ -1089,7 +1334,7 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 			nv_wr32(priv, 0x41a188, i >> 6);
 		nv_wr32(priv, 0x41a184, oclass->gpccs.ucode->code.data[i]);
 	}
-	nv_wr32(priv, 0x000260, r000260);
+	nouveau_mc(priv)->unk260(nouveau_mc(priv), 1);
 
 	/* load register lists */
 	nvc0_graph_init_csdata(priv, cclass->hub, 0x409000, 0x000, 0x000000);
@@ -1224,6 +1469,9 @@ nvc0_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400134, 0xffffffff);
 
 	nv_wr32(priv, 0x400054, 0x34ce3464);
+
+	nvc0_graph_zbc_init(priv);
+
 	return nvc0_graph_init_ctxctl(priv);
 }
 
@@ -1287,7 +1535,7 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nouveau_device *device = nv_device(parent);
 	struct nvc0_graph_priv *priv;
 	bool use_ext_fw, enable;
-	int ret, i;
+	int ret, i, j;
 
 	use_ext_fw = nouveau_boolopt(device->cfgopt, "NvGrUseFW",
 				     oclass->fecs.ucode == NULL);
@@ -1333,6 +1581,11 @@ nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	for (i = 0; i < priv->gpc_nr; i++) {
 		priv->tpc_nr[i]  = nv_rd32(priv, GPC_UNIT(i, 0x2608));
 		priv->tpc_total += priv->tpc_nr[i];
+		priv->ppc_nr[i]  = oclass->ppc_nr;
+		for (j = 0; j < priv->ppc_nr[i]; j++) {
+			u8 mask = nv_rd32(priv, GPC_UNIT(i, 0x0c30 + (j * 4)));
+			priv->ppc_tpc_nr[i][j] = hweight8(mask);
+		}
 	}
 
 	/*XXX: these need figuring out... though it might not even matter */
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
index ffc289198dd8..7ed9e89c3435 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
@@ -30,10 +30,15 @@
 #include <core/gpuobj.h>
 #include <core/option.h>
 
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+
 #include <subdev/fb.h>
 #include <subdev/vm.h>
 #include <subdev/bar.h>
 #include <subdev/timer.h>
+#include <subdev/mc.h>
+#include <subdev/ltc.h>
 
 #include <engine/fifo.h>
 #include <engine/graph.h>
@@ -60,7 +65,7 @@ struct nvc0_graph_mmio {
 	u32 addr;
 	u32 data;
 	u32 shift;
-	u32 buffer;
+	int buffer;
 };
 
 struct nvc0_graph_fuc {
@@ -68,6 +73,18 @@ struct nvc0_graph_fuc {
 	u32  size;
 };
 
+struct nvc0_graph_zbc_color {
+	u32 format;
+	u32 ds[4];
+	u32 l2[4];
+};
+
+struct nvc0_graph_zbc_depth {
+	u32 format;
+	u32 ds;
+	u32 l2;
+};
+
 struct nvc0_graph_priv {
 	struct nouveau_graph base;
 
@@ -77,10 +94,15 @@ struct nvc0_graph_priv {
 	struct nvc0_graph_fuc fuc41ad;
 	bool firmware;
 
+	struct nvc0_graph_zbc_color zbc_color[NOUVEAU_LTC_MAX_ZBC_CNT];
+	struct nvc0_graph_zbc_depth zbc_depth[NOUVEAU_LTC_MAX_ZBC_CNT];
+
 	u8 rop_nr;
 	u8 gpc_nr;
 	u8 tpc_nr[GPC_MAX];
 	u8 tpc_total;
+	u8 ppc_nr[GPC_MAX];
+	u8 ppc_tpc_nr[GPC_MAX][4];
 
 	struct nouveau_gpuobj *unk4188b4;
 	struct nouveau_gpuobj *unk4188b8;
@@ -118,12 +140,20 @@ int  nvc0_graph_ctor(struct nouveau_object *, struct nouveau_object *,
 		     struct nouveau_object **);
 void nvc0_graph_dtor(struct nouveau_object *);
 int  nvc0_graph_init(struct nouveau_object *);
+void nvc0_graph_zbc_init(struct nvc0_graph_priv *);
+
 int  nve4_graph_fini(struct nouveau_object *, bool);
 int  nve4_graph_init(struct nouveau_object *);
 
-extern struct nouveau_oclass nvc0_graph_sclass[];
+int  nvf0_graph_fini(struct nouveau_object *, bool);
+
+extern struct nouveau_ofuncs nvc0_fermi_ofuncs;
 
+extern struct nouveau_oclass nvc0_graph_sclass[];
+extern struct nouveau_omthds nvc0_graph_9097_omthds[];
+extern struct nouveau_omthds nvc0_graph_90c0_omthds[];
 extern struct nouveau_oclass nvc8_graph_sclass[];
+extern struct nouveau_oclass nvf0_graph_sclass[];
 
 struct nvc0_graph_init {
 	u32 addr;
@@ -149,6 +179,9 @@ struct nvc0_graph_ucode {
 extern struct nvc0_graph_ucode nvc0_graph_fecs_ucode;
 extern struct nvc0_graph_ucode nvc0_graph_gpccs_ucode;
 
+extern struct nvc0_graph_ucode nvf0_graph_fecs_ucode;
+extern struct nvc0_graph_ucode nvf0_graph_gpccs_ucode;
+
 struct nvc0_graph_oclass {
 	struct nouveau_oclass base;
 	struct nouveau_oclass **cclass;
@@ -160,6 +193,7 @@ struct nvc0_graph_oclass {
 	struct {
 		struct nvc0_graph_ucode *ucode;
 	} gpccs;
+	int ppc_nr;
 };
 
 void nvc0_graph_mmio(struct nvc0_graph_priv *, const struct nvc0_graph_pack *);
@@ -223,9 +257,11 @@ extern const struct nvc0_graph_init nve4_graph_init_be_0[];
 extern const struct nvc0_graph_pack nve4_graph_pack_mmio[];
 
 extern const struct nvc0_graph_init nvf0_graph_init_fe_0[];
+extern const struct nvc0_graph_init nvf0_graph_init_ds_0[];
 extern const struct nvc0_graph_init nvf0_graph_init_sked_0[];
 extern const struct nvc0_graph_init nvf0_graph_init_cwd_0[];
 extern const struct nvc0_graph_init nvf0_graph_init_gpc_unk_1[];
+extern const struct nvc0_graph_init nvf0_graph_init_tex_0[];
 extern const struct nvc0_graph_init nvf0_graph_init_sm_0[];
 
 extern const struct nvc0_graph_init nv108_graph_init_gpc_unk_0[];
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
index 30cab0b2eba1..93d58e5b82c2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
@@ -33,9 +33,9 @@ static struct nouveau_oclass
 nvc1_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
+	{ FERMI_A, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_B, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
index a6bf783e1256..692e1eda0eb4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
@@ -33,10 +33,10 @@ struct nouveau_oclass
 nvc8_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
-	{ 0x9297, &nouveau_object_ofuncs },
+	{ FERMI_A, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_B, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_C, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ FERMI_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
index 2a6a94e2a041..41e8445c7eea 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
@@ -133,4 +133,5 @@ nvd7_graph_oclass = &(struct nvc0_graph_oclass) {
 	.mmio = nvd7_graph_pack_mmio,
 	.fecs.ucode = &nvd7_graph_fecs_ucode,
 	.gpccs.ucode = &nvd7_graph_gpccs_ucode,
+	.ppc_nr = 1,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
index 51e0c075ad34..0c71f5c67ae0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
@@ -22,6 +22,8 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
+#include <subdev/pwr.h>
+
 #include "nvc0.h"
 #include "ctxnvc0.h"
 
@@ -33,8 +35,8 @@ static struct nouveau_oclass
 nve4_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0xa040, &nouveau_object_ofuncs },
-	{ 0xa097, &nouveau_object_ofuncs },
-	{ 0xa0c0, &nouveau_object_ofuncs },
+	{ KEPLER_A, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ KEPLER_COMPUTE_A, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
@@ -190,39 +192,20 @@ nve4_graph_pack_mmio[] = {
  ******************************************************************************/
 
 int
-nve4_graph_fini(struct nouveau_object *object, bool suspend)
-{
-	struct nvc0_graph_priv *priv = (void *)object;
-
-	/*XXX: this is a nasty hack to power on gr on certain boards
-	 *     where it's disabled by therm, somehow.  ideally it'd
-	 *     be nice to know when we should be doing this, and why,
-	 *     but, it's yet to be determined.  for now we test for
-	 *     the particular mmio error that occurs in the situation,
-	 *     and then bash therm in the way nvidia do.
-	 */
-	nv_mask(priv, 0x000200, 0x08001000, 0x08001000);
-	nv_rd32(priv, 0x000200);
-	if (nv_rd32(priv, 0x400700) == 0xbadf1000) {
-		nv_mask(priv, 0x000200, 0x08001000, 0x00000000);
-		nv_rd32(priv, 0x000200);
-		nv_mask(priv, 0x020004, 0xc0000000, 0x40000000);
-	}
-
-	return nouveau_graph_fini(&priv->base, suspend);
-}
-
-int
 nve4_graph_init(struct nouveau_object *object)
 {
 	struct nvc0_graph_oclass *oclass = (void *)object->oclass;
 	struct nvc0_graph_priv *priv = (void *)object;
+	struct nouveau_pwr *ppwr = nouveau_pwr(priv);
 	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
 	u32 data[TPC_MAX / 8] = {};
 	u8  tpcnr[GPC_MAX];
 	int gpc, tpc, rop;
 	int ret, i;
 
+	if (ppwr)
+		ppwr->pgob(ppwr, false);
+
 	ret = nouveau_graph_init(&priv->base);
 	if (ret)
 		return ret;
@@ -320,6 +303,9 @@ nve4_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400134, 0xffffffff);
 
 	nv_wr32(priv, 0x400054, 0x34ce3464);
+
+	nvc0_graph_zbc_init(priv);
+
 	return nvc0_graph_init_ctxctl(priv);
 }
 
@@ -350,11 +336,12 @@ nve4_graph_oclass = &(struct nvc0_graph_oclass) {
 		.ctor = nvc0_graph_ctor,
 		.dtor = nvc0_graph_dtor,
 		.init = nve4_graph_init,
-		.fini = nve4_graph_fini,
+		.fini = _nouveau_graph_fini,
 	},
 	.cclass = &nve4_grctx_oclass,
 	.sclass = nve4_graph_sclass,
 	.mmio = nve4_graph_pack_mmio,
 	.fecs.ucode = &nve4_graph_fecs_ucode,
 	.gpccs.ucode = &nve4_graph_gpccs_ucode,
+	.ppc_nr = 1,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
index c96762122b9b..c306c0f2fc84 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
@@ -29,12 +29,12 @@
  * Graphics object classes
  ******************************************************************************/
 
-static struct nouveau_oclass
+struct nouveau_oclass
 nvf0_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0xa140, &nouveau_object_ofuncs },
-	{ 0xa197, &nouveau_object_ofuncs },
-	{ 0xa1c0, &nouveau_object_ofuncs },
+	{ KEPLER_B, &nvc0_fermi_ofuncs, nvc0_graph_9097_omthds },
+	{ KEPLER_COMPUTE_B, &nouveau_object_ofuncs, nvc0_graph_90c0_omthds },
 	{}
 };
 
@@ -50,7 +50,7 @@ nvf0_graph_init_fe_0[] = {
 	{}
 };
 
-static const struct nvc0_graph_init
+const struct nvc0_graph_init
 nvf0_graph_init_ds_0[] = {
 	{ 0x405844,   1, 0x04, 0x00ffffff },
 	{ 0x405850,   1, 0x04, 0x00000000 },
@@ -88,7 +88,7 @@ nvf0_graph_init_gpc_unk_1[] = {
 	{}
 };
 
-static const struct nvc0_graph_init
+const struct nvc0_graph_init
 nvf0_graph_init_tex_0[] = {
 	{ 0x419ab0,   1, 0x04, 0x00000000 },
 	{ 0x419ac8,   1, 0x04, 0x00000000 },
@@ -170,7 +170,7 @@ nvf0_graph_pack_mmio[] = {
  * PGRAPH engine/subdev functions
  ******************************************************************************/
 
-static int
+int
 nvf0_graph_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nvc0_graph_priv *priv = (void *)object;
@@ -209,7 +209,7 @@ nvf0_graph_fini(struct nouveau_object *object, bool suspend)
 
 #include "fuc/hubnvf0.fuc.h"
 
-static struct nvc0_graph_ucode
+struct nvc0_graph_ucode
 nvf0_graph_fecs_ucode = {
 	.code.data = nvf0_grhub_code,
 	.code.size = sizeof(nvf0_grhub_code),
@@ -219,7 +219,7 @@ nvf0_graph_fecs_ucode = {
 
 #include "fuc/gpcnvf0.fuc.h"
 
-static struct nvc0_graph_ucode
+struct nvc0_graph_ucode
 nvf0_graph_gpccs_ucode = {
 	.code.data = nvf0_grgpc_code,
 	.code.size = sizeof(nvf0_grgpc_code),
@@ -241,4 +241,5 @@ nvf0_graph_oclass = &(struct nvc0_graph_oclass) {
 	.mmio = nvf0_graph_pack_mmio,
 	.fecs.ucode = &nvf0_graph_fecs_ucode,
 	.gpccs.ucode = &nvf0_graph_gpccs_ucode,
+	.ppc_nr = 2,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
index 7eb6d94c84e2..d88c700b2f69 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv31.c
@@ -24,7 +24,6 @@
 
 #include <core/client.h>
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/handle.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv40.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv40.c
index d4e7ec0ba68c..bdb2f20ff7b1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv40.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <subdev/fb.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv44.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv44.c
index 3d8c2133e0e8..72c7f33fd29b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv44.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv44.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/client.h>
 #include <core/engctx.h>
 #include <core/handle.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
index 37a2bd9e8078..cae33f86b11a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <subdev/vm.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
index 96f5aa92677b..e9cc8b116a24 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <subdev/vm.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/perfmon/base.c b/drivers/gpu/drm/nouveau/core/engine/perfmon/base.c
index e9c5e51943ef..63013812f7c9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/perfmon/base.c
+++ b/drivers/gpu/drm/nouveau/core/engine/perfmon/base.c
@@ -22,8 +22,11 @@
  * Authors: Ben Skeggs
  */
 
+#include <core/client.h>
 #include <core/option.h>
-#include <core/class.h>
+#include <nvif/unpack.h>
+#include <nvif/class.h>
+#include <nvif/ioctl.h>
 
 #include <subdev/clock.h>
 
@@ -101,24 +104,28 @@ nouveau_perfsig_wrap(struct nouveau_perfmon *ppm, const char *name,
  * Perfmon object classes
  ******************************************************************************/
 static int
-nouveau_perfctr_query(struct nouveau_object *object, u32 mthd,
-		      void *data, u32 size)
+nouveau_perfctr_query(struct nouveau_object *object, void *data, u32 size)
 {
+	union {
+		struct nvif_perfctr_query_v0 v0;
+	} *args = data;
 	struct nouveau_device *device = nv_device(object);
 	struct nouveau_perfmon *ppm = (void *)object->engine;
 	struct nouveau_perfdom *dom = NULL, *chk;
-	struct nv_perfctr_query *args = data;
 	const bool all = nouveau_boolopt(device->cfgopt, "NvPmShowAll", false);
 	const bool raw = nouveau_boolopt(device->cfgopt, "NvPmUnnamed", all);
 	const char *name;
 	int tmp = 0, di, si;
-	char path[64];
-
-	if (size < sizeof(*args))
-		return -EINVAL;
+	int ret;
 
-	di = (args->iter & 0xff000000) >> 24;
-	si = (args->iter & 0x00ffffff) - 1;
+	nv_ioctl(object, "perfctr query size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "perfctr query vers %d iter %08x\n",
+			 args->v0.version, args->v0.iter);
+		di = (args->v0.iter & 0xff000000) >> 24;
+		si = (args->v0.iter & 0x00ffffff) - 1;
+	} else
+		return ret;
 
 	list_for_each_entry(chk, &ppm->domains, head) {
 		if (tmp++ == di) {
@@ -132,19 +139,17 @@ nouveau_perfctr_query(struct nouveau_object *object, u32 mthd,
 
 	if (si >= 0) {
 		if (raw || !(name = dom->signal[si].name)) {
-			snprintf(path, sizeof(path), "/%s/%02x", dom->name, si);
-			name = path;
+			snprintf(args->v0.name, sizeof(args->v0.name),
+				 "/%s/%02x", dom->name, si);
+		} else {
+			strncpy(args->v0.name, name, sizeof(args->v0.name));
 		}
-
-		if (args->name)
-			strncpy(args->name, name, args->size);
-		args->size = strlen(name) + 1;
 	}
 
 	do {
 		while (++si < dom->signal_nr) {
 			if (all || dom->signal[si].name) {
-				args->iter = (di << 24) | ++si;
+				args->v0.iter = (di << 24) | ++si;
 				return 0;
 			}
 		}
@@ -153,21 +158,26 @@ nouveau_perfctr_query(struct nouveau_object *object, u32 mthd,
 		dom = list_entry(dom->head.next, typeof(*dom), head);
 	} while (&dom->head != &ppm->domains);
 
-	args->iter = 0xffffffff;
+	args->v0.iter = 0xffffffff;
 	return 0;
 }
 
 static int
-nouveau_perfctr_sample(struct nouveau_object *object, u32 mthd,
-		       void *data, u32 size)
+nouveau_perfctr_sample(struct nouveau_object *object, void *data, u32 size)
 {
+	union {
+		struct nvif_perfctr_sample none;
+	} *args = data;
 	struct nouveau_perfmon *ppm = (void *)object->engine;
 	struct nouveau_perfctr *ctr, *tmp;
 	struct nouveau_perfdom *dom;
-	struct nv_perfctr_sample *args = data;
+	int ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(object, "perfctr sample size %d\n", size);
+	if (nvif_unvers(args->none)) {
+		nv_ioctl(object, "perfctr sample\n");
+	} else
+		return ret;
 	ppm->sequence++;
 
 	list_for_each_entry(dom, &ppm->domains, head) {
@@ -206,22 +216,45 @@ nouveau_perfctr_sample(struct nouveau_object *object, u32 mthd,
 }
 
 static int
-nouveau_perfctr_read(struct nouveau_object *object, u32 mthd,
-		     void *data, u32 size)
+nouveau_perfctr_read(struct nouveau_object *object, void *data, u32 size)
 {
+	union {
+		struct nvif_perfctr_read_v0 v0;
+	} *args = data;
 	struct nouveau_perfctr *ctr = (void *)object;
-	struct nv_perfctr_read *args = data;
+	int ret;
+
+	nv_ioctl(object, "perfctr read size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(object, "perfctr read vers %d\n", args->v0.version);
+	} else
+		return ret;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
 	if (!ctr->clk)
 		return -EAGAIN;
 
-	args->clk = ctr->clk;
-	args->ctr = ctr->ctr;
+	args->v0.clk = ctr->clk;
+	args->v0.ctr = ctr->ctr;
 	return 0;
 }
 
+static int
+nouveau_perfctr_mthd(struct nouveau_object *object, u32 mthd,
+		     void *data, u32 size)
+{
+	switch (mthd) {
+	case NVIF_PERFCTR_V0_QUERY:
+		return nouveau_perfctr_query(object, data, size);
+	case NVIF_PERFCTR_V0_SAMPLE:
+		return nouveau_perfctr_sample(object, data, size);
+	case NVIF_PERFCTR_V0_READ:
+		return nouveau_perfctr_read(object, data, size);
+	default:
+		break;
+	}
+	return -EINVAL;
+}
+
 static void
 nouveau_perfctr_dtor(struct nouveau_object *object)
 {
@@ -237,19 +270,27 @@ nouveau_perfctr_ctor(struct nouveau_object *parent,
 		     struct nouveau_oclass *oclass, void *data, u32 size,
 		     struct nouveau_object **pobject)
 {
+	union {
+		struct nvif_perfctr_v0 v0;
+	} *args = data;
 	struct nouveau_perfmon *ppm = (void *)engine;
 	struct nouveau_perfdom *dom = NULL;
 	struct nouveau_perfsig *sig[4] = {};
 	struct nouveau_perfctr *ctr;
-	struct nv_perfctr_class *args = data;
 	int ret, i;
 
-	if (size < sizeof(*args))
-		return -EINVAL;
+	nv_ioctl(parent, "create perfctr size %d\n", size);
+	if (nvif_unpack(args->v0, 0, 0, false)) {
+		nv_ioctl(parent, "create perfctr vers %d logic_op %04x\n",
+			 args->v0.version, args->v0.logic_op);
+	} else
+		return ret;
 
-	for (i = 0; i < ARRAY_SIZE(args->signal) && args->signal[i].name; i++) {
-		sig[i] = nouveau_perfsig_find(ppm, args->signal[i].name,
-					      args->signal[i].size, &dom);
+	for (i = 0; i < ARRAY_SIZE(args->v0.name) && args->v0.name[i][0]; i++) {
+		sig[i] = nouveau_perfsig_find(ppm, args->v0.name[i],
+					      strnlen(args->v0.name[i],
+					      sizeof(args->v0.name[i])),
+					      &dom);
 		if (!sig[i])
 			return -EINVAL;
 	}
@@ -260,7 +301,7 @@ nouveau_perfctr_ctor(struct nouveau_object *parent,
 		return ret;
 
 	ctr->slot = -1;
-	ctr->logic_op = args->logic_op;
+	ctr->logic_op = args->v0.logic_op;
 	ctr->signal[0] = sig[0];
 	ctr->signal[1] = sig[1];
 	ctr->signal[2] = sig[2];
@@ -276,21 +317,13 @@ nouveau_perfctr_ofuncs = {
 	.dtor = nouveau_perfctr_dtor,
 	.init = nouveau_object_init,
 	.fini = nouveau_object_fini,
-};
-
-static struct nouveau_omthds
-nouveau_perfctr_omthds[] = {
-	{ NV_PERFCTR_QUERY, NV_PERFCTR_QUERY, nouveau_perfctr_query },
-	{ NV_PERFCTR_SAMPLE, NV_PERFCTR_SAMPLE, nouveau_perfctr_sample },
-	{ NV_PERFCTR_READ, NV_PERFCTR_READ, nouveau_perfctr_read },
-	{}
+	.mthd = nouveau_perfctr_mthd,
 };
 
 struct nouveau_oclass
 nouveau_perfmon_sclass[] = {
-	{ .handle = NV_PERFCTR_CLASS,
+	{ .handle = NVIF_IOCTL_NEW_V0_PERFCTR,
 	  .ofuncs = &nouveau_perfctr_ofuncs,
-	  .omthds =  nouveau_perfctr_omthds,
 	},
 	{},
 };
@@ -303,6 +336,7 @@ nouveau_perfctx_dtor(struct nouveau_object *object)
 {
 	struct nouveau_perfmon *ppm = (void *)object->engine;
 	mutex_lock(&nv_subdev(ppm)->mutex);
+	nouveau_engctx_destroy(&ppm->context->base);
 	ppm->context = NULL;
 	mutex_unlock(&nv_subdev(ppm)->mutex);
 }
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv04.c b/drivers/gpu/drm/nouveau/core/engine/software/nv04.c
index c571758e4a27..64df15c7f051 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv04.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv10.c b/drivers/gpu/drm/nouveau/core/engine/software/nv10.c
index a62f11a78430..f54a2253deca 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv10.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 
 #include <engine/software.h>
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
index f3b4d9dbf23c..a0fec205f9db 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv50.c
@@ -23,12 +23,12 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/namedb.h>
 #include <core/handle.h>
 #include <core/gpuobj.h>
 #include <core/event.h>
+#include <nvif/event.h>
 
 #include <subdev/bar.h>
 
@@ -86,10 +86,10 @@ nv50_software_mthd_vblsem_release(struct nouveau_object *object, u32 mthd,
 {
 	struct nv50_software_chan *chan = (void *)nv_engctx(object->parent);
 	u32 head = *(u32 *)args;
-	if (head >= chan->vblank.nr_event)
+	if (head >= nouveau_disp(chan)->vblank.index_nr)
 		return -EINVAL;
 
-	nouveau_event_get(chan->vblank.event[head]);
+	nvkm_notify_get(&chan->vblank.notify[head]);
 	return 0;
 }
 
@@ -124,9 +124,10 @@ nv50_software_sclass[] = {
  ******************************************************************************/
 
 static int
-nv50_software_vblsem_release(void *data, u32 type, int head)
+nv50_software_vblsem_release(struct nvkm_notify *notify)
 {
-	struct nv50_software_chan *chan = data;
+	struct nv50_software_chan *chan =
+		container_of(notify, typeof(*chan), vblank.notify[notify->index]);
 	struct nv50_software_priv *priv = (void *)nv_object(chan)->engine;
 	struct nouveau_bar *bar = nouveau_bar(priv);
 
@@ -142,7 +143,7 @@ nv50_software_vblsem_release(void *data, u32 type, int head)
 		nv_wr32(priv, 0x060014, chan->vblank.value);
 	}
 
-	return NVKM_EVENT_DROP;
+	return NVKM_NOTIFY_DROP;
 }
 
 void
@@ -151,11 +152,8 @@ nv50_software_context_dtor(struct nouveau_object *object)
 	struct nv50_software_chan *chan = (void *)object;
 	int i;
 
-	if (chan->vblank.event) {
-		for (i = 0; i < chan->vblank.nr_event; i++)
-			nouveau_event_ref(NULL, &chan->vblank.event[i]);
-		kfree(chan->vblank.event);
-	}
+	for (i = 0; i < ARRAY_SIZE(chan->vblank.notify); i++)
+		nvkm_notify_fini(&chan->vblank.notify[i]);
 
 	nouveau_software_context_destroy(&chan->base);
 }
@@ -176,15 +174,15 @@ nv50_software_context_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	chan->vblank.nr_event = pdisp ? pdisp->vblank->index_nr : 0;
-	chan->vblank.event = kzalloc(chan->vblank.nr_event *
-				     sizeof(*chan->vblank.event), GFP_KERNEL);
-	if (!chan->vblank.event)
-		return -ENOMEM;
-
-	for (i = 0; i < chan->vblank.nr_event; i++) {
-		ret = nouveau_event_new(pdisp->vblank, 1, i, pclass->vblank,
-					chan, &chan->vblank.event[i]);
+	for (i = 0; pdisp && i < pdisp->vblank.index_nr; i++) {
+		ret = nvkm_notify_init(NULL, &pdisp->vblank, pclass->vblank,
+				       false,
+				       &(struct nvif_notify_head_req_v0) {
+					.head = i,
+				       },
+				       sizeof(struct nvif_notify_head_req_v0),
+				       sizeof(struct nvif_notify_head_rep_v0),
+				       &chan->vblank.notify[i]);
 		if (ret)
 			return ret;
 	}
@@ -198,7 +196,7 @@ nv50_software_cclass = {
 	.base.handle = NV_ENGCTX(SW, 0x50),
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_software_context_ctor,
-		.dtor = _nouveau_software_context_dtor,
+		.dtor = nv50_software_context_dtor,
 		.init = _nouveau_software_context_init,
 		.fini = _nouveau_software_context_fini,
 	},
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nv50.h b/drivers/gpu/drm/nouveau/core/engine/software/nv50.h
index bb49a7a20857..41542e725b4b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nv50.h
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nv50.h
@@ -19,14 +19,13 @@ int  nv50_software_ctor(struct nouveau_object *, struct nouveau_object *,
 
 struct nv50_software_cclass {
 	struct nouveau_oclass base;
-	int (*vblank)(void *, u32, int);
+	int (*vblank)(struct nvkm_notify *);
 };
 
 struct nv50_software_chan {
 	struct nouveau_software_chan base;
 	struct {
-		struct nouveau_eventh **event;
-		int nr_event;
+		struct nvkm_notify notify[4];
 		u32 channel;
 		u32 ctxdma;
 		u64 offset;
diff --git a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
index 135c20f38356..6af370d3a06d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/software/nvc0.c
@@ -23,7 +23,6 @@
  */
 
 #include <core/os.h>
-#include <core/class.h>
 #include <core/engctx.h>
 #include <core/event.h>
 
@@ -104,9 +103,10 @@ nvc0_software_sclass[] = {
  ******************************************************************************/
 
 static int
-nvc0_software_vblsem_release(void *data, u32 type, int head)
+nvc0_software_vblsem_release(struct nvkm_notify *notify)
 {
-	struct nv50_software_chan *chan = data;
+	struct nv50_software_chan *chan =
+		container_of(notify, typeof(*chan), vblank.notify[notify->index]);
 	struct nv50_software_priv *priv = (void *)nv_object(chan)->engine;
 	struct nouveau_bar *bar = nouveau_bar(priv);
 
@@ -116,7 +116,7 @@ nvc0_software_vblsem_release(void *data, u32 type, int head)
 	nv_wr32(priv, 0x060010, lower_32_bits(chan->vblank.offset));
 	nv_wr32(priv, 0x060014, chan->vblank.value);
 
-	return NVKM_EVENT_DROP;
+	return NVKM_NOTIFY_DROP;
 }
 
 static struct nv50_software_cclass
@@ -124,7 +124,7 @@ nvc0_software_cclass = {
 	.base.handle = NV_ENGCTX(SW, 0xc0),
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_software_context_ctor,
-		.dtor = _nouveau_software_context_dtor,
+		.dtor = nv50_software_context_dtor,
 		.init = _nouveau_software_context_init,
 		.fini = _nouveau_software_context_fini,
 	},
diff --git a/drivers/gpu/drm/nouveau/core/include/core/class.h b/drivers/gpu/drm/nouveau/core/include/core/class.h
deleted file mode 100644
index e0c812bc884f..000000000000
--- a/drivers/gpu/drm/nouveau/core/include/core/class.h
+++ /dev/null
@@ -1,470 +0,0 @@
-#ifndef __NOUVEAU_CLASS_H__
-#define __NOUVEAU_CLASS_H__
-
-/* Device class
- *
- * 0080: NV_DEVICE
- */
-#define NV_DEVICE_CLASS                                              0x00000080
-
-#define NV_DEVICE_DISABLE_IDENTIFY                        0x0000000000000001ULL
-#define NV_DEVICE_DISABLE_MMIO                            0x0000000000000002ULL
-#define NV_DEVICE_DISABLE_VBIOS                           0x0000000000000004ULL
-#define NV_DEVICE_DISABLE_CORE                            0x0000000000000008ULL
-#define NV_DEVICE_DISABLE_DISP                            0x0000000000010000ULL
-#define NV_DEVICE_DISABLE_FIFO                            0x0000000000020000ULL
-#define NV_DEVICE_DISABLE_GRAPH                           0x0000000100000000ULL
-#define NV_DEVICE_DISABLE_MPEG                            0x0000000200000000ULL
-#define NV_DEVICE_DISABLE_ME                              0x0000000400000000ULL
-#define NV_DEVICE_DISABLE_VP                              0x0000000800000000ULL
-#define NV_DEVICE_DISABLE_CRYPT                           0x0000001000000000ULL
-#define NV_DEVICE_DISABLE_BSP                             0x0000002000000000ULL
-#define NV_DEVICE_DISABLE_PPP                             0x0000004000000000ULL
-#define NV_DEVICE_DISABLE_COPY0                           0x0000008000000000ULL
-#define NV_DEVICE_DISABLE_COPY1                           0x0000010000000000ULL
-#define NV_DEVICE_DISABLE_VIC                             0x0000020000000000ULL
-#define NV_DEVICE_DISABLE_VENC                            0x0000040000000000ULL
-
-struct nv_device_class {
-	u64 device;	/* device identifier, ~0 for client default */
-	u64 disable;	/* disable particular subsystems */
-	u64 debug0;	/* as above, but *internal* ids, and *NOT* ABI */
-};
-
-/* DMA object classes
- *
- * 0002: NV_DMA_FROM_MEMORY
- * 0003: NV_DMA_TO_MEMORY
- * 003d: NV_DMA_IN_MEMORY
- */
-#define NV_DMA_FROM_MEMORY_CLASS                                     0x00000002
-#define NV_DMA_TO_MEMORY_CLASS                                       0x00000003
-#define NV_DMA_IN_MEMORY_CLASS                                       0x0000003d
-
-#define NV_DMA_TARGET_MASK                                           0x000000ff
-#define NV_DMA_TARGET_VM                                             0x00000000
-#define NV_DMA_TARGET_VRAM                                           0x00000001
-#define NV_DMA_TARGET_PCI                                            0x00000002
-#define NV_DMA_TARGET_PCI_US                                         0x00000003
-#define NV_DMA_TARGET_AGP                                            0x00000004
-#define NV_DMA_ACCESS_MASK                                           0x00000f00
-#define NV_DMA_ACCESS_VM                                             0x00000000
-#define NV_DMA_ACCESS_RD                                             0x00000100
-#define NV_DMA_ACCESS_WR                                             0x00000200
-#define NV_DMA_ACCESS_RDWR                                           0x00000300
-
-/* NV50:NVC0 */
-#define NV50_DMA_CONF0_ENABLE                                        0x80000000
-#define NV50_DMA_CONF0_PRIV                                          0x00300000
-#define NV50_DMA_CONF0_PRIV_VM                                       0x00000000
-#define NV50_DMA_CONF0_PRIV_US                                       0x00100000
-#define NV50_DMA_CONF0_PRIV__S                                       0x00200000
-#define NV50_DMA_CONF0_PART                                          0x00030000
-#define NV50_DMA_CONF0_PART_VM                                       0x00000000
-#define NV50_DMA_CONF0_PART_256                                      0x00010000
-#define NV50_DMA_CONF0_PART_1KB                                      0x00020000
-#define NV50_DMA_CONF0_COMP                                          0x00000180
-#define NV50_DMA_CONF0_COMP_NONE                                     0x00000000
-#define NV50_DMA_CONF0_COMP_VM                                       0x00000180
-#define NV50_DMA_CONF0_TYPE                                          0x0000007f
-#define NV50_DMA_CONF0_TYPE_LINEAR                                   0x00000000
-#define NV50_DMA_CONF0_TYPE_VM                                       0x0000007f
-
-/* NVC0:NVD9 */
-#define NVC0_DMA_CONF0_ENABLE                                        0x80000000
-#define NVC0_DMA_CONF0_PRIV                                          0x00300000
-#define NVC0_DMA_CONF0_PRIV_VM                                       0x00000000
-#define NVC0_DMA_CONF0_PRIV_US                                       0x00100000
-#define NVC0_DMA_CONF0_PRIV__S                                       0x00200000
-#define NVC0_DMA_CONF0_UNKN /* PART? */                              0x00030000
-#define NVC0_DMA_CONF0_TYPE                                          0x000000ff
-#define NVC0_DMA_CONF0_TYPE_LINEAR                                   0x00000000
-#define NVC0_DMA_CONF0_TYPE_VM                                       0x000000ff
-
-/* NVD9- */
-#define NVD0_DMA_CONF0_ENABLE                                        0x80000000
-#define NVD0_DMA_CONF0_PAGE                                          0x00000400
-#define NVD0_DMA_CONF0_PAGE_LP                                       0x00000000
-#define NVD0_DMA_CONF0_PAGE_SP                                       0x00000400
-#define NVD0_DMA_CONF0_TYPE                                          0x000000ff
-#define NVD0_DMA_CONF0_TYPE_LINEAR                                   0x00000000
-#define NVD0_DMA_CONF0_TYPE_VM                                       0x000000ff
-
-struct nv_dma_class {
-	u32 flags;
-	u32 pad0;
-	u64 start;
-	u64 limit;
-	u32 conf0;
-};
-
-/* Perfmon counter class
- *
- * XXXX: NV_PERFCTR
- */
-#define NV_PERFCTR_CLASS                                             0x0000ffff
-#define NV_PERFCTR_QUERY                                             0x00000000
-#define NV_PERFCTR_SAMPLE                                            0x00000001
-#define NV_PERFCTR_READ                                              0x00000002
-
-struct nv_perfctr_class {
-	u16 logic_op;
-	struct {
-		char __user *name; /*XXX: use cfu when exposed to userspace */
-		u32 size;
-	} signal[4];
-};
-
-struct nv_perfctr_query {
-	u32 iter;
-	u32 size;
-	char __user *name; /*XXX: use ctu when exposed to userspace */
-};
-
-struct nv_perfctr_sample {
-};
-
-struct nv_perfctr_read {
-	u32 ctr;
-	u32 clk;
-};
-
-/* Device control class
- *
- * XXXX: NV_CONTROL
- */
-#define NV_CONTROL_CLASS                                             0x0000fffe
-
-#define NV_CONTROL_PSTATE_INFO                                       0x00000000
-#define NV_CONTROL_PSTATE_INFO_USTATE_DISABLE                              (-1)
-#define NV_CONTROL_PSTATE_INFO_USTATE_PERFMON                              (-2)
-#define NV_CONTROL_PSTATE_INFO_PSTATE_UNKNOWN                              (-1)
-#define NV_CONTROL_PSTATE_INFO_PSTATE_PERFMON                              (-2)
-#define NV_CONTROL_PSTATE_ATTR                                       0x00000001
-#define NV_CONTROL_PSTATE_ATTR_STATE_CURRENT                               (-1)
-#define NV_CONTROL_PSTATE_USER                                       0x00000002
-#define NV_CONTROL_PSTATE_USER_STATE_UNKNOWN                               (-1)
-#define NV_CONTROL_PSTATE_USER_STATE_PERFMON                               (-2)
-
-struct nv_control_pstate_info {
-	u32 count; /* out: number of power states */
-	s32 ustate; /* out: current target pstate index */
-	u32 pstate; /* out: current pstate index */
-};
-
-struct nv_control_pstate_attr {
-	s32 state; /*  in: index of pstate to query
-		    * out: pstate identifier
-		    */
-	u32 index; /*  in: index of attribute to query
-		    * out: index of next attribute, or 0 if no more
-		    */
-	char name[32];
-	char unit[16];
-	u32 min;
-	u32 max;
-};
-
-struct nv_control_pstate_user {
-	s32 state; /*  in: pstate identifier */
-};
-
-/* DMA FIFO channel classes
- *
- * 006b: NV03_CHANNEL_DMA
- * 006e: NV10_CHANNEL_DMA
- * 176e: NV17_CHANNEL_DMA
- * 406e: NV40_CHANNEL_DMA
- * 506e: NV50_CHANNEL_DMA
- * 826e: NV84_CHANNEL_DMA
- */
-#define NV03_CHANNEL_DMA_CLASS                                       0x0000006b
-#define NV10_CHANNEL_DMA_CLASS                                       0x0000006e
-#define NV17_CHANNEL_DMA_CLASS                                       0x0000176e
-#define NV40_CHANNEL_DMA_CLASS                                       0x0000406e
-#define NV50_CHANNEL_DMA_CLASS                                       0x0000506e
-#define NV84_CHANNEL_DMA_CLASS                                       0x0000826e
-
-struct nv03_channel_dma_class {
-	u32 pushbuf;
-	u32 pad0;
-	u64 offset;
-};
-
-/* Indirect FIFO channel classes
- *
- * 506f: NV50_CHANNEL_IND
- * 826f: NV84_CHANNEL_IND
- * 906f: NVC0_CHANNEL_IND
- * a06f: NVE0_CHANNEL_IND
- */
-
-#define NV50_CHANNEL_IND_CLASS                                       0x0000506f
-#define NV84_CHANNEL_IND_CLASS                                       0x0000826f
-#define NVC0_CHANNEL_IND_CLASS                                       0x0000906f
-#define NVE0_CHANNEL_IND_CLASS                                       0x0000a06f
-
-struct nv50_channel_ind_class {
-	u32 pushbuf;
-	u32 ilength;
-	u64 ioffset;
-};
-
-#define NVE0_CHANNEL_IND_ENGINE_GR                                   0x00000001
-#define NVE0_CHANNEL_IND_ENGINE_VP                                   0x00000002
-#define NVE0_CHANNEL_IND_ENGINE_PPP                                  0x00000004
-#define NVE0_CHANNEL_IND_ENGINE_BSP                                  0x00000008
-#define NVE0_CHANNEL_IND_ENGINE_CE0                                  0x00000010
-#define NVE0_CHANNEL_IND_ENGINE_CE1                                  0x00000020
-#define NVE0_CHANNEL_IND_ENGINE_ENC                                  0x00000040
-
-struct nve0_channel_ind_class {
-	u32 pushbuf;
-	u32 ilength;
-	u64 ioffset;
-	u32 engine;
-};
-
-/* 0046: NV04_DISP
- */
-
-#define NV04_DISP_CLASS                                              0x00000046
-
-#define NV04_DISP_MTHD                                               0x00000000
-#define NV04_DISP_MTHD_HEAD                                          0x00000001
-
-#define NV04_DISP_SCANOUTPOS                                         0x00000000
-
-struct nv04_display_class {
-};
-
-struct nv04_display_scanoutpos {
-	s64 time[2];
-	u32 vblanks;
-	u32 vblanke;
-	u32 vtotal;
-	u32 vline;
-	u32 hblanks;
-	u32 hblanke;
-	u32 htotal;
-	u32 hline;
-};
-
-/* 5070: NV50_DISP
- * 8270: NV84_DISP
- * 8370: NVA0_DISP
- * 8870: NV94_DISP
- * 8570: NVA3_DISP
- * 9070: NVD0_DISP
- * 9170: NVE0_DISP
- * 9270: NVF0_DISP
- * 9470: GM107_DISP
- */
-
-#define NV50_DISP_CLASS                                              0x00005070
-#define NV84_DISP_CLASS                                              0x00008270
-#define NVA0_DISP_CLASS                                              0x00008370
-#define NV94_DISP_CLASS                                              0x00008870
-#define NVA3_DISP_CLASS                                              0x00008570
-#define NVD0_DISP_CLASS                                              0x00009070
-#define NVE0_DISP_CLASS                                              0x00009170
-#define NVF0_DISP_CLASS                                              0x00009270
-#define GM107_DISP_CLASS                                             0x00009470
-
-#define NV50_DISP_MTHD                                               0x00000000
-#define NV50_DISP_MTHD_HEAD                                          0x00000003
-
-#define NV50_DISP_SCANOUTPOS                                         0x00000000
-
-#define NV50_DISP_SOR_MTHD                                           0x00010000
-#define NV50_DISP_SOR_MTHD_TYPE                                      0x0000f000
-#define NV50_DISP_SOR_MTHD_HEAD                                      0x00000018
-#define NV50_DISP_SOR_MTHD_LINK                                      0x00000004
-#define NV50_DISP_SOR_MTHD_OR                                        0x00000003
-
-#define NV50_DISP_SOR_PWR                                            0x00010000
-#define NV50_DISP_SOR_PWR_STATE                                      0x00000001
-#define NV50_DISP_SOR_PWR_STATE_ON                                   0x00000001
-#define NV50_DISP_SOR_PWR_STATE_OFF                                  0x00000000
-#define NVA3_DISP_SOR_HDA_ELD                                        0x00010100
-#define NV84_DISP_SOR_HDMI_PWR                                       0x00012000
-#define NV84_DISP_SOR_HDMI_PWR_STATE                                 0x40000000
-#define NV84_DISP_SOR_HDMI_PWR_STATE_OFF                             0x00000000
-#define NV84_DISP_SOR_HDMI_PWR_STATE_ON                              0x40000000
-#define NV84_DISP_SOR_HDMI_PWR_MAX_AC_PACKET                         0x001f0000
-#define NV84_DISP_SOR_HDMI_PWR_REKEY                                 0x0000007f
-#define NV50_DISP_SOR_LVDS_SCRIPT                                    0x00013000
-#define NV50_DISP_SOR_LVDS_SCRIPT_ID                                 0x0000ffff
-#define NV94_DISP_SOR_DP_PWR                                         0x00016000
-#define NV94_DISP_SOR_DP_PWR_STATE                                   0x00000001
-#define NV94_DISP_SOR_DP_PWR_STATE_OFF                               0x00000000
-#define NV94_DISP_SOR_DP_PWR_STATE_ON                                0x00000001
-
-#define NV50_DISP_DAC_MTHD                                           0x00020000
-#define NV50_DISP_DAC_MTHD_TYPE                                      0x0000f000
-#define NV50_DISP_DAC_MTHD_OR                                        0x00000003
-
-#define NV50_DISP_DAC_PWR                                            0x00020000
-#define NV50_DISP_DAC_PWR_HSYNC                                      0x00000001
-#define NV50_DISP_DAC_PWR_HSYNC_ON                                   0x00000000
-#define NV50_DISP_DAC_PWR_HSYNC_LO                                   0x00000001
-#define NV50_DISP_DAC_PWR_VSYNC                                      0x00000004
-#define NV50_DISP_DAC_PWR_VSYNC_ON                                   0x00000000
-#define NV50_DISP_DAC_PWR_VSYNC_LO                                   0x00000004
-#define NV50_DISP_DAC_PWR_DATA                                       0x00000010
-#define NV50_DISP_DAC_PWR_DATA_ON                                    0x00000000
-#define NV50_DISP_DAC_PWR_DATA_LO                                    0x00000010
-#define NV50_DISP_DAC_PWR_STATE                                      0x00000040
-#define NV50_DISP_DAC_PWR_STATE_ON                                   0x00000000
-#define NV50_DISP_DAC_PWR_STATE_OFF                                  0x00000040
-#define NV50_DISP_DAC_LOAD                                           0x00020100
-#define NV50_DISP_DAC_LOAD_VALUE                                     0x00000007
-
-#define NV50_DISP_PIOR_MTHD                                          0x00030000
-#define NV50_DISP_PIOR_MTHD_TYPE                                     0x0000f000
-#define NV50_DISP_PIOR_MTHD_OR                                       0x00000003
-
-#define NV50_DISP_PIOR_PWR                                           0x00030000
-#define NV50_DISP_PIOR_PWR_STATE                                     0x00000001
-#define NV50_DISP_PIOR_PWR_STATE_ON                                  0x00000001
-#define NV50_DISP_PIOR_PWR_STATE_OFF                                 0x00000000
-#define NV50_DISP_PIOR_TMDS_PWR                                      0x00032000
-#define NV50_DISP_PIOR_TMDS_PWR_STATE                                0x00000001
-#define NV50_DISP_PIOR_TMDS_PWR_STATE_ON                             0x00000001
-#define NV50_DISP_PIOR_TMDS_PWR_STATE_OFF                            0x00000000
-#define NV50_DISP_PIOR_DP_PWR                                        0x00036000
-#define NV50_DISP_PIOR_DP_PWR_STATE                                  0x00000001
-#define NV50_DISP_PIOR_DP_PWR_STATE_ON                               0x00000001
-#define NV50_DISP_PIOR_DP_PWR_STATE_OFF                              0x00000000
-
-struct nv50_display_class {
-};
-
-/* 507a: NV50_DISP_CURS
- * 827a: NV84_DISP_CURS
- * 837a: NVA0_DISP_CURS
- * 887a: NV94_DISP_CURS
- * 857a: NVA3_DISP_CURS
- * 907a: NVD0_DISP_CURS
- * 917a: NVE0_DISP_CURS
- * 927a: NVF0_DISP_CURS
- * 947a: GM107_DISP_CURS
- */
-
-#define NV50_DISP_CURS_CLASS                                         0x0000507a
-#define NV84_DISP_CURS_CLASS                                         0x0000827a
-#define NVA0_DISP_CURS_CLASS                                         0x0000837a
-#define NV94_DISP_CURS_CLASS                                         0x0000887a
-#define NVA3_DISP_CURS_CLASS                                         0x0000857a
-#define NVD0_DISP_CURS_CLASS                                         0x0000907a
-#define NVE0_DISP_CURS_CLASS                                         0x0000917a
-#define NVF0_DISP_CURS_CLASS                                         0x0000927a
-#define GM107_DISP_CURS_CLASS                                        0x0000947a
-
-struct nv50_display_curs_class {
-	u32 head;
-};
-
-/* 507b: NV50_DISP_OIMM
- * 827b: NV84_DISP_OIMM
- * 837b: NVA0_DISP_OIMM
- * 887b: NV94_DISP_OIMM
- * 857b: NVA3_DISP_OIMM
- * 907b: NVD0_DISP_OIMM
- * 917b: NVE0_DISP_OIMM
- * 927b: NVE0_DISP_OIMM
- * 947b: GM107_DISP_OIMM
- */
-
-#define NV50_DISP_OIMM_CLASS                                         0x0000507b
-#define NV84_DISP_OIMM_CLASS                                         0x0000827b
-#define NVA0_DISP_OIMM_CLASS                                         0x0000837b
-#define NV94_DISP_OIMM_CLASS                                         0x0000887b
-#define NVA3_DISP_OIMM_CLASS                                         0x0000857b
-#define NVD0_DISP_OIMM_CLASS                                         0x0000907b
-#define NVE0_DISP_OIMM_CLASS                                         0x0000917b
-#define NVF0_DISP_OIMM_CLASS                                         0x0000927b
-#define GM107_DISP_OIMM_CLASS                                        0x0000947b
-
-struct nv50_display_oimm_class {
-	u32 head;
-};
-
-/* 507c: NV50_DISP_SYNC
- * 827c: NV84_DISP_SYNC
- * 837c: NVA0_DISP_SYNC
- * 887c: NV94_DISP_SYNC
- * 857c: NVA3_DISP_SYNC
- * 907c: NVD0_DISP_SYNC
- * 917c: NVE0_DISP_SYNC
- * 927c: NVF0_DISP_SYNC
- * 947c: GM107_DISP_SYNC
- */
-
-#define NV50_DISP_SYNC_CLASS                                         0x0000507c
-#define NV84_DISP_SYNC_CLASS                                         0x0000827c
-#define NVA0_DISP_SYNC_CLASS                                         0x0000837c
-#define NV94_DISP_SYNC_CLASS                                         0x0000887c
-#define NVA3_DISP_SYNC_CLASS                                         0x0000857c
-#define NVD0_DISP_SYNC_CLASS                                         0x0000907c
-#define NVE0_DISP_SYNC_CLASS                                         0x0000917c
-#define NVF0_DISP_SYNC_CLASS                                         0x0000927c
-#define GM107_DISP_SYNC_CLASS                                        0x0000947c
-
-struct nv50_display_sync_class {
-	u32 pushbuf;
-	u32 head;
-};
-
-/* 507d: NV50_DISP_MAST
- * 827d: NV84_DISP_MAST
- * 837d: NVA0_DISP_MAST
- * 887d: NV94_DISP_MAST
- * 857d: NVA3_DISP_MAST
- * 907d: NVD0_DISP_MAST
- * 917d: NVE0_DISP_MAST
- * 927d: NVF0_DISP_MAST
- * 947d: GM107_DISP_MAST
- */
-
-#define NV50_DISP_MAST_CLASS                                         0x0000507d
-#define NV84_DISP_MAST_CLASS                                         0x0000827d
-#define NVA0_DISP_MAST_CLASS                                         0x0000837d
-#define NV94_DISP_MAST_CLASS                                         0x0000887d
-#define NVA3_DISP_MAST_CLASS                                         0x0000857d
-#define NVD0_DISP_MAST_CLASS                                         0x0000907d
-#define NVE0_DISP_MAST_CLASS                                         0x0000917d
-#define NVF0_DISP_MAST_CLASS                                         0x0000927d
-#define GM107_DISP_MAST_CLASS                                        0x0000947d
-
-struct nv50_display_mast_class {
-	u32 pushbuf;
-};
-
-/* 507e: NV50_DISP_OVLY
- * 827e: NV84_DISP_OVLY
- * 837e: NVA0_DISP_OVLY
- * 887e: NV94_DISP_OVLY
- * 857e: NVA3_DISP_OVLY
- * 907e: NVD0_DISP_OVLY
- * 917e: NVE0_DISP_OVLY
- * 927e: NVF0_DISP_OVLY
- * 947e: GM107_DISP_OVLY
- */
-
-#define NV50_DISP_OVLY_CLASS                                         0x0000507e
-#define NV84_DISP_OVLY_CLASS                                         0x0000827e
-#define NVA0_DISP_OVLY_CLASS                                         0x0000837e
-#define NV94_DISP_OVLY_CLASS                                         0x0000887e
-#define NVA3_DISP_OVLY_CLASS                                         0x0000857e
-#define NVD0_DISP_OVLY_CLASS                                         0x0000907e
-#define NVE0_DISP_OVLY_CLASS                                         0x0000917e
-#define NVF0_DISP_OVLY_CLASS                                         0x0000927e
-#define GM107_DISP_OVLY_CLASS                                        0x0000947e
-
-struct nv50_display_ovly_class {
-	u32 pushbuf;
-	u32 head;
-};
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/client.h b/drivers/gpu/drm/nouveau/core/include/core/client.h
index c66eac513803..b0ce9f6680b5 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/client.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/client.h
@@ -10,6 +10,11 @@ struct nouveau_client {
 	char name[32];
 	u32 debug;
 	struct nouveau_vm *vm;
+	bool super;
+	void *data;
+
+	int (*ntfy)(const void *, u32, const void *, u32);
+	struct nvkm_client_notify *notify[16];
 };
 
 static inline struct nouveau_client *
@@ -43,4 +48,10 @@ int  nouveau_client_init(struct nouveau_client *);
 int  nouveau_client_fini(struct nouveau_client *, bool suspend);
 const char *nouveau_client_name(void *obj);
 
+int nvkm_client_notify_new(struct nouveau_object *, struct nvkm_event *,
+			   void *data, u32 size);
+int nvkm_client_notify_del(struct nouveau_client *, int index);
+int nvkm_client_notify_get(struct nouveau_client *, int index);
+int nvkm_client_notify_put(struct nouveau_client *, int index);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index a8a9a9cf16cb..1d9d893929bb 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -4,6 +4,7 @@
 #include <core/object.h>
 #include <core/subdev.h>
 #include <core/engine.h>
+#include <core/event.h>
 
 enum nv_subdev_type {
 	NVDEV_ENGINE_DEVICE,
@@ -23,12 +24,13 @@ enum nv_subdev_type {
 	 * been created, and are allowed to assume any subdevs in the
 	 * list above them exist and have been initialised.
 	 */
+	NVDEV_SUBDEV_FUSE,
 	NVDEV_SUBDEV_MXM,
 	NVDEV_SUBDEV_MC,
 	NVDEV_SUBDEV_BUS,
 	NVDEV_SUBDEV_TIMER,
 	NVDEV_SUBDEV_FB,
-	NVDEV_SUBDEV_LTCG,
+	NVDEV_SUBDEV_LTC,
 	NVDEV_SUBDEV_IBUS,
 	NVDEV_SUBDEV_INSTMEM,
 	NVDEV_SUBDEV_VM,
@@ -69,6 +71,8 @@ struct nouveau_device {
 	struct platform_device *platformdev;
 	u64 handle;
 
+	struct nvkm_event event;
+
 	const char *cfgopt;
 	const char *dbgopt;
 	const char *name;
@@ -84,7 +88,6 @@ struct nouveau_device {
 		NV_40    = 0x40,
 		NV_50    = 0x50,
 		NV_C0    = 0xc0,
-		NV_D0    = 0xd0,
 		NV_E0    = 0xe0,
 		GM100    = 0x110,
 	} card_type;
@@ -93,8 +96,14 @@ struct nouveau_device {
 
 	struct nouveau_oclass *oclass[NVDEV_SUBDEV_NR];
 	struct nouveau_object *subdev[NVDEV_SUBDEV_NR];
+
+	struct {
+		struct notifier_block nb;
+	} acpi;
 };
 
+int nouveau_device_list(u64 *name, int size);
+
 static inline struct nouveau_device *
 nv_device(void *obj)
 {
@@ -162,12 +171,6 @@ nv_device_resource_start(struct nouveau_device *device, unsigned int bar);
 resource_size_t
 nv_device_resource_len(struct nouveau_device *device, unsigned int bar);
 
-dma_addr_t
-nv_device_map_page(struct nouveau_device *device, struct page *page);
-
-void
-nv_device_unmap_page(struct nouveau_device *device, dma_addr_t addr);
-
 int
 nv_device_get_irq(struct nouveau_device *device, bool stall);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/event.h b/drivers/gpu/drm/nouveau/core/include/core/event.h
index ba3f1a76a815..92876528972f 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/event.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/event.h
@@ -1,47 +1,35 @@
 #ifndef __NVKM_EVENT_H__
 #define __NVKM_EVENT_H__
 
-/* return codes from event handlers */
-#define NVKM_EVENT_DROP 0
-#define NVKM_EVENT_KEEP 1
-
-/* nouveau_eventh.flags bit #s */
-#define NVKM_EVENT_ENABLE 0
-
-struct nouveau_eventh {
-	struct nouveau_event *event;
-	struct list_head head;
-	unsigned long flags;
-	u32 types;
-	int index;
-	int (*func)(void *, u32, int);
-	void *priv;
+#include <core/notify.h>
+
+struct nvkm_event_func {
+	int  (*ctor)(struct nouveau_object *, void *data, u32 size,
+		     struct nvkm_notify *);
+	void (*send)(void *data, u32 size, struct nvkm_notify *);
+	void (*init)(struct nvkm_event *, int type, int index);
+	void (*fini)(struct nvkm_event *, int type, int index);
 };
 
-struct nouveau_event {
-	void *priv;
-	int (*check)(struct nouveau_event *, u32 type, int index);
-	void (*enable)(struct nouveau_event *, int type, int index);
-	void (*disable)(struct nouveau_event *, int type, int index);
+struct nvkm_event {
+	const struct nvkm_event_func *func;
 
 	int types_nr;
 	int index_nr;
 
-	spinlock_t list_lock;
-	struct list_head *list;
 	spinlock_t refs_lock;
-	int refs[];
+	spinlock_t list_lock;
+	struct list_head list;
+	int *refs;
 };
 
-int  nouveau_event_create(int types_nr, int index_nr, struct nouveau_event **);
-void nouveau_event_destroy(struct nouveau_event **);
-void nouveau_event_trigger(struct nouveau_event *, u32 types, int index);
-
-int  nouveau_event_new(struct nouveau_event *, u32 types, int index,
-		       int (*func)(void *, u32, int), void *,
-		       struct nouveau_eventh **);
-void nouveau_event_ref(struct nouveau_eventh *, struct nouveau_eventh **);
-void nouveau_event_get(struct nouveau_eventh *);
-void nouveau_event_put(struct nouveau_eventh *);
+int  nvkm_event_init(const struct nvkm_event_func *func,
+		     int types_nr, int index_nr,
+		     struct nvkm_event *);
+void nvkm_event_fini(struct nvkm_event *);
+void nvkm_event_get(struct nvkm_event *, u32 types, int index);
+void nvkm_event_put(struct nvkm_event *, u32 types, int index);
+void nvkm_event_send(struct nvkm_event *, u32 types, int index,
+		     void *data, u32 size);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/handle.h b/drivers/gpu/drm/nouveau/core/include/core/handle.h
index 363674cdf8ab..ceb67d770875 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/handle.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/handle.h
@@ -10,6 +10,9 @@ struct nouveau_handle {
 	u32 name;
 	u32 priv;
 
+	u8  route;
+	u64 token;
+
 	struct nouveau_handle *parent;
 	struct nouveau_object *object;
 };
@@ -20,6 +23,11 @@ void nouveau_handle_destroy(struct nouveau_handle *);
 int  nouveau_handle_init(struct nouveau_handle *);
 int  nouveau_handle_fini(struct nouveau_handle *, bool suspend);
 
+int  nouveau_handle_new(struct nouveau_object *, u32 parent, u32 handle,
+			u16 oclass, void *data, u32 size,
+			struct nouveau_object **);
+int  nouveau_handle_del(struct nouveau_object *, u32 parent, u32 handle);
+
 struct nouveau_object *
 nouveau_handle_ref(struct nouveau_object *, u32 name);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/core/ioctl.h b/drivers/gpu/drm/nouveau/core/include/core/ioctl.h
new file mode 100644
index 000000000000..ac7935c2474e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/core/ioctl.h
@@ -0,0 +1,6 @@
+#ifndef __NVKM_IOCTL_H__
+#define __NVKM_IOCTL_H__
+
+int nvkm_ioctl(struct nouveau_client *, bool, void *, u32, void **);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/mm.h b/drivers/gpu/drm/nouveau/core/include/core/mm.h
index 2bf7d0e32261..bfe6931544fe 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/mm.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/mm.h
@@ -6,6 +6,10 @@ struct nouveau_mm_node {
 	struct list_head fl_entry;
 	struct list_head rl_entry;
 
+#define NVKM_MM_HEAP_ANY 0x00
+	u8  heap;
+#define NVKM_MM_TYPE_NONE 0x00
+#define NVKM_MM_TYPE_HOLE 0xff
 	u8  type;
 	u32 offset;
 	u32 length;
@@ -27,10 +31,10 @@ nouveau_mm_initialised(struct nouveau_mm *mm)
 
 int  nouveau_mm_init(struct nouveau_mm *, u32 offset, u32 length, u32 block);
 int  nouveau_mm_fini(struct nouveau_mm *);
-int  nouveau_mm_head(struct nouveau_mm *, u8 type, u32 size_max, u32 size_min,
-		     u32 align, struct nouveau_mm_node **);
-int  nouveau_mm_tail(struct nouveau_mm *, u8 type, u32 size_max, u32 size_min,
-		     u32 align, struct nouveau_mm_node **);
+int  nouveau_mm_head(struct nouveau_mm *, u8 heap, u8 type, u32 size_max,
+		     u32 size_min, u32 align, struct nouveau_mm_node **);
+int  nouveau_mm_tail(struct nouveau_mm *, u8 heap, u8 type, u32 size_max,
+		     u32 size_min, u32 align, struct nouveau_mm_node **);
 void nouveau_mm_free(struct nouveau_mm *, struct nouveau_mm_node **);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/notify.h b/drivers/gpu/drm/nouveau/core/include/core/notify.h
new file mode 100644
index 000000000000..a7c3c5f578cc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/core/notify.h
@@ -0,0 +1,37 @@
+#ifndef __NVKM_NOTIFY_H__
+#define __NVKM_NOTIFY_H__
+
+struct nvkm_notify {
+	struct nvkm_event *event;
+	struct list_head head;
+#define NVKM_NOTIFY_USER 0
+#define NVKM_NOTIFY_WORK 1
+	unsigned long flags;
+	int block;
+#define NVKM_NOTIFY_DROP 0
+#define NVKM_NOTIFY_KEEP 1
+	int (*func)(struct nvkm_notify *);
+
+	/* set by nvkm_event ctor */
+	u32 types;
+	int index;
+	u32 size;
+
+	struct work_struct work;
+	/* this is const for a *very* good reason - the data might be on the
+	 * stack from an irq handler.  if you're not core/notify.c then you
+	 * should probably think twice before casting it away...
+	 */
+	const void *data;
+};
+
+int  nvkm_notify_init(struct nouveau_object *, struct nvkm_event *,
+		      int (*func)(struct nvkm_notify *), bool work,
+		      void *data, u32 size, u32 reply,
+		      struct nvkm_notify *);
+void nvkm_notify_fini(struct nvkm_notify *);
+void nvkm_notify_get(struct nvkm_notify *);
+void nvkm_notify_put(struct nvkm_notify *);
+void nvkm_notify_send(struct nvkm_notify *, void *data, u32 size);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/object.h b/drivers/gpu/drm/nouveau/core/include/core/object.h
index 62e68baef087..d7039482d6fd 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/object.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/object.h
@@ -48,6 +48,10 @@ void nouveau_object_destroy(struct nouveau_object *);
 int  nouveau_object_init(struct nouveau_object *);
 int  nouveau_object_fini(struct nouveau_object *, bool suspend);
 
+int _nouveau_object_ctor(struct nouveau_object *, struct nouveau_object *,
+			 struct nouveau_oclass *, void *, u32,
+			 struct nouveau_object **);
+
 extern struct nouveau_ofuncs nouveau_object_ofuncs;
 
 /* Don't allocate dynamically, because lockdep needs lock_class_keys to be in
@@ -78,6 +82,7 @@ struct nouveau_omthds {
 	int (*call)(struct nouveau_object *, u32, void *, u32);
 };
 
+struct nvkm_event;
 struct nouveau_ofuncs {
 	int  (*ctor)(struct nouveau_object *, struct nouveau_object *,
 		     struct nouveau_oclass *, void *data, u32 size,
@@ -85,6 +90,9 @@ struct nouveau_ofuncs {
 	void (*dtor)(struct nouveau_object *);
 	int  (*init)(struct nouveau_object *);
 	int  (*fini)(struct nouveau_object *, bool suspend);
+	int  (*mthd)(struct nouveau_object *, u32, void *, u32);
+	int  (*ntfy)(struct nouveau_object *, u32, struct nvkm_event **);
+	int  (* map)(struct nouveau_object *, u64 *, u32 *);
 	u8   (*rd08)(struct nouveau_object *, u64 offset);
 	u16  (*rd16)(struct nouveau_object *, u64 offset);
 	u32  (*rd32)(struct nouveau_object *, u64 offset);
@@ -106,10 +114,6 @@ void nouveau_object_ref(struct nouveau_object *, struct nouveau_object **);
 int nouveau_object_inc(struct nouveau_object *);
 int nouveau_object_dec(struct nouveau_object *, bool suspend);
 
-int nouveau_object_new(struct nouveau_object *, u32 parent, u32 handle,
-		       u16 oclass, void *data, u32 size,
-		       struct nouveau_object **);
-int nouveau_object_del(struct nouveau_object *, u32 parent, u32 handle);
 void nouveau_object_debug(void);
 
 static inline int
@@ -199,4 +203,21 @@ nv_memcmp(void *obj, u32 addr, const char *str, u32 len)
 	return 0;
 }
 
+#include <core/handle.h>
+
+static inline int
+nouveau_object_new(struct nouveau_object *client, u32 parent, u32 handle,
+		   u16 oclass, void *data, u32 size,
+		   struct nouveau_object **pobject)
+{
+	return nouveau_handle_new(client, parent, handle, oclass,
+				  data, size, pobject);
+}
+
+static inline int
+nouveau_object_del(struct nouveau_object *client, u32 parent, u32 handle)
+{
+	return nouveau_handle_del(client, parent, handle);
+}
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/parent.h b/drivers/gpu/drm/nouveau/core/include/core/parent.h
index 9f5ea900ff00..12da418ec70a 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/parent.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/parent.h
@@ -57,5 +57,6 @@ void _nouveau_parent_dtor(struct nouveau_object *);
 int nouveau_parent_sclass(struct nouveau_object *, u16 handle,
 			  struct nouveau_object **pengine,
 			  struct nouveau_oclass **poclass);
+int nouveau_parent_lclass(struct nouveau_object *, u32 *, int);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/printk.h b/drivers/gpu/drm/nouveau/core/include/core/printk.h
index 0f9a37bd32b0..451b6ed20b7e 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/printk.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/printk.h
@@ -21,6 +21,7 @@ nv_printk_(struct nouveau_object *, int, const char *, ...);
 #define nv_debug(o,f,a...) nv_printk((o), DEBUG, f, ##a)
 #define nv_trace(o,f,a...) nv_printk((o), TRACE, f, ##a)
 #define nv_spam(o,f,a...) nv_printk((o), SPAM, f, ##a)
+#define nv_ioctl(o,f,a...) nv_trace(nouveau_client(o), "ioctl: "f, ##a)
 
 #define nv_assert(f,a...) do {                                                 \
 	if (NV_DBG_FATAL <= CONFIG_NOUVEAU_DEBUG)                              \
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/disp.h b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
index fde842896806..7a64f347b385 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/disp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/disp.h
@@ -6,20 +6,13 @@
 #include <core/device.h>
 #include <core/event.h>
 
-enum nvkm_hpd_event {
-	NVKM_HPD_PLUG = 1,
-	NVKM_HPD_UNPLUG = 2,
-	NVKM_HPD_IRQ = 4,
-	NVKM_HPD = (NVKM_HPD_PLUG | NVKM_HPD_UNPLUG | NVKM_HPD_IRQ)
-};
-
 struct nouveau_disp {
 	struct nouveau_engine base;
 
 	struct list_head outp;
-	struct nouveau_event *hpd;
 
-	struct nouveau_event *vblank;
+	struct nvkm_event hpd;
+	struct nvkm_event vblank;
 };
 
 static inline struct nouveau_disp *
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/dmaobj.h b/drivers/gpu/drm/nouveau/core/include/engine/dmaobj.h
index b28914ed1752..1b283a7b78e6 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/dmaobj.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/dmaobj.h
@@ -12,37 +12,20 @@ struct nouveau_dmaobj {
 	u32 access;
 	u64 start;
 	u64 limit;
-	u32 conf0;
 };
 
 struct nouveau_dmaeng {
 	struct nouveau_engine base;
 
 	/* creates a "physical" dma object from a struct nouveau_dmaobj */
-	int (*bind)(struct nouveau_dmaeng *dmaeng,
+	int (*bind)(struct nouveau_dmaobj *dmaobj,
 		    struct nouveau_object *parent,
-		    struct nouveau_dmaobj *dmaobj,
 		    struct nouveau_gpuobj **);
 };
 
-#define nouveau_dmaeng_create(p,e,c,d)                                         \
-	nouveau_engine_create((p), (e), (c), true, "DMAOBJ", "dmaobj", (d))
-#define nouveau_dmaeng_destroy(p)                                              \
-	nouveau_engine_destroy(&(p)->base)
-#define nouveau_dmaeng_init(p)                                                 \
-	nouveau_engine_init(&(p)->base)
-#define nouveau_dmaeng_fini(p,s)                                               \
-	nouveau_engine_fini(&(p)->base, (s))
-
-#define _nouveau_dmaeng_dtor _nouveau_engine_dtor
-#define _nouveau_dmaeng_init _nouveau_engine_init
-#define _nouveau_dmaeng_fini _nouveau_engine_fini
-
-extern struct nouveau_oclass nv04_dmaeng_oclass;
-extern struct nouveau_oclass nv50_dmaeng_oclass;
-extern struct nouveau_oclass nvc0_dmaeng_oclass;
-extern struct nouveau_oclass nvd0_dmaeng_oclass;
-
-extern struct nouveau_oclass nouveau_dmaobj_sclass[];
+extern struct nouveau_oclass *nv04_dmaeng_oclass;
+extern struct nouveau_oclass *nv50_dmaeng_oclass;
+extern struct nouveau_oclass *nvc0_dmaeng_oclass;
+extern struct nouveau_oclass *nvd0_dmaeng_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
index b639eb2c74ff..2007453f6fce 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/fifo.h
@@ -4,12 +4,14 @@
 #include <core/namedb.h>
 #include <core/gpuobj.h>
 #include <core/engine.h>
+#include <core/event.h>
 
 struct nouveau_fifo_chan {
 	struct nouveau_namedb base;
 	struct nouveau_dmaobj *pushdma;
 	struct nouveau_gpuobj *pushgpu;
 	void __iomem *user;
+	u64 addr;
 	u32 size;
 	u16 chid;
 	atomic_t refcnt; /* NV04_NVSW_SET_REF */
@@ -40,8 +42,10 @@ void nouveau_fifo_channel_destroy(struct nouveau_fifo_chan *);
 #define _nouveau_fifo_channel_fini _nouveau_namedb_fini
 
 void _nouveau_fifo_channel_dtor(struct nouveau_object *);
+int  _nouveau_fifo_channel_map(struct nouveau_object *, u64 *, u32 *);
 u32  _nouveau_fifo_channel_rd32(struct nouveau_object *, u64);
 void _nouveau_fifo_channel_wr32(struct nouveau_object *, u64, u32);
+int  _nouveau_fifo_channel_ntfy(struct nouveau_object *, u32, struct nvkm_event **);
 
 struct nouveau_fifo_base {
 	struct nouveau_gpuobj base;
@@ -65,8 +69,8 @@ struct nouveau_fifo_base {
 struct nouveau_fifo {
 	struct nouveau_engine base;
 
-	struct nouveau_event *cevent; /* channel creation event */
-	struct nouveau_event *uevent; /* async user trigger */
+	struct nvkm_event cevent; /* channel creation event */
+	struct nvkm_event uevent; /* async user trigger */
 
 	struct nouveau_object **channel;
 	spinlock_t lock;
@@ -112,6 +116,10 @@ extern struct nouveau_oclass *nve0_fifo_oclass;
 extern struct nouveau_oclass *gk20a_fifo_oclass;
 extern struct nouveau_oclass *nv108_fifo_oclass;
 
+int  nouveau_fifo_uevent_ctor(struct nouveau_object *, void *, u32,
+			      struct nvkm_notify *);
+void nouveau_fifo_uevent(struct nouveau_fifo *);
+
 void nv04_fifo_intr(struct nouveau_subdev *);
 int  nv04_fifo_context_attach(struct nouveau_object *, struct nouveau_object *);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/graph.h b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
index 8c1d4772da0c..d5055570d01b 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/graph.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
@@ -70,6 +70,7 @@ extern struct nouveau_oclass *nvd9_graph_oclass;
 extern struct nouveau_oclass *nve4_graph_oclass;
 extern struct nouveau_oclass *gk20a_graph_oclass;
 extern struct nouveau_oclass *nvf0_graph_oclass;
+extern struct nouveau_oclass *gk110b_graph_oclass;
 extern struct nouveau_oclass *nv108_graph_oclass;
 extern struct nouveau_oclass *gm107_graph_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/perfmon.h b/drivers/gpu/drm/nouveau/core/include/engine/perfmon.h
index 49b0024910fe..88cc812baaa3 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/perfmon.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/perfmon.h
@@ -4,7 +4,6 @@
 #include <core/device.h>
 #include <core/engine.h>
 #include <core/engctx.h>
-#include <core/class.h>
 
 struct nouveau_perfdom;
 struct nouveau_perfctr;
diff --git a/drivers/gpu/drm/nouveau/core/include/nvif/class.h b/drivers/gpu/drm/nouveau/core/include/nvif/class.h
new file mode 120000
index 000000000000..f1ac4859edd4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/nvif/class.h
@@ -0,0 +1 @@
+../../../nvif/class.h
\ No newline at end of file
diff --git a/drivers/gpu/drm/nouveau/core/include/nvif/event.h b/drivers/gpu/drm/nouveau/core/include/nvif/event.h
new file mode 120000
index 000000000000..1b798538a725
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/nvif/event.h
@@ -0,0 +1 @@
+../../../nvif/event.h
\ No newline at end of file
diff --git a/drivers/gpu/drm/nouveau/core/include/nvif/ioctl.h b/drivers/gpu/drm/nouveau/core/include/nvif/ioctl.h
new file mode 120000
index 000000000000..8569c86907c5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/nvif/ioctl.h
@@ -0,0 +1 @@
+../../../nvif/ioctl.h
\ No newline at end of file
diff --git a/drivers/gpu/drm/nouveau/core/include/nvif/unpack.h b/drivers/gpu/drm/nouveau/core/include/nvif/unpack.h
new file mode 120000
index 000000000000..69d99292bca4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/nvif/unpack.h
@@ -0,0 +1 @@
+../../../nvif/unpack.h
\ No newline at end of file
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
index 9faa98e67ad8..257ddf6d36d4 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bar.h
@@ -12,7 +12,6 @@ struct nouveau_bar {
 
 	int (*alloc)(struct nouveau_bar *, struct nouveau_object *,
 		     struct nouveau_mem *, struct nouveau_object **);
-	void __iomem *iomem;
 
 	int (*kmap)(struct nouveau_bar *, struct nouveau_mem *,
 		    u32 flags, struct nouveau_vma *);
@@ -20,6 +19,9 @@ struct nouveau_bar {
 		    u32 flags, struct nouveau_vma *);
 	void (*unmap)(struct nouveau_bar *, struct nouveau_vma *);
 	void (*flush)(struct nouveau_bar *);
+
+	/* whether the BAR supports to be ioremapped WC or should be uncached */
+	bool iomap_uncached;
 };
 
 static inline struct nouveau_bar *
@@ -30,5 +32,6 @@ nouveau_bar(void *obj)
 
 extern struct nouveau_oclass nv50_bar_oclass;
 extern struct nouveau_oclass nvc0_bar_oclass;
+extern struct nouveau_oclass gk20a_bar_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h
new file mode 100644
index 000000000000..e171120cec81
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0205.h
@@ -0,0 +1,32 @@
+#ifndef __NVBIOS_M0205_H__
+#define __NVBIOS_M0205_H__
+
+struct nvbios_M0205T {
+	u16 freq;
+};
+
+u32 nvbios_M0205Te(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz);
+u32 nvbios_M0205Tp(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz,
+		   struct nvbios_M0205T *);
+
+struct nvbios_M0205E {
+	u8 type;
+};
+
+u32 nvbios_M0205Ee(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_M0205Ep(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		   struct nvbios_M0205E *);
+
+struct nvbios_M0205S {
+	u8 data;
+};
+
+u32 nvbios_M0205Se(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_M0205Sp(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr,
+		   struct nvbios_M0205S *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h
new file mode 100644
index 000000000000..67dc50d837bc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/M0209.h
@@ -0,0 +1,30 @@
+#ifndef __NVBIOS_M0209_H__
+#define __NVBIOS_M0209_H__
+
+u32 nvbios_M0209Te(struct nouveau_bios *,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz);
+
+struct nvbios_M0209E {
+	u8 v00_40;
+	u8 bits;
+	u8 modulo;
+	u8 v02_40;
+	u8 v02_07;
+	u8 v03;
+};
+
+u32 nvbios_M0209Ee(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_M0209Ep(struct nouveau_bios *, int idx,
+		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		   struct nvbios_M0209E *);
+
+struct nvbios_M0209S {
+	u32 data[0x200];
+};
+
+u32 nvbios_M0209Se(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr);
+u32 nvbios_M0209Sp(struct nouveau_bios *, int ent, int idx, u8 *ver, u8 *hdr,
+		   struct nvbios_M0209S *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h
new file mode 100644
index 000000000000..119d0874e041
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/fan.h
@@ -0,0 +1,8 @@
+#ifndef __NVBIOS_FAN_H__
+#define __NVBIOS_FAN_H__
+
+#include <subdev/bios/therm.h>
+
+u16 nvbios_fan_parse(struct nouveau_bios *bios, struct nvbios_therm_fan *fan);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
index c086ac6d677d..a685bbd04568 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/ramcfg.h
@@ -4,60 +4,118 @@
 struct nouveau_bios;
 
 struct nvbios_ramcfg {
-	unsigned rammap_11_08_01:1;
-	unsigned rammap_11_08_0c:2;
-	unsigned rammap_11_08_10:1;
-	unsigned rammap_11_11_0c:2;
+	unsigned rammap_ver;
+	unsigned rammap_hdr;
+	unsigned rammap_min;
+	unsigned rammap_max;
+	union {
+		struct {
+			unsigned rammap_10_04_02:1;
+			unsigned rammap_10_04_08:1;
+		};
+		struct {
+			unsigned rammap_11_08_01:1;
+			unsigned rammap_11_08_0c:2;
+			unsigned rammap_11_08_10:1;
+			unsigned rammap_11_09_01ff:9;
+			unsigned rammap_11_0a_03fe:9;
+			unsigned rammap_11_0a_0400:1;
+			unsigned rammap_11_0a_0800:1;
+			unsigned rammap_11_0b_01f0:5;
+			unsigned rammap_11_0b_0200:1;
+			unsigned rammap_11_0b_0400:1;
+			unsigned rammap_11_0b_0800:1;
+			unsigned rammap_11_0d:8;
+			unsigned rammap_11_0e:8;
+			unsigned rammap_11_0f:8;
+			unsigned rammap_11_11_0c:2;
+		};
+	};
 
-	unsigned ramcfg_11_01_01:1;
-	unsigned ramcfg_11_01_02:1;
-	unsigned ramcfg_11_01_04:1;
-	unsigned ramcfg_11_01_08:1;
-	unsigned ramcfg_11_01_10:1;
-	unsigned ramcfg_11_01_20:1;
-	unsigned ramcfg_11_01_40:1;
-	unsigned ramcfg_11_01_80:1;
-	unsigned ramcfg_11_02_03:2;
-	unsigned ramcfg_11_02_04:1;
-	unsigned ramcfg_11_02_08:1;
-	unsigned ramcfg_11_02_10:1;
-	unsigned ramcfg_11_02_40:1;
-	unsigned ramcfg_11_02_80:1;
-	unsigned ramcfg_11_03_0f:4;
-	unsigned ramcfg_11_03_30:2;
-	unsigned ramcfg_11_03_c0:2;
-	unsigned ramcfg_11_03_f0:4;
-	unsigned ramcfg_11_04:8;
-	unsigned ramcfg_11_06:8;
-	unsigned ramcfg_11_07_02:1;
-	unsigned ramcfg_11_07_04:1;
-	unsigned ramcfg_11_07_08:1;
-	unsigned ramcfg_11_07_10:1;
-	unsigned ramcfg_11_07_40:1;
-	unsigned ramcfg_11_07_80:1;
-	unsigned ramcfg_11_08_01:1;
-	unsigned ramcfg_11_08_02:1;
-	unsigned ramcfg_11_08_04:1;
-	unsigned ramcfg_11_08_08:1;
-	unsigned ramcfg_11_08_10:1;
-	unsigned ramcfg_11_08_20:1;
-	unsigned ramcfg_11_09:8;
+	unsigned ramcfg_ver;
+	unsigned ramcfg_hdr;
+	unsigned ramcfg_timing;
+	union {
+		struct {
+			unsigned ramcfg_10_02_01:1;
+			unsigned ramcfg_10_02_02:1;
+			unsigned ramcfg_10_02_04:1;
+			unsigned ramcfg_10_02_08:1;
+			unsigned ramcfg_10_02_10:1;
+			unsigned ramcfg_10_02_20:1;
+			unsigned ramcfg_10_02_40:1;
+			unsigned ramcfg_10_03_0f:4;
+			unsigned ramcfg_10_05:8;
+			unsigned ramcfg_10_06:8;
+			unsigned ramcfg_10_07:8;
+			unsigned ramcfg_10_08:8;
+			unsigned ramcfg_10_09_0f:4;
+			unsigned ramcfg_10_09_f0:4;
+		};
+		struct {
+			unsigned ramcfg_11_01_01:1;
+			unsigned ramcfg_11_01_02:1;
+			unsigned ramcfg_11_01_04:1;
+			unsigned ramcfg_11_01_08:1;
+			unsigned ramcfg_11_01_10:1;
+			unsigned ramcfg_11_01_20:1;
+			unsigned ramcfg_11_01_40:1;
+			unsigned ramcfg_11_01_80:1;
+			unsigned ramcfg_11_02_03:2;
+			unsigned ramcfg_11_02_04:1;
+			unsigned ramcfg_11_02_08:1;
+			unsigned ramcfg_11_02_10:1;
+			unsigned ramcfg_11_02_40:1;
+			unsigned ramcfg_11_02_80:1;
+			unsigned ramcfg_11_03_0f:4;
+			unsigned ramcfg_11_03_30:2;
+			unsigned ramcfg_11_03_c0:2;
+			unsigned ramcfg_11_03_f0:4;
+			unsigned ramcfg_11_04:8;
+			unsigned ramcfg_11_06:8;
+			unsigned ramcfg_11_07_02:1;
+			unsigned ramcfg_11_07_04:1;
+			unsigned ramcfg_11_07_08:1;
+			unsigned ramcfg_11_07_10:1;
+			unsigned ramcfg_11_07_40:1;
+			unsigned ramcfg_11_07_80:1;
+			unsigned ramcfg_11_08_01:1;
+			unsigned ramcfg_11_08_02:1;
+			unsigned ramcfg_11_08_04:1;
+			unsigned ramcfg_11_08_08:1;
+			unsigned ramcfg_11_08_10:1;
+			unsigned ramcfg_11_08_20:1;
+			unsigned ramcfg_11_09:8;
+		};
+	};
 
+	unsigned timing_ver;
+	unsigned timing_hdr;
 	unsigned timing[11];
-	unsigned timing_20_2e_03:2;
-	unsigned timing_20_2e_30:2;
-	unsigned timing_20_2e_c0:2;
-	unsigned timing_20_2f_03:2;
-	unsigned timing_20_2c_003f:6;
-	unsigned timing_20_2c_1fc0:7;
-	unsigned timing_20_30_f8:5;
-	unsigned timing_20_30_07:3;
-	unsigned timing_20_31_0007:3;
-	unsigned timing_20_31_0078:4;
-	unsigned timing_20_31_0780:4;
-	unsigned timing_20_31_0800:1;
-	unsigned timing_20_31_7000:3;
-	unsigned timing_20_31_8000:1;
+	union {
+		struct {
+			unsigned timing_10_WR:8;
+			unsigned timing_10_CL:8;
+			unsigned timing_10_ODT:3;
+			unsigned timing_10_CWL:8;
+		};
+		struct {
+			unsigned timing_20_2e_03:2;
+			unsigned timing_20_2e_30:2;
+			unsigned timing_20_2e_c0:2;
+			unsigned timing_20_2f_03:2;
+			unsigned timing_20_2c_003f:6;
+			unsigned timing_20_2c_1fc0:7;
+			unsigned timing_20_30_f8:5;
+			unsigned timing_20_30_07:3;
+			unsigned timing_20_31_0007:3;
+			unsigned timing_20_31_0078:4;
+			unsigned timing_20_31_0780:4;
+			unsigned timing_20_31_0800:1;
+			unsigned timing_20_31_7000:3;
+			unsigned timing_20_31_8000:1;
+		};
+	};
 };
 
 u8 nvbios_ramcfg_count(struct nouveau_bios *);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
index 5bdf8e4db40a..47e021d3e20d 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/rammap.h
@@ -8,9 +8,10 @@ u32 nvbios_rammapTe(struct nouveau_bios *, u8 *ver, u8 *hdr,
 
 u32 nvbios_rammapEe(struct nouveau_bios *, int idx,
 		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
+u32 nvbios_rammapEp(struct nouveau_bios *, int idx,
+		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		    struct nvbios_ramcfg *);
 u32 nvbios_rammapEm(struct nouveau_bios *, u16 mhz,
-		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len);
-u32 nvbios_rammapEp(struct nouveau_bios *, u16 mhz,
 		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
 		    struct nvbios_ramcfg *);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
index 8dc5051df55d..295d093f3b30 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/bios/therm.h
@@ -23,6 +23,12 @@ struct nvbios_therm_sensor {
 	struct nvbios_therm_threshold thrs_shutdown;
 };
 
+enum nvbios_therm_fan_type {
+	NVBIOS_THERM_FAN_UNK = 0,
+	NVBIOS_THERM_FAN_TOGGLE = 1,
+	NVBIOS_THERM_FAN_PWM = 2,
+};
+
 /* no vbios have more than 6 */
 #define NOUVEAU_TEMP_FAN_TRIP_MAX 10
 struct nouveau_therm_trip_point {
@@ -38,7 +44,9 @@ enum nvbios_therm_fan_mode {
 };
 
 struct nvbios_therm_fan {
-	u16 pwm_freq;
+	enum nvbios_therm_fan_type type;
+
+	u32 pwm_freq;
 
 	u8 min_duty;
 	u8 max_duty;
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
index c01e29c9f89a..36ed035d4d42 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
@@ -29,6 +29,7 @@ enum nv_clk_src {
 	nv_clk_src_mdiv,
 
 	nv_clk_src_core,
+	nv_clk_src_core_intm,
 	nv_clk_src_shader,
 
 	nv_clk_src_mem,
@@ -71,8 +72,15 @@ struct nouveau_clock {
 	struct list_head states;
 	int state_nr;
 
+	struct work_struct work;
+	wait_queue_head_t wait;
+	atomic_t waiting;
+
+	struct nvkm_notify pwrsrc_ntfy;
+	int pwrsrc;
 	int pstate; /* current */
-	int ustate; /* user-requested (-1 disabled, -2 perfmon) */
+	int ustate_ac; /* user-requested (-1 disabled, -2 perfmon) */
+	int ustate_dc; /* user-requested (-1 disabled, -2 perfmon) */
 	int astate; /* perfmon adjustment (base) */
 	int tstate; /* thermal adjustment (max-) */
 	int dstate; /* display adjustment (min+) */
@@ -108,8 +116,9 @@ struct nouveau_clocks {
 	int mdiv;
 };
 
-#define nouveau_clock_create(p,e,o,i,r,d)                                      \
-	nouveau_clock_create_((p), (e), (o), (i), (r), sizeof(**d), (void **)d)
+#define nouveau_clock_create(p,e,o,i,r,s,n,d)                                  \
+	nouveau_clock_create_((p), (e), (o), (i), (r), (s), (n), sizeof(**d),  \
+			      (void **)d)
 #define nouveau_clock_destroy(p) ({                                            \
 	struct nouveau_clock *clk = (p);                                       \
 	_nouveau_clock_dtor(nv_object(clk));                                   \
@@ -118,15 +127,18 @@ struct nouveau_clocks {
 	struct nouveau_clock *clk = (p);                                       \
 	_nouveau_clock_init(nv_object(clk));                                   \
 })
-#define nouveau_clock_fini(p,s)                                                \
-	nouveau_subdev_fini(&(p)->base, (s))
+#define nouveau_clock_fini(p,s) ({                                             \
+	struct nouveau_clock *clk = (p);                                       \
+	_nouveau_clock_fini(nv_object(clk), (s));                              \
+})
 
 int  nouveau_clock_create_(struct nouveau_object *, struct nouveau_object *,
 			   struct nouveau_oclass *,
-			   struct nouveau_clocks *, bool, int, void **);
+			   struct nouveau_clocks *, struct nouveau_pstate *,
+			   int, bool, int, void **);
 void _nouveau_clock_dtor(struct nouveau_object *);
-int _nouveau_clock_init(struct nouveau_object *);
-#define _nouveau_clock_fini _nouveau_subdev_fini
+int  _nouveau_clock_init(struct nouveau_object *);
+int  _nouveau_clock_fini(struct nouveau_object *, bool);
 
 extern struct nouveau_oclass nv04_clock_oclass;
 extern struct nouveau_oclass nv40_clock_oclass;
@@ -136,6 +148,7 @@ extern struct nouveau_oclass *nvaa_clock_oclass;
 extern struct nouveau_oclass nva3_clock_oclass;
 extern struct nouveau_oclass nvc0_clock_oclass;
 extern struct nouveau_oclass nve0_clock_oclass;
+extern struct nouveau_oclass gk20a_clock_oclass;
 
 int nv04_clock_pll_set(struct nouveau_clock *, u32 type, u32 freq);
 int nv04_clock_pll_calc(struct nouveau_clock *, struct nvbios_pll *,
@@ -145,7 +158,7 @@ int nv04_clock_pll_prog(struct nouveau_clock *, u32 reg1,
 int nva3_clock_pll_calc(struct nouveau_clock *, struct nvbios_pll *,
 			int clk, struct nouveau_pll_vals *);
 
-int nouveau_clock_ustate(struct nouveau_clock *, int req);
+int nouveau_clock_ustate(struct nouveau_clock *, int req, int pwr);
 int nouveau_clock_astate(struct nouveau_clock *, int req, int rel);
 int nouveau_clock_dstate(struct nouveau_clock *, int req, int rel);
 int nouveau_clock_tstate(struct nouveau_clock *, int req, int rel);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
index 871e73914b24..8d0032f15205 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
@@ -111,6 +111,7 @@ extern struct nouveau_oclass *gm107_fb_oclass;
 #include <subdev/bios/ramcfg.h>
 
 struct nouveau_ram_data {
+	struct list_head head;
 	struct nvbios_ramcfg bios;
 	u32 freq;
 };
@@ -136,6 +137,7 @@ struct nouveau_ram {
 
 	int ranks;
 	int parts;
+	int part_mask;
 
 	int  (*get)(struct nouveau_fb *, u64 size, u32 align,
 		    u32 size_nc, u32 type, struct nouveau_mem **);
@@ -144,11 +146,6 @@ struct nouveau_ram {
 	int  (*calc)(struct nouveau_fb *, u32 freq);
 	int  (*prog)(struct nouveau_fb *);
 	void (*tidy)(struct nouveau_fb *);
-	struct {
-		u8  version;
-		u32 data;
-		u8  size;
-	} rammap, ramcfg, timing;
 	u32 freq;
 	u32 mr[16];
 	u32 mr1_nuts;
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h
new file mode 100644
index 000000000000..0f7fc0c52ab2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb/regsnv04.h
@@ -0,0 +1,21 @@
+#ifndef __NOUVEAU_FB_REGS_04_H__
+#define __NOUVEAU_FB_REGS_04_H__
+
+#define NV04_PFB_BOOT_0						0x00100000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
+#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
+#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
+#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h b/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h
new file mode 100644
index 000000000000..2b1ddb2a9a7d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fuse.h
@@ -0,0 +1,30 @@
+#ifndef __NOUVEAU_FUSE_H__
+#define __NOUVEAU_FUSE_H__
+
+#include <core/subdev.h>
+#include <core/device.h>
+
+struct nouveau_fuse {
+	struct nouveau_subdev base;
+};
+
+static inline struct nouveau_fuse *
+nouveau_fuse(void *obj)
+{
+	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_FUSE];
+}
+
+#define nouveau_fuse_create(p, e, o, d)                                        \
+	nouveau_fuse_create_((p), (e), (o), sizeof(**d), (void **)d)
+
+int  nouveau_fuse_create_(struct nouveau_object *, struct nouveau_object *,
+			  struct nouveau_oclass *, int, void **);
+void _nouveau_fuse_dtor(struct nouveau_object *);
+int  _nouveau_fuse_init(struct nouveau_object *);
+#define _nouveau_fuse_fini _nouveau_subdev_fini
+
+extern struct nouveau_oclass g80_fuse_oclass;
+extern struct nouveau_oclass gf100_fuse_oclass;
+extern struct nouveau_oclass gm107_fuse_oclass;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
index 612d82ab683d..f855140dbcb7 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/gpio.h
@@ -8,16 +8,22 @@
 #include <subdev/bios.h>
 #include <subdev/bios/gpio.h>
 
-enum nvkm_gpio_event {
-	NVKM_GPIO_HI = 1,
-	NVKM_GPIO_LO = 2,
-	NVKM_GPIO_TOGGLED = (NVKM_GPIO_HI | NVKM_GPIO_LO),
+struct nvkm_gpio_ntfy_req {
+#define NVKM_GPIO_HI                                                       0x01
+#define NVKM_GPIO_LO                                                       0x02
+#define NVKM_GPIO_TOGGLED                                                  0x03
+	u8 mask;
+	u8 line;
+};
+
+struct nvkm_gpio_ntfy_rep {
+	u8 mask;
 };
 
 struct nouveau_gpio {
 	struct nouveau_subdev base;
 
-	struct nouveau_event *events;
+	struct nvkm_event event;
 
 	void (*reset)(struct nouveau_gpio *, u8 func);
 	int  (*find)(struct nouveau_gpio *, int idx, u8 tag, u8 line,
@@ -34,7 +40,7 @@ nouveau_gpio(void *obj)
 
 extern struct nouveau_oclass *nv10_gpio_oclass;
 extern struct nouveau_oclass *nv50_gpio_oclass;
-extern struct nouveau_oclass *nv92_gpio_oclass;
+extern struct nouveau_oclass *nv94_gpio_oclass;
 extern struct nouveau_oclass *nvd0_gpio_oclass;
 extern struct nouveau_oclass *nve0_gpio_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
index 825f7bb46b67..1b937c2c25ae 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/i2c.h
@@ -14,15 +14,18 @@
 #define NV_I2C_TYPE_EXTDDC(e) (0x0005 | (e) << 8)
 #define NV_I2C_TYPE_EXTAUX(e) (0x0006 | (e) << 8)
 
-enum nvkm_i2c_event {
-	NVKM_I2C_PLUG = 1,
-	NVKM_I2C_UNPLUG = 2,
-	NVKM_I2C_IRQ = 4,
-	NVKM_I2C_DONE = 8,
-	NVKM_I2C_ANY = (NVKM_I2C_PLUG |
-			NVKM_I2C_UNPLUG |
-			NVKM_I2C_IRQ |
-			NVKM_I2C_DONE),
+struct nvkm_i2c_ntfy_req {
+#define NVKM_I2C_PLUG                                                      0x01
+#define NVKM_I2C_UNPLUG                                                    0x02
+#define NVKM_I2C_IRQ                                                       0x04
+#define NVKM_I2C_DONE                                                      0x08
+#define NVKM_I2C_ANY                                                       0x0f
+	u8 mask;
+	u8 port;
+};
+
+struct nvkm_i2c_ntfy_rep {
+	u8 mask;
 };
 
 struct nouveau_i2c_port {
@@ -56,7 +59,7 @@ struct nouveau_i2c_board_info {
 
 struct nouveau_i2c {
 	struct nouveau_subdev base;
-	struct nouveau_event *ntfy;
+	struct nvkm_event event;
 
 	struct nouveau_i2c_port *(*find)(struct nouveau_i2c *, u8 index);
 	struct nouveau_i2c_port *(*find_type)(struct nouveau_i2c *, u16 type);
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/ltc.h b/drivers/gpu/drm/nouveau/core/include/subdev/ltc.h
new file mode 100644
index 000000000000..b909a7363f6b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/ltc.h
@@ -0,0 +1,35 @@
+#ifndef __NOUVEAU_LTC_H__
+#define __NOUVEAU_LTC_H__
+
+#include <core/subdev.h>
+#include <core/device.h>
+
+#define NOUVEAU_LTC_MAX_ZBC_CNT 16
+
+struct nouveau_mm_node;
+
+struct nouveau_ltc {
+	struct nouveau_subdev base;
+
+	int  (*tags_alloc)(struct nouveau_ltc *, u32 count,
+	                   struct nouveau_mm_node **);
+	void (*tags_free)(struct nouveau_ltc *, struct nouveau_mm_node **);
+	void (*tags_clear)(struct nouveau_ltc *, u32 first, u32 count);
+
+	int zbc_min;
+	int zbc_max;
+	int (*zbc_color_get)(struct nouveau_ltc *, int index, const u32[4]);
+	int (*zbc_depth_get)(struct nouveau_ltc *, int index, const u32);
+};
+
+static inline struct nouveau_ltc *
+nouveau_ltc(void *obj)
+{
+	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_LTC];
+}
+
+extern struct nouveau_oclass *gf100_ltc_oclass;
+extern struct nouveau_oclass *gk104_ltc_oclass;
+extern struct nouveau_oclass *gm107_ltc_oclass;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h b/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h
deleted file mode 100644
index c9c1950b7743..000000000000
--- a/drivers/gpu/drm/nouveau/core/include/subdev/ltcg.h
+++ /dev/null
@@ -1,41 +0,0 @@
-#ifndef __NOUVEAU_LTCG_H__
-#define __NOUVEAU_LTCG_H__
-
-#include <core/subdev.h>
-#include <core/device.h>
-
-struct nouveau_mm_node;
-
-struct nouveau_ltcg {
-	struct nouveau_subdev base;
-
-	int  (*tags_alloc)(struct nouveau_ltcg *, u32 count,
-	                   struct nouveau_mm_node **);
-	void (*tags_free)(struct nouveau_ltcg *, struct nouveau_mm_node **);
-	void (*tags_clear)(struct nouveau_ltcg *, u32 first, u32 count);
-};
-
-static inline struct nouveau_ltcg *
-nouveau_ltcg(void *obj)
-{
-	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_LTCG];
-}
-
-#define nouveau_ltcg_create(p,e,o,d)                                           \
-	nouveau_subdev_create_((p), (e), (o), 0, "PLTCG", "level2",            \
-			       sizeof(**d), (void **)d)
-#define nouveau_ltcg_destroy(p)                                                \
-	nouveau_subdev_destroy(&(p)->base)
-#define nouveau_ltcg_init(p)                                                   \
-	nouveau_subdev_init(&(p)->base)
-#define nouveau_ltcg_fini(p,s)                                                 \
-	nouveau_subdev_fini(&(p)->base, (s))
-
-#define _nouveau_ltcg_dtor _nouveau_subdev_dtor
-#define _nouveau_ltcg_init _nouveau_subdev_init
-#define _nouveau_ltcg_fini _nouveau_subdev_fini
-
-extern struct nouveau_oclass *gf100_ltcg_oclass;
-extern struct nouveau_oclass *gm107_ltcg_oclass;
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/mc.h b/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
index 72b176831be6..568e4dfc5e9e 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/mc.h
@@ -4,15 +4,11 @@
 #include <core/subdev.h>
 #include <core/device.h>
 
-struct nouveau_mc_intr {
-	u32 stat;
-	u32 unit;
-};
-
 struct nouveau_mc {
 	struct nouveau_subdev base;
 	bool use_msi;
 	unsigned int irq;
+	void (*unk260)(struct nouveau_mc *, u32);
 };
 
 static inline struct nouveau_mc *
@@ -21,30 +17,6 @@ nouveau_mc(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_MC];
 }
 
-#define nouveau_mc_create(p,e,o,d)                                             \
-	nouveau_mc_create_((p), (e), (o), sizeof(**d), (void **)d)
-#define nouveau_mc_destroy(p) ({                                               \
-	struct nouveau_mc *pmc = (p); _nouveau_mc_dtor(nv_object(pmc));        \
-})
-#define nouveau_mc_init(p) ({                                                  \
-	struct nouveau_mc *pmc = (p); _nouveau_mc_init(nv_object(pmc));        \
-})
-#define nouveau_mc_fini(p,s) ({                                                \
-	struct nouveau_mc *pmc = (p); _nouveau_mc_fini(nv_object(pmc), (s));   \
-})
-
-int  nouveau_mc_create_(struct nouveau_object *, struct nouveau_object *,
-			struct nouveau_oclass *, int, void **);
-void _nouveau_mc_dtor(struct nouveau_object *);
-int  _nouveau_mc_init(struct nouveau_object *);
-int  _nouveau_mc_fini(struct nouveau_object *, bool);
-
-struct nouveau_mc_oclass {
-	struct nouveau_oclass base;
-	const struct nouveau_mc_intr *intr;
-	void (*msi_rearm)(struct nouveau_mc *);
-};
-
 extern struct nouveau_oclass *nv04_mc_oclass;
 extern struct nouveau_oclass *nv40_mc_oclass;
 extern struct nouveau_oclass *nv44_mc_oclass;
@@ -54,5 +26,6 @@ extern struct nouveau_oclass *nv94_mc_oclass;
 extern struct nouveau_oclass *nv98_mc_oclass;
 extern struct nouveau_oclass *nvc0_mc_oclass;
 extern struct nouveau_oclass *nvc3_mc_oclass;
+extern struct nouveau_oclass *gk20a_mc_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h b/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
index c5c92cbed33f..bf3d1f611333 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/pwr.h
@@ -8,18 +8,6 @@ struct nouveau_pwr {
 	struct nouveau_subdev base;
 
 	struct {
-		u32 limit;
-		u32 *data;
-		u32  size;
-	} code;
-
-	struct {
-		u32 limit;
-		u32 *data;
-		u32  size;
-	} data;
-
-	struct {
 		u32 base;
 		u32 size;
 	} send;
@@ -35,7 +23,8 @@ struct nouveau_pwr {
 		u32 data[2];
 	} recv;
 
-	int (*message)(struct nouveau_pwr *, u32[2], u32, u32, u32, u32);
+	int  (*message)(struct nouveau_pwr *, u32[2], u32, u32, u32, u32);
+	void (*pgob)(struct nouveau_pwr *, bool);
 };
 
 static inline struct nouveau_pwr *
@@ -44,29 +33,11 @@ nouveau_pwr(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_PWR];
 }
 
-#define nouveau_pwr_create(p, e, o, d)                                         \
-	nouveau_pwr_create_((p), (e), (o), sizeof(**d), (void **)d)
-#define nouveau_pwr_destroy(p)                                                 \
-	nouveau_subdev_destroy(&(p)->base)
-#define nouveau_pwr_init(p) ({                                                 \
-	struct nouveau_pwr *ppwr = (p);                                        \
-	_nouveau_pwr_init(nv_object(ppwr));                                    \
-})
-#define nouveau_pwr_fini(p,s) ({                                               \
-	struct nouveau_pwr *ppwr = (p);                                        \
-	_nouveau_pwr_fini(nv_object(ppwr), (s));                               \
-})
-
-int nouveau_pwr_create_(struct nouveau_object *, struct nouveau_object *,
-			   struct nouveau_oclass *, int, void **);
-#define _nouveau_pwr_dtor _nouveau_subdev_dtor
-int _nouveau_pwr_init(struct nouveau_object *);
-int _nouveau_pwr_fini(struct nouveau_object *, bool);
-
-extern struct nouveau_oclass nva3_pwr_oclass;
-extern struct nouveau_oclass nvc0_pwr_oclass;
-extern struct nouveau_oclass nvd0_pwr_oclass;
-extern struct nouveau_oclass nv108_pwr_oclass;
+extern struct nouveau_oclass *nva3_pwr_oclass;
+extern struct nouveau_oclass *nvc0_pwr_oclass;
+extern struct nouveau_oclass *nvd0_pwr_oclass;
+extern struct nouveau_oclass *gk104_pwr_oclass;
+extern struct nouveau_oclass *nv108_pwr_oclass;
 
 /* interface to MEMX process running on PPWR */
 struct nouveau_memx;
@@ -76,5 +47,8 @@ void nouveau_memx_wr32(struct nouveau_memx *, u32 addr, u32 data);
 void nouveau_memx_wait(struct nouveau_memx *,
 		       u32 addr, u32 mask, u32 data, u32 nsec);
 void nouveau_memx_nsec(struct nouveau_memx *, u32 nsec);
+void nouveau_memx_wait_vblank(struct nouveau_memx *);
+void nouveau_memx_block(struct nouveau_memx *);
+void nouveau_memx_unblock(struct nouveau_memx *);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
index d4a68179e586..a437597dcafc 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/therm.h
@@ -78,5 +78,6 @@ extern struct nouveau_oclass nv50_therm_oclass;
 extern struct nouveau_oclass nv84_therm_oclass;
 extern struct nouveau_oclass nva3_therm_oclass;
 extern struct nouveau_oclass nvd0_therm_oclass;
+extern struct nouveau_oclass gm107_therm_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/os.h b/drivers/gpu/drm/nouveau/core/os.h
index d0ced94ca54c..ccfa21d72ddc 100644
--- a/drivers/gpu/drm/nouveau/core/os.h
+++ b/drivers/gpu/drm/nouveau/core/os.h
@@ -21,6 +21,8 @@
 #include <linux/interrupt.h>
 #include <linux/log2.h>
 #include <linux/pm_runtime.h>
+#include <linux/power_supply.h>
+#include <linux/clk.h>
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/base.c b/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
index 73b1ed20c8d5..b1adc69efd88 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/base.c
@@ -38,10 +38,12 @@ struct nouveau_barobj {
 static int
 nouveau_barobj_ctor(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
-		    struct nouveau_oclass *oclass, void *mem, u32 size,
+		    struct nouveau_oclass *oclass, void *data, u32 size,
 		    struct nouveau_object **pobject)
 {
+	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_bar *bar = (void *)engine;
+	struct nouveau_mem *mem = data;
 	struct nouveau_barobj *barobj;
 	int ret;
 
@@ -54,7 +56,13 @@ nouveau_barobj_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	barobj->iomem = bar->iomem + (u32)barobj->vma.offset;
+	barobj->iomem = ioremap(nv_device_resource_start(device, 3) +
+				(u32)barobj->vma.offset, mem->size << 12);
+	if (!barobj->iomem) {
+		nv_warn(bar, "PRAMIN ioremap failed\n");
+		return -ENOMEM;
+	}
+
 	return 0;
 }
 
@@ -63,8 +71,11 @@ nouveau_barobj_dtor(struct nouveau_object *object)
 {
 	struct nouveau_bar *bar = (void *)object->engine;
 	struct nouveau_barobj *barobj = (void *)object;
-	if (barobj->vma.node)
+	if (barobj->vma.node) {
+		if (barobj->iomem)
+			iounmap(barobj->iomem);
 		bar->unmap(bar, &barobj->vma);
+	}
 	nouveau_object_destroy(&barobj->base);
 }
 
@@ -99,8 +110,12 @@ nouveau_bar_alloc(struct nouveau_bar *bar, struct nouveau_object *parent,
 		  struct nouveau_mem *mem, struct nouveau_object **pobject)
 {
 	struct nouveau_object *engine = nv_object(bar);
-	return nouveau_object_ctor(parent, engine, &nouveau_barobj_oclass,
-				   mem, 0, pobject);
+	struct nouveau_object *gpuobj;
+	int ret = nouveau_object_ctor(parent, engine, &nouveau_barobj_oclass,
+				      mem, 0, &gpuobj);
+	if (ret == 0)
+		*pobject = gpuobj;
+	return ret;
 }
 
 int
@@ -108,7 +123,6 @@ nouveau_bar_create_(struct nouveau_object *parent,
 		    struct nouveau_object *engine,
 		    struct nouveau_oclass *oclass, int length, void **pobject)
 {
-	struct nouveau_device *device = nv_device(parent);
 	struct nouveau_bar *bar;
 	int ret;
 
@@ -118,18 +132,12 @@ nouveau_bar_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	if (nv_device_resource_len(device, 3) != 0)
-		bar->iomem = ioremap(nv_device_resource_start(device, 3),
-				     nv_device_resource_len(device, 3));
-
 	return 0;
 }
 
 void
 nouveau_bar_destroy(struct nouveau_bar *bar)
 {
-	if (bar->iomem)
-		iounmap(bar->iomem);
 	nouveau_subdev_destroy(&bar->base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
new file mode 100644
index 000000000000..bf877af9d3bd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/gk20a.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <subdev/bar.h>
+
+#include "priv.h"
+
+int
+gk20a_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct nouveau_bar *bar;
+	int ret;
+
+	ret = nvc0_bar_ctor(parent, engine, oclass, data, size, pobject);
+	if (ret)
+		return ret;
+
+	bar = (struct nouveau_bar *)*pobject;
+	bar->iomap_uncached = true;
+
+	return 0;
+}
+
+struct nouveau_oclass
+gk20a_bar_oclass = {
+	.handle = NV_SUBDEV(BAR, 0xea),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gk20a_bar_ctor,
+		.dtor = nvc0_bar_dtor,
+		.init = nvc0_bar_init,
+		.fini = _nouveau_bar_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
index ca8139b9ab27..05a278bab247 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
@@ -133,7 +133,7 @@ nvc0_bar_init_vm(struct nvc0_bar_priv *priv, struct nvc0_bar_priv_vm *bar_vm,
 	return 0;
 }
 
-static int
+int
 nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	      struct nouveau_oclass *oclass, void *data, u32 size,
 	      struct nouveau_object **pobject)
@@ -169,7 +169,7 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	return 0;
 }
 
-static void
+void
 nvc0_bar_dtor(struct nouveau_object *object)
 {
 	struct nvc0_bar_priv *priv = (void *)object;
@@ -188,7 +188,7 @@ nvc0_bar_dtor(struct nouveau_object *object)
 	nouveau_bar_destroy(&priv->base);
 }
 
-static int
+int
 nvc0_bar_init(struct nouveau_object *object)
 {
 	struct nvc0_bar_priv *priv = (void *)object;
@@ -200,7 +200,6 @@ nvc0_bar_init(struct nouveau_object *object)
 
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000000);
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000100);
-	nv_mask(priv, 0x100c80, 0x00000001, 0x00000000);
 
 	nv_wr32(priv, 0x001704, 0x80000000 | priv->bar[1].mem->addr >> 12);
 	if (priv->bar[0].mem)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h b/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
index ffad8f337ead..3ee8b1476d00 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/priv.h
@@ -23,4 +23,10 @@ int  nouveau_bar_alloc(struct nouveau_bar *, struct nouveau_object *,
 
 void nv84_bar_flush(struct nouveau_bar *);
 
+int nvc0_bar_ctor(struct nouveau_object *, struct nouveau_object *,
+		  struct nouveau_oclass *, void *, u32,
+		  struct nouveau_object **);
+void nvc0_bar_dtor(struct nouveau_object *);
+int nvc0_bar_init(struct nouveau_object *);
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c b/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c
new file mode 100644
index 000000000000..ac9617c5fc2a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/M0205.c
@@ -0,0 +1,136 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/M0205.h>
+
+u32
+nvbios_M0205Te(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz)
+{
+	struct bit_entry bit_M;
+	u32 data = 0x00000000;
+
+	if (!bit_entry(bios, 'M', &bit_M)) {
+		if (bit_M.version == 2 && bit_M.length > 0x08)
+			data = nv_ro32(bios, bit_M.offset + 0x05);
+		if (data) {
+			*ver = nv_ro08(bios, data + 0x00);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, data + 0x01);
+				*len = nv_ro08(bios, data + 0x02);
+				*ssz = nv_ro08(bios, data + 0x03);
+				*snr = nv_ro08(bios, data + 0x04);
+				*cnt = nv_ro08(bios, data + 0x05);
+				return data;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Tp(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz,
+	       struct nvbios_M0205T *info)
+{
+	u32 data = nvbios_M0205Te(bios, ver, hdr, cnt, len, snr, ssz);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->freq = nv_ro16(bios, data + 0x06);
+		break;
+	default:
+		break;
+	}
+	return data;
+}
+
+u32
+nvbios_M0205Ee(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u8  snr, ssz;
+	u32 data = nvbios_M0205Te(bios, ver, hdr, cnt, len, &snr, &ssz);
+	if (data && idx < *cnt) {
+		data = data + *hdr + idx * (*len + (snr * ssz));
+		*hdr = *len;
+		*cnt = snr;
+		*len = ssz;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Ep(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_M0205E *info)
+{
+	u32 data = nvbios_M0205Ee(bios, idx, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->type = nv_ro08(bios, data + 0x00) & 0x0f;
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Se(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr)
+{
+
+	u8  cnt, len;
+	u32 data = nvbios_M0205Ee(bios, ent, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + idx * len;
+		*hdr = len;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0205Sp(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr,
+	       struct nvbios_M0205S *info)
+{
+	u32 data = nvbios_M0205Se(bios, ent, idx, ver, hdr);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->data = nv_ro08(bios, data + 0x00);
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c b/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c
new file mode 100644
index 000000000000..b142a510e89f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/M0209.c
@@ -0,0 +1,137 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/M0209.h>
+
+u32
+nvbios_M0209Te(struct nouveau_bios *bios,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz)
+{
+	struct bit_entry bit_M;
+	u32 data = 0x00000000;
+
+	if (!bit_entry(bios, 'M', &bit_M)) {
+		if (bit_M.version == 2 && bit_M.length > 0x0c)
+			data = nv_ro32(bios, bit_M.offset + 0x09);
+		if (data) {
+			*ver = nv_ro08(bios, data + 0x00);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, data + 0x01);
+				*len = nv_ro08(bios, data + 0x02);
+				*ssz = nv_ro08(bios, data + 0x03);
+				*snr = 1;
+				*cnt = nv_ro08(bios, data + 0x04);
+				return data;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Ee(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	u8  snr, ssz;
+	u32 data = nvbios_M0209Te(bios, ver, hdr, cnt, len, &snr, &ssz);
+	if (data && idx < *cnt) {
+		data = data + *hdr + idx * (*len + (snr * ssz));
+		*hdr = *len;
+		*cnt = snr;
+		*len = ssz;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Ep(struct nouveau_bios *bios, int idx,
+	       u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+	       struct nvbios_M0209E *info)
+{
+	u32 data = nvbios_M0209Ee(bios, idx, ver, hdr, cnt, len);
+	memset(info, 0x00, sizeof(*info));
+	switch (!!data * *ver) {
+	case 0x10:
+		info->v00_40 = (nv_ro08(bios, data + 0x00) & 0x40) >> 6;
+		info->bits   =  nv_ro08(bios, data + 0x00) & 0x3f;
+		info->modulo =  nv_ro08(bios, data + 0x01);
+		info->v02_40 = (nv_ro08(bios, data + 0x02) & 0x40) >> 6;
+		info->v02_07 =  nv_ro08(bios, data + 0x02) & 0x07;
+		info->v03    =  nv_ro08(bios, data + 0x03);
+		return data;
+	default:
+		break;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Se(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr)
+{
+
+	u8  cnt, len;
+	u32 data = nvbios_M0209Ee(bios, ent, ver, hdr, &cnt, &len);
+	if (data && idx < cnt) {
+		data = data + *hdr + idx * len;
+		*hdr = len;
+		return data;
+	}
+	return 0x00000000;
+}
+
+u32
+nvbios_M0209Sp(struct nouveau_bios *bios, int ent, int idx, u8 *ver, u8 *hdr,
+	       struct nvbios_M0209S *info)
+{
+	struct nvbios_M0209E M0209E;
+	u8  cnt, len;
+	u32 data = nvbios_M0209Ep(bios, ent, ver, hdr, &cnt, &len, &M0209E);
+	if (data) {
+		u32 i, data = nvbios_M0209Se(bios, ent, idx, ver, hdr);
+		memset(info, 0x00, sizeof(*info));
+		switch (!!data * *ver) {
+		case 0x10:
+			for (i = 0; i < ARRAY_SIZE(info->data); i++) {
+				u32 bits = (i % M0209E.modulo) * M0209E.bits;
+				u32 mask = (1ULL << M0209E.bits) - 1;
+				u16  off = bits / 8;
+				u8   mod = bits % 8;
+				info->data[i] = nv_ro32(bios, data + off);
+				info->data[i] = info->data[i] >> mod;
+				info->data[i] = info->data[i] & mask;
+			}
+			return data;
+		default:
+			break;
+		}
+	}
+	return 0x00000000;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
index 88606bfaf847..bd8d348385b3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/dcb.c
@@ -124,6 +124,7 @@ dcb_outp_parse(struct nouveau_bios *bios, u8 idx, u8 *ver, u8 *len,
 	       struct dcb_output *outp)
 {
 	u16 dcb = dcb_outp(bios, idx, ver, len);
+	memset(outp, 0x00, sizeof(*outp));
 	if (dcb) {
 		if (*ver >= 0x20) {
 			u32 conn = nv_ro32(bios, dcb + 0x00);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c b/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c
new file mode 100644
index 000000000000..e419892240f5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/fan.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include <subdev/bios.h>
+#include <subdev/bios/bit.h>
+#include <subdev/bios/fan.h>
+
+u16
+nvbios_fan_table(struct nouveau_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
+{
+	struct bit_entry bit_P;
+	u16 fan = 0x0000;
+
+	if (!bit_entry(bios, 'P', &bit_P)) {
+		if (bit_P.version == 2 && bit_P.length >= 0x5a)
+			fan = nv_ro16(bios, bit_P.offset + 0x58);
+
+		if (fan) {
+			*ver = nv_ro08(bios, fan + 0);
+			switch (*ver) {
+			case 0x10:
+				*hdr = nv_ro08(bios, fan + 1);
+				*len = nv_ro08(bios, fan + 2);
+				*cnt = nv_ro08(bios, fan + 3);
+				return fan;
+			default:
+				break;
+			}
+		}
+	}
+
+	return 0x0000;
+}
+
+u16
+nvbios_fan_entry(struct nouveau_bios *bios, int idx, u8 *ver, u8 *hdr,
+		 u8 *cnt, u8 *len)
+{
+	u16 data = nvbios_fan_table(bios, ver, hdr, cnt, len);
+	if (data && idx < *cnt)
+		return data + *hdr + (idx * (*len));
+	return 0x0000;
+}
+
+u16
+nvbios_fan_parse(struct nouveau_bios *bios, struct nvbios_therm_fan *fan)
+{
+	u8 ver, hdr, cnt, len;
+
+	u16 data = nvbios_fan_entry(bios, 0, &ver, &hdr, &cnt, &len);
+	if (data) {
+		u8 type = nv_ro08(bios, data + 0x00);
+		switch (type) {
+		case 0:
+			fan->type = NVBIOS_THERM_FAN_TOGGLE;
+			break;
+		case 1:
+		case 2:
+			/* TODO: Understand the difference between the two! */
+			fan->type = NVBIOS_THERM_FAN_PWM;
+			break;
+		default:
+			fan->type = NVBIOS_THERM_FAN_UNK;
+		}
+
+		fan->min_duty = nv_ro08(bios, data + 0x02);
+		fan->max_duty = nv_ro08(bios, data + 0x03);
+
+		fan->pwm_freq = nv_ro32(bios, data + 0x0b) & 0xffffff;
+	}
+	return data;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c b/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
index 1811b2cb0472..585e69331ccc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/rammap.c
@@ -75,31 +75,39 @@ nvbios_rammapEe(struct nouveau_bios *bios, int idx,
 }
 
 u32
-nvbios_rammapEm(struct nouveau_bios *bios, u16 khz,
-		u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
-{
-	int idx = 0;
-	u32 data;
-	while ((data = nvbios_rammapEe(bios, idx++, ver, hdr, cnt, len))) {
-		if (khz >= nv_ro16(bios, data + 0x00) &&
-		    khz <= nv_ro16(bios, data + 0x02))
-			break;
-	}
-	return data;
-}
-
-u32
-nvbios_rammapEp(struct nouveau_bios *bios, u16 khz,
+nvbios_rammapEp(struct nouveau_bios *bios, int idx,
 		u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
 		struct nvbios_ramcfg *p)
 {
-	u32 data = nvbios_rammapEm(bios, khz, ver, hdr, cnt, len);
+	u32 data = nvbios_rammapEe(bios, idx, ver, hdr, cnt, len), temp;
 	memset(p, 0x00, sizeof(*p));
+	p->rammap_ver = *ver;
+	p->rammap_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->rammap_min      =  nv_ro16(bios, data + 0x00);
+		p->rammap_max      =  nv_ro16(bios, data + 0x02);
+		p->rammap_10_04_02 = (nv_ro08(bios, data + 0x04) & 0x02) >> 1;
+		p->rammap_10_04_08 = (nv_ro08(bios, data + 0x04) & 0x08) >> 3;
+		break;
 	case 0x11:
+		p->rammap_min      =  nv_ro16(bios, data + 0x00);
+		p->rammap_max      =  nv_ro16(bios, data + 0x02);
 		p->rammap_11_08_01 = (nv_ro08(bios, data + 0x08) & 0x01) >> 0;
 		p->rammap_11_08_0c = (nv_ro08(bios, data + 0x08) & 0x0c) >> 2;
 		p->rammap_11_08_10 = (nv_ro08(bios, data + 0x08) & 0x10) >> 4;
+		temp = nv_ro32(bios, data + 0x09);
+		p->rammap_11_09_01ff = (temp & 0x000001ff) >> 0;
+		p->rammap_11_0a_03fe = (temp & 0x0003fe00) >> 9;
+		p->rammap_11_0a_0400 = (temp & 0x00040000) >> 18;
+		p->rammap_11_0a_0800 = (temp & 0x00080000) >> 19;
+		p->rammap_11_0b_01f0 = (temp & 0x01f00000) >> 20;
+		p->rammap_11_0b_0200 = (temp & 0x02000000) >> 25;
+		p->rammap_11_0b_0400 = (temp & 0x04000000) >> 26;
+		p->rammap_11_0b_0800 = (temp & 0x08000000) >> 27;
+		p->rammap_11_0d    =  nv_ro08(bios, data + 0x0d);
+		p->rammap_11_0e    =  nv_ro08(bios, data + 0x0e);
+		p->rammap_11_0f    =  nv_ro08(bios, data + 0x0f);
 		p->rammap_11_11_0c = (nv_ro08(bios, data + 0x11) & 0x0c) >> 2;
 		break;
 	default:
@@ -110,6 +118,20 @@ nvbios_rammapEp(struct nouveau_bios *bios, u16 khz,
 }
 
 u32
+nvbios_rammapEm(struct nouveau_bios *bios, u16 mhz,
+		u8 *ver, u8 *hdr, u8 *cnt, u8 *len,
+		struct nvbios_ramcfg *info)
+{
+	int idx = 0;
+	u32 data;
+	while ((data = nvbios_rammapEp(bios, idx++, ver, hdr, cnt, len, info))) {
+		if (mhz >= info->rammap_min && mhz <= info->rammap_max)
+			break;
+	}
+	return data;
+}
+
+u32
 nvbios_rammapSe(struct nouveau_bios *bios, u32 data,
 		u8 ever, u8 ehdr, u8 ecnt, u8 elen, int idx,
 		u8 *ver, u8 *hdr)
@@ -129,8 +151,28 @@ nvbios_rammapSp(struct nouveau_bios *bios, u32 data,
 		u8 *ver, u8 *hdr, struct nvbios_ramcfg *p)
 {
 	data = nvbios_rammapSe(bios, data, ever, ehdr, ecnt, elen, idx, ver, hdr);
+	p->ramcfg_ver = *ver;
+	p->ramcfg_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->ramcfg_timing   =  nv_ro08(bios, data + 0x01);
+		p->ramcfg_10_02_01 = (nv_ro08(bios, data + 0x02) & 0x01) >> 0;
+		p->ramcfg_10_02_02 = (nv_ro08(bios, data + 0x02) & 0x02) >> 1;
+		p->ramcfg_10_02_04 = (nv_ro08(bios, data + 0x02) & 0x04) >> 2;
+		p->ramcfg_10_02_08 = (nv_ro08(bios, data + 0x02) & 0x08) >> 3;
+		p->ramcfg_10_02_10 = (nv_ro08(bios, data + 0x02) & 0x10) >> 4;
+		p->ramcfg_10_02_20 = (nv_ro08(bios, data + 0x02) & 0x20) >> 5;
+		p->ramcfg_10_02_40 = (nv_ro08(bios, data + 0x02) & 0x40) >> 6;
+		p->ramcfg_10_03_0f = (nv_ro08(bios, data + 0x03) & 0x0f) >> 0;
+		p->ramcfg_10_05    = (nv_ro08(bios, data + 0x05) & 0xff) >> 0;
+		p->ramcfg_10_06    = (nv_ro08(bios, data + 0x06) & 0xff) >> 0;
+		p->ramcfg_10_07    = (nv_ro08(bios, data + 0x07) & 0xff) >> 0;
+		p->ramcfg_10_08    = (nv_ro08(bios, data + 0x08) & 0xff) >> 0;
+		p->ramcfg_10_09_0f = (nv_ro08(bios, data + 0x09) & 0x0f) >> 0;
+		p->ramcfg_10_09_f0 = (nv_ro08(bios, data + 0x09) & 0xf0) >> 4;
+		break;
 	case 0x11:
+		p->ramcfg_timing   =  nv_ro08(bios, data + 0x00);
 		p->ramcfg_11_01_01 = (nv_ro08(bios, data + 0x01) & 0x01) >> 0;
 		p->ramcfg_11_01_02 = (nv_ro08(bios, data + 0x01) & 0x02) >> 1;
 		p->ramcfg_11_01_04 = (nv_ro08(bios, data + 0x01) & 0x04) >> 2;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c b/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
index 350d44ab2ba2..46d955eb51eb 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/timing.c
@@ -89,7 +89,15 @@ nvbios_timingEp(struct nouveau_bios *bios, int idx,
 		struct nvbios_ramcfg *p)
 {
 	u16 data = nvbios_timingEe(bios, idx, ver, hdr, cnt, len), temp;
+	p->timing_ver = *ver;
+	p->timing_hdr = *hdr;
 	switch (!!data * *ver) {
+	case 0x10:
+		p->timing_10_WR = nv_ro08(bios, data + 0x00);
+		p->timing_10_CL = nv_ro08(bios, data + 0x02);
+		p->timing_10_ODT = nv_ro08(bios, data + 0x0e) & 0x07;
+		p->timing_10_CWL = nv_ro08(bios, data + 0x13);
+		break;
 	case 0x20:
 		p->timing[0] = nv_ro32(bios, data + 0x00);
 		p->timing[1] = nv_ro32(bios, data + 0x04);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/base.c b/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
index 22351f594d2a..e51b72d47129 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/base.c
@@ -90,16 +90,20 @@ nouveau_cstate_prog(struct nouveau_clock *clk,
 		cstate = &pstate->base;
 	}
 
-	ret = nouveau_therm_cstate(ptherm, pstate->fanspeed, +1);
-	if (ret && ret != -ENODEV) {
-		nv_error(clk, "failed to raise fan speed: %d\n", ret);
-		return ret;
+	if (ptherm) {
+		ret = nouveau_therm_cstate(ptherm, pstate->fanspeed, +1);
+		if (ret && ret != -ENODEV) {
+			nv_error(clk, "failed to raise fan speed: %d\n", ret);
+			return ret;
+		}
 	}
 
-	ret = volt->set_id(volt, cstate->voltage, +1);
-	if (ret && ret != -ENODEV) {
-		nv_error(clk, "failed to raise voltage: %d\n", ret);
-		return ret;
+	if (volt) {
+		ret = volt->set_id(volt, cstate->voltage, +1);
+		if (ret && ret != -ENODEV) {
+			nv_error(clk, "failed to raise voltage: %d\n", ret);
+			return ret;
+		}
 	}
 
 	ret = clk->calc(clk, cstate);
@@ -108,13 +112,17 @@ nouveau_cstate_prog(struct nouveau_clock *clk,
 		clk->tidy(clk);
 	}
 
-	ret = volt->set_id(volt, cstate->voltage, -1);
-	if (ret && ret != -ENODEV)
-		nv_error(clk, "failed to lower voltage: %d\n", ret);
+	if (volt) {
+		ret = volt->set_id(volt, cstate->voltage, -1);
+		if (ret && ret != -ENODEV)
+			nv_error(clk, "failed to lower voltage: %d\n", ret);
+	}
 
-	ret = nouveau_therm_cstate(ptherm, pstate->fanspeed, -1);
-	if (ret && ret != -ENODEV)
-		nv_error(clk, "failed to lower fan speed: %d\n", ret);
+	if (ptherm) {
+		ret = nouveau_therm_cstate(ptherm, pstate->fanspeed, -1);
+		if (ret && ret != -ENODEV)
+			nv_error(clk, "failed to lower fan speed: %d\n", ret);
+	}
 
 	return 0;
 }
@@ -194,16 +202,23 @@ nouveau_pstate_prog(struct nouveau_clock *clk, int pstatei)
 	return nouveau_cstate_prog(clk, pstate, 0);
 }
 
-static int
-nouveau_pstate_calc(struct nouveau_clock *clk)
+static void
+nouveau_pstate_work(struct work_struct *work)
 {
-	int pstate, ret = 0;
+	struct nouveau_clock *clk = container_of(work, typeof(*clk), work);
+	int pstate;
 
-	nv_trace(clk, "P %d U %d A %d T %d D %d\n", clk->pstate,
-		 clk->ustate, clk->astate, clk->tstate, clk->dstate);
+	if (!atomic_xchg(&clk->waiting, 0))
+		return;
+	clk->pwrsrc = power_supply_is_system_supplied();
 
-	if (clk->state_nr && clk->ustate != -1) {
-		pstate = (clk->ustate < 0) ? clk->astate : clk->ustate;
+	nv_trace(clk, "P %d PWR %d U(AC) %d U(DC) %d A %d T %d D %d\n",
+		 clk->pstate, clk->pwrsrc, clk->ustate_ac, clk->ustate_dc,
+		 clk->astate, clk->tstate, clk->dstate);
+
+	pstate = clk->pwrsrc ? clk->ustate_ac : clk->ustate_dc;
+	if (clk->state_nr && pstate != -1) {
+		pstate = (pstate < 0) ? clk->astate : pstate;
 		pstate = min(pstate, clk->state_nr - 1 - clk->tstate);
 		pstate = max(pstate, clk->dstate);
 	} else {
@@ -211,9 +226,26 @@ nouveau_pstate_calc(struct nouveau_clock *clk)
 	}
 
 	nv_trace(clk, "-> %d\n", pstate);
-	if (pstate != clk->pstate)
-		ret = nouveau_pstate_prog(clk, pstate);
-	return ret;
+	if (pstate != clk->pstate) {
+		int ret = nouveau_pstate_prog(clk, pstate);
+		if (ret) {
+			nv_error(clk, "error setting pstate %d: %d\n",
+				 pstate, ret);
+		}
+	}
+
+	wake_up_all(&clk->wait);
+	nvkm_notify_get(&clk->pwrsrc_ntfy);
+}
+
+static int
+nouveau_pstate_calc(struct nouveau_clock *clk, bool wait)
+{
+	atomic_set(&clk->waiting, 1);
+	schedule_work(&clk->work);
+	if (wait)
+		wait_event(clk->wait, !atomic_read(&clk->waiting));
+	return 0;
 }
 
 static void
@@ -361,17 +393,40 @@ nouveau_clock_ustate_update(struct nouveau_clock *clk, int req)
 		req = i;
 	}
 
-	clk->ustate = req;
-	return 0;
+	return req + 2;
+}
+
+static int
+nouveau_clock_nstate(struct nouveau_clock *clk, const char *mode, int arglen)
+{
+	int ret = 1;
+
+	if (strncasecmpz(mode, "disabled", arglen)) {
+		char save = mode[arglen];
+		long v;
+
+		((char *)mode)[arglen] = '\0';
+		if (!kstrtol(mode, 0, &v)) {
+			ret = nouveau_clock_ustate_update(clk, v);
+			if (ret < 0)
+				ret = 1;
+		}
+		((char *)mode)[arglen] = save;
+	}
+
+	return ret - 2;
 }
 
 int
-nouveau_clock_ustate(struct nouveau_clock *clk, int req)
+nouveau_clock_ustate(struct nouveau_clock *clk, int req, int pwr)
 {
 	int ret = nouveau_clock_ustate_update(clk, req);
-	if (ret)
-		return ret;
-	return nouveau_pstate_calc(clk);
+	if (ret >= 0) {
+		if (ret -= 2, pwr) clk->ustate_ac = ret;
+		else		   clk->ustate_dc = ret;
+		return nouveau_pstate_calc(clk, true);
+	}
+	return ret;
 }
 
 int
@@ -381,7 +436,7 @@ nouveau_clock_astate(struct nouveau_clock *clk, int req, int rel)
 	if ( rel) clk->astate += rel;
 	clk->astate = min(clk->astate, clk->state_nr - 1);
 	clk->astate = max(clk->astate, 0);
-	return nouveau_pstate_calc(clk);
+	return nouveau_pstate_calc(clk, true);
 }
 
 int
@@ -391,7 +446,7 @@ nouveau_clock_tstate(struct nouveau_clock *clk, int req, int rel)
 	if ( rel) clk->tstate += rel;
 	clk->tstate = min(clk->tstate, 0);
 	clk->tstate = max(clk->tstate, -(clk->state_nr - 1));
-	return nouveau_pstate_calc(clk);
+	return nouveau_pstate_calc(clk, true);
 }
 
 int
@@ -401,12 +456,30 @@ nouveau_clock_dstate(struct nouveau_clock *clk, int req, int rel)
 	if ( rel) clk->dstate += rel;
 	clk->dstate = min(clk->dstate, clk->state_nr - 1);
 	clk->dstate = max(clk->dstate, 0);
-	return nouveau_pstate_calc(clk);
+	return nouveau_pstate_calc(clk, true);
+}
+
+static int
+nouveau_clock_pwrsrc(struct nvkm_notify *notify)
+{
+	struct nouveau_clock *clk =
+		container_of(notify, typeof(*clk), pwrsrc_ntfy);
+	nouveau_pstate_calc(clk, false);
+	return NVKM_NOTIFY_DROP;
 }
 
 /******************************************************************************
  * subdev base class implementation
  *****************************************************************************/
+
+int
+_nouveau_clock_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_clock *clk = (void *)object;
+	nvkm_notify_put(&clk->pwrsrc_ntfy);
+	return nouveau_subdev_fini(&clk->base, suspend);
+}
+
 int
 _nouveau_clock_init(struct nouveau_object *object)
 {
@@ -414,6 +487,10 @@ _nouveau_clock_init(struct nouveau_object *object)
 	struct nouveau_clocks *clock = clk->domains;
 	int ret;
 
+	ret = nouveau_subdev_init(&clk->base);
+	if (ret)
+		return ret;
+
 	memset(&clk->bstate, 0x00, sizeof(clk->bstate));
 	INIT_LIST_HEAD(&clk->bstate.list);
 	clk->bstate.pstate = 0xff;
@@ -434,7 +511,7 @@ _nouveau_clock_init(struct nouveau_object *object)
 	clk->tstate = 0;
 	clk->dstate = 0;
 	clk->pstate = -1;
-	nouveau_pstate_calc(clk);
+	nouveau_pstate_calc(clk, true);
 	return 0;
 }
 
@@ -444,6 +521,8 @@ _nouveau_clock_dtor(struct nouveau_object *object)
 	struct nouveau_clock *clk = (void *)object;
 	struct nouveau_pstate *pstate, *temp;
 
+	nvkm_notify_fini(&clk->pwrsrc_ntfy);
+
 	list_for_each_entry_safe(pstate, temp, &clk->states, head) {
 		nouveau_pstate_del(pstate);
 	}
@@ -456,6 +535,7 @@ nouveau_clock_create_(struct nouveau_object *parent,
 		      struct nouveau_object *engine,
 		      struct nouveau_oclass *oclass,
 		      struct nouveau_clocks *clocks,
+		      struct nouveau_pstate *pstates, int nb_pstates,
 		      bool allow_reclock,
 		      int length, void **object)
 {
@@ -472,29 +552,46 @@ nouveau_clock_create_(struct nouveau_object *parent,
 
 	INIT_LIST_HEAD(&clk->states);
 	clk->domains = clocks;
-	clk->ustate = -1;
+	clk->ustate_ac = -1;
+	clk->ustate_dc = -1;
+
+	INIT_WORK(&clk->work, nouveau_pstate_work);
+	init_waitqueue_head(&clk->wait);
+	atomic_set(&clk->waiting, 0);
 
-	idx = 0;
-	do {
-		ret = nouveau_pstate_new(clk, idx++);
-	} while (ret == 0);
+	/* If no pstates are provided, try and fetch them from the BIOS */
+	if (!pstates) {
+		idx = 0;
+		do {
+			ret = nouveau_pstate_new(clk, idx++);
+		} while (ret == 0);
+	} else {
+		for (idx = 0; idx < nb_pstates; idx++)
+			list_add_tail(&pstates[idx].head, &clk->states);
+		clk->state_nr = nb_pstates;
+	}
 
 	clk->allow_reclock = allow_reclock;
 
+	ret = nvkm_notify_init(NULL, &device->event, nouveau_clock_pwrsrc, true,
+			       NULL, 0, 0, &clk->pwrsrc_ntfy);
+	if (ret)
+		return ret;
+
 	mode = nouveau_stropt(device->cfgopt, "NvClkMode", &arglen);
 	if (mode) {
-		if (!strncasecmpz(mode, "disabled", arglen)) {
-			clk->ustate = -1;
-		} else {
-			char save = mode[arglen];
-			long v;
-
-			((char *)mode)[arglen] = '\0';
-			if (!kstrtol(mode, 0, &v))
-				nouveau_clock_ustate_update(clk, v);
-			((char *)mode)[arglen] = save;
-		}
+		clk->ustate_ac = nouveau_clock_nstate(clk, mode, arglen);
+		clk->ustate_dc = nouveau_clock_nstate(clk, mode, arglen);
 	}
 
+	mode = nouveau_stropt(device->cfgopt, "NvClkModeAC", &arglen);
+	if (mode)
+		clk->ustate_ac = nouveau_clock_nstate(clk, mode, arglen);
+
+	mode = nouveau_stropt(device->cfgopt, "NvClkModeDC", &arglen);
+	if (mode)
+		clk->ustate_dc = nouveau_clock_nstate(clk, mode, arglen);
+
+
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c
new file mode 100644
index 000000000000..425a8d5e9129
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/gk20a.c
@@ -0,0 +1,665 @@
+/*
+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Shamelessly ripped off from ChromeOS's gk20a/clk_pllg.c
+ *
+ */
+
+#define MHZ (1000 * 1000)
+
+#define MASK(w)	((1 << w) - 1)
+
+#define SYS_GPCPLL_CFG_BASE			0x00137000
+#define GPC_BCASE_GPCPLL_CFG_BASE		0x00132800
+
+#define GPCPLL_CFG		(SYS_GPCPLL_CFG_BASE + 0)
+#define GPCPLL_CFG_ENABLE	BIT(0)
+#define GPCPLL_CFG_IDDQ		BIT(1)
+#define GPCPLL_CFG_LOCK_DET_OFF	BIT(4)
+#define GPCPLL_CFG_LOCK		BIT(17)
+
+#define GPCPLL_COEFF		(SYS_GPCPLL_CFG_BASE + 4)
+#define GPCPLL_COEFF_M_SHIFT	0
+#define GPCPLL_COEFF_M_WIDTH	8
+#define GPCPLL_COEFF_N_SHIFT	8
+#define GPCPLL_COEFF_N_WIDTH	8
+#define GPCPLL_COEFF_P_SHIFT	16
+#define GPCPLL_COEFF_P_WIDTH	6
+
+#define GPCPLL_CFG2			(SYS_GPCPLL_CFG_BASE + 0xc)
+#define GPCPLL_CFG2_SETUP2_SHIFT	16
+#define GPCPLL_CFG2_PLL_STEPA_SHIFT	24
+
+#define GPCPLL_CFG3			(SYS_GPCPLL_CFG_BASE + 0x18)
+#define GPCPLL_CFG3_PLL_STEPB_SHIFT	16
+
+#define GPCPLL_NDIV_SLOWDOWN			(SYS_GPCPLL_CFG_BASE + 0x1c)
+#define GPCPLL_NDIV_SLOWDOWN_NDIV_LO_SHIFT	0
+#define GPCPLL_NDIV_SLOWDOWN_NDIV_MID_SHIFT	8
+#define GPCPLL_NDIV_SLOWDOWN_STEP_SIZE_LO2MID_SHIFT	16
+#define GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT	22
+#define GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT	31
+
+#define SEL_VCO				(SYS_GPCPLL_CFG_BASE + 0x100)
+#define SEL_VCO_GPC2CLK_OUT_SHIFT	0
+
+#define GPC2CLK_OUT			(SYS_GPCPLL_CFG_BASE + 0x250)
+#define GPC2CLK_OUT_SDIV14_INDIV4_WIDTH	1
+#define GPC2CLK_OUT_SDIV14_INDIV4_SHIFT	31
+#define GPC2CLK_OUT_SDIV14_INDIV4_MODE	1
+#define GPC2CLK_OUT_VCODIV_WIDTH	6
+#define GPC2CLK_OUT_VCODIV_SHIFT	8
+#define GPC2CLK_OUT_VCODIV1		0
+#define GPC2CLK_OUT_VCODIV_MASK		(MASK(GPC2CLK_OUT_VCODIV_WIDTH) << \
+					GPC2CLK_OUT_VCODIV_SHIFT)
+#define	GPC2CLK_OUT_BYPDIV_WIDTH	6
+#define GPC2CLK_OUT_BYPDIV_SHIFT	0
+#define GPC2CLK_OUT_BYPDIV31		0x3c
+#define GPC2CLK_OUT_INIT_MASK	((MASK(GPC2CLK_OUT_SDIV14_INDIV4_WIDTH) << \
+		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT)\
+		| (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << GPC2CLK_OUT_VCODIV_SHIFT)\
+		| (MASK(GPC2CLK_OUT_BYPDIV_WIDTH) << GPC2CLK_OUT_BYPDIV_SHIFT))
+#define GPC2CLK_OUT_INIT_VAL	((GPC2CLK_OUT_SDIV14_INDIV4_MODE << \
+		GPC2CLK_OUT_SDIV14_INDIV4_SHIFT) \
+		| (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT) \
+		| (GPC2CLK_OUT_BYPDIV31 << GPC2CLK_OUT_BYPDIV_SHIFT))
+
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG	(GPC_BCASE_GPCPLL_CFG_BASE + 0xa0)
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT	24
+#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK \
+	    (0x1 << GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT)
+
+#include <subdev/clock.h>
+#include <subdev/timer.h>
+
+#ifdef __KERNEL__
+#include <nouveau_platform.h>
+#endif
+
+static const u8 pl_to_div[] = {
+/* PL:   0, 1, 2, 3, 4, 5, 6,  7,  8,  9, 10, 11, 12, 13, 14 */
+/* p: */ 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 12, 16, 20, 24, 32,
+};
+
+/* All frequencies in Mhz */
+struct gk20a_clk_pllg_params {
+	u32 min_vco, max_vco;
+	u32 min_u, max_u;
+	u32 min_m, max_m;
+	u32 min_n, max_n;
+	u32 min_pl, max_pl;
+};
+
+static const struct gk20a_clk_pllg_params gk20a_pllg_params = {
+	.min_vco = 1000, .max_vco = 1700,
+	.min_u = 12, .max_u = 38,
+	.min_m = 1, .max_m = 255,
+	.min_n = 8, .max_n = 255,
+	.min_pl = 1, .max_pl = 32,
+};
+
+struct gk20a_clock_priv {
+	struct nouveau_clock base;
+	const struct gk20a_clk_pllg_params *params;
+	u32 m, n, pl;
+	u32 parent_rate;
+};
+#define to_gk20a_clock(base) container_of(base, struct gk20a_clock_priv, base)
+
+static void
+gk20a_pllg_read_mnp(struct gk20a_clock_priv *priv)
+{
+	u32 val;
+
+	val = nv_rd32(priv, GPCPLL_COEFF);
+	priv->m = (val >> GPCPLL_COEFF_M_SHIFT) & MASK(GPCPLL_COEFF_M_WIDTH);
+	priv->n = (val >> GPCPLL_COEFF_N_SHIFT) & MASK(GPCPLL_COEFF_N_WIDTH);
+	priv->pl = (val >> GPCPLL_COEFF_P_SHIFT) & MASK(GPCPLL_COEFF_P_WIDTH);
+}
+
+static u32
+gk20a_pllg_calc_rate(struct gk20a_clock_priv *priv)
+{
+	u32 rate;
+	u32 divider;
+
+	rate = priv->parent_rate * priv->n;
+	divider = priv->m * pl_to_div[priv->pl];
+	do_div(rate, divider);
+
+	return rate / 2;
+}
+
+static int
+gk20a_pllg_calc_mnp(struct gk20a_clock_priv *priv, unsigned long rate)
+{
+	u32 target_clk_f, ref_clk_f, target_freq;
+	u32 min_vco_f, max_vco_f;
+	u32 low_pl, high_pl, best_pl;
+	u32 target_vco_f, vco_f;
+	u32 best_m, best_n;
+	u32 u_f;
+	u32 m, n, n2;
+	u32 delta, lwv, best_delta = ~0;
+	u32 pl;
+
+	target_clk_f = rate * 2 / MHZ;
+	ref_clk_f = priv->parent_rate / MHZ;
+
+	max_vco_f = priv->params->max_vco;
+	min_vco_f = priv->params->min_vco;
+	best_m = priv->params->max_m;
+	best_n = priv->params->min_n;
+	best_pl = priv->params->min_pl;
+
+	target_vco_f = target_clk_f + target_clk_f / 50;
+	if (max_vco_f < target_vco_f)
+		max_vco_f = target_vco_f;
+
+	/* min_pl <= high_pl <= max_pl */
+	high_pl = (max_vco_f + target_vco_f - 1) / target_vco_f;
+	high_pl = min(high_pl, priv->params->max_pl);
+	high_pl = max(high_pl, priv->params->min_pl);
+
+	/* min_pl <= low_pl <= max_pl */
+	low_pl = min_vco_f / target_vco_f;
+	low_pl = min(low_pl, priv->params->max_pl);
+	low_pl = max(low_pl, priv->params->min_pl);
+
+	/* Find Indices of high_pl and low_pl */
+	for (pl = 0; pl < ARRAY_SIZE(pl_to_div) - 1; pl++) {
+		if (pl_to_div[pl] >= low_pl) {
+			low_pl = pl;
+			break;
+		}
+	}
+	for (pl = 0; pl < ARRAY_SIZE(pl_to_div) - 1; pl++) {
+		if (pl_to_div[pl] >= high_pl) {
+			high_pl = pl;
+			break;
+		}
+	}
+
+	nv_debug(priv, "low_PL %d(div%d), high_PL %d(div%d)", low_pl,
+		 pl_to_div[low_pl], high_pl, pl_to_div[high_pl]);
+
+	/* Select lowest possible VCO */
+	for (pl = low_pl; pl <= high_pl; pl++) {
+		target_vco_f = target_clk_f * pl_to_div[pl];
+		for (m = priv->params->min_m; m <= priv->params->max_m; m++) {
+			u_f = ref_clk_f / m;
+
+			if (u_f < priv->params->min_u)
+				break;
+			if (u_f > priv->params->max_u)
+				continue;
+
+			n = (target_vco_f * m) / ref_clk_f;
+			n2 = ((target_vco_f * m) + (ref_clk_f - 1)) / ref_clk_f;
+
+			if (n > priv->params->max_n)
+				break;
+
+			for (; n <= n2; n++) {
+				if (n < priv->params->min_n)
+					continue;
+				if (n > priv->params->max_n)
+					break;
+
+				vco_f = ref_clk_f * n / m;
+
+				if (vco_f >= min_vco_f && vco_f <= max_vco_f) {
+					lwv = (vco_f + (pl_to_div[pl] / 2))
+						/ pl_to_div[pl];
+					delta = abs(lwv - target_clk_f);
+
+					if (delta < best_delta) {
+						best_delta = delta;
+						best_m = m;
+						best_n = n;
+						best_pl = pl;
+
+						if (best_delta == 0)
+							goto found_match;
+					}
+				}
+			}
+		}
+	}
+
+found_match:
+	WARN_ON(best_delta == ~0);
+
+	if (best_delta != 0)
+		nv_debug(priv, "no best match for target @ %dMHz on gpc_pll",
+			 target_clk_f);
+
+	priv->m = best_m;
+	priv->n = best_n;
+	priv->pl = best_pl;
+
+	target_freq = gk20a_pllg_calc_rate(priv) / MHZ;
+
+	nv_debug(priv, "actual target freq %d MHz, M %d, N %d, PL %d(div%d)\n",
+		 target_freq, priv->m, priv->n, priv->pl, pl_to_div[priv->pl]);
+
+	return 0;
+}
+
+static int
+gk20a_pllg_slide(struct gk20a_clock_priv *priv, u32 n)
+{
+	u32 val;
+	int ramp_timeout;
+
+	/* get old coefficients */
+	val = nv_rd32(priv, GPCPLL_COEFF);
+	/* do nothing if NDIV is the same */
+	if (n == ((val >> GPCPLL_COEFF_N_SHIFT) & MASK(GPCPLL_COEFF_N_WIDTH)))
+		return 0;
+
+	/* setup */
+	nv_mask(priv, GPCPLL_CFG2, 0xff << GPCPLL_CFG2_PLL_STEPA_SHIFT,
+		0x2b << GPCPLL_CFG2_PLL_STEPA_SHIFT);
+	nv_mask(priv, GPCPLL_CFG3, 0xff << GPCPLL_CFG3_PLL_STEPB_SHIFT,
+		0xb << GPCPLL_CFG3_PLL_STEPB_SHIFT);
+
+	/* pll slowdown mode */
+	nv_mask(priv, GPCPLL_NDIV_SLOWDOWN,
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT),
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT));
+
+	/* new ndiv ready for ramp */
+	val = nv_rd32(priv, GPCPLL_COEFF);
+	val &= ~(MASK(GPCPLL_COEFF_N_WIDTH) << GPCPLL_COEFF_N_SHIFT);
+	val |= (n & MASK(GPCPLL_COEFF_N_WIDTH)) << GPCPLL_COEFF_N_SHIFT;
+	udelay(1);
+	nv_wr32(priv, GPCPLL_COEFF, val);
+
+	/* dynamic ramp to new ndiv */
+	val = nv_rd32(priv, GPCPLL_NDIV_SLOWDOWN);
+	val |= 0x1 << GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT;
+	udelay(1);
+	nv_wr32(priv, GPCPLL_NDIV_SLOWDOWN, val);
+
+	for (ramp_timeout = 500; ramp_timeout > 0; ramp_timeout--) {
+		udelay(1);
+		val = nv_rd32(priv, GPC_BCAST_NDIV_SLOWDOWN_DEBUG);
+		if (val & GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK)
+			break;
+	}
+
+	/* exit slowdown mode */
+	nv_mask(priv, GPCPLL_NDIV_SLOWDOWN,
+		BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT) |
+		BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), 0);
+	nv_rd32(priv, GPCPLL_NDIV_SLOWDOWN);
+
+	if (ramp_timeout <= 0) {
+		nv_error(priv, "gpcpll dynamic ramp timeout\n");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static void
+_gk20a_pllg_enable(struct gk20a_clock_priv *priv)
+{
+	nv_mask(priv, GPCPLL_CFG, GPCPLL_CFG_ENABLE, GPCPLL_CFG_ENABLE);
+	nv_rd32(priv, GPCPLL_CFG);
+}
+
+static void
+_gk20a_pllg_disable(struct gk20a_clock_priv *priv)
+{
+	nv_mask(priv, GPCPLL_CFG, GPCPLL_CFG_ENABLE, 0);
+	nv_rd32(priv, GPCPLL_CFG);
+}
+
+static int
+_gk20a_pllg_program_mnp(struct gk20a_clock_priv *priv, bool allow_slide)
+{
+	u32 val, cfg;
+	u32 m_old, pl_old, n_lo;
+
+	/* get old coefficients */
+	val = nv_rd32(priv, GPCPLL_COEFF);
+	m_old = (val >> GPCPLL_COEFF_M_SHIFT) & MASK(GPCPLL_COEFF_M_WIDTH);
+	pl_old = (val >> GPCPLL_COEFF_P_SHIFT) & MASK(GPCPLL_COEFF_P_WIDTH);
+
+	/* do NDIV slide if there is no change in M and PL */
+	cfg = nv_rd32(priv, GPCPLL_CFG);
+	if (allow_slide && priv->m == m_old && priv->pl == pl_old &&
+	    (cfg & GPCPLL_CFG_ENABLE)) {
+		return gk20a_pllg_slide(priv, priv->n);
+	}
+
+	/* slide down to NDIV_LO */
+	n_lo = DIV_ROUND_UP(m_old * priv->params->min_vco,
+			    priv->parent_rate / MHZ);
+	if (allow_slide && (cfg & GPCPLL_CFG_ENABLE)) {
+		int ret = gk20a_pllg_slide(priv, n_lo);
+
+		if (ret)
+			return ret;
+	}
+
+	/* split FO-to-bypass jump in halfs by setting out divider 1:2 */
+	nv_mask(priv, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK,
+		0x2 << GPC2CLK_OUT_VCODIV_SHIFT);
+
+	/* put PLL in bypass before programming it */
+	val = nv_rd32(priv, SEL_VCO);
+	val &= ~(BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+	udelay(2);
+	nv_wr32(priv, SEL_VCO, val);
+
+	/* get out from IDDQ */
+	val = nv_rd32(priv, GPCPLL_CFG);
+	if (val & GPCPLL_CFG_IDDQ) {
+		val &= ~GPCPLL_CFG_IDDQ;
+		nv_wr32(priv, GPCPLL_CFG, val);
+		nv_rd32(priv, GPCPLL_CFG);
+		udelay(2);
+	}
+
+	_gk20a_pllg_disable(priv);
+
+	nv_debug(priv, "%s: m=%d n=%d pl=%d\n", __func__, priv->m, priv->n,
+		 priv->pl);
+
+	n_lo = DIV_ROUND_UP(priv->m * priv->params->min_vco,
+			    priv->parent_rate / MHZ);
+	val = priv->m << GPCPLL_COEFF_M_SHIFT;
+	val |= (allow_slide ? n_lo : priv->n) << GPCPLL_COEFF_N_SHIFT;
+	val |= priv->pl << GPCPLL_COEFF_P_SHIFT;
+	nv_wr32(priv, GPCPLL_COEFF, val);
+
+	_gk20a_pllg_enable(priv);
+
+	val = nv_rd32(priv, GPCPLL_CFG);
+	if (val & GPCPLL_CFG_LOCK_DET_OFF) {
+		val &= ~GPCPLL_CFG_LOCK_DET_OFF;
+		nv_wr32(priv, GPCPLL_CFG, val);
+	}
+
+	if (!nouveau_timer_wait_eq(priv, 300000, GPCPLL_CFG, GPCPLL_CFG_LOCK,
+				   GPCPLL_CFG_LOCK)) {
+		nv_error(priv, "%s: timeout waiting for pllg lock\n", __func__);
+		return -ETIMEDOUT;
+	}
+
+	/* switch to VCO mode */
+	nv_mask(priv, SEL_VCO, 0, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT));
+
+	/* restore out divider 1:1 */
+	val = nv_rd32(priv, GPC2CLK_OUT);
+	val &= ~GPC2CLK_OUT_VCODIV_MASK;
+	udelay(2);
+	nv_wr32(priv, GPC2CLK_OUT, val);
+
+	/* slide up to new NDIV */
+	return allow_slide ? gk20a_pllg_slide(priv, priv->n) : 0;
+}
+
+static int
+gk20a_pllg_program_mnp(struct gk20a_clock_priv *priv)
+{
+	int err;
+
+	err = _gk20a_pllg_program_mnp(priv, true);
+	if (err)
+		err = _gk20a_pllg_program_mnp(priv, false);
+
+	return err;
+}
+
+static void
+gk20a_pllg_disable(struct gk20a_clock_priv *priv)
+{
+	u32 val;
+
+	/* slide to VCO min */
+	val = nv_rd32(priv, GPCPLL_CFG);
+	if (val & GPCPLL_CFG_ENABLE) {
+		u32 coeff, m, n_lo;
+
+		coeff = nv_rd32(priv, GPCPLL_COEFF);
+		m = (coeff >> GPCPLL_COEFF_M_SHIFT) & MASK(GPCPLL_COEFF_M_WIDTH);
+		n_lo = DIV_ROUND_UP(m * priv->params->min_vco,
+				    priv->parent_rate / MHZ);
+		gk20a_pllg_slide(priv, n_lo);
+	}
+
+	/* put PLL in bypass before disabling it */
+	nv_mask(priv, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0);
+
+	_gk20a_pllg_disable(priv);
+}
+
+#define GK20A_CLK_GPC_MDIV 1000
+
+static struct nouveau_clocks
+gk20a_domains[] = {
+	{ nv_clk_src_crystal, 0xff },
+	{ nv_clk_src_gpc, 0xff, 0, "core", GK20A_CLK_GPC_MDIV },
+	{ nv_clk_src_max }
+};
+
+static struct nouveau_pstate
+gk20a_pstates[] = {
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 72000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 108000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 180000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 252000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 324000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 396000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 468000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 540000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 612000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 648000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 684000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 708000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 756000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 804000,
+		},
+	},
+	{
+		.base = {
+			.domain[nv_clk_src_gpc] = 852000,
+		},
+	},
+};
+
+static int
+gk20a_clock_read(struct nouveau_clock *clk, enum nv_clk_src src)
+{
+	struct gk20a_clock_priv *priv = (void *)clk;
+
+	switch (src) {
+	case nv_clk_src_crystal:
+		return nv_device(clk)->crystal;
+	case nv_clk_src_gpc:
+		gk20a_pllg_read_mnp(priv);
+		return gk20a_pllg_calc_rate(priv) / GK20A_CLK_GPC_MDIV;
+	default:
+		nv_error(clk, "invalid clock source %d\n", src);
+		return -EINVAL;
+	}
+}
+
+static int
+gk20a_clock_calc(struct nouveau_clock *clk, struct nouveau_cstate *cstate)
+{
+	struct gk20a_clock_priv *priv = (void *)clk;
+
+	return gk20a_pllg_calc_mnp(priv, cstate->domain[nv_clk_src_gpc] *
+					 GK20A_CLK_GPC_MDIV);
+}
+
+static int
+gk20a_clock_prog(struct nouveau_clock *clk)
+{
+	struct gk20a_clock_priv *priv = (void *)clk;
+
+	return gk20a_pllg_program_mnp(priv);
+}
+
+static void
+gk20a_clock_tidy(struct nouveau_clock *clk)
+{
+}
+
+static int
+gk20a_clock_fini(struct nouveau_object *object, bool suspend)
+{
+	struct gk20a_clock_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_clock_fini(&priv->base, false);
+
+	gk20a_pllg_disable(priv);
+
+	return ret;
+}
+
+static int
+gk20a_clock_init(struct nouveau_object *object)
+{
+	struct gk20a_clock_priv *priv = (void *)object;
+	int ret;
+
+	nv_mask(priv, GPC2CLK_OUT, GPC2CLK_OUT_INIT_MASK, GPC2CLK_OUT_INIT_VAL);
+
+	ret = nouveau_clock_init(&priv->base);
+	if (ret)
+		return ret;
+
+	ret = gk20a_clock_prog(&priv->base);
+	if (ret) {
+		nv_error(priv, "cannot initialize clock\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static int
+gk20a_clock_ctor(struct nouveau_object *parent,  struct nouveau_object *engine,
+		 struct nouveau_oclass *oclass, void *data, u32 size,
+		 struct nouveau_object **pobject)
+{
+	struct gk20a_clock_priv *priv;
+	struct nouveau_platform_device *plat;
+	int ret;
+	int i;
+
+	/* Finish initializing the pstates */
+	for (i = 0; i < ARRAY_SIZE(gk20a_pstates); i++) {
+		INIT_LIST_HEAD(&gk20a_pstates[i].list);
+		gk20a_pstates[i].pstate = i + 1;
+	}
+
+	ret = nouveau_clock_create(parent, engine, oclass, gk20a_domains,
+			gk20a_pstates, ARRAY_SIZE(gk20a_pstates), true, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->params = &gk20a_pllg_params;
+
+	plat = nv_device_to_platform(nv_device(parent));
+	priv->parent_rate = clk_get_rate(plat->gpu->clk);
+	nv_info(priv, "parent clock rate: %d Mhz\n", priv->parent_rate / MHZ);
+
+	priv->base.read = gk20a_clock_read;
+	priv->base.calc = gk20a_clock_calc;
+	priv->base.prog = gk20a_clock_prog;
+	priv->base.tidy = gk20a_clock_tidy;
+
+	return 0;
+}
+
+struct nouveau_oclass
+gk20a_clock_oclass = {
+	.handle = NV_SUBDEV(CLOCK, 0xea),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gk20a_clock_ctor,
+		.dtor = _nouveau_subdev_dtor,
+		.init = gk20a_clock_init,
+		.fini = gk20a_clock_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
index eb2d4425a49e..4c48232686be 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
@@ -82,8 +82,8 @@ nv04_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv04_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nv04_domain, false,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nv04_domain, NULL, 0,
+				   false, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
index 8a9e16839791..08368fe97029 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
@@ -213,8 +213,8 @@ nv40_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv40_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nv40_domain, true,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nv40_domain, NULL, 0,
+				   true, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
index 8c132772ba9e..5070ebc260f8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
@@ -507,7 +507,7 @@ nv50_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int ret;
 
 	ret = nouveau_clock_create(parent, engine, oclass, pclass->domains,
-				   false, &priv);
+				   NULL, 0, false, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
index 9fb58354a80b..094551d8ad9b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
@@ -20,8 +20,10 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  * Authors: Ben Skeggs
+ *          Roy Spliet
  */
 
+#include <engine/fifo.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
 #include <subdev/timer.h>
@@ -42,9 +44,17 @@ static u32
 read_vco(struct nva3_clock_priv *priv, int clk)
 {
 	u32 sctl = nv_rd32(priv, 0x4120 + (clk * 4));
-	if ((sctl & 0x00000030) != 0x00000030)
+
+	switch (sctl & 0x00000030) {
+	case 0x00000000:
+		return nv_device(priv)->crystal;
+	case 0x00000020:
 		return read_pll(priv, 0x41, 0x00e820);
-	return read_pll(priv, 0x42, 0x00e8a0);
+	case 0x00000030:
+		return read_pll(priv, 0x42, 0x00e8a0);
+	default:
+		return 0;
+	}
 }
 
 static u32
@@ -66,14 +76,25 @@ read_clk(struct nva3_clock_priv *priv, int clk, bool ignore_en)
 	if (!ignore_en && !(sctl & 0x00000100))
 		return 0;
 
+	/* out_alt */
+	if (sctl & 0x00000400)
+		return 108000;
+
+	/* vco_out */
 	switch (sctl & 0x00003000) {
 	case 0x00000000:
-		return nv_device(priv)->crystal;
+		if (!(sctl & 0x00000200))
+			return nv_device(priv)->crystal;
+		return 0;
 	case 0x00002000:
 		if (sctl & 0x00000040)
 			return 108000;
 		return 100000;
 	case 0x00003000:
+		/* vco_enable */
+		if (!(sctl & 0x00000001))
+			return 0;
+
 		sclk = read_vco(priv, clk);
 		sdiv = ((sctl & 0x003f0000) >> 16) + 2;
 		return (sclk * 2) / sdiv;
@@ -95,7 +116,9 @@ read_pll(struct nva3_clock_priv *priv, int clk, u32 pll)
 			N = (coef & 0x0000ff00) >> 8;
 			P = (coef & 0x003f0000) >> 16;
 
-			/* no post-divider on these.. */
+			/* no post-divider on these..
+			 * XXX: it looks more like two post-"dividers" that
+			 * cross each other out in the default RPLL config */
 			if ((pll & 0x00ff00) == 0x00e800)
 				P = 1;
 
@@ -114,13 +137,13 @@ static int
 nva3_clock_read(struct nouveau_clock *clk, enum nv_clk_src src)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
+	u32 hsrc;
 
 	switch (src) {
 	case nv_clk_src_crystal:
 		return nv_device(priv)->crystal;
-	case nv_clk_src_href:
-		return 100000;
 	case nv_clk_src_core:
+	case nv_clk_src_core_intm:
 		return read_pll(priv, 0x00, 0x4200);
 	case nv_clk_src_shader:
 		return read_pll(priv, 0x01, 0x4220);
@@ -132,24 +155,33 @@ nva3_clock_read(struct nouveau_clock *clk, enum nv_clk_src src)
 		return read_clk(priv, 0x21, false);
 	case nv_clk_src_daemon:
 		return read_clk(priv, 0x25, false);
+	case nv_clk_src_host:
+		hsrc = (nv_rd32(priv, 0xc040) & 0x30000000) >> 28;
+		switch (hsrc) {
+		case 0:
+			return read_clk(priv, 0x1d, false);
+		case 2:
+		case 3:
+			return 277000;
+		default:
+			nv_error(clk, "unknown HOST clock source %d\n", hsrc);
+			return -EINVAL;
+		}
 	default:
 		nv_error(clk, "invalid clock source %d\n", src);
 		return -EINVAL;
 	}
+
+	return 0;
 }
 
 int
-nva3_clock_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
+nva3_clk_info(struct nouveau_clock *clock, int clk, u32 khz,
 		struct nva3_clock_info *info)
 {
-	struct nouveau_bios *bios = nouveau_bios(clock);
 	struct nva3_clock_priv *priv = (void *)clock;
-	struct nvbios_pll limits;
-	u32 oclk, sclk, sdiv;
-	int P, N, M, diff;
-	int ret;
+	u32 oclk, sclk, sdiv, diff;
 
-	info->pll = 0;
 	info->clk = 0;
 
 	switch (khz) {
@@ -164,43 +196,69 @@ nva3_clock_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
 		return khz;
 	default:
 		sclk = read_vco(priv, clk);
-		sdiv = min((sclk * 2) / (khz - 2999), (u32)65);
-		/* if the clock has a PLL attached, and we can get a within
-		 * [-2, 3) MHz of a divider, we'll disable the PLL and use
-		 * the divider instead.
-		 *
-		 * divider can go as low as 2, limited here because NVIDIA
+		sdiv = min((sclk * 2) / khz, (u32)65);
+		oclk = (sclk * 2) / sdiv;
+		diff = ((khz + 3000) - oclk);
+
+		/* When imprecise, play it safe and aim for a clock lower than
+		 * desired rather than higher */
+		if (diff < 0) {
+			sdiv++;
+			oclk = (sclk * 2) / sdiv;
+		}
+
+		/* divider can go as low as 2, limited here because NVIDIA
 		 * and the VBIOS on my NVA8 seem to prefer using the PLL
 		 * for 810MHz - is there a good reason?
-		 */
+		 * XXX: PLLs with refclk 810MHz?  */
 		if (sdiv > 4) {
-			oclk = (sclk * 2) / sdiv;
-			diff = khz - oclk;
-			if (!pll || (diff >= -2000 && diff < 3000)) {
-				info->clk = (((sdiv - 2) << 16) | 0x00003100);
-				return oclk;
-			}
+			info->clk = (((sdiv - 2) << 16) | 0x00003100);
+			return oclk;
 		}
 
-		if (!pll)
-			return -ERANGE;
 		break;
 	}
 
+	return -ERANGE;
+}
+
+int
+nva3_pll_info(struct nouveau_clock *clock, int clk, u32 pll, u32 khz,
+		struct nva3_clock_info *info)
+{
+	struct nouveau_bios *bios = nouveau_bios(clock);
+	struct nva3_clock_priv *priv = (void *)clock;
+	struct nvbios_pll limits;
+	int P, N, M, diff;
+	int ret;
+
+	info->pll = 0;
+
+	/* If we can get a within [-2, 3) MHz of a divider, we'll disable the
+	 * PLL and use the divider instead. */
+	ret = nva3_clk_info(clock, clk, khz, info);
+	diff = khz - ret;
+	if (!pll || (diff >= -2000 && diff < 3000)) {
+		goto out;
+	}
+
+	/* Try with PLL */
 	ret = nvbios_pll_parse(bios, pll, &limits);
 	if (ret)
 		return ret;
 
-	limits.refclk = read_clk(priv, clk - 0x10, true);
-	if (!limits.refclk)
+	ret = nva3_clk_info(clock, clk - 0x10, limits.refclk, info);
+	if (ret != limits.refclk)
 		return -EINVAL;
 
 	ret = nva3_pll_calc(nv_subdev(priv), &limits, khz, &N, NULL, &M, &P);
 	if (ret >= 0) {
-		info->clk = nv_rd32(priv, 0x4120 + (clk * 4));
 		info->pll = (P << 16) | (N << 8) | M;
 	}
 
+out:
+	info->fb_delay = max(((khz + 7566) / 15133), (u32) 18);
+
 	return ret ? ret : -ERANGE;
 }
 
@@ -208,13 +266,76 @@ static int
 calc_clk(struct nva3_clock_priv *priv, struct nouveau_cstate *cstate,
 	 int clk, u32 pll, int idx)
 {
-	int ret = nva3_clock_info(&priv->base, clk, pll, cstate->domain[idx],
+	int ret = nva3_pll_info(&priv->base, clk, pll, cstate->domain[idx],
 				  &priv->eng[idx]);
 	if (ret >= 0)
 		return 0;
 	return ret;
 }
 
+static int
+calc_host(struct nva3_clock_priv *priv, struct nouveau_cstate *cstate)
+{
+	int ret = 0;
+	u32 kHz = cstate->domain[nv_clk_src_host];
+	struct nva3_clock_info *info = &priv->eng[nv_clk_src_host];
+
+	if (kHz == 277000) {
+		info->clk = 0;
+		info->host_out = NVA3_HOST_277;
+		return 0;
+	}
+
+	info->host_out = NVA3_HOST_CLK;
+
+	ret = nva3_clk_info(&priv->base, 0x1d, kHz, info);
+	if (ret >= 0)
+		return 0;
+	return ret;
+}
+
+int
+nva3_clock_pre(struct nouveau_clock *clk, unsigned long *flags)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+
+	/* halt and idle execution engines */
+	nv_mask(clk, 0x020060, 0x00070000, 0x00000000);
+	nv_mask(clk, 0x002504, 0x00000001, 0x00000001);
+	/* Wait until the interrupt handler is finished */
+	if (!nv_wait(clk, 0x000100, 0xffffffff, 0x00000000))
+		return -EBUSY;
+
+	if (pfifo)
+		pfifo->pause(pfifo, flags);
+
+	if (!nv_wait(clk, 0x002504, 0x00000010, 0x00000010))
+		return -EIO;
+	if (!nv_wait(clk, 0x00251c, 0x0000003f, 0x0000003f))
+		return -EIO;
+
+	return 0;
+}
+
+void
+nva3_clock_post(struct nouveau_clock *clk, unsigned long *flags)
+{
+	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+
+	if (pfifo && flags)
+		pfifo->start(pfifo, flags);
+
+	nv_mask(clk, 0x002504, 0x00000001, 0x00000000);
+	nv_mask(clk, 0x020060, 0x00070000, 0x00040000);
+}
+
+static void
+disable_clk_src(struct nva3_clock_priv *priv, u32 src)
+{
+	nv_mask(priv, src, 0x00000100, 0x00000000);
+	nv_mask(priv, src, 0x00000001, 0x00000000);
+}
+
 static void
 prog_pll(struct nva3_clock_priv *priv, int clk, u32 pll, int idx)
 {
@@ -223,24 +344,35 @@ prog_pll(struct nva3_clock_priv *priv, int clk, u32 pll, int idx)
 	const u32 src1 = 0x004160 + (clk * 4);
 	const u32 ctrl = pll + 0;
 	const u32 coef = pll + 4;
+	u32 bypass;
 
 	if (info->pll) {
-		nv_mask(priv, src0, 0x00000101, 0x00000101);
+		/* Always start from a non-PLL clock */
+		bypass = nv_rd32(priv, ctrl)  & 0x00000008;
+		if (!bypass) {
+			nv_mask(priv, src1, 0x00000101, 0x00000101);
+			nv_mask(priv, ctrl, 0x00000008, 0x00000008);
+			udelay(20);
+		}
+
+		nv_mask(priv, src0, 0x003f3141, 0x00000101 | info->clk);
 		nv_wr32(priv, coef, info->pll);
 		nv_mask(priv, ctrl, 0x00000015, 0x00000015);
 		nv_mask(priv, ctrl, 0x00000010, 0x00000000);
-		nv_wait(priv, ctrl, 0x00020000, 0x00020000);
+		if (!nv_wait(priv, ctrl, 0x00020000, 0x00020000)) {
+			nv_mask(priv, ctrl, 0x00000010, 0x00000010);
+			nv_mask(priv, src0, 0x00000101, 0x00000000);
+			return;
+		}
 		nv_mask(priv, ctrl, 0x00000010, 0x00000010);
 		nv_mask(priv, ctrl, 0x00000008, 0x00000000);
-		nv_mask(priv, src1, 0x00000100, 0x00000000);
-		nv_mask(priv, src1, 0x00000001, 0x00000000);
+		disable_clk_src(priv, src1);
 	} else {
 		nv_mask(priv, src1, 0x003f3141, 0x00000101 | info->clk);
 		nv_mask(priv, ctrl, 0x00000018, 0x00000018);
 		udelay(20);
 		nv_mask(priv, ctrl, 0x00000001, 0x00000000);
-		nv_mask(priv, src0, 0x00000100, 0x00000000);
-		nv_mask(priv, src0, 0x00000001, 0x00000000);
+		disable_clk_src(priv, src0);
 	}
 }
 
@@ -251,18 +383,72 @@ prog_clk(struct nva3_clock_priv *priv, int clk, int idx)
 	nv_mask(priv, 0x004120 + (clk * 4), 0x003f3141, 0x00000101 | info->clk);
 }
 
+static void
+prog_host(struct nva3_clock_priv *priv)
+{
+	struct nva3_clock_info *info = &priv->eng[nv_clk_src_host];
+	u32 hsrc = (nv_rd32(priv, 0xc040));
+
+	switch (info->host_out) {
+	case NVA3_HOST_277:
+		if ((hsrc & 0x30000000) == 0) {
+			nv_wr32(priv, 0xc040, hsrc | 0x20000000);
+			disable_clk_src(priv, 0x4194);
+		}
+		break;
+	case NVA3_HOST_CLK:
+		prog_clk(priv, 0x1d, nv_clk_src_host);
+		if ((hsrc & 0x30000000) >= 0x20000000) {
+			nv_wr32(priv, 0xc040, hsrc & ~0x30000000);
+		}
+		break;
+	default:
+		break;
+	}
+
+	/* This seems to be a clock gating factor on idle, always set to 64 */
+	nv_wr32(priv, 0xc044, 0x3e);
+}
+
+static void
+prog_core(struct nva3_clock_priv *priv, int idx)
+{
+	struct nva3_clock_info *info = &priv->eng[idx];
+	u32 fb_delay = nv_rd32(priv, 0x10002c);
+
+	if (fb_delay < info->fb_delay)
+		nv_wr32(priv, 0x10002c, info->fb_delay);
+
+	prog_pll(priv, 0x00, 0x004200, idx);
+
+	if (fb_delay > info->fb_delay)
+		nv_wr32(priv, 0x10002c, info->fb_delay);
+}
+
 static int
 nva3_clock_calc(struct nouveau_clock *clk, struct nouveau_cstate *cstate)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
+	struct nva3_clock_info *core = &priv->eng[nv_clk_src_core];
 	int ret;
 
 	if ((ret = calc_clk(priv, cstate, 0x10, 0x4200, nv_clk_src_core)) ||
 	    (ret = calc_clk(priv, cstate, 0x11, 0x4220, nv_clk_src_shader)) ||
 	    (ret = calc_clk(priv, cstate, 0x20, 0x0000, nv_clk_src_disp)) ||
-	    (ret = calc_clk(priv, cstate, 0x21, 0x0000, nv_clk_src_vdec)))
+	    (ret = calc_clk(priv, cstate, 0x21, 0x0000, nv_clk_src_vdec)) ||
+	    (ret = calc_host(priv, cstate)))
 		return ret;
 
+	/* XXX: Should be reading the highest bit in the VBIOS clock to decide
+	 * whether to use a PLL or not... but using a PLL defeats the purpose */
+	if (core->pll) {
+		ret = nva3_clk_info(clk, 0x10,
+				cstate->domain[nv_clk_src_core_intm],
+				&priv->eng[nv_clk_src_core_intm]);
+		if (ret < 0)
+			return ret;
+	}
+
 	return 0;
 }
 
@@ -270,11 +456,31 @@ static int
 nva3_clock_prog(struct nouveau_clock *clk)
 {
 	struct nva3_clock_priv *priv = (void *)clk;
-	prog_pll(priv, 0x00, 0x004200, nv_clk_src_core);
+	struct nva3_clock_info *core = &priv->eng[nv_clk_src_core];
+	int ret = 0;
+	unsigned long flags;
+	unsigned long *f = &flags;
+
+	ret = nva3_clock_pre(clk, f);
+	if (ret)
+		goto out;
+
+	if (core->pll)
+		prog_core(priv, nv_clk_src_core_intm);
+
+	prog_core(priv,  nv_clk_src_core);
 	prog_pll(priv, 0x01, 0x004220, nv_clk_src_shader);
 	prog_clk(priv, 0x20, nv_clk_src_disp);
 	prog_clk(priv, 0x21, nv_clk_src_vdec);
-	return 0;
+	prog_host(priv);
+
+out:
+	if (ret == -EBUSY)
+		f = NULL;
+
+	nva3_clock_post(clk, f);
+
+	return ret;
 }
 
 static void
@@ -284,13 +490,14 @@ nva3_clock_tidy(struct nouveau_clock *clk)
 
 static struct nouveau_clocks
 nva3_domain[] = {
-	{ nv_clk_src_crystal, 0xff },
-	{ nv_clk_src_href   , 0xff },
-	{ nv_clk_src_core   , 0x00, 0, "core", 1000 },
-	{ nv_clk_src_shader , 0x01, 0, "shader", 1000 },
-	{ nv_clk_src_mem    , 0x02, 0, "memory", 1000 },
-	{ nv_clk_src_vdec   , 0x03 },
-	{ nv_clk_src_disp   , 0x04 },
+	{ nv_clk_src_crystal  , 0xff },
+	{ nv_clk_src_core     , 0x00, 0, "core", 1000 },
+	{ nv_clk_src_shader   , 0x01, 0, "shader", 1000 },
+	{ nv_clk_src_mem      , 0x02, 0, "memory", 1000 },
+	{ nv_clk_src_vdec     , 0x03 },
+	{ nv_clk_src_disp     , 0x04 },
+	{ nv_clk_src_host     , 0x05 },
+	{ nv_clk_src_core_intm, 0x06 },
 	{ nv_clk_src_max }
 };
 
@@ -302,8 +509,8 @@ nva3_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nva3_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nva3_domain, false,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nva3_domain, NULL, 0,
+				   false, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
index 6229a509b42e..a45a1038b12f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.h
@@ -6,9 +6,15 @@
 struct nva3_clock_info {
 	u32 clk;
 	u32 pll;
+	enum {
+		NVA3_HOST_277,
+		NVA3_HOST_CLK,
+	} host_out;
+	u32 fb_delay;
 };
 
-int nva3_clock_info(struct nouveau_clock *, int, u32, u32,
+int nva3_pll_info(struct nouveau_clock *, int, u32, u32,
 		    struct nva3_clock_info *);
-
+int nva3_clock_pre(struct nouveau_clock *clk, unsigned long *flags);
+void nva3_clock_post(struct nouveau_clock *clk, unsigned long *flags);
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
index 6a65fc9e9663..54aeab8005a0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nvaa.c
@@ -28,6 +28,7 @@
 #include <subdev/timer.h>
 #include <subdev/clock.h>
 
+#include "nva3.h"
 #include "pll.h"
 
 struct nvaa_clock_priv {
@@ -299,25 +300,14 @@ static int
 nvaa_clock_prog(struct nouveau_clock *clk)
 {
 	struct nvaa_clock_priv *priv = (void *)clk;
-	struct nouveau_fifo *pfifo = nouveau_fifo(clk);
+	u32 pllmask = 0, mast;
 	unsigned long flags;
-	u32 pllmask = 0, mast, ptherm_gate;
-	int ret = -EBUSY;
-
-	/* halt and idle execution engines */
-	ptherm_gate = nv_mask(clk, 0x020060, 0x00070000, 0x00000000);
-	nv_mask(clk, 0x002504, 0x00000001, 0x00000001);
-	/* Wait until the interrupt handler is finished */
-	if (!nv_wait(clk, 0x000100, 0xffffffff, 0x00000000))
-		goto resume;
-
-	if (pfifo)
-		pfifo->pause(pfifo, &flags);
+	unsigned long *f = &flags;
+	int ret = 0;
 
-	if (!nv_wait(clk, 0x002504, 0x00000010, 0x00000010))
-		goto resume;
-	if (!nv_wait(clk, 0x00251c, 0x0000003f, 0x0000003f))
-		goto resume;
+	ret = nva3_clock_pre(clk, f);
+	if (ret)
+		goto out;
 
 	/* First switch to safe clocks: href */
 	mast = nv_mask(clk, 0xc054, 0x03400e70, 0x03400640);
@@ -375,15 +365,8 @@ nvaa_clock_prog(struct nouveau_clock *clk)
 	}
 
 	nv_wr32(clk, 0xc054, mast);
-	ret = 0;
 
 resume:
-	if (pfifo)
-		pfifo->start(pfifo, &flags);
-
-	nv_mask(clk, 0x002504, 0x00000001, 0x00000000);
-	nv_wr32(clk, 0x020060, ptherm_gate);
-
 	/* Disable some PLLs and dividers when unused */
 	if (priv->csrc != nv_clk_src_core) {
 		nv_wr32(clk, 0x4040, 0x00000000);
@@ -395,6 +378,12 @@ resume:
 		nv_mask(clk, 0x4020, 0x80000000, 0x00000000);
 	}
 
+out:
+	if (ret == -EBUSY)
+		f = NULL;
+
+	nva3_clock_post(clk, f);
+
 	return ret;
 }
 
@@ -421,8 +410,8 @@ nvaa_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nvaa_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nvaa_domains, true,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nvaa_domains, NULL,
+				   0, true, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
index dbf8517f54da..1234abaab2db 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
@@ -437,8 +437,8 @@ nvc0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nvc0_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nvc0_domain, false,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nvc0_domain, NULL, 0,
+				   false, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nve0.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nve0.c
index 0e62a3240144..7eccad57512e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nve0.c
@@ -475,8 +475,8 @@ nve0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nve0_clock_priv *priv;
 	int ret;
 
-	ret = nouveau_clock_create(parent, engine, oclass, nve0_domain, true,
-				   &priv);
+	ret = nouveau_clock_create(parent, engine, oclass, nve0_domain, NULL, 0,
+				   true, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
index 4fe49cf4c99a..6103484fea72 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/fbmem.h
@@ -26,22 +26,8 @@
 
 #include <core/device.h>
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+#include <subdev/fb/regsnv04.h>
+
 #define NV04_PFB_DEBUG_0					0x00100080
 #	define NV04_PFB_DEBUG_0_PAGE_MODE			0x00000001
 #	define NV04_PFB_DEBUG_0_REFRESH_OFF			0x00000010
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
index 66fe959b4f74..7fbbe05d5c60 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/gddr5.c
@@ -40,7 +40,7 @@ nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts)
 	int WL, CL, WR, at[2], dt, ds;
 	int rq = ram->freq < 1000000; /* XXX */
 
-	switch (ram->ramcfg.version) {
+	switch (ram->next->bios.ramcfg_ver) {
 	case 0x11:
 		pd =  ram->next->bios.ramcfg_11_01_80;
 		lf =  ram->next->bios.ramcfg_11_01_40;
@@ -54,7 +54,7 @@ nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts)
 		return -ENOSYS;
 	}
 
-	switch (ram->timing.version) {
+	switch (ram->next->bios.timing_ver) {
 	case 0x20:
 		WL = (ram->next->bios.timing[1] & 0x00000f80) >> 7;
 		CL = (ram->next->bios.timing[1] & 0x0000001f);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
index a16024a74771..fde42e4d1b56 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/gk20a.c
@@ -27,6 +27,20 @@ struct gk20a_fb_priv {
 };
 
 static int
+gk20a_fb_init(struct nouveau_object *object)
+{
+	struct gk20a_fb_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_fb_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_mask(priv, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */
+	return 0;
+}
+
+static int
 gk20a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -48,7 +62,7 @@ gk20a_fb_oclass = &(struct nouveau_fb_impl) {
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = gk20a_fb_ctor,
 		.dtor = _nouveau_fb_dtor,
-		.init = _nouveau_fb_init,
+		.init = gk20a_fb_init,
 		.fini = _nouveau_fb_fini,
 	},
 	.memtype = nvc0_fb_memtype_valid,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
index f003c1b1893f..2209ade63339 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
@@ -45,7 +45,7 @@ nv20_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00000000; /* Z16 */
 		else              tile->zcomp = 0x04000000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
index f34f4223210b..e2a66c355c50 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
@@ -32,7 +32,7 @@ nv25_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00100000; /* Z16 */
 		else              tile->zcomp = 0x00200000; /* Z24S8 */
 		tile->zcomp |= tile->tag->offset;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
index 69093f7151f0..cbec402ba5b9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
@@ -51,7 +51,7 @@ nv30_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x01000000; /* Z16 */
 		else           tile->zcomp |= 0x02000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
index 161b06e8fc3f..b2cf8c69fb2e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
@@ -32,7 +32,7 @@ nv35_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x04000000; /* Z16 */
 		else           tile->zcomp |= 0x08000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
index 2dd3d0aab6bb..b4cdae2a3b2f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
@@ -32,7 +32,7 @@ nv36_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
-	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	if (!nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x10000000; /* Z16 */
 		else           tile->zcomp |= 0x20000000; /* Z24S8 */
 		tile->zcomp |= ((tile->tag->offset           ) >> 6);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
index 95a115ab0c86..52814258c212 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
@@ -33,7 +33,7 @@ nv40_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 	u32 tiles = DIV_ROUND_UP(size, 0x80);
 	u32 tags  = round_up(tiles / pfb->ram->parts, 0x100);
 	if ( (flags & 2) &&
-	    !nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
+	    !nouveau_mm_head(&pfb->tags, 0, 1, tags, tags, 1, &tile->tag)) {
 		tile->zcomp  = 0x28000000; /* Z24S8_SPLIT_GRAD */
 		tile->zcomp |= ((tile->tag->offset           ) >> 8);
 		tile->zcomp |= ((tile->tag->offset + tags - 1) >> 8) << 13;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 1fc55c1e91a1..4150b0d10af8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -250,9 +250,11 @@ nv50_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	priv->r100c08_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (priv->r100c08_page) {
-		priv->r100c08 = nv_device_map_page(device, priv->r100c08_page);
-		if (!priv->r100c08)
-			nv_warn(priv, "failed 0x100c08 page map\n");
+		priv->r100c08 = dma_map_page(nv_device_base(device),
+					     priv->r100c08_page, 0, PAGE_SIZE,
+					     DMA_BIDIRECTIONAL);
+		if (dma_mapping_error(nv_device_base(device), priv->r100c08))
+			return -EFAULT;
 	} else {
 		nv_warn(priv, "failed 0x100c08 page alloc\n");
 	}
@@ -268,7 +270,8 @@ nv50_fb_dtor(struct nouveau_object *object)
 	struct nv50_fb_priv *priv = (void *)object;
 
 	if (priv->r100c08_page) {
-		nv_device_unmap_page(device, priv->r100c08);
+		dma_unmap_page(nv_device_base(device), priv->r100c08, PAGE_SIZE,
+			       DMA_BIDIRECTIONAL);
 		__free_page(priv->r100c08_page);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
index 0670ae33ee45..32f28dc73ef2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
@@ -60,6 +60,7 @@ nvc0_fb_init(struct nouveau_object *object)
 
 	if (priv->r100c10_page)
 		nv_wr32(priv, 0x100c10, priv->r100c10 >> 8);
+	nv_mask(priv, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */
 	return 0;
 }
 
@@ -70,7 +71,8 @@ nvc0_fb_dtor(struct nouveau_object *object)
 	struct nvc0_fb_priv *priv = (void *)object;
 
 	if (priv->r100c10_page) {
-		nv_device_unmap_page(device, priv->r100c10);
+		dma_unmap_page(nv_device_base(device), priv->r100c10, PAGE_SIZE,
+			       DMA_BIDIRECTIONAL);
 		__free_page(priv->r100c10_page);
 	}
 
@@ -93,8 +95,10 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	priv->r100c10_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (priv->r100c10_page) {
-		priv->r100c10 = nv_device_map_page(device, priv->r100c10_page);
-		if (!priv->r100c10)
+		priv->r100c10 = dma_map_page(nv_device_base(device),
+					     priv->r100c10_page, 0, PAGE_SIZE,
+					     DMA_BIDIRECTIONAL);
+		if (dma_mapping_error(nv_device_base(device), priv->r100c10))
 			return -EFAULT;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
index 82273f832e42..60322e906dd4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
@@ -35,6 +35,7 @@ extern struct nouveau_oclass nve0_ram_oclass;
 extern struct nouveau_oclass gk20a_ram_oclass;
 extern struct nouveau_oclass gm107_ram_oclass;
 
+int nouveau_sddr2_calc(struct nouveau_ram *ram);
 int nouveau_sddr3_calc(struct nouveau_ram *ram);
 int nouveau_gddr5_calc(struct nouveau_ram *ram, bool nuts);
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h b/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
index 2af9cfd2c60f..d1fbbe4b00a2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramfuc.h
@@ -12,16 +12,32 @@ struct ramfuc {
 struct ramfuc_reg {
 	int sequence;
 	bool force;
-	u32 addr[2];
+	u32 addr;
+	u32 stride; /* in bytes */
+	u32 mask;
 	u32 data;
 };
 
 static inline struct ramfuc_reg
+ramfuc_stride(u32 addr, u32 stride, u32 mask)
+{
+	return (struct ramfuc_reg) {
+		.sequence = 0,
+		.addr = addr,
+		.stride = stride,
+		.mask = mask,
+		.data = 0xdeadbeef,
+	};
+}
+
+static inline struct ramfuc_reg
 ramfuc_reg2(u32 addr1, u32 addr2)
 {
 	return (struct ramfuc_reg) {
 		.sequence = 0,
-		.addr = { addr1, addr2 },
+		.addr = addr1,
+		.stride = addr2 - addr1,
+		.mask = 0x3,
 		.data = 0xdeadbeef,
 	};
 }
@@ -29,7 +45,13 @@ ramfuc_reg2(u32 addr1, u32 addr2)
 static noinline struct ramfuc_reg
 ramfuc_reg(u32 addr)
 {
-	return ramfuc_reg2(addr, addr);
+	return (struct ramfuc_reg) {
+		.sequence = 0,
+		.addr = addr,
+		.stride = 0,
+		.mask = 0x1,
+		.data = 0xdeadbeef,
+	};
 }
 
 static inline int
@@ -62,18 +84,25 @@ static inline u32
 ramfuc_rd32(struct ramfuc *ram, struct ramfuc_reg *reg)
 {
 	if (reg->sequence != ram->sequence)
-		reg->data = nv_rd32(ram->pfb, reg->addr[0]);
+		reg->data = nv_rd32(ram->pfb, reg->addr);
 	return reg->data;
 }
 
 static inline void
 ramfuc_wr32(struct ramfuc *ram, struct ramfuc_reg *reg, u32 data)
 {
+	unsigned int mask, off = 0;
+
 	reg->sequence = ram->sequence;
 	reg->data = data;
-	if (reg->addr[0] != reg->addr[1])
-		nouveau_memx_wr32(ram->memx, reg->addr[1], reg->data);
-	nouveau_memx_wr32(ram->memx, reg->addr[0], reg->data);
+
+	for (mask = reg->mask; mask > 0; mask = (mask & ~1) >> 1) {
+		if (mask & 1) {
+			nouveau_memx_wr32(ram->memx, reg->addr+off, reg->data);
+		}
+
+		off += reg->stride;
+	}
 }
 
 static inline void
@@ -105,14 +134,35 @@ ramfuc_nsec(struct ramfuc *ram, u32 nsec)
 	nouveau_memx_nsec(ram->memx, nsec);
 }
 
-#define ram_init(s,p)       ramfuc_init(&(s)->base, (p))
-#define ram_exec(s,e)       ramfuc_exec(&(s)->base, (e))
-#define ram_have(s,r)       ((s)->r_##r.addr[0] != 0x000000)
-#define ram_rd32(s,r)       ramfuc_rd32(&(s)->base, &(s)->r_##r)
-#define ram_wr32(s,r,d)     ramfuc_wr32(&(s)->base, &(s)->r_##r, (d))
-#define ram_nuke(s,r)       ramfuc_nuke(&(s)->base, &(s)->r_##r)
-#define ram_mask(s,r,m,d)   ramfuc_mask(&(s)->base, &(s)->r_##r, (m), (d))
-#define ram_wait(s,r,m,d,n) ramfuc_wait(&(s)->base, (r), (m), (d), (n))
-#define ram_nsec(s,n)       ramfuc_nsec(&(s)->base, (n))
+static inline void
+ramfuc_wait_vblank(struct ramfuc *ram)
+{
+	nouveau_memx_wait_vblank(ram->memx);
+}
+
+static inline void
+ramfuc_block(struct ramfuc *ram)
+{
+	nouveau_memx_block(ram->memx);
+}
+
+static inline void
+ramfuc_unblock(struct ramfuc *ram)
+{
+	nouveau_memx_unblock(ram->memx);
+}
+
+#define ram_init(s,p)        ramfuc_init(&(s)->base, (p))
+#define ram_exec(s,e)        ramfuc_exec(&(s)->base, (e))
+#define ram_have(s,r)        ((s)->r_##r.addr != 0x000000)
+#define ram_rd32(s,r)        ramfuc_rd32(&(s)->base, &(s)->r_##r)
+#define ram_wr32(s,r,d)      ramfuc_wr32(&(s)->base, &(s)->r_##r, (d))
+#define ram_nuke(s,r)        ramfuc_nuke(&(s)->base, &(s)->r_##r)
+#define ram_mask(s,r,m,d)    ramfuc_mask(&(s)->base, &(s)->r_##r, (m), (d))
+#define ram_wait(s,r,m,d,n)  ramfuc_wait(&(s)->base, (r), (m), (d), (n))
+#define ram_nsec(s,n)        ramfuc_nsec(&(s)->base, (n))
+#define ram_wait_vblank(s)   ramfuc_wait_vblank(&(s)->base)
+#define ram_block(s)         ramfuc_block(&(s)->base)
+#define ram_unblock(s)       ramfuc_unblock(&(s)->base)
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
index e781080d3327..1972268d1410 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
@@ -22,22 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+#include <subdev/fb/regsnv04.h>
 
 #include "priv.h"
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
index e5d12c24cc43..64a983c96625 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
@@ -280,7 +280,7 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 		if (align == 16) {
 			int n = (max >> 4) * comp;
 
-			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
+			ret = nouveau_mm_head(tags, 0, 1, n, n, 1, &mem->tag);
 			if (ret)
 				mem->tag = NULL;
 		}
@@ -296,9 +296,9 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 	type = nv50_fb_memtype[type];
 	do {
 		if (back)
-			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
+			ret = nouveau_mm_tail(heap, 0, type, max, min, align, &r);
 		else
-			ret = nouveau_mm_head(heap, type, max, min, align, &r);
+			ret = nouveau_mm_head(heap, 0, type, max, min, align, &r);
 		if (ret) {
 			mutex_unlock(&pfb->base.mutex);
 			pfb->ram->put(pfb, &mem);
@@ -319,27 +319,22 @@ nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 static u32
 nv50_fb_vram_rblock(struct nouveau_fb *pfb, struct nouveau_ram *ram)
 {
-	int i, parts, colbits, rowbitsa, rowbitsb, banks;
+	int colbits, rowbitsa, rowbitsb, banks;
 	u64 rowsize, predicted;
-	u32 r0, r4, rt, ru, rblock_size;
+	u32 r0, r4, rt, rblock_size;
 
 	r0 = nv_rd32(pfb, 0x100200);
 	r4 = nv_rd32(pfb, 0x100204);
 	rt = nv_rd32(pfb, 0x100250);
-	ru = nv_rd32(pfb, 0x001540);
-	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
-
-	for (i = 0, parts = 0; i < 8; i++) {
-		if (ru & (0x00010000 << i))
-			parts++;
-	}
+	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt,
+			nv_rd32(pfb, 0x001540));
 
 	colbits  =  (r4 & 0x0000f000) >> 12;
 	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
 	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
 	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
 
-	rowsize = parts * banks * (1 << colbits) * 8;
+	rowsize = ram->parts * banks * (1 << colbits) * 8;
 	predicted = rowsize << rowbitsa;
 	if (r0 & 0x00000004)
 		predicted += rowsize << rowbitsb;
@@ -376,6 +371,9 @@ nv50_ram_create_(struct nouveau_object *parent, struct nouveau_object *engine,
 	ram->size = nv_rd32(pfb, 0x10020c);
 	ram->size = (ram->size & 0xffffff00) | ((ram->size & 0x000000ff) << 32);
 
+	ram->part_mask = (nv_rd32(pfb, 0x001540) & 0x00ff0000) >> 16;
+	ram->parts = hweight8(ram->part_mask);
+
 	switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
 	case 0: ram->type = NV_MEM_TYPE_DDR1; break;
 	case 1:
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
index 8076fb195dd5..3601deca0bd5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnva3.c
@@ -79,20 +79,27 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	struct nva3_ram *ram = (void *)pfb->ram;
 	struct nva3_ramfuc *fuc = &ram->fuc;
 	struct nva3_clock_info mclk;
-	u8  ver, cnt, len, strap;
+	struct nouveau_ram_data *next;
+	u8  ver, hdr, cnt, len, strap;
 	u32 data;
-	struct {
-		u32 data;
-		u8  size;
-	} rammap, ramcfg, timing;
 	u32 r004018, r100760, ctrl;
 	u32 unk714, unk718, unk71c;
-	int ret;
+	int ret, i;
+
+	next = &ram->base.target;
+	next->freq = freq;
+	ram->base.next = next;
 
 	/* lookup memory config data relevant to the target frequency */
-	rammap.data = nvbios_rammapEm(bios, freq / 1000, &ver, &rammap.size,
-				     &cnt, &ramcfg.size);
-	if (!rammap.data || ver != 0x10 || rammap.size < 0x0e) {
+	i = 0;
+	while ((data = nvbios_rammapEp(bios, i++, &ver, &hdr, &cnt, &len,
+				      &next->bios))) {
+		if (freq / 1000 >= next->bios.rammap_min &&
+		    freq / 1000 <= next->bios.rammap_max)
+			break;
+	}
+
+	if (!data || ver != 0x10 || hdr < 0x0e) {
 		nv_error(pfb, "invalid/missing rammap entry\n");
 		return -EINVAL;
 	}
@@ -104,26 +111,25 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 		return -EINVAL;
 	}
 
-	ramcfg.data = rammap.data + rammap.size + (strap * ramcfg.size);
-	if (!ramcfg.data || ver != 0x10 || ramcfg.size < 0x0e) {
+	data = nvbios_rammapSp(bios, data, ver, hdr, cnt, len, strap,
+			       &ver, &hdr, &next->bios);
+	if (!data || ver != 0x10 || hdr < 0x0e) {
 		nv_error(pfb, "invalid/missing ramcfg entry\n");
 		return -EINVAL;
 	}
 
 	/* lookup memory timings, if bios says they're present */
-	strap = nv_ro08(bios, ramcfg.data + 0x01);
-	if (strap != 0xff) {
-		timing.data = nvbios_timingEe(bios, strap, &ver, &timing.size,
-					     &cnt, &len);
-		if (!timing.data || ver != 0x10 || timing.size < 0x19) {
+	if (next->bios.ramcfg_timing != 0xff) {
+		data = nvbios_timingEp(bios, next->bios.ramcfg_timing,
+				       &ver, &hdr, &cnt, &len,
+				       &next->bios);
+		if (!data || ver != 0x10 || hdr < 0x19) {
 			nv_error(pfb, "invalid/missing timing entry\n");
 			return -EINVAL;
 		}
-	} else {
-		timing.data = 0;
 	}
 
-	ret = nva3_clock_info(nouveau_clock(pfb), 0x12, 0x4000, freq, &mclk);
+	ret = nva3_pll_info(nouveau_clock(pfb), 0x12, 0x4000, freq, &mclk);
 	if (ret < 0) {
 		nv_error(pfb, "failed mclk calculation\n");
 		return ret;
@@ -163,17 +169,17 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 		ram_mask(fuc, 0x004168, 0x003f3141, ctrl);
 	}
 
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)) {
+	if (next->bios.ramcfg_10_02_10) {
 		ram_mask(fuc, 0x111104, 0x00000600, 0x00000000);
 	} else {
 		ram_mask(fuc, 0x111100, 0x40000000, 0x40000000);
 		ram_mask(fuc, 0x111104, 0x00000180, 0x00000000);
 	}
 
-	if (!(nv_ro08(bios, rammap.data + 0x04) & 0x02))
+	if (!next->bios.rammap_10_04_02)
 		ram_mask(fuc, 0x100200, 0x00000800, 0x00000000);
 	ram_wr32(fuc, 0x611200, 0x00003300);
-	if (!(nv_ro08(bios, ramcfg.data + 0x02) & 0x10))
+	if (!next->bios.ramcfg_10_02_10)
 		ram_wr32(fuc, 0x111100, 0x4c020000); /*XXX*/
 
 	ram_wr32(fuc, 0x1002d4, 0x00000001);
@@ -202,17 +208,16 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 		ram_wr32(fuc, 0x004018, 0x0000d000 | r004018);
 	}
 
-	if ( (nv_ro08(bios, rammap.data + 0x04) & 0x08)) {
-		u32 unk5a0 = (nv_ro16(bios, ramcfg.data + 0x05) << 8) |
-			      nv_ro08(bios, ramcfg.data + 0x05);
-		u32 unk5a4 = (nv_ro16(bios, ramcfg.data + 0x07));
-		u32 unk804 = (nv_ro08(bios, ramcfg.data + 0x09) & 0xf0) << 16 |
-			     (nv_ro08(bios, ramcfg.data + 0x03) & 0x0f) << 16 |
-			     (nv_ro08(bios, ramcfg.data + 0x09) & 0x0f) |
-			     0x80000000;
-		ram_wr32(fuc, 0x1005a0, unk5a0);
-		ram_wr32(fuc, 0x1005a4, unk5a4);
-		ram_wr32(fuc, 0x10f804, unk804);
+	if (next->bios.rammap_10_04_08) {
+		ram_wr32(fuc, 0x1005a0, next->bios.ramcfg_10_06 << 16 |
+					next->bios.ramcfg_10_05 << 8 |
+					next->bios.ramcfg_10_05);
+		ram_wr32(fuc, 0x1005a4, next->bios.ramcfg_10_08 << 8 |
+					next->bios.ramcfg_10_07);
+		ram_wr32(fuc, 0x10f804, next->bios.ramcfg_10_09_f0 << 20 |
+					next->bios.ramcfg_10_03_0f << 16 |
+					next->bios.ramcfg_10_09_0f |
+					0x80000000);
 		ram_mask(fuc, 0x10053c, 0x00001000, 0x00000000);
 	} else {
 		ram_mask(fuc, 0x10053c, 0x00001000, 0x00001000);
@@ -250,27 +255,26 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x100220[0], 0x00000000, 0x00000000);
 	ram_mask(fuc, 0x100220[8], 0x00000000, 0x00000000);
 
-	data = (nv_ro08(bios, ramcfg.data + 0x02) & 0x08) ? 0x00000000 : 0x00001000;
-	ram_mask(fuc, 0x100200, 0x00001000, data);
+	ram_mask(fuc, 0x100200, 0x00001000, !next->bios.ramcfg_10_02_08 << 12);
 
 	unk714 = ram_rd32(fuc, 0x100714) & ~0xf0000010;
 	unk718 = ram_rd32(fuc, 0x100718) & ~0x00000100;
 	unk71c = ram_rd32(fuc, 0x10071c) & ~0x00000100;
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x20))
+	if (next->bios.ramcfg_10_02_20)
 		unk714 |= 0xf0000000;
-	if (!(nv_ro08(bios, ramcfg.data + 0x02) & 0x04))
+	if (!next->bios.ramcfg_10_02_04)
 		unk714 |= 0x00000010;
 	ram_wr32(fuc, 0x100714, unk714);
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x01)
+	if (next->bios.ramcfg_10_02_01)
 		unk71c |= 0x00000100;
 	ram_wr32(fuc, 0x10071c, unk71c);
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x02)
+	if (next->bios.ramcfg_10_02_02)
 		unk718 |= 0x00000100;
 	ram_wr32(fuc, 0x100718, unk718);
 
-	if (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)
+	if (next->bios.ramcfg_10_02_10)
 		ram_wr32(fuc, 0x111100, 0x48000000); /*XXX*/
 
 	ram_mask(fuc, mr[0], 0x100, 0x100);
@@ -282,9 +286,9 @@ nva3_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	ram_nsec(fuc, 12000);
 
 	ram_wr32(fuc, 0x611200, 0x00003330);
-	if ( (nv_ro08(bios, rammap.data + 0x04) & 0x02))
+	if (next->bios.rammap_10_04_02)
 		ram_mask(fuc, 0x100200, 0x00000800, 0x00000800);
-	if ( (nv_ro08(bios, ramcfg.data + 0x02) & 0x10)) {
+	if (next->bios.ramcfg_10_02_10) {
 		ram_mask(fuc, 0x111104, 0x00000180, 0x00000180);
 		ram_mask(fuc, 0x111100, 0x40000000, 0x00000000);
 	} else {
@@ -404,11 +408,11 @@ nva3_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	ram->fuc.r_0x100714 = ramfuc_reg(0x100714);
 	ram->fuc.r_0x100718 = ramfuc_reg(0x100718);
 	ram->fuc.r_0x10071c = ramfuc_reg(0x10071c);
-	ram->fuc.r_0x100760 = ramfuc_reg(0x100760);
-	ram->fuc.r_0x1007a0 = ramfuc_reg(0x1007a0);
-	ram->fuc.r_0x1007e0 = ramfuc_reg(0x1007e0);
+	ram->fuc.r_0x100760 = ramfuc_stride(0x100760, 4, ram->base.part_mask);
+	ram->fuc.r_0x1007a0 = ramfuc_stride(0x1007a0, 4, ram->base.part_mask);
+	ram->fuc.r_0x1007e0 = ramfuc_stride(0x1007e0, 4, ram->base.part_mask);
 	ram->fuc.r_0x10f804 = ramfuc_reg(0x10f804);
-	ram->fuc.r_0x1110e0 = ramfuc_reg(0x1110e0);
+	ram->fuc.r_0x1110e0 = ramfuc_stride(0x1110e0, 4, ram->base.part_mask);
 	ram->fuc.r_0x111100 = ramfuc_reg(0x111100);
 	ram->fuc.r_0x111104 = ramfuc_reg(0x111104);
 	ram->fuc.r_0x611200 = ramfuc_reg(0x611200);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
index 5a6a5027f749..735cb9580abe 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
@@ -26,7 +26,7 @@
 #include <subdev/bios/pll.h>
 #include <subdev/bios/rammap.h>
 #include <subdev/bios/timing.h>
-#include <subdev/ltcg.h>
+#include <subdev/ltc.h>
 
 #include <subdev/clock.h>
 #include <subdev/clock/pll.h>
@@ -133,6 +133,7 @@ nvc0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	struct nouveau_bios *bios = nouveau_bios(pfb);
 	struct nvc0_ram *ram = (void *)pfb->ram;
 	struct nvc0_ramfuc *fuc = &ram->fuc;
+	struct nvbios_ramcfg cfg;
 	u8  ver, cnt, len, strap;
 	struct {
 		u32 data;
@@ -145,7 +146,7 @@ nvc0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 
 	/* lookup memory config data relevant to the target frequency */
 	rammap.data = nvbios_rammapEm(bios, freq / 1000, &ver, &rammap.size,
-				     &cnt, &ramcfg.size);
+				     &cnt, &ramcfg.size, &cfg);
 	if (!rammap.data || ver != 0x10 || rammap.size < 0x0e) {
 		nv_error(pfb, "invalid/missing rammap entry\n");
 		return -EINVAL;
@@ -425,7 +426,7 @@ extern const u8 nvc0_pte_storage_type_map[256];
 void
 nvc0_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
 {
-	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+	struct nouveau_ltc *ltc = nouveau_ltc(pfb);
 	struct nouveau_mem *mem = *pmem;
 
 	*pmem = NULL;
@@ -434,7 +435,7 @@ nvc0_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
 
 	mutex_lock(&pfb->base.mutex);
 	if (mem->tag)
-		ltcg->tags_free(ltcg, &mem->tag);
+		ltc->tags_free(ltc, &mem->tag);
 	__nv50_ram_put(pfb, mem);
 	mutex_unlock(&pfb->base.mutex);
 
@@ -468,12 +469,12 @@ nvc0_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 
 	mutex_lock(&pfb->base.mutex);
 	if (comp) {
-		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+		struct nouveau_ltc *ltc = nouveau_ltc(pfb);
 
 		/* compression only works with lpages */
 		if (align == (1 << (17 - 12))) {
 			int n = size >> 5;
-			ltcg->tags_alloc(ltcg, n, &mem->tag);
+			ltc->tags_alloc(ltc, n, &mem->tag);
 		}
 
 		if (unlikely(!mem->tag))
@@ -483,9 +484,9 @@ nvc0_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
 
 	do {
 		if (back)
-			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
+			ret = nouveau_mm_tail(mm, 0, 1, size, ncmin, align, &r);
 		else
-			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
+			ret = nouveau_mm_head(mm, 0, 1, size, ncmin, align, &r);
 		if (ret) {
 			mutex_unlock(&pfb->base.mutex);
 			pfb->ram->put(pfb, &mem);
@@ -554,15 +555,15 @@ nvc0_ram_create_(struct nouveau_object *parent, struct nouveau_object *engine,
 	} else {
 		/* otherwise, address lowest common amount from 0GiB */
 		ret = nouveau_mm_init(&pfb->vram, rsvd_head,
-				      (bsize << 8) * parts, 1);
+				      (bsize << 8) * parts - rsvd_head, 1);
 		if (ret)
 			return ret;
 
 		/* and the rest starting from (8GiB + common_size) */
 		offset = (0x0200000000ULL >> 12) + (bsize << 8);
-		length = (ram->size >> 12) - (bsize << 8) - rsvd_tail;
+		length = (ram->size >> 12) - ((bsize * parts) << 8) - rsvd_tail;
 
-		ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 1);
 		if (ret)
 			nouveau_mm_fini(&pfb->vram);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
index c5b46e302319..6bae474abb44 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnve0.c
@@ -29,6 +29,8 @@
 #include <subdev/bios/init.h>
 #include <subdev/bios/rammap.h>
 #include <subdev/bios/timing.h>
+#include <subdev/bios/M0205.h>
+#include <subdev/bios/M0209.h>
 
 #include <subdev/clock.h>
 #include <subdev/clock/pll.h>
@@ -41,14 +43,6 @@
 
 #include "ramfuc.h"
 
-/* binary driver only executes this path if the condition (a) is true
- * for any configuration (combination of rammap+ramcfg+timing) that
- * can be reached on a given card.  for now, we will execute the branch
- * unconditionally in the hope that a "false everywhere" in the bios
- * tables doesn't actually mean "don't touch this".
- */
-#define NOTE00(a) 1
-
 struct nve0_ramfuc {
 	struct ramfuc base;
 
@@ -134,10 +128,12 @@ struct nve0_ram {
 	struct nouveau_ram base;
 	struct nve0_ramfuc fuc;
 
+	struct list_head cfg;
 	u32 parts;
 	u32 pmask;
 	u32 pnuts;
 
+	struct nvbios_ramcfg diff;
 	int from;
 	int mode;
 	int N1, fN1, M1, P1;
@@ -241,7 +237,7 @@ nve0_ram_nuts(struct nve0_ram *ram, struct ramfuc_reg *reg,
 {
 	struct nve0_fb_priv *priv = (void *)nouveau_fb(ram);
 	struct ramfuc *fuc = &ram->fuc.base;
-	u32 addr = 0x110000 + (reg->addr[0] & 0xfff);
+	u32 addr = 0x110000 + (reg->addr & 0xfff);
 	u32 mask = _mask | _copy;
 	u32 data = (_data & _mask) | (reg->data & _copy);
 	u32 i;
@@ -268,6 +264,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	u32 mask, data;
 
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
+	ram_block(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	/* MR1: turn termination on early, for some reason.. */
@@ -478,7 +475,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f2e8, 0xffffffff, next->bios.timing[9]);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_08_20)) {
+	if (ram->diff.ramcfg_11_08_20) {
 		if (next->bios.ramcfg_11_08_20)
 			data |= 0x01000000;
 		mask |= 0x01000000;
@@ -486,11 +483,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f200, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_02_03 != 0)) {
+	if (ram->diff.ramcfg_11_02_03) {
 		data |= next->bios.ramcfg_11_02_03 << 8;
 		mask |= 0x00000300;
 	}
-	if (NOTE00(ramcfg_01_10)) {
+	if (ram->diff.ramcfg_11_01_10) {
 		if (next->bios.ramcfg_11_01_10)
 			data |= 0x70000000;
 		mask |= 0x70000000;
@@ -498,11 +495,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f604, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(timing_30_07 != 0)) {
+	if (ram->diff.timing_20_30_07) {
 		data |= next->bios.timing_20_30_07 << 28;
 		mask |= 0x70000000;
 	}
-	if (NOTE00(ramcfg_01_01)) {
+	if (ram->diff.ramcfg_11_01_01) {
 		if (next->bios.ramcfg_11_01_01)
 			data |= 0x00000100;
 		mask |= 0x00000100;
@@ -510,11 +507,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f614, mask, data);
 
 	data = mask = 0x00000000;
-	if (NOTE00(timing_30_07 != 0)) {
+	if (ram->diff.timing_20_30_07) {
 		data |= next->bios.timing_20_30_07 << 28;
 		mask |= 0x70000000;
 	}
-	if (NOTE00(ramcfg_01_02)) {
+	if (ram->diff.ramcfg_11_01_02) {
 		if (next->bios.ramcfg_11_01_02)
 			data |= 0x00000100;
 		mask |= 0x00000100;
@@ -548,11 +545,11 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	ram_wr32(fuc, 0x10f870, 0x11111111 * next->bios.ramcfg_11_03_0f);
 
 	data = mask = 0x00000000;
-	if (NOTE00(ramcfg_02_03 != 0)) {
+	if (ram->diff.ramcfg_11_02_03) {
 		data |= next->bios.ramcfg_11_02_03;
 		mask |= 0x00000003;
 	}
-	if (NOTE00(ramcfg_01_10)) {
+	if (ram->diff.ramcfg_11_01_10) {
 		if (next->bios.ramcfg_11_01_10)
 			data |= 0x00000004;
 		mask |= 0x00000004;
@@ -666,6 +663,7 @@ nve0_ram_calc_gddr5(struct nouveau_fb *pfb, u32 freq)
 	if (next->bios.ramcfg_11_07_02)
 		nve0_ram_train(fuc, 0x80020000, 0x01000000);
 
+	ram_unblock(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
@@ -695,6 +693,7 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
 	u32 mask, data;
 
 	ram_mask(fuc, 0x10f808, 0x40000000, 0x40000000);
+	ram_block(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0000);
 
 	if (vc == 1 && ram_have(fuc, gpio2E)) {
@@ -917,6 +916,7 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
 	ram_mask(fuc, 0x10f200, 0x80000000, 0x00000000);
 	ram_nsec(fuc, 1000);
 
+	ram_unblock(fuc);
 	ram_wr32(fuc, 0x62c000, 0x0f0f0f00);
 
 	if (next->bios.rammap_11_08_01)
@@ -932,58 +932,24 @@ nve0_ram_calc_sddr3(struct nouveau_fb *pfb, u32 freq)
  ******************************************************************************/
 
 static int
-nve0_ram_calc_data(struct nouveau_fb *pfb, u32 freq,
+nve0_ram_calc_data(struct nouveau_fb *pfb, u32 khz,
 		   struct nouveau_ram_data *data)
 {
-	struct nouveau_bios *bios = nouveau_bios(pfb);
 	struct nve0_ram *ram = (void *)pfb->ram;
-	u8 strap, cnt, len;
-
-	/* lookup memory config data relevant to the target frequency */
-	ram->base.rammap.data = nvbios_rammapEp(bios, freq / 1000,
-					       &ram->base.rammap.version,
-					       &ram->base.rammap.size,
-					       &cnt, &len, &data->bios);
-	if (!ram->base.rammap.data || ram->base.rammap.version != 0x11 ||
-	     ram->base.rammap.size < 0x09) {
-		nv_error(pfb, "invalid/missing rammap entry\n");
-		return -EINVAL;
-	}
-
-	/* locate specific data set for the attached memory */
-	strap = nvbios_ramcfg_index(nv_subdev(pfb));
-	ram->base.ramcfg.data = nvbios_rammapSp(bios, ram->base.rammap.data,
-						ram->base.rammap.version,
-						ram->base.rammap.size,
-						cnt, len, strap,
-						&ram->base.ramcfg.version,
-						&ram->base.ramcfg.size,
-						&data->bios);
-	if (!ram->base.ramcfg.data || ram->base.ramcfg.version != 0x11 ||
-	     ram->base.ramcfg.size < 0x08) {
-		nv_error(pfb, "invalid/missing ramcfg entry\n");
-		return -EINVAL;
-	}
-
-	/* lookup memory timings, if bios says they're present */
-	strap = nv_ro08(bios, ram->base.ramcfg.data + 0x00);
-	if (strap != 0xff) {
-		ram->base.timing.data =
-			nvbios_timingEp(bios, strap, &ram->base.timing.version,
-				       &ram->base.timing.size, &cnt, &len,
-				       &data->bios);
-		if (!ram->base.timing.data ||
-		     ram->base.timing.version != 0x20 ||
-		     ram->base.timing.size < 0x33) {
-			nv_error(pfb, "invalid/missing timing entry\n");
-			return -EINVAL;
+	struct nouveau_ram_data *cfg;
+	u32 mhz = khz / 1000;
+
+	list_for_each_entry(cfg, &ram->cfg, head) {
+		if (mhz >= cfg->bios.rammap_min &&
+		    mhz <= cfg->bios.rammap_max) {
+			*data = *cfg;
+			data->freq = khz;
+			return 0;
 		}
-	} else {
-		ram->base.timing.data = 0;
 	}
 
-	data->freq = freq;
-	return 0;
+	nv_error(ram, "ramcfg data for %dMHz not found\n", mhz);
+	return -EINVAL;
 }
 
 static int
@@ -1106,13 +1072,99 @@ nve0_ram_calc(struct nouveau_fb *pfb, u32 freq)
 	return nve0_ram_calc_xits(pfb, ram->base.next);
 }
 
+static void
+nve0_ram_prog_0(struct nouveau_fb *pfb, u32 freq)
+{
+	struct nve0_ram *ram = (void *)pfb->ram;
+	struct nouveau_ram_data *cfg;
+	u32 mhz = freq / 1000;
+	u32 mask, data;
+
+	list_for_each_entry(cfg, &ram->cfg, head) {
+		if (mhz >= cfg->bios.rammap_min &&
+		    mhz <= cfg->bios.rammap_max)
+			break;
+	}
+
+	if (&cfg->head == &ram->cfg)
+		return;
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_03fe) {
+		data |= cfg->bios.rammap_11_0a_03fe << 12;
+		mask |= 0x001ff000;
+	}
+	if (ram->diff.rammap_11_09_01ff) {
+		data |= cfg->bios.rammap_11_09_01ff;
+		mask |= 0x000001ff;
+	}
+	nv_mask(pfb, 0x10f468, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_0400) {
+		data |= cfg->bios.rammap_11_0a_0400;
+		mask |= 0x00000001;
+	}
+	nv_mask(pfb, 0x10f420, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0a_0800) {
+		data |= cfg->bios.rammap_11_0a_0800;
+		mask |= 0x00000001;
+	}
+	nv_mask(pfb, 0x10f430, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0b_01f0) {
+		data |= cfg->bios.rammap_11_0b_01f0;
+		mask |= 0x0000001f;
+	}
+	nv_mask(pfb, 0x10f400, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0b_0200) {
+		data |= cfg->bios.rammap_11_0b_0200 << 9;
+		mask |= 0x00000200;
+	}
+	nv_mask(pfb, 0x10f410, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0d) {
+		data |= cfg->bios.rammap_11_0d << 16;
+		mask |= 0x00ff0000;
+	}
+	if (ram->diff.rammap_11_0f) {
+		data |= cfg->bios.rammap_11_0f << 8;
+		mask |= 0x0000ff00;
+	}
+	nv_mask(pfb, 0x10f440, mask, data);
+
+	if (mask = 0, data = 0, ram->diff.rammap_11_0e) {
+		data |= cfg->bios.rammap_11_0e << 8;
+		mask |= 0x0000ff00;
+	}
+	if (ram->diff.rammap_11_0b_0800) {
+		data |= cfg->bios.rammap_11_0b_0800 << 7;
+		mask |= 0x00000080;
+	}
+	if (ram->diff.rammap_11_0b_0400) {
+		data |= cfg->bios.rammap_11_0b_0400 << 5;
+		mask |= 0x00000020;
+	}
+	nv_mask(pfb, 0x10f444, mask, data);
+}
+
 static int
 nve0_ram_prog(struct nouveau_fb *pfb)
 {
 	struct nouveau_device *device = nv_device(pfb);
 	struct nve0_ram *ram = (void *)pfb->ram;
 	struct nve0_ramfuc *fuc = &ram->fuc;
-	ram_exec(fuc, nouveau_boolopt(device->cfgopt, "NvMemExec", true));
+	struct nouveau_ram_data *next = ram->base.next;
+
+	if (!nouveau_boolopt(device->cfgopt, "NvMemExec", true)) {
+		ram_exec(fuc, false);
+		return (ram->base.next == &ram->base.xition);
+	}
+
+	nve0_ram_prog_0(pfb, 1000);
+	ram_exec(fuc, true);
+	nve0_ram_prog_0(pfb, next->freq);
+
 	return (ram->base.next == &ram->base.xition);
 }
 
@@ -1125,24 +1177,147 @@ nve0_ram_tidy(struct nouveau_fb *pfb)
 	ram_exec(fuc, false);
 }
 
+struct nve0_ram_train {
+	u16 mask;
+	struct nvbios_M0209S remap;
+	struct nvbios_M0209S type00;
+	struct nvbios_M0209S type01;
+	struct nvbios_M0209S type04;
+	struct nvbios_M0209S type06;
+	struct nvbios_M0209S type07;
+	struct nvbios_M0209S type08;
+	struct nvbios_M0209S type09;
+};
+
+static int
+nve0_ram_train_type(struct nouveau_fb *pfb, int i, u8 ramcfg,
+		    struct nve0_ram_train *train)
+{
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nvbios_M0205E M0205E;
+	struct nvbios_M0205S M0205S;
+	struct nvbios_M0209E M0209E;
+	struct nvbios_M0209S *remap = &train->remap;
+	struct nvbios_M0209S *value;
+	u8  ver, hdr, cnt, len;
+	u32 data;
+
+	/* determine type of data for this index */
+	if (!(data = nvbios_M0205Ep(bios, i, &ver, &hdr, &cnt, &len, &M0205E)))
+		return -ENOENT;
+
+	switch (M0205E.type) {
+	case 0x00: value = &train->type00; break;
+	case 0x01: value = &train->type01; break;
+	case 0x04: value = &train->type04; break;
+	case 0x06: value = &train->type06; break;
+	case 0x07: value = &train->type07; break;
+	case 0x08: value = &train->type08; break;
+	case 0x09: value = &train->type09; break;
+	default:
+		return 0;
+	}
+
+	/* training data index determined by ramcfg strap */
+	if (!(data = nvbios_M0205Sp(bios, i, ramcfg, &ver, &hdr, &M0205S)))
+		return -EINVAL;
+	i = M0205S.data;
+
+	/* training data format information */
+	if (!(data = nvbios_M0209Ep(bios, i, &ver, &hdr, &cnt, &len, &M0209E)))
+		return -EINVAL;
+
+	/* ... and the raw data */
+	if (!(data = nvbios_M0209Sp(bios, i, 0, &ver, &hdr, value)))
+		return -EINVAL;
+
+	if (M0209E.v02_07 == 2) {
+		/* of course! why wouldn't we have a pointer to another entry
+		 * in the same table, and use the first one as an array of
+		 * remap indices...
+		 */
+		if (!(data = nvbios_M0209Sp(bios, M0209E.v03, 0, &ver, &hdr,
+					    remap)))
+			return -EINVAL;
+
+		for (i = 0; i < ARRAY_SIZE(value->data); i++)
+			value->data[i] = remap->data[value->data[i]];
+	} else
+	if (M0209E.v02_07 != 1)
+		return -EINVAL;
+
+	train->mask |= 1 << M0205E.type;
+	return 0;
+}
+
+static int
+nve0_ram_train_init_0(struct nouveau_fb *pfb, struct nve0_ram_train *train)
+{
+	int i, j;
+
+	if ((train->mask & 0x03d3) != 0x03d3) {
+		nv_warn(pfb, "missing link training data\n");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < 0x30; i++) {
+		for (j = 0; j < 8; j += 4) {
+			nv_wr32(pfb, 0x10f968 + j, 0x00000000 | (i << 8));
+			nv_wr32(pfb, 0x10f920 + j, 0x00000000 |
+						   train->type08.data[i] << 4 |
+						   train->type06.data[i]);
+			nv_wr32(pfb, 0x10f918 + j, train->type00.data[i]);
+			nv_wr32(pfb, 0x10f920 + j, 0x00000100 |
+						   train->type09.data[i] << 4 |
+						   train->type07.data[i]);
+			nv_wr32(pfb, 0x10f918 + j, train->type01.data[i]);
+		}
+	}
+
+	for (j = 0; j < 8; j += 4) {
+		for (i = 0; i < 0x100; i++) {
+			nv_wr32(pfb, 0x10f968 + j, i);
+			nv_wr32(pfb, 0x10f900 + j, train->type04.data[i]);
+		}
+	}
+
+	return 0;
+}
+
+static int
+nve0_ram_train_init(struct nouveau_fb *pfb)
+{
+	u8 ramcfg = nvbios_ramcfg_index(nv_subdev(pfb));
+	struct nve0_ram_train *train;
+	int ret = -ENOMEM, i;
+
+	if ((train = kzalloc(sizeof(*train), GFP_KERNEL))) {
+		for (i = 0; i < 0x100; i++) {
+			ret = nve0_ram_train_type(pfb, i, ramcfg, train);
+			if (ret && ret != -ENOENT)
+				break;
+		}
+	}
+
+	switch (pfb->ram->type) {
+	case NV_MEM_TYPE_GDDR5:
+		ret = nve0_ram_train_init_0(pfb, train);
+		break;
+	default:
+		ret = 0;
+		break;
+	}
+
+	kfree(train);
+	return ret;
+}
+
 int
 nve0_ram_init(struct nouveau_object *object)
 {
 	struct nouveau_fb *pfb = (void *)object->parent;
 	struct nve0_ram *ram   = (void *)object;
 	struct nouveau_bios *bios = nouveau_bios(pfb);
-	static const u8  train0[] = {
-		0x00, 0xff, 0xff, 0x00, 0xff, 0x00,
-		0x00, 0xff, 0xff, 0x00, 0xff, 0x00,
-	};
-	static const u32 train1[] = {
-		0x00000000, 0xffffffff,
-		0x55555555, 0xaaaaaaaa,
-		0x33333333, 0xcccccccc,
-		0xf0f0f0f0, 0x0f0f0f0f,
-		0x00ff00ff, 0xff00ff00,
-		0x0000ffff, 0xffff0000,
-	};
 	u8  ver, hdr, cnt, len, snr, ssz;
 	u32 data, save;
 	int ret, i;
@@ -1168,51 +1343,107 @@ nve0_ram_init(struct nouveau_object *object)
 
 	cnt  = nv_ro08(bios, data + 0x14); /* guess at count */
 	data = nv_ro32(bios, data + 0x10); /* guess u32... */
-	save = nv_rd32(pfb, 0x10f65c);
-	for (i = 0; i < cnt; i++) {
-		nv_mask(pfb, 0x10f65c, 0x000000f0, i << 4);
-		nvbios_exec(&(struct nvbios_init) {
-				.subdev = nv_subdev(pfb),
-				.bios = bios,
-				.offset = nv_ro32(bios, data), /* guess u32 */
-				.execute = 1,
-			    });
-		data += 4;
-	}
-	nv_wr32(pfb, 0x10f65c, save);
+	save = nv_rd32(pfb, 0x10f65c) & 0x000000f0;
+	for (i = 0; i < cnt; i++, data += 4) {
+		if (i != save >> 4) {
+			nv_mask(pfb, 0x10f65c, 0x000000f0, i << 4);
+			nvbios_exec(&(struct nvbios_init) {
+					.subdev = nv_subdev(pfb),
+					.bios = bios,
+					.offset = nv_ro32(bios, data),
+					.execute = 1,
+				    });
+		}
+	}
+	nv_mask(pfb, 0x10f65c, 0x000000f0, save);
 	nv_mask(pfb, 0x10f584, 0x11000000, 0x00000000);
+	nv_wr32(pfb, 0x10ecc0, 0xffffffff);
+	nv_mask(pfb, 0x10f160, 0x00000010, 0x00000010);
 
-	switch (ram->base.type) {
-	case NV_MEM_TYPE_GDDR5:
-		for (i = 0; i < 0x30; i++) {
-			nv_wr32(pfb, 0x10f968, 0x00000000 | (i << 8));
-			nv_wr32(pfb, 0x10f920, 0x00000000 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f918,              train1[i % 12]);
-			nv_wr32(pfb, 0x10f920, 0x00000100 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f918,              train1[i % 12]);
-
-			nv_wr32(pfb, 0x10f96c, 0x00000000 | (i << 8));
-			nv_wr32(pfb, 0x10f924, 0x00000000 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f91c,              train1[i % 12]);
-			nv_wr32(pfb, 0x10f924, 0x00000100 | train0[i % 12]);
-			nv_wr32(pfb, 0x10f91c,              train1[i % 12]);
-		}
+	return nve0_ram_train_init(pfb);
+}
 
-		for (i = 0; i < 0x100; i++) {
-			nv_wr32(pfb, 0x10f968, i);
-			nv_wr32(pfb, 0x10f900, train1[2 + (i & 1)]);
-		}
+static int
+nve0_ram_ctor_data(struct nve0_ram *ram, u8 ramcfg, int i)
+{
+	struct nouveau_fb *pfb = (void *)nv_object(ram)->parent;
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nouveau_ram_data *cfg;
+	struct nvbios_ramcfg *d = &ram->diff;
+	struct nvbios_ramcfg *p, *n;
+	u8  ver, hdr, cnt, len;
+	u32 data;
+	int ret;
 
-		for (i = 0; i < 0x100; i++) {
-			nv_wr32(pfb, 0x10f96c, i);
-			nv_wr32(pfb, 0x10f900, train1[2 + (i & 1)]);
-		}
-		break;
-	default:
-		break;
+	if (!(cfg = kmalloc(sizeof(*cfg), GFP_KERNEL)))
+		return -ENOMEM;
+	p = &list_last_entry(&ram->cfg, typeof(*cfg), head)->bios;
+	n = &cfg->bios;
+
+	/* memory config data for a range of target frequencies */
+	data = nvbios_rammapEp(bios, i, &ver, &hdr, &cnt, &len, &cfg->bios);
+	if (ret = -ENOENT, !data)
+		goto done;
+	if (ret = -ENOSYS, ver != 0x11 || hdr < 0x12)
+		goto done;
+
+	/* ... and a portion specific to the attached memory */
+	data = nvbios_rammapSp(bios, data, ver, hdr, cnt, len, ramcfg,
+			       &ver, &hdr, &cfg->bios);
+	if (ret = -EINVAL, !data)
+		goto done;
+	if (ret = -ENOSYS, ver != 0x11 || hdr < 0x0a)
+		goto done;
+
+	/* lookup memory timings, if bios says they're present */
+	if (cfg->bios.ramcfg_timing != 0xff) {
+		data = nvbios_timingEp(bios, cfg->bios.ramcfg_timing,
+				       &ver, &hdr, &cnt, &len,
+				       &cfg->bios);
+		if (ret = -EINVAL, !data)
+			goto done;
+		if (ret = -ENOSYS, ver != 0x20 || hdr < 0x33)
+			goto done;
 	}
 
-	return 0;
+	list_add_tail(&cfg->head, &ram->cfg);
+	if (ret = 0, i == 0)
+		goto done;
+
+	d->rammap_11_0a_03fe |= p->rammap_11_0a_03fe != n->rammap_11_0a_03fe;
+	d->rammap_11_09_01ff |= p->rammap_11_09_01ff != n->rammap_11_09_01ff;
+	d->rammap_11_0a_0400 |= p->rammap_11_0a_0400 != n->rammap_11_0a_0400;
+	d->rammap_11_0a_0800 |= p->rammap_11_0a_0800 != n->rammap_11_0a_0800;
+	d->rammap_11_0b_01f0 |= p->rammap_11_0b_01f0 != n->rammap_11_0b_01f0;
+	d->rammap_11_0b_0200 |= p->rammap_11_0b_0200 != n->rammap_11_0b_0200;
+	d->rammap_11_0d |= p->rammap_11_0d != n->rammap_11_0d;
+	d->rammap_11_0f |= p->rammap_11_0f != n->rammap_11_0f;
+	d->rammap_11_0e |= p->rammap_11_0e != n->rammap_11_0e;
+	d->rammap_11_0b_0800 |= p->rammap_11_0b_0800 != n->rammap_11_0b_0800;
+	d->rammap_11_0b_0400 |= p->rammap_11_0b_0400 != n->rammap_11_0b_0400;
+	d->ramcfg_11_01_01 |= p->ramcfg_11_01_01 != n->ramcfg_11_01_01;
+	d->ramcfg_11_01_02 |= p->ramcfg_11_01_02 != n->ramcfg_11_01_02;
+	d->ramcfg_11_01_10 |= p->ramcfg_11_01_10 != n->ramcfg_11_01_10;
+	d->ramcfg_11_02_03 |= p->ramcfg_11_02_03 != n->ramcfg_11_02_03;
+	d->ramcfg_11_08_20 |= p->ramcfg_11_08_20 != n->ramcfg_11_08_20;
+	d->timing_20_30_07 |= p->timing_20_30_07 != n->timing_20_30_07;
+done:
+	if (ret)
+		kfree(cfg);
+	return ret;
+}
+
+static void
+nve0_ram_dtor(struct nouveau_object *object)
+{
+	struct nve0_ram *ram = (void *)object;
+	struct nouveau_ram_data *cfg, *tmp;
+
+	list_for_each_entry_safe(cfg, tmp, &ram->cfg, head) {
+		kfree(cfg);
+	}
+
+	nouveau_ram_destroy(&ram->base);
 }
 
 static int
@@ -1226,6 +1457,7 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct dcb_gpio_func func;
 	struct nve0_ram *ram;
 	int ret, i;
+	u8  ramcfg = nvbios_ramcfg_index(nv_subdev(pfb));
 	u32 tmp;
 
 	ret = nvc0_ram_create(parent, engine, oclass, 0x022554, &ram);
@@ -1233,6 +1465,8 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	INIT_LIST_HEAD(&ram->cfg);
+
 	switch (ram->base.type) {
 	case NV_MEM_TYPE_DDR3:
 	case NV_MEM_TYPE_GDDR5:
@@ -1264,7 +1498,26 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		}
 	}
 
-	// parse bios data for both pll's
+	/* parse bios data for all rammap table entries up-front, and
+	 * build information on whether certain fields differ between
+	 * any of the entries.
+	 *
+	 * the binary driver appears to completely ignore some fields
+	 * when all entries contain the same value.  at first, it was
+	 * hoped that these were mere optimisations and the bios init
+	 * tables had configured as per the values here, but there is
+	 * evidence now to suggest that this isn't the case and we do
+	 * need to treat this condition as a "don't touch" indicator.
+	 */
+	for (i = 0; !ret; i++) {
+		ret = nve0_ram_ctor_data(ram, ramcfg, i);
+		if (ret && ret != -ENOENT) {
+			nv_error(pfb, "failed to parse ramcfg data\n");
+			return ret;
+		}
+	}
+
+	/* parse bios data for both pll's */
 	ret = nvbios_pll_parse(bios, 0x0c, &ram->fuc.refpll);
 	if (ret) {
 		nv_error(pfb, "mclk refpll data not found\n");
@@ -1277,6 +1530,7 @@ nve0_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 	}
 
+	/* lookup memory voltage gpios */
 	ret = gpio->find(gpio, 0, 0x18, DCB_GPIO_UNUSED, &func);
 	if (ret == 0) {
 		ram->fuc.r_gpioMV = ramfuc_reg(0x00d610 + (func.line * 0x04));
@@ -1385,7 +1639,7 @@ nve0_ram_oclass = {
 	.handle = 0,
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nve0_ram_ctor,
-		.dtor = _nouveau_ram_dtor,
+		.dtor = nve0_ram_dtor,
 		.init = nve0_ram_init,
 		.fini = _nouveau_ram_fini,
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c
new file mode 100644
index 000000000000..bb1eb8f3e639
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr2.c
@@ -0,0 +1,94 @@
+/*
+ * Copyright 2014 Roy Spliet
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Roy Spliet <rspliet@eclipso.eu>
+ *          Ben Skeggs
+ */
+
+#include "priv.h"
+
+struct ramxlat {
+	int id;
+	u8 enc;
+};
+
+static inline int
+ramxlat(const struct ramxlat *xlat, int id)
+{
+	while (xlat->id >= 0) {
+		if (xlat->id == id)
+			return xlat->enc;
+		xlat++;
+	}
+	return -EINVAL;
+}
+
+static const struct ramxlat
+ramddr2_cl[] = {
+	{ 2, 2 }, { 3, 3 }, { 4, 4 }, { 5, 5 }, { 6, 6 },
+	/* The following are available in some, but not all DDR2 docs */
+	{ 7, 7 },
+	{ -1 }
+};
+
+static const struct ramxlat
+ramddr2_wr[] = {
+	{ 2, 1 }, { 3, 2 }, { 4, 3 }, { 5, 4 }, { 6, 5 },
+	/* The following are available in some, but not all DDR2 docs */
+	{ 7, 6 },
+	{ -1 }
+};
+
+int
+nouveau_sddr2_calc(struct nouveau_ram *ram)
+{
+	int CL, WR, DLL = 0, ODT = 0;
+
+	switch (ram->next->bios.timing_ver) {
+	case 0x10:
+		CL  = ram->next->bios.timing_10_CL;
+		WR  = ram->next->bios.timing_10_WR;
+		DLL = !ram->next->bios.ramcfg_10_02_40;
+		ODT = ram->next->bios.timing_10_ODT & 3;
+		break;
+	case 0x20:
+		CL  = (ram->next->bios.timing[1] & 0x0000001f);
+		WR  = (ram->next->bios.timing[2] & 0x007f0000) >> 16;
+		break;
+	default:
+		return -ENOSYS;
+	}
+
+	CL  = ramxlat(ramddr2_cl, CL);
+	WR  = ramxlat(ramddr2_wr, WR);
+	if (CL < 0 || WR < 0)
+		return -EINVAL;
+
+	ram->mr[0] &= ~0xf70;
+	ram->mr[0] |= (WR & 0x07) << 9;
+	ram->mr[0] |= (CL & 0x07) << 4;
+
+	ram->mr[1] &= ~0x045;
+	ram->mr[1] |= (ODT & 0x1) << 2;
+	ram->mr[1] |= (ODT & 0x2) << 5;
+	ram->mr[1] |= !DLL;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
index ebd4cd9c35d9..83949b11833a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/sddr3.c
@@ -20,9 +20,9 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  *
  * Authors: Ben Skeggs <bskeggs@redhat.com>
+ * 	    Roy Spliet <rspliet@eclipso.eu>
  */
 
-#include <subdev/bios.h>
 #include "priv.h"
 
 struct ramxlat {
@@ -69,31 +69,52 @@ ramddr3_cwl[] = {
 int
 nouveau_sddr3_calc(struct nouveau_ram *ram)
 {
-	struct nouveau_bios *bios = nouveau_bios(ram);
-	int WL, CL, WR;
+	int CWL, CL, WR, DLL = 0, ODT = 0;
 
-	switch (!!ram->timing.data * ram->timing.version) {
+	switch (ram->next->bios.timing_ver) {
+	case 0x10:
+		if (ram->next->bios.timing_hdr < 0x17) {
+			/* XXX: NV50: Get CWL from the timing register */
+			return -ENOSYS;
+		}
+		CWL = ram->next->bios.timing_10_CWL;
+		CL  = ram->next->bios.timing_10_CL;
+		WR  = ram->next->bios.timing_10_WR;
+		DLL = !ram->next->bios.ramcfg_10_02_40;
+		ODT = ram->next->bios.timing_10_ODT;
+		break;
 	case 0x20:
-		WL = (nv_ro16(bios, ram->timing.data + 0x04) & 0x0f80) >> 7;
-		CL =  nv_ro08(bios, ram->timing.data + 0x04) & 0x1f;
-		WR =  nv_ro08(bios, ram->timing.data + 0x0a) & 0x7f;
+		CWL = (ram->next->bios.timing[1] & 0x00000f80) >> 7;
+		CL  = (ram->next->bios.timing[1] & 0x0000001f) >> 0;
+		WR  = (ram->next->bios.timing[2] & 0x007f0000) >> 16;
+		/* XXX: Get these values from the VBIOS instead */
+		DLL = !(ram->mr[1] & 0x1);
+		ODT =   (ram->mr[1] & 0x004) >> 2 |
+			(ram->mr[1] & 0x040) >> 5 |
+		        (ram->mr[1] & 0x200) >> 7;
 		break;
 	default:
 		return -ENOSYS;
 	}
 
-	WL = ramxlat(ramddr3_cwl, WL);
-	CL = ramxlat(ramddr3_cl, CL);
-	WR = ramxlat(ramddr3_wr, WR);
-	if (WL < 0 || CL < 0 || WR < 0)
+	CWL = ramxlat(ramddr3_cwl, CWL);
+	CL  = ramxlat(ramddr3_cl, CL);
+	WR  = ramxlat(ramddr3_wr, WR);
+	if (CL < 0 || CWL < 0 || WR < 0)
 		return -EINVAL;
 
-	ram->mr[0] &= ~0xe74;
+	ram->mr[0] &= ~0xf74;
 	ram->mr[0] |= (WR & 0x07) << 9;
 	ram->mr[0] |= (CL & 0x0e) << 3;
 	ram->mr[0] |= (CL & 0x01) << 2;
 
+	ram->mr[1] &= ~0x245;
+	ram->mr[1] |= (ODT & 0x1) << 2;
+	ram->mr[1] |= (ODT & 0x2) << 5;
+	ram->mr[1] |= (ODT & 0x4) << 7;
+	ram->mr[1] |= !DLL;
+
 	ram->mr[2] &= ~0x038;
-	ram->mr[2] |= (WL & 0x07) << 3;
+	ram->mr[2] |= (CWL & 0x07) << 3;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c
new file mode 100644
index 000000000000..9e8e92127715
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/base.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include <subdev/fuse.h>
+
+int
+_nouveau_fuse_init(struct nouveau_object *object)
+{
+	struct nouveau_fuse *fuse = (void *)object;
+	return nouveau_subdev_init(&fuse->base);
+}
+
+void
+_nouveau_fuse_dtor(struct nouveau_object *object)
+{
+	struct nouveau_fuse *fuse = (void *)object;
+	nouveau_subdev_destroy(&fuse->base);
+}
+
+int
+nouveau_fuse_create_(struct nouveau_object *parent,
+		     struct nouveau_object *engine,
+		     struct nouveau_oclass *oclass, int length, void **pobject)
+{
+	struct nouveau_fuse *fuse;
+	int ret;
+
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "FUSE",
+				     "fuse", length, pobject);
+	fuse = *pobject;
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c
new file mode 100644
index 000000000000..a374ade485be
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/g80.c
@@ -0,0 +1,81 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct g80_fuse_priv {
+	struct nouveau_fuse base;
+
+	spinlock_t fuse_enable_lock;
+};
+
+static u32
+g80_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct g80_fuse_priv *priv = (void *)object;
+	unsigned long flags;
+	u32 fuse_enable, val;
+
+	spin_lock_irqsave(&priv->fuse_enable_lock, flags);
+
+	/* racy if another part of nouveau start writing to this reg */
+	fuse_enable = nv_mask(priv, 0x1084, 0x800, 0x800);
+	val = nv_rd32(priv, 0x21000 + addr);
+	nv_wr32(priv, 0x1084, fuse_enable);
+
+	spin_unlock_irqrestore(&priv->fuse_enable_lock, flags);
+
+	return val;
+}
+
+
+static int
+g80_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct g80_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	spin_lock_init(&priv->fuse_enable_lock);
+
+	return 0;
+}
+
+struct nouveau_oclass
+g80_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0x50),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = g80_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = g80_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c
new file mode 100644
index 000000000000..5ed03f54b3d4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/gf100.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gf100_fuse_priv {
+	struct nouveau_fuse base;
+
+	spinlock_t fuse_enable_lock;
+};
+
+static u32
+gf100_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct gf100_fuse_priv *priv = (void *)object;
+	unsigned long flags;
+	u32 fuse_enable, unk, val;
+
+	spin_lock_irqsave(&priv->fuse_enable_lock, flags);
+
+	/* racy if another part of nouveau start writing to these regs */
+	fuse_enable = nv_mask(priv, 0x22400, 0x800, 0x800);
+	unk = nv_mask(priv, 0x21000, 0x1, 0x1);
+	val = nv_rd32(priv, 0x21100 + addr);
+	nv_wr32(priv, 0x21000, unk);
+	nv_wr32(priv, 0x22400, fuse_enable);
+
+	spin_unlock_irqrestore(&priv->fuse_enable_lock, flags);
+
+	return val;
+}
+
+
+static int
+gf100_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct gf100_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	spin_lock_init(&priv->fuse_enable_lock);
+
+	return 0;
+}
+
+struct nouveau_oclass
+gf100_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0xC0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gf100_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = gf100_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c
new file mode 100644
index 000000000000..4f1a636c6538
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/gm107.c
@@ -0,0 +1,66 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gm107_fuse_priv {
+	struct nouveau_fuse base;
+};
+
+static u32
+gm107_fuse_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct gf100_fuse_priv *priv = (void *)object;
+
+	return nv_rd32(priv, 0x21100 + addr);
+}
+
+
+static int
+gm107_fuse_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	       struct nouveau_oclass *oclass, void *data, u32 size,
+	       struct nouveau_object **pobject)
+{
+	struct gm107_fuse_priv *priv;
+	int ret;
+
+	ret = nouveau_fuse_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+struct nouveau_oclass
+gm107_fuse_oclass = {
+	.handle = NV_SUBDEV(FUSE, 0x117),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm107_fuse_ctor,
+		.dtor = _nouveau_fuse_dtor,
+		.init = _nouveau_fuse_init,
+		.fini = _nouveau_fuse_fini,
+		.rd32 = gm107_fuse_rd32,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h
new file mode 100644
index 000000000000..d2085411a5cb
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fuse/priv.h
@@ -0,0 +1,9 @@
+#ifndef __NVKM_FUSE_PRIV_H__
+#define __NVKM_FUSE_PRIV_H__
+
+#include <subdev/fuse.h>
+
+int _nouveau_fuse_init(struct nouveau_object *object);
+void _nouveau_fuse_dtor(struct nouveau_object *object);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
index 45e0202f3151..7ad99b763f4c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/base.c
@@ -106,39 +106,60 @@ nouveau_gpio_get(struct nouveau_gpio *gpio, int idx, u8 tag, u8 line)
 }
 
 static void
-nouveau_gpio_intr_disable(struct nouveau_event *event, int type, int index)
+nouveau_gpio_intr_fini(struct nvkm_event *event, int type, int index)
 {
-	struct nouveau_gpio *gpio = nouveau_gpio(event->priv);
+	struct nouveau_gpio *gpio = container_of(event, typeof(*gpio), event);
 	const struct nouveau_gpio_impl *impl = (void *)nv_object(gpio)->oclass;
 	impl->intr_mask(gpio, type, 1 << index, 0);
 }
 
 static void
-nouveau_gpio_intr_enable(struct nouveau_event *event, int type, int index)
+nouveau_gpio_intr_init(struct nvkm_event *event, int type, int index)
 {
-	struct nouveau_gpio *gpio = nouveau_gpio(event->priv);
+	struct nouveau_gpio *gpio = container_of(event, typeof(*gpio), event);
 	const struct nouveau_gpio_impl *impl = (void *)nv_object(gpio)->oclass;
 	impl->intr_mask(gpio, type, 1 << index, 1 << index);
 }
 
+static int
+nouveau_gpio_intr_ctor(struct nouveau_object *object, void *data, u32 size,
+		       struct nvkm_notify *notify)
+{
+	struct nvkm_gpio_ntfy_req *req = data;
+	if (!WARN_ON(size != sizeof(*req))) {
+		notify->size  = sizeof(struct nvkm_gpio_ntfy_rep);
+		notify->types = req->mask;
+		notify->index = req->line;
+		return 0;
+	}
+	return -EINVAL;
+}
+
 static void
 nouveau_gpio_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_gpio *gpio = nouveau_gpio(subdev);
 	const struct nouveau_gpio_impl *impl = (void *)nv_object(gpio)->oclass;
-	u32 hi, lo, e, i;
+	u32 hi, lo, i;
 
 	impl->intr_stat(gpio, &hi, &lo);
 
-	for (i = 0; e = 0, (hi | lo) && i < impl->lines; i++) {
-		if (hi & (1 << i))
-			e |= NVKM_GPIO_HI;
-		if (lo & (1 << i))
-			e |= NVKM_GPIO_LO;
-		nouveau_event_trigger(gpio->events, e, i);
+	for (i = 0; (hi | lo) && i < impl->lines; i++) {
+		struct nvkm_gpio_ntfy_rep rep = {
+			.mask = (NVKM_GPIO_HI * !!(hi & (1 << i))) |
+				(NVKM_GPIO_LO * !!(lo & (1 << i))),
+		};
+		nvkm_event_send(&gpio->event, rep.mask, i, &rep, sizeof(rep));
 	}
 }
 
+static const struct nvkm_event_func
+nouveau_gpio_intr_func = {
+	.ctor = nouveau_gpio_intr_ctor,
+	.init = nouveau_gpio_intr_init,
+	.fini = nouveau_gpio_intr_fini,
+};
+
 int
 _nouveau_gpio_fini(struct nouveau_object *object, bool suspend)
 {
@@ -183,7 +204,7 @@ void
 _nouveau_gpio_dtor(struct nouveau_object *object)
 {
 	struct nouveau_gpio *gpio = (void *)object;
-	nouveau_event_destroy(&gpio->events);
+	nvkm_event_fini(&gpio->event);
 	nouveau_subdev_destroy(&gpio->base);
 }
 
@@ -208,13 +229,11 @@ nouveau_gpio_create_(struct nouveau_object *parent,
 	gpio->get  = nouveau_gpio_get;
 	gpio->reset = impl->reset;
 
-	ret = nouveau_event_create(2, impl->lines, &gpio->events);
+	ret = nvkm_event_init(&nouveau_gpio_intr_func, 2, impl->lines,
+			      &gpio->event);
 	if (ret)
 		return ret;
 
-	gpio->events->priv = gpio;
-	gpio->events->enable = nouveau_gpio_intr_enable;
-	gpio->events->disable = nouveau_gpio_intr_disable;
 	nv_subdev(gpio)->intr = nouveau_gpio_intr;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv92.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv94.c
index 252083d376f5..cae404ccadac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nv92.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nv94.c
@@ -25,7 +25,7 @@
 #include "priv.h"
 
 void
-nv92_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
+nv94_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
 {
 	u32 intr0 = nv_rd32(gpio, 0x00e054);
 	u32 intr1 = nv_rd32(gpio, 0x00e074);
@@ -38,7 +38,7 @@ nv92_gpio_intr_stat(struct nouveau_gpio *gpio, u32 *hi, u32 *lo)
 }
 
 void
-nv92_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
+nv94_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
 {
 	u32 inte0 = nv_rd32(gpio, 0x00e050);
 	u32 inte1 = nv_rd32(gpio, 0x00e070);
@@ -57,8 +57,8 @@ nv92_gpio_intr_mask(struct nouveau_gpio *gpio, u32 type, u32 mask, u32 data)
 }
 
 struct nouveau_oclass *
-nv92_gpio_oclass = &(struct nouveau_gpio_impl) {
-	.base.handle = NV_SUBDEV(GPIO, 0x92),
+nv94_gpio_oclass = &(struct nouveau_gpio_impl) {
+	.base.handle = NV_SUBDEV(GPIO, 0x94),
 	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = _nouveau_gpio_ctor,
 		.dtor = _nouveau_gpio_dtor,
@@ -66,8 +66,8 @@ nv92_gpio_oclass = &(struct nouveau_gpio_impl) {
 		.fini = _nouveau_gpio_fini,
 	},
 	.lines = 32,
-	.intr_stat = nv92_gpio_intr_stat,
-	.intr_mask = nv92_gpio_intr_mask,
+	.intr_stat = nv94_gpio_intr_stat,
+	.intr_mask = nv94_gpio_intr_mask,
 	.drive = nv50_gpio_drive,
 	.sense = nv50_gpio_sense,
 	.reset = nv50_gpio_reset,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
index a4682b0956ad..480d6d2af770 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/nvd0.c
@@ -77,8 +77,8 @@ nvd0_gpio_oclass = &(struct nouveau_gpio_impl) {
 		.fini = _nouveau_gpio_fini,
 	},
 	.lines = 32,
-	.intr_stat = nv92_gpio_intr_stat,
-	.intr_mask = nv92_gpio_intr_mask,
+	.intr_stat = nv94_gpio_intr_stat,
+	.intr_mask = nv94_gpio_intr_mask,
 	.drive = nvd0_gpio_drive,
 	.sense = nvd0_gpio_sense,
 	.reset = nvd0_gpio_reset,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
index e1724dfc86ae..bff98b86e2b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/gpio/priv.h
@@ -56,8 +56,8 @@ void nv50_gpio_reset(struct nouveau_gpio *, u8);
 int  nv50_gpio_drive(struct nouveau_gpio *, int, int, int);
 int  nv50_gpio_sense(struct nouveau_gpio *, int);
 
-void nv92_gpio_intr_stat(struct nouveau_gpio *, u32 *, u32 *);
-void nv92_gpio_intr_mask(struct nouveau_gpio *, u32, u32, u32);
+void nv94_gpio_intr_stat(struct nouveau_gpio *, u32 *, u32 *);
+void nv94_gpio_intr_mask(struct nouveau_gpio *, u32, u32, u32);
 
 void nvd0_gpio_reset(struct nouveau_gpio *, u8);
 int  nvd0_gpio_drive(struct nouveau_gpio *, int, int, int);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
index 09ba2cc851cf..2b1bf545e488 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/i2c/base.c
@@ -23,6 +23,7 @@
  */
 
 #include <core/option.h>
+#include <core/object.h>
 #include <core/event.h>
 
 #include <subdev/bios.h>
@@ -326,9 +327,9 @@ nouveau_i2c_identify(struct nouveau_i2c *i2c, int index, const char *what,
 }
 
 static void
-nouveau_i2c_intr_disable(struct nouveau_event *event, int type, int index)
+nouveau_i2c_intr_fini(struct nvkm_event *event, int type, int index)
 {
-	struct nouveau_i2c *i2c = nouveau_i2c(event->priv);
+	struct nouveau_i2c *i2c = container_of(event, typeof(*i2c), event);
 	struct nouveau_i2c_port *port = i2c->find(i2c, index);
 	const struct nouveau_i2c_impl *impl = (void *)nv_object(i2c)->oclass;
 	if (port && port->aux >= 0)
@@ -336,15 +337,29 @@ nouveau_i2c_intr_disable(struct nouveau_event *event, int type, int index)
 }
 
 static void
-nouveau_i2c_intr_enable(struct nouveau_event *event, int type, int index)
+nouveau_i2c_intr_init(struct nvkm_event *event, int type, int index)
 {
-	struct nouveau_i2c *i2c = nouveau_i2c(event->priv);
+	struct nouveau_i2c *i2c = container_of(event, typeof(*i2c), event);
 	struct nouveau_i2c_port *port = i2c->find(i2c, index);
 	const struct nouveau_i2c_impl *impl = (void *)nv_object(i2c)->oclass;
 	if (port && port->aux >= 0)
 		impl->aux_mask(i2c, type, 1 << port->aux, 1 << port->aux);
 }
 
+static int
+nouveau_i2c_intr_ctor(struct nouveau_object *object, void *data, u32 size,
+		      struct nvkm_notify *notify)
+{
+	struct nvkm_i2c_ntfy_req *req = data;
+	if (!WARN_ON(size != sizeof(*req))) {
+		notify->size  = sizeof(struct nvkm_i2c_ntfy_rep);
+		notify->types = req->mask;
+		notify->index = req->port;
+		return 0;
+	}
+	return -EINVAL;
+}
+
 static void
 nouveau_i2c_intr(struct nouveau_subdev *subdev)
 {
@@ -364,13 +379,26 @@ nouveau_i2c_intr(struct nouveau_subdev *subdev)
 				if (lo & (1 << port->aux)) e |= NVKM_I2C_UNPLUG;
 				if (rq & (1 << port->aux)) e |= NVKM_I2C_IRQ;
 				if (tx & (1 << port->aux)) e |= NVKM_I2C_DONE;
-
-				nouveau_event_trigger(i2c->ntfy, e, port->index);
+				if (e) {
+					struct nvkm_i2c_ntfy_rep rep = {
+						.mask = e,
+					};
+					nvkm_event_send(&i2c->event, rep.mask,
+							port->index, &rep,
+							sizeof(rep));
+				}
 			}
 		}
 	}
 }
 
+static const struct nvkm_event_func
+nouveau_i2c_intr_func = {
+	.ctor = nouveau_i2c_intr_ctor,
+	.init = nouveau_i2c_intr_init,
+	.fini = nouveau_i2c_intr_fini,
+};
+
 int
 _nouveau_i2c_fini(struct nouveau_object *object, bool suspend)
 {
@@ -431,7 +459,7 @@ _nouveau_i2c_dtor(struct nouveau_object *object)
 	struct nouveau_i2c *i2c = (void *)object;
 	struct nouveau_i2c_port *port, *temp;
 
-	nouveau_event_destroy(&i2c->ntfy);
+	nvkm_event_fini(&i2c->event);
 
 	list_for_each_entry_safe(port, temp, &i2c->ports, head) {
 		nouveau_object_ref(NULL, (struct nouveau_object **)&port);
@@ -547,13 +575,10 @@ nouveau_i2c_create_(struct nouveau_object *parent,
 		}
 	}
 
-	ret = nouveau_event_create(4, index, &i2c->ntfy);
+	ret = nvkm_event_init(&nouveau_i2c_intr_func, 4, index, &i2c->event);
 	if (ret)
 		return ret;
 
-	i2c->ntfy->priv = i2c;
-	i2c->ntfy->enable = nouveau_i2c_intr_enable;
-	i2c->ntfy->disable = nouveau_i2c_intr_disable;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
index 7b64befee48f..e8b1401c59c0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv04.c
@@ -69,7 +69,7 @@ nv04_instobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = nouveau_mm_head(&priv->heap, 1, args->size, args->size,
+	ret = nouveau_mm_head(&priv->heap, 0, 1, args->size, args->size,
 			      args->align, &node->mem);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c
new file mode 100644
index 000000000000..7fa331516f84
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/base.c
@@ -0,0 +1,126 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "priv.h"
+
+static int
+nvkm_ltc_tags_alloc(struct nouveau_ltc *ltc, u32 n,
+		    struct nouveau_mm_node **pnode)
+{
+	struct nvkm_ltc_priv *priv = (void *)ltc;
+	int ret;
+
+	ret = nouveau_mm_head(&priv->tags, 0, 1, n, n, 1, pnode);
+	if (ret)
+		*pnode = NULL;
+
+	return ret;
+}
+
+static void
+nvkm_ltc_tags_free(struct nouveau_ltc *ltc, struct nouveau_mm_node **pnode)
+{
+	struct nvkm_ltc_priv *priv = (void *)ltc;
+	nouveau_mm_free(&priv->tags, pnode);
+}
+
+static void
+nvkm_ltc_tags_clear(struct nouveau_ltc *ltc, u32 first, u32 count)
+{
+	const struct nvkm_ltc_impl *impl = (void *)nv_oclass(ltc);
+	struct nvkm_ltc_priv *priv = (void *)ltc;
+	const u32 limit = first + count - 1;
+
+	BUG_ON((first > limit) || (limit >= priv->num_tags));
+
+	impl->cbc_clear(priv, first, limit);
+	impl->cbc_wait(priv);
+}
+
+static int
+nvkm_ltc_zbc_color_get(struct nouveau_ltc *ltc, int index, const u32 color[4])
+{
+	const struct nvkm_ltc_impl *impl = (void *)nv_oclass(ltc);
+	struct nvkm_ltc_priv *priv = (void *)ltc;
+	memcpy(priv->zbc_color[index], color, sizeof(priv->zbc_color[index]));
+	impl->zbc_clear_color(priv, index, color);
+	return index;
+}
+
+static int
+nvkm_ltc_zbc_depth_get(struct nouveau_ltc *ltc, int index, const u32 depth)
+{
+	const struct nvkm_ltc_impl *impl = (void *)nv_oclass(ltc);
+	struct nvkm_ltc_priv *priv = (void *)ltc;
+	priv->zbc_depth[index] = depth;
+	impl->zbc_clear_depth(priv, index, depth);
+	return index;
+}
+
+int
+_nvkm_ltc_init(struct nouveau_object *object)
+{
+	const struct nvkm_ltc_impl *impl = (void *)nv_oclass(object);
+	struct nvkm_ltc_priv *priv = (void *)object;
+	int ret, i;
+
+	ret = nouveau_subdev_init(&priv->base.base);
+	if (ret)
+		return ret;
+
+	for (i = priv->base.zbc_min; i <= priv->base.zbc_max; i++) {
+		impl->zbc_clear_color(priv, i, priv->zbc_color[i]);
+		impl->zbc_clear_depth(priv, i, priv->zbc_depth[i]);
+	}
+
+	return 0;
+}
+
+int
+nvkm_ltc_create_(struct nouveau_object *parent, struct nouveau_object *engine,
+		 struct nouveau_oclass *oclass, int length, void **pobject)
+{
+	const struct nvkm_ltc_impl *impl = (void *)oclass;
+	struct nvkm_ltc_priv *priv;
+	int ret;
+
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "PLTCG",
+				     "l2c", length, pobject);
+	priv = *pobject;
+	if (ret)
+		return ret;
+
+	memset(priv->zbc_color, 0x00, sizeof(priv->zbc_color));
+	memset(priv->zbc_depth, 0x00, sizeof(priv->zbc_depth));
+
+	priv->base.base.intr = impl->intr;
+	priv->base.tags_alloc = nvkm_ltc_tags_alloc;
+	priv->base.tags_free = nvkm_ltc_tags_free;
+	priv->base.tags_clear = nvkm_ltc_tags_clear;
+	priv->base.zbc_min = 1; /* reserve 0 for disabled */
+	priv->base.zbc_max = min(impl->zbc, NOUVEAU_LTC_MAX_ZBC_CNT) - 1;
+	priv->base.zbc_color_get = nvkm_ltc_zbc_color_get;
+	priv->base.zbc_depth_get = nvkm_ltc_zbc_depth_get;
+	return 0;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gf100.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c
index f2f3338a967a..2db0977284f8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gf100.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/gf100.c
@@ -25,86 +25,126 @@
 #include <subdev/fb.h>
 #include <subdev/timer.h>
 
-#include "gf100.h"
+#include "priv.h"
+
+void
+gf100_ltc_cbc_clear(struct nvkm_ltc_priv *priv, u32 start, u32 limit)
+{
+	nv_wr32(priv, 0x17e8cc, start);
+	nv_wr32(priv, 0x17e8d0, limit);
+	nv_wr32(priv, 0x17e8c8, 0x00000004);
+}
+
+void
+gf100_ltc_cbc_wait(struct nvkm_ltc_priv *priv)
+{
+	int c, s;
+	for (c = 0; c < priv->ltc_nr; c++) {
+		for (s = 0; s < priv->lts_nr; s++)
+			nv_wait(priv, 0x1410c8 + c * 0x2000 + s * 0x400, ~0, 0);
+	}
+}
+
+void
+gf100_ltc_zbc_clear_color(struct nvkm_ltc_priv *priv, int i, const u32 color[4])
+{
+	nv_mask(priv, 0x17ea44, 0x0000000f, i);
+	nv_wr32(priv, 0x17ea48, color[0]);
+	nv_wr32(priv, 0x17ea4c, color[1]);
+	nv_wr32(priv, 0x17ea50, color[2]);
+	nv_wr32(priv, 0x17ea54, color[3]);
+}
+
+void
+gf100_ltc_zbc_clear_depth(struct nvkm_ltc_priv *priv, int i, const u32 depth)
+{
+	nv_mask(priv, 0x17ea44, 0x0000000f, i);
+	nv_wr32(priv, 0x17ea58, depth);
+}
+
+static const struct nouveau_bitfield
+gf100_ltc_lts_intr_name[] = {
+	{ 0x00000001, "IDLE_ERROR_IQ" },
+	{ 0x00000002, "IDLE_ERROR_CBC" },
+	{ 0x00000004, "IDLE_ERROR_TSTG" },
+	{ 0x00000008, "IDLE_ERROR_DSTG" },
+	{ 0x00000010, "EVICTED_CB" },
+	{ 0x00000020, "ILLEGAL_COMPSTAT" },
+	{ 0x00000040, "BLOCKLINEAR_CB" },
+	{ 0x00000100, "ECC_SEC_ERROR" },
+	{ 0x00000200, "ECC_DED_ERROR" },
+	{ 0x00000400, "DEBUG" },
+	{ 0x00000800, "ATOMIC_TO_Z" },
+	{ 0x00001000, "ILLEGAL_ATOMIC" },
+	{ 0x00002000, "BLKACTIVITY_ERR" },
+	{}
+};
 
 static void
-gf100_ltcg_lts_isr(struct gf100_ltcg_priv *priv, int ltc, int lts)
+gf100_ltc_lts_intr(struct nvkm_ltc_priv *priv, int ltc, int lts)
 {
 	u32 base = 0x141000 + (ltc * 0x2000) + (lts * 0x400);
-	u32 stat = nv_rd32(priv, base + 0x020);
+	u32 intr = nv_rd32(priv, base + 0x020);
+	u32 stat = intr & 0x0000ffff;
 
 	if (stat) {
-		nv_info(priv, "LTC%d_LTS%d: 0x%08x\n", ltc, lts, stat);
-		nv_wr32(priv, base + 0x020, stat);
+		nv_info(priv, "LTC%d_LTS%d:", ltc, lts);
+		nouveau_bitfield_print(gf100_ltc_lts_intr_name, stat);
+		pr_cont("\n");
 	}
+
+	nv_wr32(priv, base + 0x020, intr);
 }
 
-static void
-gf100_ltcg_intr(struct nouveau_subdev *subdev)
+void
+gf100_ltc_intr(struct nouveau_subdev *subdev)
 {
-	struct gf100_ltcg_priv *priv = (void *)subdev;
+	struct nvkm_ltc_priv *priv = (void *)subdev;
 	u32 mask;
 
 	mask = nv_rd32(priv, 0x00017c);
 	while (mask) {
 		u32 lts, ltc = __ffs(mask);
 		for (lts = 0; lts < priv->lts_nr; lts++)
-			gf100_ltcg_lts_isr(priv, ltc, lts);
+			gf100_ltc_lts_intr(priv, ltc, lts);
 		mask &= ~(1 << ltc);
 	}
-
-	/* we do something horribly wrong and upset PMFB a lot, so mask off
-	 * interrupts from it after the first one until it's fixed
-	 */
-	nv_mask(priv, 0x000640, 0x02000000, 0x00000000);
 }
 
-int
-gf100_ltcg_tags_alloc(struct nouveau_ltcg *ltcg, u32 n,
-		     struct nouveau_mm_node **pnode)
+static int
+gf100_ltc_init(struct nouveau_object *object)
 {
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
+	struct nvkm_ltc_priv *priv = (void *)object;
+	u32 lpg128 = !(nv_rd32(priv, 0x100c80) & 0x00000001);
 	int ret;
 
-	ret = nouveau_mm_head(&priv->tags, 1, n, n, 1, pnode);
+	ret = nvkm_ltc_init(priv);
 	if (ret)
-		*pnode = NULL;
+		return ret;
 
-	return ret;
+	nv_mask(priv, 0x17e820, 0x00100000, 0x00000000); /* INTR_EN &= ~0x10 */
+	nv_wr32(priv, 0x17e8d8, priv->ltc_nr);
+	nv_wr32(priv, 0x17e8d4, priv->tag_base);
+	nv_mask(priv, 0x17e8c0, 0x00000002, lpg128 ? 0x00000002 : 0x00000000);
+	return 0;
 }
 
 void
-gf100_ltcg_tags_free(struct nouveau_ltcg *ltcg, struct nouveau_mm_node **pnode)
-{
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-
-	nouveau_mm_free(&priv->tags, pnode);
-}
-
-static void
-gf100_ltcg_tags_clear(struct nouveau_ltcg *ltcg, u32 first, u32 count)
+gf100_ltc_dtor(struct nouveau_object *object)
 {
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-	u32 last = first + count - 1;
-	int p, i;
+	struct nouveau_fb *pfb = nouveau_fb(object);
+	struct nvkm_ltc_priv *priv = (void *)object;
 
-	BUG_ON((first > last) || (last >= priv->num_tags));
-
-	nv_wr32(priv, 0x17e8cc, first);
-	nv_wr32(priv, 0x17e8d0, last);
-	nv_wr32(priv, 0x17e8c8, 0x4); /* trigger clear */
+	nouveau_mm_fini(&priv->tags);
+	nouveau_mm_free(&pfb->vram, &priv->tag_ram);
 
-	/* wait until it's finished with clearing */
-	for (p = 0; p < priv->ltc_nr; ++p) {
-		for (i = 0; i < priv->lts_nr; ++i)
-			nv_wait(priv, 0x1410c8 + p * 0x2000 + i * 0x400, ~0, 0);
-	}
+	nvkm_ltc_destroy(priv);
 }
 
 /* TODO: Figure out tag memory details and drop the over-cautious allocation.
  */
 int
-gf100_ltcg_init_tag_ram(struct nouveau_fb *pfb, struct gf100_ltcg_priv *priv)
+gf100_ltc_init_tag_ram(struct nouveau_fb *pfb, struct nvkm_ltc_priv *priv)
 {
 	u32 tag_size, tag_margin, tag_align;
 	int ret;
@@ -130,34 +170,34 @@ gf100_ltcg_init_tag_ram(struct nouveau_fb *pfb, struct gf100_ltcg_priv *priv)
 	tag_size += tag_align;
 	tag_size  = (tag_size + 0xfff) >> 12; /* round up */
 
-	ret = nouveau_mm_tail(&pfb->vram, 1, tag_size, tag_size, 1,
+	ret = nouveau_mm_tail(&pfb->vram, 1, 1, tag_size, tag_size, 1,
 	                      &priv->tag_ram);
 	if (ret) {
 		priv->num_tags = 0;
 	} else {
-		u64 tag_base = (priv->tag_ram->offset << 12) + tag_margin;
+		u64 tag_base = ((u64)priv->tag_ram->offset << 12) + tag_margin;
 
 		tag_base += tag_align - 1;
 		ret = do_div(tag_base, tag_align);
 
 		priv->tag_base = tag_base;
 	}
-	ret = nouveau_mm_init(&priv->tags, 0, priv->num_tags, 1);
 
+	ret = nouveau_mm_init(&priv->tags, 0, priv->num_tags, 1);
 	return ret;
 }
 
-static int
-gf100_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+int
+gf100_ltc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
 {
-	struct gf100_ltcg_priv *priv;
 	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nvkm_ltc_priv *priv;
 	u32 parts, mask;
 	int ret, i;
 
-	ret = nouveau_ltcg_create(parent, engine, oclass, &priv);
+	ret = nvkm_ltc_create(parent, engine, oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -170,57 +210,27 @@ gf100_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	}
 	priv->lts_nr = nv_rd32(priv, 0x17e8dc) >> 28;
 
-	ret = gf100_ltcg_init_tag_ram(pfb, priv);
-	if (ret)
-		return ret;
-
-	priv->base.tags_alloc = gf100_ltcg_tags_alloc;
-	priv->base.tags_free  = gf100_ltcg_tags_free;
-	priv->base.tags_clear = gf100_ltcg_tags_clear;
-
-	nv_subdev(priv)->intr = gf100_ltcg_intr;
-	return 0;
-}
-
-void
-gf100_ltcg_dtor(struct nouveau_object *object)
-{
-	struct nouveau_ltcg *ltcg = (struct nouveau_ltcg *)object;
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-	struct nouveau_fb *pfb = nouveau_fb(ltcg->base.base.parent);
-
-	nouveau_mm_fini(&priv->tags);
-	nouveau_mm_free(&pfb->vram, &priv->tag_ram);
-
-	nouveau_ltcg_destroy(ltcg);
-}
-
-static int
-gf100_ltcg_init(struct nouveau_object *object)
-{
-	struct nouveau_ltcg *ltcg = (struct nouveau_ltcg *)object;
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-	int ret;
-
-	ret = nouveau_ltcg_init(ltcg);
+	ret = gf100_ltc_init_tag_ram(pfb, priv);
 	if (ret)
 		return ret;
 
-	nv_mask(priv, 0x17e820, 0x00100000, 0x00000000); /* INTR_EN &= ~0x10 */
-	nv_wr32(priv, 0x17e8d8, priv->ltc_nr);
-	if (nv_device(ltcg)->card_type >= NV_E0)
-		nv_wr32(priv, 0x17e000, priv->ltc_nr);
-	nv_wr32(priv, 0x17e8d4, priv->tag_base);
+	nv_subdev(priv)->intr = gf100_ltc_intr;
 	return 0;
 }
 
 struct nouveau_oclass *
-gf100_ltcg_oclass = &(struct nouveau_oclass) {
-	.handle = NV_SUBDEV(LTCG, 0xc0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = gf100_ltcg_ctor,
-		.dtor = gf100_ltcg_dtor,
-		.init = gf100_ltcg_init,
-		.fini = _nouveau_ltcg_fini,
+gf100_ltc_oclass = &(struct nvkm_ltc_impl) {
+	.base.handle = NV_SUBDEV(LTC, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gf100_ltc_ctor,
+		.dtor = gf100_ltc_dtor,
+		.init = gf100_ltc_init,
+		.fini = _nvkm_ltc_fini,
 	},
-};
+	.intr = gf100_ltc_intr,
+	.cbc_clear = gf100_ltc_cbc_clear,
+	.cbc_wait = gf100_ltc_cbc_wait,
+	.zbc = 16,
+	.zbc_clear_color = gf100_ltc_zbc_clear_color,
+	.zbc_clear_depth = gf100_ltc_zbc_clear_depth,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/gk104.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/gk104.c
new file mode 100644
index 000000000000..b39b5d0eb8f9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/gk104.c
@@ -0,0 +1,60 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+gk104_ltc_init(struct nouveau_object *object)
+{
+	struct nvkm_ltc_priv *priv = (void *)object;
+	u32 lpg128 = !(nv_rd32(priv, 0x100c80) & 0x00000001);
+	int ret;
+
+	ret = nvkm_ltc_init(priv);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, 0x17e8d8, priv->ltc_nr);
+	nv_wr32(priv, 0x17e000, priv->ltc_nr);
+	nv_wr32(priv, 0x17e8d4, priv->tag_base);
+	nv_mask(priv, 0x17e8c0, 0x00000002, lpg128 ? 0x00000002 : 0x00000000);
+	return 0;
+}
+
+struct nouveau_oclass *
+gk104_ltc_oclass = &(struct nvkm_ltc_impl) {
+	.base.handle = NV_SUBDEV(LTC, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gf100_ltc_ctor,
+		.dtor = gf100_ltc_dtor,
+		.init = gk104_ltc_init,
+		.fini = _nvkm_ltc_fini,
+	},
+	.intr = gf100_ltc_intr,
+	.cbc_clear = gf100_ltc_cbc_clear,
+	.cbc_wait = gf100_ltc_cbc_wait,
+	.zbc = 16,
+	.zbc_clear_color = gf100_ltc_zbc_clear_color,
+	.zbc_clear_depth = gf100_ltc_zbc_clear_depth,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c
index e79d0e81de40..89fc4238f50c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gm107.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/gm107.c
@@ -25,10 +25,45 @@
 #include <subdev/fb.h>
 #include <subdev/timer.h>
 
-#include "gf100.h"
+#include "priv.h"
 
 static void
-gm107_ltcg_lts_isr(struct gf100_ltcg_priv *priv, int ltc, int lts)
+gm107_ltc_cbc_clear(struct nvkm_ltc_priv *priv, u32 start, u32 limit)
+{
+	nv_wr32(priv, 0x17e270, start);
+	nv_wr32(priv, 0x17e274, limit);
+	nv_wr32(priv, 0x17e26c, 0x00000004);
+}
+
+static void
+gm107_ltc_cbc_wait(struct nvkm_ltc_priv *priv)
+{
+	int c, s;
+	for (c = 0; c < priv->ltc_nr; c++) {
+		for (s = 0; s < priv->lts_nr; s++)
+			nv_wait(priv, 0x14046c + c * 0x2000 + s * 0x200, ~0, 0);
+	}
+}
+
+static void
+gm107_ltc_zbc_clear_color(struct nvkm_ltc_priv *priv, int i, const u32 color[4])
+{
+	nv_mask(priv, 0x17e338, 0x0000000f, i);
+	nv_wr32(priv, 0x17e33c, color[0]);
+	nv_wr32(priv, 0x17e340, color[1]);
+	nv_wr32(priv, 0x17e344, color[2]);
+	nv_wr32(priv, 0x17e348, color[3]);
+}
+
+static void
+gm107_ltc_zbc_clear_depth(struct nvkm_ltc_priv *priv, int i, const u32 depth)
+{
+	nv_mask(priv, 0x17e338, 0x0000000f, i);
+	nv_wr32(priv, 0x17e34c, depth);
+}
+
+static void
+gm107_ltc_lts_isr(struct nvkm_ltc_priv *priv, int ltc, int lts)
 {
 	u32 base = 0x140000 + (ltc * 0x2000) + (lts * 0x400);
 	u32 stat = nv_rd32(priv, base + 0x00c);
@@ -40,56 +75,48 @@ gm107_ltcg_lts_isr(struct gf100_ltcg_priv *priv, int ltc, int lts)
 }
 
 static void
-gm107_ltcg_intr(struct nouveau_subdev *subdev)
+gm107_ltc_intr(struct nouveau_subdev *subdev)
 {
-	struct gf100_ltcg_priv *priv = (void *)subdev;
+	struct nvkm_ltc_priv *priv = (void *)subdev;
 	u32 mask;
 
 	mask = nv_rd32(priv, 0x00017c);
 	while (mask) {
 		u32 lts, ltc = __ffs(mask);
 		for (lts = 0; lts < priv->lts_nr; lts++)
-			gm107_ltcg_lts_isr(priv, ltc, lts);
+			gm107_ltc_lts_isr(priv, ltc, lts);
 		mask &= ~(1 << ltc);
 	}
-
-	/* we do something horribly wrong and upset PMFB a lot, so mask off
-	 * interrupts from it after the first one until it's fixed
-	 */
-	nv_mask(priv, 0x000640, 0x02000000, 0x00000000);
 }
 
-static void
-gm107_ltcg_tags_clear(struct nouveau_ltcg *ltcg, u32 first, u32 count)
+static int
+gm107_ltc_init(struct nouveau_object *object)
 {
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-	u32 last = first + count - 1;
-	int p, i;
-
-	BUG_ON((first > last) || (last >= priv->num_tags));
+	struct nvkm_ltc_priv *priv = (void *)object;
+	u32 lpg128 = !(nv_rd32(priv, 0x100c80) & 0x00000001);
+	int ret;
 
-	nv_wr32(priv, 0x17e270, first);
-	nv_wr32(priv, 0x17e274, last);
-	nv_wr32(priv, 0x17e26c, 0x4); /* trigger clear */
+	ret = nvkm_ltc_init(priv);
+	if (ret)
+		return ret;
 
-	/* wait until it's finished with clearing */
-	for (p = 0; p < priv->ltc_nr; ++p) {
-		for (i = 0; i < priv->lts_nr; ++i)
-			nv_wait(priv, 0x14046c + p * 0x2000 + i * 0x200, ~0, 0);
-	}
+	nv_wr32(priv, 0x17e27c, priv->ltc_nr);
+	nv_wr32(priv, 0x17e278, priv->tag_base);
+	nv_mask(priv, 0x17e264, 0x00000002, lpg128 ? 0x00000002 : 0x00000000);
+	return 0;
 }
 
 static int
-gm107_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+gm107_ltc_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
 {
-	struct gf100_ltcg_priv *priv;
 	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nvkm_ltc_priv *priv;
 	u32 parts, mask;
 	int ret, i;
 
-	ret = nouveau_ltcg_create(parent, engine, oclass, &priv);
+	ret = nvkm_ltc_create(parent, engine, oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -102,41 +129,26 @@ gm107_ltcg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	}
 	priv->lts_nr = nv_rd32(priv, 0x17e280) >> 28;
 
-	ret = gf100_ltcg_init_tag_ram(pfb, priv);
-	if (ret)
-		return ret;
-
-	priv->base.tags_alloc = gf100_ltcg_tags_alloc;
-	priv->base.tags_free  = gf100_ltcg_tags_free;
-	priv->base.tags_clear = gm107_ltcg_tags_clear;
-
-	nv_subdev(priv)->intr = gm107_ltcg_intr;
-	return 0;
-}
-
-static int
-gm107_ltcg_init(struct nouveau_object *object)
-{
-	struct nouveau_ltcg *ltcg = (struct nouveau_ltcg *)object;
-	struct gf100_ltcg_priv *priv = (struct gf100_ltcg_priv *)ltcg;
-	int ret;
-
-	ret = nouveau_ltcg_init(ltcg);
+	ret = gf100_ltc_init_tag_ram(pfb, priv);
 	if (ret)
 		return ret;
 
-	nv_wr32(priv, 0x17e27c, priv->ltc_nr);
-	nv_wr32(priv, 0x17e278, priv->tag_base);
 	return 0;
 }
 
 struct nouveau_oclass *
-gm107_ltcg_oclass = &(struct nouveau_oclass) {
-	.handle = NV_SUBDEV(LTCG, 0xff),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = gm107_ltcg_ctor,
-		.dtor = gf100_ltcg_dtor,
-		.init = gm107_ltcg_init,
-		.fini = _nouveau_ltcg_fini,
+gm107_ltc_oclass = &(struct nvkm_ltc_impl) {
+	.base.handle = NV_SUBDEV(LTC, 0xff),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm107_ltc_ctor,
+		.dtor = gf100_ltc_dtor,
+		.init = gm107_ltc_init,
+		.fini = _nvkm_ltc_fini,
 	},
-};
+	.intr = gm107_ltc_intr,
+	.cbc_clear = gm107_ltc_cbc_clear,
+	.cbc_wait = gm107_ltc_cbc_wait,
+	.zbc = 16,
+	.zbc_clear_color = gm107_ltc_zbc_clear_color,
+	.zbc_clear_depth = gm107_ltc_zbc_clear_depth,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h b/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h
new file mode 100644
index 000000000000..41f179d93da6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltc/priv.h
@@ -0,0 +1,71 @@
+#ifndef __NVKM_LTC_PRIV_H__
+#define __NVKM_LTC_PRIV_H__
+
+#include <subdev/ltc.h>
+#include <subdev/fb.h>
+
+#include <core/enum.h>
+
+struct nvkm_ltc_priv {
+	struct nouveau_ltc base;
+	u32 ltc_nr;
+	u32 lts_nr;
+
+	u32 num_tags;
+	u32 tag_base;
+	struct nouveau_mm tags;
+	struct nouveau_mm_node *tag_ram;
+
+	u32 zbc_color[NOUVEAU_LTC_MAX_ZBC_CNT][4];
+	u32 zbc_depth[NOUVEAU_LTC_MAX_ZBC_CNT];
+};
+
+#define nvkm_ltc_create(p,e,o,d)                                               \
+	nvkm_ltc_create_((p), (e), (o), sizeof(**d), (void **)d)
+#define nvkm_ltc_destroy(p) ({                                                 \
+	struct nvkm_ltc_priv *_priv = (p);                                     \
+	_nvkm_ltc_dtor(nv_object(_priv));                                      \
+})
+#define nvkm_ltc_init(p) ({                                                    \
+	struct nvkm_ltc_priv *_priv = (p);                                     \
+	_nvkm_ltc_init(nv_object(_priv));                                      \
+})
+#define nvkm_ltc_fini(p,s) ({                                                  \
+	struct nvkm_ltc_priv *_priv = (p);                                     \
+	_nvkm_ltc_fini(nv_object(_priv), (s));                                 \
+})
+
+int  nvkm_ltc_create_(struct nouveau_object *, struct nouveau_object *,
+		      struct nouveau_oclass *, int, void **);
+
+#define _nvkm_ltc_dtor _nouveau_subdev_dtor
+int _nvkm_ltc_init(struct nouveau_object *);
+#define _nvkm_ltc_fini _nouveau_subdev_fini
+
+int  gf100_ltc_ctor(struct nouveau_object *, struct nouveau_object *,
+		    struct nouveau_oclass *, void *, u32,
+		    struct nouveau_object **);
+void gf100_ltc_dtor(struct nouveau_object *);
+int  gf100_ltc_init_tag_ram(struct nouveau_fb *, struct nvkm_ltc_priv *);
+int  gf100_ltc_tags_alloc(struct nouveau_ltc *, u32, struct nouveau_mm_node **);
+void gf100_ltc_tags_free(struct nouveau_ltc *, struct nouveau_mm_node **);
+
+struct nvkm_ltc_impl {
+	struct nouveau_oclass base;
+	void (*intr)(struct nouveau_subdev *);
+
+	void (*cbc_clear)(struct nvkm_ltc_priv *, u32 start, u32 limit);
+	void (*cbc_wait)(struct nvkm_ltc_priv *);
+
+	int zbc;
+	void (*zbc_clear_color)(struct nvkm_ltc_priv *, int, const u32[4]);
+	void (*zbc_clear_depth)(struct nvkm_ltc_priv *, int, const u32);
+};
+
+void gf100_ltc_intr(struct nouveau_subdev *);
+void gf100_ltc_cbc_clear(struct nvkm_ltc_priv *, u32, u32);
+void gf100_ltc_cbc_wait(struct nvkm_ltc_priv *);
+void gf100_ltc_zbc_clear_color(struct nvkm_ltc_priv *, int, const u32[4]);
+void gf100_ltc_zbc_clear_depth(struct nvkm_ltc_priv *, int, const u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gf100.h b/drivers/gpu/drm/nouveau/core/subdev/ltcg/gf100.h
deleted file mode 100644
index 87b10b8412ea..000000000000
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/gf100.h
+++ /dev/null
@@ -1,21 +0,0 @@
-#ifndef __NVKM_LTCG_PRIV_GF100_H__
-#define __NVKM_LTCG_PRIV_GF100_H__
-
-#include <subdev/ltcg.h>
-
-struct gf100_ltcg_priv {
-	struct nouveau_ltcg base;
-	u32 ltc_nr;
-	u32 lts_nr;
-	u32 num_tags;
-	u32 tag_base;
-	struct nouveau_mm tags;
-	struct nouveau_mm_node *tag_ram;
-};
-
-void gf100_ltcg_dtor(struct nouveau_object *);
-int  gf100_ltcg_init_tag_ram(struct nouveau_fb *, struct gf100_ltcg_priv *);
-int  gf100_ltcg_tags_alloc(struct nouveau_ltcg *, u32, struct nouveau_mm_node **);
-void gf100_ltcg_tags_free(struct nouveau_ltcg *, struct nouveau_mm_node **);
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/base.c b/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
index 8a5555192fa5..ca7cee3a314a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/base.c
@@ -22,9 +22,17 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/mc.h>
+#include "priv.h"
 #include <core/option.h>
 
+static inline void
+nouveau_mc_unk260(struct nouveau_mc *pmc, u32 data)
+{
+	const struct nouveau_mc_oclass *impl = (void *)nv_oclass(pmc);
+	if (impl->unk260)
+		impl->unk260(pmc, data);
+}
+
 static inline u32
 nouveau_mc_intr_mask(struct nouveau_mc *pmc)
 {
@@ -114,6 +122,8 @@ nouveau_mc_create_(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	pmc->unk260 = nouveau_mc_unk260;
+
 	if (nv_device_is_pci(device))
 		switch (device->pdev->device & 0x0ff0) {
 		case 0x00f0:
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/gk20a.c b/drivers/gpu/drm/nouveau/core/subdev/mc/gk20a.c
new file mode 100644
index 000000000000..b8d6cb435d0a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/gk20a.c
@@ -0,0 +1,38 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "nv04.h"
+
+struct nouveau_oclass *
+gk20a_mc_oclass = &(struct nouveau_mc_oclass) {
+	.base.handle = NV_SUBDEV(MC, 0xea),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv04_mc_ctor,
+		.dtor = _nouveau_mc_dtor,
+		.init = nv50_mc_init,
+		.fini = _nouveau_mc_fini,
+	},
+	.intr = nvc0_mc_intr,
+	.msi_rearm = nv40_mc_msi_rearm,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.h b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.h
index 81a408e7d034..4d9ea46c47c2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv04.h
@@ -1,7 +1,7 @@
 #ifndef __NVKM_MC_NV04_H__
 #define __NVKM_MC_NV04_H__
 
-#include <subdev/mc.h>
+#include "priv.h"
 
 struct nv04_mc_priv {
 	struct nouveau_mc base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
index f9c6a678b47d..15d41dc176ff 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
@@ -41,7 +41,7 @@ nvc0_mc_intr[] = {
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },	/* PMGR->GPIO */
 	{ 0x00200000, NVDEV_SUBDEV_I2C },	/* PMGR->I2C/AUX */
 	{ 0x01000000, NVDEV_SUBDEV_PWR },
-	{ 0x02000000, NVDEV_SUBDEV_LTCG },
+	{ 0x02000000, NVDEV_SUBDEV_LTC },
 	{ 0x08000000, NVDEV_SUBDEV_FB },
 	{ 0x10000000, NVDEV_SUBDEV_BUS },
 	{ 0x40000000, NVDEV_SUBDEV_IBUS },
@@ -56,6 +56,12 @@ nvc0_mc_msi_rearm(struct nouveau_mc *pmc)
 	nv_wr32(priv, 0x088704, 0x00000000);
 }
 
+void
+nvc0_mc_unk260(struct nouveau_mc *pmc, u32 data)
+{
+	nv_wr32(pmc, 0x000260, data);
+}
+
 struct nouveau_oclass *
 nvc0_mc_oclass = &(struct nouveau_mc_oclass) {
 	.base.handle = NV_SUBDEV(MC, 0xc0),
@@ -67,4 +73,5 @@ nvc0_mc_oclass = &(struct nouveau_mc_oclass) {
 	},
 	.intr = nvc0_mc_intr,
 	.msi_rearm = nvc0_mc_msi_rearm,
+	.unk260 = nvc0_mc_unk260,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc3.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc3.c
index 837e545aeb9f..68b5f61aadb5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc3.c
@@ -35,4 +35,5 @@ nvc3_mc_oclass = &(struct nouveau_mc_oclass) {
 	},
 	.intr = nvc0_mc_intr,
 	.msi_rearm = nv40_mc_msi_rearm,
+	.unk260 = nvc0_mc_unk260,
 }.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/priv.h b/drivers/gpu/drm/nouveau/core/subdev/mc/priv.h
new file mode 100644
index 000000000000..911e66392587
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/priv.h
@@ -0,0 +1,38 @@
+#ifndef __NVKM_MC_PRIV_H__
+#define __NVKM_MC_PRIV_H__
+
+#include <subdev/mc.h>
+
+#define nouveau_mc_create(p,e,o,d)                                             \
+	nouveau_mc_create_((p), (e), (o), sizeof(**d), (void **)d)
+#define nouveau_mc_destroy(p) ({                                               \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_dtor(nv_object(pmc));        \
+})
+#define nouveau_mc_init(p) ({                                                  \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_init(nv_object(pmc));        \
+})
+#define nouveau_mc_fini(p,s) ({                                                \
+	struct nouveau_mc *pmc = (p); _nouveau_mc_fini(nv_object(pmc), (s));   \
+})
+
+int  nouveau_mc_create_(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, int, void **);
+void _nouveau_mc_dtor(struct nouveau_object *);
+int  _nouveau_mc_init(struct nouveau_object *);
+int  _nouveau_mc_fini(struct nouveau_object *, bool);
+
+struct nouveau_mc_intr {
+	u32 stat;
+	u32 unit;
+};
+
+struct nouveau_mc_oclass {
+	struct nouveau_oclass base;
+	const struct nouveau_mc_intr *intr;
+	void (*msi_rearm)(struct nouveau_mc *);
+	void (*unk260)(struct nouveau_mc *, u32);
+};
+
+void nvc0_mc_unk260(struct nouveau_mc *, u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
index d4fd3bc9c66f..0ab55f27ec45 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/base.c
@@ -22,9 +22,18 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/pwr.h>
 #include <subdev/timer.h>
 
+#include "priv.h"
+
+static void
+nouveau_pwr_pgob(struct nouveau_pwr *ppwr, bool enable)
+{
+	const struct nvkm_pwr_impl *impl = (void *)nv_oclass(ppwr);
+	if (impl->pgob)
+		impl->pgob(ppwr, enable);
+}
+
 static int
 nouveau_pwr_send(struct nouveau_pwr *ppwr, u32 reply[2],
 		 u32 process, u32 message, u32 data0, u32 data1)
@@ -177,6 +186,7 @@ _nouveau_pwr_fini(struct nouveau_object *object, bool suspend)
 int
 _nouveau_pwr_init(struct nouveau_object *object)
 {
+	const struct nvkm_pwr_impl *impl = (void *)object->oclass;
 	struct nouveau_pwr *ppwr = (void *)object;
 	int ret, i;
 
@@ -186,24 +196,27 @@ _nouveau_pwr_init(struct nouveau_object *object)
 
 	nv_subdev(ppwr)->intr = nouveau_pwr_intr;
 	ppwr->message = nouveau_pwr_send;
+	ppwr->pgob = nouveau_pwr_pgob;
 
 	/* prevent previous ucode from running, wait for idle, reset */
 	nv_wr32(ppwr, 0x10a014, 0x0000ffff); /* INTR_EN_CLR = ALL */
 	nv_wait(ppwr, 0x10a04c, 0xffffffff, 0x00000000);
 	nv_mask(ppwr, 0x000200, 0x00002000, 0x00000000);
 	nv_mask(ppwr, 0x000200, 0x00002000, 0x00002000);
+	nv_rd32(ppwr, 0x000200);
+	nv_wait(ppwr, 0x10a10c, 0x00000006, 0x00000000);
 
 	/* upload data segment */
 	nv_wr32(ppwr, 0x10a1c0, 0x01000000);
-	for (i = 0; i < ppwr->data.size / 4; i++)
-		nv_wr32(ppwr, 0x10a1c4, ppwr->data.data[i]);
+	for (i = 0; i < impl->data.size / 4; i++)
+		nv_wr32(ppwr, 0x10a1c4, impl->data.data[i]);
 
 	/* upload code segment */
 	nv_wr32(ppwr, 0x10a180, 0x01000000);
-	for (i = 0; i < ppwr->code.size / 4; i++) {
+	for (i = 0; i < impl->code.size / 4; i++) {
 		if ((i & 0x3f) == 0)
 			nv_wr32(ppwr, 0x10a188, i >> 6);
-		nv_wr32(ppwr, 0x10a184, ppwr->code.data[i]);
+		nv_wr32(ppwr, 0x10a184, impl->code.data[i]);
 	}
 
 	/* start it running */
@@ -245,3 +258,15 @@ nouveau_pwr_create_(struct nouveau_object *parent,
 	init_waitqueue_head(&ppwr->recv.wait);
 	return 0;
 }
+
+int
+_nouveau_pwr_ctor(struct nouveau_object *parent,
+		  struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nouveau_pwr *ppwr;
+	int ret = nouveau_pwr_create(parent, engine, oclass, &ppwr);
+	*pobject = nv_object(ppwr);
+	return ret;
+}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc
new file mode 100644
index 000000000000..214a6d9e088d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/arith.fuc
@@ -0,0 +1,94 @@
+/*
+ * Copyright 2014 Martin Peres <martin.peres@free.fr>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the folloing conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+/******************************************************************************
+ * arith data segment
+ *****************************************************************************/
+#ifdef INCLUDE_PROC
+#endif
+
+#ifdef INCLUDE_DATA
+#endif
+
+/******************************************************************************
+ * arith code segment
+ *****************************************************************************/
+#ifdef INCLUDE_CODE
+
+// does a 32x32 -> 64 multiplication
+//
+// A * B = A_lo * B_lo
+//        + ( A_hi * B_lo ) << 16
+//        + ( A_lo * B_hi ) << 16
+//        + ( A_hi * B_hi ) << 32
+//
+// $r15 - current
+// $r14 - A
+// $r13 - B
+// $r12 - mul_lo (return)
+// $r11 - mul_hi (return)
+// $r0  - zero
+mulu32_32_64:
+	push $r1 // A_hi
+	push $r2 // B_hi
+	push $r3 // tmp0
+	push $r4 // tmp1
+
+	shr b32 $r1 $r14 16
+	shr b32 $r2 $r13 16
+
+	clear b32 $r12
+	clear b32 $r11
+
+	// A_lo * B_lo
+	mulu $r12 $r14 $r13
+
+	// ( A_hi * B_lo ) << 16
+	mulu $r3 $r1 $r13 // tmp0 = A_hi * B_lo
+	mov b32 $r4 $r3
+	and $r3 0xffff // tmp0 = tmp0_lo
+	shl b32 $r3 16
+	shr b32 $r4 16 // tmp1 = tmp0_hi
+	add b32 $r12 $r3
+	adc b32 $r11 $r4
+
+	// ( A_lo * B_hi ) << 16
+	mulu $r3 $r14 $r2 // tmp0 = A_lo * B_hi
+	mov b32 $r4 $r3
+	and $r3 0xffff // tmp0 = tmp0_lo
+	shl b32 $r3 16
+	shr b32 $r4 16 // tmp1 = tmp0_hi
+	add b32 $r12 $r3
+	adc b32 $r11 $r4
+
+	// ( A_hi * B_hi ) << 32
+	mulu $r3 $r1 $r2 // tmp0 = A_hi * B_hi
+	add b32 $r11 $r3
+
+	pop $r4
+	pop $r3
+	pop $r2
+	pop $r1
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
index 8f29badd785f..5cf5be63cbef 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/kernel.fuc
@@ -98,12 +98,16 @@ wr32:
 // $r14 - ns
 // $r0  - zero
 nsec:
+	push $r9
+	push $r8
 	nv_iord($r8, NV_PPWR_TIMER_LOW)
 	nsec_loop:
 		nv_iord($r9, NV_PPWR_TIMER_LOW)
 		sub b32 $r9 $r8
 		cmp b32 $r9 $r14
 		bra l #nsec_loop
+	pop $r8
+	pop $r9
 	ret
 
 // busy-wait for a period of time
@@ -115,6 +119,8 @@ nsec:
 // $r11 - timeout (ns)
 // $r0  - zero
 wait:
+	push $r9
+	push $r8
 	nv_iord($r8, NV_PPWR_TIMER_LOW)
 	wait_loop:
 		nv_rd32($r10, $r14)
@@ -126,6 +132,8 @@ wait:
 		cmp b32 $r9 $r11
 		bra l #wait_loop
 	wait_done:
+	pop $r8
+	pop $r9
 	ret
 
 // $r15 - current (kern)
@@ -242,12 +250,89 @@ intr:
 	bclr $flags $p0
 	iret
 
-// request the current process be sent a message after a timeout expires
+// calculate the number of ticks in the specified nanoseconds delay
+//
+// $r15 - current
+// $r14 - ns
+// $r14 - ticks (return)
+// $r0  - zero
+ticks_from_ns:
+	push $r12
+	push $r11
+
+	/* try not losing precision (multiply then divide) */
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+
+	/* use an immeditate, it's ok because HW_TICKS_PER_US < 16 bits */
+	div $r12 $r12 1000
+
+	/* check if there wasn't any overflow */
+	cmpu b32 $r11 0
+	bra e #ticks_from_ns_quit
+
+	/* let's divide then multiply, too bad for the precision! */
+	div $r14 $r14 1000
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+
+	/* this cannot overflow as long as HW_TICKS_PER_US < 1000 */
+
+ticks_from_ns_quit:
+	mov b32 $r14 $r12
+	pop $r11
+	pop $r12
+	ret
+
+// calculate the number of ticks in the specified microsecond delay
+//
+// $r15 - current
+// $r14 - us
+// $r14 - ticks (return)
+// $r0  - zero
+ticks_from_us:
+	push $r12
+	push $r11
+
+	/* simply multiply $us by HW_TICKS_PER_US */
+	imm32($r13, HW_TICKS_PER_US)
+	call #mulu32_32_64
+	mov b32 $r14 $r12
+
+	/* check if there wasn't any overflow */
+	cmpu b32 $r11 0
+	bra e #ticks_from_us_quit
+
+	/* Overflow! */
+	clear b32 $r14
+
+ticks_from_us_quit:
+	pop $r11
+	pop $r12
+	ret
+
+// calculate the number of ticks in the specified microsecond delay
 //
 // $r15 - current
 // $r14 - ticks
+// $r14 - us (return)
+// $r0  - zero
+ticks_to_us:
+	/* simply divide $ticks by HW_TICKS_PER_US */
+	imm32($r13, HW_TICKS_PER_US)
+	div $r14 $r14 $r13
+
+	ret
+
+// request the current process be sent a message after a timeout expires
+//
+// $r15 - current
+// $r14 - ticks (make sure it is < 2^31 to avoid any possible overflow)
 // $r0  - zero
 timer:
+	push $r9
+	push $r8
+
 	// interrupts off to prevent racing with timer isr
 	bclr $flags ie0
 
@@ -255,13 +340,22 @@ timer:
 	ld b32 $r8 D[$r15 + #proc_time]
 	cmp b32 $r8 0
 	bra g #timer_done
-	st b32 D[$r15 + #proc_time] $r14
 
-	// halt watchdog timer temporarily and check for a pending
-	// interrupt.  if there's one already pending, we can just
-	// bail since the timer isr will queue the next soonest
-	// right after it's done
+	// halt watchdog timer temporarily
+	clear b32 $r8
 	nv_iowr(NV_PPWR_WATCHDOG_ENABLE, $r8)
+
+	// find out how much time elapsed since the last update
+	// of the watchdog and add this time to the wanted ticks
+	nv_iord($r8, NV_PPWR_WATCHDOG_TIME)
+	ld b32 $r9 D[$r0 + #time_prev]
+	sub b32 $r9 $r8
+	add b32 $r14 $r9
+	st b32 D[$r15 + #proc_time] $r14
+
+	// check for a pending interrupt.  if there's one already
+	// pending, we can just bail since the timer isr will
+	// queue the next soonest right after it's done
 	nv_iord($r8, NV_PPWR_INTR)
 	and $r8 NV_PPWR_INTR_WATCHDOG
 	bra nz #timer_enable
@@ -272,10 +366,10 @@ timer:
 	cmp b32 $r14 $r0
 	bra e #timer_reset
 	cmp b32 $r14 $r8
-	bra l #timer_done
-	timer_reset:
-	nv_iowr(NV_PPWR_WATCHDOG_TIME, $r14)
-	st b32 D[$r0 + #time_prev] $r14
+	bra g #timer_enable
+		timer_reset:
+		nv_iowr(NV_PPWR_WATCHDOG_TIME, $r14)
+		st b32 D[$r0 + #time_prev] $r14
 
 	// re-enable the watchdog timer
 	timer_enable:
@@ -285,6 +379,9 @@ timer:
 	// interrupts back on
 	timer_done:
 	bset $flags ie0
+
+	pop $r8
+	pop $r9
 	ret
 
 // send message to another process
@@ -371,6 +468,9 @@ send:
 // $r14 - process
 // $r0  - zero
 recv:
+	push $r9
+	push $r8
+
 	ld b32 $r8 D[$r14 + #proc_qget]
 	ld b32 $r9 D[$r14 + #proc_qput]
 	bclr $flags $p1
@@ -403,6 +503,8 @@ recv:
 		bset $flags $p1
 		pop $r15
 	recv_done:
+	pop $r8
+	pop $r9
 	ret
 
 init:
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
index e2a63ac5422b..96fc984dafdc 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/macros.fuc
@@ -242,7 +242,7 @@
 */	push reg /*
 */	pop $r13 /*
 */	pop $r14 /*
-*/	call(wr32) /*
+*/	call(wr32)
 #else
 #define nv_wr32(addr,reg) /*
 */ 	sethi $r0 0x14000000 /*
@@ -250,3 +250,23 @@
 */	st b32 D[$r0] reg /*
 */	clear b32 $r0
 #endif
+
+#define st(size, addr, reg) /*
+*/	movw $r0 addr /*
+*/	st size D[$r0] reg /*
+*/	clear b32 $r0
+
+#define ld(size, reg, addr) /*
+*/	movw $r0 addr /*
+*/	ld size reg D[$r0] /*
+*/	clear b32 $r0
+
+// does a 64+64 -> 64 unsigned addition (C = A + B)
+#define addu64(reg_a_c_hi, reg_a_c_lo, b_hi, b_lo) /*
+*/    add b32 reg_a_c_lo b_lo /*
+*/    adc b32 reg_a_c_hi b_hi
+
+// does a 64+64 -> 64 substraction (C = A - B)
+#define subu64(reg_a_c_hi, reg_a_c_lo, b_hi, b_lo) /*
+*/    sub b32 reg_a_c_lo b_lo /*
+*/    sbb b32 reg_a_c_hi b_hi
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
index d43741eccb11..e89789a53b80 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/memx.fuc
@@ -43,17 +43,23 @@ process(PROC_MEMX, #memx_init, #memx_recv)
 */	.b32 func
 
 memx_func_head:
-handler(ENTER , 0x0001, 0x0000, #memx_func_enter)
+handler(ENTER , 0x0000, 0x0000, #memx_func_enter)
 memx_func_next:
 handler(LEAVE , 0x0000, 0x0000, #memx_func_leave)
 handler(WR32  , 0x0000, 0x0002, #memx_func_wr32)
 handler(WAIT  , 0x0004, 0x0000, #memx_func_wait)
 handler(DELAY , 0x0001, 0x0000, #memx_func_delay)
+handler(VBLANK, 0x0001, 0x0000, #memx_func_wait_vblank)
 memx_func_tail:
 
 .equ #memx_func_size #memx_func_next - #memx_func_head
 .equ #memx_func_num (#memx_func_tail - #memx_func_head) / #memx_func_size
 
+memx_ts_start:
+.b32 0
+memx_ts_end:
+.b32 0
+
 memx_data_head:
 .skip 0x0800
 memx_data_tail:
@@ -67,19 +73,44 @@ memx_data_tail:
 //
 // $r15 - current (memx)
 // $r4  - packet length
-//	+00: bitmask of heads to wait for vblank on
 // $r3  - opcode desciption
 // $r0  - zero
 memx_func_enter:
+#if NVKM_PPWR_CHIPSET == GT215
+	movw $r8 0x1610
+	nv_rd32($r7, $r8)
+	imm32($r6, 0xfffffffc)
+	and $r7 $r6
+	movw $r6 0x2
+	or $r7 $r6
+	nv_wr32($r8, $r7)
+#else
+	movw $r6 0x001620
+	imm32($r7, ~0x00000aa2);
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+
+	imm32($r7, ~0x00000001)
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+
+	movw $r6 0x0026f0
+	nv_rd32($r8, $r6)
+	and $r8 $r7
+	nv_wr32($r6, $r8)
+#endif
+
 	mov $r6 NV_PPWR_OUTPUT_SET_FB_PAUSE
 	nv_iowr(NV_PPWR_OUTPUT_SET, $r6)
 	memx_func_enter_wait:
 		nv_iord($r6, NV_PPWR_OUTPUT)
 		and $r6 NV_PPWR_OUTPUT_FB_PAUSE
 		bra z #memx_func_enter_wait
-	//XXX: TODO
-	ld b32 $r6 D[$r1 + 0x00]
-	add b32 $r1 0x04
+
+	nv_iord($r6, NV_PPWR_TIMER_LOW)
+	st b32 D[$r0 + #memx_ts_start] $r6
 	ret
 
 // description
@@ -89,14 +120,93 @@ memx_func_enter:
 // $r3  - opcode desciption
 // $r0  - zero
 memx_func_leave:
+	nv_iord($r6, NV_PPWR_TIMER_LOW)
+	st b32 D[$r0 + #memx_ts_end] $r6
+
 	mov $r6 NV_PPWR_OUTPUT_CLR_FB_PAUSE
 	nv_iowr(NV_PPWR_OUTPUT_CLR, $r6)
 	memx_func_leave_wait:
 		nv_iord($r6, NV_PPWR_OUTPUT)
 		and $r6 NV_PPWR_OUTPUT_FB_PAUSE
 		bra nz #memx_func_leave_wait
+
+#if NVKM_PPWR_CHIPSET == GT215
+	movw $r8 0x1610
+	nv_rd32($r7, $r8)
+	imm32($r6, 0xffffffcc)
+	and $r7 $r6
+	nv_wr32($r8, $r7)
+#else
+	movw $r6 0x0026f0
+	imm32($r7, 0x00000001)
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+
+	movw $r6 0x001620
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+
+	imm32($r7, 0x00000aa2);
+	nv_rd32($r8, $r6)
+	or $r8 $r7
+	nv_wr32($r6, $r8)
+#endif
+	ret
+
+#if NVKM_PPWR_CHIPSET < GF119
+// description
+//
+// $r15 - current (memx)
+// $r4  - packet length
+//	+00: head to wait for vblank on
+// $r3  - opcode desciption
+// $r0  - zero
+memx_func_wait_vblank:
+	ld b32 $r6 D[$r1 + 0x00]
+	cmp b32 $r6 0x0
+	bra z #memx_func_wait_vblank_head0
+	cmp b32 $r6 0x1
+	bra z #memx_func_wait_vblank_head1
+	bra #memx_func_wait_vblank_fini
+
+	memx_func_wait_vblank_head1:
+	movw $r7 0x20
+	bra #memx_func_wait_vblank_0
+
+	memx_func_wait_vblank_head0:
+	movw $r7 0x8
+
+	memx_func_wait_vblank_0:
+		nv_iord($r6, NV_PPWR_INPUT)
+		and $r6 $r7
+		bra nz #memx_func_wait_vblank_0
+
+	memx_func_wait_vblank_1:
+		nv_iord($r6, NV_PPWR_INPUT)
+		and $r6 $r7
+		bra z #memx_func_wait_vblank_1
+
+	memx_func_wait_vblank_fini:
+	add b32 $r1 0x4
+	ret
+
+#else
+
+// XXX: currently no-op
+//
+// $r15 - current (memx)
+// $r4  - packet length
+//	+00: head to wait for vblank on
+// $r3  - opcode desciption
+// $r0  - zero
+memx_func_wait_vblank:
+	add b32 $r1 0x4
 	ret
 
+#endif
+
 // description
 //
 // $r15 - current (memx)
@@ -160,14 +270,17 @@ memx_exec:
 	push $r13
 	mov b32 $r1 $r12
 	mov b32 $r2 $r11
+
 	memx_exec_next:
-		// fetch the packet header, and locate opcode info
+		// fetch the packet header
 		ld b32 $r3 D[$r1]
 		add b32 $r1 4
-		shr b32 $r4 $r3 16
-		mulu $r3 #memx_func_size
+		extr $r4 $r3 16:31
+		extr $r3 $r3 0:15
 
 		// execute the opcode handler
+		sub b32 $r3 1
+		mulu $r3 #memx_func_size
 		ld b32 $r5 D[$r3 + #memx_func_head + #memx_func]
 		call $r5
 
@@ -176,6 +289,10 @@ memx_exec:
 		bra l #memx_exec_next
 
 	// send completion reply
+	ld b32 $r11 D[$r0 + #memx_ts_start]
+	ld b32 $r12 D[$r0 + #memx_ts_end]
+	sub b32 $r12 $r11
+	nv_iord($r11, NV_PPWR_INPUT)
 	pop $r13
 	pop $r14
 	call(send)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
index 17a8a383d91a..b439519ec866 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GK208
+#define HW_TICKS_PER_US 324
 
 #define NVKM_FALCON_PC24
 #define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nv108_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nv108_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
index 39a5dc150a05..4d278a96b2bb 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nv108.fuc.h
@@ -24,8 +24,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000379,
-	0x0000032a,
+	0x00000453,
+	0x00000404,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x0000046f,
-	0x00000461,
+	0x0000061c,
+	0x0000060e,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000473,
-	0x00000471,
+	0x00000620,
+	0x0000061e,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x00000877,
-	0x0000071e,
+	0x00000a24,
+	0x000008cb,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x00000898,
-	0x00000879,
+	0x00000a45,
+	0x00000a26,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x000008a3,
-	0x000008a1,
+	0x00000a50,
+	0x00000a4e,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,31 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x000003a9,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x000003c7,
+	0x00000483,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x00000500,
+	0x00000003,
 	0x00000002,
-	0x000003df,
-	0x00040003,
+	0x00000580,
+	0x00040004,
+	0x00000000,
+	0x0000059d,
+	0x00010005,
+	0x00000000,
+	0x000005b7,
+	0x00010006,
+	0x00000000,
+	0x0000057b,
+/* 0x03b8: memx_func_tail */
+/* 0x03b8: memx_ts_start */
 	0x00000000,
-	0x00000407,
-	0x00010004,
+/* 0x03bc: memx_ts_end */
 	0x00000000,
-	0x00000421,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
+/* 0x03c0: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +764,8 @@ uint32_t nv108_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+/* 0x0bc0: memx_data_tail */
+/* 0x0bc0: i2c_scl_map */
 	0x00000400,
 	0x00000800,
 	0x00001000,
@@ -769,7 +776,7 @@ uint32_t nv108_pwr_data[] = {
 	0x00020000,
 	0x00040000,
 	0x00080000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0be8: i2c_sda_map */
 	0x00100000,
 	0x00200000,
 	0x00400000,
@@ -781,10 +788,69 @@ uint32_t nv108_pwr_data[] = {
 	0x10000000,
 	0x20000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
 
 uint32_t nv108_pwr_code[] = {
-	0x02910ef5,
+	0x031c0ef5,
 /* 0x0004: rd32 */
 	0xf607a040,
 	0x04bd000e,
@@ -812,15 +878,18 @@ uint32_t nv108_pwr_code[] = {
 	0x7000d4f1,
 	0xf8f61bf4,
 /* 0x005d: nsec */
-	0xcf2c0800,
-/* 0x0062: nsec_loop */
+	0xf990f900,
+	0xcf2c0880,
+/* 0x0066: nsec_loop */
 	0x2c090088,
 	0xbb0099cf,
 	0x9ea60298,
-	0xf8f61ef4,
-/* 0x0071: wait */
-	0xcf2c0800,
-/* 0x0076: wait_loop */
+	0xfcf61ef4,
+	0xf890fc80,
+/* 0x0079: wait */
+	0xf990f900,
+	0xcf2c0880,
+/* 0x0082: wait_loop */
 	0xeeb20088,
 	0x0000047e,
 	0xadfddab2,
@@ -828,28 +897,29 @@ uint32_t nv108_pwr_code[] = {
 	0x2c09100b,
 	0xbb0099cf,
 	0x9ba60298,
-/* 0x0093: wait_done */
-	0xf8e61ef4,
-/* 0x0095: intr_watchdog */
+/* 0x009f: wait_done */
+	0xfce61ef4,
+	0xf890fc80,
+/* 0x00a5: intr_watchdog */
 	0x03e99800,
 	0xf40096b0,
 	0x0a98280b,
 	0x029abb9a,
 	0x0d0e1cf4,
-	0x01de7e01,
+	0x02617e01,
 	0xf494bd00,
-/* 0x00b2: intr_watchdog_next_time */
+/* 0x00c2: intr_watchdog_next_time */
 	0x0a98140e,
 	0x00a6b09b,
 	0xa6080bf4,
 	0x061cf49a,
-/* 0x00c0: intr_watchdog_next_time_set */
-/* 0x00c3: intr_watchdog_next_proc */
+/* 0x00d0: intr_watchdog_next_time_set */
+/* 0x00d3: intr_watchdog_next_proc */
 	0xb59b09b5,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc81bf402,
-/* 0x00d2: intr */
+/* 0x00e2: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -865,13 +935,13 @@ uint32_t nv108_pwr_code[] = {
 	0xc40088cf,
 	0x0bf40289,
 	0x9b00b51f,
-	0x957e580e,
+	0xa57e580e,
 	0x09980000,
 	0x0096b09b,
 	0x000d0bf4,
 	0x0009f634,
 	0x09b504bd,
-/* 0x0125: intr_skip_watchdog */
+/* 0x0135: intr_skip_watchdog */
 	0x0089e49a,
 	0x360bf408,
 	0xcf068849,
@@ -881,20 +951,20 @@ uint32_t nv108_pwr_code[] = {
 	0xc0f900cc,
 	0xf14f484e,
 	0x0d5453e3,
-	0x023f7e00,
+	0x02c27e00,
 	0x40c0fc00,
 	0x0cf604c0,
-/* 0x0157: intr_subintr_skip_fifo */
+/* 0x0167: intr_subintr_skip_fifo */
 	0x4004bd00,
 	0x09f60688,
-/* 0x015f: intr_skip_subintr */
+/* 0x016f: intr_skip_subintr */
 	0xc404bd00,
 	0x0bf42089,
 	0xbfa4f107,
-/* 0x0169: intr_skip_pause */
+/* 0x0179: intr_skip_pause */
 	0x4089c4ff,
 	0xf1070bf4,
-/* 0x0173: intr_skip_user0 */
+/* 0x0183: intr_skip_user0 */
 	0x00ffbfa4,
 	0x0008f604,
 	0x80fc04bd,
@@ -904,551 +974,682 @@ uint32_t nv108_pwr_code[] = {
 	0xfca0fcb0,
 	0xfc80fc90,
 	0x0032f400,
-/* 0x0196: timer */
-	0x32f401f8,
-	0x03f89810,
-	0xf40086b0,
-	0xfeb53a1c,
-	0xf6380003,
+/* 0x01a6: ticks_from_ns */
+	0xc0f901f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0x7721f500,
+	0xe8ccec03,
+	0x00b4b003,
+	0xec120bf4,
+	0xf103e8ee,
+	0xf00144d7,
+	0x21f500d3,
+/* 0x01ce: ticks_from_ns_quit */
+	0xceb20377,
+	0xc0fcb0fc,
+/* 0x01d6: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0x7721f500,
+	0xb0ceb203,
+	0x0bf400b4,
+/* 0x01ef: ticks_from_us_quit */
+	0xfce4bd05,
+	0xf8c0fcb0,
+/* 0x01f5: ticks_to_us */
+	0x44d7f100,
+	0x00d3f001,
+	0xf8ecedff,
+/* 0x0201: timer */
+	0xf990f900,
+	0x1032f480,
+	0xb003f898,
+	0x1cf40086,
+	0x0084bd4a,
+	0x0008f638,
+	0x340804bd,
+	0x980088cf,
+	0x98bb9a09,
+	0x00e9bb02,
+	0x0803feb5,
+	0x0088cf08,
+	0xf40284f0,
+	0x34081c1b,
+	0xa60088cf,
+	0x080bf4e0,
+	0x1cf4e8a6,
+/* 0x0245: timer_reset */
+	0xf634000d,
+	0x04bd000e,
+/* 0x024f: timer_enable */
+	0x089a0eb5,
+	0xf6380001,
 	0x04bd0008,
-	0x88cf0808,
-	0x0284f000,
-	0x081c1bf4,
-	0x0088cf34,
-	0x0bf4e0a6,
-	0xf4e8a608,
-/* 0x01c6: timer_reset */
-	0x3400161e,
-	0xbd000ef6,
-	0x9a0eb504,
-/* 0x01d0: timer_enable */
-	0x38000108,
-	0xbd0008f6,
-/* 0x01d9: timer_done */
-	0x1031f404,
-/* 0x01de: send_proc */
-	0x80f900f8,
-	0xe89890f9,
-	0x04e99805,
-	0xa60486f0,
-	0x2a0bf489,
-	0x940398c4,
-	0x80b60488,
-	0x008ebb18,
-	0xb500fa98,
-	0x8db5008a,
-	0x028cb501,
-	0xb6038bb5,
-	0x94f00190,
-	0x04e9b507,
-/* 0x0217: send_done */
-	0xfc0231f4,
-	0xf880fc90,
-/* 0x021d: find */
-	0x0880f900,
-	0x0131f458,
-/* 0x0224: find_loop */
-	0xa6008a98,
-	0x100bf4ae,
-	0xb15880b6,
-	0xf4026886,
-	0x32f4f11b,
-/* 0x0239: find_done */
-	0xfc8eb201,
-/* 0x023f: send */
-	0x7e00f880,
-	0xf400021d,
-	0x00f89b01,
-/* 0x0248: recv */
-	0x9805e898,
-	0x32f404e9,
-	0xf489a601,
-	0x89c43c0b,
-	0x0180b603,
-	0xb50784f0,
-	0xea9805e8,
-	0xfef0f902,
-	0xf0f9018f,
-	0x9994efb2,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x028f: recv_done */
-	0xf8f0fc01,
-/* 0x0291: init */
-	0x01084100,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0x00e04100,
-	0x000013f0,
-	0x0001f61c,
-	0xff0104bd,
-	0x01f61400,
-	0x0104bd00,
-	0x0015f102,
-	0xf6100008,
-	0x04bd0001,
-	0xf000d241,
-	0x10fe0013,
-	0x1031f400,
-	0x38000101,
+/* 0x0258: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x0261: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89a60486,
+	0xc42a0bf4,
+	0x88940398,
+	0x1880b604,
+	0x98008ebb,
+	0x8ab500fa,
+	0x018db500,
+	0xb5028cb5,
+	0x90b6038b,
+	0x0794f001,
+	0xf404e9b5,
+/* 0x029a: send_done */
+	0x90fc0231,
+	0x00f880fc,
+/* 0x02a0: find */
+	0x580880f9,
+/* 0x02a7: find_loop */
+	0x980131f4,
+	0xaea6008a,
+	0xb6100bf4,
+	0x86b15880,
+	0x1bf40268,
+	0x0132f4f1,
+/* 0x02bc: find_done */
+	0x80fc8eb2,
+/* 0x02c2: send */
+	0xa07e00f8,
+	0x01f40002,
+/* 0x02cb: recv */
+	0xf900f89b,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0x0bf489a6,
+	0x0389c43c,
+	0xf00180b6,
+	0xe8b50784,
+	0x02ea9805,
+	0x8ffef0f9,
+	0xb2f0f901,
+	0x049994ef,
+	0xb600e9bb,
+	0xeb9818e0,
+	0x02ec9803,
+	0x9801ed98,
+	0xa5f900ee,
+	0xf8fef0fc,
+	0x0131f400,
+/* 0x0316: recv_done */
+	0x80fcf0fc,
+	0x00f890fc,
+/* 0x031c: init */
+	0xcf010841,
+	0x11e70011,
+	0x14b60109,
+	0x0014fe08,
+	0xf000e041,
+	0x1c000013,
 	0xbd0001f6,
-/* 0x02db: init_proc */
-	0x98580f04,
-	0x16b001f1,
-	0xfa0bf400,
-	0xf0b615f9,
-	0xf20ef458,
-/* 0x02ec: host_send */
-	0xcf04b041,
-	0xa0420011,
-	0x0022cf04,
-	0x0bf412a6,
-	0x071ec42e,
-	0xb704ee94,
-	0x980270e0,
-	0xec9803eb,
-	0x01ed9802,
-	0x7e00ee98,
-	0xb600023f,
-	0x1ec40110,
-	0x04b0400f,
-	0xbd000ef6,
-	0xc70ef404,
-/* 0x0328: host_send_done */
-/* 0x032a: host_recv */
-	0x494100f8,
-	0x5413f14e,
-	0xf4e1a652,
-/* 0x0336: host_recv_wait */
-	0xcc41b90b,
-	0x0011cf04,
-	0xcf04c842,
-	0x16f00022,
-	0xf412a608,
-	0x23c4ef0b,
-	0x0434b607,
-	0x02f030b7,
-	0xb5033bb5,
-	0x3db5023c,
-	0x003eb501,
-	0xf00120b6,
-	0xc8400f24,
-	0x0002f604,
-	0x400204bd,
-	0x02f60000,
-	0xf804bd00,
-/* 0x0379: host_init */
-	0x00804100,
-	0xf11014b6,
-	0x40027015,
-	0x01f604d0,
+	0x00ff0104,
+	0x0001f614,
+	0x020104bd,
+	0x080015f1,
+	0x01f61000,
 	0x4104bd00,
+	0x13f000e2,
+	0x0010fe00,
+	0x011031f4,
+	0xf6380001,
+	0x04bd0001,
+/* 0x0366: init_proc */
+	0xf198580f,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0377: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb2301dff,
+	0xff34f134,
+	0x1034b6ff,
+	0xbb1045b6,
+	0xb4bb00c3,
+	0x30e2ff01,
+	0x34f134b2,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0x12ff01b4,
+	0x00b3bb30,
+	0x30fc40fc,
+	0x10fc20fc,
+/* 0x03c6: host_send */
+	0xb04100f8,
+	0x0011cf04,
+	0xcf04a042,
+	0x12a60022,
+	0xc42e0bf4,
+	0xee94071e,
+	0x70e0b704,
+	0x03eb9802,
+	0x9802ec98,
+	0xee9801ed,
+	0x02c27e00,
+	0x0110b600,
+	0x400f1ec4,
+	0x0ef604b0,
+	0xf404bd00,
+/* 0x0402: host_send_done */
+	0x00f8c70e,
+/* 0x0404: host_recv */
+	0xf14e4941,
+	0xa6525413,
+	0xb90bf4e1,
+/* 0x0410: host_recv_wait */
+	0xcf04cc41,
+	0xc8420011,
+	0x0022cf04,
+	0xa60816f0,
+	0xef0bf412,
+	0xb60723c4,
+	0x30b70434,
+	0x3bb502f0,
+	0x023cb503,
+	0xb5013db5,
+	0x20b6003e,
+	0x0f24f001,
+	0xf604c840,
+	0x04bd0002,
+	0x00004002,
+	0xbd0002f6,
+/* 0x0453: host_init */
+	0x4100f804,
 	0x14b60080,
-	0xf015f110,
-	0x04dc4002,
+	0x7015f110,
+	0x04d04002,
+	0xbd0001f6,
+	0x00804104,
+	0xf11014b6,
+	0x4002f015,
+	0x01f604dc,
+	0x0104bd00,
+	0x04c44001,
 	0xbd0001f6,
-	0x40010104,
-	0x01f604c4,
-	0xf804bd00,
-/* 0x03a9: memx_func_enter */
-	0x40040600,
-	0x06f607e0,
-/* 0x03b3: memx_func_enter_wait */
-	0x4604bd00,
-	0x66cf07c0,
-	0x0464f000,
-	0x98f70bf4,
-	0x10b60016,
-/* 0x03c7: memx_func_leave */
-	0x0600f804,
-	0x07e44004,
-	0xbd0006f6,
-/* 0x03d1: memx_func_leave_wait */
-	0x07c04604,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x03df: memx_func_wr32 */
-	0x9800f8f7,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
+/* 0x0483: memx_func_enter */
+	0xf100f804,
+	0xf1162067,
+	0xf1f55d77,
+	0xb2ffff73,
+	0x00047e6e,
+	0xfdd8b200,
+	0x60f90487,
+	0xd0fc80f9,
+	0x2e7ee0fc,
+	0x77f10000,
+	0x73f1fffe,
+	0x6eb2ffff,
+	0x0000047e,
+	0x87fdd8b2,
+	0xf960f904,
+	0xfcd0fc80,
+	0x002e7ee0,
+	0xf067f100,
+	0x7e6eb226,
+	0xb2000004,
+	0x0487fdd8,
+	0x80f960f9,
 	0xe0fcd0fc,
 	0x00002e7e,
-	0x140003f1,
-	0xa00506fd,
-	0xb604bd05,
-	0x1bf40242,
-/* 0x0407: memx_func_wait */
-	0x0800f8dd,
-	0x0088cf2c,
-	0x98001e98,
-	0x1c98011d,
-	0x031b9802,
-	0x7e1010b6,
-	0xf8000071,
-/* 0x0421: memx_func_delay */
-	0x001e9800,
-	0x7e0410b6,
-	0xf800005d,
-/* 0x042d: memx_exec */
-	0xf9e0f900,
-	0xb2c1b2d0,
-/* 0x0435: memx_exec_next */
-	0x001398b2,
-	0x950410b6,
-	0x30f01034,
-	0xde35980c,
-	0x12a655f9,
-	0xfced1ef4,
+	0xe0400406,
+	0x0006f607,
+/* 0x04ea: memx_func_enter_wait */
+	0xc04604bd,
+	0x0066cf07,
+	0xf40464f0,
+	0x2c06f70b,
+	0xb50066cf,
+	0x00f8ee06,
+/* 0x0500: memx_func_leave */
+	0x66cf2c06,
+	0xef06b500,
+	0xe4400406,
+	0x0006f607,
+/* 0x0512: memx_func_leave_wait */
+	0xc04604bd,
+	0x0066cf07,
+	0xf40464f0,
+	0x67f1f71b,
+	0x77f126f0,
+	0x73f00001,
+	0x7e6eb200,
+	0xb2000004,
+	0x0587fdd8,
+	0x80f960f9,
+	0xe0fcd0fc,
+	0x00002e7e,
+	0x162067f1,
+	0x047e6eb2,
+	0xd8b20000,
+	0xf90587fd,
+	0xfc80f960,
 	0x7ee0fcd0,
-	0xf800023f,
-/* 0x0455: memx_info */
-	0x03ac4c00,
-	0x7e08004b,
-	0xf800023f,
-/* 0x0461: memx_recv */
-	0x01d6b000,
-	0xb0c90bf4,
-	0x0bf400d6,
-/* 0x046f: memx_init */
-	0xf800f8eb,
-/* 0x0471: perf_recv */
-/* 0x0473: perf_init */
-	0xf800f800,
-/* 0x0475: i2c_drive_scl */
-	0x0036b000,
-	0x400d0bf4,
-	0x01f607e0,
-	0xf804bd00,
-/* 0x0485: i2c_drive_scl_lo */
-	0x07e44000,
-	0xbd0001f6,
-/* 0x048f: i2c_drive_sda */
-	0xb000f804,
-	0x0bf40036,
-	0x07e0400d,
-	0xbd0002f6,
-/* 0x049f: i2c_drive_sda_lo */
-	0x4000f804,
-	0x02f607e4,
-	0xf804bd00,
-/* 0x04a9: i2c_sense_scl */
-	0x0132f400,
-	0xcf07c443,
-	0x31fd0033,
-	0x060bf404,
-/* 0x04bb: i2c_sense_scl_done */
-	0xf80131f4,
-/* 0x04bd: i2c_sense_sda */
-	0x0132f400,
-	0xcf07c443,
-	0x32fd0033,
-	0x060bf404,
-/* 0x04cf: i2c_sense_sda_done */
-	0xf80131f4,
-/* 0x04d1: i2c_raise_scl */
-	0x4440f900,
-	0x01030898,
-	0x0004757e,
-/* 0x04dc: i2c_raise_scl_wait */
-	0x7e03e84e,
-	0x7e00005d,
-	0xf40004a9,
-	0x42b60901,
-	0xef1bf401,
-/* 0x04f0: i2c_raise_scl_done */
-	0x00f840fc,
-/* 0x04f4: i2c_start */
-	0x0004a97e,
-	0x7e0d11f4,
-	0xf40004bd,
-	0x0ef40611,
-/* 0x0505: i2c_start_rep */
-	0x7e00032e,
-	0x03000475,
-	0x048f7e01,
-	0x0076bb00,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0xd17e50fc,
-	0x64b60004,
-	0x1d11f404,
-/* 0x0530: i2c_start_send */
-	0x8f7e0003,
-	0x884e0004,
-	0x005d7e13,
-	0x7e000300,
-	0x4e000475,
-	0x5d7e1388,
-/* 0x054a: i2c_start_out */
-	0x00f80000,
-/* 0x054c: i2c_stop */
-	0x757e0003,
-	0x00030004,
-	0x00048f7e,
-	0x7e03e84e,
-	0x0300005d,
-	0x04757e01,
-	0x13884e00,
+	0xf100002e,
+	0xf00aa277,
+	0x6eb20073,
+	0x0000047e,
+	0x87fdd8b2,
+	0xf960f905,
+	0xfcd0fc80,
+	0x002e7ee0,
+/* 0x057b: memx_func_wait_vblank */
+	0xb600f800,
+	0x00f80410,
+/* 0x0580: memx_func_wr32 */
+	0x98001698,
+	0x10b60115,
+	0xf960f908,
+	0xfcd0fc50,
+	0x002e7ee0,
+	0x0242b600,
+	0xf8e81bf4,
+/* 0x059d: memx_func_wait */
+	0xcf2c0800,
+	0x1e980088,
+	0x011d9800,
+	0x98021c98,
+	0x10b6031b,
+	0x00797e10,
+/* 0x05b7: memx_func_delay */
+	0x9800f800,
+	0x10b6001e,
+	0x005d7e04,
+/* 0x05c3: memx_exec */
+	0xf900f800,
+	0xb2d0f9e0,
+/* 0x05cb: memx_exec_next */
+	0x98b2b2c1,
+	0x10b60013,
+	0xf034e704,
+	0xe033e701,
+	0x0132b601,
+	0x980c30f0,
+	0x55f9de35,
+	0x1ef412a6,
+	0xee0b98e5,
+	0xbbef0c98,
+	0xc44b02cb,
+	0x00bbcf07,
+	0xe0fcd0fc,
+	0x0002c27e,
+/* 0x0602: memx_info */
+	0xc04c00f8,
+	0x08004b03,
+	0x0002c27e,
+/* 0x060e: memx_recv */
+	0xd6b000f8,
+	0xb20bf401,
+	0xf400d6b0,
+	0x00f8eb0b,
+/* 0x061c: memx_init */
+/* 0x061e: perf_recv */
+	0x00f800f8,
+/* 0x0620: perf_init */
+/* 0x0622: i2c_drive_scl */
+	0x36b000f8,
+	0x0d0bf400,
+	0xf607e040,
+	0x04bd0001,
+/* 0x0632: i2c_drive_scl_lo */
+	0xe44000f8,
+	0x0001f607,
+	0x00f804bd,
+/* 0x063c: i2c_drive_sda */
+	0xf40036b0,
+	0xe0400d0b,
+	0x0002f607,
+	0x00f804bd,
+/* 0x064c: i2c_drive_sda_lo */
+	0xf607e440,
+	0x04bd0002,
+/* 0x0656: i2c_sense_scl */
+	0x32f400f8,
+	0x07c44301,
+	0xfd0033cf,
+	0x0bf40431,
+	0x0131f406,
+/* 0x0668: i2c_sense_scl_done */
+/* 0x066a: i2c_sense_sda */
+	0x32f400f8,
+	0x07c44301,
+	0xfd0033cf,
+	0x0bf40432,
+	0x0131f406,
+/* 0x067c: i2c_sense_sda_done */
+/* 0x067e: i2c_raise_scl */
+	0x40f900f8,
+	0x03089844,
+	0x06227e01,
+/* 0x0689: i2c_raise_scl_wait */
+	0x03e84e00,
 	0x00005d7e,
-	0x8f7e0103,
-	0x884e0004,
-	0x005d7e13,
-/* 0x057b: i2c_bitw */
-	0x7e00f800,
-	0x4e00048f,
-	0x5d7e03e8,
-	0x76bb0000,
+	0x0006567e,
+	0xb60901f4,
+	0x1bf40142,
+/* 0x069d: i2c_raise_scl_done */
+	0xf840fcef,
+/* 0x06a1: i2c_start */
+	0x06567e00,
+	0x0d11f400,
+	0x00066a7e,
+	0xf40611f4,
+/* 0x06b2: i2c_start_rep */
+	0x00032e0e,
+	0x0006227e,
+	0x3c7e0103,
+	0x76bb0006,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb60004d1,
+	0xb600067e,
 	0x11f40464,
-	0x13884e17,
-	0x00005d7e,
-	0x757e0003,
-	0x884e0004,
-	0x005d7e13,
-/* 0x05b9: i2c_bitw_out */
-/* 0x05bb: i2c_bitr */
-	0x0300f800,
-	0x048f7e01,
+/* 0x06dd: i2c_start_send */
+	0x7e00031d,
+	0x4e00063c,
+	0x5d7e1388,
+	0x00030000,
+	0x0006227e,
+	0x7e13884e,
+/* 0x06f7: i2c_start_out */
+	0xf800005d,
+/* 0x06f9: i2c_stop */
+	0x7e000300,
+	0x03000622,
+	0x063c7e00,
 	0x03e84e00,
 	0x00005d7e,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x04d17e50,
-	0x0464b600,
-	0x7e1a11f4,
-	0x030004bd,
-	0x04757e00,
-	0x13884e00,
-	0x00005d7e,
-	0xf4013cf0,
-/* 0x05fe: i2c_bitr_done */
-	0x00f80131,
-/* 0x0600: i2c_get_byte */
-	0x08040005,
-/* 0x0604: i2c_get_byte_next */
-	0xbb0154b6,
+	0x227e0103,
+	0x884e0006,
+	0x005d7e13,
+	0x7e010300,
+	0x4e00063c,
+	0x5d7e1388,
+	0x00f80000,
+/* 0x0728: i2c_bitw */
+	0x00063c7e,
+	0x7e03e84e,
+	0xbb00005d,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x0005bb7e,
+	0x00067e7e,
 	0xf40464b6,
-	0x53fd2a11,
-	0x0142b605,
-	0x03d81bf4,
-	0x0076bb01,
+	0x884e1711,
+	0x005d7e13,
+	0x7e000300,
+	0x4e000622,
+	0x5d7e1388,
+/* 0x0766: i2c_bitw_out */
+	0x00f80000,
+/* 0x0768: i2c_bitr */
+	0x3c7e0103,
+	0xe84e0006,
+	0x005d7e03,
+	0x0076bb00,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
-	0x7b7e50fc,
-	0x64b60005,
-/* 0x064d: i2c_get_byte_done */
-/* 0x064f: i2c_put_byte */
-	0x0400f804,
-/* 0x0651: i2c_put_byte_next */
-	0x0142b608,
-	0xbb3854ff,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x00057b7e,
-	0xf40464b6,
-	0x46b03411,
-	0xd81bf400,
+	0x7e7e50fc,
+	0x64b60006,
+	0x1a11f404,
+	0x00066a7e,
+	0x227e0003,
+	0x884e0006,
+	0x005d7e13,
+	0x013cf000,
+/* 0x07ab: i2c_bitr_done */
+	0xf80131f4,
+/* 0x07ad: i2c_get_byte */
+	0x04000500,
+/* 0x07b1: i2c_get_byte_next */
+	0x0154b608,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x05bb7e50,
+	0x07687e50,
 	0x0464b600,
-	0xbb0f11f4,
-	0x36b00076,
-	0x061bf401,
-/* 0x06a7: i2c_put_byte_done */
-	0xf80132f4,
-/* 0x06a9: i2c_addr */
-	0x0076bb00,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0xf47e50fc,
-	0x64b60004,
-	0x2911f404,
-	0x012ec3e7,
-	0xfd0134b6,
-	0x76bb0553,
+	0xfd2a11f4,
+	0x42b60553,
+	0xd81bf401,
+	0x76bb0103,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb600064f,
-/* 0x06ee: i2c_addr_done */
+	0xb6000728,
+/* 0x07fa: i2c_get_byte_done */
 	0x00f80464,
-/* 0x06f0: i2c_acquire_addr */
-	0xb6f8cec7,
-	0xe0b705e4,
-	0x00f8d014,
-/* 0x06fc: i2c_acquire */
-	0x0006f07e,
-	0x0000047e,
-	0x7e03d9f0,
-	0xf800002e,
-/* 0x070d: i2c_release */
-	0x06f07e00,
-	0x00047e00,
-	0x03daf000,
-	0x00002e7e,
-/* 0x071e: i2c_recv */
-	0x32f400f8,
-	0xf8c1c701,
-	0xb00214b6,
-	0x1ff52816,
-	0x13b80137,
-	0x98000bd4,
-	0x13b80032,
-	0x98000bac,
-	0x31f40031,
-	0xf9d0f902,
-	0xf1d0f9e0,
-	0xf1000067,
-	0x92100063,
-	0x76bb0167,
-	0x0465b600,
-	0x659450f9,
-	0x0256bb04,
-	0x75fd50bd,
-	0x7e50fc04,
-	0xb60006fc,
-	0xd0fc0464,
-	0xf500d6b0,
-	0x0500b01b,
-	0x0076bb00,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0xa97e50fc,
-	0x64b60006,
-	0xcc11f504,
-	0xe0c5c700,
+/* 0x07fc: i2c_put_byte */
+/* 0x07fe: i2c_put_byte_next */
+	0x42b60804,
+	0x3854ff01,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x064f7e50,
+	0x07287e50,
 	0x0464b600,
-	0x00a911f5,
-	0x76bb0105,
+	0xb03411f4,
+	0x1bf40046,
+	0x0076bbd8,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x687e50fc,
+	0x64b60007,
+	0x0f11f404,
+	0xb00076bb,
+	0x1bf40136,
+	0x0132f406,
+/* 0x0854: i2c_put_byte_done */
+/* 0x0856: i2c_addr */
+	0x76bb00f8,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb60006a9,
-	0x11f50464,
-	0x76bb0087,
+	0xb60006a1,
+	0x11f40464,
+	0x2ec3e729,
+	0x0134b601,
+	0xbb0553fd,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x0007fc7e,
+/* 0x089b: i2c_addr_done */
+	0xf80464b6,
+/* 0x089d: i2c_acquire_addr */
+	0xf8cec700,
+	0xb705e4b6,
+	0xf8d014e0,
+/* 0x08a9: i2c_acquire */
+	0x089d7e00,
+	0x00047e00,
+	0x03d9f000,
+	0x00002e7e,
+/* 0x08ba: i2c_release */
+	0x9d7e00f8,
+	0x047e0008,
+	0xdaf00000,
+	0x002e7e03,
+/* 0x08cb: i2c_recv */
+	0xf400f800,
+	0xc1c70132,
+	0x0214b6f8,
+	0xf52816b0,
+	0xb801371f,
+	0x000be813,
+	0xb8003298,
+	0x000bc013,
+	0xf4003198,
+	0xd0f90231,
+	0xd0f9e0f9,
+	0x000067f1,
+	0x100063f1,
+	0xbb016792,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x0008a97e,
+	0xfc0464b6,
+	0x00d6b0d0,
+	0x00b01bf5,
+	0x76bb0005,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0x7e50fc04,
-	0xb6000600,
-	0x11f40464,
-	0xe05bcb67,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x054c7e50,
-	0x0464b600,
-	0x74bd5bb2,
-/* 0x0823: i2c_recv_not_rd08 */
-	0xb0410ef4,
-	0x1bf401d6,
-	0x7e00053b,
-	0xf40006a9,
-	0xc5c73211,
-	0x064f7ee0,
-	0x2811f400,
-	0xa97e0005,
-	0x11f40006,
-	0xe0b5c71f,
-	0x00064f7e,
-	0x7e1511f4,
-	0xbd00054c,
-	0x08c5c774,
-	0xf4091bf4,
-	0x0ef40232,
-/* 0x0861: i2c_recv_not_wr08 */
-/* 0x0861: i2c_recv_done */
-	0xf8cec703,
-	0x00070d7e,
-	0xd0fce0fc,
-	0xb20912f4,
-	0x023f7e7c,
-/* 0x0875: i2c_recv_exit */
-/* 0x0877: i2c_init */
-	0xf800f800,
-/* 0x0879: test_recv */
-	0x04584100,
-	0xb60011cf,
-	0x58400110,
-	0x0001f604,
-	0xe7f104bd,
-	0xe3f1d900,
-	0x967e134f,
-	0x00f80001,
-/* 0x0898: test_init */
-	0x7e08004e,
-	0xf8000196,
-/* 0x08a1: idle_recv */
-/* 0x08a3: idle */
-	0xf400f800,
-	0x54410031,
+	0xb6000856,
+	0x11f50464,
+	0xc5c700cc,
+	0x0076bbe0,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0xfc7e50fc,
+	0x64b60007,
+	0xa911f504,
+	0xbb010500,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x0008567e,
+	0xf50464b6,
+	0xbb008711,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x0007ad7e,
+	0xf40464b6,
+	0x5bcb6711,
+	0x0076bbe0,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0xf97e50fc,
+	0x64b60006,
+	0xbd5bb204,
+	0x410ef474,
+/* 0x09d0: i2c_recv_not_rd08 */
+	0xf401d6b0,
+	0x00053b1b,
+	0x0008567e,
+	0xc73211f4,
+	0xfc7ee0c5,
+	0x11f40007,
+	0x7e000528,
+	0xf4000856,
+	0xb5c71f11,
+	0x07fc7ee0,
+	0x1511f400,
+	0x0006f97e,
+	0xc5c774bd,
+	0x091bf408,
+	0xf40232f4,
+/* 0x0a0e: i2c_recv_not_wr08 */
+/* 0x0a0e: i2c_recv_done */
+	0xcec7030e,
+	0x08ba7ef8,
+	0xfce0fc00,
+	0x0912f4d0,
+	0xc27e7cb2,
+/* 0x0a22: i2c_recv_exit */
+	0x00f80002,
+/* 0x0a24: i2c_init */
+/* 0x0a26: test_recv */
+	0x584100f8,
 	0x0011cf04,
 	0x400110b6,
-	0x01f60454,
-/* 0x08b7: idle_loop */
-	0x0104bd00,
-	0x0232f458,
-/* 0x08bc: idle_proc */
-/* 0x08bc: idle_proc_exec */
-	0x1eb210f9,
-	0x0002487e,
-	0x11f410fc,
-	0x0231f409,
-/* 0x08cf: idle_proc_next */
-	0xb6f00ef4,
-	0x1fa65810,
-	0xf4e81bf4,
-	0x28f4e002,
-	0xc60ef400,
+	0x01f60458,
+	0xf104bd00,
+	0xf1d900e7,
+	0x7e134fe3,
+	0xf8000201,
+/* 0x0a45: test_init */
+	0x08004e00,
+	0x0002017e,
+/* 0x0a4e: idle_recv */
+	0x00f800f8,
+/* 0x0a50: idle */
+	0x410031f4,
+	0x11cf0454,
+	0x0110b600,
+	0xf6045440,
+	0x04bd0001,
+/* 0x0a64: idle_loop */
+	0x32f45801,
+/* 0x0a69: idle_proc */
+/* 0x0a69: idle_proc_exec */
+	0xb210f902,
+	0x02cb7e1e,
+	0xf410fc00,
+	0x31f40911,
+	0xf00ef402,
+/* 0x0a7c: idle_proc_next */
+	0xa65810b6,
+	0xe81bf41f,
+	0xf4e002f4,
+	0x0ef40028,
+	0x000000c6,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
index 6744fcc06151..daa06c1c655e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GT215
+#define HW_TICKS_PER_US 203 // should be 202.5
 
 //#define NVKM_FALCON_PC24
 //#define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nva3_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nva3_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
index 254205cd5166..64e97baabc3c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nva3.fuc.h
@@ -24,8 +24,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000430,
-	0x000003cd,
+	0x00000512,
+	0x000004af,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x0000054e,
-	0x00000540,
+	0x000006e0,
+	0x000006d2,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000552,
-	0x00000550,
+	0x000006e4,
+	0x000006e2,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x00000982,
-	0x00000825,
+	0x00000b14,
+	0x000009b7,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x000009ab,
-	0x00000984,
+	0x00000b3d,
+	0x00000b16,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x000009b7,
-	0x000009b5,
+	0x00000b49,
+	0x00000b47,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,27 +227,31 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x0000046f,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000496,
+	0x00000551,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x000005a8,
+	0x00000003,
 	0x00000002,
-	0x000004b7,
-	0x00040003,
+	0x0000063a,
+	0x00040004,
 	0x00000000,
-	0x000004df,
-	0x00010004,
+	0x00000656,
+	0x00010005,
 	0x00000000,
-	0x000004fc,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
+	0x00000673,
+	0x00010006,
 	0x00000000,
+	0x000005f8,
+/* 0x03b8: memx_func_tail */
+/* 0x03b8: memx_ts_start */
 	0x00000000,
+/* 0x03bc: memx_ts_end */
 	0x00000000,
+/* 0x03c0: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +761,11 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0bc0: memx_data_tail */
+/* 0x0bc0: i2c_scl_map */
 	0x00001000,
 	0x00004000,
 	0x00010000,
@@ -769,7 +776,7 @@ uint32_t nva3_pwr_data[] = {
 	0x01000000,
 	0x04000000,
 	0x10000000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0be8: i2c_sda_map */
 	0x00002000,
 	0x00008000,
 	0x00020000,
@@ -780,7 +787,7 @@ uint32_t nva3_pwr_data[] = {
 	0x02000000,
 	0x08000000,
 	0x20000000,
-/* 0x0bfc: i2c_ctrl */
+/* 0x0c10: i2c_ctrl */
 	0x0000e138,
 	0x0000e150,
 	0x0000e168,
@@ -841,15 +848,10 @@ uint32_t nva3_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
 
 uint32_t nva3_pwr_code[] = {
-	0x030d0ef5,
+	0x039e0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xd00604b6,
@@ -885,19 +887,22 @@ uint32_t nva3_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x007f: nsec */
-	0xf000f8f2,
+	0xf900f8f2,
+	0xf080f990,
 	0x84b62c87,
 	0x0088cf06,
-/* 0x0088: nsec_loop */
+/* 0x008c: nsec_loop */
 	0xb62c97f0,
 	0x99cf0694,
 	0x0298bb00,
 	0xf4069eb8,
-	0x00f8f11e,
-/* 0x009c: wait */
+	0x80fcf11e,
+	0x00f890fc,
+/* 0x00a4: wait */
+	0x80f990f9,
 	0xb62c87f0,
 	0x88cf0684,
-/* 0x00a5: wait_loop */
+/* 0x00b1: wait_loop */
 	0x02eeb900,
 	0xb90421f4,
 	0xadfd02da,
@@ -907,28 +912,29 @@ uint32_t nva3_pwr_code[] = {
 	0x0099cf06,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00c9: wait_done */
-/* 0x00cb: intr_watchdog */
-	0x9800f8df,
+/* 0x00d5: wait_done */
+	0xfc80fcdf,
+/* 0x00db: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x025421f5,
+	0x02dd21f5,
 	0x0ef494bd,
-/* 0x00e9: intr_watchdog_next_time */
+/* 0x00f9: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00f8: intr_watchdog_next_time_set */
-/* 0x00fb: intr_watchdog_next_proc */
+/* 0x0108: intr_watchdog_next_time_set */
+/* 0x010b: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x010a: intr */
+/* 0x011a: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -948,13 +954,13 @@ uint32_t nva3_pwr_code[] = {
 	0xf40289c4,
 	0x0080230b,
 	0x58e7f09b,
-	0x98cb21f4,
+	0x98db21f4,
 	0x96b09b09,
 	0x110bf400,
 	0xb63407f0,
 	0x09d00604,
 	0x8004bd00,
-/* 0x016e: intr_skip_watchdog */
+/* 0x017e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f148,
@@ -967,22 +973,22 @@ uint32_t nva3_pwr_code[] = {
 	0x48e7f1c0,
 	0x53e3f14f,
 	0x00d7f054,
-	0x02b921f5,
+	0x034221f5,
 	0x07f1c0fc,
 	0x04b604c0,
 	0x000cd006,
-/* 0x01ae: intr_subintr_skip_fifo */
+/* 0x01be: intr_subintr_skip_fifo */
 	0x07f104bd,
 	0x04b60688,
 	0x0009d006,
-/* 0x01ba: intr_skip_subintr */
+/* 0x01ca: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x01c4: intr_skip_pause */
+/* 0x01d4: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x01ce: intr_skip_user0 */
+/* 0x01de: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0604b604,
 	0xbd0008d0,
@@ -993,597 +999,733 @@ uint32_t nva3_pwr_code[] = {
 	0x90fca0fc,
 	0x00fc80fc,
 	0xf80032f4,
-/* 0x01f5: timer */
-	0x1032f401,
-	0xb003f898,
-	0x1cf40086,
-	0x03fe8051,
+/* 0x0205: ticks_from_ns */
+	0xf9c0f901,
+	0xcbd7f1b0,
+	0x00d3f000,
+	0x041321f5,
+	0x03e8ccec,
+	0xf400b4b0,
+	0xeeec120b,
+	0xd7f103e8,
+	0xd3f000cb,
+	0x1321f500,
+/* 0x022d: ticks_from_ns_quit */
+	0x02ceb904,
+	0xc0fcb0fc,
+/* 0x0236: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f000cb,
+	0x1321f500,
+	0x02ceb904,
+	0xf400b4b0,
+	0xe4bd050b,
+/* 0x0250: ticks_from_us_quit */
+	0xc0fcb0fc,
+/* 0x0256: ticks_to_us */
+	0xd7f100f8,
+	0xd3f000cb,
+	0xecedff00,
+/* 0x0262: timer */
+	0x90f900f8,
+	0x32f480f9,
+	0x03f89810,
+	0xf40086b0,
+	0x84bd651c,
 	0xb63807f0,
 	0x08d00604,
 	0xf004bd00,
-	0x84b60887,
+	0x84b63487,
 	0x0088cf06,
-	0xf40284f0,
-	0x87f0261b,
-	0x0684b634,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x0233: timer_reset */
-	0xf01f1ef4,
-	0x04b63407,
-	0x000ed006,
-	0x0e8004bd,
-/* 0x0241: timer_enable */
-	0x0187f09a,
+	0xbb9a0998,
+	0xe9bb0298,
+	0x03fe8000,
+	0xb60887f0,
+	0x88cf0684,
+	0x0284f000,
+	0xf0261bf4,
+	0x84b63487,
+	0x0088cf06,
+	0xf406e0b8,
+	0xe8b8090b,
+	0x111cf406,
+/* 0x02b8: timer_reset */
+	0xb63407f0,
+	0x0ed00604,
+	0x8004bd00,
+/* 0x02c6: timer_enable */
+	0x87f09a0e,
+	0x3807f001,
+	0xd00604b6,
+	0x04bd0008,
+/* 0x02d4: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x02dd: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89b80486,
+	0x2a0bf406,
+	0x940398c4,
+	0x80b60488,
+	0x008ebb18,
+	0x8000fa98,
+	0x8d80008a,
+	0x028c8001,
+	0xb6038b80,
+	0x94f00190,
+	0x04e98007,
+/* 0x0317: send_done */
+	0xfc0231f4,
+	0xf880fc90,
+/* 0x031d: find */
+	0xf080f900,
+	0x31f45887,
+/* 0x0325: find_loop */
+	0x008a9801,
+	0xf406aeb8,
+	0x80b6100b,
+	0x6886b158,
+	0xf01bf402,
+/* 0x033b: find_done */
+	0xb90132f4,
+	0x80fc028e,
+/* 0x0342: send */
+	0x21f500f8,
+	0x01f4031d,
+/* 0x034b: recv */
+	0xf900f897,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0xf40689b8,
+	0x89c43d0b,
+	0x0180b603,
+	0x800784f0,
+	0xea9805e8,
+	0xfef0f902,
+	0xf0f9018f,
+	0x9402efb9,
+	0xe9bb0499,
+	0x18e0b600,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0xf0fca5f9,
+	0xf400f8fe,
+	0xf0fc0131,
+/* 0x0398: recv_done */
+	0x90fc80fc,
+/* 0x039e: init */
+	0x17f100f8,
+	0x14b60108,
+	0x0011cf06,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xd00604b6,
+	0x04bd0001,
+	0xf0ff17f0,
+	0x04b61407,
+	0x0001d006,
+	0x17f004bd,
+	0x0015f102,
+	0x1007f008,
+	0xd00604b6,
+	0x04bd0001,
+	0x011a17f1,
+	0xfe0013f0,
+	0x31f40010,
+	0x0117f010,
 	0xb63807f0,
-	0x08d00604,
-/* 0x024f: timer_done */
-	0xf404bd00,
-	0x00f81031,
-/* 0x0254: send_proc */
-	0x90f980f9,
-	0x9805e898,
-	0x86f004e9,
-	0x0689b804,
-	0xc42a0bf4,
-	0x88940398,
-	0x1880b604,
-	0x98008ebb,
-	0x8a8000fa,
-	0x018d8000,
-	0x80028c80,
-	0x90b6038b,
-	0x0794f001,
-	0xf404e980,
-/* 0x028e: send_done */
-	0x90fc0231,
-	0x00f880fc,
-/* 0x0294: find */
-	0x87f080f9,
-	0x0131f458,
-/* 0x029c: find_loop */
-	0xb8008a98,
-	0x0bf406ae,
-	0x5880b610,
-	0x026886b1,
-	0xf4f01bf4,
-/* 0x02b2: find_done */
-	0x8eb90132,
-	0xf880fc02,
-/* 0x02b9: send */
-	0x9421f500,
-	0x9701f402,
-/* 0x02c2: recv */
-	0xe89800f8,
-	0x04e99805,
-	0xb80132f4,
-	0x0bf40689,
-	0x0389c43d,
-	0xf00180b6,
-	0xe8800784,
-	0x02ea9805,
-	0x8ffef0f9,
-	0xb9f0f901,
-	0x999402ef,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x030b: recv_done */
-	0xf8f0fc01,
-/* 0x030d: init */
-	0x0817f100,
-	0x0614b601,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0xe017f100,
-	0x0013f000,
-	0xb61c07f0,
 	0x01d00604,
 	0xf004bd00,
-	0x07f0ff17,
-	0x0604b614,
-	0xbd0001d0,
-	0x0217f004,
-	0x080015f1,
-	0xb61007f0,
-	0x01d00604,
-	0xf104bd00,
-	0xf0010a17,
-	0x10fe0013,
-	0x1031f400,
-	0xf00117f0,
-	0x04b63807,
-	0x0001d006,
-	0xf7f004bd,
-/* 0x0371: init_proc */
-	0x01f19858,
-	0xf40016b0,
-	0x15f9fa0b,
-	0xf458f0b6,
-/* 0x0382: host_send */
-	0x17f1f20e,
-	0x14b604b0,
-	0x0011cf06,
-	0x04a027f1,
-	0xcf0624b6,
-	0x12b80022,
-	0x320bf406,
-	0x94071ec4,
-	0xe0b704ee,
-	0xeb980270,
-	0x02ec9803,
-	0x9801ed98,
-	0x21f500ee,
-	0x10b602b9,
-	0x0f1ec401,
-	0x04b007f1,
-	0xd00604b6,
-	0x04bd000e,
-/* 0x03cb: host_send_done */
-	0xf8ba0ef4,
-/* 0x03cd: host_recv */
-	0x4917f100,
-	0x5413f14e,
-	0x06e1b852,
-/* 0x03db: host_recv_wait */
-	0xf1aa0bf4,
-	0xb604cc17,
-	0x11cf0614,
-	0xc827f100,
-	0x0624b604,
-	0xf00022cf,
-	0x12b80816,
-	0xe60bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
+/* 0x0402: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0413: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0464: host_send */
+	0x04b017f1,
+	0xcf0614b6,
+	0x27f10011,
+	0x24b604a0,
+	0x0022cf06,
+	0xf40612b8,
+	0x1ec4320b,
+	0x04ee9407,
+	0x0270e0b7,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0x034221f5,
+	0xc40110b6,
+	0x07f10f1e,
+	0x04b604b0,
+	0x000ed006,
+	0x0ef404bd,
+/* 0x04ad: host_send_done */
+/* 0x04af: host_recv */
+	0xf100f8ba,
+	0xf14e4917,
+	0xb8525413,
+	0x0bf406e1,
+/* 0x04bd: host_recv_wait */
+	0xcc17f1aa,
+	0x0614b604,
+	0xf10011cf,
+	0xb604c827,
+	0x22cf0624,
+	0x0816f000,
+	0xf40612b8,
+	0x23c4e60b,
+	0x0434b607,
+	0x02f030b7,
+	0x80033b80,
+	0x3d80023c,
+	0x003e8001,
+	0xf00120b6,
+	0x07f10f24,
+	0x04b604c8,
+	0x0002d006,
+	0x27f004bd,
+	0x0007f040,
 	0xd00604b6,
 	0x04bd0002,
-	0xf04027f0,
-	0x04b60007,
-	0x0002d006,
-	0x00f804bd,
-/* 0x0430: host_init */
-	0x008017f1,
-	0xf11014b6,
-	0xf1027015,
-	0xb604d007,
-	0x01d00604,
-	0xf104bd00,
-	0xb6008017,
-	0x15f11014,
-	0x07f102f0,
-	0x04b604dc,
-	0x0001d006,
-	0x17f004bd,
-	0xc407f101,
+/* 0x0512: host_init */
+	0x17f100f8,
+	0x14b60080,
+	0x7015f110,
+	0xd007f102,
 	0x0604b604,
 	0xbd0001d0,
-/* 0x046f: memx_func_enter */
-	0xf000f804,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xd00604b6,
+	0x04bd0001,
+	0xf10117f0,
+	0xb604c407,
+	0x01d00604,
+	0xf804bd00,
+/* 0x0551: memx_func_enter */
+	0x1087f100,
+	0x028eb916,
+	0xb90421f4,
+	0x67f102d7,
+	0x63f1fffc,
+	0x76fdffff,
+	0x0267f104,
+	0x0576fd00,
+	0x70f980f9,
+	0xe0fcd0fc,
+	0xf03f21f4,
 	0x07f10467,
 	0x04b607e0,
 	0x0006d006,
-/* 0x047e: memx_func_enter_wait */
+/* 0x058a: memx_func_enter_wait */
 	0x67f104bd,
 	0x64b607c0,
 	0x0066cf06,
 	0xf40464f0,
-	0x1698f30b,
-	0x0410b600,
-/* 0x0496: memx_func_leave */
-	0x67f000f8,
-	0xe407f104,
-	0x0604b607,
-	0xbd0006d0,
-/* 0x04a5: memx_func_leave_wait */
-	0xc067f104,
+	0x67f0f30b,
+	0x0664b62c,
+	0x800066cf,
+	0x00f8ee06,
+/* 0x05a8: memx_func_leave */
+	0xb62c67f0,
+	0x66cf0664,
+	0xef068000,
+	0xf10467f0,
+	0xb607e407,
+	0x06d00604,
+/* 0x05c3: memx_func_leave_wait */
+	0xf104bd00,
+	0xb607c067,
+	0x66cf0664,
+	0x0464f000,
+	0xf1f31bf4,
+	0xb9161087,
+	0x21f4028e,
+	0x02d7b904,
+	0xffcc67f1,
+	0xffff63f1,
+	0xf90476fd,
+	0xfc70f980,
+	0xf4e0fcd0,
+	0x00f83f21,
+/* 0x05f8: memx_func_wait_vblank */
+	0xb0001698,
+	0x0bf40066,
+	0x0166b013,
+	0xf4060bf4,
+/* 0x060a: memx_func_wait_vblank_head1 */
+	0x77f12e0e,
+	0x0ef40020,
+/* 0x0611: memx_func_wait_vblank_head0 */
+	0x0877f107,
+/* 0x0615: memx_func_wait_vblank_0 */
+	0xc467f100,
 	0x0664b607,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x04b7: memx_func_wr32 */
-	0x9800f8f3,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
-	0xe0fcd0fc,
-	0xf13f21f4,
-	0xfd140003,
-	0x05800506,
-	0xb604bd00,
-	0x1bf40242,
-/* 0x04df: memx_func_wait */
-	0xf000f8dd,
-	0x84b62c87,
-	0x0088cf06,
-	0x98001e98,
-	0x1c98011d,
-	0x031b9802,
-	0xf41010b6,
-	0x00f89c21,
-/* 0x04fc: memx_func_delay */
-	0xb6001e98,
-	0x21f40410,
-/* 0x0507: memx_exec */
-	0xf900f87f,
-	0xb9d0f9e0,
-	0xb2b902c1,
-/* 0x0511: memx_exec_next */
-	0x00139802,
-	0x950410b6,
-	0x30f01034,
-	0xde35980c,
-	0x12b855f9,
-	0xec1ef406,
-	0xe0fcd0fc,
-	0x02b921f5,
-/* 0x0532: memx_info */
-	0xc7f100f8,
-	0xb7f103ac,
-	0x21f50800,
-	0x00f802b9,
-/* 0x0540: memx_recv */
-	0xf401d6b0,
-	0xd6b0c40b,
-	0xe90bf400,
-/* 0x054e: memx_init */
-	0x00f800f8,
-/* 0x0550: perf_recv */
-/* 0x0552: perf_init */
+	0xfd0066cf,
+	0x1bf40467,
+/* 0x0625: memx_func_wait_vblank_1 */
+	0xc467f1f3,
+	0x0664b607,
+	0xfd0066cf,
+	0x0bf40467,
+/* 0x0635: memx_func_wait_vblank_fini */
+	0x0410b6f3,
+/* 0x063a: memx_func_wr32 */
+	0x169800f8,
+	0x01159800,
+	0xf90810b6,
+	0xfc50f960,
+	0xf4e0fcd0,
+	0x42b63f21,
+	0xe91bf402,
+/* 0x0656: memx_func_wait */
+	0x87f000f8,
+	0x0684b62c,
+	0x980088cf,
+	0x1d98001e,
+	0x021c9801,
+	0xb6031b98,
+	0x21f41010,
+/* 0x0673: memx_func_delay */
+	0x9800f8a4,
+	0x10b6001e,
+	0x7f21f404,
+/* 0x067e: memx_exec */
+	0xe0f900f8,
+	0xc1b9d0f9,
+	0x02b2b902,
+/* 0x0688: memx_exec_next */
+	0xb6001398,
+	0x34e70410,
+	0x33e701f0,
+	0x32b601e0,
+	0x0c30f001,
+	0xf9de3598,
+	0x0612b855,
+	0x98e41ef4,
+	0x0c98ee0b,
+	0x02cbbbef,
+	0x07c4b7f1,
+	0xcf06b4b6,
+	0xd0fc00bb,
+	0x21f5e0fc,
+	0x00f80342,
+/* 0x06c4: memx_info */
+	0x03c0c7f1,
+	0x0800b7f1,
+	0x034221f5,
+/* 0x06d2: memx_recv */
+	0xd6b000f8,
+	0xa90bf401,
+	0xf400d6b0,
+	0x00f8e90b,
+/* 0x06e0: memx_init */
+/* 0x06e2: perf_recv */
 	0x00f800f8,
-/* 0x0554: i2c_drive_scl */
-	0xf40036b0,
-	0x07f1110b,
-	0x04b607e0,
-	0x0001d006,
-	0x00f804bd,
-/* 0x0568: i2c_drive_scl_lo */
-	0x07e407f1,
-	0xd00604b6,
-	0x04bd0001,
-/* 0x0576: i2c_drive_sda */
+/* 0x06e4: perf_init */
+/* 0x06e6: i2c_drive_scl */
 	0x36b000f8,
 	0x110bf400,
 	0x07e007f1,
 	0xd00604b6,
-	0x04bd0002,
-/* 0x058a: i2c_drive_sda_lo */
+	0x04bd0001,
+/* 0x06fa: i2c_drive_scl_lo */
 	0x07f100f8,
 	0x04b607e4,
+	0x0001d006,
+	0x00f804bd,
+/* 0x0708: i2c_drive_sda */
+	0xf40036b0,
+	0x07f1110b,
+	0x04b607e0,
 	0x0002d006,
 	0x00f804bd,
-/* 0x0598: i2c_sense_scl */
-	0xf10132f4,
-	0xb607c437,
-	0x33cf0634,
-	0x0431fd00,
-	0xf4060bf4,
-/* 0x05ae: i2c_sense_scl_done */
-	0x00f80131,
-/* 0x05b0: i2c_sense_sda */
-	0xf10132f4,
-	0xb607c437,
-	0x33cf0634,
-	0x0432fd00,
-	0xf4060bf4,
-/* 0x05c6: i2c_sense_sda_done */
-	0x00f80131,
-/* 0x05c8: i2c_raise_scl */
-	0x47f140f9,
-	0x37f00898,
-	0x5421f501,
-/* 0x05d5: i2c_raise_scl_wait */
-	0xe8e7f105,
-	0x7f21f403,
-	0x059821f5,
-	0xb60901f4,
-	0x1bf40142,
-/* 0x05e9: i2c_raise_scl_done */
-	0xf840fcef,
-/* 0x05ed: i2c_start */
-	0x9821f500,
-	0x0d11f405,
-	0x05b021f5,
-	0xf40611f4,
-/* 0x05fe: i2c_start_rep */
-	0x37f0300e,
-	0x5421f500,
-	0x0137f005,
-	0x057621f5,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0xc821f550,
-	0x0464b605,
-/* 0x062b: i2c_start_send */
-	0xf01f11f4,
+/* 0x071c: i2c_drive_sda_lo */
+	0x07e407f1,
+	0xd00604b6,
+	0x04bd0002,
+/* 0x072a: i2c_sense_scl */
+	0x32f400f8,
+	0xc437f101,
+	0x0634b607,
+	0xfd0033cf,
+	0x0bf40431,
+	0x0131f406,
+/* 0x0740: i2c_sense_scl_done */
+/* 0x0742: i2c_sense_sda */
+	0x32f400f8,
+	0xc437f101,
+	0x0634b607,
+	0xfd0033cf,
+	0x0bf40432,
+	0x0131f406,
+/* 0x0758: i2c_sense_sda_done */
+/* 0x075a: i2c_raise_scl */
+	0x40f900f8,
+	0x089847f1,
+	0xf50137f0,
+/* 0x0767: i2c_raise_scl_wait */
+	0xf106e621,
+	0xf403e8e7,
+	0x21f57f21,
+	0x01f4072a,
+	0x0142b609,
+/* 0x077b: i2c_raise_scl_done */
+	0xfcef1bf4,
+/* 0x077f: i2c_start */
+	0xf500f840,
+	0xf4072a21,
+	0x21f50d11,
+	0x11f40742,
+	0x300ef406,
+/* 0x0790: i2c_start_rep */
+	0xf50037f0,
+	0xf006e621,
+	0x21f50137,
+	0x76bb0708,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb6075a21,
+	0x11f40464,
+/* 0x07bd: i2c_start_send */
+	0x0037f01f,
+	0x070821f5,
+	0x1388e7f1,
+	0xf07f21f4,
 	0x21f50037,
-	0xe7f10576,
+	0xe7f106e6,
 	0x21f41388,
-	0x0037f07f,
-	0x055421f5,
-	0x1388e7f1,
-/* 0x0647: i2c_start_out */
-	0xf87f21f4,
-/* 0x0649: i2c_stop */
-	0x0037f000,
-	0x055421f5,
-	0xf50037f0,
-	0xf1057621,
-	0xf403e8e7,
-	0x37f07f21,
-	0x5421f501,
-	0x88e7f105,
-	0x7f21f413,
+/* 0x07d9: i2c_start_out */
+/* 0x07db: i2c_stop */
+	0xf000f87f,
+	0x21f50037,
+	0x37f006e6,
+	0x0821f500,
+	0xe8e7f107,
+	0x7f21f403,
 	0xf50137f0,
-	0xf1057621,
+	0xf106e621,
 	0xf41388e7,
-	0x00f87f21,
-/* 0x067c: i2c_bitw */
-	0x057621f5,
-	0x03e8e7f1,
-	0xbb7f21f4,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x05c821f5,
-	0xf40464b6,
-	0xe7f11811,
-	0x21f41388,
-	0x0037f07f,
-	0x055421f5,
-	0x1388e7f1,
-/* 0x06bb: i2c_bitw_out */
-	0xf87f21f4,
-/* 0x06bd: i2c_bitr */
-	0x0137f000,
-	0x057621f5,
-	0x03e8e7f1,
-	0xbb7f21f4,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x05c821f5,
-	0xf40464b6,
-	0x21f51b11,
-	0x37f005b0,
-	0x5421f500,
-	0x88e7f105,
+	0x37f07f21,
+	0x0821f501,
+	0x88e7f107,
 	0x7f21f413,
-	0xf4013cf0,
-/* 0x0702: i2c_bitr_done */
-	0x00f80131,
-/* 0x0704: i2c_get_byte */
-	0xf00057f0,
-/* 0x070a: i2c_get_byte_next */
-	0x54b60847,
-	0x0076bb01,
+/* 0x080e: i2c_bitw */
+	0x21f500f8,
+	0xe7f10708,
+	0x21f403e8,
+	0x0076bb7f,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606bd,
-	0x2b11f404,
-	0xb60553fd,
-	0x1bf40142,
-	0x0137f0d8,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x7c21f550,
-	0x0464b606,
-/* 0x0754: i2c_get_byte_done */
-/* 0x0756: i2c_put_byte */
-	0x47f000f8,
-/* 0x0759: i2c_put_byte_next */
-	0x0142b608,
-	0xbb3854ff,
+	0x64b6075a,
+	0x1811f404,
+	0x1388e7f1,
+	0xf07f21f4,
+	0x21f50037,
+	0xe7f106e6,
+	0x21f41388,
+/* 0x084d: i2c_bitw_out */
+/* 0x084f: i2c_bitr */
+	0xf000f87f,
+	0x21f50137,
+	0xe7f10708,
+	0x21f403e8,
+	0x0076bb7f,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x21f550fc,
+	0x64b6075a,
+	0x1b11f404,
+	0x074221f5,
+	0xf50037f0,
+	0xf106e621,
+	0xf41388e7,
+	0x3cf07f21,
+	0x0131f401,
+/* 0x0894: i2c_bitr_done */
+/* 0x0896: i2c_get_byte */
+	0x57f000f8,
+	0x0847f000,
+/* 0x089c: i2c_get_byte_next */
+	0xbb0154b6,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x067c21f5,
+	0x084f21f5,
 	0xf40464b6,
-	0x46b03411,
-	0xd81bf400,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0xbd21f550,
-	0x0464b606,
-	0xbb0f11f4,
-	0x36b00076,
-	0x061bf401,
-/* 0x07af: i2c_put_byte_done */
-	0xf80132f4,
-/* 0x07b1: i2c_addr */
-	0x0076bb00,
+	0x53fd2b11,
+	0x0142b605,
+	0xf0d81bf4,
+	0x76bb0137,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb6080e21,
+/* 0x08e6: i2c_get_byte_done */
+	0x00f80464,
+/* 0x08e8: i2c_put_byte */
+/* 0x08eb: i2c_put_byte_next */
+	0xb60847f0,
+	0x54ff0142,
+	0x0076bb38,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b605ed,
-	0x2911f404,
-	0x012ec3e7,
-	0xfd0134b6,
-	0x76bb0553,
+	0x64b6080e,
+	0x3411f404,
+	0xf40046b0,
+	0x76bbd81b,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6075621,
-/* 0x07f6: i2c_addr_done */
-	0x00f80464,
-/* 0x07f8: i2c_acquire_addr */
-	0xb6f8cec7,
-	0xe0b702e4,
-	0xee980bfc,
-/* 0x0807: i2c_acquire */
-	0xf500f800,
-	0xf407f821,
-	0xd9f00421,
-	0x3f21f403,
-/* 0x0816: i2c_release */
-	0x21f500f8,
-	0x21f407f8,
-	0x03daf004,
-	0xf83f21f4,
-/* 0x0825: i2c_recv */
-	0x0132f400,
-	0xb6f8c1c7,
-	0x16b00214,
-	0x3a1ff528,
-	0xd413a001,
-	0x0032980b,
-	0x0bac13a0,
-	0xf4003198,
-	0xd0f90231,
-	0xd0f9e0f9,
-	0x000067f1,
-	0x100063f1,
-	0xbb016792,
+	0xb6084f21,
+	0x11f40464,
+	0x0076bb0f,
+	0xf40136b0,
+	0x32f4061b,
+/* 0x0941: i2c_put_byte_done */
+/* 0x0943: i2c_addr */
+	0xbb00f801,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x080721f5,
-	0xfc0464b6,
-	0x00d6b0d0,
-	0x00b31bf5,
-	0xbb0057f0,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x07b121f5,
-	0xf50464b6,
-	0xc700d011,
-	0x76bbe0c5,
-	0x0465b600,
-	0x659450f9,
-	0x0256bb04,
-	0x75fd50bd,
-	0xf550fc04,
-	0xb6075621,
-	0x11f50464,
-	0x57f000ad,
+	0x077f21f5,
+	0xf40464b6,
+	0xc3e72911,
+	0x34b6012e,
+	0x0553fd01,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0xe821f550,
+	0x0464b608,
+/* 0x0988: i2c_addr_done */
+/* 0x098a: i2c_acquire_addr */
+	0xcec700f8,
+	0x02e4b6f8,
+	0x0c10e0b7,
+	0xf800ee98,
+/* 0x0999: i2c_acquire */
+	0x8a21f500,
+	0x0421f409,
+	0xf403d9f0,
+	0x00f83f21,
+/* 0x09a8: i2c_release */
+	0x098a21f5,
+	0xf00421f4,
+	0x21f403da,
+/* 0x09b7: i2c_recv */
+	0xf400f83f,
+	0xc1c70132,
+	0x0214b6f8,
+	0xf52816b0,
+	0xa0013a1f,
+	0x980be813,
+	0x13a00032,
+	0x31980bc0,
+	0x0231f400,
+	0xe0f9d0f9,
+	0x67f1d0f9,
+	0x63f10000,
+	0x67921000,
 	0x0076bb01,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b607b1,
-	0x8a11f504,
+	0x64b60999,
+	0xb0d0fc04,
+	0x1bf500d6,
+	0x57f000b3,
 	0x0076bb00,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b60704,
-	0x6a11f404,
-	0xbbe05bcb,
+	0x64b60943,
+	0xd011f504,
+	0xe0c5c700,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0xe821f550,
+	0x0464b608,
+	0x00ad11f5,
+	0xbb0157f0,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x064921f5,
-	0xb90464b6,
-	0x74bd025b,
-/* 0x092b: i2c_recv_not_rd08 */
-	0xb0430ef4,
-	0x1bf401d6,
-	0x0057f03d,
-	0x07b121f5,
-	0xc73311f4,
-	0x21f5e0c5,
-	0x11f40756,
-	0x0057f029,
-	0x07b121f5,
-	0xc71f11f4,
-	0x21f5e0b5,
-	0x11f40756,
-	0x4921f515,
-	0xc774bd06,
-	0x1bf408c5,
-	0x0232f409,
-/* 0x096b: i2c_recv_not_wr08 */
-/* 0x096b: i2c_recv_done */
-	0xc7030ef4,
-	0x21f5f8ce,
-	0xe0fc0816,
-	0x12f4d0fc,
-	0x027cb90a,
-	0x02b921f5,
-/* 0x0980: i2c_recv_exit */
-/* 0x0982: i2c_init */
-	0x00f800f8,
-/* 0x0984: test_recv */
-	0x05d817f1,
+	0x094321f5,
+	0xf50464b6,
+	0xbb008a11,
+	0x65b60076,
+	0x9450f904,
+	0x56bb0465,
+	0xfd50bd02,
+	0x50fc0475,
+	0x089621f5,
+	0xf40464b6,
+	0x5bcb6a11,
+	0x0076bbe0,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x21f550fc,
+	0x64b607db,
+	0x025bb904,
+	0x0ef474bd,
+/* 0x0abd: i2c_recv_not_rd08 */
+	0x01d6b043,
+	0xf03d1bf4,
+	0x21f50057,
+	0x11f40943,
+	0xe0c5c733,
+	0x08e821f5,
+	0xf02911f4,
+	0x21f50057,
+	0x11f40943,
+	0xe0b5c71f,
+	0x08e821f5,
+	0xf51511f4,
+	0xbd07db21,
+	0x08c5c774,
+	0xf4091bf4,
+	0x0ef40232,
+/* 0x0afd: i2c_recv_not_wr08 */
+/* 0x0afd: i2c_recv_done */
+	0xf8cec703,
+	0x09a821f5,
+	0xd0fce0fc,
+	0xb90a12f4,
+	0x21f5027c,
+/* 0x0b12: i2c_recv_exit */
+	0x00f80342,
+/* 0x0b14: i2c_init */
+/* 0x0b16: test_recv */
+	0x17f100f8,
+	0x14b605d8,
+	0x0011cf06,
+	0xf10110b6,
+	0xb605d807,
+	0x01d00604,
+	0xf104bd00,
+	0xf1d900e7,
+	0xf5134fe3,
+	0xf8026221,
+/* 0x0b3d: test_init */
+	0x00e7f100,
+	0x6221f508,
+/* 0x0b47: idle_recv */
+	0xf800f802,
+/* 0x0b49: idle */
+	0x0031f400,
+	0x05d417f1,
 	0xcf0614b6,
 	0x10b60011,
-	0xd807f101,
+	0xd407f101,
 	0x0604b605,
 	0xbd0001d0,
-	0x00e7f104,
-	0x4fe3f1d9,
-	0xf521f513,
-/* 0x09ab: test_init */
-	0xf100f801,
-	0xf50800e7,
-	0xf801f521,
-/* 0x09b5: idle_recv */
-/* 0x09b7: idle */
-	0xf400f800,
-	0x17f10031,
-	0x14b605d4,
-	0x0011cf06,
-	0xf10110b6,
-	0xb605d407,
-	0x01d00604,
-/* 0x09d3: idle_loop */
-	0xf004bd00,
-	0x32f45817,
-/* 0x09d9: idle_proc */
-/* 0x09d9: idle_proc_exec */
-	0xb910f902,
-	0x21f5021e,
-	0x10fc02c2,
-	0xf40911f4,
-	0x0ef40231,
-/* 0x09ed: idle_proc_next */
-	0x5810b6ef,
-	0xf4061fb8,
-	0x02f4e61b,
-	0x0028f4dd,
-	0x00bb0ef4,
+/* 0x0b65: idle_loop */
+	0x5817f004,
+/* 0x0b6b: idle_proc */
+/* 0x0b6b: idle_proc_exec */
+	0xf90232f4,
+	0x021eb910,
+	0x034b21f5,
+	0x11f410fc,
+	0x0231f409,
+/* 0x0b7f: idle_proc_next */
+	0xb6ef0ef4,
+	0x1fb85810,
+	0xe61bf406,
+	0xf4dd02f4,
+	0x0ef40028,
+	0x000000bb,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
index 48f79434a449..21bf8cc7618f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GF100
+#define HW_TICKS_PER_US 203 // should be 202.5
 
 //#define NVKM_FALCON_PC24
 //#define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nvc0_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nvc0_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
index 7ac87405d01b..ca30fa4011b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvc0.fuc.h
@@ -24,8 +24,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x00000430,
-	0x000003cd,
+	0x00000512,
+	0x000004af,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x0000054e,
-	0x00000540,
+	0x0000074b,
+	0x0000073d,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x00000552,
-	0x00000550,
+	0x0000074f,
+	0x0000074d,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x00000982,
-	0x00000825,
+	0x00000b7f,
+	0x00000a22,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x000009ab,
-	0x00000984,
+	0x00000ba8,
+	0x00000b81,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x000009b7,
-	0x000009b5,
+	0x00000bb4,
+	0x00000bb2,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,25 +227,31 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x0000046f,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000496,
+	0x00000551,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x000005db,
+	0x00000003,
 	0x00000002,
-	0x000004b7,
-	0x00040003,
+	0x000006a5,
+	0x00040004,
+	0x00000000,
+	0x000006c1,
+	0x00010005,
+	0x00000000,
+	0x000006de,
+	0x00010006,
 	0x00000000,
-	0x000004df,
-	0x00010004,
+	0x00000663,
+/* 0x03b8: memx_func_tail */
+/* 0x03b8: memx_ts_start */
 	0x00000000,
-	0x000004fc,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
+/* 0x03bc: memx_ts_end */
 	0x00000000,
+/* 0x03c0: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +763,9 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+	0x00000000,
+/* 0x0bc0: memx_data_tail */
+/* 0x0bc0: i2c_scl_map */
 	0x00001000,
 	0x00004000,
 	0x00010000,
@@ -769,7 +776,7 @@ uint32_t nvc0_pwr_data[] = {
 	0x01000000,
 	0x04000000,
 	0x10000000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0be8: i2c_sda_map */
 	0x00002000,
 	0x00008000,
 	0x00020000,
@@ -780,7 +787,7 @@ uint32_t nvc0_pwr_data[] = {
 	0x02000000,
 	0x08000000,
 	0x20000000,
-/* 0x0bfc: i2c_ctrl */
+/* 0x0c10: i2c_ctrl */
 	0x0000e138,
 	0x0000e150,
 	0x0000e168,
@@ -841,15 +848,10 @@ uint32_t nvc0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
 
 uint32_t nvc0_pwr_code[] = {
-	0x030d0ef5,
+	0x039e0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xd00604b6,
@@ -885,19 +887,22 @@ uint32_t nvc0_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x007f: nsec */
-	0xf000f8f2,
+	0xf900f8f2,
+	0xf080f990,
 	0x84b62c87,
 	0x0088cf06,
-/* 0x0088: nsec_loop */
+/* 0x008c: nsec_loop */
 	0xb62c97f0,
 	0x99cf0694,
 	0x0298bb00,
 	0xf4069eb8,
-	0x00f8f11e,
-/* 0x009c: wait */
+	0x80fcf11e,
+	0x00f890fc,
+/* 0x00a4: wait */
+	0x80f990f9,
 	0xb62c87f0,
 	0x88cf0684,
-/* 0x00a5: wait_loop */
+/* 0x00b1: wait_loop */
 	0x02eeb900,
 	0xb90421f4,
 	0xadfd02da,
@@ -907,28 +912,29 @@ uint32_t nvc0_pwr_code[] = {
 	0x0099cf06,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00c9: wait_done */
-/* 0x00cb: intr_watchdog */
-	0x9800f8df,
+/* 0x00d5: wait_done */
+	0xfc80fcdf,
+/* 0x00db: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x025421f5,
+	0x02dd21f5,
 	0x0ef494bd,
-/* 0x00e9: intr_watchdog_next_time */
+/* 0x00f9: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00f8: intr_watchdog_next_time_set */
-/* 0x00fb: intr_watchdog_next_proc */
+/* 0x0108: intr_watchdog_next_time_set */
+/* 0x010b: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x010a: intr */
+/* 0x011a: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -948,13 +954,13 @@ uint32_t nvc0_pwr_code[] = {
 	0xf40289c4,
 	0x0080230b,
 	0x58e7f09b,
-	0x98cb21f4,
+	0x98db21f4,
 	0x96b09b09,
 	0x110bf400,
 	0xb63407f0,
 	0x09d00604,
 	0x8004bd00,
-/* 0x016e: intr_skip_watchdog */
+/* 0x017e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f148,
@@ -967,22 +973,22 @@ uint32_t nvc0_pwr_code[] = {
 	0x48e7f1c0,
 	0x53e3f14f,
 	0x00d7f054,
-	0x02b921f5,
+	0x034221f5,
 	0x07f1c0fc,
 	0x04b604c0,
 	0x000cd006,
-/* 0x01ae: intr_subintr_skip_fifo */
+/* 0x01be: intr_subintr_skip_fifo */
 	0x07f104bd,
 	0x04b60688,
 	0x0009d006,
-/* 0x01ba: intr_skip_subintr */
+/* 0x01ca: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x01c4: intr_skip_pause */
+/* 0x01d4: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x01ce: intr_skip_user0 */
+/* 0x01de: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0604b604,
 	0xbd0008d0,
@@ -993,597 +999,733 @@ uint32_t nvc0_pwr_code[] = {
 	0x90fca0fc,
 	0x00fc80fc,
 	0xf80032f4,
-/* 0x01f5: timer */
-	0x1032f401,
-	0xb003f898,
-	0x1cf40086,
-	0x03fe8051,
+/* 0x0205: ticks_from_ns */
+	0xf9c0f901,
+	0xcbd7f1b0,
+	0x00d3f000,
+	0x041321f5,
+	0x03e8ccec,
+	0xf400b4b0,
+	0xeeec120b,
+	0xd7f103e8,
+	0xd3f000cb,
+	0x1321f500,
+/* 0x022d: ticks_from_ns_quit */
+	0x02ceb904,
+	0xc0fcb0fc,
+/* 0x0236: ticks_from_us */
+	0xc0f900f8,
+	0xd7f1b0f9,
+	0xd3f000cb,
+	0x1321f500,
+	0x02ceb904,
+	0xf400b4b0,
+	0xe4bd050b,
+/* 0x0250: ticks_from_us_quit */
+	0xc0fcb0fc,
+/* 0x0256: ticks_to_us */
+	0xd7f100f8,
+	0xd3f000cb,
+	0xecedff00,
+/* 0x0262: timer */
+	0x90f900f8,
+	0x32f480f9,
+	0x03f89810,
+	0xf40086b0,
+	0x84bd651c,
 	0xb63807f0,
 	0x08d00604,
 	0xf004bd00,
-	0x84b60887,
+	0x84b63487,
 	0x0088cf06,
-	0xf40284f0,
-	0x87f0261b,
-	0x0684b634,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x0233: timer_reset */
-	0xf01f1ef4,
-	0x04b63407,
-	0x000ed006,
-	0x0e8004bd,
-/* 0x0241: timer_enable */
-	0x0187f09a,
+	0xbb9a0998,
+	0xe9bb0298,
+	0x03fe8000,
+	0xb60887f0,
+	0x88cf0684,
+	0x0284f000,
+	0xf0261bf4,
+	0x84b63487,
+	0x0088cf06,
+	0xf406e0b8,
+	0xe8b8090b,
+	0x111cf406,
+/* 0x02b8: timer_reset */
+	0xb63407f0,
+	0x0ed00604,
+	0x8004bd00,
+/* 0x02c6: timer_enable */
+	0x87f09a0e,
+	0x3807f001,
+	0xd00604b6,
+	0x04bd0008,
+/* 0x02d4: timer_done */
+	0xfc1031f4,
+	0xf890fc80,
+/* 0x02dd: send_proc */
+	0xf980f900,
+	0x05e89890,
+	0xf004e998,
+	0x89b80486,
+	0x2a0bf406,
+	0x940398c4,
+	0x80b60488,
+	0x008ebb18,
+	0x8000fa98,
+	0x8d80008a,
+	0x028c8001,
+	0xb6038b80,
+	0x94f00190,
+	0x04e98007,
+/* 0x0317: send_done */
+	0xfc0231f4,
+	0xf880fc90,
+/* 0x031d: find */
+	0xf080f900,
+	0x31f45887,
+/* 0x0325: find_loop */
+	0x008a9801,
+	0xf406aeb8,
+	0x80b6100b,
+	0x6886b158,
+	0xf01bf402,
+/* 0x033b: find_done */
+	0xb90132f4,
+	0x80fc028e,
+/* 0x0342: send */
+	0x21f500f8,
+	0x01f4031d,
+/* 0x034b: recv */
+	0xf900f897,
+	0x9880f990,
+	0xe99805e8,
+	0x0132f404,
+	0xf40689b8,
+	0x89c43d0b,
+	0x0180b603,
+	0x800784f0,
+	0xea9805e8,
+	0xfef0f902,
+	0xf0f9018f,
+	0x9402efb9,
+	0xe9bb0499,
+	0x18e0b600,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0xf0fca5f9,
+	0xf400f8fe,
+	0xf0fc0131,
+/* 0x0398: recv_done */
+	0x90fc80fc,
+/* 0x039e: init */
+	0x17f100f8,
+	0x14b60108,
+	0x0011cf06,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xd00604b6,
+	0x04bd0001,
+	0xf0ff17f0,
+	0x04b61407,
+	0x0001d006,
+	0x17f004bd,
+	0x0015f102,
+	0x1007f008,
+	0xd00604b6,
+	0x04bd0001,
+	0x011a17f1,
+	0xfe0013f0,
+	0x31f40010,
+	0x0117f010,
 	0xb63807f0,
-	0x08d00604,
-/* 0x024f: timer_done */
-	0xf404bd00,
-	0x00f81031,
-/* 0x0254: send_proc */
-	0x90f980f9,
-	0x9805e898,
-	0x86f004e9,
-	0x0689b804,
-	0xc42a0bf4,
-	0x88940398,
-	0x1880b604,
-	0x98008ebb,
-	0x8a8000fa,
-	0x018d8000,
-	0x80028c80,
-	0x90b6038b,
-	0x0794f001,
-	0xf404e980,
-/* 0x028e: send_done */
-	0x90fc0231,
-	0x00f880fc,
-/* 0x0294: find */
-	0x87f080f9,
-	0x0131f458,
-/* 0x029c: find_loop */
-	0xb8008a98,
-	0x0bf406ae,
-	0x5880b610,
-	0x026886b1,
-	0xf4f01bf4,
-/* 0x02b2: find_done */
-	0x8eb90132,
-	0xf880fc02,
-/* 0x02b9: send */
-	0x9421f500,
-	0x9701f402,
-/* 0x02c2: recv */
-	0xe89800f8,
-	0x04e99805,
-	0xb80132f4,
-	0x0bf40689,
-	0x0389c43d,
-	0xf00180b6,
-	0xe8800784,
-	0x02ea9805,
-	0x8ffef0f9,
-	0xb9f0f901,
-	0x999402ef,
-	0x00e9bb04,
-	0x9818e0b6,
-	0xec9803eb,
-	0x01ed9802,
-	0xf900ee98,
-	0xfef0fca5,
-	0x31f400f8,
-/* 0x030b: recv_done */
-	0xf8f0fc01,
-/* 0x030d: init */
-	0x0817f100,
-	0x0614b601,
-	0xe70011cf,
-	0xb6010911,
-	0x14fe0814,
-	0xe017f100,
-	0x0013f000,
-	0xb61c07f0,
 	0x01d00604,
 	0xf004bd00,
-	0x07f0ff17,
-	0x0604b614,
-	0xbd0001d0,
-	0x0217f004,
-	0x080015f1,
-	0xb61007f0,
-	0x01d00604,
-	0xf104bd00,
-	0xf0010a17,
-	0x10fe0013,
-	0x1031f400,
-	0xf00117f0,
-	0x04b63807,
-	0x0001d006,
-	0xf7f004bd,
-/* 0x0371: init_proc */
-	0x01f19858,
-	0xf40016b0,
-	0x15f9fa0b,
-	0xf458f0b6,
-/* 0x0382: host_send */
-	0x17f1f20e,
-	0x14b604b0,
-	0x0011cf06,
-	0x04a027f1,
-	0xcf0624b6,
-	0x12b80022,
-	0x320bf406,
-	0x94071ec4,
-	0xe0b704ee,
-	0xeb980270,
-	0x02ec9803,
-	0x9801ed98,
-	0x21f500ee,
-	0x10b602b9,
-	0x0f1ec401,
-	0x04b007f1,
-	0xd00604b6,
-	0x04bd000e,
-/* 0x03cb: host_send_done */
-	0xf8ba0ef4,
-/* 0x03cd: host_recv */
-	0x4917f100,
-	0x5413f14e,
-	0x06e1b852,
-/* 0x03db: host_recv_wait */
-	0xf1aa0bf4,
-	0xb604cc17,
-	0x11cf0614,
-	0xc827f100,
-	0x0624b604,
-	0xf00022cf,
-	0x12b80816,
-	0xe60bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
+/* 0x0402: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x0413: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0464: host_send */
+	0x04b017f1,
+	0xcf0614b6,
+	0x27f10011,
+	0x24b604a0,
+	0x0022cf06,
+	0xf40612b8,
+	0x1ec4320b,
+	0x04ee9407,
+	0x0270e0b7,
+	0x9803eb98,
+	0xed9802ec,
+	0x00ee9801,
+	0x034221f5,
+	0xc40110b6,
+	0x07f10f1e,
+	0x04b604b0,
+	0x000ed006,
+	0x0ef404bd,
+/* 0x04ad: host_send_done */
+/* 0x04af: host_recv */
+	0xf100f8ba,
+	0xf14e4917,
+	0xb8525413,
+	0x0bf406e1,
+/* 0x04bd: host_recv_wait */
+	0xcc17f1aa,
+	0x0614b604,
+	0xf10011cf,
+	0xb604c827,
+	0x22cf0624,
+	0x0816f000,
+	0xf40612b8,
+	0x23c4e60b,
+	0x0434b607,
+	0x02f030b7,
+	0x80033b80,
+	0x3d80023c,
+	0x003e8001,
+	0xf00120b6,
+	0x07f10f24,
+	0x04b604c8,
+	0x0002d006,
+	0x27f004bd,
+	0x0007f040,
 	0xd00604b6,
 	0x04bd0002,
-	0xf04027f0,
-	0x04b60007,
-	0x0002d006,
-	0x00f804bd,
-/* 0x0430: host_init */
-	0x008017f1,
-	0xf11014b6,
-	0xf1027015,
-	0xb604d007,
-	0x01d00604,
-	0xf104bd00,
-	0xb6008017,
-	0x15f11014,
-	0x07f102f0,
-	0x04b604dc,
-	0x0001d006,
-	0x17f004bd,
-	0xc407f101,
+/* 0x0512: host_init */
+	0x17f100f8,
+	0x14b60080,
+	0x7015f110,
+	0xd007f102,
 	0x0604b604,
 	0xbd0001d0,
-/* 0x046f: memx_func_enter */
-	0xf000f804,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xd00604b6,
+	0x04bd0001,
+	0xf10117f0,
+	0xb604c407,
+	0x01d00604,
+	0xf804bd00,
+/* 0x0551: memx_func_enter */
+	0x2067f100,
+	0x5d77f116,
+	0xff73f1f5,
+	0x026eb9ff,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f904,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0xfffe77f1,
+	0xffff73f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0487fd02,
+	0x80f960f9,
+	0xe0fcd0fc,
+	0xf13f21f4,
+	0xb926f067,
+	0x21f4026e,
+	0x02d8b904,
+	0xf90487fd,
+	0xfc80f960,
+	0xf4e0fcd0,
+	0x67f03f21,
+	0xe007f104,
+	0x0604b607,
+	0xbd0006d0,
+/* 0x05bd: memx_func_enter_wait */
+	0xc067f104,
+	0x0664b607,
+	0xf00066cf,
+	0x0bf40464,
+	0x2c67f0f3,
+	0xcf0664b6,
+	0x06800066,
+/* 0x05db: memx_func_leave */
+	0xf000f8ee,
+	0x64b62c67,
+	0x0066cf06,
+	0xf0ef0680,
 	0x07f10467,
-	0x04b607e0,
+	0x04b607e4,
 	0x0006d006,
-/* 0x047e: memx_func_enter_wait */
+/* 0x05f6: memx_func_leave_wait */
 	0x67f104bd,
 	0x64b607c0,
 	0x0066cf06,
 	0xf40464f0,
-	0x1698f30b,
-	0x0410b600,
-/* 0x0496: memx_func_leave */
-	0x67f000f8,
-	0xe407f104,
-	0x0604b607,
-	0xbd0006d0,
-/* 0x04a5: memx_func_leave_wait */
-	0xc067f104,
-	0x0664b607,
-	0xf00066cf,
-	0x1bf40464,
-/* 0x04b7: memx_func_wr32 */
-	0x9800f8f3,
-	0x15980016,
-	0x0810b601,
-	0x50f960f9,
+	0x67f1f31b,
+	0x77f126f0,
+	0x73f00001,
+	0x026eb900,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f905,
+	0xfcd0fc80,
+	0x3f21f4e0,
+	0x162067f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0587fd02,
+	0x80f960f9,
 	0xe0fcd0fc,
 	0xf13f21f4,
-	0xfd140003,
-	0x05800506,
-	0xb604bd00,
-	0x1bf40242,
-/* 0x04df: memx_func_wait */
-	0xf000f8dd,
-	0x84b62c87,
-	0x0088cf06,
-	0x98001e98,
-	0x1c98011d,
-	0x031b9802,
-	0xf41010b6,
-	0x00f89c21,
-/* 0x04fc: memx_func_delay */
-	0xb6001e98,
-	0x21f40410,
-/* 0x0507: memx_exec */
-	0xf900f87f,
-	0xb9d0f9e0,
-	0xb2b902c1,
-/* 0x0511: memx_exec_next */
-	0x00139802,
-	0x950410b6,
-	0x30f01034,
-	0xde35980c,
-	0x12b855f9,
-	0xec1ef406,
-	0xe0fcd0fc,
-	0x02b921f5,
-/* 0x0532: memx_info */
-	0xc7f100f8,
-	0xb7f103ac,
-	0x21f50800,
-	0x00f802b9,
-/* 0x0540: memx_recv */
-	0xf401d6b0,
-	0xd6b0c40b,
-	0xe90bf400,
-/* 0x054e: memx_init */
-	0x00f800f8,
-/* 0x0550: perf_recv */
-/* 0x0552: perf_init */
-	0x00f800f8,
-/* 0x0554: i2c_drive_scl */
-	0xf40036b0,
-	0x07f1110b,
-	0x04b607e0,
-	0x0001d006,
-	0x00f804bd,
-/* 0x0568: i2c_drive_scl_lo */
-	0x07e407f1,
-	0xd00604b6,
-	0x04bd0001,
-/* 0x0576: i2c_drive_sda */
-	0x36b000f8,
-	0x110bf400,
-	0x07e007f1,
-	0xd00604b6,
-	0x04bd0002,
-/* 0x058a: i2c_drive_sda_lo */
-	0x07f100f8,
-	0x04b607e4,
-	0x0002d006,
-	0x00f804bd,
-/* 0x0598: i2c_sense_scl */
-	0xf10132f4,
-	0xb607c437,
-	0x33cf0634,
-	0x0431fd00,
-	0xf4060bf4,
-/* 0x05ae: i2c_sense_scl_done */
-	0x00f80131,
-/* 0x05b0: i2c_sense_sda */
-	0xf10132f4,
-	0xb607c437,
-	0x33cf0634,
-	0x0432fd00,
-	0xf4060bf4,
-/* 0x05c6: i2c_sense_sda_done */
-	0x00f80131,
-/* 0x05c8: i2c_raise_scl */
-	0x47f140f9,
-	0x37f00898,
-	0x5421f501,
-/* 0x05d5: i2c_raise_scl_wait */
-	0xe8e7f105,
-	0x7f21f403,
-	0x059821f5,
-	0xb60901f4,
-	0x1bf40142,
-/* 0x05e9: i2c_raise_scl_done */
-	0xf840fcef,
-/* 0x05ed: i2c_start */
-	0x9821f500,
-	0x0d11f405,
-	0x05b021f5,
-	0xf40611f4,
-/* 0x05fe: i2c_start_rep */
-	0x37f0300e,
-	0x5421f500,
-	0x0137f005,
-	0x057621f5,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0xc821f550,
-	0x0464b605,
-/* 0x062b: i2c_start_send */
-	0xf01f11f4,
-	0x21f50037,
-	0xe7f10576,
-	0x21f41388,
-	0x0037f07f,
-	0x055421f5,
-	0x1388e7f1,
-/* 0x0647: i2c_start_out */
+	0xf00aa277,
+	0x6eb90073,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90587,
+	0xd0fc80f9,
+	0x21f4e0fc,
+/* 0x0663: memx_func_wait_vblank */
+	0x9800f83f,
+	0x66b00016,
+	0x130bf400,
+	0xf40166b0,
+	0x0ef4060b,
+/* 0x0675: memx_func_wait_vblank_head1 */
+	0x2077f12e,
+	0x070ef400,
+/* 0x067c: memx_func_wait_vblank_head0 */
+	0x000877f1,
+/* 0x0680: memx_func_wait_vblank_0 */
+	0x07c467f1,
+	0xcf0664b6,
+	0x67fd0066,
+	0xf31bf404,
+/* 0x0690: memx_func_wait_vblank_1 */
+	0x07c467f1,
+	0xcf0664b6,
+	0x67fd0066,
+	0xf30bf404,
+/* 0x06a0: memx_func_wait_vblank_fini */
+	0xf80410b6,
+/* 0x06a5: memx_func_wr32 */
+	0x00169800,
+	0xb6011598,
+	0x60f90810,
+	0xd0fc50f9,
+	0x21f4e0fc,
+	0x0242b63f,
+	0xf8e91bf4,
+/* 0x06c1: memx_func_wait */
+	0x2c87f000,
+	0xcf0684b6,
+	0x1e980088,
+	0x011d9800,
+	0x98021c98,
+	0x10b6031b,
+	0xa421f410,
+/* 0x06de: memx_func_delay */
+	0x1e9800f8,
+	0x0410b600,
 	0xf87f21f4,
-/* 0x0649: i2c_stop */
-	0x0037f000,
-	0x055421f5,
+/* 0x06e9: memx_exec */
+	0xf9e0f900,
+	0x02c1b9d0,
+/* 0x06f3: memx_exec_next */
+	0x9802b2b9,
+	0x10b60013,
+	0xf034e704,
+	0xe033e701,
+	0x0132b601,
+	0x980c30f0,
+	0x55f9de35,
+	0xf40612b8,
+	0x0b98e41e,
+	0xef0c98ee,
+	0xf102cbbb,
+	0xb607c4b7,
+	0xbbcf06b4,
+	0xfcd0fc00,
+	0x4221f5e0,
+/* 0x072f: memx_info */
+	0xf100f803,
+	0xf103c0c7,
+	0xf50800b7,
+	0xf8034221,
+/* 0x073d: memx_recv */
+	0x01d6b000,
+	0xb0a90bf4,
+	0x0bf400d6,
+/* 0x074b: memx_init */
+	0xf800f8e9,
+/* 0x074d: perf_recv */
+/* 0x074f: perf_init */
+	0xf800f800,
+/* 0x0751: i2c_drive_scl */
+	0x0036b000,
+	0xf1110bf4,
+	0xb607e007,
+	0x01d00604,
+	0xf804bd00,
+/* 0x0765: i2c_drive_scl_lo */
+	0xe407f100,
+	0x0604b607,
+	0xbd0001d0,
+/* 0x0773: i2c_drive_sda */
+	0xb000f804,
+	0x0bf40036,
+	0xe007f111,
+	0x0604b607,
+	0xbd0002d0,
+/* 0x0787: i2c_drive_sda_lo */
+	0xf100f804,
+	0xb607e407,
+	0x02d00604,
+	0xf804bd00,
+/* 0x0795: i2c_sense_scl */
+	0x0132f400,
+	0x07c437f1,
+	0xcf0634b6,
+	0x31fd0033,
+	0x060bf404,
+/* 0x07ab: i2c_sense_scl_done */
+	0xf80131f4,
+/* 0x07ad: i2c_sense_sda */
+	0x0132f400,
+	0x07c437f1,
+	0xcf0634b6,
+	0x32fd0033,
+	0x060bf404,
+/* 0x07c3: i2c_sense_sda_done */
+	0xf80131f4,
+/* 0x07c5: i2c_raise_scl */
+	0xf140f900,
+	0xf0089847,
+	0x21f50137,
+/* 0x07d2: i2c_raise_scl_wait */
+	0xe7f10751,
+	0x21f403e8,
+	0x9521f57f,
+	0x0901f407,
+	0xf40142b6,
+/* 0x07e6: i2c_raise_scl_done */
+	0x40fcef1b,
+/* 0x07ea: i2c_start */
+	0x21f500f8,
+	0x11f40795,
+	0xad21f50d,
+	0x0611f407,
+/* 0x07fb: i2c_start_rep */
+	0xf0300ef4,
+	0x21f50037,
+	0x37f00751,
+	0x7321f501,
+	0x0076bb07,
+	0xf90465b6,
+	0x04659450,
+	0xbd0256bb,
+	0x0475fd50,
+	0x21f550fc,
+	0x64b607c5,
+	0x1f11f404,
+/* 0x0828: i2c_start_send */
 	0xf50037f0,
-	0xf1057621,
-	0xf403e8e7,
+	0xf1077321,
+	0xf41388e7,
 	0x37f07f21,
-	0x5421f501,
-	0x88e7f105,
+	0x5121f500,
+	0x88e7f107,
 	0x7f21f413,
-	0xf50137f0,
-	0xf1057621,
-	0xf41388e7,
-	0x00f87f21,
-/* 0x067c: i2c_bitw */
-	0x057621f5,
+/* 0x0844: i2c_start_out */
+/* 0x0846: i2c_stop */
+	0x37f000f8,
+	0x5121f500,
+	0x0037f007,
+	0x077321f5,
 	0x03e8e7f1,
-	0xbb7f21f4,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x05c821f5,
-	0xf40464b6,
-	0xe7f11811,
+	0xf07f21f4,
+	0x21f50137,
+	0xe7f10751,
 	0x21f41388,
-	0x0037f07f,
-	0x055421f5,
+	0x0137f07f,
+	0x077321f5,
 	0x1388e7f1,
-/* 0x06bb: i2c_bitw_out */
 	0xf87f21f4,
-/* 0x06bd: i2c_bitr */
-	0x0137f000,
-	0x057621f5,
-	0x03e8e7f1,
-	0xbb7f21f4,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x05c821f5,
-	0xf40464b6,
-	0x21f51b11,
-	0x37f005b0,
-	0x5421f500,
-	0x88e7f105,
+/* 0x0879: i2c_bitw */
+	0x7321f500,
+	0xe8e7f107,
+	0x7f21f403,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0xc521f550,
+	0x0464b607,
+	0xf11811f4,
+	0xf41388e7,
+	0x37f07f21,
+	0x5121f500,
+	0x88e7f107,
 	0x7f21f413,
-	0xf4013cf0,
-/* 0x0702: i2c_bitr_done */
-	0x00f80131,
-/* 0x0704: i2c_get_byte */
-	0xf00057f0,
-/* 0x070a: i2c_get_byte_next */
-	0x54b60847,
+/* 0x08b8: i2c_bitw_out */
+/* 0x08ba: i2c_bitr */
+	0x37f000f8,
+	0x7321f501,
+	0xe8e7f107,
+	0x7f21f403,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0xc521f550,
+	0x0464b607,
+	0xf51b11f4,
+	0xf007ad21,
+	0x21f50037,
+	0xe7f10751,
+	0x21f41388,
+	0x013cf07f,
+/* 0x08ff: i2c_bitr_done */
+	0xf80131f4,
+/* 0x0901: i2c_get_byte */
+	0x0057f000,
+/* 0x0907: i2c_get_byte_next */
+	0xb60847f0,
+	0x76bb0154,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb608ba21,
+	0x11f40464,
+	0x0553fd2b,
+	0xf40142b6,
+	0x37f0d81b,
 	0x0076bb01,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b606bd,
-	0x2b11f404,
-	0xb60553fd,
-	0x1bf40142,
-	0x0137f0d8,
-	0xb60076bb,
-	0x50f90465,
-	0xbb046594,
-	0x50bd0256,
-	0xfc0475fd,
-	0x7c21f550,
-	0x0464b606,
-/* 0x0754: i2c_get_byte_done */
-/* 0x0756: i2c_put_byte */
-	0x47f000f8,
-/* 0x0759: i2c_put_byte_next */
-	0x0142b608,
-	0xbb3854ff,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x067c21f5,
-	0xf40464b6,
-	0x46b03411,
-	0xd81bf400,
+	0x64b60879,
+/* 0x0951: i2c_get_byte_done */
+/* 0x0953: i2c_put_byte */
+	0xf000f804,
+/* 0x0956: i2c_put_byte_next */
+	0x42b60847,
+	0x3854ff01,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xbd21f550,
-	0x0464b606,
-	0xbb0f11f4,
-	0x36b00076,
-	0x061bf401,
-/* 0x07af: i2c_put_byte_done */
-	0xf80132f4,
-/* 0x07b1: i2c_addr */
-	0x0076bb00,
+	0x7921f550,
+	0x0464b608,
+	0xb03411f4,
+	0x1bf40046,
+	0x0076bbd8,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b605ed,
-	0x2911f404,
-	0x012ec3e7,
-	0xfd0134b6,
-	0x76bb0553,
+	0x64b608ba,
+	0x0f11f404,
+	0xb00076bb,
+	0x1bf40136,
+	0x0132f406,
+/* 0x09ac: i2c_put_byte_done */
+/* 0x09ae: i2c_addr */
+	0x76bb00f8,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6075621,
-/* 0x07f6: i2c_addr_done */
-	0x00f80464,
-/* 0x07f8: i2c_acquire_addr */
-	0xb6f8cec7,
-	0xe0b702e4,
-	0xee980bfc,
-/* 0x0807: i2c_acquire */
-	0xf500f800,
-	0xf407f821,
-	0xd9f00421,
-	0x3f21f403,
-/* 0x0816: i2c_release */
-	0x21f500f8,
-	0x21f407f8,
-	0x03daf004,
-	0xf83f21f4,
-/* 0x0825: i2c_recv */
-	0x0132f400,
-	0xb6f8c1c7,
-	0x16b00214,
-	0x3a1ff528,
-	0xd413a001,
-	0x0032980b,
-	0x0bac13a0,
-	0xf4003198,
-	0xd0f90231,
-	0xd0f9e0f9,
-	0x000067f1,
-	0x100063f1,
-	0xbb016792,
+	0xb607ea21,
+	0x11f40464,
+	0x2ec3e729,
+	0x0134b601,
+	0xbb0553fd,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x080721f5,
-	0xfc0464b6,
-	0x00d6b0d0,
-	0x00b31bf5,
-	0xbb0057f0,
+	0x095321f5,
+/* 0x09f3: i2c_addr_done */
+	0xf80464b6,
+/* 0x09f5: i2c_acquire_addr */
+	0xf8cec700,
+	0xb702e4b6,
+	0x980c10e0,
+	0x00f800ee,
+/* 0x0a04: i2c_acquire */
+	0x09f521f5,
+	0xf00421f4,
+	0x21f403d9,
+/* 0x0a13: i2c_release */
+	0xf500f83f,
+	0xf409f521,
+	0xdaf00421,
+	0x3f21f403,
+/* 0x0a22: i2c_recv */
+	0x32f400f8,
+	0xf8c1c701,
+	0xb00214b6,
+	0x1ff52816,
+	0x13a0013a,
+	0x32980be8,
+	0xc013a000,
+	0x0031980b,
+	0xf90231f4,
+	0xf9e0f9d0,
+	0x0067f1d0,
+	0x0063f100,
+	0x01679210,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x0421f550,
+	0x0464b60a,
+	0xd6b0d0fc,
+	0xb31bf500,
+	0x0057f000,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0xae21f550,
+	0x0464b609,
+	0x00d011f5,
+	0xbbe0c5c7,
 	0x65b60076,
 	0x9450f904,
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x07b121f5,
+	0x095321f5,
 	0xf50464b6,
-	0xc700d011,
-	0x76bbe0c5,
+	0xf000ad11,
+	0x76bb0157,
 	0x0465b600,
 	0x659450f9,
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6075621,
+	0xb609ae21,
 	0x11f50464,
-	0x57f000ad,
-	0x0076bb01,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0x21f550fc,
-	0x64b607b1,
-	0x8a11f504,
-	0x0076bb00,
-	0xf90465b6,
-	0x04659450,
-	0xbd0256bb,
-	0x0475fd50,
-	0x21f550fc,
-	0x64b60704,
-	0x6a11f404,
-	0xbbe05bcb,
-	0x65b60076,
-	0x9450f904,
-	0x56bb0465,
-	0xfd50bd02,
-	0x50fc0475,
-	0x064921f5,
-	0xb90464b6,
-	0x74bd025b,
-/* 0x092b: i2c_recv_not_rd08 */
-	0xb0430ef4,
-	0x1bf401d6,
-	0x0057f03d,
-	0x07b121f5,
-	0xc73311f4,
-	0x21f5e0c5,
-	0x11f40756,
-	0x0057f029,
-	0x07b121f5,
-	0xc71f11f4,
-	0x21f5e0b5,
-	0x11f40756,
-	0x4921f515,
-	0xc774bd06,
-	0x1bf408c5,
-	0x0232f409,
-/* 0x096b: i2c_recv_not_wr08 */
-/* 0x096b: i2c_recv_done */
-	0xc7030ef4,
-	0x21f5f8ce,
-	0xe0fc0816,
-	0x12f4d0fc,
-	0x027cb90a,
-	0x02b921f5,
-/* 0x0980: i2c_recv_exit */
-/* 0x0982: i2c_init */
+	0x76bb008a,
+	0x0465b600,
+	0x659450f9,
+	0x0256bb04,
+	0x75fd50bd,
+	0xf550fc04,
+	0xb6090121,
+	0x11f40464,
+	0xe05bcb6a,
+	0xb60076bb,
+	0x50f90465,
+	0xbb046594,
+	0x50bd0256,
+	0xfc0475fd,
+	0x4621f550,
+	0x0464b608,
+	0xbd025bb9,
+	0x430ef474,
+/* 0x0b28: i2c_recv_not_rd08 */
+	0xf401d6b0,
+	0x57f03d1b,
+	0xae21f500,
+	0x3311f409,
+	0xf5e0c5c7,
+	0xf4095321,
+	0x57f02911,
+	0xae21f500,
+	0x1f11f409,
+	0xf5e0b5c7,
+	0xf4095321,
+	0x21f51511,
+	0x74bd0846,
+	0xf408c5c7,
+	0x32f4091b,
+	0x030ef402,
+/* 0x0b68: i2c_recv_not_wr08 */
+/* 0x0b68: i2c_recv_done */
+	0xf5f8cec7,
+	0xfc0a1321,
+	0xf4d0fce0,
+	0x7cb90a12,
+	0x4221f502,
+/* 0x0b7d: i2c_recv_exit */
+/* 0x0b7f: i2c_init */
+	0xf800f803,
+/* 0x0b81: test_recv */
+	0xd817f100,
+	0x0614b605,
+	0xb60011cf,
+	0x07f10110,
+	0x04b605d8,
+	0x0001d006,
+	0xe7f104bd,
+	0xe3f1d900,
+	0x21f5134f,
+	0x00f80262,
+/* 0x0ba8: test_init */
+	0x0800e7f1,
+	0x026221f5,
+/* 0x0bb2: idle_recv */
 	0x00f800f8,
-/* 0x0984: test_recv */
-	0x05d817f1,
-	0xcf0614b6,
-	0x10b60011,
-	0xd807f101,
-	0x0604b605,
-	0xbd0001d0,
-	0x00e7f104,
-	0x4fe3f1d9,
-	0xf521f513,
-/* 0x09ab: test_init */
-	0xf100f801,
-	0xf50800e7,
-	0xf801f521,
-/* 0x09b5: idle_recv */
-/* 0x09b7: idle */
-	0xf400f800,
-	0x17f10031,
-	0x14b605d4,
-	0x0011cf06,
-	0xf10110b6,
-	0xb605d407,
-	0x01d00604,
-/* 0x09d3: idle_loop */
-	0xf004bd00,
-	0x32f45817,
-/* 0x09d9: idle_proc */
-/* 0x09d9: idle_proc_exec */
-	0xb910f902,
-	0x21f5021e,
-	0x10fc02c2,
-	0xf40911f4,
-	0x0ef40231,
-/* 0x09ed: idle_proc_next */
-	0x5810b6ef,
-	0xf4061fb8,
-	0x02f4e61b,
-	0x0028f4dd,
-	0x00bb0ef4,
+/* 0x0bb4: idle */
+	0xf10031f4,
+	0xb605d417,
+	0x11cf0614,
+	0x0110b600,
+	0x05d407f1,
+	0xd00604b6,
+	0x04bd0001,
+/* 0x0bd0: idle_loop */
+	0xf45817f0,
+/* 0x0bd6: idle_proc */
+/* 0x0bd6: idle_proc_exec */
+	0x10f90232,
+	0xf5021eb9,
+	0xfc034b21,
+	0x0911f410,
+	0xf40231f4,
+/* 0x0bea: idle_proc_next */
+	0x10b6ef0e,
+	0x061fb858,
+	0xf4e61bf4,
+	0x28f4dd02,
+	0xbb0ef400,
+	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
index 8a89dfe41ce1..b85443261569 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc
@@ -23,6 +23,7 @@
  */
 
 #define NVKM_PPWR_CHIPSET GF119
+#define HW_TICKS_PER_US 324
 
 //#define NVKM_FALCON_PC24
 #define NVKM_FALCON_UNSHIFTED_IO
@@ -34,6 +35,7 @@
 .section #nvd0_pwr_data
 #define INCLUDE_PROC
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -44,6 +46,7 @@
 
 #define INCLUDE_DATA
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
@@ -56,6 +59,7 @@
 .section #nvd0_pwr_code
 #define INCLUDE_CODE
 #include "kernel.fuc"
+#include "arith.fuc"
 #include "host.fuc"
 #include "memx.fuc"
 #include "perf.fuc"
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
index cd9ff1a73284..12d86f72ad10 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/nvd0.fuc.h
@@ -24,8 +24,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 /* 0x0058: proc_list_head */
 	0x54534f48,
-	0x000003be,
-	0x00000367,
+	0x0000049d,
+	0x00000446,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -46,8 +46,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x584d454d,
-	0x000004c4,
-	0x000004b6,
+	0x00000678,
+	0x0000066a,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -68,8 +68,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x46524550,
-	0x000004c8,
-	0x000004c6,
+	0x0000067c,
+	0x0000067a,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -90,8 +90,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x5f433249,
-	0x000008e3,
-	0x00000786,
+	0x00000a97,
+	0x0000093a,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -112,8 +112,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x54534554,
-	0x00000906,
-	0x000008e5,
+	0x00000aba,
+	0x00000a99,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -134,8 +134,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x454c4449,
-	0x00000912,
-	0x00000910,
+	0x00000ac6,
+	0x00000ac4,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -227,24 +227,31 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 /* 0x0370: memx_func_head */
-	0x00010000,
-	0x00000000,
-	0x000003f4,
-/* 0x037c: memx_func_next */
 	0x00000001,
 	0x00000000,
-	0x00000415,
+	0x000004d3,
+/* 0x037c: memx_func_next */
 	0x00000002,
+	0x00000000,
+	0x00000554,
+	0x00000003,
 	0x00000002,
-	0x00000430,
-	0x00040003,
+	0x000005d8,
+	0x00040004,
+	0x00000000,
+	0x000005f4,
+	0x00010005,
+	0x00000000,
+	0x0000060e,
+	0x00010006,
+	0x00000000,
+	0x000005d3,
+/* 0x03b8: memx_func_tail */
+/* 0x03b8: memx_ts_start */
 	0x00000000,
-	0x00000458,
-	0x00010004,
+/* 0x03bc: memx_ts_end */
 	0x00000000,
-	0x00000472,
-/* 0x03ac: memx_func_tail */
-/* 0x03ac: memx_data_head */
+/* 0x03c0: memx_data_head */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -757,8 +764,8 @@ uint32_t nvd0_pwr_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0bac: memx_data_tail */
-/* 0x0bac: i2c_scl_map */
+/* 0x0bc0: memx_data_tail */
+/* 0x0bc0: i2c_scl_map */
 	0x00000400,
 	0x00000800,
 	0x00001000,
@@ -769,7 +776,7 @@ uint32_t nvd0_pwr_data[] = {
 	0x00020000,
 	0x00040000,
 	0x00080000,
-/* 0x0bd4: i2c_sda_map */
+/* 0x0be8: i2c_sda_map */
 	0x00100000,
 	0x00200000,
 	0x00400000,
@@ -781,10 +788,69 @@ uint32_t nvd0_pwr_data[] = {
 	0x10000000,
 	0x20000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
 
 uint32_t nvd0_pwr_code[] = {
-	0x02bf0ef5,
+	0x034d0ef5,
 /* 0x0004: rd32 */
 	0x07a007f1,
 	0xbd000ed0,
@@ -814,17 +880,20 @@ uint32_t nvd0_pwr_code[] = {
 	0xd4f100dd,
 	0x1bf47000,
 /* 0x0067: nsec */
-	0xf000f8f5,
+	0xf900f8f5,
+	0xf080f990,
 	0x88cf2c87,
-/* 0x006d: nsec_loop */
+/* 0x0071: nsec_loop */
 	0x2c97f000,
 	0xbb0099cf,
 	0x9eb80298,
 	0xf41ef406,
-/* 0x007e: wait */
-	0x87f000f8,
+	0x90fc80fc,
+/* 0x0086: wait */
+	0x90f900f8,
+	0x87f080f9,
 	0x0088cf2c,
-/* 0x0084: wait_loop */
+/* 0x0090: wait_loop */
 	0xf402eeb9,
 	0xdab90421,
 	0x04adfd02,
@@ -833,28 +902,29 @@ uint32_t nvd0_pwr_code[] = {
 	0x0099cf2c,
 	0xb80298bb,
 	0x1ef4069b,
-/* 0x00a5: wait_done */
-/* 0x00a7: intr_watchdog */
-	0x9800f8e2,
+/* 0x00b1: wait_done */
+	0xfc80fce2,
+/* 0x00b7: intr_watchdog */
+	0x9800f890,
 	0x96b003e9,
 	0x2a0bf400,
 	0xbb9a0a98,
 	0x1cf4029a,
 	0x01d7f00f,
-	0x020621f5,
+	0x028c21f5,
 	0x0ef494bd,
-/* 0x00c5: intr_watchdog_next_time */
+/* 0x00d5: intr_watchdog_next_time */
 	0x9b0a9815,
 	0xf400a6b0,
 	0x9ab8090b,
 	0x061cf406,
-/* 0x00d4: intr_watchdog_next_time_set */
-/* 0x00d7: intr_watchdog_next_proc */
+/* 0x00e4: intr_watchdog_next_time_set */
+/* 0x00e7: intr_watchdog_next_proc */
 	0x809b0980,
 	0xe0b603e9,
 	0x68e6b158,
 	0xc61bf402,
-/* 0x00e6: intr */
+/* 0x00f6: intr */
 	0x00f900f8,
 	0x80f904bd,
 	0xa0f990f9,
@@ -872,12 +942,12 @@ uint32_t nvd0_pwr_code[] = {
 	0x0bf40289,
 	0x9b008020,
 	0xf458e7f0,
-	0x0998a721,
+	0x0998b721,
 	0x0096b09b,
 	0xf00e0bf4,
 	0x09d03407,
 	0x8004bd00,
-/* 0x013e: intr_skip_watchdog */
+/* 0x014e: intr_skip_watchdog */
 	0x89e49a09,
 	0x0bf40800,
 	0x8897f13c,
@@ -889,20 +959,20 @@ uint32_t nvd0_pwr_code[] = {
 	0xf14f48e7,
 	0xf05453e3,
 	0x21f500d7,
-	0xc0fc026b,
+	0xc0fc02f1,
 	0x04c007f1,
 	0xbd000cd0,
-/* 0x0175: intr_subintr_skip_fifo */
+/* 0x0185: intr_subintr_skip_fifo */
 	0x8807f104,
 	0x0009d006,
-/* 0x017e: intr_skip_subintr */
+/* 0x018e: intr_skip_subintr */
 	0x89c404bd,
 	0x070bf420,
 	0xffbfa4f1,
-/* 0x0188: intr_skip_pause */
+/* 0x0198: intr_skip_pause */
 	0xf44089c4,
 	0xa4f1070b,
-/* 0x0192: intr_skip_user0 */
+/* 0x01a2: intr_skip_user0 */
 	0x07f0ffbf,
 	0x0008d004,
 	0x80fc04bd,
@@ -912,324 +982,436 @@ uint32_t nvd0_pwr_code[] = {
 	0xfca0fcb0,
 	0xfc80fc90,
 	0x0032f400,
-/* 0x01b6: timer */
-	0x32f401f8,
-	0x03f89810,
-	0xf40086b0,
-	0xfe80421c,
-	0x3807f003,
+/* 0x01c6: ticks_from_ns */
+	0xc0f901f8,
+	0xd7f1b0f9,
+	0xd3f00144,
+	0xb321f500,
+	0xe8ccec03,
+	0x00b4b003,
+	0xec120bf4,
+	0xf103e8ee,
+	0xf00144d7,
+	0x21f500d3,
+/* 0x01ee: ticks_from_ns_quit */
+	0xceb903b3,
+	0xfcb0fc02,
+/* 0x01f7: ticks_from_us */
+	0xf900f8c0,
+	0xf1b0f9c0,
+	0xf00144d7,
+	0x21f500d3,
+	0xceb903b3,
+	0x00b4b002,
+	0xbd050bf4,
+/* 0x0211: ticks_from_us_quit */
+	0xfcb0fce4,
+/* 0x0217: ticks_to_us */
+	0xf100f8c0,
+	0xf00144d7,
+	0xedff00d3,
+/* 0x0223: timer */
+	0xf900f8ec,
+	0xf480f990,
+	0xf8981032,
+	0x0086b003,
+	0xbd531cf4,
+	0x3807f084,
 	0xbd0008d0,
-	0x0887f004,
-	0xf00088cf,
-	0x1bf40284,
-	0x3487f020,
-	0xb80088cf,
-	0x0bf406e0,
-	0x06e8b809,
-/* 0x01eb: timer_reset */
-	0xf0191ef4,
-	0x0ed03407,
-	0x8004bd00,
-/* 0x01f6: timer_enable */
-	0x87f09a0e,
-	0x3807f001,
-	0xbd0008d0,
-/* 0x0201: timer_done */
-	0x1031f404,
-/* 0x0206: send_proc */
-	0x80f900f8,
-	0xe89890f9,
+	0x3487f004,
+	0x980088cf,
+	0x98bb9a09,
+	0x00e9bb02,
+	0xf003fe80,
+	0x88cf0887,
+	0x0284f000,
+	0xf0201bf4,
+	0x88cf3487,
+	0x06e0b800,
+	0xb8090bf4,
+	0x1cf406e8,
+/* 0x026d: timer_reset */
+	0x3407f00e,
+	0xbd000ed0,
+	0x9a0e8004,
+/* 0x0278: timer_enable */
+	0xf00187f0,
+	0x08d03807,
+/* 0x0283: timer_done */
+	0xf404bd00,
+	0x80fc1031,
+	0x00f890fc,
+/* 0x028c: send_proc */
+	0x90f980f9,
+	0x9805e898,
+	0x86f004e9,
+	0x0689b804,
+	0xc42a0bf4,
+	0x88940398,
+	0x1880b604,
+	0x98008ebb,
+	0x8a8000fa,
+	0x018d8000,
+	0x80028c80,
+	0x90b6038b,
+	0x0794f001,
+	0xf404e980,
+/* 0x02c6: send_done */
+	0x90fc0231,
+	0x00f880fc,
+/* 0x02cc: find */
+	0x87f080f9,
+	0x0131f458,
+/* 0x02d4: find_loop */
+	0xb8008a98,
+	0x0bf406ae,
+	0x5880b610,
+	0x026886b1,
+	0xf4f01bf4,
+/* 0x02ea: find_done */
+	0x8eb90132,
+	0xf880fc02,
+/* 0x02f1: send */
+	0xcc21f500,
+	0x9701f402,
+/* 0x02fa: recv */
+	0x90f900f8,
+	0xe89880f9,
 	0x04e99805,
-	0xb80486f0,
+	0xb80132f4,
 	0x0bf40689,
-	0x0398c42a,
-	0xb6048894,
-	0x8ebb1880,
-	0x00fa9800,
-	0x80008a80,
-	0x8c80018d,
-	0x038b8002,
-	0xf00190b6,
-	0xe9800794,
-	0x0231f404,
-/* 0x0240: send_done */
-	0x80fc90fc,
-/* 0x0246: find */
-	0x80f900f8,
-	0xf45887f0,
-/* 0x024e: find_loop */
-	0x8a980131,
-	0x06aeb800,
-	0xb6100bf4,
-	0x86b15880,
-	0x1bf40268,
-	0x0132f4f0,
-/* 0x0264: find_done */
-	0xfc028eb9,
-/* 0x026b: send */
-	0xf500f880,
-	0xf4024621,
-	0x00f89701,
-/* 0x0274: recv */
-	0x9805e898,
-	0x32f404e9,
-	0x0689b801,
-	0xc43d0bf4,
-	0x80b60389,
-	0x0784f001,
-	0x9805e880,
-	0xf0f902ea,
-	0xf9018ffe,
-	0x02efb9f0,
-	0xbb049994,
-	0xe0b600e9,
-	0x03eb9818,
-	0x9802ec98,
-	0xee9801ed,
-	0xfca5f900,
-	0x00f8fef0,
-	0xfc0131f4,
-/* 0x02bd: recv_done */
-/* 0x02bf: init */
-	0xf100f8f0,
-	0xcf010817,
-	0x11e70011,
-	0x14b60109,
-	0x0014fe08,
-	0x00e017f1,
-	0xf00013f0,
-	0x01d01c07,
-	0xf004bd00,
-	0x07f0ff17,
-	0x0001d014,
-	0x17f004bd,
-	0x0015f102,
-	0x1007f008,
-	0xbd0001d0,
-	0xe617f104,
-	0x0013f000,
-	0xf40010fe,
-	0x17f01031,
-	0x3807f001,
-	0xbd0001d0,
-	0x58f7f004,
-/* 0x0314: init_proc */
-	0xb001f198,
-	0x0bf40016,
-	0xb615f9fa,
-	0x0ef458f0,
-/* 0x0325: host_send */
-	0xb017f1f2,
-	0x0011cf04,
-	0x04a027f1,
-	0xb80022cf,
-	0x0bf40612,
-	0x071ec42f,
-	0xb704ee94,
-	0x980270e0,
+	0x0389c43d,
+	0xf00180b6,
+	0xe8800784,
+	0x02ea9805,
+	0x8ffef0f9,
+	0xb9f0f901,
+	0x999402ef,
+	0x00e9bb04,
+	0x9818e0b6,
 	0xec9803eb,
 	0x01ed9802,
-	0xf500ee98,
-	0xb6026b21,
-	0x1ec40110,
-	0xb007f10f,
-	0x000ed004,
-	0x0ef404bd,
-/* 0x0365: host_send_done */
-/* 0x0367: host_recv */
-	0xf100f8c3,
-	0xf14e4917,
-	0xb8525413,
-	0x0bf406e1,
-/* 0x0375: host_recv_wait */
-	0xcc17f1b3,
-	0x0011cf04,
-	0x04c827f1,
-	0xf00022cf,
-	0x12b80816,
-	0xec0bf406,
-	0xb60723c4,
-	0x30b70434,
-	0x3b8002f0,
-	0x023c8003,
-	0x80013d80,
-	0x20b6003e,
-	0x0f24f001,
-	0x04c807f1,
-	0xbd0002d0,
-	0x4027f004,
-	0xd00007f0,
-	0x04bd0002,
-/* 0x03be: host_init */
+	0xf900ee98,
+	0xfef0fca5,
+	0x31f400f8,
+/* 0x0347: recv_done */
+	0xfcf0fc01,
+	0xf890fc80,
+/* 0x034d: init */
+	0x0817f100,
+	0x0011cf01,
+	0x010911e7,
+	0xfe0814b6,
+	0x17f10014,
+	0x13f000e0,
+	0x1c07f000,
+	0xbd0001d0,
+	0xff17f004,
+	0xd01407f0,
+	0x04bd0001,
+	0xf10217f0,
+	0xf0080015,
+	0x01d01007,
+	0xf104bd00,
+	0xf000f617,
+	0x10fe0013,
+	0x1031f400,
+	0xf00117f0,
+	0x01d03807,
+	0xf004bd00,
+/* 0x03a2: init_proc */
+	0xf19858f7,
+	0x0016b001,
+	0xf9fa0bf4,
+	0x58f0b615,
+/* 0x03b3: mulu32_32_64 */
+	0xf9f20ef4,
+	0xf920f910,
+	0x9540f930,
+	0xd29510e1,
+	0xbdc4bd10,
+	0xc0edffb4,
+	0xb9301dff,
+	0x34f10234,
+	0x34b6ffff,
+	0x1045b610,
+	0xbb00c3bb,
+	0xe2ff01b4,
+	0x0234b930,
+	0xffff34f1,
+	0xb61034b6,
+	0xc3bb1045,
+	0x01b4bb00,
+	0xbb3012ff,
+	0x40fc00b3,
+	0x20fc30fc,
+	0x00f810fc,
+/* 0x0404: host_send */
+	0x04b017f1,
+	0xf10011cf,
+	0xcf04a027,
+	0x12b80022,
+	0x2f0bf406,
+	0x94071ec4,
+	0xe0b704ee,
+	0xeb980270,
+	0x02ec9803,
+	0x9801ed98,
+	0x21f500ee,
+	0x10b602f1,
+	0x0f1ec401,
+	0x04b007f1,
+	0xbd000ed0,
+	0xc30ef404,
+/* 0x0444: host_send_done */
+/* 0x0446: host_recv */
 	0x17f100f8,
-	0x14b60080,
-	0x7015f110,
-	0xd007f102,
-	0x0001d004,
-	0x17f104bd,
-	0x14b60080,
-	0xf015f110,
-	0xdc07f102,
-	0x0001d004,
-	0x17f004bd,
-	0xc407f101,
-	0x0001d004,
-	0x00f804bd,
-/* 0x03f4: memx_func_enter */
+	0x13f14e49,
+	0xe1b85254,
+	0xb30bf406,
+/* 0x0454: host_recv_wait */
+	0x04cc17f1,
+	0xf10011cf,
+	0xcf04c827,
+	0x16f00022,
+	0x0612b808,
+	0xc4ec0bf4,
+	0x34b60723,
+	0xf030b704,
+	0x033b8002,
+	0x80023c80,
+	0x3e80013d,
+	0x0120b600,
+	0xf10f24f0,
+	0xd004c807,
+	0x04bd0002,
+	0xf04027f0,
+	0x02d00007,
+	0xf804bd00,
+/* 0x049d: host_init */
+	0x8017f100,
+	0x1014b600,
+	0x027015f1,
+	0x04d007f1,
+	0xbd0001d0,
+	0x8017f104,
+	0x1014b600,
+	0x02f015f1,
+	0x04dc07f1,
+	0xbd0001d0,
+	0x0117f004,
+	0x04c407f1,
+	0xbd0001d0,
+/* 0x04d3: memx_func_enter */
+	0xf100f804,
+	0xf1162067,
+	0xf1f55d77,
+	0xb9ffff73,
+	0x21f4026e,
+	0x02d8b904,
+	0xf90487fd,
+	0xfc80f960,
+	0xf4e0fcd0,
+	0x77f13321,
+	0x73f1fffe,
+	0x6eb9ffff,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90487,
+	0xd0fc80f9,
+	0x21f4e0fc,
+	0xf067f133,
+	0x026eb926,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f904,
+	0xfcd0fc80,
+	0x3321f4e0,
 	0xf10467f0,
 	0xd007e007,
 	0x04bd0006,
-/* 0x0400: memx_func_enter_wait */
+/* 0x053c: memx_func_enter_wait */
 	0x07c067f1,
 	0xf00066cf,
 	0x0bf40464,
-	0x001698f6,
-	0xf80410b6,
-/* 0x0415: memx_func_leave */
-	0x0467f000,
+	0x2c67f0f6,
+	0x800066cf,
+	0x00f8ee06,
+/* 0x0554: memx_func_leave */
+	0xcf2c67f0,
+	0x06800066,
+	0x0467f0ef,
 	0x07e407f1,
 	0xbd0006d0,
-/* 0x0421: memx_func_leave_wait */
+/* 0x0569: memx_func_leave_wait */
 	0xc067f104,
 	0x0066cf07,
 	0xf40464f0,
-	0x00f8f61b,
-/* 0x0430: memx_func_wr32 */
+	0x67f1f61b,
+	0x77f126f0,
+	0x73f00001,
+	0x026eb900,
+	0xb90421f4,
+	0x87fd02d8,
+	0xf960f905,
+	0xfcd0fc80,
+	0x3321f4e0,
+	0x162067f1,
+	0xf4026eb9,
+	0xd8b90421,
+	0x0587fd02,
+	0x80f960f9,
+	0xe0fcd0fc,
+	0xf13321f4,
+	0xf00aa277,
+	0x6eb90073,
+	0x0421f402,
+	0xfd02d8b9,
+	0x60f90587,
+	0xd0fc80f9,
+	0x21f4e0fc,
+/* 0x05d3: memx_func_wait_vblank */
+	0xb600f833,
+	0x00f80410,
+/* 0x05d8: memx_func_wr32 */
 	0x98001698,
 	0x10b60115,
 	0xf960f908,
 	0xfcd0fc50,
 	0x3321f4e0,
-	0x140003f1,
-	0x800506fd,
-	0x04bd0005,
 	0xf40242b6,
-	0x00f8dd1b,
-/* 0x0458: memx_func_wait */
+	0x00f8e91b,
+/* 0x05f4: memx_func_wait */
 	0xcf2c87f0,
 	0x1e980088,
 	0x011d9800,
 	0x98021c98,
 	0x10b6031b,
-	0x7e21f410,
-/* 0x0472: memx_func_delay */
+	0x8621f410,
+/* 0x060e: memx_func_delay */
 	0x1e9800f8,
 	0x0410b600,
 	0xf86721f4,
-/* 0x047d: memx_exec */
+/* 0x0619: memx_exec */
 	0xf9e0f900,
 	0x02c1b9d0,
-/* 0x0487: memx_exec_next */
+/* 0x0623: memx_exec_next */
 	0x9802b2b9,
 	0x10b60013,
-	0x10349504,
+	0xf034e704,
+	0xe033e701,
+	0x0132b601,
 	0x980c30f0,
 	0x55f9de35,
 	0xf40612b8,
-	0xd0fcec1e,
+	0x0b98e41e,
+	0xef0c98ee,
+	0xf102cbbb,
+	0xcf07c4b7,
+	0xd0fc00bb,
 	0x21f5e0fc,
-	0x00f8026b,
-/* 0x04a8: memx_info */
-	0x03acc7f1,
+	0x00f802f1,
+/* 0x065c: memx_info */
+	0x03c0c7f1,
 	0x0800b7f1,
-	0x026b21f5,
-/* 0x04b6: memx_recv */
+	0x02f121f5,
+/* 0x066a: memx_recv */
 	0xd6b000f8,
-	0xc40bf401,
+	0xac0bf401,
 	0xf400d6b0,
 	0x00f8e90b,
-/* 0x04c4: memx_init */
-/* 0x04c6: perf_recv */
+/* 0x0678: memx_init */
+/* 0x067a: perf_recv */
 	0x00f800f8,
-/* 0x04c8: perf_init */
-/* 0x04ca: i2c_drive_scl */
+/* 0x067c: perf_init */
+/* 0x067e: i2c_drive_scl */
 	0x36b000f8,
 	0x0e0bf400,
 	0x07e007f1,
 	0xbd0001d0,
-/* 0x04db: i2c_drive_scl_lo */
+/* 0x068f: i2c_drive_scl_lo */
 	0xf100f804,
 	0xd007e407,
 	0x04bd0001,
-/* 0x04e6: i2c_drive_sda */
+/* 0x069a: i2c_drive_sda */
 	0x36b000f8,
 	0x0e0bf400,
 	0x07e007f1,
 	0xbd0002d0,
-/* 0x04f7: i2c_drive_sda_lo */
+/* 0x06ab: i2c_drive_sda_lo */
 	0xf100f804,
 	0xd007e407,
 	0x04bd0002,
-/* 0x0502: i2c_sense_scl */
+/* 0x06b6: i2c_sense_scl */
 	0x32f400f8,
 	0xc437f101,
 	0x0033cf07,
 	0xf40431fd,
 	0x31f4060b,
-/* 0x0515: i2c_sense_scl_done */
-/* 0x0517: i2c_sense_sda */
+/* 0x06c9: i2c_sense_scl_done */
+/* 0x06cb: i2c_sense_sda */
 	0xf400f801,
 	0x37f10132,
 	0x33cf07c4,
 	0x0432fd00,
 	0xf4060bf4,
-/* 0x052a: i2c_sense_sda_done */
+/* 0x06de: i2c_sense_sda_done */
 	0x00f80131,
-/* 0x052c: i2c_raise_scl */
+/* 0x06e0: i2c_raise_scl */
 	0x47f140f9,
 	0x37f00898,
-	0xca21f501,
-/* 0x0539: i2c_raise_scl_wait */
-	0xe8e7f104,
+	0x7e21f501,
+/* 0x06ed: i2c_raise_scl_wait */
+	0xe8e7f106,
 	0x6721f403,
-	0x050221f5,
+	0x06b621f5,
 	0xb60901f4,
 	0x1bf40142,
-/* 0x054d: i2c_raise_scl_done */
+/* 0x0701: i2c_raise_scl_done */
 	0xf840fcef,
-/* 0x0551: i2c_start */
-	0x0221f500,
-	0x0d11f405,
-	0x051721f5,
+/* 0x0705: i2c_start */
+	0xb621f500,
+	0x0d11f406,
+	0x06cb21f5,
 	0xf40611f4,
-/* 0x0562: i2c_start_rep */
+/* 0x0716: i2c_start_rep */
 	0x37f0300e,
-	0xca21f500,
-	0x0137f004,
-	0x04e621f5,
+	0x7e21f500,
+	0x0137f006,
+	0x069a21f5,
 	0xb60076bb,
 	0x50f90465,
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x2c21f550,
-	0x0464b605,
-/* 0x058f: i2c_start_send */
+	0xe021f550,
+	0x0464b606,
+/* 0x0743: i2c_start_send */
 	0xf01f11f4,
 	0x21f50037,
-	0xe7f104e6,
+	0xe7f1069a,
 	0x21f41388,
 	0x0037f067,
-	0x04ca21f5,
+	0x067e21f5,
 	0x1388e7f1,
-/* 0x05ab: i2c_start_out */
+/* 0x075f: i2c_start_out */
 	0xf86721f4,
-/* 0x05ad: i2c_stop */
+/* 0x0761: i2c_stop */
 	0x0037f000,
-	0x04ca21f5,
+	0x067e21f5,
 	0xf50037f0,
-	0xf104e621,
+	0xf1069a21,
 	0xf403e8e7,
 	0x37f06721,
-	0xca21f501,
-	0x88e7f104,
+	0x7e21f501,
+	0x88e7f106,
 	0x6721f413,
 	0xf50137f0,
-	0xf104e621,
+	0xf1069a21,
 	0xf41388e7,
 	0x00f86721,
-/* 0x05e0: i2c_bitw */
-	0x04e621f5,
+/* 0x0794: i2c_bitw */
+	0x069a21f5,
 	0x03e8e7f1,
 	0xbb6721f4,
 	0x65b60076,
@@ -1237,18 +1419,18 @@ uint32_t nvd0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x052c21f5,
+	0x06e021f5,
 	0xf40464b6,
 	0xe7f11811,
 	0x21f41388,
 	0x0037f067,
-	0x04ca21f5,
+	0x067e21f5,
 	0x1388e7f1,
-/* 0x061f: i2c_bitw_out */
+/* 0x07d3: i2c_bitw_out */
 	0xf86721f4,
-/* 0x0621: i2c_bitr */
+/* 0x07d5: i2c_bitr */
 	0x0137f000,
-	0x04e621f5,
+	0x069a21f5,
 	0x03e8e7f1,
 	0xbb6721f4,
 	0x65b60076,
@@ -1256,19 +1438,19 @@ uint32_t nvd0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x052c21f5,
+	0x06e021f5,
 	0xf40464b6,
 	0x21f51b11,
-	0x37f00517,
-	0xca21f500,
-	0x88e7f104,
+	0x37f006cb,
+	0x7e21f500,
+	0x88e7f106,
 	0x6721f413,
 	0xf4013cf0,
-/* 0x0666: i2c_bitr_done */
+/* 0x081a: i2c_bitr_done */
 	0x00f80131,
-/* 0x0668: i2c_get_byte */
+/* 0x081c: i2c_get_byte */
 	0xf00057f0,
-/* 0x066e: i2c_get_byte_next */
+/* 0x0822: i2c_get_byte_next */
 	0x54b60847,
 	0x0076bb01,
 	0xf90465b6,
@@ -1276,7 +1458,7 @@ uint32_t nvd0_pwr_code[] = {
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b60621,
+	0x64b607d5,
 	0x2b11f404,
 	0xb60553fd,
 	0x1bf40142,
@@ -1286,12 +1468,12 @@ uint32_t nvd0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xe021f550,
-	0x0464b605,
-/* 0x06b8: i2c_get_byte_done */
-/* 0x06ba: i2c_put_byte */
+	0x9421f550,
+	0x0464b607,
+/* 0x086c: i2c_get_byte_done */
+/* 0x086e: i2c_put_byte */
 	0x47f000f8,
-/* 0x06bd: i2c_put_byte_next */
+/* 0x0871: i2c_put_byte_next */
 	0x0142b608,
 	0xbb3854ff,
 	0x65b60076,
@@ -1299,7 +1481,7 @@ uint32_t nvd0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x05e021f5,
+	0x079421f5,
 	0xf40464b6,
 	0x46b03411,
 	0xd81bf400,
@@ -1308,21 +1490,21 @@ uint32_t nvd0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x2121f550,
-	0x0464b606,
+	0xd521f550,
+	0x0464b607,
 	0xbb0f11f4,
 	0x36b00076,
 	0x061bf401,
-/* 0x0713: i2c_put_byte_done */
+/* 0x08c7: i2c_put_byte_done */
 	0xf80132f4,
-/* 0x0715: i2c_addr */
+/* 0x08c9: i2c_addr */
 	0x0076bb00,
 	0xf90465b6,
 	0x04659450,
 	0xbd0256bb,
 	0x0475fd50,
 	0x21f550fc,
-	0x64b60551,
+	0x64b60705,
 	0x2911f404,
 	0x012ec3e7,
 	0xfd0134b6,
@@ -1332,30 +1514,30 @@ uint32_t nvd0_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb606ba21,
-/* 0x075a: i2c_addr_done */
+	0xb6086e21,
+/* 0x090e: i2c_addr_done */
 	0x00f80464,
-/* 0x075c: i2c_acquire_addr */
+/* 0x0910: i2c_acquire_addr */
 	0xb6f8cec7,
 	0xe0b705e4,
 	0x00f8d014,
-/* 0x0768: i2c_acquire */
-	0x075c21f5,
+/* 0x091c: i2c_acquire */
+	0x091021f5,
 	0xf00421f4,
 	0x21f403d9,
-/* 0x0777: i2c_release */
+/* 0x092b: i2c_release */
 	0xf500f833,
-	0xf4075c21,
+	0xf4091021,
 	0xdaf00421,
 	0x3321f403,
-/* 0x0786: i2c_recv */
+/* 0x093a: i2c_recv */
 	0x32f400f8,
 	0xf8c1c701,
 	0xb00214b6,
 	0x1ff52816,
 	0x13a0013a,
-	0x32980bd4,
-	0xac13a000,
+	0x32980be8,
+	0xc013a000,
 	0x0031980b,
 	0xf90231f4,
 	0xf9e0f9d0,
@@ -1367,8 +1549,8 @@ uint32_t nvd0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x6821f550,
-	0x0464b607,
+	0x1c21f550,
+	0x0464b609,
 	0xd6b0d0fc,
 	0xb31bf500,
 	0x0057f000,
@@ -1377,8 +1559,8 @@ uint32_t nvd0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0x1521f550,
-	0x0464b607,
+	0xc921f550,
+	0x0464b608,
 	0x00d011f5,
 	0xbbe0c5c7,
 	0x65b60076,
@@ -1386,7 +1568,7 @@ uint32_t nvd0_pwr_code[] = {
 	0x56bb0465,
 	0xfd50bd02,
 	0x50fc0475,
-	0x06ba21f5,
+	0x086e21f5,
 	0xf50464b6,
 	0xf000ad11,
 	0x76bb0157,
@@ -1395,7 +1577,7 @@ uint32_t nvd0_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6071521,
+	0xb608c921,
 	0x11f50464,
 	0x76bb008a,
 	0x0465b600,
@@ -1403,7 +1585,7 @@ uint32_t nvd0_pwr_code[] = {
 	0x0256bb04,
 	0x75fd50bd,
 	0xf550fc04,
-	0xb6066821,
+	0xb6081c21,
 	0x11f40464,
 	0xe05bcb6a,
 	0xb60076bb,
@@ -1411,38 +1593,38 @@ uint32_t nvd0_pwr_code[] = {
 	0xbb046594,
 	0x50bd0256,
 	0xfc0475fd,
-	0xad21f550,
-	0x0464b605,
+	0x6121f550,
+	0x0464b607,
 	0xbd025bb9,
 	0x430ef474,
-/* 0x088c: i2c_recv_not_rd08 */
+/* 0x0a40: i2c_recv_not_rd08 */
 	0xf401d6b0,
 	0x57f03d1b,
-	0x1521f500,
-	0x3311f407,
+	0xc921f500,
+	0x3311f408,
 	0xf5e0c5c7,
-	0xf406ba21,
+	0xf4086e21,
 	0x57f02911,
-	0x1521f500,
-	0x1f11f407,
+	0xc921f500,
+	0x1f11f408,
 	0xf5e0b5c7,
-	0xf406ba21,
+	0xf4086e21,
 	0x21f51511,
-	0x74bd05ad,
+	0x74bd0761,
 	0xf408c5c7,
 	0x32f4091b,
 	0x030ef402,
-/* 0x08cc: i2c_recv_not_wr08 */
-/* 0x08cc: i2c_recv_done */
+/* 0x0a80: i2c_recv_not_wr08 */
+/* 0x0a80: i2c_recv_done */
 	0xf5f8cec7,
-	0xfc077721,
+	0xfc092b21,
 	0xf4d0fce0,
 	0x7cb90a12,
-	0x6b21f502,
-/* 0x08e1: i2c_recv_exit */
-/* 0x08e3: i2c_init */
+	0xf121f502,
+/* 0x0a95: i2c_recv_exit */
+/* 0x0a97: i2c_init */
 	0xf800f802,
-/* 0x08e5: test_recv */
+/* 0x0a99: test_recv */
 	0xd817f100,
 	0x0011cf05,
 	0xf10110b6,
@@ -1450,29 +1632,29 @@ uint32_t nvd0_pwr_code[] = {
 	0x04bd0001,
 	0xd900e7f1,
 	0x134fe3f1,
-	0x01b621f5,
-/* 0x0906: test_init */
+	0x022321f5,
+/* 0x0aba: test_init */
 	0xe7f100f8,
 	0x21f50800,
-	0x00f801b6,
-/* 0x0910: idle_recv */
-/* 0x0912: idle */
+	0x00f80223,
+/* 0x0ac4: idle_recv */
+/* 0x0ac6: idle */
 	0x31f400f8,
 	0xd417f100,
 	0x0011cf05,
 	0xf10110b6,
 	0xd005d407,
 	0x04bd0001,
-/* 0x0928: idle_loop */
+/* 0x0adc: idle_loop */
 	0xf45817f0,
-/* 0x092e: idle_proc */
-/* 0x092e: idle_proc_exec */
+/* 0x0ae2: idle_proc */
+/* 0x0ae2: idle_proc_exec */
 	0x10f90232,
 	0xf5021eb9,
-	0xfc027421,
+	0xfc02fa21,
 	0x0911f410,
 	0xf40231f4,
-/* 0x0942: idle_proc_next */
+/* 0x0af6: idle_proc_next */
 	0x10b6ef0e,
 	0x061fb858,
 	0xf4e61bf4,
@@ -1521,4 +1703,23 @@ uint32_t nvd0_pwr_code[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
index 574acfa44c8c..522e3079f824 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/fuc/os.h
@@ -19,11 +19,12 @@
 #define MEMX_MSG_EXEC 1
 
 /* MEMX: script opcode definitions */
-#define MEMX_ENTER  0
-#define MEMX_LEAVE  1
-#define MEMX_WR32   2
-#define MEMX_WAIT   3
-#define MEMX_DELAY  4
+#define MEMX_ENTER  1
+#define MEMX_LEAVE  2
+#define MEMX_WR32   3
+#define MEMX_WAIT   4
+#define MEMX_DELAY  5
+#define MEMX_VBLANK 6
 
 /* I2C_: message identifiers */
 #define I2C__MSG_RD08 0
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/gk104.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/gk104.c
new file mode 100644
index 000000000000..d76612999b9f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/gk104.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+#define nvd0_pwr_code gk104_pwr_code
+#define nvd0_pwr_data gk104_pwr_data
+#include "fuc/nvd0.fuc.h"
+
+static void
+gk104_pwr_pgob(struct nouveau_pwr *ppwr, bool enable)
+{
+	nv_mask(ppwr, 0x000200, 0x00001000, 0x00000000);
+	nv_rd32(ppwr, 0x000200);
+	nv_mask(ppwr, 0x000200, 0x08000000, 0x08000000);
+	msleep(50);
+
+	nv_mask(ppwr, 0x10a78c, 0x00000002, 0x00000002);
+	nv_mask(ppwr, 0x10a78c, 0x00000001, 0x00000001);
+	nv_mask(ppwr, 0x10a78c, 0x00000001, 0x00000000);
+
+	nv_mask(ppwr, 0x020004, 0xc0000000, enable ? 0xc0000000 : 0x40000000);
+	msleep(50);
+
+	nv_mask(ppwr, 0x10a78c, 0x00000002, 0x00000000);
+	nv_mask(ppwr, 0x10a78c, 0x00000001, 0x00000001);
+	nv_mask(ppwr, 0x10a78c, 0x00000001, 0x00000000);
+
+	nv_mask(ppwr, 0x000200, 0x08000000, 0x00000000);
+	nv_mask(ppwr, 0x000200, 0x00001000, 0x00001000);
+	nv_rd32(ppwr, 0x000200);
+}
+
+struct nouveau_oclass *
+gk104_pwr_oclass = &(struct nvkm_pwr_impl) {
+	.base.handle = NV_SUBDEV(PWR, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_pwr_ctor,
+		.dtor = _nouveau_pwr_dtor,
+		.init = _nouveau_pwr_init,
+		.fini = _nouveau_pwr_fini,
+	},
+	.code.data = gk104_pwr_code,
+	.code.size = sizeof(gk104_pwr_code),
+	.data.data = gk104_pwr_data,
+	.data.size = sizeof(gk104_pwr_data),
+	.pgob = gk104_pwr_pgob,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
index 03de3107d29f..65eaa2546cad 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/memx.c
@@ -1,8 +1,7 @@
 #ifndef __NVKM_PWR_MEMX_H__
 #define __NVKM_PWR_MEMX_H__
 
-#include <subdev/pwr.h>
-#include <subdev/pwr/fuc/os.h>
+#include "priv.h"
 
 struct nouveau_memx {
 	struct nouveau_pwr *ppwr;
@@ -21,10 +20,11 @@ memx_out(struct nouveau_memx *memx)
 	struct nouveau_pwr *ppwr = memx->ppwr;
 	int i;
 
-	if (memx->c.size) {
+	if (memx->c.mthd) {
 		nv_wr32(ppwr, 0x10a1c4, (memx->c.size << 16) | memx->c.mthd);
 		for (i = 0; i < memx->c.size; i++)
 			nv_wr32(ppwr, 0x10a1c4, memx->c.data[i]);
+		memx->c.mthd = 0;
 		memx->c.size = 0;
 	}
 }
@@ -33,7 +33,7 @@ static void
 memx_cmd(struct nouveau_memx *memx, u32 mthd, u32 size, u32 data[])
 {
 	if ((memx->c.size + size >= ARRAY_SIZE(memx->c.data)) ||
-	    (memx->c.size && memx->c.mthd != mthd))
+	    (memx->c.mthd && memx->c.mthd != mthd))
 		memx_out(memx);
 	memcpy(&memx->c.data[memx->c.size], data, size * sizeof(data[0]));
 	memx->c.size += size;
@@ -63,8 +63,7 @@ nouveau_memx_init(struct nouveau_pwr *ppwr, struct nouveau_memx **pmemx)
 		nv_wr32(ppwr, 0x10a580, 0x00000003);
 	} while (nv_rd32(ppwr, 0x10a580) != 0x00000003);
 	nv_wr32(ppwr, 0x10a1c0, 0x01000000 | memx->base);
-	nv_wr32(ppwr, 0x10a1c4, 0x00010000 | MEMX_ENTER);
-	nv_wr32(ppwr, 0x10a1c4, 0x00000000);
+
 	return 0;
 }
 
@@ -79,7 +78,6 @@ nouveau_memx_fini(struct nouveau_memx **pmemx, bool exec)
 	memx_out(memx);
 
 	/* release data segment access */
-	nv_wr32(ppwr, 0x10a1c4, 0x00000000 | MEMX_LEAVE);
 	finish = nv_rd32(ppwr, 0x10a1c0) & 0x00ffffff;
 	nv_wr32(ppwr, 0x10a580, 0x00000000);
 
@@ -89,6 +87,8 @@ nouveau_memx_fini(struct nouveau_memx **pmemx, bool exec)
 				 memx->base, finish);
 	}
 
+	nv_debug(memx->ppwr, "Exec took %uns, PPWR_IN %08x\n",
+		 reply[0], reply[1]);
 	kfree(memx);
 	return 0;
 }
@@ -118,4 +118,51 @@ nouveau_memx_nsec(struct nouveau_memx *memx, u32 nsec)
 	memx_out(memx); /* fuc can't handle multiple */
 }
 
+void
+nouveau_memx_wait_vblank(struct nouveau_memx *memx)
+{
+	struct nouveau_pwr *ppwr = memx->ppwr;
+	u32 heads, x, y, px = 0;
+	int i, head_sync;
+
+	if (nv_device(ppwr)->chipset < 0xd0) {
+		heads = nv_rd32(ppwr, 0x610050);
+		for (i = 0; i < 2; i++) {
+			/* Heuristic: sync to head with biggest resolution */
+			if (heads & (2 << (i << 3))) {
+				x = nv_rd32(ppwr, 0x610b40 + (0x540 * i));
+				y = (x & 0xffff0000) >> 16;
+				x &= 0x0000ffff;
+				if ((x * y) > px) {
+					px = (x * y);
+					head_sync = i;
+				}
+			}
+		}
+	}
+
+	if (px == 0) {
+		nv_debug(memx->ppwr, "WAIT VBLANK !NO ACTIVE HEAD\n");
+		return;
+	}
+
+	nv_debug(memx->ppwr, "WAIT VBLANK HEAD%d\n", head_sync);
+	memx_cmd(memx, MEMX_VBLANK, 1, (u32[]){ head_sync });
+	memx_out(memx); /* fuc can't handle multiple */
+}
+
+void
+nouveau_memx_block(struct nouveau_memx *memx)
+{
+	nv_debug(memx->ppwr, "   HOST BLOCKED\n");
+	memx_cmd(memx, MEMX_ENTER, 0, NULL);
+}
+
+void
+nouveau_memx_unblock(struct nouveau_memx *memx)
+{
+	nv_debug(memx->ppwr, "   HOST UNBLOCKED\n");
+	memx_cmd(memx, MEMX_LEAVE, 0, NULL);
+}
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/nv108.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/nv108.c
index 52c85414866a..04ff7c3c34e9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/nv108.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/nv108.c
@@ -22,41 +22,20 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/pwr.h>
-
+#include "priv.h"
 #include "fuc/nv108.fuc.h"
 
-struct nv108_pwr_priv {
-	struct nouveau_pwr base;
-};
-
-static int
-nv108_pwr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	       struct nouveau_oclass *oclass, void *data, u32 size,
-	       struct nouveau_object **pobject)
-{
-	struct nv108_pwr_priv *priv;
-	int ret;
-
-	ret = nouveau_pwr_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	priv->base.code.data = nv108_pwr_code;
-	priv->base.code.size = sizeof(nv108_pwr_code);
-	priv->base.data.data = nv108_pwr_data;
-	priv->base.data.size = sizeof(nv108_pwr_data);
-	return 0;
-}
-
-struct nouveau_oclass
-nv108_pwr_oclass = {
-	.handle = NV_SUBDEV(PWR, 0x00),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nv108_pwr_ctor,
+struct nouveau_oclass *
+nv108_pwr_oclass = &(struct nvkm_pwr_impl) {
+	.base.handle = NV_SUBDEV(PWR, 0x00),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_pwr_ctor,
 		.dtor = _nouveau_pwr_dtor,
 		.init = _nouveau_pwr_init,
 		.fini = _nouveau_pwr_fini,
 	},
-};
+	.code.data = nv108_pwr_code,
+	.code.size = sizeof(nv108_pwr_code),
+	.data.data = nv108_pwr_data,
+	.data.size = sizeof(nv108_pwr_data),
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/nva3.c
index c132b7ca9747..998d53076b8b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/nva3.c
@@ -22,50 +22,29 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/pwr.h>
-
+#include "priv.h"
 #include "fuc/nva3.fuc.h"
 
-struct nva3_pwr_priv {
-	struct nouveau_pwr base;
-};
-
 static int
 nva3_pwr_init(struct nouveau_object *object)
 {
-	struct nva3_pwr_priv *priv = (void *)object;
-	nv_mask(priv, 0x022210, 0x00000001, 0x00000000);
-	nv_mask(priv, 0x022210, 0x00000001, 0x00000001);
-	return nouveau_pwr_init(&priv->base);
-}
-
-static int
-nva3_pwr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	      struct nouveau_oclass *oclass, void *data, u32 size,
-	      struct nouveau_object **pobject)
-{
-	struct nva3_pwr_priv *priv;
-	int ret;
-
-	ret = nouveau_pwr_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	priv->base.code.data = nva3_pwr_code;
-	priv->base.code.size = sizeof(nva3_pwr_code);
-	priv->base.data.data = nva3_pwr_data;
-	priv->base.data.size = sizeof(nva3_pwr_data);
-	return 0;
+	struct nouveau_pwr *ppwr = (void *)object;
+	nv_mask(ppwr, 0x022210, 0x00000001, 0x00000000);
+	nv_mask(ppwr, 0x022210, 0x00000001, 0x00000001);
+	return nouveau_pwr_init(ppwr);
 }
 
-struct nouveau_oclass
-nva3_pwr_oclass = {
-	.handle = NV_SUBDEV(PWR, 0xa3),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nva3_pwr_ctor,
+struct nouveau_oclass *
+nva3_pwr_oclass = &(struct nvkm_pwr_impl) {
+	.base.handle = NV_SUBDEV(PWR, 0xa3),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_pwr_ctor,
 		.dtor = _nouveau_pwr_dtor,
 		.init = nva3_pwr_init,
 		.fini = _nouveau_pwr_fini,
 	},
-};
+	.code.data = nva3_pwr_code,
+	.code.size = sizeof(nva3_pwr_code),
+	.data.data = nva3_pwr_data,
+	.data.size = sizeof(nva3_pwr_data),
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/nvc0.c
index 495f6857428d..9a773e66efa4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/nvc0.c
@@ -22,41 +22,20 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/pwr.h>
-
+#include "priv.h"
 #include "fuc/nvc0.fuc.h"
 
-struct nvc0_pwr_priv {
-	struct nouveau_pwr base;
-};
-
-static int
-nvc0_pwr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	      struct nouveau_oclass *oclass, void *data, u32 size,
-	      struct nouveau_object **pobject)
-{
-	struct nvc0_pwr_priv *priv;
-	int ret;
-
-	ret = nouveau_pwr_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	priv->base.code.data = nvc0_pwr_code;
-	priv->base.code.size = sizeof(nvc0_pwr_code);
-	priv->base.data.data = nvc0_pwr_data;
-	priv->base.data.size = sizeof(nvc0_pwr_data);
-	return 0;
-}
-
-struct nouveau_oclass
-nvc0_pwr_oclass = {
-	.handle = NV_SUBDEV(PWR, 0xc0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_pwr_ctor,
+struct nouveau_oclass *
+nvc0_pwr_oclass = &(struct nvkm_pwr_impl) {
+	.base.handle = NV_SUBDEV(PWR, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_pwr_ctor,
 		.dtor = _nouveau_pwr_dtor,
 		.init = _nouveau_pwr_init,
 		.fini = _nouveau_pwr_fini,
 	},
-};
+	.code.data = nvc0_pwr_code,
+	.code.size = sizeof(nvc0_pwr_code),
+	.data.data = nvc0_pwr_data,
+	.data.size = sizeof(nvc0_pwr_data),
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/pwr/nvd0.c
index 043aa142fe82..2b29be5d08ac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/pwr/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/nvd0.c
@@ -22,41 +22,20 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/pwr.h>
-
+#include "priv.h"
 #include "fuc/nvd0.fuc.h"
 
-struct nvd0_pwr_priv {
-	struct nouveau_pwr base;
-};
-
-static int
-nvd0_pwr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	      struct nouveau_oclass *oclass, void *data, u32 size,
-	      struct nouveau_object **pobject)
-{
-	struct nvd0_pwr_priv *priv;
-	int ret;
-
-	ret = nouveau_pwr_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	priv->base.code.data = nvd0_pwr_code;
-	priv->base.code.size = sizeof(nvd0_pwr_code);
-	priv->base.data.data = nvd0_pwr_data;
-	priv->base.data.size = sizeof(nvd0_pwr_data);
-	return 0;
-}
-
-struct nouveau_oclass
-nvd0_pwr_oclass = {
-	.handle = NV_SUBDEV(PWR, 0xd0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvd0_pwr_ctor,
+struct nouveau_oclass *
+nvd0_pwr_oclass = &(struct nvkm_pwr_impl) {
+	.base.handle = NV_SUBDEV(PWR, 0xd0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_pwr_ctor,
 		.dtor = _nouveau_pwr_dtor,
 		.init = _nouveau_pwr_init,
 		.fini = _nouveau_pwr_fini,
 	},
-};
+	.code.data = nvd0_pwr_code,
+	.code.size = sizeof(nvd0_pwr_code),
+	.data.data = nvd0_pwr_data,
+	.data.size = sizeof(nvd0_pwr_data),
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/pwr/priv.h b/drivers/gpu/drm/nouveau/core/subdev/pwr/priv.h
new file mode 100644
index 000000000000..3814a341db32
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/pwr/priv.h
@@ -0,0 +1,44 @@
+#ifndef __NVKM_PWR_PRIV_H__
+#define __NVKM_PWR_PRIV_H__
+
+#include <subdev/pwr.h>
+#include <subdev/pwr/fuc/os.h>
+
+#define nouveau_pwr_create(p, e, o, d)                                         \
+	nouveau_pwr_create_((p), (e), (o), sizeof(**d), (void **)d)
+#define nouveau_pwr_destroy(p)                                                 \
+	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_pwr_init(p) ({                                                 \
+	struct nouveau_pwr *_ppwr = (p);                                       \
+	_nouveau_pwr_init(nv_object(_ppwr));                                   \
+})
+#define nouveau_pwr_fini(p,s) ({                                               \
+	struct nouveau_pwr *_ppwr = (p);                                       \
+	_nouveau_pwr_fini(nv_object(_ppwr), (s));                              \
+})
+
+int nouveau_pwr_create_(struct nouveau_object *, struct nouveau_object *,
+			struct nouveau_oclass *, int, void **);
+
+int _nouveau_pwr_ctor(struct nouveau_object *, struct nouveau_object *,
+		      struct nouveau_oclass *, void *, u32,
+		      struct nouveau_object **);
+#define _nouveau_pwr_dtor _nouveau_subdev_dtor
+int _nouveau_pwr_init(struct nouveau_object *);
+int _nouveau_pwr_fini(struct nouveau_object *, bool);
+
+struct nvkm_pwr_impl {
+	struct nouveau_oclass base;
+	struct {
+		u32 *data;
+		u32  size;
+	} code;
+	struct {
+		u32 *data;
+		u32  size;
+	} data;
+
+	void (*pgob)(struct nouveau_pwr *, bool);
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
index 016990a8252c..3656d605168f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fan.c
@@ -31,6 +31,8 @@
 #include <subdev/gpio.h>
 #include <subdev/timer.h>
 
+#include <subdev/bios/fan.h>
+
 static int
 nouveau_fan_update(struct nouveau_fan *fan, bool immediate, int target)
 {
@@ -275,8 +277,11 @@ nouveau_therm_fan_ctor(struct nouveau_therm *therm)
 	/* other random init... */
 	nouveau_therm_fan_set_defaults(therm);
 	nvbios_perf_fan_parse(bios, &priv->fan->perf);
-	if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
-		nv_error(therm, "parsing the thermal table failed\n");
+	if (!nvbios_fan_parse(bios, &priv->fan->bios)) {
+		nv_debug(therm, "parsing the fan table failed\n");
+		if (nvbios_therm_fan_parse(bios, &priv->fan->bios))
+			nv_error(therm, "parsing both fan tables failed\n");
+	}
 	nouveau_therm_fan_safety_checks(therm);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
index 9a5c07340263..c629d7f2a6a4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/fanpwm.c
@@ -25,6 +25,8 @@
 
 #include <core/option.h>
 #include <subdev/gpio.h>
+#include <subdev/bios.h>
+#include <subdev/bios/fan.h>
 
 #include "priv.h"
 
@@ -86,11 +88,15 @@ nouveau_fanpwm_create(struct nouveau_therm *therm, struct dcb_gpio_func *func)
 {
 	struct nouveau_device *device = nv_device(therm);
 	struct nouveau_therm_priv *tpriv = (void *)therm;
+	struct nouveau_bios *bios = nouveau_bios(therm);
 	struct nouveau_fanpwm_priv *priv;
+	struct nvbios_therm_fan fan;
 	u32 divs, duty;
 
+	nvbios_fan_parse(bios, &fan);
+
 	if (!nouveau_boolopt(device->cfgopt, "NvFanPWM", func->param) ||
-	    !therm->pwm_ctrl ||
+	    !therm->pwm_ctrl || fan.type == NVBIOS_THERM_FAN_TOGGLE ||
 	     therm->pwm_get(therm, func->line, &divs, &duty) == -ENODEV)
 		return -ENODEV;
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c b/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c
new file mode 100644
index 000000000000..668cf3322285
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/gm107.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014 Martin Peres
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "priv.h"
+
+struct gm107_therm_priv {
+	struct nouveau_therm_priv base;
+};
+
+static int
+gm107_fan_pwm_ctrl(struct nouveau_therm *therm, int line, bool enable)
+{
+	/* nothing to do, it seems hardwired */
+	return 0;
+}
+
+static int
+gm107_fan_pwm_get(struct nouveau_therm *therm, int line, u32 *divs, u32 *duty)
+{
+	*divs = nv_rd32(therm, 0x10eb20) & 0x1fff;
+	*duty = nv_rd32(therm, 0x10eb24) & 0x1fff;
+	return 0;
+}
+
+static int
+gm107_fan_pwm_set(struct nouveau_therm *therm, int line, u32 divs, u32 duty)
+{
+	nv_mask(therm, 0x10eb10, 0x1fff, divs); /* keep the high bits */
+	nv_wr32(therm, 0x10eb14, duty | 0x80000000);
+	return 0;
+}
+
+static int
+gm107_fan_pwm_clock(struct nouveau_therm *therm, int line)
+{
+	return nv_device(therm)->crystal * 1000;
+}
+
+static int
+gm107_therm_ctor(struct nouveau_object *parent,
+		struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct gm107_therm_priv *priv;
+	int ret;
+
+	ret = nouveau_therm_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.base.pwm_ctrl = gm107_fan_pwm_ctrl;
+	priv->base.base.pwm_get = gm107_fan_pwm_get;
+	priv->base.base.pwm_set = gm107_fan_pwm_set;
+	priv->base.base.pwm_clock = gm107_fan_pwm_clock;
+	priv->base.base.temp_get = nv84_temp_get;
+	priv->base.base.fan_sense = nva3_therm_fan_sense;
+	priv->base.sensor.program_alarms = nouveau_therm_program_alarms_polling;
+	return nouveau_therm_preinit(&priv->base.base);
+}
+
+struct nouveau_oclass
+gm107_therm_oclass = {
+	.handle = NV_SUBDEV(THERM, 0x117),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = gm107_therm_ctor,
+		.dtor = _nouveau_therm_dtor,
+		.init = nvd0_therm_init,
+		.fini = nv84_therm_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
index 1d15c52fad0c..14e2e09bfc24 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nv84.c
@@ -24,6 +24,7 @@
  */
 
 #include "priv.h"
+#include <subdev/fuse.h>
 
 struct nv84_therm_priv {
 	struct nouveau_therm_priv base;
@@ -32,7 +33,25 @@ struct nv84_therm_priv {
 int
 nv84_temp_get(struct nouveau_therm *therm)
 {
-	return nv_rd32(therm, 0x20400);
+	struct nouveau_fuse *fuse = nouveau_fuse(therm);
+
+	if (nv_ro32(fuse, 0x1a8) == 1)
+		return nv_rd32(therm, 0x20400);
+	else
+		return -ENODEV;
+}
+
+void
+nv84_sensor_setup(struct nouveau_therm *therm)
+{
+	struct nouveau_fuse *fuse = nouveau_fuse(therm);
+
+	/* enable temperature reading for cards with insane defaults */
+	if (nv_ro32(fuse, 0x1a8) == 1) {
+		nv_mask(therm, 0x20008, 0x80008000, 0x80000000);
+		nv_mask(therm, 0x2000c, 0x80000003, 0x00000000);
+		mdelay(20); /* wait for the temperature to stabilize */
+	}
 }
 
 static void
@@ -171,6 +190,21 @@ nv84_therm_intr(struct nouveau_subdev *subdev)
 }
 
 static int
+nv84_therm_init(struct nouveau_object *object)
+{
+	struct nv84_therm_priv *priv = (void *)object;
+	int ret;
+
+	ret = nouveau_therm_init(&priv->base.base);
+	if (ret)
+		return ret;
+
+	nv84_sensor_setup(&priv->base.base);
+
+	return 0;
+}
+
+static int
 nv84_therm_ctor(struct nouveau_object *parent,
 		struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
@@ -228,7 +262,7 @@ nv84_therm_oclass = {
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_therm_ctor,
 		.dtor = _nouveau_therm_dtor,
-		.init = _nouveau_therm_init,
+		.init = nv84_therm_init,
 		.fini = nv84_therm_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
index 0478b2e3fb1d..7893357a7e9f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nva3.c
@@ -51,6 +51,8 @@ nva3_therm_init(struct nouveau_object *object)
 	if (ret)
 		return ret;
 
+	nv84_sensor_setup(&priv->base.base);
+
 	/* enable fan tach, count revolutions per-second */
 	nv_mask(priv, 0x00e720, 0x00000003, 0x00000002);
 	if (tach->func != DCB_GPIO_UNUSED) {
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
index bbf117be572f..b70f7cc649b8 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/nvd0.c
@@ -114,7 +114,7 @@ nvd0_fan_pwm_clock(struct nouveau_therm *therm, int line)
 		return nv_device(therm)->crystal * 1000 / 10;
 }
 
-static int
+int
 nvd0_therm_init(struct nouveau_object *object)
 {
 	struct nvd0_therm_priv *priv = (void *)object;
@@ -150,6 +150,8 @@ nvd0_therm_ctor(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
+	nv84_sensor_setup(&priv->base.base);
+
 	priv->base.base.pwm_ctrl = nvd0_fan_pwm_ctrl;
 	priv->base.base.pwm_get = nvd0_fan_pwm_get;
 	priv->base.base.pwm_set = nvd0_fan_pwm_set;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
index 916fca5c7816..7dba8c281a0b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/therm/priv.h
@@ -145,10 +145,13 @@ int nv50_fan_pwm_get(struct nouveau_therm *, int, u32 *, u32 *);
 int nv50_fan_pwm_set(struct nouveau_therm *, int, u32, u32);
 int nv50_fan_pwm_clock(struct nouveau_therm *, int);
 int nv84_temp_get(struct nouveau_therm *therm);
+void nv84_sensor_setup(struct nouveau_therm *therm);
 int nv84_therm_fini(struct nouveau_object *object, bool suspend);
 
 int nva3_therm_fan_sense(struct nouveau_therm *);
 
+int nvd0_therm_init(struct nouveau_object *object);
+
 int nouveau_fanpwm_create(struct nouveau_therm *, struct dcb_gpio_func *);
 int nouveau_fantog_create(struct nouveau_therm *, struct dcb_gpio_func *);
 int nouveau_fannil_create(struct nouveau_therm *);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
index 7dd680ff2f6f..f75a683bd47a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
@@ -296,7 +296,7 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 	int ret;
 
 	mutex_lock(&nv_subdev(vmm)->mutex);
-	ret = nouveau_mm_head(&vm->mm, page_shift, msize, msize, align,
+	ret = nouveau_mm_head(&vm->mm, 0, page_shift, msize, msize, align,
 			     &vma->node);
 	if (unlikely(ret != 0)) {
 		mutex_unlock(&nv_subdev(vmm)->mutex);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
index 668cf964e4a9..2d0988755530 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
@@ -28,7 +28,7 @@
 #include <subdev/timer.h>
 #include <subdev/fb.h>
 #include <subdev/vm.h>
-#include <subdev/ltcg.h>
+#include <subdev/ltc.h>
 #include <subdev/bar.h>
 
 struct nvc0_vmmgr_priv {
@@ -116,12 +116,12 @@ nvc0_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	pte <<= 3;
 
 	if (mem->tag) {
-		struct nouveau_ltcg *ltcg =
-			nouveau_ltcg(vma->vm->vmm->base.base.parent);
+		struct nouveau_ltc *ltc =
+			nouveau_ltc(vma->vm->vmm->base.base.parent);
 		u32 tag = mem->tag->offset + (delta >> 17);
 		phys |= (u64)tag << (32 + 12);
 		next |= (u64)1   << (32 + 12);
-		ltcg->tags_clear(ltcg, tag, cnt);
+		ltc->tags_clear(ltc, tag, cnt);
 	}
 
 	while (cnt--) {
diff --git a/drivers/gpu/drm/nouveau/dispnv04/arb.c b/drivers/gpu/drm/nouveau/dispnv04/arb.c
index 2a15b98b4d2b..c6361422a0b2 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/arb.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/arb.c
@@ -198,12 +198,12 @@ nv04_update_arb(struct drm_device *dev, int VClk, int bpp,
 		int *burst, int *lwm)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nv_fifo_info fifo_data;
 	struct nv_sim_state sim_data;
 	int MClk = nouveau_hw_get_clock(dev, PLL_MEMORY);
 	int NVClk = nouveau_hw_get_clock(dev, PLL_CORE);
-	uint32_t cfg1 = nv_rd32(device, NV04_PFB_CFG1);
+	uint32_t cfg1 = nvif_rd32(device, NV04_PFB_CFG1);
 
 	sim_data.pclk_khz = VClk;
 	sim_data.mclk_khz = MClk;
@@ -221,13 +221,13 @@ nv04_update_arb(struct drm_device *dev, int VClk, int bpp,
 		sim_data.mem_latency = 3;
 		sim_data.mem_page_miss = 10;
 	} else {
-		sim_data.memory_type = nv_rd32(device, NV04_PFB_CFG0) & 0x1;
-		sim_data.memory_width = (nv_rd32(device, NV_PEXTDEV_BOOT_0) & 0x10) ? 128 : 64;
+		sim_data.memory_type = nvif_rd32(device, NV04_PFB_CFG0) & 0x1;
+		sim_data.memory_width = (nvif_rd32(device, NV_PEXTDEV_BOOT_0) & 0x10) ? 128 : 64;
 		sim_data.mem_latency = cfg1 & 0xf;
 		sim_data.mem_page_miss = ((cfg1 >> 4) & 0xf) + ((cfg1 >> 31) & 0x1);
 	}
 
-	if (nv_device(drm->device)->card_type == NV_04)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_TNT)
 		nv04_calc_arb(&fifo_data, &sim_data);
 	else
 		nv10_calc_arb(&fifo_data, &sim_data);
@@ -254,7 +254,7 @@ nouveau_calc_arb(struct drm_device *dev, int vclk, int bpp, int *burst, int *lwm
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
-	if (nv_device(drm->device)->card_type < NV_20)
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_KELVIN)
 		nv04_update_arb(dev, vclk, bpp, burst, lwm);
 	else if ((dev->pdev->device & 0xfff0) == 0x0240 /*CHIPSET_C51*/ ||
 		 (dev->pdev->device & 0xfff0) == 0x03d0 /*CHIPSET_C512*/) {
diff --git a/drivers/gpu/drm/nouveau/dispnv04/crtc.c b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
index 41be3424c906..fca6a1f9c20c 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/crtc.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/crtc.c
@@ -111,8 +111,8 @@ static void nv_crtc_calc_state_ext(struct drm_crtc *crtc, struct drm_display_mod
 {
 	struct drm_device *dev = crtc->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_bios *bios = nouveau_bios(drm->device);
-	struct nouveau_clock *clk = nouveau_clock(drm->device);
+	struct nouveau_bios *bios = nvkm_bios(&drm->device);
+	struct nouveau_clock *clk = nvkm_clock(&drm->device);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	struct nv04_mode_state *state = &nv04_display(dev)->mode_reg;
 	struct nv04_crtc_reg *regp = &state->crtc_reg[nv_crtc->index];
@@ -136,7 +136,7 @@ static void nv_crtc_calc_state_ext(struct drm_crtc *crtc, struct drm_display_mod
 	 * has yet been observed in allowing the use a single stage pll on all
 	 * nv43 however.  the behaviour of single stage use is untested on nv40
 	 */
-	if (nv_device(drm->device)->chipset > 0x40 && dot_clock <= (pll_lim.vco1.max_freq / 2))
+	if (drm->device.info.chipset > 0x40 && dot_clock <= (pll_lim.vco1.max_freq / 2))
 		memset(&pll_lim.vco2, 0, sizeof(pll_lim.vco2));
 
 
@@ -146,10 +146,10 @@ static void nv_crtc_calc_state_ext(struct drm_crtc *crtc, struct drm_display_mod
 	state->pllsel &= PLLSEL_VPLL1_MASK | PLLSEL_VPLL2_MASK | PLLSEL_TV_MASK;
 
 	/* The blob uses this always, so let's do the same */
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		state->pllsel |= NV_PRAMDAC_PLL_COEFF_SELECT_USE_VPLL2_TRUE;
 	/* again nv40 and some nv43 act more like nv3x as described above */
-	if (nv_device(drm->device)->chipset < 0x41)
+	if (drm->device.info.chipset < 0x41)
 		state->pllsel |= NV_PRAMDAC_PLL_COEFF_SELECT_SOURCE_PROG_MPLL |
 				 NV_PRAMDAC_PLL_COEFF_SELECT_SOURCE_PROG_NVPLL;
 	state->pllsel |= nv_crtc->index ? PLLSEL_VPLL2_MASK : PLLSEL_VPLL1_MASK;
@@ -275,7 +275,7 @@ nv_crtc_mode_set_vga(struct drm_crtc *crtc, struct drm_display_mode *mode)
 		horizEnd = horizTotal - 2;
 		horizBlankEnd = horizTotal + 4;
 #if 0
-		if (dev->overlayAdaptor && nv_device(drm->device)->card_type >= NV_10)
+		if (dev->overlayAdaptor && drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS)
 			/* This reportedly works around some video overlay bandwidth problems */
 			horizTotal += 2;
 #endif
@@ -509,7 +509,7 @@ nv_crtc_mode_set_regs(struct drm_crtc *crtc, struct drm_display_mode * mode)
 	regp->cursor_cfg = NV_PCRTC_CURSOR_CONFIG_CUR_LINES_64 |
 			     NV_PCRTC_CURSOR_CONFIG_CUR_PIXELS_64 |
 			     NV_PCRTC_CURSOR_CONFIG_ADDRESS_SPACE_PNVM;
-	if (nv_device(drm->device)->chipset >= 0x11)
+	if (drm->device.info.chipset >= 0x11)
 		regp->cursor_cfg |= NV_PCRTC_CURSOR_CONFIG_CUR_BPP_32;
 	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
 		regp->cursor_cfg |= NV_PCRTC_CURSOR_CONFIG_DOUBLE_SCAN_ENABLE;
@@ -550,26 +550,26 @@ nv_crtc_mode_set_regs(struct drm_crtc *crtc, struct drm_display_mode * mode)
 	 * 1 << 30 on 0x60.830), for no apparent reason */
 	regp->CRTC[NV_CIO_CRE_59] = off_chip_digital;
 
-	if (nv_device(drm->device)->card_type >= NV_30)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 		regp->CRTC[0x9f] = off_chip_digital ? 0x11 : 0x1;
 
 	regp->crtc_830 = mode->crtc_vdisplay - 3;
 	regp->crtc_834 = mode->crtc_vdisplay - 1;
 
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		/* This is what the blob does */
 		regp->crtc_850 = NVReadCRTC(dev, 0, NV_PCRTC_850);
 
-	if (nv_device(drm->device)->card_type >= NV_30)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 		regp->gpio_ext = NVReadCRTC(dev, 0, NV_PCRTC_GPIO_EXT);
 
-	if (nv_device(drm->device)->card_type >= NV_10)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS)
 		regp->crtc_cfg = NV10_PCRTC_CONFIG_START_ADDRESS_HSYNC;
 	else
 		regp->crtc_cfg = NV04_PCRTC_CONFIG_START_ADDRESS_HSYNC;
 
 	/* Some misc regs */
-	if (nv_device(drm->device)->card_type == NV_40) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE) {
 		regp->CRTC[NV_CIO_CRE_85] = 0xFF;
 		regp->CRTC[NV_CIO_CRE_86] = 0x1;
 	}
@@ -581,7 +581,7 @@ nv_crtc_mode_set_regs(struct drm_crtc *crtc, struct drm_display_mode * mode)
 
 	/* Generic PRAMDAC regs */
 
-	if (nv_device(drm->device)->card_type >= NV_10)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS)
 		/* Only bit that bios and blob set. */
 		regp->nv10_cursync = (1 << 25);
 
@@ -590,7 +590,7 @@ nv_crtc_mode_set_regs(struct drm_crtc *crtc, struct drm_display_mode * mode)
 				NV_PRAMDAC_GENERAL_CONTROL_PIXMIX_ON;
 	if (crtc->primary->fb->depth == 16)
 		regp->ramdac_gen_ctrl |= NV_PRAMDAC_GENERAL_CONTROL_ALT_MODE_SEL;
-	if (nv_device(drm->device)->chipset >= 0x11)
+	if (drm->device.info.chipset >= 0x11)
 		regp->ramdac_gen_ctrl |= NV_PRAMDAC_GENERAL_CONTROL_PIPE_LONG;
 
 	regp->ramdac_630 = 0; /* turn off green mode (tv test pattern?) */
@@ -653,7 +653,7 @@ nv_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 
 	nv_crtc_mode_set_vga(crtc, adjusted_mode);
 	/* calculated in nv04_dfp_prepare, nv40 needs it written before calculating PLLs */
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK, nv04_display(dev)->mode_reg.sel_clk);
 	nv_crtc_mode_set_regs(crtc, adjusted_mode);
 	nv_crtc_calc_state_ext(crtc, mode, adjusted_mode->clock);
@@ -714,7 +714,7 @@ static void nv_crtc_prepare(struct drm_crtc *crtc)
 
 	/* Some more preparation. */
 	NVWriteCRTC(dev, nv_crtc->index, NV_PCRTC_CONFIG, NV_PCRTC_CONFIG_START_ADDRESS_NON_VGA);
-	if (nv_device(drm->device)->card_type == NV_40) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE) {
 		uint32_t reg900 = NVReadRAMDAC(dev, nv_crtc->index, NV_PRAMDAC_900);
 		NVWriteRAMDAC(dev, nv_crtc->index, NV_PRAMDAC_900, reg900 & ~0x10000);
 	}
@@ -888,7 +888,7 @@ nv04_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_FF_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_FFLWM__INDEX);
 
-	if (nv_device(drm->device)->card_type >= NV_20) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_KELVIN) {
 		regp->CRTC[NV_CIO_CRE_47] = arb_lwm >> 8;
 		crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_47);
 	}
@@ -915,9 +915,9 @@ nv04_crtc_mode_set_base_atomic(struct drm_crtc *crtc,
 	struct drm_device *dev = drm->dev;
 
 	if (state == ENTER_ATOMIC_MODE_SET)
-		nouveau_fbcon_save_disable_accel(dev);
+		nouveau_fbcon_accel_save_disable(dev);
 	else
-		nouveau_fbcon_restore_accel(dev);
+		nouveau_fbcon_accel_restore(dev);
 
 	return nv04_crtc_do_mode_set_base(crtc, fb, x, y, true);
 }
@@ -969,7 +969,7 @@ static void nv11_cursor_upload(struct drm_device *dev, struct nouveau_bo *src,
 		{
 			struct nouveau_drm *drm = nouveau_drm(dev);
 
-			if (nv_device(drm->device)->chipset == 0x11) {
+			if (drm->device.info.chipset == 0x11) {
 				pixel = ((pixel & 0x000000ff) << 24) |
 					((pixel & 0x0000ff00) << 8) |
 					((pixel & 0x00ff0000) >> 8) |
@@ -1010,7 +1010,7 @@ nv04_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	if (ret)
 		goto out;
 
-	if (nv_device(drm->device)->chipset >= 0x11)
+	if (drm->device.info.chipset >= 0x11)
 		nv11_cursor_upload(dev, cursor, nv_crtc->cursor.nvbo);
 	else
 		nv04_cursor_upload(dev, cursor, nv_crtc->cursor.nvbo);
@@ -1127,7 +1127,7 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num)
 	drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256);
 
 	ret = nouveau_bo_new(dev, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &nv_crtc->cursor.nvbo);
+			     0, 0x0000, NULL, NULL, &nv_crtc->cursor.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
 		if (!ret) {
diff --git a/drivers/gpu/drm/nouveau/dispnv04/cursor.c b/drivers/gpu/drm/nouveau/dispnv04/cursor.c
index a810303169de..4e61173c3353 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/cursor.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/cursor.c
@@ -55,7 +55,7 @@ nv04_cursor_set_offset(struct nouveau_crtc *nv_crtc, uint32_t offset)
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_HCUR_ADDR0_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_HCUR_ADDR1_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_HCUR_ADDR2_INDEX);
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		nv_fix_nv40_hw_cursor(dev, nv_crtc->index);
 }
 
diff --git a/drivers/gpu/drm/nouveau/dispnv04/dac.c b/drivers/gpu/drm/nouveau/dispnv04/dac.c
index a96dda48718e..2d8056cde996 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/dac.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/dac.c
@@ -65,8 +65,8 @@ int nv04_dac_output_offset(struct drm_encoder *encoder)
 
 static int sample_load_twice(struct drm_device *dev, bool sense[2])
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	struct nouveau_timer *ptimer = nouveau_timer(device);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	struct nouveau_timer *ptimer = nvkm_timer(device);
 	int i;
 
 	for (i = 0; i < 2; i++) {
@@ -95,15 +95,15 @@ static int sample_load_twice(struct drm_device *dev, bool sense[2])
 
 		udelay(100);
 		/* when level triggers, sense is _LO_ */
-		sense_a = nv_rd08(device, NV_PRMCIO_INP0) & 0x10;
+		sense_a = nvif_rd08(device, NV_PRMCIO_INP0) & 0x10;
 
 		/* take another reading until it agrees with sense_a... */
 		do {
 			udelay(100);
-			sense_b = nv_rd08(device, NV_PRMCIO_INP0) & 0x10;
+			sense_b = nvif_rd08(device, NV_PRMCIO_INP0) & 0x10;
 			if (sense_a != sense_b) {
 				sense_b_prime =
-					nv_rd08(device, NV_PRMCIO_INP0) & 0x10;
+					nvif_rd08(device, NV_PRMCIO_INP0) & 0x10;
 				if (sense_b == sense_b_prime) {
 					/* ... unless two consecutive subsequent
 					 * samples agree; sense_a is replaced */
@@ -128,7 +128,7 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder,
 						 struct drm_connector *connector)
 {
 	struct drm_device *dev = encoder->dev;
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	uint8_t saved_seq1, saved_pi, saved_rpc1, saved_cr_mode;
 	uint8_t saved_palette0[3], saved_palette_mask;
@@ -164,11 +164,11 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder,
 	saved_rpc1 = NVReadVgaCrtc(dev, 0, NV_CIO_CRE_RPC1_INDEX);
 	NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_RPC1_INDEX, saved_rpc1 & ~0xc0);
 
-	nv_wr08(device, NV_PRMDIO_READ_MODE_ADDRESS, 0x0);
+	nvif_wr08(device, NV_PRMDIO_READ_MODE_ADDRESS, 0x0);
 	for (i = 0; i < 3; i++)
-		saved_palette0[i] = nv_rd08(device, NV_PRMDIO_PALETTE_DATA);
-	saved_palette_mask = nv_rd08(device, NV_PRMDIO_PIXEL_MASK);
-	nv_wr08(device, NV_PRMDIO_PIXEL_MASK, 0);
+		saved_palette0[i] = nvif_rd08(device, NV_PRMDIO_PALETTE_DATA);
+	saved_palette_mask = nvif_rd08(device, NV_PRMDIO_PIXEL_MASK);
+	nvif_wr08(device, NV_PRMDIO_PIXEL_MASK, 0);
 
 	saved_rgen_ctrl = NVReadRAMDAC(dev, 0, NV_PRAMDAC_GENERAL_CONTROL);
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_GENERAL_CONTROL,
@@ -181,11 +181,11 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder,
 	do {
 		bool sense_pair[2];
 
-		nv_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS, 0);
-		nv_wr08(device, NV_PRMDIO_PALETTE_DATA, 0);
-		nv_wr08(device, NV_PRMDIO_PALETTE_DATA, 0);
+		nvif_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS, 0);
+		nvif_wr08(device, NV_PRMDIO_PALETTE_DATA, 0);
+		nvif_wr08(device, NV_PRMDIO_PALETTE_DATA, 0);
 		/* testing blue won't find monochrome monitors.  I don't care */
-		nv_wr08(device, NV_PRMDIO_PALETTE_DATA, blue);
+		nvif_wr08(device, NV_PRMDIO_PALETTE_DATA, blue);
 
 		i = 0;
 		/* take sample pairs until both samples in the pair agree */
@@ -208,11 +208,11 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder,
 	} while (++blue < 0x18 && sense);
 
 out:
-	nv_wr08(device, NV_PRMDIO_PIXEL_MASK, saved_palette_mask);
+	nvif_wr08(device, NV_PRMDIO_PIXEL_MASK, saved_palette_mask);
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_GENERAL_CONTROL, saved_rgen_ctrl);
-	nv_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS, 0);
+	nvif_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS, 0);
 	for (i = 0; i < 3; i++)
-		nv_wr08(device, NV_PRMDIO_PALETTE_DATA, saved_palette0[i]);
+		nvif_wr08(device, NV_PRMDIO_PALETTE_DATA, saved_palette0[i]);
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL, saved_rtest_ctrl);
 	NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_PIXEL_INDEX, saved_pi);
 	NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_RPC1_INDEX, saved_rpc1);
@@ -231,8 +231,8 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nouveau_dev(dev);
-	struct nouveau_gpio *gpio = nouveau_gpio(device);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	struct nouveau_gpio *gpio = nvkm_gpio(device);
 	struct dcb_output *dcb = nouveau_encoder(encoder)->dcb;
 	uint32_t sample, testval, regoffset = nv04_dac_output_offset(encoder);
 	uint32_t saved_powerctrl_2 = 0, saved_powerctrl_4 = 0, saved_routput,
@@ -256,12 +256,12 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + regoffset,
 		      saved_rtest_ctrl & ~NV_PRAMDAC_TEST_CONTROL_PWRDWN_DAC_OFF);
 
-	saved_powerctrl_2 = nv_rd32(device, NV_PBUS_POWERCTRL_2);
+	saved_powerctrl_2 = nvif_rd32(device, NV_PBUS_POWERCTRL_2);
 
-	nv_wr32(device, NV_PBUS_POWERCTRL_2, saved_powerctrl_2 & 0xd7ffffff);
+	nvif_wr32(device, NV_PBUS_POWERCTRL_2, saved_powerctrl_2 & 0xd7ffffff);
 	if (regoffset == 0x68) {
-		saved_powerctrl_4 = nv_rd32(device, NV_PBUS_POWERCTRL_4);
-		nv_wr32(device, NV_PBUS_POWERCTRL_4, saved_powerctrl_4 & 0xffffffcf);
+		saved_powerctrl_4 = nvif_rd32(device, NV_PBUS_POWERCTRL_4);
+		nvif_wr32(device, NV_PBUS_POWERCTRL_4, saved_powerctrl_4 & 0xffffffcf);
 	}
 
 	if (gpio) {
@@ -283,7 +283,7 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 	/* nv driver and nv31 use 0xfffffeee, nv34 and 6600 use 0xfffffece */
 	routput = (saved_routput & 0xfffffece) | head << 8;
 
-	if (nv_device(drm->device)->card_type >= NV_40) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CURIE) {
 		if (dcb->type == DCB_OUTPUT_TV)
 			routput |= 0x1a << 16;
 		else
@@ -316,8 +316,8 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_DACCLK + regoffset, saved_routput);
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + regoffset, saved_rtest_ctrl);
 	if (regoffset == 0x68)
-		nv_wr32(device, NV_PBUS_POWERCTRL_4, saved_powerctrl_4);
-	nv_wr32(device, NV_PBUS_POWERCTRL_2, saved_powerctrl_2);
+		nvif_wr32(device, NV_PBUS_POWERCTRL_4, saved_powerctrl_4);
+	nvif_wr32(device, NV_PBUS_POWERCTRL_2, saved_powerctrl_2);
 
 	if (gpio) {
 		gpio->set(gpio, 0, DCB_GPIO_TVDAC1, 0xff, saved_gpio1);
@@ -398,7 +398,7 @@ static void nv04_dac_mode_set(struct drm_encoder *encoder,
 	}
 
 	/* This could use refinement for flatpanels, but it should work this way */
-	if (nv_device(drm->device)->chipset < 0x44)
+	if (drm->device.info.chipset < 0x44)
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + nv04_dac_output_offset(encoder), 0xf0000000);
 	else
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + nv04_dac_output_offset(encoder), 0x00100000);
diff --git a/drivers/gpu/drm/nouveau/dispnv04/dfp.c b/drivers/gpu/drm/nouveau/dispnv04/dfp.c
index e57babb206d3..42a5435259f7 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/dfp.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/dfp.c
@@ -281,7 +281,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 			      struct drm_display_mode *adjusted_mode)
 {
 	struct drm_device *dev = encoder->dev;
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
 	struct nv04_crtc_reg *regp = &nv04_display(dev)->mode_reg.crtc_reg[nv_crtc->index];
@@ -335,7 +335,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 		regp->fp_control |= NV_PRAMDAC_FP_TG_CONTROL_MODE_NATIVE;
 	else /* gpu needs to scale */
 		regp->fp_control |= NV_PRAMDAC_FP_TG_CONTROL_MODE_SCALE;
-	if (nv_rd32(device, NV_PEXTDEV_BOOT_0) & NV_PEXTDEV_BOOT_0_STRAP_FP_IFACE_12BIT)
+	if (nvif_rd32(device, NV_PEXTDEV_BOOT_0) & NV_PEXTDEV_BOOT_0_STRAP_FP_IFACE_12BIT)
 		regp->fp_control |= NV_PRAMDAC_FP_TG_CONTROL_WIDTH_12;
 	if (nv_encoder->dcb->location != DCB_LOC_ON_CHIP &&
 	    output_mode->clock > 165000)
@@ -416,7 +416,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 	if ((nv_connector->dithering_mode == DITHERING_MODE_ON) ||
 	    (nv_connector->dithering_mode == DITHERING_MODE_AUTO &&
 	     encoder->crtc->primary->fb->depth > connector->display_info.bpc * 3)) {
-		if (nv_device(drm->device)->chipset == 0x11)
+		if (drm->device.info.chipset == 0x11)
 			regp->dither = savep->dither | 0x00010000;
 		else {
 			int i;
@@ -427,7 +427,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 			}
 		}
 	} else {
-		if (nv_device(drm->device)->chipset != 0x11) {
+		if (drm->device.info.chipset != 0x11) {
 			/* reset them */
 			int i;
 			for (i = 0; i < 3; i++) {
@@ -463,7 +463,7 @@ static void nv04_dfp_commit(struct drm_encoder *encoder)
 		NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_TG_CONTROL);
 
 	/* This could use refinement for flatpanels, but it should work this way */
-	if (nv_device(drm->device)->chipset < 0x44)
+	if (drm->device.info.chipset < 0x44)
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + nv04_dac_output_offset(encoder), 0xf0000000);
 	else
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + nv04_dac_output_offset(encoder), 0x00100000);
@@ -485,7 +485,7 @@ static void nv04_dfp_update_backlight(struct drm_encoder *encoder, int mode)
 {
 #ifdef __powerpc__
 	struct drm_device *dev = encoder->dev;
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 
 	/* BIOS scripts usually take care of the backlight, thanks
 	 * Apple for your consistency.
@@ -623,7 +623,7 @@ static void nv04_tmds_slave_init(struct drm_encoder *encoder)
 	struct drm_device *dev = encoder->dev;
 	struct dcb_output *dcb = nouveau_encoder(encoder)->dcb;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct nouveau_i2c_port *port = i2c->find(i2c, 2);
 	struct nouveau_i2c_board_info info[] = {
 		{
diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c
index 4342fdaee707..3d0afa1c6cff 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/disp.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c
@@ -22,9 +22,6 @@
  * Author: Ben Skeggs
  */
 
-#include <core/object.h>
-#include <core/class.h>
-
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
@@ -34,8 +31,6 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 
-#include <subdev/i2c.h>
-
 int
 nv04_display_early_init(struct drm_device *dev)
 {
@@ -58,7 +53,7 @@ int
 nv04_display_create(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct dcb_table *dcb = &drm->vbios.dcb;
 	struct drm_connector *connector, *ct;
 	struct drm_encoder *encoder;
@@ -70,6 +65,8 @@ nv04_display_create(struct drm_device *dev)
 	if (!disp)
 		return -ENOMEM;
 
+	nvif_object_map(nvif_object(&drm->device));
+
 	nouveau_display(dev)->priv = disp;
 	nouveau_display(dev)->dtor = nv04_display_destroy;
 	nouveau_display(dev)->init = nv04_display_init;
@@ -144,6 +141,7 @@ void
 nv04_display_destroy(struct drm_device *dev)
 {
 	struct nv04_display *disp = nv04_display(dev);
+	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct drm_encoder *encoder;
 	struct drm_crtc *crtc;
 
@@ -170,6 +168,8 @@ nv04_display_destroy(struct drm_device *dev)
 
 	nouveau_display(dev)->priv = NULL;
 	kfree(disp);
+
+	nvif_object_unmap(nvif_object(&drm->device));
 }
 
 int
diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.h b/drivers/gpu/drm/nouveau/dispnv04/disp.h
index 4245fc3dab70..17b899d9aba3 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/disp.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/disp.h
@@ -131,7 +131,7 @@ nv_two_heads(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	const int impl = dev->pdev->device & 0x0ff0;
 
-	if (nv_device(drm->device)->card_type >= NV_10 && impl != 0x0100 &&
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS && impl != 0x0100 &&
 	    impl != 0x0150 && impl != 0x01a0 && impl != 0x0200)
 		return true;
 
@@ -150,7 +150,7 @@ nv_two_reg_pll(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	const int impl = dev->pdev->device & 0x0ff0;
 
-	if (impl == 0x0310 || impl == 0x0340 || nv_device(drm->device)->card_type >= NV_40)
+	if (impl == 0x0310 || impl == 0x0340 || drm->device.info.family >= NV_DEVICE_INFO_V0_CURIE)
 		return true;
 	return false;
 }
@@ -171,8 +171,8 @@ static inline void
 nouveau_bios_run_init_table(struct drm_device *dev, u16 table,
 			    struct dcb_output *outp, int crtc)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	struct nouveau_bios *bios = nouveau_bios(device);
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_bios *bios = nvkm_bios(&drm->device);
 	struct nvbios_init init = {
 		.subdev = nv_subdev(bios),
 		.bios = bios,
diff --git a/drivers/gpu/drm/nouveau/dispnv04/hw.c b/drivers/gpu/drm/nouveau/dispnv04/hw.c
index aca76af115b3..3d4c19300768 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/hw.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/hw.c
@@ -27,9 +27,6 @@
 #include "hw.h"
 
 #include <subdev/bios/pll.h>
-#include <subdev/fb.h>
-#include <subdev/clock.h>
-#include <subdev/timer.h>
 
 #define CHIPSET_NFORCE 0x01a0
 #define CHIPSET_NFORCE2 0x01f0
@@ -92,7 +89,7 @@ NVSetOwner(struct drm_device *dev, int owner)
 	if (owner == 1)
 		owner *= 3;
 
-	if (nv_device(drm->device)->chipset == 0x11) {
+	if (drm->device.info.chipset == 0x11) {
 		/* This might seem stupid, but the blob does it and
 		 * omitting it often locks the system up.
 		 */
@@ -103,7 +100,7 @@ NVSetOwner(struct drm_device *dev, int owner)
 	/* CR44 is always changed on CRTC0 */
 	NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_44, owner);
 
-	if (nv_device(drm->device)->chipset == 0x11) {	/* set me harder */
+	if (drm->device.info.chipset == 0x11) {	/* set me harder */
 		NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_2E, owner);
 		NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_2E, owner);
 	}
@@ -152,7 +149,7 @@ nouveau_hw_decode_pll(struct drm_device *dev, uint32_t reg1, uint32_t pll1,
 		pllvals->NM1 = pll1 & 0xffff;
 		if (nv_two_reg_pll(dev) && pll2 & NV31_RAMDAC_ENABLE_VCO2)
 			pllvals->NM2 = pll2 & 0xffff;
-		else if (nv_device(drm->device)->chipset == 0x30 || nv_device(drm->device)->chipset == 0x35) {
+		else if (drm->device.info.chipset == 0x30 || drm->device.info.chipset == 0x35) {
 			pllvals->M1 &= 0xf; /* only 4 bits */
 			if (pll1 & NV30_RAMDAC_ENABLE_VCO2) {
 				pllvals->M2 = (pll1 >> 4) & 0x7;
@@ -168,8 +165,8 @@ nouveau_hw_get_pllvals(struct drm_device *dev, enum nvbios_pll_type plltype,
 		       struct nouveau_pll_vals *pllvals)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_bios *bios = nouveau_bios(device);
+	struct nvif_device *device = &drm->device;
+	struct nouveau_bios *bios = nvkm_bios(device);
 	uint32_t reg1, pll1, pll2 = 0;
 	struct nvbios_pll pll_lim;
 	int ret;
@@ -178,16 +175,16 @@ nouveau_hw_get_pllvals(struct drm_device *dev, enum nvbios_pll_type plltype,
 	if (ret || !(reg1 = pll_lim.reg))
 		return -ENOENT;
 
-	pll1 = nv_rd32(device, reg1);
+	pll1 = nvif_rd32(device, reg1);
 	if (reg1 <= 0x405c)
-		pll2 = nv_rd32(device, reg1 + 4);
+		pll2 = nvif_rd32(device, reg1 + 4);
 	else if (nv_two_reg_pll(dev)) {
 		uint32_t reg2 = reg1 + (reg1 == NV_RAMDAC_VPLL2 ? 0x5c : 0x70);
 
-		pll2 = nv_rd32(device, reg2);
+		pll2 = nvif_rd32(device, reg2);
 	}
 
-	if (nv_device(drm->device)->card_type == 0x40 && reg1 >= NV_PRAMDAC_VPLL_COEFF) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CELSIUS && reg1 >= NV_PRAMDAC_VPLL_COEFF) {
 		uint32_t ramdac580 = NVReadRAMDAC(dev, 0, NV_PRAMDAC_580);
 
 		/* check whether vpll has been forced into single stage mode */
@@ -255,9 +252,9 @@ nouveau_hw_fix_bad_vpll(struct drm_device *dev, int head)
 	 */
 
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_clock *clk = nouveau_clock(device);
-	struct nouveau_bios *bios = nouveau_bios(device);
+	struct nvif_device *device = &drm->device;
+	struct nouveau_clock *clk = nvkm_clock(device);
+	struct nouveau_bios *bios = nvkm_bios(device);
 	struct nvbios_pll pll_lim;
 	struct nouveau_pll_vals pv;
 	enum nvbios_pll_type pll = head ? PLL_VPLL1 : PLL_VPLL0;
@@ -394,21 +391,21 @@ nv_save_state_ramdac(struct drm_device *dev, int head,
 	struct nv04_crtc_reg *regp = &state->crtc_reg[head];
 	int i;
 
-	if (nv_device(drm->device)->card_type >= NV_10)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS)
 		regp->nv10_cursync = NVReadRAMDAC(dev, head, NV_RAMDAC_NV10_CURSYNC);
 
 	nouveau_hw_get_pllvals(dev, head ? PLL_VPLL1 : PLL_VPLL0, &regp->pllvals);
 	state->pllsel = NVReadRAMDAC(dev, 0, NV_PRAMDAC_PLL_COEFF_SELECT);
 	if (nv_two_heads(dev))
 		state->sel_clk = NVReadRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK);
-	if (nv_device(drm->device)->chipset == 0x11)
+	if (drm->device.info.chipset == 0x11)
 		regp->dither = NVReadRAMDAC(dev, head, NV_RAMDAC_DITHER_NV11);
 
 	regp->ramdac_gen_ctrl = NVReadRAMDAC(dev, head, NV_PRAMDAC_GENERAL_CONTROL);
 
 	if (nv_gf4_disp_arch(dev))
 		regp->ramdac_630 = NVReadRAMDAC(dev, head, NV_PRAMDAC_630);
-	if (nv_device(drm->device)->chipset >= 0x30)
+	if (drm->device.info.chipset >= 0x30)
 		regp->ramdac_634 = NVReadRAMDAC(dev, head, NV_PRAMDAC_634);
 
 	regp->tv_setup = NVReadRAMDAC(dev, head, NV_PRAMDAC_TV_SETUP);
@@ -450,7 +447,7 @@ nv_save_state_ramdac(struct drm_device *dev, int head,
 	if (nv_gf4_disp_arch(dev))
 		regp->ramdac_8c0 = NVReadRAMDAC(dev, head, NV_PRAMDAC_8C0);
 
-	if (nv_device(drm->device)->card_type == NV_40) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE) {
 		regp->ramdac_a20 = NVReadRAMDAC(dev, head, NV_PRAMDAC_A20);
 		regp->ramdac_a24 = NVReadRAMDAC(dev, head, NV_PRAMDAC_A24);
 		regp->ramdac_a34 = NVReadRAMDAC(dev, head, NV_PRAMDAC_A34);
@@ -466,26 +463,26 @@ nv_load_state_ramdac(struct drm_device *dev, int head,
 		     struct nv04_mode_state *state)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_clock *clk = nouveau_clock(drm->device);
+	struct nouveau_clock *clk = nvkm_clock(&drm->device);
 	struct nv04_crtc_reg *regp = &state->crtc_reg[head];
 	uint32_t pllreg = head ? NV_RAMDAC_VPLL2 : NV_PRAMDAC_VPLL_COEFF;
 	int i;
 
-	if (nv_device(drm->device)->card_type >= NV_10)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS)
 		NVWriteRAMDAC(dev, head, NV_RAMDAC_NV10_CURSYNC, regp->nv10_cursync);
 
 	clk->pll_prog(clk, pllreg, &regp->pllvals);
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_PLL_COEFF_SELECT, state->pllsel);
 	if (nv_two_heads(dev))
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK, state->sel_clk);
-	if (nv_device(drm->device)->chipset == 0x11)
+	if (drm->device.info.chipset == 0x11)
 		NVWriteRAMDAC(dev, head, NV_RAMDAC_DITHER_NV11, regp->dither);
 
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_GENERAL_CONTROL, regp->ramdac_gen_ctrl);
 
 	if (nv_gf4_disp_arch(dev))
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_630, regp->ramdac_630);
-	if (nv_device(drm->device)->chipset >= 0x30)
+	if (drm->device.info.chipset >= 0x30)
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_634, regp->ramdac_634);
 
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_TV_SETUP, regp->tv_setup);
@@ -522,7 +519,7 @@ nv_load_state_ramdac(struct drm_device *dev, int head,
 	if (nv_gf4_disp_arch(dev))
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_8C0, regp->ramdac_8c0);
 
-	if (nv_device(drm->device)->card_type == NV_40) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE) {
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_A20, regp->ramdac_a20);
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_A24, regp->ramdac_a24);
 		NVWriteRAMDAC(dev, head, NV_PRAMDAC_A34, regp->ramdac_a34);
@@ -603,10 +600,10 @@ nv_save_state_ext(struct drm_device *dev, int head,
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_FFLWM__INDEX);
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_21);
 
-	if (nv_device(drm->device)->card_type >= NV_20)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_KELVIN)
 		rd_cio_state(dev, head, regp, NV_CIO_CRE_47);
 
-	if (nv_device(drm->device)->card_type >= NV_30)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 		rd_cio_state(dev, head, regp, 0x9f);
 
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_49);
@@ -615,14 +612,14 @@ nv_save_state_ext(struct drm_device *dev, int head,
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR2_INDEX);
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_ILACE__INDEX);
 
-	if (nv_device(drm->device)->card_type >= NV_10) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS) {
 		regp->crtc_830 = NVReadCRTC(dev, head, NV_PCRTC_830);
 		regp->crtc_834 = NVReadCRTC(dev, head, NV_PCRTC_834);
 
-		if (nv_device(drm->device)->card_type >= NV_30)
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 			regp->gpio_ext = NVReadCRTC(dev, head, NV_PCRTC_GPIO_EXT);
 
-		if (nv_device(drm->device)->card_type == NV_40)
+		if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 			regp->crtc_850 = NVReadCRTC(dev, head, NV_PCRTC_850);
 
 		if (nv_two_heads(dev))
@@ -634,7 +631,7 @@ nv_save_state_ext(struct drm_device *dev, int head,
 
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_SCRATCH3__INDEX);
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_SCRATCH4__INDEX);
-	if (nv_device(drm->device)->card_type >= NV_10) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS) {
 		rd_cio_state(dev, head, regp, NV_CIO_CRE_EBR_INDEX);
 		rd_cio_state(dev, head, regp, NV_CIO_CRE_CSB);
 		rd_cio_state(dev, head, regp, NV_CIO_CRE_4B);
@@ -663,14 +660,13 @@ nv_load_state_ext(struct drm_device *dev, int head,
 		  struct nv04_mode_state *state)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_timer *ptimer = nouveau_timer(device);
-	struct nouveau_fb *pfb = nouveau_fb(device);
+	struct nvif_device *device = &drm->device;
+	struct nouveau_timer *ptimer = nvkm_timer(device);
 	struct nv04_crtc_reg *regp = &state->crtc_reg[head];
 	uint32_t reg900;
 	int i;
 
-	if (nv_device(drm->device)->card_type >= NV_10) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS) {
 		if (nv_two_heads(dev))
 			/* setting ENGINE_CTRL (EC) *must* come before
 			 * CIO_CRE_LCD, as writing CRE_LCD sets bits 16 & 17 in
@@ -678,24 +674,24 @@ nv_load_state_ext(struct drm_device *dev, int head,
 			 */
 			NVWriteCRTC(dev, head, NV_PCRTC_ENGINE_CTRL, regp->crtc_eng_ctrl);
 
-		nv_wr32(device, NV_PVIDEO_STOP, 1);
-		nv_wr32(device, NV_PVIDEO_INTR_EN, 0);
-		nv_wr32(device, NV_PVIDEO_OFFSET_BUFF(0), 0);
-		nv_wr32(device, NV_PVIDEO_OFFSET_BUFF(1), 0);
-		nv_wr32(device, NV_PVIDEO_LIMIT(0), pfb->ram->size - 1);
-		nv_wr32(device, NV_PVIDEO_LIMIT(1), pfb->ram->size - 1);
-		nv_wr32(device, NV_PVIDEO_UVPLANE_LIMIT(0), pfb->ram->size - 1);
-		nv_wr32(device, NV_PVIDEO_UVPLANE_LIMIT(1), pfb->ram->size - 1);
-		nv_wr32(device, NV_PBUS_POWERCTRL_2, 0);
+		nvif_wr32(device, NV_PVIDEO_STOP, 1);
+		nvif_wr32(device, NV_PVIDEO_INTR_EN, 0);
+		nvif_wr32(device, NV_PVIDEO_OFFSET_BUFF(0), 0);
+		nvif_wr32(device, NV_PVIDEO_OFFSET_BUFF(1), 0);
+		nvif_wr32(device, NV_PVIDEO_LIMIT(0), device->info.ram_size - 1);
+		nvif_wr32(device, NV_PVIDEO_LIMIT(1), device->info.ram_size - 1);
+		nvif_wr32(device, NV_PVIDEO_UVPLANE_LIMIT(0), device->info.ram_size - 1);
+		nvif_wr32(device, NV_PVIDEO_UVPLANE_LIMIT(1), device->info.ram_size - 1);
+		nvif_wr32(device, NV_PBUS_POWERCTRL_2, 0);
 
 		NVWriteCRTC(dev, head, NV_PCRTC_CURSOR_CONFIG, regp->cursor_cfg);
 		NVWriteCRTC(dev, head, NV_PCRTC_830, regp->crtc_830);
 		NVWriteCRTC(dev, head, NV_PCRTC_834, regp->crtc_834);
 
-		if (nv_device(drm->device)->card_type >= NV_30)
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 			NVWriteCRTC(dev, head, NV_PCRTC_GPIO_EXT, regp->gpio_ext);
 
-		if (nv_device(drm->device)->card_type == NV_40) {
+		if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE) {
 			NVWriteCRTC(dev, head, NV_PCRTC_850, regp->crtc_850);
 
 			reg900 = NVReadRAMDAC(dev, head, NV_PRAMDAC_900);
@@ -718,23 +714,23 @@ nv_load_state_ext(struct drm_device *dev, int head,
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_FF_INDEX);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_FFLWM__INDEX);
 
-	if (nv_device(drm->device)->card_type >= NV_20)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_KELVIN)
 		wr_cio_state(dev, head, regp, NV_CIO_CRE_47);
 
-	if (nv_device(drm->device)->card_type >= NV_30)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE)
 		wr_cio_state(dev, head, regp, 0x9f);
 
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_49);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR0_INDEX);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR1_INDEX);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR2_INDEX);
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		nv_fix_nv40_hw_cursor(dev, head);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_ILACE__INDEX);
 
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_SCRATCH3__INDEX);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_SCRATCH4__INDEX);
-	if (nv_device(drm->device)->card_type >= NV_10) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS) {
 		wr_cio_state(dev, head, regp, NV_CIO_CRE_EBR_INDEX);
 		wr_cio_state(dev, head, regp, NV_CIO_CRE_CSB);
 		wr_cio_state(dev, head, regp, NV_CIO_CRE_4B);
@@ -742,7 +738,7 @@ nv_load_state_ext(struct drm_device *dev, int head,
 	}
 	/* NV11 and NV20 stop at 0x52. */
 	if (nv_gf4_disp_arch(dev)) {
-		if (nv_device(drm->device)->card_type < NV_20) {
+		if (drm->device.info.family < NV_DEVICE_INFO_V0_KELVIN) {
 			/* Not waiting for vertical retrace before modifying
 			   CRE_53/CRE_54 causes lockups. */
 			nouveau_timer_wait_eq(ptimer, 650000000, NV_PRMCIO_INP0__COLOR, 0x8, 0x8);
@@ -769,15 +765,15 @@ static void
 nv_save_state_palette(struct drm_device *dev, int head,
 		      struct nv04_mode_state *state)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	int head_offset = head * NV_PRMDIO_SIZE, i;
 
-	nv_wr08(device, NV_PRMDIO_PIXEL_MASK + head_offset,
+	nvif_wr08(device, NV_PRMDIO_PIXEL_MASK + head_offset,
 				NV_PRMDIO_PIXEL_MASK_MASK);
-	nv_wr08(device, NV_PRMDIO_READ_MODE_ADDRESS + head_offset, 0x0);
+	nvif_wr08(device, NV_PRMDIO_READ_MODE_ADDRESS + head_offset, 0x0);
 
 	for (i = 0; i < 768; i++) {
-		state->crtc_reg[head].DAC[i] = nv_rd08(device,
+		state->crtc_reg[head].DAC[i] = nvif_rd08(device,
 				NV_PRMDIO_PALETTE_DATA + head_offset);
 	}
 
@@ -788,15 +784,15 @@ void
 nouveau_hw_load_state_palette(struct drm_device *dev, int head,
 			      struct nv04_mode_state *state)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	int head_offset = head * NV_PRMDIO_SIZE, i;
 
-	nv_wr08(device, NV_PRMDIO_PIXEL_MASK + head_offset,
+	nvif_wr08(device, NV_PRMDIO_PIXEL_MASK + head_offset,
 				NV_PRMDIO_PIXEL_MASK_MASK);
-	nv_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS + head_offset, 0x0);
+	nvif_wr08(device, NV_PRMDIO_WRITE_MODE_ADDRESS + head_offset, 0x0);
 
 	for (i = 0; i < 768; i++) {
-		nv_wr08(device, NV_PRMDIO_PALETTE_DATA + head_offset,
+		nvif_wr08(device, NV_PRMDIO_PALETTE_DATA + head_offset,
 				state->crtc_reg[head].DAC[i]);
 	}
 
@@ -808,7 +804,7 @@ void nouveau_hw_save_state(struct drm_device *dev, int head,
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
-	if (nv_device(drm->device)->chipset == 0x11)
+	if (drm->device.info.chipset == 0x11)
 		/* NB: no attempt is made to restore the bad pll later on */
 		nouveau_hw_fix_bad_vpll(dev, head);
 	nv_save_state_ramdac(dev, head, state);
diff --git a/drivers/gpu/drm/nouveau/dispnv04/hw.h b/drivers/gpu/drm/nouveau/dispnv04/hw.h
index eeb70d912d99..7f53c571f31f 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/hw.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/hw.h
@@ -60,41 +60,41 @@ extern void nouveau_calc_arb(struct drm_device *, int vclk, int bpp,
 static inline uint32_t NVReadCRTC(struct drm_device *dev,
 					int head, uint32_t reg)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	uint32_t val;
 	if (head)
 		reg += NV_PCRTC0_SIZE;
-	val = nv_rd32(device, reg);
+	val = nvif_rd32(device, reg);
 	return val;
 }
 
 static inline void NVWriteCRTC(struct drm_device *dev,
 					int head, uint32_t reg, uint32_t val)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	if (head)
 		reg += NV_PCRTC0_SIZE;
-	nv_wr32(device, reg, val);
+	nvif_wr32(device, reg, val);
 }
 
 static inline uint32_t NVReadRAMDAC(struct drm_device *dev,
 					int head, uint32_t reg)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	uint32_t val;
 	if (head)
 		reg += NV_PRAMDAC0_SIZE;
-	val = nv_rd32(device, reg);
+	val = nvif_rd32(device, reg);
 	return val;
 }
 
 static inline void NVWriteRAMDAC(struct drm_device *dev,
 					int head, uint32_t reg, uint32_t val)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	if (head)
 		reg += NV_PRAMDAC0_SIZE;
-	nv_wr32(device, reg, val);
+	nvif_wr32(device, reg, val);
 }
 
 static inline uint8_t nv_read_tmds(struct drm_device *dev,
@@ -120,18 +120,18 @@ static inline void nv_write_tmds(struct drm_device *dev,
 static inline void NVWriteVgaCrtc(struct drm_device *dev,
 					int head, uint8_t index, uint8_t value)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	nv_wr08(device, NV_PRMCIO_CRX__COLOR + head * NV_PRMCIO_SIZE, index);
-	nv_wr08(device, NV_PRMCIO_CR__COLOR + head * NV_PRMCIO_SIZE, value);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	nvif_wr08(device, NV_PRMCIO_CRX__COLOR + head * NV_PRMCIO_SIZE, index);
+	nvif_wr08(device, NV_PRMCIO_CR__COLOR + head * NV_PRMCIO_SIZE, value);
 }
 
 static inline uint8_t NVReadVgaCrtc(struct drm_device *dev,
 					int head, uint8_t index)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	uint8_t val;
-	nv_wr08(device, NV_PRMCIO_CRX__COLOR + head * NV_PRMCIO_SIZE, index);
-	val = nv_rd08(device, NV_PRMCIO_CR__COLOR + head * NV_PRMCIO_SIZE);
+	nvif_wr08(device, NV_PRMCIO_CRX__COLOR + head * NV_PRMCIO_SIZE, index);
+	val = nvif_rd08(device, NV_PRMCIO_CR__COLOR + head * NV_PRMCIO_SIZE);
 	return val;
 }
 
@@ -165,74 +165,74 @@ static inline uint8_t NVReadVgaCrtc5758(struct drm_device *dev, int head, uint8_
 static inline uint8_t NVReadPRMVIO(struct drm_device *dev,
 					int head, uint32_t reg)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	uint8_t val;
 
 	/* Only NV4x have two pvio ranges; other twoHeads cards MUST call
 	 * NVSetOwner for the relevant head to be programmed */
-	if (head && nv_device(drm->device)->card_type == NV_40)
+	if (head && drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		reg += NV_PRMVIO_SIZE;
 
-	val = nv_rd08(device, reg);
+	val = nvif_rd08(device, reg);
 	return val;
 }
 
 static inline void NVWritePRMVIO(struct drm_device *dev,
 					int head, uint32_t reg, uint8_t value)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
 	/* Only NV4x have two pvio ranges; other twoHeads cards MUST call
 	 * NVSetOwner for the relevant head to be programmed */
-	if (head && nv_device(drm->device)->card_type == NV_40)
+	if (head && drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		reg += NV_PRMVIO_SIZE;
 
-	nv_wr08(device, reg, value);
+	nvif_wr08(device, reg, value);
 }
 
 static inline void NVSetEnablePalette(struct drm_device *dev, int head, bool enable)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	nv_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
-	nv_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, enable ? 0 : 0x20);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	nvif_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
+	nvif_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, enable ? 0 : 0x20);
 }
 
 static inline bool NVGetEnablePalette(struct drm_device *dev, int head)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	nv_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
-	return !(nv_rd08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE) & 0x20);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	nvif_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
+	return !(nvif_rd08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE) & 0x20);
 }
 
 static inline void NVWriteVgaAttr(struct drm_device *dev,
 					int head, uint8_t index, uint8_t value)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	if (NVGetEnablePalette(dev, head))
 		index &= ~0x20;
 	else
 		index |= 0x20;
 
-	nv_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
-	nv_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, index);
-	nv_wr08(device, NV_PRMCIO_AR__WRITE + head * NV_PRMCIO_SIZE, value);
+	nvif_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
+	nvif_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, index);
+	nvif_wr08(device, NV_PRMCIO_AR__WRITE + head * NV_PRMCIO_SIZE, value);
 }
 
 static inline uint8_t NVReadVgaAttr(struct drm_device *dev,
 					int head, uint8_t index)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	uint8_t val;
 	if (NVGetEnablePalette(dev, head))
 		index &= ~0x20;
 	else
 		index |= 0x20;
 
-	nv_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
-	nv_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, index);
-	val = nv_rd08(device, NV_PRMCIO_AR__READ + head * NV_PRMCIO_SIZE);
+	nvif_rd08(device, NV_PRMCIO_INP0__COLOR + head * NV_PRMCIO_SIZE);
+	nvif_wr08(device, NV_PRMCIO_ARX + head * NV_PRMCIO_SIZE, index);
+	val = nvif_rd08(device, NV_PRMCIO_AR__READ + head * NV_PRMCIO_SIZE);
 	return val;
 }
 
@@ -259,11 +259,11 @@ static inline void NVVgaProtect(struct drm_device *dev, int head, bool protect)
 static inline bool
 nv_heads_tied(struct drm_device *dev)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
-	if (nv_device(drm->device)->chipset == 0x11)
-		return !!(nv_rd32(device, NV_PBUS_DEBUG_1) & (1 << 28));
+	if (drm->device.info.chipset == 0x11)
+		return !!(nvif_rd32(device, NV_PBUS_DEBUG_1) & (1 << 28));
 
 	return NVReadVgaCrtc(dev, 0, NV_CIO_CRE_44) & 0x4;
 }
@@ -318,7 +318,7 @@ NVLockVgaCrtcs(struct drm_device *dev, bool lock)
 	NVWriteVgaCrtc(dev, 0, NV_CIO_SR_LOCK_INDEX,
 		       lock ? NV_CIO_SR_LOCK_VALUE : NV_CIO_SR_UNLOCK_RW_VALUE);
 	/* NV11 has independently lockable extended crtcs, except when tied */
-	if (nv_device(drm->device)->chipset == 0x11 && !nv_heads_tied(dev))
+	if (drm->device.info.chipset == 0x11 && !nv_heads_tied(dev))
 		NVWriteVgaCrtc(dev, 1, NV_CIO_SR_LOCK_INDEX,
 			       lock ? NV_CIO_SR_LOCK_VALUE :
 				      NV_CIO_SR_UNLOCK_RW_VALUE);
@@ -335,7 +335,7 @@ static inline int nv_cursor_width(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
-	return nv_device(drm->device)->card_type >= NV_10 ? NV10_CURSOR_SIZE : NV04_CURSOR_SIZE;
+	return drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS ? NV10_CURSOR_SIZE : NV04_CURSOR_SIZE;
 }
 
 static inline void
@@ -357,7 +357,7 @@ nv_set_crtc_base(struct drm_device *dev, int head, uint32_t offset)
 
 	NVWriteCRTC(dev, head, NV_PCRTC_START, offset);
 
-	if (nv_device(drm->device)->card_type == NV_04) {
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_TNT) {
 		/*
 		 * Hilarious, the 24th bit doesn't want to stick to
 		 * PCRTC_START...
@@ -382,7 +382,7 @@ nv_show_cursor(struct drm_device *dev, int head, bool show)
 		*curctl1 &= ~MASK(NV_CIO_CRE_HCUR_ADDR1_ENABLE);
 	NVWriteVgaCrtc(dev, head, NV_CIO_CRE_HCUR_ADDR1_INDEX, *curctl1);
 
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		nv_fix_nv40_hw_cursor(dev, head);
 }
 
@@ -398,7 +398,7 @@ nv_pitch_align(struct drm_device *dev, uint32_t width, int bpp)
 		bpp = 8;
 
 	/* Alignment requirements taken from the Haiku driver */
-	if (nv_device(drm->device)->card_type == NV_04)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_TNT)
 		mask = 128 / bpp - 1;
 	else
 		mask = 512 / bpp - 1;
diff --git a/drivers/gpu/drm/nouveau/dispnv04/overlay.c b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
index ab03f7719d2d..1e9056a8df94 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/overlay.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/overlay.c
@@ -96,8 +96,9 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		  uint32_t src_x, uint32_t src_y,
 		  uint32_t src_w, uint32_t src_h)
 {
-	struct nouveau_device *dev = nouveau_dev(plane->dev);
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	struct nouveau_bo *cur = nv_plane->cur;
@@ -117,7 +118,7 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (format > 0xffff)
 		return -ERANGE;
 
-	if (dev->chipset >= 0x30) {
+	if (dev->info.chipset >= 0x30) {
 		if (crtc_w < (src_w >> 1) || crtc_h < (src_h >> 1))
 			return -ERANGE;
 	} else {
@@ -131,17 +132,17 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	nv_plane->cur = nv_fb->nvbo;
 
-	nv_mask(dev, NV_PCRTC_ENGINE_CTRL + soff, NV_CRTC_FSEL_OVERLAY, NV_CRTC_FSEL_OVERLAY);
-	nv_mask(dev, NV_PCRTC_ENGINE_CTRL + soff2, NV_CRTC_FSEL_OVERLAY, 0);
+	nvif_mask(dev, NV_PCRTC_ENGINE_CTRL + soff, NV_CRTC_FSEL_OVERLAY, NV_CRTC_FSEL_OVERLAY);
+	nvif_mask(dev, NV_PCRTC_ENGINE_CTRL + soff2, NV_CRTC_FSEL_OVERLAY, 0);
 
-	nv_wr32(dev, NV_PVIDEO_BASE(flip), 0);
-	nv_wr32(dev, NV_PVIDEO_OFFSET_BUFF(flip), nv_fb->nvbo->bo.offset);
-	nv_wr32(dev, NV_PVIDEO_SIZE_IN(flip), src_h << 16 | src_w);
-	nv_wr32(dev, NV_PVIDEO_POINT_IN(flip), src_y << 16 | src_x);
-	nv_wr32(dev, NV_PVIDEO_DS_DX(flip), (src_w << 20) / crtc_w);
-	nv_wr32(dev, NV_PVIDEO_DT_DY(flip), (src_h << 20) / crtc_h);
-	nv_wr32(dev, NV_PVIDEO_POINT_OUT(flip), crtc_y << 16 | crtc_x);
-	nv_wr32(dev, NV_PVIDEO_SIZE_OUT(flip), crtc_h << 16 | crtc_w);
+	nvif_wr32(dev, NV_PVIDEO_BASE(flip), 0);
+	nvif_wr32(dev, NV_PVIDEO_OFFSET_BUFF(flip), nv_fb->nvbo->bo.offset);
+	nvif_wr32(dev, NV_PVIDEO_SIZE_IN(flip), src_h << 16 | src_w);
+	nvif_wr32(dev, NV_PVIDEO_POINT_IN(flip), src_y << 16 | src_x);
+	nvif_wr32(dev, NV_PVIDEO_DS_DX(flip), (src_w << 20) / crtc_w);
+	nvif_wr32(dev, NV_PVIDEO_DT_DY(flip), (src_h << 20) / crtc_h);
+	nvif_wr32(dev, NV_PVIDEO_POINT_OUT(flip), crtc_y << 16 | crtc_x);
+	nvif_wr32(dev, NV_PVIDEO_SIZE_OUT(flip), crtc_h << 16 | crtc_w);
 
 	if (fb->pixel_format != DRM_FORMAT_UYVY)
 		format |= NV_PVIDEO_FORMAT_COLOR_LE_CR8YB8CB8YA8;
@@ -153,14 +154,14 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		format |= NV_PVIDEO_FORMAT_DISPLAY_COLOR_KEY;
 
 	if (fb->pixel_format == DRM_FORMAT_NV12) {
-		nv_wr32(dev, NV_PVIDEO_UVPLANE_BASE(flip), 0);
-		nv_wr32(dev, NV_PVIDEO_UVPLANE_OFFSET_BUFF(flip),
+		nvif_wr32(dev, NV_PVIDEO_UVPLANE_BASE(flip), 0);
+		nvif_wr32(dev, NV_PVIDEO_UVPLANE_OFFSET_BUFF(flip),
 			nv_fb->nvbo->bo.offset + fb->offsets[1]);
 	}
-	nv_wr32(dev, NV_PVIDEO_FORMAT(flip), format);
-	nv_wr32(dev, NV_PVIDEO_STOP, 0);
+	nvif_wr32(dev, NV_PVIDEO_FORMAT(flip), format);
+	nvif_wr32(dev, NV_PVIDEO_STOP, 0);
 	/* TODO: wait for vblank? */
-	nv_wr32(dev, NV_PVIDEO_BUFFER, flip ? 0x10 : 0x1);
+	nvif_wr32(dev, NV_PVIDEO_BUFFER, flip ? 0x10 : 0x1);
 	nv_plane->flip = !flip;
 
 	if (cur)
@@ -172,10 +173,11 @@ nv10_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 static int
 nv10_disable_plane(struct drm_plane *plane)
 {
-	struct nouveau_device *dev = nouveau_dev(plane->dev);
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 
-	nv_wr32(dev, NV_PVIDEO_STOP, 1);
+	nvif_wr32(dev, NV_PVIDEO_STOP, 1);
 	if (nv_plane->cur) {
 		nouveau_bo_unpin(nv_plane->cur);
 		nv_plane->cur = NULL;
@@ -195,24 +197,24 @@ nv_destroy_plane(struct drm_plane *plane)
 static void
 nv10_set_params(struct nouveau_plane *plane)
 {
-	struct nouveau_device *dev = nouveau_dev(plane->base.dev);
+	struct nvif_device *dev = &nouveau_drm(plane->base.dev)->device;
 	u32 luma = (plane->brightness - 512) << 16 | plane->contrast;
 	u32 chroma = ((sin_mul(plane->hue, plane->saturation) & 0xffff) << 16) |
 		(cos_mul(plane->hue, plane->saturation) & 0xffff);
 	u32 format = 0;
 
-	nv_wr32(dev, NV_PVIDEO_LUMINANCE(0), luma);
-	nv_wr32(dev, NV_PVIDEO_LUMINANCE(1), luma);
-	nv_wr32(dev, NV_PVIDEO_CHROMINANCE(0), chroma);
-	nv_wr32(dev, NV_PVIDEO_CHROMINANCE(1), chroma);
-	nv_wr32(dev, NV_PVIDEO_COLOR_KEY, plane->colorkey & 0xffffff);
+	nvif_wr32(dev, NV_PVIDEO_LUMINANCE(0), luma);
+	nvif_wr32(dev, NV_PVIDEO_LUMINANCE(1), luma);
+	nvif_wr32(dev, NV_PVIDEO_CHROMINANCE(0), chroma);
+	nvif_wr32(dev, NV_PVIDEO_CHROMINANCE(1), chroma);
+	nvif_wr32(dev, NV_PVIDEO_COLOR_KEY, plane->colorkey & 0xffffff);
 
 	if (plane->cur) {
 		if (plane->iturbt_709)
 			format |= NV_PVIDEO_FORMAT_MATRIX_ITURBT709;
 		if (plane->colorkey & (1 << 24))
 			format |= NV_PVIDEO_FORMAT_DISPLAY_COLOR_KEY;
-		nv_mask(dev, NV_PVIDEO_FORMAT(plane->flip),
+		nvif_mask(dev, NV_PVIDEO_FORMAT(plane->flip),
 			NV_PVIDEO_FORMAT_MATRIX_ITURBT709 |
 			NV_PVIDEO_FORMAT_DISPLAY_COLOR_KEY,
 			format);
@@ -224,7 +226,8 @@ nv_set_property(struct drm_plane *plane,
 		struct drm_property *property,
 		uint64_t value)
 {
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 
 	if (property == nv_plane->props.colorkey)
 		nv_plane->colorkey = value;
@@ -256,7 +259,7 @@ static const struct drm_plane_funcs nv10_plane_funcs = {
 static void
 nv10_overlay_init(struct drm_device *device)
 {
-	struct nouveau_device *dev = nouveau_dev(device);
+	struct nouveau_drm *drm = nouveau_drm(device);
 	struct nouveau_plane *plane = kzalloc(sizeof(struct nouveau_plane), GFP_KERNEL);
 	int num_formats = ARRAY_SIZE(formats);
 	int ret;
@@ -264,7 +267,7 @@ nv10_overlay_init(struct drm_device *device)
 	if (!plane)
 		return;
 
-	switch (dev->chipset) {
+	switch (drm->device.info.chipset) {
 	case 0x10:
 	case 0x11:
 	case 0x15:
@@ -333,7 +336,7 @@ cleanup:
 	drm_plane_cleanup(&plane->base);
 err:
 	kfree(plane);
-	nv_error(dev, "Failed to create plane\n");
+	NV_ERROR(drm, "Failed to create plane\n");
 }
 
 static int
@@ -343,8 +346,9 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		  uint32_t src_x, uint32_t src_y,
 		  uint32_t src_w, uint32_t src_h)
 {
-	struct nouveau_device *dev = nouveau_dev(plane->dev);
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
+	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
 	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
 	struct nouveau_bo *cur = nv_plane->cur;
 	uint32_t overlay = 1;
@@ -375,43 +379,43 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 	nv_plane->cur = nv_fb->nvbo;
 
-	nv_wr32(dev, NV_PVIDEO_OE_STATE, 0);
-	nv_wr32(dev, NV_PVIDEO_SU_STATE, 0);
-	nv_wr32(dev, NV_PVIDEO_RM_STATE, 0);
+	nvif_wr32(dev, NV_PVIDEO_OE_STATE, 0);
+	nvif_wr32(dev, NV_PVIDEO_SU_STATE, 0);
+	nvif_wr32(dev, NV_PVIDEO_RM_STATE, 0);
 
 	for (i = 0; i < 2; i++) {
-		nv_wr32(dev, NV_PVIDEO_BUFF0_START_ADDRESS + 4 * i,
+		nvif_wr32(dev, NV_PVIDEO_BUFF0_START_ADDRESS + 4 * i,
 			nv_fb->nvbo->bo.offset);
-		nv_wr32(dev, NV_PVIDEO_BUFF0_PITCH_LENGTH + 4 * i, pitch);
-		nv_wr32(dev, NV_PVIDEO_BUFF0_OFFSET + 4 * i, 0);
+		nvif_wr32(dev, NV_PVIDEO_BUFF0_PITCH_LENGTH + 4 * i, pitch);
+		nvif_wr32(dev, NV_PVIDEO_BUFF0_OFFSET + 4 * i, 0);
 	}
-	nv_wr32(dev, NV_PVIDEO_WINDOW_START, crtc_y << 16 | crtc_x);
-	nv_wr32(dev, NV_PVIDEO_WINDOW_SIZE, crtc_h << 16 | crtc_w);
-	nv_wr32(dev, NV_PVIDEO_STEP_SIZE,
+	nvif_wr32(dev, NV_PVIDEO_WINDOW_START, crtc_y << 16 | crtc_x);
+	nvif_wr32(dev, NV_PVIDEO_WINDOW_SIZE, crtc_h << 16 | crtc_w);
+	nvif_wr32(dev, NV_PVIDEO_STEP_SIZE,
 		(uint32_t)(((src_h - 1) << 11) / (crtc_h - 1)) << 16 | (uint32_t)(((src_w - 1) << 11) / (crtc_w - 1)));
 
 	/* It should be possible to convert hue/contrast to this */
-	nv_wr32(dev, NV_PVIDEO_RED_CSC_OFFSET, 0x69 - brightness);
-	nv_wr32(dev, NV_PVIDEO_GREEN_CSC_OFFSET, 0x3e + brightness);
-	nv_wr32(dev, NV_PVIDEO_BLUE_CSC_OFFSET, 0x89 - brightness);
-	nv_wr32(dev, NV_PVIDEO_CSC_ADJUST, 0);
+	nvif_wr32(dev, NV_PVIDEO_RED_CSC_OFFSET, 0x69 - brightness);
+	nvif_wr32(dev, NV_PVIDEO_GREEN_CSC_OFFSET, 0x3e + brightness);
+	nvif_wr32(dev, NV_PVIDEO_BLUE_CSC_OFFSET, 0x89 - brightness);
+	nvif_wr32(dev, NV_PVIDEO_CSC_ADJUST, 0);
 
-	nv_wr32(dev, NV_PVIDEO_CONTROL_Y, 0x001); /* (BLUR_ON, LINE_HALF) */
-	nv_wr32(dev, NV_PVIDEO_CONTROL_X, 0x111); /* (WEIGHT_HEAVY, SHARPENING_ON, SMOOTHING_ON) */
+	nvif_wr32(dev, NV_PVIDEO_CONTROL_Y, 0x001); /* (BLUR_ON, LINE_HALF) */
+	nvif_wr32(dev, NV_PVIDEO_CONTROL_X, 0x111); /* (WEIGHT_HEAVY, SHARPENING_ON, SMOOTHING_ON) */
 
-	nv_wr32(dev, NV_PVIDEO_FIFO_BURST_LENGTH, 0x03);
-	nv_wr32(dev, NV_PVIDEO_FIFO_THRES_SIZE, 0x38);
+	nvif_wr32(dev, NV_PVIDEO_FIFO_BURST_LENGTH, 0x03);
+	nvif_wr32(dev, NV_PVIDEO_FIFO_THRES_SIZE, 0x38);
 
-	nv_wr32(dev, NV_PVIDEO_KEY, nv_plane->colorkey);
+	nvif_wr32(dev, NV_PVIDEO_KEY, nv_plane->colorkey);
 
 	if (nv_plane->colorkey & (1 << 24))
 		overlay |= 0x10;
 	if (fb->pixel_format == DRM_FORMAT_YUYV)
 		overlay |= 0x100;
 
-	nv_wr32(dev, NV_PVIDEO_OVERLAY, overlay);
+	nvif_wr32(dev, NV_PVIDEO_OVERLAY, overlay);
 
-	nv_wr32(dev, NV_PVIDEO_SU_STATE, nv_rd32(dev, NV_PVIDEO_SU_STATE) ^ (1 << 16));
+	nvif_wr32(dev, NV_PVIDEO_SU_STATE, nvif_rd32(dev, NV_PVIDEO_SU_STATE) ^ (1 << 16));
 
 	if (cur)
 		nouveau_bo_unpin(cur);
@@ -422,13 +426,14 @@ nv04_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 static int
 nv04_disable_plane(struct drm_plane *plane)
 {
-	struct nouveau_device *dev = nouveau_dev(plane->dev);
-	struct nouveau_plane *nv_plane = (struct nouveau_plane *)plane;
-
-	nv_mask(dev, NV_PVIDEO_OVERLAY, 1, 0);
-	nv_wr32(dev, NV_PVIDEO_OE_STATE, 0);
-	nv_wr32(dev, NV_PVIDEO_SU_STATE, 0);
-	nv_wr32(dev, NV_PVIDEO_RM_STATE, 0);
+	struct nvif_device *dev = &nouveau_drm(plane->dev)->device;
+	struct nouveau_plane *nv_plane =
+		container_of(plane, struct nouveau_plane, base);
+
+	nvif_mask(dev, NV_PVIDEO_OVERLAY, 1, 0);
+	nvif_wr32(dev, NV_PVIDEO_OE_STATE, 0);
+	nvif_wr32(dev, NV_PVIDEO_SU_STATE, 0);
+	nvif_wr32(dev, NV_PVIDEO_RM_STATE, 0);
 	if (nv_plane->cur) {
 		nouveau_bo_unpin(nv_plane->cur);
 		nv_plane->cur = NULL;
@@ -447,7 +452,7 @@ static const struct drm_plane_funcs nv04_plane_funcs = {
 static void
 nv04_overlay_init(struct drm_device *device)
 {
-	struct nouveau_device *dev = nouveau_dev(device);
+	struct nouveau_drm *drm = nouveau_drm(device);
 	struct nouveau_plane *plane = kzalloc(sizeof(struct nouveau_plane), GFP_KERNEL);
 	int ret;
 
@@ -483,15 +488,15 @@ cleanup:
 	drm_plane_cleanup(&plane->base);
 err:
 	kfree(plane);
-	nv_error(dev, "Failed to create plane\n");
+	NV_ERROR(drm, "Failed to create plane\n");
 }
 
 void
 nouveau_overlay_init(struct drm_device *device)
 {
-	struct nouveau_device *dev = nouveau_dev(device);
-	if (dev->chipset < 0x10)
+	struct nvif_device *dev = &nouveau_drm(device)->device;
+	if (dev->info.chipset < 0x10)
 		nv04_overlay_init(device);
-	else if (dev->chipset <= 0x40)
+	else if (dev->info.chipset <= 0x40)
 		nv10_overlay_init(device);
 }
diff --git a/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c b/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c
index 8667620b703a..8061d8d0ce79 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv04.c
@@ -35,8 +35,6 @@
 
 #include <drm/i2c/ch7006.h>
 
-#include <subdev/i2c.h>
-
 static struct nouveau_i2c_board_info nv04_tv_encoder_info[] = {
 	{
 		{
@@ -56,7 +54,7 @@ static struct nouveau_i2c_board_info nv04_tv_encoder_info[] = {
 int nv04_tv_identify(struct drm_device *dev, int i2c_index)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 
 	return i2c->identify(i2c, i2c_index, "TV encoder",
 			     nv04_tv_encoder_info, NULL, NULL);
@@ -206,7 +204,7 @@ nv04_tv_create(struct drm_connector *connector, struct dcb_output *entry)
 	struct drm_encoder *encoder;
 	struct drm_device *dev = connector->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct nouveau_i2c_port *port = i2c->find(i2c, entry->i2c_index);
 	int type, ret;
 
diff --git a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c
index 195bd8e86c6a..72d2ab04db47 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c
@@ -34,11 +34,6 @@
 #include "hw.h"
 #include "tvnv17.h"
 
-#include <core/device.h>
-
-#include <subdev/bios/gpio.h>
-#include <subdev/gpio.h>
-
 MODULE_PARM_DESC(tv_norm, "Default TV norm.\n"
 		 "\t\tSupported: PAL, PAL-M, PAL-N, PAL-Nc, NTSC-M, NTSC-J,\n"
 		 "\t\t\thd480i, hd480p, hd576i, hd576p, hd720p, hd1080i.\n"
@@ -51,7 +46,7 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
+	struct nouveau_gpio *gpio = nvkm_gpio(&drm->device);
 	uint32_t testval, regoffset = nv04_dac_output_offset(encoder);
 	uint32_t gpio0, gpio1, fp_htotal, fp_hsync_start, fp_hsync_end,
 		fp_control, test_ctrl, dacclk, ctv_14, ctv_1c, ctv_6c;
@@ -135,17 +130,17 @@ static bool
 get_tv_detect_quirks(struct drm_device *dev, uint32_t *pin_mask)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_object *device = drm->device;
+	struct nvif_device *device = &drm->device;
 
 	/* Zotac FX5200 */
-	if (nv_device_match(device, 0x0322, 0x19da, 0x1035) ||
-	    nv_device_match(device, 0x0322, 0x19da, 0x2035)) {
+	if (nv_device_match(nvkm_object(device), 0x0322, 0x19da, 0x1035) ||
+	    nv_device_match(nvkm_object(device), 0x0322, 0x19da, 0x2035)) {
 		*pin_mask = 0xc;
 		return false;
 	}
 
 	/* MSI nForce2 IGP */
-	if (nv_device_match(device, 0x01f0, 0x1462, 0x5710)) {
+	if (nv_device_match(nvkm_object(device), 0x01f0, 0x1462, 0x5710)) {
 		*pin_mask = 0xc;
 		return false;
 	}
@@ -167,8 +162,8 @@ nv17_tv_detect(struct drm_encoder *encoder, struct drm_connector *connector)
 		return connector_status_disconnected;
 
 	if (reliable) {
-		if (nv_device(drm->device)->chipset == 0x42 ||
-		    nv_device(drm->device)->chipset == 0x43)
+		if (drm->device.info.chipset == 0x42 ||
+		    drm->device.info.chipset == 0x43)
 			tv_enc->pin_mask =
 				nv42_tv_sample_load(encoder) >> 28 & 0xe;
 		else
@@ -375,7 +370,7 @@ static void  nv17_tv_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
+	struct nouveau_gpio *gpio = nvkm_gpio(&drm->device);
 	struct nv17_tv_state *regs = &to_tv_enc(encoder)->state;
 	struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
 
@@ -448,7 +443,7 @@ static void nv17_tv_prepare(struct drm_encoder *encoder)
 	/* Set the DACCLK register */
 	dacclk = (NVReadRAMDAC(dev, 0, dacclk_off) & ~0x30) | 0x1;
 
-	if (nv_device(drm->device)->card_type == NV_40)
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CURIE)
 		dacclk |= 0x1a << 16;
 
 	if (tv_norm->kind == CTV_ENC_MODE) {
@@ -505,7 +500,7 @@ static void nv17_tv_mode_set(struct drm_encoder *encoder,
 			tv_regs->ptv_614 = 0x13;
 		}
 
-		if (nv_device(drm->device)->card_type >= NV_30) {
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_RANKINE) {
 			tv_regs->ptv_500 = 0xe8e0;
 			tv_regs->ptv_504 = 0x1710;
 			tv_regs->ptv_604 = 0x0;
@@ -600,7 +595,7 @@ static void nv17_tv_commit(struct drm_encoder *encoder)
 	nv17_tv_state_load(dev, &to_tv_enc(encoder)->state);
 
 	/* This could use refinement for flatpanels, but it should work */
-	if (nv_device(drm->device)->chipset < 0x44)
+	if (drm->device.info.chipset < 0x44)
 		NVWriteRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL +
 					nv04_dac_output_offset(encoder),
 					0xf0000000);
diff --git a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h
index 7b331543a41b..225894cdcac2 100644
--- a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h
+++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.h
@@ -130,14 +130,14 @@ void nv17_ctv_update_rescaler(struct drm_encoder *encoder);
 static inline void nv_write_ptv(struct drm_device *dev, uint32_t reg,
 				uint32_t val)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	nv_wr32(device, reg, val);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	nvif_wr32(device, reg, val);
 }
 
 static inline uint32_t nv_read_ptv(struct drm_device *dev, uint32_t reg)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
-	return nv_rd32(device, reg);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
+	return nvif_rd32(device, reg);
 }
 
 static inline void nv_write_tv_enc(struct drm_device *dev, uint8_t reg,
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index b13f441c6431..a24faa5e2a2a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -21,16 +21,10 @@
  *
  */
 
-#include <core/object.h>
-#include <core/client.h>
-#include <core/device.h>
-#include <core/class.h>
-#include <core/mm.h>
-
-#include <subdev/fb.h>
-#include <subdev/timer.h>
-#include <subdev/instmem.h>
-#include <engine/graph.h>
+#include <nvif/client.h>
+#include <nvif/driver.h>
+#include <nvif/ioctl.h>
+#include <nvif/class.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
@@ -47,20 +41,20 @@ nouveau_abi16_get(struct drm_file *file_priv, struct drm_device *dev)
 		struct nouveau_abi16 *abi16;
 		cli->abi16 = abi16 = kzalloc(sizeof(*abi16), GFP_KERNEL);
 		if (cli->abi16) {
+			struct nv_device_v0 args = {
+				.device = ~0ULL,
+			};
+
 			INIT_LIST_HEAD(&abi16->channels);
-			abi16->client = nv_object(cli);
 
 			/* allocate device object targeting client's default
 			 * device (ie. the one that belongs to the fd it
 			 * opened)
 			 */
-			if (nouveau_object_new(abi16->client, NVDRM_CLIENT,
-					       NVDRM_DEVICE, 0x0080,
-					       &(struct nv_device_class) {
-						.device = ~0ULL,
-					       },
-					       sizeof(struct nv_device_class),
-					       &abi16->device) == 0)
+			if (nvif_device_init(&cli->base.base, NULL,
+					     NOUVEAU_ABI16_DEVICE, NV_DEVICE,
+					     &args, sizeof(args),
+					     &abi16->device) == 0)
 				return cli->abi16;
 
 			kfree(cli->abi16);
@@ -75,7 +69,7 @@ nouveau_abi16_get(struct drm_file *file_priv, struct drm_device *dev)
 int
 nouveau_abi16_put(struct nouveau_abi16 *abi16, int ret)
 {
-	struct nouveau_cli *cli = (void *)abi16->client;
+	struct nouveau_cli *cli = (void *)nvif_client(&abi16->device.base);
 	mutex_unlock(&cli->mutex);
 	return ret;
 }
@@ -83,21 +77,19 @@ nouveau_abi16_put(struct nouveau_abi16 *abi16, int ret)
 u16
 nouveau_abi16_swclass(struct nouveau_drm *drm)
 {
-	switch (nv_device(drm->device)->card_type) {
-	case NV_04:
+	switch (drm->device.info.family) {
+	case NV_DEVICE_INFO_V0_TNT:
 		return 0x006e;
-	case NV_10:
-	case NV_11:
-	case NV_20:
-	case NV_30:
-	case NV_40:
+	case NV_DEVICE_INFO_V0_CELSIUS:
+	case NV_DEVICE_INFO_V0_KELVIN:
+	case NV_DEVICE_INFO_V0_RANKINE:
+	case NV_DEVICE_INFO_V0_CURIE:
 		return 0x016e;
-	case NV_50:
+	case NV_DEVICE_INFO_V0_TESLA:
 		return 0x506e;
-	case NV_C0:
-	case NV_D0:
-	case NV_E0:
-	case GM100:
+	case NV_DEVICE_INFO_V0_FERMI:
+	case NV_DEVICE_INFO_V0_KEPLER:
+	case NV_DEVICE_INFO_V0_MAXWELL:
 		return 0x906e;
 	}
 
@@ -140,7 +132,7 @@ nouveau_abi16_chan_fini(struct nouveau_abi16 *abi16,
 
 	/* destroy channel object, all children will be killed too */
 	if (chan->chan) {
-		abi16->handles &= ~(1ULL << (chan->chan->handle & 0xffff));
+		abi16->handles &= ~(1ULL << (chan->chan->object->handle & 0xffff));
 		nouveau_channel_del(&chan->chan);
 	}
 
@@ -151,7 +143,7 @@ nouveau_abi16_chan_fini(struct nouveau_abi16 *abi16,
 void
 nouveau_abi16_fini(struct nouveau_abi16 *abi16)
 {
-	struct nouveau_cli *cli = (void *)abi16->client;
+	struct nouveau_cli *cli = (void *)nvif_client(&abi16->device.base);
 	struct nouveau_abi16_chan *chan, *temp;
 
 	/* cleanup channels */
@@ -160,7 +152,7 @@ nouveau_abi16_fini(struct nouveau_abi16 *abi16)
 	}
 
 	/* destroy the device object */
-	nouveau_object_del(abi16->client, NVDRM_CLIENT, NVDRM_DEVICE);
+	nvif_device_fini(&abi16->device);
 
 	kfree(cli->abi16);
 	cli->abi16 = NULL;
@@ -169,30 +161,31 @@ nouveau_abi16_fini(struct nouveau_abi16 *abi16)
 int
 nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS)
 {
+	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_timer *ptimer = nouveau_timer(device);
-	struct nouveau_graph *graph = (void *)nouveau_engine(device, NVDEV_ENGINE_GR);
+	struct nvif_device *device = &drm->device;
+	struct nouveau_timer *ptimer = nvkm_timer(device);
+	struct nouveau_graph *graph = nvkm_gr(device);
 	struct drm_nouveau_getparam *getparam = data;
 
 	switch (getparam->param) {
 	case NOUVEAU_GETPARAM_CHIPSET_ID:
-		getparam->value = device->chipset;
+		getparam->value = device->info.chipset;
 		break;
 	case NOUVEAU_GETPARAM_PCI_VENDOR:
-		if (nv_device_is_pci(device))
+		if (nv_device_is_pci(nvkm_device(device)))
 			getparam->value = dev->pdev->vendor;
 		else
 			getparam->value = 0;
 		break;
 	case NOUVEAU_GETPARAM_PCI_DEVICE:
-		if (nv_device_is_pci(device))
+		if (nv_device_is_pci(nvkm_device(device)))
 			getparam->value = dev->pdev->device;
 		else
 			getparam->value = 0;
 		break;
 	case NOUVEAU_GETPARAM_BUS_TYPE:
-		if (!nv_device_is_pci(device))
+		if (!nv_device_is_pci(nvkm_device(device)))
 			getparam->value = 3;
 		else
 		if (drm_pci_device_is_agp(dev))
@@ -225,7 +218,7 @@ nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS)
 		getparam->value = graph->units ? graph->units(graph) : 0;
 		break;
 	default:
-		nv_debug(device, "unknown parameter %lld\n", getparam->param);
+		NV_PRINTK(debug, cli, "unknown parameter %lld\n", getparam->param);
 		return -EINVAL;
 	}
 
@@ -246,10 +239,7 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
 	struct nouveau_abi16_chan *chan;
-	struct nouveau_client *client;
-	struct nouveau_device *device;
-	struct nouveau_instmem *imem;
-	struct nouveau_fb *pfb;
+	struct nvif_device *device;
 	int ret;
 
 	if (unlikely(!abi16))
@@ -258,21 +248,18 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	if (!drm->channel)
 		return nouveau_abi16_put(abi16, -ENODEV);
 
-	client = nv_client(abi16->client);
-	device = nv_device(abi16->device);
-	imem   = nouveau_instmem(device);
-	pfb    = nouveau_fb(device);
+	device = &abi16->device;
 
 	/* hack to allow channel engine type specification on kepler */
-	if (device->card_type >= NV_E0) {
+	if (device->info.family >= NV_DEVICE_INFO_V0_KEPLER) {
 		if (init->fb_ctxdma_handle != ~0)
-			init->fb_ctxdma_handle = NVE0_CHANNEL_IND_ENGINE_GR;
+			init->fb_ctxdma_handle = KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_GR;
 		else
 			init->fb_ctxdma_handle = init->tt_ctxdma_handle;
 
 		/* allow flips to be executed if this is a graphics channel */
 		init->tt_ctxdma_handle = 0;
-		if (init->fb_ctxdma_handle == NVE0_CHANNEL_IND_ENGINE_GR)
+		if (init->fb_ctxdma_handle == KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_GR)
 			init->tt_ctxdma_handle = 1;
 	}
 
@@ -293,13 +280,14 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	abi16->handles |= (1ULL << init->channel);
 
 	/* create channel object and initialise dma and fence management */
-	ret = nouveau_channel_new(drm, cli, NVDRM_DEVICE, NVDRM_CHAN |
-				  init->channel, init->fb_ctxdma_handle,
+	ret = nouveau_channel_new(drm, device,
+				  NOUVEAU_ABI16_CHAN(init->channel),
+				  init->fb_ctxdma_handle,
 				  init->tt_ctxdma_handle, &chan->chan);
 	if (ret)
 		goto done;
 
-	if (device->card_type >= NV_50)
+	if (device->info.family >= NV_DEVICE_INFO_V0_TESLA)
 		init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM |
 					NOUVEAU_GEM_DOMAIN_GART;
 	else
@@ -308,10 +296,10 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	else
 		init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_GART;
 
-	if (device->card_type < NV_10) {
+	if (device->info.family < NV_DEVICE_INFO_V0_CELSIUS) {
 		init->subchan[0].handle = 0x00000000;
 		init->subchan[0].grclass = 0x0000;
-		init->subchan[1].handle = NvSw;
+		init->subchan[1].handle = chan->chan->nvsw.handle;
 		init->subchan[1].grclass = 0x506e;
 		init->nr_subchan = 2;
 	}
@@ -324,8 +312,8 @@ nouveau_abi16_ioctl_channel_alloc(ABI16_IOCTL_ARGS)
 	if (ret)
 		goto done;
 
-	if (device->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(chan->ntfy, client->vm,
+	if (device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_bo_vma_add(chan->ntfy, cli->vm,
 					&chan->ntfy_vma);
 		if (ret)
 			goto done;
@@ -343,6 +331,18 @@ done:
 	return nouveau_abi16_put(abi16, ret);
 }
 
+static struct nouveau_abi16_chan *
+nouveau_abi16_chan(struct nouveau_abi16 *abi16, int channel)
+{
+	struct nouveau_abi16_chan *chan;
+
+	list_for_each_entry(chan, &abi16->channels, head) {
+		if (chan->chan->object->handle == NOUVEAU_ABI16_CHAN(channel))
+			return chan;
+	}
+
+	return NULL;
+}
 
 int
 nouveau_abi16_ioctl_channel_free(ABI16_IOCTL_ARGS)
@@ -350,28 +350,38 @@ nouveau_abi16_ioctl_channel_free(ABI16_IOCTL_ARGS)
 	struct drm_nouveau_channel_free *req = data;
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
 	struct nouveau_abi16_chan *chan;
-	int ret = -ENOENT;
 
 	if (unlikely(!abi16))
 		return -ENOMEM;
 
-	list_for_each_entry(chan, &abi16->channels, head) {
-		if (chan->chan->handle == (NVDRM_CHAN | req->channel)) {
-			nouveau_abi16_chan_fini(abi16, chan);
-			return nouveau_abi16_put(abi16, 0);
-		}
-	}
-
-	return nouveau_abi16_put(abi16, ret);
+	chan = nouveau_abi16_chan(abi16, req->channel);
+	if (!chan)
+		return nouveau_abi16_put(abi16, -ENOENT);
+	nouveau_abi16_chan_fini(abi16, chan);
+	return nouveau_abi16_put(abi16, 0);
 }
 
 int
 nouveau_abi16_ioctl_grobj_alloc(ABI16_IOCTL_ARGS)
 {
 	struct drm_nouveau_grobj_alloc *init = data;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_new_v0 new;
+	} args = {
+		.ioctl.owner = NVIF_IOCTL_V0_OWNER_ANY,
+		.ioctl.type = NVIF_IOCTL_V0_NEW,
+		.ioctl.path_nr = 3,
+		.ioctl.path[2] = NOUVEAU_ABI16_CLIENT,
+		.ioctl.path[1] = NOUVEAU_ABI16_DEVICE,
+		.ioctl.path[0] = NOUVEAU_ABI16_CHAN(init->channel),
+		.new.route = NVDRM_OBJECT_ABI16,
+		.new.handle = init->handle,
+		.new.oclass = init->class,
+	};
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_object *object;
+	struct nvif_client *client;
 	int ret;
 
 	if (unlikely(!abi16))
@@ -379,6 +389,7 @@ nouveau_abi16_ioctl_grobj_alloc(ABI16_IOCTL_ARGS)
 
 	if (init->handle == ~0)
 		return nouveau_abi16_put(abi16, -EINVAL);
+	client = nvif_client(nvif_object(&abi16->device));
 
 	/* compatibility with userspace that assumes 506e for all chipsets */
 	if (init->class == 0x506e) {
@@ -387,8 +398,7 @@ nouveau_abi16_ioctl_grobj_alloc(ABI16_IOCTL_ARGS)
 			return nouveau_abi16_put(abi16, 0);
 	}
 
-	ret = nouveau_object_new(abi16->client, NVDRM_CHAN | init->channel,
-				  init->handle, init->class, NULL, 0, &object);
+	ret = nvif_client_ioctl(client, &args, sizeof(args));
 	return nouveau_abi16_put(abi16, ret);
 }
 
@@ -396,29 +406,38 @@ int
 nouveau_abi16_ioctl_notifierobj_alloc(ABI16_IOCTL_ARGS)
 {
 	struct drm_nouveau_notifierobj_alloc *info = data;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_new_v0 new;
+		struct nv_dma_v0 ctxdma;
+	} args = {
+		.ioctl.owner = NVIF_IOCTL_V0_OWNER_ANY,
+		.ioctl.type = NVIF_IOCTL_V0_NEW,
+		.ioctl.path_nr = 3,
+		.ioctl.path[2] = NOUVEAU_ABI16_CLIENT,
+		.ioctl.path[1] = NOUVEAU_ABI16_DEVICE,
+		.ioctl.path[0] = NOUVEAU_ABI16_CHAN(info->channel),
+		.new.route = NVDRM_OBJECT_ABI16,
+		.new.handle = info->handle,
+		.new.oclass = NV_DMA_IN_MEMORY,
+	};
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
-	struct nouveau_abi16_chan *chan = NULL, *temp;
+	struct nouveau_abi16_chan *chan;
 	struct nouveau_abi16_ntfy *ntfy;
-	struct nouveau_object *object;
-	struct nv_dma_class args = {};
+	struct nvif_device *device = &abi16->device;
+	struct nvif_client *client;
 	int ret;
 
 	if (unlikely(!abi16))
 		return -ENOMEM;
 
 	/* completely unnecessary for these chipsets... */
-	if (unlikely(nv_device(abi16->device)->card_type >= NV_C0))
+	if (unlikely(device->info.family >= NV_DEVICE_INFO_V0_FERMI))
 		return nouveau_abi16_put(abi16, -EINVAL);
+	client = nvif_client(nvif_object(&abi16->device));
 
-	list_for_each_entry(temp, &abi16->channels, head) {
-		if (temp->chan->handle == (NVDRM_CHAN | info->channel)) {
-			chan = temp;
-			break;
-		}
-	}
-
+	chan = nouveau_abi16_chan(abi16, info->channel);
 	if (!chan)
 		return nouveau_abi16_put(abi16, -ENOENT);
 
@@ -429,31 +448,34 @@ nouveau_abi16_ioctl_notifierobj_alloc(ABI16_IOCTL_ARGS)
 	list_add(&ntfy->head, &chan->notifiers);
 	ntfy->handle = info->handle;
 
-	ret = nouveau_mm_head(&chan->heap, 1, info->size, info->size, 1,
+	ret = nouveau_mm_head(&chan->heap, 0, 1, info->size, info->size, 1,
 			      &ntfy->node);
 	if (ret)
 		goto done;
 
-	args.start = ntfy->node->offset;
-	args.limit = ntfy->node->offset + ntfy->node->length - 1;
-	if (device->card_type >= NV_50) {
-		args.flags  = NV_DMA_TARGET_VM | NV_DMA_ACCESS_VM;
-		args.start += chan->ntfy_vma.offset;
-		args.limit += chan->ntfy_vma.offset;
+	args.ctxdma.start = ntfy->node->offset;
+	args.ctxdma.limit = ntfy->node->offset + ntfy->node->length - 1;
+	if (device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		args.ctxdma.target = NV_DMA_V0_TARGET_VM;
+		args.ctxdma.access = NV_DMA_V0_ACCESS_VM;
+		args.ctxdma.start += chan->ntfy_vma.offset;
+		args.ctxdma.limit += chan->ntfy_vma.offset;
 	} else
 	if (drm->agp.stat == ENABLED) {
-		args.flags  = NV_DMA_TARGET_AGP | NV_DMA_ACCESS_RDWR;
-		args.start += drm->agp.base + chan->ntfy->bo.offset;
-		args.limit += drm->agp.base + chan->ntfy->bo.offset;
+		args.ctxdma.target = NV_DMA_V0_TARGET_AGP;
+		args.ctxdma.access = NV_DMA_V0_ACCESS_RDWR;
+		args.ctxdma.start += drm->agp.base + chan->ntfy->bo.offset;
+		args.ctxdma.limit += drm->agp.base + chan->ntfy->bo.offset;
+		client->super = true;
 	} else {
-		args.flags  = NV_DMA_TARGET_VM | NV_DMA_ACCESS_RDWR;
-		args.start += chan->ntfy->bo.offset;
-		args.limit += chan->ntfy->bo.offset;
+		args.ctxdma.target = NV_DMA_V0_TARGET_VM;
+		args.ctxdma.access = NV_DMA_V0_ACCESS_RDWR;
+		args.ctxdma.start += chan->ntfy->bo.offset;
+		args.ctxdma.limit += chan->ntfy->bo.offset;
 	}
 
-	ret = nouveau_object_new(abi16->client, chan->chan->handle,
-				 ntfy->handle, 0x003d, &args,
-				 sizeof(args), &object);
+	ret = nvif_client_ioctl(client, &args, sizeof(args));
+	client->super = false;
 	if (ret)
 		goto done;
 
@@ -469,28 +491,36 @@ int
 nouveau_abi16_ioctl_gpuobj_free(ABI16_IOCTL_ARGS)
 {
 	struct drm_nouveau_gpuobj_free *fini = data;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_del del;
+	} args = {
+		.ioctl.owner = NVDRM_OBJECT_ABI16,
+		.ioctl.type = NVIF_IOCTL_V0_DEL,
+		.ioctl.path_nr = 4,
+		.ioctl.path[3] = NOUVEAU_ABI16_CLIENT,
+		.ioctl.path[2] = NOUVEAU_ABI16_DEVICE,
+		.ioctl.path[1] = NOUVEAU_ABI16_CHAN(fini->channel),
+		.ioctl.path[0] = fini->handle,
+	};
 	struct nouveau_abi16 *abi16 = nouveau_abi16_get(file_priv, dev);
-	struct nouveau_abi16_chan *chan = NULL, *temp;
+	struct nouveau_abi16_chan *chan;
 	struct nouveau_abi16_ntfy *ntfy;
+	struct nvif_client *client;
 	int ret;
 
 	if (unlikely(!abi16))
 		return -ENOMEM;
 
-	list_for_each_entry(temp, &abi16->channels, head) {
-		if (temp->chan->handle == (NVDRM_CHAN | fini->channel)) {
-			chan = temp;
-			break;
-		}
-	}
-
+	chan = nouveau_abi16_chan(abi16, fini->channel);
 	if (!chan)
 		return nouveau_abi16_put(abi16, -ENOENT);
+	client = nvif_client(nvif_object(&abi16->device));
 
 	/* synchronize with the user channel and destroy the gpu object */
 	nouveau_channel_idle(chan->chan);
 
-	ret = nouveau_object_del(abi16->client, chan->chan->handle, fini->handle);
+	ret = nvif_client_ioctl(client, &args, sizeof(args));
 	if (ret)
 		return nouveau_abi16_put(abi16, ret);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.h b/drivers/gpu/drm/nouveau/nouveau_abi16.h
index 90004081a501..39844e6bfbff 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.h
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.h
@@ -28,8 +28,7 @@ struct nouveau_abi16_chan {
 };
 
 struct nouveau_abi16 {
-	struct nouveau_object *client;
-	struct nouveau_object *device;
+	struct nvif_device device;
 	struct list_head channels;
 	u64 handles;
 };
diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index 279206997e5c..622424692b3b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -46,7 +46,6 @@ static struct nouveau_dsm_priv {
 	bool dsm_detected;
 	bool optimus_detected;
 	acpi_handle dhandle;
-	acpi_handle other_handle;
 	acpi_handle rom_handle;
 } nouveau_dsm_priv;
 
@@ -222,10 +221,9 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev)
 	if (!dhandle)
 		return false;
 
-	if (!acpi_has_method(dhandle, "_DSM")) {
-		nouveau_dsm_priv.other_handle = dhandle;
+	if (!acpi_has_method(dhandle, "_DSM"))
 		return false;
-	}
+
 	if (acpi_check_dsm(dhandle, nouveau_dsm_muid, 0x00000102,
 			   1 << NOUVEAU_DSM_POWER))
 		retval |= NOUVEAU_DSM_HAS_MUX;
@@ -301,16 +299,6 @@ static bool nouveau_dsm_detect(void)
 		printk(KERN_INFO "VGA switcheroo: detected DSM switching method %s handle\n",
 			acpi_method_name);
 		nouveau_dsm_priv.dsm_detected = true;
-		/*
-		 * On some systems hotplug events are generated for the device
-		 * being switched off when _DSM is executed.  They cause ACPI
-		 * hotplug to trigger and attempt to remove the device from
-		 * the system, which causes it to break down.  Prevent that from
-		 * happening by setting the no_hotplug flag for the involved
-		 * ACPI device objects.
-		 */
-		acpi_bus_no_hotplug(nouveau_dsm_priv.dhandle);
-		acpi_bus_no_hotplug(nouveau_dsm_priv.other_handle);
 		ret = true;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_agp.c b/drivers/gpu/drm/nouveau/nouveau_agp.c
index 51666daddb94..1f6f6ba6847a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_agp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_agp.c
@@ -1,7 +1,5 @@
 #include <linux/module.h>
 
-#include <core/device.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_agp.h"
 #include "nouveau_reg.h"
@@ -29,7 +27,7 @@ static struct nouveau_agpmode_quirk nouveau_agpmode_quirk_list[] = {
 static unsigned long
 get_agp_mode(struct nouveau_drm *drm, const struct drm_agp_info *info)
 {
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct nouveau_agpmode_quirk *quirk = nouveau_agpmode_quirk_list;
 	int agpmode = nouveau_agpmode;
 	unsigned long mode = info->mode;
@@ -38,7 +36,7 @@ get_agp_mode(struct nouveau_drm *drm, const struct drm_agp_info *info)
 	 * FW seems to be broken on nv18, it makes the card lock up
 	 * randomly.
 	 */
-	if (device->chipset == 0x18)
+	if (device->info.chipset == 0x18)
 		mode &= ~PCI_AGP_COMMAND_FW;
 
 	/*
@@ -47,10 +45,10 @@ get_agp_mode(struct nouveau_drm *drm, const struct drm_agp_info *info)
 	while (agpmode == -1 && quirk->hostbridge_vendor) {
 		if (info->id_vendor == quirk->hostbridge_vendor &&
 		    info->id_device == quirk->hostbridge_device &&
-		    device->pdev->vendor == quirk->chip_vendor &&
-		    device->pdev->device == quirk->chip_device) {
+		    nvkm_device(device)->pdev->vendor == quirk->chip_vendor &&
+		    nvkm_device(device)->pdev->device == quirk->chip_device) {
 			agpmode = quirk->mode;
-			nv_info(device, "Forcing agp mode to %dX. Use agpmode to override.\n",
+			NV_INFO(drm, "Forcing agp mode to %dX. Use agpmode to override.\n",
 				agpmode);
 			break;
 		}
@@ -104,7 +102,7 @@ void
 nouveau_agp_reset(struct nouveau_drm *drm)
 {
 #if __OS_HAS_AGP
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct drm_device *dev = drm->dev;
 	u32 save[2];
 	int ret;
@@ -115,7 +113,7 @@ nouveau_agp_reset(struct nouveau_drm *drm)
 	/* First of all, disable fast writes, otherwise if it's
 	 * already enabled in the AGP bridge and we disable the card's
 	 * AGP controller we might be locking ourselves out of it. */
-	if ((nv_rd32(device, NV04_PBUS_PCI_NV_19) |
+	if ((nvif_rd32(device, NV04_PBUS_PCI_NV_19) |
 	     dev->agp->mode) & PCI_AGP_COMMAND_FW) {
 		struct drm_agp_info info;
 		struct drm_agp_mode mode;
@@ -134,15 +132,15 @@ nouveau_agp_reset(struct nouveau_drm *drm)
 
 
 	/* clear busmaster bit, and disable AGP */
-	save[0] = nv_mask(device, NV04_PBUS_PCI_NV_1, 0x00000004, 0x00000000);
-	nv_wr32(device, NV04_PBUS_PCI_NV_19, 0);
+	save[0] = nvif_mask(device, NV04_PBUS_PCI_NV_1, 0x00000004, 0x00000000);
+	nvif_wr32(device, NV04_PBUS_PCI_NV_19, 0);
 
 	/* reset PGRAPH, PFIFO and PTIMER */
-	save[1] = nv_mask(device, 0x000200, 0x00011100, 0x00000000);
-	nv_mask(device, 0x000200, 0x00011100, save[1]);
+	save[1] = nvif_mask(device, 0x000200, 0x00011100, 0x00000000);
+	nvif_mask(device, 0x000200, 0x00011100, save[1]);
 
 	/* and restore bustmaster bit (gives effect of resetting AGP) */
-	nv_wr32(device, NV04_PBUS_PCI_NV_1, save[0]);
+	nvif_wr32(device, NV04_PBUS_PCI_NV_1, save[0]);
 #endif
 }
 
@@ -150,7 +148,6 @@ void
 nouveau_agp_init(struct nouveau_drm *drm)
 {
 #if __OS_HAS_AGP
-	struct nouveau_device *device = nv_device(drm->device);
 	struct drm_device *dev = drm->dev;
 	struct drm_agp_info info;
 	struct drm_agp_mode mode;
@@ -162,13 +159,13 @@ nouveau_agp_init(struct nouveau_drm *drm)
 
 	ret = drm_agp_acquire(dev);
 	if (ret) {
-		nv_error(device, "unable to acquire AGP: %d\n", ret);
+		NV_ERROR(drm, "unable to acquire AGP: %d\n", ret);
 		return;
 	}
 
 	ret = drm_agp_info(dev, &info);
 	if (ret) {
-		nv_error(device, "unable to get AGP info: %d\n", ret);
+		NV_ERROR(drm, "unable to get AGP info: %d\n", ret);
 		return;
 	}
 
@@ -177,7 +174,7 @@ nouveau_agp_init(struct nouveau_drm *drm)
 
 	ret = drm_agp_enable(dev, mode);
 	if (ret) {
-		nv_error(device, "unable to enable AGP: %d\n", ret);
+		NV_ERROR(drm, "unable to enable AGP: %d\n", ret);
 		return;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
index 2c1e4aad7da3..e566c5b53651 100644
--- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
+++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
@@ -40,8 +40,8 @@ static int
 nv40_get_intensity(struct backlight_device *bd)
 {
 	struct nouveau_drm *drm = bl_get_data(bd);
-	struct nouveau_device *device = nv_device(drm->device);
-	int val = (nv_rd32(device, NV40_PMC_BACKLIGHT) &
+	struct nvif_device *device = &drm->device;
+	int val = (nvif_rd32(device, NV40_PMC_BACKLIGHT) &
 				   NV40_PMC_BACKLIGHT_MASK) >> 16;
 
 	return val;
@@ -51,11 +51,11 @@ static int
 nv40_set_intensity(struct backlight_device *bd)
 {
 	struct nouveau_drm *drm = bl_get_data(bd);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int val = bd->props.brightness;
-	int reg = nv_rd32(device, NV40_PMC_BACKLIGHT);
+	int reg = nvif_rd32(device, NV40_PMC_BACKLIGHT);
 
-	nv_wr32(device, NV40_PMC_BACKLIGHT,
+	nvif_wr32(device, NV40_PMC_BACKLIGHT,
 		 (val << 16) | (reg & ~NV40_PMC_BACKLIGHT_MASK));
 
 	return 0;
@@ -71,11 +71,11 @@ static int
 nv40_backlight_init(struct drm_connector *connector)
 {
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct backlight_properties props;
 	struct backlight_device *bd;
 
-	if (!(nv_rd32(device, NV40_PMC_BACKLIGHT) & NV40_PMC_BACKLIGHT_MASK))
+	if (!(nvif_rd32(device, NV40_PMC_BACKLIGHT) & NV40_PMC_BACKLIGHT_MASK))
 		return 0;
 
 	memset(&props, 0, sizeof(struct backlight_properties));
@@ -97,12 +97,12 @@ nv50_get_intensity(struct backlight_device *bd)
 {
 	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int or = nv_encoder->or;
 	u32 div = 1025;
 	u32 val;
 
-	val  = nv_rd32(device, NV50_PDISP_SOR_PWM_CTL(or));
+	val  = nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(or));
 	val &= NV50_PDISP_SOR_PWM_CTL_VAL;
 	return ((val * 100) + (div / 2)) / div;
 }
@@ -112,12 +112,12 @@ nv50_set_intensity(struct backlight_device *bd)
 {
 	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int or = nv_encoder->or;
 	u32 div = 1025;
 	u32 val = (bd->props.brightness * div) / 100;
 
-	nv_wr32(device, NV50_PDISP_SOR_PWM_CTL(or),
+	nvif_wr32(device, NV50_PDISP_SOR_PWM_CTL(or),
 			NV50_PDISP_SOR_PWM_CTL_NEW | val);
 	return 0;
 }
@@ -133,12 +133,12 @@ nva3_get_intensity(struct backlight_device *bd)
 {
 	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int or = nv_encoder->or;
 	u32 div, val;
 
-	div  = nv_rd32(device, NV50_PDISP_SOR_PWM_DIV(or));
-	val  = nv_rd32(device, NV50_PDISP_SOR_PWM_CTL(or));
+	div  = nvif_rd32(device, NV50_PDISP_SOR_PWM_DIV(or));
+	val  = nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(or));
 	val &= NVA3_PDISP_SOR_PWM_CTL_VAL;
 	if (div && div >= val)
 		return ((val * 100) + (div / 2)) / div;
@@ -151,14 +151,14 @@ nva3_set_intensity(struct backlight_device *bd)
 {
 	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int or = nv_encoder->or;
 	u32 div, val;
 
-	div = nv_rd32(device, NV50_PDISP_SOR_PWM_DIV(or));
+	div = nvif_rd32(device, NV50_PDISP_SOR_PWM_DIV(or));
 	val = (bd->props.brightness * div) / 100;
 	if (div) {
-		nv_wr32(device, NV50_PDISP_SOR_PWM_CTL(or), val |
+		nvif_wr32(device, NV50_PDISP_SOR_PWM_CTL(or), val |
 				NV50_PDISP_SOR_PWM_CTL_NEW |
 				NVA3_PDISP_SOR_PWM_CTL_UNK);
 		return 0;
@@ -177,7 +177,7 @@ static int
 nv50_backlight_init(struct drm_connector *connector)
 {
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct nouveau_encoder *nv_encoder;
 	struct backlight_properties props;
 	struct backlight_device *bd;
@@ -190,12 +190,12 @@ nv50_backlight_init(struct drm_connector *connector)
 			return -ENODEV;
 	}
 
-	if (!nv_rd32(device, NV50_PDISP_SOR_PWM_CTL(nv_encoder->or)))
+	if (!nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(nv_encoder->or)))
 		return 0;
 
-	if (device->chipset <= 0xa0 ||
-	    device->chipset == 0xaa ||
-	    device->chipset == 0xac)
+	if (device->info.chipset <= 0xa0 ||
+	    device->info.chipset == 0xaa ||
+	    device->info.chipset == 0xac)
 		ops = &nv50_bl_ops;
 	else
 		ops = &nva3_bl_ops;
@@ -218,7 +218,7 @@ int
 nouveau_backlight_init(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct drm_connector *connector;
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
@@ -226,13 +226,12 @@ nouveau_backlight_init(struct drm_device *dev)
 		    connector->connector_type != DRM_MODE_CONNECTOR_eDP)
 			continue;
 
-		switch (device->card_type) {
-		case NV_40:
+		switch (device->info.family) {
+		case NV_DEVICE_INFO_V0_CURIE:
 			return nv40_backlight_init(connector);
-		case NV_50:
-		case NV_C0:
-		case NV_D0:
-		case NV_E0:
+		case NV_DEVICE_INFO_V0_TESLA:
+		case NV_DEVICE_INFO_V0_FERMI:
+		case NV_DEVICE_INFO_V0_KEPLER:
 			return nv50_backlight_init(connector);
 		default:
 			break;
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 8268a4ccac15..dae2c96deef8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -22,8 +22,6 @@
  * SOFTWARE.
  */
 
-#include <subdev/bios.h>
-
 #include <drm/drmP.h>
 
 #include "nouveau_drm.h"
@@ -217,7 +215,7 @@ int call_lvds_script(struct drm_device *dev, struct dcb_output *dcbent, int head
 	 */
 
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct nvbios *bios = &drm->vbios;
 	uint8_t lvds_ver = bios->data[bios->fp.lvdsmanufacturerpointer];
 	uint32_t sel_clk_binding, sel_clk;
@@ -240,7 +238,7 @@ int call_lvds_script(struct drm_device *dev, struct dcb_output *dcbent, int head
 	NV_INFO(drm, "Calling LVDS script %d:\n", script);
 
 	/* don't let script change pll->head binding */
-	sel_clk_binding = nv_rd32(device, NV_PRAMDAC_SEL_CLK) & 0x50000;
+	sel_clk_binding = nvif_rd32(device, NV_PRAMDAC_SEL_CLK) & 0x50000;
 
 	if (lvds_ver < 0x30)
 		ret = call_lvds_manufacturer_script(dev, dcbent, head, script);
@@ -252,7 +250,7 @@ int call_lvds_script(struct drm_device *dev, struct dcb_output *dcbent, int head
 	sel_clk = NVReadRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK) & ~0x50000;
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK, sel_clk | sel_clk_binding);
 	/* some scripts set a value in NV_PBUS_POWERCTRL_2 and break video overlay */
-	nv_wr32(device, NV_PBUS_POWERCTRL_2, 0);
+	nvif_wr32(device, NV_PBUS_POWERCTRL_2, 0);
 
 	return ret;
 }
@@ -320,7 +318,7 @@ static int parse_lvds_manufacturer_table_header(struct drm_device *dev, struct n
 static int
 get_fp_strap(struct drm_device *dev, struct nvbios *bios)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 
 	/*
 	 * The fp strap is normally dictated by the "User Strap" in
@@ -334,10 +332,10 @@ get_fp_strap(struct drm_device *dev, struct nvbios *bios)
 	if (bios->major_version < 5 && bios->data[0x48] & 0x4)
 		return NVReadVgaCrtc5758(dev, 0, 0xf) & 0xf;
 
-	if (device->card_type >= NV_50)
-		return (nv_rd32(device, NV_PEXTDEV_BOOT_0) >> 24) & 0xf;
+	if (device->info.family >= NV_DEVICE_INFO_V0_TESLA)
+		return (nvif_rd32(device, NV_PEXTDEV_BOOT_0) >> 24) & 0xf;
 	else
-		return (nv_rd32(device, NV_PEXTDEV_BOOT_0) >> 16) & 0xf;
+		return (nvif_rd32(device, NV_PEXTDEV_BOOT_0) >> 16) & 0xf;
 }
 
 static int parse_fp_mode_table(struct drm_device *dev, struct nvbios *bios)
@@ -636,7 +634,7 @@ int run_tmds_table(struct drm_device *dev, struct dcb_output *dcbent, int head,
 	 */
 
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct nvbios *bios = &drm->vbios;
 	int cv = bios->chip_version;
 	uint16_t clktable = 0, scriptptr;
@@ -670,7 +668,7 @@ int run_tmds_table(struct drm_device *dev, struct dcb_output *dcbent, int head,
 	}
 
 	/* don't let script change pll->head binding */
-	sel_clk_binding = nv_rd32(device, NV_PRAMDAC_SEL_CLK) & 0x50000;
+	sel_clk_binding = nvif_rd32(device, NV_PRAMDAC_SEL_CLK) & 0x50000;
 	run_digital_op_script(dev, scriptptr, dcbent, head, pxclk >= 165000);
 	sel_clk = NVReadRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK) & ~0x50000;
 	NVWriteRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK, sel_clk | sel_clk_binding);
@@ -1253,7 +1251,7 @@ olddcb_table(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	u8 *dcb = NULL;
 
-	if (nv_device(drm->device)->card_type > NV_04)
+	if (drm->device.info.family > NV_DEVICE_INFO_V0_TNT)
 		dcb = ROMPTR(dev, drm->vbios.data[0x36]);
 	if (!dcb) {
 		NV_WARN(drm, "No DCB data found in VBIOS\n");
@@ -1399,6 +1397,7 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 		  uint32_t conn, uint32_t conf, struct dcb_output *entry)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	int link = 0;
 
 	entry->type = conn & 0xf;
 	entry->i2c_index = (conn >> 4) & 0xf;
@@ -1444,6 +1443,7 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 			if (conf & 0x4)
 				entry->lvdsconf.use_power_scripts = true;
 			entry->lvdsconf.sor.link = (conf & 0x00000030) >> 4;
+			link = entry->lvdsconf.sor.link;
 		}
 		if (conf & mask) {
 			/*
@@ -1492,17 +1492,18 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 			entry->dpconf.link_nr = 1;
 			break;
 		}
+		link = entry->dpconf.sor.link;
 		break;
 	case DCB_OUTPUT_TMDS:
 		if (dcb->version >= 0x40) {
 			entry->tmdsconf.sor.link = (conf & 0x00000030) >> 4;
 			entry->extdev = (conf & 0x0000ff00) >> 8;
+			link = entry->tmdsconf.sor.link;
 		}
 		else if (dcb->version >= 0x30)
 			entry->tmdsconf.slave_addr = (conf & 0x00000700) >> 8;
 		else if (dcb->version >= 0x22)
 			entry->tmdsconf.slave_addr = (conf & 0x00000070) >> 4;
-
 		break;
 	case DCB_OUTPUT_EOL:
 		/* weird g80 mobile type that "nv" treats as a terminator */
@@ -1526,6 +1527,8 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 	if (conf & 0x100000)
 		entry->i2c_upper_default = true;
 
+	entry->hasht = (entry->location << 4) | entry->type;
+	entry->hashm = (entry->heads << 8) | (link << 6) | entry->or;
 	return true;
 }
 
@@ -1908,7 +1911,7 @@ static int load_nv17_hwsq_ucode_entry(struct drm_device *dev, struct nvbios *bio
 	 */
 
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	uint8_t bytes_to_write;
 	uint16_t hwsq_entry_offset;
 	int i;
@@ -1931,15 +1934,15 @@ static int load_nv17_hwsq_ucode_entry(struct drm_device *dev, struct nvbios *bio
 	hwsq_entry_offset = hwsq_offset + 2 + entry * bytes_to_write;
 
 	/* set sequencer control */
-	nv_wr32(device, 0x00001304, ROM32(bios->data[hwsq_entry_offset]));
+	nvif_wr32(device, 0x00001304, ROM32(bios->data[hwsq_entry_offset]));
 	bytes_to_write -= 4;
 
 	/* write ucode */
 	for (i = 0; i < bytes_to_write; i += 4)
-		nv_wr32(device, 0x00001400 + i, ROM32(bios->data[hwsq_entry_offset + i + 4]));
+		nvif_wr32(device, 0x00001400 + i, ROM32(bios->data[hwsq_entry_offset + i + 4]));
 
 	/* twiddle NV_PBUS_DEBUG_4 */
-	nv_wr32(device, NV_PBUS_DEBUG_4, nv_rd32(device, NV_PBUS_DEBUG_4) | 0x18);
+	nvif_wr32(device, NV_PBUS_DEBUG_4, nvif_rd32(device, NV_PBUS_DEBUG_4) | 0x18);
 
 	return 0;
 }
@@ -2002,7 +2005,7 @@ uint8_t *nouveau_bios_embedded_edid(struct drm_device *dev)
 static bool NVInitVBIOS(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_bios *bios = nouveau_bios(drm->device);
+	struct nouveau_bios *bios = nvkm_bios(&drm->device);
 	struct nvbios *legacy = &drm->vbios;
 
 	memset(legacy, 0, sizeof(struct nvbios));
@@ -2054,7 +2057,7 @@ nouveau_bios_posted(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	unsigned htotal;
 
-	if (nv_device(drm->device)->card_type >= NV_50)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA)
 		return true;
 
 	htotal  = NVReadVgaCrtc(dev, 0, 0x06);
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index b6dc85c614be..3d474ac03f88 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -27,13 +27,9 @@
  *	    Jeremy Kolb  <jkolb@brandeis.edu>
  */
 
-#include <core/engine.h>
+#include <linux/dma-mapping.h>
 #include <linux/swiotlb.h>
 
-#include <subdev/fb.h>
-#include <subdev/vm.h>
-#include <subdev/bar.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
@@ -52,7 +48,7 @@ nv10_bo_update_tile_region(struct drm_device *dev, struct nouveau_drm_tile *reg,
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	int i = reg - drm->tile.reg;
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	struct nouveau_fb_tile *tile = &pfb->tile.region[i];
 	struct nouveau_engine *engine;
 
@@ -92,13 +88,13 @@ nv10_bo_get_tile_region(struct drm_device *dev, int i)
 
 static void
 nv10_bo_put_tile_region(struct drm_device *dev, struct nouveau_drm_tile *tile,
-			struct nouveau_fence *fence)
+			struct fence *fence)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 
 	if (tile) {
 		spin_lock(&drm->tile.lock);
-		tile->fence = nouveau_fence_ref(fence);
+		tile->fence = (struct nouveau_fence *)fence_get(fence);
 		tile->used = false;
 		spin_unlock(&drm->tile.lock);
 	}
@@ -109,7 +105,7 @@ nv10_bo_set_tiling(struct drm_device *dev, u32 addr,
 		   u32 size, u32 pitch, u32 flags)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	struct nouveau_drm_tile *tile, *found = NULL;
 	int i;
 
@@ -153,23 +149,23 @@ nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
 		       int *align, int *size)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 
-	if (device->card_type < NV_50) {
+	if (device->info.family < NV_DEVICE_INFO_V0_TESLA) {
 		if (nvbo->tile_mode) {
-			if (device->chipset >= 0x40) {
+			if (device->info.chipset >= 0x40) {
 				*align = 65536;
 				*size = roundup(*size, 64 * nvbo->tile_mode);
 
-			} else if (device->chipset >= 0x30) {
+			} else if (device->info.chipset >= 0x30) {
 				*align = 32768;
 				*size = roundup(*size, 64 * nvbo->tile_mode);
 
-			} else if (device->chipset >= 0x20) {
+			} else if (device->info.chipset >= 0x20) {
 				*align = 16384;
 				*size = roundup(*size, 64 * nvbo->tile_mode);
 
-			} else if (device->chipset >= 0x10) {
+			} else if (device->info.chipset >= 0x10) {
 				*align = 16384;
 				*size = roundup(*size, 32 * nvbo->tile_mode);
 			}
@@ -185,7 +181,7 @@ nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
 int
 nouveau_bo_new(struct drm_device *dev, int size, int align,
 	       uint32_t flags, uint32_t tile_mode, uint32_t tile_flags,
-	       struct sg_table *sg,
+	       struct sg_table *sg, struct reservation_object *robj,
 	       struct nouveau_bo **pnvbo)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
@@ -196,12 +192,12 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	int lpg_shift = 12;
 	int max_size;
 
-	if (drm->client.base.vm)
-		lpg_shift = drm->client.base.vm->vmm->lpg_shift;
+	if (drm->client.vm)
+		lpg_shift = drm->client.vm->vmm->lpg_shift;
 	max_size = INT_MAX & ~((1 << lpg_shift) - 1);
 
 	if (size <= 0 || size > max_size) {
-		nv_warn(drm, "skipped size %x\n", (u32)size);
+		NV_WARN(drm, "skipped size %x\n", (u32)size);
 		return -EINVAL;
 	}
 
@@ -219,9 +215,9 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	nvbo->bo.bdev = &drm->ttm.bdev;
 
 	nvbo->page_shift = 12;
-	if (drm->client.base.vm) {
+	if (drm->client.vm) {
 		if (!(flags & TTM_PL_FLAG_TT) && size > 256 * 1024)
-			nvbo->page_shift = drm->client.base.vm->vmm->lpg_shift;
+			nvbo->page_shift = drm->client.vm->vmm->lpg_shift;
 	}
 
 	nouveau_bo_fixup_align(nvbo, flags, &align, &size);
@@ -234,7 +230,7 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	ret = ttm_bo_init(&drm->ttm.bdev, &nvbo->bo, size,
 			  type, &nvbo->placement,
 			  align >> PAGE_SHIFT, false, NULL, acc_size, sg,
-			  nouveau_bo_del_ttm);
+			  robj, nouveau_bo_del_ttm);
 	if (ret) {
 		/* ttm will call nouveau_bo_del_ttm if it fails.. */
 		return ret;
@@ -245,27 +241,26 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 }
 
 static void
-set_placement_list(uint32_t *pl, unsigned *n, uint32_t type, uint32_t flags)
+set_placement_list(struct ttm_place *pl, unsigned *n, uint32_t type, uint32_t flags)
 {
 	*n = 0;
 
 	if (type & TTM_PL_FLAG_VRAM)
-		pl[(*n)++] = TTM_PL_FLAG_VRAM | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_VRAM | flags;
 	if (type & TTM_PL_FLAG_TT)
-		pl[(*n)++] = TTM_PL_FLAG_TT | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_TT | flags;
 	if (type & TTM_PL_FLAG_SYSTEM)
-		pl[(*n)++] = TTM_PL_FLAG_SYSTEM | flags;
+		pl[(*n)++].flags = TTM_PL_FLAG_SYSTEM | flags;
 }
 
 static void
 set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
-	u32 vram_pages = pfb->ram->size >> PAGE_SHIFT;
+	u32 vram_pages = drm->device.info.ram_size >> PAGE_SHIFT;
+	unsigned i, fpfn, lpfn;
 
-	if ((nv_device(drm->device)->card_type == NV_10 ||
-	     nv_device(drm->device)->card_type == NV_11) &&
+	if (drm->device.info.family == NV_DEVICE_INFO_V0_CELSIUS &&
 	    nvbo->tile_mode && (type & TTM_PL_FLAG_VRAM) &&
 	    nvbo->bo.mem.num_pages < vram_pages / 4) {
 		/*
@@ -275,11 +270,19 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 		 * at the same time.
 		 */
 		if (nvbo->tile_flags & NOUVEAU_GEM_TILE_ZETA) {
-			nvbo->placement.fpfn = vram_pages / 2;
-			nvbo->placement.lpfn = ~0;
+			fpfn = vram_pages / 2;
+			lpfn = ~0;
 		} else {
-			nvbo->placement.fpfn = 0;
-			nvbo->placement.lpfn = vram_pages / 2;
+			fpfn = 0;
+			lpfn = vram_pages / 2;
+		}
+		for (i = 0; i < nvbo->placement.num_placement; ++i) {
+			nvbo->placements[i].fpfn = fpfn;
+			nvbo->placements[i].lpfn = lpfn;
+		}
+		for (i = 0; i < nvbo->placement.num_busy_placement; ++i) {
+			nvbo->busy_placements[i].fpfn = fpfn;
+			nvbo->busy_placements[i].lpfn = lpfn;
 		}
 	}
 }
@@ -309,7 +312,7 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype)
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	int ret;
 
-	ret = ttm_bo_reserve(bo, false, false, false, 0);
+	ret = ttm_bo_reserve(bo, false, false, false, NULL);
 	if (ret)
 		goto out;
 
@@ -350,7 +353,7 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo)
 	struct ttm_buffer_object *bo = &nvbo->bo;
 	int ret, ref;
 
-	ret = ttm_bo_reserve(bo, false, false, false, 0);
+	ret = ttm_bo_reserve(bo, false, false, false, NULL);
 	if (ret)
 		return ret;
 
@@ -385,7 +388,7 @@ nouveau_bo_map(struct nouveau_bo *nvbo)
 {
 	int ret;
 
-	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, 0);
+	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, NULL);
 	if (ret)
 		return ret;
 
@@ -500,21 +503,28 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		man->default_caching = TTM_PL_FLAG_CACHED;
 		break;
 	case TTM_PL_VRAM:
-		if (nv_device(drm->device)->card_type >= NV_50) {
+		man->flags = TTM_MEMTYPE_FLAG_FIXED |
+			     TTM_MEMTYPE_FLAG_MAPPABLE;
+		man->available_caching = TTM_PL_FLAG_UNCACHED |
+					 TTM_PL_FLAG_WC;
+		man->default_caching = TTM_PL_FLAG_WC;
+
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+			/* Some BARs do not support being ioremapped WC */
+			if (nvkm_bar(&drm->device)->iomap_uncached) {
+				man->available_caching = TTM_PL_FLAG_UNCACHED;
+				man->default_caching = TTM_PL_FLAG_UNCACHED;
+			}
+
 			man->func = &nouveau_vram_manager;
 			man->io_reserve_fastpath = false;
 			man->use_io_reserve_lru = true;
 		} else {
 			man->func = &ttm_bo_manager_func;
 		}
-		man->flags = TTM_MEMTYPE_FLAG_FIXED |
-			     TTM_MEMTYPE_FLAG_MAPPABLE;
-		man->available_caching = TTM_PL_FLAG_UNCACHED |
-					 TTM_PL_FLAG_WC;
-		man->default_caching = TTM_PL_FLAG_WC;
 		break;
 	case TTM_PL_TT:
-		if (nv_device(drm->device)->card_type >= NV_50)
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA)
 			man->func = &nouveau_gart_manager;
 		else
 		if (drm->agp.stat != ENABLED)
@@ -763,9 +773,9 @@ nv50_bo_move_init(struct nouveau_channel *chan, u32 handle)
 		BEGIN_NV04(chan, NvSubCopy, 0x0000, 1);
 		OUT_RING  (chan, handle);
 		BEGIN_NV04(chan, NvSubCopy, 0x0180, 3);
-		OUT_RING  (chan, NvNotify0);
-		OUT_RING  (chan, NvDmaFB);
-		OUT_RING  (chan, NvDmaFB);
+		OUT_RING  (chan, chan->drm->ntfy.handle);
+		OUT_RING  (chan, chan->vram.handle);
+		OUT_RING  (chan, chan->vram.handle);
 	}
 
 	return ret;
@@ -852,7 +862,7 @@ nv04_bo_move_init(struct nouveau_channel *chan, u32 handle)
 		BEGIN_NV04(chan, NvSubCopy, 0x0000, 1);
 		OUT_RING  (chan, handle);
 		BEGIN_NV04(chan, NvSubCopy, 0x0180, 1);
-		OUT_RING  (chan, NvNotify0);
+		OUT_RING  (chan, chan->drm->ntfy.handle);
 	}
 
 	return ret;
@@ -864,7 +874,7 @@ nouveau_bo_mem_ctxdma(struct ttm_buffer_object *bo,
 {
 	if (mem->mem_type == TTM_PL_TT)
 		return NvDmaTT;
-	return NvDmaFB;
+	return chan->vram.handle;
 }
 
 static int
@@ -922,12 +932,12 @@ nouveau_bo_move_prep(struct nouveau_drm *drm, struct ttm_buffer_object *bo,
 	u64 size = (u64)mem->num_pages << PAGE_SHIFT;
 	int ret;
 
-	ret = nouveau_vm_get(nv_client(drm)->vm, size, old_node->page_shift,
+	ret = nouveau_vm_get(drm->client.vm, size, old_node->page_shift,
 			     NV_MEM_ACCESS_RW, &old_node->vma[0]);
 	if (ret)
 		return ret;
 
-	ret = nouveau_vm_get(nv_client(drm)->vm, size, new_node->page_shift,
+	ret = nouveau_vm_get(drm->client.vm, size, new_node->page_shift,
 			     NV_MEM_ACCESS_RW, &old_node->vma[1]);
 	if (ret) {
 		nouveau_vm_put(&old_node->vma[0]);
@@ -945,6 +955,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct nouveau_channel *chan = drm->ttm.chan;
+	struct nouveau_cli *cli = (void *)nvif_client(&chan->device->base);
 	struct nouveau_fence *fence;
 	int ret;
 
@@ -952,20 +963,21 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 	 * old nouveau_mem node, these will get cleaned up after ttm has
 	 * destroyed the ttm_mem_reg
 	 */
-	if (nv_device(drm->device)->card_type >= NV_50) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
 		ret = nouveau_bo_move_prep(drm, bo, new_mem);
 		if (ret)
 			return ret;
 	}
 
-	mutex_lock_nested(&chan->cli->mutex, SINGLE_DEPTH_NESTING);
-	ret = nouveau_fence_sync(bo->sync_obj, chan);
+	mutex_lock_nested(&cli->mutex, SINGLE_DEPTH_NESTING);
+	ret = nouveau_fence_sync(nouveau_bo(bo), chan, true, intr);
 	if (ret == 0) {
 		ret = drm->ttm.move(chan, bo, &bo->mem, new_mem);
 		if (ret == 0) {
 			ret = nouveau_fence_new(chan, false, &fence);
 			if (ret == 0) {
-				ret = ttm_bo_move_accel_cleanup(bo, fence,
+				ret = ttm_bo_move_accel_cleanup(bo,
+								&fence->base,
 								evict,
 								no_wait_gpu,
 								new_mem);
@@ -973,7 +985,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 			}
 		}
 	}
-	mutex_unlock(&chan->cli->mutex);
+	mutex_unlock(&cli->mutex);
 	return ret;
 }
 
@@ -1005,9 +1017,7 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 	int ret;
 
 	do {
-		struct nouveau_object *object;
 		struct nouveau_channel *chan;
-		u32 handle = (mthd->engine << 16) | mthd->oclass;
 
 		if (mthd->engine)
 			chan = drm->cechan;
@@ -1016,13 +1026,14 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 		if (chan == NULL)
 			continue;
 
-		ret = nouveau_object_new(nv_object(drm), chan->handle, handle,
-					 mthd->oclass, NULL, 0, &object);
+		ret = nvif_object_init(chan->object, NULL,
+				       mthd->oclass | (mthd->engine << 16),
+				       mthd->oclass, NULL, 0,
+				       &drm->ttm.copy);
 		if (ret == 0) {
-			ret = mthd->init(chan, handle);
+			ret = mthd->init(chan, drm->ttm.copy.handle);
 			if (ret) {
-				nouveau_object_del(nv_object(drm),
-						   chan->handle, handle);
+				nvif_object_fini(&drm->ttm.copy);
 				continue;
 			}
 
@@ -1040,12 +1051,15 @@ static int
 nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 		      bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
-	u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING;
+	struct ttm_place placement_memtype = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+	};
 	struct ttm_placement placement;
 	struct ttm_mem_reg tmp_mem;
 	int ret;
 
-	placement.fpfn = placement.lpfn = 0;
 	placement.num_placement = placement.num_busy_placement = 1;
 	placement.placement = placement.busy_placement = &placement_memtype;
 
@@ -1073,12 +1087,15 @@ static int
 nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
 		      bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
-	u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING;
+	struct ttm_place placement_memtype = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING
+	};
 	struct ttm_placement placement;
 	struct ttm_mem_reg tmp_mem;
 	int ret;
 
-	placement.fpfn = placement.lpfn = 0;
 	placement.num_placement = placement.num_busy_placement = 1;
 	placement.placement = placement.busy_placement = &placement_memtype;
 
@@ -1135,7 +1152,7 @@ nouveau_bo_vm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem,
 	if (new_mem->mem_type != TTM_PL_VRAM)
 		return 0;
 
-	if (nv_device(drm->device)->card_type >= NV_10) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_CELSIUS) {
 		*new_tile = nv10_bo_set_tiling(dev, offset, new_mem->size,
 						nvbo->tile_mode,
 						nvbo->tile_flags);
@@ -1151,8 +1168,9 @@ nouveau_bo_vm_cleanup(struct ttm_buffer_object *bo,
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct drm_device *dev = drm->dev;
+	struct fence *fence = reservation_object_get_excl(bo->resv);
 
-	nv10_bo_put_tile_region(dev, *old_tile, bo->sync_obj);
+	nv10_bo_put_tile_region(dev, *old_tile, fence);
 	*old_tile = new_tile;
 }
 
@@ -1166,7 +1184,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	struct nouveau_drm_tile *new_tile = NULL;
 	int ret = 0;
 
-	if (nv_device(drm->device)->card_type < NV_50) {
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 		ret = nouveau_bo_vm_bind(bo, new_mem, &new_tile);
 		if (ret)
 			return ret;
@@ -1196,14 +1214,12 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	}
 
 	/* Fallback to software copy. */
-	spin_lock(&bo->bdev->fence_lock);
 	ret = ttm_bo_wait(bo, true, intr, no_wait_gpu);
-	spin_unlock(&bo->bdev->fence_lock);
 	if (ret == 0)
 		ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
 
 out:
-	if (nv_device(drm->device)->card_type < NV_50) {
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 		if (ret)
 			nouveau_bo_vm_cleanup(bo, NULL, &new_tile);
 		else
@@ -1227,7 +1243,6 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
 	struct nouveau_mem *node = mem->mm_node;
-	struct drm_device *dev = drm->dev;
 	int ret;
 
 	mem->bus.addr = NULL;
@@ -1246,19 +1261,19 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 		if (drm->agp.stat == ENABLED) {
 			mem->bus.offset = mem->start << PAGE_SHIFT;
 			mem->bus.base = drm->agp.base;
-			mem->bus.is_iomem = !dev->agp->cant_use_aperture;
+			mem->bus.is_iomem = !drm->dev->agp->cant_use_aperture;
 		}
 #endif
-		if (nv_device(drm->device)->card_type < NV_50 || !node->memtype)
+		if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA || !node->memtype)
 			/* untiled */
 			break;
 		/* fallthrough, tiled memory */
 	case TTM_PL_VRAM:
 		mem->bus.offset = mem->start << PAGE_SHIFT;
-		mem->bus.base = nv_device_resource_start(nouveau_dev(dev), 1);
+		mem->bus.base = nv_device_resource_start(nvkm_device(&drm->device), 1);
 		mem->bus.is_iomem = true;
-		if (nv_device(drm->device)->card_type >= NV_50) {
-			struct nouveau_bar *bar = nouveau_bar(drm->device);
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+			struct nouveau_bar *bar = nvkm_bar(&drm->device);
 
 			ret = bar->umap(bar, node, NV_MEM_ACCESS_RW,
 					&node->bar_vma);
@@ -1278,7 +1293,7 @@ static void
 nouveau_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bdev);
-	struct nouveau_bar *bar = nouveau_bar(drm->device);
+	struct nouveau_bar *bar = nvkm_bar(&drm->device);
 	struct nouveau_mem *node = mem->mm_node;
 
 	if (!node->bar_vma.node)
@@ -1292,15 +1307,15 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
-	struct nouveau_device *device = nv_device(drm->device);
-	u32 mappable = nv_device_resource_len(device, 1) >> PAGE_SHIFT;
-	int ret;
+	struct nvif_device *device = &drm->device;
+	u32 mappable = nv_device_resource_len(nvkm_device(device), 1) >> PAGE_SHIFT;
+	int i, ret;
 
 	/* as long as the bo isn't in vram, and isn't tiled, we've got
 	 * nothing to do here.
 	 */
 	if (bo->mem.mem_type != TTM_PL_VRAM) {
-		if (nv_device(drm->device)->card_type < NV_50 ||
+		if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA ||
 		    !nouveau_bo_tile_layout(nvbo))
 			return 0;
 
@@ -1315,13 +1330,20 @@ nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 	}
 
 	/* make sure bo is in mappable vram */
-	if (nv_device(drm->device)->card_type >= NV_50 ||
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA ||
 	    bo->mem.start + bo->mem.num_pages < mappable)
 		return 0;
 
+	for (i = 0; i < nvbo->placement.num_placement; ++i) {
+		nvbo->placements[i].fpfn = 0;
+		nvbo->placements[i].lpfn = mappable;
+	}
+
+	for (i = 0; i < nvbo->placement.num_busy_placement; ++i) {
+		nvbo->busy_placements[i].fpfn = 0;
+		nvbo->busy_placements[i].lpfn = mappable;
+	}
 
-	nvbo->placement.fpfn = 0;
-	nvbo->placement.lpfn = mappable;
 	nouveau_bo_placement_set(nvbo, TTM_PL_FLAG_VRAM, 0);
 	return nouveau_bo_validate(nvbo, false, false);
 }
@@ -1333,6 +1355,7 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 	struct nouveau_drm *drm;
 	struct nouveau_device *device;
 	struct drm_device *dev;
+	struct device *pdev;
 	unsigned i;
 	int r;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
@@ -1349,8 +1372,9 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 	}
 
 	drm = nouveau_bdev(ttm->bdev);
-	device = nv_device(drm->device);
+	device = nvkm_device(&drm->device);
 	dev = drm->dev;
+	pdev = nv_device_base(device);
 
 #if __OS_HAS_AGP
 	if (drm->agp.stat == ENABLED) {
@@ -1370,17 +1394,22 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm)
 	}
 
 	for (i = 0; i < ttm->num_pages; i++) {
-		ttm_dma->dma_address[i] = nv_device_map_page(device,
-							     ttm->pages[i]);
-		if (!ttm_dma->dma_address[i]) {
+		dma_addr_t addr;
+
+		addr = dma_map_page(pdev, ttm->pages[i], 0, PAGE_SIZE,
+				    DMA_BIDIRECTIONAL);
+
+		if (dma_mapping_error(pdev, addr)) {
 			while (--i) {
-				nv_device_unmap_page(device,
-						     ttm_dma->dma_address[i]);
+				dma_unmap_page(pdev, ttm_dma->dma_address[i],
+					       PAGE_SIZE, DMA_BIDIRECTIONAL);
 				ttm_dma->dma_address[i] = 0;
 			}
 			ttm_pool_unpopulate(ttm);
 			return -EFAULT;
 		}
+
+		ttm_dma->dma_address[i] = addr;
 	}
 	return 0;
 }
@@ -1392,6 +1421,7 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	struct nouveau_drm *drm;
 	struct nouveau_device *device;
 	struct drm_device *dev;
+	struct device *pdev;
 	unsigned i;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
@@ -1399,8 +1429,9 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 		return;
 
 	drm = nouveau_bdev(ttm->bdev);
-	device = nv_device(drm->device);
+	device = nvkm_device(&drm->device);
 	dev = drm->dev;
+	pdev = nv_device_base(device);
 
 #if __OS_HAS_AGP
 	if (drm->agp.stat == ENABLED) {
@@ -1418,7 +1449,8 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 
 	for (i = 0; i < ttm->num_pages; i++) {
 		if (ttm_dma->dma_address[i]) {
-			nv_device_unmap_page(device, ttm_dma->dma_address[i]);
+			dma_unmap_page(pdev, ttm_dma->dma_address[i], PAGE_SIZE,
+				       DMA_BIDIRECTIONAL);
 		}
 	}
 
@@ -1426,47 +1458,14 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
 }
 
 void
-nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence)
-{
-	struct nouveau_fence *new_fence = nouveau_fence_ref(fence);
-	struct nouveau_fence *old_fence = NULL;
-
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	old_fence = nvbo->bo.sync_obj;
-	nvbo->bo.sync_obj = new_fence;
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
-
-	nouveau_fence_unref(&old_fence);
-}
-
-static void
-nouveau_bo_fence_unref(void **sync_obj)
-{
-	nouveau_fence_unref((struct nouveau_fence **)sync_obj);
-}
-
-static void *
-nouveau_bo_fence_ref(void *sync_obj)
+nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence, bool exclusive)
 {
-	return nouveau_fence_ref(sync_obj);
-}
-
-static bool
-nouveau_bo_fence_signalled(void *sync_obj)
-{
-	return nouveau_fence_done(sync_obj);
-}
+	struct reservation_object *resv = nvbo->bo.resv;
 
-static int
-nouveau_bo_fence_wait(void *sync_obj, bool lazy, bool intr)
-{
-	return nouveau_fence_wait(sync_obj, lazy, intr);
-}
-
-static int
-nouveau_bo_fence_flush(void *sync_obj)
-{
-	return 0;
+	if (exclusive)
+		reservation_object_add_excl_fence(resv, &fence->base);
+	else if (fence)
+		reservation_object_add_shared_fence(resv, &fence->base);
 }
 
 struct ttm_bo_driver nouveau_bo_driver = {
@@ -1479,11 +1478,6 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.move_notify = nouveau_bo_move_ntfy,
 	.move = nouveau_bo_move,
 	.verify_access = nouveau_bo_verify_access,
-	.sync_obj_signaled = nouveau_bo_fence_signalled,
-	.sync_obj_wait = nouveau_bo_fence_wait,
-	.sync_obj_flush = nouveau_bo_fence_flush,
-	.sync_obj_unref = nouveau_bo_fence_unref,
-	.sync_obj_ref = nouveau_bo_fence_ref,
 	.fault_reserve_notify = &nouveau_ttm_fault_reserve_notify,
 	.io_mem_reserve = &nouveau_ttm_io_mem_reserve,
 	.io_mem_free = &nouveau_ttm_io_mem_free,
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index ff17c1f432fc..22d2c764d80b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -1,6 +1,8 @@
 #ifndef __NOUVEAU_BO_H__
 #define __NOUVEAU_BO_H__
 
+#include <drm/drm_gem.h>
+
 struct nouveau_channel;
 struct nouveau_fence;
 struct nouveau_vma;
@@ -9,8 +11,8 @@ struct nouveau_bo {
 	struct ttm_buffer_object bo;
 	struct ttm_placement placement;
 	u32 valid_domains;
-	u32 placements[3];
-	u32 busy_placements[3];
+	struct ttm_place placements[3];
+	struct ttm_place busy_placements[3];
 	struct ttm_bo_kmap_obj kmap;
 	struct list_head head;
 
@@ -68,6 +70,7 @@ extern struct ttm_bo_driver nouveau_bo_driver;
 void nouveau_bo_move_init(struct nouveau_drm *);
 int  nouveau_bo_new(struct drm_device *, int size, int align, u32 flags,
 		    u32 tile_mode, u32 tile_flags, struct sg_table *sg,
+		    struct reservation_object *robj,
 		    struct nouveau_bo **);
 int  nouveau_bo_pin(struct nouveau_bo *, u32 flags);
 int  nouveau_bo_unpin(struct nouveau_bo *);
@@ -78,7 +81,7 @@ u16  nouveau_bo_rd16(struct nouveau_bo *, unsigned index);
 void nouveau_bo_wr16(struct nouveau_bo *, unsigned index, u16 val);
 u32  nouveau_bo_rd32(struct nouveau_bo *, unsigned index);
 void nouveau_bo_wr32(struct nouveau_bo *, unsigned index, u32 val);
-void nouveau_bo_fence(struct nouveau_bo *, struct nouveau_fence *);
+void nouveau_bo_fence(struct nouveau_bo *, struct nouveau_fence *, bool exclusive);
 int  nouveau_bo_validate(struct nouveau_bo *, bool interruptible,
 			 bool no_wait_gpu);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index ccb6b452d6d0..fd3dbd59d73e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -22,16 +22,11 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/object.h>
-#include <core/client.h>
-#include <core/device.h>
-#include <core/class.h>
-
-#include <subdev/fb.h>
-#include <subdev/vm.h>
-#include <subdev/instmem.h>
+#include <nvif/os.h>
+#include <nvif/class.h>
 
-#include <engine/software.h>
+/*XXX*/
+#include <core/client.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
@@ -41,13 +36,13 @@
 #include "nouveau_abi16.h"
 
 MODULE_PARM_DESC(vram_pushbuf, "Create DMA push buffers in VRAM");
-static int nouveau_vram_pushbuf;
+int nouveau_vram_pushbuf;
 module_param_named(vram_pushbuf, nouveau_vram_pushbuf, int, 0400);
 
 int
 nouveau_channel_idle(struct nouveau_channel *chan)
 {
-	struct nouveau_cli *cli = chan->cli;
+	struct nouveau_cli *cli = (void *)nvif_client(chan->object);
 	struct nouveau_fence *fence = NULL;
 	int ret;
 
@@ -58,8 +53,8 @@ nouveau_channel_idle(struct nouveau_channel *chan)
 	}
 
 	if (ret)
-		NV_ERROR(cli, "failed to idle channel 0x%08x [%s]\n",
-			 chan->handle, cli->base.name);
+		NV_PRINTK(error, cli, "failed to idle channel 0x%08x [%s]\n",
+			  chan->object->handle, nvkm_client(&cli->base)->name);
 	return ret;
 }
 
@@ -68,36 +63,34 @@ nouveau_channel_del(struct nouveau_channel **pchan)
 {
 	struct nouveau_channel *chan = *pchan;
 	if (chan) {
-		struct nouveau_object *client = nv_object(chan->cli);
 		if (chan->fence) {
 			nouveau_channel_idle(chan);
 			nouveau_fence(chan->drm)->context_del(chan);
 		}
-		nouveau_object_del(client, NVDRM_DEVICE, chan->handle);
-		nouveau_object_del(client, NVDRM_DEVICE, chan->push.handle);
+		nvif_object_fini(&chan->nvsw);
+		nvif_object_fini(&chan->gart);
+		nvif_object_fini(&chan->vram);
+		nvif_object_ref(NULL, &chan->object);
+		nvif_object_fini(&chan->push.ctxdma);
 		nouveau_bo_vma_del(chan->push.buffer, &chan->push.vma);
 		nouveau_bo_unmap(chan->push.buffer);
 		if (chan->push.buffer && chan->push.buffer->pin_refcnt)
 			nouveau_bo_unpin(chan->push.buffer);
 		nouveau_bo_ref(NULL, &chan->push.buffer);
+		nvif_device_ref(NULL, &chan->device);
 		kfree(chan);
 	}
 	*pchan = NULL;
 }
 
 static int
-nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
-		     u32 parent, u32 handle, u32 size,
-		     struct nouveau_channel **pchan)
+nouveau_channel_prep(struct nouveau_drm *drm, struct nvif_device *device,
+		     u32 handle, u32 size, struct nouveau_channel **pchan)
 {
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_instmem *imem = nouveau_instmem(device);
-	struct nouveau_vmmgr *vmm = nouveau_vmmgr(device);
-	struct nouveau_fb *pfb = nouveau_fb(device);
-	struct nouveau_client *client = &cli->base;
-	struct nv_dma_class args = {};
+	struct nouveau_cli *cli = (void *)nvif_client(&device->base);
+	struct nouveau_vmmgr *vmm = nvkm_vmmgr(device);
+	struct nv_dma_v0 args = {};
 	struct nouveau_channel *chan;
-	struct nouveau_object *push;
 	u32 target;
 	int ret;
 
@@ -105,16 +98,15 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
 	if (!chan)
 		return -ENOMEM;
 
-	chan->cli = cli;
+	nvif_device_ref(device, &chan->device);
 	chan->drm = drm;
-	chan->handle = handle;
 
 	/* allocate memory for dma push buffer */
 	target = TTM_PL_FLAG_TT;
 	if (nouveau_vram_pushbuf)
 		target = TTM_PL_FLAG_VRAM;
 
-	ret = nouveau_bo_new(drm->dev, size, 0, target, 0, 0, NULL,
+	ret = nouveau_bo_new(drm->dev, size, 0, target, 0, 0, NULL, NULL,
 			    &chan->push.buffer);
 	if (ret == 0) {
 		ret = nouveau_bo_pin(chan->push.buffer, target);
@@ -132,51 +124,54 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
 	 * we be able to call out to other (indirect) push buffers
 	 */
 	chan->push.vma.offset = chan->push.buffer->bo.offset;
-	chan->push.handle = NVDRM_PUSH | (handle & 0xffff);
 
-	if (device->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(chan->push.buffer, client->vm,
+	if (device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_bo_vma_add(chan->push.buffer, cli->vm,
 					&chan->push.vma);
 		if (ret) {
 			nouveau_channel_del(pchan);
 			return ret;
 		}
 
-		args.flags = NV_DMA_TARGET_VM | NV_DMA_ACCESS_VM;
+		args.target = NV_DMA_V0_TARGET_VM;
+		args.access = NV_DMA_V0_ACCESS_VM;
 		args.start = 0;
-		args.limit = client->vm->vmm->limit - 1;
+		args.limit = cli->vm->vmm->limit - 1;
 	} else
 	if (chan->push.buffer->bo.mem.mem_type == TTM_PL_VRAM) {
-		u64 limit = pfb->ram->size - imem->reserved - 1;
-		if (device->card_type == NV_04) {
+		if (device->info.family == NV_DEVICE_INFO_V0_TNT) {
 			/* nv04 vram pushbuf hack, retarget to its location in
 			 * the framebuffer bar rather than direct vram access..
 			 * nfi why this exists, it came from the -nv ddx.
 			 */
-			args.flags = NV_DMA_TARGET_PCI | NV_DMA_ACCESS_RDWR;
-			args.start = nv_device_resource_start(device, 1);
-			args.limit = args.start + limit;
+			args.target = NV_DMA_V0_TARGET_PCI;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
+			args.start = nv_device_resource_start(nvkm_device(device), 1);
+			args.limit = args.start + device->info.ram_user - 1;
 		} else {
-			args.flags = NV_DMA_TARGET_VRAM | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_VRAM;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = 0;
-			args.limit = limit;
+			args.limit = device->info.ram_user - 1;
 		}
 	} else {
 		if (chan->drm->agp.stat == ENABLED) {
-			args.flags = NV_DMA_TARGET_AGP | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_AGP;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = chan->drm->agp.base;
 			args.limit = chan->drm->agp.base +
 				     chan->drm->agp.size - 1;
 		} else {
-			args.flags = NV_DMA_TARGET_VM | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_VM;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = 0;
 			args.limit = vmm->limit - 1;
 		}
 	}
 
-	ret = nouveau_object_new(nv_object(chan->cli), parent,
-				 chan->push.handle, 0x0002,
-				 &args, sizeof(args), &push);
+	ret = nvif_object_init(nvif_object(device), NULL, NVDRM_PUSH |
+			       (handle & 0xffff), NV_DMA_FROM_MEMORY,
+			       &args, sizeof(args), &chan->push.ctxdma);
 	if (ret) {
 		nouveau_channel_del(pchan);
 		return ret;
@@ -186,38 +181,56 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
 }
 
 static int
-nouveau_channel_ind(struct nouveau_drm *drm, struct nouveau_cli *cli,
-		    u32 parent, u32 handle, u32 engine,
-		    struct nouveau_channel **pchan)
+nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device,
+		    u32 handle, u32 engine, struct nouveau_channel **pchan)
 {
-	static const u16 oclasses[] = { NVE0_CHANNEL_IND_CLASS,
-					NVC0_CHANNEL_IND_CLASS,
-					NV84_CHANNEL_IND_CLASS,
-					NV50_CHANNEL_IND_CLASS,
+	static const u16 oclasses[] = { KEPLER_CHANNEL_GPFIFO_A,
+					FERMI_CHANNEL_GPFIFO,
+					G82_CHANNEL_GPFIFO,
+					NV50_CHANNEL_GPFIFO,
 					0 };
 	const u16 *oclass = oclasses;
-	struct nve0_channel_ind_class args;
+	union {
+		struct nv50_channel_gpfifo_v0 nv50;
+		struct kepler_channel_gpfifo_a_v0 kepler;
+	} args, *retn;
 	struct nouveau_channel *chan;
+	u32 size;
 	int ret;
 
 	/* allocate dma push buffer */
-	ret = nouveau_channel_prep(drm, cli, parent, handle, 0x12000, &chan);
+	ret = nouveau_channel_prep(drm, device, handle, 0x12000, &chan);
 	*pchan = chan;
 	if (ret)
 		return ret;
 
 	/* create channel object */
-	args.pushbuf = chan->push.handle;
-	args.ioffset = 0x10000 + chan->push.vma.offset;
-	args.ilength = 0x02000;
-	args.engine  = engine;
-
 	do {
-		ret = nouveau_object_new(nv_object(cli), parent, handle,
-					 *oclass++, &args, sizeof(args),
-					 &chan->object);
-		if (ret == 0)
+		if (oclass[0] >= KEPLER_CHANNEL_GPFIFO_A) {
+			args.kepler.version = 0;
+			args.kepler.engine  = engine;
+			args.kepler.pushbuf = chan->push.ctxdma.handle;
+			args.kepler.ilength = 0x02000;
+			args.kepler.ioffset = 0x10000 + chan->push.vma.offset;
+			size = sizeof(args.kepler);
+		} else {
+			args.nv50.version = 0;
+			args.nv50.pushbuf = chan->push.ctxdma.handle;
+			args.nv50.ilength = 0x02000;
+			args.nv50.ioffset = 0x10000 + chan->push.vma.offset;
+			size = sizeof(args.nv50);
+		}
+
+		ret = nvif_object_new(nvif_object(device), handle, *oclass++,
+				      &args, size, &chan->object);
+		if (ret == 0) {
+			retn = chan->object->data;
+			if (chan->object->oclass >= KEPLER_CHANNEL_GPFIFO_A)
+				chan->chid = retn->kepler.chid;
+			else
+				chan->chid = retn->nv50.chid;
 			return ret;
+		}
 	} while (*oclass);
 
 	nouveau_channel_del(pchan);
@@ -225,35 +238,38 @@ nouveau_channel_ind(struct nouveau_drm *drm, struct nouveau_cli *cli,
 }
 
 static int
-nouveau_channel_dma(struct nouveau_drm *drm, struct nouveau_cli *cli,
-		    u32 parent, u32 handle, struct nouveau_channel **pchan)
+nouveau_channel_dma(struct nouveau_drm *drm, struct nvif_device *device,
+		    u32 handle, struct nouveau_channel **pchan)
 {
-	static const u16 oclasses[] = { NV40_CHANNEL_DMA_CLASS,
-					NV17_CHANNEL_DMA_CLASS,
-					NV10_CHANNEL_DMA_CLASS,
-					NV03_CHANNEL_DMA_CLASS,
+	static const u16 oclasses[] = { NV40_CHANNEL_DMA,
+					NV17_CHANNEL_DMA,
+					NV10_CHANNEL_DMA,
+					NV03_CHANNEL_DMA,
 					0 };
 	const u16 *oclass = oclasses;
-	struct nv03_channel_dma_class args;
+	struct nv03_channel_dma_v0 args, *retn;
 	struct nouveau_channel *chan;
 	int ret;
 
 	/* allocate dma push buffer */
-	ret = nouveau_channel_prep(drm, cli, parent, handle, 0x10000, &chan);
+	ret = nouveau_channel_prep(drm, device, handle, 0x10000, &chan);
 	*pchan = chan;
 	if (ret)
 		return ret;
 
 	/* create channel object */
-	args.pushbuf = chan->push.handle;
+	args.version = 0;
+	args.pushbuf = chan->push.ctxdma.handle;
 	args.offset = chan->push.vma.offset;
 
 	do {
-		ret = nouveau_object_new(nv_object(cli), parent, handle,
-					 *oclass++, &args, sizeof(args),
-					 &chan->object);
-		if (ret == 0)
+		ret = nvif_object_new(nvif_object(device), handle, *oclass++,
+				      &args, sizeof(args), &chan->object);
+		if (ret == 0) {
+			retn = chan->object->data;
+			chan->chid = retn->chid;
 			return ret;
+		}
 	} while (ret && *oclass);
 
 	nouveau_channel_del(pchan);
@@ -263,60 +279,64 @@ nouveau_channel_dma(struct nouveau_drm *drm, struct nouveau_cli *cli,
 static int
 nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 {
-	struct nouveau_client *client = nv_client(chan->cli);
-	struct nouveau_device *device = nv_device(chan->drm->device);
-	struct nouveau_instmem *imem = nouveau_instmem(device);
-	struct nouveau_vmmgr *vmm = nouveau_vmmgr(device);
-	struct nouveau_fb *pfb = nouveau_fb(device);
+	struct nvif_device *device = chan->device;
+	struct nouveau_cli *cli = (void *)nvif_client(&device->base);
+	struct nouveau_vmmgr *vmm = nvkm_vmmgr(device);
 	struct nouveau_software_chan *swch;
-	struct nouveau_object *object;
-	struct nv_dma_class args = {};
+	struct nv_dma_v0 args = {};
 	int ret, i;
+	bool save;
+
+	nvif_object_map(chan->object);
 
 	/* allocate dma objects to cover all allowed vram, and gart */
-	if (device->card_type < NV_C0) {
-		if (device->card_type >= NV_50) {
-			args.flags = NV_DMA_TARGET_VM | NV_DMA_ACCESS_VM;
+	if (device->info.family < NV_DEVICE_INFO_V0_FERMI) {
+		if (device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+			args.target = NV_DMA_V0_TARGET_VM;
+			args.access = NV_DMA_V0_ACCESS_VM;
 			args.start = 0;
-			args.limit = client->vm->vmm->limit - 1;
+			args.limit = cli->vm->vmm->limit - 1;
 		} else {
-			args.flags = NV_DMA_TARGET_VRAM | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_VRAM;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = 0;
-			args.limit = pfb->ram->size - imem->reserved - 1;
+			args.limit = device->info.ram_user - 1;
 		}
 
-		ret = nouveau_object_new(nv_object(client), chan->handle, vram,
-					 0x003d, &args, sizeof(args), &object);
+		ret = nvif_object_init(chan->object, NULL, vram,
+				       NV_DMA_IN_MEMORY, &args,
+				       sizeof(args), &chan->vram);
 		if (ret)
 			return ret;
 
-		if (device->card_type >= NV_50) {
-			args.flags = NV_DMA_TARGET_VM | NV_DMA_ACCESS_VM;
+		if (device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+			args.target = NV_DMA_V0_TARGET_VM;
+			args.access = NV_DMA_V0_ACCESS_VM;
 			args.start = 0;
-			args.limit = client->vm->vmm->limit - 1;
+			args.limit = cli->vm->vmm->limit - 1;
 		} else
 		if (chan->drm->agp.stat == ENABLED) {
-			args.flags = NV_DMA_TARGET_AGP | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_AGP;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = chan->drm->agp.base;
 			args.limit = chan->drm->agp.base +
 				     chan->drm->agp.size - 1;
 		} else {
-			args.flags = NV_DMA_TARGET_VM | NV_DMA_ACCESS_RDWR;
+			args.target = NV_DMA_V0_TARGET_VM;
+			args.access = NV_DMA_V0_ACCESS_RDWR;
 			args.start = 0;
 			args.limit = vmm->limit - 1;
 		}
 
-		ret = nouveau_object_new(nv_object(client), chan->handle, gart,
-					 0x003d, &args, sizeof(args), &object);
+		ret = nvif_object_init(chan->object, NULL, gart,
+				       NV_DMA_IN_MEMORY, &args,
+				       sizeof(args), &chan->gart);
 		if (ret)
 			return ret;
-
-		chan->vram = vram;
-		chan->gart = gart;
 	}
 
 	/* initialise dma tracking parameters */
-	switch (nv_hclass(chan->object) & 0x00ff) {
+	switch (chan->object->oclass & 0x00ff) {
 	case 0x006b:
 	case 0x006e:
 		chan->user_put = 0x40;
@@ -347,13 +367,13 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 		OUT_RING(chan, 0x00000000);
 
 	/* allocate software object class (used for fences on <= nv05) */
-	if (device->card_type < NV_10) {
-		ret = nouveau_object_new(nv_object(client), chan->handle,
-					 NvSw, 0x006e, NULL, 0, &object);
+	if (device->info.family < NV_DEVICE_INFO_V0_CELSIUS) {
+		ret = nvif_object_init(chan->object, NULL, 0x006e, 0x006e,
+				       NULL, 0, &chan->nvsw);
 		if (ret)
 			return ret;
 
-		swch = (void *)object->parent;
+		swch = (void *)nvkm_object(&chan->nvsw)->parent;
 		swch->flip = nouveau_flip_complete;
 		swch->flip_data = chan;
 
@@ -362,37 +382,48 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 			return ret;
 
 		BEGIN_NV04(chan, NvSubSw, 0x0000, 1);
-		OUT_RING  (chan, NvSw);
+		OUT_RING  (chan, chan->nvsw.handle);
 		FIRE_RING (chan);
 	}
 
 	/* initialise synchronisation */
-	return nouveau_fence(chan->drm)->context_new(chan);
+	save = cli->base.super;
+	cli->base.super = true; /* hack until fencenv50 fixed */
+	ret = nouveau_fence(chan->drm)->context_new(chan);
+	cli->base.super = save;
+	return ret;
 }
 
 int
-nouveau_channel_new(struct nouveau_drm *drm, struct nouveau_cli *cli,
-		    u32 parent, u32 handle, u32 arg0, u32 arg1,
+nouveau_channel_new(struct nouveau_drm *drm, struct nvif_device *device,
+		    u32 handle, u32 arg0, u32 arg1,
 		    struct nouveau_channel **pchan)
 {
+	struct nouveau_cli *cli = (void *)nvif_client(&device->base);
+	bool super;
 	int ret;
 
-	ret = nouveau_channel_ind(drm, cli, parent, handle, arg0, pchan);
+	/* hack until fencenv50 is fixed, and agp access relaxed */
+	super = cli->base.super;
+	cli->base.super = true;
+
+	ret = nouveau_channel_ind(drm, device, handle, arg0, pchan);
 	if (ret) {
-		NV_DEBUG(cli, "ib channel create, %d\n", ret);
-		ret = nouveau_channel_dma(drm, cli, parent, handle, pchan);
+		NV_PRINTK(debug, cli, "ib channel create, %d\n", ret);
+		ret = nouveau_channel_dma(drm, device, handle, pchan);
 		if (ret) {
-			NV_DEBUG(cli, "dma channel create, %d\n", ret);
-			return ret;
+			NV_PRINTK(debug, cli, "dma channel create, %d\n", ret);
+			goto done;
 		}
 	}
 
 	ret = nouveau_channel_init(*pchan, arg0, arg1);
 	if (ret) {
-		NV_ERROR(cli, "channel failed to initialise, %d\n", ret);
+		NV_PRINTK(error, cli, "channel failed to initialise, %d\n", ret);
 		nouveau_channel_del(pchan);
-		return ret;
 	}
 
-	return 0;
+done:
+	cli->base.super = super;
+	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.h b/drivers/gpu/drm/nouveau/nouveau_chan.h
index 40f97e2c47b6..8309c24ee698 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.h
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.h
@@ -1,20 +1,23 @@
 #ifndef __NOUVEAU_CHAN_H__
 #define __NOUVEAU_CHAN_H__
 
-struct nouveau_cli;
+#include <nvif/object.h>
+struct nvif_device;
 
 struct nouveau_channel {
-	struct nouveau_cli *cli;
+	struct nvif_device *device;
 	struct nouveau_drm *drm;
 
-	u32 handle;
-	u32 vram;
-	u32 gart;
+	int chid;
+
+	struct nvif_object vram;
+	struct nvif_object gart;
+	struct nvif_object nvsw;
 
 	struct {
 		struct nouveau_bo *buffer;
 		struct nouveau_vma vma;
-		u32 handle;
+		struct nvif_object ctxdma;
 	} push;
 
 	/* TODO: this will be reworked in the near future */
@@ -34,14 +37,16 @@ struct nouveau_channel {
 	u32 user_get;
 	u32 user_put;
 
-	struct nouveau_object *object;
+	struct nvif_object *object;
 };
 
 
-int  nouveau_channel_new(struct nouveau_drm *, struct nouveau_cli *,
-			 u32 parent, u32 handle, u32 arg0, u32 arg1,
+int  nouveau_channel_new(struct nouveau_drm *, struct nvif_device *,
+			 u32 handle, u32 arg0, u32 arg1,
 			 struct nouveau_channel **);
 void nouveau_channel_del(struct nouveau_channel **);
 int  nouveau_channel_idle(struct nouveau_channel *);
 
+extern int nouveau_vram_pushbuf;
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 1fa222e8f007..c8ac9482cf2e 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -42,20 +42,18 @@
 #include "nouveau_encoder.h"
 #include "nouveau_crtc.h"
 
-#include <subdev/i2c.h>
-#include <subdev/gpio.h>
-#include <engine/disp.h>
+#include <nvif/event.h>
 
 MODULE_PARM_DESC(tv_disable, "Disable TV-out detection");
-static int nouveau_tv_disable = 0;
+int nouveau_tv_disable = 0;
 module_param_named(tv_disable, nouveau_tv_disable, int, 0400);
 
 MODULE_PARM_DESC(ignorelid, "Ignore ACPI lid status");
-static int nouveau_ignorelid = 0;
+int nouveau_ignorelid = 0;
 module_param_named(ignorelid, nouveau_ignorelid, int, 0400);
 
 MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (default: enabled)");
-static int nouveau_duallink = 1;
+int nouveau_duallink = 1;
 module_param_named(duallink, nouveau_duallink, int, 0400);
 
 struct nouveau_encoder *
@@ -63,7 +61,7 @@ find_encoder(struct drm_connector *connector, int type)
 {
 	struct drm_device *dev = connector->dev;
 	struct nouveau_encoder *nv_encoder;
-	struct drm_mode_object *obj;
+	struct drm_encoder *enc;
 	int i, id;
 
 	for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) {
@@ -71,10 +69,10 @@ find_encoder(struct drm_connector *connector, int type)
 		if (!id)
 			break;
 
-		obj = drm_mode_object_find(dev, id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		enc = drm_encoder_find(dev, id);
+		if (!enc)
 			continue;
-		nv_encoder = nouveau_encoder(obj_to_encoder(obj));
+		nv_encoder = nouveau_encoder(enc);
 
 		if (type == DCB_OUTPUT_ANY ||
 		    (nv_encoder->dcb && nv_encoder->dcb->type == type))
@@ -102,9 +100,9 @@ static void
 nouveau_connector_destroy(struct drm_connector *connector)
 {
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
-	nouveau_event_ref(NULL, &nv_connector->hpd);
+	nvif_notify_fini(&nv_connector->hpd);
 	kfree(nv_connector->edid);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	if (nv_connector->aux.transfer)
 		drm_dp_aux_unregister(&nv_connector->aux);
@@ -117,9 +115,9 @@ nouveau_connector_ddc_detect(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_gpio *gpio = nouveau_gpio(drm->device);
+	struct nouveau_gpio *gpio = nvkm_gpio(&drm->device);
 	struct nouveau_encoder *nv_encoder;
-	struct drm_mode_object *obj;
+	struct drm_encoder *encoder;
 	int i, panel = -ENODEV;
 
 	/* eDP panels need powering on by us (if the VBIOS doesn't default it
@@ -139,10 +137,10 @@ nouveau_connector_ddc_detect(struct drm_connector *connector)
 		if (id == 0)
 			break;
 
-		obj = drm_mode_object_find(dev, id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(dev, id);
+		if (!encoder)
 			continue;
-		nv_encoder = nouveau_encoder(obj_to_encoder(obj));
+		nv_encoder = nouveau_encoder(encoder);
 
 		if (nv_encoder->dcb->type == DCB_OUTPUT_DP) {
 			int ret = nouveau_dp_detect(nv_encoder);
@@ -206,7 +204,7 @@ nouveau_connector_set_encoder(struct drm_connector *connector,
 		return;
 	nv_connector->detected_encoder = nv_encoder;
 
-	if (nv_device(drm->device)->card_type >= NV_50) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
 		connector->interlace_allowed = true;
 		connector->doublescan_allowed = true;
 	} else
@@ -216,9 +214,8 @@ nouveau_connector_set_encoder(struct drm_connector *connector,
 		connector->interlace_allowed = false;
 	} else {
 		connector->doublescan_allowed = true;
-		if (nv_device(drm->device)->card_type == NV_20 ||
-		    ((nv_device(drm->device)->card_type == NV_10 ||
-		      nv_device(drm->device)->card_type == NV_11) &&
+		if (drm->device.info.family == NV_DEVICE_INFO_V0_KELVIN ||
+		    (drm->device.info.family == NV_DEVICE_INFO_V0_CELSIUS &&
 		     (dev->pdev->device & 0x0ff0) != 0x0100 &&
 		     (dev->pdev->device & 0x0ff0) != 0x0150))
 			/* HW is broken */
@@ -802,11 +799,11 @@ get_tmds_link_bandwidth(struct drm_connector *connector)
 	struct dcb_output *dcb = nv_connector->detected_encoder->dcb;
 
 	if (dcb->location != DCB_LOC_ON_CHIP ||
-	    nv_device(drm->device)->chipset >= 0x46)
+	    drm->device.info.chipset >= 0x46)
 		return 165000;
-	else if (nv_device(drm->device)->chipset >= 0x40)
+	else if (drm->device.info.chipset >= 0x40)
 		return 155000;
-	else if (nv_device(drm->device)->chipset >= 0x18)
+	else if (drm->device.info.chipset >= 0x18)
 		return 135000;
 	else
 		return 112000;
@@ -939,18 +936,19 @@ nouveau_connector_funcs_dp = {
 	.force = nouveau_connector_force
 };
 
-static void
-nouveau_connector_hotplug_work(struct work_struct *work)
+static int
+nouveau_connector_hotplug(struct nvif_notify *notify)
 {
 	struct nouveau_connector *nv_connector =
-		container_of(work, typeof(*nv_connector), work);
+		container_of(notify, typeof(*nv_connector), hpd);
 	struct drm_connector *connector = &nv_connector->base;
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
+	const struct nvif_notify_conn_rep_v0 *rep = notify->data;
 	const char *name = connector->name;
 
-	if (nv_connector->status & NVKM_HPD_IRQ) {
+	if (rep->mask & NVIF_NOTIFY_CONN_V0_IRQ) {
 	} else {
-		bool plugged = (nv_connector->status != NVKM_HPD_UNPLUG);
+		bool plugged = (rep->mask != NVIF_NOTIFY_CONN_V0_UNPLUG);
 
 		NV_DEBUG(drm, "%splugged %s\n", plugged ? "" : "un", name);
 
@@ -961,16 +959,7 @@ nouveau_connector_hotplug_work(struct work_struct *work)
 		drm_helper_hpd_irq_event(connector->dev);
 	}
 
-	nouveau_event_get(nv_connector->hpd);
-}
-
-static int
-nouveau_connector_hotplug(void *data, u32 type, int index)
-{
-	struct nouveau_connector *nv_connector = data;
-	nv_connector->status = type;
-	schedule_work(&nv_connector->work);
-	return NVKM_EVENT_DROP;
+	return NVIF_NOTIFY_KEEP;
 }
 
 static ssize_t
@@ -1040,7 +1029,6 @@ nouveau_connector_create(struct drm_device *dev, int index)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_display *disp = nouveau_display(dev);
 	struct nouveau_connector *nv_connector = NULL;
-	struct nouveau_disp *pdisp = nouveau_disp(drm->device);
 	struct drm_connector *connector;
 	int type, ret = 0;
 	bool dummy;
@@ -1194,7 +1182,7 @@ nouveau_connector_create(struct drm_device *dev, int index)
 
 	switch (nv_connector->type) {
 	case DCB_CONNECTOR_VGA:
-		if (nv_device(drm->device)->card_type >= NV_50) {
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
 			drm_object_attach_property(&connector->base,
 					dev->mode_config.scaling_mode_property,
 					nv_connector->scaling_mode);
@@ -1226,16 +1214,20 @@ nouveau_connector_create(struct drm_device *dev, int index)
 		break;
 	}
 
-	ret = nouveau_event_new(pdisp->hpd, NVKM_HPD, index,
-				nouveau_connector_hotplug,
-				nv_connector, &nv_connector->hpd);
+	ret = nvif_notify_init(&disp->disp, NULL, nouveau_connector_hotplug,
+				true, NV04_DISP_NTFY_CONN,
+			       &(struct nvif_notify_conn_req_v0) {
+				.mask = NVIF_NOTIFY_CONN_V0_ANY,
+				.conn = index,
+			       },
+			       sizeof(struct nvif_notify_conn_req_v0),
+			       sizeof(struct nvif_notify_conn_rep_v0),
+			       &nv_connector->hpd);
 	if (ret)
 		connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 	else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
-	INIT_WORK(&nv_connector->work, nouveau_connector_hotplug_work);
-
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return connector;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index 8861b6c579ad..629a380c7085 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -27,14 +27,12 @@
 #ifndef __NOUVEAU_CONNECTOR_H__
 #define __NOUVEAU_CONNECTOR_H__
 
+#include <nvif/notify.h>
+
 #include <drm/drm_edid.h>
 #include <drm/drm_dp_helper.h>
 #include "nouveau_crtc.h"
 
-#include <core/event.h>
-
-#include <subdev/bios.h>
-
 struct nouveau_i2c_port;
 
 enum nouveau_underscan_type {
@@ -67,9 +65,7 @@ struct nouveau_connector {
 	u8 index;
 	u8 *dcb;
 
-	struct nouveau_eventh *hpd;
-	u32 status;
-	struct work_struct work;
+	struct nvif_notify hpd;
 
 	struct drm_dp_aux aux;
 
@@ -109,4 +105,8 @@ nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc)
 struct drm_connector *
 nouveau_connector_create(struct drm_device *, int index);
 
+extern int nouveau_tv_disable;
+extern int nouveau_ignorelid;
+extern int nouveau_duallink;
+
 #endif /* __NOUVEAU_CONNECTOR_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_crtc.h b/drivers/gpu/drm/nouveau/nouveau_crtc.h
index a0534489d23f..f19cb1c5fc5a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_crtc.h
+++ b/drivers/gpu/drm/nouveau/nouveau_crtc.h
@@ -27,10 +27,13 @@
 #ifndef __NOUVEAU_CRTC_H__
 #define __NOUVEAU_CRTC_H__
 
+#include <nvif/notify.h>
+
 struct nouveau_crtc {
 	struct drm_crtc base;
 
 	int index;
+	struct nvif_notify vblank;
 
 	uint32_t dpms_saved_fp_control;
 	uint32_t fp_users;
@@ -46,7 +49,7 @@ struct nouveau_crtc {
 		int cpp;
 		bool blanked;
 		uint32_t offset;
-		uint32_t tile_flags;
+		uint32_t handle;
 	} fb;
 
 	struct {
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 47ad74255bf1..a88e6927f571 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -27,6 +27,8 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 
+#include <nvif/class.h>
+
 #include "nouveau_fbcon.h"
 #include "dispnv04/hw.h"
 #include "nouveau_crtc.h"
@@ -37,35 +39,42 @@
 
 #include "nouveau_fence.h"
 
-#include <engine/disp.h>
-
-#include <core/class.h>
+#include <nvif/event.h>
 
 static int
-nouveau_display_vblank_handler(void *data, u32 type, int head)
+nouveau_display_vblank_handler(struct nvif_notify *notify)
 {
-	struct nouveau_drm *drm = data;
-	drm_handle_vblank(drm->dev, head);
-	return NVKM_EVENT_KEEP;
+	struct nouveau_crtc *nv_crtc =
+		container_of(notify, typeof(*nv_crtc), vblank);
+	drm_handle_vblank(nv_crtc->base.dev, nv_crtc->index);
+	return NVIF_NOTIFY_KEEP;
 }
 
 int
 nouveau_display_vblank_enable(struct drm_device *dev, int head)
 {
-	struct nouveau_display *disp = nouveau_display(dev);
-	if (disp) {
-		nouveau_event_get(disp->vblank[head]);
-		return 0;
+	struct drm_crtc *crtc;
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		if (nv_crtc->index == head) {
+			nvif_notify_get(&nv_crtc->vblank);
+			return 0;
+		}
 	}
-	return -EIO;
+	return -EINVAL;
 }
 
 void
 nouveau_display_vblank_disable(struct drm_device *dev, int head)
 {
-	struct nouveau_display *disp = nouveau_display(dev);
-	if (disp)
-		nouveau_event_put(disp->vblank[head]);
+	struct drm_crtc *crtc;
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		if (nv_crtc->index == head) {
+			nvif_notify_put(&nv_crtc->vblank);
+			return;
+		}
+	}
 }
 
 static inline int
@@ -86,17 +95,22 @@ int
 nouveau_display_scanoutpos_head(struct drm_crtc *crtc, int *vpos, int *hpos,
 				ktime_t *stime, ktime_t *etime)
 {
-	const u32 mthd = NV04_DISP_SCANOUTPOS + nouveau_crtc(crtc)->index;
+	struct {
+		struct nv04_disp_mthd_v0 base;
+		struct nv04_disp_scanoutpos_v0 scan;
+	} args = {
+		.base.method = NV04_DISP_SCANOUTPOS,
+		.base.head = nouveau_crtc(crtc)->index,
+	};
 	struct nouveau_display *disp = nouveau_display(crtc->dev);
-	struct nv04_display_scanoutpos args;
 	int ret, retry = 1;
 
 	do {
-		ret = nv_exec(disp->core, mthd, &args, sizeof(args));
+		ret = nvif_mthd(&disp->disp, 0, &args, sizeof(args));
 		if (ret != 0)
 			return 0;
 
-		if (args.vline) {
+		if (args.scan.vline) {
 			ret |= DRM_SCANOUTPOS_ACCURATE;
 			ret |= DRM_SCANOUTPOS_VALID;
 			break;
@@ -105,13 +119,14 @@ nouveau_display_scanoutpos_head(struct drm_crtc *crtc, int *vpos, int *hpos,
 		if (retry) ndelay(crtc->linedur_ns);
 	} while (retry--);
 
-	*hpos = args.hline;
-	*vpos = calc(args.vblanks, args.vblanke, args.vtotal, args.vline);
-	if (stime) *stime = ns_to_ktime(args.time[0]);
-	if (etime) *etime = ns_to_ktime(args.time[1]);
+	*hpos = args.scan.hline;
+	*vpos = calc(args.scan.vblanks, args.scan.vblanke,
+		     args.scan.vtotal, args.scan.vline);
+	if (stime) *stime = ns_to_ktime(args.scan.time[0]);
+	if (etime) *etime = ns_to_ktime(args.scan.time[1]);
 
 	if (*vpos < 0)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 	return ret;
 }
 
@@ -151,16 +166,13 @@ nouveau_display_vblstamp(struct drm_device *dev, int head, int *max_error,
 static void
 nouveau_display_vblank_fini(struct drm_device *dev)
 {
-	struct nouveau_display *disp = nouveau_display(dev);
-	int i;
+	struct drm_crtc *crtc;
 
 	drm_vblank_cleanup(dev);
 
-	if (disp->vblank) {
-		for (i = 0; i < dev->mode_config.num_crtc; i++)
-			nouveau_event_ref(NULL, &disp->vblank[i]);
-		kfree(disp->vblank);
-		disp->vblank = NULL;
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		nvif_notify_fini(&nv_crtc->vblank);
 	}
 }
 
@@ -168,19 +180,20 @@ static int
 nouveau_display_vblank_init(struct drm_device *dev)
 {
 	struct nouveau_display *disp = nouveau_display(dev);
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_disp *pdisp = nouveau_disp(drm->device);
-	int ret, i;
-
-	disp->vblank = kzalloc(dev->mode_config.num_crtc *
-			       sizeof(*disp->vblank), GFP_KERNEL);
-	if (!disp->vblank)
-		return -ENOMEM;
+	struct drm_crtc *crtc;
+	int ret;
 
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		ret = nouveau_event_new(pdisp->vblank, 1, i,
-					nouveau_display_vblank_handler,
-					drm, &disp->vblank[i]);
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		ret = nvif_notify_init(&disp->disp, NULL,
+				       nouveau_display_vblank_handler, false,
+				       NV04_DISP_NTFY_VBLANK,
+				       &(struct nvif_notify_head_req_v0) {
+					.head = nv_crtc->index,
+				       },
+				       sizeof(struct nvif_notify_head_req_v0),
+				       sizeof(struct nvif_notify_head_rep_v0),
+				       &nv_crtc->vblank);
 		if (ret) {
 			nouveau_display_vblank_fini(dev);
 			return ret;
@@ -200,6 +213,10 @@ static void
 nouveau_user_framebuffer_destroy(struct drm_framebuffer *drm_fb)
 {
 	struct nouveau_framebuffer *fb = nouveau_framebuffer(drm_fb);
+	struct nouveau_display *disp = nouveau_display(drm_fb->dev);
+
+	if (disp->fb_dtor)
+		disp->fb_dtor(drm_fb);
 
 	if (fb->nvbo)
 		drm_gem_object_unreference_unlocked(&fb->nvbo->gem);
@@ -229,63 +246,24 @@ nouveau_framebuffer_init(struct drm_device *dev,
 			 struct drm_mode_fb_cmd2 *mode_cmd,
 			 struct nouveau_bo *nvbo)
 {
-	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_display *disp = nouveau_display(dev);
 	struct drm_framebuffer *fb = &nv_fb->base;
 	int ret;
 
 	drm_helper_mode_fill_fb_struct(fb, mode_cmd);
 	nv_fb->nvbo = nvbo;
 
-	if (nv_device(drm->device)->card_type >= NV_50) {
-		u32 tile_flags = nouveau_bo_tile_layout(nvbo);
-		if (tile_flags == 0x7a00 ||
-		    tile_flags == 0xfe00)
-			nv_fb->r_dma = NvEvoFB32;
-		else
-		if (tile_flags == 0x7000)
-			nv_fb->r_dma = NvEvoFB16;
-		else
-			nv_fb->r_dma = NvEvoVRAM_LP;
-
-		switch (fb->depth) {
-		case  8: nv_fb->r_format = 0x1e00; break;
-		case 15: nv_fb->r_format = 0xe900; break;
-		case 16: nv_fb->r_format = 0xe800; break;
-		case 24:
-		case 32: nv_fb->r_format = 0xcf00; break;
-		case 30: nv_fb->r_format = 0xd100; break;
-		default:
-			 NV_ERROR(drm, "unknown depth %d\n", fb->depth);
-			 return -EINVAL;
-		}
-
-		if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG) {
-			NV_ERROR(drm, "framebuffer requires contiguous bo\n");
-			return -EINVAL;
-		}
-
-		if (nv_device(drm->device)->chipset == 0x50)
-			nv_fb->r_format |= (tile_flags << 8);
-
-		if (!tile_flags) {
-			if (nv_device(drm->device)->card_type < NV_D0)
-				nv_fb->r_pitch = 0x00100000 | fb->pitches[0];
-			else
-				nv_fb->r_pitch = 0x01000000 | fb->pitches[0];
-		} else {
-			u32 mode = nvbo->tile_mode;
-			if (nv_device(drm->device)->card_type >= NV_C0)
-				mode >>= 4;
-			nv_fb->r_pitch = ((fb->pitches[0] / 4) << 4) | mode;
-		}
-	}
-
 	ret = drm_framebuffer_init(dev, fb, &nouveau_framebuffer_funcs);
-	if (ret) {
+	if (ret)
 		return ret;
+
+	if (disp->fb_ctor) {
+		ret = disp->fb_ctor(fb);
+		if (ret)
+			disp->fb_dtor(fb);
 	}
 
-	return 0;
+	return ret;
 }
 
 static struct drm_framebuffer *
@@ -393,7 +371,7 @@ nouveau_display_init(struct drm_device *dev)
 	/* enable hotplug interrupts */
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
-		if (conn->hpd) nouveau_event_get(conn->hpd);
+		nvif_notify_get(&conn->hpd);
 	}
 
 	return ret;
@@ -404,37 +382,32 @@ nouveau_display_fini(struct drm_device *dev)
 {
 	struct nouveau_display *disp = nouveau_display(dev);
 	struct drm_connector *connector;
+	int head;
+
+	/* Make sure that drm and hw vblank irqs get properly disabled. */
+	for (head = 0; head < dev->mode_config.num_crtc; head++)
+		drm_vblank_off(dev, head);
 
 	/* disable hotplug interrupts */
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct nouveau_connector *conn = nouveau_connector(connector);
-		if (conn->hpd) nouveau_event_put(conn->hpd);
+		nvif_notify_put(&conn->hpd);
 	}
 
 	drm_kms_helper_poll_disable(dev);
 	disp->fini(dev);
 }
 
-int
-nouveau_display_create(struct drm_device *dev)
+static void
+nouveau_display_create_properties(struct drm_device *dev)
 {
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nouveau_dev(dev);
-	struct nouveau_display *disp;
-	int ret, gen;
-
-	disp = drm->display = kzalloc(sizeof(*disp), GFP_KERNEL);
-	if (!disp)
-		return -ENOMEM;
-
-	drm_mode_config_init(dev);
-	drm_mode_create_scaling_mode_property(dev);
-	drm_mode_create_dvi_i_properties(dev);
+	struct nouveau_display *disp = nouveau_display(dev);
+	int gen;
 
-	if (nv_device(drm->device)->card_type < NV_50)
+	if (disp->disp.oclass < NV50_DISP)
 		gen = 0;
 	else
-	if (nv_device(drm->device)->card_type < NV_D0)
+	if (disp->disp.oclass < GF110_DISP)
 		gen = 1;
 	else
 		gen = 2;
@@ -449,26 +422,43 @@ nouveau_display_create(struct drm_device *dev)
 	disp->underscan_vborder_property =
 		drm_property_create_range(dev, 0, "underscan vborder", 0, 128);
 
-	if (gen >= 1) {
-		/* -90..+90 */
-		disp->vibrant_hue_property =
-			drm_property_create_range(dev, 0, "vibrant hue", 0, 180);
+	if (gen < 1)
+		return;
 
-		/* -100..+100 */
-		disp->color_vibrance_property =
-			drm_property_create_range(dev, 0, "color vibrance", 0, 200);
-	}
+	/* -90..+90 */
+	disp->vibrant_hue_property =
+		drm_property_create_range(dev, 0, "vibrant hue", 0, 180);
+
+	/* -100..+100 */
+	disp->color_vibrance_property =
+		drm_property_create_range(dev, 0, "color vibrance", 0, 200);
+}
+
+int
+nouveau_display_create(struct drm_device *dev)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_display *disp;
+	int ret;
+
+	disp = drm->display = kzalloc(sizeof(*disp), GFP_KERNEL);
+	if (!disp)
+		return -ENOMEM;
+
+	drm_mode_config_init(dev);
+	drm_mode_create_scaling_mode_property(dev);
+	drm_mode_create_dvi_i_properties(dev);
 
 	dev->mode_config.funcs = &nouveau_mode_config_funcs;
-	dev->mode_config.fb_base = nv_device_resource_start(device, 1);
+	dev->mode_config.fb_base = nv_device_resource_start(nvkm_device(&drm->device), 1);
 
 	dev->mode_config.min_width = 0;
 	dev->mode_config.min_height = 0;
-	if (nv_device(drm->device)->card_type < NV_10) {
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_CELSIUS) {
 		dev->mode_config.max_width = 2048;
 		dev->mode_config.max_height = 2048;
 	} else
-	if (nv_device(drm->device)->card_type < NV_50) {
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 		dev->mode_config.max_width = 4096;
 		dev->mode_config.max_height = 4096;
 	} else {
@@ -479,7 +469,7 @@ nouveau_display_create(struct drm_device *dev)
 	dev->mode_config.preferred_depth = 24;
 	dev->mode_config.prefer_shadow = 1;
 
-	if (nv_device(drm->device)->chipset < 0x11)
+	if (drm->device.info.chipset < 0x11)
 		dev->mode_config.async_page_flip = false;
 	else
 		dev->mode_config.async_page_flip = true;
@@ -487,29 +477,30 @@ nouveau_display_create(struct drm_device *dev)
 	drm_kms_helper_poll_init(dev);
 	drm_kms_helper_poll_disable(dev);
 
-	if (drm->vbios.dcb.entries) {
+	if (nouveau_modeset != 2 && drm->vbios.dcb.entries) {
 		static const u16 oclass[] = {
-			GM107_DISP_CLASS,
-			NVF0_DISP_CLASS,
-			NVE0_DISP_CLASS,
-			NVD0_DISP_CLASS,
-			NVA3_DISP_CLASS,
-			NV94_DISP_CLASS,
-			NVA0_DISP_CLASS,
-			NV84_DISP_CLASS,
-			NV50_DISP_CLASS,
-			NV04_DISP_CLASS,
+			GM107_DISP,
+			GK110_DISP,
+			GK104_DISP,
+			GF110_DISP,
+			GT214_DISP,
+			GT206_DISP,
+			GT200_DISP,
+			G82_DISP,
+			NV50_DISP,
+			NV04_DISP,
 		};
 		int i;
 
 		for (i = 0, ret = -ENODEV; ret && i < ARRAY_SIZE(oclass); i++) {
-			ret = nouveau_object_new(nv_object(drm), NVDRM_DEVICE,
-						 NVDRM_DISPLAY, oclass[i],
-						 NULL, 0, &disp->core);
+			ret = nvif_object_init(nvif_object(&drm->device), NULL,
+					       NVDRM_DISPLAY, oclass[i],
+					       NULL, 0, &disp->disp);
 		}
 
 		if (ret == 0) {
-			if (nv_mclass(disp->core) < NV50_DISP_CLASS)
+			nouveau_display_create_properties(dev);
+			if (disp->disp.oclass < NV50_DISP)
 				ret = nv04_display_create(dev);
 			else
 				ret = nv50_display_create(dev);
@@ -542,7 +533,6 @@ void
 nouveau_display_destroy(struct drm_device *dev)
 {
 	struct nouveau_display *disp = nouveau_display(dev);
-	struct nouveau_drm *drm = nouveau_drm(dev);
 
 	nouveau_backlight_exit(dev);
 	nouveau_display_vblank_fini(dev);
@@ -553,21 +543,19 @@ nouveau_display_destroy(struct drm_device *dev)
 	if (disp->dtor)
 		disp->dtor(dev);
 
-	nouveau_object_del(nv_object(drm), NVDRM_DEVICE, NVDRM_DISPLAY);
+	nvif_object_fini(&disp->disp);
 
 	nouveau_drm(dev)->display = NULL;
 	kfree(disp);
 }
 
 int
-nouveau_display_suspend(struct drm_device *dev)
+nouveau_display_suspend(struct drm_device *dev, bool runtime)
 {
-	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct drm_crtc *crtc;
 
 	nouveau_display_fini(dev);
 
-	NV_INFO(drm, "unpinning framebuffer(s)...\n");
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_framebuffer *nouveau_fb;
 
@@ -589,12 +577,13 @@ nouveau_display_suspend(struct drm_device *dev)
 }
 
 void
-nouveau_display_repin(struct drm_device *dev)
+nouveau_display_resume(struct drm_device *dev, bool runtime)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct drm_crtc *crtc;
-	int ret;
+	int ret, head;
 
+	/* re-pin fb/cursors */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_framebuffer *nouveau_fb;
 
@@ -602,7 +591,9 @@ nouveau_display_repin(struct drm_device *dev)
 		if (!nouveau_fb || !nouveau_fb->nvbo)
 			continue;
 
-		nouveau_bo_pin(nouveau_fb->nvbo, TTM_PL_FLAG_VRAM);
+		ret = nouveau_bo_pin(nouveau_fb->nvbo, TTM_PL_FLAG_VRAM);
+		if (ret)
+			NV_ERROR(drm, "Could not pin framebuffer\n");
 	}
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -614,12 +605,7 @@ nouveau_display_repin(struct drm_device *dev)
 		if (ret)
 			NV_ERROR(drm, "Could not pin/map cursor.\n");
 	}
-}
 
-void
-nouveau_display_resume(struct drm_device *dev)
-{
-	struct drm_crtc *crtc;
 	nouveau_display_init(dev);
 
 	/* Force CLUT to get re-loaded during modeset */
@@ -629,6 +615,17 @@ nouveau_display_resume(struct drm_device *dev)
 		nv_crtc->lut.depth = 0;
 	}
 
+	/* Make sure that drm and hw vblank irqs get resumed if needed. */
+	for (head = 0; head < dev->mode_config.num_crtc; head++)
+		drm_vblank_on(dev, head);
+
+	/* This should ensure we don't hit a locking problem when someone
+	 * wakes us up via a connector.  We should never go into suspend
+	 * while the display is on anyways.
+	 */
+	if (runtime)
+		return;
+
 	drm_helper_resume_force_mode(dev);
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -660,7 +657,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
 	/* Synchronize with the old framebuffer */
-	ret = nouveau_fence_sync(old_bo->bo.sync_obj, chan);
+	ret = nouveau_fence_sync(old_bo, chan, false, false);
 	if (ret)
 		goto fail;
 
@@ -669,7 +666,7 @@ nouveau_page_flip_emit(struct nouveau_channel *chan,
 	if (ret)
 		goto fail;
 
-	if (nv_device(drm->device)->card_type < NV_C0)
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_FERMI)
 		BEGIN_NV04(chan, NvSubSw, NV_SW_PAGE_FLIP, 1);
 	else
 		BEGIN_NVC0(chan, FermiSw, NV_SW_PAGE_FLIP, 1);
@@ -698,12 +695,15 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	struct nouveau_bo *old_bo = nouveau_framebuffer(crtc->primary->fb)->nvbo;
 	struct nouveau_bo *new_bo = nouveau_framebuffer(fb)->nvbo;
 	struct nouveau_page_flip_state *s;
-	struct nouveau_channel *chan = drm->channel;
+	struct nouveau_channel *chan;
+	struct nouveau_cli *cli;
 	struct nouveau_fence *fence;
 	int ret;
 
-	if (!drm->channel)
+	chan = drm->channel;
+	if (!chan)
 		return -ENODEV;
+	cli = (void *)nvif_client(&chan->device->base);
 
 	s = kzalloc(sizeof(*s), GFP_KERNEL);
 	if (!s)
@@ -715,20 +715,25 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			goto fail_free;
 	}
 
-	mutex_lock(&chan->cli->mutex);
-
-	/* synchronise rendering channel with the kernel's channel */
-	spin_lock(&new_bo->bo.bdev->fence_lock);
-	fence = nouveau_fence_ref(new_bo->bo.sync_obj);
-	spin_unlock(&new_bo->bo.bdev->fence_lock);
-	ret = nouveau_fence_sync(fence, chan);
-	nouveau_fence_unref(&fence);
+	mutex_lock(&cli->mutex);
+	ret = ttm_bo_reserve(&new_bo->bo, true, false, false, NULL);
 	if (ret)
 		goto fail_unpin;
 
-	ret = ttm_bo_reserve(&old_bo->bo, true, false, false, NULL);
-	if (ret)
+	/* synchronise rendering channel with the kernel's channel */
+	ret = nouveau_fence_sync(new_bo, chan, false, true);
+	if (ret) {
+		ttm_bo_unreserve(&new_bo->bo);
 		goto fail_unpin;
+	}
+
+	if (new_bo != old_bo) {
+		ttm_bo_unreserve(&new_bo->bo);
+
+		ret = ttm_bo_reserve(&old_bo->bo, true, false, false, NULL);
+		if (ret)
+			goto fail_unpin;
+	}
 
 	/* Initialize a page flip struct */
 	*s = (struct nouveau_page_flip_state)
@@ -740,7 +745,7 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	drm_vblank_get(dev, nouveau_crtc(crtc)->index);
 
 	/* Emit a page flip */
-	if (nv_device(drm->device)->card_type >= NV_50) {
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
 		ret = nv50_display_flip_next(crtc, fb, chan, swap_interval);
 		if (ret)
 			goto fail_unreserve;
@@ -769,12 +774,12 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	ret = nouveau_page_flip_emit(chan, old_bo, new_bo, s, &fence);
 	if (ret)
 		goto fail_unreserve;
-	mutex_unlock(&chan->cli->mutex);
+	mutex_unlock(&cli->mutex);
 
 	/* Update the crtc struct and cleanup */
 	crtc->primary->fb = fb;
 
-	nouveau_bo_fence(old_bo, fence);
+	nouveau_bo_fence(old_bo, fence, false);
 	ttm_bo_unreserve(&old_bo->bo);
 	if (old_bo != new_bo)
 		nouveau_bo_unpin(old_bo);
@@ -785,7 +790,7 @@ fail_unreserve:
 	drm_vblank_put(dev, nouveau_crtc(crtc)->index);
 	ttm_bo_unreserve(&old_bo->bo);
 fail_unpin:
-	mutex_unlock(&chan->cli->mutex);
+	mutex_unlock(&cli->mutex);
 	if (old_bo != new_bo)
 		nouveau_bo_unpin(new_bo);
 fail_free:
@@ -815,7 +820,7 @@ nouveau_finish_page_flip(struct nouveau_channel *chan,
 	s = list_first_entry(&fctx->flip, struct nouveau_page_flip_state, head);
 	if (s->event) {
 		/* Vblank timestamps/counts are only correct on >= NV-50 */
-		if (nv_device(drm->device)->card_type >= NV_50)
+		if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA)
 			crtcid = s->crtc;
 
 		drm_send_vblank_event(dev, crtcid, s->event);
@@ -841,7 +846,7 @@ nouveau_flip_complete(void *data)
 	struct nouveau_page_flip_state state;
 
 	if (!nouveau_finish_page_flip(chan, &state)) {
-		if (nv_device(drm->device)->card_type < NV_50) {
+		if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 			nv_set_crtc_base(drm->dev, state.crtc, state.offset +
 					 state.y * state.pitch +
 					 state.x * state.bpp / 8);
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.h b/drivers/gpu/drm/nouveau/nouveau_display.h
index a71cf77e55b2..be3d5947c6be 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.h
+++ b/drivers/gpu/drm/nouveau/nouveau_display.h
@@ -9,9 +9,11 @@ struct nouveau_framebuffer {
 	struct drm_framebuffer base;
 	struct nouveau_bo *nvbo;
 	struct nouveau_vma vma;
-	u32 r_dma;
+	u32 r_handle;
 	u32 r_format;
 	u32 r_pitch;
+	struct nvif_object h_base[4];
+	struct nvif_object h_core;
 };
 
 static inline struct nouveau_framebuffer *
@@ -36,8 +38,10 @@ struct nouveau_display {
 	int  (*init)(struct drm_device *);
 	void (*fini)(struct drm_device *);
 
-	struct nouveau_object *core;
-	struct nouveau_eventh **vblank;
+	int  (*fb_ctor)(struct drm_framebuffer *);
+	void (*fb_dtor)(struct drm_framebuffer *);
+
+	struct nvif_object disp;
 
 	struct drm_property *dithering_mode;
 	struct drm_property *dithering_depth;
@@ -59,9 +63,8 @@ int  nouveau_display_create(struct drm_device *dev);
 void nouveau_display_destroy(struct drm_device *dev);
 int  nouveau_display_init(struct drm_device *dev);
 void nouveau_display_fini(struct drm_device *dev);
-int  nouveau_display_suspend(struct drm_device *dev);
-void nouveau_display_repin(struct drm_device *dev);
-void nouveau_display_resume(struct drm_device *dev);
+int  nouveau_display_suspend(struct drm_device *dev, bool runtime);
+void nouveau_display_resume(struct drm_device *dev, bool runtime);
 int  nouveau_display_vblank_enable(struct drm_device *, int);
 void nouveau_display_vblank_disable(struct drm_device *, int);
 int  nouveau_display_scanoutpos(struct drm_device *, int, unsigned int,
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c
index c177272152e2..8508603cc8c3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.c
@@ -24,8 +24,6 @@
  *
  */
 
-#include <core/client.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 
@@ -54,9 +52,9 @@ READ_GET(struct nouveau_channel *chan, uint64_t *prev_get, int *timeout)
 {
 	uint64_t val;
 
-	val = nv_ro32(chan->object, chan->user_get);
+	val = nvif_rd32(chan, chan->user_get);
         if (chan->user_get_hi)
-                val |= (uint64_t)nv_ro32(chan->object, chan->user_get_hi) << 32;
+                val |= (uint64_t)nvif_rd32(chan, chan->user_get_hi) << 32;
 
 	/* reset counter as long as GET is still advancing, this is
 	 * to avoid misdetecting a GPU lockup if the GPU happens to
@@ -84,12 +82,13 @@ void
 nv50_dma_push(struct nouveau_channel *chan, struct nouveau_bo *bo,
 	      int delta, int length)
 {
+	struct nouveau_cli *cli = (void *)nvif_client(&chan->device->base);
 	struct nouveau_bo *pb = chan->push.buffer;
 	struct nouveau_vma *vma;
 	int ip = (chan->dma.ib_put * 2) + chan->dma.ib_base;
 	u64 offset;
 
-	vma = nouveau_bo_vma_find(bo, nv_client(chan->cli)->vm);
+	vma = nouveau_bo_vma_find(bo, cli->vm);
 	BUG_ON(!vma);
 	offset = vma->offset + delta;
 
@@ -104,7 +103,7 @@ nv50_dma_push(struct nouveau_channel *chan, struct nouveau_bo *bo,
 	/* Flush writes. */
 	nouveau_bo_rd32(pb, 0);
 
-	nv_wo32(chan->object, 0x8c, chan->dma.ib_put);
+	nvif_wr32(chan, 0x8c, chan->dma.ib_put);
 	chan->dma.ib_free--;
 }
 
@@ -114,7 +113,7 @@ nv50_dma_push_wait(struct nouveau_channel *chan, int count)
 	uint32_t cnt = 0, prev_get = 0;
 
 	while (chan->dma.ib_free < count) {
-		uint32_t get = nv_ro32(chan->object, 0x88);
+		uint32_t get = nvif_rd32(chan, 0x88);
 		if (get != prev_get) {
 			prev_get = get;
 			cnt = 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.h b/drivers/gpu/drm/nouveau/nouveau_dma.h
index dc0e0c5cadb4..8da0a272c45a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.h
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.h
@@ -58,31 +58,14 @@ enum {
 	FermiSw		= 5, /* DO NOT CHANGE (well.. 6/7 will work...) */
 };
 
-/* Object handles. */
+/* Object handles - for stuff that's doesn't use handle == oclass. */
 enum {
-	NvM2MF		= 0x80000001,
 	NvDmaFB		= 0x80000002,
 	NvDmaTT		= 0x80000003,
 	NvNotify0       = 0x80000006,
-	Nv2D		= 0x80000007,
-	NvCtxSurf2D	= 0x80000008,
-	NvRop		= 0x80000009,
-	NvImagePatt	= 0x8000000a,
-	NvClipRect	= 0x8000000b,
-	NvGdiRect	= 0x8000000c,
-	NvImageBlit	= 0x8000000d,
-	NvSw		= 0x8000000e,
 	NvSema		= 0x8000000f,
 	NvEvoSema0	= 0x80000010,
 	NvEvoSema1	= 0x80000011,
-	NvNotify1       = 0x80000012,
-
-	/* G80+ display objects */
-	NvEvoVRAM	= 0x01000000,
-	NvEvoFB16	= 0x01000001,
-	NvEvoFB32	= 0x01000002,
-	NvEvoVRAM_LP	= 0x01000003,
-	NvEvoSync	= 0xcafe0000
 };
 
 #define NV_MEMORY_TO_MEMORY_FORMAT                                    0x00000039
@@ -157,7 +140,7 @@ BEGIN_IMC0(struct nouveau_channel *chan, int subc, int mthd, u16 data)
 #define WRITE_PUT(val) do {                                                    \
 	mb();                                                   \
 	nouveau_bo_rd32(chan->push.buffer, 0);                                 \
-	nv_wo32(chan->object, chan->user_put, ((val) << 2) + chan->push.vma.offset);  \
+	nvif_wr32(chan, chan->user_put, ((val) << 2) + chan->push.vma.offset); \
 } while (0)
 
 static inline void
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index 5675ffc175ae..c5137cccce7d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -30,11 +30,6 @@
 #include "nouveau_encoder.h"
 #include "nouveau_crtc.h"
 
-#include <core/class.h>
-
-#include <subdev/gpio.h>
-#include <subdev/i2c.h>
-
 static void
 nouveau_dp_probe_oui(struct drm_device *dev, struct nouveau_i2c_port *auxch,
 		     u8 *dpcd)
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index c9428c943afb..57238076049f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -27,21 +27,14 @@
 #include <linux/pci.h>
 #include <linux/pm_runtime.h>
 #include <linux/vga_switcheroo.h>
+
 #include "drmP.h"
 #include "drm_crtc_helper.h"
+
 #include <core/device.h>
-#include <core/client.h>
 #include <core/gpuobj.h>
-#include <core/class.h>
 #include <core/option.h>
 
-#include <engine/device.h>
-#include <engine/disp.h>
-#include <engine/fifo.h>
-#include <engine/software.h>
-
-#include <subdev/vm.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_ttm.h"
@@ -57,6 +50,8 @@
 #include "nouveau_fbcon.h"
 #include "nouveau_fence.h"
 #include "nouveau_debugfs.h"
+#include "nouveau_usif.h"
+#include "nouveau_connector.h"
 
 MODULE_PARM_DESC(config, "option string to pass to driver core");
 static char *nouveau_config;
@@ -79,7 +74,9 @@ MODULE_PARM_DESC(runpm, "disable (0), force enable (1), optimus only default (-1
 int nouveau_runtime_pm = -1;
 module_param_named(runpm, nouveau_runtime_pm, int, 0400);
 
-static struct drm_driver driver;
+static struct drm_driver driver_stub;
+static struct drm_driver driver_pci;
+static struct drm_driver driver_platform;
 
 static u64
 nouveau_pci_name(struct pci_dev *pdev)
@@ -109,40 +106,37 @@ static int
 nouveau_cli_create(u64 name, const char *sname,
 		   int size, void **pcli)
 {
-	struct nouveau_cli *cli;
-	int ret;
-
-	*pcli = NULL;
-	ret = nouveau_client_create_(sname, name, nouveau_config,
-				     nouveau_debug, size, pcli);
-	cli = *pcli;
-	if (ret) {
-		if (cli)
-			nouveau_client_destroy(&cli->base);
-		*pcli = NULL;
+	struct nouveau_cli *cli = *pcli = kzalloc(size, GFP_KERNEL);
+	if (cli) {
+		int ret = nvif_client_init(NULL, NULL, sname, name,
+					   nouveau_config, nouveau_debug,
+					  &cli->base);
+		if (ret == 0) {
+			mutex_init(&cli->mutex);
+			usif_client_init(cli);
+		}
 		return ret;
 	}
-
-	mutex_init(&cli->mutex);
-	return 0;
+	return -ENOMEM;
 }
 
 static void
 nouveau_cli_destroy(struct nouveau_cli *cli)
 {
-	struct nouveau_object *client = nv_object(cli);
-	nouveau_vm_ref(NULL, &cli->base.vm, NULL);
-	nouveau_client_fini(&cli->base, false);
-	atomic_set(&client->refcount, 1);
-	nouveau_object_ref(NULL, &client);
+	nouveau_vm_ref(NULL, &nvkm_client(&cli->base)->vm, NULL);
+	nvif_client_fini(&cli->base);
+	usif_client_fini(cli);
 }
 
 static void
 nouveau_accel_fini(struct nouveau_drm *drm)
 {
-	nouveau_gpuobj_ref(NULL, &drm->notify);
 	nouveau_channel_del(&drm->channel);
+	nvif_object_fini(&drm->ntfy);
+	nouveau_gpuobj_ref(NULL, &drm->notify);
+	nvif_object_fini(&drm->nvsw);
 	nouveau_channel_del(&drm->cechan);
+	nvif_object_fini(&drm->ttm.copy);
 	if (drm->fence)
 		nouveau_fence(drm)->dtor(drm);
 }
@@ -150,46 +144,71 @@ nouveau_accel_fini(struct nouveau_drm *drm)
 static void
 nouveau_accel_init(struct nouveau_drm *drm)
 {
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_object *object;
+	struct nvif_device *device = &drm->device;
 	u32 arg0, arg1;
-	int ret;
+	u32 sclass[16];
+	int ret, i;
 
-	if (nouveau_noaccel || !nouveau_fifo(device) /*XXX*/)
+	if (nouveau_noaccel)
 		return;
 
 	/* initialise synchronisation routines */
-	if      (device->card_type < NV_10) ret = nv04_fence_create(drm);
-	else if (device->card_type < NV_11 ||
-		 device->chipset   <  0x17) ret = nv10_fence_create(drm);
-	else if (device->card_type < NV_50) ret = nv17_fence_create(drm);
-	else if (device->chipset   <  0x84) ret = nv50_fence_create(drm);
-	else if (device->card_type < NV_C0) ret = nv84_fence_create(drm);
-	else                                ret = nvc0_fence_create(drm);
+	/*XXX: this is crap, but the fence/channel stuff is a little
+	 *     backwards in some places.  this will be fixed.
+	 */
+	ret = nvif_object_sclass(&device->base, sclass, ARRAY_SIZE(sclass));
+	if (ret < 0)
+		return;
+
+	for (ret = -ENOSYS, i = 0; ret && i < ARRAY_SIZE(sclass); i++) {
+		switch (sclass[i]) {
+		case NV03_CHANNEL_DMA:
+			ret = nv04_fence_create(drm);
+			break;
+		case NV10_CHANNEL_DMA:
+			ret = nv10_fence_create(drm);
+			break;
+		case NV17_CHANNEL_DMA:
+		case NV40_CHANNEL_DMA:
+			ret = nv17_fence_create(drm);
+			break;
+		case NV50_CHANNEL_GPFIFO:
+			ret = nv50_fence_create(drm);
+			break;
+		case G82_CHANNEL_GPFIFO:
+			ret = nv84_fence_create(drm);
+			break;
+		case FERMI_CHANNEL_GPFIFO:
+		case KEPLER_CHANNEL_GPFIFO_A:
+			ret = nvc0_fence_create(drm);
+			break;
+		default:
+			break;
+		}
+	}
+
 	if (ret) {
 		NV_ERROR(drm, "failed to initialise sync subsystem, %d\n", ret);
 		nouveau_accel_fini(drm);
 		return;
 	}
 
-	if (device->card_type >= NV_E0) {
-		ret = nouveau_channel_new(drm, &drm->client, NVDRM_DEVICE,
-					  NVDRM_CHAN + 1,
-					  NVE0_CHANNEL_IND_ENGINE_CE0 |
-					  NVE0_CHANNEL_IND_ENGINE_CE1, 0,
-					  &drm->cechan);
+	if (device->info.family >= NV_DEVICE_INFO_V0_KEPLER) {
+		ret = nouveau_channel_new(drm, &drm->device, NVDRM_CHAN + 1,
+					  KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_CE0|
+					  KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_CE1,
+					  0, &drm->cechan);
 		if (ret)
 			NV_ERROR(drm, "failed to create ce channel, %d\n", ret);
 
-		arg0 = NVE0_CHANNEL_IND_ENGINE_GR;
+		arg0 = KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_GR;
 		arg1 = 1;
 	} else
-	if (device->chipset >= 0xa3 &&
-	    device->chipset != 0xaa &&
-	    device->chipset != 0xac) {
-		ret = nouveau_channel_new(drm, &drm->client, NVDRM_DEVICE,
-					  NVDRM_CHAN + 1, NvDmaFB, NvDmaTT,
-					  &drm->cechan);
+	if (device->info.chipset >= 0xa3 &&
+	    device->info.chipset != 0xaa &&
+	    device->info.chipset != 0xac) {
+		ret = nouveau_channel_new(drm, &drm->device, NVDRM_CHAN + 1,
+					  NvDmaFB, NvDmaTT, &drm->cechan);
 		if (ret)
 			NV_ERROR(drm, "failed to create ce channel, %d\n", ret);
 
@@ -200,30 +219,30 @@ nouveau_accel_init(struct nouveau_drm *drm)
 		arg1 = NvDmaTT;
 	}
 
-	ret = nouveau_channel_new(drm, &drm->client, NVDRM_DEVICE, NVDRM_CHAN,
-				  arg0, arg1, &drm->channel);
+	ret = nouveau_channel_new(drm, &drm->device, NVDRM_CHAN, arg0, arg1,
+				 &drm->channel);
 	if (ret) {
 		NV_ERROR(drm, "failed to create kernel channel, %d\n", ret);
 		nouveau_accel_fini(drm);
 		return;
 	}
 
-	ret = nouveau_object_new(nv_object(drm), NVDRM_CHAN, NVDRM_NVSW,
-				 nouveau_abi16_swclass(drm), NULL, 0, &object);
+	ret = nvif_object_init(drm->channel->object, NULL, NVDRM_NVSW,
+			       nouveau_abi16_swclass(drm), NULL, 0, &drm->nvsw);
 	if (ret == 0) {
-		struct nouveau_software_chan *swch = (void *)object->parent;
+		struct nouveau_software_chan *swch;
 		ret = RING_SPACE(drm->channel, 2);
 		if (ret == 0) {
-			if (device->card_type < NV_C0) {
+			if (device->info.family < NV_DEVICE_INFO_V0_FERMI) {
 				BEGIN_NV04(drm->channel, NvSubSw, 0, 1);
 				OUT_RING  (drm->channel, NVDRM_NVSW);
 			} else
-			if (device->card_type < NV_E0) {
+			if (device->info.family < NV_DEVICE_INFO_V0_KEPLER) {
 				BEGIN_NVC0(drm->channel, FermiSw, 0, 1);
 				OUT_RING  (drm->channel, 0x001f0000);
 			}
 		}
-		swch = (void *)object->parent;
+		swch = (void *)nvkm_object(&drm->nvsw)->parent;
 		swch->flip = nouveau_flip_complete;
 		swch->flip_data = drm->channel;
 	}
@@ -234,24 +253,24 @@ nouveau_accel_init(struct nouveau_drm *drm)
 		return;
 	}
 
-	if (device->card_type < NV_C0) {
-		ret = nouveau_gpuobj_new(drm->device, NULL, 32, 0, 0,
-					&drm->notify);
+	if (device->info.family < NV_DEVICE_INFO_V0_FERMI) {
+		ret = nouveau_gpuobj_new(nvkm_object(&drm->device), NULL, 32,
+					 0, 0, &drm->notify);
 		if (ret) {
 			NV_ERROR(drm, "failed to allocate notifier, %d\n", ret);
 			nouveau_accel_fini(drm);
 			return;
 		}
 
-		ret = nouveau_object_new(nv_object(drm),
-					 drm->channel->handle, NvNotify0,
-					 0x003d, &(struct nv_dma_class) {
-						.flags = NV_DMA_TARGET_VRAM |
-							 NV_DMA_ACCESS_RDWR,
+		ret = nvif_object_init(drm->channel->object, NULL, NvNotify0,
+				       NV_DMA_IN_MEMORY,
+				       &(struct nv_dma_v0) {
+						.target = NV_DMA_V0_TARGET_VRAM,
+						.access = NV_DMA_V0_ACCESS_RDWR,
 						.start = drm->notify->addr,
 						.limit = drm->notify->addr + 31
-						}, sizeof(struct nv_dma_class),
-					 &object);
+				       }, sizeof(struct nv_dma_v0),
+				       &drm->ntfy);
 		if (ret) {
 			nouveau_accel_fini(drm);
 			return;
@@ -294,7 +313,8 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 #ifdef CONFIG_X86
 	boot = pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
 #endif
-	remove_conflicting_framebuffers(aper, "nouveaufb", boot);
+	if (nouveau_modeset != 2)
+		remove_conflicting_framebuffers(aper, "nouveaufb", boot);
 	kfree(aper);
 
 	ret = nouveau_device_create(pdev, NOUVEAU_BUS_PCI,
@@ -305,7 +325,7 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	ret = drm_get_pci_dev(pdev, pent, &driver);
+	ret = drm_get_pci_dev(pdev, pent, &driver_pci);
 	if (ret) {
 		nouveau_object_ref(NULL, (struct nouveau_object **)&device);
 		return ret;
@@ -348,7 +368,6 @@ static int
 nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 {
 	struct pci_dev *pdev = dev->pdev;
-	struct nouveau_device *device;
 	struct nouveau_drm *drm;
 	int ret;
 
@@ -359,7 +378,8 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 
 	dev->dev_private = drm;
 	drm->dev = dev;
-	nouveau_client(drm)->debug = nouveau_dbgopt(nouveau_debug, "DRM");
+	nvkm_client(&drm->client.base)->debug =
+		nouveau_dbgopt(nouveau_debug, "DRM");
 
 	INIT_LIST_HEAD(&drm->clients);
 	spin_lock_init(&drm->tile.lock);
@@ -370,33 +390,34 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 	 * (possibly) execute vbios init tables (see nouveau_agp.h)
 	 */
 	if (pdev && drm_pci_device_is_agp(dev) && dev->agp) {
+		const u64 enables = NV_DEVICE_V0_DISABLE_IDENTIFY |
+				    NV_DEVICE_V0_DISABLE_MMIO;
 		/* dummy device object, doesn't init anything, but allows
 		 * agp code access to registers
 		 */
-		ret = nouveau_object_new(nv_object(drm), NVDRM_CLIENT,
-					 NVDRM_DEVICE, 0x0080,
-					 &(struct nv_device_class) {
+		ret = nvif_device_init(&drm->client.base.base, NULL,
+				       NVDRM_DEVICE, NV_DEVICE,
+				       &(struct nv_device_v0) {
 						.device = ~0,
-						.disable =
-						 ~(NV_DEVICE_DISABLE_MMIO |
-						   NV_DEVICE_DISABLE_IDENTIFY),
+						.disable = ~enables,
 						.debug0 = ~0,
-					 }, sizeof(struct nv_device_class),
-					 &drm->device);
+				       }, sizeof(struct nv_device_v0),
+				       &drm->device);
 		if (ret)
 			goto fail_device;
 
 		nouveau_agp_reset(drm);
-		nouveau_object_del(nv_object(drm), NVDRM_CLIENT, NVDRM_DEVICE);
+		nvif_device_fini(&drm->device);
 	}
 
-	ret = nouveau_object_new(nv_object(drm), NVDRM_CLIENT, NVDRM_DEVICE,
-				 0x0080, &(struct nv_device_class) {
+	ret = nvif_device_init(&drm->client.base.base, NULL, NVDRM_DEVICE,
+			       NV_DEVICE,
+			       &(struct nv_device_v0) {
 					.device = ~0,
 					.disable = 0,
 					.debug0 = 0,
-				 }, sizeof(struct nv_device_class),
-				 &drm->device);
+			       }, sizeof(struct nv_device_v0),
+			       &drm->device);
 	if (ret)
 		goto fail_device;
 
@@ -406,18 +427,19 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 	 * nosnoop capability.  hopefully won't cause issues until a
 	 * better fix is found - assuming there is one...
 	 */
-	device = nv_device(drm->device);
-	if (nv_device(drm->device)->chipset == 0xc1)
-		nv_mask(device, 0x00088080, 0x00000800, 0x00000000);
+	if (drm->device.info.chipset == 0xc1)
+		nvif_mask(&drm->device, 0x00088080, 0x00000800, 0x00000000);
 
 	nouveau_vga_init(drm);
 	nouveau_agp_init(drm);
 
-	if (device->card_type >= NV_50) {
-		ret = nouveau_vm_new(nv_device(drm->device), 0, (1ULL << 40),
-				     0x1000, &drm->client.base.vm);
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_vm_new(nvkm_device(&drm->device), 0, (1ULL << 40),
+				     0x1000, &drm->client.vm);
 		if (ret)
 			goto fail_device;
+
+		nvkm_client(&drm->client.base)->vm = drm->client.vm;
 	}
 
 	ret = nouveau_ttm_init(drm);
@@ -463,6 +485,7 @@ fail_ttm:
 	nouveau_agp_fini(drm);
 	nouveau_vga_fini(drm);
 fail_device:
+	nvif_device_fini(&drm->device);
 	nouveau_cli_destroy(&drm->client);
 	return ret;
 }
@@ -488,26 +511,37 @@ nouveau_drm_unload(struct drm_device *dev)
 	nouveau_agp_fini(drm);
 	nouveau_vga_fini(drm);
 
+	nvif_device_fini(&drm->device);
 	if (drm->hdmi_device)
 		pci_dev_put(drm->hdmi_device);
 	nouveau_cli_destroy(&drm->client);
 	return 0;
 }
 
-static void
-nouveau_drm_remove(struct pci_dev *pdev)
+void
+nouveau_drm_device_remove(struct drm_device *dev)
 {
-	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_client *client;
 	struct nouveau_object *device;
 
 	dev->irq_enabled = false;
-	device = drm->client.base.device;
+	client = nvkm_client(&drm->client.base);
+	device = client->device;
 	drm_put_dev(dev);
 
 	nouveau_object_ref(NULL, &device);
 	nouveau_object_debug();
 }
+EXPORT_SYMBOL(nouveau_drm_device_remove);
+
+static void
+nouveau_drm_remove(struct pci_dev *pdev)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+
+	nouveau_drm_device_remove(dev);
+}
 
 static int
 nouveau_do_suspend(struct drm_device *dev, bool runtime)
@@ -516,9 +550,11 @@ nouveau_do_suspend(struct drm_device *dev, bool runtime)
 	struct nouveau_cli *cli;
 	int ret;
 
-	if (dev->mode_config.num_crtc && !runtime) {
+	if (dev->mode_config.num_crtc) {
+		NV_INFO(drm, "suspending console...\n");
+		nouveau_fbcon_set_suspend(dev, 1);
 		NV_INFO(drm, "suspending display...\n");
-		ret = nouveau_display_suspend(dev);
+		ret = nouveau_display_suspend(dev, runtime);
 		if (ret)
 			return ret;
 	}
@@ -548,13 +584,13 @@ nouveau_do_suspend(struct drm_device *dev, bool runtime)
 	}
 
 	list_for_each_entry(cli, &drm->clients, head) {
-		ret = nouveau_client_fini(&cli->base, true);
+		ret = nvif_client_suspend(&cli->base);
 		if (ret)
 			goto fail_client;
 	}
 
 	NV_INFO(drm, "suspending kernel object tree...\n");
-	ret = nouveau_client_fini(&drm->client.base, true);
+	ret = nvif_client_suspend(&drm->client.base);
 	if (ret)
 		goto fail_client;
 
@@ -563,7 +599,7 @@ nouveau_do_suspend(struct drm_device *dev, bool runtime)
 
 fail_client:
 	list_for_each_entry_continue_reverse(cli, &drm->clients, head) {
-		nouveau_client_init(&cli->base);
+		nvif_client_resume(&cli->base);
 	}
 
 	if (drm->fence && nouveau_fence(drm)->resume)
@@ -572,7 +608,7 @@ fail_client:
 fail_display:
 	if (dev->mode_config.num_crtc) {
 		NV_INFO(drm, "resuming display...\n");
-		nouveau_display_resume(dev);
+		nouveau_display_resume(dev, runtime);
 	}
 	return ret;
 }
@@ -587,21 +623,19 @@ int nouveau_pmops_suspend(struct device *dev)
 	    drm_dev->switch_power_state == DRM_SWITCH_POWER_DYNAMIC_OFF)
 		return 0;
 
-	if (drm_dev->mode_config.num_crtc)
-		nouveau_fbcon_set_suspend(drm_dev, 1);
-
 	ret = nouveau_do_suspend(drm_dev, false);
 	if (ret)
 		return ret;
 
 	pci_save_state(pdev);
 	pci_disable_device(pdev);
+	pci_ignore_hotplug(pdev);
 	pci_set_power_state(pdev, PCI_D3hot);
 	return 0;
 }
 
 static int
-nouveau_do_resume(struct drm_device *dev)
+nouveau_do_resume(struct drm_device *dev, bool runtime)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_cli *cli;
@@ -611,7 +645,7 @@ nouveau_do_resume(struct drm_device *dev)
 	nouveau_agp_reset(drm);
 
 	NV_INFO(drm, "resuming kernel object tree...\n");
-	nouveau_client_init(&drm->client.base);
+	nvif_client_resume(&drm->client.base);
 	nouveau_agp_init(drm);
 
 	NV_INFO(drm, "resuming client object trees...\n");
@@ -619,14 +653,16 @@ nouveau_do_resume(struct drm_device *dev)
 		nouveau_fence(drm)->resume(drm);
 
 	list_for_each_entry(cli, &drm->clients, head) {
-		nouveau_client_init(&cli->base);
+		nvif_client_resume(&cli->base);
 	}
 
 	nouveau_run_vbios_init(dev);
 
 	if (dev->mode_config.num_crtc) {
 		NV_INFO(drm, "resuming display...\n");
-		nouveau_display_repin(dev);
+		nouveau_display_resume(dev, runtime);
+		NV_INFO(drm, "resuming console...\n");
+		nouveau_fbcon_set_suspend(dev, 0);
 	}
 
 	return 0;
@@ -649,47 +685,21 @@ int nouveau_pmops_resume(struct device *dev)
 		return ret;
 	pci_set_master(pdev);
 
-	ret = nouveau_do_resume(drm_dev);
-	if (ret)
-		return ret;
-
-	if (drm_dev->mode_config.num_crtc) {
-		nouveau_display_resume(drm_dev);
-		nouveau_fbcon_set_suspend(drm_dev, 0);
-	}
-
-	return 0;
+	return nouveau_do_resume(drm_dev, false);
 }
 
 static int nouveau_pmops_freeze(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	int ret;
-
-	if (drm_dev->mode_config.num_crtc)
-		nouveau_fbcon_set_suspend(drm_dev, 1);
-
-	ret = nouveau_do_suspend(drm_dev, false);
-	return ret;
+	return nouveau_do_suspend(drm_dev, false);
 }
 
 static int nouveau_pmops_thaw(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	int ret;
-
-	ret = nouveau_do_resume(drm_dev);
-	if (ret)
-		return ret;
-
-	if (drm_dev->mode_config.num_crtc) {
-		nouveau_display_resume(drm_dev);
-		nouveau_fbcon_set_suspend(drm_dev, 0);
-	}
-
-	return 0;
+	return nouveau_do_resume(drm_dev, false);
 }
 
 
@@ -715,13 +725,17 @@ nouveau_drm_open(struct drm_device *dev, struct drm_file *fpriv)
 	if (ret)
 		goto out_suspend;
 
-	if (nv_device(drm->device)->card_type >= NV_50) {
-		ret = nouveau_vm_new(nv_device(drm->device), 0, (1ULL << 40),
-				     0x1000, &cli->base.vm);
+	cli->base.super = false;
+
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_vm_new(nvkm_device(&drm->device), 0, (1ULL << 40),
+				     0x1000, &cli->vm);
 		if (ret) {
 			nouveau_cli_destroy(cli);
 			goto out_suspend;
 		}
+
+		nvkm_client(&cli->base)->vm = cli->vm;
 	}
 
 	fpriv->driver_priv = cli;
@@ -779,24 +793,31 @@ nouveau_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(NOUVEAU_GEM_INFO, nouveau_gem_ioctl_info, DRM_UNLOCKED|DRM_AUTH|DRM_RENDER_ALLOW),
 };
 
-long nouveau_drm_ioctl(struct file *filp,
-		       unsigned int cmd, unsigned long arg)
+long
+nouveau_drm_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 {
-	struct drm_file *file_priv = filp->private_data;
-	struct drm_device *dev;
+	struct drm_file *filp = file->private_data;
+	struct drm_device *dev = filp->minor->dev;
 	long ret;
-	dev = file_priv->minor->dev;
 
 	ret = pm_runtime_get_sync(dev->dev);
 	if (ret < 0 && ret != -EACCES)
 		return ret;
 
-	ret = drm_ioctl(filp, cmd, arg);
+	switch (_IOC_NR(cmd) - DRM_COMMAND_BASE) {
+	case DRM_NOUVEAU_NVIF:
+		ret = usif_ioctl(filp, (void __user *)arg, _IOC_SIZE(cmd));
+		break;
+	default:
+		ret = drm_ioctl(file, cmd, arg);
+		break;
+	}
 
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
 	return ret;
 }
+
 static const struct file_operations
 nouveau_driver_fops = {
 	.owner = THIS_MODULE,
@@ -813,7 +834,7 @@ nouveau_driver_fops = {
 };
 
 static struct drm_driver
-driver = {
+driver_stub = {
 	.driver_features =
 		DRIVER_USE_AGP |
 		DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME | DRIVER_RENDER,
@@ -921,7 +942,7 @@ static int nouveau_pmops_runtime_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	struct nouveau_device *device = nouveau_dev(drm_dev);
+	struct nvif_device *device = &nouveau_drm(drm_dev)->device;
 	int ret;
 
 	if (nouveau_runtime_pm == 0)
@@ -934,10 +955,10 @@ static int nouveau_pmops_runtime_resume(struct device *dev)
 		return ret;
 	pci_set_master(pdev);
 
-	ret = nouveau_do_resume(drm_dev);
+	ret = nouveau_do_resume(drm_dev, true);
 	drm_kms_helper_poll_enable(drm_dev);
 	/* do magic */
-	nv_mask(device, 0x88488, (1 << 25), (1 << 25));
+	nvif_mask(device, 0x88488, (1 << 25), (1 << 25));
 	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON);
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	nv_debug_level(NORMAL);
@@ -984,6 +1005,23 @@ static int nouveau_pmops_runtime_idle(struct device *dev)
 	return 1;
 }
 
+static void nouveau_display_options(void)
+{
+	DRM_DEBUG_DRIVER("Loading Nouveau with parameters:\n");
+
+	DRM_DEBUG_DRIVER("... tv_disable   : %d\n", nouveau_tv_disable);
+	DRM_DEBUG_DRIVER("... ignorelid    : %d\n", nouveau_ignorelid);
+	DRM_DEBUG_DRIVER("... duallink     : %d\n", nouveau_duallink);
+	DRM_DEBUG_DRIVER("... nofbaccel    : %d\n", nouveau_nofbaccel);
+	DRM_DEBUG_DRIVER("... config       : %s\n", nouveau_config);
+	DRM_DEBUG_DRIVER("... debug        : %s\n", nouveau_debug);
+	DRM_DEBUG_DRIVER("... noaccel      : %d\n", nouveau_noaccel);
+	DRM_DEBUG_DRIVER("... modeset      : %d\n", nouveau_modeset);
+	DRM_DEBUG_DRIVER("... runpm        : %d\n", nouveau_runtime_pm);
+	DRM_DEBUG_DRIVER("... vram_pushbuf : %d\n", nouveau_vram_pushbuf);
+	DRM_DEBUG_DRIVER("... pstate       : %d\n", nouveau_pstate);
+}
+
 static const struct dev_pm_ops nouveau_pm_ops = {
 	.suspend = nouveau_pmops_suspend,
 	.resume = nouveau_pmops_resume,
@@ -1005,28 +1043,52 @@ nouveau_drm_pci_driver = {
 	.driver.pm = &nouveau_pm_ops,
 };
 
-int nouveau_drm_platform_probe(struct platform_device *pdev)
+struct drm_device *
+nouveau_platform_device_create_(struct platform_device *pdev, int size,
+				void **pobject)
 {
-	struct nouveau_device *device;
-	int ret;
+	struct drm_device *drm;
+	int err;
 
-	ret = nouveau_device_create(pdev, NOUVEAU_BUS_PLATFORM,
+	err = nouveau_device_create_(pdev, NOUVEAU_BUS_PLATFORM,
 				    nouveau_platform_name(pdev),
 				    dev_name(&pdev->dev), nouveau_config,
-				    nouveau_debug, &device);
-
-	ret = drm_platform_init(&driver, pdev);
-	if (ret) {
-		nouveau_object_ref(NULL, (struct nouveau_object **)&device);
-		return ret;
+				    nouveau_debug, size, pobject);
+	if (err)
+		return ERR_PTR(err);
+
+	drm = drm_dev_alloc(&driver_platform, &pdev->dev);
+	if (!drm) {
+		err = -ENOMEM;
+		goto err_free;
 	}
 
-	return ret;
+	err = drm_dev_set_unique(drm, "%s", dev_name(&pdev->dev));
+	if (err < 0)
+		goto err_free;
+
+	drm->platformdev = pdev;
+	platform_set_drvdata(pdev, drm);
+
+	return drm;
+
+err_free:
+	nouveau_object_ref(NULL, (struct nouveau_object **)pobject);
+
+	return ERR_PTR(err);
 }
+EXPORT_SYMBOL(nouveau_platform_device_create_);
 
 static int __init
 nouveau_drm_init(void)
 {
+	driver_pci = driver_stub;
+	driver_pci.set_busid = drm_pci_set_busid;
+	driver_platform = driver_stub;
+	driver_platform.set_busid = drm_platform_set_busid;
+
+	nouveau_display_options();
+
 	if (nouveau_modeset == -1) {
 #ifdef CONFIG_VGA_CONSOLE
 		if (vgacon_text_force())
@@ -1038,7 +1100,7 @@ nouveau_drm_init(void)
 		return 0;
 
 	nouveau_register_dsm_handler();
-	return drm_pci_init(&driver, &nouveau_drm_pci_driver);
+	return drm_pci_init(&driver_pci, &nouveau_drm_pci_driver);
 }
 
 static void __exit
@@ -1047,7 +1109,7 @@ nouveau_drm_exit(void)
 	if (!nouveau_modeset)
 		return;
 
-	drm_pci_exit(&driver, &nouveau_drm_pci_driver);
+	drm_pci_exit(&driver_pci, &nouveau_drm_pci_driver);
 	nouveau_unregister_dsm_handler();
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index 7efbafaf7c1d..8ae36f265fb8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -9,7 +9,7 @@
 #define DRIVER_DATE		"20120801"
 
 #define DRIVER_MAJOR		1
-#define DRIVER_MINOR		1
+#define DRIVER_MINOR		2
 #define DRIVER_PATCHLEVEL	1
 
 /*
@@ -21,15 +21,19 @@
  *        to control registers on the MPs to enable performance counters,
  *        and to control the warp error enable mask (OpenGL requires out of
  *        bounds access to local memory to be silently ignored / return 0).
+ * 1.1.2:
+ *      - fixes multiple bugs in flip completion events and timestamping
+ * 1.2.0:
+ * 	- object api exposed to userspace
+ * 	- fermi,kepler,maxwell zbc
+ * 1.2.1:
+ *      - allow concurrent access to bo's mapped read/write.
  */
 
-#include <core/client.h>
-#include <core/event.h>
-
-#include <subdev/vm.h>
+#include <nvif/client.h>
+#include <nvif/device.h>
 
 #include <drmP.h>
-#include <drm/nouveau_drm.h>
 
 #include <drm/ttm/ttm_bo_api.h>
 #include <drm/ttm/ttm_bo_driver.h>
@@ -38,7 +42,10 @@
 #include <drm/ttm/ttm_module.h>
 #include <drm/ttm/ttm_page_alloc.h>
 
+#include "uapi/drm/nouveau_drm.h"
+
 struct nouveau_channel;
+struct platform_device;
 
 #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT)
 
@@ -50,6 +57,17 @@ struct nouveau_drm_tile {
 	bool used;
 };
 
+enum nouveau_drm_object_route {
+	NVDRM_OBJECT_NVIF = 0,
+	NVDRM_OBJECT_USIF,
+	NVDRM_OBJECT_ABI16,
+};
+
+enum nouveau_drm_notify_route {
+	NVDRM_NOTIFY_NVIF = 0,
+	NVDRM_NOTIFY_USIF
+};
+
 enum nouveau_drm_handle {
 	NVDRM_CLIENT  = 0xffffffff,
 	NVDRM_DEVICE  = 0xdddddddd,
@@ -61,10 +79,13 @@ enum nouveau_drm_handle {
 };
 
 struct nouveau_cli {
-	struct nouveau_client base;
+	struct nvif_client base;
+	struct nouveau_vm *vm; /*XXX*/
 	struct list_head head;
 	struct mutex mutex;
 	void *abi16;
+	struct list_head objects;
+	struct list_head notifys;
 };
 
 static inline struct nouveau_cli *
@@ -73,13 +94,16 @@ nouveau_cli(struct drm_file *fpriv)
 	return fpriv ? fpriv->driver_priv : NULL;
 }
 
+#include <nvif/object.h>
+#include <nvif/device.h>
+
 extern int nouveau_runtime_pm;
 
 struct nouveau_drm {
 	struct nouveau_cli client;
 	struct drm_device *dev;
 
-	struct nouveau_object *device;
+	struct nvif_device device;
 	struct list_head clients;
 
 	struct {
@@ -102,6 +126,7 @@ struct nouveau_drm {
 			    struct ttm_buffer_object *,
 			    struct ttm_mem_reg *, struct ttm_mem_reg *);
 		struct nouveau_channel *chan;
+		struct nvif_object copy;
 		int mtrr;
 	} ttm;
 
@@ -119,6 +144,8 @@ struct nouveau_drm {
 	struct nouveau_channel *channel;
 	struct nouveau_gpuobj *notify;
 	struct nouveau_fbdev *fbcon;
+	struct nvif_object nvsw;
+	struct nvif_object ntfy;
 
 	/* nv10-nv40 tiling regions */
 	struct {
@@ -148,20 +175,25 @@ nouveau_drm(struct drm_device *dev)
 	return dev->dev_private;
 }
 
-static inline struct nouveau_device *
-nouveau_dev(struct drm_device *dev)
-{
-	return nv_device(nouveau_drm(dev)->device);
-}
-
 int nouveau_pmops_suspend(struct device *);
 int nouveau_pmops_resume(struct device *);
 
-#define NV_FATAL(cli, fmt, args...) nv_fatal((cli), fmt, ##args)
-#define NV_ERROR(cli, fmt, args...) nv_error((cli), fmt, ##args)
-#define NV_WARN(cli, fmt, args...) nv_warn((cli), fmt, ##args)
-#define NV_INFO(cli, fmt, args...) nv_info((cli), fmt, ##args)
-#define NV_DEBUG(cli, fmt, args...) nv_debug((cli), fmt, ##args)
+#define nouveau_platform_device_create(p, u)                                   \
+	nouveau_platform_device_create_(p, sizeof(**u), (void **)u)
+struct drm_device *
+nouveau_platform_device_create_(struct platform_device *pdev,
+				int size, void **pobject);
+void nouveau_drm_device_remove(struct drm_device *dev);
+
+#define NV_PRINTK(l,c,f,a...) do {                                             \
+	struct nouveau_cli *_cli = (c);                                        \
+	nv_##l(_cli->base.base.priv, f, ##a);                                  \
+} while(0)
+#define NV_FATAL(drm,f,a...) NV_PRINTK(fatal, &(drm)->client, f, ##a)
+#define NV_ERROR(drm,f,a...) NV_PRINTK(error, &(drm)->client, f, ##a)
+#define NV_WARN(drm,f,a...) NV_PRINTK(warn, &(drm)->client, f, ##a)
+#define NV_INFO(drm,f,a...) NV_PRINTK(info, &(drm)->client, f, ##a)
+#define NV_DEBUG(drm,f,a...) NV_PRINTK(debug, &(drm)->client, f, ##a)
 
 extern int nouveau_modeset;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 191665ee7f52..593ef8a2a069 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -51,13 +51,8 @@
 
 #include "nouveau_crtc.h"
 
-#include <core/client.h>
-#include <core/device.h>
-
-#include <subdev/fb.h>
-
 MODULE_PARM_DESC(nofbaccel, "Disable fbcon acceleration");
-static int nouveau_nofbaccel = 0;
+int nouveau_nofbaccel = 0;
 module_param_named(nofbaccel, nouveau_nofbaccel, int, 0400);
 
 static void
@@ -65,7 +60,7 @@ nouveau_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect)
 {
 	struct nouveau_fbdev *fbcon = info->par;
 	struct nouveau_drm *drm = nouveau_drm(fbcon->dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int ret;
 
 	if (info->state != FBINFO_STATE_RUNNING)
@@ -74,10 +69,10 @@ nouveau_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect)
 	ret = -ENODEV;
 	if (!in_interrupt() && !(info->flags & FBINFO_HWACCEL_DISABLED) &&
 	    mutex_trylock(&drm->client.mutex)) {
-		if (device->card_type < NV_50)
+		if (device->info.family < NV_DEVICE_INFO_V0_TESLA)
 			ret = nv04_fbcon_fillrect(info, rect);
 		else
-		if (device->card_type < NV_C0)
+		if (device->info.family < NV_DEVICE_INFO_V0_FERMI)
 			ret = nv50_fbcon_fillrect(info, rect);
 		else
 			ret = nvc0_fbcon_fillrect(info, rect);
@@ -97,7 +92,7 @@ nouveau_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *image)
 {
 	struct nouveau_fbdev *fbcon = info->par;
 	struct nouveau_drm *drm = nouveau_drm(fbcon->dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int ret;
 
 	if (info->state != FBINFO_STATE_RUNNING)
@@ -106,10 +101,10 @@ nouveau_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *image)
 	ret = -ENODEV;
 	if (!in_interrupt() && !(info->flags & FBINFO_HWACCEL_DISABLED) &&
 	    mutex_trylock(&drm->client.mutex)) {
-		if (device->card_type < NV_50)
+		if (device->info.family < NV_DEVICE_INFO_V0_TESLA)
 			ret = nv04_fbcon_copyarea(info, image);
 		else
-		if (device->card_type < NV_C0)
+		if (device->info.family < NV_DEVICE_INFO_V0_FERMI)
 			ret = nv50_fbcon_copyarea(info, image);
 		else
 			ret = nvc0_fbcon_copyarea(info, image);
@@ -129,7 +124,7 @@ nouveau_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 {
 	struct nouveau_fbdev *fbcon = info->par;
 	struct nouveau_drm *drm = nouveau_drm(fbcon->dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	int ret;
 
 	if (info->state != FBINFO_STATE_RUNNING)
@@ -138,10 +133,10 @@ nouveau_fbcon_imageblit(struct fb_info *info, const struct fb_image *image)
 	ret = -ENODEV;
 	if (!in_interrupt() && !(info->flags & FBINFO_HWACCEL_DISABLED) &&
 	    mutex_trylock(&drm->client.mutex)) {
-		if (device->card_type < NV_50)
+		if (device->info.family < NV_DEVICE_INFO_V0_TESLA)
 			ret = nv04_fbcon_imageblit(info, image);
 		else
-		if (device->card_type < NV_C0)
+		if (device->info.family < NV_DEVICE_INFO_V0_FERMI)
 			ret = nv50_fbcon_imageblit(info, image);
 		else
 			ret = nvc0_fbcon_imageblit(info, image);
@@ -212,6 +207,65 @@ static struct fb_ops nouveau_fbcon_sw_ops = {
 	.fb_debug_leave = drm_fb_helper_debug_leave,
 };
 
+void
+nouveau_fbcon_accel_save_disable(struct drm_device *dev)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	if (drm->fbcon) {
+		drm->fbcon->saved_flags = drm->fbcon->helper.fbdev->flags;
+		drm->fbcon->helper.fbdev->flags |= FBINFO_HWACCEL_DISABLED;
+	}
+}
+
+void
+nouveau_fbcon_accel_restore(struct drm_device *dev)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	if (drm->fbcon) {
+		drm->fbcon->helper.fbdev->flags = drm->fbcon->saved_flags;
+	}
+}
+
+static void
+nouveau_fbcon_accel_fini(struct drm_device *dev)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_fbdev *fbcon = drm->fbcon;
+	if (fbcon && drm->channel) {
+		console_lock();
+		fbcon->helper.fbdev->flags |= FBINFO_HWACCEL_DISABLED;
+		console_unlock();
+		nouveau_channel_idle(drm->channel);
+		nvif_object_fini(&fbcon->twod);
+		nvif_object_fini(&fbcon->blit);
+		nvif_object_fini(&fbcon->gdi);
+		nvif_object_fini(&fbcon->patt);
+		nvif_object_fini(&fbcon->rop);
+		nvif_object_fini(&fbcon->clip);
+		nvif_object_fini(&fbcon->surf2d);
+	}
+}
+
+static void
+nouveau_fbcon_accel_init(struct drm_device *dev)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nouveau_fbdev *fbcon = drm->fbcon;
+	struct fb_info *info = fbcon->helper.fbdev;
+	int ret;
+
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA)
+		ret = nv04_fbcon_accel_init(info);
+	else
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_FERMI)
+		ret = nv50_fbcon_accel_init(info);
+	else
+		ret = nvc0_fbcon_accel_init(info);
+
+	if (ret == 0)
+		info->fbops = &nouveau_fbcon_ops;
+}
+
 static void nouveau_fbcon_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 				    u16 blue, int regno)
 {
@@ -254,10 +308,11 @@ static int
 nouveau_fbcon_create(struct drm_fb_helper *helper,
 		     struct drm_fb_helper_surface_size *sizes)
 {
-	struct nouveau_fbdev *fbcon = (struct nouveau_fbdev *)helper;
+	struct nouveau_fbdev *fbcon =
+		container_of(helper, struct nouveau_fbdev, helper);
 	struct drm_device *dev = fbcon->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
 	struct nouveau_framebuffer *nouveau_fb;
@@ -299,8 +354,8 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	}
 
 	chan = nouveau_nofbaccel ? NULL : drm->channel;
-	if (chan && device->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(nvbo, nv_client(chan->cli)->vm,
+	if (chan && device->info.family >= NV_DEVICE_INFO_V0_TESLA) {
+		ret = nouveau_bo_vma_add(nvbo, drm->client.vm,
 					&fbcon->nouveau_fb.vma);
 		if (ret) {
 			NV_ERROR(drm, "failed to map fb into chan: %d\n", ret);
@@ -357,20 +412,8 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 
 	mutex_unlock(&dev->struct_mutex);
 
-	if (chan) {
-		ret = -ENODEV;
-		if (device->card_type < NV_50)
-			ret = nv04_fbcon_accel_init(info);
-		else
-		if (device->card_type < NV_C0)
-			ret = nv50_fbcon_accel_init(info);
-		else
-			ret = nvc0_fbcon_accel_init(info);
-
-		if (ret == 0)
-			info->fbops = &nouveau_fbcon_ops;
-	}
-
+	if (chan)
+		nouveau_fbcon_accel_init(dev);
 	nouveau_fbcon_zfill(dev, fbcon);
 
 	/* To allow resizeing without swapping buffers */
@@ -438,18 +481,27 @@ void nouveau_fbcon_gpu_lockup(struct fb_info *info)
 	info->flags |= FBINFO_HWACCEL_DISABLED;
 }
 
-static struct drm_fb_helper_funcs nouveau_fbcon_helper_funcs = {
+static const struct drm_fb_helper_funcs nouveau_fbcon_helper_funcs = {
 	.gamma_set = nouveau_fbcon_gamma_set,
 	.gamma_get = nouveau_fbcon_gamma_get,
 	.fb_probe = nouveau_fbcon_create,
 };
 
+static void
+nouveau_fbcon_set_suspend_work(struct work_struct *work)
+{
+	struct nouveau_fbdev *fbcon = container_of(work, typeof(*fbcon), work);
+	console_lock();
+	nouveau_fbcon_accel_restore(fbcon->dev);
+	nouveau_fbcon_zfill(fbcon->dev, fbcon);
+	fb_set_suspend(fbcon->helper.fbdev, FBINFO_STATE_RUNNING);
+	console_unlock();
+}
 
 int
 nouveau_fbcon_init(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
 	struct nouveau_fbdev *fbcon;
 	int preferred_bpp;
 	int ret;
@@ -462,9 +514,11 @@ nouveau_fbcon_init(struct drm_device *dev)
 	if (!fbcon)
 		return -ENOMEM;
 
+	INIT_WORK(&fbcon->work, nouveau_fbcon_set_suspend_work);
 	fbcon->dev = dev;
 	drm->fbcon = fbcon;
-	fbcon->helper.funcs = &nouveau_fbcon_helper_funcs;
+
+	drm_fb_helper_prepare(dev, &fbcon->helper, &nouveau_fbcon_helper_funcs);
 
 	ret = drm_fb_helper_init(dev, &fbcon->helper,
 				 dev->mode_config.num_crtc, 4);
@@ -475,10 +529,10 @@ nouveau_fbcon_init(struct drm_device *dev)
 
 	drm_fb_helper_single_add_all_connectors(&fbcon->helper);
 
-	if (pfb->ram->size <= 32 * 1024 * 1024)
+	if (drm->device.info.ram_size <= 32 * 1024 * 1024)
 		preferred_bpp = 8;
 	else
-	if (pfb->ram->size <= 64 * 1024 * 1024)
+	if (drm->device.info.ram_size <= 64 * 1024 * 1024)
 		preferred_bpp = 16;
 	else
 		preferred_bpp = 32;
@@ -498,43 +552,25 @@ nouveau_fbcon_fini(struct drm_device *dev)
 	if (!drm->fbcon)
 		return;
 
+	nouveau_fbcon_accel_fini(dev);
 	nouveau_fbcon_destroy(dev, drm->fbcon);
 	kfree(drm->fbcon);
 	drm->fbcon = NULL;
 }
 
 void
-nouveau_fbcon_save_disable_accel(struct drm_device *dev)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	if (drm->fbcon) {
-		drm->fbcon->saved_flags = drm->fbcon->helper.fbdev->flags;
-		drm->fbcon->helper.fbdev->flags |= FBINFO_HWACCEL_DISABLED;
-	}
-}
-
-void
-nouveau_fbcon_restore_accel(struct drm_device *dev)
-{
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	if (drm->fbcon) {
-		drm->fbcon->helper.fbdev->flags = drm->fbcon->saved_flags;
-	}
-}
-
-void
 nouveau_fbcon_set_suspend(struct drm_device *dev, int state)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	if (drm->fbcon) {
+		if (state == FBINFO_STATE_RUNNING) {
+			schedule_work(&drm->fbcon->work);
+			return;
+		}
+		flush_work(&drm->fbcon->work);
 		console_lock();
-		if (state == 1)
-			nouveau_fbcon_save_disable_accel(dev);
 		fb_set_suspend(drm->fbcon->helper.fbdev, state);
-		if (state == 0) {
-			nouveau_fbcon_restore_accel(dev);
-			nouveau_fbcon_zfill(dev, drm->fbcon);
-		}
+		nouveau_fbcon_accel_save_disable(dev);
 		console_unlock();
 	}
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.h b/drivers/gpu/drm/nouveau/nouveau_fbcon.h
index fcff797d2084..6208e70e4a1c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.h
@@ -36,7 +36,15 @@ struct nouveau_fbdev {
 	struct nouveau_framebuffer nouveau_fb;
 	struct list_head fbdev_list;
 	struct drm_device *dev;
+	struct work_struct work;
 	unsigned int saved_flags;
+	struct nvif_object surf2d;
+	struct nvif_object clip;
+	struct nvif_object rop;
+	struct nvif_object patt;
+	struct nvif_object gdi;
+	struct nvif_object blit;
+	struct nvif_object twod;
 };
 
 void nouveau_fbcon_restore(void);
@@ -61,9 +69,12 @@ void nouveau_fbcon_gpu_lockup(struct fb_info *info);
 int nouveau_fbcon_init(struct drm_device *dev);
 void nouveau_fbcon_fini(struct drm_device *dev);
 void nouveau_fbcon_set_suspend(struct drm_device *dev, int state);
-void nouveau_fbcon_save_disable_accel(struct drm_device *dev);
-void nouveau_fbcon_restore_accel(struct drm_device *dev);
+void nouveau_fbcon_accel_save_disable(struct drm_device *dev);
+void nouveau_fbcon_accel_restore(struct drm_device *dev);
 
 void nouveau_fbcon_output_poll_changed(struct drm_device *dev);
+
+extern int nouveau_nofbaccel;
+
 #endif /* __NV50_FBCON_H__ */
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index ab5ea3b0d666..515cd9aebb99 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -28,130 +28,243 @@
 
 #include <linux/ktime.h>
 #include <linux/hrtimer.h>
+#include <trace/events/fence.h>
+
+#include <nvif/notify.h>
+#include <nvif/event.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
 
-#include <engine/fifo.h>
+static const struct fence_ops nouveau_fence_ops_uevent;
+static const struct fence_ops nouveau_fence_ops_legacy;
 
-struct fence_work {
-	struct work_struct base;
-	struct list_head head;
-	void (*func)(void *);
-	void *data;
-};
+static inline struct nouveau_fence *
+from_fence(struct fence *fence)
+{
+	return container_of(fence, struct nouveau_fence, base);
+}
+
+static inline struct nouveau_fence_chan *
+nouveau_fctx(struct nouveau_fence *fence)
+{
+	return container_of(fence->base.lock, struct nouveau_fence_chan, lock);
+}
 
 static void
 nouveau_fence_signal(struct nouveau_fence *fence)
 {
-	struct fence_work *work, *temp;
+	fence_signal_locked(&fence->base);
+	list_del(&fence->head);
+
+	if (test_bit(FENCE_FLAG_USER_BITS, &fence->base.flags)) {
+		struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
 
-	list_for_each_entry_safe(work, temp, &fence->work, head) {
-		schedule_work(&work->base);
-		list_del(&work->head);
+		if (!--fctx->notify_ref)
+			nvif_notify_put(&fctx->notify);
 	}
 
-	fence->channel = NULL;
-	list_del(&fence->head);
+	fence_put(&fence->base);
+}
+
+static struct nouveau_fence *
+nouveau_local_fence(struct fence *fence, struct nouveau_drm *drm) {
+	struct nouveau_fence_priv *priv = (void*)drm->fence;
+
+	if (fence->ops != &nouveau_fence_ops_legacy &&
+	    fence->ops != &nouveau_fence_ops_uevent)
+		return NULL;
+
+	if (fence->context < priv->context_base ||
+	    fence->context >= priv->context_base + priv->contexts)
+		return NULL;
+
+	return from_fence(fence);
 }
 
 void
 nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 {
-	struct nouveau_fence *fence, *fnext;
-	spin_lock(&fctx->lock);
-	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
+	struct nouveau_fence *fence;
+
+	nvif_notify_fini(&fctx->notify);
+
+	spin_lock_irq(&fctx->lock);
+	while (!list_empty(&fctx->pending)) {
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+
+		nouveau_fence_signal(fence);
+		fence->channel = NULL;
+	}
+	spin_unlock_irq(&fctx->lock);
+}
+
+static void
+nouveau_fence_context_put(struct kref *fence_ref)
+{
+	kfree(container_of(fence_ref, struct nouveau_fence_chan, fence_ref));
+}
+
+void
+nouveau_fence_context_free(struct nouveau_fence_chan *fctx)
+{
+	kref_put(&fctx->fence_ref, nouveau_fence_context_put);
+}
+
+static void
+nouveau_fence_update(struct nouveau_channel *chan, struct nouveau_fence_chan *fctx)
+{
+	struct nouveau_fence *fence;
+
+	u32 seq = fctx->read(chan);
+
+	while (!list_empty(&fctx->pending)) {
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+
+		if ((int)(seq - fence->base.seqno) < 0)
+			return;
+
 		nouveau_fence_signal(fence);
 	}
-	spin_unlock(&fctx->lock);
+}
+
+static int
+nouveau_fence_wait_uevent_handler(struct nvif_notify *notify)
+{
+	struct nouveau_fence_chan *fctx =
+		container_of(notify, typeof(*fctx), notify);
+	unsigned long flags;
+
+	spin_lock_irqsave(&fctx->lock, flags);
+	if (!list_empty(&fctx->pending)) {
+		struct nouveau_fence *fence;
+
+		fence = list_entry(fctx->pending.next, typeof(*fence), head);
+		nouveau_fence_update(fence->channel, fctx);
+	}
+	spin_unlock_irqrestore(&fctx->lock, flags);
+
+	/* Always return keep here. NVIF refcount is handled with nouveau_fence_update */
+	return NVIF_NOTIFY_KEEP;
 }
 
 void
-nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
+nouveau_fence_context_new(struct nouveau_channel *chan, struct nouveau_fence_chan *fctx)
 {
+	struct nouveau_fence_priv *priv = (void*)chan->drm->fence;
+	struct nouveau_cli *cli = (void *)nvif_client(chan->object);
+	int ret;
+
 	INIT_LIST_HEAD(&fctx->flip);
 	INIT_LIST_HEAD(&fctx->pending);
 	spin_lock_init(&fctx->lock);
+	fctx->context = priv->context_base + chan->chid;
+
+	if (chan == chan->drm->cechan)
+		strcpy(fctx->name, "copy engine channel");
+	else if (chan == chan->drm->channel)
+		strcpy(fctx->name, "generic kernel channel");
+	else
+		strcpy(fctx->name, nvkm_client(&cli->base)->name);
+
+	kref_init(&fctx->fence_ref);
+	if (!priv->uevent)
+		return;
+
+	ret = nvif_notify_init(chan->object, NULL,
+			 nouveau_fence_wait_uevent_handler, false,
+			 G82_CHANNEL_DMA_V0_NTFY_UEVENT,
+			 &(struct nvif_notify_uevent_req) { },
+			 sizeof(struct nvif_notify_uevent_req),
+			 sizeof(struct nvif_notify_uevent_rep),
+			 &fctx->notify);
+
+	WARN_ON(ret);
 }
 
+struct nouveau_fence_work {
+	struct work_struct work;
+	struct fence_cb cb;
+	void (*func)(void *);
+	void *data;
+};
+
 static void
 nouveau_fence_work_handler(struct work_struct *kwork)
 {
-	struct fence_work *work = container_of(kwork, typeof(*work), base);
+	struct nouveau_fence_work *work = container_of(kwork, typeof(*work), work);
 	work->func(work->data);
 	kfree(work);
 }
 
+static void nouveau_fence_work_cb(struct fence *fence, struct fence_cb *cb)
+{
+	struct nouveau_fence_work *work = container_of(cb, typeof(*work), cb);
+
+	schedule_work(&work->work);
+}
+
 void
-nouveau_fence_work(struct nouveau_fence *fence,
+nouveau_fence_work(struct fence *fence,
 		   void (*func)(void *), void *data)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fence_chan *fctx;
-	struct fence_work *work = NULL;
+	struct nouveau_fence_work *work;
 
-	if (nouveau_fence_done(fence)) {
-		func(data);
-		return;
-	}
+	if (fence_is_signaled(fence))
+		goto err;
 
-	fctx = chan->fence;
 	work = kmalloc(sizeof(*work), GFP_KERNEL);
 	if (!work) {
-		WARN_ON(nouveau_fence_wait(fence, false, false));
-		func(data);
-		return;
-	}
-
-	spin_lock(&fctx->lock);
-	if (!fence->channel) {
-		spin_unlock(&fctx->lock);
-		kfree(work);
-		func(data);
-		return;
+		/*
+		 * this might not be a nouveau fence any more,
+		 * so force a lazy wait here
+		 */
+		WARN_ON(nouveau_fence_wait((struct nouveau_fence *)fence,
+					   true, false));
+		goto err;
 	}
 
-	INIT_WORK(&work->base, nouveau_fence_work_handler);
+	INIT_WORK(&work->work, nouveau_fence_work_handler);
 	work->func = func;
 	work->data = data;
-	list_add(&work->head, &fence->work);
-	spin_unlock(&fctx->lock);
-}
-
-static void
-nouveau_fence_update(struct nouveau_channel *chan)
-{
-	struct nouveau_fence_chan *fctx = chan->fence;
-	struct nouveau_fence *fence, *fnext;
 
-	spin_lock(&fctx->lock);
-	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		if (fctx->read(chan) < fence->sequence)
-			break;
+	if (fence_add_callback(fence, &work->cb, nouveau_fence_work_cb) < 0)
+		goto err_free;
+	return;
 
-		nouveau_fence_signal(fence);
-		nouveau_fence_unref(&fence);
-	}
-	spin_unlock(&fctx->lock);
+err_free:
+	kfree(work);
+err:
+	func(data);
 }
 
 int
 nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
+	struct nouveau_fence_priv *priv = (void*)chan->drm->fence;
 	int ret;
 
 	fence->channel  = chan;
 	fence->timeout  = jiffies + (15 * HZ);
-	fence->sequence = ++fctx->sequence;
 
+	if (priv->uevent)
+		fence_init(&fence->base, &nouveau_fence_ops_uevent,
+			   &fctx->lock, fctx->context, ++fctx->sequence);
+	else
+		fence_init(&fence->base, &nouveau_fence_ops_legacy,
+			   &fctx->lock, fctx->context, ++fctx->sequence);
+	kref_get(&fctx->fence_ref);
+
+	trace_fence_emit(&fence->base);
 	ret = fctx->emit(fence);
 	if (!ret) {
-		kref_get(&fence->kref);
-		spin_lock(&fctx->lock);
+		fence_get(&fence->base);
+		spin_lock_irq(&fctx->lock);
+		nouveau_fence_update(chan, fctx);
 		list_add_tail(&fence->head, &fctx->pending);
-		spin_unlock(&fctx->lock);
+		spin_unlock_irq(&fctx->lock);
 	}
 
 	return ret;
@@ -160,104 +273,70 @@ nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 bool
 nouveau_fence_done(struct nouveau_fence *fence)
 {
-	if (fence->channel)
-		nouveau_fence_update(fence->channel);
-	return !fence->channel;
-}
+	if (fence->base.ops == &nouveau_fence_ops_legacy ||
+	    fence->base.ops == &nouveau_fence_ops_uevent) {
+		struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+		unsigned long flags;
 
-static int
-nouveau_fence_wait_uevent_handler(void *data, u32 type, int index)
-{
-	struct nouveau_fence_priv *priv = data;
-	wake_up_all(&priv->waiting);
-	return NVKM_EVENT_KEEP;
-}
+		if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+			return true;
 
-static int
-nouveau_fence_wait_uevent(struct nouveau_fence *fence, bool intr)
+		spin_lock_irqsave(&fctx->lock, flags);
+		nouveau_fence_update(fence->channel, fctx);
+		spin_unlock_irqrestore(&fctx->lock, flags);
+	}
+	return fence_is_signaled(&fence->base);
+}
 
+static long
+nouveau_fence_wait_legacy(struct fence *f, bool intr, long wait)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fifo *pfifo = nouveau_fifo(chan->drm->device);
-	struct nouveau_fence_priv *priv = chan->drm->fence;
-	struct nouveau_eventh *handler;
-	int ret = 0;
+	struct nouveau_fence *fence = from_fence(f);
+	unsigned long sleep_time = NSEC_PER_MSEC / 1000;
+	unsigned long t = jiffies, timeout = t + wait;
 
-	ret = nouveau_event_new(pfifo->uevent, 1, 0,
-				nouveau_fence_wait_uevent_handler,
-				priv, &handler);
-	if (ret)
-		return ret;
+	while (!nouveau_fence_done(fence)) {
+		ktime_t kt;
 
-	nouveau_event_get(handler);
-
-	if (fence->timeout) {
-		unsigned long timeout = fence->timeout - jiffies;
-
-		if (time_before(jiffies, fence->timeout)) {
-			if (intr) {
-				ret = wait_event_interruptible_timeout(
-						priv->waiting,
-						nouveau_fence_done(fence),
-						timeout);
-			} else {
-				ret = wait_event_timeout(priv->waiting,
-						nouveau_fence_done(fence),
-						timeout);
-			}
-		}
+		t = jiffies;
 
-		if (ret >= 0) {
-			fence->timeout = jiffies + ret;
-			if (time_after_eq(jiffies, fence->timeout))
-				ret = -EBUSY;
-		}
-	} else {
-		if (intr) {
-			ret = wait_event_interruptible(priv->waiting,
-					nouveau_fence_done(fence));
-		} else {
-			wait_event(priv->waiting, nouveau_fence_done(fence));
+		if (wait != MAX_SCHEDULE_TIMEOUT && time_after_eq(t, timeout)) {
+			__set_current_state(TASK_RUNNING);
+			return 0;
 		}
+
+		__set_current_state(intr ? TASK_INTERRUPTIBLE :
+					   TASK_UNINTERRUPTIBLE);
+
+		kt = ktime_set(0, sleep_time);
+		schedule_hrtimeout(&kt, HRTIMER_MODE_REL);
+		sleep_time *= 2;
+		if (sleep_time > NSEC_PER_MSEC)
+			sleep_time = NSEC_PER_MSEC;
+
+		if (intr && signal_pending(current))
+			return -ERESTARTSYS;
 	}
 
-	nouveau_event_ref(NULL, &handler);
-	if (unlikely(ret < 0))
-		return ret;
+	__set_current_state(TASK_RUNNING);
 
-	return 0;
+	return timeout - t;
 }
 
-int
-nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
+static int
+nouveau_fence_wait_busy(struct nouveau_fence *fence, bool intr)
 {
-	struct nouveau_channel *chan = fence->channel;
-	struct nouveau_fence_priv *priv = chan ? chan->drm->fence : NULL;
-	unsigned long sleep_time = NSEC_PER_MSEC / 1000;
-	ktime_t t;
 	int ret = 0;
 
-	while (priv && priv->uevent && lazy && !nouveau_fence_done(fence)) {
-		ret = nouveau_fence_wait_uevent(fence, intr);
-		if (ret < 0)
-			return ret;
-	}
-
 	while (!nouveau_fence_done(fence)) {
-		if (fence->timeout && time_after_eq(jiffies, fence->timeout)) {
+		if (time_after_eq(jiffies, fence->timeout)) {
 			ret = -EBUSY;
 			break;
 		}
 
-		__set_current_state(intr ? TASK_INTERRUPTIBLE :
-					   TASK_UNINTERRUPTIBLE);
-		if (lazy) {
-			t = ktime_set(0, sleep_time);
-			schedule_hrtimeout(&t, HRTIMER_MODE_REL);
-			sleep_time *= 2;
-			if (sleep_time > NSEC_PER_MSEC)
-				sleep_time = NSEC_PER_MSEC;
-		}
+		__set_current_state(intr ?
+				    TASK_INTERRUPTIBLE :
+				    TASK_UNINTERRUPTIBLE);
 
 		if (intr && signal_pending(current)) {
 			ret = -ERESTARTSYS;
@@ -270,47 +349,86 @@ nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
 }
 
 int
-nouveau_fence_sync(struct nouveau_fence *fence, struct nouveau_channel *chan)
+nouveau_fence_wait(struct nouveau_fence *fence, bool lazy, bool intr)
+{
+	long ret;
+
+	if (!lazy)
+		return nouveau_fence_wait_busy(fence, intr);
+
+	ret = fence_wait_timeout(&fence->base, intr, 15 * HZ);
+	if (ret < 0)
+		return ret;
+	else if (!ret)
+		return -EBUSY;
+	else
+		return 0;
+}
+
+int
+nouveau_fence_sync(struct nouveau_bo *nvbo, struct nouveau_channel *chan, bool exclusive, bool intr)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
-	struct nouveau_channel *prev;
-	int ret = 0;
+	struct fence *fence;
+	struct reservation_object *resv = nvbo->bo.resv;
+	struct reservation_object_list *fobj;
+	struct nouveau_fence *f;
+	int ret = 0, i;
 
-	prev = fence ? fence->channel : NULL;
-	if (prev) {
-		if (unlikely(prev != chan && !nouveau_fence_done(fence))) {
-			ret = fctx->sync(fence, prev, chan);
-			if (unlikely(ret))
-				ret = nouveau_fence_wait(fence, true, false);
-		}
+	if (!exclusive) {
+		ret = reservation_object_reserve_shared(resv);
+
+		if (ret)
+			return ret;
 	}
 
-	return ret;
-}
+	fobj = reservation_object_get_list(resv);
+	fence = reservation_object_get_excl(resv);
 
-static void
-nouveau_fence_del(struct kref *kref)
-{
-	struct nouveau_fence *fence = container_of(kref, typeof(*fence), kref);
-	kfree(fence);
+	if (fence && (!exclusive || !fobj || !fobj->shared_count)) {
+		struct nouveau_channel *prev = NULL;
+
+		f = nouveau_local_fence(fence, chan->drm);
+		if (f)
+			prev = f->channel;
+
+		if (!prev || (prev != chan && (ret = fctx->sync(f, prev, chan))))
+			ret = fence_wait(fence, intr);
+
+		return ret;
+	}
+
+	if (!exclusive || !fobj)
+		return ret;
+
+	for (i = 0; i < fobj->shared_count && !ret; ++i) {
+		struct nouveau_channel *prev = NULL;
+
+		fence = rcu_dereference_protected(fobj->shared[i],
+						reservation_object_held(resv));
+
+		f = nouveau_local_fence(fence, chan->drm);
+		if (f)
+			prev = f->channel;
+
+		if (!prev || (prev != chan && (ret = fctx->sync(f, prev, chan))))
+			ret = fence_wait(fence, intr);
+
+		if (ret)
+			break;
+	}
+
+	return ret;
 }
 
 void
 nouveau_fence_unref(struct nouveau_fence **pfence)
 {
 	if (*pfence)
-		kref_put(&(*pfence)->kref, nouveau_fence_del);
+		fence_put(&(*pfence)->base);
 	*pfence = NULL;
 }
 
-struct nouveau_fence *
-nouveau_fence_ref(struct nouveau_fence *fence)
-{
-	if (fence)
-		kref_get(&fence->kref);
-	return fence;
-}
-
 int
 nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 		  struct nouveau_fence **pfence)
@@ -325,9 +443,7 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	if (!fence)
 		return -ENOMEM;
 
-	INIT_LIST_HEAD(&fence->work);
 	fence->sysmem = sysmem;
-	kref_init(&fence->kref);
 
 	ret = nouveau_fence_emit(fence, chan);
 	if (ret)
@@ -336,3 +452,101 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	*pfence = fence;
 	return ret;
 }
+
+static const char *nouveau_fence_get_get_driver_name(struct fence *fence)
+{
+	return "nouveau";
+}
+
+static const char *nouveau_fence_get_timeline_name(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+
+	return fence->channel ? fctx->name : "dead channel";
+}
+
+/*
+ * In an ideal world, read would not assume the channel context is still alive.
+ * This function may be called from another device, running into free memory as a
+ * result. The drm node should still be there, so we can derive the index from
+ * the fence context.
+ */
+static bool nouveau_fence_is_signaled(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+	struct nouveau_channel *chan = fence->channel;
+
+	return (int)(fctx->read(chan) - fence->base.seqno) >= 0;
+}
+
+static bool nouveau_fence_no_signaling(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+
+	/*
+	 * caller should have a reference on the fence,
+	 * else fence could get freed here
+	 */
+	WARN_ON(atomic_read(&fence->base.refcount.refcount) <= 1);
+
+	/*
+	 * This needs uevents to work correctly, but fence_add_callback relies on
+	 * being able to enable signaling. It will still get signaled eventually,
+	 * just not right away.
+	 */
+	if (nouveau_fence_is_signaled(f)) {
+		list_del(&fence->head);
+
+		fence_put(&fence->base);
+		return false;
+	}
+
+	return true;
+}
+
+static void nouveau_fence_release(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+
+	kref_put(&fctx->fence_ref, nouveau_fence_context_put);
+	fence_free(&fence->base);
+}
+
+static const struct fence_ops nouveau_fence_ops_legacy = {
+	.get_driver_name = nouveau_fence_get_get_driver_name,
+	.get_timeline_name = nouveau_fence_get_timeline_name,
+	.enable_signaling = nouveau_fence_no_signaling,
+	.signaled = nouveau_fence_is_signaled,
+	.wait = nouveau_fence_wait_legacy,
+	.release = nouveau_fence_release
+};
+
+static bool nouveau_fence_enable_signaling(struct fence *f)
+{
+	struct nouveau_fence *fence = from_fence(f);
+	struct nouveau_fence_chan *fctx = nouveau_fctx(fence);
+	bool ret;
+
+	if (!fctx->notify_ref++)
+		nvif_notify_get(&fctx->notify);
+
+	ret = nouveau_fence_no_signaling(f);
+	if (ret)
+		set_bit(FENCE_FLAG_USER_BITS, &fence->base.flags);
+	else if (!--fctx->notify_ref)
+		nvif_notify_put(&fctx->notify);
+
+	return ret;
+}
+
+static const struct fence_ops nouveau_fence_ops_uevent = {
+	.get_driver_name = nouveau_fence_get_get_driver_name,
+	.get_timeline_name = nouveau_fence_get_timeline_name,
+	.enable_signaling = nouveau_fence_enable_signaling,
+	.signaled = nouveau_fence_is_signaled,
+	.wait = fence_default_wait,
+	.release = NULL
+};
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index c57bb61da58c..943b0b17b1fc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -1,33 +1,37 @@
 #ifndef __NOUVEAU_FENCE_H__
 #define __NOUVEAU_FENCE_H__
 
+#include <linux/fence.h>
+#include <nvif/notify.h>
+
 struct nouveau_drm;
+struct nouveau_bo;
 
 struct nouveau_fence {
+	struct fence base;
+
 	struct list_head head;
-	struct list_head work;
-	struct kref kref;
 
 	bool sysmem;
 
 	struct nouveau_channel *channel;
 	unsigned long timeout;
-	u32 sequence;
 };
 
 int  nouveau_fence_new(struct nouveau_channel *, bool sysmem,
 		       struct nouveau_fence **);
-struct nouveau_fence *
-nouveau_fence_ref(struct nouveau_fence *);
 void nouveau_fence_unref(struct nouveau_fence **);
 
 int  nouveau_fence_emit(struct nouveau_fence *, struct nouveau_channel *);
 bool nouveau_fence_done(struct nouveau_fence *);
-void nouveau_fence_work(struct nouveau_fence *, void (*)(void *), void *);
+void nouveau_fence_work(struct fence *, void (*)(void *), void *);
 int  nouveau_fence_wait(struct nouveau_fence *, bool lazy, bool intr);
-int  nouveau_fence_sync(struct nouveau_fence *, struct nouveau_channel *);
+int  nouveau_fence_sync(struct nouveau_bo *, struct nouveau_channel *, bool exclusive, bool intr);
 
 struct nouveau_fence_chan {
+	spinlock_t lock;
+	struct kref fence_ref;
+
 	struct list_head pending;
 	struct list_head flip;
 
@@ -38,8 +42,12 @@ struct nouveau_fence_chan {
 	int  (*emit32)(struct nouveau_channel *, u64, u32);
 	int  (*sync32)(struct nouveau_channel *, u64, u32);
 
-	spinlock_t lock;
 	u32 sequence;
+	u32 context;
+	char name[32];
+
+	struct nvif_notify notify;
+	int notify_ref;
 };
 
 struct nouveau_fence_priv {
@@ -49,14 +57,15 @@ struct nouveau_fence_priv {
 	int  (*context_new)(struct nouveau_channel *);
 	void (*context_del)(struct nouveau_channel *);
 
-	wait_queue_head_t waiting;
+	u32 contexts, context_base;
 	bool uevent;
 };
 
 #define nouveau_fence(drm) ((struct nouveau_fence_priv *)(drm)->fence)
 
-void nouveau_fence_context_new(struct nouveau_fence_chan *);
+void nouveau_fence_context_new(struct nouveau_channel *, struct nouveau_fence_chan *);
 void nouveau_fence_context_del(struct nouveau_fence_chan *);
+void nouveau_fence_context_free(struct nouveau_fence_chan *);
 
 int nv04_fence_create(struct nouveau_drm *);
 int nv04_fence_mthd(struct nouveau_channel *, u32, u32, u32);
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index c90c0dc0afe8..36951ee4b157 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -24,8 +24,6 @@
  *
  */
 
-#include <subdev/fb.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
@@ -58,14 +56,14 @@ nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
 	struct nouveau_vma *vma;
 	int ret;
 
-	if (!cli->base.vm)
+	if (!cli->vm)
 		return 0;
 
-	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, 0);
+	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, NULL);
 	if (ret)
 		return ret;
 
-	vma = nouveau_bo_vma_find(nvbo, cli->base.vm);
+	vma = nouveau_bo_vma_find(nvbo, cli->vm);
 	if (!vma) {
 		vma = kzalloc(sizeof(*vma), GFP_KERNEL);
 		if (!vma) {
@@ -73,7 +71,7 @@ nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
 			goto out;
 		}
 
-		ret = nouveau_bo_vma_add(nvbo, cli->base.vm, vma);
+		ret = nouveau_bo_vma_add(nvbo, cli->vm, vma);
 		if (ret) {
 			kfree(vma);
 			goto out;
@@ -100,17 +98,23 @@ static void
 nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 {
 	const bool mapped = nvbo->bo.mem.mem_type != TTM_PL_SYSTEM;
-	struct nouveau_fence *fence = NULL;
+	struct reservation_object *resv = nvbo->bo.resv;
+	struct reservation_object_list *fobj;
+	struct fence *fence = NULL;
+
+	fobj = reservation_object_get_list(resv);
 
 	list_del(&vma->head);
 
-	if (mapped) {
-		spin_lock(&nvbo->bo.bdev->fence_lock);
-		fence = nouveau_fence_ref(nvbo->bo.sync_obj);
-		spin_unlock(&nvbo->bo.bdev->fence_lock);
-	}
+	if (fobj && fobj->shared_count > 1)
+		ttm_bo_wait(&nvbo->bo, true, false, false);
+	else if (fobj && fobj->shared_count == 1)
+		fence = rcu_dereference_protected(fobj->shared[0],
+						reservation_object_held(resv));
+	else
+		fence = reservation_object_get_excl(nvbo->bo.resv);
 
-	if (fence) {
+	if (fence && mapped) {
 		nouveau_fence_work(fence, nouveau_gem_object_delete, vma);
 	} else {
 		if (mapped)
@@ -118,7 +122,6 @@ nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 		nouveau_vm_put(vma);
 		kfree(vma);
 	}
-	nouveau_fence_unref(&fence);
 }
 
 void
@@ -129,14 +132,14 @@ nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 	struct nouveau_vma *vma;
 	int ret;
 
-	if (!cli->base.vm)
+	if (!cli->vm)
 		return;
 
-	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, 0);
+	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, NULL);
 	if (ret)
 		return;
 
-	vma = nouveau_bo_vma_find(nvbo, cli->base.vm);
+	vma = nouveau_bo_vma_find(nvbo, cli->vm);
 	if (vma) {
 		if (--vma->refcount == 0)
 			nouveau_gem_object_unmap(nvbo, vma);
@@ -162,7 +165,7 @@ nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
 		flags |= TTM_PL_FLAG_SYSTEM;
 
 	ret = nouveau_bo_new(dev, size, align, flags, tile_mode,
-			     tile_flags, NULL, pnvbo);
+			     tile_flags, NULL, NULL, pnvbo);
 	if (ret)
 		return ret;
 	nvbo = *pnvbo;
@@ -173,7 +176,7 @@ nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
 	 */
 	nvbo->valid_domains = NOUVEAU_GEM_DOMAIN_VRAM |
 			      NOUVEAU_GEM_DOMAIN_GART;
-	if (nv_device(drm->device)->card_type >= NV_50)
+	if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA)
 		nvbo->valid_domains &= domain;
 
 	/* Initialize the embedded gem-object. We return a single gem-reference
@@ -202,8 +205,8 @@ nouveau_gem_info(struct drm_file *file_priv, struct drm_gem_object *gem,
 		rep->domain = NOUVEAU_GEM_DOMAIN_VRAM;
 
 	rep->offset = nvbo->bo.offset;
-	if (cli->base.vm) {
-		vma = nouveau_bo_vma_find(nvbo, cli->base.vm);
+	if (cli->vm) {
+		vma = nouveau_bo_vma_find(nvbo, cli->vm);
 		if (!vma)
 			return -EINVAL;
 
@@ -223,13 +226,13 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data,
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	struct drm_nouveau_gem_new *req = data;
 	struct nouveau_bo *nvbo = NULL;
 	int ret = 0;
 
 	if (!pfb->memtype_valid(pfb, req->info.tile_flags)) {
-		NV_ERROR(cli, "bad page flags: 0x%08x\n", req->info.tile_flags);
+		NV_PRINTK(error, cli, "bad page flags: 0x%08x\n", req->info.tile_flags);
 		return -EINVAL;
 	}
 
@@ -290,24 +293,23 @@ nouveau_gem_set_domain(struct drm_gem_object *gem, uint32_t read_domains,
 }
 
 struct validate_op {
-	struct list_head vram_list;
-	struct list_head gart_list;
-	struct list_head both_list;
+	struct list_head list;
 	struct ww_acquire_ctx ticket;
 };
 
 static void
-validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
-		   struct ww_acquire_ctx *ticket)
+validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence,
+			struct drm_nouveau_gem_pushbuf_bo *pbbo)
 {
-	struct list_head *entry, *tmp;
 	struct nouveau_bo *nvbo;
+	struct drm_nouveau_gem_pushbuf_bo *b;
 
-	list_for_each_safe(entry, tmp, list) {
-		nvbo = list_entry(entry, struct nouveau_bo, entry);
+	while (!list_empty(&op->list)) {
+		nvbo = list_entry(op->list.next, struct nouveau_bo, entry);
+		b = &pbbo[nvbo->pbbo_index];
 
 		if (likely(fence))
-			nouveau_bo_fence(nvbo, fence);
+			nouveau_bo_fence(nvbo, fence, !!b->write_domains);
 
 		if (unlikely(nvbo->validate_mapped)) {
 			ttm_bo_kunmap(&nvbo->kmap);
@@ -316,23 +318,16 @@ validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
 
 		list_del(&nvbo->entry);
 		nvbo->reserved_by = NULL;
-		ttm_bo_unreserve_ticket(&nvbo->bo, ticket);
+		ttm_bo_unreserve_ticket(&nvbo->bo, &op->ticket);
 		drm_gem_object_unreference_unlocked(&nvbo->gem);
 	}
 }
 
 static void
-validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence)
+validate_fini(struct validate_op *op, struct nouveau_fence *fence,
+	      struct drm_nouveau_gem_pushbuf_bo *pbbo)
 {
-	validate_fini_list(&op->vram_list, fence, &op->ticket);
-	validate_fini_list(&op->gart_list, fence, &op->ticket);
-	validate_fini_list(&op->both_list, fence, &op->ticket);
-}
-
-static void
-validate_fini(struct validate_op *op, struct nouveau_fence *fence)
-{
-	validate_fini_no_ticket(op, fence);
+	validate_fini_no_ticket(op, fence, pbbo);
 	ww_acquire_fini(&op->ticket);
 }
 
@@ -346,11 +341,14 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 	int trycnt = 0;
 	int ret, i;
 	struct nouveau_bo *res_bo = NULL;
+	LIST_HEAD(gart_list);
+	LIST_HEAD(vram_list);
+	LIST_HEAD(both_list);
 
 	ww_acquire_init(&op->ticket, &reservation_ww_class);
 retry:
 	if (++trycnt > 100000) {
-		NV_ERROR(cli, "%s failed and gave up.\n", __func__);
+		NV_PRINTK(error, cli, "%s failed and gave up.\n", __func__);
 		return -EINVAL;
 	}
 
@@ -361,10 +359,9 @@ retry:
 
 		gem = drm_gem_object_lookup(dev, file_priv, b->handle);
 		if (!gem) {
-			NV_ERROR(cli, "Unknown handle 0x%08x\n", b->handle);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -ENOENT;
+			NV_PRINTK(error, cli, "Unknown handle 0x%08x\n", b->handle);
+			ret = -ENOENT;
+			break;
 		}
 		nvbo = nouveau_gem_object(gem);
 		if (nvbo == res_bo) {
@@ -374,17 +371,19 @@ retry:
 		}
 
 		if (nvbo->reserved_by && nvbo->reserved_by == file_priv) {
-			NV_ERROR(cli, "multiple instances of buffer %d on "
+			NV_PRINTK(error, cli, "multiple instances of buffer %d on "
 				      "validation list\n", b->handle);
 			drm_gem_object_unreference_unlocked(gem);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -EINVAL;
+			ret = -EINVAL;
+			break;
 		}
 
 		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, &op->ticket);
 		if (ret) {
-			validate_fini_no_ticket(op, NULL);
+			list_splice_tail_init(&vram_list, &op->list);
+			list_splice_tail_init(&gart_list, &op->list);
+			list_splice_tail_init(&both_list, &op->list);
+			validate_fini_no_ticket(op, NULL, NULL);
 			if (unlikely(ret == -EDEADLK)) {
 				ret = ttm_bo_reserve_slowpath(&nvbo->bo, true,
 							      &op->ticket);
@@ -392,12 +391,9 @@ retry:
 					res_bo = nvbo;
 			}
 			if (unlikely(ret)) {
-				ww_acquire_done(&op->ticket);
-				ww_acquire_fini(&op->ticket);
-				drm_gem_object_unreference_unlocked(gem);
 				if (ret != -ERESTARTSYS)
-					NV_ERROR(cli, "fail reserve\n");
-				return ret;
+					NV_PRINTK(error, cli, "fail reserve\n");
+				break;
 			}
 		}
 
@@ -406,45 +402,32 @@ retry:
 		nvbo->pbbo_index = i;
 		if ((b->valid_domains & NOUVEAU_GEM_DOMAIN_VRAM) &&
 		    (b->valid_domains & NOUVEAU_GEM_DOMAIN_GART))
-			list_add_tail(&nvbo->entry, &op->both_list);
+			list_add_tail(&nvbo->entry, &both_list);
 		else
 		if (b->valid_domains & NOUVEAU_GEM_DOMAIN_VRAM)
-			list_add_tail(&nvbo->entry, &op->vram_list);
+			list_add_tail(&nvbo->entry, &vram_list);
 		else
 		if (b->valid_domains & NOUVEAU_GEM_DOMAIN_GART)
-			list_add_tail(&nvbo->entry, &op->gart_list);
+			list_add_tail(&nvbo->entry, &gart_list);
 		else {
-			NV_ERROR(cli, "invalid valid domains: 0x%08x\n",
+			NV_PRINTK(error, cli, "invalid valid domains: 0x%08x\n",
 				 b->valid_domains);
-			list_add_tail(&nvbo->entry, &op->both_list);
-			ww_acquire_done(&op->ticket);
-			validate_fini(op, NULL);
-			return -EINVAL;
+			list_add_tail(&nvbo->entry, &both_list);
+			ret = -EINVAL;
+			break;
 		}
 		if (nvbo == res_bo)
 			goto retry;
 	}
 
 	ww_acquire_done(&op->ticket);
-	return 0;
-}
-
-static int
-validate_sync(struct nouveau_channel *chan, struct nouveau_bo *nvbo)
-{
-	struct nouveau_fence *fence = NULL;
-	int ret = 0;
-
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	fence = nouveau_fence_ref(nvbo->bo.sync_obj);
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
-
-	if (fence) {
-		ret = nouveau_fence_sync(fence, chan);
-		nouveau_fence_unref(&fence);
-	}
-
+	list_splice_tail(&vram_list, &op->list);
+	list_splice_tail(&gart_list, &op->list);
+	list_splice_tail(&both_list, &op->list);
+	if (ret)
+		validate_fini(op, NULL, NULL);
 	return ret;
+
 }
 
 static int
@@ -465,24 +448,25 @@ validate_list(struct nouveau_channel *chan, struct nouveau_cli *cli,
 					     b->write_domains,
 					     b->valid_domains);
 		if (unlikely(ret)) {
-			NV_ERROR(cli, "fail set_domain\n");
+			NV_PRINTK(error, cli, "fail set_domain\n");
 			return ret;
 		}
 
 		ret = nouveau_bo_validate(nvbo, true, false);
 		if (unlikely(ret)) {
 			if (ret != -ERESTARTSYS)
-				NV_ERROR(cli, "fail ttm_validate\n");
+				NV_PRINTK(error, cli, "fail ttm_validate\n");
 			return ret;
 		}
 
-		ret = validate_sync(chan, nvbo);
+		ret = nouveau_fence_sync(nvbo, chan, !!b->write_domains, true);
 		if (unlikely(ret)) {
-			NV_ERROR(cli, "fail post-validate sync\n");
+			if (ret != -ERESTARTSYS)
+				NV_PRINTK(error, cli, "fail post-validate sync\n");
 			return ret;
 		}
 
-		if (nv_device(drm->device)->card_type < NV_50) {
+		if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) {
 			if (nvbo->bo.offset == b->presumed.offset &&
 			    ((nvbo->bo.mem.mem_type == TTM_PL_VRAM &&
 			      b->presumed.domain & NOUVEAU_GEM_DOMAIN_VRAM) ||
@@ -515,11 +499,9 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 			     struct validate_op *op, int *apply_relocs)
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
-	int ret, relocs = 0;
+	int ret;
 
-	INIT_LIST_HEAD(&op->vram_list);
-	INIT_LIST_HEAD(&op->gart_list);
-	INIT_LIST_HEAD(&op->both_list);
+	INIT_LIST_HEAD(&op->list);
 
 	if (nr_buffers == 0)
 		return 0;
@@ -527,38 +509,18 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan,
 	ret = validate_init(chan, file_priv, pbbo, nr_buffers, op);
 	if (unlikely(ret)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(cli, "validate_init\n");
+			NV_PRINTK(error, cli, "validate_init\n");
 		return ret;
 	}
 
-	ret = validate_list(chan, cli, &op->vram_list, pbbo, user_buffers);
+	ret = validate_list(chan, cli, &op->list, pbbo, user_buffers);
 	if (unlikely(ret < 0)) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(cli, "validate vram_list\n");
-		validate_fini(op, NULL);
+			NV_PRINTK(error, cli, "validating bo list\n");
+		validate_fini(op, NULL, NULL);
 		return ret;
 	}
-	relocs += ret;
-
-	ret = validate_list(chan, cli, &op->gart_list, pbbo, user_buffers);
-	if (unlikely(ret < 0)) {
-		if (ret != -ERESTARTSYS)
-			NV_ERROR(cli, "validate gart_list\n");
-		validate_fini(op, NULL);
-		return ret;
-	}
-	relocs += ret;
-
-	ret = validate_list(chan, cli, &op->both_list, pbbo, user_buffers);
-	if (unlikely(ret < 0)) {
-		if (ret != -ERESTARTSYS)
-			NV_ERROR(cli, "validate both_list\n");
-		validate_fini(op, NULL);
-		return ret;
-	}
-	relocs += ret;
-
-	*apply_relocs = relocs;
+	*apply_relocs = ret;
 	return 0;
 }
 
@@ -613,7 +575,7 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 		uint32_t data;
 
 		if (unlikely(r->bo_index > req->nr_buffers)) {
-			NV_ERROR(cli, "reloc bo index invalid\n");
+			NV_PRINTK(error, cli, "reloc bo index invalid\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -623,7 +585,7 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 			continue;
 
 		if (unlikely(r->reloc_bo_index > req->nr_buffers)) {
-			NV_ERROR(cli, "reloc container bo index invalid\n");
+			NV_PRINTK(error, cli, "reloc container bo index invalid\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -631,7 +593,7 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 
 		if (unlikely(r->reloc_bo_offset + 4 >
 			     nvbo->bo.mem.num_pages << PAGE_SHIFT)) {
-			NV_ERROR(cli, "reloc outside of bo\n");
+			NV_PRINTK(error, cli, "reloc outside of bo\n");
 			ret = -EINVAL;
 			break;
 		}
@@ -640,7 +602,7 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 			ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages,
 					  &nvbo->kmap);
 			if (ret) {
-				NV_ERROR(cli, "failed kmap for reloc\n");
+				NV_PRINTK(error, cli, "failed kmap for reloc\n");
 				break;
 			}
 			nvbo->validate_mapped = true;
@@ -661,11 +623,9 @@ nouveau_gem_pushbuf_reloc_apply(struct nouveau_cli *cli,
 				data |= r->vor;
 		}
 
-		spin_lock(&nvbo->bo.bdev->fence_lock);
-		ret = ttm_bo_wait(&nvbo->bo, false, false, false);
-		spin_unlock(&nvbo->bo.bdev->fence_lock);
+		ret = ttm_bo_wait(&nvbo->bo, true, false, false);
 		if (ret) {
-			NV_ERROR(cli, "reloc wait_idle failed: %d\n", ret);
+			NV_PRINTK(error, cli, "reloc wait_idle failed: %d\n", ret);
 			break;
 		}
 
@@ -696,7 +656,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		return -ENOMEM;
 
 	list_for_each_entry(temp, &abi16->channels, head) {
-		if (temp->chan->handle == (NVDRM_CHAN | req->channel)) {
+		if (temp->chan->object->handle == (NVDRM_CHAN | req->channel)) {
 			chan = temp->chan;
 			break;
 		}
@@ -711,19 +671,19 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		goto out_next;
 
 	if (unlikely(req->nr_push > NOUVEAU_GEM_MAX_PUSH)) {
-		NV_ERROR(cli, "pushbuf push count exceeds limit: %d max %d\n",
+		NV_PRINTK(error, cli, "pushbuf push count exceeds limit: %d max %d\n",
 			 req->nr_push, NOUVEAU_GEM_MAX_PUSH);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
 
 	if (unlikely(req->nr_buffers > NOUVEAU_GEM_MAX_BUFFERS)) {
-		NV_ERROR(cli, "pushbuf bo count exceeds limit: %d max %d\n",
+		NV_PRINTK(error, cli, "pushbuf bo count exceeds limit: %d max %d\n",
 			 req->nr_buffers, NOUVEAU_GEM_MAX_BUFFERS);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
 
 	if (unlikely(req->nr_relocs > NOUVEAU_GEM_MAX_RELOCS)) {
-		NV_ERROR(cli, "pushbuf reloc count exceeds limit: %d max %d\n",
+		NV_PRINTK(error, cli, "pushbuf reloc count exceeds limit: %d max %d\n",
 			 req->nr_relocs, NOUVEAU_GEM_MAX_RELOCS);
 		return nouveau_abi16_put(abi16, -EINVAL);
 	}
@@ -741,7 +701,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	/* Ensure all push buffers are on validate list */
 	for (i = 0; i < req->nr_push; i++) {
 		if (push[i].bo_index >= req->nr_buffers) {
-			NV_ERROR(cli, "push %d buffer not in list\n", i);
+			NV_PRINTK(error, cli, "push %d buffer not in list\n", i);
 			ret = -EINVAL;
 			goto out_prevalid;
 		}
@@ -752,7 +712,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 					   req->nr_buffers, &op, &do_reloc);
 	if (ret) {
 		if (ret != -ERESTARTSYS)
-			NV_ERROR(cli, "validate: %d\n", ret);
+			NV_PRINTK(error, cli, "validate: %d\n", ret);
 		goto out_prevalid;
 	}
 
@@ -760,7 +720,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	if (do_reloc) {
 		ret = nouveau_gem_pushbuf_reloc_apply(cli, req, bo);
 		if (ret) {
-			NV_ERROR(cli, "reloc apply: %d\n", ret);
+			NV_PRINTK(error, cli, "reloc apply: %d\n", ret);
 			goto out;
 		}
 	}
@@ -768,7 +728,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	if (chan->dma.ib_max) {
 		ret = nouveau_dma_wait(chan, req->nr_push + 1, 16);
 		if (ret) {
-			NV_ERROR(cli, "nv50cal_space: %d\n", ret);
+			NV_PRINTK(error, cli, "nv50cal_space: %d\n", ret);
 			goto out;
 		}
 
@@ -780,10 +740,10 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 				      push[i].length);
 		}
 	} else
-	if (nv_device(drm->device)->chipset >= 0x25) {
+	if (drm->device.info.chipset >= 0x25) {
 		ret = RING_SPACE(chan, req->nr_push * 2);
 		if (ret) {
-			NV_ERROR(cli, "cal_space: %d\n", ret);
+			NV_PRINTK(error, cli, "cal_space: %d\n", ret);
 			goto out;
 		}
 
@@ -797,7 +757,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	} else {
 		ret = RING_SPACE(chan, req->nr_push * (2 + NOUVEAU_DMA_SKIPS));
 		if (ret) {
-			NV_ERROR(cli, "jmp_space: %d\n", ret);
+			NV_PRINTK(error, cli, "jmp_space: %d\n", ret);
 			goto out;
 		}
 
@@ -835,13 +795,13 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 
 	ret = nouveau_fence_new(chan, false, &fence);
 	if (ret) {
-		NV_ERROR(cli, "error fencing pushbuf: %d\n", ret);
+		NV_PRINTK(error, cli, "error fencing pushbuf: %d\n", ret);
 		WIND_RING(chan);
 		goto out;
 	}
 
 out:
-	validate_fini(&op, fence);
+	validate_fini(&op, fence, bo);
 	nouveau_fence_unref(&fence);
 
 out_prevalid:
@@ -853,7 +813,7 @@ out_next:
 		req->suffix0 = 0x00000000;
 		req->suffix1 = 0x00000000;
 	} else
-	if (nv_device(drm->device)->chipset >= 0x25) {
+	if (drm->device.info.chipset >= 0x25) {
 		req->suffix0 = 0x00020000;
 		req->suffix1 = 0x00000000;
 	} else {
@@ -886,17 +846,29 @@ nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void *data,
 	struct drm_gem_object *gem;
 	struct nouveau_bo *nvbo;
 	bool no_wait = !!(req->flags & NOUVEAU_GEM_CPU_PREP_NOWAIT);
-	int ret = -EINVAL;
+	bool write = !!(req->flags & NOUVEAU_GEM_CPU_PREP_WRITE);
+	int ret;
 
 	gem = drm_gem_object_lookup(dev, file_priv, req->handle);
 	if (!gem)
 		return -ENOENT;
 	nvbo = nouveau_gem_object(gem);
 
-	spin_lock(&nvbo->bo.bdev->fence_lock);
-	ret = ttm_bo_wait(&nvbo->bo, true, true, no_wait);
-	spin_unlock(&nvbo->bo.bdev->fence_lock);
+	if (no_wait)
+		ret = reservation_object_test_signaled_rcu(nvbo->bo.resv, write) ? 0 : -EBUSY;
+	else {
+		long lret;
+
+		lret = reservation_object_wait_timeout_rcu(nvbo->bo.resv, write, true, 30 * HZ);
+		if (!lret)
+			ret = -EBUSY;
+		else if (lret > 0)
+			ret = 0;
+		else
+			ret = lret;
+	}
 	drm_gem_object_unreference_unlocked(gem);
+
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.h b/drivers/gpu/drm/nouveau/nouveau_gem.h
index ddab762d81fe..e4049faca780 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.h
@@ -39,7 +39,7 @@ struct reservation_object *nouveau_gem_prime_res_obj(struct drm_gem_object *);
 extern void nouveau_gem_prime_unpin(struct drm_gem_object *);
 extern struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *);
 extern struct drm_gem_object *nouveau_gem_prime_import_sg_table(
-	struct drm_device *, size_t size, struct sg_table *);
+	struct drm_device *, struct dma_buf_attachment *, struct sg_table *);
 extern void *nouveau_gem_prime_vmap(struct drm_gem_object *);
 extern void nouveau_gem_prime_vunmap(struct drm_gem_object *, void *);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_hwmon.c b/drivers/gpu/drm/nouveau/nouveau_hwmon.c
index 19fd767bab10..afb36d66e78d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_hwmon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_hwmon.c
@@ -34,17 +34,13 @@
 #include "nouveau_drm.h"
 #include "nouveau_hwmon.h"
 
-#include <subdev/gpio.h>
-#include <subdev/timer.h>
-#include <subdev/therm.h>
-
 #if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE))
 static ssize_t
 nouveau_hwmon_show_temp(struct device *d, struct device_attribute *a, char *buf)
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int temp = therm->temp_get(therm);
 
 	if (temp < 0)
@@ -70,7 +66,7 @@ nouveau_hwmon_temp1_auto_point1_temp(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	      therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST) * 1000);
@@ -82,7 +78,7 @@ nouveau_hwmon_set_temp1_auto_point1_temp(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -103,7 +99,7 @@ nouveau_hwmon_temp1_auto_point1_temp_hyst(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	 therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_FAN_BOOST_HYST) * 1000);
@@ -115,7 +111,7 @@ nouveau_hwmon_set_temp1_auto_point1_temp_hyst(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -135,7 +131,7 @@ nouveau_hwmon_max_temp(struct device *d, struct device_attribute *a, char *buf)
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	       therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_DOWN_CLK) * 1000);
@@ -146,7 +142,7 @@ nouveau_hwmon_set_max_temp(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -166,7 +162,7 @@ nouveau_hwmon_max_temp_hyst(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_DOWN_CLK_HYST) * 1000);
@@ -177,7 +173,7 @@ nouveau_hwmon_set_max_temp_hyst(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -198,7 +194,7 @@ nouveau_hwmon_critical_temp(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	       therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_CRITICAL) * 1000);
@@ -210,7 +206,7 @@ nouveau_hwmon_set_critical_temp(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -231,7 +227,7 @@ nouveau_hwmon_critical_temp_hyst(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_CRITICAL_HYST) * 1000);
@@ -244,7 +240,7 @@ nouveau_hwmon_set_critical_temp_hyst(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -264,7 +260,7 @@ nouveau_hwmon_emergency_temp(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	       therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN) * 1000);
@@ -276,7 +272,7 @@ nouveau_hwmon_set_emergency_temp(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -297,7 +293,7 @@ nouveau_hwmon_emergency_temp_hyst(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 	  therm->attr_get(therm, NOUVEAU_THERM_ATTR_THRS_SHUTDOWN_HYST) * 1000);
@@ -310,7 +306,7 @@ nouveau_hwmon_set_emergency_temp_hyst(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 
 	if (kstrtol(buf, 10, &value) == -EINVAL)
@@ -350,7 +346,7 @@ nouveau_hwmon_show_fan1_input(struct device *d, struct device_attribute *attr,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", therm->fan_sense(therm));
 }
@@ -363,7 +359,7 @@ nouveau_hwmon_get_pwm1_enable(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int ret;
 
 	ret = therm->attr_get(therm, NOUVEAU_THERM_ATTR_FAN_MODE);
@@ -379,7 +375,7 @@ nouveau_hwmon_set_pwm1_enable(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 	int ret;
 
@@ -402,7 +398,7 @@ nouveau_hwmon_get_pwm1(struct device *d, struct device_attribute *a, char *buf)
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int ret;
 
 	ret = therm->fan_get(therm);
@@ -418,7 +414,7 @@ nouveau_hwmon_set_pwm1(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int ret = -ENODEV;
 	long value;
 
@@ -442,7 +438,7 @@ nouveau_hwmon_get_pwm1_min(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int ret;
 
 	ret = therm->attr_get(therm, NOUVEAU_THERM_ATTR_FAN_MIN_DUTY);
@@ -458,7 +454,7 @@ nouveau_hwmon_set_pwm1_min(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 	int ret;
 
@@ -482,7 +478,7 @@ nouveau_hwmon_get_pwm1_max(struct device *d,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	int ret;
 
 	ret = therm->attr_get(therm, NOUVEAU_THERM_ATTR_FAN_MAX_DUTY);
@@ -498,7 +494,7 @@ nouveau_hwmon_set_pwm1_max(struct device *d, struct device_attribute *a,
 {
 	struct drm_device *dev = dev_get_drvdata(d);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	long value;
 	int ret;
 
@@ -565,7 +561,7 @@ nouveau_hwmon_init(struct drm_device *dev)
 {
 #if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE))
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_therm *therm = nouveau_therm(drm->device);
+	struct nouveau_therm *therm = nvkm_therm(&drm->device);
 	struct nouveau_hwmon *hwmon;
 	struct device *hwmon_dev;
 	int ret = 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_nvif.c b/drivers/gpu/drm/nouveau/nouveau_nvif.c
new file mode 100644
index 000000000000..6544b84f0303
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_nvif.c
@@ -0,0 +1,136 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+/*******************************************************************************
+ * NVIF client driver - NVKM directly linked
+ ******************************************************************************/
+
+#include <core/client.h>
+#include <core/notify.h>
+#include <core/ioctl.h>
+
+#include <nvif/client.h>
+#include <nvif/driver.h>
+#include <nvif/notify.h>
+#include <nvif/event.h>
+#include <nvif/ioctl.h>
+
+#include "nouveau_drm.h"
+#include "nouveau_usif.h"
+
+static void
+nvkm_client_unmap(void *priv, void __iomem *ptr, u32 size)
+{
+	iounmap(ptr);
+}
+
+static void __iomem *
+nvkm_client_map(void *priv, u64 handle, u32 size)
+{
+	return ioremap(handle, size);
+}
+
+static int
+nvkm_client_ioctl(void *priv, bool super, void *data, u32 size, void **hack)
+{
+	return nvkm_ioctl(priv, super, data, size, hack);
+}
+
+static int
+nvkm_client_resume(void *priv)
+{
+	return nouveau_client_init(priv);
+}
+
+static int
+nvkm_client_suspend(void *priv)
+{
+	return nouveau_client_fini(priv, true);
+}
+
+static void
+nvkm_client_fini(void *priv)
+{
+	struct nouveau_object *client = priv;
+	nouveau_client_fini(nv_client(client), false);
+	atomic_set(&client->refcount, 1);
+	nouveau_object_ref(NULL, &client);
+}
+
+static int
+nvkm_client_ntfy(const void *header, u32 length, const void *data, u32 size)
+{
+	const union {
+		struct nvif_notify_req_v0 v0;
+	} *args = header;
+	u8 route;
+
+	if (length == sizeof(args->v0) && args->v0.version == 0) {
+		route = args->v0.route;
+	} else {
+		WARN_ON(1);
+		return NVKM_NOTIFY_DROP;
+	}
+
+	switch (route) {
+	case NVDRM_NOTIFY_NVIF:
+		return nvif_notify(header, length, data, size);
+	case NVDRM_NOTIFY_USIF:
+		return usif_notify(header, length, data, size);
+	default:
+		WARN_ON(1);
+		break;
+	}
+
+	return NVKM_NOTIFY_DROP;
+}
+
+static int
+nvkm_client_init(const char *name, u64 device, const char *cfg,
+		 const char *dbg, void **ppriv)
+{
+	struct nouveau_client *client;
+	int ret;
+
+	ret = nouveau_client_create(name, device, cfg, dbg, &client);
+	*ppriv = client;
+	if (ret)
+		return ret;
+
+	client->ntfy = nvkm_client_ntfy;
+	return 0;
+}
+
+const struct nvif_driver
+nvif_driver_nvkm = {
+	.name = "nvkm",
+	.init = nvkm_client_init,
+	.fini = nvkm_client_fini,
+	.suspend = nvkm_client_suspend,
+	.resume = nvkm_client_resume,
+	.ioctl = nvkm_client_ioctl,
+	.map = nvkm_client_map,
+	.unmap = nvkm_client_unmap,
+	.keep = false,
+};
diff --git a/drivers/gpu/drm/nouveau/nouveau_platform.c b/drivers/gpu/drm/nouveau/nouveau_platform.c
new file mode 100644
index 000000000000..246a824c16ca
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_platform.c
@@ -0,0 +1,183 @@
+/*
+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/of.h>
+#include <linux/reset.h>
+#include <linux/regulator/consumer.h>
+#include <soc/tegra/pmc.h>
+
+#include "nouveau_drm.h"
+#include "nouveau_platform.h"
+
+static int nouveau_platform_power_up(struct nouveau_platform_gpu *gpu)
+{
+	int err;
+
+	err = regulator_enable(gpu->vdd);
+	if (err)
+		goto err_power;
+
+	err = clk_prepare_enable(gpu->clk);
+	if (err)
+		goto err_clk;
+	err = clk_prepare_enable(gpu->clk_pwr);
+	if (err)
+		goto err_clk_pwr;
+	clk_set_rate(gpu->clk_pwr, 204000000);
+	udelay(10);
+
+	reset_control_assert(gpu->rst);
+	udelay(10);
+
+	err = tegra_powergate_remove_clamping(TEGRA_POWERGATE_3D);
+	if (err)
+		goto err_clamp;
+	udelay(10);
+
+	reset_control_deassert(gpu->rst);
+	udelay(10);
+
+	return 0;
+
+err_clamp:
+	clk_disable_unprepare(gpu->clk_pwr);
+err_clk_pwr:
+	clk_disable_unprepare(gpu->clk);
+err_clk:
+	regulator_disable(gpu->vdd);
+err_power:
+	return err;
+}
+
+static int nouveau_platform_power_down(struct nouveau_platform_gpu *gpu)
+{
+	int err;
+
+	reset_control_assert(gpu->rst);
+	udelay(10);
+
+	clk_disable_unprepare(gpu->clk_pwr);
+	clk_disable_unprepare(gpu->clk);
+	udelay(10);
+
+	err = regulator_disable(gpu->vdd);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int nouveau_platform_probe(struct platform_device *pdev)
+{
+	struct nouveau_platform_gpu *gpu;
+	struct nouveau_platform_device *device;
+	struct drm_device *drm;
+	int err;
+
+	gpu = devm_kzalloc(&pdev->dev, sizeof(*gpu), GFP_KERNEL);
+	if (!gpu)
+		return -ENOMEM;
+
+	gpu->vdd = devm_regulator_get(&pdev->dev, "vdd");
+	if (IS_ERR(gpu->vdd))
+		return PTR_ERR(gpu->vdd);
+
+	gpu->rst = devm_reset_control_get(&pdev->dev, "gpu");
+	if (IS_ERR(gpu->rst))
+		return PTR_ERR(gpu->rst);
+
+	gpu->clk = devm_clk_get(&pdev->dev, "gpu");
+	if (IS_ERR(gpu->clk))
+		return PTR_ERR(gpu->clk);
+
+	gpu->clk_pwr = devm_clk_get(&pdev->dev, "pwr");
+	if (IS_ERR(gpu->clk_pwr))
+		return PTR_ERR(gpu->clk_pwr);
+
+	err = nouveau_platform_power_up(gpu);
+	if (err)
+		return err;
+
+	drm = nouveau_platform_device_create(pdev, &device);
+	if (IS_ERR(drm)) {
+		err = PTR_ERR(drm);
+		goto power_down;
+	}
+
+	device->gpu = gpu;
+
+	err = drm_dev_register(drm, 0);
+	if (err < 0)
+		goto err_unref;
+
+	return 0;
+
+err_unref:
+	drm_dev_unref(drm);
+
+	return 0;
+
+power_down:
+	nouveau_platform_power_down(gpu);
+
+	return err;
+}
+
+static int nouveau_platform_remove(struct platform_device *pdev)
+{
+	struct drm_device *drm_dev = platform_get_drvdata(pdev);
+	struct nouveau_drm *drm = nouveau_drm(drm_dev);
+	struct nouveau_device *device = nvkm_device(&drm->device);
+	struct nouveau_platform_gpu *gpu = nv_device_to_platform(device)->gpu;
+
+	nouveau_drm_device_remove(drm_dev);
+
+	return nouveau_platform_power_down(gpu);
+}
+
+#if IS_ENABLED(CONFIG_OF)
+static const struct of_device_id nouveau_platform_match[] = {
+	{ .compatible = "nvidia,gk20a" },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(of, nouveau_platform_match);
+#endif
+
+struct platform_driver nouveau_platform_driver = {
+	.driver = {
+		.name = "nouveau",
+		.of_match_table = of_match_ptr(nouveau_platform_match),
+	},
+	.probe = nouveau_platform_probe,
+	.remove = nouveau_platform_remove,
+};
+
+module_platform_driver(nouveau_platform_driver);
+
+MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_DESCRIPTION(DRIVER_DESC);
+MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/nouveau/nouveau_platform.h b/drivers/gpu/drm/nouveau/nouveau_platform.h
new file mode 100644
index 000000000000..91f66504900e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_platform.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __NOUVEAU_PLATFORM_H__
+#define __NOUVEAU_PLATFORM_H__
+
+#include "core/device.h"
+
+struct reset_control;
+struct clk;
+struct regulator;
+
+struct nouveau_platform_gpu {
+	struct reset_control *rst;
+	struct clk *clk;
+	struct clk *clk_pwr;
+
+	struct regulator *vdd;
+};
+
+struct nouveau_platform_device {
+	struct nouveau_device device;
+
+	struct nouveau_platform_gpu *gpu;
+};
+
+#define nv_device_to_platform(d)                                               \
+	container_of(d, struct nouveau_platform_device, device)
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index 1f51008e4d26..228226ab27fc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -23,6 +23,7 @@
  */
 
 #include <drm/drmP.h>
+#include <linux/dma-buf.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_gem.h"
@@ -56,17 +57,20 @@ void nouveau_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 }
 
 struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
-							 size_t size,
+							 struct dma_buf_attachment *attach,
 							 struct sg_table *sg)
 {
 	struct nouveau_bo *nvbo;
+	struct reservation_object *robj = attach->dmabuf->resv;
 	u32 flags = 0;
 	int ret;
 
 	flags = TTM_PL_FLAG_TT;
 
-	ret = nouveau_bo_new(dev, size, 0, flags, 0, 0,
-			     sg, &nvbo);
+	ww_mutex_lock(&robj->lock, NULL);
+	ret = nouveau_bo_new(dev, attach->dmabuf->size, 0, flags, 0, 0,
+			     sg, robj, &nvbo);
+	ww_mutex_unlock(&robj->lock);
 	if (ret)
 		return ERR_PTR(ret);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index a4d22e5eb176..01707e7deaf5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -1,8 +1,6 @@
 #include <linux/pagemap.h>
 #include <linux/slab.h>
 
-#include <subdev/fb.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_ttm.h"
 
@@ -104,7 +102,7 @@ nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev,
 		return NULL;
 
 	nvbe->dev = drm->dev;
-	if (nv_device(drm->device)->card_type < NV_50)
+	if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA)
 		nvbe->ttm.ttm.func = &nv04_sgdma_backend;
 	else
 		nvbe->ttm.ttm.func = &nv50_sgdma_backend;
diff --git a/drivers/gpu/drm/nouveau/nouveau_sysfs.c b/drivers/gpu/drm/nouveau/nouveau_sysfs.c
index 75dda2b07176..8fbbf3093d86 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sysfs.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sysfs.c
@@ -22,10 +22,15 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
+#include <nvif/os.h>
+#include <nvif/class.h>
+#include <nvif/ioctl.h>
+
 #include "nouveau_sysfs.h"
 
-#include <core/object.h>
-#include <core/class.h>
+MODULE_PARM_DESC(pstate, "enable sysfs pstate file, which will be moved in the future");
+int nouveau_pstate;
+module_param_named(pstate, nouveau_pstate, int, 0400);
 
 static inline struct drm_device *
 drm_device(struct device *d)
@@ -43,38 +48,42 @@ static ssize_t
 nouveau_sysfs_pstate_get(struct device *d, struct device_attribute *a, char *b)
 {
 	struct nouveau_sysfs *sysfs = nouveau_sysfs(drm_device(d));
-	struct nv_control_pstate_info info;
+	struct nvif_control_pstate_info_v0 info = {};
 	size_t cnt = PAGE_SIZE;
 	char *buf = b;
 	int ret, i;
 
-	ret = nv_exec(sysfs->ctrl, NV_CONTROL_PSTATE_INFO, &info, sizeof(info));
+	ret = nvif_mthd(&sysfs->ctrl, NVIF_CONTROL_PSTATE_INFO,
+			&info, sizeof(info));
 	if (ret)
 		return ret;
 
 	for (i = 0; i < info.count + 1; i++) {
 		const s32 state = i < info.count ? i :
-			NV_CONTROL_PSTATE_ATTR_STATE_CURRENT;
-		struct nv_control_pstate_attr attr = {
+			NVIF_CONTROL_PSTATE_ATTR_V0_STATE_CURRENT;
+		struct nvif_control_pstate_attr_v0 attr = {
 			.state = state,
 			.index = 0,
 		};
 
-		ret = nv_exec(sysfs->ctrl, NV_CONTROL_PSTATE_ATTR,
-			     &attr, sizeof(attr));
+		ret = nvif_mthd(&sysfs->ctrl, NVIF_CONTROL_PSTATE_ATTR,
+				&attr, sizeof(attr));
 		if (ret)
 			return ret;
 
 		if (i < info.count)
 			snappendf(buf, cnt, "%02x:", attr.state);
 		else
-			snappendf(buf, cnt, "--:");
+			snappendf(buf, cnt, "%s:", info.pwrsrc == 0 ? "DC" :
+						   info.pwrsrc == 1 ? "AC" :
+						   "--");
 
 		attr.index = 0;
 		do {
 			attr.state = state;
-			ret = nv_exec(sysfs->ctrl, NV_CONTROL_PSTATE_ATTR,
-				     &attr, sizeof(attr));
+			ret = nvif_mthd(&sysfs->ctrl,
+					NVIF_CONTROL_PSTATE_ATTR,
+					&attr, sizeof(attr));
 			if (ret)
 				return ret;
 
@@ -84,9 +93,20 @@ nouveau_sysfs_pstate_get(struct device *d, struct device_attribute *a, char *b)
 			snappendf(buf, cnt, " %s", attr.unit);
 		} while (attr.index);
 
-		if ((state >= 0 && info.pstate == state) ||
-		    (state <  0 && info.ustate < 0))
-			snappendf(buf, cnt, " *");
+		if (state >= 0) {
+			if (info.ustate_ac == state)
+				snappendf(buf, cnt, " AC");
+			if (info.ustate_dc == state)
+				snappendf(buf, cnt, " DC");
+			if (info.pstate == state)
+				snappendf(buf, cnt, " *");
+		} else {
+			if (info.ustate_ac < -1)
+				snappendf(buf, cnt, " AC");
+			if (info.ustate_dc < -1)
+				snappendf(buf, cnt, " DC");
+		}
+
 		snappendf(buf, cnt, "\n");
 	}
 
@@ -98,26 +118,36 @@ nouveau_sysfs_pstate_set(struct device *d, struct device_attribute *a,
 			 const char *buf, size_t count)
 {
 	struct nouveau_sysfs *sysfs = nouveau_sysfs(drm_device(d));
-	struct nv_control_pstate_user args;
+	struct nvif_control_pstate_user_v0 args = { .pwrsrc = -EINVAL };
 	long value, ret;
 	char *tmp;
 
 	if ((tmp = strchr(buf, '\n')))
 		*tmp = '\0';
 
+	if (!strncasecmp(buf, "dc:", 3)) {
+		args.pwrsrc = 0;
+		buf += 3;
+	} else
+	if (!strncasecmp(buf, "ac:", 3)) {
+		args.pwrsrc = 1;
+		buf += 3;
+	}
+
 	if (!strcasecmp(buf, "none"))
-		args.state = NV_CONTROL_PSTATE_USER_STATE_UNKNOWN;
+		args.ustate = NVIF_CONTROL_PSTATE_USER_V0_STATE_UNKNOWN;
 	else
 	if (!strcasecmp(buf, "auto"))
-		args.state = NV_CONTROL_PSTATE_USER_STATE_PERFMON;
+		args.ustate = NVIF_CONTROL_PSTATE_USER_V0_STATE_PERFMON;
 	else {
 		ret = kstrtol(buf, 16, &value);
 		if (ret)
 			return ret;
-		args.state = value;
+		args.ustate = value;
 	}
 
-	ret = nv_exec(sysfs->ctrl, NV_CONTROL_PSTATE_USER, &args, sizeof(args));
+	ret = nvif_mthd(&sysfs->ctrl, NVIF_CONTROL_PSTATE_USER,
+			&args, sizeof(args));
 	if (ret < 0)
 		return ret;
 
@@ -132,11 +162,11 @@ nouveau_sysfs_fini(struct drm_device *dev)
 {
 	struct nouveau_sysfs *sysfs = nouveau_sysfs(dev);
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 
-	if (sysfs->ctrl) {
-		device_remove_file(nv_device_base(device), &dev_attr_pstate);
-		nouveau_object_del(nv_object(drm), NVDRM_DEVICE, NVDRM_CONTROL);
+	if (sysfs && sysfs->ctrl.priv) {
+		device_remove_file(nv_device_base(nvkm_device(device)), &dev_attr_pstate);
+		nvif_object_fini(&sysfs->ctrl);
 	}
 
 	drm->sysfs = NULL;
@@ -147,18 +177,22 @@ int
 nouveau_sysfs_init(struct drm_device *dev)
 {
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_device *device = nv_device(drm->device);
+	struct nvif_device *device = &drm->device;
 	struct nouveau_sysfs *sysfs;
 	int ret;
 
+	if (!nouveau_pstate)
+		return 0;
+
 	sysfs = drm->sysfs = kzalloc(sizeof(*sysfs), GFP_KERNEL);
 	if (!sysfs)
 		return -ENOMEM;
 
-	ret = nouveau_object_new(nv_object(drm), NVDRM_DEVICE, NVDRM_CONTROL,
-				 NV_CONTROL_CLASS, NULL, 0, &sysfs->ctrl);
+	ret = nvif_object_init(nvif_object(device), NULL, NVDRM_CONTROL,
+			       NVIF_IOCTL_NEW_V0_CONTROL, NULL, 0,
+			      &sysfs->ctrl);
 	if (ret == 0)
-		device_create_file(nv_device_base(device), &dev_attr_pstate);
+		device_create_file(nv_device_base(nvkm_device(device)), &dev_attr_pstate);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_sysfs.h b/drivers/gpu/drm/nouveau/nouveau_sysfs.h
index 74b47f1e01ed..4e5ea9241b28 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sysfs.h
+++ b/drivers/gpu/drm/nouveau/nouveau_sysfs.h
@@ -4,7 +4,7 @@
 #include "nouveau_drm.h"
 
 struct nouveau_sysfs {
-	struct nouveau_object *ctrl;
+	struct nvif_object ctrl;
 };
 
 static inline struct nouveau_sysfs *
@@ -16,4 +16,6 @@ nouveau_sysfs(struct drm_device *dev)
 int  nouveau_sysfs_init(struct drm_device *);
 void nouveau_sysfs_fini(struct drm_device *);
 
+extern int nouveau_pstate;
+
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index ab0228f640a5..753a6def61e7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -24,10 +24,6 @@
  * USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <subdev/fb.h>
-#include <subdev/vm.h>
-#include <subdev/instmem.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_ttm.h"
 #include "nouveau_gem.h"
@@ -36,7 +32,7 @@ static int
 nouveau_vram_manager_init(struct ttm_mem_type_manager *man, unsigned long psize)
 {
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	man->priv = pfb;
 	return 0;
 }
@@ -67,7 +63,7 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 			 struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	nouveau_mem_node_cleanup(mem->mm_node);
 	pfb->ram->put(pfb, (struct nouveau_mem **)&mem->mm_node);
 }
@@ -75,11 +71,11 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 static int
 nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_buffer_object *bo,
-			 struct ttm_placement *placement,
+			 const struct ttm_place *place,
 			 struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
-	struct nouveau_fb *pfb = nouveau_fb(drm->device);
+	struct nouveau_fb *pfb = nvkm_fb(&drm->device);
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nouveau_mem *node;
 	u32 size_nc = 0;
@@ -161,7 +157,7 @@ nouveau_gart_manager_del(struct ttm_mem_type_manager *man,
 static int
 nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_buffer_object *bo,
-			 struct ttm_placement *placement,
+			 const struct ttm_place *place,
 			 struct ttm_mem_reg *mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
@@ -174,14 +170,13 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 
 	node->page_shift = 12;
 
-	switch (nv_device(drm->device)->card_type) {
-	case NV_50:
-		if (nv_device(drm->device)->chipset != 0x50)
+	switch (drm->device.info.family) {
+	case NV_DEVICE_INFO_V0_TESLA:
+		if (drm->device.info.chipset != 0x50)
 			node->memtype = (nvbo->tile_flags & 0x7f00) >> 8;
 		break;
-	case NV_C0:
-	case NV_D0:
-	case NV_E0:
+	case NV_DEVICE_INFO_V0_FERMI:
+	case NV_DEVICE_INFO_V0_KEPLER:
 		node->memtype = (nvbo->tile_flags & 0xff00) >> 8;
 		break;
 	default:
@@ -206,12 +201,13 @@ const struct ttm_mem_type_manager_func nouveau_gart_manager = {
 	nouveau_gart_manager_debug
 };
 
+/*XXX*/
 #include <core/subdev/vm/nv04.h>
 static int
 nv04_gart_manager_init(struct ttm_mem_type_manager *man, unsigned long psize)
 {
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
-	struct nouveau_vmmgr *vmm = nouveau_vmmgr(drm->device);
+	struct nouveau_vmmgr *vmm = nvkm_vmmgr(&drm->device);
 	struct nv04_vmmgr_priv *priv = (void *)vmm;
 	struct nouveau_vm *vm = NULL;
 	nouveau_vm_ref(priv->vm, &vm, NULL);
@@ -241,7 +237,7 @@ nv04_gart_manager_del(struct ttm_mem_type_manager *man, struct ttm_mem_reg *mem)
 static int
 nv04_gart_manager_new(struct ttm_mem_type_manager *man,
 		      struct ttm_buffer_object *bo,
-		      struct ttm_placement *placement,
+		      const struct ttm_place *place,
 		      struct ttm_mem_reg *mem)
 {
 	struct nouveau_mem *node;
@@ -285,7 +281,7 @@ nouveau_ttm_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct nouveau_drm *drm = nouveau_drm(file_priv->minor->dev);
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET))
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 
 	return ttm_bo_mmap(filp, vma, &drm->ttm.bdev);
 }
@@ -354,12 +350,11 @@ int
 nouveau_ttm_init(struct nouveau_drm *drm)
 {
 	struct drm_device *dev = drm->dev;
-	struct nouveau_device *device = nv_device(drm->device);
 	u32 bits;
 	int ret;
 
-	bits = nouveau_vmmgr(drm->device)->dma_bits;
-	if (nv_device_is_pci(device)) {
+	bits = nvkm_vmmgr(&drm->device)->dma_bits;
+	if (nv_device_is_pci(nvkm_device(&drm->device))) {
 		if (drm->agp.stat == ENABLED ||
 		     !pci_dma_supported(dev->pdev, DMA_BIT_MASK(bits)))
 			bits = 32;
@@ -391,8 +386,7 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 	}
 
 	/* VRAM init */
-	drm->gem.vram_available  = nouveau_fb(drm->device)->ram->size;
-	drm->gem.vram_available -= nouveau_instmem(drm->device)->reserved;
+	drm->gem.vram_available = drm->device.info.ram_user;
 
 	ret = ttm_bo_init_mm(&drm->ttm.bdev, TTM_PL_VRAM,
 			      drm->gem.vram_available >> PAGE_SHIFT);
@@ -401,12 +395,12 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 		return ret;
 	}
 
-	drm->ttm.mtrr = arch_phys_wc_add(nv_device_resource_start(device, 1),
-					 nv_device_resource_len(device, 1));
+	drm->ttm.mtrr = arch_phys_wc_add(nv_device_resource_start(nvkm_device(&drm->device), 1),
+					 nv_device_resource_len(nvkm_device(&drm->device), 1));
 
 	/* GART init */
 	if (drm->agp.stat != ENABLED) {
-		drm->gem.gart_available = nouveau_vmmgr(drm->device)->limit;
+		drm->gem.gart_available = nvkm_vmmgr(&drm->device)->limit;
 	} else {
 		drm->gem.gart_available = drm->agp.size;
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_usif.c b/drivers/gpu/drm/nouveau/nouveau_usif.c
new file mode 100644
index 000000000000..cb1182d7e80e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_usif.c
@@ -0,0 +1,384 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nouveau_drm.h"
+#include "nouveau_usif.h"
+
+#include <nvif/notify.h>
+#include <nvif/unpack.h>
+#include <nvif/client.h>
+#include <nvif/event.h>
+#include <nvif/ioctl.h>
+
+struct usif_notify_p {
+	struct drm_pending_event base;
+	struct {
+		struct drm_event base;
+		u8 data[];
+	} e;
+};
+
+struct usif_notify {
+	struct list_head head;
+	atomic_t enabled;
+	u32 handle;
+	u16 reply;
+	u8  route;
+	u64 token;
+	struct usif_notify_p *p;
+};
+
+static inline struct usif_notify *
+usif_notify_find(struct drm_file *filp, u32 handle)
+{
+	struct nouveau_cli *cli = nouveau_cli(filp);
+	struct usif_notify *ntfy;
+	list_for_each_entry(ntfy, &cli->notifys, head) {
+		if (ntfy->handle == handle)
+			return ntfy;
+	}
+	return NULL;
+}
+
+static inline void
+usif_notify_dtor(struct usif_notify *ntfy)
+{
+	list_del(&ntfy->head);
+	kfree(ntfy);
+}
+
+int
+usif_notify(const void *header, u32 length, const void *data, u32 size)
+{
+	struct usif_notify *ntfy = NULL;
+	const union {
+		struct nvif_notify_rep_v0 v0;
+	} *rep = header;
+	struct drm_device *dev;
+	struct drm_file *filp;
+	unsigned long flags;
+
+	if (length == sizeof(rep->v0) && rep->v0.version == 0) {
+		if (WARN_ON(!(ntfy = (void *)(unsigned long)rep->v0.token)))
+			return NVIF_NOTIFY_DROP;
+		BUG_ON(rep->v0.route != NVDRM_NOTIFY_USIF);
+	} else
+	if (WARN_ON(1))
+		return NVIF_NOTIFY_DROP;
+
+	if (WARN_ON(!ntfy->p || ntfy->reply != (length + size)))
+		return NVIF_NOTIFY_DROP;
+	filp = ntfy->p->base.file_priv;
+	dev = filp->minor->dev;
+
+	memcpy(&ntfy->p->e.data[0], header, length);
+	memcpy(&ntfy->p->e.data[length], data, size);
+	switch (rep->v0.version) {
+	case 0: {
+		struct nvif_notify_rep_v0 *rep = (void *)ntfy->p->e.data;
+		rep->route = ntfy->route;
+		rep->token = ntfy->token;
+	}
+		break;
+	default:
+		BUG_ON(1);
+		break;
+	}
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (!WARN_ON(filp->event_space < ntfy->p->e.base.length)) {
+		list_add_tail(&ntfy->p->base.link, &filp->event_list);
+		filp->event_space -= ntfy->p->e.base.length;
+	}
+	wake_up_interruptible(&filp->event_wait);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+	atomic_set(&ntfy->enabled, 0);
+	return NVIF_NOTIFY_DROP;
+}
+
+static int
+usif_notify_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(f);
+	struct nvif_client *client = &cli->base;
+	union {
+		struct nvif_ioctl_ntfy_new_v0 v0;
+	} *args = data;
+	union {
+		struct nvif_notify_req_v0 v0;
+	} *req;
+	struct usif_notify *ntfy;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		if (usif_notify_find(f, args->v0.index))
+			return -EEXIST;
+	} else
+		return ret;
+	req = data;
+
+	if (!(ntfy = kmalloc(sizeof(*ntfy), GFP_KERNEL)))
+		return -ENOMEM;
+	atomic_set(&ntfy->enabled, 0);
+
+	if (nvif_unpack(req->v0, 0, 0, true)) {
+		ntfy->reply = sizeof(struct nvif_notify_rep_v0) + req->v0.reply;
+		ntfy->route = req->v0.route;
+		ntfy->token = req->v0.token;
+		req->v0.route = NVDRM_NOTIFY_USIF;
+		req->v0.token = (unsigned long)(void *)ntfy;
+		ret = nvif_client_ioctl(client, argv, argc);
+		req->v0.token = ntfy->token;
+		req->v0.route = ntfy->route;
+		ntfy->handle = args->v0.index;
+	}
+
+	if (ret == 0)
+		list_add(&ntfy->head, &cli->notifys);
+	if (ret)
+		kfree(ntfy);
+	return ret;
+}
+
+static int
+usif_notify_del(struct drm_file *f, void *data, u32 size, void *argv, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(f);
+	struct nvif_client *client = &cli->base;
+	union {
+		struct nvif_ioctl_ntfy_del_v0 v0;
+	} *args = data;
+	struct usif_notify *ntfy;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		if (!(ntfy = usif_notify_find(f, args->v0.index)))
+			return -ENOENT;
+	} else
+		return ret;
+
+	ret = nvif_client_ioctl(client, argv, argc);
+	if (ret == 0)
+		usif_notify_dtor(ntfy);
+	return ret;
+}
+
+static int
+usif_notify_get(struct drm_file *f, void *data, u32 size, void *argv, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(f);
+	struct nvif_client *client = &cli->base;
+	union {
+		struct nvif_ioctl_ntfy_del_v0 v0;
+	} *args = data;
+	struct usif_notify *ntfy;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		if (!(ntfy = usif_notify_find(f, args->v0.index)))
+			return -ENOENT;
+	} else
+		return ret;
+
+	if (atomic_xchg(&ntfy->enabled, 1))
+		return 0;
+
+	ntfy->p = kmalloc(sizeof(*ntfy->p) + ntfy->reply, GFP_KERNEL);
+	if (ret = -ENOMEM, !ntfy->p)
+		goto done;
+	ntfy->p->base.event = &ntfy->p->e.base;
+	ntfy->p->base.file_priv = f;
+	ntfy->p->base.pid = current->pid;
+	ntfy->p->base.destroy =(void(*)(struct drm_pending_event *))kfree;
+	ntfy->p->e.base.type = DRM_NOUVEAU_EVENT_NVIF;
+	ntfy->p->e.base.length = sizeof(ntfy->p->e.base) + ntfy->reply;
+
+	ret = nvif_client_ioctl(client, argv, argc);
+done:
+	if (ret) {
+		atomic_set(&ntfy->enabled, 0);
+		kfree(ntfy->p);
+	}
+	return ret;
+}
+
+static int
+usif_notify_put(struct drm_file *f, void *data, u32 size, void *argv, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(f);
+	struct nvif_client *client = &cli->base;
+	union {
+		struct nvif_ioctl_ntfy_put_v0 v0;
+	} *args = data;
+	struct usif_notify *ntfy;
+	int ret;
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		if (!(ntfy = usif_notify_find(f, args->v0.index)))
+			return -ENOENT;
+	} else
+		return ret;
+
+	ret = nvif_client_ioctl(client, argv, argc);
+	if (ret == 0 && atomic_xchg(&ntfy->enabled, 0))
+		kfree(ntfy->p);
+	return ret;
+}
+
+struct usif_object {
+	struct list_head head;
+	struct list_head ntfy;
+	u8  route;
+	u64 token;
+};
+
+static void
+usif_object_dtor(struct usif_object *object)
+{
+	list_del(&object->head);
+	kfree(object);
+}
+
+static int
+usif_object_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(f);
+	struct nvif_client *client = &cli->base;
+	union {
+		struct nvif_ioctl_new_v0 v0;
+	} *args = data;
+	struct usif_object *object;
+	int ret;
+
+	if (!(object = kmalloc(sizeof(*object), GFP_KERNEL)))
+		return -ENOMEM;
+	list_add(&object->head, &cli->objects);
+
+	if (nvif_unpack(args->v0, 0, 0, true)) {
+		object->route = args->v0.route;
+		object->token = args->v0.token;
+		args->v0.route = NVDRM_OBJECT_USIF;
+		args->v0.token = (unsigned long)(void *)object;
+		ret = nvif_client_ioctl(client, argv, argc);
+		args->v0.token = object->token;
+		args->v0.route = object->route;
+	}
+
+	if (ret)
+		usif_object_dtor(object);
+	return ret;
+}
+
+int
+usif_ioctl(struct drm_file *filp, void __user *user, u32 argc)
+{
+	struct nouveau_cli *cli = nouveau_cli(filp);
+	struct nvif_client *client = &cli->base;
+	void *data = kmalloc(argc, GFP_KERNEL);
+	u32   size = argc;
+	union {
+		struct nvif_ioctl_v0 v0;
+	} *argv = data;
+	struct usif_object *object;
+	u8 owner;
+	int ret;
+
+	if (ret = -ENOMEM, !argv)
+		goto done;
+	if (ret = -EFAULT, copy_from_user(argv, user, size))
+		goto done;
+
+	if (nvif_unpack(argv->v0, 0, 0, true)) {
+		/* block access to objects not created via this interface */
+		owner = argv->v0.owner;
+		argv->v0.owner = NVDRM_OBJECT_USIF;
+	} else
+		goto done;
+
+	mutex_lock(&cli->mutex);
+	switch (argv->v0.type) {
+	case NVIF_IOCTL_V0_NEW:
+		/* ... except if we're creating children */
+		argv->v0.owner = NVIF_IOCTL_V0_OWNER_ANY;
+		ret = usif_object_new(filp, data, size, argv, argc);
+		break;
+	case NVIF_IOCTL_V0_NTFY_NEW:
+		ret = usif_notify_new(filp, data, size, argv, argc);
+		break;
+	case NVIF_IOCTL_V0_NTFY_DEL:
+		ret = usif_notify_del(filp, data, size, argv, argc);
+		break;
+	case NVIF_IOCTL_V0_NTFY_GET:
+		ret = usif_notify_get(filp, data, size, argv, argc);
+		break;
+	case NVIF_IOCTL_V0_NTFY_PUT:
+		ret = usif_notify_put(filp, data, size, argv, argc);
+		break;
+	default:
+		ret = nvif_client_ioctl(client, argv, argc);
+		break;
+	}
+	if (argv->v0.route == NVDRM_OBJECT_USIF) {
+		object = (void *)(unsigned long)argv->v0.token;
+		argv->v0.route = object->route;
+		argv->v0.token = object->token;
+		if (ret == 0 && argv->v0.type == NVIF_IOCTL_V0_DEL) {
+			list_del(&object->head);
+			kfree(object);
+		}
+	} else {
+		argv->v0.route = NVIF_IOCTL_V0_ROUTE_HIDDEN;
+		argv->v0.token = 0;
+	}
+	argv->v0.owner = owner;
+	mutex_unlock(&cli->mutex);
+
+	if (copy_to_user(user, argv, argc))
+		ret = -EFAULT;
+done:
+	kfree(argv);
+	return ret;
+}
+
+void
+usif_client_fini(struct nouveau_cli *cli)
+{
+	struct usif_object *object, *otemp;
+	struct usif_notify *notify, *ntemp;
+
+	list_for_each_entry_safe(notify, ntemp, &cli->notifys, head) {
+		usif_notify_dtor(notify);
+	}
+
+	list_for_each_entry_safe(object, otemp, &cli->objects, head) {
+		usif_object_dtor(object);
+	}
+}
+
+void
+usif_client_init(struct nouveau_cli *cli)
+{
+	INIT_LIST_HEAD(&cli->objects);
+	INIT_LIST_HEAD(&cli->notifys);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_usif.h b/drivers/gpu/drm/nouveau/nouveau_usif.h
new file mode 100644
index 000000000000..c037e3ae8c70
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_usif.h
@@ -0,0 +1,9 @@
+#ifndef __NOUVEAU_USIF_H__
+#define __NOUVEAU_USIF_H__
+
+void usif_client_init(struct nouveau_cli *);
+void usif_client_fini(struct nouveau_cli *);
+int  usif_ioctl(struct drm_file *, void __user *, u32);
+int  usif_notify(const void *, u32, const void *, u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_vga.c b/drivers/gpu/drm/nouveau/nouveau_vga.c
index 4f4c3fec6916..c7592ec8ecb8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_vga.c
+++ b/drivers/gpu/drm/nouveau/nouveau_vga.c
@@ -12,14 +12,16 @@
 static unsigned int
 nouveau_vga_set_decode(void *priv, bool state)
 {
-	struct nouveau_device *device = nouveau_dev(priv);
+	struct nvif_device *device = &nouveau_drm(priv)->device;
 
-	if (device->card_type == NV_40 && device->chipset >= 0x4c)
-		nv_wr32(device, 0x088060, state);
-	else if (device->chipset >= 0x40)
-		nv_wr32(device, 0x088054, state);
+	if (device->info.family == NV_DEVICE_INFO_V0_CURIE &&
+	    device->info.chipset >= 0x4c)
+		nvif_wr32(device, 0x088060, state);
 	else
-		nv_wr32(device, 0x001854, state);
+	if (device->info.chipset >= 0x40)
+		nvif_wr32(device, 0x088054, state);
+	else
+		nvif_wr32(device, 0x001854, state);
 
 	if (state)
 		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
@@ -106,7 +108,16 @@ void
 nouveau_vga_fini(struct nouveau_drm *drm)
 {
 	struct drm_device *dev = drm->dev;
+	bool runtime = false;
+
+	if (nouveau_runtime_pm == 1)
+		runtime = true;
+	if ((nouveau_runtime_pm == -1) && (nouveau_is_optimus() || nouveau_is_v1_dsm()))
+		runtime = true;
+
 	vga_switcheroo_unregister_client(dev->pdev);
+	if (runtime && nouveau_is_v1_dsm() && !nouveau_is_optimus())
+		vga_switcheroo_fini_domain_pm_ops(drm->dev->dev);
 	vga_client_register(dev->pdev, NULL, NULL, NULL);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv04_fbcon.c b/drivers/gpu/drm/nouveau/nv04_fbcon.c
index 8fe32bbed99a..4ef602c5469d 100644
--- a/drivers/gpu/drm/nouveau/nv04_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv04_fbcon.c
@@ -22,8 +22,6 @@
  * DEALINGS IN THE SOFTWARE.
  */
 
-#include <core/object.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fbcon.h"
@@ -141,8 +139,7 @@ nv04_fbcon_accel_init(struct fb_info *info)
 	struct drm_device *dev = nfbdev->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_channel *chan = drm->channel;
-	struct nouveau_device *device = nv_device(drm->device);
-	struct nouveau_object *object;
+	struct nvif_device *device = &drm->device;
 	int surface_fmt, pattern_fmt, rect_fmt;
 	int ret;
 
@@ -174,35 +171,35 @@ nv04_fbcon_accel_init(struct fb_info *info)
 		return -EINVAL;
 	}
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvCtxSurf2D,
-				 device->card_type >= NV_10 ? 0x0062 : 0x0042,
-				 NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x0062,
+			       device->info.family >= NV_DEVICE_INFO_V0_CELSIUS ?
+			       0x0062 : 0x0042, NULL, 0, &nfbdev->surf2d);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvClipRect,
-				 0x0019, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x0019, 0x0019, NULL, 0,
+			       &nfbdev->clip);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvRop,
-				 0x0043, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x0043, 0x0043, NULL, 0,
+			       &nfbdev->rop);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvImagePatt,
-				 0x0044, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x0044, 0x0044, NULL, 0,
+			       &nfbdev->patt);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvGdiRect,
-				 0x004a, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x004a, 0x004a, NULL, 0,
+			       &nfbdev->gdi);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, NvImageBlit,
-				 device->chipset >= 0x11 ? 0x009f : 0x005f,
-				 NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x005f,
+			       device->info.chipset >= 0x11 ? 0x009f : 0x005f,
+			       NULL, 0, &nfbdev->blit);
 	if (ret)
 		return ret;
 
@@ -212,10 +209,10 @@ nv04_fbcon_accel_init(struct fb_info *info)
 	}
 
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0000, 1);
-	OUT_RING(chan, NvCtxSurf2D);
+	OUT_RING(chan, nfbdev->surf2d.handle);
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0184, 2);
-	OUT_RING(chan, NvDmaFB);
-	OUT_RING(chan, NvDmaFB);
+	OUT_RING(chan, chan->vram.handle);
+	OUT_RING(chan, chan->vram.handle);
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0300, 4);
 	OUT_RING(chan, surface_fmt);
 	OUT_RING(chan, info->fix.line_length | (info->fix.line_length << 16));
@@ -223,12 +220,12 @@ nv04_fbcon_accel_init(struct fb_info *info)
 	OUT_RING(chan, info->fix.smem_start - dev->mode_config.fb_base);
 
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0000, 1);
-	OUT_RING(chan, NvRop);
+	OUT_RING(chan, nfbdev->rop.handle);
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0300, 1);
 	OUT_RING(chan, 0x55);
 
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0000, 1);
-	OUT_RING(chan, NvImagePatt);
+	OUT_RING(chan, nfbdev->patt.handle);
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0300, 8);
 	OUT_RING(chan, pattern_fmt);
 #ifdef __BIG_ENDIAN
@@ -244,18 +241,18 @@ nv04_fbcon_accel_init(struct fb_info *info)
 	OUT_RING(chan, ~0);
 
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0000, 1);
-	OUT_RING(chan, NvClipRect);
+	OUT_RING(chan, nfbdev->clip.handle);
 	BEGIN_NV04(chan, NvSubCtxSurf2D, 0x0300, 2);
 	OUT_RING(chan, 0);
 	OUT_RING(chan, (info->var.yres_virtual << 16) | info->var.xres_virtual);
 
 	BEGIN_NV04(chan, NvSubImageBlit, 0x0000, 1);
-	OUT_RING(chan, NvImageBlit);
+	OUT_RING(chan, nfbdev->blit.handle);
 	BEGIN_NV04(chan, NvSubImageBlit, 0x019c, 1);
-	OUT_RING(chan, NvCtxSurf2D);
+	OUT_RING(chan, nfbdev->surf2d.handle);
 	BEGIN_NV04(chan, NvSubImageBlit, 0x02fc, 1);
 	OUT_RING(chan, 3);
-	if (device->chipset >= 0x11 /*XXX: oclass == 0x009f*/) {
+	if (device->info.chipset >= 0x11 /*XXX: oclass == 0x009f*/) {
 		BEGIN_NV04(chan, NvSubImageBlit, 0x0120, 3);
 		OUT_RING(chan, 0);
 		OUT_RING(chan, 1);
@@ -263,12 +260,12 @@ nv04_fbcon_accel_init(struct fb_info *info)
 	}
 
 	BEGIN_NV04(chan, NvSubGdiRect, 0x0000, 1);
-	OUT_RING(chan, NvGdiRect);
+	OUT_RING(chan, nfbdev->gdi.handle);
 	BEGIN_NV04(chan, NvSubGdiRect, 0x0198, 1);
-	OUT_RING(chan, NvCtxSurf2D);
+	OUT_RING(chan, nfbdev->surf2d.handle);
 	BEGIN_NV04(chan, NvSubGdiRect, 0x0188, 2);
-	OUT_RING(chan, NvImagePatt);
-	OUT_RING(chan, NvRop);
+	OUT_RING(chan, nfbdev->patt.handle);
+	OUT_RING(chan, nfbdev->rop.handle);
 	BEGIN_NV04(chan, NvSubGdiRect, 0x0304, 1);
 	OUT_RING(chan, 1);
 	BEGIN_NV04(chan, NvSubGdiRect, 0x0300, 1);
diff --git a/drivers/gpu/drm/nouveau/nv04_fence.c b/drivers/gpu/drm/nouveau/nv04_fence.c
index 94eadd1dd10a..f9859deb108a 100644
--- a/drivers/gpu/drm/nouveau/nv04_fence.c
+++ b/drivers/gpu/drm/nouveau/nv04_fence.c
@@ -22,8 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <engine/fifo.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
@@ -43,7 +41,7 @@ nv04_fence_emit(struct nouveau_fence *fence)
 	int ret = RING_SPACE(chan, 2);
 	if (ret == 0) {
 		BEGIN_NV04(chan, NvSubSw, 0x0150, 1);
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, fence->base.seqno);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -59,7 +57,7 @@ nv04_fence_sync(struct nouveau_fence *fence,
 static u32
 nv04_fence_read(struct nouveau_channel *chan)
 {
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
+	struct nouveau_fifo_chan *fifo = nvkm_fifo_chan(chan);;
 	return atomic_read(&fifo->refcnt);
 }
 
@@ -69,7 +67,7 @@ nv04_fence_context_del(struct nouveau_channel *chan)
 	struct nv04_fence_chan *fctx = chan->fence;
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 static int
@@ -77,7 +75,7 @@ nv04_fence_context_new(struct nouveau_channel *chan)
 {
 	struct nv04_fence_chan *fctx = kzalloc(sizeof(*fctx), GFP_KERNEL);
 	if (fctx) {
-		nouveau_fence_context_new(&fctx->base);
+		nouveau_fence_context_new(chan, &fctx->base);
 		fctx->base.emit = nv04_fence_emit;
 		fctx->base.sync = nv04_fence_sync;
 		fctx->base.read = nv04_fence_read;
@@ -107,5 +105,7 @@ nv04_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv04_fence_destroy;
 	priv->base.context_new = nv04_fence_context_new;
 	priv->base.context_del = nv04_fence_context_del;
+	priv->base.contexts = 15;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.c b/drivers/gpu/drm/nouveau/nv10_fence.c
index 06f434f03fba..5e1ea1cdce75 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.c
+++ b/drivers/gpu/drm/nouveau/nv10_fence.c
@@ -22,9 +22,6 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-#include <core/object.h>
-#include <core/class.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nv10_fence.h"
@@ -36,7 +33,7 @@ nv10_fence_emit(struct nouveau_fence *fence)
 	int ret = RING_SPACE(chan, 2);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV10_SUBCHAN_REF_CNT, 1);
-		OUT_RING  (chan, fence->sequence);
+		OUT_RING  (chan, fence->base.seqno);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -53,16 +50,20 @@ nv10_fence_sync(struct nouveau_fence *fence,
 u32
 nv10_fence_read(struct nouveau_channel *chan)
 {
-	return nv_ro32(chan->object, 0x0048);
+	return nvif_rd32(chan, 0x0048);
 }
 
 void
 nv10_fence_context_del(struct nouveau_channel *chan)
 {
 	struct nv10_fence_chan *fctx = chan->fence;
+	int i;
 	nouveau_fence_context_del(&fctx->base);
+	for (i = 0; i < ARRAY_SIZE(fctx->head); i++)
+		nvif_object_fini(&fctx->head[i]);
+	nvif_object_fini(&fctx->sema);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 int
@@ -74,7 +75,7 @@ nv10_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv10_fence_sync;
@@ -105,6 +106,8 @@ nv10_fence_create(struct nouveau_drm *drm)
 	priv->base.dtor = nv10_fence_destroy;
 	priv->base.context_new = nv10_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 31;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_fence.h b/drivers/gpu/drm/nouveau/nv10_fence.h
index e5d9204826c2..a87259f3983a 100644
--- a/drivers/gpu/drm/nouveau/nv10_fence.h
+++ b/drivers/gpu/drm/nouveau/nv10_fence.h
@@ -1,12 +1,13 @@
 #ifndef __NV10_FENCE_H_
 #define __NV10_FENCE_H_
 
-#include <core/os.h>
 #include "nouveau_fence.h"
 #include "nouveau_bo.h"
 
 struct nv10_fence_chan {
 	struct nouveau_fence_chan base;
+	struct nvif_object sema;
+	struct nvif_object head[4];
 };
 
 struct nv10_fence_priv {
diff --git a/drivers/gpu/drm/nouveau/nv17_fence.c b/drivers/gpu/drm/nouveau/nv17_fence.c
index 22aa9963ea6f..40b461c7d5c5 100644
--- a/drivers/gpu/drm/nouveau/nv17_fence.c
+++ b/drivers/gpu/drm/nouveau/nv17_fence.c
@@ -22,8 +22,8 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-#include <core/object.h>
-#include <core/class.h>
+#include <nvif/os.h>
+#include <nvif/class.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
@@ -33,11 +33,13 @@ int
 nv17_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
+	struct nouveau_cli *cli = (void *)nvif_client(&prev->device->base);
 	struct nv10_fence_priv *priv = chan->drm->fence;
+	struct nv10_fence_chan *fctx = chan->fence;
 	u32 value;
 	int ret;
 
-	if (!mutex_trylock(&prev->cli->mutex))
+	if (!mutex_trylock(&cli->mutex))
 		return -EBUSY;
 
 	spin_lock(&priv->lock);
@@ -48,7 +50,7 @@ nv17_fence_sync(struct nouveau_fence *fence,
 	ret = RING_SPACE(prev, 5);
 	if (!ret) {
 		BEGIN_NV04(prev, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
-		OUT_RING  (prev, NvSema);
+		OUT_RING  (prev, fctx->sema.handle);
 		OUT_RING  (prev, 0);
 		OUT_RING  (prev, value + 0);
 		OUT_RING  (prev, value + 1);
@@ -57,14 +59,14 @@ nv17_fence_sync(struct nouveau_fence *fence,
 
 	if (!ret && !(ret = RING_SPACE(chan, 5))) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 4);
-		OUT_RING  (chan, NvSema);
+		OUT_RING  (chan, fctx->sema.handle);
 		OUT_RING  (chan, 0);
 		OUT_RING  (chan, value + 1);
 		OUT_RING  (chan, value + 2);
 		FIRE_RING (chan);
 	}
 
-	mutex_unlock(&prev->cli->mutex);
+	mutex_unlock(&cli->mutex);
 	return 0;
 }
 
@@ -74,7 +76,6 @@ nv17_fence_context_new(struct nouveau_channel *chan)
 	struct nv10_fence_priv *priv = chan->drm->fence;
 	struct nv10_fence_chan *fctx;
 	struct ttm_mem_reg *mem = &priv->bo->bo.mem;
-	struct nouveau_object *object;
 	u32 start = mem->start * PAGE_SIZE;
 	u32 limit = start + mem->size - 1;
 	int ret = 0;
@@ -83,20 +84,19 @@ nv17_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv17_fence_sync;
 
-	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-				 NvSema, 0x0002,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
+	ret = nvif_object_init(chan->object, NULL, NvSema, NV_DMA_FROM_MEMORY,
+			       &(struct nv_dma_v0) {
+					.target = NV_DMA_V0_TARGET_VRAM,
+					.access = NV_DMA_V0_ACCESS_RDWR,
 					.start = start,
 					.limit = limit,
-				 }, sizeof(struct nv_dma_class),
-				 &object);
+			       }, sizeof(struct nv_dma_v0),
+			       &fctx->sema);
 	if (ret)
 		nv10_fence_context_del(chan);
 	return ret;
@@ -124,10 +124,12 @@ nv17_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv17_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 31;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &priv->bo);
+			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
 		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
 		if (!ret) {
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 4c534b7b04da..eb8b36714fa1 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -28,6 +28,8 @@
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_dp_helper.h>
 
+#include <nvif/class.h>
+
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_gem.h"
@@ -37,15 +39,6 @@
 #include "nouveau_fence.h"
 #include "nv50_display.h"
 
-#include <core/client.h>
-#include <core/gpuobj.h>
-#include <core/class.h>
-
-#include <subdev/timer.h>
-#include <subdev/bar.h>
-#include <subdev/fb.h>
-#include <subdev/i2c.h>
-
 #define EVO_DMA_NR 9
 
 #define EVO_MASTER  (0x00)
@@ -60,45 +53,34 @@
 #define EVO_FLIP_SEM0(c)  EVO_SYNC((c) + 1, 0x00)
 #define EVO_FLIP_SEM1(c)  EVO_SYNC((c) + 1, 0x10)
 
-#define EVO_CORE_HANDLE      (0xd1500000)
-#define EVO_CHAN_HANDLE(t,i) (0xd15c0000 | (((t) & 0x00ff) << 8) | (i))
-#define EVO_CHAN_OCLASS(t,c) ((nv_hclass(c) & 0xff00) | ((t) & 0x00ff))
-#define EVO_PUSH_HANDLE(t,i) (0xd15b0000 | (i) |                               \
-			      (((NV50_DISP_##t##_CLASS) & 0x00ff) << 8))
-
 /******************************************************************************
  * EVO channel
  *****************************************************************************/
 
 struct nv50_chan {
-	struct nouveau_object *user;
-	u32 handle;
+	struct nvif_object user;
 };
 
 static int
-nv50_chan_create(struct nouveau_object *core, u32 bclass, u8 head,
+nv50_chan_create(struct nvif_object *disp, const u32 *oclass, u8 head,
 		 void *data, u32 size, struct nv50_chan *chan)
 {
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	const u32 oclass = EVO_CHAN_OCLASS(bclass, core);
-	const u32 handle = EVO_CHAN_HANDLE(bclass, head);
-	int ret;
-
-	ret = nouveau_object_new(client, EVO_CORE_HANDLE, handle,
-				 oclass, data, size, &chan->user);
-	if (ret)
-		return ret;
-
-	chan->handle = handle;
-	return 0;
+	while (oclass[0]) {
+		int ret = nvif_object_init(disp, NULL, (oclass[0] << 16) | head,
+					   oclass[0], data, size,
+					  &chan->user);
+		if (oclass++, ret == 0) {
+			nvif_object_map(&chan->user);
+			return ret;
+		}
+	}
+	return -ENOSYS;
 }
 
 static void
-nv50_chan_destroy(struct nouveau_object *core, struct nv50_chan *chan)
+nv50_chan_destroy(struct nv50_chan *chan)
 {
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	if (chan->handle)
-		nouveau_object_del(client, EVO_CORE_HANDLE, chan->handle);
+	nvif_object_fini(&chan->user);
 }
 
 /******************************************************************************
@@ -110,16 +92,70 @@ struct nv50_pioc {
 };
 
 static void
-nv50_pioc_destroy(struct nouveau_object *core, struct nv50_pioc *pioc)
+nv50_pioc_destroy(struct nv50_pioc *pioc)
 {
-	nv50_chan_destroy(core, &pioc->base);
+	nv50_chan_destroy(&pioc->base);
 }
 
 static int
-nv50_pioc_create(struct nouveau_object *core, u32 bclass, u8 head,
+nv50_pioc_create(struct nvif_object *disp, const u32 *oclass, u8 head,
 		 void *data, u32 size, struct nv50_pioc *pioc)
 {
-	return nv50_chan_create(core, bclass, head, data, size, &pioc->base);
+	return nv50_chan_create(disp, oclass, head, data, size, &pioc->base);
+}
+
+/******************************************************************************
+ * Cursor Immediate
+ *****************************************************************************/
+
+struct nv50_curs {
+	struct nv50_pioc base;
+};
+
+static int
+nv50_curs_create(struct nvif_object *disp, int head, struct nv50_curs *curs)
+{
+	struct nv50_disp_cursor_v0 args = {
+		.head = head,
+	};
+	static const u32 oclass[] = {
+		GK104_DISP_CURSOR,
+		GF110_DISP_CURSOR,
+		GT214_DISP_CURSOR,
+		G82_DISP_CURSOR,
+		NV50_DISP_CURSOR,
+		0
+	};
+
+	return nv50_pioc_create(disp, oclass, head, &args, sizeof(args),
+			       &curs->base);
+}
+
+/******************************************************************************
+ * Overlay Immediate
+ *****************************************************************************/
+
+struct nv50_oimm {
+	struct nv50_pioc base;
+};
+
+static int
+nv50_oimm_create(struct nvif_object *disp, int head, struct nv50_oimm *oimm)
+{
+	struct nv50_disp_cursor_v0 args = {
+		.head = head,
+	};
+	static const u32 oclass[] = {
+		GK104_DISP_OVERLAY,
+		GF110_DISP_OVERLAY,
+		GT214_DISP_OVERLAY,
+		G82_DISP_OVERLAY,
+		NV50_DISP_OVERLAY,
+		0
+	};
+
+	return nv50_pioc_create(disp, oclass, head, &args, sizeof(args),
+			       &oimm->base);
 }
 
 /******************************************************************************
@@ -131,6 +167,9 @@ struct nv50_dmac {
 	dma_addr_t handle;
 	u32 *ptr;
 
+	struct nvif_object sync;
+	struct nvif_object vram;
+
 	/* Protects against concurrent pushbuf access to this channel, lock is
 	 * grabbed by evo_wait (if the pushbuf reservation is successful) and
 	 * dropped again by evo_kick. */
@@ -138,207 +177,113 @@ struct nv50_dmac {
 };
 
 static void
-nv50_dmac_destroy(struct nouveau_object *core, struct nv50_dmac *dmac)
+nv50_dmac_destroy(struct nv50_dmac *dmac, struct nvif_object *disp)
 {
+	nvif_object_fini(&dmac->vram);
+	nvif_object_fini(&dmac->sync);
+
+	nv50_chan_destroy(&dmac->base);
+
 	if (dmac->ptr) {
-		struct pci_dev *pdev = nv_device(core)->pdev;
+		struct pci_dev *pdev = nvkm_device(nvif_device(disp))->pdev;
 		pci_free_consistent(pdev, PAGE_SIZE, dmac->ptr, dmac->handle);
 	}
-
-	nv50_chan_destroy(core, &dmac->base);
-}
-
-static int
-nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
-{
-	struct nouveau_fb *pfb = nouveau_fb(core);
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	struct nouveau_object *object;
-	int ret = nouveau_object_new(client, parent, NvEvoVRAM_LP,
-				     NV_DMA_IN_MEMORY_CLASS,
-				     &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NV50_DMA_CONF0_ENABLE |
-					         NV50_DMA_CONF0_PART_256,
-				     }, sizeof(struct nv_dma_class), &object);
-	if (ret)
-		return ret;
-
-	ret = nouveau_object_new(client, parent, NvEvoFB16,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NV50_DMA_CONF0_ENABLE | 0x70 |
-					         NV50_DMA_CONF0_PART_256,
-				 }, sizeof(struct nv_dma_class), &object);
-	if (ret)
-		return ret;
-
-	ret = nouveau_object_new(client, parent, NvEvoFB32,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NV50_DMA_CONF0_ENABLE | 0x7a |
-					         NV50_DMA_CONF0_PART_256,
-				 }, sizeof(struct nv_dma_class), &object);
-	return ret;
-}
-
-static int
-nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
-{
-	struct nouveau_fb *pfb = nouveau_fb(core);
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	struct nouveau_object *object;
-	int ret = nouveau_object_new(client, parent, NvEvoVRAM_LP,
-				     NV_DMA_IN_MEMORY_CLASS,
-				     &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NVC0_DMA_CONF0_ENABLE,
-				     }, sizeof(struct nv_dma_class), &object);
-	if (ret)
-		return ret;
-
-	ret = nouveau_object_new(client, parent, NvEvoFB16,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
-				 }, sizeof(struct nv_dma_class), &object);
-	if (ret)
-		return ret;
-
-	ret = nouveau_object_new(client, parent, NvEvoFB32,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
-				 }, sizeof(struct nv_dma_class), &object);
-	return ret;
-}
-
-static int
-nvd0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
-{
-	struct nouveau_fb *pfb = nouveau_fb(core);
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	struct nouveau_object *object;
-	int ret = nouveau_object_new(client, parent, NvEvoVRAM_LP,
-				     NV_DMA_IN_MEMORY_CLASS,
-				     &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NVD0_DMA_CONF0_ENABLE |
-						 NVD0_DMA_CONF0_PAGE_LP,
-				     }, sizeof(struct nv_dma_class), &object);
-	if (ret)
-		return ret;
-
-	ret = nouveau_object_new(client, parent, NvEvoFB32,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
-					.start = 0,
-					.limit = pfb->ram->size - 1,
-					.conf0 = NVD0_DMA_CONF0_ENABLE | 0xfe |
-						 NVD0_DMA_CONF0_PAGE_LP,
-				 }, sizeof(struct nv_dma_class), &object);
-	return ret;
 }
 
 static int
-nv50_dmac_create(struct nouveau_object *core, u32 bclass, u8 head,
+nv50_dmac_create(struct nvif_object *disp, const u32 *oclass, u8 head,
 		 void *data, u32 size, u64 syncbuf,
 		 struct nv50_dmac *dmac)
 {
-	struct nouveau_fb *pfb = nouveau_fb(core);
-	struct nouveau_object *client = nv_pclass(core, NV_CLIENT_CLASS);
-	struct nouveau_object *object;
-	u32 pushbuf = *(u32 *)data;
+	struct nvif_device *device = nvif_device(disp);
+	struct nv50_disp_core_channel_dma_v0 *args = data;
+	struct nvif_object pushbuf;
 	int ret;
 
 	mutex_init(&dmac->lock);
 
-	dmac->ptr = pci_alloc_consistent(nv_device(core)->pdev, PAGE_SIZE,
-					&dmac->handle);
+	dmac->ptr = pci_alloc_consistent(nvkm_device(device)->pdev,
+					 PAGE_SIZE, &dmac->handle);
 	if (!dmac->ptr)
 		return -ENOMEM;
 
-	ret = nouveau_object_new(client, NVDRM_DEVICE, pushbuf,
-				 NV_DMA_FROM_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_PCI_US |
-						 NV_DMA_ACCESS_RD,
+	ret = nvif_object_init(nvif_object(device), NULL,
+			       args->pushbuf, NV_DMA_FROM_MEMORY,
+			       &(struct nv_dma_v0) {
+					.target = NV_DMA_V0_TARGET_PCI_US,
+					.access = NV_DMA_V0_ACCESS_RD,
 					.start = dmac->handle + 0x0000,
 					.limit = dmac->handle + 0x0fff,
-				 }, sizeof(struct nv_dma_class), &object);
+			       }, sizeof(struct nv_dma_v0), &pushbuf);
 	if (ret)
 		return ret;
 
-	ret = nv50_chan_create(core, bclass, head, data, size, &dmac->base);
+	ret = nv50_chan_create(disp, oclass, head, data, size, &dmac->base);
+	nvif_object_fini(&pushbuf);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(client, dmac->base.handle, NvEvoSync,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
+	ret = nvif_object_init(&dmac->base.user, NULL, 0xf0000000,
+			       NV_DMA_IN_MEMORY,
+			       &(struct nv_dma_v0) {
+					.target = NV_DMA_V0_TARGET_VRAM,
+					.access = NV_DMA_V0_ACCESS_RDWR,
 					.start = syncbuf + 0x0000,
 					.limit = syncbuf + 0x0fff,
-				 }, sizeof(struct nv_dma_class), &object);
+			       }, sizeof(struct nv_dma_v0),
+			       &dmac->sync);
 	if (ret)
 		return ret;
 
-	ret = nouveau_object_new(client, dmac->base.handle, NvEvoVRAM,
-				 NV_DMA_IN_MEMORY_CLASS,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
+	ret = nvif_object_init(&dmac->base.user, NULL, 0xf0000001,
+			       NV_DMA_IN_MEMORY,
+			       &(struct nv_dma_v0) {
+					.target = NV_DMA_V0_TARGET_VRAM,
+					.access = NV_DMA_V0_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram->size - 1,
-				 }, sizeof(struct nv_dma_class), &object);
+					.limit = device->info.ram_user - 1,
+			       }, sizeof(struct nv_dma_v0),
+			       &dmac->vram);
 	if (ret)
 		return ret;
 
-	if (nv_device(core)->card_type < NV_C0)
-		ret = nv50_dmac_create_fbdma(core, dmac->base.handle);
-	else
-	if (nv_device(core)->card_type < NV_D0)
-		ret = nvc0_dmac_create_fbdma(core, dmac->base.handle);
-	else
-		ret = nvd0_dmac_create_fbdma(core, dmac->base.handle);
 	return ret;
 }
 
+/******************************************************************************
+ * Core
+ *****************************************************************************/
+
 struct nv50_mast {
 	struct nv50_dmac base;
 };
 
-struct nv50_curs {
-	struct nv50_pioc base;
-};
+static int
+nv50_core_create(struct nvif_object *disp, u64 syncbuf, struct nv50_mast *core)
+{
+	struct nv50_disp_core_channel_dma_v0 args = {
+		.pushbuf = 0xb0007d00,
+	};
+	static const u32 oclass[] = {
+		GM107_DISP_CORE_CHANNEL_DMA,
+		GK110_DISP_CORE_CHANNEL_DMA,
+		GK104_DISP_CORE_CHANNEL_DMA,
+		GF110_DISP_CORE_CHANNEL_DMA,
+		GT214_DISP_CORE_CHANNEL_DMA,
+		GT206_DISP_CORE_CHANNEL_DMA,
+		GT200_DISP_CORE_CHANNEL_DMA,
+		G82_DISP_CORE_CHANNEL_DMA,
+		NV50_DISP_CORE_CHANNEL_DMA,
+		0
+	};
+
+	return nv50_dmac_create(disp, oclass, 0, &args, sizeof(args), syncbuf,
+			       &core->base);
+}
+
+/******************************************************************************
+ * Base
+ *****************************************************************************/
 
 struct nv50_sync {
 	struct nv50_dmac base;
@@ -346,13 +291,58 @@ struct nv50_sync {
 	u32 data;
 };
 
+static int
+nv50_base_create(struct nvif_object *disp, int head, u64 syncbuf,
+		 struct nv50_sync *base)
+{
+	struct nv50_disp_base_channel_dma_v0 args = {
+		.pushbuf = 0xb0007c00 | head,
+		.head = head,
+	};
+	static const u32 oclass[] = {
+		GK110_DISP_BASE_CHANNEL_DMA,
+		GK104_DISP_BASE_CHANNEL_DMA,
+		GF110_DISP_BASE_CHANNEL_DMA,
+		GT214_DISP_BASE_CHANNEL_DMA,
+		GT200_DISP_BASE_CHANNEL_DMA,
+		G82_DISP_BASE_CHANNEL_DMA,
+		NV50_DISP_BASE_CHANNEL_DMA,
+		0
+	};
+
+	return nv50_dmac_create(disp, oclass, head, &args, sizeof(args),
+				syncbuf, &base->base);
+}
+
+/******************************************************************************
+ * Overlay
+ *****************************************************************************/
+
 struct nv50_ovly {
 	struct nv50_dmac base;
 };
 
-struct nv50_oimm {
-	struct nv50_pioc base;
-};
+static int
+nv50_ovly_create(struct nvif_object *disp, int head, u64 syncbuf,
+		 struct nv50_ovly *ovly)
+{
+	struct nv50_disp_overlay_channel_dma_v0 args = {
+		.pushbuf = 0xb0007e00 | head,
+		.head = head,
+	};
+	static const u32 oclass[] = {
+		GK104_DISP_OVERLAY_CONTROL_DMA,
+		GF110_DISP_OVERLAY_CONTROL_DMA,
+		GT214_DISP_OVERLAY_CHANNEL_DMA,
+		GT200_DISP_OVERLAY_CHANNEL_DMA,
+		G82_DISP_OVERLAY_CHANNEL_DMA,
+		NV50_DISP_OVERLAY_CHANNEL_DMA,
+		0
+	};
+
+	return nv50_dmac_create(disp, oclass, head, &args, sizeof(args),
+				syncbuf, &ovly->base);
+}
 
 struct nv50_head {
 	struct nouveau_crtc base;
@@ -369,13 +359,19 @@ struct nv50_head {
 #define nv50_ovly(c) (&nv50_head(c)->ovly)
 #define nv50_oimm(c) (&nv50_head(c)->oimm)
 #define nv50_chan(c) (&(c)->base.base)
-#define nv50_vers(c) nv_mclass(nv50_chan(c)->user)
+#define nv50_vers(c) nv50_chan(c)->user.oclass
+
+struct nv50_fbdma {
+	struct list_head head;
+	struct nvif_object core;
+	struct nvif_object base[4];
+};
 
 struct nv50_disp {
-	struct nouveau_object *core;
+	struct nvif_object *disp;
 	struct nv50_mast mast;
 
-	u32 modeset;
+	struct list_head fbdma;
 
 	struct nouveau_bo *sync;
 };
@@ -401,16 +397,16 @@ static u32 *
 evo_wait(void *evoc, int nr)
 {
 	struct nv50_dmac *dmac = evoc;
-	u32 put = nv_ro32(dmac->base.user, 0x0000) / 4;
+	u32 put = nvif_rd32(&dmac->base.user, 0x0000) / 4;
 
 	mutex_lock(&dmac->lock);
 	if (put + nr >= (PAGE_SIZE / 4) - 8) {
 		dmac->ptr[put] = 0x20000000;
 
-		nv_wo32(dmac->base.user, 0x0000, 0x00000000);
-		if (!nv_wait(dmac->base.user, 0x0004, ~0, 0x00000000)) {
+		nvif_wr32(&dmac->base.user, 0x0000, 0x00000000);
+		if (!nvkm_wait(&dmac->base.user, 0x0004, ~0, 0x00000000)) {
 			mutex_unlock(&dmac->lock);
-			NV_ERROR(dmac->base.user, "channel stalled\n");
+			nv_error(nvkm_object(&dmac->base.user), "channel stalled\n");
 			return NULL;
 		}
 
@@ -424,7 +420,7 @@ static void
 evo_kick(u32 *push, void *evoc)
 {
 	struct nv50_dmac *dmac = evoc;
-	nv_wo32(dmac->base.user, 0x0000, (push - dmac->ptr) << 2);
+	nvif_wr32(&dmac->base.user, 0x0000, (push - dmac->ptr) << 2);
 	mutex_unlock(&dmac->lock);
 }
 
@@ -443,7 +439,7 @@ evo_sync_wait(void *data)
 static int
 evo_sync(struct drm_device *dev)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nv50_disp *disp = nv50_disp(dev);
 	struct nv50_mast *mast = nv50_mast(dev);
 	u32 *push = evo_wait(mast, 8);
@@ -455,7 +451,7 @@ evo_sync(struct drm_device *dev)
 		evo_data(push, 0x00000000);
 		evo_data(push, 0x00000000);
 		evo_kick(push, mast);
-		if (nv_wait_cb(device, evo_sync_wait, disp->sync))
+		if (nv_wait_cb(nvkm_device(device), evo_sync_wait, disp->sync))
 			return 0;
 	}
 
@@ -490,7 +486,7 @@ nv50_display_flip_wait(void *data)
 void
 nv50_display_flip_stop(struct drm_crtc *crtc)
 {
-	struct nouveau_device *device = nouveau_dev(crtc->dev);
+	struct nvif_device *device = &nouveau_drm(crtc->dev)->device;
 	struct nv50_display_flip flip = {
 		.disp = nv50_disp(crtc->dev),
 		.chan = nv50_sync(crtc),
@@ -510,7 +506,7 @@ nv50_display_flip_stop(struct drm_crtc *crtc)
 		evo_kick(push, flip.chan);
 	}
 
-	nv_wait_cb(device, nv50_display_flip_wait, &flip);
+	nv_wait_cb(nvkm_device(device), nv50_display_flip_wait, &flip);
 }
 
 int
@@ -534,7 +530,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	if (unlikely(push == NULL))
 		return -EBUSY;
 
-	if (chan && nv_mclass(chan->object) < NV84_CHANNEL_IND_CLASS) {
+	if (chan && chan->object->oclass < G82_CHANNEL_GPFIFO) {
 		ret = RING_SPACE(chan, 8);
 		if (ret)
 			return ret;
@@ -548,14 +544,14 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		OUT_RING  (chan, sync->addr);
 		OUT_RING  (chan, sync->data);
 	} else
-	if (chan && nv_mclass(chan->object) < NVC0_CHANNEL_IND_CLASS) {
+	if (chan && chan->object->oclass < FERMI_CHANNEL_GPFIFO) {
 		u64 addr = nv84_fence_crtc(chan, nv_crtc->index) + sync->addr;
 		ret = RING_SPACE(chan, 12);
 		if (ret)
 			return ret;
 
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-		OUT_RING  (chan, chan->vram);
+		OUT_RING  (chan, chan->vram.handle);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(addr ^ 0x10));
 		OUT_RING  (chan, lower_32_bits(addr ^ 0x10));
@@ -606,16 +602,16 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	evo_data(push, sync->addr);
 	evo_data(push, sync->data++);
 	evo_data(push, sync->data);
-	evo_data(push, NvEvoSync);
+	evo_data(push, sync->base.sync.handle);
 	evo_mthd(push, 0x00a0, 2);
 	evo_data(push, 0x00000000);
 	evo_data(push, 0x00000000);
 	evo_mthd(push, 0x00c0, 1);
-	evo_data(push, nv_fb->r_dma);
+	evo_data(push, nv_fb->r_handle);
 	evo_mthd(push, 0x0110, 2);
 	evo_data(push, 0x00000000);
 	evo_data(push, 0x00000000);
-	if (nv50_vers(sync) < NVD0_DISP_SYNC_CLASS) {
+	if (nv50_vers(sync) < GF110_DISP_BASE_CHANNEL_DMA) {
 		evo_mthd(push, 0x0800, 5);
 		evo_data(push, nv_fb->nvbo->bo.offset >> 8);
 		evo_data(push, 0);
@@ -667,11 +663,11 @@ nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
 
 	push = evo_wait(mast, 4);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x08a0 + (nv_crtc->index * 0x0400), 1);
 			evo_data(push, mode);
 		} else
-		if (nv50_vers(mast) < NVE0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GK104_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0490 + (nv_crtc->index * 0x0300), 1);
 			evo_data(push, mode);
 		} else {
@@ -762,7 +758,7 @@ nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update)
 
 	push = evo_wait(mast, 8);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			/*XXX: SCALE_CTRL_ACTIVE??? */
 			evo_mthd(push, 0x08d8 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, (oY << 16) | oX);
@@ -795,6 +791,22 @@ nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update)
 }
 
 static int
+nv50_crtc_set_raster_vblank_dmi(struct nouveau_crtc *nv_crtc, u32 usec)
+{
+	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
+	u32 *push;
+
+	push = evo_wait(mast, 8);
+	if (!push)
+		return -ENOMEM;
+
+	evo_mthd(push, 0x0828 + (nv_crtc->index * 0x400), 1);
+	evo_data(push, usec);
+	evo_kick(push, mast);
+	return 0;
+}
+
+static int
 nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update)
 {
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
@@ -807,7 +819,7 @@ nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update)
 
 	push = evo_wait(mast, 16);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x08a8 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, (hue << 20) | (vib << 8));
 		} else {
@@ -835,7 +847,7 @@ nv50_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb,
 
 	push = evo_wait(mast, 16);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0860 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, nvfb->nvbo->bo.offset >> 8);
 			evo_mthd(push, 0x0868 + (nv_crtc->index * 0x400), 3);
@@ -844,9 +856,9 @@ nv50_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb,
 			evo_data(push, nvfb->r_format);
 			evo_mthd(push, 0x08c0 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, (y << 16) | x);
-			if (nv50_vers(mast) > NV50_DISP_MAST_CLASS) {
+			if (nv50_vers(mast) > NV50_DISP_CORE_CHANNEL_DMA) {
 				evo_mthd(push, 0x0874 + (nv_crtc->index * 0x400), 1);
-				evo_data(push, nvfb->r_dma);
+				evo_data(push, nvfb->r_handle);
 			}
 		} else {
 			evo_mthd(push, 0x0460 + (nv_crtc->index * 0x300), 1);
@@ -855,7 +867,7 @@ nv50_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb,
 			evo_data(push, (fb->height << 16) | fb->width);
 			evo_data(push, nvfb->r_pitch);
 			evo_data(push, nvfb->r_format);
-			evo_data(push, nvfb->r_dma);
+			evo_data(push, nvfb->r_handle);
 			evo_mthd(push, 0x04b0 + (nv_crtc->index * 0x300), 1);
 			evo_data(push, (y << 16) | x);
 		}
@@ -867,7 +879,7 @@ nv50_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb,
 		evo_kick(push, mast);
 	}
 
-	nv_crtc->fb.tile_flags = nvfb->r_dma;
+	nv_crtc->fb.handle = nvfb->r_handle;
 	return 0;
 }
 
@@ -877,23 +889,23 @@ nv50_crtc_cursor_show(struct nouveau_crtc *nv_crtc)
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
 	u32 *push = evo_wait(mast, 16);
 	if (push) {
-		if (nv50_vers(mast) < NV84_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < G82_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0x85000000);
 			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
 		} else
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0x85000000);
 			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
 			evo_mthd(push, 0x089c + (nv_crtc->index * 0x400), 1);
-			evo_data(push, NvEvoVRAM);
+			evo_data(push, mast->base.vram.handle);
 		} else {
 			evo_mthd(push, 0x0480 + (nv_crtc->index * 0x300), 2);
 			evo_data(push, 0x85000000);
 			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
 			evo_mthd(push, 0x048c + (nv_crtc->index * 0x300), 1);
-			evo_data(push, NvEvoVRAM);
+			evo_data(push, mast->base.vram.handle);
 		}
 		evo_kick(push, mast);
 	}
@@ -905,11 +917,11 @@ nv50_crtc_cursor_hide(struct nouveau_crtc *nv_crtc)
 	struct nv50_mast *mast = nv50_mast(nv_crtc->base.dev);
 	u32 *push = evo_wait(mast, 16);
 	if (push) {
-		if (nv50_vers(mast) < NV84_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < G82_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, 0x05000000);
 		} else
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0880 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, 0x05000000);
 			evo_mthd(push, 0x089c + (nv_crtc->index * 0x400), 1);
@@ -960,13 +972,13 @@ nv50_crtc_prepare(struct drm_crtc *crtc)
 
 	push = evo_wait(mast, 6);
 	if (push) {
-		if (nv50_vers(mast) < NV84_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < G82_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0874 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, 0x00000000);
 			evo_mthd(push, 0x0840 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, 0x40000000);
 		} else
-		if (nv50_vers(mast) <  NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) <  GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0874 + (nv_crtc->index * 0x400), 1);
 			evo_data(push, 0x00000000);
 			evo_mthd(push, 0x0840 + (nv_crtc->index * 0x400), 1);
@@ -997,31 +1009,31 @@ nv50_crtc_commit(struct drm_crtc *crtc)
 
 	push = evo_wait(mast, 32);
 	if (push) {
-		if (nv50_vers(mast) < NV84_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < G82_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0874 + (nv_crtc->index * 0x400), 1);
-			evo_data(push, NvEvoVRAM_LP);
+			evo_data(push, nv_crtc->fb.handle);
 			evo_mthd(push, 0x0840 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0xc0000000);
 			evo_data(push, nv_crtc->lut.nvbo->bo.offset >> 8);
 		} else
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0874 + (nv_crtc->index * 0x400), 1);
-			evo_data(push, nv_crtc->fb.tile_flags);
+			evo_data(push, nv_crtc->fb.handle);
 			evo_mthd(push, 0x0840 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0xc0000000);
 			evo_data(push, nv_crtc->lut.nvbo->bo.offset >> 8);
 			evo_mthd(push, 0x085c + (nv_crtc->index * 0x400), 1);
-			evo_data(push, NvEvoVRAM);
+			evo_data(push, mast->base.vram.handle);
 		} else {
 			evo_mthd(push, 0x0474 + (nv_crtc->index * 0x300), 1);
-			evo_data(push, nv_crtc->fb.tile_flags);
+			evo_data(push, nv_crtc->fb.handle);
 			evo_mthd(push, 0x0440 + (nv_crtc->index * 0x300), 4);
 			evo_data(push, 0x83000000);
 			evo_data(push, nv_crtc->lut.nvbo->bo.offset >> 8);
 			evo_data(push, 0x00000000);
 			evo_data(push, 0x00000000);
 			evo_mthd(push, 0x045c + (nv_crtc->index * 0x300), 1);
-			evo_data(push, NvEvoVRAM);
+			evo_data(push, mast->base.vram.handle);
 			evo_mthd(push, 0x0430 + (nv_crtc->index * 0x300), 1);
 			evo_data(push, 0xffffff00);
 		}
@@ -1070,7 +1082,7 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1;
 	u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks;
 	u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks;
-	u32 vblan2e = 0, vblan2s = 1;
+	u32 vblan2e = 0, vblan2s = 1, vblankus = 0;
 	u32 *push;
 	int ret;
 
@@ -1087,6 +1099,11 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	vblanke = vsynce + vbackp;
 	vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace;
 	vblanks = vactive - vfrontp - 1;
+	/* XXX: Safe underestimate, even "0" works */
+	vblankus = (vactive - mode->vdisplay - 2) * hactive;
+	vblankus *= 1000;
+	vblankus /= mode->clock;
+
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		vblan2e = vactive + vsynce + vbackp;
 		vblan2s = vblan2e + (mode->vdisplay * vscan / ilace);
@@ -1099,7 +1116,7 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 
 	push = evo_wait(mast, 64);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0804 + (nv_crtc->index * 0x400), 2);
 			evo_data(push, 0x00800000 | mode->clock);
 			evo_data(push, (ilace == 2) ? 2 : 0);
@@ -1140,6 +1157,11 @@ nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
 	nv_connector = nouveau_crtc_connector_get(nv_crtc);
 	nv50_crtc_set_dither(nv_crtc, false);
 	nv50_crtc_set_scale(nv_crtc, false);
+
+	/* G94 only accepts this after setting scale */
+	if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA)
+		nv50_crtc_set_raster_vblank_dmi(nv_crtc, vblankus);
+
 	nv50_crtc_set_color_vibrance(nv_crtc, false);
 	nv50_crtc_set_image(nv_crtc, crtc->primary->fb, x, y, false);
 	return 0;
@@ -1192,7 +1214,7 @@ nv50_crtc_lut_load(struct drm_crtc *crtc)
 		u16 g = nv_crtc->lut.g[i] >> 2;
 		u16 b = nv_crtc->lut.b[i] >> 2;
 
-		if (nv_mclass(disp->core) < NVD0_DISP_CLASS) {
+		if (disp->disp->oclass < GF110_DISP) {
 			writew(r + 0x0000, lut + (i * 0x08) + 0);
 			writew(g + 0x0000, lut + (i * 0x08) + 2);
 			writew(b + 0x0000, lut + (i * 0x08) + 4);
@@ -1259,8 +1281,8 @@ nv50_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 {
 	struct nv50_curs *curs = nv50_curs(crtc);
 	struct nv50_chan *chan = nv50_chan(curs);
-	nv_wo32(chan->user, 0x0084, (y << 16) | (x & 0xffff));
-	nv_wo32(chan->user, 0x0080, 0x00000000);
+	nvif_wr32(&chan->user, 0x0084, (y << 16) | (x & 0xffff));
+	nvif_wr32(&chan->user, 0x0080, 0x00000000);
 	return 0;
 }
 
@@ -1287,11 +1309,16 @@ nv50_crtc_destroy(struct drm_crtc *crtc)
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	struct nv50_disp *disp = nv50_disp(crtc->dev);
 	struct nv50_head *head = nv50_head(crtc);
+	struct nv50_fbdma *fbdma;
 
-	nv50_dmac_destroy(disp->core, &head->ovly.base);
-	nv50_pioc_destroy(disp->core, &head->oimm.base);
-	nv50_dmac_destroy(disp->core, &head->sync.base);
-	nv50_pioc_destroy(disp->core, &head->curs.base);
+	list_for_each_entry(fbdma, &disp->fbdma, head) {
+		nvif_object_fini(&fbdma->base[nv_crtc->index]);
+	}
+
+	nv50_dmac_destroy(&head->ovly.base, disp->disp);
+	nv50_pioc_destroy(&head->oimm.base);
+	nv50_dmac_destroy(&head->sync.base, disp->disp);
+	nv50_pioc_destroy(&head->curs.base);
 
 	/*XXX: this shouldn't be necessary, but the core doesn't call
 	 *     disconnect() during the cleanup paths
@@ -1346,7 +1373,7 @@ nv50_cursor_set_offset(struct nouveau_crtc *nv_crtc, uint32_t offset)
 }
 
 static int
-nv50_crtc_create(struct drm_device *dev, struct nouveau_object *core, int index)
+nv50_crtc_create(struct drm_device *dev, int index)
 {
 	struct nv50_disp *disp = nv50_disp(dev);
 	struct nv50_head *head;
@@ -1377,7 +1404,7 @@ nv50_crtc_create(struct drm_device *dev, struct nouveau_object *core, int index)
 	drm_mode_crtc_set_gamma_size(crtc, 256);
 
 	ret = nouveau_bo_new(dev, 8192, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &head->base.lut.nvbo);
+			     0, 0x0000, NULL, NULL, &head->base.lut.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(head->base.lut.nvbo, TTM_PL_FLAG_VRAM);
 		if (!ret) {
@@ -1395,16 +1422,12 @@ nv50_crtc_create(struct drm_device *dev, struct nouveau_object *core, int index)
 	nv50_crtc_lut_load(crtc);
 
 	/* allocate cursor resources */
-	ret = nv50_pioc_create(disp->core, NV50_DISP_CURS_CLASS, index,
-			      &(struct nv50_display_curs_class) {
-					.head = index,
-			      }, sizeof(struct nv50_display_curs_class),
-			      &head->curs.base);
+	ret = nv50_curs_create(disp->disp, index, &head->curs);
 	if (ret)
 		goto out;
 
 	ret = nouveau_bo_new(dev, 64 * 64 * 4, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &head->base.cursor.nvbo);
+			     0, 0x0000, NULL, NULL, &head->base.cursor.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(head->base.cursor.nvbo, TTM_PL_FLAG_VRAM);
 		if (!ret) {
@@ -1420,12 +1443,8 @@ nv50_crtc_create(struct drm_device *dev, struct nouveau_object *core, int index)
 		goto out;
 
 	/* allocate page flip / sync resources */
-	ret = nv50_dmac_create(disp->core, NV50_DISP_SYNC_CLASS, index,
-			      &(struct nv50_display_sync_class) {
-					.pushbuf = EVO_PUSH_HANDLE(SYNC, index),
-					.head = index,
-			      }, sizeof(struct nv50_display_sync_class),
-			      disp->sync->bo.offset, &head->sync.base);
+	ret = nv50_base_create(disp->disp, index, disp->sync->bo.offset,
+			      &head->sync);
 	if (ret)
 		goto out;
 
@@ -1433,20 +1452,12 @@ nv50_crtc_create(struct drm_device *dev, struct nouveau_object *core, int index)
 	head->sync.data = 0x00000000;
 
 	/* allocate overlay resources */
-	ret = nv50_pioc_create(disp->core, NV50_DISP_OIMM_CLASS, index,
-			      &(struct nv50_display_oimm_class) {
-					.head = index,
-			      }, sizeof(struct nv50_display_oimm_class),
-			      &head->oimm.base);
+	ret = nv50_oimm_create(disp->disp, index, &head->oimm);
 	if (ret)
 		goto out;
 
-	ret = nv50_dmac_create(disp->core, NV50_DISP_OVLY_CLASS, index,
-			      &(struct nv50_display_ovly_class) {
-					.pushbuf = EVO_PUSH_HANDLE(OVLY, index),
-					.head = index,
-			      }, sizeof(struct nv50_display_ovly_class),
-			      disp->sync->bo.offset, &head->ovly.base);
+	ret = nv50_ovly_create(disp->disp, index, disp->sync->bo.offset,
+			      &head->ovly);
 	if (ret)
 		goto out;
 
@@ -1464,16 +1475,23 @@ nv50_dac_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	int or = nv_encoder->or;
-	u32 dpms_ctrl;
-
-	dpms_ctrl = 0x00000000;
-	if (mode == DRM_MODE_DPMS_STANDBY || mode == DRM_MODE_DPMS_OFF)
-		dpms_ctrl |= 0x00000001;
-	if (mode == DRM_MODE_DPMS_SUSPEND || mode == DRM_MODE_DPMS_OFF)
-		dpms_ctrl |= 0x00000004;
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_dac_pwr_v0 pwr;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_DAC_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = nv_encoder->dcb->hashm,
+		.pwr.state = 1,
+		.pwr.data  = 1,
+		.pwr.vsync = (mode != DRM_MODE_DPMS_SUSPEND &&
+			      mode != DRM_MODE_DPMS_OFF),
+		.pwr.hsync = (mode != DRM_MODE_DPMS_STANDBY &&
+			      mode != DRM_MODE_DPMS_OFF),
+	};
 
-	nv_call(disp->core, NV50_DISP_DAC_PWR + or, dpms_ctrl);
+	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 }
 
 static bool
@@ -1514,7 +1532,7 @@ nv50_dac_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	push = evo_wait(mast, 8);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			u32 syncs = 0x00000000;
 
 			if (mode->flags & DRM_MODE_FLAG_NHSYNC)
@@ -1563,7 +1581,7 @@ nv50_dac_disconnect(struct drm_encoder *encoder)
 
 		push = evo_wait(mast, 4);
 		if (push) {
-			if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+			if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 				evo_mthd(push, 0x0400 + (or * 0x080), 1);
 				evo_data(push, 0x00000000);
 			} else {
@@ -1580,14 +1598,25 @@ nv50_dac_disconnect(struct drm_encoder *encoder)
 static enum drm_connector_status
 nv50_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
 {
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	int ret, or = nouveau_encoder(encoder)->or;
-	u32 load = nouveau_drm(encoder->dev)->vbios.dactestval;
-	if (load == 0)
-		load = 340;
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_dac_load_v0 load;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_DAC_LOAD,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = nv_encoder->dcb->hashm,
+	};
+	int ret;
 
-	ret = nv_exec(disp->core, NV50_DISP_DAC_LOAD + or, &load, sizeof(load));
-	if (ret || !load)
+	args.load.data = nouveau_drm(encoder->dev)->vbios.dactestval;
+	if (args.load.data == 0)
+		args.load.data = 340;
+
+	ret = nvif_mthd(disp->disp, 0, &args, sizeof(args));
+	if (ret || !args.load.load)
 		return connector_status_disconnected;
 
 	return connector_status_connected;
@@ -1619,7 +1648,7 @@ static int
 nv50_dac_create(struct drm_connector *connector, struct dcb_output *dcbe)
 {
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
 	int type = DRM_MODE_ENCODER_DAC;
@@ -1648,27 +1677,50 @@ static void
 nv50_audio_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
 	struct nouveau_connector *nv_connector;
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
+	struct __packed {
+		struct {
+			struct nv50_disp_mthd_v1 mthd;
+			struct nv50_disp_sor_hda_eld_v0 eld;
+		} base;
+		u8 data[sizeof(nv_connector->base.eld)];
+	} args = {
+		.base.mthd.version = 1,
+		.base.mthd.method  = NV50_DISP_MTHD_V1_SOR_HDA_ELD,
+		.base.mthd.hasht   = nv_encoder->dcb->hasht,
+		.base.mthd.hashm   = (0xf0ff & nv_encoder->dcb->hashm) |
+				     (0x0100 << nv_crtc->index),
+	};
 
 	nv_connector = nouveau_encoder_connector_get(nv_encoder);
 	if (!drm_detect_monitor_audio(nv_connector->edid))
 		return;
 
 	drm_edid_to_eld(&nv_connector->base, nv_connector->edid);
+	memcpy(args.data, nv_connector->base.eld, sizeof(args.data));
 
-	nv_exec(disp->core, NVA3_DISP_SOR_HDA_ELD + nv_encoder->or,
-			    nv_connector->base.eld,
-			    nv_connector->base.eld[2] * 4);
+	nvif_mthd(disp->disp, 0, &args, sizeof(args.base) + args.data[2] * 4);
 }
 
 static void
-nv50_audio_disconnect(struct drm_encoder *encoder)
+nv50_audio_disconnect(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_hda_eld_v0 eld;
+	} args = {
+		.base.version = 1,
+		.base.method  = NV50_DISP_MTHD_V1_SOR_HDA_ELD,
+		.base.hasht   = nv_encoder->dcb->hasht,
+		.base.hashm   = (0xf0ff & nv_encoder->dcb->hashm) |
+				(0x0100 << nv_crtc->index),
+	};
 
-	nv_exec(disp->core, NVA3_DISP_SOR_HDA_ELD + nv_encoder->or, NULL, 0);
+	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 }
 
 /******************************************************************************
@@ -1679,10 +1731,20 @@ nv50_hdmi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector;
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	const u32 moff = (nv_crtc->index << 3) | nv_encoder->or;
-	u32 rekey = 56; /* binary driver, and tegra constant */
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_hdmi_pwr_v0 pwr;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_SOR_HDMI_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = (0xf0ff & nv_encoder->dcb->hashm) |
+			       (0x0100 << nv_crtc->index),
+		.pwr.state = 1,
+		.pwr.rekey = 56, /* binary driver, and tegra, constant */
+	};
+	struct nouveau_connector *nv_connector;
 	u32 max_ac_packet;
 
 	nv_connector = nouveau_encoder_connector_get(nv_encoder);
@@ -1690,14 +1752,11 @@ nv50_hdmi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
 		return;
 
 	max_ac_packet  = mode->htotal - mode->hdisplay;
-	max_ac_packet -= rekey;
+	max_ac_packet -= args.pwr.rekey;
 	max_ac_packet -= 18; /* constant from tegra */
-	max_ac_packet /= 32;
-
-	nv_call(disp->core, NV84_DISP_SOR_HDMI_PWR + moff,
-			    NV84_DISP_SOR_HDMI_PWR_STATE_ON |
-			    (max_ac_packet << 16) | rekey);
+	args.pwr.max_ac_packet = max_ac_packet / 32;
 
+	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 	nv50_audio_mode_set(encoder, mode);
 }
 
@@ -1706,11 +1765,18 @@ nv50_hdmi_disconnect(struct drm_encoder *encoder, struct nouveau_crtc *nv_crtc)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	const u32 moff = (nv_crtc->index << 3) | nv_encoder->or;
-
-	nv50_audio_disconnect(encoder);
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_hdmi_pwr_v0 pwr;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_SOR_HDMI_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = (0xf0ff & nv_encoder->dcb->hashm) |
+			       (0x0100 << nv_crtc->index),
+	};
 
-	nv_call(disp->core, NV84_DISP_SOR_HDMI_PWR + moff, 0x00000000);
+	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 }
 
 /******************************************************************************
@@ -1720,10 +1786,29 @@ static void
 nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nv50_disp *disp = nv50_disp(encoder->dev);
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_pwr_v0 pwr;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_SOR_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = nv_encoder->dcb->hashm,
+		.pwr.state = mode == DRM_MODE_DPMS_ON,
+	};
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_dp_pwr_v0 pwr;
+	} link = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_SOR_DP_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = nv_encoder->dcb->hashm,
+		.pwr.state = mode == DRM_MODE_DPMS_ON,
+	};
 	struct drm_device *dev = encoder->dev;
-	struct nv50_disp *disp = nv50_disp(dev);
 	struct drm_encoder *partner;
-	u32 mthd;
 
 	nv_encoder->last_dpms = mode;
 
@@ -1741,18 +1826,13 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 		}
 	}
 
-	mthd  = (ffs(nv_encoder->dcb->heads) - 1) << 3;
-	mthd |= (ffs(nv_encoder->dcb->sorconf.link) - 1) << 2;
-	mthd |= nv_encoder->or;
-
 	if (nv_encoder->dcb->type == DCB_OUTPUT_DP) {
-		nv_call(disp->core, NV50_DISP_SOR_PWR | mthd, 1);
-		mthd |= NV94_DISP_SOR_DP_PWR;
+		args.pwr.state = 1;
+		nvif_mthd(disp->disp, 0, &args, sizeof(args));
+		nvif_mthd(disp->disp, 0, &link, sizeof(link));
 	} else {
-		mthd |= NV50_DISP_SOR_PWR;
+		nvif_mthd(disp->disp, 0, &args, sizeof(args));
 	}
-
-	nv_call(disp->core, mthd, (mode == DRM_MODE_DPMS_ON));
 }
 
 static bool
@@ -1781,7 +1861,7 @@ nv50_sor_ctrl(struct nouveau_encoder *nv_encoder, u32 mask, u32 data)
 	struct nv50_mast *mast = nv50_mast(nv_encoder->base.base.dev);
 	u32 temp = (nv_encoder->ctrl & ~mask) | (data & mask), *push;
 	if (temp != nv_encoder->ctrl && (push = evo_wait(mast, 2))) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			evo_mthd(push, 0x0600 + (nv_encoder->or * 0x40), 1);
 			evo_data(push, (nv_encoder->ctrl = temp));
 		} else {
@@ -1804,6 +1884,7 @@ nv50_sor_disconnect(struct drm_encoder *encoder)
 	if (nv_crtc) {
 		nv50_crtc_prepare(&nv_crtc->base);
 		nv50_sor_ctrl(nv_encoder, 1 << nv_crtc->index, 0);
+		nv50_audio_disconnect(encoder, nv_crtc);
 		nv50_hdmi_disconnect(&nv_encoder->base.base, nv_crtc);
 	}
 }
@@ -1817,15 +1898,24 @@ static void
 nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 		  struct drm_display_mode *mode)
 {
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_sor_lvds_script_v0 lvds;
+	} lvds = {
+		.base.version = 1,
+		.base.method  = NV50_DISP_MTHD_V1_SOR_LVDS_SCRIPT,
+		.base.hasht   = nv_encoder->dcb->hasht,
+		.base.hashm   = nv_encoder->dcb->hashm,
+	};
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
 	struct nv50_mast *mast = nv50_mast(encoder->dev);
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
 	struct nouveau_connector *nv_connector;
 	struct nvbios *bios = &drm->vbios;
-	u32 lvds = 0, mask, ctrl;
+	u32 mask, ctrl;
 	u8 owner = 1 << nv_crtc->index;
 	u8 proto = 0xf;
 	u8 depth = 0x0;
@@ -1851,31 +1941,31 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 
 		if (bios->fp_no_ddc) {
 			if (bios->fp.dual_link)
-				lvds |= 0x0100;
+				lvds.lvds.script |= 0x0100;
 			if (bios->fp.if_is_24bit)
-				lvds |= 0x0200;
+				lvds.lvds.script |= 0x0200;
 		} else {
 			if (nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) {
 				if (((u8 *)nv_connector->edid)[121] == 2)
-					lvds |= 0x0100;
+					lvds.lvds.script |= 0x0100;
 			} else
 			if (mode->clock >= bios->fp.duallink_transition_clk) {
-				lvds |= 0x0100;
+				lvds.lvds.script |= 0x0100;
 			}
 
-			if (lvds & 0x0100) {
+			if (lvds.lvds.script & 0x0100) {
 				if (bios->fp.strapless_is_24bit & 2)
-					lvds |= 0x0200;
+					lvds.lvds.script |= 0x0200;
 			} else {
 				if (bios->fp.strapless_is_24bit & 1)
-					lvds |= 0x0200;
+					lvds.lvds.script |= 0x0200;
 			}
 
 			if (nv_connector->base.display_info.bpc == 8)
-				lvds |= 0x0200;
+				lvds.lvds.script |= 0x0200;
 		}
 
-		nv_call(disp->core, NV50_DISP_SOR_LVDS_SCRIPT + nv_encoder->or, lvds);
+		nvif_mthd(disp->disp, 0, &lvds, sizeof(lvds));
 		break;
 	case DCB_OUTPUT_DP:
 		if (nv_connector->base.display_info.bpc == 6) {
@@ -1894,6 +1984,7 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 			proto = 0x8;
 		else
 			proto = 0x9;
+		nv50_audio_mode_set(encoder, mode);
 		break;
 	default:
 		BUG_ON(1);
@@ -1902,7 +1993,7 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 
 	nv50_sor_dpms(&nv_encoder->base.base, DRM_MODE_DPMS_ON);
 
-	if (nv50_vers(mast) >= NVD0_DISP_CLASS) {
+	if (nv50_vers(mast) >= GF110_DISP) {
 		u32 *push = evo_wait(mast, 3);
 		if (push) {
 			u32 magic = 0x31ec6000 | (nv_crtc->index << 25);
@@ -1961,7 +2052,7 @@ static int
 nv50_sor_create(struct drm_connector *connector, struct dcb_output *dcbe)
 {
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
 	int type;
@@ -2002,9 +2093,19 @@ nv50_pior_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct nv50_disp *disp = nv50_disp(encoder->dev);
-	u32 mthd = (nv_encoder->dcb->type << 12) | nv_encoder->or;
-	u32 ctrl = (mode == DRM_MODE_DPMS_ON);
-	nv_call(disp->core, NV50_DISP_PIOR_PWR + mthd, ctrl);
+	struct {
+		struct nv50_disp_mthd_v1 base;
+		struct nv50_disp_pior_pwr_v0 pwr;
+	} args = {
+		.base.version = 1,
+		.base.method = NV50_DISP_MTHD_V1_PIOR_PWR,
+		.base.hasht  = nv_encoder->dcb->hasht,
+		.base.hashm  = nv_encoder->dcb->hashm,
+		.pwr.state = mode == DRM_MODE_DPMS_ON,
+		.pwr.type = nv_encoder->dcb->type,
+	};
+
+	nvif_mthd(disp->disp, 0, &args, sizeof(args));
 }
 
 static bool
@@ -2067,7 +2168,7 @@ nv50_pior_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	push = evo_wait(mast, 8);
 	if (push) {
-		if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+		if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 			u32 ctrl = (depth << 16) | (proto << 8) | owner;
 			if (mode->flags & DRM_MODE_FLAG_NHSYNC)
 				ctrl |= 0x00001000;
@@ -2096,7 +2197,7 @@ nv50_pior_disconnect(struct drm_encoder *encoder)
 
 		push = evo_wait(mast, 4);
 		if (push) {
-			if (nv50_vers(mast) < NVD0_DISP_MAST_CLASS) {
+			if (nv50_vers(mast) < GF110_DISP_CORE_CHANNEL_DMA) {
 				evo_mthd(push, 0x0700 + (or * 0x040), 1);
 				evo_data(push, 0x00000000);
 			}
@@ -2132,7 +2233,7 @@ static int
 nv50_pior_create(struct drm_connector *connector, struct dcb_output *dcbe)
 {
 	struct nouveau_drm *drm = nouveau_drm(connector->dev);
-	struct nouveau_i2c *i2c = nouveau_i2c(drm->device);
+	struct nouveau_i2c *i2c = nvkm_i2c(&drm->device);
 	struct nouveau_i2c_port *ddc = NULL;
 	struct nouveau_encoder *nv_encoder;
 	struct drm_encoder *encoder;
@@ -2169,8 +2270,151 @@ nv50_pior_create(struct drm_connector *connector, struct dcb_output *dcbe)
 }
 
 /******************************************************************************
+ * Framebuffer
+ *****************************************************************************/
+
+static void
+nv50_fbdma_fini(struct nv50_fbdma *fbdma)
+{
+	int i;
+	for (i = 0; i < ARRAY_SIZE(fbdma->base); i++)
+		nvif_object_fini(&fbdma->base[i]);
+	nvif_object_fini(&fbdma->core);
+	list_del(&fbdma->head);
+	kfree(fbdma);
+}
+
+static int
+nv50_fbdma_init(struct drm_device *dev, u32 name, u64 offset, u64 length, u8 kind)
+{
+	struct nouveau_drm *drm = nouveau_drm(dev);
+	struct nv50_disp *disp = nv50_disp(dev);
+	struct nv50_mast *mast = nv50_mast(dev);
+	struct __attribute__ ((packed)) {
+		struct nv_dma_v0 base;
+		union {
+			struct nv50_dma_v0 nv50;
+			struct gf100_dma_v0 gf100;
+			struct gf110_dma_v0 gf110;
+		};
+	} args = {};
+	struct nv50_fbdma *fbdma;
+	struct drm_crtc *crtc;
+	u32 size = sizeof(args.base);
+	int ret;
+
+	list_for_each_entry(fbdma, &disp->fbdma, head) {
+		if (fbdma->core.handle == name)
+			return 0;
+	}
+
+	fbdma = kzalloc(sizeof(*fbdma), GFP_KERNEL);
+	if (!fbdma)
+		return -ENOMEM;
+	list_add(&fbdma->head, &disp->fbdma);
+
+	args.base.target = NV_DMA_V0_TARGET_VRAM;
+	args.base.access = NV_DMA_V0_ACCESS_RDWR;
+	args.base.start = offset;
+	args.base.limit = offset + length - 1;
+
+	if (drm->device.info.chipset < 0x80) {
+		args.nv50.part = NV50_DMA_V0_PART_256;
+		size += sizeof(args.nv50);
+	} else
+	if (drm->device.info.chipset < 0xc0) {
+		args.nv50.part = NV50_DMA_V0_PART_256;
+		args.nv50.kind = kind;
+		size += sizeof(args.nv50);
+	} else
+	if (drm->device.info.chipset < 0xd0) {
+		args.gf100.kind = kind;
+		size += sizeof(args.gf100);
+	} else {
+		args.gf110.page = GF110_DMA_V0_PAGE_LP;
+		args.gf110.kind = kind;
+		size += sizeof(args.gf110);
+	}
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nv50_head *head = nv50_head(crtc);
+		int ret = nvif_object_init(&head->sync.base.base.user, NULL,
+					    name, NV_DMA_IN_MEMORY, &args, size,
+					   &fbdma->base[head->base.index]);
+		if (ret) {
+			nv50_fbdma_fini(fbdma);
+			return ret;
+		}
+	}
+
+	ret = nvif_object_init(&mast->base.base.user, NULL, name,
+				NV_DMA_IN_MEMORY, &args, size,
+			       &fbdma->core);
+	if (ret) {
+		nv50_fbdma_fini(fbdma);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void
+nv50_fb_dtor(struct drm_framebuffer *fb)
+{
+}
+
+static int
+nv50_fb_ctor(struct drm_framebuffer *fb)
+{
+	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
+	struct nouveau_drm *drm = nouveau_drm(fb->dev);
+	struct nouveau_bo *nvbo = nv_fb->nvbo;
+	struct nv50_disp *disp = nv50_disp(fb->dev);
+	u8 kind = nouveau_bo_tile_layout(nvbo) >> 8;
+	u8 tile = nvbo->tile_mode;
+
+	if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG) {
+		NV_ERROR(drm, "framebuffer requires contiguous bo\n");
+		return -EINVAL;
+	}
+
+	if (drm->device.info.chipset >= 0xc0)
+		tile >>= 4; /* yep.. */
+
+	switch (fb->depth) {
+	case  8: nv_fb->r_format = 0x1e00; break;
+	case 15: nv_fb->r_format = 0xe900; break;
+	case 16: nv_fb->r_format = 0xe800; break;
+	case 24:
+	case 32: nv_fb->r_format = 0xcf00; break;
+	case 30: nv_fb->r_format = 0xd100; break;
+	default:
+		 NV_ERROR(drm, "unknown depth %d\n", fb->depth);
+		 return -EINVAL;
+	}
+
+	if (disp->disp->oclass < G82_DISP) {
+		nv_fb->r_pitch   = kind ? (((fb->pitches[0] / 4) << 4) | tile) :
+					    (fb->pitches[0] | 0x00100000);
+		nv_fb->r_format |= kind << 16;
+	} else
+	if (disp->disp->oclass < GF110_DISP) {
+		nv_fb->r_pitch  = kind ? (((fb->pitches[0] / 4) << 4) | tile) :
+					   (fb->pitches[0] | 0x00100000);
+	} else {
+		nv_fb->r_pitch  = kind ? (((fb->pitches[0] / 4) << 4) | tile) :
+					   (fb->pitches[0] | 0x01000000);
+	}
+	nv_fb->r_handle = 0xffff0000 | kind;
+
+	return nv50_fbdma_init(fb->dev, nv_fb->r_handle, 0,
+			       drm->device.info.ram_user, kind);
+}
+
+/******************************************************************************
  * Init
  *****************************************************************************/
+
 void
 nv50_display_fini(struct drm_device *dev)
 {
@@ -2193,7 +2437,7 @@ nv50_display_init(struct drm_device *dev)
 	}
 
 	evo_mthd(push, 0x0088, 1);
-	evo_data(push, NvEvoSync);
+	evo_data(push, nv50_mast(dev)->base.sync.handle);
 	evo_kick(push, nv50_mast(dev));
 	return 0;
 }
@@ -2202,8 +2446,13 @@ void
 nv50_display_destroy(struct drm_device *dev)
 {
 	struct nv50_disp *disp = nv50_disp(dev);
+	struct nv50_fbdma *fbdma, *fbtmp;
+
+	list_for_each_entry_safe(fbdma, fbtmp, &disp->fbdma, head) {
+		nv50_fbdma_fini(fbdma);
+	}
 
-	nv50_dmac_destroy(disp->core, &disp->mast.base);
+	nv50_dmac_destroy(&disp->mast.base, disp->disp);
 
 	nouveau_bo_unmap(disp->sync);
 	if (disp->sync)
@@ -2217,7 +2466,7 @@ nv50_display_destroy(struct drm_device *dev)
 int
 nv50_display_create(struct drm_device *dev)
 {
-	struct nouveau_device *device = nouveau_dev(dev);
+	struct nvif_device *device = &nouveau_drm(dev)->device;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct dcb_table *dcb = &drm->vbios.dcb;
 	struct drm_connector *connector, *tmp;
@@ -2228,16 +2477,19 @@ nv50_display_create(struct drm_device *dev)
 	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
 	if (!disp)
 		return -ENOMEM;
+	INIT_LIST_HEAD(&disp->fbdma);
 
 	nouveau_display(dev)->priv = disp;
 	nouveau_display(dev)->dtor = nv50_display_destroy;
 	nouveau_display(dev)->init = nv50_display_init;
 	nouveau_display(dev)->fini = nv50_display_fini;
-	disp->core = nouveau_display(dev)->core;
+	nouveau_display(dev)->fb_ctor = nv50_fb_ctor;
+	nouveau_display(dev)->fb_dtor = nv50_fb_dtor;
+	disp->disp = &nouveau_display(dev)->disp;
 
 	/* small shared memory area we use for notifiers and semaphores */
 	ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &disp->sync);
+			     0, 0x0000, NULL, NULL, &disp->sync);
 	if (!ret) {
 		ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM);
 		if (!ret) {
@@ -2253,22 +2505,19 @@ nv50_display_create(struct drm_device *dev)
 		goto out;
 
 	/* allocate master evo channel */
-	ret = nv50_dmac_create(disp->core, NV50_DISP_MAST_CLASS, 0,
-			      &(struct nv50_display_mast_class) {
-					.pushbuf = EVO_PUSH_HANDLE(MAST, 0),
-			      }, sizeof(struct nv50_display_mast_class),
-			      disp->sync->bo.offset, &disp->mast.base);
+	ret = nv50_core_create(disp->disp, disp->sync->bo.offset,
+			      &disp->mast);
 	if (ret)
 		goto out;
 
 	/* create crtc objects to represent the hw heads */
-	if (nv_mclass(disp->core) >= NVD0_DISP_CLASS)
-		crtcs = nv_rd32(device, 0x022448);
+	if (disp->disp->oclass >= GF110_DISP)
+		crtcs = nvif_rd32(device, 0x022448);
 	else
 		crtcs = 2;
 
 	for (i = 0; i < crtcs; i++) {
-		ret = nv50_crtc_create(dev, disp->core, i);
+		ret = nv50_crtc_create(dev, i);
 		if (ret)
 			goto out;
 	}
diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c
index 52068a0910dc..394c89abcc97 100644
--- a/drivers/gpu/drm/nouveau/nv50_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c
@@ -154,7 +154,6 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	struct drm_device *dev = nfbdev->dev;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_channel *chan = drm->channel;
-	struct nouveau_object *object;
 	int ret, format;
 
 	switch (info->var.bits_per_pixel) {
@@ -184,8 +183,8 @@ nv50_fbcon_accel_init(struct fb_info *info)
 		return -EINVAL;
 	}
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, Nv2D,
-				 0x502d, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x502d, 0x502d, NULL, 0,
+			       &nfbdev->twod);
 	if (ret)
 		return ret;
 
@@ -196,11 +195,11 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	}
 
 	BEGIN_NV04(chan, NvSub2D, 0x0000, 1);
-	OUT_RING(chan, Nv2D);
+	OUT_RING(chan, nfbdev->twod.handle);
 	BEGIN_NV04(chan, NvSub2D, 0x0184, 3);
-	OUT_RING(chan, NvDmaFB);
-	OUT_RING(chan, NvDmaFB);
-	OUT_RING(chan, NvDmaFB);
+	OUT_RING(chan, chan->vram.handle);
+	OUT_RING(chan, chan->vram.handle);
+	OUT_RING(chan, chan->vram.handle);
 	BEGIN_NV04(chan, NvSub2D, 0x0290, 1);
 	OUT_RING(chan, 0);
 	BEGIN_NV04(chan, NvSub2D, 0x0888, 1);
diff --git a/drivers/gpu/drm/nouveau/nv50_fence.c b/drivers/gpu/drm/nouveau/nv50_fence.c
index 0ee363840035..22d242b37962 100644
--- a/drivers/gpu/drm/nouveau/nv50_fence.c
+++ b/drivers/gpu/drm/nouveau/nv50_fence.c
@@ -22,8 +22,8 @@
  * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-#include <core/object.h>
-#include <core/class.h>
+#include <nvif/os.h>
+#include <nvif/class.h>
 
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
@@ -38,7 +38,6 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 	struct nv10_fence_priv *priv = chan->drm->fence;
 	struct nv10_fence_chan *fctx;
 	struct ttm_mem_reg *mem = &priv->bo->bo.mem;
-	struct nouveau_object *object;
 	u32 start = mem->start * PAGE_SIZE;
 	u32 limit = start + mem->size - 1;
 	int ret, i;
@@ -47,20 +46,19 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv10_fence_emit;
 	fctx->base.read = nv10_fence_read;
 	fctx->base.sync = nv17_fence_sync;
 
-	ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-				 NvSema, 0x003d,
-				 &(struct nv_dma_class) {
-					.flags = NV_DMA_TARGET_VRAM |
-						 NV_DMA_ACCESS_RDWR,
+	ret = nvif_object_init(chan->object, NULL, NvSema, NV_DMA_IN_MEMORY,
+			       &(struct nv_dma_v0) {
+					.target = NV_DMA_V0_TARGET_VRAM,
+					.access = NV_DMA_V0_ACCESS_RDWR,
 					.start = start,
 					.limit = limit,
-				 }, sizeof(struct nv_dma_class),
-				 &object);
+			       }, sizeof(struct nv_dma_v0),
+			       &fctx->sema);
 
 	/* dma objects for display sync channel semaphore blocks */
 	for (i = 0; !ret && i < dev->mode_config.num_crtc; i++) {
@@ -68,15 +66,14 @@ nv50_fence_context_new(struct nouveau_channel *chan)
 		u32 start = bo->bo.mem.start * PAGE_SIZE;
 		u32 limit = start + bo->bo.mem.size - 1;
 
-		ret = nouveau_object_new(nv_object(chan->cli), chan->handle,
-					 NvEvoSema0 + i, 0x003d,
-					 &(struct nv_dma_class) {
-						.flags = NV_DMA_TARGET_VRAM |
-							 NV_DMA_ACCESS_RDWR,
+		ret = nvif_object_init(chan->object, NULL, NvEvoSema0 + i,
+				       NV_DMA_IN_MEMORY, &(struct nv_dma_v0) {
+						.target = NV_DMA_V0_TARGET_VRAM,
+						.access = NV_DMA_V0_ACCESS_RDWR,
 						.start = start,
 						.limit = limit,
-					 }, sizeof(struct nv_dma_class),
-					 &object);
+				       }, sizeof(struct nv_dma_v0),
+				       &fctx->head[i]);
 	}
 
 	if (ret)
@@ -98,10 +95,12 @@ nv50_fence_create(struct nouveau_drm *drm)
 	priv->base.resume = nv17_fence_resume;
 	priv->base.context_new = nv50_fence_context_new;
 	priv->base.context_del = nv10_fence_context_del;
+	priv->base.contexts = 127;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	spin_lock_init(&priv->lock);
 
 	ret = nouveau_bo_new(drm->dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, NULL, &priv->bo);
+			     0, 0x0000, NULL, NULL, &priv->bo);
 	if (!ret) {
 		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
 		if (!ret) {
diff --git a/drivers/gpu/drm/nouveau/nv84_fence.c b/drivers/gpu/drm/nouveau/nv84_fence.c
index 9fd475c89820..d6c6c87c3f07 100644
--- a/drivers/gpu/drm/nouveau/nv84_fence.c
+++ b/drivers/gpu/drm/nouveau/nv84_fence.c
@@ -22,12 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/object.h>
-#include <core/client.h>
-#include <core/class.h>
-
-#include <engine/fifo.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
@@ -47,7 +41,7 @@ nv84_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 	int ret = RING_SPACE(chan, 8);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-		OUT_RING  (chan, chan->vram);
+		OUT_RING  (chan, chan->vram.handle);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
 		OUT_RING  (chan, upper_32_bits(virtual));
 		OUT_RING  (chan, lower_32_bits(virtual));
@@ -65,7 +59,7 @@ nv84_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
 	int ret = RING_SPACE(chan, 7);
 	if (ret == 0) {
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
-		OUT_RING  (chan, chan->vram);
+		OUT_RING  (chan, chan->vram.handle);
 		BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
 		OUT_RING  (chan, upper_32_bits(virtual));
 		OUT_RING  (chan, lower_32_bits(virtual));
@@ -81,15 +75,14 @@ nv84_fence_emit(struct nouveau_fence *fence)
 {
 	struct nouveau_channel *chan = fence->channel;
 	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	u64 addr = fifo->chid * 16;
+	u64 addr = chan->chid * 16;
 
 	if (fence->sysmem)
 		addr += fctx->vma_gart.offset;
 	else
 		addr += fctx->vma.offset;
 
-	return fctx->base.emit32(chan, addr, fence->sequence);
+	return fctx->base.emit32(chan, addr, fence->base.seqno);
 }
 
 static int
@@ -97,23 +90,21 @@ nv84_fence_sync(struct nouveau_fence *fence,
 		struct nouveau_channel *prev, struct nouveau_channel *chan)
 {
 	struct nv84_fence_chan *fctx = chan->fence;
-	struct nouveau_fifo_chan *fifo = (void *)prev->object;
-	u64 addr = fifo->chid * 16;
+	u64 addr = prev->chid * 16;
 
 	if (fence->sysmem)
 		addr += fctx->vma_gart.offset;
 	else
 		addr += fctx->vma.offset;
 
-	return fctx->base.sync32(chan, addr, fence->sequence);
+	return fctx->base.sync32(chan, addr, fence->base.seqno);
 }
 
 static u32
 nv84_fence_read(struct nouveau_channel *chan)
 {
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
 	struct nv84_fence_priv *priv = chan->drm->fence;
-	return nouveau_bo_rd32(priv->bo, fifo->chid * 16/4);
+	return nouveau_bo_rd32(priv->bo, chan->chid * 16/4);
 }
 
 static void
@@ -129,18 +120,18 @@ nv84_fence_context_del(struct nouveau_channel *chan)
 		nouveau_bo_vma_del(bo, &fctx->dispc_vma[i]);
 	}
 
+	nouveau_bo_wr32(priv->bo, chan->chid * 16 / 4, fctx->base.sequence);
 	nouveau_bo_vma_del(priv->bo, &fctx->vma_gart);
 	nouveau_bo_vma_del(priv->bo, &fctx->vma);
 	nouveau_fence_context_del(&fctx->base);
 	chan->fence = NULL;
-	kfree(fctx);
+	nouveau_fence_context_free(&fctx->base);
 }
 
 int
 nv84_fence_context_new(struct nouveau_channel *chan)
 {
-	struct nouveau_fifo_chan *fifo = (void *)chan->object;
-	struct nouveau_client *client = nouveau_client(fifo);
+	struct nouveau_cli *cli = (void *)nvif_client(&chan->device->base);
 	struct nv84_fence_priv *priv = chan->drm->fence;
 	struct nv84_fence_chan *fctx;
 	int ret, i;
@@ -149,27 +140,26 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 	if (!fctx)
 		return -ENOMEM;
 
-	nouveau_fence_context_new(&fctx->base);
+	nouveau_fence_context_new(chan, &fctx->base);
 	fctx->base.emit = nv84_fence_emit;
 	fctx->base.sync = nv84_fence_sync;
 	fctx->base.read = nv84_fence_read;
 	fctx->base.emit32 = nv84_fence_emit32;
 	fctx->base.sync32 = nv84_fence_sync32;
+	fctx->base.sequence = nv84_fence_read(chan);
 
-	ret = nouveau_bo_vma_add(priv->bo, client->vm, &fctx->vma);
+	ret = nouveau_bo_vma_add(priv->bo, cli->vm, &fctx->vma);
 	if (ret == 0) {
-		ret = nouveau_bo_vma_add(priv->bo_gart, client->vm,
+		ret = nouveau_bo_vma_add(priv->bo_gart, cli->vm,
 					&fctx->vma_gart);
 	}
 
 	/* map display semaphore buffers into channel's vm */
 	for (i = 0; !ret && i < chan->drm->dev->mode_config.num_crtc; i++) {
 		struct nouveau_bo *bo = nv50_display_crtc_sema(chan->drm->dev, i);
-		ret = nouveau_bo_vma_add(bo, client->vm, &fctx->dispc_vma[i]);
+		ret = nouveau_bo_vma_add(bo, cli->vm, &fctx->dispc_vma[i]);
 	}
 
-	nouveau_bo_wr32(priv->bo, fifo->chid * 16/4, 0x00000000);
-
 	if (ret)
 		nv84_fence_context_del(chan);
 	return ret;
@@ -178,13 +168,12 @@ nv84_fence_context_new(struct nouveau_channel *chan)
 static bool
 nv84_fence_suspend(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
 	struct nv84_fence_priv *priv = drm->fence;
 	int i;
 
-	priv->suspend = vmalloc((pfifo->max + 1) * sizeof(u32));
+	priv->suspend = vmalloc(priv->base.contexts * sizeof(u32));
 	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
+		for (i = 0; i < priv->base.contexts; i++)
 			priv->suspend[i] = nouveau_bo_rd32(priv->bo, i*4);
 	}
 
@@ -194,12 +183,11 @@ nv84_fence_suspend(struct nouveau_drm *drm)
 static void
 nv84_fence_resume(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
 	struct nv84_fence_priv *priv = drm->fence;
 	int i;
 
 	if (priv->suspend) {
-		for (i = 0; i <= pfifo->max; i++)
+		for (i = 0; i < priv->base.contexts; i++)
 			nouveau_bo_wr32(priv->bo, i*4, priv->suspend[i]);
 		vfree(priv->suspend);
 		priv->suspend = NULL;
@@ -225,7 +213,7 @@ nv84_fence_destroy(struct nouveau_drm *drm)
 int
 nv84_fence_create(struct nouveau_drm *drm)
 {
-	struct nouveau_fifo *pfifo = nouveau_fifo(drm->device);
+	struct nouveau_fifo *pfifo = nvkm_fifo(&drm->device);
 	struct nv84_fence_priv *priv;
 	int ret;
 
@@ -239,11 +227,12 @@ nv84_fence_create(struct nouveau_drm *drm)
 	priv->base.context_new = nv84_fence_context_new;
 	priv->base.context_del = nv84_fence_context_del;
 
-	init_waitqueue_head(&priv->base.waiting);
+	priv->base.contexts = pfifo->max + 1;
+	priv->base.context_base = fence_context_alloc(priv->base.contexts);
 	priv->base.uevent = true;
 
-	ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
-			     TTM_PL_FLAG_VRAM, 0, 0, NULL, &priv->bo);
+	ret = nouveau_bo_new(drm->dev, 16 * priv->base.contexts, 0,
+			     TTM_PL_FLAG_VRAM, 0, 0, NULL, NULL, &priv->bo);
 	if (ret == 0) {
 		ret = nouveau_bo_pin(priv->bo, TTM_PL_FLAG_VRAM);
 		if (ret == 0) {
@@ -256,8 +245,8 @@ nv84_fence_create(struct nouveau_drm *drm)
 	}
 
 	if (ret == 0)
-		ret = nouveau_bo_new(drm->dev, 16 * (pfifo->max + 1), 0,
-				     TTM_PL_FLAG_TT, 0, 0, NULL,
+		ret = nouveau_bo_new(drm->dev, 16 * priv->base.contexts, 0,
+				     TTM_PL_FLAG_TT, 0, 0, NULL, NULL,
 				     &priv->bo_gart);
 	if (ret == 0) {
 		ret = nouveau_bo_pin(priv->bo_gart, TTM_PL_FLAG_TT);
diff --git a/drivers/gpu/drm/nouveau/nvc0_fbcon.c b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
index 9dcd30f3e1e0..61246677e8dc 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
@@ -154,11 +154,10 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	struct nouveau_framebuffer *fb = &nfbdev->nouveau_fb;
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	struct nouveau_channel *chan = drm->channel;
-	struct nouveau_object *object;
 	int ret, format;
 
-	ret = nouveau_object_new(nv_object(chan->cli), NVDRM_CHAN, Nv2D,
-				 0x902d, NULL, 0, &object);
+	ret = nvif_object_init(chan->object, NULL, 0x902d, 0x902d, NULL, 0,
+			       &nfbdev->twod);
 	if (ret)
 		return ret;
 
@@ -197,7 +196,7 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	}
 
 	BEGIN_NVC0(chan, NvSub2D, 0x0000, 1);
-	OUT_RING  (chan, 0x0000902d);
+	OUT_RING  (chan, nfbdev->twod.handle);
 	BEGIN_NVC0(chan, NvSub2D, 0x0290, 1);
 	OUT_RING  (chan, 0);
 	BEGIN_NVC0(chan, NvSub2D, 0x0888, 1);
diff --git a/drivers/gpu/drm/nouveau/nvc0_fence.c b/drivers/gpu/drm/nouveau/nvc0_fence.c
index 9566267fbc42..becf19abda2d 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fence.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fence.c
@@ -22,12 +22,6 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/object.h>
-#include <core/client.h>
-#include <core/class.h>
-
-#include <engine/fifo.h>
-
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
 #include "nouveau_fence.h"
diff --git a/drivers/gpu/drm/nouveau/nvif/class.h b/drivers/gpu/drm/nouveau/nvif/class.h
new file mode 100644
index 000000000000..e5a27df0672b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/class.h
@@ -0,0 +1,567 @@
+#ifndef __NVIF_CLASS_H__
+#define __NVIF_CLASS_H__
+
+/*******************************************************************************
+ * class identifiers
+ ******************************************************************************/
+
+/* the below match nvidia-assigned (either in hw, or sw) class numbers */
+#define NV_DEVICE                                                    0x00000080
+
+#define NV_DMA_FROM_MEMORY                                           0x00000002
+#define NV_DMA_TO_MEMORY                                             0x00000003
+#define NV_DMA_IN_MEMORY                                             0x0000003d
+
+#define NV04_DISP                                                    0x00000046
+
+#define NV03_CHANNEL_DMA                                             0x0000006b
+#define NV10_CHANNEL_DMA                                             0x0000006e
+#define NV17_CHANNEL_DMA                                             0x0000176e
+#define NV40_CHANNEL_DMA                                             0x0000406e
+#define NV50_CHANNEL_DMA                                             0x0000506e
+#define G82_CHANNEL_DMA                                              0x0000826e
+
+#define NV50_CHANNEL_GPFIFO                                          0x0000506f
+#define G82_CHANNEL_GPFIFO                                           0x0000826f
+#define FERMI_CHANNEL_GPFIFO                                         0x0000906f
+#define KEPLER_CHANNEL_GPFIFO_A                                      0x0000a06f
+
+#define NV50_DISP                                                    0x00005070
+#define G82_DISP                                                     0x00008270
+#define GT200_DISP                                                   0x00008370
+#define GT214_DISP                                                   0x00008570
+#define GT206_DISP                                                   0x00008870
+#define GF110_DISP                                                   0x00009070
+#define GK104_DISP                                                   0x00009170
+#define GK110_DISP                                                   0x00009270
+#define GM107_DISP                                                   0x00009470
+
+#define NV50_DISP_CURSOR                                             0x0000507a
+#define G82_DISP_CURSOR                                              0x0000827a
+#define GT214_DISP_CURSOR                                            0x0000857a
+#define GF110_DISP_CURSOR                                            0x0000907a
+#define GK104_DISP_CURSOR                                            0x0000917a
+
+#define NV50_DISP_OVERLAY                                            0x0000507b
+#define G82_DISP_OVERLAY                                             0x0000827b
+#define GT214_DISP_OVERLAY                                           0x0000857b
+#define GF110_DISP_OVERLAY                                           0x0000907b
+#define GK104_DISP_OVERLAY                                           0x0000917b
+
+#define NV50_DISP_BASE_CHANNEL_DMA                                   0x0000507c
+#define G82_DISP_BASE_CHANNEL_DMA                                    0x0000827c
+#define GT200_DISP_BASE_CHANNEL_DMA                                  0x0000837c
+#define GT214_DISP_BASE_CHANNEL_DMA                                  0x0000857c
+#define GF110_DISP_BASE_CHANNEL_DMA                                  0x0000907c
+#define GK104_DISP_BASE_CHANNEL_DMA                                  0x0000917c
+#define GK110_DISP_BASE_CHANNEL_DMA                                  0x0000927c
+
+#define NV50_DISP_CORE_CHANNEL_DMA                                   0x0000507d
+#define G82_DISP_CORE_CHANNEL_DMA                                    0x0000827d
+#define GT200_DISP_CORE_CHANNEL_DMA                                  0x0000837d
+#define GT214_DISP_CORE_CHANNEL_DMA                                  0x0000857d
+#define GT206_DISP_CORE_CHANNEL_DMA                                  0x0000887d
+#define GF110_DISP_CORE_CHANNEL_DMA                                  0x0000907d
+#define GK104_DISP_CORE_CHANNEL_DMA                                  0x0000917d
+#define GK110_DISP_CORE_CHANNEL_DMA                                  0x0000927d
+#define GM107_DISP_CORE_CHANNEL_DMA                                  0x0000947d
+
+#define NV50_DISP_OVERLAY_CHANNEL_DMA                                0x0000507e
+#define G82_DISP_OVERLAY_CHANNEL_DMA                                 0x0000827e
+#define GT200_DISP_OVERLAY_CHANNEL_DMA                               0x0000837e
+#define GT214_DISP_OVERLAY_CHANNEL_DMA                               0x0000857e
+#define GF110_DISP_OVERLAY_CONTROL_DMA                               0x0000907e
+#define GK104_DISP_OVERLAY_CONTROL_DMA                               0x0000917e
+
+#define FERMI_A                                                      0x00009097
+#define FERMI_B                                                      0x00009197
+#define FERMI_C                                                      0x00009297
+
+#define KEPLER_A                                                     0x0000a097
+#define KEPLER_B                                                     0x0000a197
+#define KEPLER_C                                                     0x0000a297
+
+#define MAXWELL_A                                                    0x0000b097
+
+#define FERMI_COMPUTE_A                                              0x000090c0
+#define FERMI_COMPUTE_B                                              0x000091c0
+
+#define KEPLER_COMPUTE_A                                             0x0000a0c0
+#define KEPLER_COMPUTE_B                                             0x0000a1c0
+
+#define MAXWELL_COMPUTE_A                                            0x0000b0c0
+
+
+/*******************************************************************************
+ * client
+ ******************************************************************************/
+
+#define NV_CLIENT_DEVLIST                                                  0x00
+
+struct nv_client_devlist_v0 {
+	__u8  version;
+	__u8  count;
+	__u8  pad02[6];
+	__u64 device[];
+};
+
+
+/*******************************************************************************
+ * device
+ ******************************************************************************/
+
+struct nv_device_v0 {
+	__u8  version;
+	__u8  pad01[7];
+	__u64 device;	/* device identifier, ~0 for client default */
+#define NV_DEVICE_V0_DISABLE_IDENTIFY                     0x0000000000000001ULL
+#define NV_DEVICE_V0_DISABLE_MMIO                         0x0000000000000002ULL
+#define NV_DEVICE_V0_DISABLE_VBIOS                        0x0000000000000004ULL
+#define NV_DEVICE_V0_DISABLE_CORE                         0x0000000000000008ULL
+#define NV_DEVICE_V0_DISABLE_DISP                         0x0000000000010000ULL
+#define NV_DEVICE_V0_DISABLE_FIFO                         0x0000000000020000ULL
+#define NV_DEVICE_V0_DISABLE_GRAPH                        0x0000000100000000ULL
+#define NV_DEVICE_V0_DISABLE_MPEG                         0x0000000200000000ULL
+#define NV_DEVICE_V0_DISABLE_ME                           0x0000000400000000ULL
+#define NV_DEVICE_V0_DISABLE_VP                           0x0000000800000000ULL
+#define NV_DEVICE_V0_DISABLE_CRYPT                        0x0000001000000000ULL
+#define NV_DEVICE_V0_DISABLE_BSP                          0x0000002000000000ULL
+#define NV_DEVICE_V0_DISABLE_PPP                          0x0000004000000000ULL
+#define NV_DEVICE_V0_DISABLE_COPY0                        0x0000008000000000ULL
+#define NV_DEVICE_V0_DISABLE_COPY1                        0x0000010000000000ULL
+#define NV_DEVICE_V0_DISABLE_VIC                          0x0000020000000000ULL
+#define NV_DEVICE_V0_DISABLE_VENC                         0x0000040000000000ULL
+	__u64 disable;	/* disable particular subsystems */
+	__u64 debug0;	/* as above, but *internal* ids, and *NOT* ABI */
+};
+
+#define NV_DEVICE_V0_INFO                                                  0x00
+
+struct nv_device_info_v0 {
+	__u8  version;
+#define NV_DEVICE_INFO_V0_IGP                                              0x00
+#define NV_DEVICE_INFO_V0_PCI                                              0x01
+#define NV_DEVICE_INFO_V0_AGP                                              0x02
+#define NV_DEVICE_INFO_V0_PCIE                                             0x03
+#define NV_DEVICE_INFO_V0_SOC                                              0x04
+	__u8  platform;
+	__u16 chipset;	/* from NV_PMC_BOOT_0 */
+	__u8  revision;	/* from NV_PMC_BOOT_0 */
+#define NV_DEVICE_INFO_V0_TNT                                              0x01
+#define NV_DEVICE_INFO_V0_CELSIUS                                          0x02
+#define NV_DEVICE_INFO_V0_KELVIN                                           0x03
+#define NV_DEVICE_INFO_V0_RANKINE                                          0x04
+#define NV_DEVICE_INFO_V0_CURIE                                            0x05
+#define NV_DEVICE_INFO_V0_TESLA                                            0x06
+#define NV_DEVICE_INFO_V0_FERMI                                            0x07
+#define NV_DEVICE_INFO_V0_KEPLER                                           0x08
+#define NV_DEVICE_INFO_V0_MAXWELL                                          0x09
+	__u8  family;
+	__u8  pad06[2];
+	__u64 ram_size;
+	__u64 ram_user;
+};
+
+
+/*******************************************************************************
+ * context dma
+ ******************************************************************************/
+
+struct nv_dma_v0 {
+	__u8  version;
+#define NV_DMA_V0_TARGET_VM                                                0x00
+#define NV_DMA_V0_TARGET_VRAM                                              0x01
+#define NV_DMA_V0_TARGET_PCI                                               0x02
+#define NV_DMA_V0_TARGET_PCI_US                                            0x03
+#define NV_DMA_V0_TARGET_AGP                                               0x04
+	__u8  target;
+#define NV_DMA_V0_ACCESS_VM                                                0x00
+#define NV_DMA_V0_ACCESS_RD                                                0x01
+#define NV_DMA_V0_ACCESS_WR                                                0x02
+#define NV_DMA_V0_ACCESS_RDWR                 (NV_DMA_V0_ACCESS_RD | NV_DMA_V0_ACCESS_WR)
+	__u8  access;
+	__u8  pad03[5];
+	__u64 start;
+	__u64 limit;
+	/* ... chipset-specific class data */
+};
+
+struct nv50_dma_v0 {
+	__u8  version;
+#define NV50_DMA_V0_PRIV_VM                                                0x00
+#define NV50_DMA_V0_PRIV_US                                                0x01
+#define NV50_DMA_V0_PRIV__S                                                0x02
+	__u8  priv;
+#define NV50_DMA_V0_PART_VM                                                0x00
+#define NV50_DMA_V0_PART_256                                               0x01
+#define NV50_DMA_V0_PART_1KB                                               0x02
+	__u8  part;
+#define NV50_DMA_V0_COMP_NONE                                              0x00
+#define NV50_DMA_V0_COMP_1                                                 0x01
+#define NV50_DMA_V0_COMP_2                                                 0x02
+#define NV50_DMA_V0_COMP_VM                                                0x03
+	__u8  comp;
+#define NV50_DMA_V0_KIND_PITCH                                             0x00
+#define NV50_DMA_V0_KIND_VM                                                0x7f
+	__u8  kind;
+	__u8  pad05[3];
+};
+
+struct gf100_dma_v0 {
+	__u8  version;
+#define GF100_DMA_V0_PRIV_VM                                               0x00
+#define GF100_DMA_V0_PRIV_US                                               0x01
+#define GF100_DMA_V0_PRIV__S                                               0x02
+	__u8  priv;
+#define GF100_DMA_V0_KIND_PITCH                                            0x00
+#define GF100_DMA_V0_KIND_VM                                               0xff
+	__u8  kind;
+	__u8  pad03[5];
+};
+
+struct gf110_dma_v0 {
+	__u8  version;
+#define GF110_DMA_V0_PAGE_LP                                               0x00
+#define GF110_DMA_V0_PAGE_SP                                               0x01
+	__u8  page;
+#define GF110_DMA_V0_KIND_PITCH                                            0x00
+#define GF110_DMA_V0_KIND_VM                                               0xff
+	__u8  kind;
+	__u8  pad03[5];
+};
+
+
+/*******************************************************************************
+ * perfmon
+ ******************************************************************************/
+
+struct nvif_perfctr_v0 {
+	__u8  version;
+	__u8  pad01[1];
+	__u16 logic_op;
+	__u8  pad04[4];
+	char  name[4][64];
+};
+
+#define NVIF_PERFCTR_V0_QUERY                                              0x00
+#define NVIF_PERFCTR_V0_SAMPLE                                             0x01
+#define NVIF_PERFCTR_V0_READ                                               0x02
+
+struct nvif_perfctr_query_v0 {
+	__u8  version;
+	__u8  pad01[3];
+	__u32 iter;
+	char  name[64];
+};
+
+struct nvif_perfctr_sample {
+};
+
+struct nvif_perfctr_read_v0 {
+	__u8  version;
+	__u8  pad01[7];
+	__u32 ctr;
+	__u32 clk;
+};
+
+
+/*******************************************************************************
+ * device control
+ ******************************************************************************/
+
+#define NVIF_CONTROL_PSTATE_INFO                                           0x00
+#define NVIF_CONTROL_PSTATE_ATTR                                           0x01
+#define NVIF_CONTROL_PSTATE_USER                                           0x02
+
+struct nvif_control_pstate_info_v0 {
+	__u8  version;
+	__u8  count; /* out: number of power states */
+#define NVIF_CONTROL_PSTATE_INFO_V0_USTATE_DISABLE                         (-1)
+#define NVIF_CONTROL_PSTATE_INFO_V0_USTATE_PERFMON                         (-2)
+	__s8  ustate_ac; /* out: target pstate index */
+	__s8  ustate_dc; /* out: target pstate index */
+	__s8  pwrsrc; /* out: current power source */
+#define NVIF_CONTROL_PSTATE_INFO_V0_PSTATE_UNKNOWN                         (-1)
+#define NVIF_CONTROL_PSTATE_INFO_V0_PSTATE_PERFMON                         (-2)
+	__s8  pstate; /* out: current pstate index */
+	__u8  pad06[2];
+};
+
+struct nvif_control_pstate_attr_v0 {
+	__u8  version;
+#define NVIF_CONTROL_PSTATE_ATTR_V0_STATE_CURRENT                          (-1)
+	__s8  state; /*  in: index of pstate to query
+		      * out: pstate identifier
+		      */
+	__u8  index; /*  in: index of attribute to query
+		      * out: index of next attribute, or 0 if no more
+		      */
+	__u8  pad03[5];
+	__u32 min;
+	__u32 max;
+	char  name[32];
+	char  unit[16];
+};
+
+struct nvif_control_pstate_user_v0 {
+	__u8  version;
+#define NVIF_CONTROL_PSTATE_USER_V0_STATE_UNKNOWN                          (-1)
+#define NVIF_CONTROL_PSTATE_USER_V0_STATE_PERFMON                          (-2)
+	__s8  ustate; /*  in: pstate identifier */
+	__s8  pwrsrc; /*  in: target power source */
+	__u8  pad03[5];
+};
+
+
+/*******************************************************************************
+ * DMA FIFO channels
+ ******************************************************************************/
+
+struct nv03_channel_dma_v0 {
+	__u8  version;
+	__u8  chid;
+	__u8  pad02[2];
+	__u32 pushbuf;
+	__u64 offset;
+};
+
+#define G82_CHANNEL_DMA_V0_NTFY_UEVENT                                     0x00
+
+/*******************************************************************************
+ * GPFIFO channels
+ ******************************************************************************/
+
+struct nv50_channel_gpfifo_v0 {
+	__u8  version;
+	__u8  chid;
+	__u8  pad01[6];
+	__u32 pushbuf;
+	__u32 ilength;
+	__u64 ioffset;
+};
+
+struct kepler_channel_gpfifo_a_v0 {
+	__u8  version;
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_GR                               0x01
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_VP                               0x02
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_PPP                              0x04
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_BSP                              0x08
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_CE0                              0x10
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_CE1                              0x20
+#define KEPLER_CHANNEL_GPFIFO_A_V0_ENGINE_ENC                              0x40
+	__u8  engine;
+	__u16 chid;
+	__u8  pad04[4];
+	__u32 pushbuf;
+	__u32 ilength;
+	__u64 ioffset;
+};
+
+/*******************************************************************************
+ * legacy display
+ ******************************************************************************/
+
+#define NV04_DISP_NTFY_VBLANK                                              0x00
+#define NV04_DISP_NTFY_CONN                                                0x01
+
+struct nv04_disp_mthd_v0 {
+	__u8  version;
+#define NV04_DISP_SCANOUTPOS                                               0x00
+	__u8  method;
+	__u8  head;
+	__u8  pad03[5];
+};
+
+struct nv04_disp_scanoutpos_v0 {
+	__u8  version;
+	__u8  pad01[7];
+	__s64 time[2];
+	__u16 vblanks;
+	__u16 vblanke;
+	__u16 vtotal;
+	__u16 vline;
+	__u16 hblanks;
+	__u16 hblanke;
+	__u16 htotal;
+	__u16 hline;
+};
+
+/*******************************************************************************
+ * display
+ ******************************************************************************/
+
+#define NV50_DISP_MTHD                                                     0x00
+
+struct nv50_disp_mthd_v0 {
+	__u8  version;
+#define NV50_DISP_SCANOUTPOS                                               0x00
+	__u8  method;
+	__u8  head;
+	__u8  pad03[5];
+};
+
+struct nv50_disp_mthd_v1 {
+	__u8  version;
+#define NV50_DISP_MTHD_V1_DAC_PWR                                          0x10
+#define NV50_DISP_MTHD_V1_DAC_LOAD                                         0x11
+#define NV50_DISP_MTHD_V1_SOR_PWR                                          0x20
+#define NV50_DISP_MTHD_V1_SOR_HDA_ELD                                      0x21
+#define NV50_DISP_MTHD_V1_SOR_HDMI_PWR                                     0x22
+#define NV50_DISP_MTHD_V1_SOR_LVDS_SCRIPT                                  0x23
+#define NV50_DISP_MTHD_V1_SOR_DP_PWR                                       0x24
+#define NV50_DISP_MTHD_V1_PIOR_PWR                                         0x30
+	__u8  method;
+	__u16 hasht;
+	__u16 hashm;
+	__u8  pad06[2];
+};
+
+struct nv50_disp_dac_pwr_v0 {
+	__u8  version;
+	__u8  state;
+	__u8  data;
+	__u8  vsync;
+	__u8  hsync;
+	__u8  pad05[3];
+};
+
+struct nv50_disp_dac_load_v0 {
+	__u8  version;
+	__u8  load;
+	__u8  pad02[2];
+	__u32 data;
+};
+
+struct nv50_disp_sor_pwr_v0 {
+	__u8  version;
+	__u8  state;
+	__u8  pad02[6];
+};
+
+struct nv50_disp_sor_hda_eld_v0 {
+	__u8  version;
+	__u8  pad01[7];
+	__u8  data[];
+};
+
+struct nv50_disp_sor_hdmi_pwr_v0 {
+	__u8  version;
+	__u8  state;
+	__u8  max_ac_packet;
+	__u8  rekey;
+	__u8  pad04[4];
+};
+
+struct nv50_disp_sor_lvds_script_v0 {
+	__u8  version;
+	__u8  pad01[1];
+	__u16 script;
+	__u8  pad04[4];
+};
+
+struct nv50_disp_sor_dp_pwr_v0 {
+	__u8  version;
+	__u8  state;
+	__u8  pad02[6];
+};
+
+struct nv50_disp_pior_pwr_v0 {
+	__u8  version;
+	__u8  state;
+	__u8  type;
+	__u8  pad03[5];
+};
+
+/* core */
+struct nv50_disp_core_channel_dma_v0 {
+	__u8  version;
+	__u8  pad01[3];
+	__u32 pushbuf;
+};
+
+#define NV50_DISP_CORE_CHANNEL_DMA_V0_NTFY_UEVENT                          0x00
+
+/* cursor immediate */
+struct nv50_disp_cursor_v0 {
+	__u8  version;
+	__u8  head;
+	__u8  pad02[6];
+};
+
+#define NV50_DISP_CURSOR_V0_NTFY_UEVENT                                    0x00
+
+/* base */
+struct nv50_disp_base_channel_dma_v0 {
+	__u8  version;
+	__u8  pad01[2];
+	__u8  head;
+	__u32 pushbuf;
+};
+
+#define NV50_DISP_BASE_CHANNEL_DMA_V0_NTFY_UEVENT                          0x00
+
+/* overlay */
+struct nv50_disp_overlay_channel_dma_v0 {
+	__u8  version;
+	__u8  pad01[2];
+	__u8  head;
+	__u32 pushbuf;
+};
+
+#define NV50_DISP_OVERLAY_CHANNEL_DMA_V0_NTFY_UEVENT                       0x00
+
+/* overlay immediate */
+struct nv50_disp_overlay_v0 {
+	__u8  version;
+	__u8  head;
+	__u8  pad02[6];
+};
+
+#define NV50_DISP_OVERLAY_V0_NTFY_UEVENT                                   0x00
+
+/*******************************************************************************
+ * fermi
+ ******************************************************************************/
+
+#define FERMI_A_ZBC_COLOR                                                  0x00
+#define FERMI_A_ZBC_DEPTH                                                  0x01
+
+struct fermi_a_zbc_color_v0 {
+	__u8  version;
+#define FERMI_A_ZBC_COLOR_V0_FMT_ZERO                                      0x01
+#define FERMI_A_ZBC_COLOR_V0_FMT_UNORM_ONE                                 0x02
+#define FERMI_A_ZBC_COLOR_V0_FMT_RF32_GF32_BF32_AF32                       0x04
+#define FERMI_A_ZBC_COLOR_V0_FMT_R16_G16_B16_A16                           0x08
+#define FERMI_A_ZBC_COLOR_V0_FMT_RN16_GN16_BN16_AN16                       0x0c
+#define FERMI_A_ZBC_COLOR_V0_FMT_RS16_GS16_BS16_AS16                       0x10
+#define FERMI_A_ZBC_COLOR_V0_FMT_RU16_GU16_BU16_AU16                       0x14
+#define FERMI_A_ZBC_COLOR_V0_FMT_RF16_GF16_BF16_AF16                       0x16
+#define FERMI_A_ZBC_COLOR_V0_FMT_A8R8G8B8                                  0x18
+#define FERMI_A_ZBC_COLOR_V0_FMT_A8RL8GL8BL8                               0x1c
+#define FERMI_A_ZBC_COLOR_V0_FMT_A2B10G10R10                               0x20
+#define FERMI_A_ZBC_COLOR_V0_FMT_AU2BU10GU10RU10                           0x24
+#define FERMI_A_ZBC_COLOR_V0_FMT_A8B8G8R8                                  0x28
+#define FERMI_A_ZBC_COLOR_V0_FMT_A8BL8GL8RL8                               0x2c
+#define FERMI_A_ZBC_COLOR_V0_FMT_AN8BN8GN8RN8                              0x30
+#define FERMI_A_ZBC_COLOR_V0_FMT_AS8BS8GS8RS8                              0x34
+#define FERMI_A_ZBC_COLOR_V0_FMT_AU8BU8GU8RU8                              0x38
+#define FERMI_A_ZBC_COLOR_V0_FMT_A2R10G10B10                               0x3c
+#define FERMI_A_ZBC_COLOR_V0_FMT_BF10GF11RF11                              0x40
+	__u8  format;
+	__u8  index;
+	__u8  pad03[5];
+	__u32 ds[4];
+	__u32 l2[4];
+};
+
+struct fermi_a_zbc_depth_v0 {
+	__u8  version;
+#define FERMI_A_ZBC_DEPTH_V0_FMT_FP32                                      0x01
+	__u8  format;
+	__u8  index;
+	__u8  pad03[5];
+	__u32 ds;
+	__u32 l2;
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/client.c b/drivers/gpu/drm/nouveau/nvif/client.c
new file mode 100644
index 000000000000..3c4df1fc26dc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/client.c
@@ -0,0 +1,129 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "client.h"
+#include "driver.h"
+#include "ioctl.h"
+
+int
+nvif_client_ioctl(struct nvif_client *client, void *data, u32 size)
+{
+	return client->driver->ioctl(client->base.priv, client->super, data, size, NULL);
+}
+
+int
+nvif_client_suspend(struct nvif_client *client)
+{
+	return client->driver->suspend(client->base.priv);
+}
+
+int
+nvif_client_resume(struct nvif_client *client)
+{
+	return client->driver->resume(client->base.priv);
+}
+
+void
+nvif_client_fini(struct nvif_client *client)
+{
+	if (client->driver) {
+		client->driver->fini(client->base.priv);
+		client->driver = NULL;
+		client->base.parent = NULL;
+		nvif_object_fini(&client->base);
+	}
+}
+
+const struct nvif_driver *
+nvif_drivers[] = {
+#ifdef __KERNEL__
+	&nvif_driver_nvkm,
+#else
+	&nvif_driver_drm,
+	&nvif_driver_lib,
+#endif
+	NULL
+};
+
+int
+nvif_client_init(void (*dtor)(struct nvif_client *), const char *driver,
+		 const char *name, u64 device, const char *cfg, const char *dbg,
+		 struct nvif_client *client)
+{
+	int ret, i;
+
+	ret = nvif_object_init(NULL, (void*)dtor, 0, 0, NULL, 0, &client->base);
+	if (ret)
+		return ret;
+
+	client->base.parent = &client->base;
+	client->base.handle = ~0;
+	client->object = &client->base;
+	client->super = true;
+
+	for (i = 0, ret = -EINVAL; (client->driver = nvif_drivers[i]); i++) {
+		if (!driver || !strcmp(client->driver->name, driver)) {
+			ret = client->driver->init(name, device, cfg, dbg,
+						  &client->base.priv);
+			if (!ret || driver)
+				break;
+		}
+	}
+
+	if (ret)
+		nvif_client_fini(client);
+	return ret;
+}
+
+static void
+nvif_client_del(struct nvif_client *client)
+{
+	nvif_client_fini(client);
+	kfree(client);
+}
+
+int
+nvif_client_new(const char *driver, const char *name, u64 device,
+		const char *cfg, const char *dbg,
+		struct nvif_client **pclient)
+{
+	struct nvif_client *client = kzalloc(sizeof(*client), GFP_KERNEL);
+	if (client) {
+		int ret = nvif_client_init(nvif_client_del, driver, name,
+					   device, cfg, dbg, client);
+		if (ret) {
+			kfree(client);
+			client = NULL;
+		}
+		*pclient = client;
+		return ret;
+	}
+	return -ENOMEM;
+}
+
+void
+nvif_client_ref(struct nvif_client *client, struct nvif_client **pclient)
+{
+	nvif_object_ref(&client->base, (struct nvif_object **)pclient);
+}
diff --git a/drivers/gpu/drm/nouveau/nvif/client.h b/drivers/gpu/drm/nouveau/nvif/client.h
new file mode 100644
index 000000000000..28352f0882ec
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/client.h
@@ -0,0 +1,39 @@
+#ifndef __NVIF_CLIENT_H__
+#define __NVIF_CLIENT_H__
+
+#include "object.h"
+
+struct nvif_client {
+	struct nvif_object base;
+	struct nvif_object *object; /*XXX: hack for nvif_object() */
+	const struct nvif_driver *driver;
+	bool super;
+};
+
+static inline struct nvif_client *
+nvif_client(struct nvif_object *object)
+{
+	while (object && object->parent != object)
+		object = object->parent;
+	return (void *)object;
+}
+
+int  nvif_client_init(void (*dtor)(struct nvif_client *), const char *,
+		      const char *, u64, const char *, const char *,
+		      struct nvif_client *);
+void nvif_client_fini(struct nvif_client *);
+int  nvif_client_new(const char *, const char *, u64, const char *,
+		     const char *, struct nvif_client **);
+void nvif_client_ref(struct nvif_client *, struct nvif_client **);
+int  nvif_client_ioctl(struct nvif_client *, void *, u32);
+int  nvif_client_suspend(struct nvif_client *);
+int  nvif_client_resume(struct nvif_client *);
+
+/*XXX*/
+#include <core/client.h>
+#define nvkm_client(a) ({ \
+	struct nvif_client *_client = nvif_client(nvif_object(a)); \
+	nouveau_client(_client->base.priv); \
+})
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/device.c b/drivers/gpu/drm/nouveau/nvif/device.c
new file mode 100644
index 000000000000..f477579725e3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/device.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "device.h"
+
+void
+nvif_device_fini(struct nvif_device *device)
+{
+	nvif_object_fini(&device->base);
+}
+
+int
+nvif_device_init(struct nvif_object *parent, void (*dtor)(struct nvif_device *),
+		 u32 handle, u32 oclass, void *data, u32 size,
+		 struct nvif_device *device)
+{
+	int ret = nvif_object_init(parent, (void *)dtor, handle, oclass,
+				   data, size, &device->base);
+	if (ret == 0) {
+		device->object = &device->base;
+		device->info.version = 0;
+		ret = nvif_object_mthd(&device->base, NV_DEVICE_V0_INFO,
+				       &device->info, sizeof(device->info));
+	}
+	return ret;
+}
+
+static void
+nvif_device_del(struct nvif_device *device)
+{
+	nvif_device_fini(device);
+	kfree(device);
+}
+
+int
+nvif_device_new(struct nvif_object *parent, u32 handle, u32 oclass,
+		void *data, u32 size, struct nvif_device **pdevice)
+{
+	struct nvif_device *device = kzalloc(sizeof(*device), GFP_KERNEL);
+	if (device) {
+		int ret = nvif_device_init(parent, nvif_device_del, handle,
+					   oclass, data, size, device);
+		if (ret) {
+			kfree(device);
+			device = NULL;
+		}
+		*pdevice = device;
+		return ret;
+	}
+	return -ENOMEM;
+}
+
+void
+nvif_device_ref(struct nvif_device *device, struct nvif_device **pdevice)
+{
+	nvif_object_ref(&device->base, (struct nvif_object **)pdevice);
+}
diff --git a/drivers/gpu/drm/nouveau/nvif/device.h b/drivers/gpu/drm/nouveau/nvif/device.h
new file mode 100644
index 000000000000..43180f9fe630
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/device.h
@@ -0,0 +1,62 @@
+#ifndef __NVIF_DEVICE_H__
+#define __NVIF_DEVICE_H__
+
+#include "object.h"
+#include "class.h"
+
+struct nvif_device {
+	struct nvif_object base;
+	struct nvif_object *object; /*XXX: hack for nvif_object() */
+	struct nv_device_info_v0 info;
+};
+
+static inline struct nvif_device *
+nvif_device(struct nvif_object *object)
+{
+	while (object && object->oclass != 0x0080 /*XXX: NV_DEVICE_CLASS*/ )
+		object = object->parent;
+	return (void *)object;
+}
+
+int  nvif_device_init(struct nvif_object *, void (*dtor)(struct nvif_device *),
+		      u32 handle, u32 oclass, void *, u32,
+		      struct nvif_device *);
+void nvif_device_fini(struct nvif_device *);
+int  nvif_device_new(struct nvif_object *, u32 handle, u32 oclass,
+		     void *, u32, struct nvif_device **);
+void nvif_device_ref(struct nvif_device *, struct nvif_device **);
+
+/*XXX*/
+#include <subdev/bios.h>
+#include <subdev/fb.h>
+#include <subdev/vm.h>
+#include <subdev/bar.h>
+#include <subdev/gpio.h>
+#include <subdev/clock.h>
+#include <subdev/i2c.h>
+#include <subdev/timer.h>
+#include <subdev/therm.h>
+
+#define nvkm_device(a) nv_device(nvkm_object((a)))
+#define nvkm_bios(a) nouveau_bios(nvkm_device(a))
+#define nvkm_fb(a) nouveau_fb(nvkm_device(a))
+#define nvkm_vmmgr(a) nouveau_vmmgr(nvkm_device(a))
+#define nvkm_bar(a) nouveau_bar(nvkm_device(a))
+#define nvkm_gpio(a) nouveau_gpio(nvkm_device(a))
+#define nvkm_clock(a) nouveau_clock(nvkm_device(a))
+#define nvkm_i2c(a) nouveau_i2c(nvkm_device(a))
+#define nvkm_timer(a) nouveau_timer(nvkm_device(a))
+#define nvkm_wait(a,b,c,d) nv_wait(nvkm_timer(a), (b), (c), (d))
+#define nvkm_wait_cb(a,b,c) nv_wait_cb(nvkm_timer(a), (b), (c))
+#define nvkm_therm(a) nouveau_therm(nvkm_device(a))
+
+#include <engine/device.h>
+#include <engine/fifo.h>
+#include <engine/graph.h>
+#include <engine/software.h>
+
+#define nvkm_fifo(a) nouveau_fifo(nvkm_device(a))
+#define nvkm_fifo_chan(a) ((struct nouveau_fifo_chan *)nvkm_object(a))
+#define nvkm_gr(a) ((struct nouveau_graph *)nouveau_engine(nvkm_object(a), NVDEV_ENGINE_GR))
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/driver.h b/drivers/gpu/drm/nouveau/nvif/driver.h
new file mode 100644
index 000000000000..ac4bdb3ea506
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/driver.h
@@ -0,0 +1,21 @@
+#ifndef __NVIF_DRIVER_H__
+#define __NVIF_DRIVER_H__
+
+struct nvif_driver {
+	const char *name;
+	int (*init)(const char *name, u64 device, const char *cfg,
+		    const char *dbg, void **priv);
+	void (*fini)(void *priv);
+	int (*suspend)(void *priv);
+	int (*resume)(void *priv);
+	int (*ioctl)(void *priv, bool super, void *data, u32 size, void **hack);
+	void __iomem *(*map)(void *priv, u64 handle, u32 size);
+	void (*unmap)(void *priv, void __iomem *ptr, u32 size);
+	bool keep;
+};
+
+extern const struct nvif_driver nvif_driver_nvkm;
+extern const struct nvif_driver nvif_driver_drm;
+extern const struct nvif_driver nvif_driver_lib;
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/event.h b/drivers/gpu/drm/nouveau/nvif/event.h
new file mode 100644
index 000000000000..21764499b4be
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/event.h
@@ -0,0 +1,62 @@
+#ifndef __NVIF_EVENT_H__
+#define __NVIF_EVENT_H__
+
+struct nvif_notify_req_v0 {
+	__u8  version;
+	__u8  reply;
+	__u8  pad02[5];
+#define NVIF_NOTIFY_V0_ROUTE_NVIF                                          0x00
+	__u8  route;
+	__u64 token;	/* must be unique */
+	__u8  data[];	/* request data (below) */
+};
+
+struct nvif_notify_rep_v0 {
+	__u8  version;
+	__u8  pad01[6];
+	__u8  route;
+	__u64 token;
+	__u8  data[];	/* reply data (below) */
+};
+
+struct nvif_notify_head_req_v0 {
+	/* nvif_notify_req ... */
+	__u8  version;
+	__u8  head;
+	__u8  pad02[6];
+};
+
+struct nvif_notify_head_rep_v0 {
+	/* nvif_notify_rep ... */
+	__u8  version;
+	__u8  pad01[7];
+};
+
+struct nvif_notify_conn_req_v0 {
+	/* nvif_notify_req ... */
+	__u8  version;
+#define NVIF_NOTIFY_CONN_V0_PLUG                                           0x01
+#define NVIF_NOTIFY_CONN_V0_UNPLUG                                         0x02
+#define NVIF_NOTIFY_CONN_V0_IRQ                                            0x04
+#define NVIF_NOTIFY_CONN_V0_ANY                                            0x07
+	__u8  mask;
+	__u8  conn;
+	__u8  pad03[5];
+};
+
+struct nvif_notify_conn_rep_v0 {
+	/* nvif_notify_rep ... */
+	__u8  version;
+	__u8  mask;
+	__u8  pad02[6];
+};
+
+struct nvif_notify_uevent_req {
+	/* nvif_notify_req ... */
+};
+
+struct nvif_notify_uevent_rep {
+	/* nvif_notify_rep ... */
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/ioctl.h b/drivers/gpu/drm/nouveau/nvif/ioctl.h
new file mode 100644
index 000000000000..4cd8e323b23d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/ioctl.h
@@ -0,0 +1,128 @@
+#ifndef __NVIF_IOCTL_H__
+#define __NVIF_IOCTL_H__
+
+struct nvif_ioctl_v0 {
+	__u8  version;
+#define NVIF_IOCTL_V0_OWNER_NVIF                                           0x00
+#define NVIF_IOCTL_V0_OWNER_ANY                                            0xff
+	__u8  owner;
+#define NVIF_IOCTL_V0_NOP                                                  0x00
+#define NVIF_IOCTL_V0_SCLASS                                               0x01
+#define NVIF_IOCTL_V0_NEW                                                  0x02
+#define NVIF_IOCTL_V0_DEL                                                  0x03
+#define NVIF_IOCTL_V0_MTHD                                                 0x04
+#define NVIF_IOCTL_V0_RD                                                   0x05
+#define NVIF_IOCTL_V0_WR                                                   0x06
+#define NVIF_IOCTL_V0_MAP                                                  0x07
+#define NVIF_IOCTL_V0_UNMAP                                                0x08
+#define NVIF_IOCTL_V0_NTFY_NEW                                             0x09
+#define NVIF_IOCTL_V0_NTFY_DEL                                             0x0a
+#define NVIF_IOCTL_V0_NTFY_GET                                             0x0b
+#define NVIF_IOCTL_V0_NTFY_PUT                                             0x0c
+	__u8  type;
+	__u8  path_nr;
+#define NVIF_IOCTL_V0_ROUTE_NVIF                                           0x00
+#define NVIF_IOCTL_V0_ROUTE_HIDDEN                                         0xff
+	__u8  pad04[3];
+	__u8  route;
+	__u64 token;
+	__u32 path[8];		/* in reverse */
+	__u8  data[];		/* ioctl data (below) */
+};
+
+struct nvif_ioctl_nop {
+};
+
+struct nvif_ioctl_sclass_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  count;
+	__u8  pad02[6];
+	__u32 oclass[];
+};
+
+struct nvif_ioctl_new_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  pad01[6];
+	__u8  route;
+	__u64 token;
+	__u32 handle;
+/* these class numbers are made up by us, and not nvidia-assigned */
+#define NVIF_IOCTL_NEW_V0_PERFCTR                                    0x0000ffff
+#define NVIF_IOCTL_NEW_V0_CONTROL                                    0x0000fffe
+	__u32 oclass;
+	__u8  data[];		/* class data (class.h) */
+};
+
+struct nvif_ioctl_del {
+};
+
+struct nvif_ioctl_rd_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  size;
+	__u8  pad02[2];
+	__u32 data;
+	__u64 addr;
+};
+
+struct nvif_ioctl_wr_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  size;
+	__u8  pad02[2];
+	__u32 data;
+	__u64 addr;
+};
+
+struct nvif_ioctl_map_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  pad01[3];
+	__u32 length;
+	__u64 handle;
+};
+
+struct nvif_ioctl_unmap {
+};
+
+struct nvif_ioctl_ntfy_new_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  event;
+	__u8  index;
+	__u8  pad03[5];
+	__u8  data[];		/* event request data (event.h) */
+};
+
+struct nvif_ioctl_ntfy_del_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  index;
+	__u8  pad02[6];
+};
+
+struct nvif_ioctl_ntfy_get_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  index;
+	__u8  pad02[6];
+};
+
+struct nvif_ioctl_ntfy_put_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  index;
+	__u8  pad02[6];
+};
+
+struct nvif_ioctl_mthd_v0 {
+	/* nvif_ioctl ... */
+	__u8  version;
+	__u8  method;
+	__u8  pad02[6];
+	__u8  data[];		/* method data (class.h) */
+};
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/list.h b/drivers/gpu/drm/nouveau/nvif/list.h
new file mode 100644
index 000000000000..8af5d144ecb0
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/list.h
@@ -0,0 +1,353 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ * Copyright © 2010 Francisco Jerez <currojerez@riseup.net>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+/* Modified by Ben Skeggs <bskeggs@redhat.com> to match kernel list APIs */
+
+#ifndef _XORG_LIST_H_
+#define _XORG_LIST_H_
+
+/**
+ * @file Classic doubly-link circular list implementation.
+ * For real usage examples of the linked list, see the file test/list.c
+ *
+ * Example:
+ * We need to keep a list of struct foo in the parent struct bar, i.e. what
+ * we want is something like this.
+ *
+ *     struct bar {
+ *          ...
+ *          struct foo *list_of_foos; -----> struct foo {}, struct foo {}, struct foo{}
+ *          ...
+ *     }
+ *
+ * We need one list head in bar and a list element in all list_of_foos (both are of
+ * data type 'struct list_head').
+ *
+ *     struct bar {
+ *          ...
+ *          struct list_head list_of_foos;
+ *          ...
+ *     }
+ *
+ *     struct foo {
+ *          ...
+ *          struct list_head entry;
+ *          ...
+ *     }
+ *
+ * Now we initialize the list head:
+ *
+ *     struct bar bar;
+ *     ...
+ *     INIT_LIST_HEAD(&bar.list_of_foos);
+ *
+ * Then we create the first element and add it to this list:
+ *
+ *     struct foo *foo = malloc(...);
+ *     ....
+ *     list_add(&foo->entry, &bar.list_of_foos);
+ *
+ * Repeat the above for each element you want to add to the list. Deleting
+ * works with the element itself.
+ *      list_del(&foo->entry);
+ *      free(foo);
+ *
+ * Note: calling list_del(&bar.list_of_foos) will set bar.list_of_foos to an empty
+ * list again.
+ *
+ * Looping through the list requires a 'struct foo' as iterator and the
+ * name of the field the subnodes use.
+ *
+ * struct foo *iterator;
+ * list_for_each_entry(iterator, &bar.list_of_foos, entry) {
+ *      if (iterator->something == ...)
+ *             ...
+ * }
+ *
+ * Note: You must not call list_del() on the iterator if you continue the
+ * loop. You need to run the safe for-each loop instead:
+ *
+ * struct foo *iterator, *next;
+ * list_for_each_entry_safe(iterator, next, &bar.list_of_foos, entry) {
+ *      if (...)
+ *              list_del(&iterator->entry);
+ * }
+ *
+ */
+
+/**
+ * The linkage struct for list nodes. This struct must be part of your
+ * to-be-linked struct. struct list_head is required for both the head of the
+ * list and for each list node.
+ *
+ * Position and name of the struct list_head field is irrelevant.
+ * There are no requirements that elements of a list are of the same type.
+ * There are no requirements for a list head, any struct list_head can be a list
+ * head.
+ */
+struct list_head {
+    struct list_head *next, *prev;
+};
+
+/**
+ * Initialize the list as an empty list.
+ *
+ * Example:
+ * INIT_LIST_HEAD(&bar->list_of_foos);
+ *
+ * @param The list to initialized.
+ */
+#define LIST_HEAD_INIT(name) { &(name), &(name) }
+
+#define LIST_HEAD(name) \
+	struct list_head name = LIST_HEAD_INIT(name)
+
+static inline void
+INIT_LIST_HEAD(struct list_head *list)
+{
+    list->next = list->prev = list;
+}
+
+static inline void
+__list_add(struct list_head *entry,
+                struct list_head *prev, struct list_head *next)
+{
+    next->prev = entry;
+    entry->next = next;
+    entry->prev = prev;
+    prev->next = entry;
+}
+
+/**
+ * Insert a new element after the given list head. The new element does not
+ * need to be initialised as empty list.
+ * The list changes from:
+ *      head → some element → ...
+ * to
+ *      head → new element → older element → ...
+ *
+ * Example:
+ * struct foo *newfoo = malloc(...);
+ * list_add(&newfoo->entry, &bar->list_of_foos);
+ *
+ * @param entry The new element to prepend to the list.
+ * @param head The existing list.
+ */
+static inline void
+list_add(struct list_head *entry, struct list_head *head)
+{
+    __list_add(entry, head, head->next);
+}
+
+/**
+ * Append a new element to the end of the list given with this list head.
+ *
+ * The list changes from:
+ *      head → some element → ... → lastelement
+ * to
+ *      head → some element → ... → lastelement → new element
+ *
+ * Example:
+ * struct foo *newfoo = malloc(...);
+ * list_add_tail(&newfoo->entry, &bar->list_of_foos);
+ *
+ * @param entry The new element to prepend to the list.
+ * @param head The existing list.
+ */
+static inline void
+list_add_tail(struct list_head *entry, struct list_head *head)
+{
+    __list_add(entry, head->prev, head);
+}
+
+static inline void
+__list_del(struct list_head *prev, struct list_head *next)
+{
+    next->prev = prev;
+    prev->next = next;
+}
+
+/**
+ * Remove the element from the list it is in. Using this function will reset
+ * the pointers to/from this element so it is removed from the list. It does
+ * NOT free the element itself or manipulate it otherwise.
+ *
+ * Using list_del on a pure list head (like in the example at the top of
+ * this file) will NOT remove the first element from
+ * the list but rather reset the list as empty list.
+ *
+ * Example:
+ * list_del(&foo->entry);
+ *
+ * @param entry The element to remove.
+ */
+static inline void
+list_del(struct list_head *entry)
+{
+    __list_del(entry->prev, entry->next);
+}
+
+static inline void
+list_del_init(struct list_head *entry)
+{
+    __list_del(entry->prev, entry->next);
+    INIT_LIST_HEAD(entry);
+}
+
+static inline void list_move_tail(struct list_head *list,
+				  struct list_head *head)
+{
+	__list_del(list->prev, list->next);
+	list_add_tail(list, head);
+}
+
+/**
+ * Check if the list is empty.
+ *
+ * Example:
+ * list_empty(&bar->list_of_foos);
+ *
+ * @return True if the list contains one or more elements or False otherwise.
+ */
+static inline bool
+list_empty(struct list_head *head)
+{
+    return head->next == head;
+}
+
+/**
+ * Returns a pointer to the container of this list element.
+ *
+ * Example:
+ * struct foo* f;
+ * f = container_of(&foo->entry, struct foo, entry);
+ * assert(f == foo);
+ *
+ * @param ptr Pointer to the struct list_head.
+ * @param type Data type of the list element.
+ * @param member Member name of the struct list_head field in the list element.
+ * @return A pointer to the data struct containing the list head.
+ */
+#ifndef container_of
+#define container_of(ptr, type, member) \
+    (type *)((char *)(ptr) - (char *) &((type *)0)->member)
+#endif
+
+/**
+ * Alias of container_of
+ */
+#define list_entry(ptr, type, member) \
+    container_of(ptr, type, member)
+
+/**
+ * Retrieve the first list entry for the given list pointer.
+ *
+ * Example:
+ * struct foo *first;
+ * first = list_first_entry(&bar->list_of_foos, struct foo, list_of_foos);
+ *
+ * @param ptr The list head
+ * @param type Data type of the list element to retrieve
+ * @param member Member name of the struct list_head field in the list element.
+ * @return A pointer to the first list element.
+ */
+#define list_first_entry(ptr, type, member) \
+    list_entry((ptr)->next, type, member)
+
+/**
+ * Retrieve the last list entry for the given listpointer.
+ *
+ * Example:
+ * struct foo *first;
+ * first = list_last_entry(&bar->list_of_foos, struct foo, list_of_foos);
+ *
+ * @param ptr The list head
+ * @param type Data type of the list element to retrieve
+ * @param member Member name of the struct list_head field in the list element.
+ * @return A pointer to the last list element.
+ */
+#define list_last_entry(ptr, type, member) \
+    list_entry((ptr)->prev, type, member)
+
+#define __container_of(ptr, sample, member)				\
+    (void *)container_of((ptr), typeof(*(sample)), member)
+
+/**
+ * Loop through the list given by head and set pos to struct in the list.
+ *
+ * Example:
+ * struct foo *iterator;
+ * list_for_each_entry(iterator, &bar->list_of_foos, entry) {
+ *      [modify iterator]
+ * }
+ *
+ * This macro is not safe for node deletion. Use list_for_each_entry_safe
+ * instead.
+ *
+ * @param pos Iterator variable of the type of the list elements.
+ * @param head List head
+ * @param member Member name of the struct list_head in the list elements.
+ *
+ */
+#define list_for_each_entry(pos, head, member)				\
+    for (pos = __container_of((head)->next, pos, member);		\
+	 &pos->member != (head);					\
+	 pos = __container_of(pos->member.next, pos, member))
+
+/**
+ * Loop through the list, keeping a backup pointer to the element. This
+ * macro allows for the deletion of a list element while looping through the
+ * list.
+ *
+ * See list_for_each_entry for more details.
+ */
+#define list_for_each_entry_safe(pos, tmp, head, member)		\
+    for (pos = __container_of((head)->next, pos, member),		\
+	 tmp = __container_of(pos->member.next, pos, member);		\
+	 &pos->member != (head);					\
+	 pos = tmp, tmp = __container_of(pos->member.next, tmp, member))
+
+
+#define list_for_each_entry_reverse(pos, head, member)			\
+	for (pos = __container_of((head)->prev, pos, member);		\
+	     &pos->member != (head);					\
+	     pos = __container_of(pos->member.prev, pos, member))
+
+#define list_for_each_entry_continue(pos, head, member)			\
+	for (pos = __container_of(pos->member.next, pos, member);	\
+	     &pos->member != (head);					\
+	     pos = __container_of(pos->member.next, pos, member))
+
+#define list_for_each_entry_continue_reverse(pos, head, member)		\
+	for (pos = __container_of(pos->member.prev, pos, member);	\
+	     &pos->member != (head);					\
+	     pos = __container_of(pos->member.prev, pos, member))
+
+#define list_for_each_entry_from(pos, head, member)			\
+	for (;								\
+	     &pos->member != (head);					\
+	     pos = __container_of(pos->member.next, pos, member))
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/notify.c b/drivers/gpu/drm/nouveau/nvif/notify.c
new file mode 100644
index 000000000000..0898c3155292
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/notify.c
@@ -0,0 +1,248 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include <nvif/client.h>
+#include <nvif/driver.h>
+#include <nvif/notify.h>
+#include <nvif/object.h>
+#include <nvif/ioctl.h>
+#include <nvif/event.h>
+
+static inline int
+nvif_notify_put_(struct nvif_notify *notify)
+{
+	struct nvif_object *object = notify->object;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_ntfy_put_v0 ntfy;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_NTFY_PUT,
+		.ntfy.index = notify->index,
+	};
+
+	if (atomic_inc_return(&notify->putcnt) != 1)
+		return 0;
+
+	return nvif_object_ioctl(object, &args, sizeof(args), NULL);
+}
+
+int
+nvif_notify_put(struct nvif_notify *notify)
+{
+	if (likely(notify->object) &&
+	    test_and_clear_bit(NVIF_NOTIFY_USER, &notify->flags)) {
+		int ret = nvif_notify_put_(notify);
+		if (test_bit(NVIF_NOTIFY_WORK, &notify->flags))
+			flush_work(&notify->work);
+		return ret;
+	}
+	return 0;
+}
+
+static inline int
+nvif_notify_get_(struct nvif_notify *notify)
+{
+	struct nvif_object *object = notify->object;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_ntfy_get_v0 ntfy;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_NTFY_GET,
+		.ntfy.index = notify->index,
+	};
+
+	if (atomic_dec_return(&notify->putcnt) != 0)
+		return 0;
+
+	return nvif_object_ioctl(object, &args, sizeof(args), NULL);
+}
+
+int
+nvif_notify_get(struct nvif_notify *notify)
+{
+	if (likely(notify->object) &&
+	    !test_and_set_bit(NVIF_NOTIFY_USER, &notify->flags))
+		return nvif_notify_get_(notify);
+	return 0;
+}
+
+static inline int
+nvif_notify_func(struct nvif_notify *notify, bool keep)
+{
+	int ret = notify->func(notify);
+	if (ret == NVIF_NOTIFY_KEEP ||
+	    !test_and_clear_bit(NVKM_NOTIFY_USER, &notify->flags)) {
+		if (!keep)
+			atomic_dec(&notify->putcnt);
+		else
+			nvif_notify_get_(notify);
+	}
+	return ret;
+}
+
+static void
+nvif_notify_work(struct work_struct *work)
+{
+	struct nvif_notify *notify = container_of(work, typeof(*notify), work);
+	nvif_notify_func(notify, true);
+}
+
+int
+nvif_notify(const void *header, u32 length, const void *data, u32 size)
+{
+	struct nvif_notify *notify = NULL;
+	const union {
+		struct nvif_notify_rep_v0 v0;
+	} *args = header;
+	int ret = NVIF_NOTIFY_DROP;
+
+	if (length == sizeof(args->v0) && args->v0.version == 0) {
+		if (WARN_ON(args->v0.route))
+			return NVIF_NOTIFY_DROP;
+		notify = (void *)(unsigned long)args->v0.token;
+	}
+
+	if (!WARN_ON(notify == NULL)) {
+		struct nvif_client *client = nvif_client(notify->object);
+		if (!WARN_ON(notify->size != size)) {
+			atomic_inc(&notify->putcnt);
+			if (test_bit(NVIF_NOTIFY_WORK, &notify->flags)) {
+				memcpy((void *)notify->data, data, size);
+				schedule_work(&notify->work);
+				return NVIF_NOTIFY_DROP;
+			}
+			notify->data = data;
+			ret = nvif_notify_func(notify, client->driver->keep);
+			notify->data = NULL;
+		}
+	}
+
+	return ret;
+}
+
+int
+nvif_notify_fini(struct nvif_notify *notify)
+{
+	struct nvif_object *object = notify->object;
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_ntfy_del_v0 ntfy;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_NTFY_DEL,
+		.ntfy.index = notify->index,
+	};
+	int ret = nvif_notify_put(notify);
+	if (ret >= 0 && object) {
+		ret = nvif_object_ioctl(object, &args, sizeof(args), NULL);
+		if (ret == 0) {
+			nvif_object_ref(NULL, &notify->object);
+			kfree((void *)notify->data);
+		}
+	}
+	return ret;
+}
+
+int
+nvif_notify_init(struct nvif_object *object, void (*dtor)(struct nvif_notify *),
+		 int (*func)(struct nvif_notify *), bool work, u8 event,
+		 void *data, u32 size, u32 reply, struct nvif_notify *notify)
+{
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_ntfy_new_v0 ntfy;
+		struct nvif_notify_req_v0 req;
+	} *args;
+	int ret = -ENOMEM;
+
+	notify->object = NULL;
+	nvif_object_ref(object, &notify->object);
+	notify->flags = 0;
+	atomic_set(&notify->putcnt, 1);
+	notify->dtor = dtor;
+	notify->func = func;
+	notify->data = NULL;
+	notify->size = reply;
+	if (work) {
+		INIT_WORK(&notify->work, nvif_notify_work);
+		set_bit(NVIF_NOTIFY_WORK, &notify->flags);
+		notify->data = kmalloc(notify->size, GFP_KERNEL);
+		if (!notify->data)
+			goto done;
+	}
+
+	if (!(args = kmalloc(sizeof(*args) + size, GFP_KERNEL)))
+		goto done;
+	args->ioctl.version = 0;
+	args->ioctl.type = NVIF_IOCTL_V0_NTFY_NEW;
+	args->ntfy.version = 0;
+	args->ntfy.event = event;
+	args->req.version = 0;
+	args->req.reply = notify->size;
+	args->req.route = 0;
+	args->req.token = (unsigned long)(void *)notify;
+
+	memcpy(args->req.data, data, size);
+	ret = nvif_object_ioctl(object, args, sizeof(*args) + size, NULL);
+	notify->index = args->ntfy.index;
+	kfree(args);
+done:
+	if (ret)
+		nvif_notify_fini(notify);
+	return ret;
+}
+
+static void
+nvif_notify_del(struct nvif_notify *notify)
+{
+	nvif_notify_fini(notify);
+	kfree(notify);
+}
+
+void
+nvif_notify_ref(struct nvif_notify *notify, struct nvif_notify **pnotify)
+{
+	BUG_ON(notify != NULL);
+	if (*pnotify)
+		(*pnotify)->dtor(*pnotify);
+	*pnotify = notify;
+}
+
+int
+nvif_notify_new(struct nvif_object *object, int (*func)(struct nvif_notify *),
+		bool work, u8 type, void *data, u32 size, u32 reply,
+		struct nvif_notify **pnotify)
+{
+	struct nvif_notify *notify = kzalloc(sizeof(*notify), GFP_KERNEL);
+	if (notify) {
+		int ret = nvif_notify_init(object, nvif_notify_del, func, work,
+					   type, data, size, reply, notify);
+		if (ret) {
+			kfree(notify);
+			notify = NULL;
+		}
+		*pnotify = notify;
+		return ret;
+	}
+	return -ENOMEM;
+}
diff --git a/drivers/gpu/drm/nouveau/nvif/notify.h b/drivers/gpu/drm/nouveau/nvif/notify.h
new file mode 100644
index 000000000000..9ebfa3b45e76
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/notify.h
@@ -0,0 +1,39 @@
+#ifndef __NVIF_NOTIFY_H__
+#define __NVIF_NOTIFY_H__
+
+struct nvif_notify {
+	struct nvif_object *object;
+	int index;
+
+#define NVIF_NOTIFY_USER 0
+#define NVIF_NOTIFY_WORK 1
+	unsigned long flags;
+	atomic_t putcnt;
+	void (*dtor)(struct nvif_notify *);
+#define NVIF_NOTIFY_DROP 0
+#define NVIF_NOTIFY_KEEP 1
+	int  (*func)(struct nvif_notify *);
+
+	/* this is const for a *very* good reason - the data might be on the
+	 * stack from an irq handler.  if you're not nvif/notify.c then you
+	 * should probably think twice before casting it away...
+	 */
+	const void *data;
+	u32 size;
+	struct work_struct work;
+};
+
+int  nvif_notify_init(struct nvif_object *, void (*dtor)(struct nvif_notify *),
+		      int (*func)(struct nvif_notify *), bool work, u8 type,
+		      void *data, u32 size, u32 reply, struct nvif_notify *);
+int  nvif_notify_fini(struct nvif_notify *);
+int  nvif_notify_get(struct nvif_notify *);
+int  nvif_notify_put(struct nvif_notify *);
+int  nvif_notify(const void *, u32, const void *, u32);
+
+int  nvif_notify_new(struct nvif_object *, int (*func)(struct nvif_notify *),
+		     bool work, u8 type, void *data, u32 size, u32 reply,
+		     struct nvif_notify **);
+void nvif_notify_ref(struct nvif_notify *, struct nvif_notify **);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/object.c b/drivers/gpu/drm/nouveau/nvif/object.c
new file mode 100644
index 000000000000..dd85b56f6aa5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/object.c
@@ -0,0 +1,304 @@
+/*
+ * Copyright 2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "object.h"
+#include "client.h"
+#include "driver.h"
+#include "ioctl.h"
+
+int
+nvif_object_ioctl(struct nvif_object *object, void *data, u32 size, void **hack)
+{
+	struct nvif_client *client = nvif_client(object);
+	union {
+		struct nvif_ioctl_v0 v0;
+	} *args = data;
+
+	if (size >= sizeof(*args) && args->v0.version == 0) {
+		args->v0.owner = NVIF_IOCTL_V0_OWNER_ANY;
+		args->v0.path_nr = 0;
+		while (args->v0.path_nr < ARRAY_SIZE(args->v0.path)) {
+			args->v0.path[args->v0.path_nr++] = object->handle;
+			if (object->parent == object)
+				break;
+			object = object->parent;
+		}
+	} else
+		return -ENOSYS;
+
+	return client->driver->ioctl(client->base.priv, client->super, data, size, hack);
+}
+
+int
+nvif_object_sclass(struct nvif_object *object, u32 *oclass, int count)
+{
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_sclass_v0 sclass;
+	} *args;
+	u32 size = count * sizeof(args->sclass.oclass[0]);
+	int ret;
+
+	if (!(args = kmalloc(sizeof(*args) + size, GFP_KERNEL)))
+		return -ENOMEM;
+	args->ioctl.version = 0;
+	args->ioctl.type = NVIF_IOCTL_V0_SCLASS;
+	args->sclass.version = 0;
+	args->sclass.count = count;
+
+	memcpy(args->sclass.oclass, oclass, size);
+	ret = nvif_object_ioctl(object, args, sizeof(*args) + size, NULL);
+	ret = ret ? ret : args->sclass.count;
+	memcpy(oclass, args->sclass.oclass, size);
+	kfree(args);
+	return ret;
+}
+
+u32
+nvif_object_rd(struct nvif_object *object, int size, u64 addr)
+{
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_rd_v0 rd;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_RD,
+		.rd.size = size,
+		.rd.addr = addr,
+	};
+	int ret = nvif_object_ioctl(object, &args, sizeof(args), NULL);
+	if (ret) {
+		/*XXX: warn? */
+		return 0;
+	}
+	return args.rd.data;
+}
+
+void
+nvif_object_wr(struct nvif_object *object, int size, u64 addr, u32 data)
+{
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_wr_v0 wr;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_WR,
+		.wr.size = size,
+		.wr.addr = addr,
+		.wr.data = data,
+	};
+	int ret = nvif_object_ioctl(object, &args, sizeof(args), NULL);
+	if (ret) {
+		/*XXX: warn? */
+	}
+}
+
+int
+nvif_object_mthd(struct nvif_object *object, u32 mthd, void *data, u32 size)
+{
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_mthd_v0 mthd;
+	} *args;
+	u8 stack[128];
+	int ret;
+
+	if (sizeof(*args) + size > sizeof(stack)) {
+		if (!(args = kmalloc(sizeof(*args) + size, GFP_KERNEL)))
+			return -ENOMEM;
+	} else {
+		args = (void *)stack;
+	}
+	args->ioctl.version = 0;
+	args->ioctl.type = NVIF_IOCTL_V0_MTHD;
+	args->mthd.version = 0;
+	args->mthd.method = mthd;
+
+	memcpy(args->mthd.data, data, size);
+	ret = nvif_object_ioctl(object, args, sizeof(*args) + size, NULL);
+	memcpy(data, args->mthd.data, size);
+	if (args != (void *)stack)
+		kfree(args);
+	return ret;
+}
+
+void
+nvif_object_unmap(struct nvif_object *object)
+{
+	if (object->map.size) {
+		struct nvif_client *client = nvif_client(object);
+		struct {
+			struct nvif_ioctl_v0 ioctl;
+			struct nvif_ioctl_unmap unmap;
+		} args = {
+			.ioctl.type = NVIF_IOCTL_V0_UNMAP,
+		};
+
+		if (object->map.ptr) {
+			client->driver->unmap(client, object->map.ptr,
+						      object->map.size);
+			object->map.ptr = NULL;
+		}
+
+		nvif_object_ioctl(object, &args, sizeof(args), NULL);
+		object->map.size = 0;
+	}
+}
+
+int
+nvif_object_map(struct nvif_object *object)
+{
+	struct nvif_client *client = nvif_client(object);
+	struct {
+		struct nvif_ioctl_v0 ioctl;
+		struct nvif_ioctl_map_v0 map;
+	} args = {
+		.ioctl.type = NVIF_IOCTL_V0_MAP,
+	};
+	int ret = nvif_object_ioctl(object, &args, sizeof(args), NULL);
+	if (ret == 0) {
+		object->map.size = args.map.length;
+		object->map.ptr = client->driver->map(client, args.map.handle,
+						      object->map.size);
+		if (ret = -ENOMEM, object->map.ptr)
+			return 0;
+		nvif_object_unmap(object);
+	}
+	return ret;
+}
+
+struct ctor {
+	struct nvif_ioctl_v0 ioctl;
+	struct nvif_ioctl_new_v0 new;
+};
+
+void
+nvif_object_fini(struct nvif_object *object)
+{
+	struct ctor *ctor = container_of(object->data, typeof(*ctor), new.data);
+	if (object->parent) {
+		struct {
+			struct nvif_ioctl_v0 ioctl;
+			struct nvif_ioctl_del del;
+		} args = {
+			.ioctl.type = NVIF_IOCTL_V0_DEL,
+		};
+
+		nvif_object_unmap(object);
+		nvif_object_ioctl(object, &args, sizeof(args), NULL);
+		if (object->data) {
+			object->size = 0;
+			object->data = NULL;
+			kfree(ctor);
+		}
+		nvif_object_ref(NULL, &object->parent);
+	}
+}
+
+int
+nvif_object_init(struct nvif_object *parent, void (*dtor)(struct nvif_object *),
+		 u32 handle, u32 oclass, void *data, u32 size,
+		 struct nvif_object *object)
+{
+	struct ctor *ctor;
+	int ret = 0;
+
+	object->parent = NULL;
+	object->object = object;
+	nvif_object_ref(parent, &object->parent);
+	kref_init(&object->refcount);
+	object->handle = handle;
+	object->oclass = oclass;
+	object->data = NULL;
+	object->size = 0;
+	object->dtor = dtor;
+	object->map.ptr = NULL;
+	object->map.size = 0;
+
+	if (object->parent) {
+		if (!(ctor = kmalloc(sizeof(*ctor) + size, GFP_KERNEL))) {
+			nvif_object_fini(object);
+			return -ENOMEM;
+		}
+		object->data = ctor->new.data;
+		object->size = size;
+		memcpy(object->data, data, size);
+
+		ctor->ioctl.version = 0;
+		ctor->ioctl.type = NVIF_IOCTL_V0_NEW;
+		ctor->new.version = 0;
+		ctor->new.route = NVIF_IOCTL_V0_ROUTE_NVIF;
+		ctor->new.token = (unsigned long)(void *)object;
+		ctor->new.handle = handle;
+		ctor->new.oclass = oclass;
+
+		ret = nvif_object_ioctl(parent, ctor, sizeof(*ctor) +
+					object->size, &object->priv);
+	}
+
+	if (ret)
+		nvif_object_fini(object);
+	return ret;
+}
+
+static void
+nvif_object_del(struct nvif_object *object)
+{
+	nvif_object_fini(object);
+	kfree(object);
+}
+
+int
+nvif_object_new(struct nvif_object *parent, u32 handle, u32 oclass,
+		void *data, u32 size, struct nvif_object **pobject)
+{
+	struct nvif_object *object = kzalloc(sizeof(*object), GFP_KERNEL);
+	if (object) {
+		int ret = nvif_object_init(parent, nvif_object_del, handle,
+					   oclass, data, size, object);
+		if (ret) {
+			kfree(object);
+			object = NULL;
+		}
+		*pobject = object;
+		return ret;
+	}
+	return -ENOMEM;
+}
+
+static void
+nvif_object_put(struct kref *kref)
+{
+	struct nvif_object *object =
+		container_of(kref, typeof(*object), refcount);
+	object->dtor(object);
+}
+
+void
+nvif_object_ref(struct nvif_object *object, struct nvif_object **pobject)
+{
+	if (object)
+		kref_get(&object->refcount);
+	if (*pobject)
+		kref_put(&(*pobject)->refcount, nvif_object_put);
+	*pobject = object;
+}
diff --git a/drivers/gpu/drm/nouveau/nvif/object.h b/drivers/gpu/drm/nouveau/nvif/object.h
new file mode 100644
index 000000000000..fe519179b76c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/object.h
@@ -0,0 +1,75 @@
+#ifndef __NVIF_OBJECT_H__
+#define __NVIF_OBJECT_H__
+
+#include <nvif/os.h>
+
+struct nvif_object {
+	struct nvif_object *parent;
+	struct nvif_object *object; /*XXX: hack for nvif_object() */
+	struct kref refcount;
+	u32 handle;
+	u32 oclass;
+	void *data;
+	u32   size;
+	void *priv; /*XXX: hack */
+	void (*dtor)(struct nvif_object *);
+	struct {
+		void __iomem *ptr;
+		u32 size;
+	} map;
+};
+
+int  nvif_object_init(struct nvif_object *, void (*dtor)(struct nvif_object *),
+		      u32 handle, u32 oclass, void *, u32,
+		      struct nvif_object *);
+void nvif_object_fini(struct nvif_object *);
+int  nvif_object_new(struct nvif_object *, u32 handle, u32 oclass,
+		     void *, u32, struct nvif_object **);
+void nvif_object_ref(struct nvif_object *, struct nvif_object **);
+int  nvif_object_ioctl(struct nvif_object *, void *, u32, void **);
+int  nvif_object_sclass(struct nvif_object *, u32 *, int);
+u32  nvif_object_rd(struct nvif_object *, int, u64);
+void nvif_object_wr(struct nvif_object *, int, u64, u32);
+int  nvif_object_mthd(struct nvif_object *, u32, void *, u32);
+int  nvif_object_map(struct nvif_object *);
+void nvif_object_unmap(struct nvif_object *);
+
+#define nvif_object(a) (a)->object
+
+#define ioread8_native ioread8
+#define iowrite8_native iowrite8
+#define nvif_rd(a,b,c) ({                                                      \
+	struct nvif_object *_object = nvif_object(a);                          \
+	u32 _data;                                                             \
+	if (likely(_object->map.ptr))                                          \
+		_data = ioread##b##_native((u8 __iomem *)_object->map.ptr + (c));      \
+	else                                                                   \
+		_data = nvif_object_rd(_object, (b) / 8, (c));                 \
+	_data;                                                                 \
+})
+#define nvif_wr(a,b,c,d) ({                                                    \
+	struct nvif_object *_object = nvif_object(a);                          \
+	if (likely(_object->map.ptr))                                          \
+		iowrite##b##_native((d), (u8 __iomem *)_object->map.ptr + (c));        \
+	else                                                                   \
+		nvif_object_wr(_object, (b) / 8, (c), (d));                    \
+})
+#define nvif_rd08(a,b) ({ u8  _v = nvif_rd((a), 8, (b)); _v; })
+#define nvif_rd16(a,b) ({ u16 _v = nvif_rd((a), 16, (b)); _v; })
+#define nvif_rd32(a,b) ({ u32 _v = nvif_rd((a), 32, (b)); _v; })
+#define nvif_wr08(a,b,c) nvif_wr((a), 8, (b), (u8)(c))
+#define nvif_wr16(a,b,c) nvif_wr((a), 16, (b), (u16)(c))
+#define nvif_wr32(a,b,c) nvif_wr((a), 32, (b), (u32)(c))
+#define nvif_mask(a,b,c,d) ({                                                  \
+	u32 _v = nvif_rd32(nvif_object(a), (b));                               \
+	nvif_wr32(nvif_object(a), (b), (_v & ~(c)) | (d));                     \
+	_v;                                                                    \
+})
+
+#define nvif_mthd(a,b,c,d) nvif_object_mthd(nvif_object(a), (b), (c), (d))
+
+/*XXX*/
+#include <core/object.h>
+#define nvkm_object(a) ((struct nouveau_object *)nvif_object(a)->priv)
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nvif/os.h b/drivers/gpu/drm/nouveau/nvif/os.h
new file mode 120000
index 000000000000..bd744b2cf5cf
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/os.h
@@ -0,0 +1 @@
+../core/os.h
\ No newline at end of file
diff --git a/drivers/gpu/drm/nouveau/nvif/unpack.h b/drivers/gpu/drm/nouveau/nvif/unpack.h
new file mode 100644
index 000000000000..5933188b4a77
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nvif/unpack.h
@@ -0,0 +1,24 @@
+#ifndef __NVIF_UNPACK_H__
+#define __NVIF_UNPACK_H__
+
+#define nvif_unvers(d) ({                                                      \
+	ret = (size == sizeof(d)) ? 0 : -ENOSYS;                               \
+	(ret == 0);                                                            \
+})
+
+#define nvif_unpack(d,vl,vh,m) ({                                              \
+	if ((vl) == 0 || ret == -ENOSYS) {                                     \
+		int _size = sizeof(d);                                         \
+		if (_size <= size && (d).version >= (vl) &&                    \
+				     (d).version <= (vh)) {                    \
+			data = (u8 *)data + _size;                             \
+			size = size - _size;                                   \
+			ret = ((m) || !size) ? 0 : -E2BIG;                     \
+		} else {                                                       \
+			ret = -ENOSYS;                                         \
+		}                                                              \
+	}                                                                      \
+	(ret == 0);                                                            \
+})
+
+#endif
diff --git a/drivers/gpu/drm/omapdrm/omap_connector.c b/drivers/gpu/drm/omapdrm/omap_connector.c
index 86f4ead0441d..a94b11f7859d 100644
--- a/drivers/gpu/drm/omapdrm/omap_connector.c
+++ b/drivers/gpu/drm/omapdrm/omap_connector.c
@@ -32,8 +32,16 @@ struct omap_connector {
 	struct drm_connector base;
 	struct omap_dss_device *dssdev;
 	struct drm_encoder *encoder;
+	bool hdmi_mode;
 };
 
+bool omap_connector_get_hdmi_mode(struct drm_connector *connector)
+{
+	struct omap_connector *omap_connector = to_omap_connector(connector);
+
+	return omap_connector->hdmi_mode;
+}
+
 void copy_timings_omap_to_drm(struct drm_display_mode *mode,
 		struct omap_video_timings *timings)
 {
@@ -130,7 +138,7 @@ static void omap_connector_destroy(struct drm_connector *connector)
 	struct omap_dss_device *dssdev = omap_connector->dssdev;
 
 	DBG("%s", omap_connector->dssdev->name);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(omap_connector);
 
@@ -162,10 +170,14 @@ static int omap_connector_get_modes(struct drm_connector *connector)
 			drm_mode_connector_update_edid_property(
 					connector, edid);
 			n = drm_add_edid_modes(connector, edid);
+
+			omap_connector->hdmi_mode =
+				drm_detect_hdmi_monitor(edid);
 		} else {
 			drm_mode_connector_update_edid_property(
 					connector, NULL);
 		}
+
 		kfree(edid);
 	} else {
 		struct drm_display_mode *mode = drm_mode_create(dev);
@@ -307,7 +319,7 @@ struct drm_connector *omap_connector_init(struct drm_device *dev,
 	connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	return connector;
 
diff --git a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c
index f926b4caf449..56c60552abba 100644
--- a/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c
+++ b/drivers/gpu/drm/omapdrm/omap_dmm_tiler.c
@@ -199,7 +199,7 @@ static struct dmm_txn *dmm_txn_init(struct dmm *dmm, struct tcm *tcm)
 static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area,
 		struct page **pages, uint32_t npages, uint32_t roll)
 {
-	dma_addr_t pat_pa = 0;
+	dma_addr_t pat_pa = 0, data_pa = 0;
 	uint32_t *data;
 	struct pat *pat;
 	struct refill_engine *engine = txn->engine_handle;
@@ -223,7 +223,9 @@ static void dmm_txn_append(struct dmm_txn *txn, struct pat_area *area,
 			.lut_id = engine->tcm->lut_id,
 		};
 
-	data = alloc_dma(txn, 4*i, &pat->data_pa);
+	data = alloc_dma(txn, 4*i, &data_pa);
+	/* FIXME: what if data_pa is more than 32-bit ? */
+	pat->data_pa = data_pa;
 
 	while (i--) {
 		int n = i + roll;
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
index 002b9721e85a..862ba03c236c 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.c
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -629,6 +629,7 @@ static struct drm_driver omap_drm_driver = {
 		.lastclose = dev_lastclose,
 		.preclose = dev_preclose,
 		.postclose = dev_postclose,
+		.set_busid = drm_platform_set_busid,
 		.get_vblank_counter = drm_vblank_count,
 		.enable_vblank = omap_irq_enable_vblank,
 		.disable_vblank = omap_irq_disable_vblank,
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
index 284b80fc3c54..60e47b33c801 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.h
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -26,6 +26,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/omap_drm.h>
+#include <drm/drm_gem.h>
 #include <linux/platform_data/omap_drm.h>
 
 
@@ -119,13 +120,6 @@ struct omap_drm_private {
 	struct omap_drm_irq error_handler;
 };
 
-/* this should probably be in drm-core to standardize amongst drivers */
-#define DRM_ROTATE_0	0
-#define DRM_ROTATE_90	1
-#define DRM_ROTATE_180	2
-#define DRM_ROTATE_270	3
-#define DRM_REFLECT_X	4
-#define DRM_REFLECT_Y	5
 
 #ifdef CONFIG_DEBUG_FS
 int omap_debugfs_init(struct drm_minor *minor);
@@ -194,6 +188,7 @@ struct drm_encoder *omap_connector_attached_encoder(
 		struct drm_connector *connector);
 void omap_connector_flush(struct drm_connector *connector,
 		int x, int y, int w, int h);
+bool omap_connector_get_hdmi_mode(struct drm_connector *connector);
 
 void copy_timings_omap_to_drm(struct drm_display_mode *mode,
 		struct omap_video_timings *timings);
diff --git a/drivers/gpu/drm/omapdrm/omap_encoder.c b/drivers/gpu/drm/omapdrm/omap_encoder.c
index 5290a88c681d..7445fb1491ae 100644
--- a/drivers/gpu/drm/omapdrm/omap_encoder.c
+++ b/drivers/gpu/drm/omapdrm/omap_encoder.c
@@ -17,6 +17,8 @@
  * this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <drm/drm_edid.h>
+
 #include "omap_drv.h"
 
 #include "drm_crtc.h"
@@ -89,6 +91,31 @@ static void omap_encoder_mode_set(struct drm_encoder *encoder,
 				struct drm_display_mode *mode,
 				struct drm_display_mode *adjusted_mode)
 {
+	struct drm_device *dev = encoder->dev;
+	struct omap_encoder *omap_encoder = to_omap_encoder(encoder);
+	struct omap_dss_device *dssdev = omap_encoder->dssdev;
+	struct drm_connector *connector;
+	bool hdmi_mode;
+	int r;
+
+	hdmi_mode = false;
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (connector->encoder == encoder) {
+			hdmi_mode = omap_connector_get_hdmi_mode(connector);
+			break;
+		}
+	}
+
+	if (dssdev->driver->set_hdmi_mode)
+		dssdev->driver->set_hdmi_mode(dssdev, hdmi_mode);
+
+	if (hdmi_mode && dssdev->driver->set_hdmi_infoframe) {
+		struct hdmi_avi_infoframe avi;
+
+		r = drm_hdmi_avi_infoframe_from_display_mode(&avi, adjusted_mode);
+		if (r == 0)
+			dssdev->driver->set_hdmi_infoframe(dssdev, &avi);
+	}
 }
 
 static void omap_encoder_prepare(struct drm_encoder *encoder)
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index 1388ca7f87e8..8436c6857cda 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -281,7 +281,7 @@ fail:
 	return ret;
 }
 
-static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs omap_fb_helper_funcs = {
 	.fb_probe = omap_fbdev_create,
 };
 
@@ -325,7 +325,7 @@ struct drm_fb_helper *omap_fbdev_init(struct drm_device *dev)
 
 	helper = &fbdev->base;
 
-	helper->funcs = &omap_fb_helper_funcs;
+	drm_fb_helper_prepare(dev, helper, &omap_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(dev, helper,
 			priv->num_crtcs, priv->num_connectors);
diff --git a/drivers/gpu/drm/omapdrm/omap_gem.c b/drivers/gpu/drm/omapdrm/omap_gem.c
index 95dbce286a41..e4849413ee80 100644
--- a/drivers/gpu/drm/omapdrm/omap_gem.c
+++ b/drivers/gpu/drm/omapdrm/omap_gem.c
@@ -233,11 +233,7 @@ static int omap_gem_attach_pages(struct drm_gem_object *obj)
 
 	WARN_ON(omap_obj->pages);
 
-	/* TODO: __GFP_DMA32 .. but somehow GFP_HIGHMEM is coming from the
-	 * mapping_gfp_mask(mapping) which conflicts w/ GFP_DMA32.. probably
-	 * we actually want CMA memory for it all anyways..
-	 */
-	pages = drm_gem_get_pages(obj, GFP_KERNEL);
+	pages = drm_gem_get_pages(obj);
 	if (IS_ERR(pages)) {
 		dev_err(obj->dev->dev, "could not get pages: %ld\n", PTR_ERR(pages));
 		return PTR_ERR(pages);
@@ -791,7 +787,7 @@ int omap_gem_get_paddr(struct drm_gem_object *obj,
 			omap_obj->paddr = tiler_ssptr(block);
 			omap_obj->block = block;
 
-			DBG("got paddr: %08x", omap_obj->paddr);
+			DBG("got paddr: %pad", &omap_obj->paddr);
 		}
 
 		omap_obj->paddr_cnt++;
@@ -985,9 +981,9 @@ void omap_gem_describe(struct drm_gem_object *obj, struct seq_file *m)
 
 	off = drm_vma_node_start(&obj->vma_node);
 
-	seq_printf(m, "%08x: %2d (%2d) %08llx %08Zx (%2d) %p %4d",
+	seq_printf(m, "%08x: %2d (%2d) %08llx %pad (%2d) %p %4d",
 			omap_obj->flags, obj->name, obj->refcount.refcount.counter,
-			off, omap_obj->paddr, omap_obj->paddr_cnt,
+			off, &omap_obj->paddr, omap_obj->paddr_cnt,
 			omap_obj->vaddr, omap_obj->roll);
 
 	if (omap_obj->flags & OMAP_BO_TILED) {
@@ -1183,9 +1179,7 @@ int omap_gem_op_sync(struct drm_gem_object *obj, enum omap_gem_op op)
 			}
 		}
 		spin_unlock(&sync_lock);
-
-		if (waiter)
-			kfree(waiter);
+		kfree(waiter);
 	}
 	return ret;
 }
@@ -1347,6 +1341,7 @@ struct drm_gem_object *omap_gem_new(struct drm_device *dev,
 	struct omap_drm_private *priv = dev->dev_private;
 	struct omap_gem_object *omap_obj;
 	struct drm_gem_object *obj = NULL;
+	struct address_space *mapping;
 	size_t size;
 	int ret;
 
@@ -1404,14 +1399,16 @@ struct drm_gem_object *omap_gem_new(struct drm_device *dev,
 		omap_obj->height = gsize.tiled.height;
 	}
 
-	ret = 0;
-	if (flags & (OMAP_BO_DMA|OMAP_BO_EXT_MEM))
+	if (flags & (OMAP_BO_DMA|OMAP_BO_EXT_MEM)) {
 		drm_gem_private_object_init(dev, obj, size);
-	else
+	} else {
 		ret = drm_gem_object_init(dev, obj, size);
+		if (ret)
+			goto fail;
 
-	if (ret)
-		goto fail;
+		mapping = file_inode(obj->filp)->i_mapping;
+		mapping_set_gfp_mask(mapping, GFP_USER | __GFP_DMA32);
+	}
 
 	return obj;
 
@@ -1467,8 +1464,8 @@ void omap_gem_init(struct drm_device *dev)
 			entry->paddr = tiler_ssptr(block);
 			entry->block = block;
 
-			DBG("%d:%d: %dx%d: paddr=%08x stride=%d", i, j, w, h,
-					entry->paddr,
+			DBG("%d:%d: %dx%d: paddr=%pad stride=%d", i, j, w, h,
+					&entry->paddr,
 					usergart[i].stride_pfn << PAGE_SHIFT);
 		}
 	}
diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c
index 3cf31ee59aac..891a4dc608af 100644
--- a/drivers/gpu/drm/omapdrm/omap_plane.c
+++ b/drivers/gpu/drm/omapdrm/omap_plane.c
@@ -142,8 +142,8 @@ static void omap_plane_pre_apply(struct omap_drm_apply *apply)
 	DBG("%dx%d -> %dx%d (%d)", info->width, info->height,
 			info->out_width, info->out_height,
 			info->screen_width);
-	DBG("%d,%d %08x %08x", info->pos_x, info->pos_y,
-			info->paddr, info->p_uv_addr);
+	DBG("%d,%d %pad %pad", info->pos_x, info->pos_y,
+			&info->paddr, &info->p_uv_addr);
 
 	/* TODO: */
 	ilace = false;
@@ -308,16 +308,13 @@ void omap_plane_install_properties(struct drm_plane *plane,
 	if (priv->has_dmm) {
 		prop = priv->rotation_prop;
 		if (!prop) {
-			const struct drm_prop_enum_list props[] = {
-					{ DRM_ROTATE_0,   "rotate-0" },
-					{ DRM_ROTATE_90,  "rotate-90" },
-					{ DRM_ROTATE_180, "rotate-180" },
-					{ DRM_ROTATE_270, "rotate-270" },
-					{ DRM_REFLECT_X,  "reflect-x" },
-					{ DRM_REFLECT_Y,  "reflect-y" },
-			};
-			prop = drm_property_create_bitmask(dev, 0, "rotation",
-					props, ARRAY_SIZE(props));
+			prop = drm_mode_create_rotation_property(dev,
+								 BIT(DRM_ROTATE_0) |
+								 BIT(DRM_ROTATE_90) |
+								 BIT(DRM_ROTATE_180) |
+								 BIT(DRM_ROTATE_270) |
+								 BIT(DRM_REFLECT_X) |
+								 BIT(DRM_REFLECT_Y));
 			if (prop == NULL)
 				return;
 			priv->rotation_prop = prop;
diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig
index 4ec874da5668..bee9f72b3a93 100644
--- a/drivers/gpu/drm/panel/Kconfig
+++ b/drivers/gpu/drm/panel/Kconfig
@@ -5,7 +5,7 @@ config DRM_PANEL
 	  Panel registration and lookup framework.
 
 menu "Display Panels"
-	depends on DRM_PANEL
+	depends on DRM && DRM_PANEL
 
 config DRM_PANEL_SIMPLE
 	tristate "support for simple panels"
@@ -18,14 +18,11 @@ config DRM_PANEL_SIMPLE
 
 config DRM_PANEL_LD9040
 	tristate "LD9040 RGB/SPI panel"
-	depends on DRM && DRM_PANEL
-	depends on OF
-	select SPI
+	depends on OF && SPI
 	select VIDEOMODE_HELPERS
 
 config DRM_PANEL_S6E8AA0
 	tristate "S6E8AA0 DSI video mode panel"
-	depends on DRM && DRM_PANEL
 	depends on OF
 	select DRM_MIPI_DSI
 	select VIDEOMODE_HELPERS
diff --git a/drivers/gpu/drm/panel/panel-ld9040.c b/drivers/gpu/drm/panel/panel-ld9040.c
index db1601fdbe29..42ac67b21e9f 100644
--- a/drivers/gpu/drm/panel/panel-ld9040.c
+++ b/drivers/gpu/drm/panel/panel-ld9040.c
@@ -110,7 +110,10 @@ struct ld9040 {
 	int error;
 };
 
-#define panel_to_ld9040(p) container_of(p, struct ld9040, panel)
+static inline struct ld9040 *panel_to_ld9040(struct drm_panel *panel)
+{
+	return container_of(panel, struct ld9040, panel);
+}
 
 static int ld9040_clear_error(struct ld9040 *ctx)
 {
@@ -216,6 +219,11 @@ static int ld9040_power_off(struct ld9040 *ctx)
 
 static int ld9040_disable(struct drm_panel *panel)
 {
+	return 0;
+}
+
+static int ld9040_unprepare(struct drm_panel *panel)
+{
 	struct ld9040 *ctx = panel_to_ld9040(panel);
 
 	msleep(120);
@@ -228,7 +236,7 @@ static int ld9040_disable(struct drm_panel *panel)
 	return ld9040_power_off(ctx);
 }
 
-static int ld9040_enable(struct drm_panel *panel)
+static int ld9040_prepare(struct drm_panel *panel)
 {
 	struct ld9040 *ctx = panel_to_ld9040(panel);
 	int ret;
@@ -242,11 +250,16 @@ static int ld9040_enable(struct drm_panel *panel)
 	ret = ld9040_clear_error(ctx);
 
 	if (ret < 0)
-		ld9040_disable(panel);
+		ld9040_unprepare(panel);
 
 	return ret;
 }
 
+static int ld9040_enable(struct drm_panel *panel)
+{
+	return 0;
+}
+
 static int ld9040_get_modes(struct drm_panel *panel)
 {
 	struct drm_connector *connector = panel->connector;
@@ -273,6 +286,8 @@ static int ld9040_get_modes(struct drm_panel *panel)
 
 static const struct drm_panel_funcs ld9040_drm_funcs = {
 	.disable = ld9040_disable,
+	.unprepare = ld9040_unprepare,
+	.prepare = ld9040_prepare,
 	.enable = ld9040_enable,
 	.get_modes = ld9040_get_modes,
 };
diff --git a/drivers/gpu/drm/panel/panel-s6e8aa0.c b/drivers/gpu/drm/panel/panel-s6e8aa0.c
index 06e57a26db7a..b5217fe37f02 100644
--- a/drivers/gpu/drm/panel/panel-s6e8aa0.c
+++ b/drivers/gpu/drm/panel/panel-s6e8aa0.c
@@ -120,7 +120,10 @@ struct s6e8aa0 {
 	int error;
 };
 
-#define panel_to_s6e8aa0(p) container_of(p, struct s6e8aa0, panel)
+static inline struct s6e8aa0 *panel_to_s6e8aa0(struct drm_panel *panel)
+{
+	return container_of(panel, struct s6e8aa0, panel);
+}
 
 static int s6e8aa0_clear_error(struct s6e8aa0 *ctx)
 {
@@ -133,14 +136,14 @@ static int s6e8aa0_clear_error(struct s6e8aa0 *ctx)
 static void s6e8aa0_dcs_write(struct s6e8aa0 *ctx, const void *data, size_t len)
 {
 	struct mipi_dsi_device *dsi = to_mipi_dsi_device(ctx->dev);
-	int ret;
+	ssize_t ret;
 
 	if (ctx->error < 0)
 		return;
 
-	ret = mipi_dsi_dcs_write(dsi, dsi->channel, data, len);
+	ret = mipi_dsi_dcs_write(dsi, data, len);
 	if (ret < 0) {
-		dev_err(ctx->dev, "error %d writing dcs seq: %*ph\n", ret, len,
+		dev_err(ctx->dev, "error %zd writing dcs seq: %*ph\n", ret, len,
 			data);
 		ctx->error = ret;
 	}
@@ -154,7 +157,7 @@ static int s6e8aa0_dcs_read(struct s6e8aa0 *ctx, u8 cmd, void *data, size_t len)
 	if (ctx->error < 0)
 		return ctx->error;
 
-	ret = mipi_dsi_dcs_read(dsi, dsi->channel, cmd, data, len);
+	ret = mipi_dsi_dcs_read(dsi, cmd, data, len);
 	if (ret < 0) {
 		dev_err(ctx->dev, "error %d reading dcs seq(%#x)\n", ret, cmd);
 		ctx->error = ret;
@@ -889,6 +892,11 @@ static int s6e8aa0_power_off(struct s6e8aa0 *ctx)
 
 static int s6e8aa0_disable(struct drm_panel *panel)
 {
+	return 0;
+}
+
+static int s6e8aa0_unprepare(struct drm_panel *panel)
+{
 	struct s6e8aa0 *ctx = panel_to_s6e8aa0(panel);
 
 	s6e8aa0_dcs_write_seq_static(ctx, MIPI_DCS_ENTER_SLEEP_MODE);
@@ -900,7 +908,7 @@ static int s6e8aa0_disable(struct drm_panel *panel)
 	return s6e8aa0_power_off(ctx);
 }
 
-static int s6e8aa0_enable(struct drm_panel *panel)
+static int s6e8aa0_prepare(struct drm_panel *panel)
 {
 	struct s6e8aa0 *ctx = panel_to_s6e8aa0(panel);
 	int ret;
@@ -913,11 +921,16 @@ static int s6e8aa0_enable(struct drm_panel *panel)
 	ret = ctx->error;
 
 	if (ret < 0)
-		s6e8aa0_disable(panel);
+		s6e8aa0_unprepare(panel);
 
 	return ret;
 }
 
+static int s6e8aa0_enable(struct drm_panel *panel)
+{
+	return 0;
+}
+
 static int s6e8aa0_get_modes(struct drm_panel *panel)
 {
 	struct drm_connector *connector = panel->connector;
@@ -944,6 +957,8 @@ static int s6e8aa0_get_modes(struct drm_panel *panel)
 
 static const struct drm_panel_funcs s6e8aa0_drm_funcs = {
 	.disable = s6e8aa0_disable,
+	.unprepare = s6e8aa0_unprepare,
+	.prepare = s6e8aa0_prepare,
 	.enable = s6e8aa0_enable,
 	.get_modes = s6e8aa0_get_modes,
 };
diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c
index a25136132c31..23de22f8c820 100644
--- a/drivers/gpu/drm/panel/panel-simple.c
+++ b/drivers/gpu/drm/panel/panel-simple.c
@@ -37,14 +37,35 @@ struct panel_desc {
 	const struct drm_display_mode *modes;
 	unsigned int num_modes;
 
+	unsigned int bpc;
+
 	struct {
 		unsigned int width;
 		unsigned int height;
 	} size;
+
+	/**
+	 * @prepare: the time (in milliseconds) that it takes for the panel to
+	 *           become ready and start receiving video data
+	 * @enable: the time (in milliseconds) that it takes for the panel to
+	 *          display the first valid frame after starting to receive
+	 *          video data
+	 * @disable: the time (in milliseconds) that it takes for the panel to
+	 *           turn the display off (no content is visible)
+	 * @unprepare: the time (in milliseconds) that it takes for the panel
+	 *             to power itself down completely
+	 */
+	struct {
+		unsigned int prepare;
+		unsigned int enable;
+		unsigned int disable;
+		unsigned int unprepare;
+	} delay;
 };
 
 struct panel_simple {
 	struct drm_panel base;
+	bool prepared;
 	bool enabled;
 
 	const struct panel_desc *desc;
@@ -87,6 +108,7 @@ static int panel_simple_get_fixed_modes(struct panel_simple *panel)
 		num++;
 	}
 
+	connector->display_info.bpc = panel->desc->bpc;
 	connector->display_info.width_mm = panel->desc->size.width;
 	connector->display_info.height_mm = panel->desc->size.height;
 
@@ -105,21 +127,40 @@ static int panel_simple_disable(struct drm_panel *panel)
 		backlight_update_status(p->backlight);
 	}
 
+	if (p->desc->delay.disable)
+		msleep(p->desc->delay.disable);
+
+	p->enabled = false;
+
+	return 0;
+}
+
+static int panel_simple_unprepare(struct drm_panel *panel)
+{
+	struct panel_simple *p = to_panel_simple(panel);
+
+	if (!p->prepared)
+		return 0;
+
 	if (p->enable_gpio)
 		gpiod_set_value_cansleep(p->enable_gpio, 0);
 
 	regulator_disable(p->supply);
-	p->enabled = false;
+
+	if (p->desc->delay.unprepare)
+		msleep(p->desc->delay.unprepare);
+
+	p->prepared = false;
 
 	return 0;
 }
 
-static int panel_simple_enable(struct drm_panel *panel)
+static int panel_simple_prepare(struct drm_panel *panel)
 {
 	struct panel_simple *p = to_panel_simple(panel);
 	int err;
 
-	if (p->enabled)
+	if (p->prepared)
 		return 0;
 
 	err = regulator_enable(p->supply);
@@ -131,6 +172,24 @@ static int panel_simple_enable(struct drm_panel *panel)
 	if (p->enable_gpio)
 		gpiod_set_value_cansleep(p->enable_gpio, 1);
 
+	if (p->desc->delay.prepare)
+		msleep(p->desc->delay.prepare);
+
+	p->prepared = true;
+
+	return 0;
+}
+
+static int panel_simple_enable(struct drm_panel *panel)
+{
+	struct panel_simple *p = to_panel_simple(panel);
+
+	if (p->enabled)
+		return 0;
+
+	if (p->desc->delay.enable)
+		msleep(p->desc->delay.enable);
+
 	if (p->backlight) {
 		p->backlight->props.power = FB_BLANK_UNBLANK;
 		backlight_update_status(p->backlight);
@@ -164,6 +223,8 @@ static int panel_simple_get_modes(struct drm_panel *panel)
 
 static const struct drm_panel_funcs panel_simple_funcs = {
 	.disable = panel_simple_disable,
+	.unprepare = panel_simple_unprepare,
+	.prepare = panel_simple_prepare,
 	.enable = panel_simple_enable,
 	.get_modes = panel_simple_get_modes,
 };
@@ -179,22 +240,21 @@ static int panel_simple_probe(struct device *dev, const struct panel_desc *desc)
 		return -ENOMEM;
 
 	panel->enabled = false;
+	panel->prepared = false;
 	panel->desc = desc;
 
 	panel->supply = devm_regulator_get(dev, "power");
 	if (IS_ERR(panel->supply))
 		return PTR_ERR(panel->supply);
 
-	panel->enable_gpio = devm_gpiod_get(dev, "enable");
+	panel->enable_gpio = devm_gpiod_get_optional(dev, "enable");
 	if (IS_ERR(panel->enable_gpio)) {
 		err = PTR_ERR(panel->enable_gpio);
-		if (err != -ENOENT) {
-			dev_err(dev, "failed to request GPIO: %d\n", err);
-			return err;
-		}
+		dev_err(dev, "failed to request GPIO: %d\n", err);
+		return err;
+	}
 
-		panel->enable_gpio = NULL;
-	} else {
+	if (panel->enable_gpio) {
 		err = gpiod_direction_output(panel->enable_gpio, 0);
 		if (err < 0) {
 			dev_err(dev, "failed to setup GPIO: %d\n", err);
@@ -285,6 +345,31 @@ static const struct drm_display_mode auo_b101aw03_mode = {
 static const struct panel_desc auo_b101aw03 = {
 	.modes = &auo_b101aw03_mode,
 	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 223,
+		.height = 125,
+	},
+};
+
+static const struct drm_display_mode auo_b101xtn01_mode = {
+	.clock = 72000,
+	.hdisplay = 1366,
+	.hsync_start = 1366 + 20,
+	.hsync_end = 1366 + 20 + 70,
+	.htotal = 1366 + 20 + 70,
+	.vdisplay = 768,
+	.vsync_start = 768 + 14,
+	.vsync_end = 768 + 14 + 42,
+	.vtotal = 768 + 14 + 42,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC,
+};
+
+static const struct panel_desc auo_b101xtn01 = {
+	.modes = &auo_b101xtn01_mode,
+	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 223,
 		.height = 125,
@@ -307,12 +392,40 @@ static const struct drm_display_mode auo_b133xtn01_mode = {
 static const struct panel_desc auo_b133xtn01 = {
 	.modes = &auo_b133xtn01_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 293,
 		.height = 165,
 	},
 };
 
+static const struct drm_display_mode auo_b133htn01_mode = {
+	.clock = 150660,
+	.hdisplay = 1920,
+	.hsync_start = 1920 + 172,
+	.hsync_end = 1920 + 172 + 80,
+	.htotal = 1920 + 172 + 80 + 60,
+	.vdisplay = 1080,
+	.vsync_start = 1080 + 25,
+	.vsync_end = 1080 + 25 + 10,
+	.vtotal = 1080 + 25 + 10 + 10,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc auo_b133htn01 = {
+	.modes = &auo_b133htn01_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 293,
+		.height = 165,
+	},
+	.delay = {
+		.prepare = 105,
+		.enable = 20,
+		.unprepare = 50,
+	},
+};
+
 static const struct drm_display_mode chunghwa_claa101wa01a_mode = {
 	.clock = 72070,
 	.hdisplay = 1366,
@@ -329,6 +442,7 @@ static const struct drm_display_mode chunghwa_claa101wa01a_mode = {
 static const struct panel_desc chunghwa_claa101wa01a = {
 	.modes = &chunghwa_claa101wa01a_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 220,
 		.height = 120,
@@ -351,6 +465,7 @@ static const struct drm_display_mode chunghwa_claa101wb01_mode = {
 static const struct panel_desc chunghwa_claa101wb01 = {
 	.modes = &chunghwa_claa101wb01_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 223,
 		.height = 125,
@@ -374,6 +489,7 @@ static const struct drm_display_mode edt_et057090dhu_mode = {
 static const struct panel_desc edt_et057090dhu = {
 	.modes = &edt_et057090dhu_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 115,
 		.height = 86,
@@ -397,12 +513,82 @@ static const struct drm_display_mode edt_etm0700g0dh6_mode = {
 static const struct panel_desc edt_etm0700g0dh6 = {
 	.modes = &edt_etm0700g0dh6_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 152,
 		.height = 91,
 	},
 };
 
+static const struct drm_display_mode foxlink_fl500wvr00_a0t_mode = {
+	.clock = 32260,
+	.hdisplay = 800,
+	.hsync_start = 800 + 168,
+	.hsync_end = 800 + 168 + 64,
+	.htotal = 800 + 168 + 64 + 88,
+	.vdisplay = 480,
+	.vsync_start = 480 + 37,
+	.vsync_end = 480 + 37 + 2,
+	.vtotal = 480 + 37 + 2 + 8,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc foxlink_fl500wvr00_a0t = {
+	.modes = &foxlink_fl500wvr00_a0t_mode,
+	.num_modes = 1,
+	.size = {
+		.width = 108,
+		.height = 65,
+	},
+};
+
+static const struct drm_display_mode innolux_n116bge_mode = {
+	.clock = 71000,
+	.hdisplay = 1366,
+	.hsync_start = 1366 + 64,
+	.hsync_end = 1366 + 64 + 6,
+	.htotal = 1366 + 64 + 6 + 64,
+	.vdisplay = 768,
+	.vsync_start = 768 + 8,
+	.vsync_end = 768 + 8 + 4,
+	.vtotal = 768 + 8 + 4 + 8,
+	.vrefresh = 60,
+	.flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC,
+};
+
+static const struct panel_desc innolux_n116bge = {
+	.modes = &innolux_n116bge_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 256,
+		.height = 144,
+	},
+};
+
+static const struct drm_display_mode innolux_n156bge_l21_mode = {
+	.clock = 69300,
+	.hdisplay = 1366,
+	.hsync_start = 1366 + 16,
+	.hsync_end = 1366 + 16 + 34,
+	.htotal = 1366 + 16 + 34 + 50,
+	.vdisplay = 768,
+	.vsync_start = 768 + 2,
+	.vsync_end = 768 + 2 + 6,
+	.vtotal = 768 + 2 + 6 + 12,
+	.vrefresh = 60,
+};
+
+static const struct panel_desc innolux_n156bge_l21 = {
+	.modes = &innolux_n156bge_l21_mode,
+	.num_modes = 1,
+	.bpc = 6,
+	.size = {
+		.width = 344,
+		.height = 193,
+	},
+};
+
 static const struct drm_display_mode lg_lp129qe_mode = {
 	.clock = 285250,
 	.hdisplay = 2560,
@@ -419,6 +605,7 @@ static const struct drm_display_mode lg_lp129qe_mode = {
 static const struct panel_desc lg_lp129qe = {
 	.modes = &lg_lp129qe_mode,
 	.num_modes = 1,
+	.bpc = 8,
 	.size = {
 		.width = 272,
 		.height = 181,
@@ -441,6 +628,7 @@ static const struct drm_display_mode samsung_ltn101nt05_mode = {
 static const struct panel_desc samsung_ltn101nt05 = {
 	.modes = &samsung_ltn101nt05_mode,
 	.num_modes = 1,
+	.bpc = 6,
 	.size = {
 		.width = 1024,
 		.height = 600,
@@ -452,6 +640,12 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "auo,b101aw03",
 		.data = &auo_b101aw03,
 	}, {
+		.compatible = "auo,b101xtn01",
+		.data = &auo_b101xtn01,
+	}, {
+		.compatible = "auo,b133htn01",
+		.data = &auo_b133htn01,
+	}, {
 		.compatible = "auo,b133xtn01",
 		.data = &auo_b133xtn01,
 	}, {
@@ -470,14 +664,21 @@ static const struct of_device_id platform_of_match[] = {
 		.compatible = "edt,etm0700g0dh6",
 		.data = &edt_etm0700g0dh6,
 	}, {
+		.compatible = "foxlink,fl500wvr00-a0t",
+		.data = &foxlink_fl500wvr00_a0t,
+	}, {
+		.compatible = "innolux,n116bge",
+		.data = &innolux_n116bge,
+	}, {
+		.compatible = "innolux,n156bge-l21",
+		.data = &innolux_n156bge_l21,
+	}, {
 		.compatible = "lg,lp129qe",
 		.data = &lg_lp129qe,
 	}, {
 		.compatible = "samsung,ltn101nt05",
 		.data = &samsung_ltn101nt05,
 	}, {
-		.compatible = "simple-panel",
-	}, {
 		/* sentinel */
 	}
 };
@@ -545,7 +746,7 @@ static const struct panel_desc_dsi lg_ld070wx3_sl01 = {
 			.height = 151,
 		},
 	},
-	.flags = MIPI_DSI_MODE_VIDEO,
+	.flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_CLOCK_NON_CONTINUOUS,
 	.format = MIPI_DSI_FMT_RGB888,
 	.lanes = 4,
 };
@@ -599,7 +800,8 @@ static const struct panel_desc_dsi panasonic_vvx10f004b00 = {
 			.height = 136,
 		},
 	},
-	.flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_SYNC_PULSE,
+	.flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_SYNC_PULSE |
+		 MIPI_DSI_CLOCK_NON_CONTINUOUS,
 	.format = MIPI_DSI_FMT_RGB888,
 	.lanes = 4,
 };
diff --git a/drivers/gpu/drm/qxl/Makefile b/drivers/gpu/drm/qxl/Makefile
index ea046ba691d2..bacc4aff1201 100644
--- a/drivers/gpu/drm/qxl/Makefile
+++ b/drivers/gpu/drm/qxl/Makefile
@@ -4,6 +4,6 @@
 
 ccflags-y := -Iinclude/drm
 
-qxl-y := qxl_drv.o qxl_kms.o qxl_display.o qxl_ttm.o qxl_fb.o qxl_object.o qxl_gem.o qxl_cmd.o qxl_image.o qxl_draw.o qxl_debugfs.o qxl_irq.o qxl_dumb.o qxl_ioctl.o qxl_fence.o qxl_release.o
+qxl-y := qxl_drv.o qxl_kms.o qxl_display.o qxl_ttm.o qxl_fb.o qxl_object.o qxl_gem.o qxl_cmd.o qxl_image.o qxl_draw.o qxl_debugfs.o qxl_irq.o qxl_dumb.o qxl_ioctl.o qxl_release.o qxl_prime.o
 
 obj-$(CONFIG_DRM_QXL)+= qxl.o
diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c
index eb89653a7a17..97823644d347 100644
--- a/drivers/gpu/drm/qxl/qxl_cmd.c
+++ b/drivers/gpu/drm/qxl/qxl_cmd.c
@@ -620,17 +620,10 @@ static int qxl_reap_surf(struct qxl_device *qdev, struct qxl_bo *surf, bool stal
 	if (ret == -EBUSY)
 		return -EBUSY;
 
-	if (surf->fence.num_active_releases > 0 && stall == false) {
-		qxl_bo_unreserve(surf);
-		return -EBUSY;
-	}
-
 	if (stall)
 		mutex_unlock(&qdev->surf_evict_mutex);
 
-	spin_lock(&surf->tbo.bdev->fence_lock);
 	ret = ttm_bo_wait(&surf->tbo, true, true, !stall);
-	spin_unlock(&surf->tbo.bdev->fence_lock);
 
 	if (stall)
 		mutex_lock(&qdev->surf_evict_mutex);
diff --git a/drivers/gpu/drm/qxl/qxl_debugfs.c b/drivers/gpu/drm/qxl/qxl_debugfs.c
index c3c2bbdc6674..6911b8c44492 100644
--- a/drivers/gpu/drm/qxl/qxl_debugfs.c
+++ b/drivers/gpu/drm/qxl/qxl_debugfs.c
@@ -58,9 +58,17 @@ qxl_debugfs_buffers_info(struct seq_file *m, void *data)
 	struct qxl_bo *bo;
 
 	list_for_each_entry(bo, &qdev->gem.objects, list) {
-		seq_printf(m, "size %ld, pc %d, sync obj %p, num releases %d\n",
-			   (unsigned long)bo->gem_base.size, bo->pin_count,
-			   bo->tbo.sync_obj, bo->fence.num_active_releases);
+		struct reservation_object_list *fobj;
+		int rel;
+
+		rcu_read_lock();
+		fobj = rcu_dereference(bo->tbo.resv->fence);
+		rel = fobj ? fobj->shared_count : 0;
+		rcu_read_unlock();
+
+		seq_printf(m, "size %ld, pc %d, num releases %d\n",
+			   (unsigned long)bo->gem_base.size,
+			   bo->pin_count, rel);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
index 5d7ea2461852..0d1396266857 100644
--- a/drivers/gpu/drm/qxl/qxl_display.c
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -187,6 +187,54 @@ static void qxl_crtc_destroy(struct drm_crtc *crtc)
 	kfree(qxl_crtc);
 }
 
+static int qxl_crtc_page_flip(struct drm_crtc *crtc,
+                              struct drm_framebuffer *fb,
+                              struct drm_pending_vblank_event *event,
+                              uint32_t page_flip_flags)
+{
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct qxl_framebuffer *qfb_src = to_qxl_framebuffer(fb);
+	struct qxl_framebuffer *qfb_old = to_qxl_framebuffer(crtc->primary->fb);
+	struct qxl_bo *bo_old = gem_to_qxl_bo(qfb_old->obj);
+	struct qxl_bo *bo = gem_to_qxl_bo(qfb_src->obj);
+	unsigned long flags;
+	struct drm_clip_rect norect = {
+	    .x1 = 0,
+	    .y1 = 0,
+	    .x2 = fb->width,
+	    .y2 = fb->height
+	};
+	int inc = 1;
+	int one_clip_rect = 1;
+	int ret = 0;
+
+	crtc->primary->fb = fb;
+	bo_old->is_primary = false;
+	bo->is_primary = true;
+
+	ret = qxl_bo_reserve(bo, false);
+	if (ret)
+		return ret;
+
+	qxl_draw_dirty_fb(qdev, qfb_src, bo, 0, 0,
+			  &norect, one_clip_rect, inc);
+
+	drm_vblank_get(dev, qcrtc->index);
+
+	if (event) {
+		spin_lock_irqsave(&dev->event_lock, flags);
+		drm_send_vblank_event(dev, qcrtc->index, event);
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	}
+	drm_vblank_put(dev, qcrtc->index);
+
+	qxl_bo_unreserve(bo);
+
+	return 0;
+}
+
 static int
 qxl_hide_cursor(struct qxl_device *qdev)
 {
@@ -374,6 +422,7 @@ static const struct drm_crtc_funcs qxl_crtc_funcs = {
 	.cursor_move = qxl_crtc_cursor_move,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = qxl_crtc_destroy,
+	.page_flip = qxl_crtc_page_flip,
 };
 
 static void qxl_user_framebuffer_destroy(struct drm_framebuffer *fb)
@@ -523,7 +572,6 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	struct qxl_framebuffer *qfb;
 	struct qxl_bo *bo, *old_bo = NULL;
 	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
-	uint32_t width, height, base_offset;
 	bool recreate_primary = false;
 	int ret;
 	int surf_id;
@@ -553,9 +601,10 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	if (qcrtc->index == 0)
 		recreate_primary = true;
 
-	width = mode->hdisplay;
-	height = mode->vdisplay;
-	base_offset = 0;
+	if (bo->surf.stride * bo->surf.height > qdev->vram_size) {
+		DRM_ERROR("Mode doesn't fit in vram size (vgamem)");
+		return -EINVAL;
+        }
 
 	ret = qxl_bo_reserve(bo, false);
 	if (ret != 0)
@@ -569,10 +618,10 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	if (recreate_primary) {
 		qxl_io_destroy_primary(qdev);
 		qxl_io_log(qdev,
-			   "recreate primary: %dx%d (was %dx%d,%d,%d)\n",
-			   width, height, bo->surf.width,
-			   bo->surf.height, bo->surf.stride, bo->surf.format);
-		qxl_io_create_primary(qdev, base_offset, bo);
+			   "recreate primary: %dx%d,%d,%d\n",
+			   bo->surf.width, bo->surf.height,
+			   bo->surf.stride, bo->surf.format);
+		qxl_io_create_primary(qdev, 0, bo);
 		bo->is_primary = true;
 	}
 
@@ -835,7 +884,7 @@ static void qxl_conn_destroy(struct drm_connector *connector)
 	struct qxl_output *qxl_output =
 		drm_connector_to_qxl_output(connector);
 
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(qxl_output);
 }
@@ -902,7 +951,7 @@ static int qdev_output_init(struct drm_device *dev, int num_output)
 
 	drm_object_attach_property(&connector->base,
 				   qdev->hotplug_mode_update_property, 0);
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
index 6e936634d65c..1d9b80c91a15 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.c
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -38,7 +38,7 @@
 #include "qxl_object.h"
 
 extern int qxl_max_ioctls;
-static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
+static const struct pci_device_id pciidlist[] = {
 	{ 0x1b36, 0x100, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_DISPLAY_VGA << 8,
 	  0xffff00, 0 },
 	{ 0x1b36, 0x100, PCI_ANY_ID, PCI_ANY_ID, PCI_CLASS_DISPLAY_OTHER << 8,
@@ -84,6 +84,7 @@ static const struct file_operations qxl_fops = {
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
 	.poll = drm_poll,
+	.read = drm_read,
 	.mmap = qxl_mmap,
 };
 
@@ -195,6 +196,20 @@ static int qxl_pm_restore(struct device *dev)
 	return qxl_drm_resume(drm_dev, false);
 }
 
+static u32 qxl_noop_get_vblank_counter(struct drm_device *dev, int crtc)
+{
+	return dev->vblank[crtc].count.counter;
+}
+
+static int qxl_noop_enable_vblank(struct drm_device *dev, int crtc)
+{
+	return 0;
+}
+
+static void qxl_noop_disable_vblank(struct drm_device *dev, int crtc)
+{
+}
+
 static const struct dev_pm_ops qxl_pm_ops = {
 	.suspend = qxl_pm_suspend,
 	.resume = qxl_pm_resume,
@@ -212,10 +227,15 @@ static struct pci_driver qxl_pci_driver = {
 };
 
 static struct drm_driver qxl_driver = {
-	.driver_features = DRIVER_GEM | DRIVER_MODESET |
+	.driver_features = DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME |
 			   DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED,
 	.load = qxl_driver_load,
 	.unload = qxl_driver_unload,
+	.get_vblank_counter = qxl_noop_get_vblank_counter,
+	.enable_vblank = qxl_noop_enable_vblank,
+	.disable_vblank = qxl_noop_disable_vblank,
+
+	.set_busid = drm_pci_set_busid,
 
 	.dumb_create = qxl_mode_dumb_create,
 	.dumb_map_offset = qxl_mode_dumb_mmap,
@@ -224,6 +244,17 @@ static struct drm_driver qxl_driver = {
 	.debugfs_init = qxl_debugfs_init,
 	.debugfs_cleanup = qxl_debugfs_takedown,
 #endif
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_pin = qxl_gem_prime_pin,
+	.gem_prime_unpin = qxl_gem_prime_unpin,
+	.gem_prime_get_sg_table = qxl_gem_prime_get_sg_table,
+	.gem_prime_import_sg_table = qxl_gem_prime_import_sg_table,
+	.gem_prime_vmap = qxl_gem_prime_vmap,
+	.gem_prime_vunmap = qxl_gem_prime_vunmap,
+	.gem_prime_mmap = qxl_gem_prime_mmap,
 	.gem_free_object = qxl_gem_object_free,
 	.gem_open_object = qxl_gem_object_open,
 	.gem_close_object = qxl_gem_object_close,
diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h
index 36ed40ba773f..7c6cafe21f5f 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.h
+++ b/drivers/gpu/drm/qxl/qxl_drv.h
@@ -31,6 +31,7 @@
  * Definitions taken from spice-protocol, plus kernel driver specific bits.
  */
 
+#include <linux/fence.h>
 #include <linux/workqueue.h>
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
@@ -42,6 +43,8 @@
 #include <ttm/ttm_placement.h>
 #include <ttm/ttm_module.h>
 
+#include <drm/drm_gem.h>
+
 /* just for ttm_validate_buffer */
 #include <ttm/ttm_execbuf_util.h>
 
@@ -95,31 +98,24 @@ enum {
 	QXL_INTERRUPT_IO_CMD |\
 	QXL_INTERRUPT_CLIENT_MONITORS_CONFIG)
 
-struct qxl_fence {
-	struct qxl_device *qdev;
-	uint32_t num_active_releases;
-	uint32_t *release_ids;
-	struct radix_tree_root tree;
-};
-
 struct qxl_bo {
 	/* Protected by gem.mutex */
 	struct list_head		list;
 	/* Protected by tbo.reserved */
-	u32				placements[3];
+	struct ttm_place		placements[3];
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
 	unsigned			pin_count;
 	void				*kptr;
 	int                             type;
+
 	/* Constant after initialization */
 	struct drm_gem_object		gem_base;
 	bool is_primary; /* is this now a primary surface */
 	bool hw_surf_alloc;
 	struct qxl_surface surf;
 	uint32_t surface_id;
-	struct qxl_fence fence; /* per bo fence  - list of releases */
 	struct qxl_release *surf_create;
 };
 #define gem_to_qxl_bo(gobj) container_of((gobj), struct qxl_bo, gem_base)
@@ -191,6 +187,8 @@ enum {
  * spice-protocol/qxl_dev.h */
 #define QXL_MAX_RES 96
 struct qxl_release {
+	struct fence base;
+
 	int id;
 	int type;
 	uint32_t release_offset;
@@ -284,7 +282,9 @@ struct qxl_device {
 	uint8_t		slot_gen_bits;
 	uint64_t	va_slot_mask;
 
+	spinlock_t	release_lock;
 	struct idr	release_idr;
+	uint32_t	release_seqno;
 	spinlock_t release_idr_lock;
 	struct mutex	async_io_mutex;
 	unsigned int last_sent_io_cmd;
@@ -532,6 +532,18 @@ int qxl_garbage_collect(struct qxl_device *qdev);
 int qxl_debugfs_init(struct drm_minor *minor);
 void qxl_debugfs_takedown(struct drm_minor *minor);
 
+/* qxl_prime.c */
+int qxl_gem_prime_pin(struct drm_gem_object *obj);
+void qxl_gem_prime_unpin(struct drm_gem_object *obj);
+struct sg_table *qxl_gem_prime_get_sg_table(struct drm_gem_object *obj);
+struct drm_gem_object *qxl_gem_prime_import_sg_table(
+	struct drm_device *dev, struct dma_buf_attachment *attach,
+	struct sg_table *sgt);
+void *qxl_gem_prime_vmap(struct drm_gem_object *obj);
+void qxl_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
+int qxl_gem_prime_mmap(struct drm_gem_object *obj,
+				struct vm_area_struct *vma);
+
 /* qxl_irq.c */
 int qxl_irq_init(struct qxl_device *qdev);
 irqreturn_t qxl_irq_handler(int irq, void *arg);
@@ -561,10 +573,4 @@ qxl_surface_lookup(struct drm_device *dev, int surface_id);
 void qxl_surface_evict(struct qxl_device *qdev, struct qxl_bo *surf, bool freeing);
 int qxl_update_surface(struct qxl_device *qdev, struct qxl_bo *surf);
 
-/* qxl_fence.c */
-void qxl_fence_add_release_locked(struct qxl_fence *qfence, uint32_t rel_id);
-int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id);
-int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence);
-void qxl_fence_fini(struct qxl_fence *qfence);
-
 #endif
diff --git a/drivers/gpu/drm/qxl/qxl_fb.c b/drivers/gpu/drm/qxl/qxl_fb.c
index f437b30ce689..3d7c1d00a424 100644
--- a/drivers/gpu/drm/qxl/qxl_fb.c
+++ b/drivers/gpu/drm/qxl/qxl_fb.c
@@ -625,7 +625,8 @@ static int qxl_fb_find_or_create_single(
 		struct drm_fb_helper *helper,
 		struct drm_fb_helper_surface_size *sizes)
 {
-	struct qxl_fbdev *qfbdev = (struct qxl_fbdev *)helper;
+	struct qxl_fbdev *qfbdev =
+		container_of(helper, struct qxl_fbdev, helper);
 	int new_fb = 0;
 	int ret;
 
@@ -660,7 +661,7 @@ static int qxl_fbdev_destroy(struct drm_device *dev, struct qxl_fbdev *qfbdev)
 	return 0;
 }
 
-static struct drm_fb_helper_funcs qxl_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs qxl_fb_helper_funcs = {
 	.fb_probe = qxl_fb_find_or_create_single,
 };
 
@@ -676,9 +677,12 @@ int qxl_fbdev_init(struct qxl_device *qdev)
 
 	qfbdev->qdev = qdev;
 	qdev->mode_info.qfbdev = qfbdev;
-	qfbdev->helper.funcs = &qxl_fb_helper_funcs;
 	spin_lock_init(&qfbdev->delayed_ops_lock);
 	INIT_LIST_HEAD(&qfbdev->delayed_ops);
+
+	drm_fb_helper_prepare(qdev->ddev, &qfbdev->helper,
+			      &qxl_fb_helper_funcs);
+
 	ret = drm_fb_helper_init(qdev->ddev, &qfbdev->helper,
 				 qxl_num_crtc /* num_crtc - QXL supports just 1 */,
 				 QXLFB_CONN_LIMIT);
diff --git a/drivers/gpu/drm/qxl/qxl_fence.c b/drivers/gpu/drm/qxl/qxl_fence.c
deleted file mode 100644
index ae59e91cfb9a..000000000000
--- a/drivers/gpu/drm/qxl/qxl_fence.c
+++ /dev/null
@@ -1,91 +0,0 @@
-/*
- * Copyright 2013 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Dave Airlie
- *          Alon Levy
- */
-
-
-#include "qxl_drv.h"
-
-/* QXL fencing-
-
-   When we submit operations to the GPU we pass a release reference to the GPU
-   with them, the release reference is then added to the release ring when
-   the GPU is finished with that particular operation and has removed it from
-   its tree.
-
-   So we have can have multiple outstanding non linear fences per object.
-
-   From a TTM POV we only care if the object has any outstanding releases on
-   it.
-
-   we wait until all outstanding releases are processeed.
-
-   sync object is just a list of release ids that represent that fence on
-   that buffer.
-
-   we just add new releases onto the sync object attached to the object.
-
-   This currently uses a radix tree to store the list of release ids.
-
-   For some reason every so often qxl hw fails to release, things go wrong.
-*/
-/* must be called with the fence lock held */
-void qxl_fence_add_release_locked(struct qxl_fence *qfence, uint32_t rel_id)
-{
-	radix_tree_insert(&qfence->tree, rel_id, qfence);
-	qfence->num_active_releases++;
-}
-
-int qxl_fence_remove_release(struct qxl_fence *qfence, uint32_t rel_id)
-{
-	void *ret;
-	int retval = 0;
-	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
-
-	spin_lock(&bo->tbo.bdev->fence_lock);
-
-	ret = radix_tree_delete(&qfence->tree, rel_id);
-	if (ret == qfence)
-		qfence->num_active_releases--;
-	else {
-		DRM_DEBUG("didn't find fence in radix tree for %d\n", rel_id);
-		retval = -ENOENT;
-	}
-	spin_unlock(&bo->tbo.bdev->fence_lock);
-	return retval;
-}
-
-
-int qxl_fence_init(struct qxl_device *qdev, struct qxl_fence *qfence)
-{
-	qfence->qdev = qdev;
-	qfence->num_active_releases = 0;
-	INIT_RADIX_TREE(&qfence->tree, GFP_ATOMIC);
-	return 0;
-}
-
-void qxl_fence_fini(struct qxl_fence *qfence)
-{
-	kfree(qfence->release_ids);
-	qfence->num_active_releases = 0;
-}
diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
index fd88eb4a3f79..b2977a181935 100644
--- a/drivers/gpu/drm/qxl/qxl_kms.c
+++ b/drivers/gpu/drm/qxl/qxl_kms.c
@@ -223,6 +223,7 @@ static int qxl_device_init(struct qxl_device *qdev,
 
 	idr_init(&qdev->release_idr);
 	spin_lock_init(&qdev->release_idr_lock);
+	spin_lock_init(&qdev->release_lock);
 
 	idr_init(&qdev->surf_id_idr);
 	spin_lock_init(&qdev->surf_id_idr_lock);
@@ -297,6 +298,9 @@ int qxl_driver_unload(struct drm_device *dev)
 
 	if (qdev == NULL)
 		return 0;
+
+	drm_vblank_cleanup(dev);
+
 	qxl_modeset_fini(qdev);
 	qxl_device_fini(qdev);
 
@@ -324,15 +328,20 @@ int qxl_driver_load(struct drm_device *dev, unsigned long flags)
 	if (r)
 		goto out;
 
+	r = drm_vblank_init(dev, 1);
+	if (r)
+		goto unload;
+
 	r = qxl_modeset_init(qdev);
-	if (r) {
-		qxl_driver_unload(dev);
-		goto out;
-	}
+	if (r)
+		goto unload;
 
 	drm_kms_helper_poll_init(qdev->ddev);
 
 	return 0;
+unload:
+	qxl_driver_unload(dev);
+
 out:
 	kfree(qdev);
 	return r;
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index b95f144f0b49..cdeaf08fdc74 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -36,7 +36,6 @@ static void qxl_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
 
 	qxl_surface_evict(qdev, bo, false);
-	qxl_fence_fini(&bo->fence);
 	mutex_lock(&qdev->gem.mutex);
 	list_del_init(&bo->list);
 	mutex_unlock(&qdev->gem.mutex);
@@ -55,21 +54,24 @@ void qxl_ttm_placement_from_domain(struct qxl_bo *qbo, u32 domain, bool pinned)
 {
 	u32 c = 0;
 	u32 pflag = pinned ? TTM_PL_FLAG_NO_EVICT : 0;
+	unsigned i;
 
-	qbo->placement.fpfn = 0;
-	qbo->placement.lpfn = 0;
 	qbo->placement.placement = qbo->placements;
 	qbo->placement.busy_placement = qbo->placements;
 	if (domain == QXL_GEM_DOMAIN_VRAM)
-		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_VRAM | pflag;
 	if (domain == QXL_GEM_DOMAIN_SURFACE)
-		qbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV0 | pflag;
+		qbo->placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_PRIV0 | pflag;
 	if (domain == QXL_GEM_DOMAIN_CPU)
-		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM | pflag;
 	if (!c)
-		qbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		qbo->placements[c++].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 	qbo->placement.num_placement = c;
 	qbo->placement.num_busy_placement = c;
+	for (i = 0; i < c; ++i) {
+		qbo->placements[i].fpfn = 0;
+		qbo->placements[i].lpfn = 0;
+	}
 }
 
 
@@ -99,7 +101,6 @@ int qxl_bo_create(struct qxl_device *qdev,
 	bo->type = domain;
 	bo->pin_count = pinned ? 1 : 0;
 	bo->surface_id = 0;
-	qxl_fence_init(qdev, &bo->fence);
 	INIT_LIST_HEAD(&bo->list);
 
 	if (surf)
@@ -109,7 +110,7 @@ int qxl_bo_create(struct qxl_device *qdev,
 
 	r = ttm_bo_init(&qdev->mman.bdev, &bo->tbo, size, type,
 			&bo->placement, 0, !kernel, NULL, size,
-			NULL, &qxl_ttm_bo_destroy);
+			NULL, NULL, &qxl_ttm_bo_destroy);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS)
 			dev_err(qdev->dev,
@@ -259,7 +260,7 @@ int qxl_bo_unpin(struct qxl_bo *bo)
 	if (bo->pin_count)
 		return 0;
 	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
 	if (unlikely(r != 0))
 		dev_err(qdev->dev, "%p validate failed for unpin\n", bo);
diff --git a/drivers/gpu/drm/qxl/qxl_object.h b/drivers/gpu/drm/qxl/qxl_object.h
index d458a140c024..37af1bc0dd00 100644
--- a/drivers/gpu/drm/qxl/qxl_object.h
+++ b/drivers/gpu/drm/qxl/qxl_object.h
@@ -31,7 +31,7 @@ static inline int qxl_bo_reserve(struct qxl_bo *bo, bool no_wait)
 {
 	int r;
 
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, NULL);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS) {
 			struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
@@ -67,7 +67,7 @@ static inline int qxl_bo_wait(struct qxl_bo *bo, u32 *mem_type,
 {
 	int r;
 
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, NULL);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS) {
 			struct qxl_device *qdev = (struct qxl_device *)bo->gem_base.dev->dev_private;
@@ -76,12 +76,10 @@ static inline int qxl_bo_wait(struct qxl_bo *bo, u32 *mem_type,
 		}
 		return r;
 	}
-	spin_lock(&bo->tbo.bdev->fence_lock);
 	if (mem_type)
 		*mem_type = bo->tbo.mem.mem_type;
-	if (bo->tbo.sync_obj)
-		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
-	spin_unlock(&bo->tbo.bdev->fence_lock);
+
+	r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
diff --git a/drivers/gpu/drm/qxl/qxl_prime.c b/drivers/gpu/drm/qxl/qxl_prime.c
new file mode 100644
index 000000000000..3d031b50a8fd
--- /dev/null
+++ b/drivers/gpu/drm/qxl/qxl_prime.c
@@ -0,0 +1,72 @@
+/*
+ * Copyright 2014 Canonical
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Andreas Pokorny
+ */
+
+#include "qxl_drv.h"
+
+/* Empty Implementations as there should not be any other driver for a virtual
+ * device that might share buffers with qxl */
+
+int qxl_gem_prime_pin(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return -ENOSYS;
+}
+
+void qxl_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+}
+
+
+struct sg_table *qxl_gem_prime_get_sg_table(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+struct drm_gem_object *qxl_gem_prime_import_sg_table(
+	struct drm_device *dev, struct dma_buf_attachment *attach,
+	struct sg_table *table)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+void *qxl_gem_prime_vmap(struct drm_gem_object *obj)
+{
+	WARN_ONCE(1, "not implemented");
+	return ERR_PTR(-ENOSYS);
+}
+
+void qxl_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
+{
+	WARN_ONCE(1, "not implemented");
+}
+
+int qxl_gem_prime_mmap(struct drm_gem_object *obj,
+		       struct vm_area_struct *area)
+{
+	WARN_ONCE(1, "not implemented");
+	return ENOSYS;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c
index 14e776f1d14e..446e71ca36cb 100644
--- a/drivers/gpu/drm/qxl/qxl_release.c
+++ b/drivers/gpu/drm/qxl/qxl_release.c
@@ -21,6 +21,7 @@
  */
 #include "qxl_drv.h"
 #include "qxl_object.h"
+#include <trace/events/fence.h>
 
 /*
  * drawable cmd cache - allocate a bunch of VRAM pages, suballocate
@@ -39,6 +40,88 @@
 static const int release_size_per_bo[] = { RELEASE_SIZE, SURFACE_RELEASE_SIZE, RELEASE_SIZE };
 static const int releases_per_bo[] = { RELEASES_PER_BO, SURFACE_RELEASES_PER_BO, RELEASES_PER_BO };
 
+static const char *qxl_get_driver_name(struct fence *fence)
+{
+	return "qxl";
+}
+
+static const char *qxl_get_timeline_name(struct fence *fence)
+{
+	return "release";
+}
+
+static bool qxl_nop_signaling(struct fence *fence)
+{
+	/* fences are always automatically signaled, so just pretend we did this.. */
+	return true;
+}
+
+static long qxl_fence_wait(struct fence *fence, bool intr, signed long timeout)
+{
+	struct qxl_device *qdev;
+	struct qxl_release *release;
+	int count = 0, sc = 0;
+	bool have_drawable_releases;
+	unsigned long cur, end = jiffies + timeout;
+
+	qdev = container_of(fence->lock, struct qxl_device, release_lock);
+	release = container_of(fence, struct qxl_release, base);
+	have_drawable_releases = release->type == QXL_RELEASE_DRAWABLE;
+
+retry:
+	sc++;
+
+	if (fence_is_signaled(fence))
+		goto signaled;
+
+	qxl_io_notify_oom(qdev);
+
+	for (count = 0; count < 11; count++) {
+		if (!qxl_queue_garbage_collect(qdev, true))
+			break;
+
+		if (fence_is_signaled(fence))
+			goto signaled;
+	}
+
+	if (fence_is_signaled(fence))
+		goto signaled;
+
+	if (have_drawable_releases || sc < 4) {
+		if (sc > 2)
+			/* back off */
+			usleep_range(500, 1000);
+
+		if (time_after(jiffies, end))
+			return 0;
+
+		if (have_drawable_releases && sc > 300) {
+			FENCE_WARN(fence, "failed to wait on release %d "
+					  "after spincount %d\n",
+					  fence->context & ~0xf0000000, sc);
+			goto signaled;
+		}
+		goto retry;
+	}
+	/*
+	 * yeah, original sync_obj_wait gave up after 3 spins when
+	 * have_drawable_releases is not set.
+	 */
+
+signaled:
+	cur = jiffies;
+	if (time_after(cur, end))
+		return 0;
+	return end - cur;
+}
+
+static const struct fence_ops qxl_fence_ops = {
+	.get_driver_name = qxl_get_driver_name,
+	.get_timeline_name = qxl_get_timeline_name,
+	.enable_signaling = qxl_nop_signaling,
+	.wait = qxl_fence_wait,
+};
+
 static uint64_t
 qxl_release_alloc(struct qxl_device *qdev, int type,
 		  struct qxl_release **ret)
@@ -46,13 +129,13 @@ qxl_release_alloc(struct qxl_device *qdev, int type,
 	struct qxl_release *release;
 	int handle;
 	size_t size = sizeof(*release);
-	int idr_ret;
 
 	release = kmalloc(size, GFP_KERNEL);
 	if (!release) {
 		DRM_ERROR("Out of memory\n");
 		return 0;
 	}
+	release->base.ops = NULL;
 	release->type = type;
 	release->release_offset = 0;
 	release->surface_release_id = 0;
@@ -60,44 +143,61 @@ qxl_release_alloc(struct qxl_device *qdev, int type,
 
 	idr_preload(GFP_KERNEL);
 	spin_lock(&qdev->release_idr_lock);
-	idr_ret = idr_alloc(&qdev->release_idr, release, 1, 0, GFP_NOWAIT);
+	handle = idr_alloc(&qdev->release_idr, release, 1, 0, GFP_NOWAIT);
+	release->base.seqno = ++qdev->release_seqno;
 	spin_unlock(&qdev->release_idr_lock);
 	idr_preload_end();
-	handle = idr_ret;
-	if (idr_ret < 0)
-		goto release_fail;
+	if (handle < 0) {
+		kfree(release);
+		*ret = NULL;
+		return handle;
+	}
 	*ret = release;
 	QXL_INFO(qdev, "allocated release %lld\n", handle);
 	release->id = handle;
-release_fail:
-
 	return handle;
 }
 
+static void
+qxl_release_free_list(struct qxl_release *release)
+{
+	while (!list_empty(&release->bos)) {
+		struct qxl_bo_list *entry;
+		struct qxl_bo *bo;
+
+		entry = container_of(release->bos.next,
+				     struct qxl_bo_list, tv.head);
+		bo = to_qxl_bo(entry->tv.bo);
+		qxl_bo_unref(&bo);
+		list_del(&entry->tv.head);
+		kfree(entry);
+	}
+}
+
 void
 qxl_release_free(struct qxl_device *qdev,
 		 struct qxl_release *release)
 {
-	struct qxl_bo_list *entry, *tmp;
 	QXL_INFO(qdev, "release %d, type %d\n", release->id,
 		 release->type);
 
 	if (release->surface_release_id)
 		qxl_surface_id_dealloc(qdev, release->surface_release_id);
 
-	list_for_each_entry_safe(entry, tmp, &release->bos, tv.head) {
-		struct qxl_bo *bo = to_qxl_bo(entry->tv.bo);
-		QXL_INFO(qdev, "release %llx\n",
-			drm_vma_node_offset_addr(&entry->tv.bo->vma_node)
-						- DRM_FILE_OFFSET);
-		qxl_fence_remove_release(&bo->fence, release->id);
-		qxl_bo_unref(&bo);
-		kfree(entry);
-	}
 	spin_lock(&qdev->release_idr_lock);
 	idr_remove(&qdev->release_idr, release->id);
 	spin_unlock(&qdev->release_idr_lock);
-	kfree(release);
+
+	if (release->base.ops) {
+		WARN_ON(list_empty(&release->bos));
+		qxl_release_free_list(release);
+
+		fence_signal(&release->base);
+		fence_put(&release->base);
+	} else {
+		qxl_release_free_list(release);
+		kfree(release);
+	}
 }
 
 static int qxl_release_bo_alloc(struct qxl_device *qdev,
@@ -126,6 +226,7 @@ int qxl_release_list_add(struct qxl_release *release, struct qxl_bo *bo)
 
 	qxl_bo_ref(bo);
 	entry->tv.bo = &bo->tbo;
+	entry->tv.shared = false;
 	list_add_tail(&entry->tv.head, &release->bos);
 	return 0;
 }
@@ -142,6 +243,10 @@ static int qxl_release_validate_bo(struct qxl_bo *bo)
 			return ret;
 	}
 
+	ret = reservation_object_reserve_shared(bo->tbo.resv);
+	if (ret)
+		return ret;
+
 	/* allocate a surface for reserved + validated buffers */
 	ret = qxl_bo_check_id(bo->gem_base.dev->dev_private, bo);
 	if (ret)
@@ -159,7 +264,7 @@ int qxl_release_reserve_list(struct qxl_release *release, bool no_intr)
 	if (list_is_singular(&release->bos))
 		return 0;
 
-	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos);
+	ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos, !no_intr);
 	if (ret)
 		return ret;
 
@@ -199,6 +304,8 @@ int qxl_alloc_surface_release_reserved(struct qxl_device *qdev,
 
 		/* stash the release after the create command */
 		idr_ret = qxl_release_alloc(qdev, QXL_RELEASE_SURFACE_CMD, release);
+		if (idr_ret < 0)
+			return idr_ret;
 		bo = qxl_bo_ref(to_qxl_bo(entry->tv.bo));
 
 		(*release)->release_offset = create_rel->release_offset + 64;
@@ -239,6 +346,11 @@ int qxl_alloc_release_reserved(struct qxl_device *qdev, unsigned long size,
 	}
 
 	idr_ret = qxl_release_alloc(qdev, type, release);
+	if (idr_ret < 0) {
+		if (rbo)
+			*rbo = NULL;
+		return idr_ret;
+	}
 
 	mutex_lock(&qdev->release_mutex);
 	if (qdev->current_release_bo_offset[cur_idx] + 1 >= releases_per_bo[cur_idx]) {
@@ -319,40 +431,44 @@ void qxl_release_unmap(struct qxl_device *qdev,
 
 void qxl_release_fence_buffer_objects(struct qxl_release *release)
 {
-	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
 	struct ttm_bo_global *glob;
 	struct ttm_bo_device *bdev;
 	struct ttm_bo_driver *driver;
 	struct qxl_bo *qbo;
+	struct ttm_validate_buffer *entry;
+	struct qxl_device *qdev;
 
 	/* if only one object on the release its the release itself
 	   since these objects are pinned no need to reserve */
-	if (list_is_singular(&release->bos))
+	if (list_is_singular(&release->bos) || list_empty(&release->bos))
 		return;
 
 	bo = list_first_entry(&release->bos, struct ttm_validate_buffer, head)->bo;
 	bdev = bo->bdev;
+	qdev = container_of(bdev, struct qxl_device, mman.bdev);
+
+	/*
+	 * Since we never really allocated a context and we don't want to conflict,
+	 * set the highest bits. This will break if we really allow exporting of dma-bufs.
+	 */
+	fence_init(&release->base, &qxl_fence_ops, &qdev->release_lock,
+		   release->id | 0xf0000000, release->base.seqno);
+	trace_fence_emit(&release->base);
+
 	driver = bdev->driver;
 	glob = bo->glob;
 
 	spin_lock(&glob->lru_lock);
-	spin_lock(&bdev->fence_lock);
 
 	list_for_each_entry(entry, &release->bos, head) {
 		bo = entry->bo;
 		qbo = to_qxl_bo(bo);
 
-		if (!entry->bo->sync_obj)
-			entry->bo->sync_obj = &qbo->fence;
-
-		qxl_fence_add_release_locked(&qbo->fence, release->id);
-
+		reservation_object_add_shared_fence(bo->resv, &release->base);
 		ttm_bo_add_to_lru(bo);
 		__ttm_bo_unreserve(bo);
-		entry->reserved = false;
 	}
-	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
 	ww_acquire_fini(&release->ticket);
 }
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index 71a1baeac14e..0cbc4c987164 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -127,7 +127,7 @@ int qxl_mmap(struct file *filp, struct vm_area_struct *vma)
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET)) {
 		pr_info("%s: vma->vm_pgoff (%ld) < DRM_FILE_PAGE_OFFSET\n",
 			__func__, vma->vm_pgoff);
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	}
 
 	file_priv = filp->private_data;
@@ -188,11 +188,13 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo,
 				struct ttm_placement *placement)
 {
 	struct qxl_bo *qbo;
-	static u32 placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+	static struct ttm_place placements = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+	};
 
 	if (!qxl_ttm_bo_is_qxl_bo(bo)) {
-		placement->fpfn = 0;
-		placement->lpfn = 0;
 		placement->placement = &placements;
 		placement->busy_placement = &placements;
 		placement->num_placement = 1;
@@ -355,92 +357,6 @@ static int qxl_bo_move(struct ttm_buffer_object *bo,
 	return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem);
 }
 
-
-static int qxl_sync_obj_wait(void *sync_obj,
-			     bool lazy, bool interruptible)
-{
-	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
-	int count = 0, sc = 0;
-	struct qxl_bo *bo = container_of(qfence, struct qxl_bo, fence);
-
-	if (qfence->num_active_releases == 0)
-		return 0;
-
-retry:
-	if (sc == 0) {
-		if (bo->type == QXL_GEM_DOMAIN_SURFACE)
-			qxl_update_surface(qfence->qdev, bo);
-	} else if (sc >= 1) {
-		qxl_io_notify_oom(qfence->qdev);
-	}
-
-	sc++;
-
-	for (count = 0; count < 10; count++) {
-		bool ret;
-		ret = qxl_queue_garbage_collect(qfence->qdev, true);
-		if (ret == false)
-			break;
-
-		if (qfence->num_active_releases == 0)
-			return 0;
-	}
-
-	if (qfence->num_active_releases) {
-		bool have_drawable_releases = false;
-		void **slot;
-		struct radix_tree_iter iter;
-		int release_id;
-
-		radix_tree_for_each_slot(slot, &qfence->tree, &iter, 0) {
-			struct qxl_release *release;
-
-			release_id = iter.index;
-			release = qxl_release_from_id_locked(qfence->qdev, release_id);
-			if (release == NULL)
-				continue;
-
-			if (release->type == QXL_RELEASE_DRAWABLE)
-				have_drawable_releases = true;
-		}
-
-		qxl_queue_garbage_collect(qfence->qdev, true);
-
-		if (have_drawable_releases || sc < 4) {
-			if (sc > 2)
-				/* back off */
-				usleep_range(500, 1000);
-			if (have_drawable_releases && sc > 300) {
-				WARN(1, "sync obj %d still has outstanding releases %d %d %d %ld %d\n", sc, bo->surface_id, bo->is_primary, bo->pin_count, (unsigned long)bo->gem_base.size, qfence->num_active_releases);
-				return -EBUSY;
-			}
-			goto retry;
-		}
-	}
-	return 0;
-}
-
-static int qxl_sync_obj_flush(void *sync_obj)
-{
-	return 0;
-}
-
-static void qxl_sync_obj_unref(void **sync_obj)
-{
-	*sync_obj = NULL;
-}
-
-static void *qxl_sync_obj_ref(void *sync_obj)
-{
-	return sync_obj;
-}
-
-static bool qxl_sync_obj_signaled(void *sync_obj)
-{
-	struct qxl_fence *qfence = (struct qxl_fence *)sync_obj;
-	return (qfence->num_active_releases == 0);
-}
-
 static void qxl_bo_move_notify(struct ttm_buffer_object *bo,
 			       struct ttm_mem_reg *new_mem)
 {
@@ -467,16 +383,9 @@ static struct ttm_bo_driver qxl_bo_driver = {
 	.verify_access = &qxl_verify_access,
 	.io_mem_reserve = &qxl_ttm_io_mem_reserve,
 	.io_mem_free = &qxl_ttm_io_mem_free,
-	.sync_obj_signaled = &qxl_sync_obj_signaled,
-	.sync_obj_wait = &qxl_sync_obj_wait,
-	.sync_obj_flush = &qxl_sync_obj_flush,
-	.sync_obj_unref = &qxl_sync_obj_unref,
-	.sync_obj_ref = &qxl_sync_obj_ref,
 	.move_notify = &qxl_bo_move_notify,
 };
 
-
-
 int qxl_ttm_init(struct qxl_device *qdev)
 {
 	int r;
diff --git a/drivers/gpu/drm/r128/r128_cce.c b/drivers/gpu/drm/r128/r128_cce.c
index 59459fe4e8c5..2c45ac9c1dc3 100644
--- a/drivers/gpu/drm/r128/r128_cce.c
+++ b/drivers/gpu/drm/r128/r128_cce.c
@@ -452,7 +452,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 	dev_priv->span_pitch_offset_c = (((dev_priv->depth_pitch / 8) << 21) |
 					 (dev_priv->span_offset >> 5));
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -460,21 +460,21 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 		return -EINVAL;
 	}
 
-	dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio) {
 		DRM_ERROR("could not find mmio region!\n");
 		dev->dev_private = (void *)dev_priv;
 		r128_do_cleanup_cce(dev);
 		return -EINVAL;
 	}
-	dev_priv->cce_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cce_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cce_ring) {
 		DRM_ERROR("could not find cce ring region!\n");
 		dev->dev_private = (void *)dev_priv;
 		r128_do_cleanup_cce(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -482,7 +482,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -492,7 +492,7 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 
 	if (!dev_priv->is_pci) {
 		dev_priv->agp_textures =
-		    drm_core_findmap(dev, init->agp_textures_offset);
+		    drm_legacy_findmap(dev, init->agp_textures_offset);
 		if (!dev_priv->agp_textures) {
 			DRM_ERROR("could not find agp texture region!\n");
 			dev->dev_private = (void *)dev_priv;
@@ -507,9 +507,9 @@ static int r128_do_init_cce(struct drm_device *dev, drm_r128_init_t *init)
 
 #if __OS_HAS_AGP
 	if (!dev_priv->is_pci) {
-		drm_core_ioremap_wc(dev_priv->cce_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cce_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cce_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -603,11 +603,11 @@ int r128_do_cleanup_cce(struct drm_device *dev)
 #if __OS_HAS_AGP
 		if (!dev_priv->is_pci) {
 			if (dev_priv->cce_ring != NULL)
-				drm_core_ioremapfree(dev_priv->cce_ring, dev);
+				drm_legacy_ioremapfree(dev_priv->cce_ring, dev);
 			if (dev_priv->ring_rptr != NULL)
-				drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+				drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			if (dev->agp_buffer_map != NULL) {
-				drm_core_ioremapfree(dev->agp_buffer_map, dev);
+				drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 				dev->agp_buffer_map = NULL;
 			}
 		} else
diff --git a/drivers/gpu/drm/r128/r128_drv.c b/drivers/gpu/drm/r128/r128_drv.c
index 5bd307cd8da1..c57b4de63caf 100644
--- a/drivers/gpu/drm/r128/r128_drv.c
+++ b/drivers/gpu/drm/r128/r128_drv.c
@@ -46,7 +46,7 @@ static const struct file_operations r128_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = r128_compat_ioctl,
@@ -62,6 +62,7 @@ static struct drm_driver driver = {
 	.load = r128_driver_load,
 	.preclose = r128_driver_preclose,
 	.lastclose = r128_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.get_vblank_counter = r128_get_vblank_counter,
 	.enable_vblank = r128_enable_vblank,
 	.disable_vblank = r128_disable_vblank,
diff --git a/drivers/gpu/drm/r128/r128_drv.h b/drivers/gpu/drm/r128/r128_drv.h
index 5bf3f5ff805d..723e5d6f10a4 100644
--- a/drivers/gpu/drm/r128/r128_drv.h
+++ b/drivers/gpu/drm/r128/r128_drv.h
@@ -35,6 +35,9 @@
 #ifndef __R128_DRV_H__
 #define __R128_DRV_H__
 
+#include <drm/ati_pcigart.h>
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 #define DRIVER_AUTHOR		"Gareth Hughes, VA Linux Systems Inc."
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index dbcbfe80aac0..d01b87991422 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -60,7 +60,7 @@ radeon-y := radeon_drv.o
 
 # add UMS driver
 radeon-$(CONFIG_DRM_RADEON_UMS)+= radeon_cp.o radeon_state.o radeon_mem.o \
-	radeon_irq.o r300_cmdbuf.o r600_cp.o r600_blit.o
+	radeon_irq.o r300_cmdbuf.o r600_cp.o r600_blit.o drm_buffer.o
 
 # add KMS driver
 radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
@@ -72,15 +72,15 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \
 	rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \
 	r200.o radeon_legacy_tv.o r600_cs.o r600_blit_shaders.o \
-	radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o dce3_1_afmt.o \
+	radeon_pm.o atombios_dp.o r600_hdmi.o dce3_1_afmt.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
-	si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \
+	si_blit_shaders.o radeon_prime.o cik.o cik_blit_shaders.o \
 	r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \
 	rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \
 	trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \
-	ci_dpm.o dce6_afmt.o radeon_vm.o
+	ci_dpm.o dce6_afmt.o radeon_vm.o radeon_ucode.o radeon_ib.o radeon_mn.o
 
 # add async DMA block
 radeon-y += \
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index 15da7ef344a4..ec1593a6a561 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -1217,7 +1217,7 @@ free:
 	return ret;
 }
 
-int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
+int atom_execute_table_scratch_unlocked(struct atom_context *ctx, int index, uint32_t * params)
 {
 	int r;
 
@@ -1238,6 +1238,15 @@ int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
 	return r;
 }
 
+int atom_execute_table(struct atom_context *ctx, int index, uint32_t * params)
+{
+	int r;
+	mutex_lock(&ctx->scratch_mutex);
+	r = atom_execute_table_scratch_unlocked(ctx, index, params);
+	mutex_unlock(&ctx->scratch_mutex);
+	return r;
+}
+
 static int atom_iio_len[] = { 1, 2, 3, 3, 3, 3, 4, 4, 4, 3 };
 
 static void atom_index_iio(struct atom_context *ctx, int base)
diff --git a/drivers/gpu/drm/radeon/atom.h b/drivers/gpu/drm/radeon/atom.h
index feba6b8d36b3..6d014ddb6b78 100644
--- a/drivers/gpu/drm/radeon/atom.h
+++ b/drivers/gpu/drm/radeon/atom.h
@@ -125,6 +125,7 @@ struct card_info {
 struct atom_context {
 	struct card_info *card;
 	struct mutex mutex;
+	struct mutex scratch_mutex;
 	void *bios;
 	uint32_t cmd_table, data_table;
 	uint16_t *iio;
@@ -145,6 +146,7 @@ extern int atom_debug;
 
 struct atom_context *atom_parse(struct card_info *, void *);
 int atom_execute_table(struct atom_context *, int, uint32_t *);
+int atom_execute_table_scratch_unlocked(struct atom_context *, int, uint32_t *);
 int atom_asic_init(struct atom_context *);
 void atom_destroy(struct atom_context *);
 bool atom_parse_data_header(struct atom_context *ctx, int index, uint16_t *size,
diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index b1e11f8434e2..11ba9d21b89b 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -100,6 +100,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 	memset(&args, 0, sizeof(args));
 
 	mutex_lock(&chan->mutex);
+	mutex_lock(&rdev->mode_info.atom_context->scratch_mutex);
 
 	base = (unsigned char *)(rdev->mode_info.atom_context->scratch + 1);
 
@@ -113,7 +114,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 	if (ASIC_IS_DCE4(rdev))
 		args.v2.ucHPD_ID = chan->rec.hpd;
 
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+	atom_execute_table_scratch_unlocked(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
 	*ack = args.v1.ucReplyStatus;
 
@@ -147,6 +148,7 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan,
 
 	r = recv_bytes;
 done:
+	mutex_unlock(&rdev->mode_info.atom_context->scratch_mutex);
 	mutex_unlock(&chan->mutex);
 
 	return r;
@@ -232,8 +234,8 @@ void radeon_dp_aux_init(struct radeon_connector *radeon_connector)
 
 /***** general DP utility functions *****/
 
-#define DP_VOLTAGE_MAX         DP_TRAIN_VOLTAGE_SWING_1200
-#define DP_PRE_EMPHASIS_MAX    DP_TRAIN_PRE_EMPHASIS_9_5
+#define DP_VOLTAGE_MAX         DP_TRAIN_VOLTAGE_SWING_LEVEL_3
+#define DP_PRE_EMPHASIS_MAX    DP_TRAIN_PRE_EMPH_LEVEL_3
 
 static void dp_get_adjust_train(u8 link_status[DP_LINK_STATUS_SIZE],
 				int lane_count,
@@ -405,16 +407,13 @@ bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector)
 	u8 msg[DP_DPCD_SIZE];
 	int ret;
 
-	char dpcd_hex_dump[DP_DPCD_SIZE * 3];
-
 	ret = drm_dp_dpcd_read(&radeon_connector->ddc_bus->aux, DP_DPCD_REV, msg,
 			       DP_DPCD_SIZE);
 	if (ret > 0) {
 		memcpy(dig_connector->dpcd, msg, DP_DPCD_SIZE);
 
-		hex_dump_to_buffer(dig_connector->dpcd, sizeof(dig_connector->dpcd),
-				   32, 1, dpcd_hex_dump, sizeof(dpcd_hex_dump), false);
-		DRM_DEBUG_KMS("DPCD: %s\n", dpcd_hex_dump);
+		DRM_DEBUG_KMS("DPCD: %*ph\n", (int)sizeof(dig_connector->dpcd),
+			      dig_connector->dpcd);
 
 		radeon_dp_probe_oui(radeon_connector);
 
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index 7d68203a3737..b8cd7975f797 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -291,29 +291,6 @@ static void radeon_atom_backlight_exit(struct radeon_encoder *encoder)
 bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index,
 				struct drm_display_mode *mode);
 
-
-static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
-		return true;
-	default:
-		return false;
-	}
-}
-
 static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 				   const struct drm_display_mode *mode,
 				   struct drm_display_mode *adjusted_mode)
@@ -331,12 +308,10 @@ static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 	    && (mode->crtc_vsync_start < (mode->crtc_vdisplay + 2)))
 		adjusted_mode->crtc_vsync_start = adjusted_mode->crtc_vdisplay + 2;
 
-	/* get the native mode for LVDS */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT))
+	/* get the native mode for scaling */
+	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT)) {
 		radeon_panel_mode_fixup(encoder, adjusted_mode);
-
-	/* get the native mode for TV */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
+	} else if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
 		struct radeon_encoder_atom_dac *tv_dac = radeon_encoder->enc_priv;
 		if (tv_dac) {
 			if (tv_dac->tv_std == TV_STD_NTSC ||
@@ -346,6 +321,8 @@ static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
 			else
 				radeon_atom_get_tv_timings(rdev, 1, adjusted_mode);
 		}
+	} else if (radeon_encoder->rmx_type != RMX_OFF) {
+		radeon_panel_mode_fixup(encoder, adjusted_mode);
 	}
 
 	if (ASIC_IS_DCE3(rdev) &&
@@ -716,7 +693,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 			if (radeon_connector->use_digital &&
 			    (radeon_connector->audio == RADEON_AUDIO_ENABLE))
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else if (radeon_connector->use_digital)
@@ -735,7 +712,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 		if (radeon_audio != 0) {
 			if (radeon_connector->audio == RADEON_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
@@ -755,7 +732,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder)
 		} else if (radeon_audio != 0) {
 			if (radeon_connector->audio == RADEON_AUDIO_ENABLE)
 				return ATOM_ENCODER_MODE_HDMI;
-			else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			else if (drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 				 (radeon_connector->audio == RADEON_AUDIO_AUTO))
 				return ATOM_ENCODER_MODE_HDMI;
 			else
diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c
index 9c570fb15b8c..4157780585a0 100644
--- a/drivers/gpu/drm/radeon/atombios_i2c.c
+++ b/drivers/gpu/drm/radeon/atombios_i2c.c
@@ -48,6 +48,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 	memset(&args, 0, sizeof(args));
 
 	mutex_lock(&chan->mutex);
+	mutex_lock(&rdev->mode_info.atom_context->scratch_mutex);
 
 	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
 
@@ -82,7 +83,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 	args.ucSlaveAddr = slave_addr << 1;
 	args.ucLineNumber = chan->rec.i2c_id;
 
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+	atom_execute_table_scratch_unlocked(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
 	/* error */
 	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
@@ -95,6 +96,7 @@ static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
 		radeon_atom_copy_swap(buf, base, num, false);
 
 done:
+	mutex_unlock(&rdev->mode_info.atom_context->scratch_mutex);
 	mutex_unlock(&chan->mutex);
 
 	return r;
diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c
index f81d7ca134db..0b2929de9f41 100644
--- a/drivers/gpu/drm/radeon/btc_dpm.c
+++ b/drivers/gpu/drm/radeon/btc_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "btcd.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
@@ -2099,7 +2100,6 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
 	bool disable_mclk_switching;
 	u32 mclk, sclk;
 	u16 vddc, vddci;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 
 	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
 	    btc_dpm_vblank_too_short(rdev))
@@ -2141,39 +2141,6 @@ static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
 			ps->low.vddci = max_limits->vddci;
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	if (max_sclk_vddc) {
-		if (ps->low.sclk > max_sclk_vddc)
-			ps->low.sclk = max_sclk_vddc;
-		if (ps->medium.sclk > max_sclk_vddc)
-			ps->medium.sclk = max_sclk_vddc;
-		if (ps->high.sclk > max_sclk_vddc)
-			ps->high.sclk = max_sclk_vddc;
-	}
-	if (max_mclk_vddci) {
-		if (ps->low.mclk > max_mclk_vddci)
-			ps->low.mclk = max_mclk_vddci;
-		if (ps->medium.mclk > max_mclk_vddci)
-			ps->medium.mclk = max_mclk_vddci;
-		if (ps->high.mclk > max_mclk_vddci)
-			ps->high.mclk = max_mclk_vddci;
-	}
-	if (max_mclk_vddc) {
-		if (ps->low.mclk > max_mclk_vddc)
-			ps->low.mclk = max_mclk_vddc;
-		if (ps->medium.mclk > max_mclk_vddc)
-			ps->medium.mclk = max_mclk_vddc;
-		if (ps->high.mclk > max_mclk_vddc)
-			ps->high.mclk = max_mclk_vddc;
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	if (disable_mclk_switching) {
diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c
index 584090ac3eb9..11a55e9dad7f 100644
--- a/drivers/gpu/drm/radeon/ci_dpm.c
+++ b/drivers/gpu/drm/radeon/ci_dpm.c
@@ -24,6 +24,7 @@
 #include <linux/firmware.h>
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "radeon_ucode.h"
 #include "cikd.h"
 #include "r600_dpm.h"
@@ -162,8 +163,6 @@ static const struct ci_pt_config_reg didt_config_ci[] =
 };
 
 extern u8 rv770_get_memory_module_index(struct radeon_device *rdev);
-extern void btc_get_max_clock_from_voltage_dependency_table(struct radeon_clock_voltage_dependency_table *table,
-							    u32 *max_clock);
 extern int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
 				       u32 arb_freq_src, u32 arb_freq_dest);
 extern u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock);
@@ -748,7 +747,6 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev,
 	struct radeon_clock_and_voltage_limits *max_limits;
 	bool disable_mclk_switching;
 	u32 sclk, mclk;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 	int i;
 
 	if (rps->vce_active) {
@@ -784,29 +782,6 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev,
 		}
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	for (i = 0; i < ps->performance_level_count; i++) {
-		if (max_sclk_vddc) {
-			if (ps->performance_levels[i].sclk > max_sclk_vddc)
-				ps->performance_levels[i].sclk = max_sclk_vddc;
-		}
-		if (max_mclk_vddci) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddci)
-				ps->performance_levels[i].mclk = max_mclk_vddci;
-		}
-		if (max_mclk_vddc) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddc)
-				ps->performance_levels[i].mclk = max_mclk_vddc;
-		}
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	if (disable_mclk_switching) {
@@ -869,6 +844,9 @@ static int ci_set_thermal_temperature_range(struct radeon_device *rdev,
 	WREG32_SMC(CG_THERMAL_CTRL, tmp);
 #endif
 
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
 	return 0;
 }
 
@@ -940,7 +918,18 @@ static void ci_get_leakage_voltages(struct radeon_device *rdev)
 	pi->vddc_leakage.count = 0;
 	pi->vddci_leakage.count = 0;
 
-	if (radeon_atom_get_leakage_id_from_vbios(rdev, &leakage_id) == 0) {
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) {
+		for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) {
+			virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i;
+			if (radeon_atom_get_voltage_evv(rdev, virtual_voltage_id, &vddc) != 0)
+				continue;
+			if (vddc != 0 && vddc != virtual_voltage_id) {
+				pi->vddc_leakage.actual_voltage[pi->vddc_leakage.count] = vddc;
+				pi->vddc_leakage.leakage_id[pi->vddc_leakage.count] = virtual_voltage_id;
+				pi->vddc_leakage.count++;
+			}
+		}
+	} else if (radeon_atom_get_leakage_id_from_vbios(rdev, &leakage_id) == 0) {
 		for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) {
 			virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i;
 			if (radeon_atom_get_leakage_vddc_based_on_leakage_params(rdev, &vddc, &vddci,
@@ -5279,9 +5268,13 @@ int ci_dpm_init(struct radeon_device *rdev)
 void ci_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
 						    struct seq_file *m)
 {
+	struct ci_power_info *pi = ci_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
 	u32 sclk = ci_get_average_sclk_freq(rdev);
 	u32 mclk = ci_get_average_mclk_freq(rdev);
 
+	seq_printf(m, "uvd    %sabled\n", pi->uvd_enabled ? "en" : "dis");
+	seq_printf(m, "vce    %sabled\n", rps->vce_active ? "en" : "dis");
 	seq_printf(m, "power level avg    sclk: %u mclk: %u\n",
 		   sclk, mclk);
 }
diff --git a/drivers/gpu/drm/radeon/ci_smc.c b/drivers/gpu/drm/radeon/ci_smc.c
index 8debc9d47362..b630edc2fd0c 100644
--- a/drivers/gpu/drm/radeon/ci_smc.c
+++ b/drivers/gpu/drm/radeon/ci_smc.c
@@ -213,24 +213,37 @@ int ci_load_smc_ucode(struct radeon_device *rdev, u32 limit)
 	if (!rdev->smc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-		ucode_start_address = BONAIRE_SMC_UCODE_START;
-		ucode_size = BONAIRE_SMC_UCODE_SIZE;
-		break;
-	case CHIP_HAWAII:
-		ucode_start_address = HAWAII_SMC_UCODE_START;
-		ucode_size = HAWAII_SMC_UCODE_SIZE;
-		break;
-	default:
-		DRM_ERROR("unknown asic in smc ucode loader\n");
-		BUG();
+	if (rdev->new_fw) {
+		const struct smc_firmware_header_v1_0 *hdr =
+			(const struct smc_firmware_header_v1_0 *)rdev->smc_fw->data;
+
+		radeon_ucode_print_smc_hdr(&hdr->header);
+
+		ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
+		src = (const u8 *)
+			(rdev->smc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+			ucode_start_address = BONAIRE_SMC_UCODE_START;
+			ucode_size = BONAIRE_SMC_UCODE_SIZE;
+			break;
+		case CHIP_HAWAII:
+			ucode_start_address = HAWAII_SMC_UCODE_START;
+			ucode_size = HAWAII_SMC_UCODE_SIZE;
+			break;
+		default:
+			DRM_ERROR("unknown asic in smc ucode loader\n");
+			BUG();
+		}
+
+		src = (const u8 *)rdev->smc_fw->data;
 	}
 
 	if (ucode_size & 3)
 		return -EINVAL;
 
-	src = (const u8 *)rdev->smc_fw->data;
 	spin_lock_irqsave(&rdev->smc_idx_lock, flags);
 	WREG32(SMC_IND_INDEX_0, ucode_start_address);
 	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index c0ea66192fe0..89c01fa6dd8e 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -42,6 +42,16 @@ MODULE_FIRMWARE("radeon/BONAIRE_mc2.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin");
 MODULE_FIRMWARE("radeon/BONAIRE_smc.bin");
+
+MODULE_FIRMWARE("radeon/bonaire_pfp.bin");
+MODULE_FIRMWARE("radeon/bonaire_me.bin");
+MODULE_FIRMWARE("radeon/bonaire_ce.bin");
+MODULE_FIRMWARE("radeon/bonaire_mec.bin");
+MODULE_FIRMWARE("radeon/bonaire_mc.bin");
+MODULE_FIRMWARE("radeon/bonaire_rlc.bin");
+MODULE_FIRMWARE("radeon/bonaire_sdma.bin");
+MODULE_FIRMWARE("radeon/bonaire_smc.bin");
+
 MODULE_FIRMWARE("radeon/HAWAII_pfp.bin");
 MODULE_FIRMWARE("radeon/HAWAII_me.bin");
 MODULE_FIRMWARE("radeon/HAWAII_ce.bin");
@@ -51,18 +61,45 @@ MODULE_FIRMWARE("radeon/HAWAII_mc2.bin");
 MODULE_FIRMWARE("radeon/HAWAII_rlc.bin");
 MODULE_FIRMWARE("radeon/HAWAII_sdma.bin");
 MODULE_FIRMWARE("radeon/HAWAII_smc.bin");
+
+MODULE_FIRMWARE("radeon/hawaii_pfp.bin");
+MODULE_FIRMWARE("radeon/hawaii_me.bin");
+MODULE_FIRMWARE("radeon/hawaii_ce.bin");
+MODULE_FIRMWARE("radeon/hawaii_mec.bin");
+MODULE_FIRMWARE("radeon/hawaii_mc.bin");
+MODULE_FIRMWARE("radeon/hawaii_rlc.bin");
+MODULE_FIRMWARE("radeon/hawaii_sdma.bin");
+MODULE_FIRMWARE("radeon/hawaii_smc.bin");
+
 MODULE_FIRMWARE("radeon/KAVERI_pfp.bin");
 MODULE_FIRMWARE("radeon/KAVERI_me.bin");
 MODULE_FIRMWARE("radeon/KAVERI_ce.bin");
 MODULE_FIRMWARE("radeon/KAVERI_mec.bin");
 MODULE_FIRMWARE("radeon/KAVERI_rlc.bin");
 MODULE_FIRMWARE("radeon/KAVERI_sdma.bin");
+
+MODULE_FIRMWARE("radeon/kaveri_pfp.bin");
+MODULE_FIRMWARE("radeon/kaveri_me.bin");
+MODULE_FIRMWARE("radeon/kaveri_ce.bin");
+MODULE_FIRMWARE("radeon/kaveri_mec.bin");
+MODULE_FIRMWARE("radeon/kaveri_mec2.bin");
+MODULE_FIRMWARE("radeon/kaveri_rlc.bin");
+MODULE_FIRMWARE("radeon/kaveri_sdma.bin");
+
 MODULE_FIRMWARE("radeon/KABINI_pfp.bin");
 MODULE_FIRMWARE("radeon/KABINI_me.bin");
 MODULE_FIRMWARE("radeon/KABINI_ce.bin");
 MODULE_FIRMWARE("radeon/KABINI_mec.bin");
 MODULE_FIRMWARE("radeon/KABINI_rlc.bin");
 MODULE_FIRMWARE("radeon/KABINI_sdma.bin");
+
+MODULE_FIRMWARE("radeon/kabini_pfp.bin");
+MODULE_FIRMWARE("radeon/kabini_me.bin");
+MODULE_FIRMWARE("radeon/kabini_ce.bin");
+MODULE_FIRMWARE("radeon/kabini_mec.bin");
+MODULE_FIRMWARE("radeon/kabini_rlc.bin");
+MODULE_FIRMWARE("radeon/kabini_sdma.bin");
+
 MODULE_FIRMWARE("radeon/MULLINS_pfp.bin");
 MODULE_FIRMWARE("radeon/MULLINS_me.bin");
 MODULE_FIRMWARE("radeon/MULLINS_ce.bin");
@@ -70,6 +107,13 @@ MODULE_FIRMWARE("radeon/MULLINS_mec.bin");
 MODULE_FIRMWARE("radeon/MULLINS_rlc.bin");
 MODULE_FIRMWARE("radeon/MULLINS_sdma.bin");
 
+MODULE_FIRMWARE("radeon/mullins_pfp.bin");
+MODULE_FIRMWARE("radeon/mullins_me.bin");
+MODULE_FIRMWARE("radeon/mullins_ce.bin");
+MODULE_FIRMWARE("radeon/mullins_mec.bin");
+MODULE_FIRMWARE("radeon/mullins_rlc.bin");
+MODULE_FIRMWARE("radeon/mullins_sdma.bin");
+
 extern int r600_ih_ring_alloc(struct radeon_device *rdev);
 extern void r600_ih_ring_fini(struct radeon_device *rdev);
 extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
@@ -1760,27 +1804,44 @@ static void cik_srbm_select(struct radeon_device *rdev,
  */
 int ci_mc_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
+	const __be32 *fw_data = NULL;
+	const __le32 *new_fw_data = NULL;
 	u32 running, blackout = 0;
-	u32 *io_mc_regs;
+	u32 *io_mc_regs = NULL;
+	const __le32 *new_io_mc_regs = NULL;
 	int i, regs_size, ucode_size;
 
 	if (!rdev->mc_fw)
 		return -EINVAL;
 
-	ucode_size = rdev->mc_fw->size / 4;
+	if (rdev->new_fw) {
+		const struct mc_firmware_header_v1_0 *hdr =
+			(const struct mc_firmware_header_v1_0 *)rdev->mc_fw->data;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-		io_mc_regs = (u32 *)&bonaire_io_mc_regs;
-		regs_size = BONAIRE_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_HAWAII:
-		io_mc_regs = (u32 *)&hawaii_io_mc_regs;
-		regs_size = HAWAII_IO_MC_REGS_SIZE;
-		break;
-	default:
-		return -EINVAL;
+		radeon_ucode_print_mc_hdr(&hdr->header);
+
+		regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2);
+		new_io_mc_regs = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes));
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		new_fw_data = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		ucode_size = rdev->mc_fw->size / 4;
+
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+			io_mc_regs = (u32 *)&bonaire_io_mc_regs;
+			regs_size = BONAIRE_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_HAWAII:
+			io_mc_regs = (u32 *)&hawaii_io_mc_regs;
+			regs_size = HAWAII_IO_MC_REGS_SIZE;
+			break;
+		default:
+			return -EINVAL;
+		}
+		fw_data = (const __be32 *)rdev->mc_fw->data;
 	}
 
 	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
@@ -1797,13 +1858,21 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 
 		/* load mc io regs */
 		for (i = 0; i < regs_size; i++) {
-			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
-			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			if (rdev->new_fw) {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(new_io_mc_regs++));
+				WREG32(MC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++));
+			} else {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+				WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			}
 		}
 		/* load the MC ucode */
-		fw_data = (const __be32 *)rdev->mc_fw->data;
-		for (i = 0; i < ucode_size; i++)
-			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		for (i = 0; i < ucode_size; i++) {
+			if (rdev->new_fw)
+				WREG32(MC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++));
+			else
+				WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		}
 
 		/* put the engine back into the active state */
 		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
@@ -1841,17 +1910,21 @@ int ci_mc_load_microcode(struct radeon_device *rdev)
 static int cik_init_microcode(struct radeon_device *rdev)
 {
 	const char *chip_name;
+	const char *new_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size,
 		mec_req_size, rlc_req_size, mc_req_size = 0,
 		sdma_req_size, smc_req_size = 0, mc2_req_size = 0;
 	char fw_name[30];
+	int new_fw = 0;
 	int err;
+	int num_fw;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
 		chip_name = "BONAIRE";
+		new_chip_name = "bonaire";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
@@ -1861,9 +1934,11 @@ static int cik_init_microcode(struct radeon_device *rdev)
 		mc2_req_size = BONAIRE_MC2_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
 		smc_req_size = ALIGN(BONAIRE_SMC_UCODE_SIZE, 4);
+		num_fw = 8;
 		break;
 	case CHIP_HAWAII:
 		chip_name = "HAWAII";
+		new_chip_name = "hawaii";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
@@ -1873,142 +1948,285 @@ static int cik_init_microcode(struct radeon_device *rdev)
 		mc2_req_size = HAWAII_MC2_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
 		smc_req_size = ALIGN(HAWAII_SMC_UCODE_SIZE, 4);
+		num_fw = 8;
 		break;
 	case CHIP_KAVERI:
 		chip_name = "KAVERI";
+		new_chip_name = "kaveri";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = KV_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 7;
 		break;
 	case CHIP_KABINI:
 		chip_name = "KABINI";
+		new_chip_name = "kabini";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = KB_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 6;
 		break;
 	case CHIP_MULLINS:
 		chip_name = "MULLINS";
+		new_chip_name = "mullins";
 		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
 		me_req_size = CIK_ME_UCODE_SIZE * 4;
 		ce_req_size = CIK_CE_UCODE_SIZE * 4;
 		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
 		rlc_req_size = ML_RLC_UCODE_SIZE * 4;
 		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		num_fw = 6;
 		break;
 	default: BUG();
 	}
 
-	DRM_INFO("Loading %s Microcode\n", chip_name);
+	DRM_INFO("Loading %s Microcode\n", new_chip_name);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name);
 	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->pfp_fw->size != pfp_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->pfp_fw->size, fw_name);
-		err = -EINVAL;
-		goto out;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+		err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->pfp_fw->size != pfp_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->pfp_fw->size, fw_name);
+			err = -EINVAL;
+			goto out;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->pfp_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", new_chip_name);
 	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->me_fw->size != me_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->me_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+		err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->me_fw->size != me_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->me_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->me_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", new_chip_name);
 	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->ce_fw->size != ce_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->ce_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+		err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->ce_fw->size != ce_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->ce_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->ce_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", new_chip_name);
 	err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->mec_fw->size != mec_req_size) {
-		printk(KERN_ERR
-		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->mec_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+		err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->mec_fw->size != mec_req_size) {
+			printk(KERN_ERR
+			       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mec_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->mec_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+	if (rdev->family == CHIP_KAVERI) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec2.bin", new_chip_name);
+		err = request_firmware(&rdev->mec2_fw, fw_name, rdev->dev);
+		if (err) {
+			goto out;
+		} else {
+			err = radeon_ucode_validate(rdev->mec2_fw);
+			if (err) {
+				goto out;
+			} else {
+				new_fw++;
+			}
+		}
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", new_chip_name);
 	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->rlc_fw->size != rlc_req_size) {
-		printk(KERN_ERR
-		       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->rlc_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+		err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->rlc_fw->size != rlc_req_size) {
+			printk(KERN_ERR
+			       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->rlc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->rlc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", new_chip_name);
 	err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->sdma_fw->size != sdma_req_size) {
-		printk(KERN_ERR
-		       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->sdma_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+		err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->sdma_fw->size != sdma_req_size) {
+			printk(KERN_ERR
+			       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->sdma_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->sdma_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "cik_fw: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
 	/* No SMC, MC ucode on APUs */
 	if (!(rdev->flags & RADEON_IS_IGP)) {
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name);
 		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 		if (err) {
-			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
 			err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
-			if (err)
+			if (err) {
+				snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+				err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+				if (err)
+					goto out;
+			}
+			if ((rdev->mc_fw->size != mc_req_size) &&
+			    (rdev->mc_fw->size != mc2_req_size)){
+				printk(KERN_ERR
+				       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
+				       rdev->mc_fw->size, fw_name);
+				err = -EINVAL;
+			}
+			DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
+		} else {
+			err = radeon_ucode_validate(rdev->mc_fw);
+			if (err) {
+				printk(KERN_ERR
+				       "cik_fw: validation failed for firmware \"%s\"\n",
+				       fw_name);
 				goto out;
+			} else {
+				new_fw++;
+			}
 		}
-		if ((rdev->mc_fw->size != mc_req_size) &&
-		    (rdev->mc_fw->size != mc2_req_size)){
-			printk(KERN_ERR
-			       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
-			       rdev->mc_fw->size, fw_name);
-			err = -EINVAL;
-		}
-		DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
 
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 		if (err) {
-			printk(KERN_ERR
-			       "smc: error loading firmware \"%s\"\n",
-			       fw_name);
-			release_firmware(rdev->smc_fw);
-			rdev->smc_fw = NULL;
-			err = 0;
-		} else if (rdev->smc_fw->size != smc_req_size) {
-			printk(KERN_ERR
-			       "cik_smc: Bogus length %zu in firmware \"%s\"\n",
-			       rdev->smc_fw->size, fw_name);
-			err = -EINVAL;
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+			err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+			if (err) {
+				printk(KERN_ERR
+				       "smc: error loading firmware \"%s\"\n",
+				       fw_name);
+				release_firmware(rdev->smc_fw);
+				rdev->smc_fw = NULL;
+				err = 0;
+			} else if (rdev->smc_fw->size != smc_req_size) {
+				printk(KERN_ERR
+				       "cik_smc: Bogus length %zu in firmware \"%s\"\n",
+				       rdev->smc_fw->size, fw_name);
+				err = -EINVAL;
+			}
+		} else {
+			err = radeon_ucode_validate(rdev->smc_fw);
+			if (err) {
+				printk(KERN_ERR
+				       "cik_fw: validation failed for firmware \"%s\"\n",
+				       fw_name);
+				goto out;
+			} else {
+				new_fw++;
+			}
 		}
 	}
 
+	if (new_fw == 0) {
+		rdev->new_fw = false;
+	} else if (new_fw < num_fw) {
+		printk(KERN_ERR "ci_fw: mixing new and old firmware!\n");
+		err = -EINVAL;
+	} else {
+		rdev->new_fw = true;
+	}
+
 out:
 	if (err) {
 		if (err != -EINVAL)
@@ -2021,8 +2239,14 @@ out:
 		rdev->me_fw = NULL;
 		release_firmware(rdev->ce_fw);
 		rdev->ce_fw = NULL;
+		release_firmware(rdev->mec_fw);
+		rdev->mec_fw = NULL;
+		release_firmware(rdev->mec2_fw);
+		rdev->mec2_fw = NULL;
 		release_firmware(rdev->rlc_fw);
 		rdev->rlc_fw = NULL;
+		release_firmware(rdev->sdma_fw);
+		rdev->sdma_fw = NULL;
 		release_firmware(rdev->mc_fw);
 		rdev->mc_fw = NULL;
 		release_firmware(rdev->smc_fw);
@@ -3259,7 +3483,7 @@ static void cik_gpu_init(struct radeon_device *rdev)
 	u32 mc_shared_chmap, mc_arb_ramcfg;
 	u32 hdp_host_path_cntl;
 	u32 tmp;
-	int i, j, k;
+	int i, j;
 
 	switch (rdev->family) {
 	case CHIP_BONAIRE:
@@ -3320,6 +3544,7 @@ static void cik_gpu_init(struct radeon_device *rdev)
 			   (rdev->pdev->device == 0x130B) ||
 			   (rdev->pdev->device == 0x130E) ||
 			   (rdev->pdev->device == 0x1315) ||
+			   (rdev->pdev->device == 0x1318) ||
 			   (rdev->pdev->device == 0x131B)) {
 			rdev->config.cik.max_cu_per_sh = 4;
 			rdev->config.cik.max_backends_per_se = 1;
@@ -3448,12 +3673,11 @@ static void cik_gpu_init(struct radeon_device *rdev)
 		     rdev->config.cik.max_sh_per_se,
 		     rdev->config.cik.max_backends_per_se);
 
+	rdev->config.cik.active_cus = 0;
 	for (i = 0; i < rdev->config.cik.max_shader_engines; i++) {
 		for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) {
-			for (k = 0; k < rdev->config.cik.max_cu_per_sh; k++) {
-				rdev->config.cik.active_cus +=
-					hweight32(cik_get_cu_active_bitmap(rdev, i, j));
-			}
+			rdev->config.cik.active_cus +=
+				hweight32(cik_get_cu_active_bitmap(rdev, i, j));
 		}
 	}
 
@@ -3577,7 +3801,7 @@ int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
 	radeon_ring_write(ring, ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
@@ -3666,8 +3890,6 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | DATA_SEL(1) | INT_SEL(2));
 	radeon_ring_write(ring, fence->seq);
 	radeon_ring_write(ring, 0);
-	/* HDP flush */
-	cik_hdp_flush_cp_ring_emit(rdev, fence->ring);
 }
 
 /**
@@ -3696,10 +3918,19 @@ void cik_fence_compute_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, upper_32_bits(addr));
 	radeon_ring_write(ring, fence->seq);
 	radeon_ring_write(ring, 0);
-	/* HDP flush */
-	cik_hdp_flush_cp_ring_emit(rdev, fence->ring);
 }
 
+/**
+ * cik_semaphore_ring_emit - emit a semaphore on the CP ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon ring buffer object
+ * @semaphore: radeon semaphore object
+ * @emit_wait: Is this a sempahore wait?
+ *
+ * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP
+ * from running ahead of semaphore waits.
+ */
 bool cik_semaphore_ring_emit(struct radeon_device *rdev,
 			     struct radeon_ring *ring,
 			     struct radeon_semaphore *semaphore,
@@ -3712,6 +3943,12 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, lower_32_bits(addr));
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel);
 
+	if (emit_wait && ring->idx == RADEON_RING_TYPE_GFX_INDEX) {
+		/* Prevent the PFP from running ahead of the semaphore wait */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+
 	return true;
 }
 
@@ -3722,18 +3959,19 @@ bool cik_semaphore_ring_emit(struct radeon_device *rdev,
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the CP DMA engine (CIK+).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_cpdma(struct radeon_device *rdev,
-		   uint64_t src_offset, uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence)
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.blit_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes, control;
@@ -3743,7 +3981,7 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -3752,10 +3990,10 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -3777,17 +4015,17 @@ int cik_copy_cpdma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
 
 /*
@@ -3883,7 +4121,7 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2);
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_scratch_free(rdev, scratch);
 		radeon_ib_free(rdev, &ib);
@@ -3969,7 +4207,6 @@ static void cik_cp_gfx_enable(struct radeon_device *rdev, bool enable)
  */
 static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
@@ -3977,31 +4214,72 @@ static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
 
 	cik_cp_gfx_enable(rdev, false);
 
-	/* PFP */
-	fw_data = (const __be32 *)rdev->pfp_fw->data;
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
-		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-
-	/* CE */
-	fw_data = (const __be32 *)rdev->ce_fw->data;
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
-		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_CE_UCODE_ADDR, 0);
-
-	/* ME */
-	fw_data = (const __be32 *)rdev->me_fw->data;
-	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
-		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_ME_RAM_WADDR, 0);
-
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	WREG32(CP_ME_RAM_WADDR, 0);
-	WREG32(CP_ME_RAM_RADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *pfp_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+		const struct gfx_firmware_header_v1_0 *ce_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+		const struct gfx_firmware_header_v1_0 *me_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&pfp_hdr->header);
+		radeon_ucode_print_gfx_hdr(&ce_hdr->header);
+		radeon_ucode_print_gfx_hdr(&me_hdr->header);
+
+		/* PFP */
+		fw_data = (const __le32 *)
+			(rdev->pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, le32_to_cpu(pfp_hdr->header.ucode_version));
+
+		/* CE */
+		fw_data = (const __le32 *)
+			(rdev->ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, le32_to_cpu(ce_hdr->header.ucode_version));
+
+		/* ME */
+		fw_data = (const __be32 *)
+			(rdev->me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, le32_to_cpu(me_hdr->header.ucode_version));
+		WREG32(CP_ME_RAM_RADDR, le32_to_cpu(me_hdr->header.ucode_version));
+	} else {
+		const __be32 *fw_data;
+
+		/* PFP */
+		fw_data = (const __be32 *)rdev->pfp_fw->data;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
+			WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __be32 *)rdev->ce_fw->data;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
+			WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)rdev->me_fw->data;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
+			WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	}
+
 	return 0;
 }
 
@@ -4035,8 +4313,8 @@ static int cik_cp_gfx_start(struct radeon_device *rdev)
 	/* init the CE partitions.  CE only used for gfx on CIK */
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
 	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
-	radeon_ring_write(ring, 0xc000);
-	radeon_ring_write(ring, 0xc000);
+	radeon_ring_write(ring, 0x8000);
+	radeon_ring_write(ring, 0x8000);
 
 	/* setup clear context state */
 	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
@@ -4061,7 +4339,7 @@ static int cik_cp_gfx_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	return 0;
 }
@@ -4261,7 +4539,6 @@ static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable)
  */
 static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->mec_fw)
@@ -4269,20 +4546,55 @@ static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
 
 	cik_cp_compute_enable(rdev, false);
 
-	/* MEC1 */
-	fw_data = (const __be32 *)rdev->mec_fw->data;
-	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
-	for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
-		WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *mec_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->mec_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&mec_hdr->header);
+
+		/* MEC1 */
+		fw_data = (const __le32 *)
+			(rdev->mec_fw->data + le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_MEC_ME1_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME1_UCODE_ADDR, le32_to_cpu(mec_hdr->header.ucode_version));
 
-	if (rdev->family == CHIP_KAVERI) {
 		/* MEC2 */
+		if (rdev->family == CHIP_KAVERI) {
+			const struct gfx_firmware_header_v1_0 *mec2_hdr =
+				(const struct gfx_firmware_header_v1_0 *)rdev->mec2_fw->data;
+
+			fw_data = (const __le32 *)
+				(rdev->mec2_fw->data +
+				 le32_to_cpu(mec2_hdr->header.ucode_array_offset_bytes));
+			fw_size = le32_to_cpu(mec2_hdr->header.ucode_size_bytes) / 4;
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			for (i = 0; i < fw_size; i++)
+				WREG32(CP_MEC_ME2_UCODE_DATA, le32_to_cpup(fw_data++));
+			WREG32(CP_MEC_ME2_UCODE_ADDR, le32_to_cpu(mec2_hdr->header.ucode_version));
+		}
+	} else {
+		const __be32 *fw_data;
+
+		/* MEC1 */
 		fw_data = (const __be32 *)rdev->mec_fw->data;
-		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
 		for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
-			WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
-		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+
+		if (rdev->family == CHIP_KAVERI) {
+			/* MEC2 */
+			fw_data = (const __be32 *)rdev->mec_fw->data;
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+			for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+				WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
+			WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		}
 	}
 
 	return 0;
@@ -4375,7 +4687,7 @@ static int cik_mec_init(struct radeon_device *rdev)
 		r = radeon_bo_create(rdev,
 				     rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2,
 				     PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
 				     &rdev->mec.hpd_eop_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create HDP EOP bo failed\n", r);
@@ -4489,7 +4801,7 @@ struct bonaire_mqd
  */
 static int cik_cp_compute_resume(struct radeon_device *rdev)
 {
-	int r, i, idx;
+	int r, i, j, idx;
 	u32 tmp;
 	bool use_doorbell = true;
 	u64 hqd_gpu_addr;
@@ -4545,8 +4857,8 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 			r = radeon_bo_create(rdev,
 					     sizeof(struct bonaire_mqd),
 					     PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_GTT, NULL,
-					     &rdev->ring[idx].mqd_obj);
+					     RADEON_GEM_DOMAIN_GTT, 0, NULL,
+					     NULL, &rdev->ring[idx].mqd_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create MQD bo failed\n", r);
 				return r;
@@ -4608,7 +4920,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev)
 		mqd->queue_state.cp_hqd_pq_wptr= 0;
 		if (RREG32(CP_HQD_ACTIVE) & 1) {
 			WREG32(CP_HQD_DEQUEUE_REQUEST, 1);
-			for (i = 0; i < rdev->usec_timeout; i++) {
+			for (j = 0; j < rdev->usec_timeout; j++) {
 				if (!(RREG32(CP_HQD_ACTIVE) & 1))
 					break;
 				udelay(1);
@@ -5402,7 +5714,6 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -5436,20 +5747,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(0x15D8, 0);
 	WREG32(0x15DC, 0);
 
-	/* empty context1-15 */
-	/* FIXME start with 4G, once using 2 level pt switch to full
-	 * vm size space
-	 */
+	/* restore context1-15 */
 	/* set vm size, must be a multiple of 4 */
 	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
 	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn);
 	for (i = 1; i < 16; i++) {
 		if (i < 8)
 			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 		else
 			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-15 */
@@ -5514,6 +5822,17 @@ static int cik_pcie_gart_enable(struct radeon_device *rdev)
  */
 static void cik_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 16; ++i) {
+		uint32_t reg;
+		if (i < 8)
+			reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2);
+		else
+			reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2);
+		rdev->vm_manager.saved_table_addr[i] = RREG32(reg);
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -5642,12 +5961,13 @@ static void cik_vm_decode_fault(struct radeon_device *rdev,
 void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 {
 	struct radeon_ring *ring = &rdev->ring[ridx];
+	int usepfp = (ridx == RADEON_RING_TYPE_GFX_INDEX);
 
 	if (vm == NULL)
 		return;
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	if (vm->id < 8) {
 		radeon_ring_write(ring,
@@ -5661,14 +5981,14 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* update SH_MEM_* regs */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, VMID(vm->id));
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SH_MEM_BASES >> 2);
 	radeon_ring_write(ring, 0);
@@ -5679,7 +5999,7 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 	radeon_ring_write(ring, 0); /* SH_MEM_APE1_LIMIT */
 
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
 	radeon_ring_write(ring, 0);
@@ -5690,14 +6010,14 @@ void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* bits 0-15 are the VM contexts0-15 */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 1 << vm->id);
 
 	/* compute doesn't have PFP */
-	if (ridx == RADEON_RING_TYPE_GFX_INDEX) {
+	if (usepfp) {
 		/* sync PFP to ME, otherwise we might get invalid PFP reads */
 		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
 		radeon_ring_write(ring, 0x0);
@@ -5865,28 +6185,10 @@ static void cik_rlc_start(struct radeon_device *rdev)
 static int cik_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i, size, tmp;
-	const __be32 *fw_data;
 
 	if (!rdev->rlc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_BONAIRE:
-	case CHIP_HAWAII:
-	default:
-		size = BONAIRE_RLC_UCODE_SIZE;
-		break;
-	case CHIP_KAVERI:
-		size = KV_RLC_UCODE_SIZE;
-		break;
-	case CHIP_KABINI:
-		size = KB_RLC_UCODE_SIZE;
-		break;
-	case CHIP_MULLINS:
-		size = ML_RLC_UCODE_SIZE;
-		break;
-	}
-
 	cik_rlc_stop(rdev);
 
 	/* disable CG */
@@ -5910,11 +6212,45 @@ static int cik_rlc_resume(struct radeon_device *rdev)
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
-	fw_data = (const __be32 *)rdev->rlc_fw->data;
+	if (rdev->new_fw) {
+		const struct rlc_firmware_header_v1_0 *hdr =
+			(const struct rlc_firmware_header_v1_0 *)rdev->rlc_fw->data;
+		const __le32 *fw_data = (const __le32 *)
+			(rdev->rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		radeon_ucode_print_rlc_hdr(&hdr->header);
+
+		size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
 		WREG32(RLC_GPM_UCODE_ADDR, 0);
-	for (i = 0; i < size; i++)
-		WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(RLC_GPM_UCODE_ADDR, 0);
+		for (i = 0; i < size; i++)
+			WREG32(RLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(RLC_GPM_UCODE_ADDR, le32_to_cpu(hdr->header.ucode_version));
+	} else {
+		const __be32 *fw_data;
+
+		switch (rdev->family) {
+		case CHIP_BONAIRE:
+		case CHIP_HAWAII:
+		default:
+			size = BONAIRE_RLC_UCODE_SIZE;
+			break;
+		case CHIP_KAVERI:
+			size = KV_RLC_UCODE_SIZE;
+			break;
+		case CHIP_KABINI:
+			size = KB_RLC_UCODE_SIZE;
+			break;
+		case CHIP_MULLINS:
+			size = ML_RLC_UCODE_SIZE;
+			break;
+		}
+
+		fw_data = (const __be32 *)rdev->rlc_fw->data;
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+		for (i = 0; i < size; i++)
+			WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+	}
 
 	/* XXX - find out what chips support lbpw */
 	cik_enable_lbpw(rdev, false);
@@ -6348,11 +6684,10 @@ static void cik_enable_gds_pg(struct radeon_device *rdev, bool enable)
 
 void cik_init_cp_pg_table(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	volatile u32 *dst_ptr;
 	int me, i, max_me = 4;
 	u32 bo_offset = 0;
-	u32 table_offset;
+	u32 table_offset, table_size;
 
 	if (rdev->family == CHIP_KAVERI)
 		max_me = 5;
@@ -6363,24 +6698,71 @@ void cik_init_cp_pg_table(struct radeon_device *rdev)
 	/* write the cp table buffer */
 	dst_ptr = rdev->rlc.cp_table_ptr;
 	for (me = 0; me < max_me; me++) {
-		if (me == 0) {
-			fw_data = (const __be32 *)rdev->ce_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
-		} else if (me == 1) {
-			fw_data = (const __be32 *)rdev->pfp_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
-		} else if (me == 2) {
-			fw_data = (const __be32 *)rdev->me_fw->data;
-			table_offset = CP_ME_TABLE_OFFSET;
+		if (rdev->new_fw) {
+			const __le32 *fw_data;
+			const struct gfx_firmware_header_v1_0 *hdr;
+
+			if (me == 0) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->ce_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 1) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->pfp_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 2) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->me_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else if (me == 3) {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->mec_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->mec_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			} else {
+				hdr = (const struct gfx_firmware_header_v1_0 *)rdev->mec2_fw->data;
+				fw_data = (const __le32 *)
+					(rdev->mec2_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+				table_offset = le32_to_cpu(hdr->jt_offset);
+				table_size = le32_to_cpu(hdr->jt_size);
+			}
+
+			for (i = 0; i < table_size; i ++) {
+				dst_ptr[bo_offset + i] =
+					cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
+			}
+			bo_offset += table_size;
 		} else {
-			fw_data = (const __be32 *)rdev->mec_fw->data;
-			table_offset = CP_MEC_TABLE_OFFSET;
-		}
+			const __be32 *fw_data;
+			table_size = CP_ME_TABLE_SIZE;
+
+			if (me == 0) {
+				fw_data = (const __be32 *)rdev->ce_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else if (me == 1) {
+				fw_data = (const __be32 *)rdev->pfp_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else if (me == 2) {
+				fw_data = (const __be32 *)rdev->me_fw->data;
+				table_offset = CP_ME_TABLE_OFFSET;
+			} else {
+				fw_data = (const __be32 *)rdev->mec_fw->data;
+				table_offset = CP_MEC_TABLE_OFFSET;
+			}
 
-		for (i = 0; i < CP_ME_TABLE_SIZE; i ++) {
-			dst_ptr[bo_offset + i] = cpu_to_le32(be32_to_cpu(fw_data[table_offset + i]));
+			for (i = 0; i < table_size; i ++) {
+				dst_ptr[bo_offset + i] =
+					cpu_to_le32(be32_to_cpu(fw_data[table_offset + i]));
+			}
+			bo_offset += table_size;
 		}
-		bo_offset += CP_ME_TABLE_SIZE;
 	}
 }
 
@@ -7367,17 +7749,17 @@ static inline u32 cik_get_ih_wptr(struct radeon_device *rdev)
 		wptr = RREG32(IH_RB_WPTR);
 
 	if (wptr & RB_OVERFLOW) {
+		wptr &= ~RB_OVERFLOW;
 		/* When a ring buffer overflow happen start parsing interrupt
 		 * from the last not overwritten vector (wptr + 16). Hopefully
 		 * this should allow us to catchup.
 		 */
-		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
-			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			 wptr, rdev->ih.rptr, (wptr + 16) & rdev->ih.ptr_mask);
 		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
-		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -7618,7 +8000,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
@@ -7866,15 +8249,17 @@ restart_ih:
 		/* wptr/rptr are in bytes! */
 		rptr += 16;
 		rptr &= rdev->ih.ptr_mask;
+		WREG32(IH_RB_RPTR, rptr);
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
-	if (queue_reset)
-		schedule_work(&rdev->reset_work);
+	if (queue_reset) {
+		rdev->needs_reset = true;
+		wake_up_all(&rdev->fence_queue);
+	}
 	if (queue_thermal)
 		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
-	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
 
 	/* make sure wptr hasn't changed while processing */
@@ -7900,6 +8285,7 @@ restart_ih:
 static int cik_startup(struct radeon_device *rdev)
 {
 	struct radeon_ring *ring;
+	u32 nop;
 	int r;
 
 	/* enable pcie gen2/3 link */
@@ -8033,9 +8419,18 @@ static int cik_startup(struct radeon_device *rdev)
 	}
 	cik_irq_set(rdev);
 
+	if (rdev->family == CHIP_HAWAII) {
+		if (rdev->new_fw)
+			nop = PACKET3(PACKET3_NOP, 0x3FFF);
+		else
+			nop = RADEON_CP_PACKET2;
+	} else {
+		nop = PACKET3(PACKET3_NOP, 0x3FFF);
+	}
+
 	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 
@@ -8043,7 +8438,7 @@ static int cik_startup(struct radeon_device *rdev)
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 	ring->me = 1; /* first MEC */
@@ -8054,7 +8449,7 @@ static int cik_startup(struct radeon_device *rdev)
 	/* type-2 packets are deprecated on MEC, use type-3 instead */
 	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
 	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
-			     PACKET3(PACKET3_NOP, 0x3FFF));
+			     nop);
 	if (r)
 		return r;
 	/* dGPU only have 1 MEC */
@@ -9052,6 +9447,9 @@ void dce8_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
@@ -9168,6 +9566,9 @@ static void cik_pcie_gen3_enable(struct radeon_device *rdev)
 	int ret, i;
 	u16 tmp16;
 
+	if (pci_is_root_bus(rdev->pdev->bus))
+		return;
+
 	if (radeon_pcie_gen2 == 0)
 		return;
 
@@ -9394,7 +9795,8 @@ static void cik_program_aspm(struct radeon_device *rdev)
 			if (orig != data)
 				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
 
-			if (!disable_clkreq) {
+			if (!disable_clkreq &&
+			    !pci_is_root_bus(rdev->pdev->bus)) {
 				struct pci_dev *root = rdev->pdev->bus->self;
 				u32 lnkcap;
 
diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c
index 8e9d0f1d858e..d748963af08b 100644
--- a/drivers/gpu/drm/radeon/cik_sdma.c
+++ b/drivers/gpu/drm/radeon/cik_sdma.c
@@ -24,6 +24,7 @@
 #include <linux/firmware.h>
 #include <drm/drmP.h>
 #include "radeon.h"
+#include "radeon_ucode.h"
 #include "radeon_asic.h"
 #include "radeon_trace.h"
 #include "cikd.h"
@@ -118,6 +119,7 @@ void cik_sdma_set_wptr(struct radeon_device *rdev,
 		reg = SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET;
 
 	WREG32(reg, (ring->wptr << 2) & 0x3fffc);
+	(void)RREG32(reg);
 }
 
 /**
@@ -419,7 +421,6 @@ static int cik_sdma_rlc_resume(struct radeon_device *rdev)
  */
 static int cik_sdma_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
 	if (!rdev->sdma_fw)
@@ -428,19 +429,48 @@ static int cik_sdma_load_microcode(struct radeon_device *rdev)
 	/* halt the MEs */
 	cik_sdma_enable(rdev, false);
 
-	/* sdma0 */
-	fw_data = (const __be32 *)rdev->sdma_fw->data;
-	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
-	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
-		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
-	WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
-
-	/* sdma1 */
-	fw_data = (const __be32 *)rdev->sdma_fw->data;
-	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
-	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
-		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
-	WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	if (rdev->new_fw) {
+		const struct sdma_firmware_header_v1_0 *hdr =
+			(const struct sdma_firmware_header_v1_0 *)rdev->sdma_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_sdma_hdr(&hdr->header);
+
+		/* sdma0 */
+		fw_data = (const __le32 *)
+			(rdev->sdma_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, le32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+		/* sdma1 */
+		fw_data = (const __le32 *)
+			(rdev->sdma_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, le32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	} else {
+		const __be32 *fw_data;
+
+		/* sdma0 */
+		fw_data = (const __be32 *)rdev->sdma_fw->data;
+		WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+		/* sdma1 */
+		fw_data = (const __be32 *)rdev->sdma_fw->data;
+		WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+			WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+	}
 
 	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
 	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
@@ -459,13 +489,6 @@ int cik_sdma_resume(struct radeon_device *rdev)
 {
 	int r;
 
-	/* Reset dma */
-	WREG32(SRBM_SOFT_RESET, SOFT_RESET_SDMA | SOFT_RESET_SDMA1);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-	RREG32(SRBM_SOFT_RESET);
-
 	r = cik_sdma_load_microcode(rdev);
 	if (r)
 		return r;
@@ -507,18 +530,19 @@ void cik_sdma_fini(struct radeon_device *rdev)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (CIK).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int cik_copy_dma(struct radeon_device *rdev,
-		 uint64_t src_offset, uint64_t dst_offset,
-		 unsigned num_gpu_pages,
-		 struct radeon_fence **fence)
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+				  uint64_t src_offset, uint64_t dst_offset,
+				  unsigned num_gpu_pages,
+				  struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes;
@@ -528,7 +552,7 @@ int cik_copy_dma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -537,10 +561,10 @@ int cik_copy_dma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -559,17 +583,17 @@ int cik_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
 
 /**
@@ -587,16 +611,19 @@ int cik_sdma_ring_test(struct radeon_device *rdev,
 {
 	unsigned i;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	unsigned index;
 	u32 tmp;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ring_lock(rdev, ring, 5);
 	if (r) {
@@ -604,14 +631,14 @@ int cik_sdma_ring_test(struct radeon_device *rdev,
 		return r;
 	}
 	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
-	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
-	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr));
+	radeon_ring_write(ring, lower_32_bits(gpu_addr));
+	radeon_ring_write(ring, upper_32_bits(gpu_addr));
 	radeon_ring_write(ring, 1); /* number of DWs to follow */
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -640,17 +667,20 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib ib;
 	unsigned i;
+	unsigned index;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
 	u32 tmp = 0;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
 	if (r) {
@@ -659,13 +689,13 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 
 	ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
-	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr);
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr);
 	ib.ptr[3] = 1;
 	ib.ptr[4] = 0xDEADBEEF;
 	ib.length_dw = 5;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
@@ -677,7 +707,7 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -719,7 +749,43 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * cik_sdma_vm_set_page - update the page tables using sDMA
+ * cik_sdma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using sDMA (CIK).
+ */
+void cik_sdma_vm_copy_pages(struct radeon_device *rdev,
+			    struct radeon_ib *ib,
+			    uint64_t pe, uint64_t src,
+			    unsigned count)
+{
+	while (count) {
+		unsigned bytes = count * 8;
+		if (bytes > 0x1FFFF8)
+			bytes = 0x1FFFF8;
+
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY,
+			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+		ib->ptr[ib->length_dw++] = bytes;
+		ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(src);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+
+		pe += bytes;
+		src += bytes;
+		count -= bytes / 8;
+	}
+}
+
+/**
+ * cik_sdma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -729,84 +795,103 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using sDMA (CIK).
+ * Update PTEs by writing them manually using sDMA (CIK).
  */
-void cik_sdma_vm_set_page(struct radeon_device *rdev,
-			  struct radeon_ib *ib,
-			  uint64_t pe,
-			  uint64_t addr, unsigned count,
-			  uint32_t incr, uint32_t flags)
+void cik_sdma_vm_write_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if (flags == R600_PTE_GART) {
-		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
-		while (count) {
-			unsigned bytes = count * 8;
-			if (bytes > 0x1FFFF8)
-				bytes = 0x1FFFF8;
-
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-			ib->ptr[ib->length_dw++] = bytes;
-			ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
-			ib->ptr[ib->length_dw++] = lower_32_bits(src);
-			ib->ptr[ib->length_dw++] = upper_32_bits(src);
-			ib->ptr[ib->length_dw++] = lower_32_bits(pe);
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-
-			pe += bytes;
-			src += bytes;
-			count -= bytes / 8;
-		}
-	} else if (flags & R600_PTE_SYSTEM) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-			ib->ptr[ib->length_dw++] = ndw;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE,
+			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+		ib->ptr[ib->length_dw++] = ndw;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
 				value = radeon_vm_map_gart(rdev, addr);
 				value &= 0xFFFFFFFFFFFFF000ULL;
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count;
-			if (ndw > 0x7FFFF)
-				ndw = 0x7FFFF;
-
-			if (flags & R600_PTE_VALID)
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = ndw; /* number of entries */
-			pe += ndw * 8;
-			addr += ndw * incr;
-			count -= ndw;
 		}
 	}
+}
+
+/**
+ * cik_sdma_vm_set_pages - update the page tables using sDMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using sDMA (CIK).
+ */
+void cik_sdma_vm_set_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count;
+		if (ndw > 0x7FFFF)
+			ndw = 0x7FFFF;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = ndw; /* number of entries */
+
+		pe += ndw * 8;
+		addr += ndw * incr;
+		count -= ndw;
+	}
+}
+
+/**
+ * cik_sdma_vm_pad_ib - pad the IB to the required number of dw
+ *
+ * @ib: indirect buffer to fill with padding
+ *
+ */
+void cik_sdma_vm_pad_ib(struct radeon_ib *ib)
+{
 	while (ib->length_dw & 0x7)
 		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
 }
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c
index 47d31e915758..9aad0327e4d1 100644
--- a/drivers/gpu/drm/radeon/cypress_dpm.c
+++ b/drivers/gpu/drm/radeon/cypress_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "evergreend.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
diff --git a/drivers/gpu/drm/radeon/dce3_1_afmt.c b/drivers/gpu/drm/radeon/dce3_1_afmt.c
index 51800e340a57..2fe8cfc966d9 100644
--- a/drivers/gpu/drm/radeon/dce3_1_afmt.c
+++ b/drivers/gpu/drm/radeon/dce3_1_afmt.c
@@ -32,7 +32,7 @@ static void dce3_2_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
@@ -49,8 +49,8 @@ static void dce3_2_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 
 	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
 	if (sad_count < 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -165,7 +165,7 @@ void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *m
 
 	/* disable audio prior to setting up hw */
 	dig->afmt->pin = r600_audio_get_pin(rdev);
-	r600_audio_enable(rdev, dig->afmt->pin, false);
+	r600_audio_enable(rdev, dig->afmt->pin, 0);
 
 	r600_audio_set_dto(encoder, mode->clock);
 
@@ -240,5 +240,5 @@ void dce3_1_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *m
 	r600_hdmi_audio_workaround(encoder);
 
 	/* enable audio after to setting up hw */
-	r600_audio_enable(rdev, dig->afmt->pin, true);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
diff --git a/drivers/gpu/drm/radeon/dce6_afmt.c b/drivers/gpu/drm/radeon/dce6_afmt.c
index 0a65dc7e93e7..f312edf4d50e 100644
--- a/drivers/gpu/drm/radeon/dce6_afmt.c
+++ b/drivers/gpu/drm/radeon/dce6_afmt.c
@@ -136,13 +136,13 @@ void dce6_afmt_write_latency_fields(struct drm_encoder *encoder,
 			tmp = VIDEO_LIPSYNC(connector->video_latency[1]) |
 				AUDIO_LIPSYNC(connector->audio_latency[1]);
 		else
-			tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255);
+			tmp = VIDEO_LIPSYNC(0) | AUDIO_LIPSYNC(0);
 	} else {
 		if (connector->latency_present[0])
 			tmp = VIDEO_LIPSYNC(connector->video_latency[0]) |
 				AUDIO_LIPSYNC(connector->audio_latency[0]);
 		else
-			tmp = VIDEO_LIPSYNC(255) | AUDIO_LIPSYNC(255);
+			tmp = VIDEO_LIPSYNC(0) | AUDIO_LIPSYNC(0);
 	}
 	WREG32_ENDPOINT(offset, AZ_F0_CODEC_PIN_CONTROL_RESPONSE_LIPSYNC, tmp);
 }
@@ -155,7 +155,7 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 offset, tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	if (!dig || !dig->afmt || !dig->afmt->pin)
@@ -164,8 +164,10 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	offset = dig->afmt->pin->offset;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
-		if (connector->encoder == encoder)
+		if (connector->encoder == encoder) {
 			radeon_connector = to_radeon_connector(connector);
+			break;
+		}
 	}
 
 	if (!radeon_connector) {
@@ -173,10 +175,10 @@ void dce6_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
-	if (sad_count <= 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
+	if (sad_count < 0) {
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -225,8 +227,10 @@ void dce6_afmt_write_sad_regs(struct drm_encoder *encoder)
 	offset = dig->afmt->pin->offset;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
-		if (connector->encoder == encoder)
+		if (connector->encoder == encoder) {
 			radeon_connector = to_radeon_connector(connector);
+			break;
+		}
 	}
 
 	if (!radeon_connector) {
@@ -234,7 +238,7 @@ void dce6_afmt_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
+	sad_count = drm_edid_to_sad(radeon_connector_edid(connector), &sads);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 		return;
@@ -280,13 +284,13 @@ static int dce6_audio_chipset_supported(struct radeon_device *rdev)
 
 void dce6_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable)
+		       u8 enable_mask)
 {
 	if (!pin)
 		return;
 
-	WREG32_ENDPOINT(pin->offset, AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL,
-			enable ? AUDIO_ENABLED : 0);
+	WREG32_ENDPOINT(pin->offset, AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL,
+			enable_mask ? AUDIO_ENABLED : 0);
 }
 
 static const u32 pin_offsets[7] =
diff --git a/drivers/gpu/drm/drm_buffer.c b/drivers/gpu/drm/radeon/drm_buffer.c
index 0406110f83ed..f4e0f3a3d7b1 100644
--- a/drivers/gpu/drm/drm_buffer.c
+++ b/drivers/gpu/drm/radeon/drm_buffer.c
@@ -33,7 +33,7 @@
  */
 
 #include <linux/export.h>
-#include <drm/drm_buffer.h>
+#include "drm_buffer.h"
 
 /**
  * Allocate the drm buffer object.
@@ -80,17 +80,12 @@ int drm_buffer_alloc(struct drm_buffer **buf, int size)
 
 error_out:
 
-	/* Only last element can be null pointer so check for it first. */
-	if ((*buf)->data[idx])
-		kfree((*buf)->data[idx]);
-
-	for (--idx; idx >= 0; --idx)
+	for (; idx >= 0; --idx)
 		kfree((*buf)->data[idx]);
 
 	kfree(*buf);
 	return -ENOMEM;
 }
-EXPORT_SYMBOL(drm_buffer_alloc);
 
 /**
  * Copy the user data to the begin of the buffer and reset the processing
@@ -127,7 +122,6 @@ int drm_buffer_copy_from_user(struct drm_buffer *buf,
 	buf->iterator = 0;
 	return 0;
 }
-EXPORT_SYMBOL(drm_buffer_copy_from_user);
 
 /**
  * Free the drm buffer object
@@ -145,7 +139,6 @@ void drm_buffer_free(struct drm_buffer *buf)
 		kfree(buf);
 	}
 }
-EXPORT_SYMBOL(drm_buffer_free);
 
 /**
  * Read an object from buffer that may be split to multiple parts. If object
@@ -182,4 +175,3 @@ void *drm_buffer_read_object(struct drm_buffer *buf,
 	drm_buffer_advance(buf, objsize);
 	return obj;
 }
-EXPORT_SYMBOL(drm_buffer_read_object);
diff --git a/drivers/gpu/drm/radeon/drm_buffer.h b/drivers/gpu/drm/radeon/drm_buffer.h
new file mode 100644
index 000000000000..c80d3a340b94
--- /dev/null
+++ b/drivers/gpu/drm/radeon/drm_buffer.h
@@ -0,0 +1,148 @@
+/**************************************************************************
+ *
+ * Copyright 2010 Pauli Nieminen.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ **************************************************************************/
+/*
+ * Multipart buffer for coping data which is larger than the page size.
+ *
+ * Authors:
+ * Pauli Nieminen <suokkos-at-gmail-dot-com>
+ */
+
+#ifndef _DRM_BUFFER_H_
+#define _DRM_BUFFER_H_
+
+#include <drm/drmP.h>
+
+struct drm_buffer {
+	int iterator;
+	int size;
+	char *data[];
+};
+
+
+/**
+ * Return the index of page that buffer is currently pointing at.
+ */
+static inline int drm_buffer_page(struct drm_buffer *buf)
+{
+	return buf->iterator / PAGE_SIZE;
+}
+/**
+ * Return the index of the current byte in the page
+ */
+static inline int drm_buffer_index(struct drm_buffer *buf)
+{
+	return buf->iterator & (PAGE_SIZE - 1);
+}
+/**
+ * Return number of bytes that is left to process
+ */
+static inline int drm_buffer_unprocessed(struct drm_buffer *buf)
+{
+	return buf->size - buf->iterator;
+}
+
+/**
+ * Advance the buffer iterator number of bytes that is given.
+ */
+static inline void drm_buffer_advance(struct drm_buffer *buf, int bytes)
+{
+	buf->iterator += bytes;
+}
+
+/**
+ * Allocate the drm buffer object.
+ *
+ *   buf: A pointer to a pointer where the object is stored.
+ *   size: The number of bytes to allocate.
+ */
+extern int drm_buffer_alloc(struct drm_buffer **buf, int size);
+
+/**
+ * Copy the user data to the begin of the buffer and reset the processing
+ * iterator.
+ *
+ *   user_data: A pointer the data that is copied to the buffer.
+ *   size: The Number of bytes to copy.
+ */
+extern int drm_buffer_copy_from_user(struct drm_buffer *buf,
+		void __user *user_data, int size);
+
+/**
+ * Free the drm buffer object
+ */
+extern void drm_buffer_free(struct drm_buffer *buf);
+
+/**
+ * Read an object from buffer that may be split to multiple parts. If object
+ * is not split function just returns the pointer to object in buffer. But in
+ * case of split object data is copied to given stack object that is suplied
+ * by caller.
+ *
+ * The processing location of the buffer is also advanced to the next byte
+ * after the object.
+ *
+ *   objsize: The size of the objet in bytes.
+ *   stack_obj: A pointer to a memory location where object can be copied.
+ */
+extern void *drm_buffer_read_object(struct drm_buffer *buf,
+		int objsize, void *stack_obj);
+
+/**
+ * Returns the pointer to the dword which is offset number of elements from the
+ * current processing location.
+ *
+ * Caller must make sure that dword is not split in the buffer. This
+ * requirement is easily met if all the sizes of objects in buffer are
+ * multiples of dword and PAGE_SIZE is multiple dword.
+ *
+ * Call to this function doesn't change the processing location.
+ *
+ *   offset: The index of the dword relative to the internat iterator.
+ */
+static inline void *drm_buffer_pointer_to_dword(struct drm_buffer *buffer,
+		int offset)
+{
+	int iter = buffer->iterator + offset * 4;
+	return &buffer->data[iter / PAGE_SIZE][iter & (PAGE_SIZE - 1)];
+}
+/**
+ * Returns the pointer to the dword which is offset number of elements from
+ * the current processing location.
+ *
+ * Call to this function doesn't change the processing location.
+ *
+ *   offset: The index of the byte relative to the internat iterator.
+ */
+static inline void *drm_buffer_pointer_to_byte(struct drm_buffer *buffer,
+		int offset)
+{
+	int iter = buffer->iterator + offset;
+	return &buffer->data[iter / PAGE_SIZE][iter & (PAGE_SIZE - 1)];
+}
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 15e4f28015e1..85995b4e3338 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -22,7 +22,6 @@
  * Authors: Alex Deucher
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
 #include "radeon.h"
@@ -2346,6 +2345,9 @@ void evergreen_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
@@ -2424,7 +2426,6 @@ static int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
@@ -2554,6 +2555,7 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav
 					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 					tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
 					WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
+					WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
 				}
 			} else {
 				tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]);
@@ -2677,7 +2679,7 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s
 		if (save->crtc_enabled[i]) {
 			if (ASIC_IS_DCE6(rdev)) {
 				tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]);
-				tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
+				tmp &= ~EVERGREEN_CRTC_BLANK_DATA_EN;
 				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
 				WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
 				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
@@ -2870,7 +2872,7 @@ static int evergreen_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cp_me = 0xff;
 	WREG32(CP_ME_CNTL, cp_me);
@@ -2913,7 +2915,7 @@ static int evergreen_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	return 0;
 }
@@ -3007,7 +3009,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	u32 vgt_cache_invalidation;
 	u32 hdp_host_path_cntl, tmp;
 	u32 disabled_rb_mask;
-	int i, j, num_shader_engines, ps_thread_count;
+	int i, j, ps_thread_count;
 
 	switch (rdev->family) {
 	case CHIP_CYPRESS:
@@ -3305,8 +3307,6 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	rdev->config.evergreen.tile_config |=
 		((gb_addr_config & 0x30000000) >> 28) << 12;
 
-	num_shader_engines = (gb_addr_config & NUM_SHADER_ENGINES(3) >> 12) + 1;
-
 	if ((rdev->family >= CHIP_CEDAR) && (rdev->family <= CHIP_HEMLOCK)) {
 		u32 efuse_straps_4;
 		u32 efuse_straps_3;
@@ -4023,7 +4023,8 @@ int sumo_rlc_init(struct radeon_device *rdev)
 		/* save restore block */
 		if (rdev->rlc.save_restore_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj);
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     NULL, &rdev->rlc.save_restore_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
 				return r;
@@ -4101,7 +4102,8 @@ int sumo_rlc_init(struct radeon_device *rdev)
 
 		if (rdev->rlc.clear_state_obj == NULL) {
 			r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     NULL, &rdev->rlc.clear_state_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
 				sumo_rlc_fini(rdev);
@@ -4175,8 +4177,10 @@ int sumo_rlc_init(struct radeon_device *rdev)
 
 	if (rdev->rlc.cp_table_size) {
 		if (rdev->rlc.cp_table_obj == NULL) {
-			r = radeon_bo_create(rdev, rdev->rlc.cp_table_size, PAGE_SIZE, true,
-					     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.cp_table_obj);
+			r = radeon_bo_create(rdev, rdev->rlc.cp_table_size,
+					     PAGE_SIZE, true,
+					     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+					     NULL, &rdev->rlc.cp_table_obj);
 			if (r) {
 				dev_warn(rdev->dev, "(%d) create RLC cp table bo failed\n", r);
 				sumo_rlc_fini(rdev);
@@ -4746,17 +4750,17 @@ static u32 evergreen_get_ih_wptr(struct radeon_device *rdev)
 		wptr = RREG32(IH_RB_WPTR);
 
 	if (wptr & RB_OVERFLOW) {
+		wptr &= ~RB_OVERFLOW;
 		/* When a ring buffer overflow happen start parsing interrupt
 		 * from the last not overwritten vector (wptr + 16). Hopefully
 		 * this should allow us to catchup.
 		 */
-		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
-			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			 wptr, rdev->ih.rptr, (wptr + 16) & rdev->ih.ptr_mask);
 		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
-		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -4961,7 +4965,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
@@ -5133,6 +5138,7 @@ restart_ih:
 		/* wptr/rptr are in bytes! */
 		rptr += 16;
 		rptr &= rdev->ih.ptr_mask;
+		WREG32(IH_RB_RPTR, rptr);
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
@@ -5141,7 +5147,6 @@ restart_ih:
 	if (queue_thermal && rdev->pm.dpm_enabled)
 		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
-	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
 
 	/* make sure wptr hasn't changed while processing */
diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c
index 478caefe0fef..66bcfadeedd1 100644
--- a/drivers/gpu/drm/radeon/evergreen_dma.c
+++ b/drivers/gpu/drm/radeon/evergreen_dma.c
@@ -104,12 +104,14 @@ void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int evergreen_copy_dma(struct radeon_device *rdev,
-		       uint64_t src_offset, uint64_t dst_offset,
-		       unsigned num_gpu_pages,
-		       struct radeon_fence **fence)
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+					uint64_t src_offset,
+					uint64_t dst_offset,
+					unsigned num_gpu_pages,
+					struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
@@ -119,7 +121,7 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -128,10 +130,10 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -148,17 +150,17 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index 1ec0e6e83f9f..53abd9b17a50 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -38,6 +38,37 @@ extern void dce6_afmt_select_pin(struct drm_encoder *encoder);
 extern void dce6_afmt_write_latency_fields(struct drm_encoder *encoder,
 					   struct drm_display_mode *mode);
 
+/* enable the audio stream */
+static void dce4_audio_enable(struct radeon_device *rdev,
+			      struct r600_audio_pin *pin,
+			      u8 enable_mask)
+{
+	u32 tmp = RREG32(AZ_HOT_PLUG_CONTROL);
+
+	if (!pin)
+		return;
+
+	if (enable_mask) {
+		tmp |= AUDIO_ENABLED;
+		if (enable_mask & 1)
+			tmp |= PIN0_AUDIO_ENABLED;
+		if (enable_mask & 2)
+			tmp |= PIN1_AUDIO_ENABLED;
+		if (enable_mask & 4)
+			tmp |= PIN2_AUDIO_ENABLED;
+		if (enable_mask & 8)
+			tmp |= PIN3_AUDIO_ENABLED;
+	} else {
+		tmp &= ~(AUDIO_ENABLED |
+			 PIN0_AUDIO_ENABLED |
+			 PIN1_AUDIO_ENABLED |
+			 PIN2_AUDIO_ENABLED |
+			 PIN3_AUDIO_ENABLED);
+	}
+
+	WREG32(AZ_HOT_PLUG_CONTROL, tmp);
+}
+
 /*
  * update the N and CTS parameters for a given pixel clock rate
  */
@@ -102,7 +133,7 @@ static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector = NULL;
 	u32 tmp;
-	u8 *sadb;
+	u8 *sadb = NULL;
 	int sad_count;
 
 	list_for_each_entry(connector, &encoder->dev->mode_config.connector_list, head) {
@@ -117,10 +148,10 @@ static void dce4_afmt_write_speaker_allocation(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_speaker_allocation(radeon_connector->edid, &sadb);
-	if (sad_count <= 0) {
-		DRM_ERROR("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
-		return;
+	sad_count = drm_edid_to_speaker_allocation(radeon_connector_edid(connector), &sadb);
+	if (sad_count < 0) {
+		DRM_DEBUG("Couldn't read Speaker Allocation Data Block: %d\n", sad_count);
+		sad_count = 0;
 	}
 
 	/* program the speaker allocation */
@@ -172,7 +203,7 @@ static void evergreen_hdmi_write_sad_regs(struct drm_encoder *encoder)
 		return;
 	}
 
-	sad_count = drm_edid_to_sad(radeon_connector->edid, &sads);
+	sad_count = drm_edid_to_sad(radeon_connector_edid(connector), &sads);
 	if (sad_count <= 0) {
 		DRM_ERROR("Couldn't read SADs: %d\n", sad_count);
 		return;
@@ -318,10 +349,10 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	/* disable audio prior to setting up hw */
 	if (ASIC_IS_DCE6(rdev)) {
 		dig->afmt->pin = dce6_audio_get_pin(rdev);
-		dce6_audio_enable(rdev, dig->afmt->pin, false);
+		dce6_audio_enable(rdev, dig->afmt->pin, 0);
 	} else {
 		dig->afmt->pin = r600_audio_get_pin(rdev);
-		r600_audio_enable(rdev, dig->afmt->pin, false);
+		dce4_audio_enable(rdev, dig->afmt->pin, 0);
 	}
 
 	evergreen_audio_set_dto(encoder, mode->clock);
@@ -463,13 +494,15 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 
 	/* enable audio after to setting up hw */
 	if (ASIC_IS_DCE6(rdev))
-		dce6_audio_enable(rdev, dig->afmt->pin, true);
+		dce6_audio_enable(rdev, dig->afmt->pin, 1);
 	else
-		r600_audio_enable(rdev, dig->afmt->pin, true);
+		dce4_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
 
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
 {
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 
@@ -482,6 +515,14 @@ void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	if (!enable && !dig->afmt->enabled)
 		return;
 
+	if (!enable && dig->afmt->pin) {
+		if (ASIC_IS_DCE6(rdev))
+			dce6_audio_enable(rdev, dig->afmt->pin, 0);
+		else
+			dce4_audio_enable(rdev, dig->afmt->pin, 0);
+		dig->afmt->pin = NULL;
+	}
+
 	dig->afmt->enabled = enable;
 
 	DRM_DEBUG("%sabling HDMI interface @ 0x%04X for encoder 0x%x\n",
diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c
index 9ef8c38f2d66..9b42001295ba 100644
--- a/drivers/gpu/drm/radeon/kv_dpm.c
+++ b/drivers/gpu/drm/radeon/kv_dpm.c
@@ -33,6 +33,8 @@
 #define KV_MINIMUM_ENGINE_CLOCK         800
 #define SMC_RAM_END                     0x40000
 
+static int kv_enable_nb_dpm(struct radeon_device *rdev,
+			    bool enable);
 static void kv_init_graphics_levels(struct radeon_device *rdev);
 static int kv_calculate_ds_divider(struct radeon_device *rdev);
 static int kv_calculate_nbps_level_settings(struct radeon_device *rdev);
@@ -1295,6 +1297,9 @@ void kv_dpm_disable(struct radeon_device *rdev)
 {
 	kv_smc_bapm_enable(rdev, false);
 
+	if (rdev->family == CHIP_MULLINS)
+		kv_enable_nb_dpm(rdev, false);
+
 	/* powerup blocks */
 	kv_dpm_powergate_acp(rdev, false);
 	kv_dpm_powergate_samu(rdev, false);
@@ -1438,14 +1443,14 @@ static int kv_update_uvd_dpm(struct radeon_device *rdev, bool gate)
 	return kv_enable_uvd_dpm(rdev, !gate);
 }
 
-static u8 kv_get_vce_boot_level(struct radeon_device *rdev)
+static u8 kv_get_vce_boot_level(struct radeon_device *rdev, u32 evclk)
 {
 	u8 i;
 	struct radeon_vce_clock_voltage_dependency_table *table =
 		&rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table;
 
 	for (i = 0; i < table->count; i++) {
-		if (table->entries[i].evclk >= 0) /* XXX */
+		if (table->entries[i].evclk >= evclk)
 			break;
 	}
 
@@ -1468,7 +1473,7 @@ static int kv_update_vce_dpm(struct radeon_device *rdev,
 		if (pi->caps_stable_p_state)
 			pi->vce_boot_level = table->count - 1;
 		else
-			pi->vce_boot_level = kv_get_vce_boot_level(rdev);
+			pi->vce_boot_level = kv_get_vce_boot_level(rdev, radeon_new_state->evclk);
 
 		ret = kv_copy_bytes_to_smc(rdev,
 					   pi->dpm_table_start +
@@ -1769,15 +1774,24 @@ static int kv_update_dfs_bypass_settings(struct radeon_device *rdev,
 	return ret;
 }
 
-static int kv_enable_nb_dpm(struct radeon_device *rdev)
+static int kv_enable_nb_dpm(struct radeon_device *rdev,
+			    bool enable)
 {
 	struct kv_power_info *pi = kv_get_pi(rdev);
 	int ret = 0;
 
-	if (pi->enable_nb_dpm && !pi->nb_dpm_enabled) {
-		ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_NBDPM_Enable);
-		if (ret == 0)
-			pi->nb_dpm_enabled = true;
+	if (enable) {
+		if (pi->enable_nb_dpm && !pi->nb_dpm_enabled) {
+			ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_NBDPM_Enable);
+			if (ret == 0)
+				pi->nb_dpm_enabled = true;
+		}
+	} else {
+		if (pi->enable_nb_dpm && pi->nb_dpm_enabled) {
+			ret = kv_notify_message_to_smu(rdev, PPSMC_MSG_NBDPM_Disable);
+			if (ret == 0)
+				pi->nb_dpm_enabled = false;
+		}
 	}
 
 	return ret;
@@ -1864,7 +1878,7 @@ int kv_dpm_set_power_state(struct radeon_device *rdev)
 			}
 			kv_update_sclk_t(rdev);
 			if (rdev->family == CHIP_MULLINS)
-				kv_enable_nb_dpm(rdev);
+				kv_enable_nb_dpm(rdev, true);
 		}
 	} else {
 		if (pi->enable_dpm) {
@@ -1889,7 +1903,7 @@ int kv_dpm_set_power_state(struct radeon_device *rdev)
 			}
 			kv_update_acp_boot_level(rdev);
 			kv_update_sclk_t(rdev);
-			kv_enable_nb_dpm(rdev);
+			kv_enable_nb_dpm(rdev, true);
 		}
 	}
 
@@ -2711,7 +2725,11 @@ int kv_dpm_init(struct radeon_device *rdev)
 
         pi->sram_end = SMC_RAM_END;
 
-	pi->enable_nb_dpm = true;
+	/* Enabling nb dpm on an asrock system prevents dpm from working */
+	if (rdev->pdev->subsystem_vendor == 0x1849)
+		pi->enable_nb_dpm = false;
+	else
+		pi->enable_nb_dpm = true;
 
 	pi->caps_power_containment = true;
 	pi->caps_cac = true;
@@ -2726,7 +2744,19 @@ int kv_dpm_init(struct radeon_device *rdev)
 	pi->caps_sclk_ds = true;
 	pi->enable_auto_thermal_throttling = true;
 	pi->disable_nb_ps3_in_battery = false;
-	pi->bapm_enable = true;
+	if (radeon_bapm == -1) {
+		/* There are stability issues reported on with
+		 * bapm enabled on an asrock system.
+		 */
+		if (rdev->pdev->subsystem_vendor == 0x1849)
+			pi->bapm_enable = false;
+		else
+			pi->bapm_enable = true;
+	} else if (radeon_bapm == 0) {
+		pi->bapm_enable = false;
+	} else {
+		pi->bapm_enable = true;
+	}
 	pi->voltage_drop_t = 0;
 	pi->caps_sclk_throttle_low_notification = false;
 	pi->caps_fps = false; /* true? */
@@ -2770,6 +2800,8 @@ void kv_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
 		tmp = (RREG32_SMC(SMU_VOLTAGE_STATUS) & SMU_VOLTAGE_CURRENT_LEVEL_MASK) >>
 			SMU_VOLTAGE_CURRENT_LEVEL_SHIFT;
 		vddc = kv_convert_8bit_index_to_voltage(rdev, (u16)tmp);
+		seq_printf(m, "uvd    %sabled\n", pi->uvd_power_gated ? "dis" : "en");
+		seq_printf(m, "vce    %sabled\n", pi->vce_power_gated ? "dis" : "en");
 		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
 			   current_index, sclk, vddc);
 	}
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 5a33ca681867..3faee58946dd 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -1229,7 +1229,6 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -1272,7 +1271,7 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 		WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0);
 		WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), rdev->vm_manager.max_pfn);
 		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			rdev->gart.table_addr >> 12);
+		       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-7 */
@@ -1304,6 +1303,13 @@ static int cayman_pcie_gart_enable(struct radeon_device *rdev)
 
 static void cayman_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 8; ++i) {
+		rdev->vm_manager.saved_table_addr[i] = RREG32(
+			VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2));
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -1506,7 +1512,7 @@ static int cayman_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cayman_cp_enable(rdev, true);
 
@@ -1548,7 +1554,7 @@ static int cayman_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	/* XXX init other rings */
 
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c
index 6378e0276691..f26f0a9fb522 100644
--- a/drivers/gpu/drm/radeon/ni_dma.c
+++ b/drivers/gpu/drm/radeon/ni_dma.c
@@ -191,12 +191,6 @@ int cayman_dma_resume(struct radeon_device *rdev)
 	u32 reg_offset, wb_offset;
 	int i, r;
 
-	/* Reset dma */
-	WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
 	for (i = 0; i < 2; i++) {
 		if (i == 0) {
 			ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
@@ -307,7 +301,43 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * cayman_dma_vm_set_page - update the page tables using the DMA
+ * cayman_dma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr where to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using the DMA (cayman/TN).
+ */
+void cayman_dma_vm_copy_pages(struct radeon_device *rdev,
+			      struct radeon_ib *ib,
+			      uint64_t pe, uint64_t src,
+			      unsigned count)
+{
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
+						      0, 0, ndw);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
+
+		pe += ndw * 4;
+		src += ndw * 4;
+		count -= ndw / 2;
+	}
+}
+
+/**
+ * cayman_dma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -315,71 +345,103 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @addr: dst addr to write into pe
  * @count: number of page entries to update
  * @incr: increase next addr by incr bytes
- * @flags: hw access flags 
+ * @flags: hw access flags
  *
- * Update the page tables using the DMA (cayman/TN).
+ * Update PTEs by writing them manually using the DMA (cayman/TN).
  */
-void cayman_dma_vm_set_page(struct radeon_device *rdev,
-			    struct radeon_ib *ib,
-			    uint64_t pe,
-			    uint64_t addr, unsigned count,
-			    uint32_t incr, uint32_t flags)
+void cayman_dma_vm_write_pages(struct radeon_device *rdev,
+			       struct radeon_ib *ib,
+			       uint64_t pe,
+			       uint64_t addr, unsigned count,
+			       uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if ((flags & R600_PTE_SYSTEM) || (count == 1)) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, ndw);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
-				if (flags & R600_PTE_SYSTEM) {
-					value = radeon_vm_map_gart(rdev, addr);
-					value &= 0xFFFFFFFFFFFFF000ULL;
-				} else if (flags & R600_PTE_VALID) {
-					value = addr;
-				} else {
-					value = 0;
-				}
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			if (flags & R600_PTE_VALID)
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE,
+						      0, 0, ndw);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
+				value = radeon_vm_map_gart(rdev, addr);
+				value &= 0xFFFFFFFFFFFFF000ULL;
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			pe += ndw * 4;
-			addr += (ndw / 2) * incr;
-			count -= ndw / 2;
 		}
 	}
+}
+
+/**
+ * cayman_dma_vm_set_pages - update the page tables using the DMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: hw access flags
+ *
+ * Update the page tables using the DMA (cayman/TN).
+ */
+void cayman_dma_vm_set_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+
+		pe += ndw * 4;
+		addr += (ndw / 2) * incr;
+		count -= ndw / 2;
+	}
+}
+
+/**
+ * cayman_dma_vm_pad_ib - pad the IB to the required number of dw
+ *
+ * @ib: indirect buffer to fill with padding
+ *
+ */
+void cayman_dma_vm_pad_ib(struct radeon_ib *ib)
+{
 	while (ib->length_dw & 0x7)
 		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0);
 }
diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c
index 01fc4888e6fe..6d2f16cf2c1c 100644
--- a/drivers/gpu/drm/radeon/ni_dpm.c
+++ b/drivers/gpu/drm/radeon/ni_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "nid.h"
 #include "r600_dpm.h"
 #include "ni_dpm.h"
@@ -789,7 +790,6 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
 	bool disable_mclk_switching;
 	u32 mclk;
 	u16 vddci;
-	u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc;
 	int i;
 
 	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
@@ -816,29 +816,6 @@ static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
 		}
 	}
 
-	/* limit clocks to max supported clocks based on voltage dependency tables */
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
-							&max_sclk_vddc);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
-							&max_mclk_vddci);
-	btc_get_max_clock_from_voltage_dependency_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
-							&max_mclk_vddc);
-
-	for (i = 0; i < ps->performance_level_count; i++) {
-		if (max_sclk_vddc) {
-			if (ps->performance_levels[i].sclk > max_sclk_vddc)
-				ps->performance_levels[i].sclk = max_sclk_vddc;
-		}
-		if (max_mclk_vddci) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddci)
-				ps->performance_levels[i].mclk = max_mclk_vddci;
-		}
-		if (max_mclk_vddc) {
-			if (ps->performance_levels[i].mclk > max_mclk_vddc)
-				ps->performance_levels[i].mclk = max_mclk_vddc;
-		}
-	}
-
 	/* XXX validate the min clocks required for display */
 
 	/* adjust low state */
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 1544efcf1c3a..b53b31a7b76f 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -652,7 +652,6 @@ int r100_pci_gart_enable(struct radeon_device *rdev)
 {
 	uint32_t tmp;
 
-	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32(RADEON_AIC_CNTL, tmp);
@@ -683,7 +682,7 @@ void r100_pci_gart_disable(struct radeon_device *rdev)
 }
 
 void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
-			    uint64_t addr)
+			    uint64_t addr, uint32_t flags)
 {
 	u32 *gtt = rdev->gart.ptr;
 	gtt[i] = cpu_to_le32(lower_32_bits(addr));
@@ -822,6 +821,20 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc)
 		return RREG32(RADEON_CRTC2_CRNT_FRAME);
 }
 
+/**
+ * r100_ring_hdp_flush - flush Host Data Path via the ring buffer
+ * rdev: radeon device structure
+ * ring: ring buffer struct for emitting packets
+ */
+static void r100_ring_hdp_flush(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(ring, rdev->config.r100.hdp_cntl |
+				RADEON_HDP_READ_BUFFER_INVALIDATE);
+	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(ring, rdev->config.r100.hdp_cntl);
+}
+
 /* Who ever call radeon_fence_emit should call ring_lock and ask
  * for enough space (today caller are ib schedule and buffer move) */
 void r100_fence_ring_emit(struct radeon_device *rdev,
@@ -838,11 +851,7 @@ void r100_fence_ring_emit(struct radeon_device *rdev,
 	/* Wait until IDLE & CLEAN */
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
 	radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl |
-				RADEON_HDP_READ_BUFFER_INVALIDATE);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl);
+	r100_ring_hdp_flush(rdev, ring);
 	/* Emit fence sequence & fire IRQ */
 	radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0));
 	radeon_ring_write(ring, fence->seq);
@@ -860,13 +869,14 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
 	return false;
 }
 
-int r100_copy_blit(struct radeon_device *rdev,
-		   uint64_t src_offset,
-		   uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence)
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+				    uint64_t src_offset,
+				    uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
 	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	struct radeon_fence *fence;
 	uint32_t cur_pages;
 	uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE;
 	uint32_t pitch;
@@ -887,7 +897,7 @@ int r100_copy_blit(struct radeon_device *rdev,
 	r = radeon_ring_lock(rdev, ring, ndw);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw);
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 	}
 	while (num_gpu_pages > 0) {
 		cur_pages = num_gpu_pages;
@@ -927,11 +937,13 @@ int r100_copy_blit(struct radeon_device *rdev,
 			  RADEON_WAIT_2D_IDLECLEAN |
 			  RADEON_WAIT_HOST_IDLECLEAN |
 			  RADEON_WAIT_DMA_GUI_IDLE);
-	if (fence) {
-		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return ERR_PTR(r);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
-	return r;
+	radeon_ring_unlock_commit(rdev, ring, false);
+	return fence;
 }
 
 static int r100_cp_wait_for_idle(struct radeon_device *rdev)
@@ -963,7 +975,7 @@ void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			  RADEON_ISYNC_ANY3D_IDLE2D |
 			  RADEON_ISYNC_WAIT_IDLEGUI |
 			  RADEON_ISYNC_CPSCRATCH_IDLEGUI);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 
@@ -1401,7 +1413,6 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p,
  */
 int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 {
-	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	struct radeon_crtc *radeon_crtc;
 	struct radeon_cs_packet p3reloc, waitreloc;
@@ -1441,12 +1452,11 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p)
 	header = radeon_get_ib_value(p, h_idx);
 	crtc_id = radeon_get_ib_value(p, h_idx + 5);
 	reg = R100_CP_PACKET0_GET_REG(header);
-	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
+	crtc = drm_crtc_find(p->rdev->ddev, crtc_id);
+	if (!crtc) {
 		DRM_ERROR("cannot find crtc %d\n", crtc_id);
 		return -ENOENT;
 	}
-	crtc = obj_to_crtc(obj);
 	radeon_crtc = to_radeon_crtc(crtc);
 	crtc_id = radeon_crtc->crtc_id;
 
@@ -3197,6 +3207,9 @@ void r100_bandwidth_update(struct radeon_device *rdev)
 	uint32_t pixel_bytes1 = 0;
 	uint32_t pixel_bytes2 = 0;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled) {
@@ -3631,7 +3644,7 @@ int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 	radeon_ring_write(ring, PACKET0(scratch, 0));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF) {
@@ -3693,7 +3706,7 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[6] = PACKET2(0);
 	ib.ptr[7] = PACKET2(0);
 	ib.length_dw = 8;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 		goto free_ib;
@@ -4067,39 +4080,6 @@ int r100_init(struct radeon_device *rdev)
 	return 0;
 }
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
-		      bool always_indirect)
-{
-	if (reg < rdev->rmmio_size && !always_indirect)
-		return readl(((void __iomem *)rdev->rmmio) + reg);
-	else {
-		unsigned long flags;
-		uint32_t ret;
-
-		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		ret = readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
-
-		return ret;
-	}
-}
-
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
-		  bool always_indirect)
-{
-	if (reg < rdev->rmmio_size && !always_indirect)
-		writel(v, ((void __iomem *)rdev->rmmio) + reg);
-	else {
-		unsigned long flags;
-
-		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
-	}
-}
-
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg)
 {
 	if (reg < rdev->rio_mem_size)
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index 58f0473aa73f..732d4938aab7 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -80,13 +80,14 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0)
 	return vtx_size;
 }
 
-int r200_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset,
-		  uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset,
+				   uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv)
 {
 	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	struct radeon_fence *fence;
 	uint32_t size;
 	uint32_t cur_size;
 	int i, num_loops;
@@ -98,7 +99,7 @@ int r200_copy_dma(struct radeon_device *rdev,
 	r = radeon_ring_lock(rdev, ring, num_loops * 4 + 64);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 	/* Must wait for 2D idle & clean before DMA or hangs might happen */
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
@@ -118,11 +119,13 @@ int r200_copy_dma(struct radeon_device *rdev,
 	}
 	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
 	radeon_ring_write(ring, RADEON_WAIT_DMA_GUI_IDLE);
-	if (fence) {
-		r = radeon_fence_emit(rdev, fence, RADEON_RING_TYPE_GFX_INDEX);
+	r = radeon_fence_emit(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return ERR_PTR(r);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
-	return r;
+	radeon_ring_unlock_commit(rdev, ring, false);
+	return fence;
 }
 
 
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 3c21d77a483d..1bc4704034ce 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -69,17 +69,23 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev)
 	mb();
 }
 
+#define R300_PTE_UNSNOOPED (1 << 0)
 #define R300_PTE_WRITEABLE (1 << 2)
 #define R300_PTE_READABLE  (1 << 3)
 
 void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
-			      uint64_t addr)
+			      uint64_t addr, uint32_t flags)
 {
 	void __iomem *ptr = rdev->gart.ptr;
 
 	addr = (lower_32_bits(addr) >> 8) |
-	       ((upper_32_bits(addr) & 0xff) << 24) |
-	       R300_PTE_WRITEABLE | R300_PTE_READABLE;
+		((upper_32_bits(addr) & 0xff) << 24);
+	if (flags & RADEON_GART_PAGE_READ)
+		addr |= R300_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		addr |= R300_PTE_WRITEABLE;
+	if (!(flags & RADEON_GART_PAGE_SNOOP))
+		addr |= R300_PTE_UNSNOOPED;
 	/* on x86 we want this to be CPU endian, on powerpc
 	 * on powerpc without HW swappers, it'll get swapped on way
 	 * into VRAM - so no need for cpu_to_le32 on VRAM tables */
@@ -120,7 +126,6 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
 	WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp);
@@ -290,7 +295,7 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring,
 			  R300_GEOMETRY_ROUND_NEAREST |
 			  R300_COLOR_ROUND_NEAREST);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 static void r300_errata(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c
index 84b1d5367a11..9418e388b045 100644
--- a/drivers/gpu/drm/radeon/r300_cmdbuf.c
+++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c
@@ -34,10 +34,10 @@
  */
 
 #include <drm/drmP.h>
-#include <drm/drm_buffer.h>
 #include <drm/radeon_drm.h>
 #include "radeon_drv.h"
 #include "r300_reg.h"
+#include "drm_buffer.h"
 
 #include <asm/unaligned.h>
 
diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c
index 802b19220a21..2828605aef3f 100644
--- a/drivers/gpu/drm/radeon/r420.c
+++ b/drivers/gpu/drm/radeon/r420.c
@@ -219,7 +219,7 @@ static void r420_cp_errata_init(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET0(R300_CP_RESYNC_ADDR, 1));
 	radeon_ring_write(ring, rdev->config.r300.resync_scratch);
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 static void r420_cp_errata_fini(struct radeon_device *rdev)
@@ -232,7 +232,7 @@ static void r420_cp_errata_fini(struct radeon_device *rdev)
 	radeon_ring_lock(rdev, ring, 8);
 	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
 	radeon_ring_write(ring, R300_RB3D_DC_FINISH);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	radeon_scratch_free(rdev, rdev->config.r300.resync_scratch);
 }
 
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 3c69f58e46ef..56b02927cd3d 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -122,6 +122,94 @@ u32 r600_get_xclk(struct radeon_device *rdev)
 
 int r600_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
 {
+	unsigned fb_div = 0, ref_div, vclk_div = 0, dclk_div = 0;
+	int r;
+
+	/* bypass vclk and dclk with bclk */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(1) | DCLK_SRC_SEL(1),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	/* assert BYPASS_EN, deassert UPLL_RESET, UPLL_SLEEP and UPLL_CTLREQ */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_BYPASS_EN_MASK, ~(
+		 UPLL_RESET_MASK | UPLL_SLEEP_MASK | UPLL_CTLREQ_MASK));
+
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(GFX_MACRO_BYPASS_CNTL, UPLL_BYPASS_CNTL,
+			 ~UPLL_BYPASS_CNTL);
+
+	if (!vclk || !dclk) {
+		/* keep the Bypass mode, put PLL to sleep */
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_SLEEP_MASK, ~UPLL_SLEEP_MASK);
+		return 0;
+	}
+
+	if (rdev->clock.spll.reference_freq == 10000)
+		ref_div = 34;
+	else
+		ref_div = 4;
+
+	r = radeon_uvd_calc_upll_dividers(rdev, vclk, dclk, 50000, 160000,
+					  ref_div + 1, 0xFFF, 2, 30, ~0,
+					  &fb_div, &vclk_div, &dclk_div);
+	if (r)
+		return r;
+
+	if (rdev->family >= CHIP_RV670 && rdev->family < CHIP_RS780)
+		fb_div >>= 1;
+	else
+		fb_div |= 1;
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+        if (r)
+                return r;
+
+	/* assert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_RESET_MASK, ~UPLL_RESET_MASK);
+
+	/* For RS780 we have to choose ref clk */
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(CG_UPLL_FUNC_CNTL, UPLL_REFCLK_SRC_SEL_MASK,
+			 ~UPLL_REFCLK_SRC_SEL_MASK);
+
+	/* set the required fb, ref and post divder values */
+	WREG32_P(CG_UPLL_FUNC_CNTL,
+		 UPLL_FB_DIV(fb_div) |
+		 UPLL_REF_DIV(ref_div),
+		 ~(UPLL_FB_DIV_MASK | UPLL_REF_DIV_MASK));
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 UPLL_SW_HILEN(vclk_div >> 1) |
+		 UPLL_SW_LOLEN((vclk_div >> 1) + (vclk_div & 1)) |
+		 UPLL_SW_HILEN2(dclk_div >> 1) |
+		 UPLL_SW_LOLEN2((dclk_div >> 1) + (dclk_div & 1)) |
+		 UPLL_DIVEN_MASK | UPLL_DIVEN2_MASK,
+		 ~UPLL_SW_MASK);
+
+	/* give the PLL some time to settle */
+	mdelay(15);
+
+	/* deassert PLL_RESET */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_RESET_MASK);
+
+	mdelay(15);
+
+	/* deassert BYPASS EN */
+	WREG32_P(CG_UPLL_FUNC_CNTL, 0, ~UPLL_BYPASS_EN_MASK);
+
+	if (rdev->family >= CHIP_RS780)
+		WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~UPLL_BYPASS_CNTL);
+
+	r = radeon_uvd_send_upll_ctlreq(rdev, CG_UPLL_FUNC_CNTL);
+	if (r)
+		return r;
+
+	/* switch VCLK and DCLK selection */
+	WREG32_P(CG_UPLL_FUNC_CNTL_2,
+		 VCLK_SRC_SEL(2) | DCLK_SRC_SEL(2),
+		 ~(VCLK_SRC_SEL_MASK | DCLK_SRC_SEL_MASK));
+
+	mdelay(100);
+
 	return 0;
 }
 
@@ -968,7 +1056,6 @@ static int r600_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
@@ -993,6 +1080,8 @@ static int r600_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(MC_VM_L1_TLB_MCB_WR_GFX_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_PDMA_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_WR_PDMA_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_RD_UVD_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_WR_UVD_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_SEM_CNTL, tmp | ENABLE_SEMAPHORE_MODE);
 	WREG32(MC_VM_L1_TLB_MCB_WR_SEM_CNTL, tmp | ENABLE_SEMAPHORE_MODE);
 	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
@@ -1043,6 +1132,8 @@ static void r600_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(MC_VM_L1_TLB_MCB_WR_SYS_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_HDP_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_WR_HDP_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_RD_UVD_CNTL, tmp);
+	WREG32(MC_VM_L1_TLB_MCB_WR_UVD_CNTL, tmp);
 	radeon_gart_table_vram_unpin(rdev);
 }
 
@@ -1339,7 +1430,7 @@ int r600_vram_scratch_init(struct radeon_device *rdev)
 	if (rdev->vram_scratch.robj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     NULL, &rdev->vram_scratch.robj);
+				     0, NULL, NULL, &rdev->vram_scratch.robj);
 		if (r) {
 			return r;
 		}
@@ -1813,7 +1904,6 @@ static void r600_gpu_init(struct radeon_device *rdev)
 {
 	u32 tiling_config;
 	u32 ramcfg;
-	u32 cc_rb_backend_disable;
 	u32 cc_gc_shader_pipe_config;
 	u32 tmp;
 	int i, j;
@@ -1940,29 +2030,20 @@ static void r600_gpu_init(struct radeon_device *rdev)
 	}
 	tiling_config |= BANK_SWAPS(1);
 
-	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000;
-	tmp = R6XX_MAX_BACKENDS -
-		r600_count_pipe_bits((cc_rb_backend_disable >> 16) & R6XX_MAX_BACKENDS_MASK);
-	if (tmp < rdev->config.r600.max_backends) {
-		rdev->config.r600.max_backends = tmp;
-	}
-
 	cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0x00ffff00;
-	tmp = R6XX_MAX_PIPES -
-		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R6XX_MAX_PIPES_MASK);
-	if (tmp < rdev->config.r600.max_pipes) {
-		rdev->config.r600.max_pipes = tmp;
-	}
-	tmp = R6XX_MAX_SIMDS -
-		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK);
-	if (tmp < rdev->config.r600.max_simds) {
-		rdev->config.r600.max_simds = tmp;
-	}
 	tmp = rdev->config.r600.max_simds -
 		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R6XX_MAX_SIMDS_MASK);
 	rdev->config.r600.active_simds = tmp;
 
 	disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R6XX_MAX_BACKENDS_MASK;
+	tmp = 0;
+	for (i = 0; i < rdev->config.r600.max_backends; i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < rdev->config.r600.max_backends; i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 	tmp = (tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT;
 	tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.r600.max_backends,
 					R6XX_MAX_BACKENDS, disabled_rb_mask);
@@ -2548,7 +2629,7 @@ int r600_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
 	radeon_ring_write(ring, 0);
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	cp_me = 0xff;
 	WREG32(R_0086D8_CP_ME_CNTL, cp_me);
@@ -2684,7 +2765,7 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
 	radeon_ring_write(ring, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF)
@@ -2754,6 +2835,17 @@ void r600_fence_ring_emit(struct radeon_device *rdev,
 	}
 }
 
+/**
+ * r600_semaphore_ring_emit - emit a semaphore on the CP ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon ring buffer object
+ * @semaphore: radeon semaphore object
+ * @emit_wait: Is this a sempahore wait?
+ *
+ * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP
+ * from running ahead of semaphore waits.
+ */
 bool r600_semaphore_ring_emit(struct radeon_device *rdev,
 			      struct radeon_ring *ring,
 			      struct radeon_semaphore *semaphore,
@@ -2769,6 +2861,13 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, lower_32_bits(addr));
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel);
 
+	/* PFP_SYNC_ME packet only exists on 7xx+, only enable it on eg+ */
+	if (emit_wait && (rdev->family >= CHIP_CEDAR)) {
+		/* Prevent the PFP from running ahead of the semaphore wait */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+
 	return true;
 }
 
@@ -2785,12 +2884,13 @@ bool r600_semaphore_ring_emit(struct radeon_device *rdev,
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_cpdma(struct radeon_device *rdev,
-		    uint64_t src_offset, uint64_t dst_offset,
-		    unsigned num_gpu_pages,
-		    struct radeon_fence **fence)
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+				     uint64_t src_offset, uint64_t dst_offset,
+				     unsigned num_gpu_pages,
+				     struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.blit_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes, tmp;
@@ -2800,7 +2900,7 @@ int r600_copy_cpdma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -2809,10 +2909,10 @@ int r600_copy_cpdma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
@@ -2839,17 +2939,17 @@ int r600_copy_cpdma(struct radeon_device *rdev,
 	radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	radeon_ring_write(ring, WAIT_CP_DMA_IDLE_bit);
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
 
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
@@ -2900,6 +3000,18 @@ static int r600_startup(struct radeon_device *rdev)
 		return r;
 	}
 
+	if (rdev->has_uvd) {
+		r = uvd_v1_0_resume(rdev);
+		if (!r) {
+			r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX);
+			if (r) {
+				dev_err(rdev->dev, "failed initializing UVD fences (%d).\n", r);
+			}
+		}
+		if (r)
+			rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+	}
+
 	/* Enable IRQ */
 	if (!rdev->irq.installed) {
 		r = radeon_irq_kms_init(rdev);
@@ -2928,6 +3040,18 @@ static int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
+	if (rdev->has_uvd) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		if (ring->ring_size) {
+			r = radeon_ring_init(rdev, ring, ring->ring_size, 0,
+					     RADEON_CP_PACKET2);
+			if (!r)
+				r = uvd_v1_0_init(rdev);
+			if (r)
+				DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+		}
+	}
+
 	r = radeon_ib_pool_init(rdev);
 	if (r) {
 		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
@@ -2987,6 +3111,10 @@ int r600_suspend(struct radeon_device *rdev)
 	radeon_pm_suspend(rdev);
 	r600_audio_fini(rdev);
 	r600_cp_stop(rdev);
+	if (rdev->has_uvd) {
+		uvd_v1_0_fini(rdev);
+		radeon_uvd_suspend(rdev);
+	}
 	r600_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	r600_pcie_gart_disable(rdev);
@@ -3066,6 +3194,14 @@ int r600_init(struct radeon_device *rdev)
 	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL;
 	r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024);
 
+	if (rdev->has_uvd) {
+		r = radeon_uvd_init(rdev);
+		if (!r) {
+			rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_obj = NULL;
+			r600_ring_init(rdev, &rdev->ring[R600_RING_TYPE_UVD_INDEX], 4096);
+		}
+	}
+
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
 
@@ -3095,6 +3231,10 @@ void r600_fini(struct radeon_device *rdev)
 	r600_audio_fini(rdev);
 	r600_cp_fini(rdev);
 	r600_irq_fini(rdev);
+	if (rdev->has_uvd) {
+		uvd_v1_0_fini(rdev);
+		radeon_uvd_fini(rdev);
+	}
 	radeon_wb_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
@@ -3166,7 +3306,7 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	ib.ptr[2] = 0xDEADBEEF;
 	ib.length_dw = 3;
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 		goto free_ib;
@@ -3227,8 +3367,8 @@ int r600_ih_ring_alloc(struct radeon_device *rdev)
 	if (rdev->ih.ring_obj == NULL) {
 		r = radeon_bo_create(rdev, rdev->ih.ring_size,
 				     PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT,
-				     NULL, &rdev->ih.ring_obj);
+				     RADEON_GEM_DOMAIN_GTT, 0,
+				     NULL, NULL, &rdev->ih.ring_obj);
 		if (r) {
 			DRM_ERROR("radeon: failed to create ih ring buffer (%d).\n", r);
 			return r;
@@ -3785,17 +3925,17 @@ static u32 r600_get_ih_wptr(struct radeon_device *rdev)
 		wptr = RREG32(IH_RB_WPTR);
 
 	if (wptr & RB_OVERFLOW) {
+		wptr &= ~RB_OVERFLOW;
 		/* When a ring buffer overflow happen start parsing interrupt
 		 * from the last not overwritten vector (wptr + 16). Hopefully
 		 * this should allow us to catchup.
 		 */
-		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
-			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			 wptr, rdev->ih.rptr, (wptr + 16) & rdev->ih.ptr_mask);
 		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
-		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -3924,11 +4064,13 @@ restart_ih:
 			break;
 		case 9: /* D1 pflip */
 			DRM_DEBUG("IH: D1 flip\n");
-			radeon_crtc_handle_flip(rdev, 0);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, 0);
 			break;
 		case 11: /* D2 pflip */
 			DRM_DEBUG("IH: D2 flip\n");
-			radeon_crtc_handle_flip(rdev, 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, 1);
 			break;
 		case 19: /* HPD/DAC hotplug */
 			switch (src_data) {
@@ -4039,6 +4181,7 @@ restart_ih:
 		/* wptr/rptr are in bytes! */
 		rptr += 16;
 		rptr &= rdev->ih.ptr_mask;
+		WREG32(IH_RB_RPTR, rptr);
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
@@ -4047,7 +4190,6 @@ restart_ih:
 	if (queue_thermal && rdev->pm.dpm_enabled)
 		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
-	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
 
 	/* make sure wptr hasn't changed while processing */
@@ -4089,16 +4231,15 @@ int r600_debugfs_mc_info_init(struct radeon_device *rdev)
 }
 
 /**
- * r600_ioctl_wait_idle - flush host path cache on wait idle ioctl
+ * r600_mmio_hdp_flush - flush Host Data Path cache via MMIO
  * rdev: radeon device structure
- * bo: buffer object struct which userspace is waiting for idle
  *
- * Some R6XX/R7XX doesn't seems to take into account HDP flush performed
- * through ring buffer, this leads to corruption in rendering, see
- * http://bugzilla.kernel.org/show_bug.cgi?id=15186 to avoid this we
- * directly perform HDP flush by writing register through MMIO.
+ * Some R6XX/R7XX don't seem to take into account HDP flushes performed
+ * through the ring buffer. This leads to corruption in rendering, see
+ * http://bugzilla.kernel.org/show_bug.cgi?id=15186 . To avoid this, we
+ * directly perform the HDP flush by writing the register through MMIO.
  */
-void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo)
+void r600_mmio_hdp_flush(struct radeon_device *rdev)
 {
 	/* r7xx hw bug.  write to HDP_DEBUG1 followed by fb read
 	 * rather than write to HDP_REG_COHERENCY_FLUSH_CNTL.
diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c
deleted file mode 100644
index bffac10c4296..000000000000
--- a/drivers/gpu/drm/radeon/r600_audio.c
+++ /dev/null
@@ -1,207 +0,0 @@
-/*
- * Copyright 2008 Advanced Micro Devices, Inc.
- * Copyright 2008 Red Hat Inc.
- * Copyright 2009 Christian König.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Christian König
- */
-#include <drm/drmP.h>
-#include "radeon.h"
-#include "radeon_reg.h"
-#include "radeon_asic.h"
-#include "atom.h"
-
-/*
- * check if enc_priv stores radeon_encoder_atom_dig
- */
-static bool radeon_dig_encoder(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		return true;
-	}
-	return false;
-}
-
-/*
- * check if the chipset is supported
- */
-static int r600_audio_chipset_supported(struct radeon_device *rdev)
-{
-	return ASIC_IS_DCE2(rdev) && !ASIC_IS_NODCE(rdev);
-}
-
-struct r600_audio_pin r600_audio_status(struct radeon_device *rdev)
-{
-	struct r600_audio_pin status;
-	uint32_t value;
-
-	value = RREG32(R600_AUDIO_RATE_BPS_CHANNEL);
-
-	/* number of channels */
-	status.channels = (value & 0x7) + 1;
-
-	/* bits per sample */
-	switch ((value & 0xF0) >> 4) {
-	case 0x0:
-		status.bits_per_sample = 8;
-		break;
-	case 0x1:
-		status.bits_per_sample = 16;
-		break;
-	case 0x2:
-		status.bits_per_sample = 20;
-		break;
-	case 0x3:
-		status.bits_per_sample = 24;
-		break;
-	case 0x4:
-		status.bits_per_sample = 32;
-		break;
-	default:
-		dev_err(rdev->dev, "Unknown bits per sample 0x%x, using 16\n",
-			(int)value);
-		status.bits_per_sample = 16;
-	}
-
-	/* current sampling rate in HZ */
-	if (value & 0x4000)
-		status.rate = 44100;
-	else
-		status.rate = 48000;
-	status.rate *= ((value >> 11) & 0x7) + 1;
-	status.rate /= ((value >> 8) & 0x7) + 1;
-
-	value = RREG32(R600_AUDIO_STATUS_BITS);
-
-	/* iec 60958 status bits */
-	status.status_bits = value & 0xff;
-
-	/* iec 60958 category code */
-	status.category_code = (value >> 8) & 0xff;
-
-	return status;
-}
-
-/*
- * update all hdmi interfaces with current audio parameters
- */
-void r600_audio_update_hdmi(struct work_struct *work)
-{
-	struct radeon_device *rdev = container_of(work, struct radeon_device,
-						  audio_work);
-	struct drm_device *dev = rdev->ddev;
-	struct r600_audio_pin audio_status = r600_audio_status(rdev);
-	struct drm_encoder *encoder;
-	bool changed = false;
-
-	if (rdev->audio.pin[0].channels != audio_status.channels ||
-	    rdev->audio.pin[0].rate != audio_status.rate ||
-	    rdev->audio.pin[0].bits_per_sample != audio_status.bits_per_sample ||
-	    rdev->audio.pin[0].status_bits != audio_status.status_bits ||
-	    rdev->audio.pin[0].category_code != audio_status.category_code) {
-		rdev->audio.pin[0] = audio_status;
-		changed = true;
-	}
-
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		if (!radeon_dig_encoder(encoder))
-			continue;
-		if (changed || r600_hdmi_buffer_status_changed(encoder))
-			r600_hdmi_update_audio_settings(encoder);
-	}
-}
-
-/* enable the audio stream */
-void r600_audio_enable(struct radeon_device *rdev,
-		       struct r600_audio_pin *pin,
-		       bool enable)
-{
-	u32 value = 0;
-
-	if (!pin)
-		return;
-
-	if (ASIC_IS_DCE4(rdev)) {
-		if (enable) {
-			value |= 0x81000000; /* Required to enable audio */
-			value |= 0x0e1000f0; /* fglrx sets that too */
-		}
-		WREG32(EVERGREEN_AUDIO_ENABLE, value);
-	} else {
-		WREG32_P(R600_AUDIO_ENABLE,
-			 enable ? 0x81000000 : 0x0, ~0x81000000);
-	}
-}
-
-/*
- * initialize the audio vars
- */
-int r600_audio_init(struct radeon_device *rdev)
-{
-	if (!radeon_audio || !r600_audio_chipset_supported(rdev))
-		return 0;
-
-	rdev->audio.enabled = true;
-
-	rdev->audio.num_pins = 1;
-	rdev->audio.pin[0].channels = -1;
-	rdev->audio.pin[0].rate = -1;
-	rdev->audio.pin[0].bits_per_sample = -1;
-	rdev->audio.pin[0].status_bits = 0;
-	rdev->audio.pin[0].category_code = 0;
-	rdev->audio.pin[0].id = 0;
-	/* disable audio.  it will be set up later */
-	r600_audio_enable(rdev, &rdev->audio.pin[0], false);
-
-	return 0;
-}
-
-/*
- * release the audio timer
- * TODO: How to do this correctly on SMP systems?
- */
-void r600_audio_fini(struct radeon_device *rdev)
-{
-	if (!rdev->audio.enabled)
-		return;
-
-	r600_audio_enable(rdev, &rdev->audio.pin[0], false);
-
-	rdev->audio.enabled = false;
-}
-
-struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev)
-{
-	/* only one pin on 6xx-NI */
-	return &rdev->audio.pin[0];
-}
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index 8c9b7e26533c..09e3f39925fa 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -1949,15 +1949,15 @@ int r600_do_cleanup_cp(struct drm_device *dev)
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
 		if (dev_priv->cp_ring != NULL) {
-			drm_core_ioremapfree(dev_priv->cp_ring, dev);
+			drm_legacy_ioremapfree(dev_priv->cp_ring, dev);
 			dev_priv->cp_ring = NULL;
 		}
 		if (dev_priv->ring_rptr != NULL) {
-			drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+			drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			dev_priv->ring_rptr = NULL;
 		}
 		if (dev->agp_buffer_map != NULL) {
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 			dev->agp_buffer_map = NULL;
 		}
 	} else
@@ -1968,7 +1968,7 @@ int r600_do_cleanup_cp(struct drm_device *dev)
 			r600_page_table_cleanup(dev, &dev_priv->gart_info);
 
 		if (dev_priv->gart_info.gart_table_location == DRM_ATI_GART_FB) {
-			drm_core_ioremapfree(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremapfree(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr = NULL;
 		}
 	}
@@ -2052,27 +2052,27 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	dev_priv->buffers_offset = init->buffers_offset;
 	dev_priv->gart_textures_offset = init->gart_textures_offset;
 
-	master_priv->sarea = drm_getsarea(dev);
+	master_priv->sarea = drm_legacy_getsarea(dev);
 	if (!master_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 
-	dev_priv->cp_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cp_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cp_ring) {
 		DRM_ERROR("could not find cp ring region!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		r600_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		r600_do_cleanup_cp(dev);
@@ -2081,7 +2081,7 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 	if (init->gart_textures_offset) {
 		dev_priv->gart_textures =
-		    drm_core_findmap(dev, init->gart_textures_offset);
+		    drm_legacy_findmap(dev, init->gart_textures_offset);
 		if (!dev_priv->gart_textures) {
 			DRM_ERROR("could not find GART texture region!\n");
 			r600_do_cleanup_cp(dev);
@@ -2092,9 +2092,9 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 #if __OS_HAS_AGP
 	/* XXX */
 	if (dev_priv->flags & RADEON_IS_AGP) {
-		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cp_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cp_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -2235,7 +2235,7 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 		dev_priv->gart_info.mapping.size =
 			dev_priv->gart_info.table_size;
 
-		drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
+		drm_legacy_ioremap_wc(&dev_priv->gart_info.mapping, dev);
 		if (!dev_priv->gart_info.mapping.handle) {
 			DRM_ERROR("ioremap failed.\n");
 			r600_do_cleanup_cp(dev);
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 12511bb5fd6f..c47537a1ddba 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -825,7 +825,6 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 			       uint32_t *vline_start_end,
 			       uint32_t *vline_status)
 {
-	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	struct radeon_crtc *radeon_crtc;
 	struct radeon_cs_packet p3reloc, wait_reg_mem;
@@ -887,12 +886,11 @@ int r600_cs_common_vline_parse(struct radeon_cs_parser *p,
 	crtc_id = radeon_get_ib_value(p, h_idx + 2 + 7 + 1);
 	reg = R600_CP_PACKET0_GET_REG(header);
 
-	obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
+	crtc = drm_crtc_find(p->rdev->ddev, crtc_id);
+	if (!crtc) {
 		DRM_ERROR("cannot find crtc %d\n", crtc_id);
 		return -ENOENT;
 	}
-	crtc = obj_to_crtc(obj);
 	radeon_crtc = to_radeon_crtc(crtc);
 	crtc_id = radeon_crtc->crtc_id;
 
diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c
index 4969cef44a19..cf0df45d455e 100644
--- a/drivers/gpu/drm/radeon/r600_dma.c
+++ b/drivers/gpu/drm/radeon/r600_dma.c
@@ -124,15 +124,6 @@ int r600_dma_resume(struct radeon_device *rdev)
 	u32 rb_bufsz;
 	int r;
 
-	/* Reset dma */
-	if (rdev->family >= CHIP_RV770)
-		WREG32(SRBM_SOFT_RESET, RV770_SOFT_RESET_DMA);
-	else
-		WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA);
-	RREG32(SRBM_SOFT_RESET);
-	udelay(50);
-	WREG32(SRBM_SOFT_RESET, 0);
-
 	WREG32(DMA_SEM_INCOMPLETE_TIMER_CNTL, 0);
 	WREG32(DMA_SEM_WAIT_FAIL_TIMER_CNTL, 0);
 
@@ -241,16 +232,19 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 {
 	unsigned i;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	unsigned index;
 	u32 tmp;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
+
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	rdev->wb.wb[index/4] = cpu_to_le32(tmp);
 
 	r = radeon_ring_lock(rdev, ring, 4);
 	if (r) {
@@ -258,13 +252,13 @@ int r600_dma_ring_test(struct radeon_device *rdev,
 		return r;
 	}
 	radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1));
-	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
-	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff);
+	radeon_ring_write(ring, lower_32_bits(gpu_addr));
+	radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xff);
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -344,17 +338,17 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	struct radeon_ib ib;
 	unsigned i;
+	unsigned index;
 	int r;
-	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
 	u32 tmp = 0;
+	u64 gpu_addr;
 
-	if (!ptr) {
-		DRM_ERROR("invalid vram scratch pointer\n");
-		return -EINVAL;
-	}
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		index = R600_WB_DMA_RING_TEST_OFFSET;
+	else
+		index = CAYMAN_WB_DMA1_RING_TEST_OFFSET;
 
-	tmp = 0xCAFEDEAD;
-	writel(tmp, ptr);
+	gpu_addr = rdev->wb.gpu_addr + index;
 
 	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
 	if (r) {
@@ -363,12 +357,12 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 
 	ib.ptr[0] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 1);
-	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
-	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xff;
+	ib.ptr[1] = lower_32_bits(gpu_addr);
+	ib.ptr[2] = upper_32_bits(gpu_addr) & 0xff;
 	ib.ptr[3] = 0xDEADBEEF;
 	ib.length_dw = 4;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
@@ -380,7 +374,7 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	for (i = 0; i < rdev->usec_timeout; i++) {
-		tmp = readl(ptr);
+		tmp = le32_to_cpu(rdev->wb.wb[index/4]);
 		if (tmp == 0xDEADBEEF)
 			break;
 		DRM_UDELAY(1);
@@ -436,18 +430,19 @@ void r600_dma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r6xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int r600_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset, uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
@@ -457,7 +452,7 @@ int r600_copy_dma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -466,10 +461,10 @@ int r600_copy_dma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -486,15 +481,15 @@ int r600_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c
index 9c61b74ef441..f6309bd23e01 100644
--- a/drivers/gpu/drm/radeon/r600_dpm.c
+++ b/drivers/gpu/drm/radeon/r600_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "r600d.h"
 #include "r600_dpm.h"
 #include "atom.h"
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 26ef8ced6f89..b90dc0eb08e6 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -72,6 +72,169 @@ static const struct radeon_hdmi_acr r600_hdmi_predefined_acr[] = {
 
 
 /*
+ * check if the chipset is supported
+ */
+static int r600_audio_chipset_supported(struct radeon_device *rdev)
+{
+	return ASIC_IS_DCE2(rdev) && !ASIC_IS_NODCE(rdev);
+}
+
+static struct r600_audio_pin r600_audio_status(struct radeon_device *rdev)
+{
+	struct r600_audio_pin status;
+	uint32_t value;
+
+	value = RREG32(R600_AUDIO_RATE_BPS_CHANNEL);
+
+	/* number of channels */
+	status.channels = (value & 0x7) + 1;
+
+	/* bits per sample */
+	switch ((value & 0xF0) >> 4) {
+	case 0x0:
+		status.bits_per_sample = 8;
+		break;
+	case 0x1:
+		status.bits_per_sample = 16;
+		break;
+	case 0x2:
+		status.bits_per_sample = 20;
+		break;
+	case 0x3:
+		status.bits_per_sample = 24;
+		break;
+	case 0x4:
+		status.bits_per_sample = 32;
+		break;
+	default:
+		dev_err(rdev->dev, "Unknown bits per sample 0x%x, using 16\n",
+			(int)value);
+		status.bits_per_sample = 16;
+	}
+
+	/* current sampling rate in HZ */
+	if (value & 0x4000)
+		status.rate = 44100;
+	else
+		status.rate = 48000;
+	status.rate *= ((value >> 11) & 0x7) + 1;
+	status.rate /= ((value >> 8) & 0x7) + 1;
+
+	value = RREG32(R600_AUDIO_STATUS_BITS);
+
+	/* iec 60958 status bits */
+	status.status_bits = value & 0xff;
+
+	/* iec 60958 category code */
+	status.category_code = (value >> 8) & 0xff;
+
+	return status;
+}
+
+/*
+ * update all hdmi interfaces with current audio parameters
+ */
+void r600_audio_update_hdmi(struct work_struct *work)
+{
+	struct radeon_device *rdev = container_of(work, struct radeon_device,
+						  audio_work);
+	struct drm_device *dev = rdev->ddev;
+	struct r600_audio_pin audio_status = r600_audio_status(rdev);
+	struct drm_encoder *encoder;
+	bool changed = false;
+
+	if (rdev->audio.pin[0].channels != audio_status.channels ||
+	    rdev->audio.pin[0].rate != audio_status.rate ||
+	    rdev->audio.pin[0].bits_per_sample != audio_status.bits_per_sample ||
+	    rdev->audio.pin[0].status_bits != audio_status.status_bits ||
+	    rdev->audio.pin[0].category_code != audio_status.category_code) {
+		rdev->audio.pin[0] = audio_status;
+		changed = true;
+	}
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (!radeon_encoder_is_digital(encoder))
+			continue;
+		if (changed || r600_hdmi_buffer_status_changed(encoder))
+			r600_hdmi_update_audio_settings(encoder);
+	}
+}
+
+/* enable the audio stream */
+void r600_audio_enable(struct radeon_device *rdev,
+		       struct r600_audio_pin *pin,
+		       u8 enable_mask)
+{
+	u32 tmp = RREG32(AZ_HOT_PLUG_CONTROL);
+
+	if (!pin)
+		return;
+
+	if (enable_mask) {
+		tmp |= AUDIO_ENABLED;
+		if (enable_mask & 1)
+			tmp |= PIN0_AUDIO_ENABLED;
+		if (enable_mask & 2)
+			tmp |= PIN1_AUDIO_ENABLED;
+		if (enable_mask & 4)
+			tmp |= PIN2_AUDIO_ENABLED;
+		if (enable_mask & 8)
+			tmp |= PIN3_AUDIO_ENABLED;
+	} else {
+		tmp &= ~(AUDIO_ENABLED |
+			 PIN0_AUDIO_ENABLED |
+			 PIN1_AUDIO_ENABLED |
+			 PIN2_AUDIO_ENABLED |
+			 PIN3_AUDIO_ENABLED);
+	}
+
+	WREG32(AZ_HOT_PLUG_CONTROL, tmp);
+}
+
+/*
+ * initialize the audio vars
+ */
+int r600_audio_init(struct radeon_device *rdev)
+{
+	if (!radeon_audio || !r600_audio_chipset_supported(rdev))
+		return 0;
+
+	rdev->audio.enabled = true;
+
+	rdev->audio.num_pins = 1;
+	rdev->audio.pin[0].channels = -1;
+	rdev->audio.pin[0].rate = -1;
+	rdev->audio.pin[0].bits_per_sample = -1;
+	rdev->audio.pin[0].status_bits = 0;
+	rdev->audio.pin[0].category_code = 0;
+	rdev->audio.pin[0].id = 0;
+	/* disable audio.  it will be set up later */
+	r600_audio_enable(rdev, &rdev->audio.pin[0], 0);
+
+	return 0;
+}
+
+/*
+ * release the audio timer
+ * TODO: How to do this correctly on SMP systems?
+ */
+void r600_audio_fini(struct radeon_device *rdev)
+{
+	if (!rdev->audio.enabled)
+		return;
+
+	r600_audio_enable(rdev, &rdev->audio.pin[0], 0);
+
+	rdev->audio.enabled = false;
+}
+
+struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev)
+{
+	/* only one pin on 6xx-NI */
+	return &rdev->audio.pin[0];
+}
+
+/*
  * calculate CTS and N values if they are not found in the table
  */
 static void r600_hdmi_calc_cts(uint32_t clock, int *CTS, int *N, int freq)
@@ -357,7 +520,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 
 	/* disable audio prior to setting up hw */
 	dig->afmt->pin = r600_audio_get_pin(rdev);
-	r600_audio_enable(rdev, dig->afmt->pin, false);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 
 	r600_audio_set_dto(encoder, mode->clock);
 
@@ -443,7 +606,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 	WREG32(HDMI0_RAMP_CONTROL3 + offset, 0x00000001);
 
 	/* enable audio after to setting up hw */
-	r600_audio_enable(rdev, dig->afmt->pin, true);
+	r600_audio_enable(rdev, dig->afmt->pin, 0xf);
 }
 
 /**
@@ -528,6 +691,11 @@ void r600_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	if (!enable && !dig->afmt->enabled)
 		return;
 
+	if (!enable && dig->afmt->pin) {
+		r600_audio_enable(rdev, dig->afmt->pin, 0);
+		dig->afmt->pin = NULL;
+	}
+
 	/* Older chipsets require setting HDMI and routing manually */
 	if (!ASIC_IS_DCE3(rdev)) {
 		if (enable)
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index f94e7a9afe75..1e8495cca41e 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -44,13 +44,6 @@
 #define R6XX_MAX_PIPES				8
 #define R6XX_MAX_PIPES_MASK			0xff
 
-/* PTE flags */
-#define PTE_VALID				(1 << 0)
-#define PTE_SYSTEM				(1 << 1)
-#define PTE_SNOOPED				(1 << 2)
-#define PTE_READABLE				(1 << 5)
-#define PTE_WRITEABLE				(1 << 6)
-
 /* tiling bits */
 #define     ARRAY_LINEAR_GENERAL              0x00000000
 #define     ARRAY_LINEAR_ALIGNED              0x00000001
@@ -330,11 +323,12 @@
 #define	HDP_TILING_CONFIG				0x2F3C
 #define HDP_DEBUG1                                      0x2F34
 
+#define MC_CONFIG					0x2000
 #define MC_VM_AGP_TOP					0x2184
 #define MC_VM_AGP_BOT					0x2188
 #define	MC_VM_AGP_BASE					0x218C
 #define MC_VM_FB_LOCATION				0x2180
-#define MC_VM_L1_TLB_MCD_RD_A_CNTL			0x219C
+#define MC_VM_L1_TLB_MCB_RD_UVD_CNTL			0x2124
 #define 	ENABLE_L1_TLB					(1 << 0)
 #define		ENABLE_L1_FRAGMENT_PROCESSING			(1 << 1)
 #define		ENABLE_L1_STRICT_ORDERING			(1 << 2)
@@ -354,12 +348,14 @@
 #define		EFFECTIVE_L1_QUEUE_SIZE(x)			(((x) & 7) << 15)
 #define		EFFECTIVE_L1_QUEUE_SIZE_MASK			0x00038000
 #define		EFFECTIVE_L1_QUEUE_SIZE_SHIFT			15
+#define MC_VM_L1_TLB_MCD_RD_A_CNTL			0x219C
 #define MC_VM_L1_TLB_MCD_RD_B_CNTL			0x21A0
 #define MC_VM_L1_TLB_MCB_RD_GFX_CNTL			0x21FC
 #define MC_VM_L1_TLB_MCB_RD_HDP_CNTL			0x2204
 #define MC_VM_L1_TLB_MCB_RD_PDMA_CNTL			0x2208
 #define MC_VM_L1_TLB_MCB_RD_SEM_CNTL			0x220C
 #define	MC_VM_L1_TLB_MCB_RD_SYS_CNTL			0x2200
+#define MC_VM_L1_TLB_MCB_WR_UVD_CNTL			0x212c
 #define MC_VM_L1_TLB_MCD_WR_A_CNTL			0x21A4
 #define MC_VM_L1_TLB_MCD_WR_B_CNTL			0x21A8
 #define MC_VM_L1_TLB_MCB_WR_GFX_CNTL			0x2210
@@ -373,6 +369,8 @@
 #define MC_VM_SYSTEM_APERTURE_HIGH_ADDR			0x2194
 #define MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR		0x2198
 
+#define RS_DQ_RD_RET_CONF				0x2348
+
 #define	PA_CL_ENHANCE					0x8A14
 #define		CLIP_VTX_REORDER_ENA				(1 << 0)
 #define		NUM_CLIP_SEQ(x)					((x) << 1)
@@ -929,6 +927,23 @@
 #       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
 #       define SELECTABLE_DEEMPHASIS                      (1 << 6)
 
+/* Audio */
+#define AZ_HOT_PLUG_CONTROL               0x7300
+#       define AZ_FORCE_CODEC_WAKE        (1 << 0)
+#       define JACK_DETECTION_ENABLE      (1 << 4)
+#       define UNSOLICITED_RESPONSE_ENABLE (1 << 8)
+#       define CODEC_HOT_PLUG_ENABLE      (1 << 12)
+#       define AUDIO_ENABLED              (1 << 31)
+/* DCE3 adds */
+#       define PIN0_JACK_DETECTION_ENABLE (1 << 4)
+#       define PIN1_JACK_DETECTION_ENABLE (1 << 5)
+#       define PIN2_JACK_DETECTION_ENABLE (1 << 6)
+#       define PIN3_JACK_DETECTION_ENABLE (1 << 7)
+#       define PIN0_AUDIO_ENABLED         (1 << 24)
+#       define PIN1_AUDIO_ENABLED         (1 << 25)
+#       define PIN2_AUDIO_ENABLED         (1 << 26)
+#       define PIN3_AUDIO_ENABLED         (1 << 27)
+
 /* Audio clocks DCE 2.0/3.0 */
 #define AUDIO_DTO                         0x7340
 #       define AUDIO_DTO_PHASE(x)         (((x) & 0xffff) << 0)
@@ -1483,6 +1498,7 @@
 #define UVD_CGC_GATE					0xf4a8
 #define UVD_LMI_CTRL2					0xf4f4
 #define UVD_MASTINT_EN					0xf500
+#define UVD_FW_START					0xf51C
 #define UVD_LMI_ADDR_EXT				0xf594
 #define UVD_LMI_CTRL					0xf598
 #define UVD_LMI_SWAP_CNTL				0xf5b4
@@ -1495,6 +1511,13 @@
 #define UVD_MPC_SET_MUX					0xf5f4
 #define UVD_MPC_SET_ALU					0xf5f8
 
+#define UVD_VCPU_CACHE_OFFSET0				0xf608
+#define UVD_VCPU_CACHE_SIZE0				0xf60c
+#define UVD_VCPU_CACHE_OFFSET1				0xf610
+#define UVD_VCPU_CACHE_SIZE1				0xf614
+#define UVD_VCPU_CACHE_OFFSET2				0xf618
+#define UVD_VCPU_CACHE_SIZE2				0xf61c
+
 #define UVD_VCPU_CNTL					0xf660
 #define UVD_SOFT_RESET					0xf680
 #define		RBC_SOFT_RESET					(1<<0)
@@ -1524,9 +1547,35 @@
 
 #define UVD_CONTEXT_ID					0xf6f4
 
+/* rs780 only */
+#define	GFX_MACRO_BYPASS_CNTL				0x30c0
+#define		SPLL_BYPASS_CNTL			(1 << 0)
+#define		UPLL_BYPASS_CNTL			(1 << 1)
+
+#define CG_UPLL_FUNC_CNTL				0x7e0
+#	define UPLL_RESET_MASK				0x00000001
+#	define UPLL_SLEEP_MASK				0x00000002
+#	define UPLL_BYPASS_EN_MASK			0x00000004
 #	define UPLL_CTLREQ_MASK				0x00000008
+#	define UPLL_FB_DIV(x)				((x) << 4)
+#	define UPLL_FB_DIV_MASK				0x0000FFF0
+#	define UPLL_REF_DIV(x)				((x) << 16)
+#	define UPLL_REF_DIV_MASK			0x003F0000
+#	define UPLL_REFCLK_SRC_SEL_MASK			0x20000000
 #	define UPLL_CTLACK_MASK				0x40000000
 #	define UPLL_CTLACK2_MASK			0x80000000
+#define CG_UPLL_FUNC_CNTL_2				0x7e4
+#	define UPLL_SW_HILEN(x)				((x) << 0)
+#	define UPLL_SW_LOLEN(x)				((x) << 4)
+#	define UPLL_SW_HILEN2(x)			((x) << 8)
+#	define UPLL_SW_LOLEN2(x)			((x) << 12)
+#	define UPLL_DIVEN_MASK				0x00010000
+#	define UPLL_DIVEN2_MASK				0x00020000
+#	define UPLL_SW_MASK				0x0003FFFF
+#	define VCLK_SRC_SEL(x)				((x) << 20)
+#	define VCLK_SRC_SEL_MASK			0x01F00000
+#	define DCLK_SRC_SEL(x)				((x) << 25)
+#	define DCLK_SRC_SEL_MASK			0x3E000000
 
 /*
  * PM4
@@ -1597,6 +1646,7 @@
 		 */
 #              define PACKET3_CP_DMA_CMD_SAIC      (1 << 28)
 #              define PACKET3_CP_DMA_CMD_DAIC      (1 << 29)
+#define	PACKET3_PFP_SYNC_ME				0x42 /* r7xx+ only */
 #define	PACKET3_SURFACE_SYNC				0x43
 #              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
 #              define PACKET3_FULL_CACHE_ENA       (1 << 20) /* r7xx+ only */
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 60c47f829122..a9717b3fbf1b 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -64,6 +64,9 @@
 #include <linux/wait.h>
 #include <linux/list.h>
 #include <linux/kref.h>
+#include <linux/interval_tree.h>
+#include <linux/hashtable.h>
+#include <linux/fence.h>
 
 #include <ttm/ttm_bo_api.h>
 #include <ttm/ttm_bo_driver.h>
@@ -71,6 +74,8 @@
 #include <ttm/ttm_module.h>
 #include <ttm/ttm_execbuf_util.h>
 
+#include <drm/drm_gem.h>
+
 #include "radeon_family.h"
 #include "radeon_mode.h"
 #include "radeon_reg.h"
@@ -103,6 +108,9 @@ extern int radeon_hard_reset;
 extern int radeon_vm_size;
 extern int radeon_vm_block_size;
 extern int radeon_deep_color;
+extern int radeon_use_pflipirq;
+extern int radeon_bapm;
+extern int radeon_backlight;
 
 /*
  * Copy from radeon_drv.h so we don't have to include both and have conflicting
@@ -116,9 +124,6 @@ extern int radeon_deep_color;
 #define RADEONFB_CONN_LIMIT			4
 #define RADEON_BIOS_NUM_SCRATCH			8
 
-/* fence seq are set to this number when signaled */
-#define RADEON_FENCE_SIGNALED_SEQ		0LL
-
 /* internal ring indices */
 /* r1xx+ has gfx CP ring */
 #define RADEON_RING_TYPE_GFX_INDEX		0
@@ -304,6 +309,9 @@ int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *r
 							 u16 *vddc, u16 *vddci,
 							 u16 virtual_voltage_id,
 							 u16 vbios_voltage_id);
+int radeon_atom_get_voltage_evv(struct radeon_device *rdev,
+				u16 virtual_voltage_id,
+				u16 *voltage);
 int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
 				      u8 voltage_type,
 				      u16 nominal_voltage,
@@ -317,6 +325,9 @@ int radeon_atom_get_voltage_table(struct radeon_device *rdev,
 				  struct atom_voltage_table *voltage_table);
 bool radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
 				 u8 voltage_type, u8 voltage_mode);
+int radeon_atom_get_svi2_info(struct radeon_device *rdev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id);
 void radeon_atom_update_memory_dll(struct radeon_device *rdev,
 				   u32 mem_clock);
 void radeon_atom_set_ac_timing(struct radeon_device *rdev,
@@ -340,28 +351,32 @@ extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
  * Fences.
  */
 struct radeon_fence_driver {
+	struct radeon_device		*rdev;
 	uint32_t			scratch_reg;
 	uint64_t			gpu_addr;
 	volatile uint32_t		*cpu_addr;
 	/* sync_seq is protected by ring emission lock */
 	uint64_t			sync_seq[RADEON_NUM_RINGS];
 	atomic64_t			last_seq;
-	bool				initialized;
+	bool				initialized, delayed_irq;
+	struct delayed_work		lockup_work;
 };
 
 struct radeon_fence {
+	struct fence base;
+
 	struct radeon_device		*rdev;
-	struct kref			kref;
-	/* protected by radeon_fence.lock */
 	uint64_t			seq;
 	/* RB, DMA, etc. */
 	unsigned			ring;
+
+	wait_queue_t			fence_wake;
 };
 
 int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring);
 int radeon_fence_driver_init(struct radeon_device *rdev);
 void radeon_fence_driver_fini(struct radeon_device *rdev);
-void radeon_fence_driver_force_completion(struct radeon_device *rdev);
+void radeon_fence_driver_force_completion(struct radeon_device *rdev, int ring);
 int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, int ring);
 void radeon_fence_process(struct radeon_device *rdev, int ring);
 bool radeon_fence_signaled(struct radeon_fence *fence);
@@ -441,14 +456,12 @@ struct radeon_mman {
 struct radeon_bo_va {
 	/* protected by bo being reserved */
 	struct list_head		bo_list;
-	uint64_t			soffset;
-	uint64_t			eoffset;
 	uint32_t			flags;
-	bool				valid;
+	uint64_t			addr;
 	unsigned			ref_count;
 
 	/* protected by vm mutex */
-	struct list_head		vm_list;
+	struct interval_tree_node	it;
 	struct list_head		vm_status;
 
 	/* constant after initialization */
@@ -461,10 +474,11 @@ struct radeon_bo {
 	struct list_head		list;
 	/* Protected by tbo.reserved */
 	u32				initial_domain;
-	u32				placements[3];
+	struct ttm_place		placements[3];
 	struct ttm_placement		placement;
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
+	u32				flags;
 	unsigned			pin_count;
 	void				*kptr;
 	u32				tiling_flags;
@@ -480,6 +494,9 @@ struct radeon_bo {
 
 	struct ttm_bo_kmap_obj		dma_buf_vmap;
 	pid_t				pid;
+
+	struct radeon_mn		*mn;
+	struct interval_tree_node	mn_it;
 };
 #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base)
 
@@ -543,9 +560,9 @@ struct radeon_gem {
 
 int radeon_gem_init(struct radeon_device *rdev);
 void radeon_gem_fini(struct radeon_device *rdev);
-int radeon_gem_object_create(struct radeon_device *rdev, int size,
+int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
 				int alignment, int initial_domain,
-				bool discardable, bool kernel,
+				u32 flags, bool kernel,
 				struct drm_gem_object **obj);
 
 int radeon_mode_dumb_create(struct drm_file *file_priv,
@@ -571,8 +588,12 @@ bool radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring,
 				  struct radeon_semaphore *semaphore);
 bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring,
 				struct radeon_semaphore *semaphore);
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-			      struct radeon_fence *fence);
+void radeon_semaphore_sync_fence(struct radeon_semaphore *semaphore,
+				 struct radeon_fence *fence);
+int radeon_semaphore_sync_resv(struct radeon_device *rdev,
+			       struct radeon_semaphore *semaphore,
+			       struct reservation_object *resv,
+			       bool shared);
 int radeon_semaphore_sync_rings(struct radeon_device *rdev,
 				struct radeon_semaphore *semaphore,
 				int waiting_ring);
@@ -590,6 +611,12 @@ struct radeon_mc;
 #define RADEON_GPU_PAGE_SHIFT 12
 #define RADEON_GPU_PAGE_ALIGN(a) (((a) + RADEON_GPU_PAGE_MASK) & ~RADEON_GPU_PAGE_MASK)
 
+#define RADEON_GART_PAGE_DUMMY  0
+#define RADEON_GART_PAGE_VALID	(1 << 0)
+#define RADEON_GART_PAGE_READ	(1 << 1)
+#define RADEON_GART_PAGE_WRITE	(1 << 2)
+#define RADEON_GART_PAGE_SNOOP	(1 << 3)
+
 struct radeon_gart {
 	dma_addr_t			table_addr;
 	struct radeon_bo		*robj;
@@ -614,8 +641,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			int pages);
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 		     int pages, struct page **pagelist,
-		     dma_addr_t *dma_addr);
-void radeon_gart_restore(struct radeon_device *rdev);
+		     dma_addr_t *dma_addr, uint32_t flags);
 
 
 /*
@@ -688,7 +714,7 @@ struct radeon_flip_work {
 	uint64_t			base;
 	struct drm_pending_vblank_event *event;
 	struct radeon_bo		*old_rbo;
-	struct radeon_fence		*fence;
+	struct fence			*fence;
 };
 
 struct r500_irq_stat_regs {
@@ -766,6 +792,7 @@ struct radeon_irq {
 int radeon_irq_kms_init(struct radeon_device *rdev);
 void radeon_irq_kms_fini(struct radeon_device *rdev);
 void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring);
+bool radeon_irq_kms_sw_irq_get_delayed(struct radeon_device *rdev, int ring);
 void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring);
 void radeon_irq_kms_pflip_irq_get(struct radeon_device *rdev, int crtc);
 void radeon_irq_kms_pflip_irq_put(struct radeon_device *rdev, int crtc);
@@ -855,9 +882,9 @@ struct radeon_mec {
 #define R600_PTE_FRAG_64KB	(4 << 7)
 #define R600_PTE_FRAG_256KB	(6 << 7)
 
-/* flags used for GART page table entries on R600+ */
-#define R600_PTE_GART	( R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED \
-			| R600_PTE_READABLE | R600_PTE_WRITEABLE)
+/* flags needed to be set so we can copy directly from the GART table */
+#define R600_PTE_GART_MASK	( R600_PTE_READABLE | R600_PTE_WRITEABLE | \
+				  R600_PTE_SYSTEM | R600_PTE_VALID )
 
 struct radeon_vm_pt {
 	struct radeon_bo		*bo;
@@ -865,9 +892,12 @@ struct radeon_vm_pt {
 };
 
 struct radeon_vm {
-	struct list_head		va;
+	struct rb_root			va;
 	unsigned			id;
 
+	/* BOs moved, but not yet updated in the PT */
+	struct list_head		invalidated;
+
 	/* BOs freed, but not yet updated in the PT */
 	struct list_head		freed;
 
@@ -899,6 +929,8 @@ struct radeon_vm_manager {
 	u64				vram_base_offset;
 	/* is vm enabled? */
 	bool				enabled;
+	/* for hw to save the PD addr on suspend/resume */
+	uint32_t			saved_table_addr[RADEON_NUM_VM];
 };
 
 /*
@@ -952,7 +984,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring,
 		  unsigned size);
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib);
 int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
-		       struct radeon_ib *const_ib);
+		       struct radeon_ib *const_ib, bool hdp_flush);
 int radeon_ib_pool_init(struct radeon_device *rdev);
 void radeon_ib_pool_fini(struct radeon_device *rdev);
 int radeon_ib_ring_tests(struct radeon_device *rdev);
@@ -962,8 +994,10 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
 void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp);
 int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
 int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp);
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp);
+void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp,
+			bool hdp_flush);
+void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp,
+			       bool hdp_flush);
 void radeon_ring_undo(struct radeon_ring *ring);
 void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp);
 int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
@@ -1099,6 +1133,8 @@ struct radeon_wb {
 #define R600_WB_EVENT_OFFSET     3072
 #define CIK_WB_CP1_WPTR_OFFSET     3328
 #define CIK_WB_CP2_WPTR_OFFSET     3584
+#define R600_WB_DMA_RING_TEST_OFFSET 3588
+#define CAYMAN_WB_DMA1_RING_TEST_OFFSET 3592
 
 /**
  * struct radeon_pm - power management datas
@@ -1621,7 +1657,8 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 			      uint32_t handle, struct radeon_fence **fence);
 int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 			       uint32_t handle, struct radeon_fence **fence);
-void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo);
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
+				       uint32_t allowed_domains);
 void radeon_uvd_free_handles(struct radeon_device *rdev,
 			     struct drm_file *filp);
 int radeon_uvd_cs_parse(struct radeon_cs_parser *parser);
@@ -1710,6 +1747,11 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 			   struct radeon_ring *cpB);
 void radeon_test_syncing(struct radeon_device *rdev);
 
+/*
+ * MMU Notifier
+ */
+int radeon_mn_register(struct radeon_bo *bo, unsigned long addr);
+void radeon_mn_unregister(struct radeon_bo *bo);
 
 /*
  * Debugfs
@@ -1740,6 +1782,7 @@ struct radeon_asic_ring {
 	/* command emmit functions */
 	void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
 	void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence);
+	void (*hdp_flush)(struct radeon_device *rdev, struct radeon_ring *ring);
 	bool (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
 			       struct radeon_semaphore *semaphore, bool emit_wait);
 	void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
@@ -1763,13 +1806,8 @@ struct radeon_asic {
 	int (*suspend)(struct radeon_device *rdev);
 	void (*vga_set_state)(struct radeon_device *rdev, bool state);
 	int (*asic_reset)(struct radeon_device *rdev);
-	/* ioctl hw specific callback. Some hw might want to perform special
-	 * operation on specific ioctl. For instance on wait idle some hw
-	 * might want to perform and HDP flush through MMIO as it seems that
-	 * some R6XX/R7XX hw doesn't take HDP flush into account if programmed
-	 * through ring.
-	 */
-	void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo);
+	/* Flush the HDP cache via MMIO */
+	void (*mmio_hdp_flush)(struct radeon_device *rdev);
 	/* check if 3D engine is idle */
 	bool (*gui_idle)(struct radeon_device *rdev);
 	/* wait for mc_idle */
@@ -1782,16 +1820,26 @@ struct radeon_asic {
 	struct {
 		void (*tlb_flush)(struct radeon_device *rdev);
 		void (*set_page)(struct radeon_device *rdev, unsigned i,
-				 uint64_t addr);
+				 uint64_t addr, uint32_t flags);
 	} gart;
 	struct {
 		int (*init)(struct radeon_device *rdev);
 		void (*fini)(struct radeon_device *rdev);
-		void (*set_page)(struct radeon_device *rdev,
-				 struct radeon_ib *ib,
-				 uint64_t pe,
-				 uint64_t addr, unsigned count,
-				 uint32_t incr, uint32_t flags);
+		void (*copy_pages)(struct radeon_device *rdev,
+				   struct radeon_ib *ib,
+				   uint64_t pe, uint64_t src,
+				   unsigned count);
+		void (*write_pages)(struct radeon_device *rdev,
+				    struct radeon_ib *ib,
+				    uint64_t pe,
+				    uint64_t addr, unsigned count,
+				    uint32_t incr, uint32_t flags);
+		void (*set_pages)(struct radeon_device *rdev,
+				  struct radeon_ib *ib,
+				  uint64_t pe,
+				  uint64_t addr, unsigned count,
+				  uint32_t incr, uint32_t flags);
+		void (*pad_ib)(struct radeon_ib *ib);
 	} vm;
 	/* ring specific callbacks */
 	struct radeon_asic_ring *ring[RADEON_NUM_RINGS];
@@ -1818,24 +1866,24 @@ struct radeon_asic {
 	} display;
 	/* copy functions for bo handling */
 	struct {
-		int (*blit)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence **fence);
+		struct radeon_fence *(*blit)(struct radeon_device *rdev,
+					     uint64_t src_offset,
+					     uint64_t dst_offset,
+					     unsigned num_gpu_pages,
+					     struct reservation_object *resv);
 		u32 blit_ring_index;
-		int (*dma)(struct radeon_device *rdev,
-			   uint64_t src_offset,
-			   uint64_t dst_offset,
-			   unsigned num_gpu_pages,
-			   struct radeon_fence **fence);
+		struct radeon_fence *(*dma)(struct radeon_device *rdev,
+					    uint64_t src_offset,
+					    uint64_t dst_offset,
+					    unsigned num_gpu_pages,
+					    struct reservation_object *resv);
 		u32 dma_ring_index;
 		/* method used for bo copy */
-		int (*copy)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence **fence);
+		struct radeon_fence *(*copy)(struct radeon_device *rdev,
+					     uint64_t src_offset,
+					     uint64_t dst_offset,
+					     unsigned num_gpu_pages,
+					     struct reservation_object *resv);
 		/* ring used for bo copies */
 		u32 copy_ring_index;
 	} copy;
@@ -2117,6 +2165,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
 int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 			    struct drm_file *filp);
+int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *filp);
 int radeon_gem_pin_ioctl(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv);
 int radeon_gem_unpin_ioctl(struct drm_device *dev, void *data,
@@ -2273,6 +2323,7 @@ struct radeon_device {
 	struct radeon_mman		mman;
 	struct radeon_fence_driver	fence_drv[RADEON_NUM_RINGS];
 	wait_queue_head_t		fence_queue;
+	unsigned			fence_context;
 	struct mutex			ring_lock;
 	struct radeon_ring		ring[RADEON_NUM_RINGS];
 	bool				ib_pool_ready;
@@ -2291,7 +2342,7 @@ struct radeon_device {
 	bool				need_dma32;
 	bool				accel_working;
 	bool				fastfb_working; /* IGP feature*/
-	bool				needs_reset;
+	bool				needs_reset, in_reset;
 	struct radeon_surface_reg surface_regs[RADEON_GEM_MAX_SURFACES];
 	const struct firmware *me_fw;	/* all family ME firmware */
 	const struct firmware *pfp_fw;	/* r6/700 PFP firmware */
@@ -2299,10 +2350,12 @@ struct radeon_device {
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	const struct firmware *ce_fw;	/* SI CE firmware */
 	const struct firmware *mec_fw;	/* CIK MEC firmware */
+	const struct firmware *mec2_fw;	/* KV MEC2 firmware */
 	const struct firmware *sdma_fw;	/* CIK SDMA firmware */
 	const struct firmware *smc_fw;	/* SMC firmware */
 	const struct firmware *uvd_fw;	/* UVD firmware */
 	const struct firmware *vce_fw;	/* VCE firmware */
+	bool new_fw;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
@@ -2310,7 +2363,6 @@ struct radeon_device {
 	struct radeon_mec mec;
 	struct work_struct hotplug_work;
 	struct work_struct audio_work;
-	struct work_struct reset_work;
 	int num_crtc; /* number of crtcs */
 	struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */
 	bool has_uvd;
@@ -2342,6 +2394,14 @@ struct radeon_device {
 
 	struct dev_pm_domain vga_pm_domain;
 	bool have_disp_power_ref;
+	u32 px_quirk_flags;
+
+	/* tracking pinned memory */
+	u64 vram_pin_size;
+	u64 gart_pin_size;
+
+	struct mutex	mn_lock;
+	DECLARE_HASHTABLE(mn_hash, 7);
 };
 
 bool radeon_is_px(struct drm_device *dev);
@@ -2352,10 +2412,42 @@ int radeon_device_init(struct radeon_device *rdev,
 void radeon_device_fini(struct radeon_device *rdev);
 int radeon_gpu_wait_for_idle(struct radeon_device *rdev);
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
-		      bool always_indirect);
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
-		  bool always_indirect);
+#define RADEON_MIN_MMIO_SIZE 0x10000
+
+static inline uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
+				    bool always_indirect)
+{
+	/* The mmio size is 64kb at minimum. Allows the if to be optimized out. */
+	if ((reg < rdev->rmmio_size || reg < RADEON_MIN_MMIO_SIZE) && !always_indirect)
+		return readl(((void __iomem *)rdev->rmmio) + reg);
+	else {
+		unsigned long flags;
+		uint32_t ret;
+
+		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		ret = readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
+
+		return ret;
+	}
+}
+
+static inline void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
+				bool always_indirect)
+{
+	if ((reg < rdev->rmmio_size || reg < RADEON_MIN_MMIO_SIZE) && !always_indirect)
+		writel(v, ((void __iomem *)rdev->rmmio) + reg);
+	else {
+		unsigned long flags;
+
+		spin_lock_irqsave(&rdev->mmio_idx_lock, flags);
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+		spin_unlock_irqrestore(&rdev->mmio_idx_lock, flags);
+	}
+}
+
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
 void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 
@@ -2365,7 +2457,17 @@ void cik_mm_wdoorbell(struct radeon_device *rdev, u32 index, u32 v);
 /*
  * Cast helper
  */
-#define to_radeon_fence(p) ((struct radeon_fence *)(p))
+extern const struct fence_ops radeon_fence_ops;
+
+static inline struct radeon_fence *to_radeon_fence(struct fence *f)
+{
+	struct radeon_fence *__f = container_of(f, struct radeon_fence, base);
+
+	if (__f->base.ops == &radeon_fence_ops)
+		return __f;
+
+	return NULL;
+}
 
 /*
  * Registers read & write functions.
@@ -2685,18 +2787,25 @@ void radeon_atombios_fini(struct radeon_device *rdev);
 /*
  * RING helpers.
  */
-#if DRM_DEBUG_CODE == 0
+
+/**
+ * radeon_ring_write - write a value to the ring
+ *
+ * @ring: radeon_ring structure holding ring information
+ * @v: dword (dw) value to write
+ *
+ * Write a value to the requested ring buffer (all asics).
+ */
 static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
 {
+	if (ring->count_dw <= 0)
+		DRM_ERROR("radeon: writing more dwords to the ring than expected!\n");
+
 	ring->ring[ring->wptr++] = v;
 	ring->wptr &= ring->ptr_mask;
 	ring->count_dw--;
 	ring->ring_free_dw--;
 }
-#else
-/* With debugging this is just too big to inline */
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
-#endif
 
 /*
  * ASICs macro.
@@ -2709,10 +2818,13 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
 #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
 #define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
-#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
+#define radeon_gart_set_page(rdev, i, p, f) (rdev)->asic->gart.set_page((rdev), (i), (p), (f))
 #define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev))
 #define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev))
-#define radeon_asic_vm_set_page(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_copy_pages(rdev, ib, pe, src, count) ((rdev)->asic->vm.copy_pages((rdev), (ib), (pe), (src), (count)))
+#define radeon_asic_vm_write_pages(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.write_pages((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_set_pages(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_pages((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
+#define radeon_asic_vm_pad_ib(rdev, ib) ((rdev)->asic->vm.pad_ib((ib)))
 #define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_start((rdev), (cp))
 #define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ring_test((rdev), (cp))
 #define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)]->ib_test((rdev), (cp))
@@ -2732,9 +2844,9 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_hdmi_setmode(rdev, e, m) (rdev)->asic->display.hdmi_setmode((e), (m))
 #define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)]->emit_fence((rdev), (fence))
 #define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)]->emit_semaphore((rdev), (cp), (semaphore), (emit_wait))
-#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
-#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f))
-#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f))
+#define radeon_copy_blit(rdev, s, d, np, resv) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (resv))
+#define radeon_copy_dma(rdev, s, d, np, resv) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (resv))
+#define radeon_copy(rdev, s, d, np, resv) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (resv))
 #define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index
 #define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index
 #define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index
@@ -2808,6 +2920,10 @@ extern void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enabl
 extern void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable);
 extern void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain);
 extern bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo);
+extern int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+				     uint32_t flags);
+extern bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm);
+extern bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm);
 extern void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base);
 extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 extern int radeon_resume_kms(struct drm_device *dev, bool resume, bool fbcon);
@@ -2840,6 +2956,8 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 				    struct radeon_vm *vm);
 int radeon_vm_clear_freed(struct radeon_device *rdev,
 			  struct radeon_vm *vm);
+int radeon_vm_clear_invalids(struct radeon_device *rdev,
+			     struct radeon_vm *vm);
 int radeon_vm_bo_update(struct radeon_device *rdev,
 			struct radeon_bo_va *bo_va,
 			struct ttm_mem_reg *mem);
@@ -2863,10 +2981,10 @@ struct r600_audio_pin *r600_audio_get_pin(struct radeon_device *rdev);
 struct r600_audio_pin *dce6_audio_get_pin(struct radeon_device *rdev);
 void r600_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable);
+		       u8 enable_mask);
 void dce6_audio_enable(struct radeon_device *rdev,
 		       struct r600_audio_pin *pin,
-		       bool enable);
+		       u8 enable_mask);
 
 /*
  * R600 vram scratch functions
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 34b9aa9e3c06..850de57069be 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -194,7 +194,7 @@ static struct radeon_asic r100_asic = {
 	.resume = &r100_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r100_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r100_mc_wait_for_idle,
 	.gart = {
@@ -260,7 +260,7 @@ static struct radeon_asic r200_asic = {
 	.resume = &r100_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r100_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r100_mc_wait_for_idle,
 	.gart = {
@@ -340,7 +340,7 @@ static struct radeon_asic r300_asic = {
 	.resume = &r300_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -406,7 +406,7 @@ static struct radeon_asic r300_asic_pcie = {
 	.resume = &r300_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -472,7 +472,7 @@ static struct radeon_asic r420_asic = {
 	.resume = &r420_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r300_mc_wait_for_idle,
 	.gart = {
@@ -538,7 +538,7 @@ static struct radeon_asic rs400_asic = {
 	.resume = &rs400_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &r300_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs400_mc_wait_for_idle,
 	.gart = {
@@ -604,7 +604,7 @@ static struct radeon_asic rs600_asic = {
 	.resume = &rs600_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs600_mc_wait_for_idle,
 	.gart = {
@@ -672,7 +672,7 @@ static struct radeon_asic rs690_asic = {
 	.resume = &rs690_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rs690_mc_wait_for_idle,
 	.gart = {
@@ -740,7 +740,7 @@ static struct radeon_asic rv515_asic = {
 	.resume = &rv515_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &rv515_mc_wait_for_idle,
 	.gart = {
@@ -806,7 +806,7 @@ static struct radeon_asic r520_asic = {
 	.resume = &r520_resume,
 	.vga_set_state = &r100_vga_set_state,
 	.asic_reset = &rs600_asic_reset,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = NULL,
 	.gui_idle = &r100_gui_idle,
 	.mc_wait_for_idle = &r520_mc_wait_for_idle,
 	.gart = {
@@ -898,7 +898,7 @@ static struct radeon_asic r600_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -963,6 +963,19 @@ static struct radeon_asic r600_asic = {
 	},
 };
 
+static struct radeon_asic_ring rv6xx_uvd_ring = {
+	.ib_execute = &uvd_v1_0_ib_execute,
+	.emit_fence = &uvd_v1_0_fence_emit,
+	.emit_semaphore = &uvd_v1_0_semaphore_emit,
+	.cs_parse = &radeon_uvd_cs_parse,
+	.ring_test = &uvd_v1_0_ring_test,
+	.ib_test = &uvd_v1_0_ib_test,
+	.is_lockup = &radeon_ring_test_lockup,
+	.get_rptr = &uvd_v1_0_get_rptr,
+	.get_wptr = &uvd_v1_0_get_wptr,
+	.set_wptr = &uvd_v1_0_set_wptr,
+};
+
 static struct radeon_asic rv6xx_asic = {
 	.init = &r600_init,
 	.fini = &r600_fini,
@@ -970,7 +983,7 @@ static struct radeon_asic rv6xx_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -982,6 +995,7 @@ static struct radeon_asic rv6xx_asic = {
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring,
 		[R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring,
+		[R600_RING_TYPE_UVD_INDEX] = &rv6xx_uvd_ring,
 	},
 	.irq = {
 		.set = &r600_irq_set,
@@ -1060,7 +1074,7 @@ static struct radeon_asic rs780_asic = {
 	.resume = &r600_resume,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1072,6 +1086,7 @@ static struct radeon_asic rs780_asic = {
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &r600_gfx_ring,
 		[R600_RING_TYPE_DMA_INDEX] = &r600_dma_ring,
+		[R600_RING_TYPE_UVD_INDEX] = &rv6xx_uvd_ring,
 	},
 	.irq = {
 		.set = &r600_irq_set,
@@ -1163,7 +1178,7 @@ static struct radeon_asic rv770_asic = {
 	.resume = &rv770_resume,
 	.asic_reset = &r600_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &r600_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1281,7 +1296,7 @@ static struct radeon_asic evergreen_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1373,7 +1388,7 @@ static struct radeon_asic sumo_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1464,7 +1479,7 @@ static struct radeon_asic btc_asic = {
 	.resume = &evergreen_resume,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1599,7 +1614,7 @@ static struct radeon_asic cayman_asic = {
 	.resume = &cayman_resume,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &rv770_get_xclk,
@@ -1611,7 +1626,10 @@ static struct radeon_asic cayman_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.set_page = &cayman_dma_vm_set_page,
+		.copy_pages = &cayman_dma_vm_copy_pages,
+		.write_pages = &cayman_dma_vm_write_pages,
+		.set_pages = &cayman_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring,
@@ -1699,7 +1717,7 @@ static struct radeon_asic trinity_asic = {
 	.resume = &cayman_resume,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &r600_get_xclk,
@@ -1711,7 +1729,10 @@ static struct radeon_asic trinity_asic = {
 	.vm = {
 		.init = &cayman_vm_init,
 		.fini = &cayman_vm_fini,
-		.set_page = &cayman_dma_vm_set_page,
+		.copy_pages = &cayman_dma_vm_copy_pages,
+		.write_pages = &cayman_dma_vm_write_pages,
+		.set_pages = &cayman_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &cayman_gfx_ring,
@@ -1829,7 +1850,7 @@ static struct radeon_asic si_asic = {
 	.resume = &si_resume,
 	.asic_reset = &si_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.mmio_hdp_flush = r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &si_get_xclk,
@@ -1841,7 +1862,10 @@ static struct radeon_asic si_asic = {
 	.vm = {
 		.init = &si_vm_init,
 		.fini = &si_vm_fini,
-		.set_page = &si_dma_vm_set_page,
+		.copy_pages = &si_dma_vm_copy_pages,
+		.write_pages = &si_dma_vm_write_pages,
+		.set_pages = &si_dma_vm_set_pages,
+		.pad_ib = &cayman_dma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &si_gfx_ring,
@@ -1987,7 +2011,7 @@ static struct radeon_asic ci_asic = {
 	.resume = &cik_resume,
 	.asic_reset = &cik_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = &r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &cik_get_xclk,
@@ -1999,7 +2023,10 @@ static struct radeon_asic ci_asic = {
 	.vm = {
 		.init = &cik_vm_init,
 		.fini = &cik_vm_fini,
-		.set_page = &cik_sdma_vm_set_page,
+		.copy_pages = &cik_sdma_vm_copy_pages,
+		.write_pages = &cik_sdma_vm_write_pages,
+		.set_pages = &cik_sdma_vm_set_pages,
+		.pad_ib = &cik_sdma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring,
@@ -2091,7 +2118,7 @@ static struct radeon_asic kv_asic = {
 	.resume = &cik_resume,
 	.asic_reset = &cik_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
-	.ioctl_wait_idle = NULL,
+	.mmio_hdp_flush = &r600_mmio_hdp_flush,
 	.gui_idle = &r600_gui_idle,
 	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
 	.get_xclk = &cik_get_xclk,
@@ -2103,7 +2130,10 @@ static struct radeon_asic kv_asic = {
 	.vm = {
 		.init = &cik_vm_init,
 		.fini = &cik_vm_fini,
-		.set_page = &cik_sdma_vm_set_page,
+		.copy_pages = &cik_sdma_vm_copy_pages,
+		.write_pages = &cik_sdma_vm_write_pages,
+		.set_pages = &cik_sdma_vm_set_pages,
+		.pad_ib = &cik_sdma_vm_pad_ib,
 	},
 	.ring = {
 		[RADEON_RING_TYPE_GFX_INDEX] = &ci_gfx_ring,
@@ -2281,7 +2311,15 @@ int radeon_asic_init(struct radeon_device *rdev)
 	case CHIP_RS780:
 	case CHIP_RS880:
 		rdev->asic = &rs780_asic;
-		rdev->has_uvd = true;
+		/* 760G/780V/880V don't have UVD */
+		if ((rdev->pdev->device == 0x9616)||
+		    (rdev->pdev->device == 0x9611)||
+		    (rdev->pdev->device == 0x9613)||
+		    (rdev->pdev->device == 0x9711)||
+		    (rdev->pdev->device == 0x9713))
+			rdev->has_uvd = false;
+		else
+			rdev->has_uvd = true;
 		break;
 	case CHIP_RV770:
 	case CHIP_RV730:
@@ -2457,7 +2495,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
@@ -2476,7 +2514,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CP_LS |
@@ -2502,7 +2540,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
@@ -2530,7 +2568,7 @@ int radeon_asic_init(struct radeon_device *rdev)
 			rdev->cg_flags =
 				RADEON_CG_SUPPORT_GFX_MGCG |
 				RADEON_CG_SUPPORT_GFX_MGLS |
-				RADEON_CG_SUPPORT_GFX_CGCG |
+				/*RADEON_CG_SUPPORT_GFX_CGCG |*/
 				RADEON_CG_SUPPORT_GFX_CGLS |
 				RADEON_CG_SUPPORT_GFX_CGTS |
 				RADEON_CG_SUPPORT_GFX_CGTS_LS |
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index 01e7c0ad8f01..d8ace5b28a5b 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -68,7 +68,7 @@ int r100_asic_reset(struct radeon_device *rdev);
 u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
 void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
-			    uint64_t addr);
+			    uint64_t addr, uint32_t flags);
 void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
 int r100_irq_set(struct radeon_device *rdev);
 int r100_irq_process(struct radeon_device *rdev);
@@ -81,11 +81,11 @@ bool r100_semaphore_ring_emit(struct radeon_device *rdev,
 int r100_cs_parse(struct radeon_cs_parser *p);
 void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg);
-int r100_copy_blit(struct radeon_device *rdev,
-		   uint64_t src_offset,
-		   uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *r100_copy_blit(struct radeon_device *rdev,
+				    uint64_t src_offset,
+				    uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
 			 uint32_t offset, uint32_t obj_size);
@@ -152,11 +152,11 @@ void r100_gfx_set_wptr(struct radeon_device *rdev,
 /*
  * r200,rv250,rs300,rv280
  */
-extern int r200_copy_dma(struct radeon_device *rdev,
-			 uint64_t src_offset,
-			 uint64_t dst_offset,
-			 unsigned num_gpu_pages,
-			 struct radeon_fence **fence);
+struct radeon_fence *r200_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset,
+				   uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv);
 void r200_set_safe_registers(struct radeon_device *rdev);
 
 /*
@@ -173,7 +173,7 @@ extern void r300_fence_ring_emit(struct radeon_device *rdev,
 extern int r300_cs_parse(struct radeon_cs_parser *p);
 extern void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev);
 extern void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
-				     uint64_t addr);
+				     uint64_t addr, uint32_t flags);
 extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes);
 extern int rv370_get_pcie_lanes(struct radeon_device *rdev);
 extern void r300_set_reg_safe(struct radeon_device *rdev);
@@ -209,7 +209,7 @@ extern int rs400_suspend(struct radeon_device *rdev);
 extern int rs400_resume(struct radeon_device *rdev);
 void rs400_gart_tlb_flush(struct radeon_device *rdev);
 void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
-			 uint64_t addr);
+			 uint64_t addr, uint32_t flags);
 uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 int rs400_gart_init(struct radeon_device *rdev);
@@ -233,7 +233,7 @@ void rs600_irq_disable(struct radeon_device *rdev);
 u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void rs600_gart_tlb_flush(struct radeon_device *rdev);
 void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
-			 uint64_t addr);
+			 uint64_t addr, uint32_t flags);
 uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 void rs600_bandwidth_update(struct radeon_device *rdev);
@@ -340,18 +340,20 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
 int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
 int r600_dma_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
-int r600_copy_cpdma(struct radeon_device *rdev,
-		    uint64_t src_offset, uint64_t dst_offset,
-		    unsigned num_gpu_pages, struct radeon_fence **fence);
-int r600_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages, struct radeon_fence **fence);
+struct radeon_fence *r600_copy_cpdma(struct radeon_device *rdev,
+				     uint64_t src_offset, uint64_t dst_offset,
+				     unsigned num_gpu_pages,
+				     struct reservation_object *resv);
+struct radeon_fence *r600_copy_dma(struct radeon_device *rdev,
+				   uint64_t src_offset, uint64_t dst_offset,
+				   unsigned num_gpu_pages,
+				   struct reservation_object *resv);
 void r600_hpd_init(struct radeon_device *rdev);
 void r600_hpd_fini(struct radeon_device *rdev);
 bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 void r600_hpd_set_polarity(struct radeon_device *rdev,
 			   enum radeon_hpd_id hpd);
-extern void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo);
+extern void r600_mmio_hdp_flush(struct radeon_device *rdev);
 extern bool r600_gui_idle(struct radeon_device *rdev);
 extern void r600_pm_misc(struct radeon_device *rdev);
 extern void r600_pm_init_profile(struct radeon_device *rdev);
@@ -389,7 +391,6 @@ void r600_disable_interrupts(struct radeon_device *rdev);
 void r600_rlc_stop(struct radeon_device *rdev);
 /* r600 audio */
 int r600_audio_init(struct radeon_device *rdev);
-struct r600_audio_pin r600_audio_status(struct radeon_device *rdev);
 void r600_audio_fini(struct radeon_device *rdev);
 void r600_audio_set_dto(struct drm_encoder *encoder, u32 clock);
 void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder, void *buffer,
@@ -461,10 +462,10 @@ bool rv770_page_flip_pending(struct radeon_device *rdev, int crtc);
 void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
 void r700_cp_stop(struct radeon_device *rdev);
 void r700_cp_fini(struct radeon_device *rdev);
-int rv770_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 u32 rv770_get_xclk(struct radeon_device *rdev);
 int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
 int rv770_get_temp(struct radeon_device *rdev);
@@ -535,10 +536,10 @@ void evergreen_dma_fence_ring_emit(struct radeon_device *rdev,
 				   struct radeon_fence *fence);
 void evergreen_dma_ring_ib_execute(struct radeon_device *rdev,
 				   struct radeon_ib *ib);
-int evergreen_copy_dma(struct radeon_device *rdev,
-		       uint64_t src_offset, uint64_t dst_offset,
-		       unsigned num_gpu_pages,
-		       struct radeon_fence **fence);
+struct radeon_fence *evergreen_copy_dma(struct radeon_device *rdev,
+					uint64_t src_offset, uint64_t dst_offset,
+					unsigned num_gpu_pages,
+					struct reservation_object *resv);
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
 void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
 int evergreen_get_temp(struct radeon_device *rdev);
@@ -606,11 +607,22 @@ void cayman_dma_ring_ib_execute(struct radeon_device *rdev,
 				struct radeon_ib *ib);
 bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
-void cayman_dma_vm_set_page(struct radeon_device *rdev,
-			    struct radeon_ib *ib,
-			    uint64_t pe,
-			    uint64_t addr, unsigned count,
-			    uint32_t incr, uint32_t flags);
+
+void cayman_dma_vm_copy_pages(struct radeon_device *rdev,
+			      struct radeon_ib *ib,
+			      uint64_t pe, uint64_t src,
+			      unsigned count);
+void cayman_dma_vm_write_pages(struct radeon_device *rdev,
+			       struct radeon_ib *ib,
+			       uint64_t pe,
+			       uint64_t addr, unsigned count,
+			       uint32_t incr, uint32_t flags);
+void cayman_dma_vm_set_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags);
+void cayman_dma_vm_pad_ib(struct radeon_ib *ib);
 
 void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 
@@ -689,15 +701,26 @@ int si_vm_init(struct radeon_device *rdev);
 void si_vm_fini(struct radeon_device *rdev);
 void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
-int si_copy_dma(struct radeon_device *rdev,
-		uint64_t src_offset, uint64_t dst_offset,
-		unsigned num_gpu_pages,
-		struct radeon_fence **fence);
-void si_dma_vm_set_page(struct radeon_device *rdev,
-			struct radeon_ib *ib,
-			uint64_t pe,
-			uint64_t addr, unsigned count,
-			uint32_t incr, uint32_t flags);
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+				 uint64_t src_offset, uint64_t dst_offset,
+				 unsigned num_gpu_pages,
+				 struct reservation_object *resv);
+
+void si_dma_vm_copy_pages(struct radeon_device *rdev,
+			  struct radeon_ib *ib,
+			  uint64_t pe, uint64_t src,
+			  unsigned count);
+void si_dma_vm_write_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags);
+void si_dma_vm_set_pages(struct radeon_device *rdev,
+			 struct radeon_ib *ib,
+			 uint64_t pe,
+			 uint64_t addr, unsigned count,
+			 uint32_t incr, uint32_t flags);
+
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
@@ -737,14 +760,14 @@ bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
 				  struct radeon_semaphore *semaphore,
 				  bool emit_wait);
 void cik_sdma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int cik_copy_dma(struct radeon_device *rdev,
-		 uint64_t src_offset, uint64_t dst_offset,
-		 unsigned num_gpu_pages,
-		 struct radeon_fence **fence);
-int cik_copy_cpdma(struct radeon_device *rdev,
-		   uint64_t src_offset, uint64_t dst_offset,
-		   unsigned num_gpu_pages,
-		   struct radeon_fence **fence);
+struct radeon_fence *cik_copy_dma(struct radeon_device *rdev,
+				  uint64_t src_offset, uint64_t dst_offset,
+				  unsigned num_gpu_pages,
+				  struct reservation_object *resv);
+struct radeon_fence *cik_copy_cpdma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv);
 int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
 int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
@@ -771,11 +794,23 @@ int cik_irq_process(struct radeon_device *rdev);
 int cik_vm_init(struct radeon_device *rdev);
 void cik_vm_fini(struct radeon_device *rdev);
 void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
-void cik_sdma_vm_set_page(struct radeon_device *rdev,
-			  struct radeon_ib *ib,
-			  uint64_t pe,
-			  uint64_t addr, unsigned count,
-			  uint32_t incr, uint32_t flags);
+
+void cik_sdma_vm_copy_pages(struct radeon_device *rdev,
+			    struct radeon_ib *ib,
+			    uint64_t pe, uint64_t src,
+			    unsigned count);
+void cik_sdma_vm_write_pages(struct radeon_device *rdev,
+			     struct radeon_ib *ib,
+			     uint64_t pe,
+			     uint64_t addr, unsigned count,
+			     uint32_t incr, uint32_t flags);
+void cik_sdma_vm_set_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags);
+void cik_sdma_vm_pad_ib(struct radeon_ib *ib);
+
 void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 u32 cik_gfx_get_rptr(struct radeon_device *rdev,
@@ -848,6 +883,7 @@ uint32_t uvd_v1_0_get_wptr(struct radeon_device *rdev,
                            struct radeon_ring *ring);
 void uvd_v1_0_set_wptr(struct radeon_device *rdev,
                        struct radeon_ring *ring);
+int uvd_v1_0_resume(struct radeon_device *rdev);
 
 int uvd_v1_0_init(struct radeon_device *rdev);
 void uvd_v1_0_fini(struct radeon_device *rdev);
@@ -855,6 +891,8 @@ int uvd_v1_0_start(struct radeon_device *rdev);
 void uvd_v1_0_stop(struct radeon_device *rdev);
 
 int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+void uvd_v1_0_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence);
 int uvd_v1_0_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
 bool uvd_v1_0_semaphore_emit(struct radeon_device *rdev,
 			     struct radeon_ring *ring,
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index 173f378428a9..df69b92ba164 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -447,11 +447,18 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev,
 		}
 	}
 
+	/* Fujitsu D3003-S2 board lists DVI-I as DVI-I and VGA */
+	if ((dev->pdev->device == 0x9805) &&
+	    (dev->pdev->subsystem_vendor == 0x1734) &&
+	    (dev->pdev->subsystem_device == 0x11bd)) {
+		if (*connector_type == DRM_MODE_CONNECTOR_VGA)
+			return false;
+	}
 
 	return true;
 }
 
-const int supported_devices_connector_convert[] = {
+static const int supported_devices_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_VGA,
 	DRM_MODE_CONNECTOR_DVII,
@@ -470,7 +477,7 @@ const int supported_devices_connector_convert[] = {
 	DRM_MODE_CONNECTOR_DisplayPort
 };
 
-const uint16_t supported_devices_connector_object_id_convert[] = {
+static const uint16_t supported_devices_connector_object_id_convert[] = {
 	CONNECTOR_OBJECT_ID_NONE,
 	CONNECTOR_OBJECT_ID_VGA,
 	CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_I, /* not all boards support DL */
@@ -487,7 +494,7 @@ const uint16_t supported_devices_connector_object_id_convert[] = {
 	CONNECTOR_OBJECT_ID_SVIDEO
 };
 
-const int object_connector_convert[] = {
+static const int object_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_DVII,
 	DRM_MODE_CONNECTOR_DVII,
@@ -1963,7 +1970,7 @@ static const char *thermal_controller_names[] = {
 	"adm1032",
 	"adm1030",
 	"max6649",
-	"lm64",
+	"lm63", /* lm64 */
 	"f75375",
 	"asc7xxx",
 };
@@ -1974,7 +1981,7 @@ static const char *pp_lib_thermal_controller_names[] = {
 	"adm1032",
 	"adm1030",
 	"max6649",
-	"lm64",
+	"lm63", /* lm64 */
 	"f75375",
 	"RV6xx",
 	"RV770",
@@ -2281,19 +2288,31 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 				 (controller->ucFanParameters &
 				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
 			rdev->pm.int_thermal_type = THERMAL_TYPE_KV;
-		} else if ((controller->ucType ==
-			    ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) ||
-			   (controller->ucType ==
-			    ATOM_PP_THERMALCONTROLLER_ADT7473_WITH_INTERNAL) ||
-			   (controller->ucType ==
-			    ATOM_PP_THERMALCONTROLLER_EMC2103_WITH_INTERNAL)) {
-			DRM_INFO("Special thermal controller config\n");
+		} else if (controller->ucType ==
+			   ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) {
+			DRM_INFO("External GPIO thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_EXTERNAL_GPIO;
+		} else if (controller->ucType ==
+			   ATOM_PP_THERMALCONTROLLER_ADT7473_WITH_INTERNAL) {
+			DRM_INFO("ADT7473 with internal thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_ADT7473_WITH_INTERNAL;
+		} else if (controller->ucType ==
+			   ATOM_PP_THERMALCONTROLLER_EMC2103_WITH_INTERNAL) {
+			DRM_INFO("EMC2103 with internal thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_EMC2103_WITH_INTERNAL;
 		} else if (controller->ucType < ARRAY_SIZE(pp_lib_thermal_controller_names)) {
 			DRM_INFO("Possible %s thermal controller at 0x%02x %s fan control\n",
 				 pp_lib_thermal_controller_names[controller->ucType],
 				 controller->ucI2cAddress >> 1,
 				 (controller->ucFanParameters &
 				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_EXTERNAL;
 			i2c_bus = radeon_lookup_i2c_gpio(rdev, controller->ucI2cLine);
 			rdev->pm.i2c_bus = radeon_i2c_lookup(rdev, &i2c_bus);
 			if (rdev->pm.i2c_bus) {
@@ -3236,6 +3255,41 @@ int radeon_atom_get_leakage_vddc_based_on_leakage_params(struct radeon_device *r
 	return 0;
 }
 
+union get_voltage_info {
+	struct  _GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2 in;
+	struct  _GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 evv_out;
+};
+
+int radeon_atom_get_voltage_evv(struct radeon_device *rdev,
+				u16 virtual_voltage_id,
+				u16 *voltage)
+{
+	int index = GetIndexIntoMasterTable(COMMAND, GetVoltageInfo);
+	u32 entry_id;
+	u32 count = rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count;
+	union get_voltage_info args;
+
+	for (entry_id = 0; entry_id < count; entry_id++) {
+		if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].v ==
+		    virtual_voltage_id)
+			break;
+	}
+
+	if (entry_id >= count)
+		return -EINVAL;
+
+	args.in.ucVoltageType = VOLTAGE_TYPE_VDDC;
+	args.in.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE;
+	args.in.ulSCLKFreq =
+		cpu_to_le32(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].clk);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	*voltage = le16_to_cpu(args.evv_out.usVoltageLevel);
+
+	return 0;
+}
+
 int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
 					  u16 voltage_level, u8 voltage_type,
 					  u32 *gpio_value, u32 *gpio_mask)
@@ -3397,6 +3451,50 @@ radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
 	return false;
 }
 
+int radeon_atom_get_svi2_info(struct radeon_device *rdev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 3:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v3(&voltage_info->v3,
+								      voltage_type,
+								      VOLTAGE_OBJ_SVID2);
+				if (voltage_object) {
+					*svd_gpio_id = voltage_object->v3.asSVID2Obj.ucSVDGpioId;
+					*svc_gpio_id = voltage_object->v3.asSVID2Obj.ucSVCGpioId;
+				} else {
+					return -EINVAL;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return 0;
+}
+
 int radeon_atom_get_max_voltage(struct radeon_device *rdev,
 				u8 voltage_type, u16 *max_voltage)
 {
diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index a9fb0d016d38..8bc7d0bbd3c8 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -33,7 +33,6 @@ static struct radeon_atpx_priv {
 	bool atpx_detected;
 	/* handle for device - and atpx */
 	acpi_handle dhandle;
-	acpi_handle other_handle;
 	struct radeon_atpx atpx;
 } radeon_atpx_priv;
 
@@ -453,10 +452,9 @@ static bool radeon_atpx_pci_probe_handle(struct pci_dev *pdev)
 		return false;
 
 	status = acpi_get_handle(dhandle, "ATPX", &atpx_handle);
-	if (ACPI_FAILURE(status)) {
-		radeon_atpx_priv.other_handle = dhandle;
+	if (ACPI_FAILURE(status))
 		return false;
-	}
+
 	radeon_atpx_priv.dhandle = dhandle;
 	radeon_atpx_priv.atpx.handle = atpx_handle;
 	return true;
@@ -540,16 +538,6 @@ static bool radeon_atpx_detect(void)
 		printk(KERN_INFO "VGA switcheroo: detected switching method %s handle\n",
 		       acpi_method_name);
 		radeon_atpx_priv.atpx_detected = true;
-		/*
-		 * On some systems hotplug events are generated for the device
-		 * being switched off when ATPX is executed.  They cause ACPI
-		 * hotplug to trigger and attempt to remove the device from
-		 * the system, which causes it to break down.  Prevent that from
-		 * happening by setting the no_hotplug flag for the involved
-		 * ACPI device objects.
-		 */
-		acpi_bus_no_hotplug(radeon_atpx_priv.dhandle);
-		acpi_bus_no_hotplug(radeon_atpx_priv.other_handle);
 		return true;
 	}
 	return false;
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index 6e05a2e75a46..9e7f23dd14bd 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -45,33 +45,29 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
 	for (i = 0; i < n; i++) {
 		switch (flag) {
 		case RADEON_BENCHMARK_COPY_DMA:
-			r = radeon_copy_dma(rdev, saddr, daddr,
-					    size / RADEON_GPU_PAGE_SIZE,
-					    &fence);
+			fence = radeon_copy_dma(rdev, saddr, daddr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 			break;
 		case RADEON_BENCHMARK_COPY_BLIT:
-			r = radeon_copy_blit(rdev, saddr, daddr,
-					     size / RADEON_GPU_PAGE_SIZE,
-					     &fence);
+			fence = radeon_copy_blit(rdev, saddr, daddr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
 			break;
 		default:
 			DRM_ERROR("Unknown copy method\n");
-			r = -EINVAL;
+			return -EINVAL;
 		}
-		if (r)
-			goto exit_do_move;
+		if (IS_ERR(fence))
+			return PTR_ERR(fence);
+
 		r = radeon_fence_wait(fence, false);
-		if (r)
-			goto exit_do_move;
 		radeon_fence_unref(&fence);
+		if (r)
+			return r;
 	}
 	end_jiffies = jiffies;
-	r = jiffies_to_msecs(end_jiffies - start_jiffies);
-
-exit_do_move:
-	if (fence)
-		radeon_fence_unref(&fence);
-	return r;
+	return jiffies_to_msecs(end_jiffies - start_jiffies);
 }
 
 
@@ -97,7 +93,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	int time;
 
 	n = RADEON_BENCHMARK_ITERATIONS;
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, NULL, &sobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, 0, NULL, NULL, &sobj);
 	if (r) {
 		goto out_cleanup;
 	}
@@ -109,7 +105,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	if (r) {
 		goto out_cleanup;
 	}
-	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, NULL, &dobj);
+	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, ddomain, 0, NULL, NULL, &dobj);
 	if (r) {
 		goto out_cleanup;
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c
index 6a03624fadaa..63ccb8fa799c 100644
--- a/drivers/gpu/drm/radeon/radeon_bios.c
+++ b/drivers/gpu/drm/radeon/radeon_bios.c
@@ -658,12 +658,10 @@ bool radeon_get_bios(struct radeon_device *rdev)
 		r = igp_read_bios_from_vram(rdev);
 	if (r == false)
 		r = radeon_read_bios(rdev);
-	if (r == false) {
+	if (r == false)
 		r = radeon_read_disabled_bios(rdev);
-	}
-	if (r == false) {
+	if (r == false)
 		r = radeon_read_platform_bios(rdev);
-	}
 	if (r == false || rdev->bios == NULL) {
 		DRM_ERROR("Unable to locate a BIOS ROM\n");
 		rdev->bios = NULL;
diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c
index 6651177110f0..3e5f6b71f3ad 100644
--- a/drivers/gpu/drm/radeon/radeon_combios.c
+++ b/drivers/gpu/drm/radeon/radeon_combios.c
@@ -116,7 +116,7 @@ enum radeon_combios_connector {
 	CONNECTOR_UNSUPPORTED_LEGACY
 };
 
-const int legacy_connector_convert[] = {
+static const int legacy_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown,
 	DRM_MODE_CONNECTOR_DVID,
 	DRM_MODE_CONNECTOR_VGA,
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 44831197e82e..300c4b3d4669 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -107,7 +107,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 	case DRM_MODE_CONNECTOR_DVII:
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (radeon_connector->use_digital) {
-			if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (connector->display_info.bpc)
 					bpc = connector->display_info.bpc;
 			}
@@ -115,7 +115,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	case DRM_MODE_CONNECTOR_DVID:
 	case DRM_MODE_CONNECTOR_HDMIA:
-		if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -124,7 +124,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		dig_connector = radeon_connector->con_priv;
 		if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) ||
 		    (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) ||
-		    drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		    drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			if (connector->display_info.bpc)
 				bpc = connector->display_info.bpc;
 		}
@@ -148,7 +148,7 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 		break;
 	}
 
-	if (drm_detect_hdmi_monitor(radeon_connector->edid)) {
+	if (drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 		/* hdmi deep color only implemented on DCE4+ */
 		if ((bpc > 8) && !ASIC_IS_DCE4(rdev)) {
 			DRM_DEBUG("%s: HDMI deep color %d bpc unsupported. Using 8 bpc.\n",
@@ -197,10 +197,19 @@ int radeon_get_monitor_bpc(struct drm_connector *connector)
 						  connector->name, bpc);
 			}
 		}
+		else if (bpc > 8) {
+			/* max_tmds_clock missing, but hdmi spec mandates it for deep color. */
+			DRM_DEBUG("%s: Required max tmds clock for HDMI deep color missing. Using 8 bpc.\n",
+					  connector->name);
+			bpc = 8;
+		}
 	}
 
-	if ((radeon_deep_color == 0) && (bpc > 8))
+	if ((radeon_deep_color == 0) && (bpc > 8)) {
+		DRM_DEBUG("%s: Deep color disabled. Set radeon module param deep_color=1 to enable.\n",
+				  connector->name);
 		bpc = 8;
+	}
 
 	DRM_DEBUG("%s: Display bpc=%d, returned bpc=%d\n",
 			  connector->name, connector->display_info.bpc, bpc);
@@ -216,7 +225,6 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 	struct drm_encoder *best_encoder = NULL;
 	struct drm_encoder *encoder = NULL;
 	struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
-	struct drm_mode_object *obj;
 	bool connected;
 	int i;
 
@@ -226,14 +234,11 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev,
-					   connector->encoder_ids[i],
-					   DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev,
+					   connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
-
 		if ((encoder == best_encoder) && (status == connector_status_connected))
 			connected = true;
 		else
@@ -249,7 +254,6 @@ radeon_connector_update_scratch_regs(struct drm_connector *connector, enum drm_c
 
 static struct drm_encoder *radeon_find_encoder(struct drm_connector *connector, int encoder_type)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	int i;
 
@@ -257,34 +261,134 @@ static struct drm_encoder *radeon_find_encoder(struct drm_connector *connector,
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		if (encoder->encoder_type == encoder_type)
 			return encoder;
 	}
 	return NULL;
 }
 
+struct edid *radeon_connector_edid(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	struct drm_property_blob *edid_blob = connector->edid_blob_ptr;
+
+	if (radeon_connector->edid) {
+		return radeon_connector->edid;
+	} else if (edid_blob) {
+		struct edid *edid = kmemdup(edid_blob->data, edid_blob->length, GFP_KERNEL);
+		if (edid)
+			radeon_connector->edid = edid;
+	}
+	return radeon_connector->edid;
+}
+
+static void radeon_connector_get_edid(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+	if (radeon_connector->edid)
+		return;
+
+	/* on hw with routers, select right port */
+	if (radeon_connector->router.ddc_valid)
+		radeon_router_select_ddc_port(radeon_connector);
+
+	if ((radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
+	     ENCODER_OBJECT_ID_NONE) &&
+	    radeon_connector->ddc_bus->has_aux) {
+		radeon_connector->edid = drm_get_edid(connector,
+						      &radeon_connector->ddc_bus->aux.ddc);
+	} else if ((connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
+		   (connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
+		struct radeon_connector_atom_dig *dig = radeon_connector->con_priv;
+
+		if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT ||
+		     dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) &&
+		    radeon_connector->ddc_bus->has_aux)
+			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+							      &radeon_connector->ddc_bus->aux.ddc);
+		else if (radeon_connector->ddc_bus)
+			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+							      &radeon_connector->ddc_bus->adapter);
+	} else if (radeon_connector->ddc_bus) {
+		radeon_connector->edid = drm_get_edid(&radeon_connector->base,
+						      &radeon_connector->ddc_bus->adapter);
+	}
+
+	if (!radeon_connector->edid) {
+		if (rdev->is_atom_bios) {
+			/* some laptops provide a hardcoded edid in rom for LCDs */
+			if (((connector->connector_type == DRM_MODE_CONNECTOR_LVDS) ||
+			     (connector->connector_type == DRM_MODE_CONNECTOR_eDP)))
+				radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
+		} else {
+			/* some servers provide a hardcoded edid in rom for KVMs */
+			radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
+		}
+	}
+}
+
+static void radeon_connector_free_edid(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+	if (radeon_connector->edid) {
+		kfree(radeon_connector->edid);
+		radeon_connector->edid = NULL;
+	}
+}
+
+static int radeon_ddc_get_modes(struct drm_connector *connector)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	int ret;
+
+	if (radeon_connector->edid) {
+		drm_mode_connector_update_edid_property(connector, radeon_connector->edid);
+		ret = drm_add_edid_modes(connector, radeon_connector->edid);
+		drm_edid_to_eld(connector, radeon_connector->edid);
+		return ret;
+	}
+	drm_mode_connector_update_edid_property(connector, NULL);
+	return 0;
+}
+
 static struct drm_encoder *radeon_best_single_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
 	/* pick the encoder ids */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
+static void radeon_get_native_mode(struct drm_connector *connector)
+{
+	struct drm_encoder *encoder = radeon_best_single_encoder(connector);
+	struct radeon_encoder *radeon_encoder;
+
+	if (encoder == NULL)
+		return;
+
+	radeon_encoder = to_radeon_encoder(encoder);
+
+	if (!list_empty(&connector->probed_modes)) {
+		struct drm_display_mode *preferred_mode =
+			list_first_entry(&connector->probed_modes,
+					 struct drm_display_mode, head);
+
+		radeon_encoder->native_mode = *preferred_mode;
+	} else {
+		radeon_encoder->native_mode.clock = 0;
+	}
+}
+
 /*
  * radeon_connector_analog_encoder_conflict_solve
  * - search for other connectors sharing this encoder
@@ -585,6 +689,35 @@ static int radeon_connector_set_property(struct drm_connector *connector, struct
 		radeon_property_change_mode(&radeon_encoder->base);
 	}
 
+	if (property == dev->mode_config.scaling_mode_property) {
+		enum radeon_rmx_type rmx_type;
+
+		if (connector->encoder)
+			radeon_encoder = to_radeon_encoder(connector->encoder);
+		else {
+			struct drm_connector_helper_funcs *connector_funcs = connector->helper_private;
+			radeon_encoder = to_radeon_encoder(connector_funcs->best_encoder(connector));
+		}
+
+		switch (val) {
+		default:
+		case DRM_MODE_SCALE_NONE: rmx_type = RMX_OFF; break;
+		case DRM_MODE_SCALE_CENTER: rmx_type = RMX_CENTER; break;
+		case DRM_MODE_SCALE_ASPECT: rmx_type = RMX_ASPECT; break;
+		case DRM_MODE_SCALE_FULLSCREEN: rmx_type = RMX_FULL; break;
+		}
+		if (radeon_encoder->rmx_type == rmx_type)
+			return 0;
+
+		if ((rmx_type != DRM_MODE_SCALE_NONE) &&
+		    (radeon_encoder->native_mode.clock == 0))
+			return 0;
+
+		radeon_encoder->rmx_type = rmx_type;
+
+		radeon_property_change_mode(&radeon_encoder->base);
+	}
+
 	return 0;
 }
 
@@ -625,22 +758,20 @@ static void radeon_fixup_lvds_native_mode(struct drm_encoder *encoder,
 
 static int radeon_lvds_get_modes(struct drm_connector *connector)
 {
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder;
 	int ret = 0;
 	struct drm_display_mode *mode;
 
-	if (radeon_connector->ddc_bus) {
-		ret = radeon_ddc_get_modes(radeon_connector);
-		if (ret > 0) {
-			encoder = radeon_best_single_encoder(connector);
-			if (encoder) {
-				radeon_fixup_lvds_native_mode(encoder, connector);
-				/* add scaled modes */
-				radeon_add_common_modes(encoder, connector);
-			}
-			return ret;
+	radeon_connector_get_edid(connector);
+	ret = radeon_ddc_get_modes(connector);
+	if (ret > 0) {
+		encoder = radeon_best_single_encoder(connector);
+		if (encoder) {
+			radeon_fixup_lvds_native_mode(encoder, connector);
+			/* add scaled modes */
+			radeon_add_common_modes(encoder, connector);
 		}
+		return ret;
 	}
 
 	encoder = radeon_best_single_encoder(connector);
@@ -715,16 +846,9 @@ radeon_lvds_detect(struct drm_connector *connector, bool force)
 	}
 
 	/* check for edid as well */
+	radeon_connector_get_edid(connector);
 	if (radeon_connector->edid)
 		ret = connector_status_connected;
-	else {
-		if (radeon_connector->ddc_bus) {
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-			if (radeon_connector->edid)
-				ret = connector_status_connected;
-		}
-	}
 	/* check acpi lid status ??? */
 
 	radeon_connector_update_scratch_regs(connector, ret);
@@ -737,10 +861,9 @@ static void radeon_connector_destroy(struct drm_connector *connector)
 {
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 
-	if (radeon_connector->edid)
-		kfree(radeon_connector->edid);
+	radeon_connector_free_edid(connector);
 	kfree(radeon_connector->con_priv);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -797,10 +920,12 @@ static const struct drm_connector_funcs radeon_lvds_connector_funcs = {
 
 static int radeon_vga_get_modes(struct drm_connector *connector)
 {
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	int ret;
 
-	ret = radeon_ddc_get_modes(radeon_connector);
+	radeon_connector_get_edid(connector);
+	ret = radeon_ddc_get_modes(connector);
+
+	radeon_get_native_mode(connector);
 
 	return ret;
 }
@@ -843,28 +968,26 @@ radeon_vga_detect(struct drm_connector *connector, bool force)
 		dret = radeon_ddc_probe(radeon_connector, false);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
-		if (radeon_connector->edid) {
-			kfree(radeon_connector->edid);
-			radeon_connector->edid = NULL;
-		}
-		radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter);
+		radeon_connector_free_edid(connector);
+		radeon_connector_get_edid(connector);
 
 		if (!radeon_connector->edid) {
 			DRM_ERROR("%s: probed a monitor but no|invalid EDID\n",
 					connector->name);
 			ret = connector_status_connected;
 		} else {
-			radeon_connector->use_digital = !!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
+			radeon_connector->use_digital =
+				!!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
 
 			/* some oems have boards with separate digital and analog connectors
 			 * with a shared ddc line (often vga + hdmi)
 			 */
 			if (radeon_connector->use_digital && radeon_connector->shared_ddc) {
-				kfree(radeon_connector->edid);
-				radeon_connector->edid = NULL;
+				radeon_connector_free_edid(connector);
 				ret = connector_status_disconnected;
-			} else
+			} else {
 				ret = connector_status_connected;
+			}
 		}
 	} else {
 
@@ -999,15 +1122,6 @@ static const struct drm_connector_funcs radeon_tv_connector_funcs = {
 	.set_property = radeon_connector_set_property,
 };
 
-static int radeon_dvi_get_modes(struct drm_connector *connector)
-{
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	int ret;
-
-	ret = radeon_ddc_get_modes(radeon_connector);
-	return ret;
-}
-
 static bool radeon_check_hpd_status_unchanged(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
@@ -1048,7 +1162,6 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 	struct drm_encoder *encoder = NULL;
 	struct drm_encoder_helper_funcs *encoder_funcs;
-	struct drm_mode_object *obj;
 	int i, r;
 	enum drm_connector_status ret = connector_status_disconnected;
 	bool dret = false, broken_edid = false;
@@ -1066,18 +1179,16 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 		dret = radeon_ddc_probe(radeon_connector, false);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
-		if (radeon_connector->edid) {
-			kfree(radeon_connector->edid);
-			radeon_connector->edid = NULL;
-		}
-		radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter);
+		radeon_connector_free_edid(connector);
+		radeon_connector_get_edid(connector);
 
 		if (!radeon_connector->edid) {
 			DRM_ERROR("%s: probed a monitor but no|invalid EDID\n",
 					connector->name);
 			/* rs690 seems to have a problem with connectors not existing and always
 			 * return a block of 0's. If we see this just stop polling on this output */
-			if ((rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) && radeon_connector->base.null_edid_counter) {
+			if ((rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) &&
+			    radeon_connector->base.null_edid_counter) {
 				ret = connector_status_disconnected;
 				DRM_ERROR("%s: detected RS690 floating bus bug, stopping ddc detect\n",
 					  connector->name);
@@ -1087,18 +1198,18 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 				broken_edid = true; /* defer use_digital to later */
 			}
 		} else {
-			radeon_connector->use_digital = !!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
+			radeon_connector->use_digital =
+				!!(radeon_connector->edid->input & DRM_EDID_INPUT_DIGITAL);
 
 			/* some oems have boards with separate digital and analog connectors
 			 * with a shared ddc line (often vga + hdmi)
 			 */
 			if ((!radeon_connector->use_digital) && radeon_connector->shared_ddc) {
-				kfree(radeon_connector->edid);
-				radeon_connector->edid = NULL;
+				radeon_connector_free_edid(connector);
 				ret = connector_status_disconnected;
-			} else
+			} else {
 				ret = connector_status_connected;
-
+			}
 			/* This gets complicated.  We have boards with VGA + HDMI with a
 			 * shared DDC line and we have boards with DVI-D + HDMI with a shared
 			 * DDC line.  The latter is more complex because with DVI<->HDMI adapters
@@ -1118,8 +1229,7 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 						if (list_connector->connector_type != DRM_MODE_CONNECTOR_VGA) {
 							/* hpd is our only option in this case */
 							if (!radeon_hpd_sense(rdev, radeon_connector->hpd.hpd)) {
-								kfree(radeon_connector->edid);
-								radeon_connector->edid = NULL;
+								radeon_connector_free_edid(connector);
 								ret = connector_status_disconnected;
 							}
 						}
@@ -1153,14 +1263,11 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 			if (connector->encoder_ids[i] == 0)
 				break;
 
-			obj = drm_mode_object_find(connector->dev,
-						   connector->encoder_ids[i],
-						   DRM_MODE_OBJECT_ENCODER);
-			if (!obj)
+			encoder = drm_encoder_find(connector->dev,
+						   connector->encoder_ids[i]);
+			if (!encoder)
 				continue;
 
-			encoder = obj_to_encoder(obj);
-
 			if (encoder->encoder_type != DRM_MODE_ENCODER_DAC &&
 			    encoder->encoder_type != DRM_MODE_ENCODER_TVDAC)
 				continue;
@@ -1225,19 +1332,16 @@ static struct drm_encoder *radeon_dvi_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
 	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	int i;
 	for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) {
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
-
 		if (radeon_connector->use_digital == true) {
 			if (encoder->encoder_type == DRM_MODE_ENCODER_TMDS)
 				return encoder;
@@ -1252,13 +1356,8 @@ static struct drm_encoder *radeon_dvi_encoder(struct drm_connector *connector)
 
 	/* then check use digitial */
 	/* pick the first one */
-	if (enc_id) {
-		obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
-			return NULL;
-		encoder = obj_to_encoder(obj);
-		return encoder;
-	}
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
 	return NULL;
 }
 
@@ -1291,7 +1390,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D) ||
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B))
 			return MODE_OK;
-		else if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+		else if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 			/* HDMI 1.3+ supports max clock of 340 Mhz */
 			if (mode->clock > 340000)
 				return MODE_CLOCK_HIGH;
@@ -1310,7 +1409,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 }
 
 static const struct drm_connector_helper_funcs radeon_dvi_connector_helper_funcs = {
-	.get_modes = radeon_dvi_get_modes,
+	.get_modes = radeon_vga_get_modes,
 	.mode_valid = radeon_dvi_mode_valid,
 	.best_encoder = radeon_dvi_encoder,
 };
@@ -1339,7 +1438,8 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			if (!radeon_dig_connector->edp_on)
 				atombios_set_edp_panel_power(connector,
 							     ATOM_TRANSMITTER_ACTION_POWER_ON);
-			ret = radeon_ddc_get_modes(radeon_connector);
+			radeon_connector_get_edid(connector);
+			ret = radeon_ddc_get_modes(connector);
 			if (!radeon_dig_connector->edp_on)
 				atombios_set_edp_panel_power(connector,
 							     ATOM_TRANSMITTER_ACTION_POWER_OFF);
@@ -1350,7 +1450,8 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 				if (encoder)
 					radeon_atom_ext_encoder_setup_ddc(encoder);
 			}
-			ret = radeon_ddc_get_modes(radeon_connector);
+			radeon_connector_get_edid(connector);
+			ret = radeon_ddc_get_modes(connector);
 		}
 
 		if (ret > 0) {
@@ -1383,7 +1484,10 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			if (encoder)
 				radeon_atom_ext_encoder_setup_ddc(encoder);
 		}
-		ret = radeon_ddc_get_modes(radeon_connector);
+		radeon_connector_get_edid(connector);
+		ret = radeon_ddc_get_modes(connector);
+
+		radeon_get_native_mode(connector);
 	}
 
 	return ret;
@@ -1391,7 +1495,6 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 
 u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	struct radeon_encoder *radeon_encoder;
 	int i;
@@ -1400,11 +1503,10 @@ u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *conn
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		radeon_encoder = to_radeon_encoder(encoder);
 
 		switch (radeon_encoder->encoder_id) {
@@ -1419,9 +1521,8 @@ u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *conn
 	return ENCODER_OBJECT_ID_NONE;
 }
 
-bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
+static bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
 {
-	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	struct radeon_encoder *radeon_encoder;
 	int i;
@@ -1431,11 +1532,10 @@ bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
 		if (connector->encoder_ids[i] == 0)
 			break;
 
-		obj = drm_mode_object_find(connector->dev, connector->encoder_ids[i], DRM_MODE_OBJECT_ENCODER);
-		if (!obj)
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
 			continue;
 
-		encoder = obj_to_encoder(obj);
 		radeon_encoder = to_radeon_encoder(encoder);
 		if (radeon_encoder->caps & ATOM_ENCODER_CAP_RECORD_HBR2)
 			found = true;
@@ -1478,10 +1578,7 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 		goto out;
 	}
 
-	if (radeon_connector->edid) {
-		kfree(radeon_connector->edid);
-		radeon_connector->edid = NULL;
-	}
+	radeon_connector_free_edid(connector);
 
 	if ((connector->connector_type == DRM_MODE_CONNECTOR_eDP) ||
 	    (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) {
@@ -1587,7 +1684,7 @@ static int radeon_dp_mode_valid(struct drm_connector *connector,
 		    (radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) {
 			return radeon_dp_mode_valid_helper(connector, mode);
 		} else {
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				/* HDMI 1.3+ supports max clock of 340 Mhz */
 				if (mode->clock > 340000)
 					return MODE_CLOCK_HIGH;
@@ -1747,6 +1844,9 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			drm_object_attach_property(&radeon_connector->base.base,
+						   dev->mode_config.scaling_mode_property,
+						   DRM_MODE_SCALE_NONE);
 			break;
 		case DRM_MODE_CONNECTOR_DVII:
 		case DRM_MODE_CONNECTOR_DVID:
@@ -1768,6 +1868,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 						      0);
 
 			drm_object_attach_property(&radeon_connector->base.base,
+						      dev->mode_config.scaling_mode_property,
+						      DRM_MODE_SCALE_NONE);
+
+			drm_object_attach_property(&radeon_connector->base.base,
 						   rdev->mode_info.dither_property,
 						   RADEON_FMT_DITHER_DISABLE);
 
@@ -1817,6 +1921,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			if (ASIC_IS_AVIVO(rdev))
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			/* no HPD on analog connectors */
 			radeon_connector->hpd.hpd = RADEON_HPD_NONE;
 			connector->polled = DRM_CONNECTOR_POLL_CONNECT;
@@ -1835,6 +1943,10 @@ radeon_add_atom_connector(struct drm_device *dev,
 			drm_object_attach_property(&radeon_connector->base.base,
 						      rdev->mode_info.load_detect_property,
 						      1);
+			if (ASIC_IS_AVIVO(rdev))
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			/* no HPD on analog connectors */
 			radeon_connector->hpd.hpd = RADEON_HPD_NONE;
 			connector->interlace_allowed = true;
@@ -1868,17 +1980,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-			}
 			if (connector_type == DRM_MODE_CONNECTOR_DVII) {
 				radeon_connector->dac_load_detect = true;
 				drm_object_attach_property(&radeon_connector->base.base,
@@ -1918,17 +2031,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-			}
 			subpixel_order = SubPixelHorizontalRGB;
 			connector->interlace_allowed = true;
 			if (connector_type == DRM_MODE_CONNECTOR_HDMIB)
@@ -1965,18 +2079,18 @@ radeon_add_atom_connector(struct drm_device *dev,
 				drm_object_attach_property(&radeon_connector->base.base,
 							      rdev->mode_info.underscan_vborder_property,
 							      0);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   rdev->mode_info.dither_property,
+							   RADEON_FMT_DITHER_DISABLE);
+				drm_object_attach_property(&radeon_connector->base.base,
+							   dev->mode_config.scaling_mode_property,
+							   DRM_MODE_SCALE_NONE);
 			}
 			if (ASIC_IS_DCE2(rdev) && (radeon_audio != 0)) {
 				drm_object_attach_property(&radeon_connector->base.base,
 							   rdev->mode_info.audio_property,
 							   RADEON_AUDIO_AUTO);
 			}
-			if (ASIC_IS_AVIVO(rdev)) {
-				drm_object_attach_property(&radeon_connector->base.base,
-							   rdev->mode_info.dither_property,
-							   RADEON_FMT_DITHER_DISABLE);
-
-			}
 			connector->interlace_allowed = true;
 			/* in theory with a DP to VGA converter... */
 			connector->doublescan_allowed = false;
@@ -2050,7 +2164,7 @@ radeon_add_atom_connector(struct drm_device *dev,
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 
 	connector->display_info.subpixel_order = subpixel_order;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	if (has_aux)
 		radeon_dp_aux_init(radeon_connector);
@@ -2211,5 +2325,5 @@ radeon_add_legacy_connector(struct drm_device *dev,
 	} else
 		connector->polled = DRM_CONNECTOR_POLL_HPD;
 	connector->display_info.subpixel_order = subpixel_order;
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c
index bb0d5c3a8311..ea134a7d51a5 100644
--- a/drivers/gpu/drm/radeon/radeon_cp.c
+++ b/drivers/gpu/drm/radeon/radeon_cp.c
@@ -1298,27 +1298,27 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	dev_priv->buffers_offset = init->buffers_offset;
 	dev_priv->gart_textures_offset = init->gart_textures_offset;
 
-	master_priv->sarea = drm_getsarea(dev);
+	master_priv->sarea = drm_legacy_getsarea(dev);
 	if (!master_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 
-	dev_priv->cp_ring = drm_core_findmap(dev, init->ring_offset);
+	dev_priv->cp_ring = drm_legacy_findmap(dev, init->ring_offset);
 	if (!dev_priv->cp_ring) {
 		DRM_ERROR("could not find cp ring region!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
-	dev_priv->ring_rptr = drm_core_findmap(dev, init->ring_rptr_offset);
+	dev_priv->ring_rptr = drm_legacy_findmap(dev, init->ring_rptr_offset);
 	if (!dev_priv->ring_rptr) {
 		DRM_ERROR("could not find ring read pointer!\n");
 		radeon_do_cleanup_cp(dev);
 		return -EINVAL;
 	}
 	dev->agp_buffer_token = init->buffers_offset;
-	dev->agp_buffer_map = drm_core_findmap(dev, init->buffers_offset);
+	dev->agp_buffer_map = drm_legacy_findmap(dev, init->buffers_offset);
 	if (!dev->agp_buffer_map) {
 		DRM_ERROR("could not find dma buffer region!\n");
 		radeon_do_cleanup_cp(dev);
@@ -1327,7 +1327,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 	if (init->gart_textures_offset) {
 		dev_priv->gart_textures =
-		    drm_core_findmap(dev, init->gart_textures_offset);
+		    drm_legacy_findmap(dev, init->gart_textures_offset);
 		if (!dev_priv->gart_textures) {
 			DRM_ERROR("could not find GART texture region!\n");
 			radeon_do_cleanup_cp(dev);
@@ -1337,9 +1337,9 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
-		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
-		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
-		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap_wc(dev_priv->cp_ring, dev);
+		drm_legacy_ioremap_wc(dev_priv->ring_rptr, dev);
+		drm_legacy_ioremap_wc(dev->agp_buffer_map, dev);
 		if (!dev_priv->cp_ring->handle ||
 		    !dev_priv->ring_rptr->handle ||
 		    !dev->agp_buffer_map->handle) {
@@ -1475,7 +1475,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 			dev_priv->gart_info.mapping.size =
 			    dev_priv->gart_info.table_size;
 
-			drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremap_wc(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr =
 			    dev_priv->gart_info.mapping.handle;
 
@@ -1569,15 +1569,15 @@ static int radeon_do_cleanup_cp(struct drm_device * dev)
 #if __OS_HAS_AGP
 	if (dev_priv->flags & RADEON_IS_AGP) {
 		if (dev_priv->cp_ring != NULL) {
-			drm_core_ioremapfree(dev_priv->cp_ring, dev);
+			drm_legacy_ioremapfree(dev_priv->cp_ring, dev);
 			dev_priv->cp_ring = NULL;
 		}
 		if (dev_priv->ring_rptr != NULL) {
-			drm_core_ioremapfree(dev_priv->ring_rptr, dev);
+			drm_legacy_ioremapfree(dev_priv->ring_rptr, dev);
 			dev_priv->ring_rptr = NULL;
 		}
 		if (dev->agp_buffer_map != NULL) {
-			drm_core_ioremapfree(dev->agp_buffer_map, dev);
+			drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 			dev->agp_buffer_map = NULL;
 		}
 	} else
@@ -1597,7 +1597,7 @@ static int radeon_do_cleanup_cp(struct drm_device * dev)
 
 		if (dev_priv->gart_info.gart_table_location == DRM_ATI_GART_FB)
 		{
-			drm_core_ioremapfree(&dev_priv->gart_info.mapping, dev);
+			drm_legacy_ioremapfree(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr = NULL;
 		}
 	}
@@ -2106,9 +2106,9 @@ int radeon_driver_load(struct drm_device *dev, unsigned long flags)
 	else
 		dev_priv->flags |= RADEON_IS_PCI;
 
-	ret = drm_addmap(dev, pci_resource_start(dev->pdev, 2),
-			 pci_resource_len(dev->pdev, 2), _DRM_REGISTERS,
-			 _DRM_READ_ONLY | _DRM_DRIVER, &dev_priv->mmio);
+	ret = drm_legacy_addmap(dev, pci_resource_start(dev->pdev, 2),
+				pci_resource_len(dev->pdev, 2), _DRM_REGISTERS,
+				_DRM_READ_ONLY | _DRM_DRIVER, &dev_priv->mmio);
 	if (ret != 0)
 		return ret;
 
@@ -2135,8 +2135,8 @@ int radeon_master_create(struct drm_device *dev, struct drm_master *master)
 
 	/* prebuild the SAREA */
 	sareapage = max_t(unsigned long, SAREA_MAX, PAGE_SIZE);
-	ret = drm_addmap(dev, 0, sareapage, _DRM_SHM, _DRM_CONTAINS_LOCK,
-			 &master_priv->sarea);
+	ret = drm_legacy_addmap(dev, 0, sareapage, _DRM_SHM, _DRM_CONTAINS_LOCK,
+				&master_priv->sarea);
 	if (ret) {
 		DRM_ERROR("SAREA setup failed\n");
 		kfree(master_priv);
@@ -2162,7 +2162,7 @@ void radeon_master_destroy(struct drm_device *dev, struct drm_master *master)
 
 	master_priv->sarea_priv = NULL;
 	if (master_priv->sarea)
-		drm_rmmap_locked(dev, master_priv->sarea);
+		drm_legacy_rmmap_locked(dev, master_priv->sarea);
 
 	kfree(master_priv);
 
@@ -2181,9 +2181,9 @@ int radeon_driver_firstopen(struct drm_device *dev)
 	dev_priv->gart_info.table_size = RADEON_PCIGART_TABLE_SIZE;
 
 	dev_priv->fb_aper_offset = pci_resource_start(dev->pdev, 0);
-	ret = drm_addmap(dev, dev_priv->fb_aper_offset,
-			 pci_resource_len(dev->pdev, 0), _DRM_FRAME_BUFFER,
-			 _DRM_WRITE_COMBINING, &map);
+	ret = drm_legacy_addmap(dev, dev_priv->fb_aper_offset,
+				pci_resource_len(dev->pdev, 0),
+				_DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING, &map);
 	if (ret != 0)
 		return ret;
 
@@ -2196,7 +2196,7 @@ int radeon_driver_unload(struct drm_device *dev)
 
 	DRM_DEBUG("\n");
 
-	drm_rmmap(dev, dev_priv->mmio);
+	drm_legacy_rmmap(dev, dev_priv->mmio);
 
 	kfree(dev_priv);
 
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index ae763f60c8a0..a3e7aed7e680 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -78,7 +78,8 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 	struct radeon_cs_chunk *chunk;
 	struct radeon_cs_buckets buckets;
 	unsigned i, j;
-	bool duplicate;
+	bool duplicate, need_mmap_lock = false;
+	int r;
 
 	if (p->chunk_relocs_idx == -1) {
 		return 0;
@@ -132,13 +133,17 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 		 * the buffers used for read only, which doubles the range
 		 * to 0 to 31. 32 is reserved for the kernel driver.
 		 */
-		priority = (r->flags & 0xf) * 2 + !!r->write_domain;
+		priority = (r->flags & RADEON_RELOC_PRIO_MASK) * 2
+			   + !!r->write_domain;
 
 		/* the first reloc of an UVD job is the msg and that must be in
-		   VRAM, also but everything into VRAM on AGP cards to avoid
-		   image corruptions */
+		   VRAM, also but everything into VRAM on AGP cards and older
+		   IGP chips to avoid image corruptions */
 		if (p->ring == R600_RING_TYPE_UVD_INDEX &&
-		    (i == 0 || drm_pci_device_is_agp(p->rdev->ddev))) {
+		    (i == 0 || drm_pci_device_is_agp(p->rdev->ddev) ||
+		     p->rdev->family == CHIP_RS780 ||
+		     p->rdev->family == CHIP_RS880)) {
+
 			/* TODO: is this still needed for NI+ ? */
 			p->relocs[i].prefered_domains =
 				RADEON_GEM_DOMAIN_VRAM;
@@ -164,7 +169,21 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			p->relocs[i].allowed_domains = domain;
 		}
 
+		if (radeon_ttm_tt_has_userptr(p->relocs[i].robj->tbo.ttm)) {
+			uint32_t domain = p->relocs[i].prefered_domains;
+			if (!(domain & RADEON_GEM_DOMAIN_GTT)) {
+				DRM_ERROR("Only RADEON_GEM_DOMAIN_GTT is "
+					  "allowed for userptr BOs\n");
+				return -EINVAL;
+			}
+			need_mmap_lock = true;
+			domain = RADEON_GEM_DOMAIN_GTT;
+			p->relocs[i].prefered_domains = domain;
+			p->relocs[i].allowed_domains = domain;
+		}
+
 		p->relocs[i].tv.bo = &p->relocs[i].robj->tbo;
+		p->relocs[i].tv.shared = !r->write_domain;
 		p->relocs[i].handle = r->handle;
 
 		radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head,
@@ -176,8 +195,15 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 	if (p->cs_flags & RADEON_CS_USE_VM)
 		p->vm_bos = radeon_vm_get_bos(p->rdev, p->ib.vm,
 					      &p->validated);
+	if (need_mmap_lock)
+		down_read(&current->mm->mmap_sem);
 
-	return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring);
+	r = radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring);
+
+	if (need_mmap_lock)
+		up_read(&current->mm->mmap_sem);
+
+	return r;
 }
 
 static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
@@ -223,17 +249,24 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority
 	return 0;
 }
 
-static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
+static int radeon_cs_sync_rings(struct radeon_cs_parser *p)
 {
-	int i;
+	int i, r = 0;
 
 	for (i = 0; i < p->nrelocs; i++) {
+		struct reservation_object *resv;
+
 		if (!p->relocs[i].robj)
 			continue;
 
-		radeon_semaphore_sync_to(p->ib.semaphore,
-					 p->relocs[i].robj->tbo.sync_obj);
+		resv = p->relocs[i].robj->tbo.resv;
+		r = radeon_semaphore_sync_resv(p->rdev, p->ib.semaphore, resv,
+					       p->relocs[i].tv.shared);
+
+		if (r)
+			break;
 	}
+	return r;
 }
 
 /* XXX: note that this is called from the legacy UMS CS ioctl as well */
@@ -402,7 +435,7 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 
 		ttm_eu_fence_buffer_objects(&parser->ticket,
 					    &parser->validated,
-					    parser->ib.fence);
+					    &parser->ib.fence->base);
 	} else if (backoff) {
 		ttm_eu_backoff_reservation(&parser->ticket,
 					   &parser->validated);
@@ -417,7 +450,7 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo
 	kfree(parser->track);
 	kfree(parser->relocs);
 	kfree(parser->relocs_ptr);
-	kfree(parser->vm_bos);
+	drm_free_large(parser->vm_bos);
 	for (i = 0; i < parser->nchunks; i++)
 		drm_free_large(parser->chunks[i].kdata);
 	kfree(parser->chunks);
@@ -443,14 +476,20 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev,
 		return r;
 	}
 
+	r = radeon_cs_sync_rings(parser);
+	if (r) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to sync rings: %i\n", r);
+		return r;
+	}
+
 	if (parser->ring == R600_RING_TYPE_UVD_INDEX)
 		radeon_uvd_note_usage(rdev);
 	else if ((parser->ring == TN_RING_TYPE_VCE1_INDEX) ||
 		 (parser->ring == TN_RING_TYPE_VCE2_INDEX))
 		radeon_vce_note_usage(rdev);
 
-	radeon_cs_sync_rings(parser);
-	r = radeon_ib_schedule(rdev, &parser->ib, NULL);
+	r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
 	if (r) {
 		DRM_ERROR("Failed to schedule IB !\n");
 	}
@@ -500,7 +539,8 @@ static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p,
 		if (r)
 			return r;
 	}
-	return 0;
+
+	return radeon_vm_clear_invalids(rdev, vm);
 }
 
 static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
@@ -535,14 +575,20 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
 	if (r) {
 		goto out;
 	}
-	radeon_cs_sync_rings(parser);
-	radeon_semaphore_sync_to(parser->ib.semaphore, vm->fence);
+
+	r = radeon_cs_sync_rings(parser);
+	if (r) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Failed to sync rings: %i\n", r);
+		goto out;
+	}
+	radeon_semaphore_sync_fence(parser->ib.semaphore, vm->fence);
 
 	if ((rdev->family >= CHIP_TAHITI) &&
 	    (parser->chunk_const_ib_idx != -1)) {
-		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib);
+		r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib, true);
 	} else {
-		r = radeon_ib_schedule(rdev, &parser->ib, NULL);
+		r = radeon_ib_schedule(rdev, &parser->ib, NULL, true);
 	}
 
 out:
@@ -627,6 +673,13 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		up_read(&rdev->exclusive_lock);
 		return -EBUSY;
 	}
+	if (rdev->in_reset) {
+		up_read(&rdev->exclusive_lock);
+		r = radeon_gpu_reset(rdev);
+		if (!r)
+			r = -EAGAIN;
+		return r;
+	}
 	/* initialize parser */
 	memset(&parser, 0, sizeof(struct radeon_cs_parser));
 	parser.filp = filp;
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 697add2cd4e3..995a8b1770dd 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -103,6 +103,35 @@ static const char radeon_family_name[][16] = {
 	"LAST",
 };
 
+#define RADEON_PX_QUIRK_DISABLE_PX  (1 << 0)
+#define RADEON_PX_QUIRK_LONG_WAKEUP (1 << 1)
+
+struct radeon_px_quirk {
+	u32 chip_vendor;
+	u32 chip_device;
+	u32 subsys_vendor;
+	u32 subsys_device;
+	u32 px_quirk_flags;
+};
+
+static struct radeon_px_quirk radeon_px_quirk_list[] = {
+	/* Acer aspire 5560g (CPU: AMD A4-3305M; GPU: AMD Radeon HD 6480g + 7470m)
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=74551
+	 */
+	{ PCI_VENDOR_ID_ATI, 0x6760, 0x1025, 0x0672, RADEON_PX_QUIRK_DISABLE_PX },
+	/* Asus K73TA laptop with AMD A6-3400M APU and Radeon 6550 GPU
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=51381
+	 */
+	{ PCI_VENDOR_ID_ATI, 0x6741, 0x1043, 0x108c, RADEON_PX_QUIRK_DISABLE_PX },
+	/* Asus K53TK laptop with AMD A6-3420M APU and Radeon 7670m GPU
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=51381
+	 */
+	{ PCI_VENDOR_ID_ATI, 0x6840, 0x1043, 0x2122, RADEON_PX_QUIRK_DISABLE_PX },
+	/* macbook pro 8.2 */
+	{ PCI_VENDOR_ID_ATI, 0x6741, PCI_VENDOR_ID_APPLE, 0x00e2, RADEON_PX_QUIRK_LONG_WAKEUP },
+	{ 0, 0, 0, 0, 0 },
+};
+
 bool radeon_is_px(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
@@ -112,6 +141,26 @@ bool radeon_is_px(struct drm_device *dev)
 	return false;
 }
 
+static void radeon_device_handle_px_quirks(struct radeon_device *rdev)
+{
+	struct radeon_px_quirk *p = radeon_px_quirk_list;
+
+	/* Apply PX quirks */
+	while (p && p->chip_device != 0) {
+		if (rdev->pdev->vendor == p->chip_vendor &&
+		    rdev->pdev->device == p->chip_device &&
+		    rdev->pdev->subsystem_vendor == p->subsys_vendor &&
+		    rdev->pdev->subsystem_device == p->subsys_device) {
+			rdev->px_quirk_flags = p->px_quirk_flags;
+			break;
+		}
+		++p;
+	}
+
+	if (rdev->px_quirk_flags & RADEON_PX_QUIRK_DISABLE_PX)
+		rdev->flags &= ~RADEON_IS_PX;
+}
+
 /**
  * radeon_program_register_sequence - program an array of registers.
  *
@@ -385,7 +434,8 @@ int radeon_wb_init(struct radeon_device *rdev)
 
 	if (rdev->wb.wb_obj == NULL) {
 		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL, &rdev->wb.wb_obj);
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
+				     &rdev->wb.wb_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create WB bo failed\n", r);
 			return r;
@@ -902,6 +952,7 @@ int radeon_atombios_init(struct radeon_device *rdev)
 	}
 
 	mutex_init(&rdev->mode_info.atom_context->mutex);
+	mutex_init(&rdev->mode_info.atom_context->scratch_mutex);
 	radeon_atom_initialize_bios_scratch_regs(rdev->ddev);
 	atom_allocate_fb_scratch(rdev->mode_info.atom_context);
 	return 0;
@@ -1077,7 +1128,19 @@ static void radeon_check_arguments(struct radeon_device *rdev)
 	/* defines number of bits in page table versus page directory,
 	 * a page is 4KB so we have 12 bits offset, minimum 9 bits in the
 	 * page table and the remaining bits are in the page directory */
-	if (radeon_vm_block_size < 9) {
+	if (radeon_vm_block_size == -1) {
+
+		/* Total bits covered by PD + PTs */
+		unsigned bits = ilog2(radeon_vm_size) + 18;
+
+		/* Make sure the PD is 4K in size up to 8GB address space.
+		   Above that split equal between PD and PTs */
+		if (radeon_vm_size <= 8)
+			radeon_vm_block_size = bits - 9;
+		else
+			radeon_vm_block_size = (bits + 3) / 2;
+
+	} else if (radeon_vm_block_size < 9) {
 		dev_warn(rdev->dev, "VM page table size (%d) too small\n",
 			 radeon_vm_block_size);
 		radeon_vm_block_size = 9;
@@ -1092,25 +1155,6 @@ static void radeon_check_arguments(struct radeon_device *rdev)
 }
 
 /**
- * radeon_switcheroo_quirk_long_wakeup - return true if longer d3 delay is
- * needed for waking up.
- *
- * @pdev: pci dev pointer
- */
-static bool radeon_switcheroo_quirk_long_wakeup(struct pci_dev *pdev)
-{
-
-	/* 6600m in a macbook pro */
-	if (pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE &&
-	    pdev->subsystem_device == 0x00e2) {
-		printk(KERN_INFO "radeon: quirking longer d3 wakeup delay\n");
-		return true;
-	}
-
-	return false;
-}
-
-/**
  * radeon_switcheroo_set_state - set switcheroo state
  *
  * @pdev: pci dev pointer
@@ -1122,6 +1166,7 @@ static bool radeon_switcheroo_quirk_long_wakeup(struct pci_dev *pdev)
 static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state)
 {
 	struct drm_device *dev = pci_get_drvdata(pdev);
+	struct radeon_device *rdev = dev->dev_private;
 
 	if (radeon_is_px(dev) && state == VGA_SWITCHEROO_OFF)
 		return;
@@ -1133,7 +1178,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		/* don't suspend or resume card normally */
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
-		if (d3_delay < 20 && radeon_switcheroo_quirk_long_wakeup(pdev))
+		if (d3_delay < 20 && (rdev->px_quirk_flags & RADEON_PX_QUIRK_LONG_WAKEUP))
 			dev->pdev->d3_delay = 20;
 
 		radeon_resume_kms(dev, true, true);
@@ -1213,6 +1258,7 @@ int radeon_device_init(struct radeon_device *rdev,
 	for (i = 0; i < RADEON_NUM_RINGS; i++) {
 		rdev->ring[i].idx = i;
 	}
+	rdev->fence_context = fence_context_alloc(RADEON_NUM_RINGS);
 
 	DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X).\n",
 		radeon_family_name[rdev->family], pdev->vendor, pdev->device,
@@ -1230,6 +1276,8 @@ int radeon_device_init(struct radeon_device *rdev,
 	init_rwsem(&rdev->pm.mclk_lock);
 	init_rwsem(&rdev->exclusive_lock);
 	init_waitqueue_head(&rdev->irq.vblank_queue);
+	mutex_init(&rdev->mn_lock);
+	hash_init(rdev->mn_hash);
 	r = radeon_gem_init(rdev);
 	if (r)
 		return r;
@@ -1337,6 +1385,9 @@ int radeon_device_init(struct radeon_device *rdev,
 	if (rdev->rio_mem == NULL)
 		DRM_ERROR("Unable to find PCI I/O BAR\n");
 
+	if (rdev->flags & RADEON_IS_PX)
+		radeon_device_handle_px_quirks(rdev);
+
 	/* if we have > 1 VGA cards, then disable the radeon VGA resources */
 	/* this will fail for cards that aren't VGA class devices, just
 	 * ignore it */
@@ -1350,11 +1401,7 @@ int radeon_device_init(struct radeon_device *rdev,
 
 	r = radeon_init(rdev);
 	if (r)
-		return r;
-
-	r = radeon_ib_ring_tests(rdev);
-	if (r)
-		DRM_ERROR("ib ring test failed (%d).\n", r);
+		goto failed;
 
 	r = radeon_gem_debugfs_init(rdev);
 	if (r) {
@@ -1370,9 +1417,13 @@ int radeon_device_init(struct radeon_device *rdev,
 		radeon_agp_disable(rdev);
 		r = radeon_init(rdev);
 		if (r)
-			return r;
+			goto failed;
 	}
 
+	r = radeon_ib_ring_tests(rdev);
+	if (r)
+		DRM_ERROR("ib ring test failed (%d).\n", r);
+
 	if ((radeon_testing & 1)) {
 		if (rdev->accel_working)
 			radeon_test_moves(rdev);
@@ -1392,6 +1443,11 @@ int radeon_device_init(struct radeon_device *rdev,
 			DRM_INFO("radeon: acceleration disabled, skipping benchmarks\n");
 	}
 	return 0;
+
+failed:
+	if (runtime)
+		vga_switcheroo_fini_domain_pm_ops(rdev->dev);
+	return r;
 }
 
 static void radeon_debugfs_remove_files(struct radeon_device *rdev);
@@ -1412,6 +1468,8 @@ void radeon_device_fini(struct radeon_device *rdev)
 	radeon_bo_evict_vram(rdev);
 	radeon_fini(rdev);
 	vga_switcheroo_unregister_client(rdev->pdev);
+	if (rdev->flags & RADEON_IS_PX)
+		vga_switcheroo_fini_domain_pm_ops(rdev->dev);
 	vga_client_register(rdev->pdev, NULL, NULL, NULL);
 	if (rdev->rio_mem)
 		pci_iounmap(rdev->pdev, rdev->rio_mem);
@@ -1443,7 +1501,6 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 	struct drm_crtc *crtc;
 	struct drm_connector *connector;
 	int i, r;
-	bool force_completion = false;
 
 	if (dev == NULL || dev->dev_private == NULL) {
 		return -ENODEV;
@@ -1487,12 +1544,9 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 		r = radeon_fence_wait_empty(rdev, i);
 		if (r) {
 			/* delay GPU reset to resume */
-			force_completion = true;
+			radeon_fence_driver_force_completion(rdev, i);
 		}
 	}
-	if (force_completion) {
-		radeon_fence_driver_force_completion(rdev);
-	}
 
 	radeon_save_bios_scratch_regs(rdev);
 
@@ -1632,13 +1686,11 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 		return 0;
 	}
 
-	rdev->needs_reset = false;
-
 	radeon_save_bios_scratch_regs(rdev);
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev);
-	radeon_pm_suspend(rdev);
 	radeon_suspend(rdev);
+	radeon_hpd_fini(rdev);
 
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		ring_sizes[i] = radeon_ring_backup(rdev, &rdev->ring[i],
@@ -1650,7 +1702,6 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 		}
 	}
 
-retry:
 	r = radeon_asic_reset(rdev);
 	if (!r) {
 		dev_info(rdev->dev, "GPU reset succeeded, trying to resume\n");
@@ -1659,40 +1710,69 @@ retry:
 
 	radeon_restore_bios_scratch_regs(rdev);
 
-	if (!r) {
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!r && ring_data[i]) {
 			radeon_ring_restore(rdev, &rdev->ring[i],
 					    ring_sizes[i], ring_data[i]);
-			ring_sizes[i] = 0;
-			ring_data[i] = NULL;
+		} else {
+			radeon_fence_driver_force_completion(rdev, i);
+			kfree(ring_data[i]);
 		}
+	}
 
-		r = radeon_ib_ring_tests(rdev);
+	if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+		/* do dpm late init */
+		r = radeon_pm_late_init(rdev);
 		if (r) {
-			dev_err(rdev->dev, "ib ring test failed (%d).\n", r);
-			if (saved) {
-				saved = false;
-				radeon_suspend(rdev);
-				goto retry;
-			}
+			rdev->pm.dpm_enabled = false;
+			DRM_ERROR("radeon_pm_late_init failed, disabling dpm\n");
 		}
 	} else {
-		radeon_fence_driver_force_completion(rdev);
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			kfree(ring_data[i]);
+		/* resume old pm late */
+		radeon_pm_resume(rdev);
+	}
+
+	/* init dig PHYs, disp eng pll */
+	if (rdev->is_atom_bios) {
+		radeon_atom_encoder_init(rdev);
+		radeon_atom_disp_eng_pll_init(rdev);
+		/* turn on the BL */
+		if (rdev->mode_info.bl_encoder) {
+			u8 bl_level = radeon_get_backlight_level(rdev,
+								 rdev->mode_info.bl_encoder);
+			radeon_set_backlight_level(rdev, rdev->mode_info.bl_encoder,
+						   bl_level);
 		}
 	}
+	/* reset hpd state */
+	radeon_hpd_init(rdev);
+
+	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
+
+	rdev->in_reset = true;
+	rdev->needs_reset = false;
+
+	downgrade_write(&rdev->exclusive_lock);
 
-	radeon_pm_resume(rdev);
 	drm_helper_resume_force_mode(rdev->ddev);
 
-	ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
-	if (r) {
+	/* set the power state here in case we are a PX system or headless */
+	if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		radeon_pm_compute_clocks(rdev);
+
+	if (!r) {
+		r = radeon_ib_ring_tests(rdev);
+		if (r && saved)
+			r = -EAGAIN;
+	} else {
 		/* bad news, how to tell it to userspace ? */
 		dev_info(rdev->dev, "GPU reset failed\n");
 	}
 
-	up_write(&rdev->exclusive_lock);
+	rdev->needs_reset = r == -EAGAIN;
+	rdev->in_reset = false;
+
+	up_read(&rdev->exclusive_lock);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index bf25061c8ac4..00ead8c2758a 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -293,6 +293,18 @@ void radeon_crtc_handle_vblank(struct radeon_device *rdev, int crtc_id)
 	if (radeon_crtc == NULL)
 		return;
 
+	/* Skip the pageflip completion check below (based on polling) on
+	 * asics which reliably support hw pageflip completion irqs. pflip
+	 * irqs are a reliable and race-free method of handling pageflip
+	 * completion detection. A use_pflipirq module parameter < 2 allows
+	 * to override this in case of asics with faulty pflip irqs.
+	 * A module parameter of 0 would only use this polling based path,
+	 * a parameter of 1 would use pflip irq only as a backup to this
+	 * path, as in Linux 3.16.
+	 */
+	if ((radeon_use_pflipirq == 2) && ASIC_IS_DCE4(rdev))
+		return;
+
 	spin_lock_irqsave(&rdev->ddev->event_lock, flags);
 	if (radeon_crtc->flip_status != RADEON_FLIP_SUBMITTED) {
 		DRM_DEBUG_DRIVER("radeon_crtc->flip_status = %d != "
@@ -390,12 +402,21 @@ static void radeon_flip_work_func(struct work_struct *__work)
 
         down_read(&rdev->exclusive_lock);
 	if (work->fence) {
-		r = radeon_fence_wait(work->fence, false);
-		if (r == -EDEADLK) {
-			up_read(&rdev->exclusive_lock);
-			r = radeon_gpu_reset(rdev);
-			down_read(&rdev->exclusive_lock);
-		}
+		struct radeon_fence *fence;
+
+		fence = to_radeon_fence(work->fence);
+		if (fence && fence->rdev == rdev) {
+			r = radeon_fence_wait(fence, false);
+			if (r == -EDEADLK) {
+				up_read(&rdev->exclusive_lock);
+				do {
+					r = radeon_gpu_reset(rdev);
+				} while (r == -EAGAIN);
+				down_read(&rdev->exclusive_lock);
+			}
+		} else
+			r = fence_wait(work->fence, false);
+
 		if (r)
 			DRM_ERROR("failed to wait on page flip fence (%d)!\n", r);
 
@@ -404,7 +425,8 @@ static void radeon_flip_work_func(struct work_struct *__work)
 		 * confused about which BO the CRTC is scanning out
 		 */
 
-		radeon_fence_unref(&work->fence);
+		fence_put(work->fence);
+		work->fence = NULL;
 	}
 
 	/* We borrow the event spin lock for protecting flip_status */
@@ -462,11 +484,6 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	obj = new_radeon_fb->obj;
 	new_rbo = gem_to_radeon_bo(obj);
 
-	spin_lock(&new_rbo->tbo.bdev->fence_lock);
-	if (new_rbo->tbo.sync_obj)
-		work->fence = radeon_fence_ref(new_rbo->tbo.sync_obj);
-	spin_unlock(&new_rbo->tbo.bdev->fence_lock);
-
 	/* pin the new buffer */
 	DRM_DEBUG_DRIVER("flip-ioctl() cur_rbo = %p, new_rbo = %p\n",
 			 work->old_rbo, new_rbo);
@@ -485,6 +502,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 		DRM_ERROR("failed to pin new rbo buffer before flip\n");
 		goto cleanup;
 	}
+	work->fence = fence_get(reservation_object_get_excl(new_rbo->tbo.resv));
 	radeon_bo_get_tiling_flags(new_rbo, &tiling_flags, NULL);
 	radeon_bo_unreserve(new_rbo);
 
@@ -566,9 +584,8 @@ pflip_cleanup:
 
 cleanup:
 	drm_gem_object_unreference_unlocked(&work->old_rbo->gem_base);
-	radeon_fence_unref(&work->fence);
+	fence_put(work->fence);
 	kfree(work);
-
 	return r;
 }
 
@@ -823,64 +840,6 @@ static bool radeon_setup_enc_conn(struct drm_device *dev)
 	return ret;
 }
 
-int radeon_ddc_get_modes(struct radeon_connector *radeon_connector)
-{
-	struct drm_device *dev = radeon_connector->base.dev;
-	struct radeon_device *rdev = dev->dev_private;
-	int ret = 0;
-
-	/* don't leak the edid if we already fetched it in detect() */
-	if (radeon_connector->edid)
-		goto got_edid;
-
-	/* on hw with routers, select right port */
-	if (radeon_connector->router.ddc_valid)
-		radeon_router_select_ddc_port(radeon_connector);
-
-	if (radeon_connector_encoder_get_dp_bridge_encoder_id(&radeon_connector->base) !=
-	    ENCODER_OBJECT_ID_NONE) {
-		if (radeon_connector->ddc_bus->has_aux)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->aux.ddc);
-	} else if ((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
-		   (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)) {
-		struct radeon_connector_atom_dig *dig = radeon_connector->con_priv;
-
-		if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT ||
-		     dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) &&
-		    radeon_connector->ddc_bus->has_aux)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->aux.ddc);
-		else if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-	} else {
-		if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-	}
-
-	if (!radeon_connector->edid) {
-		if (rdev->is_atom_bios) {
-			/* some laptops provide a hardcoded edid in rom for LCDs */
-			if (((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_LVDS) ||
-			     (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)))
-				radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
-		} else
-			/* some servers provide a hardcoded edid in rom for KVMs */
-			radeon_connector->edid = radeon_bios_get_hardcoded_edid(rdev);
-	}
-	if (radeon_connector->edid) {
-got_edid:
-		drm_mode_connector_update_edid_property(&radeon_connector->base, radeon_connector->edid);
-		ret = drm_add_edid_modes(&radeon_connector->base, radeon_connector->edid);
-		drm_edid_to_eld(&radeon_connector->base, radeon_connector->edid);
-		return ret;
-	}
-	drm_mode_connector_update_edid_property(&radeon_connector->base, NULL);
-	return 0;
-}
-
 /* avivo */
 
 /**
@@ -1749,7 +1708,7 @@ bool radeon_crtc_scaling_mode_fixup(struct drm_crtc *crtc,
 			    (!(mode->flags & DRM_MODE_FLAG_INTERLACE)) &&
 			    ((radeon_encoder->underscan_type == UNDERSCAN_ON) ||
 			     ((radeon_encoder->underscan_type == UNDERSCAN_AUTO) &&
-			      drm_detect_hdmi_monitor(radeon_connector->edid) &&
+			      drm_detect_hdmi_monitor(radeon_connector_edid(connector)) &&
 			      is_hdtv_mode(mode)))) {
 				if (radeon_encoder->underscan_hborder != 0)
 					radeon_crtc->h_border = radeon_encoder->underscan_hborder;
@@ -1963,7 +1922,7 @@ int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, unsigned int fl
 
 	/* In vblank? */
 	if (in_vbl)
-		ret |= DRM_SCANOUTPOS_INVBL;
+		ret |= DRM_SCANOUTPOS_IN_VBLANK;
 
 	/* Is vpos outside nominal vblank area, but less than
 	 * 1/100 of a frame height away from start of vblank?
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 959f0866d993..dcffa30ee2db 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -38,6 +38,8 @@
 #include <linux/module.h>
 #include <linux/pm_runtime.h>
 #include <linux/vga_switcheroo.h>
+#include <drm/drm_gem.h>
+
 #include "drm_crtc_helper.h"
 /*
  * KMS wrapper.
@@ -82,9 +84,11 @@
  *   2.38.0 - RADEON_GEM_OP (GET_INITIAL_DOMAIN, SET_INITIAL_DOMAIN),
  *            CIK: 1D and linear tiling modes contain valid PIPE_CONFIG
  *   2.39.0 - Add INFO query for number of active CUs
+ *   2.40.0 - Add RADEON_GEM_GTT_WC/UC, flush HDP cache before submitting
+ *            CS to GPU on >= r600
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	39
+#define KMS_DRIVER_MINOR	40
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -112,6 +116,9 @@ int radeon_gem_object_open(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
 void radeon_gem_object_close(struct drm_gem_object *obj,
 				struct drm_file *file_priv);
+struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
+					struct drm_gem_object *gobj,
+					int flags);
 extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      unsigned int flags,
 				      int *vpos, int *hpos, ktime_t *stime,
@@ -128,7 +135,7 @@ int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_mode_create_dumb *args);
 struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj);
 struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
-							size_t size,
+							struct dma_buf_attachment *,
 							struct sg_table *sg);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
 void radeon_gem_prime_unpin(struct drm_gem_object *obj);
@@ -174,9 +181,12 @@ int radeon_dpm = -1;
 int radeon_aspm = -1;
 int radeon_runtime_pm = -1;
 int radeon_hard_reset = 0;
-int radeon_vm_size = 4;
-int radeon_vm_block_size = 9;
+int radeon_vm_size = 8;
+int radeon_vm_block_size = -1;
 int radeon_deep_color = 0;
+int radeon_use_pflipirq = 2;
+int radeon_bapm = -1;
+int radeon_backlight = -1;
 
 MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers");
 module_param_named(no_wb, radeon_no_wb, int, 0444);
@@ -247,12 +257,21 @@ module_param_named(hard_reset, radeon_hard_reset, int, 0444);
 MODULE_PARM_DESC(vm_size, "VM address space size in gigabytes (default 4GB)");
 module_param_named(vm_size, radeon_vm_size, int, 0444);
 
-MODULE_PARM_DESC(vm_block_size, "VM page table size in bits (default 9)");
+MODULE_PARM_DESC(vm_block_size, "VM page table size in bits (default depending on vm_size)");
 module_param_named(vm_block_size, radeon_vm_block_size, int, 0444);
 
 MODULE_PARM_DESC(deep_color, "Deep Color support (1 = enable, 0 = disable (default))");
 module_param_named(deep_color, radeon_deep_color, int, 0444);
 
+MODULE_PARM_DESC(use_pflipirq, "Pflip irqs for pageflip completion (0 = disable, 1 = as fallback, 2 = exclusive (default))");
+module_param_named(use_pflipirq, radeon_use_pflipirq, int, 0444);
+
+MODULE_PARM_DESC(bapm, "BAPM support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(bapm, radeon_bapm, int, 0444);
+
+MODULE_PARM_DESC(backlight, "backlight support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(backlight, radeon_backlight, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 	radeon_PCI_IDS
 };
@@ -295,7 +314,7 @@ static const struct file_operations radeon_driver_old_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 	.read = drm_read,
 #ifdef CONFIG_COMPAT
@@ -315,6 +334,7 @@ static struct drm_driver driver_old = {
 	.preclose = radeon_driver_preclose,
 	.postclose = radeon_driver_postclose,
 	.lastclose = radeon_driver_lastclose,
+	.set_busid = drm_pci_set_busid,
 	.unload = radeon_driver_unload,
 	.suspend = radeon_suspend,
 	.resume = radeon_resume,
@@ -430,6 +450,7 @@ static int radeon_pmops_runtime_suspend(struct device *dev)
 	ret = radeon_suspend_kms(drm_dev, false, false);
 	pci_save_state(pdev);
 	pci_disable_device(pdev);
+	pci_ignore_hotplug(pdev);
 	pci_set_power_state(pdev, PCI_D3cold);
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
 
@@ -538,6 +559,7 @@ static struct drm_driver kms_driver = {
 	.preclose = radeon_driver_preclose_kms,
 	.postclose = radeon_driver_postclose_kms,
 	.lastclose = radeon_driver_lastclose_kms,
+	.set_busid = drm_pci_set_busid,
 	.unload = radeon_driver_unload_kms,
 	.get_vblank_counter = radeon_get_vblank_counter_kms,
 	.enable_vblank = radeon_enable_vblank_kms,
@@ -563,7 +585,7 @@ static struct drm_driver kms_driver = {
 
 	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
 	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
-	.gem_prime_export = drm_gem_prime_export,
+	.gem_prime_export = radeon_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = radeon_gem_prime_pin,
 	.gem_prime_unpin = radeon_gem_prime_unpin,
diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h
index dafd812e4571..46bd3938282c 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.h
+++ b/drivers/gpu/drm/radeon/radeon_drv.h
@@ -33,7 +33,9 @@
 
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
+#include <drm/drm_legacy.h>
 
+#include <drm/ati_pcigart.h>
 #include "radeon_family.h"
 
 /* General customization:
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index bd4959ca23aa..9a19e52cc655 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -158,10 +158,43 @@ radeon_get_encoder_enum(struct drm_device *dev, uint32_t supported_device, uint8
 	return ret;
 }
 
+static void radeon_encoder_add_backlight(struct radeon_encoder *radeon_encoder,
+					 struct drm_connector *connector)
+{
+	struct drm_device *dev = radeon_encoder->base.dev;
+	struct radeon_device *rdev = dev->dev_private;
+	bool use_bl = false;
+
+	if (!(radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)))
+		return;
+
+	if (radeon_backlight == 0) {
+		return;
+	} else if (radeon_backlight == 1) {
+		use_bl = true;
+	} else if (radeon_backlight == -1) {
+		/* Quirks */
+		/* Amilo Xi 2550 only works with acpi bl */
+		if ((rdev->pdev->device == 0x9583) &&
+		    (rdev->pdev->subsystem_vendor == 0x1734) &&
+		    (rdev->pdev->subsystem_device == 0x1107))
+			use_bl = false;
+		else
+			use_bl = true;
+	}
+
+	if (use_bl) {
+		if (rdev->is_atom_bios)
+			radeon_atom_backlight_init(radeon_encoder, connector);
+		else
+			radeon_legacy_backlight_init(radeon_encoder, connector);
+		rdev->mode_info.bl_encoder = radeon_encoder;
+	}
+}
+
 void
 radeon_link_encoder_connector(struct drm_device *dev)
 {
-	struct radeon_device *rdev = dev->dev_private;
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector;
 	struct drm_encoder *encoder;
@@ -174,13 +207,8 @@ radeon_link_encoder_connector(struct drm_device *dev)
 			radeon_encoder = to_radeon_encoder(encoder);
 			if (radeon_encoder->devices & radeon_connector->devices) {
 				drm_mode_connector_attach_encoder(connector, encoder);
-				if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-					if (rdev->is_atom_bios)
-						radeon_atom_backlight_init(radeon_encoder, connector);
-					else
-						radeon_legacy_backlight_init(radeon_encoder, connector);
-					rdev->mode_info.bl_encoder = radeon_encoder;
-				}
+				if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+					radeon_encoder_add_backlight(radeon_encoder, connector);
 			}
 		}
 	}
@@ -343,7 +371,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	case DRM_MODE_CONNECTOR_HDMIB:
 		if (radeon_connector->use_digital) {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (pixel_clock > 340000)
 					return true;
 				else
@@ -365,7 +393,7 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 			return false;
 		else {
 			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) {
+			if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector_edid(connector))) {
 				if (pixel_clock > 340000)
 					return true;
 				else
@@ -382,3 +410,24 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 	}
 }
 
+bool radeon_encoder_is_digital(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+		return true;
+	default:
+		return false;
+	}
+}
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 665ced3b7313..0ea1db83d573 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -127,8 +127,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	aligned_size = ALIGN(size, PAGE_SIZE);
 	ret = radeon_gem_object_create(rdev, aligned_size, 0,
 				       RADEON_GEM_DOMAIN_VRAM,
-				       false, true,
-				       &gobj);
+				       0, true, &gobj);
 	if (ret) {
 		printk(KERN_ERR "failed to allocate framebuffer (%d)\n",
 		       aligned_size);
@@ -190,7 +189,8 @@ out_unref:
 static int radeonfb_create(struct drm_fb_helper *helper,
 			   struct drm_fb_helper_surface_size *sizes)
 {
-	struct radeon_fbdev *rfbdev = (struct radeon_fbdev *)helper;
+	struct radeon_fbdev *rfbdev =
+		container_of(helper, struct radeon_fbdev, helper);
 	struct radeon_device *rdev = rfbdev->rdev;
 	struct fb_info *info;
 	struct drm_framebuffer *fb = NULL;
@@ -331,7 +331,7 @@ static int radeon_fbdev_destroy(struct drm_device *dev, struct radeon_fbdev *rfb
 	return 0;
 }
 
-static struct drm_fb_helper_funcs radeon_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs radeon_fb_helper_funcs = {
 	.gamma_set = radeon_crtc_fb_gamma_set,
 	.gamma_get = radeon_crtc_fb_gamma_get,
 	.fb_probe = radeonfb_create,
@@ -353,7 +353,9 @@ int radeon_fbdev_init(struct radeon_device *rdev)
 
 	rfbdev->rdev = rdev;
 	rdev->mode_info.rfbdev = rfbdev;
-	rfbdev->helper.funcs = &radeon_fb_helper_funcs;
+
+	drm_fb_helper_prepare(rdev->ddev, &rfbdev->helper,
+			      &radeon_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(rdev->ddev, &rfbdev->helper,
 				 rdev->num_crtc,
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 913787085dfa..995167025282 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -98,6 +98,25 @@ static u32 radeon_fence_read(struct radeon_device *rdev, int ring)
 }
 
 /**
+ * radeon_fence_schedule_check - schedule lockup check
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index we should work with
+ *
+ * Queues a delayed work item to check for lockups.
+ */
+static void radeon_fence_schedule_check(struct radeon_device *rdev, int ring)
+{
+	/*
+	 * Do not reset the timer here with mod_delayed_work,
+	 * this can livelock in an interaction with TTM delayed destroy.
+	 */
+	queue_delayed_work(system_power_efficient_wq,
+			   &rdev->fence_drv[ring].lockup_work,
+			   RADEON_FENCE_JIFFIES_TIMEOUT);
+}
+
+/**
  * radeon_fence_emit - emit a fence on the requested ring
  *
  * @rdev: radeon_device pointer
@@ -111,30 +130,70 @@ int radeon_fence_emit(struct radeon_device *rdev,
 		      struct radeon_fence **fence,
 		      int ring)
 {
+	u64 seq = ++rdev->fence_drv[ring].sync_seq[ring];
+
 	/* we are protected by the ring emission mutex */
 	*fence = kmalloc(sizeof(struct radeon_fence), GFP_KERNEL);
 	if ((*fence) == NULL) {
 		return -ENOMEM;
 	}
-	kref_init(&((*fence)->kref));
 	(*fence)->rdev = rdev;
-	(*fence)->seq = ++rdev->fence_drv[ring].sync_seq[ring];
+	(*fence)->seq = seq;
 	(*fence)->ring = ring;
+	fence_init(&(*fence)->base, &radeon_fence_ops,
+		   &rdev->fence_queue.lock, rdev->fence_context + ring, seq);
 	radeon_fence_ring_emit(rdev, ring, *fence);
 	trace_radeon_fence_emit(rdev->ddev, ring, (*fence)->seq);
+	radeon_fence_schedule_check(rdev, ring);
 	return 0;
 }
 
 /**
- * radeon_fence_process - process a fence
+ * radeon_fence_check_signaled - callback from fence_queue
+ *
+ * this function is called with fence_queue lock held, which is also used
+ * for the fence locking itself, so unlocked variants are used for
+ * fence_signal, and remove_wait_queue.
+ */
+static int radeon_fence_check_signaled(wait_queue_t *wait, unsigned mode, int flags, void *key)
+{
+	struct radeon_fence *fence;
+	u64 seq;
+
+	fence = container_of(wait, struct radeon_fence, fence_wake);
+
+	/*
+	 * We cannot use radeon_fence_process here because we're already
+	 * in the waitqueue, in a call from wake_up_all.
+	 */
+	seq = atomic64_read(&fence->rdev->fence_drv[fence->ring].last_seq);
+	if (seq >= fence->seq) {
+		int ret = fence_signal_locked(&fence->base);
+
+		if (!ret)
+			FENCE_TRACE(&fence->base, "signaled from irq context\n");
+		else
+			FENCE_TRACE(&fence->base, "was already signaled\n");
+
+		radeon_irq_kms_sw_irq_put(fence->rdev, fence->ring);
+		__remove_wait_queue(&fence->rdev->fence_queue, &fence->fence_wake);
+		fence_put(&fence->base);
+	} else
+		FENCE_TRACE(&fence->base, "pending\n");
+	return 0;
+}
+
+/**
+ * radeon_fence_activity - check for fence activity
  *
  * @rdev: radeon_device pointer
  * @ring: ring index the fence is associated with
  *
- * Checks the current fence value and wakes the fence queue
- * if the sequence number has increased (all asics).
+ * Checks the current fence value and calculates the last
+ * signalled fence value. Returns true if activity occured
+ * on the ring, and the fence_queue should be waken up.
  */
-void radeon_fence_process(struct radeon_device *rdev, int ring)
+static bool radeon_fence_activity(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq, last_seq, last_emitted;
 	unsigned count_loop = 0;
@@ -190,23 +249,77 @@ void radeon_fence_process(struct radeon_device *rdev, int ring)
 		}
 	} while (atomic64_xchg(&rdev->fence_drv[ring].last_seq, seq) > seq);
 
-	if (wake)
-		wake_up_all(&rdev->fence_queue);
+	if (seq < last_emitted)
+		radeon_fence_schedule_check(rdev, ring);
+
+	return wake;
 }
 
 /**
- * radeon_fence_destroy - destroy a fence
+ * radeon_fence_check_lockup - check for hardware lockup
  *
- * @kref: fence kref
+ * @work: delayed work item
  *
- * Frees the fence object (all asics).
+ * Checks for fence activity and if there is none probe
+ * the hardware if a lockup occured.
  */
-static void radeon_fence_destroy(struct kref *kref)
+static void radeon_fence_check_lockup(struct work_struct *work)
 {
-	struct radeon_fence *fence;
+	struct radeon_fence_driver *fence_drv;
+	struct radeon_device *rdev;
+	int ring;
+
+	fence_drv = container_of(work, struct radeon_fence_driver,
+				 lockup_work.work);
+	rdev = fence_drv->rdev;
+	ring = fence_drv - &rdev->fence_drv[0];
+
+	if (!down_read_trylock(&rdev->exclusive_lock)) {
+		/* just reschedule the check if a reset is going on */
+		radeon_fence_schedule_check(rdev, ring);
+		return;
+	}
+
+	if (fence_drv->delayed_irq && rdev->ddev->irq_enabled) {
+		unsigned long irqflags;
+
+		fence_drv->delayed_irq = false;
+		spin_lock_irqsave(&rdev->irq.lock, irqflags);
+		radeon_irq_set(rdev);
+		spin_unlock_irqrestore(&rdev->irq.lock, irqflags);
+	}
+
+	if (radeon_fence_activity(rdev, ring))
+		wake_up_all(&rdev->fence_queue);
 
-	fence = container_of(kref, struct radeon_fence, kref);
-	kfree(fence);
+	else if (radeon_ring_is_lockup(rdev, ring, &rdev->ring[ring])) {
+
+		/* good news we believe it's a lockup */
+		dev_warn(rdev->dev, "GPU lockup (current fence id "
+			 "0x%016llx last fence id 0x%016llx on ring %d)\n",
+			 (uint64_t)atomic64_read(&fence_drv->last_seq),
+			 fence_drv->sync_seq[ring], ring);
+
+		/* remember that we need an reset */
+		rdev->needs_reset = true;
+		wake_up_all(&rdev->fence_queue);
+	}
+	up_read(&rdev->exclusive_lock);
+}
+
+/**
+ * radeon_fence_process - process a fence
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index the fence is associated with
+ *
+ * Checks the current fence value and wakes the fence queue
+ * if the sequence number has increased (all asics).
+ */
+void radeon_fence_process(struct radeon_device *rdev, int ring)
+{
+	if (radeon_fence_activity(rdev, ring))
+		wake_up_all(&rdev->fence_queue);
 }
 
 /**
@@ -237,6 +350,75 @@ static bool radeon_fence_seq_signaled(struct radeon_device *rdev,
 	return false;
 }
 
+static bool radeon_fence_is_signaled(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+	unsigned ring = fence->ring;
+	u64 seq = fence->seq;
+
+	if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq) {
+		return true;
+	}
+
+	if (down_read_trylock(&rdev->exclusive_lock)) {
+		radeon_fence_process(rdev, ring);
+		up_read(&rdev->exclusive_lock);
+
+		if (atomic64_read(&rdev->fence_drv[ring].last_seq) >= seq) {
+			return true;
+		}
+	}
+	return false;
+}
+
+/**
+ * radeon_fence_enable_signaling - enable signalling on fence
+ * @fence: fence
+ *
+ * This function is called with fence_queue lock held, and adds a callback
+ * to fence_queue that checks if this fence is signaled, and if so it
+ * signals the fence and removes itself.
+ */
+static bool radeon_fence_enable_signaling(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+
+	if (atomic64_read(&rdev->fence_drv[fence->ring].last_seq) >= fence->seq)
+		return false;
+
+	if (down_read_trylock(&rdev->exclusive_lock)) {
+		radeon_irq_kms_sw_irq_get(rdev, fence->ring);
+
+		if (radeon_fence_activity(rdev, fence->ring))
+			wake_up_all_locked(&rdev->fence_queue);
+
+		/* did fence get signaled after we enabled the sw irq? */
+		if (atomic64_read(&rdev->fence_drv[fence->ring].last_seq) >= fence->seq) {
+			radeon_irq_kms_sw_irq_put(rdev, fence->ring);
+			up_read(&rdev->exclusive_lock);
+			return false;
+		}
+
+		up_read(&rdev->exclusive_lock);
+	} else {
+		/* we're probably in a lockup, lets not fiddle too much */
+		if (radeon_irq_kms_sw_irq_get_delayed(rdev, fence->ring))
+			rdev->fence_drv[fence->ring].delayed_irq = true;
+		radeon_fence_schedule_check(rdev, fence->ring);
+	}
+
+	fence->fence_wake.flags = 0;
+	fence->fence_wake.private = NULL;
+	fence->fence_wake.func = radeon_fence_check_signaled;
+	__add_wait_queue(&rdev->fence_queue, &fence->fence_wake);
+	fence_get(f);
+
+	FENCE_TRACE(&fence->base, "armed on ring %i!\n", fence->ring);
+	return true;
+}
+
 /**
  * radeon_fence_signaled - check if a fence has signaled
  *
@@ -247,14 +429,15 @@ static bool radeon_fence_seq_signaled(struct radeon_device *rdev,
  */
 bool radeon_fence_signaled(struct radeon_fence *fence)
 {
-	if (!fence) {
+	if (!fence)
 		return true;
-	}
-	if (fence->seq == RADEON_FENCE_SIGNALED_SEQ) {
-		return true;
-	}
+
 	if (radeon_fence_seq_signaled(fence->rdev, fence->seq, fence->ring)) {
-		fence->seq = RADEON_FENCE_SIGNALED_SEQ;
+		int ret;
+
+		ret = fence_signal(&fence->base);
+		if (!ret)
+			FENCE_TRACE(&fence->base, "signaled from radeon_fence_signaled\n");
 		return true;
 	}
 	return false;
@@ -283,110 +466,70 @@ static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq)
 }
 
 /**
- * radeon_fence_wait_seq - wait for a specific sequence numbers
+ * radeon_fence_wait_seq_timeout - wait for a specific sequence numbers
  *
  * @rdev: radeon device pointer
  * @target_seq: sequence number(s) we want to wait for
  * @intr: use interruptable sleep
+ * @timeout: maximum time to wait, or MAX_SCHEDULE_TIMEOUT for infinite wait
  *
  * Wait for the requested sequence number(s) to be written by any ring
  * (all asics).  Sequnce number array is indexed by ring id.
  * @intr selects whether to use interruptable (true) or non-interruptable
  * (false) sleep when waiting for the sequence number.  Helper function
  * for radeon_fence_wait_*().
- * Returns 0 if the sequence number has passed, error for all other cases.
+ * Returns remaining time if the sequence number has passed, 0 when
+ * the wait timeout, or an error for all other cases.
  * -EDEADLK is returned when a GPU lockup has been detected.
  */
-static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq,
-				 bool intr)
+static long radeon_fence_wait_seq_timeout(struct radeon_device *rdev,
+					  u64 *target_seq, bool intr,
+					  long timeout)
 {
-	uint64_t last_seq[RADEON_NUM_RINGS];
-	bool signaled;
-	int i, r;
-
-	while (!radeon_fence_any_seq_signaled(rdev, target_seq)) {
+	long r;
+	int i;
 
-		/* Save current sequence values, used to check for GPU lockups */
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			if (!target_seq[i])
-				continue;
+	if (radeon_fence_any_seq_signaled(rdev, target_seq))
+		return timeout;
 
-			last_seq[i] = atomic64_read(&rdev->fence_drv[i].last_seq);
-			trace_radeon_fence_wait_begin(rdev->ddev, i, target_seq[i]);
-			radeon_irq_kms_sw_irq_get(rdev, i);
-		}
+	/* enable IRQs and tracing */
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!target_seq[i])
+			continue;
 
-		if (intr) {
-			r = wait_event_interruptible_timeout(rdev->fence_queue, (
-				(signaled = radeon_fence_any_seq_signaled(rdev, target_seq))
-				 || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT);
-		} else {
-			r = wait_event_timeout(rdev->fence_queue, (
-				(signaled = radeon_fence_any_seq_signaled(rdev, target_seq))
-				 || rdev->needs_reset), RADEON_FENCE_JIFFIES_TIMEOUT);
-		}
+		trace_radeon_fence_wait_begin(rdev->ddev, i, target_seq[i]);
+		radeon_irq_kms_sw_irq_get(rdev, i);
+	}
 
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			if (!target_seq[i])
-				continue;
+	if (intr) {
+		r = wait_event_interruptible_timeout(rdev->fence_queue, (
+			radeon_fence_any_seq_signaled(rdev, target_seq)
+			 || rdev->needs_reset), timeout);
+	} else {
+		r = wait_event_timeout(rdev->fence_queue, (
+			radeon_fence_any_seq_signaled(rdev, target_seq)
+			 || rdev->needs_reset), timeout);
+	}
 
-			radeon_irq_kms_sw_irq_put(rdev, i);
-			trace_radeon_fence_wait_end(rdev->ddev, i, target_seq[i]);
-		}
+	if (rdev->needs_reset)
+		r = -EDEADLK;
 
-		if (unlikely(r < 0))
-			return r;
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		if (!target_seq[i])
+			continue;
 
-		if (unlikely(!signaled)) {
-			if (rdev->needs_reset)
-				return -EDEADLK;
-
-			/* we were interrupted for some reason and fence
-			 * isn't signaled yet, resume waiting */
-			if (r)
-				continue;
-
-			for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-				if (!target_seq[i])
-					continue;
-
-				if (last_seq[i] != atomic64_read(&rdev->fence_drv[i].last_seq))
-					break;
-			}
-
-			if (i != RADEON_NUM_RINGS)
-				continue;
-
-			for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-				if (!target_seq[i])
-					continue;
-
-				if (radeon_ring_is_lockup(rdev, i, &rdev->ring[i]))
-					break;
-			}
-
-			if (i < RADEON_NUM_RINGS) {
-				/* good news we believe it's a lockup */
-				dev_warn(rdev->dev, "GPU lockup (waiting for "
-					 "0x%016llx last fence id 0x%016llx on"
-					 " ring %d)\n",
-					 target_seq[i], last_seq[i], i);
-
-				/* remember that we need an reset */
-				rdev->needs_reset = true;
-				wake_up_all(&rdev->fence_queue);
-				return -EDEADLK;
-			}
-		}
+		radeon_irq_kms_sw_irq_put(rdev, i);
+		trace_radeon_fence_wait_end(rdev->ddev, i, target_seq[i]);
 	}
-	return 0;
+
+	return r;
 }
 
 /**
  * radeon_fence_wait - wait for a fence to signal
  *
  * @fence: radeon fence object
- * @intr: use interruptable sleep
+ * @intr: use interruptible sleep
  *
  * Wait for the requested fence to signal (all asics).
  * @intr selects whether to use interruptable (true) or non-interruptable
@@ -396,22 +539,26 @@ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq,
 int radeon_fence_wait(struct radeon_fence *fence, bool intr)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
-	int r;
+	long r;
 
-	if (fence == NULL) {
-		WARN(1, "Querying an invalid fence : %p !\n", fence);
-		return -EINVAL;
-	}
+	/*
+	 * This function should not be called on !radeon fences.
+	 * If this is the case, it would mean this function can
+	 * also be called on radeon fences belonging to another card.
+	 * exclusive_lock is not held in that case.
+	 */
+	if (WARN_ON_ONCE(!to_radeon_fence(&fence->base)))
+		return fence_wait(&fence->base, intr);
 
 	seq[fence->ring] = fence->seq;
-	if (seq[fence->ring] == RADEON_FENCE_SIGNALED_SEQ)
-		return 0;
-
-	r = radeon_fence_wait_seq(fence->rdev, seq, intr);
-	if (r)
+	r = radeon_fence_wait_seq_timeout(fence->rdev, seq, intr, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		return r;
+	}
 
-	fence->seq = RADEON_FENCE_SIGNALED_SEQ;
+	r = fence_signal(&fence->base);
+	if (!r)
+		FENCE_TRACE(&fence->base, "signaled from fence_wait\n");
 	return 0;
 }
 
@@ -434,7 +581,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 {
 	uint64_t seq[RADEON_NUM_RINGS];
 	unsigned i, num_rings = 0;
-	int r;
+	long r;
 
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		seq[i] = 0;
@@ -445,18 +592,14 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 
 		seq[i] = fences[i]->seq;
 		++num_rings;
-
-		/* test if something was allready signaled */
-		if (seq[i] == RADEON_FENCE_SIGNALED_SEQ)
-			return 0;
 	}
 
 	/* nothing to wait for ? */
 	if (num_rings == 0)
 		return -ENOENT;
 
-	r = radeon_fence_wait_seq(rdev, seq, intr);
-	if (r) {
+	r = radeon_fence_wait_seq_timeout(rdev, seq, intr, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		return r;
 	}
 	return 0;
@@ -475,6 +618,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev,
 int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
+	long r;
 
 	seq[ring] = atomic64_read(&rdev->fence_drv[ring].last_seq) + 1ULL;
 	if (seq[ring] >= rdev->fence_drv[ring].sync_seq[ring]) {
@@ -482,7 +626,10 @@ int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 		   already the last emited fence */
 		return -ENOENT;
 	}
-	return radeon_fence_wait_seq(rdev, seq, false);
+	r = radeon_fence_wait_seq_timeout(rdev, seq, false, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0)
+		return r;
+	return 0;
 }
 
 /**
@@ -498,18 +645,18 @@ int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 int radeon_fence_wait_empty(struct radeon_device *rdev, int ring)
 {
 	uint64_t seq[RADEON_NUM_RINGS] = {};
-	int r;
+	long r;
 
 	seq[ring] = rdev->fence_drv[ring].sync_seq[ring];
 	if (!seq[ring])
 		return 0;
 
-	r = radeon_fence_wait_seq(rdev, seq, false);
-	if (r) {
+	r = radeon_fence_wait_seq_timeout(rdev, seq, false, MAX_SCHEDULE_TIMEOUT);
+	if (r < 0) {
 		if (r == -EDEADLK)
 			return -EDEADLK;
 
-		dev_err(rdev->dev, "error waiting for ring[%d] to become idle (%d)\n",
+		dev_err(rdev->dev, "error waiting for ring[%d] to become idle (%ld)\n",
 			ring, r);
 	}
 	return 0;
@@ -525,7 +672,7 @@ int radeon_fence_wait_empty(struct radeon_device *rdev, int ring)
  */
 struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence)
 {
-	kref_get(&fence->kref);
+	fence_get(&fence->base);
 	return fence;
 }
 
@@ -542,7 +689,7 @@ void radeon_fence_unref(struct radeon_fence **fence)
 
 	*fence = NULL;
 	if (tmp) {
-		kref_put(&tmp->kref, radeon_fence_destroy);
+		fence_put(&tmp->base);
 	}
 }
 
@@ -711,6 +858,9 @@ static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring)
 		rdev->fence_drv[ring].sync_seq[i] = 0;
 	atomic64_set(&rdev->fence_drv[ring].last_seq, 0);
 	rdev->fence_drv[ring].initialized = false;
+	INIT_DELAYED_WORK(&rdev->fence_drv[ring].lockup_work,
+			  radeon_fence_check_lockup);
+	rdev->fence_drv[ring].rdev = rdev;
 }
 
 /**
@@ -758,8 +908,9 @@ void radeon_fence_driver_fini(struct radeon_device *rdev)
 		r = radeon_fence_wait_empty(rdev, ring);
 		if (r) {
 			/* no need to trigger GPU reset as we are unloading */
-			radeon_fence_driver_force_completion(rdev);
+			radeon_fence_driver_force_completion(rdev, ring);
 		}
+		cancel_delayed_work_sync(&rdev->fence_drv[ring].lockup_work);
 		wake_up_all(&rdev->fence_queue);
 		radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg);
 		rdev->fence_drv[ring].initialized = false;
@@ -771,18 +922,16 @@ void radeon_fence_driver_fini(struct radeon_device *rdev)
  * radeon_fence_driver_force_completion - force all fence waiter to complete
  *
  * @rdev: radeon device pointer
+ * @ring: the ring to complete
  *
  * In case of GPU reset failure make sure no process keep waiting on fence
  * that will never complete.
  */
-void radeon_fence_driver_force_completion(struct radeon_device *rdev)
+void radeon_fence_driver_force_completion(struct radeon_device *rdev, int ring)
 {
-	int ring;
-
-	for (ring = 0; ring < RADEON_NUM_RINGS; ring++) {
-		if (!rdev->fence_drv[ring].initialized)
-			continue;
+	if (rdev->fence_drv[ring].initialized) {
 		radeon_fence_write(rdev, rdev->fence_drv[ring].sync_seq[ring], ring);
+		cancel_delayed_work_sync(&rdev->fence_drv[ring].lockup_work);
 	}
 }
 
@@ -833,6 +982,7 @@ static int radeon_debugfs_gpu_reset(struct seq_file *m, void *data)
 	down_read(&rdev->exclusive_lock);
 	seq_printf(m, "%d\n", rdev->needs_reset);
 	rdev->needs_reset = true;
+	wake_up_all(&rdev->fence_queue);
 	up_read(&rdev->exclusive_lock);
 
 	return 0;
@@ -852,3 +1002,72 @@ int radeon_debugfs_fence_init(struct radeon_device *rdev)
 	return 0;
 #endif
 }
+
+static const char *radeon_fence_get_driver_name(struct fence *fence)
+{
+	return "radeon";
+}
+
+static const char *radeon_fence_get_timeline_name(struct fence *f)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	switch (fence->ring) {
+	case RADEON_RING_TYPE_GFX_INDEX: return "radeon.gfx";
+	case CAYMAN_RING_TYPE_CP1_INDEX: return "radeon.cp1";
+	case CAYMAN_RING_TYPE_CP2_INDEX: return "radeon.cp2";
+	case R600_RING_TYPE_DMA_INDEX: return "radeon.dma";
+	case CAYMAN_RING_TYPE_DMA1_INDEX: return "radeon.dma1";
+	case R600_RING_TYPE_UVD_INDEX: return "radeon.uvd";
+	case TN_RING_TYPE_VCE1_INDEX: return "radeon.vce1";
+	case TN_RING_TYPE_VCE2_INDEX: return "radeon.vce2";
+	default: WARN_ON_ONCE(1); return "radeon.unk";
+	}
+}
+
+static inline bool radeon_test_signaled(struct radeon_fence *fence)
+{
+	return test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags);
+}
+
+static signed long radeon_fence_default_wait(struct fence *f, bool intr,
+					     signed long t)
+{
+	struct radeon_fence *fence = to_radeon_fence(f);
+	struct radeon_device *rdev = fence->rdev;
+	bool signaled;
+
+	fence_enable_sw_signaling(&fence->base);
+
+	/*
+	 * This function has to return -EDEADLK, but cannot hold
+	 * exclusive_lock during the wait because some callers
+	 * may already hold it. This means checking needs_reset without
+	 * lock, and not fiddling with any gpu internals.
+	 *
+	 * The callback installed with fence_enable_sw_signaling will
+	 * run before our wait_event_*timeout call, so we will see
+	 * both the signaled fence and the changes to needs_reset.
+	 */
+
+	if (intr)
+		t = wait_event_interruptible_timeout(rdev->fence_queue,
+			((signaled = radeon_test_signaled(fence)) ||
+			 rdev->needs_reset), t);
+	else
+		t = wait_event_timeout(rdev->fence_queue,
+			((signaled = radeon_test_signaled(fence)) ||
+			 rdev->needs_reset), t);
+
+	if (t > 0 && !signaled)
+		return -EDEADLK;
+	return t;
+}
+
+const struct fence_ops radeon_fence_ops = {
+	.get_driver_name = radeon_fence_get_driver_name,
+	.get_timeline_name = radeon_fence_get_timeline_name,
+	.enable_signaling = radeon_fence_enable_signaling,
+	.signaled = radeon_fence_is_signaled,
+	.wait = radeon_fence_default_wait,
+	.release = NULL,
+};
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 2e723651069b..84146d5901aa 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -128,7 +128,7 @@ int radeon_gart_table_vram_alloc(struct radeon_device *rdev)
 	if (rdev->gart.robj == NULL) {
 		r = radeon_bo_create(rdev, rdev->gart.table_size,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     NULL, &rdev->gart.robj);
+				     0, NULL, NULL, &rdev->gart.robj);
 		if (r) {
 			return r;
 		}
@@ -243,7 +243,8 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
 				if (rdev->gart.ptr) {
-					radeon_gart_set_page(rdev, t, page_base);
+					radeon_gart_set_page(rdev, t, page_base,
+							     RADEON_GART_PAGE_DUMMY);
 				}
 				page_base += RADEON_GPU_PAGE_SIZE;
 			}
@@ -261,13 +262,15 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
  * @pages: number of pages to bind
  * @pagelist: pages to bind
  * @dma_addr: DMA addresses of pages
+ * @flags: RADEON_GART_PAGE_* flags
  *
  * Binds the requested pages to the gart page table
  * (all asics).
  * Returns 0 for success, -EINVAL for failure.
  */
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
-		     int pages, struct page **pagelist, dma_addr_t *dma_addr)
+		     int pages, struct page **pagelist, dma_addr_t *dma_addr,
+		     uint32_t flags)
 {
 	unsigned t;
 	unsigned p;
@@ -287,7 +290,7 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 		if (rdev->gart.ptr) {
 			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-				radeon_gart_set_page(rdev, t, page_base);
+				radeon_gart_set_page(rdev, t, page_base, flags);
 				page_base += RADEON_GPU_PAGE_SIZE;
 			}
 		}
@@ -298,33 +301,6 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 }
 
 /**
- * radeon_gart_restore - bind all pages in the gart page table
- *
- * @rdev: radeon_device pointer
- *
- * Binds all pages in the gart page table (all asics).
- * Used to rebuild the gart table on device startup or resume.
- */
-void radeon_gart_restore(struct radeon_device *rdev)
-{
-	int i, j, t;
-	u64 page_base;
-
-	if (!rdev->gart.ptr) {
-		return;
-	}
-	for (i = 0, t = 0; i < rdev->gart.num_cpu_pages; i++) {
-		page_base = rdev->gart.pages_addr[i];
-		for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-			radeon_gart_set_page(rdev, t, page_base);
-			page_base += RADEON_GPU_PAGE_SIZE;
-		}
-	}
-	mb();
-	radeon_gart_tlb_flush(rdev);
-}
-
-/**
  * radeon_gart_init - init the driver info for managing the gart
  *
  * @rdev: radeon_device pointer
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index d09650c1d720..c194497aa586 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -40,9 +40,9 @@ void radeon_gem_object_free(struct drm_gem_object *gobj)
 	}
 }
 
-int radeon_gem_object_create(struct radeon_device *rdev, int size,
+int radeon_gem_object_create(struct radeon_device *rdev, unsigned long size,
 				int alignment, int initial_domain,
-				bool discardable, bool kernel,
+				u32 flags, bool kernel,
 				struct drm_gem_object **obj)
 {
 	struct radeon_bo *robj;
@@ -55,23 +55,26 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 		alignment = PAGE_SIZE;
 	}
 
-	/* maximun bo size is the minimun btw visible vram and gtt size */
-	max_size = min(rdev->mc.visible_vram_size, rdev->mc.gtt_size);
+	/* Maximum bo size is the unpinned gtt size since we use the gtt to
+	 * handle vram to system pool migrations.
+	 */
+	max_size = rdev->mc.gtt_size - rdev->gart_pin_size;
 	if (size > max_size) {
-		printk(KERN_WARNING "%s:%d alloc size %dMb bigger than %ldMb limit\n",
-		       __func__, __LINE__, size >> 20, max_size >> 20);
+		DRM_DEBUG("Allocation size %ldMb bigger than %ldMb limit\n",
+			  size >> 20, max_size >> 20);
 		return -ENOMEM;
 	}
 
 retry:
-	r = radeon_bo_create(rdev, size, alignment, kernel, initial_domain, NULL, &robj);
+	r = radeon_bo_create(rdev, size, alignment, kernel, initial_domain,
+			     flags, NULL, NULL, &robj);
 	if (r) {
 		if (r != -ERESTARTSYS) {
 			if (initial_domain == RADEON_GEM_DOMAIN_VRAM) {
 				initial_domain |= RADEON_GEM_DOMAIN_GTT;
 				goto retry;
 			}
-			DRM_ERROR("Failed to allocate GEM object (%d, %d, %u, %d)\n",
+			DRM_ERROR("Failed to allocate GEM object (%ld, %d, %u, %d)\n",
 				  size, initial_domain, alignment, r);
 		}
 		return r;
@@ -91,7 +94,7 @@ static int radeon_gem_set_domain(struct drm_gem_object *gobj,
 {
 	struct radeon_bo *robj;
 	uint32_t domain;
-	int r;
+	long r;
 
 	/* FIXME: reeimplement */
 	robj = gem_to_radeon_bo(gobj);
@@ -107,9 +110,12 @@ static int radeon_gem_set_domain(struct drm_gem_object *gobj,
 	}
 	if (domain == RADEON_GEM_DOMAIN_CPU) {
 		/* Asking for cpu access wait for object idle */
-		r = radeon_bo_wait(robj, NULL, false);
-		if (r) {
-			printk(KERN_ERR "Failed to wait for object !\n");
+		r = reservation_object_wait_timeout_rcu(robj->tbo.resv, true, true, 30 * HZ);
+		if (!r)
+			r = -EBUSY;
+
+		if (r < 0 && r != -EINTR) {
+			printk(KERN_ERR "Failed to wait for object: %li\n", r);
 			return r;
 		}
 	}
@@ -208,18 +214,15 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 	struct radeon_device *rdev = dev->dev_private;
 	struct drm_radeon_gem_info *args = data;
 	struct ttm_mem_type_manager *man;
-	unsigned i;
 
 	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
 
 	args->vram_size = rdev->mc.real_vram_size;
 	args->vram_visible = (u64)man->size << PAGE_SHIFT;
-	if (rdev->stollen_vga_memory)
-		args->vram_visible -= radeon_bo_size(rdev->stollen_vga_memory);
-	args->vram_visible -= radeon_fbdev_total_size(rdev);
-	args->gart_size = rdev->mc.gtt_size - 4096 - RADEON_IB_POOL_SIZE*64*1024;
-	for(i = 0; i < RADEON_NUM_RINGS; ++i)
-		args->gart_size -= rdev->ring[i].ring_size;
+	args->vram_visible -= rdev->vram_pin_size;
+	args->gart_size = rdev->mc.gtt_size;
+	args->gart_size -= rdev->gart_pin_size;
+
 	return 0;
 }
 
@@ -252,8 +255,8 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 	/* create a gem object to contain this object in */
 	args->size = roundup(args->size, PAGE_SIZE);
 	r = radeon_gem_object_create(rdev, args->size, args->alignment,
-					args->initial_domain, false,
-					false, &gobj);
+				     args->initial_domain, args->flags,
+				     false, &gobj);
 	if (r) {
 		up_read(&rdev->exclusive_lock);
 		r = radeon_gem_handle_lockup(rdev, r);
@@ -272,6 +275,94 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
+int radeon_gem_userptr_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *filp)
+{
+	struct radeon_device *rdev = dev->dev_private;
+	struct drm_radeon_gem_userptr *args = data;
+	struct drm_gem_object *gobj;
+	struct radeon_bo *bo;
+	uint32_t handle;
+	int r;
+
+	if (offset_in_page(args->addr | args->size))
+		return -EINVAL;
+
+	/* reject unknown flag values */
+	if (args->flags & ~(RADEON_GEM_USERPTR_READONLY |
+	    RADEON_GEM_USERPTR_ANONONLY | RADEON_GEM_USERPTR_VALIDATE |
+	    RADEON_GEM_USERPTR_REGISTER))
+		return -EINVAL;
+
+	if (args->flags & RADEON_GEM_USERPTR_READONLY) {
+		/* readonly pages not tested on older hardware */
+		if (rdev->family < CHIP_R600)
+			return -EINVAL;
+
+	} else if (!(args->flags & RADEON_GEM_USERPTR_ANONONLY) ||
+		   !(args->flags & RADEON_GEM_USERPTR_REGISTER)) {
+
+		/* if we want to write to it we must require anonymous
+		   memory and install a MMU notifier */
+		return -EACCES;
+	}
+
+	down_read(&rdev->exclusive_lock);
+
+	/* create a gem object to contain this object in */
+	r = radeon_gem_object_create(rdev, args->size, 0,
+				     RADEON_GEM_DOMAIN_CPU, 0,
+				     false, &gobj);
+	if (r)
+		goto handle_lockup;
+
+	bo = gem_to_radeon_bo(gobj);
+	r = radeon_ttm_tt_set_userptr(bo->tbo.ttm, args->addr, args->flags);
+	if (r)
+		goto release_object;
+
+	if (args->flags & RADEON_GEM_USERPTR_REGISTER) {
+		r = radeon_mn_register(bo, args->addr);
+		if (r)
+			goto release_object;
+	}
+
+	if (args->flags & RADEON_GEM_USERPTR_VALIDATE) {
+		down_read(&current->mm->mmap_sem);
+		r = radeon_bo_reserve(bo, true);
+		if (r) {
+			up_read(&current->mm->mmap_sem);
+			goto release_object;
+		}
+
+		radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_GTT);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+		radeon_bo_unreserve(bo);
+		up_read(&current->mm->mmap_sem);
+		if (r)
+			goto release_object;
+	}
+
+	r = drm_gem_handle_create(filp, gobj, &handle);
+	/* drop reference from allocate - handle holds it now */
+	drm_gem_object_unreference_unlocked(gobj);
+	if (r)
+		goto handle_lockup;
+
+	args->handle = handle;
+	up_read(&rdev->exclusive_lock);
+	return 0;
+
+release_object:
+	drm_gem_object_unreference_unlocked(gobj);
+
+handle_lockup:
+	up_read(&rdev->exclusive_lock);
+	r = radeon_gem_handle_lockup(rdev, r);
+
+	return r;
+}
+
 int radeon_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp)
 {
@@ -315,6 +406,10 @@ int radeon_mode_dumb_mmap(struct drm_file *filp,
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
+	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm)) {
+		drm_gem_object_unreference_unlocked(gobj);
+		return -EPERM;
+	}
 	*offset_p = radeon_bo_mmap_offset(robj);
 	drm_gem_object_unreference_unlocked(gobj);
 	return 0;
@@ -357,17 +452,26 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 	struct drm_radeon_gem_wait_idle *args = data;
 	struct drm_gem_object *gobj;
 	struct radeon_bo *robj;
-	int r;
+	int r = 0;
+	uint32_t cur_placement = 0;
+	long ret;
 
 	gobj = drm_gem_object_lookup(dev, filp, args->handle);
 	if (gobj == NULL) {
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
-	r = radeon_bo_wait(robj, NULL, false);
-	/* callback hw specific functions if any */
-	if (rdev->asic->ioctl_wait_idle)
-		robj->rdev->asic->ioctl_wait_idle(rdev, robj);
+
+	ret = reservation_object_wait_timeout_rcu(robj->tbo.resv, true, true, 30 * HZ);
+	if (ret == 0)
+		r = -EBUSY;
+	else if (ret < 0)
+		r = ret;
+
+	/* Flush HDP cache via MMIO if necessary */
+	if (rdev->asic->mmio_hdp_flush &&
+	    radeon_mem_type_to_domain(cur_placement) == RADEON_GEM_DOMAIN_VRAM)
+		robj->rdev->asic->mmio_hdp_flush(rdev);
 	drm_gem_object_unreference_unlocked(gobj);
 	r = radeon_gem_handle_lockup(rdev, r);
 	return r;
@@ -461,11 +565,6 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 		args->operation = RADEON_VA_RESULT_ERROR;
 		return -EINVAL;
 	}
-	if (!(args->flags & RADEON_VM_PAGE_SNOOPED)) {
-		dev_err(&dev->pdev->dev, "only supported snooped mapping for now\n");
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
 
 	switch (args->operation) {
 	case RADEON_VA_MAP:
@@ -499,9 +598,9 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
 
 	switch (args->operation) {
 	case RADEON_VA_MAP:
-		if (bo_va->soffset) {
+		if (bo_va->it.start) {
 			args->operation = RADEON_VA_RESULT_VA_EXIST;
-			args->offset = bo_va->soffset;
+			args->offset = bo_va->it.start * RADEON_GPU_PAGE_SIZE;
 			goto out;
 		}
 		r = radeon_vm_bo_set_addr(rdev, bo_va, args->offset, args->flags);
@@ -535,6 +634,11 @@ int radeon_gem_op_ioctl(struct drm_device *dev, void *data,
 		return -ENOENT;
 	}
 	robj = gem_to_radeon_bo(gobj);
+
+	r = -EPERM;
+	if (radeon_ttm_tt_has_userptr(robj->tbo.ttm))
+		goto out;
+
 	r = radeon_bo_reserve(robj, false);
 	if (unlikely(r))
 		goto out;
@@ -572,9 +676,8 @@ int radeon_mode_dumb_create(struct drm_file *file_priv,
 	args->size = ALIGN(args->size, PAGE_SIZE);
 
 	r = radeon_gem_object_create(rdev, args->size, 0,
-				     RADEON_GEM_DOMAIN_VRAM,
-				     false, ttm_bo_type_device,
-				     &gobj);
+				     RADEON_GEM_DOMAIN_VRAM, 0,
+				     false, &gobj);
 	if (r)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/radeon/radeon_ib.c b/drivers/gpu/drm/radeon/radeon_ib.c
new file mode 100644
index 000000000000..3f39fcca4d07
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ib.c
@@ -0,0 +1,321 @@
+/*
+ * Copyright 2008 Advanced Micro Devices, Inc.
+ * Copyright 2008 Red Hat Inc.
+ * Copyright 2009 Jerome Glisse.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Dave Airlie
+ *          Alex Deucher
+ *          Jerome Glisse
+ *          Christian König
+ */
+#include <drm/drmP.h>
+#include "radeon.h"
+
+/*
+ * IB
+ * IBs (Indirect Buffers) and areas of GPU accessible memory where
+ * commands are stored.  You can put a pointer to the IB in the
+ * command ring and the hw will fetch the commands from the IB
+ * and execute them.  Generally userspace acceleration drivers
+ * produce command buffers which are send to the kernel and
+ * put in IBs for execution by the requested ring.
+ */
+static int radeon_debugfs_sa_init(struct radeon_device *rdev);
+
+/**
+ * radeon_ib_get - request an IB (Indirect Buffer)
+ *
+ * @rdev: radeon_device pointer
+ * @ring: ring index the IB is associated with
+ * @ib: IB object returned
+ * @size: requested IB size
+ *
+ * Request an IB (all asics).  IBs are allocated using the
+ * suballocator.
+ * Returns 0 on success, error on failure.
+ */
+int radeon_ib_get(struct radeon_device *rdev, int ring,
+		  struct radeon_ib *ib, struct radeon_vm *vm,
+		  unsigned size)
+{
+	int r;
+
+	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256);
+	if (r) {
+		dev_err(rdev->dev, "failed to get a new IB (%d)\n", r);
+		return r;
+	}
+
+	r = radeon_semaphore_create(rdev, &ib->semaphore);
+	if (r) {
+		return r;
+	}
+
+	ib->ring = ring;
+	ib->fence = NULL;
+	ib->ptr = radeon_sa_bo_cpu_addr(ib->sa_bo);
+	ib->vm = vm;
+	if (vm) {
+		/* ib pool is bound at RADEON_VA_IB_OFFSET in virtual address
+		 * space and soffset is the offset inside the pool bo
+		 */
+		ib->gpu_addr = ib->sa_bo->soffset + RADEON_VA_IB_OFFSET;
+	} else {
+		ib->gpu_addr = radeon_sa_bo_gpu_addr(ib->sa_bo);
+	}
+	ib->is_const_ib = false;
+
+	return 0;
+}
+
+/**
+ * radeon_ib_free - free an IB (Indirect Buffer)
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to free
+ *
+ * Free an IB (all asics).
+ */
+void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	radeon_semaphore_free(rdev, &ib->semaphore, ib->fence);
+	radeon_sa_bo_free(rdev, &ib->sa_bo, ib->fence);
+	radeon_fence_unref(&ib->fence);
+}
+
+/**
+ * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to schedule
+ * @const_ib: Const IB to schedule (SI only)
+ * @hdp_flush: Whether or not to perform an HDP cache flush
+ *
+ * Schedule an IB on the associated ring (all asics).
+ * Returns 0 on success, error on failure.
+ *
+ * On SI, there are two parallel engines fed from the primary ring,
+ * the CE (Constant Engine) and the DE (Drawing Engine).  Since
+ * resource descriptors have moved to memory, the CE allows you to
+ * prime the caches while the DE is updating register state so that
+ * the resource descriptors will be already in cache when the draw is
+ * processed.  To accomplish this, the userspace driver submits two
+ * IBs, one for the CE and one for the DE.  If there is a CE IB (called
+ * a CONST_IB), it will be put on the ring prior to the DE IB.  Prior
+ * to SI there was just a DE IB.
+ */
+int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
+		       struct radeon_ib *const_ib, bool hdp_flush)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	int r = 0;
+
+	if (!ib->length_dw || !ring->ready) {
+		/* TODO: Nothings in the ib we should report. */
+		dev_err(rdev->dev, "couldn't schedule ib\n");
+		return -EINVAL;
+	}
+
+	/* 64 dwords should be enough for fence too */
+	r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8);
+	if (r) {
+		dev_err(rdev->dev, "scheduling IB failed (%d).\n", r);
+		return r;
+	}
+
+	/* grab a vm id if necessary */
+	if (ib->vm) {
+		struct radeon_fence *vm_id_fence;
+		vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
+		radeon_semaphore_sync_fence(ib->semaphore, vm_id_fence);
+	}
+
+	/* sync with other rings */
+	r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring);
+	if (r) {
+		dev_err(rdev->dev, "failed to sync rings (%d)\n", r);
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+
+	if (ib->vm)
+		radeon_vm_flush(rdev, ib->vm, ib->ring);
+
+	if (const_ib) {
+		radeon_ring_ib_execute(rdev, const_ib->ring, const_ib);
+		radeon_semaphore_free(rdev, &const_ib->semaphore, NULL);
+	}
+	radeon_ring_ib_execute(rdev, ib->ring, ib);
+	r = radeon_fence_emit(rdev, &ib->fence, ib->ring);
+	if (r) {
+		dev_err(rdev->dev, "failed to emit fence for new IB (%d)\n", r);
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+	if (const_ib) {
+		const_ib->fence = radeon_fence_ref(ib->fence);
+	}
+
+	if (ib->vm)
+		radeon_vm_fence(rdev, ib->vm, ib->fence);
+
+	radeon_ring_unlock_commit(rdev, ring, hdp_flush);
+	return 0;
+}
+
+/**
+ * radeon_ib_pool_init - Init the IB (Indirect Buffer) pool
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Initialize the suballocator to manage a pool of memory
+ * for use as IBs (all asics).
+ * Returns 0 on success, error on failure.
+ */
+int radeon_ib_pool_init(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->ib_pool_ready) {
+		return 0;
+	}
+
+	if (rdev->family >= CHIP_BONAIRE) {
+		r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
+					      RADEON_IB_POOL_SIZE*64*1024,
+					      RADEON_GPU_PAGE_SIZE,
+					      RADEON_GEM_DOMAIN_GTT,
+					      RADEON_GEM_GTT_WC);
+	} else {
+		/* Before CIK, it's better to stick to cacheable GTT due
+		 * to the command stream checking
+		 */
+		r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
+					      RADEON_IB_POOL_SIZE*64*1024,
+					      RADEON_GPU_PAGE_SIZE,
+					      RADEON_GEM_DOMAIN_GTT, 0);
+	}
+	if (r) {
+		return r;
+	}
+
+	r = radeon_sa_bo_manager_start(rdev, &rdev->ring_tmp_bo);
+	if (r) {
+		return r;
+	}
+
+	rdev->ib_pool_ready = true;
+	if (radeon_debugfs_sa_init(rdev)) {
+		dev_err(rdev->dev, "failed to register debugfs file for SA\n");
+	}
+	return 0;
+}
+
+/**
+ * radeon_ib_pool_fini - Free the IB (Indirect Buffer) pool
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down the suballocator managing the pool of memory
+ * for use as IBs (all asics).
+ */
+void radeon_ib_pool_fini(struct radeon_device *rdev)
+{
+	if (rdev->ib_pool_ready) {
+		radeon_sa_bo_manager_suspend(rdev, &rdev->ring_tmp_bo);
+		radeon_sa_bo_manager_fini(rdev, &rdev->ring_tmp_bo);
+		rdev->ib_pool_ready = false;
+	}
+}
+
+/**
+ * radeon_ib_ring_tests - test IBs on the rings
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Test an IB (Indirect Buffer) on each ring.
+ * If the test fails, disable the ring.
+ * Returns 0 on success, error if the primary GFX ring
+ * IB test fails.
+ */
+int radeon_ib_ring_tests(struct radeon_device *rdev)
+{
+	unsigned i;
+	int r;
+
+	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
+		struct radeon_ring *ring = &rdev->ring[i];
+
+		if (!ring->ready)
+			continue;
+
+		r = radeon_ib_test(rdev, i, ring);
+		if (r) {
+			radeon_fence_driver_force_completion(rdev, i);
+			ring->ready = false;
+			rdev->needs_reset = false;
+
+			if (i == RADEON_RING_TYPE_GFX_INDEX) {
+				/* oh, oh, that's really bad */
+				DRM_ERROR("radeon: failed testing IB on GFX ring (%d).\n", r);
+		                rdev->accel_working = false;
+				return r;
+
+			} else {
+				/* still not good, but we can live with it */
+				DRM_ERROR("radeon: failed testing IB on ring %d (%d).\n", i, r);
+			}
+		}
+	}
+	return 0;
+}
+
+/*
+ * Debugfs info
+ */
+#if defined(CONFIG_DEBUG_FS)
+
+static int radeon_debugfs_sa_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct radeon_device *rdev = dev->dev_private;
+
+	radeon_sa_bo_dump_debug_info(&rdev->ring_tmp_bo, m);
+
+	return 0;
+
+}
+
+static struct drm_info_list radeon_debugfs_sa_list[] = {
+        {"radeon_sa_info", &radeon_debugfs_sa_info, 0, NULL},
+};
+
+#endif
+
+static int radeon_debugfs_sa_init(struct radeon_device *rdev)
+{
+#if defined(CONFIG_DEBUG_FS)
+	return radeon_debugfs_add_files(rdev, radeon_debugfs_sa_list, 1);
+#else
+	return 0;
+#endif
+}
diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 16807afab362..7784911d78ef 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -88,23 +88,6 @@ static void radeon_hotplug_work_func(struct work_struct *work)
 }
 
 /**
- * radeon_irq_reset_work_func - execute gpu reset
- *
- * @work: work struct
- *
- * Execute scheduled gpu reset (cayman+).
- * This function is called when the irq handler
- * thinks we need a gpu reset.
- */
-static void radeon_irq_reset_work_func(struct work_struct *work)
-{
-	struct radeon_device *rdev = container_of(work, struct radeon_device,
-						  reset_work);
-
-	radeon_gpu_reset(rdev);
-}
-
-/**
  * radeon_driver_irq_preinstall_kms - drm irq preinstall callback
  *
  * @dev: drm dev pointer
@@ -284,7 +267,6 @@ int radeon_irq_kms_init(struct radeon_device *rdev)
 
 	INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func);
 	INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi);
-	INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func);
 
 	rdev->irq.installed = true;
 	r = drm_irq_install(rdev->ddev, rdev->ddev->pdev->irq);
@@ -342,6 +324,21 @@ void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring)
 }
 
 /**
+ * radeon_irq_kms_sw_irq_get_delayed - enable software interrupt
+ *
+ * @rdev: radeon device pointer
+ * @ring: ring whose interrupt you want to enable
+ *
+ * Enables the software interrupt for a specific ring (all asics).
+ * The software interrupt is generally used to signal a fence on
+ * a particular ring.
+ */
+bool radeon_irq_kms_sw_irq_get_delayed(struct radeon_device *rdev, int ring)
+{
+	return atomic_inc_return(&rdev->irq.ring_int[ring]) == 1;
+}
+
+/**
  * radeon_irq_kms_sw_irq_put - disable software interrupt
  *
  * @rdev: radeon device pointer
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index d25ae6acfd5a..8309b11e674d 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -254,7 +254,18 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		}
 		break;
 	case RADEON_INFO_ACCEL_WORKING2:
-		*value = rdev->accel_working;
+		if (rdev->family == CHIP_HAWAII) {
+			if (rdev->accel_working) {
+				if (rdev->new_fw)
+					*value = 3;
+				else
+					*value = 2;
+			} else {
+				*value = 0;
+			}
+		} else {
+			*value = rdev->accel_working;
+		}
 		break;
 	case RADEON_INFO_TILING_CONFIG:
 		if (rdev->family >= CHIP_BONAIRE)
@@ -874,5 +885,6 @@ const struct drm_ioctl_desc radeon_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_OP, radeon_gem_op_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(RADEON_GEM_USERPTR, radeon_gem_userptr_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW),
 };
 int radeon_max_kms_ioctl = ARRAY_SIZE(radeon_ioctls_kms);
diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c
new file mode 100644
index 000000000000..a69bd441dd2d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_mn.c
@@ -0,0 +1,274 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <christian.koenig@amd.com>
+ */
+
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/mmu_notifier.h>
+#include <drm/drmP.h>
+#include <drm/drm.h>
+
+#include "radeon.h"
+
+struct radeon_mn {
+	/* constant after initialisation */
+	struct radeon_device	*rdev;
+	struct mm_struct	*mm;
+	struct mmu_notifier	mn;
+
+	/* only used on destruction */
+	struct work_struct	work;
+
+	/* protected by rdev->mn_lock */
+	struct hlist_node	node;
+
+	/* objects protected by lock */
+	struct mutex		lock;
+	struct rb_root		objects;
+};
+
+/**
+ * radeon_mn_destroy - destroy the rmn
+ *
+ * @work: previously sheduled work item
+ *
+ * Lazy destroys the notifier from a work item
+ */
+static void radeon_mn_destroy(struct work_struct *work)
+{
+	struct radeon_mn *rmn = container_of(work, struct radeon_mn, work);
+	struct radeon_device *rdev = rmn->rdev;
+	struct radeon_bo *bo, *next;
+
+	mutex_lock(&rdev->mn_lock);
+	mutex_lock(&rmn->lock);
+	hash_del(&rmn->node);
+	rbtree_postorder_for_each_entry_safe(bo, next, &rmn->objects, mn_it.rb) {
+		interval_tree_remove(&bo->mn_it, &rmn->objects);
+		bo->mn = NULL;
+	}
+	mutex_unlock(&rmn->lock);
+	mutex_unlock(&rdev->mn_lock);
+	mmu_notifier_unregister(&rmn->mn, rmn->mm);
+	kfree(rmn);
+}
+
+/**
+ * radeon_mn_release - callback to notify about mm destruction
+ *
+ * @mn: our notifier
+ * @mn: the mm this callback is about
+ *
+ * Shedule a work item to lazy destroy our notifier.
+ */
+static void radeon_mn_release(struct mmu_notifier *mn,
+			      struct mm_struct *mm)
+{
+	struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn);
+	INIT_WORK(&rmn->work, radeon_mn_destroy);
+	schedule_work(&rmn->work);
+}
+
+/**
+ * radeon_mn_invalidate_range_start - callback to notify about mm change
+ *
+ * @mn: our notifier
+ * @mn: the mm this callback is about
+ * @start: start of updated range
+ * @end: end of updated range
+ *
+ * We block for all BOs between start and end to be idle and
+ * unmap them by move them into system domain again.
+ */
+static void radeon_mn_invalidate_range_start(struct mmu_notifier *mn,
+					     struct mm_struct *mm,
+					     unsigned long start,
+					     unsigned long end)
+{
+	struct radeon_mn *rmn = container_of(mn, struct radeon_mn, mn);
+	struct interval_tree_node *it;
+
+	/* notification is exclusive, but interval is inclusive */
+	end -= 1;
+
+	mutex_lock(&rmn->lock);
+
+	it = interval_tree_iter_first(&rmn->objects, start, end);
+	while (it) {
+		struct radeon_bo *bo;
+		struct fence *fence;
+		int r;
+
+		bo = container_of(it, struct radeon_bo, mn_it);
+		it = interval_tree_iter_next(it, start, end);
+
+		r = radeon_bo_reserve(bo, true);
+		if (r) {
+			DRM_ERROR("(%d) failed to reserve user bo\n", r);
+			continue;
+		}
+
+		fence = reservation_object_get_excl(bo->tbo.resv);
+		if (fence) {
+			r = radeon_fence_wait((struct radeon_fence *)fence, false);
+			if (r)
+				DRM_ERROR("(%d) failed to wait for user bo\n", r);
+		}
+
+		radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_CPU);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+		if (r)
+			DRM_ERROR("(%d) failed to validate user bo\n", r);
+
+		radeon_bo_unreserve(bo);
+	}
+	
+	mutex_unlock(&rmn->lock);
+}
+
+static const struct mmu_notifier_ops radeon_mn_ops = {
+	.release = radeon_mn_release,
+	.invalidate_range_start = radeon_mn_invalidate_range_start,
+};
+
+/**
+ * radeon_mn_get - create notifier context
+ *
+ * @rdev: radeon device pointer
+ *
+ * Creates a notifier context for current->mm.
+ */
+static struct radeon_mn *radeon_mn_get(struct radeon_device *rdev)
+{
+	struct mm_struct *mm = current->mm;
+	struct radeon_mn *rmn;
+	int r;
+
+	down_write(&mm->mmap_sem);
+	mutex_lock(&rdev->mn_lock);
+
+	hash_for_each_possible(rdev->mn_hash, rmn, node, (unsigned long)mm)
+		if (rmn->mm == mm)
+			goto release_locks;
+
+	rmn = kzalloc(sizeof(*rmn), GFP_KERNEL);
+	if (!rmn) {
+		rmn = ERR_PTR(-ENOMEM);
+		goto release_locks;
+	}
+
+	rmn->rdev = rdev;
+	rmn->mm = mm;
+	rmn->mn.ops = &radeon_mn_ops;
+	mutex_init(&rmn->lock);
+	rmn->objects = RB_ROOT;
+	
+	r = __mmu_notifier_register(&rmn->mn, mm);
+	if (r)
+		goto free_rmn;
+
+	hash_add(rdev->mn_hash, &rmn->node, (unsigned long)mm);
+
+release_locks:
+	mutex_unlock(&rdev->mn_lock);
+	up_write(&mm->mmap_sem);
+
+	return rmn;
+
+free_rmn:
+	mutex_unlock(&rdev->mn_lock);
+	up_write(&mm->mmap_sem);
+	kfree(rmn);
+
+	return ERR_PTR(r);
+}
+
+/**
+ * radeon_mn_register - register a BO for notifier updates
+ *
+ * @bo: radeon buffer object
+ * @addr: userptr addr we should monitor
+ *
+ * Registers an MMU notifier for the given BO at the specified address.
+ * Returns 0 on success, -ERRNO if anything goes wrong.
+ */
+int radeon_mn_register(struct radeon_bo *bo, unsigned long addr)
+{
+	unsigned long end = addr + radeon_bo_size(bo) - 1;
+	struct radeon_device *rdev = bo->rdev;
+	struct radeon_mn *rmn;
+	struct interval_tree_node *it;
+
+	rmn = radeon_mn_get(rdev);
+	if (IS_ERR(rmn))
+		return PTR_ERR(rmn);
+
+	mutex_lock(&rmn->lock);
+
+	it = interval_tree_iter_first(&rmn->objects, addr, end);
+	if (it) {
+		mutex_unlock(&rmn->lock);
+		return -EEXIST;
+	}
+
+	bo->mn = rmn;
+	bo->mn_it.start = addr;
+	bo->mn_it.last = end;
+	interval_tree_insert(&bo->mn_it, &rmn->objects);
+
+	mutex_unlock(&rmn->lock);
+
+	return 0;
+}
+
+/**
+ * radeon_mn_unregister - unregister a BO for notifier updates
+ *
+ * @bo: radeon buffer object
+ *
+ * Remove any registration of MMU notifier updates from the buffer object.
+ */
+void radeon_mn_unregister(struct radeon_bo *bo)
+{
+	struct radeon_device *rdev = bo->rdev;
+	struct radeon_mn *rmn;
+
+	mutex_lock(&rdev->mn_lock);
+	rmn = bo->mn;
+	if (rmn == NULL) {
+		mutex_unlock(&rdev->mn_lock);
+		return;
+	}
+
+	mutex_lock(&rmn->lock);
+	interval_tree_remove(&bo->mn_it, &rmn->objects);
+	bo->mn = NULL;
+	mutex_unlock(&rmn->lock);
+	mutex_unlock(&rdev->mn_lock);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 0592ddb0904b..04db2fdd8692 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -685,10 +685,11 @@ extern bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 
 extern u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder);
 extern u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector);
-extern bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector);
 extern bool radeon_connector_is_dp12_capable(struct drm_connector *connector);
 extern int radeon_get_monitor_bpc(struct drm_connector *connector);
 
+extern struct edid *radeon_connector_edid(struct drm_connector *connector);
+
 extern void radeon_connector_hotplug(struct drm_connector *connector);
 extern int radeon_dp_mode_valid_helper(struct drm_connector *connector,
 				       struct drm_display_mode *mode);
@@ -738,7 +739,6 @@ extern void radeon_i2c_put_byte(struct radeon_i2c_chan *i2c,
 extern void radeon_router_select_ddc_port(struct radeon_connector *radeon_connector);
 extern void radeon_router_select_cd_port(struct radeon_connector *radeon_connector);
 extern bool radeon_ddc_probe(struct radeon_connector *radeon_connector, bool use_aux);
-extern int radeon_ddc_get_modes(struct radeon_connector *radeon_connector);
 
 extern struct drm_encoder *radeon_best_encoder(struct drm_connector *connector);
 
@@ -777,6 +777,7 @@ extern void atombios_digital_setup(struct drm_encoder *encoder, int action);
 extern int atombios_get_encoder_mode(struct drm_encoder *encoder);
 extern bool atombios_set_edp_panel_power(struct drm_connector *connector, int action);
 extern void radeon_encoder_set_active_device(struct drm_encoder *encoder);
+extern bool radeon_encoder_is_digital(struct drm_encoder *encoder);
 
 extern void radeon_crtc_load_lut(struct drm_crtc *crtc);
 extern int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y,
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 6c717b257d6d..99a960a4f302 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -46,16 +46,6 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo);
  * function are calling it.
  */
 
-static void radeon_bo_clear_va(struct radeon_bo *bo)
-{
-	struct radeon_bo_va *bo_va, *tmp;
-
-	list_for_each_entry_safe(bo_va, tmp, &bo->va, bo_list) {
-		/* remove from all vm address space */
-		radeon_vm_bo_rmv(bo->rdev, bo_va);
-	}
-}
-
 static void radeon_update_memory_usage(struct radeon_bo *bo,
 				       unsigned mem_type, int sign)
 {
@@ -85,12 +75,13 @@ static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	bo = container_of(tbo, struct radeon_bo, tbo);
 
 	radeon_update_memory_usage(bo, bo->tbo.mem.mem_type, -1);
+	radeon_mn_unregister(bo);
 
 	mutex_lock(&bo->rdev->gem.mutex);
 	list_del_init(&bo->list);
 	mutex_unlock(&bo->rdev->gem.mutex);
 	radeon_bo_clear_surface_reg(bo);
-	radeon_bo_clear_va(bo);
+	WARN_ON(!list_empty(&bo->va));
 	drm_gem_object_release(&bo->gem_base);
 	kfree(bo);
 }
@@ -106,47 +97,80 @@ void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain)
 {
 	u32 c = 0, i;
 
-	rbo->placement.fpfn = 0;
-	rbo->placement.lpfn = 0;
 	rbo->placement.placement = rbo->placements;
 	rbo->placement.busy_placement = rbo->placements;
 	if (domain & RADEON_GEM_DOMAIN_VRAM)
-		rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-					TTM_PL_FLAG_VRAM;
+		rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+					     TTM_PL_FLAG_UNCACHED |
+					     TTM_PL_FLAG_VRAM;
+
 	if (domain & RADEON_GEM_DOMAIN_GTT) {
-		if (rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_TT;
+		if (rbo->flags & RADEON_GEM_GTT_UC) {
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_TT;
+
+		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
+			   (rbo->rdev->flags & RADEON_IS_AGP)) {
+			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+				TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_TT;
 		} else {
-			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_TT;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
+						     TTM_PL_FLAG_TT;
 		}
 	}
+
 	if (domain & RADEON_GEM_DOMAIN_CPU) {
-		if (rbo->rdev->flags & RADEON_IS_AGP) {
-			rbo->placements[c++] = TTM_PL_FLAG_WC | TTM_PL_FLAG_SYSTEM;
+		if (rbo->flags & RADEON_GEM_GTT_UC) {
+			rbo->placements[c++].flags = TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_SYSTEM;
+
+		} else if ((rbo->flags & RADEON_GEM_GTT_WC) ||
+		    rbo->rdev->flags & RADEON_IS_AGP) {
+			rbo->placements[c++].flags = TTM_PL_FLAG_WC |
+				TTM_PL_FLAG_UNCACHED |
+				TTM_PL_FLAG_SYSTEM;
 		} else {
-			rbo->placements[c++] = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
+			rbo->placements[c++].flags = TTM_PL_FLAG_CACHED |
+						     TTM_PL_FLAG_SYSTEM;
 		}
 	}
 	if (!c)
-		rbo->placements[c++] = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		rbo->placements[c++].flags = TTM_PL_MASK_CACHING |
+					     TTM_PL_FLAG_SYSTEM;
+
 	rbo->placement.num_placement = c;
 	rbo->placement.num_busy_placement = c;
 
+	for (i = 0; i < c; ++i) {
+		rbo->placements[i].fpfn = 0;
+		if ((rbo->flags & RADEON_GEM_CPU_ACCESS) &&
+		    (rbo->placements[i].flags & TTM_PL_FLAG_VRAM))
+			rbo->placements[i].lpfn =
+				rbo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+		else
+			rbo->placements[i].lpfn = 0;
+	}
+
 	/*
 	 * Use two-ended allocation depending on the buffer size to
 	 * improve fragmentation quality.
 	 * 512kb was measured as the most optimal number.
 	 */
-	if (rbo->tbo.mem.size > 512 * 1024) {
+	if (!((rbo->flags & RADEON_GEM_CPU_ACCESS) &&
+	      (rbo->placements[i].flags & TTM_PL_FLAG_VRAM)) &&
+	    rbo->tbo.mem.size > 512 * 1024) {
 		for (i = 0; i < c; i++) {
-			rbo->placements[i] |= TTM_PL_FLAG_TOPDOWN;
+			rbo->placements[i].flags |= TTM_PL_FLAG_TOPDOWN;
 		}
 	}
 }
 
 int radeon_bo_create(struct radeon_device *rdev,
-		     unsigned long size, int byte_align, bool kernel, u32 domain,
-		     struct sg_table *sg, struct radeon_bo **bo_ptr)
+		     unsigned long size, int byte_align, bool kernel,
+		     u32 domain, u32 flags, struct sg_table *sg,
+		     struct reservation_object *resv,
+		     struct radeon_bo **bo_ptr)
 {
 	struct radeon_bo *bo;
 	enum ttm_bo_type type;
@@ -183,12 +207,18 @@ int radeon_bo_create(struct radeon_device *rdev,
 	bo->initial_domain = domain & (RADEON_GEM_DOMAIN_VRAM |
 	                               RADEON_GEM_DOMAIN_GTT |
 	                               RADEON_GEM_DOMAIN_CPU);
+
+	bo->flags = flags;
+	/* PCI GART is always snooped */
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		bo->flags &= ~(RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC);
+
 	radeon_ttm_placement_from_domain(bo, domain);
 	/* Kernel allocation are uninterruptible */
 	down_read(&rdev->pm.mclk_lock);
 	r = ttm_bo_init(&rdev->mman.bdev, &bo->tbo, size, type,
 			&bo->placement, page_align, !kernel, NULL,
-			acc_size, sg, &radeon_ttm_bo_destroy);
+			acc_size, sg, resv, &radeon_ttm_bo_destroy);
 	up_read(&rdev->pm.mclk_lock);
 	if (unlikely(r != 0)) {
 		return r;
@@ -232,6 +262,15 @@ void radeon_bo_kunmap(struct radeon_bo *bo)
 	ttm_bo_kunmap(&bo->kmap);
 }
 
+struct radeon_bo *radeon_bo_ref(struct radeon_bo *bo)
+{
+	if (bo == NULL)
+		return NULL;
+
+	ttm_bo_reference(&bo->tbo);
+	return bo;
+}
+
 void radeon_bo_unref(struct radeon_bo **bo)
 {
 	struct ttm_buffer_object *tbo;
@@ -241,9 +280,7 @@ void radeon_bo_unref(struct radeon_bo **bo)
 		return;
 	rdev = (*bo)->rdev;
 	tbo = &((*bo)->tbo);
-	down_read(&rdev->pm.mclk_lock);
 	ttm_bo_unref(&tbo);
-	up_read(&rdev->pm.mclk_lock);
 	if (tbo == NULL)
 		*bo = NULL;
 }
@@ -253,6 +290,9 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 {
 	int r, i;
 
+	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+		return -EPERM;
+
 	if (bo->pin_count) {
 		bo->pin_count++;
 		if (gpu_addr)
@@ -272,29 +312,31 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset,
 		return 0;
 	}
 	radeon_ttm_placement_from_domain(bo, domain);
-	if (domain == RADEON_GEM_DOMAIN_VRAM) {
+	for (i = 0; i < bo->placement.num_placement; i++) {
 		/* force to pin into visible video ram */
-		bo->placement.lpfn = bo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
-	}
-	if (max_offset) {
-		u64 lpfn = max_offset >> PAGE_SHIFT;
-
-		if (!bo->placement.lpfn)
-			bo->placement.lpfn = bo->rdev->mc.gtt_size >> PAGE_SHIFT;
+		if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
+		    !(bo->flags & RADEON_GEM_NO_CPU_ACCESS) &&
+		    (!max_offset || max_offset > bo->rdev->mc.visible_vram_size))
+			bo->placements[i].lpfn =
+				bo->rdev->mc.visible_vram_size >> PAGE_SHIFT;
+		else
+			bo->placements[i].lpfn = max_offset >> PAGE_SHIFT;
 
-		if (lpfn < bo->placement.lpfn)
-			bo->placement.lpfn = lpfn;
+		bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;
 	}
-	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] |= TTM_PL_FLAG_NO_EVICT;
+
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
 	if (likely(r == 0)) {
 		bo->pin_count = 1;
 		if (gpu_addr != NULL)
 			*gpu_addr = radeon_bo_gpu_offset(bo);
-	}
-	if (unlikely(r != 0))
+		if (domain == RADEON_GEM_DOMAIN_VRAM)
+			bo->rdev->vram_pin_size += radeon_bo_size(bo);
+		else
+			bo->rdev->gart_pin_size += radeon_bo_size(bo);
+	} else {
 		dev_err(bo->rdev->dev, "%p pin failed\n", bo);
+	}
 	return r;
 }
 
@@ -314,11 +356,19 @@ int radeon_bo_unpin(struct radeon_bo *bo)
 	bo->pin_count--;
 	if (bo->pin_count)
 		return 0;
-	for (i = 0; i < bo->placement.num_placement; i++)
-		bo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT;
+	for (i = 0; i < bo->placement.num_placement; i++) {
+		bo->placements[i].lpfn = 0;
+		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
+	}
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-	if (unlikely(r != 0))
+	if (likely(r == 0)) {
+		if (bo->tbo.mem.mem_type == TTM_PL_VRAM)
+			bo->rdev->vram_pin_size -= radeon_bo_size(bo);
+		else
+			bo->rdev->gart_pin_size -= radeon_bo_size(bo);
+	} else {
 		dev_err(bo->rdev->dev, "%p validate failed for unpin\n", bo);
+	}
 	return r;
 }
 
@@ -438,7 +488,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 	u64 bytes_moved = 0, initial_bytes_moved;
 	u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev);
 
-	r = ttm_eu_reserve_buffers(ticket, head);
+	r = ttm_eu_reserve_buffers(ticket, head, true);
 	if (unlikely(r != 0)) {
 		return r;
 	}
@@ -447,6 +497,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 		bo = lobj->robj;
 		if (!bo->pin_count) {
 			u32 domain = lobj->prefered_domains;
+			u32 allowed = lobj->allowed_domains;
 			u32 current_domain =
 				radeon_mem_type_to_domain(bo->tbo.mem.mem_type);
 
@@ -458,7 +509,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 			 * into account. We don't want to disallow buffer moves
 			 * completely.
 			 */
-			if ((lobj->allowed_domains & current_domain) != 0 &&
+			if ((allowed & current_domain) != 0 &&
 			    (domain & current_domain) == 0 && /* will be moved */
 			    bytes_moved > bytes_moved_threshold) {
 				/* don't move it */
@@ -468,7 +519,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev,
 		retry:
 			radeon_ttm_placement_from_domain(bo, domain);
 			if (ring == R600_RING_TYPE_UVD_INDEX)
-				radeon_uvd_force_into_uvd_segment(bo);
+				radeon_uvd_force_into_uvd_segment(bo, allowed);
 
 			initial_bytes_moved = atomic64_read(&rdev->num_bytes_moved);
 			r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
@@ -710,7 +761,7 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 
 	/* hurrah the memory is not visible ! */
 	radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM);
-	rbo->placement.lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT;
+	rbo->placements[0].lpfn = rdev->mc.visible_vram_size >> PAGE_SHIFT;
 	r = ttm_bo_validate(bo, &rbo->placement, false, false);
 	if (unlikely(r == -ENOMEM)) {
 		radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT);
@@ -734,12 +785,10 @@ int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait)
 	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, NULL);
 	if (unlikely(r != 0))
 		return r;
-	spin_lock(&bo->tbo.bdev->fence_lock);
 	if (mem_type)
 		*mem_type = bo->tbo.mem.mem_type;
-	if (bo->tbo.sync_obj)
-		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
-	spin_unlock(&bo->tbo.bdev->fence_lock);
+
+	r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index 5a873f31a171..1b8ec7917154 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -124,11 +124,13 @@ extern int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type,
 
 extern int radeon_bo_create(struct radeon_device *rdev,
 			    unsigned long size, int byte_align,
-			    bool kernel, u32 domain,
+			    bool kernel, u32 domain, u32 flags,
 			    struct sg_table *sg,
+			    struct reservation_object *resv,
 			    struct radeon_bo **bo_ptr);
 extern int radeon_bo_kmap(struct radeon_bo *bo, void **ptr);
 extern void radeon_bo_kunmap(struct radeon_bo *bo);
+extern struct radeon_bo *radeon_bo_ref(struct radeon_bo *bo);
 extern void radeon_bo_unref(struct radeon_bo **bo);
 extern int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr);
 extern int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain,
@@ -170,7 +172,8 @@ static inline void * radeon_sa_bo_cpu_addr(struct radeon_sa_bo *sa_bo)
 
 extern int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 				     struct radeon_sa_manager *sa_manager,
-				     unsigned size, u32 align, u32 domain);
+				     unsigned size, u32 align, u32 domain,
+				     u32 flags);
 extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev,
 				      struct radeon_sa_manager *sa_manager);
 extern int radeon_sa_bo_manager_start(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index e447e390d09a..32522cc940a1 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -460,10 +460,6 @@ static ssize_t radeon_get_dpm_state(struct device *dev,
 	struct radeon_device *rdev = ddev->dev_private;
 	enum radeon_pm_state_type pm = rdev->pm.dpm.user_state;
 
-	if  ((rdev->flags & RADEON_IS_PX) &&
-	     (ddev->switch_power_state != DRM_SWITCH_POWER_ON))
-		return snprintf(buf, PAGE_SIZE, "off\n");
-
 	return snprintf(buf, PAGE_SIZE, "%s\n",
 			(pm == POWER_STATE_TYPE_BATTERY) ? "battery" :
 			(pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance");
@@ -477,11 +473,6 @@ static ssize_t radeon_set_dpm_state(struct device *dev,
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct radeon_device *rdev = ddev->dev_private;
 
-	/* Can't set dpm state when the card is off */
-	if  ((rdev->flags & RADEON_IS_PX) &&
-	     (ddev->switch_power_state != DRM_SWITCH_POWER_ON))
-		return -EINVAL;
-
 	mutex_lock(&rdev->pm.mutex);
 	if (strncmp("battery", buf, strlen("battery")) == 0)
 		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY;
@@ -495,7 +486,12 @@ static ssize_t radeon_set_dpm_state(struct device *dev,
 		goto fail;
 	}
 	mutex_unlock(&rdev->pm.mutex);
-	radeon_pm_compute_clocks(rdev);
+
+	/* Can't set dpm state when the card is off */
+	if (!(rdev->flags & RADEON_IS_PX) ||
+	    (ddev->switch_power_state == DRM_SWITCH_POWER_ON))
+		radeon_pm_compute_clocks(rdev);
+
 fail:
 	return count;
 }
@@ -1303,10 +1299,6 @@ int radeon_pm_init(struct radeon_device *rdev)
 	case CHIP_RS780:
 	case CHIP_RS880:
 	case CHIP_RV770:
-	case CHIP_BARTS:
-	case CHIP_TURKS:
-	case CHIP_CAICOS:
-	case CHIP_CAYMAN:
 		/* DPM requires the RLC, RV770+ dGPU requires SMC */
 		if (!rdev->rlc_fw)
 			rdev->pm.pm_method = PM_METHOD_PROFILE;
@@ -1330,6 +1322,10 @@ int radeon_pm_init(struct radeon_device *rdev)
 	case CHIP_PALM:
 	case CHIP_SUMO:
 	case CHIP_SUMO2:
+	case CHIP_BARTS:
+	case CHIP_TURKS:
+	case CHIP_CAICOS:
+	case CHIP_CAYMAN:
 	case CHIP_ARUBA:
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
@@ -1400,9 +1396,7 @@ static void radeon_pm_fini_old(struct radeon_device *rdev)
 	}
 
 	radeon_hwmon_fini(rdev);
-
-	if (rdev->pm.power_state)
-		kfree(rdev->pm.power_state);
+	kfree(rdev->pm.power_state);
 }
 
 static void radeon_pm_fini_dpm(struct radeon_device *rdev)
@@ -1421,9 +1415,7 @@ static void radeon_pm_fini_dpm(struct radeon_device *rdev)
 	radeon_dpm_fini(rdev);
 
 	radeon_hwmon_fini(rdev);
-
-	if (rdev->pm.power_state)
-		kfree(rdev->pm.power_state);
+	kfree(rdev->pm.power_state);
 }
 
 void radeon_pm_fini(struct radeon_device *rdev)
@@ -1564,7 +1556,7 @@ static bool radeon_pm_in_vbl(struct radeon_device *rdev)
 		if (rdev->pm.active_crtcs & (1 << crtc)) {
 			vbl_status = radeon_get_crtc_scanoutpos(rdev->ddev, crtc, 0, &vpos, &hpos, NULL, NULL);
 			if ((vbl_status & DRM_SCANOUTPOS_VALID) &&
-			    !(vbl_status & DRM_SCANOUTPOS_INVBL))
+			    !(vbl_status & DRM_SCANOUTPOS_IN_VBLANK))
 				in_vbl = false;
 		}
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 28d71070c389..f3609c97496b 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -27,6 +27,7 @@
 
 #include "radeon.h"
 #include <drm/radeon_drm.h>
+#include <linux/dma-buf.h>
 
 struct sg_table *radeon_gem_prime_get_sg_table(struct drm_gem_object *obj)
 {
@@ -57,15 +58,18 @@ void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
 }
 
 struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
-							size_t size,
+							struct dma_buf_attachment *attach,
 							struct sg_table *sg)
 {
+	struct reservation_object *resv = attach->dmabuf->resv;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_bo *bo;
 	int ret;
 
-	ret = radeon_bo_create(rdev, size, PAGE_SIZE, false,
-			       RADEON_GEM_DOMAIN_GTT, sg, &bo);
+	ww_mutex_lock(&resv->lock, NULL);
+	ret = radeon_bo_create(rdev, attach->dmabuf->size, PAGE_SIZE, false,
+			       RADEON_GEM_DOMAIN_GTT, 0, sg, resv, &bo);
+	ww_mutex_unlock(&resv->lock);
 	if (ret)
 		return ERR_PTR(ret);
 
@@ -111,3 +115,13 @@ struct reservation_object *radeon_gem_prime_res_obj(struct drm_gem_object *obj)
 
 	return bo->tbo.resv;
 }
+
+struct dma_buf *radeon_gem_prime_export(struct drm_device *dev,
+					struct drm_gem_object *gobj,
+					int flags)
+{
+	struct radeon_bo *bo = gem_to_radeon_bo(gobj);
+	if (radeon_ttm_tt_has_userptr(bo->tbo.ttm))
+		return ERR_PTR(-EPERM);
+	return drm_gem_prime_export(dev, gobj, flags);
+}
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index f8050f5429e2..2456f69efd23 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -26,258 +26,8 @@
  *          Jerome Glisse
  *          Christian König
  */
-#include <linux/seq_file.h>
-#include <linux/slab.h>
 #include <drm/drmP.h>
-#include <drm/radeon_drm.h>
-#include "radeon_reg.h"
 #include "radeon.h"
-#include "atom.h"
-
-/*
- * IB
- * IBs (Indirect Buffers) and areas of GPU accessible memory where
- * commands are stored.  You can put a pointer to the IB in the
- * command ring and the hw will fetch the commands from the IB
- * and execute them.  Generally userspace acceleration drivers
- * produce command buffers which are send to the kernel and
- * put in IBs for execution by the requested ring.
- */
-static int radeon_debugfs_sa_init(struct radeon_device *rdev);
-
-/**
- * radeon_ib_get - request an IB (Indirect Buffer)
- *
- * @rdev: radeon_device pointer
- * @ring: ring index the IB is associated with
- * @ib: IB object returned
- * @size: requested IB size
- *
- * Request an IB (all asics).  IBs are allocated using the
- * suballocator.
- * Returns 0 on success, error on failure.
- */
-int radeon_ib_get(struct radeon_device *rdev, int ring,
-		  struct radeon_ib *ib, struct radeon_vm *vm,
-		  unsigned size)
-{
-	int r;
-
-	r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256);
-	if (r) {
-		dev_err(rdev->dev, "failed to get a new IB (%d)\n", r);
-		return r;
-	}
-
-	r = radeon_semaphore_create(rdev, &ib->semaphore);
-	if (r) {
-		return r;
-	}
-
-	ib->ring = ring;
-	ib->fence = NULL;
-	ib->ptr = radeon_sa_bo_cpu_addr(ib->sa_bo);
-	ib->vm = vm;
-	if (vm) {
-		/* ib pool is bound at RADEON_VA_IB_OFFSET in virtual address
-		 * space and soffset is the offset inside the pool bo
-		 */
-		ib->gpu_addr = ib->sa_bo->soffset + RADEON_VA_IB_OFFSET;
-	} else {
-		ib->gpu_addr = radeon_sa_bo_gpu_addr(ib->sa_bo);
-	}
-	ib->is_const_ib = false;
-
-	return 0;
-}
-
-/**
- * radeon_ib_free - free an IB (Indirect Buffer)
- *
- * @rdev: radeon_device pointer
- * @ib: IB object to free
- *
- * Free an IB (all asics).
- */
-void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib *ib)
-{
-	radeon_semaphore_free(rdev, &ib->semaphore, ib->fence);
-	radeon_sa_bo_free(rdev, &ib->sa_bo, ib->fence);
-	radeon_fence_unref(&ib->fence);
-}
-
-/**
- * radeon_ib_schedule - schedule an IB (Indirect Buffer) on the ring
- *
- * @rdev: radeon_device pointer
- * @ib: IB object to schedule
- * @const_ib: Const IB to schedule (SI only)
- *
- * Schedule an IB on the associated ring (all asics).
- * Returns 0 on success, error on failure.
- *
- * On SI, there are two parallel engines fed from the primary ring,
- * the CE (Constant Engine) and the DE (Drawing Engine).  Since
- * resource descriptors have moved to memory, the CE allows you to
- * prime the caches while the DE is updating register state so that
- * the resource descriptors will be already in cache when the draw is
- * processed.  To accomplish this, the userspace driver submits two
- * IBs, one for the CE and one for the DE.  If there is a CE IB (called
- * a CONST_IB), it will be put on the ring prior to the DE IB.  Prior
- * to SI there was just a DE IB.
- */
-int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib,
-		       struct radeon_ib *const_ib)
-{
-	struct radeon_ring *ring = &rdev->ring[ib->ring];
-	int r = 0;
-
-	if (!ib->length_dw || !ring->ready) {
-		/* TODO: Nothings in the ib we should report. */
-		dev_err(rdev->dev, "couldn't schedule ib\n");
-		return -EINVAL;
-	}
-
-	/* 64 dwords should be enough for fence too */
-	r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8);
-	if (r) {
-		dev_err(rdev->dev, "scheduling IB failed (%d).\n", r);
-		return r;
-	}
-
-	/* grab a vm id if necessary */
-	if (ib->vm) {
-		struct radeon_fence *vm_id_fence;
-		vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring);
-        	radeon_semaphore_sync_to(ib->semaphore, vm_id_fence);
-	}
-
-	/* sync with other rings */
-	r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring);
-	if (r) {
-		dev_err(rdev->dev, "failed to sync rings (%d)\n", r);
-		radeon_ring_unlock_undo(rdev, ring);
-		return r;
-	}
-
-	if (ib->vm)
-		radeon_vm_flush(rdev, ib->vm, ib->ring);
-
-	if (const_ib) {
-		radeon_ring_ib_execute(rdev, const_ib->ring, const_ib);
-		radeon_semaphore_free(rdev, &const_ib->semaphore, NULL);
-	}
-	radeon_ring_ib_execute(rdev, ib->ring, ib);
-	r = radeon_fence_emit(rdev, &ib->fence, ib->ring);
-	if (r) {
-		dev_err(rdev->dev, "failed to emit fence for new IB (%d)\n", r);
-		radeon_ring_unlock_undo(rdev, ring);
-		return r;
-	}
-	if (const_ib) {
-		const_ib->fence = radeon_fence_ref(ib->fence);
-	}
-
-	if (ib->vm)
-		radeon_vm_fence(rdev, ib->vm, ib->fence);
-
-	radeon_ring_unlock_commit(rdev, ring);
-	return 0;
-}
-
-/**
- * radeon_ib_pool_init - Init the IB (Indirect Buffer) pool
- *
- * @rdev: radeon_device pointer
- *
- * Initialize the suballocator to manage a pool of memory
- * for use as IBs (all asics).
- * Returns 0 on success, error on failure.
- */
-int radeon_ib_pool_init(struct radeon_device *rdev)
-{
-	int r;
-
-	if (rdev->ib_pool_ready) {
-		return 0;
-	}
-	r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
-				      RADEON_IB_POOL_SIZE*64*1024,
-				      RADEON_GPU_PAGE_SIZE,
-				      RADEON_GEM_DOMAIN_GTT);
-	if (r) {
-		return r;
-	}
-
-	r = radeon_sa_bo_manager_start(rdev, &rdev->ring_tmp_bo);
-	if (r) {
-		return r;
-	}
-
-	rdev->ib_pool_ready = true;
-	if (radeon_debugfs_sa_init(rdev)) {
-		dev_err(rdev->dev, "failed to register debugfs file for SA\n");
-	}
-	return 0;
-}
-
-/**
- * radeon_ib_pool_fini - Free the IB (Indirect Buffer) pool
- *
- * @rdev: radeon_device pointer
- *
- * Tear down the suballocator managing the pool of memory
- * for use as IBs (all asics).
- */
-void radeon_ib_pool_fini(struct radeon_device *rdev)
-{
-	if (rdev->ib_pool_ready) {
-		radeon_sa_bo_manager_suspend(rdev, &rdev->ring_tmp_bo);
-		radeon_sa_bo_manager_fini(rdev, &rdev->ring_tmp_bo);
-		rdev->ib_pool_ready = false;
-	}
-}
-
-/**
- * radeon_ib_ring_tests - test IBs on the rings
- *
- * @rdev: radeon_device pointer
- *
- * Test an IB (Indirect Buffer) on each ring.
- * If the test fails, disable the ring.
- * Returns 0 on success, error if the primary GFX ring
- * IB test fails.
- */
-int radeon_ib_ring_tests(struct radeon_device *rdev)
-{
-	unsigned i;
-	int r;
-
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		struct radeon_ring *ring = &rdev->ring[i];
-
-		if (!ring->ready)
-			continue;
-
-		r = radeon_ib_test(rdev, i, ring);
-		if (r) {
-			ring->ready = false;
-			rdev->needs_reset = false;
-
-			if (i == RADEON_RING_TYPE_GFX_INDEX) {
-				/* oh, oh, that's really bad */
-				DRM_ERROR("radeon: failed testing IB on GFX ring (%d).\n", r);
-		                rdev->accel_working = false;
-				return r;
-
-			} else {
-				/* still not good, but we can live with it */
-				DRM_ERROR("radeon: failed testing IB on ring %d (%d).\n", i, r);
-			}
-		}
-	}
-	return 0;
-}
 
 /*
  * Rings
@@ -295,27 +45,6 @@ int radeon_ib_ring_tests(struct radeon_device *rdev)
 static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ring *ring);
 
 /**
- * radeon_ring_write - write a value to the ring
- *
- * @ring: radeon_ring structure holding ring information
- * @v: dword (dw) value to write
- *
- * Write a value to the requested ring buffer (all asics).
- */
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
-{
-#if DRM_DEBUG_CODE
-	if (ring->count_dw <= 0) {
-		DRM_ERROR("radeon: writing more dwords to the ring than expected!\n");
-	}
-#endif
-	ring->ring[ring->wptr++] = v;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw--;
-	ring->ring_free_dw--;
-}
-
-/**
  * radeon_ring_supports_scratch_reg - check if the ring supports
  * writing to scratch registers
  *
@@ -427,17 +156,29 @@ int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *ring, unsig
  *
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
+ * @hdp_flush: Whether or not to perform an HDP cache flush
  *
  * Update the wptr (write pointer) to tell the GPU to
  * execute new commands on the ring buffer (all asics).
  */
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring,
+			bool hdp_flush)
 {
+	/* If we are emitting the HDP flush via the ring buffer, we need to
+	 * do it before padding.
+	 */
+	if (hdp_flush && rdev->asic->ring[ring->idx]->hdp_flush)
+		rdev->asic->ring[ring->idx]->hdp_flush(rdev, ring);
 	/* We pad to match fetch size */
 	while (ring->wptr & ring->align_mask) {
 		radeon_ring_write(ring, ring->nop);
 	}
 	mb();
+	/* If we are emitting the HDP flush via MMIO, we need to do it after
+	 * all CPU writes to VRAM finished.
+	 */
+	if (hdp_flush && rdev->asic->mmio_hdp_flush)
+		rdev->asic->mmio_hdp_flush(rdev);
 	radeon_ring_set_wptr(rdev, ring);
 }
 
@@ -447,12 +188,14 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
  *
  * @rdev: radeon_device pointer
  * @ring: radeon_ring structure holding ring information
+ * @hdp_flush: Whether or not to perform an HDP cache flush
  *
  * Call radeon_ring_commit() then unlock the ring (all asics).
  */
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring,
+			       bool hdp_flush)
 {
-	radeon_ring_commit(rdev, ring);
+	radeon_ring_commit(rdev, ring, hdp_flush);
 	mutex_unlock(&rdev->ring_lock);
 }
 
@@ -571,7 +314,7 @@ unsigned radeon_ring_backup(struct radeon_device *rdev, struct radeon_ring *ring
 	}
 
 	/* and then save the content of the ring */
-	*data = kmalloc_array(size, sizeof(uint32_t), GFP_KERNEL);
+	*data = drm_malloc_ab(size, sizeof(uint32_t));
 	if (!*data) {
 		mutex_unlock(&rdev->ring_lock);
 		return 0;
@@ -612,8 +355,8 @@ int radeon_ring_restore(struct radeon_device *rdev, struct radeon_ring *ring,
 		radeon_ring_write(ring, data[i]);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	kfree(data);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	drm_free_large(data);
 	return 0;
 }
 
@@ -640,7 +383,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig
 	/* Allocate ring buffer */
 	if (ring->ring_obj == NULL) {
 		r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT,
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL,
 				     NULL, &ring->ring_obj);
 		if (r) {
 			dev_err(rdev->dev, "(%d) ring create failed\n", r);
@@ -791,22 +534,6 @@ static struct drm_info_list radeon_debugfs_ring_info_list[] = {
 	{"radeon_ring_vce2", radeon_debugfs_ring_info, 0, &si_vce2_index},
 };
 
-static int radeon_debugfs_sa_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct radeon_device *rdev = dev->dev_private;
-
-	radeon_sa_bo_dump_debug_info(&rdev->ring_tmp_bo, m);
-
-	return 0;
-
-}
-
-static struct drm_info_list radeon_debugfs_sa_list[] = {
-        {"radeon_sa_info", &radeon_debugfs_sa_info, 0, NULL},
-};
-
 #endif
 
 static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ring *ring)
@@ -828,12 +555,3 @@ static int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ri
 #endif
 	return 0;
 }
-
-static int radeon_debugfs_sa_init(struct radeon_device *rdev)
-{
-#if defined(CONFIG_DEBUG_FS)
-	return radeon_debugfs_add_files(rdev, radeon_debugfs_sa_list, 1);
-#else
-	return 0;
-#endif
-}
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index adcf3e2f07da..c507896aca45 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -49,7 +49,7 @@ static void radeon_sa_bo_try_free(struct radeon_sa_manager *sa_manager);
 
 int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 			      struct radeon_sa_manager *sa_manager,
-			      unsigned size, u32 align, u32 domain)
+			      unsigned size, u32 align, u32 domain, u32 flags)
 {
 	int i, r;
 
@@ -65,7 +65,7 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 	}
 
 	r = radeon_bo_create(rdev, size, align, true,
-			     domain, NULL, &sa_manager->bo);
+			     domain, flags, NULL, NULL, &sa_manager->bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
index dbd6bcde92de..6deb08f045b7 100644
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
@@ -34,7 +34,7 @@
 int radeon_semaphore_create(struct radeon_device *rdev,
 			    struct radeon_semaphore **semaphore)
 {
-	uint32_t *cpu_addr;
+	uint64_t *cpu_addr;
 	int i, r;
 
 	*semaphore = kmalloc(sizeof(struct radeon_semaphore), GFP_KERNEL);
@@ -96,15 +96,15 @@ bool radeon_semaphore_emit_wait(struct radeon_device *rdev, int ridx,
 }
 
 /**
- * radeon_semaphore_sync_to - use the semaphore to sync to a fence
+ * radeon_semaphore_sync_fence - use the semaphore to sync to a fence
  *
  * @semaphore: semaphore object to add fence to
  * @fence: fence to sync to
  *
  * Sync to the fence using this semaphore object
  */
-void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
-			      struct radeon_fence *fence)
+void radeon_semaphore_sync_fence(struct radeon_semaphore *semaphore,
+				 struct radeon_fence *fence)
 {
         struct radeon_fence *other;
 
@@ -116,6 +116,53 @@ void radeon_semaphore_sync_to(struct radeon_semaphore *semaphore,
 }
 
 /**
+ * radeon_semaphore_sync_to - use the semaphore to sync to a reservation object
+ *
+ * @sema: semaphore object to add fence from reservation object to
+ * @resv: reservation object with embedded fence
+ * @shared: true if we should onyl sync to the exclusive fence
+ *
+ * Sync to the fence using this semaphore object
+ */
+int radeon_semaphore_sync_resv(struct radeon_device *rdev,
+			       struct radeon_semaphore *sema,
+			       struct reservation_object *resv,
+			       bool shared)
+{
+	struct reservation_object_list *flist;
+	struct fence *f;
+	struct radeon_fence *fence;
+	unsigned i;
+	int r = 0;
+
+	/* always sync to the exclusive fence */
+	f = reservation_object_get_excl(resv);
+	fence = f ? to_radeon_fence(f) : NULL;
+	if (fence && fence->rdev == rdev)
+		radeon_semaphore_sync_fence(sema, fence);
+	else if (f)
+		r = fence_wait(f, true);
+
+	flist = reservation_object_get_list(resv);
+	if (shared || !flist || r)
+		return r;
+
+	for (i = 0; i < flist->shared_count; ++i) {
+		f = rcu_dereference_protected(flist->shared[i],
+					      reservation_object_held(resv));
+		fence = to_radeon_fence(f);
+		if (fence && fence->rdev == rdev)
+			radeon_semaphore_sync_fence(sema, fence);
+		else
+			r = fence_wait(f, true);
+
+		if (r)
+			break;
+	}
+	return r;
+}
+
+/**
  * radeon_semaphore_sync_rings - sync ring to all registered fences
  *
  * @rdev: radeon_device pointer
@@ -179,7 +226,7 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev,
 			continue;
 		}
 
-		radeon_ring_commit(rdev, &rdev->ring[i]);
+		radeon_ring_commit(rdev, &rdev->ring[i], false);
 		radeon_fence_note_sync(fence, ring);
 
 		semaphore->gpu_addr += 8;
diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c
index 23bb64fd775f..535403e0c8a2 100644
--- a/drivers/gpu/drm/radeon/radeon_state.c
+++ b/drivers/gpu/drm/radeon/radeon_state.c
@@ -30,9 +30,9 @@
  */
 
 #include <drm/drmP.h>
-#include <drm/drm_buffer.h>
 #include <drm/radeon_drm.h>
 #include "radeon_drv.h"
+#include "drm_buffer.h"
 
 /* ================================================================
  * Helper functions for client state checking and fixup
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index 3a13e0d1055c..07b506b41008 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -56,13 +56,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	/* Number of tests =
 	 * (Total GTT - IB pool - writeback page - ring buffers) / test size
 	 */
-	n = rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024;
-	for (i = 0; i < RADEON_NUM_RINGS; ++i)
-		n -= rdev->ring[i].ring_size;
-	if (rdev->wb.wb_obj)
-		n -= RADEON_GPU_PAGE_SIZE;
-	if (rdev->ih.ring_obj)
-		n -= rdev->ih.ring_size;
+	n = rdev->mc.gtt_size - rdev->gart_pin_size;
 	n /= size;
 
 	gtt_obj = kzalloc(n * sizeof(*gtt_obj), GFP_KERNEL);
@@ -73,7 +67,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	}
 
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-			     NULL, &vram_obj);
+			     0, NULL, NULL, &vram_obj);
 	if (r) {
 		DRM_ERROR("Failed to create VRAM object\n");
 		goto out_cleanup;
@@ -93,7 +87,8 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		struct radeon_fence *fence = NULL;
 
 		r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_GTT, NULL, gtt_obj + i);
+				     RADEON_GEM_DOMAIN_GTT, 0, NULL, NULL,
+				     gtt_obj + i);
 		if (r) {
 			DRM_ERROR("Failed to create GTT object %d\n", i);
 			goto out_lclean;
@@ -122,11 +117,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		radeon_bo_kunmap(gtt_obj[i]);
 
 		if (ring == R600_RING_TYPE_DMA_INDEX)
-			r = radeon_copy_dma(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+			fence = radeon_copy_dma(rdev, gtt_addr, vram_addr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 		else
-			r = radeon_copy_blit(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-		if (r) {
+			fence = radeon_copy_blit(rdev, gtt_addr, vram_addr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
+		if (IS_ERR(fence)) {
 			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
+			r = PTR_ERR(fence);
 			goto out_lclean_unpin;
 		}
 
@@ -168,11 +168,16 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		radeon_bo_kunmap(vram_obj);
 
 		if (ring == R600_RING_TYPE_DMA_INDEX)
-			r = radeon_copy_dma(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
+			fence = radeon_copy_dma(rdev, vram_addr, gtt_addr,
+						size / RADEON_GPU_PAGE_SIZE,
+						NULL);
 		else
-			r = radeon_copy_blit(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
-		if (r) {
+			fence = radeon_copy_blit(rdev, vram_addr, gtt_addr,
+						 size / RADEON_GPU_PAGE_SIZE,
+						 NULL);
+		if (IS_ERR(fence)) {
 			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
+			r = PTR_ERR(fence);
 			goto out_lclean_unpin;
 		}
 
@@ -228,7 +233,7 @@ out_lclean:
 			radeon_bo_unreserve(gtt_obj[i]);
 			radeon_bo_unref(&gtt_obj[i]);
 		}
-		if (fence)
+		if (fence && !IS_ERR(fence))
 			radeon_fence_unref(&fence);
 		break;
 	}
@@ -294,7 +299,7 @@ static int radeon_test_create_and_emit_fence(struct radeon_device *rdev,
 			return r;
 		}
 		radeon_fence_emit(rdev, fence, ring->idx);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_unlock_commit(rdev, ring, false);
 	}
 	return 0;
 }
@@ -319,7 +324,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence1);
 	if (r)
@@ -331,7 +336,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fence2);
 	if (r)
@@ -350,7 +355,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 
 	r = radeon_fence_wait(fence1, false);
 	if (r) {
@@ -371,7 +376,7 @@ void radeon_test_ring_sync(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 
 	r = radeon_fence_wait(fence2, false);
 	if (r) {
@@ -414,7 +419,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringA->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringA);
+	radeon_ring_unlock_commit(rdev, ringA, false);
 
 	r = radeon_test_create_and_emit_fence(rdev, ringA, &fenceA);
 	if (r)
@@ -426,7 +431,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_wait(rdev, ringB->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
+	radeon_ring_unlock_commit(rdev, ringB, false);
 	r = radeon_test_create_and_emit_fence(rdev, ringB, &fenceB);
 	if (r)
 		goto out_cleanup;
@@ -448,7 +453,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
+	radeon_ring_unlock_commit(rdev, ringC, false);
 
 	for (i = 0; i < 30; ++i) {
 		mdelay(100);
@@ -474,7 +479,7 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev,
 		goto out_cleanup;
 	}
 	radeon_semaphore_emit_signal(rdev, ringC->idx, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
+	radeon_ring_unlock_commit(rdev, ringC, false);
 
 	mdelay(1000);
 
diff --git a/drivers/gpu/drm/radeon/radeon_trace.h b/drivers/gpu/drm/radeon/radeon_trace.h
index f749f2c3bbdb..9db74a96ef61 100644
--- a/drivers/gpu/drm/radeon/radeon_trace.h
+++ b/drivers/gpu/drm/radeon/radeon_trace.h
@@ -72,8 +72,8 @@ TRACE_EVENT(radeon_vm_bo_update,
 			     ),
 
 	    TP_fast_assign(
-			   __entry->soffset = bo_va->soffset;
-			   __entry->eoffset = bo_va->eoffset;
+			   __entry->soffset = bo_va->it.start;
+			   __entry->eoffset = bo_va->it.last + 1;
 			   __entry->flags = bo_va->flags;
 			   ),
 	    TP_printk("soffs=%010llx, eoffs=%010llx, flags=%08x",
@@ -104,6 +104,24 @@ TRACE_EVENT(radeon_vm_set_page,
 		      __entry->flags, __entry->count)
 );
 
+TRACE_EVENT(radeon_vm_flush,
+	    TP_PROTO(uint64_t pd_addr, unsigned ring, unsigned id),
+	    TP_ARGS(pd_addr, ring, id),
+	    TP_STRUCT__entry(
+			     __field(u64, pd_addr)
+			     __field(u32, ring)
+			     __field(u32, id)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pd_addr = pd_addr;
+			   __entry->ring = ring;
+			   __entry->id = id;
+			   ),
+	    TP_printk("pd_addr=%010Lx, ring=%u, id=%u",
+		      __entry->pd_addr, __entry->ring, __entry->id)
+);
+
 DECLARE_EVENT_CLASS(radeon_fence_request,
 
 	    TP_PROTO(struct drm_device *dev, int ring, u32 seqno),
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index c8a8a5144ec1..8624979afb65 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -39,6 +39,8 @@
 #include <linux/seq_file.h>
 #include <linux/slab.h>
 #include <linux/swiotlb.h>
+#include <linux/swap.h>
+#include <linux/pagemap.h>
 #include <linux/debugfs.h>
 #include "radeon_reg.h"
 #include "radeon.h"
@@ -176,12 +178,15 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 static void radeon_evict_flags(struct ttm_buffer_object *bo,
 				struct ttm_placement *placement)
 {
+	static struct ttm_place placements = {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
+	};
+
 	struct radeon_bo *rbo;
-	static u32 placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
 
 	if (!radeon_ttm_bo_is_radeon_bo(bo)) {
-		placement->fpfn = 0;
-		placement->lpfn = 0;
 		placement->placement = &placements;
 		placement->busy_placement = &placements;
 		placement->num_placement = 1;
@@ -228,6 +233,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	struct radeon_device *rdev;
 	uint64_t old_start, new_start;
 	struct radeon_fence *fence;
+	unsigned num_pages;
 	int r, ridx;
 
 	rdev = radeon_get_rdev(bo->bdev);
@@ -264,13 +270,12 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 
 	BUILD_BUG_ON((PAGE_SIZE % RADEON_GPU_PAGE_SIZE) != 0);
 
-	/* sync other rings */
-	fence = bo->sync_obj;
-	r = radeon_copy(rdev, old_start, new_start,
-			new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE), /* GPU pages */
-			&fence);
-	/* FIXME: handle copy error */
-	r = ttm_bo_move_accel_cleanup(bo, (void *)fence,
+	num_pages = new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
+	fence = radeon_copy(rdev, old_start, new_start, num_pages, bo->resv);
+	if (IS_ERR(fence))
+		return PTR_ERR(fence);
+
+	r = ttm_bo_move_accel_cleanup(bo, &fence->base,
 				      evict, no_wait_gpu, new_mem);
 	radeon_fence_unref(&fence);
 	return r;
@@ -284,20 +289,20 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	struct radeon_device *rdev;
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	struct ttm_mem_reg tmp_mem;
-	u32 placements;
+	struct ttm_place placements;
 	struct ttm_placement placement;
 	int r;
 
 	rdev = radeon_get_rdev(bo->bdev);
 	tmp_mem = *new_mem;
 	tmp_mem.mm_node = NULL;
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 1;
 	placement.placement = &placements;
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
-	placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.fpfn = 0;
+	placements.lpfn = 0;
+	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
 	if (unlikely(r)) {
@@ -332,19 +337,19 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	struct ttm_mem_reg tmp_mem;
 	struct ttm_placement placement;
-	u32 placements;
+	struct ttm_place placements;
 	int r;
 
 	rdev = radeon_get_rdev(bo->bdev);
 	tmp_mem = *new_mem;
 	tmp_mem.mm_node = NULL;
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 1;
 	placement.placement = &placements;
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
-	placements = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.fpfn = 0;
+	placements.lpfn = 0;
+	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
 	if (unlikely(r)) {
@@ -483,53 +488,131 @@ static void radeon_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_re
 {
 }
 
-static int radeon_sync_obj_wait(void *sync_obj, bool lazy, bool interruptible)
-{
-	return radeon_fence_wait((struct radeon_fence *)sync_obj, interruptible);
-}
+/*
+ * TTM backend functions.
+ */
+struct radeon_ttm_tt {
+	struct ttm_dma_tt		ttm;
+	struct radeon_device		*rdev;
+	u64				offset;
 
-static int radeon_sync_obj_flush(void *sync_obj)
+	uint64_t			userptr;
+	struct mm_struct		*usermm;
+	uint32_t			userflags;
+};
+
+/* prepare the sg table with the user pages */
+static int radeon_ttm_tt_pin_userptr(struct ttm_tt *ttm)
 {
+	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_ttm_tt *gtt = (void *)ttm;
+	unsigned pinned = 0, nents;
+	int r;
+
+	int write = !(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+	enum dma_data_direction direction = write ?
+		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+
+	if (current->mm != gtt->usermm)
+		return -EPERM;
+
+	if (gtt->userflags & RADEON_GEM_USERPTR_ANONONLY) {
+		/* check that we only pin down anonymous memory
+		   to prevent problems with writeback */
+		unsigned long end = gtt->userptr + ttm->num_pages * PAGE_SIZE;
+		struct vm_area_struct *vma;
+		vma = find_vma(gtt->usermm, gtt->userptr);
+		if (!vma || vma->vm_file || vma->vm_end < end)
+			return -EPERM;
+	}
+
+	do {
+		unsigned num_pages = ttm->num_pages - pinned;
+		uint64_t userptr = gtt->userptr + pinned * PAGE_SIZE;
+		struct page **pages = ttm->pages + pinned;
+
+		r = get_user_pages(current, current->mm, userptr, num_pages,
+				   write, 0, pages, NULL);
+		if (r < 0)
+			goto release_pages;
+
+		pinned += r;
+
+	} while (pinned < ttm->num_pages);
+
+	r = sg_alloc_table_from_pages(ttm->sg, ttm->pages, ttm->num_pages, 0,
+				      ttm->num_pages << PAGE_SHIFT,
+				      GFP_KERNEL);
+	if (r)
+		goto release_sg;
+
+	r = -ENOMEM;
+	nents = dma_map_sg(rdev->dev, ttm->sg->sgl, ttm->sg->nents, direction);
+	if (nents != ttm->sg->nents)
+		goto release_sg;
+
+	drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
+					 gtt->ttm.dma_address, ttm->num_pages);
+
 	return 0;
-}
 
-static void radeon_sync_obj_unref(void **sync_obj)
-{
-	radeon_fence_unref((struct radeon_fence **)sync_obj);
-}
+release_sg:
+	kfree(ttm->sg);
 
-static void *radeon_sync_obj_ref(void *sync_obj)
-{
-	return radeon_fence_ref((struct radeon_fence *)sync_obj);
+release_pages:
+	release_pages(ttm->pages, pinned, 0);
+	return r;
 }
 
-static bool radeon_sync_obj_signaled(void *sync_obj)
+static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
 {
-	return radeon_fence_signaled((struct radeon_fence *)sync_obj);
-}
+	struct radeon_device *rdev = radeon_get_rdev(ttm->bdev);
+	struct radeon_ttm_tt *gtt = (void *)ttm;
+	struct scatterlist *sg;
+	int i;
 
-/*
- * TTM backend functions.
- */
-struct radeon_ttm_tt {
-	struct ttm_dma_tt		ttm;
-	struct radeon_device		*rdev;
-	u64				offset;
-};
+	int write = !(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+	enum dma_data_direction direction = write ?
+		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+
+	/* free the sg table and pages again */
+	dma_unmap_sg(rdev->dev, ttm->sg->sgl, ttm->sg->nents, direction);
+
+	for_each_sg(ttm->sg->sgl, sg, ttm->sg->nents, i) {
+		struct page *page = sg_page(sg);
+
+		if (!(gtt->userflags & RADEON_GEM_USERPTR_READONLY))
+			set_page_dirty(page);
+
+		mark_page_accessed(page);
+		page_cache_release(page);
+	}
+
+	sg_free_table(ttm->sg);
+}
 
 static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
 				   struct ttm_mem_reg *bo_mem)
 {
 	struct radeon_ttm_tt *gtt = (void*)ttm;
+	uint32_t flags = RADEON_GART_PAGE_VALID | RADEON_GART_PAGE_READ |
+		RADEON_GART_PAGE_WRITE;
 	int r;
 
+	if (gtt->userptr) {
+		radeon_ttm_tt_pin_userptr(ttm);
+		flags &= ~RADEON_GART_PAGE_WRITE;
+	}
+
 	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
 	if (!ttm->num_pages) {
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
 	}
-	r = radeon_gart_bind(gtt->rdev, gtt->offset,
-			     ttm->num_pages, ttm->pages, gtt->ttm.dma_address);
+	if (ttm->caching_state == tt_cached)
+		flags |= RADEON_GART_PAGE_SNOOP;
+	r = radeon_gart_bind(gtt->rdev, gtt->offset, ttm->num_pages,
+			     ttm->pages, gtt->ttm.dma_address, flags);
 	if (r) {
 		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
 			  ttm->num_pages, (unsigned)gtt->offset);
@@ -543,6 +626,10 @@ static int radeon_ttm_backend_unbind(struct ttm_tt *ttm)
 	struct radeon_ttm_tt *gtt = (void *)ttm;
 
 	radeon_gart_unbind(gtt->rdev, gtt->offset, ttm->num_pages);
+
+	if (gtt->userptr)
+		radeon_ttm_tt_unpin_userptr(ttm);
+
 	return 0;
 }
 
@@ -588,10 +675,17 @@ static struct ttm_tt *radeon_ttm_tt_create(struct ttm_bo_device *bdev,
 	return &gtt->ttm.ttm;
 }
 
+static struct radeon_ttm_tt *radeon_ttm_tt_to_gtt(struct ttm_tt *ttm)
+{
+	if (!ttm || ttm->func != &radeon_backend_func)
+		return NULL;
+	return (struct radeon_ttm_tt *)ttm;
+}
+
 static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 {
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
 	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
 	unsigned i;
 	int r;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
@@ -599,6 +693,16 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 	if (ttm->state != tt_unpopulated)
 		return 0;
 
+	if (gtt && gtt->userptr) {
+		ttm->sg = kcalloc(1, sizeof(struct sg_table), GFP_KERNEL);
+		if (!ttm->sg)
+			return -ENOMEM;
+
+		ttm->page_flags |= TTM_PAGE_FLAG_SG;
+		ttm->state = tt_unbound;
+		return 0;
+	}
+
 	if (slave && ttm->sg) {
 		drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages,
 						 gtt->ttm.dma_address, ttm->num_pages);
@@ -644,10 +748,16 @@ static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
 static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 {
 	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
 	unsigned i;
 	bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG);
 
+	if (gtt && gtt->userptr) {
+		kfree(ttm->sg);
+		ttm->page_flags &= ~TTM_PAGE_FLAG_SG;
+		return;
+	}
+
 	if (slave)
 		return;
 
@@ -676,6 +786,40 @@ static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
 	ttm_pool_unpopulate(ttm);
 }
 
+int radeon_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
+			      uint32_t flags)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return -EINVAL;
+
+	gtt->userptr = addr;
+	gtt->usermm = current->mm;
+	gtt->userflags = flags;
+	return 0;
+}
+
+bool radeon_ttm_tt_has_userptr(struct ttm_tt *ttm)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return false;
+
+	return !!gtt->userptr;
+}
+
+bool radeon_ttm_tt_is_readonly(struct ttm_tt *ttm)
+{
+	struct radeon_ttm_tt *gtt = radeon_ttm_tt_to_gtt(ttm);
+
+	if (gtt == NULL)
+		return false;
+
+	return !!(gtt->userflags & RADEON_GEM_USERPTR_READONLY);
+}
+
 static struct ttm_bo_driver radeon_bo_driver = {
 	.ttm_tt_create = &radeon_ttm_tt_create,
 	.ttm_tt_populate = &radeon_ttm_tt_populate,
@@ -685,11 +829,6 @@ static struct ttm_bo_driver radeon_bo_driver = {
 	.evict_flags = &radeon_evict_flags,
 	.move = &radeon_bo_move,
 	.verify_access = &radeon_verify_access,
-	.sync_obj_signaled = &radeon_sync_obj_signaled,
-	.sync_obj_wait = &radeon_sync_obj_wait,
-	.sync_obj_flush = &radeon_sync_obj_flush,
-	.sync_obj_unref = &radeon_sync_obj_unref,
-	.sync_obj_ref = &radeon_sync_obj_ref,
 	.move_notify = &radeon_bo_move_notify,
 	.fault_reserve_notify = &radeon_bo_fault_reserve_notify,
 	.io_mem_reserve = &radeon_ttm_io_mem_reserve,
@@ -726,7 +865,7 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
 
 	r = radeon_bo_create(rdev, 256 * 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM,
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
 			     NULL, &rdev->stollen_vga_memory);
 	if (r) {
 		return r;
@@ -824,7 +963,7 @@ int radeon_mmap(struct file *filp, struct vm_area_struct *vma)
 	int r;
 
 	if (unlikely(vma->vm_pgoff < DRM_FILE_PAGE_OFFSET)) {
-		return drm_mmap(filp, vma);
+		return -EINVAL;
 	}
 
 	file_priv = filp->private_data;
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.c b/drivers/gpu/drm/radeon/radeon_ucode.c
new file mode 100644
index 000000000000..6beec680390c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ucode.c
@@ -0,0 +1,167 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <drm/drmP.h>
+#include "radeon.h"
+#include "radeon_ucode.h"
+
+static void radeon_ucode_print_common_hdr(const struct common_firmware_header *hdr)
+{
+	DRM_DEBUG("size_bytes: %u\n", le32_to_cpu(hdr->size_bytes));
+	DRM_DEBUG("header_size_bytes: %u\n", le32_to_cpu(hdr->header_size_bytes));
+	DRM_DEBUG("header_version_major: %u\n", le16_to_cpu(hdr->header_version_major));
+	DRM_DEBUG("header_version_minor: %u\n", le16_to_cpu(hdr->header_version_minor));
+	DRM_DEBUG("ip_version_major: %u\n", le16_to_cpu(hdr->ip_version_major));
+	DRM_DEBUG("ip_version_minor: %u\n", le16_to_cpu(hdr->ip_version_minor));
+	DRM_DEBUG("ucode_version: 0x%08x\n", le32_to_cpu(hdr->ucode_version));
+	DRM_DEBUG("ucode_size_bytes: %u\n", le32_to_cpu(hdr->ucode_size_bytes));
+	DRM_DEBUG("ucode_array_offset_bytes: %u\n",
+		  le32_to_cpu(hdr->ucode_array_offset_bytes));
+	DRM_DEBUG("crc32: 0x%08x\n", le32_to_cpu(hdr->crc32));
+}
+
+void radeon_ucode_print_mc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("MC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct mc_firmware_header_v1_0 *mc_hdr =
+			container_of(hdr, struct mc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("io_debug_size_bytes: %u\n",
+			  le32_to_cpu(mc_hdr->io_debug_size_bytes));
+		DRM_DEBUG("io_debug_array_offset_bytes: %u\n",
+			  le32_to_cpu(mc_hdr->io_debug_array_offset_bytes));
+	} else {
+		DRM_ERROR("Unknown MC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_smc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("SMC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct smc_firmware_header_v1_0 *smc_hdr =
+			container_of(hdr, struct smc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_start_addr: %u\n", le32_to_cpu(smc_hdr->ucode_start_addr));
+	} else {
+		DRM_ERROR("Unknown SMC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_gfx_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("GFX\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct gfx_firmware_header_v1_0 *gfx_hdr =
+			container_of(hdr, struct gfx_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(gfx_hdr->ucode_feature_version));
+		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(gfx_hdr->jt_offset));
+		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(gfx_hdr->jt_size));
+	} else {
+		DRM_ERROR("Unknown GFX ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_rlc_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("RLC\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct rlc_firmware_header_v1_0 *rlc_hdr =
+			container_of(hdr, struct rlc_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(rlc_hdr->ucode_feature_version));
+		DRM_DEBUG("save_and_restore_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->save_and_restore_offset));
+		DRM_DEBUG("clear_state_descriptor_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->clear_state_descriptor_offset));
+		DRM_DEBUG("avail_scratch_ram_locations: %u\n",
+			  le32_to_cpu(rlc_hdr->avail_scratch_ram_locations));
+		DRM_DEBUG("master_pkt_description_offset: %u\n",
+			  le32_to_cpu(rlc_hdr->master_pkt_description_offset));
+	} else {
+		DRM_ERROR("Unknown RLC ucode version: %u.%u\n", version_major, version_minor);
+	}
+}
+
+void radeon_ucode_print_sdma_hdr(const struct common_firmware_header *hdr)
+{
+	uint16_t version_major = le16_to_cpu(hdr->header_version_major);
+	uint16_t version_minor = le16_to_cpu(hdr->header_version_minor);
+
+	DRM_DEBUG("SDMA\n");
+	radeon_ucode_print_common_hdr(hdr);
+
+	if (version_major == 1) {
+		const struct sdma_firmware_header_v1_0 *sdma_hdr =
+			container_of(hdr, struct sdma_firmware_header_v1_0, header);
+
+		DRM_DEBUG("ucode_feature_version: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_feature_version));
+		DRM_DEBUG("ucode_change_version: %u\n",
+			  le32_to_cpu(sdma_hdr->ucode_change_version));
+		DRM_DEBUG("jt_offset: %u\n", le32_to_cpu(sdma_hdr->jt_offset));
+		DRM_DEBUG("jt_size: %u\n", le32_to_cpu(sdma_hdr->jt_size));
+	} else {
+		DRM_ERROR("Unknown SDMA ucode version: %u.%u\n",
+			  version_major, version_minor);
+	}
+}
+
+int radeon_ucode_validate(const struct firmware *fw)
+{
+	const struct common_firmware_header *hdr =
+		(const struct common_firmware_header *)fw->data;
+
+	if (fw->size == le32_to_cpu(hdr->size_bytes))
+		return 0;
+
+	return -EINVAL;
+}
+
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h
index 4e7c3269b183..dc4576e4d8ad 100644
--- a/drivers/gpu/drm/radeon/radeon_ucode.h
+++ b/drivers/gpu/drm/radeon/radeon_ucode.h
@@ -153,4 +153,75 @@
 #define HAWAII_SMC_UCODE_START       0x20000
 #define HAWAII_SMC_UCODE_SIZE        0x1FDEC
 
+struct common_firmware_header {
+	uint32_t size_bytes; /* size of the entire header+image(s) in bytes */
+	uint32_t header_size_bytes; /* size of just the header in bytes */
+	uint16_t header_version_major; /* header version */
+	uint16_t header_version_minor; /* header version */
+	uint16_t ip_version_major; /* IP version */
+	uint16_t ip_version_minor; /* IP version */
+	uint32_t ucode_version;
+	uint32_t ucode_size_bytes; /* size of ucode in bytes */
+	uint32_t ucode_array_offset_bytes; /* payload offset from the start of the header */
+	uint32_t crc32;  /* crc32 checksum of the payload */
+};
+
+/* version_major=1, version_minor=0 */
+struct mc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t io_debug_size_bytes; /* size of debug array in dwords */
+	uint32_t io_debug_array_offset_bytes; /* payload offset from the start of the header */
+};
+
+/* version_major=1, version_minor=0 */
+struct smc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_start_addr;
+};
+
+/* version_major=1, version_minor=0 */
+struct gfx_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t jt_offset; /* jt location */
+	uint32_t jt_size;  /* size of jt */
+};
+
+/* version_major=1, version_minor=0 */
+struct rlc_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t save_and_restore_offset;
+	uint32_t clear_state_descriptor_offset;
+	uint32_t avail_scratch_ram_locations;
+	uint32_t master_pkt_description_offset;
+};
+
+/* version_major=1, version_minor=0 */
+struct sdma_firmware_header_v1_0 {
+	struct common_firmware_header header;
+	uint32_t ucode_feature_version;
+	uint32_t ucode_change_version;
+	uint32_t jt_offset; /* jt location */
+	uint32_t jt_size; /* size of jt */
+};
+
+/* header is fixed size */
+union radeon_firmware_header {
+	struct common_firmware_header common;
+	struct mc_firmware_header_v1_0 mc;
+	struct smc_firmware_header_v1_0 smc;
+	struct gfx_firmware_header_v1_0 gfx;
+	struct rlc_firmware_header_v1_0 rlc;
+	struct sdma_firmware_header_v1_0 sdma;
+	uint8_t raw[0x100];
+};
+
+void radeon_ucode_print_mc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_smc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_gfx_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_rlc_hdr(const struct common_firmware_header *hdr);
+void radeon_ucode_print_sdma_hdr(const struct common_firmware_header *hdr);
+int radeon_ucode_validate(const struct firmware *fw);
+
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index a4ad270e8261..11b662469253 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -40,12 +40,18 @@
 #define UVD_IDLE_TIMEOUT_MS	1000
 
 /* Firmware Names */
+#define FIRMWARE_R600		"radeon/R600_uvd.bin"
+#define FIRMWARE_RS780		"radeon/RS780_uvd.bin"
+#define FIRMWARE_RV770		"radeon/RV770_uvd.bin"
 #define FIRMWARE_RV710		"radeon/RV710_uvd.bin"
 #define FIRMWARE_CYPRESS	"radeon/CYPRESS_uvd.bin"
 #define FIRMWARE_SUMO		"radeon/SUMO_uvd.bin"
 #define FIRMWARE_TAHITI		"radeon/TAHITI_uvd.bin"
 #define FIRMWARE_BONAIRE	"radeon/BONAIRE_uvd.bin"
 
+MODULE_FIRMWARE(FIRMWARE_R600);
+MODULE_FIRMWARE(FIRMWARE_RS780);
+MODULE_FIRMWARE(FIRMWARE_RV770);
 MODULE_FIRMWARE(FIRMWARE_RV710);
 MODULE_FIRMWARE(FIRMWARE_CYPRESS);
 MODULE_FIRMWARE(FIRMWARE_SUMO);
@@ -63,6 +69,23 @@ int radeon_uvd_init(struct radeon_device *rdev)
 	INIT_DELAYED_WORK(&rdev->uvd.idle_work, radeon_uvd_idle_work_handler);
 
 	switch (rdev->family) {
+	case CHIP_RV610:
+	case CHIP_RV630:
+	case CHIP_RV670:
+	case CHIP_RV620:
+	case CHIP_RV635:
+		fw_name = FIRMWARE_R600;
+		break;
+
+	case CHIP_RS780:
+	case CHIP_RS880:
+		fw_name = FIRMWARE_RS780;
+		break;
+
+	case CHIP_RV770:
+		fw_name = FIRMWARE_RV770;
+		break;
+
 	case CHIP_RV710:
 	case CHIP_RV730:
 	case CHIP_RV740:
@@ -115,9 +138,11 @@ int radeon_uvd_init(struct radeon_device *rdev)
 	}
 
 	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
-		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
+		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE +
+		  RADEON_GPU_PAGE_SIZE;
 	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->uvd.vcpu_bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+			     NULL, &rdev->uvd.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate UVD bo\n", r);
 		return r;
@@ -231,10 +256,30 @@ int radeon_uvd_resume(struct radeon_device *rdev)
 	return 0;
 }
 
-void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo)
+void radeon_uvd_force_into_uvd_segment(struct radeon_bo *rbo,
+				       uint32_t allowed_domains)
 {
-	rbo->placement.fpfn = 0 >> PAGE_SHIFT;
-	rbo->placement.lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+	int i;
+
+	for (i = 0; i < rbo->placement.num_placement; ++i) {
+		rbo->placements[i].fpfn = 0 >> PAGE_SHIFT;
+		rbo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+	}
+
+	/* If it must be in VRAM it must be in the first segment as well */
+	if (allowed_domains == RADEON_GEM_DOMAIN_VRAM)
+		return;
+
+	/* abort if we already have more than one placement */
+	if (rbo->placement.num_placement > 1)
+		return;
+
+	/* add another 256MB segment */
+	rbo->placements[1] = rbo->placements[0];
+	rbo->placements[1].fpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
+	rbo->placements[1].lpfn += (256 * 1024 * 1024) >> PAGE_SHIFT;
+	rbo->placement.num_placement++;
+	rbo->placement.num_busy_placement++;
 }
 
 void radeon_uvd_free_handles(struct radeon_device *rdev, struct drm_file *filp)
@@ -356,6 +401,7 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 {
 	int32_t *msg, msg_type, handle;
 	unsigned img_size = 0;
+	struct fence *f;
 	void *ptr;
 
 	int i, r;
@@ -365,8 +411,9 @@ static int radeon_uvd_cs_msg(struct radeon_cs_parser *p, struct radeon_bo *bo,
 		return -EINVAL;
 	}
 
-	if (bo->tbo.sync_obj) {
-		r = radeon_fence_wait(bo->tbo.sync_obj, false);
+	f = reservation_object_get_excl(bo->tbo.resv);
+	if (f) {
+		r = radeon_fence_wait((struct radeon_fence *)f, false);
 		if (r) {
 			DRM_ERROR("Failed waiting for UVD message (%d)!\n", r);
 			return r;
@@ -604,38 +651,16 @@ int radeon_uvd_cs_parse(struct radeon_cs_parser *p)
 }
 
 static int radeon_uvd_send_msg(struct radeon_device *rdev,
-			       int ring, struct radeon_bo *bo,
+			       int ring, uint64_t addr,
 			       struct radeon_fence **fence)
 {
-	struct ttm_validate_buffer tv;
-	struct ww_acquire_ctx ticket;
-	struct list_head head;
 	struct radeon_ib ib;
-	uint64_t addr;
 	int i, r;
 
-	memset(&tv, 0, sizeof(tv));
-	tv.bo = &bo->tbo;
-
-	INIT_LIST_HEAD(&head);
-	list_add(&tv.head, &head);
-
-	r = ttm_eu_reserve_buffers(&ticket, &head);
-	if (r)
-		return r;
-
-	radeon_ttm_placement_from_domain(bo, RADEON_GEM_DOMAIN_VRAM);
-	radeon_uvd_force_into_uvd_segment(bo);
-
-	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-	if (r) 
-		goto err;
-
 	r = radeon_ib_get(rdev, ring, &ib, NULL, 64);
 	if (r)
-		goto err;
+		return r;
 
-	addr = radeon_bo_gpu_offset(bo);
 	ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
 	ib.ptr[1] = addr;
 	ib.ptr[2] = PACKET0(UVD_GPCOM_VCPU_DATA1, 0);
@@ -646,20 +671,12 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 		ib.ptr[i] = PACKET2(0);
 	ib.length_dw = 16;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
-	if (r)
-		goto err;
-	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 
 	if (fence)
 		*fence = radeon_fence_ref(ib.fence);
 
 	radeon_ib_free(rdev, &ib);
-	radeon_bo_unref(&bo);
-	return 0;
-
-err:
-	ttm_eu_backoff_reservation(&ticket, &head);
 	return r;
 }
 
@@ -669,27 +686,18 @@ err:
 int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 			      uint32_t handle, struct radeon_fence **fence)
 {
-	struct radeon_bo *bo;
-	uint32_t *msg;
-	int r, i;
+	/* we use the last page of the vcpu bo for the UVD message */
+	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
+		RADEON_GPU_PAGE_SIZE;
 
-	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
-	if (r)
-		return r;
+	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
-	r = radeon_bo_reserve(bo, false);
-	if (r) {
-		radeon_bo_unref(&bo);
-		return r;
-	}
+	int r, i;
 
-	r = radeon_bo_kmap(bo, (void **)&msg);
-	if (r) {
-		radeon_bo_unreserve(bo);
-		radeon_bo_unref(&bo);
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
+	if (r)
 		return r;
-	}
 
 	/* stitch together an UVD create msg */
 	msg[0] = cpu_to_le32(0x00000de4);
@@ -706,36 +714,26 @@ int radeon_uvd_get_create_msg(struct radeon_device *rdev, int ring,
 	for (i = 11; i < 1024; ++i)
 		msg[i] = cpu_to_le32(0x0);
 
-	radeon_bo_kunmap(bo);
-	radeon_bo_unreserve(bo);
-
-	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+	return r;
 }
 
 int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 			       uint32_t handle, struct radeon_fence **fence)
 {
-	struct radeon_bo *bo;
-	uint32_t *msg;
-	int r, i;
+	/* we use the last page of the vcpu bo for the UVD message */
+	uint64_t offs = radeon_bo_size(rdev->uvd.vcpu_bo) -
+		RADEON_GPU_PAGE_SIZE;
 
-	r = radeon_bo_create(rdev, 1024, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &bo);
-	if (r)
-		return r;
+	uint32_t *msg = rdev->uvd.cpu_addr + offs;
+	uint64_t addr = rdev->uvd.gpu_addr + offs;
 
-	r = radeon_bo_reserve(bo, false);
-	if (r) {
-		radeon_bo_unref(&bo);
-		return r;
-	}
+	int r, i;
 
-	r = radeon_bo_kmap(bo, (void **)&msg);
-	if (r) {
-		radeon_bo_unreserve(bo);
-		radeon_bo_unref(&bo);
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, true);
+	if (r)
 		return r;
-	}
 
 	/* stitch together an UVD destroy msg */
 	msg[0] = cpu_to_le32(0x00000de4);
@@ -745,10 +743,9 @@ int radeon_uvd_get_destroy_msg(struct radeon_device *rdev, int ring,
 	for (i = 4; i < 1024; ++i)
 		msg[i] = cpu_to_le32(0x0);
 
-	radeon_bo_kunmap(bo);
-	radeon_bo_unreserve(bo);
-
-	return radeon_uvd_send_msg(rdev, ring, bo, fence);
+	r = radeon_uvd_send_msg(rdev, ring, addr, fence);
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+	return r;
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c
index aa21c31a846c..9e85757d5599 100644
--- a/drivers/gpu/drm/radeon/radeon_vce.c
+++ b/drivers/gpu/drm/radeon/radeon_vce.c
@@ -126,7 +126,8 @@ int radeon_vce_init(struct radeon_device *rdev)
 	size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) +
 	       RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE;
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL, NULL,
+			     &rdev->vce.vcpu_bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r);
 		return r;
@@ -368,7 +369,7 @@ int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring,
 	for (i = ib.length_dw; i < ib_size_dw; ++i)
 		ib.ptr[i] = 0x0;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 	        DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 	}
@@ -425,7 +426,7 @@ int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring,
 	for (i = ib.length_dw; i < ib_size_dw; ++i)
 		ib.ptr[i] = 0x0;
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 	        DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
 	}
@@ -715,7 +716,7 @@ int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	radeon_ring_write(ring, VCE_CMD_END);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = 0; i < rdev->usec_timeout; i++) {
 	        if (vce_v1_0_get_rptr(rdev, ring) != rptr)
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 725d3669014f..dfde266529e2 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -132,8 +132,8 @@ struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
 	struct radeon_cs_reloc *list;
 	unsigned i, idx;
 
-	list = kmalloc_array(vm->max_pde_used + 2,
-			     sizeof(struct radeon_cs_reloc), GFP_KERNEL);
+	list = drm_malloc_ab(vm->max_pde_used + 2,
+			     sizeof(struct radeon_cs_reloc));
 	if (!list)
 		return NULL;
 
@@ -143,6 +143,7 @@ struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
 	list[0].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 	list[0].tv.bo = &vm->page_directory->tbo;
+	list[0].tv.shared = false;
 	list[0].tiling_flags = 0;
 	list[0].handle = 0;
 	list_add(&list[0].tv.head, head);
@@ -156,6 +157,7 @@ struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev,
 		list[idx].prefered_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].allowed_domains = RADEON_GEM_DOMAIN_VRAM;
 		list[idx].tv.bo = &list[idx].robj->tbo;
+		list[idx].tv.shared = false;
 		list[idx].tiling_flags = 0;
 		list[idx].handle = 0;
 		list_add(&list[idx++].tv.head, head);
@@ -238,8 +240,8 @@ void radeon_vm_flush(struct radeon_device *rdev,
 	uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory);
 
 	/* if we can't remember our last VM flush then flush now! */
-	/* XXX figure out why we have to flush all the time */
-	if (!vm->last_flush || true || pd_addr != vm->pd_gpu_addr) {
+	if (!vm->last_flush || pd_addr != vm->pd_gpu_addr) {
+		trace_radeon_vm_flush(pd_addr, ring, vm->id);
 		vm->pd_gpu_addr = pd_addr;
 		radeon_ring_vm_flush(rdev, ring, vm);
 	}
@@ -325,17 +327,15 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
 	}
 	bo_va->vm = vm;
 	bo_va->bo = bo;
-	bo_va->soffset = 0;
-	bo_va->eoffset = 0;
+	bo_va->it.start = 0;
+	bo_va->it.last = 0;
 	bo_va->flags = 0;
-	bo_va->valid = false;
+	bo_va->addr = 0;
 	bo_va->ref_count = 1;
 	INIT_LIST_HEAD(&bo_va->bo_list);
-	INIT_LIST_HEAD(&bo_va->vm_list);
 	INIT_LIST_HEAD(&bo_va->vm_status);
 
 	mutex_lock(&vm->mutex);
-	list_add(&bo_va->vm_list, &vm->va);
 	list_add_tail(&bo_va->bo_list, &bo->va);
 	mutex_unlock(&vm->mutex);
 
@@ -343,6 +343,42 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev,
 }
 
 /**
+ * radeon_vm_set_pages - helper to call the right asic function
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: hw access flags
+ *
+ * Traces the parameters and calls the right asic functions
+ * to setup the page table using the DMA.
+ */
+static void radeon_vm_set_pages(struct radeon_device *rdev,
+				struct radeon_ib *ib,
+				uint64_t pe,
+				uint64_t addr, unsigned count,
+				uint32_t incr, uint32_t flags)
+{
+	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
+
+	if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
+		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
+		radeon_asic_vm_copy_pages(rdev, ib, pe, src, count);
+
+	} else if ((flags & R600_PTE_SYSTEM) || (count < 3)) {
+		radeon_asic_vm_write_pages(rdev, ib, pe, addr,
+					   count, incr, flags);
+
+	} else {
+		radeon_asic_vm_set_pages(rdev, ib, pe, addr,
+					 count, incr, flags);
+	}
+}
+
+/**
  * radeon_vm_clear_bo - initially clear the page dir/table
  *
  * @rdev: radeon_device pointer
@@ -361,11 +397,12 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
 
         memset(&tv, 0, sizeof(tv));
         tv.bo = &bo->tbo;
+	tv.shared = false;
 
         INIT_LIST_HEAD(&head);
         list_add(&tv.head, &head);
 
-        r = ttm_eu_reserve_buffers(&ticket, &head);
+        r = ttm_eu_reserve_buffers(&ticket, &head, true);
         if (r)
 		return r;
 
@@ -376,20 +413,21 @@ static int radeon_vm_clear_bo(struct radeon_device *rdev,
 	addr = radeon_bo_gpu_offset(bo);
 	entries = radeon_bo_size(bo) / 8;
 
-	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib,
-			  NULL, entries * 2 + 64);
+	r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, 256);
 	if (r)
                 goto error;
 
 	ib.length_dw = 0;
 
-	radeon_asic_vm_set_page(rdev, &ib, addr, 0, entries, 0, 0);
+	radeon_vm_set_pages(rdev, &ib, addr, 0, entries, 0, 0);
+	radeon_asic_vm_pad_ib(rdev, &ib);
+	WARN_ON(ib.length_dw > 64);
 
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r)
                 goto error;
 
-	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
+	ttm_eu_fence_buffer_objects(&ticket, &head, &ib.fence->base);
 	radeon_ib_free(rdev, &ib);
 
 	return 0;
@@ -419,11 +457,9 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 			  uint32_t flags)
 {
 	uint64_t size = radeon_bo_size(bo_va->bo);
-	uint64_t eoffset, last_offset = 0;
 	struct radeon_vm *vm = bo_va->vm;
-	struct radeon_bo_va *tmp;
-	struct list_head *head;
 	unsigned last_pfn, pt_idx;
+	uint64_t eoffset;
 	int r;
 
 	if (soffset) {
@@ -445,51 +481,53 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 	}
 
 	mutex_lock(&vm->mutex);
-	head = &vm->va;
-	last_offset = 0;
-	list_for_each_entry(tmp, &vm->va, vm_list) {
-		if (bo_va == tmp) {
-			/* skip over currently modified bo */
-			continue;
+	if (bo_va->it.start || bo_va->it.last) {
+		if (bo_va->addr) {
+			/* add a clone of the bo_va to clear the old address */
+			struct radeon_bo_va *tmp;
+			tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
+			if (!tmp) {
+				mutex_unlock(&vm->mutex);
+				return -ENOMEM;
+			}
+			tmp->it.start = bo_va->it.start;
+			tmp->it.last = bo_va->it.last;
+			tmp->vm = vm;
+			tmp->addr = bo_va->addr;
+			tmp->bo = radeon_bo_ref(bo_va->bo);
+			list_add(&tmp->vm_status, &vm->freed);
 		}
 
-		if (soffset >= last_offset && eoffset <= tmp->soffset) {
-			/* bo can be added before this one */
-			break;
-		}
-		if (eoffset > tmp->soffset && soffset < tmp->eoffset) {
-			/* bo and tmp overlap, invalid offset */
-			dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n",
-				bo_va->bo, (unsigned)bo_va->soffset, tmp->bo,
-				(unsigned)tmp->soffset, (unsigned)tmp->eoffset);
-			mutex_unlock(&vm->mutex);
-			return -EINVAL;
-		}
-		last_offset = tmp->eoffset;
-		head = &tmp->vm_list;
+		interval_tree_remove(&bo_va->it, &vm->va);
+		bo_va->it.start = 0;
+		bo_va->it.last = 0;
 	}
 
-	if (bo_va->soffset) {
-		/* add a clone of the bo_va to clear the old address */
-		tmp = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
-		if (!tmp) {
+	soffset /= RADEON_GPU_PAGE_SIZE;
+	eoffset /= RADEON_GPU_PAGE_SIZE;
+	if (soffset || eoffset) {
+		struct interval_tree_node *it;
+		it = interval_tree_iter_first(&vm->va, soffset, eoffset - 1);
+		if (it) {
+			struct radeon_bo_va *tmp;
+			tmp = container_of(it, struct radeon_bo_va, it);
+			/* bo and tmp overlap, invalid offset */
+			dev_err(rdev->dev, "bo %p va 0x%010Lx conflict with "
+				"(bo %p 0x%010lx 0x%010lx)\n", bo_va->bo,
+				soffset, tmp->bo, tmp->it.start, tmp->it.last);
 			mutex_unlock(&vm->mutex);
-			return -ENOMEM;
+			return -EINVAL;
 		}
-		tmp->soffset = bo_va->soffset;
-		tmp->eoffset = bo_va->eoffset;
-		tmp->vm = vm;
-		list_add(&tmp->vm_status, &vm->freed);
+		bo_va->it.start = soffset;
+		bo_va->it.last = eoffset - 1;
+		interval_tree_insert(&bo_va->it, &vm->va);
 	}
 
-	bo_va->soffset = soffset;
-	bo_va->eoffset = eoffset;
 	bo_va->flags = flags;
-	bo_va->valid = false;
-	list_move(&bo_va->vm_list, head);
+	bo_va->addr = 0;
 
-	soffset = (soffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size;
-	eoffset = (eoffset / RADEON_GPU_PAGE_SIZE) >> radeon_vm_block_size;
+	soffset >>= radeon_vm_block_size;
+	eoffset >>= radeon_vm_block_size;
 
 	BUG_ON(eoffset >= radeon_vm_num_pdes(rdev));
 
@@ -510,7 +548,8 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev,
 
 		r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8,
 				     RADEON_GPU_PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_VRAM, NULL, &pt);
+				     RADEON_GEM_DOMAIN_VRAM, 0,
+				     NULL, NULL, &pt);
 		if (r)
 			return r;
 
@@ -611,7 +650,7 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 	ndw = 64;
 
 	/* assume the worst case */
-	ndw += vm->max_pde_used * 16;
+	ndw += vm->max_pde_used * 6;
 
 	/* update too big for an IB */
 	if (ndw > 0xfffff)
@@ -640,9 +679,9 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 		    ((last_pt + incr * count) != pt)) {
 
 			if (count) {
-				radeon_asic_vm_set_page(rdev, &ib, last_pde,
-							last_pt, count, incr,
-							R600_PTE_VALID);
+				radeon_vm_set_pages(rdev, &ib, last_pde,
+						    last_pt, count, incr,
+						    R600_PTE_VALID);
 			}
 
 			count = 1;
@@ -654,13 +693,16 @@ int radeon_vm_update_page_directory(struct radeon_device *rdev,
 	}
 
 	if (count)
-		radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count,
-					incr, R600_PTE_VALID);
+		radeon_vm_set_pages(rdev, &ib, last_pde, last_pt, count,
+				    incr, R600_PTE_VALID);
 
 	if (ib.length_dw != 0) {
-		radeon_semaphore_sync_to(ib.semaphore, pd->tbo.sync_obj);
-		radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use);
-		r = radeon_ib_schedule(rdev, &ib, NULL);
+		radeon_asic_vm_pad_ib(rdev, &ib);
+
+		radeon_semaphore_sync_resv(rdev, ib.semaphore, pd->tbo.resv, false);
+		radeon_semaphore_sync_fence(ib.semaphore, vm->last_id_use);
+		WARN_ON(ib.length_dw > ndw);
+		r = radeon_ib_schedule(rdev, &ib, NULL, false);
 		if (r) {
 			radeon_ib_free(rdev, &ib);
 			return r;
@@ -725,30 +767,30 @@ static void radeon_vm_frag_ptes(struct radeon_device *rdev,
 	    (frag_start >= frag_end)) {
 
 		count = (pe_end - pe_start) / 8;
-		radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 		return;
 	}
 
 	/* handle the 4K area at the beginning */
 	if (pe_start != frag_start) {
 		count = (frag_start - pe_start) / 8;
-		radeon_asic_vm_set_page(rdev, ib, pe_start, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, pe_start, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 		addr += RADEON_GPU_PAGE_SIZE * count;
 	}
 
 	/* handle the area in the middle */
 	count = (frag_end - frag_start) / 8;
-	radeon_asic_vm_set_page(rdev, ib, frag_start, addr, count,
-				RADEON_GPU_PAGE_SIZE, flags | frag_flags);
+	radeon_vm_set_pages(rdev, ib, frag_start, addr, count,
+			    RADEON_GPU_PAGE_SIZE, flags | frag_flags);
 
 	/* handle the 4K area at the end */
 	if (frag_end != pe_end) {
 		addr += RADEON_GPU_PAGE_SIZE * count;
 		count = (pe_end - frag_end) / 8;
-		radeon_asic_vm_set_page(rdev, ib, frag_end, addr, count,
-					RADEON_GPU_PAGE_SIZE, flags);
+		radeon_vm_set_pages(rdev, ib, frag_end, addr, count,
+				    RADEON_GPU_PAGE_SIZE, flags);
 	}
 }
 
@@ -777,9 +819,6 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 	unsigned count = 0;
 	uint64_t addr;
 
-	start = start / RADEON_GPU_PAGE_SIZE;
-	end = end / RADEON_GPU_PAGE_SIZE;
-
 	/* walk over the address space and update the page tables */
 	for (addr = start; addr < end; ) {
 		uint64_t pt_idx = addr >> radeon_vm_block_size;
@@ -787,7 +826,7 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev,
 		unsigned nptes;
 		uint64_t pte;
 
-		radeon_semaphore_sync_to(ib->semaphore, pt->tbo.sync_obj);
+		radeon_semaphore_sync_resv(rdev, ib->semaphore, pt->tbo.resv, false);
 
 		if ((addr & ~mask) == (end & ~mask))
 			nptes = end - addr;
@@ -842,55 +881,76 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 {
 	struct radeon_vm *vm = bo_va->vm;
 	struct radeon_ib ib;
-	unsigned nptes, ndw;
+	unsigned nptes, ncmds, ndw;
 	uint64_t addr;
+	uint32_t flags;
 	int r;
 
-
-	if (!bo_va->soffset) {
+	if (!bo_va->it.start) {
 		dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n",
 			bo_va->bo, vm);
 		return -EINVAL;
 	}
 
-	if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL))
-		return 0;
+	list_del_init(&bo_va->vm_status);
 
 	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
 	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
+	bo_va->flags &= ~RADEON_VM_PAGE_SNOOPED;
+	if (bo_va->bo && radeon_ttm_tt_is_readonly(bo_va->bo->tbo.ttm))
+		bo_va->flags &= ~RADEON_VM_PAGE_WRITEABLE;
+
 	if (mem) {
 		addr = mem->start << PAGE_SHIFT;
 		if (mem->mem_type != TTM_PL_SYSTEM) {
 			bo_va->flags |= RADEON_VM_PAGE_VALID;
-			bo_va->valid = true;
 		}
 		if (mem->mem_type == TTM_PL_TT) {
 			bo_va->flags |= RADEON_VM_PAGE_SYSTEM;
+			if (!(bo_va->bo->flags & (RADEON_GEM_GTT_WC | RADEON_GEM_GTT_UC)))
+				bo_va->flags |= RADEON_VM_PAGE_SNOOPED;
+
 		} else {
 			addr += rdev->vm_manager.vram_base_offset;
 		}
 	} else {
 		addr = 0;
-		bo_va->valid = false;
 	}
 
+	if (addr == bo_va->addr)
+		return 0;
+	bo_va->addr = addr;
+
 	trace_radeon_vm_bo_update(bo_va);
 
-	nptes = (bo_va->eoffset - bo_va->soffset) / RADEON_GPU_PAGE_SIZE;
+	nptes = bo_va->it.last - bo_va->it.start + 1;
+
+	/* reserve space for one command every (1 << BLOCK_SIZE) entries
+	   or 2k dwords (whatever is smaller) */
+	ncmds = (nptes >> min(radeon_vm_block_size, 11)) + 1;
 
 	/* padding, etc. */
 	ndw = 64;
 
-	if (radeon_vm_block_size > 11)
-		/* reserve space for one header for every 2k dwords */
-		ndw += (nptes >> 11) * 4;
-	else
-		/* reserve space for one header for
-		    every (1 << BLOCK_SIZE) entries */
-		ndw += (nptes >> radeon_vm_block_size) * 4;
+	flags = radeon_vm_page_flags(bo_va->flags);
+	if ((flags & R600_PTE_GART_MASK) == R600_PTE_GART_MASK) {
+		/* only copy commands needed */
+		ndw += ncmds * 7;
+
+	} else if (flags & R600_PTE_SYSTEM) {
+		/* header for write data commands */
+		ndw += ncmds * 4;
+
+		/* body of write data command */
+		ndw += nptes * 2;
+
+	} else {
+		/* set page commands needed */
+		ndw += ncmds * 10;
 
-	/* reserve space for pte addresses */
-	ndw += nptes * 2;
+		/* two extra commands for begin/end of fragment */
+		ndw += 2 * 10;
+	}
 
 	/* update too big for an IB */
 	if (ndw > 0xfffff)
@@ -901,11 +961,15 @@ int radeon_vm_bo_update(struct radeon_device *rdev,
 		return r;
 	ib.length_dw = 0;
 
-	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset,
-			      addr, radeon_vm_page_flags(bo_va->flags));
+	radeon_vm_update_ptes(rdev, vm, &ib, bo_va->it.start,
+			      bo_va->it.last + 1, addr,
+			      radeon_vm_page_flags(bo_va->flags));
 
-	radeon_semaphore_sync_to(ib.semaphore, vm->fence);
-	r = radeon_ib_schedule(rdev, &ib, NULL);
+	radeon_asic_vm_pad_ib(rdev, &ib);
+	WARN_ON(ib.length_dw > ndw);
+
+	radeon_semaphore_sync_fence(ib.semaphore, vm->fence);
+	r = radeon_ib_schedule(rdev, &ib, NULL, false);
 	if (r) {
 		radeon_ib_free(rdev, &ib);
 		return r;
@@ -936,8 +1000,8 @@ int radeon_vm_clear_freed(struct radeon_device *rdev,
 	int r;
 
 	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
-		list_del(&bo_va->vm_status);
 		r = radeon_vm_bo_update(rdev, bo_va, NULL);
+		radeon_bo_unref(&bo_va->bo);
 		kfree(bo_va);
 		if (r)
 			return r;
@@ -947,6 +1011,31 @@ int radeon_vm_clear_freed(struct radeon_device *rdev,
 }
 
 /**
+ * radeon_vm_clear_invalids - clear invalidated BOs in the PT
+ *
+ * @rdev: radeon_device pointer
+ * @vm: requested vm
+ *
+ * Make sure all invalidated BOs are cleared in the PT.
+ * Returns 0 for success.
+ *
+ * PTs have to be reserved and mutex must be locked!
+ */
+int radeon_vm_clear_invalids(struct radeon_device *rdev,
+			     struct radeon_vm *vm)
+{
+	struct radeon_bo_va *bo_va, *tmp;
+	int r;
+
+	list_for_each_entry_safe(bo_va, tmp, &vm->invalidated, vm_status) {
+		r = radeon_vm_bo_update(rdev, bo_va, NULL);
+		if (r)
+			return r;
+	}
+	return 0;
+}
+
+/**
  * radeon_vm_bo_rmv - remove a bo to a specific vm
  *
  * @rdev: radeon_device pointer
@@ -964,10 +1053,11 @@ void radeon_vm_bo_rmv(struct radeon_device *rdev,
 	list_del(&bo_va->bo_list);
 
 	mutex_lock(&vm->mutex);
-	list_del(&bo_va->vm_list);
+	interval_tree_remove(&bo_va->it, &vm->va);
+	list_del(&bo_va->vm_status);
 
-	if (bo_va->soffset) {
-		bo_va->bo = NULL;
+	if (bo_va->addr) {
+		bo_va->bo = radeon_bo_ref(bo_va->bo);
 		list_add(&bo_va->vm_status, &vm->freed);
 	} else {
 		kfree(bo_va);
@@ -991,7 +1081,12 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev,
 	struct radeon_bo_va *bo_va;
 
 	list_for_each_entry(bo_va, &bo->va, bo_list) {
-		bo_va->valid = false;
+		if (bo_va->addr) {
+			mutex_lock(&bo_va->vm->mutex);
+			list_del(&bo_va->vm_status);
+			list_add(&bo_va->vm_status, &bo_va->vm->invalidated);
+			mutex_unlock(&bo_va->vm->mutex);
+		}
 	}
 }
 
@@ -1016,7 +1111,8 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	vm->last_flush = NULL;
 	vm->last_id_use = NULL;
 	mutex_init(&vm->mutex);
-	INIT_LIST_HEAD(&vm->va);
+	vm->va = RB_ROOT;
+	INIT_LIST_HEAD(&vm->invalidated);
 	INIT_LIST_HEAD(&vm->freed);
 
 	pd_size = radeon_vm_directory_size(rdev);
@@ -1031,8 +1127,8 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
 	}
 
 	r = radeon_bo_create(rdev, pd_size, align, true,
-			     RADEON_GEM_DOMAIN_VRAM, NULL,
-			     &vm->page_directory);
+			     RADEON_GEM_DOMAIN_VRAM, 0, NULL,
+			     NULL, &vm->page_directory);
 	if (r)
 		return r;
 
@@ -1060,11 +1156,11 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 	struct radeon_bo_va *bo_va, *tmp;
 	int i, r;
 
-	if (!list_empty(&vm->va)) {
+	if (!RB_EMPTY_ROOT(&vm->va)) {
 		dev_err(rdev->dev, "still active bo inside vm\n");
 	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) {
-		list_del_init(&bo_va->vm_list);
+	rbtree_postorder_for_each_entry_safe(bo_va, tmp, &vm->va, it.rb) {
+		interval_tree_remove(&bo_va->it, &vm->va);
 		r = radeon_bo_reserve(bo_va->bo, false);
 		if (!r) {
 			list_del_init(&bo_va->bo_list);
@@ -1072,8 +1168,10 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
 			kfree(bo_va);
 		}
 	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status)
+	list_for_each_entry_safe(bo_va, tmp, &vm->freed, vm_status) {
+		radeon_bo_unref(&bo_va->bo);
 		kfree(bo_va);
+	}
 
 	for (i = 0; i < radeon_vm_num_pdes(rdev); i++)
 		radeon_bo_unref(&vm->page_tables[i].bo);
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index a0f96decece3..c5799f16aa4b 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -109,7 +109,6 @@ int rs400_gart_enable(struct radeon_device *rdev)
 	uint32_t size_reg;
 	uint32_t tmp;
 
-	radeon_gart_restore(rdev);
 	tmp = RREG32_MC(RS690_AIC_CTRL_SCRATCH);
 	tmp |= RS690_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32_MC(RS690_AIC_CTRL_SCRATCH, tmp);
@@ -209,17 +208,24 @@ void rs400_gart_fini(struct radeon_device *rdev)
 	radeon_gart_table_ram_free(rdev);
 }
 
+#define RS400_PTE_UNSNOOPED (1 << 0)
 #define RS400_PTE_WRITEABLE (1 << 2)
 #define RS400_PTE_READABLE  (1 << 3)
 
-void rs400_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
+void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
+			 uint64_t addr, uint32_t flags)
 {
 	uint32_t entry;
 	u32 *gtt = rdev->gart.ptr;
 
 	entry = (lower_32_bits(addr) & PAGE_MASK) |
-		((upper_32_bits(addr) & 0xff) << 4) |
-		RS400_PTE_WRITEABLE | RS400_PTE_READABLE;
+		((upper_32_bits(addr) & 0xff) << 4);
+	if (flags & RADEON_GART_PAGE_READ)
+		entry |= RS400_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		entry |= RS400_PTE_WRITEABLE;
+	if (!(flags & RADEON_GART_PAGE_SNOOP))
+		entry |= RS400_PTE_UNSNOOPED;
 	entry = cpu_to_le32(entry);
 	gtt[i] = entry;
 }
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index d1a35cb1c91d..9acb1c3c005b 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -555,7 +555,6 @@ static int rs600_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Enable bus master */
 	tmp = RREG32(RADEON_BUS_CNTL) & ~RS600_BUS_MASTER_DIS;
 	WREG32(RADEON_BUS_CNTL, tmp);
@@ -626,15 +625,21 @@ static void rs600_gart_fini(struct radeon_device *rdev)
 	radeon_gart_table_vram_free(rdev);
 }
 
-void rs600_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
+void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
+			 uint64_t addr, uint32_t flags)
 {
 	void __iomem *ptr = (void *)rdev->gart.ptr;
 
 	addr = addr & 0xFFFFFFFFFFFFF000ULL;
-	if (addr == rdev->dummy_page.addr)
-		addr |= R600_PTE_SYSTEM | R600_PTE_SNOOPED;
-	else
-		addr |= R600_PTE_GART;
+	addr |= R600_PTE_SYSTEM;
+	if (flags & RADEON_GART_PAGE_VALID)
+		addr |= R600_PTE_VALID;
+	if (flags & RADEON_GART_PAGE_READ)
+		addr |= R600_PTE_READABLE;
+	if (flags & RADEON_GART_PAGE_WRITE)
+		addr |= R600_PTE_WRITEABLE;
+	if (flags & RADEON_GART_PAGE_SNOOP)
+		addr |= R600_PTE_SNOOPED;
 	writeq(addr, ptr + (i * 8));
 }
 
@@ -874,6 +879,9 @@ void rs600_bandwidth_update(struct radeon_device *rdev)
 	u32 d1mode_priority_a_cnt, d2mode_priority_a_cnt;
 	/* FIXME: implement full support */
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 3462b64369bf..0a2d36e81108 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -579,6 +579,9 @@ void rs690_bandwidth_update(struct radeon_device *rdev)
 	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
 	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c
index 02f7710de470..9031f4b69824 100644
--- a/drivers/gpu/drm/radeon/rs780_dpm.c
+++ b/drivers/gpu/drm/radeon/rs780_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rs780d.h"
 #include "r600_dpm.h"
 #include "rs780_dpm.h"
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 3e21e869015f..c55d653aaf5f 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -124,7 +124,7 @@ void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 	radeon_ring_write(ring, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST);
 	radeon_ring_write(ring, PACKET0(0x20C8, 0));
 	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 }
 
 int rv515_mc_wait_for_idle(struct radeon_device *rdev)
@@ -1277,6 +1277,9 @@ void rv515_bandwidth_update(struct radeon_device *rdev)
 	struct drm_display_mode *mode0 = NULL;
 	struct drm_display_mode *mode1 = NULL;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	if (rdev->mode_info.crtcs[0]->base.enabled)
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c
index e7045b085715..6a5c233361e9 100644
--- a/drivers/gpu/drm/radeon/rv6xx_dpm.c
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rv6xxd.h"
 #include "r600_dpm.h"
 #include "rv6xx_dpm.h"
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index da8703d8d455..372016e266d0 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -26,7 +26,6 @@
  *          Jerome Glisse
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <drm/drmP.h>
 #include "radeon.h"
@@ -900,7 +899,6 @@ static int rv770_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
@@ -1178,7 +1176,6 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 	u32 hdp_host_path_cntl;
 	u32 sq_dyn_gpr_size_simd_ab_0;
 	u32 gb_tiling_config = 0;
-	u32 cc_rb_backend_disable = 0;
 	u32 cc_gc_shader_pipe_config = 0;
 	u32 mc_arb_ramcfg;
 	u32 db_debug4, tmp;
@@ -1312,21 +1309,7 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 		WREG32(SPI_CONFIG_CNTL, 0);
 	}
 
-	cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000;
-	tmp = R7XX_MAX_BACKENDS - r600_count_pipe_bits(cc_rb_backend_disable >> 16);
-	if (tmp < rdev->config.rv770.max_backends) {
-		rdev->config.rv770.max_backends = tmp;
-	}
-
 	cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00;
-	tmp = R7XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 8) & R7XX_MAX_PIPES_MASK);
-	if (tmp < rdev->config.rv770.max_pipes) {
-		rdev->config.rv770.max_pipes = tmp;
-	}
-	tmp = R7XX_MAX_SIMDS - r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK);
-	if (tmp < rdev->config.rv770.max_simds) {
-		rdev->config.rv770.max_simds = tmp;
-	}
 	tmp = rdev->config.rv770.max_simds -
 		r600_count_pipe_bits((cc_gc_shader_pipe_config >> 16) & R7XX_MAX_SIMDS_MASK);
 	rdev->config.rv770.active_simds = tmp;
@@ -1349,6 +1332,14 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 	rdev->config.rv770.tiling_npipes = rdev->config.rv770.max_tile_pipes;
 
 	disabled_rb_mask = (RREG32(CC_RB_BACKEND_DISABLE) >> 16) & R7XX_MAX_BACKENDS_MASK;
+	tmp = 0;
+	for (i = 0; i < rdev->config.rv770.max_backends; i++)
+		tmp |= (1 << i);
+	/* if all the backends are disabled, fix it up here */
+	if ((disabled_rb_mask & tmp) == tmp) {
+		for (i = 0; i < rdev->config.rv770.max_backends; i++)
+			disabled_rb_mask &= ~(1 << i);
+	}
 	tmp = (gb_tiling_config & PIPE_TILING__MASK) >> PIPE_TILING__SHIFT;
 	tmp = r6xx_remap_render_backend(rdev, tmp, rdev->config.rv770.max_backends,
 					R7XX_MAX_BACKENDS, disabled_rb_mask);
diff --git a/drivers/gpu/drm/radeon/rv770_dma.c b/drivers/gpu/drm/radeon/rv770_dma.c
index bbf2e076ee45..7f34bad2e724 100644
--- a/drivers/gpu/drm/radeon/rv770_dma.c
+++ b/drivers/gpu/drm/radeon/rv770_dma.c
@@ -33,18 +33,19 @@
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (r7xx).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int rv770_copy_dma(struct radeon_device *rdev,
-		  uint64_t src_offset, uint64_t dst_offset,
-		  unsigned num_gpu_pages,
-		  struct radeon_fence **fence)
+struct radeon_fence *rv770_copy_dma(struct radeon_device *rdev,
+				    uint64_t src_offset, uint64_t dst_offset,
+				    unsigned num_gpu_pages,
+				    struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_dw, cur_size_in_dw;
@@ -54,7 +55,7 @@ int rv770_copy_dma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_dw = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT) / 4;
@@ -63,10 +64,10 @@ int rv770_copy_dma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -83,15 +84,15 @@ int rv770_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_dw * 4;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c
index 3c76e1dcdf04..755a8f96fe46 100644
--- a/drivers/gpu/drm/radeon/rv770_dpm.c
+++ b/drivers/gpu/drm/radeon/rv770_dpm.c
@@ -24,6 +24,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "rv770d.h"
 #include "r600_dpm.h"
 #include "rv770_dpm.h"
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 9e854fd016da..7d5083dc4acb 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -42,6 +42,14 @@ MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_mc2.bin");
 MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_smc.bin");
+
+MODULE_FIRMWARE("radeon/tahiti_pfp.bin");
+MODULE_FIRMWARE("radeon/tahiti_me.bin");
+MODULE_FIRMWARE("radeon/tahiti_ce.bin");
+MODULE_FIRMWARE("radeon/tahiti_mc.bin");
+MODULE_FIRMWARE("radeon/tahiti_rlc.bin");
+MODULE_FIRMWARE("radeon/tahiti_smc.bin");
+
 MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
@@ -49,6 +57,14 @@ MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin");
+
+MODULE_FIRMWARE("radeon/pitcairn_pfp.bin");
+MODULE_FIRMWARE("radeon/pitcairn_me.bin");
+MODULE_FIRMWARE("radeon/pitcairn_ce.bin");
+MODULE_FIRMWARE("radeon/pitcairn_mc.bin");
+MODULE_FIRMWARE("radeon/pitcairn_rlc.bin");
+MODULE_FIRMWARE("radeon/pitcairn_smc.bin");
+
 MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
 MODULE_FIRMWARE("radeon/VERDE_me.bin");
 MODULE_FIRMWARE("radeon/VERDE_ce.bin");
@@ -56,6 +72,14 @@ MODULE_FIRMWARE("radeon/VERDE_mc.bin");
 MODULE_FIRMWARE("radeon/VERDE_mc2.bin");
 MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
 MODULE_FIRMWARE("radeon/VERDE_smc.bin");
+
+MODULE_FIRMWARE("radeon/verde_pfp.bin");
+MODULE_FIRMWARE("radeon/verde_me.bin");
+MODULE_FIRMWARE("radeon/verde_ce.bin");
+MODULE_FIRMWARE("radeon/verde_mc.bin");
+MODULE_FIRMWARE("radeon/verde_rlc.bin");
+MODULE_FIRMWARE("radeon/verde_smc.bin");
+
 MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
 MODULE_FIRMWARE("radeon/OLAND_me.bin");
 MODULE_FIRMWARE("radeon/OLAND_ce.bin");
@@ -63,6 +87,14 @@ MODULE_FIRMWARE("radeon/OLAND_mc.bin");
 MODULE_FIRMWARE("radeon/OLAND_mc2.bin");
 MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
 MODULE_FIRMWARE("radeon/OLAND_smc.bin");
+
+MODULE_FIRMWARE("radeon/oland_pfp.bin");
+MODULE_FIRMWARE("radeon/oland_me.bin");
+MODULE_FIRMWARE("radeon/oland_ce.bin");
+MODULE_FIRMWARE("radeon/oland_mc.bin");
+MODULE_FIRMWARE("radeon/oland_rlc.bin");
+MODULE_FIRMWARE("radeon/oland_smc.bin");
+
 MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
 MODULE_FIRMWARE("radeon/HAINAN_me.bin");
 MODULE_FIRMWARE("radeon/HAINAN_ce.bin");
@@ -71,6 +103,13 @@ MODULE_FIRMWARE("radeon/HAINAN_mc2.bin");
 MODULE_FIRMWARE("radeon/HAINAN_rlc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_smc.bin");
 
+MODULE_FIRMWARE("radeon/hainan_pfp.bin");
+MODULE_FIRMWARE("radeon/hainan_me.bin");
+MODULE_FIRMWARE("radeon/hainan_ce.bin");
+MODULE_FIRMWARE("radeon/hainan_mc.bin");
+MODULE_FIRMWARE("radeon/hainan_rlc.bin");
+MODULE_FIRMWARE("radeon/hainan_smc.bin");
+
 static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh);
 static void si_pcie_gen3_enable(struct radeon_device *rdev);
 static void si_program_aspm(struct radeon_device *rdev);
@@ -1470,38 +1509,54 @@ static const u32 hainan_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
 /* ucode loading */
 int si_mc_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
+	const __be32 *fw_data = NULL;
+	const __le32 *new_fw_data = NULL;
 	u32 running, blackout = 0;
-	u32 *io_mc_regs;
+	u32 *io_mc_regs = NULL;
+	const __le32 *new_io_mc_regs = NULL;
 	int i, regs_size, ucode_size;
 
 	if (!rdev->mc_fw)
 		return -EINVAL;
 
-	ucode_size = rdev->mc_fw->size / 4;
+	if (rdev->new_fw) {
+		const struct mc_firmware_header_v1_0 *hdr =
+			(const struct mc_firmware_header_v1_0 *)rdev->mc_fw->data;
+
+		radeon_ucode_print_mc_hdr(&hdr->header);
+		regs_size = le32_to_cpu(hdr->io_debug_size_bytes) / (4 * 2);
+		new_io_mc_regs = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->io_debug_array_offset_bytes));
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		new_fw_data = (const __le32 *)
+			(rdev->mc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		ucode_size = rdev->mc_fw->size / 4;
 
-	switch (rdev->family) {
-	case CHIP_TAHITI:
-		io_mc_regs = (u32 *)&tahiti_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_PITCAIRN:
-		io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_VERDE:
-	default:
-		io_mc_regs = (u32 *)&verde_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_OLAND:
-		io_mc_regs = (u32 *)&oland_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
-	case CHIP_HAINAN:
-		io_mc_regs = (u32 *)&hainan_io_mc_regs;
-		regs_size = TAHITI_IO_MC_REGS_SIZE;
-		break;
+		switch (rdev->family) {
+		case CHIP_TAHITI:
+			io_mc_regs = (u32 *)&tahiti_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_PITCAIRN:
+			io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_VERDE:
+		default:
+			io_mc_regs = (u32 *)&verde_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_OLAND:
+			io_mc_regs = (u32 *)&oland_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		case CHIP_HAINAN:
+			io_mc_regs = (u32 *)&hainan_io_mc_regs;
+			regs_size = TAHITI_IO_MC_REGS_SIZE;
+			break;
+		}
+		fw_data = (const __be32 *)rdev->mc_fw->data;
 	}
 
 	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
@@ -1518,13 +1573,21 @@ int si_mc_load_microcode(struct radeon_device *rdev)
 
 		/* load mc io regs */
 		for (i = 0; i < regs_size; i++) {
-			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
-			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			if (rdev->new_fw) {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, le32_to_cpup(new_io_mc_regs++));
+				WREG32(MC_SEQ_IO_DEBUG_DATA, le32_to_cpup(new_io_mc_regs++));
+			} else {
+				WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+				WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+			}
 		}
 		/* load the MC ucode */
-		fw_data = (const __be32 *)rdev->mc_fw->data;
-		for (i = 0; i < ucode_size; i++)
-			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		for (i = 0; i < ucode_size; i++) {
+			if (rdev->new_fw)
+				WREG32(MC_SEQ_SUP_PGM, le32_to_cpup(new_fw_data++));
+			else
+				WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+		}
 
 		/* put the engine back into the active state */
 		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
@@ -1553,18 +1616,19 @@ int si_mc_load_microcode(struct radeon_device *rdev)
 static int si_init_microcode(struct radeon_device *rdev)
 {
 	const char *chip_name;
-	const char *rlc_chip_name;
+	const char *new_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
 	size_t smc_req_size, mc2_req_size;
 	char fw_name[30];
 	int err;
+	int new_fw = 0;
 
 	DRM_DEBUG("\n");
 
 	switch (rdev->family) {
 	case CHIP_TAHITI:
 		chip_name = "TAHITI";
-		rlc_chip_name = "TAHITI";
+		new_chip_name = "tahiti";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1575,7 +1639,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_PITCAIRN:
 		chip_name = "PITCAIRN";
-		rlc_chip_name = "PITCAIRN";
+		new_chip_name = "pitcairn";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1586,7 +1650,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_VERDE:
 		chip_name = "VERDE";
-		rlc_chip_name = "VERDE";
+		new_chip_name = "verde";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1597,7 +1661,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_OLAND:
 		chip_name = "OLAND";
-		rlc_chip_name = "OLAND";
+		new_chip_name = "oland";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1607,7 +1671,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		break;
 	case CHIP_HAINAN:
 		chip_name = "HAINAN";
-		rlc_chip_name = "HAINAN";
+		new_chip_name = "hainan";
 		pfp_req_size = SI_PFP_UCODE_SIZE * 4;
 		me_req_size = SI_PM4_UCODE_SIZE * 4;
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
@@ -1618,86 +1682,178 @@ static int si_init_microcode(struct radeon_device *rdev)
 	default: BUG();
 	}
 
-	DRM_INFO("Loading %s Microcode\n", chip_name);
+	DRM_INFO("Loading %s Microcode\n", new_chip_name);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", new_chip_name);
 	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->pfp_fw->size != pfp_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->pfp_fw->size, fw_name);
-		err = -EINVAL;
-		goto out;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+		err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->pfp_fw->size != pfp_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->pfp_fw->size, fw_name);
+			err = -EINVAL;
+			goto out;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->pfp_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", new_chip_name);
 	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->me_fw->size != me_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->me_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+		err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->me_fw->size != me_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->me_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->me_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", new_chip_name);
 	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->ce_fw->size != ce_req_size) {
-		printk(KERN_ERR
-		       "si_cp: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->ce_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+		err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->ce_fw->size != ce_req_size) {
+			printk(KERN_ERR
+			       "si_cp: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->ce_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->ce_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", new_chip_name);
 	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
-	if (err)
-		goto out;
-	if (rdev->rlc_fw->size != rlc_req_size) {
-		printk(KERN_ERR
-		       "si_rlc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->rlc_fw->size, fw_name);
-		err = -EINVAL;
+	if (err) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+		err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->rlc_fw->size != rlc_req_size) {
+			printk(KERN_ERR
+			       "si_rlc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->rlc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->rlc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", new_chip_name);
 	err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 	if (err) {
-		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
 		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
-		if (err)
+		if (err) {
+			snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+			err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+			if (err)
+				goto out;
+		}
+		if ((rdev->mc_fw->size != mc_req_size) &&
+		    (rdev->mc_fw->size != mc2_req_size)) {
+			printk(KERN_ERR
+			       "si_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+		DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
+	} else {
+		err = radeon_ucode_validate(rdev->mc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
 			goto out;
+		} else {
+			new_fw++;
+		}
 	}
-	if ((rdev->mc_fw->size != mc_req_size) &&
-	    (rdev->mc_fw->size != mc2_req_size)) {
-		printk(KERN_ERR
-		       "si_mc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->mc_fw->size, fw_name);
-		err = -EINVAL;
-	}
-	DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
 
-	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name);
 	err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
 	if (err) {
-		printk(KERN_ERR
-		       "smc: error loading firmware \"%s\"\n",
-		       fw_name);
-		release_firmware(rdev->smc_fw);
-		rdev->smc_fw = NULL;
-		err = 0;
-	} else if (rdev->smc_fw->size != smc_req_size) {
-		printk(KERN_ERR
-		       "si_smc: Bogus length %zu in firmware \"%s\"\n",
-		       rdev->smc_fw->size, fw_name);
-		err = -EINVAL;
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+		if (err) {
+			printk(KERN_ERR
+			       "smc: error loading firmware \"%s\"\n",
+			       fw_name);
+			release_firmware(rdev->smc_fw);
+			rdev->smc_fw = NULL;
+			err = 0;
+		} else if (rdev->smc_fw->size != smc_req_size) {
+			printk(KERN_ERR
+			       "si_smc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->smc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	} else {
+		err = radeon_ucode_validate(rdev->smc_fw);
+		if (err) {
+			printk(KERN_ERR
+			       "si_cp: validation failed for firmware \"%s\"\n",
+			       fw_name);
+			goto out;
+		} else {
+			new_fw++;
+		}
 	}
 
+	if (new_fw == 0) {
+		rdev->new_fw = false;
+	} else if (new_fw < 6) {
+		printk(KERN_ERR "si_fw: mixing new and old firmware!\n");
+		err = -EINVAL;
+	} else {
+		rdev->new_fw = true;
+	}
 out:
 	if (err) {
 		if (err != -EINVAL)
@@ -2228,6 +2384,9 @@ void dce6_bandwidth_update(struct radeon_device *rdev)
 	u32 num_heads = 0, lb_size;
 	int i;
 
+	if (!rdev->mode_info.mode_config_initialized)
+		return;
+
 	radeon_update_display_priority(rdev);
 
 	for (i = 0; i < rdev->num_crtc; i++) {
@@ -2901,7 +3060,7 @@ static void si_gpu_init(struct radeon_device *rdev)
 	u32 sx_debug_1;
 	u32 hdp_host_path_cntl;
 	u32 tmp;
-	int i, j, k;
+	int i, j;
 
 	switch (rdev->family) {
 	case CHIP_TAHITI:
@@ -3099,12 +3258,11 @@ static void si_gpu_init(struct radeon_device *rdev)
 		     rdev->config.si.max_sh_per_se,
 		     rdev->config.si.max_cu_per_sh);
 
+	rdev->config.si.active_cus = 0;
 	for (i = 0; i < rdev->config.si.max_shader_engines; i++) {
 		for (j = 0; j < rdev->config.si.max_sh_per_se; j++) {
-			for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) {
-				rdev->config.si.active_cus +=
-					hweight32(si_get_cu_active_bitmap(rdev, i, j));
-			}
+			rdev->config.si.active_cus +=
+				hweight32(si_get_cu_active_bitmap(rdev, i, j));
 		}
 	}
 
@@ -3282,34 +3440,77 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable)
 
 static int si_cp_load_microcode(struct radeon_device *rdev)
 {
-	const __be32 *fw_data;
 	int i;
 
-	if (!rdev->me_fw || !rdev->pfp_fw)
+	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
 		return -EINVAL;
 
 	si_cp_enable(rdev, false);
 
-	/* PFP */
-	fw_data = (const __be32 *)rdev->pfp_fw->data;
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < SI_PFP_UCODE_SIZE; i++)
-		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_PFP_UCODE_ADDR, 0);
-
-	/* CE */
-	fw_data = (const __be32 *)rdev->ce_fw->data;
-	WREG32(CP_CE_UCODE_ADDR, 0);
-	for (i = 0; i < SI_CE_UCODE_SIZE; i++)
-		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_CE_UCODE_ADDR, 0);
-
-	/* ME */
-	fw_data = (const __be32 *)rdev->me_fw->data;
-	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < SI_PM4_UCODE_SIZE; i++)
-		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
-	WREG32(CP_ME_RAM_WADDR, 0);
+	if (rdev->new_fw) {
+		const struct gfx_firmware_header_v1_0 *pfp_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->pfp_fw->data;
+		const struct gfx_firmware_header_v1_0 *ce_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->ce_fw->data;
+		const struct gfx_firmware_header_v1_0 *me_hdr =
+			(const struct gfx_firmware_header_v1_0 *)rdev->me_fw->data;
+		const __le32 *fw_data;
+		u32 fw_size;
+
+		radeon_ucode_print_gfx_hdr(&pfp_hdr->header);
+		radeon_ucode_print_gfx_hdr(&ce_hdr->header);
+		radeon_ucode_print_gfx_hdr(&me_hdr->header);
+
+		/* PFP */
+		fw_data = (const __le32 *)
+			(rdev->pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __le32 *)
+			(rdev->ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)
+			(rdev->me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+		fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < fw_size; i++)
+			WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	} else {
+		const __be32 *fw_data;
+
+		/* PFP */
+		fw_data = (const __be32 *)rdev->pfp_fw->data;
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+		for (i = 0; i < SI_PFP_UCODE_SIZE; i++)
+			WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_PFP_UCODE_ADDR, 0);
+
+		/* CE */
+		fw_data = (const __be32 *)rdev->ce_fw->data;
+		WREG32(CP_CE_UCODE_ADDR, 0);
+		for (i = 0; i < SI_CE_UCODE_SIZE; i++)
+			WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_CE_UCODE_ADDR, 0);
+
+		/* ME */
+		fw_data = (const __be32 *)rdev->me_fw->data;
+		WREG32(CP_ME_RAM_WADDR, 0);
+		for (i = 0; i < SI_PM4_UCODE_SIZE; i++)
+			WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_ME_RAM_WADDR, 0);
+	}
 
 	WREG32(CP_PFP_UCODE_ADDR, 0);
 	WREG32(CP_CE_UCODE_ADDR, 0);
@@ -3342,7 +3543,7 @@ static int si_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
 	radeon_ring_write(ring, 0xc000);
 	radeon_ring_write(ring, 0xe000);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	si_cp_enable(rdev, true);
 
@@ -3371,7 +3572,7 @@ static int si_cp_start(struct radeon_device *rdev)
 	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
 	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 	for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) {
 		ring = &rdev->ring[i];
@@ -3381,7 +3582,7 @@ static int si_cp_start(struct radeon_device *rdev)
 		radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0));
 		radeon_ring_write(ring, 0);
 
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_unlock_commit(rdev, ring, false);
 	}
 
 	return 0;
@@ -4048,7 +4249,6 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
-	radeon_gart_restore(rdev);
 	/* Setup TLB control */
 	WREG32(MC_VM_MX_L1_TLB_CNTL,
 	       (0xA << 7) |
@@ -4093,10 +4293,10 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 	for (i = 1; i < 16; i++) {
 		if (i < 8)
 			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 		else
 			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
-			       rdev->gart.table_addr >> 12);
+			       rdev->vm_manager.saved_table_addr[i]);
 	}
 
 	/* enable context1-15 */
@@ -4128,6 +4328,17 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
 
 static void si_pcie_gart_disable(struct radeon_device *rdev)
 {
+	unsigned i;
+
+	for (i = 1; i < 16; ++i) {
+		uint32_t reg;
+		if (i < 8)
+			reg = VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2);
+		else
+			reg = VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2);
+		rdev->vm_manager.saved_table_addr[i] = RREG32(reg);
+	}
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -4476,7 +4687,7 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev,
 int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 {
 	int ret = 0;
-	u32 idx = 0;
+	u32 idx = 0, i;
 	struct radeon_cs_packet pkt;
 
 	do {
@@ -4487,6 +4698,12 @@ int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
 		switch (pkt.type) {
 		case RADEON_PACKET_TYPE0:
 			dev_err(rdev->dev, "Packet0 not allowed!\n");
+			for (i = 0; i < ib->length_dw; i++) {
+				if (i == idx)
+					printk("\t0x%08x <---\n", ib->ptr[i]);
+				else
+					printk("\t0x%08x\n", ib->ptr[i]);
+			}
 			ret = -EINVAL;
 			break;
 		case RADEON_PACKET_TYPE2:
@@ -4815,7 +5032,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* write new base address */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 
 	if (vm->id < 8) {
@@ -4830,7 +5047,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* flush hdp cache */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
 	radeon_ring_write(ring, 0);
@@ -4838,7 +5055,7 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 
 	/* bits 0-15 are the VM contexts0-15 */
 	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
 				 WRITE_DATA_DST_SEL(0)));
 	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
 	radeon_ring_write(ring, 0);
@@ -5592,7 +5809,6 @@ static void si_enable_lbpw(struct radeon_device *rdev, bool enable)
 static int si_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i;
-	const __be32 *fw_data;
 
 	if (!rdev->rlc_fw)
 		return -EINVAL;
@@ -5615,10 +5831,26 @@ static int si_rlc_resume(struct radeon_device *rdev)
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
-	fw_data = (const __be32 *)rdev->rlc_fw->data;
-	for (i = 0; i < SI_RLC_UCODE_SIZE; i++) {
-		WREG32(RLC_UCODE_ADDR, i);
-		WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+	if (rdev->new_fw) {
+		const struct rlc_firmware_header_v1_0 *hdr =
+			(const struct rlc_firmware_header_v1_0 *)rdev->rlc_fw->data;
+		u32 fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+		const __le32 *fw_data = (const __le32 *)
+			(rdev->rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+		radeon_ucode_print_rlc_hdr(&hdr->header);
+
+		for (i = 0; i < fw_size; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, le32_to_cpup(fw_data++));
+		}
+	} else {
+		const __be32 *fw_data =
+			(const __be32 *)rdev->rlc_fw->data;
+		for (i = 0; i < SI_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
 	}
 	WREG32(RLC_UCODE_ADDR, 0);
 
@@ -6093,17 +6325,17 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev)
 		wptr = RREG32(IH_RB_WPTR);
 
 	if (wptr & RB_OVERFLOW) {
+		wptr &= ~RB_OVERFLOW;
 		/* When a ring buffer overflow happen start parsing interrupt
 		 * from the last not overwritten vector (wptr + 16). Hopefully
 		 * this should allow us to catchup.
 		 */
-		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
-			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n",
+			 wptr, rdev->ih.rptr, (wptr + 16) & rdev->ih.ptr_mask);
 		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
 		tmp = RREG32(IH_RB_CNTL);
 		tmp |= IH_WPTR_OVERFLOW_CLEAR;
 		WREG32(IH_RB_CNTL, tmp);
-		wptr &= ~RB_OVERFLOW;
 	}
 	return (wptr & rdev->ih.ptr_mask);
 }
@@ -6318,7 +6550,8 @@ restart_ih:
 		case 16: /* D5 page flip */
 		case 18: /* D6 page flip */
 			DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
-			radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+			if (radeon_use_pflipirq > 0)
+				radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
 			break;
 		case 42: /* HPD hotplug */
 			switch (src_data) {
@@ -6440,13 +6673,13 @@ restart_ih:
 		/* wptr/rptr are in bytes! */
 		rptr += 16;
 		rptr &= rdev->ih.ptr_mask;
+		WREG32(IH_RB_RPTR, rptr);
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
 	if (queue_thermal && rdev->pm.dpm_enabled)
 		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
-	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
 
 	/* make sure wptr hasn't changed while processing */
@@ -6964,6 +7197,9 @@ static void si_pcie_gen3_enable(struct radeon_device *rdev)
 	int ret, i;
 	u16 tmp16;
 
+	if (pci_is_root_bus(rdev->pdev->bus))
+		return;
+
 	if (radeon_pcie_gen2 == 0)
 		return;
 
@@ -7241,7 +7477,8 @@ static void si_program_aspm(struct radeon_device *rdev)
 			if (orig != data)
 				WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
 
-			if (!disable_clkreq) {
+			if (!disable_clkreq &&
+			    !pci_is_root_bus(rdev->pdev->bus)) {
 				struct pci_dev *root = rdev->pdev->bus->self;
 				u32 lnkcap;
 
diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c
index e24c94b6d14d..b58f12b762d7 100644
--- a/drivers/gpu/drm/radeon/si_dma.c
+++ b/drivers/gpu/drm/radeon/si_dma.c
@@ -56,7 +56,41 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 }
 
 /**
- * si_dma_vm_set_page - update the page tables using the DMA
+ * si_dma_vm_copy_pages - update PTEs by copying them from the GART
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @src: src addr where to copy from
+ * @count: number of page entries to update
+ *
+ * Update PTEs by copying them from the GART using the DMA (SI).
+ */
+void si_dma_vm_copy_pages(struct radeon_device *rdev,
+			  struct radeon_ib *ib,
+			  uint64_t pe, uint64_t src,
+			  unsigned count)
+{
+	while (count) {
+		unsigned bytes = count * 8;
+		if (bytes > 0xFFFF8)
+			bytes = 0xFFFF8;
+
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
+						      1, 0, 0, bytes);
+		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+		ib->ptr[ib->length_dw++] = lower_32_bits(src);
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
+
+		pe += bytes;
+		src += bytes;
+		count -= bytes / 8;
+	}
+}
+
+/**
+ * si_dma_vm_write_pages - update PTEs by writing them manually
  *
  * @rdev: radeon_device pointer
  * @ib: indirect buffer to fill with commands
@@ -66,83 +100,89 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
  * @incr: increase next addr by incr bytes
  * @flags: access flags
  *
- * Update the page tables using the DMA (SI).
+ * Update PTEs by writing them manually using the DMA (SI).
  */
-void si_dma_vm_set_page(struct radeon_device *rdev,
-			struct radeon_ib *ib,
-			uint64_t pe,
-			uint64_t addr, unsigned count,
-			uint32_t incr, uint32_t flags)
+void si_dma_vm_write_pages(struct radeon_device *rdev,
+			   struct radeon_ib *ib,
+			   uint64_t pe,
+			   uint64_t addr, unsigned count,
+			   uint32_t incr, uint32_t flags)
 {
 	uint64_t value;
 	unsigned ndw;
 
-	trace_radeon_vm_set_page(pe, addr, count, incr, flags);
-
-	if (flags == R600_PTE_GART) {
-		uint64_t src = rdev->gart.table_addr + (addr >> 12) * 8;
-		while (count) {
-			unsigned bytes = count * 8;
-			if (bytes > 0xFFFF8)
-				bytes = 0xFFFF8;
-
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_COPY,
-							      1, 0, 0, bytes);
-			ib->ptr[ib->length_dw++] = lower_32_bits(pe);
-			ib->ptr[ib->length_dw++] = lower_32_bits(src);
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = upper_32_bits(src) & 0xff;
-
-			pe += bytes;
-			src += bytes;
-			count -= bytes / 8;
-		}
-	} else if (flags & R600_PTE_SYSTEM) {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			/* for non-physically contiguous pages (system) */
-			ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
-			ib->ptr[ib->length_dw++] = pe;
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		/* for non-physically contiguous pages (system) */
+		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
+		ib->ptr[ib->length_dw++] = pe;
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+			if (flags & R600_PTE_SYSTEM) {
 				value = radeon_vm_map_gart(rdev, addr);
 				value &= 0xFFFFFFFFFFFFF000ULL;
-				addr += incr;
-				value |= flags;
-				ib->ptr[ib->length_dw++] = value;
-				ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			}
-		}
-	} else {
-		while (count) {
-			ndw = count * 2;
-			if (ndw > 0xFFFFE)
-				ndw = 0xFFFFE;
-
-			if (flags & R600_PTE_VALID)
+			} else if (flags & R600_PTE_VALID) {
 				value = addr;
-			else
+			} else {
 				value = 0;
-			/* for physically contiguous pages (vram) */
-			ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
-			ib->ptr[ib->length_dw++] = pe; /* dst addr */
-			ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
-			ib->ptr[ib->length_dw++] = flags; /* mask */
-			ib->ptr[ib->length_dw++] = 0;
-			ib->ptr[ib->length_dw++] = value; /* value */
+			}
+			addr += incr;
+			value |= flags;
+			ib->ptr[ib->length_dw++] = value;
 			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-			ib->ptr[ib->length_dw++] = incr; /* increment size */
-			ib->ptr[ib->length_dw++] = 0;
-			pe += ndw * 4;
-			addr += (ndw / 2) * incr;
-			count -= ndw / 2;
 		}
 	}
-	while (ib->length_dw & 0x7)
-		ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0);
+}
+
+/**
+ * si_dma_vm_set_pages - update the page tables using the DMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using the DMA (SI).
+ */
+void si_dma_vm_set_pages(struct radeon_device *rdev,
+			 struct radeon_ib *ib,
+			 uint64_t pe,
+			 uint64_t addr, unsigned count,
+			 uint32_t incr, uint32_t flags)
+{
+	uint64_t value;
+	unsigned ndw;
+
+	while (count) {
+		ndw = count * 2;
+		if (ndw > 0xFFFFE)
+			ndw = 0xFFFFE;
+
+		if (flags & R600_PTE_VALID)
+			value = addr;
+		else
+			value = 0;
+
+		/* for physically contiguous pages (vram) */
+		ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
+		ib->ptr[ib->length_dw++] = pe; /* dst addr */
+		ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
+		ib->ptr[ib->length_dw++] = flags; /* mask */
+		ib->ptr[ib->length_dw++] = 0;
+		ib->ptr[ib->length_dw++] = value; /* value */
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		ib->ptr[ib->length_dw++] = incr; /* increment size */
+		ib->ptr[ib->length_dw++] = 0;
+		pe += ndw * 4;
+		addr += (ndw / 2) * incr;
+		count -= ndw / 2;
+	}
 }
 
 void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
@@ -178,18 +218,19 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
  * @src_offset: src GPU address
  * @dst_offset: dst GPU address
  * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
+ * @resv: reservation object to sync to
  *
  * Copy GPU paging using the DMA engine (SI).
  * Used by the radeon ttm implementation to move pages if
  * registered as the asic copy callback.
  */
-int si_copy_dma(struct radeon_device *rdev,
-		uint64_t src_offset, uint64_t dst_offset,
-		unsigned num_gpu_pages,
-		struct radeon_fence **fence)
+struct radeon_fence *si_copy_dma(struct radeon_device *rdev,
+				 uint64_t src_offset, uint64_t dst_offset,
+				 unsigned num_gpu_pages,
+				 struct reservation_object *resv)
 {
 	struct radeon_semaphore *sem = NULL;
+	struct radeon_fence *fence;
 	int ring_index = rdev->asic->copy.dma_ring_index;
 	struct radeon_ring *ring = &rdev->ring[ring_index];
 	u32 size_in_bytes, cur_size_in_bytes;
@@ -199,7 +240,7 @@ int si_copy_dma(struct radeon_device *rdev,
 	r = radeon_semaphore_create(rdev, &sem);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
-		return r;
+		return ERR_PTR(r);
 	}
 
 	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
@@ -208,10 +249,10 @@ int si_copy_dma(struct radeon_device *rdev,
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_semaphore_sync_to(sem, *fence);
+	radeon_semaphore_sync_resv(rdev, sem, resv, false);
 	radeon_semaphore_sync_rings(rdev, sem, ring->idx);
 
 	for (i = 0; i < num_loops; i++) {
@@ -228,16 +269,16 @@ int si_copy_dma(struct radeon_device *rdev,
 		dst_offset += cur_size_in_bytes;
 	}
 
-	r = radeon_fence_emit(rdev, fence, ring->idx);
+	r = radeon_fence_emit(rdev, &fence, ring->idx);
 	if (r) {
 		radeon_ring_unlock_undo(rdev, ring);
 		radeon_semaphore_free(rdev, &sem, NULL);
-		return r;
+		return ERR_PTR(r);
 	}
 
-	radeon_ring_unlock_commit(rdev, ring);
-	radeon_semaphore_free(rdev, &sem, *fence);
+	radeon_ring_unlock_commit(rdev, ring, false);
+	radeon_semaphore_free(rdev, &sem, fence);
 
-	return r;
+	return fence;
 }
 
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
index 58918868f894..676e6c2ba90a 100644
--- a/drivers/gpu/drm/radeon/si_dpm.c
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "sid.h"
 #include "r600_dpm.h"
 #include "si_dpm.h"
@@ -3812,6 +3813,27 @@ void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
 	voltage_table->count = max_voltage_steps;
 }
 
+static int si_get_svi2_voltage_table(struct radeon_device *rdev,
+				     struct radeon_clock_voltage_dependency_table *voltage_dependency_table,
+				     struct atom_voltage_table *voltage_table)
+{
+	u32 i;
+
+	if (voltage_dependency_table == NULL)
+		return -EINVAL;
+
+	voltage_table->mask_low = 0;
+	voltage_table->phase_delay = 0;
+
+	voltage_table->count = voltage_dependency_table->count;
+	for (i = 0; i < voltage_table->count; i++) {
+		voltage_table->entries[i].value = voltage_dependency_table->entries[i].v;
+		voltage_table->entries[i].smio_low = 0;
+	}
+
+	return 0;
+}
+
 static int si_construct_voltage_tables(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
@@ -3819,15 +3841,25 @@ static int si_construct_voltage_tables(struct radeon_device *rdev)
 	struct si_power_info *si_pi = si_get_pi(rdev);
 	int ret;
 
-	ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
-					    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
-	if (ret)
-		return ret;
+	if (pi->voltage_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+						    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
+		if (ret)
+			return ret;
 
-	if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
-		si_trim_voltage_table_to_fit_state_table(rdev,
-							 SISLANDS_MAX_NO_VREG_STEPS,
-							 &eg_pi->vddc_voltage_table);
+		if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev,
+								 SISLANDS_MAX_NO_VREG_STEPS,
+								 &eg_pi->vddc_voltage_table);
+	} else if (si_pi->voltage_control_svi2) {
+		ret = si_get_svi2_voltage_table(rdev,
+						&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						&eg_pi->vddc_voltage_table);
+		if (ret)
+			return ret;
+	} else {
+		return -EINVAL;
+	}
 
 	if (eg_pi->vddci_control) {
 		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI,
@@ -3840,6 +3872,13 @@ static int si_construct_voltage_tables(struct radeon_device *rdev)
 								 SISLANDS_MAX_NO_VREG_STEPS,
 								 &eg_pi->vddci_voltage_table);
 	}
+	if (si_pi->vddci_control_svi2) {
+		ret = si_get_svi2_voltage_table(rdev,
+						&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						&eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+	}
 
 	if (pi->mvdd_control) {
 		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_MVDDC,
@@ -3893,46 +3932,55 @@ static int si_populate_smc_voltage_tables(struct radeon_device *rdev,
 	struct si_power_info *si_pi = si_get_pi(rdev);
 	u8 i;
 
-	if (eg_pi->vddc_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
-			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
-
-		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
-			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
-				table->maxVDDCIndexInPPTable = i;
-				break;
+	if (si_pi->voltage_control_svi2) {
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svc,
+			si_pi->svc_gpio_id);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svd,
+			si_pi->svd_gpio_id);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_svi_rework_plat_type,
+					   2);
+	} else {
+		if (eg_pi->vddc_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+				cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+			for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+				if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+					table->maxVDDCIndexInPPTable = i;
+					break;
+				}
 			}
 		}
-	}
 
-	if (eg_pi->vddci_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+		if (eg_pi->vddci_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
 
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
-			cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
-	}
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+				cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
+		}
 
 
-	if (si_pi->mvdd_voltage_table.count) {
-		si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
+		if (si_pi->mvdd_voltage_table.count) {
+			si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
 
-		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
-			cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
-	}
+			table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
+				cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
+		}
 
-	if (si_pi->vddc_phase_shed_control) {
-		if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
-						      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
-			si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
+		if (si_pi->vddc_phase_shed_control) {
+			if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
+							      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
+				si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
 
-			table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
-				cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
+				table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+					cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
 
-			si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
-						   (u32)si_pi->vddc_phase_shed_table.phase_delay);
-		} else {
-			si_pi->vddc_phase_shed_control = false;
+				si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
+							   (u32)si_pi->vddc_phase_shed_table.phase_delay);
+			} else {
+				si_pi->vddc_phase_shed_control = false;
+			}
 		}
 	}
 
@@ -5798,16 +5846,17 @@ int si_dpm_enable(struct radeon_device *rdev)
 {
 	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
 	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
 	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
 	int ret;
 
 	if (si_is_smc_running(rdev))
 		return -EINVAL;
-	if (pi->voltage_control)
+	if (pi->voltage_control || si_pi->voltage_control_svi2)
 		si_enable_voltage_control(rdev, true);
 	if (pi->mvdd_control)
 		si_get_mvdd_configuration(rdev);
-	if (pi->voltage_control) {
+	if (pi->voltage_control || si_pi->voltage_control_svi2) {
 		ret = si_construct_voltage_tables(rdev);
 		if (ret) {
 			DRM_ERROR("si_construct_voltage_tables failed\n");
@@ -6207,7 +6256,7 @@ static void si_parse_pplib_clock_info(struct radeon_device *rdev,
 	if ((rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) &&
 	    index == 0) {
 		/* XXX disable for A0 tahiti */
-		si_pi->ulv.supported = true;
+		si_pi->ulv.supported = false;
 		si_pi->ulv.pl = *pl;
 		si_pi->ulv.one_pcie_lane_in_ulv = false;
 		si_pi->ulv.volt_change_delay = SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT;
@@ -6406,16 +6455,32 @@ int si_dpm_init(struct radeon_device *rdev)
 	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
 
 	pi->voltage_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					    VOLTAGE_OBJ_GPIO_LUT);
+	if (!pi->voltage_control) {
+		si_pi->voltage_control_svi2 =
+			radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+						    VOLTAGE_OBJ_SVID2);
+		if (si_pi->voltage_control_svi2)
+			radeon_atom_get_svi2_info(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+						  &si_pi->svd_gpio_id, &si_pi->svc_gpio_id);
+	}
 
 	pi->mvdd_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					    VOLTAGE_OBJ_GPIO_LUT);
 
 	eg_pi->vddci_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, VOLTAGE_OBJ_GPIO_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI,
+					    VOLTAGE_OBJ_GPIO_LUT);
+	if (!eg_pi->vddci_control)
+		si_pi->vddci_control_svi2 =
+			radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI,
+						    VOLTAGE_OBJ_SVID2);
 
 	si_pi->vddc_phase_shed_control =
-		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT);
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					    VOLTAGE_OBJ_PHASE_LUT);
 
 	rv770_get_engine_memory_ss(rdev);
 
diff --git a/drivers/gpu/drm/radeon/si_dpm.h b/drivers/gpu/drm/radeon/si_dpm.h
index 4ce5032cdf49..8b5c06a0832d 100644
--- a/drivers/gpu/drm/radeon/si_dpm.h
+++ b/drivers/gpu/drm/radeon/si_dpm.h
@@ -170,6 +170,8 @@ struct si_power_info {
 	bool vddc_phase_shed_control;
 	bool pspp_notify_required;
 	bool sclk_deep_sleep_above_low;
+	bool voltage_control_svi2;
+	bool vddci_control_svi2;
 	/* smc offsets */
 	u32 sram_end;
 	u32 state_table_start;
@@ -192,6 +194,9 @@ struct si_power_info {
 	SMC_SIslands_MCRegisters smc_mc_reg_table;
 	SISLANDS_SMC_STATETABLE smc_statetable;
 	PP_SIslands_PAPMParameters papm_parm;
+	/* SVI2 */
+	u8 svd_gpio_id;
+	u8 svc_gpio_id;
 };
 
 #define SISLANDS_INITIAL_STATE_ARB_INDEX    0
diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c
index e80efcf0c230..73dbc79c959d 100644
--- a/drivers/gpu/drm/radeon/si_smc.c
+++ b/drivers/gpu/drm/radeon/si_smc.c
@@ -219,36 +219,48 @@ int si_load_smc_ucode(struct radeon_device *rdev, u32 limit)
 	if (!rdev->smc_fw)
 		return -EINVAL;
 
-	switch (rdev->family) {
-	case CHIP_TAHITI:
-		ucode_start_address = TAHITI_SMC_UCODE_START;
-		ucode_size = TAHITI_SMC_UCODE_SIZE;
-		break;
-	case CHIP_PITCAIRN:
-		ucode_start_address = PITCAIRN_SMC_UCODE_START;
-		ucode_size = PITCAIRN_SMC_UCODE_SIZE;
-		break;
-	case CHIP_VERDE:
-		ucode_start_address = VERDE_SMC_UCODE_START;
-		ucode_size = VERDE_SMC_UCODE_SIZE;
-		break;
-	case CHIP_OLAND:
-		ucode_start_address = OLAND_SMC_UCODE_START;
-		ucode_size = OLAND_SMC_UCODE_SIZE;
-		break;
-	case CHIP_HAINAN:
-		ucode_start_address = HAINAN_SMC_UCODE_START;
-		ucode_size = HAINAN_SMC_UCODE_SIZE;
-		break;
-	default:
-		DRM_ERROR("unknown asic in smc ucode loader\n");
-		BUG();
+	if (rdev->new_fw) {
+		const struct smc_firmware_header_v1_0 *hdr =
+			(const struct smc_firmware_header_v1_0 *)rdev->smc_fw->data;
+
+		radeon_ucode_print_smc_hdr(&hdr->header);
+
+		ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
+		ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
+		src = (const u8 *)
+			(rdev->smc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+	} else {
+		switch (rdev->family) {
+		case CHIP_TAHITI:
+			ucode_start_address = TAHITI_SMC_UCODE_START;
+			ucode_size = TAHITI_SMC_UCODE_SIZE;
+			break;
+		case CHIP_PITCAIRN:
+			ucode_start_address = PITCAIRN_SMC_UCODE_START;
+			ucode_size = PITCAIRN_SMC_UCODE_SIZE;
+			break;
+		case CHIP_VERDE:
+			ucode_start_address = VERDE_SMC_UCODE_START;
+			ucode_size = VERDE_SMC_UCODE_SIZE;
+			break;
+		case CHIP_OLAND:
+			ucode_start_address = OLAND_SMC_UCODE_START;
+			ucode_size = OLAND_SMC_UCODE_SIZE;
+			break;
+		case CHIP_HAINAN:
+			ucode_start_address = HAINAN_SMC_UCODE_START;
+			ucode_size = HAINAN_SMC_UCODE_SIZE;
+			break;
+		default:
+			DRM_ERROR("unknown asic in smc ucode loader\n");
+			BUG();
+		}
+		src = (const u8 *)rdev->smc_fw->data;
 	}
 
 	if (ucode_size & 3)
 		return -EINVAL;
 
-	src = (const u8 *)rdev->smc_fw->data;
 	spin_lock_irqsave(&rdev->smc_idx_lock, flags);
 	WREG32(SMC_IND_INDEX_0, ucode_start_address);
 	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index fd414d34d885..6635da9ec986 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -736,7 +736,7 @@
 #       define DESCRIPTION16(x)                          (((x) & 0xff) << 0)
 #       define DESCRIPTION17(x)                          (((x) & 0xff) << 8)
 
-#define AZ_F0_CODEC_PIN_CONTROL_HOTPLUG_CONTROL          0x54
+#define AZ_F0_CODEC_PIN_CONTROL_HOT_PLUG_CONTROL         0x54
 #       define AUDIO_ENABLED                             (1 << 31)
 
 #define AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT  0x56
diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h
index 10e945a49479..623a0b1e2d9d 100644
--- a/drivers/gpu/drm/radeon/sislands_smc.h
+++ b/drivers/gpu/drm/radeon/sislands_smc.h
@@ -241,6 +241,9 @@ typedef struct SISLANDS_SMC_STATETABLE SISLANDS_SMC_STATETABLE;
 #define SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width  0xF4
 #define SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen   0xFC
 #define SI_SMC_SOFT_REGISTER_vr_hot_gpio              0x100
+#define SI_SMC_SOFT_REGISTER_svi_rework_plat_type     0x118
+#define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svd   0x11c
+#define SI_SMC_SOFT_REGISTER_svi_rework_gpio_id_svc   0x120
 
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
 #define SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 32
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c
index 3f0e8d7b8dbe..1f8a8833e1be 100644
--- a/drivers/gpu/drm/radeon/sumo_dpm.c
+++ b/drivers/gpu/drm/radeon/sumo_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "sumod.h"
 #include "r600_dpm.h"
 #include "cypress_dpm.h"
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c
index 32e50be9c4ac..b4ec5c4e7969 100644
--- a/drivers/gpu/drm/radeon/trinity_dpm.c
+++ b/drivers/gpu/drm/radeon/trinity_dpm.c
@@ -23,6 +23,7 @@
 
 #include "drmP.h"
 #include "radeon.h"
+#include "radeon_asic.h"
 #include "trinityd.h"
 #include "r600_dpm.h"
 #include "trinity_dpm.h"
@@ -1874,16 +1875,22 @@ int trinity_dpm_init(struct radeon_device *rdev)
 	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++)
 		pi->at[i] = TRINITY_AT_DFLT;
 
-	/* There are stability issues reported on with
-	 * bapm enabled when switching between AC and battery
-	 * power.  At the same time, some MSI boards hang
-	 * if it's not enabled and dpm is enabled.  Just enable
-	 * it for MSI boards right now.
-	 */
-	if (rdev->pdev->subsystem_vendor == 0x1462)
-		pi->enable_bapm = true;
-	else
+	if (radeon_bapm == -1) {
+		/* There are stability issues reported on with
+		 * bapm enabled when switching between AC and battery
+		 * power.  At the same time, some MSI boards hang
+		 * if it's not enabled and dpm is enabled.  Just enable
+		 * it for MSI boards right now.
+		 */
+		if (rdev->pdev->subsystem_vendor == 0x1462)
+			pi->enable_bapm = true;
+		else
+			pi->enable_bapm = false;
+	} else if (radeon_bapm == 0) {
 		pi->enable_bapm = false;
+	} else {
+		pi->enable_bapm = true;
+	}
 	pi->enable_nbps_policy = true;
 	pi->enable_sclk_ds = true;
 	pi->enable_gfx_power_gating = true;
diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c
index be42c8125203..e72b3cb59358 100644
--- a/drivers/gpu/drm/radeon/uvd_v1_0.c
+++ b/drivers/gpu/drm/radeon/uvd_v1_0.c
@@ -22,6 +22,7 @@
  * Authors: Christian König <christian.koenig@amd.com>
  */
 
+#include <linux/firmware.h>
 #include <drm/drmP.h>
 #include "radeon.h"
 #include "radeon_asic.h"
@@ -70,6 +71,82 @@ void uvd_v1_0_set_wptr(struct radeon_device *rdev,
 }
 
 /**
+ * uvd_v1_0_fence_emit - emit an fence & trap command
+ *
+ * @rdev: radeon_device pointer
+ * @fence: fence to emit
+ *
+ * Write a fence and a trap command to the ring.
+ */
+void uvd_v1_0_fence_emit(struct radeon_device *rdev,
+			 struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA0, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_DATA1, 0));
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, PACKET0(UVD_GPCOM_VCPU_CMD, 0));
+	radeon_ring_write(ring, 2);
+	return;
+}
+
+/**
+ * uvd_v1_0_resume - memory controller programming
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Let the UVD memory controller know it's offsets
+ */
+int uvd_v1_0_resume(struct radeon_device *rdev)
+{
+	uint64_t addr;
+	uint32_t size;
+	int r;
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	/* programm the VCPU memory controller bits 0-27 */
+	addr = (rdev->uvd.gpu_addr >> 3) + 16;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size) >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = RADEON_UVD_STACK_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET1, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = RADEON_UVD_HEAP_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET2, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE2, size);
+
+	/* bits 28-31 */
+	addr = (rdev->uvd.gpu_addr >> 28) & 0xF;
+	WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0));
+
+	/* bits 32-39 */
+	addr = (rdev->uvd.gpu_addr >> 32) & 0xFF;
+	WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31));
+
+	WREG32(UVD_FW_START, *((uint32_t*)rdev->uvd.cpu_addr));
+
+	return 0;
+}
+
+/**
  * uvd_v1_0_init - start and test UVD block
  *
  * @rdev: radeon_device pointer
@@ -124,14 +201,38 @@ int uvd_v1_0_init(struct radeon_device *rdev)
 	radeon_ring_write(ring, PACKET0(UVD_SEMA_CNTL, 0));
 	radeon_ring_write(ring, 3);
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 
 done:
 	/* lower clocks again */
 	radeon_set_uvd_clocks(rdev, 0, 0);
 
-	if (!r)
+	if (!r) {
+		switch (rdev->family) {
+		case CHIP_RV610:
+		case CHIP_RV630:
+		case CHIP_RV620:
+			/* 64byte granularity workaround */
+			WREG32(MC_CONFIG, 0);
+			WREG32(MC_CONFIG, 1 << 4);
+			WREG32(RS_DQ_RD_RET_CONF, 0x3f);
+			WREG32(MC_CONFIG, 0x1f);
+
+			/* fall through */
+		case CHIP_RV670:
+		case CHIP_RV635:
+
+			/* write clean workaround */
+			WREG32_P(UVD_VCPU_CNTL, 0x10, ~0x10);
+			break;
+
+		default:
+			/* TODO: Do we need more? */
+			break;
+		}
+
 		DRM_INFO("UVD initialized successfully.\n");
+	}
 
 	return r;
 }
@@ -218,12 +319,12 @@ int uvd_v1_0_start(struct radeon_device *rdev)
 	/* enable UMC */
 	WREG32_P(UVD_LMI_CTRL2, 0, ~(1 << 8));
 
+	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
+
 	/* boot up the VCPU */
 	WREG32(UVD_SOFT_RESET, 0);
 	mdelay(10);
 
-	WREG32_P(UVD_RB_ARB_CTRL, 0, ~(1 << 3));
-
 	for (i = 0; i < 10; ++i) {
 		uint32_t status;
 		for (j = 0; j < 100; ++j) {
@@ -331,7 +432,7 @@ int uvd_v1_0_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 	radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0));
 	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev, ring, false);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(UVD_CONTEXT_ID);
 		if (tmp == 0xDEADBEEF)
diff --git a/drivers/gpu/drm/radeon/uvd_v2_2.c b/drivers/gpu/drm/radeon/uvd_v2_2.c
index 8bfdadd56598..89193519f8a1 100644
--- a/drivers/gpu/drm/radeon/uvd_v2_2.c
+++ b/drivers/gpu/drm/radeon/uvd_v2_2.c
@@ -72,6 +72,10 @@ int uvd_v2_2_resume(struct radeon_device *rdev)
 	uint32_t chip_id, size;
 	int r;
 
+	/* RV770 uses V1.0 MC */
+	if (rdev->family == CHIP_RV770)
+		return uvd_v1_0_resume(rdev);
+
 	r = radeon_uvd_resume(rdev);
 	if (r)
 		return r;
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig
index 2e3d7b5b0ad7..c96f6089f8bf 100644
--- a/drivers/gpu/drm/rcar-du/Kconfig
+++ b/drivers/gpu/drm/rcar-du/Kconfig
@@ -6,6 +6,7 @@ config DRM_RCAR_DU
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
 	select DRM_KMS_FB_HELPER
+	select VIDEOMODE_HELPERS
 	help
 	  Choose this option if you have an R-Car chipset.
 	  If M is selected the module will be called rcar-du-drm.
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
index 299267db2898..148b50589181 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_crtc.c  --  R-Car Display Unit CRTCs
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
index 43e7575c700c..e97ae502dec5 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_crtc.h  --  R-Car Display Unit CRTCs
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.c b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
index 792fd1d20e86..d212efa6a495 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_drv.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_drv.c  --  R-Car Display Unit DRM driver
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,6 +15,7 @@
 #include <linux/io.h>
 #include <linux/mm.h>
 #include <linux/module.h>
+#include <linux/of_device.h>
 #include <linux/platform_device.h>
 #include <linux/pm.h>
 #include <linux/slab.h>
@@ -30,6 +31,97 @@
 #include "rcar_du_regs.h"
 
 /* -----------------------------------------------------------------------------
+ * Device Information
+ */
+
+static const struct rcar_du_device_info rcar_du_r8a7779_info = {
+	.features = 0,
+	.num_crtcs = 2,
+	.routes = {
+		/* R8A7779 has two RGB outputs and one (currently unsupported)
+		 * TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_DPAD1] = {
+			.possible_crtcs = BIT(1) | BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 1,
+		},
+	},
+	.num_lvds = 0,
+};
+
+static const struct rcar_du_device_info rcar_du_r8a7790_info = {
+	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
+	.quirks = RCAR_DU_QUIRK_ALIGN_128B | RCAR_DU_QUIRK_LVDS_LANES,
+	.num_crtcs = 3,
+	.routes = {
+		/* R8A7790 has one RGB output, two LVDS outputs and one
+		 * (currently unsupported) TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(2) | BIT(1) | BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_LVDS0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 1,
+		},
+		[RCAR_DU_OUTPUT_LVDS1] = {
+			.possible_crtcs = BIT(2) | BIT(1),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 2,
+		},
+	},
+	.num_lvds = 2,
+};
+
+static const struct rcar_du_device_info rcar_du_r8a7791_info = {
+	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
+	.num_crtcs = 2,
+	.routes = {
+		/* R8A7791 has one RGB output, one LVDS output and one
+		 * (currently unsupported) TCON output.
+		 */
+		[RCAR_DU_OUTPUT_DPAD0] = {
+			.possible_crtcs = BIT(1),
+			.encoder_type = DRM_MODE_ENCODER_NONE,
+			.port = 0,
+		},
+		[RCAR_DU_OUTPUT_LVDS0] = {
+			.possible_crtcs = BIT(0),
+			.encoder_type = DRM_MODE_ENCODER_LVDS,
+			.port = 1,
+		},
+	},
+	.num_lvds = 1,
+};
+
+static const struct platform_device_id rcar_du_id_table[] = {
+	{ "rcar-du-r8a7779", (kernel_ulong_t)&rcar_du_r8a7779_info },
+	{ "rcar-du-r8a7790", (kernel_ulong_t)&rcar_du_r8a7790_info },
+	{ "rcar-du-r8a7791", (kernel_ulong_t)&rcar_du_r8a7791_info },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(platform, rcar_du_id_table);
+
+static const struct of_device_id rcar_du_of_table[] = {
+	{ .compatible = "renesas,du-r8a7779", .data = &rcar_du_r8a7779_info },
+	{ .compatible = "renesas,du-r8a7790", .data = &rcar_du_r8a7790_info },
+	{ .compatible = "renesas,du-r8a7791", .data = &rcar_du_r8a7791_info },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(of, rcar_du_of_table);
+
+/* -----------------------------------------------------------------------------
  * DRM operations
  */
 
@@ -53,12 +145,13 @@ static int rcar_du_unload(struct drm_device *dev)
 static int rcar_du_load(struct drm_device *dev, unsigned long flags)
 {
 	struct platform_device *pdev = dev->platformdev;
+	struct device_node *np = pdev->dev.of_node;
 	struct rcar_du_platform_data *pdata = pdev->dev.platform_data;
 	struct rcar_du_device *rcdu;
 	struct resource *mem;
 	int ret;
 
-	if (pdata == NULL) {
+	if (pdata == NULL && np == NULL) {
 		dev_err(dev->dev, "no platform data\n");
 		return -ENODEV;
 	}
@@ -71,7 +164,8 @@ static int rcar_du_load(struct drm_device *dev, unsigned long flags)
 
 	rcdu->dev = &pdev->dev;
 	rcdu->pdata = pdata;
-	rcdu->info = (struct rcar_du_device_info *)pdev->id_entry->driver_data;
+	rcdu->info = np ? of_match_device(rcar_du_of_table, rcdu->dev)->data
+		   : (void *)platform_get_device_id(pdev)->driver_data;
 	rcdu->ddev = dev;
 	dev->dev_private = rcdu;
 
@@ -158,6 +252,7 @@ static struct drm_driver rcar_du_driver = {
 	.unload			= rcar_du_unload,
 	.preclose		= rcar_du_preclose,
 	.lastclose		= rcar_du_lastclose,
+	.set_busid		= drm_platform_set_busid,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= rcar_du_enable_vblank,
 	.disable_vblank		= rcar_du_disable_vblank,
@@ -187,7 +282,7 @@ static struct drm_driver rcar_du_driver = {
  * Power management
  */
 
-#if CONFIG_PM_SLEEP
+#ifdef CONFIG_PM_SLEEP
 static int rcar_du_pm_suspend(struct device *dev)
 {
 	struct rcar_du_device *rcdu = dev_get_drvdata(dev);
@@ -231,77 +326,6 @@ static int rcar_du_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static const struct rcar_du_device_info rcar_du_r8a7779_info = {
-	.features = 0,
-	.num_crtcs = 2,
-	.routes = {
-		/* R8A7779 has two RGB outputs and one (currently unsupported)
-		 * TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_DPAD1] = {
-			.possible_crtcs = BIT(1) | BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-	},
-	.num_lvds = 0,
-};
-
-static const struct rcar_du_device_info rcar_du_r8a7790_info = {
-	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
-	.quirks = RCAR_DU_QUIRK_ALIGN_128B | RCAR_DU_QUIRK_LVDS_LANES,
-	.num_crtcs = 3,
-	.routes = {
-		/* R8A7790 has one RGB output, two LVDS outputs and one
-		 * (currently unsupported) TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(2) | BIT(1) | BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_LVDS0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-		[RCAR_DU_OUTPUT_LVDS1] = {
-			.possible_crtcs = BIT(2) | BIT(1),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-	},
-	.num_lvds = 2,
-};
-
-static const struct rcar_du_device_info rcar_du_r8a7791_info = {
-	.features = RCAR_DU_FEATURE_CRTC_IRQ_CLOCK | RCAR_DU_FEATURE_DEFR8,
-	.num_crtcs = 2,
-	.routes = {
-		/* R8A7791 has one RGB output, one LVDS output and one
-		 * (currently unsupported) TCON output.
-		 */
-		[RCAR_DU_OUTPUT_DPAD0] = {
-			.possible_crtcs = BIT(1),
-			.encoder_type = DRM_MODE_ENCODER_NONE,
-		},
-		[RCAR_DU_OUTPUT_LVDS0] = {
-			.possible_crtcs = BIT(0),
-			.encoder_type = DRM_MODE_ENCODER_LVDS,
-		},
-	},
-	.num_lvds = 1,
-};
-
-static const struct platform_device_id rcar_du_id_table[] = {
-	{ "rcar-du-r8a7779", (kernel_ulong_t)&rcar_du_r8a7779_info },
-	{ "rcar-du-r8a7790", (kernel_ulong_t)&rcar_du_r8a7790_info },
-	{ "rcar-du-r8a7791", (kernel_ulong_t)&rcar_du_r8a7791_info },
-	{ }
-};
-
-MODULE_DEVICE_TABLE(platform, rcar_du_id_table);
-
 static struct platform_driver rcar_du_platform_driver = {
 	.probe		= rcar_du_probe,
 	.remove		= rcar_du_remove,
@@ -309,6 +333,7 @@ static struct platform_driver rcar_du_platform_driver = {
 		.owner	= THIS_MODULE,
 		.name	= "rcar-du",
 		.pm	= &rcar_du_pm_ops,
+		.of_match_table = rcar_du_of_table,
 	},
 	.id_table	= rcar_du_id_table,
 };
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.h b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
index e31b735d3f25..8e494633c3b3 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_drv.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_drv.h  --  R-Car Display Unit DRM driver
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -37,6 +37,7 @@ struct rcar_du_lvdsenc;
  * struct rcar_du_output_routing - Output routing specification
  * @possible_crtcs: bitmask of possible CRTCs for the output
  * @encoder_type: DRM type of the internal encoder associated with the output
+ * @port: device tree port number corresponding to this output route
  *
  * The DU has 5 possible outputs (DPAD0/1, LVDS0/1, TCON). Output routing data
  * specify the valid SoC outputs, which CRTCs can drive the output, and the type
@@ -45,6 +46,7 @@ struct rcar_du_lvdsenc;
 struct rcar_du_output_routing {
 	unsigned int possible_crtcs;
 	unsigned int encoder_type;
+	unsigned int port;
 };
 
 /*
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
index 3daa7a168dc6..7c0ec95915ef 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_encoder.c  --  R-Car Display Unit Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -142,7 +142,8 @@ static const struct drm_encoder_funcs encoder_funcs = {
 int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			 enum rcar_du_encoder_type type,
 			 enum rcar_du_output output,
-			 const struct rcar_du_encoder_data *data)
+			 const struct rcar_du_encoder_data *data,
+			 struct device_node *np)
 {
 	struct rcar_du_encoder *renc;
 	unsigned int encoder_type;
@@ -189,9 +190,11 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
 
 	switch (encoder_type) {
-	case DRM_MODE_ENCODER_LVDS:
-		return rcar_du_lvds_connector_init(rcdu, renc,
-						   &data->connector.lvds.panel);
+	case DRM_MODE_ENCODER_LVDS: {
+		const struct rcar_du_panel_data *pdata =
+			data ? &data->connector.lvds.panel : NULL;
+		return rcar_du_lvds_connector_init(rcdu, renc, pdata, np);
+	}
 
 	case DRM_MODE_ENCODER_DAC:
 		return rcar_du_vga_connector_init(rcdu, renc);
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
index 0e5a65e45d0e..bd624135ef1f 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_encoder.h  --  R-Car Display Unit Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -44,6 +44,7 @@ rcar_du_connector_best_encoder(struct drm_connector *connector);
 int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			 enum rcar_du_encoder_type type,
 			 enum rcar_du_output output,
-			 const struct rcar_du_encoder_data *data);
+			 const struct rcar_du_encoder_data *data,
+			 struct device_node *np);
 
 #endif /* __RCAR_DU_ENCODER_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_group.c b/drivers/gpu/drm/rcar-du/rcar_du_group.c
index eb53cd97e8c6..4e7614b145db 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_group.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_group.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_group.c  --  R-Car Display Unit Channels Pair
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_group.h b/drivers/gpu/drm/rcar-du/rcar_du_group.h
index 5025930972ec..0c38cdcda4ca 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_group.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_group.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_group.c  --  R-Car Display Unit Planes and CRTCs Group
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
index a87edfac111f..6c24ad7d03ef 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_kms.c  --  R-Car Display Unit Mode Setting
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -17,6 +17,8 @@
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_gem_cma_helper.h>
 
+#include <linux/of_graph.h>
+
 #include "rcar_du_crtc.h"
 #include "rcar_du_drv.h"
 #include "rcar_du_encoder.h"
@@ -135,7 +137,9 @@ rcar_du_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 {
 	struct rcar_du_device *rcdu = dev->dev_private;
 	const struct rcar_du_format_info *format;
+	unsigned int max_pitch;
 	unsigned int align;
+	unsigned int bpp;
 
 	format = rcar_du_format_info(mode_cmd->pixel_format);
 	if (format == NULL) {
@@ -144,13 +148,20 @@ rcar_du_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 		return ERR_PTR(-EINVAL);
 	}
 
+	/*
+	 * The pitch and alignment constraints are expressed in pixels on the
+	 * hardware side and in bytes in the DRM API.
+	 */
+	bpp = format->planes == 2 ? 1 : format->bpp / 8;
+	max_pitch =  4096 * bpp;
+
 	if (rcar_du_needs(rcdu, RCAR_DU_QUIRK_ALIGN_128B))
 		align = 128;
 	else
-		align = 16 * format->bpp / 8;
+		align = 16 * bpp;
 
 	if (mode_cmd->pitches[0] & (align - 1) ||
-	    mode_cmd->pitches[0] >= 8192) {
+	    mode_cmd->pitches[0] >= max_pitch) {
 		dev_dbg(dev->dev, "invalid pitch value %u\n",
 			mode_cmd->pitches[0]);
 		return ERR_PTR(-EINVAL);
@@ -179,6 +190,205 @@ static const struct drm_mode_config_funcs rcar_du_mode_config_funcs = {
 	.output_poll_changed = rcar_du_output_poll_changed,
 };
 
+static int rcar_du_encoders_init_pdata(struct rcar_du_device *rcdu)
+{
+	unsigned int num_encoders = 0;
+	unsigned int i;
+	int ret;
+
+	for (i = 0; i < rcdu->pdata->num_encoders; ++i) {
+		const struct rcar_du_encoder_data *pdata =
+			&rcdu->pdata->encoders[i];
+		const struct rcar_du_output_routing *route =
+			&rcdu->info->routes[pdata->output];
+
+		if (pdata->type == RCAR_DU_ENCODER_UNUSED)
+			continue;
+
+		if (pdata->output >= RCAR_DU_OUTPUT_MAX ||
+		    route->possible_crtcs == 0) {
+			dev_warn(rcdu->dev,
+				 "encoder %u references unexisting output %u, skipping\n",
+				 i, pdata->output);
+			continue;
+		}
+
+		ret = rcar_du_encoder_init(rcdu, pdata->type, pdata->output,
+					   pdata, NULL);
+		if (ret < 0)
+			return ret;
+
+		num_encoders++;
+	}
+
+	return num_encoders;
+}
+
+static int rcar_du_encoders_init_dt_one(struct rcar_du_device *rcdu,
+					enum rcar_du_output output,
+					struct of_endpoint *ep)
+{
+	static const struct {
+		const char *compatible;
+		enum rcar_du_encoder_type type;
+	} encoders[] = {
+		{ "adi,adv7123", RCAR_DU_ENCODER_VGA },
+		{ "thine,thc63lvdm83d", RCAR_DU_ENCODER_LVDS },
+	};
+
+	enum rcar_du_encoder_type enc_type = RCAR_DU_ENCODER_NONE;
+	struct device_node *connector = NULL;
+	struct device_node *encoder = NULL;
+	struct device_node *prev = NULL;
+	struct device_node *entity_ep_node;
+	struct device_node *entity;
+	int ret;
+
+	/*
+	 * Locate the connected entity and infer its type from the number of
+	 * endpoints.
+	 */
+	entity = of_graph_get_remote_port_parent(ep->local_node);
+	if (!entity) {
+		dev_dbg(rcdu->dev, "unconnected endpoint %s, skipping\n",
+			ep->local_node->full_name);
+		return 0;
+	}
+
+	entity_ep_node = of_parse_phandle(ep->local_node, "remote-endpoint", 0);
+
+	while (1) {
+		struct device_node *ep_node;
+
+		ep_node = of_graph_get_next_endpoint(entity, prev);
+		of_node_put(prev);
+		prev = ep_node;
+
+		if (!ep_node)
+			break;
+
+		if (ep_node == entity_ep_node)
+			continue;
+
+		/*
+		 * We've found one endpoint other than the input, this must
+		 * be an encoder. Locate the connector.
+		 */
+		encoder = entity;
+		connector = of_graph_get_remote_port_parent(ep_node);
+		of_node_put(ep_node);
+
+		if (!connector) {
+			dev_warn(rcdu->dev,
+				 "no connector for encoder %s, skipping\n",
+				 encoder->full_name);
+			of_node_put(entity_ep_node);
+			of_node_put(encoder);
+			return 0;
+		}
+
+		break;
+	}
+
+	of_node_put(entity_ep_node);
+
+	if (encoder) {
+		/*
+		 * If an encoder has been found, get its type based on its
+		 * compatible string.
+		 */
+		unsigned int i;
+
+		for (i = 0; i < ARRAY_SIZE(encoders); ++i) {
+			if (of_device_is_compatible(encoder,
+						    encoders[i].compatible)) {
+				enc_type = encoders[i].type;
+				break;
+			}
+		}
+
+		if (i == ARRAY_SIZE(encoders)) {
+			dev_warn(rcdu->dev,
+				 "unknown encoder type for %s, skipping\n",
+				 encoder->full_name);
+			of_node_put(encoder);
+			of_node_put(connector);
+			return 0;
+		}
+	} else {
+		/*
+		 * If no encoder has been found the entity must be the
+		 * connector.
+		 */
+		connector = entity;
+	}
+
+	ret = rcar_du_encoder_init(rcdu, enc_type, output, NULL, connector);
+	of_node_put(encoder);
+	of_node_put(connector);
+
+	return ret < 0 ? ret : 1;
+}
+
+static int rcar_du_encoders_init_dt(struct rcar_du_device *rcdu)
+{
+	struct device_node *np = rcdu->dev->of_node;
+	struct device_node *prev = NULL;
+	unsigned int num_encoders = 0;
+
+	/*
+	 * Iterate over the endpoints and create one encoder for each output
+	 * pipeline.
+	 */
+	while (1) {
+		struct device_node *ep_node;
+		enum rcar_du_output output;
+		struct of_endpoint ep;
+		unsigned int i;
+		int ret;
+
+		ep_node = of_graph_get_next_endpoint(np, prev);
+		of_node_put(prev);
+		prev = ep_node;
+
+		if (ep_node == NULL)
+			break;
+
+		ret = of_graph_parse_endpoint(ep_node, &ep);
+		if (ret < 0) {
+			of_node_put(ep_node);
+			return ret;
+		}
+
+		/* Find the output route corresponding to the port number. */
+		for (i = 0; i < RCAR_DU_OUTPUT_MAX; ++i) {
+			if (rcdu->info->routes[i].possible_crtcs &&
+			    rcdu->info->routes[i].port == ep.port) {
+				output = i;
+				break;
+			}
+		}
+
+		if (i == RCAR_DU_OUTPUT_MAX) {
+			dev_warn(rcdu->dev,
+				 "port %u references unexisting output, skipping\n",
+				 ep.port);
+			continue;
+		}
+
+		/* Process the output pipeline. */
+		ret = rcar_du_encoders_init_dt_one(rcdu, output, &ep);
+		if (ret < 0) {
+			of_node_put(ep_node);
+			return ret;
+		}
+
+		num_encoders += ret;
+	}
+
+	return num_encoders;
+}
+
 int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 {
 	static const unsigned int mmio_offsets[] = {
@@ -188,6 +398,7 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 	struct drm_device *dev = rcdu->ddev;
 	struct drm_encoder *encoder;
 	struct drm_fbdev_cma *fbdev;
+	unsigned int num_encoders;
 	unsigned int num_groups;
 	unsigned int i;
 	int ret;
@@ -231,28 +442,15 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 	if (ret < 0)
 		return ret;
 
-	for (i = 0; i < rcdu->pdata->num_encoders; ++i) {
-		const struct rcar_du_encoder_data *pdata =
-			&rcdu->pdata->encoders[i];
-		const struct rcar_du_output_routing *route =
-			&rcdu->info->routes[pdata->output];
-
-		if (pdata->type == RCAR_DU_ENCODER_UNUSED)
-			continue;
+	if (rcdu->pdata)
+		ret = rcar_du_encoders_init_pdata(rcdu);
+	else
+		ret = rcar_du_encoders_init_dt(rcdu);
 
-		if (pdata->output >= RCAR_DU_OUTPUT_MAX ||
-		    route->possible_crtcs == 0) {
-			dev_warn(rcdu->dev,
-				 "encoder %u references unexisting output %u, skipping\n",
-				 i, pdata->output);
-			continue;
-		}
+	if (ret < 0)
+		return ret;
 
-		ret = rcar_du_encoder_init(rcdu, pdata->type, pdata->output,
-					   pdata);
-		if (ret < 0)
-			return ret;
-	}
+	num_encoders = ret;
 
 	/* Set the possible CRTCs and possible clones. There's always at least
 	 * one way for all encoders to clone each other, set all bits in the
@@ -264,7 +462,7 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu)
 			&rcdu->info->routes[renc->output];
 
 		encoder->possible_crtcs = route->possible_crtcs;
-		encoder->possible_clones = (1 << rcdu->pdata->num_encoders) - 1;
+		encoder->possible_clones = (1 << num_encoders) - 1;
 	}
 
 	/* Now that the CRTCs have been initialized register the planes. */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.h b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
index 5750e6af5655..07951d5fe38b 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_kms.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_kms.h  --  R-Car Display Unit Mode Setting
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
index 289048d1c7b2..115eed20db12 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdscon.c  --  R-Car Display Unit LVDS Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -15,6 +15,10 @@
 #include <drm/drm_crtc.h>
 #include <drm/drm_crtc_helper.h>
 
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+#include <video/videomode.h>
+
 #include "rcar_du_drv.h"
 #include "rcar_du_encoder.h"
 #include "rcar_du_kms.h"
@@ -23,7 +27,7 @@
 struct rcar_du_lvds_connector {
 	struct rcar_du_connector connector;
 
-	const struct rcar_du_panel_data *panel;
+	struct rcar_du_panel_data panel;
 };
 
 #define to_rcar_lvds_connector(c) \
@@ -40,18 +44,9 @@ static int rcar_du_lvds_connector_get_modes(struct drm_connector *connector)
 		return 0;
 
 	mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
-	mode->clock = lvdscon->panel->mode.clock;
-	mode->hdisplay = lvdscon->panel->mode.hdisplay;
-	mode->hsync_start = lvdscon->panel->mode.hsync_start;
-	mode->hsync_end = lvdscon->panel->mode.hsync_end;
-	mode->htotal = lvdscon->panel->mode.htotal;
-	mode->vdisplay = lvdscon->panel->mode.vdisplay;
-	mode->vsync_start = lvdscon->panel->mode.vsync_start;
-	mode->vsync_end = lvdscon->panel->mode.vsync_end;
-	mode->vtotal = lvdscon->panel->mode.vtotal;
-	mode->flags = lvdscon->panel->mode.flags;
-
-	drm_mode_set_name(mode);
+
+	drm_display_mode_from_videomode(&lvdscon->panel.mode, mode);
+
 	drm_mode_probed_add(connector, mode);
 
 	return 1;
@@ -64,7 +59,7 @@ static const struct drm_connector_helper_funcs connector_helper_funcs = {
 
 static void rcar_du_lvds_connector_destroy(struct drm_connector *connector)
 {
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 }
 
@@ -83,7 +78,8 @@ static const struct drm_connector_funcs connector_funcs = {
 
 int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 				struct rcar_du_encoder *renc,
-				const struct rcar_du_panel_data *panel)
+				const struct rcar_du_panel_data *panel,
+				/* TODO const */ struct device_node *np)
 {
 	struct rcar_du_lvds_connector *lvdscon;
 	struct drm_connector *connector;
@@ -93,11 +89,24 @@ int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 	if (lvdscon == NULL)
 		return -ENOMEM;
 
-	lvdscon->panel = panel;
+	if (panel) {
+		lvdscon->panel = *panel;
+	} else {
+		struct display_timing timing;
+
+		ret = of_get_display_timing(np, "panel-timing", &timing);
+		if (ret < 0)
+			return ret;
+
+		videomode_from_timing(&timing, &lvdscon->panel.mode);
+
+		of_property_read_u32(np, "width-mm", &lvdscon->panel.width_mm);
+		of_property_read_u32(np, "height-mm", &lvdscon->panel.height_mm);
+	}
 
 	connector = &lvdscon->connector.connector;
-	connector->display_info.width_mm = panel->width_mm;
-	connector->display_info.height_mm = panel->height_mm;
+	connector->display_info.width_mm = lvdscon->panel.width_mm;
+	connector->display_info.height_mm = lvdscon->panel.height_mm;
 
 	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
 				 DRM_MODE_CONNECTOR_LVDS);
@@ -105,7 +114,7 @@ int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 		return ret;
 
 	drm_connector_helper_add(connector, &connector_helper_funcs);
-	ret = drm_sysfs_connector_add(connector);
+	ret = drm_connector_register(connector);
 	if (ret < 0)
 		return ret;
 
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
index bff8683699ca..d11424d537f9 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdscon.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdscon.h  --  R-Car Display Unit LVDS Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -20,6 +20,7 @@ struct rcar_du_panel_data;
 
 int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
 				struct rcar_du_encoder *renc,
-				const struct rcar_du_panel_data *panel);
+				const struct rcar_du_panel_data *panel,
+				struct device_node *np);
 
 #endif /* __RCAR_DU_LVDSCON_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
index df30a075d793..7cfb48ce1791 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdsenc.c  --  R-Car Display Unit LVDS Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
index 7051c6de19ae..3303a55cec79 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvdsenc.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_lvdsenc.h  --  R-Car Display Unit LVDS Encoder
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.c b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
index 3fb69d9ae61b..72a7cb47bd9f 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_plane.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_plane.c  --  R-Car Display Unit Planes
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.h b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
index f94f9ce84998..3021288b1a89 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_plane.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_plane.h  --  R-Car Display Unit Planes
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
index ccfe64c7188f..564a723ede03 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.c
@@ -1,7 +1,7 @@
 /*
  * rcar_du_vgacon.c  --  R-Car Display Unit VGA Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -32,7 +32,7 @@ static const struct drm_connector_helper_funcs connector_helper_funcs = {
 
 static void rcar_du_vga_connector_destroy(struct drm_connector *connector)
 {
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 }
 
@@ -70,7 +70,7 @@ int rcar_du_vga_connector_init(struct rcar_du_device *rcdu,
 		return ret;
 
 	drm_connector_helper_add(connector, &connector_helper_funcs);
-	ret = drm_sysfs_connector_add(connector);
+	ret = drm_connector_register(connector);
 	if (ret < 0)
 		return ret;
 
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
index b12b0cf7f117..112f50316e01 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vgacon.h
@@ -1,7 +1,7 @@
 /*
  * rcar_du_vgacon.h  --  R-Car Display Unit VGA Connector
  *
- * Copyright (C) 2013 Renesas Corporation
+ * Copyright (C) 2013-2014 Renesas Electronics Corporation
  *
  * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/savage/savage_bci.c b/drivers/gpu/drm/savage/savage_bci.c
index c97cdc9ab239..d47dff95fe52 100644
--- a/drivers/gpu/drm/savage/savage_bci.c
+++ b/drivers/gpu/drm/savage/savage_bci.c
@@ -556,7 +556,7 @@ int savage_driver_load(struct drm_device *dev, unsigned long chipset)
 /*
  * Initialize mappings. On Savage4 and SavageIX the alignment
  * and size of the aperture is not suitable for automatic MTRR setup
- * in drm_addmap. Therefore we add them manually before the maps are
+ * in drm_legacy_addmap. Therefore we add them manually before the maps are
  * initialized, and tear them down on last close.
  */
 int savage_driver_firstopen(struct drm_device *dev)
@@ -624,19 +624,20 @@ int savage_driver_firstopen(struct drm_device *dev)
 		/* Automatic MTRR setup will do the right thing. */
 	}
 
-	ret = drm_addmap(dev, mmio_base, SAVAGE_MMIO_SIZE, _DRM_REGISTERS,
-			 _DRM_READ_ONLY, &dev_priv->mmio);
+	ret = drm_legacy_addmap(dev, mmio_base, SAVAGE_MMIO_SIZE,
+				_DRM_REGISTERS, _DRM_READ_ONLY,
+				&dev_priv->mmio);
 	if (ret)
 		return ret;
 
-	ret = drm_addmap(dev, fb_base, fb_size, _DRM_FRAME_BUFFER,
-			 _DRM_WRITE_COMBINING, &dev_priv->fb);
+	ret = drm_legacy_addmap(dev, fb_base, fb_size, _DRM_FRAME_BUFFER,
+				_DRM_WRITE_COMBINING, &dev_priv->fb);
 	if (ret)
 		return ret;
 
-	ret = drm_addmap(dev, aperture_base, SAVAGE_APERTURE_SIZE,
-			 _DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING,
-			 &dev_priv->aperture);
+	ret = drm_legacy_addmap(dev, aperture_base, SAVAGE_APERTURE_SIZE,
+				_DRM_FRAME_BUFFER, _DRM_WRITE_COMBINING,
+				&dev_priv->aperture);
 	return ret;
 }
 
@@ -698,14 +699,14 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	dev_priv->texture_offset = init->texture_offset;
 	dev_priv->texture_size = init->texture_size;
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		savage_do_cleanup_bci(dev);
 		return -EINVAL;
 	}
 	if (init->status_offset != 0) {
-		dev_priv->status = drm_core_findmap(dev, init->status_offset);
+		dev_priv->status = drm_legacy_findmap(dev, init->status_offset);
 		if (!dev_priv->status) {
 			DRM_ERROR("could not find shadow status region!\n");
 			savage_do_cleanup_bci(dev);
@@ -716,14 +717,14 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	}
 	if (dev_priv->dma_type == SAVAGE_DMA_AGP && init->buffers_offset) {
 		dev->agp_buffer_token = init->buffers_offset;
-		dev->agp_buffer_map = drm_core_findmap(dev,
+		dev->agp_buffer_map = drm_legacy_findmap(dev,
 						       init->buffers_offset);
 		if (!dev->agp_buffer_map) {
 			DRM_ERROR("could not find DMA buffer region!\n");
 			savage_do_cleanup_bci(dev);
 			return -EINVAL;
 		}
-		drm_core_ioremap(dev->agp_buffer_map, dev);
+		drm_legacy_ioremap(dev->agp_buffer_map, dev);
 		if (!dev->agp_buffer_map->handle) {
 			DRM_ERROR("failed to ioremap DMA buffer region!\n");
 			savage_do_cleanup_bci(dev);
@@ -732,7 +733,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 	}
 	if (init->agp_textures_offset) {
 		dev_priv->agp_textures =
-		    drm_core_findmap(dev, init->agp_textures_offset);
+		    drm_legacy_findmap(dev, init->agp_textures_offset);
 		if (!dev_priv->agp_textures) {
 			DRM_ERROR("could not find agp texture region!\n");
 			savage_do_cleanup_bci(dev);
@@ -755,7 +756,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 			savage_do_cleanup_bci(dev);
 			return -EINVAL;
 		}
-		dev_priv->cmd_dma = drm_core_findmap(dev, init->cmd_dma_offset);
+		dev_priv->cmd_dma = drm_legacy_findmap(dev, init->cmd_dma_offset);
 		if (!dev_priv->cmd_dma) {
 			DRM_ERROR("could not find command DMA region!\n");
 			savage_do_cleanup_bci(dev);
@@ -768,7 +769,7 @@ static int savage_do_init_bci(struct drm_device * dev, drm_savage_init_t * init)
 				savage_do_cleanup_bci(dev);
 				return -EINVAL;
 			}
-			drm_core_ioremap(dev_priv->cmd_dma, dev);
+			drm_legacy_ioremap(dev_priv->cmd_dma, dev);
 			if (!dev_priv->cmd_dma->handle) {
 				DRM_ERROR("failed to ioremap command "
 					  "DMA region!\n");
@@ -894,11 +895,11 @@ static int savage_do_cleanup_bci(struct drm_device * dev)
 	} else if (dev_priv->cmd_dma && dev_priv->cmd_dma->handle &&
 		   dev_priv->cmd_dma->type == _DRM_AGP &&
 		   dev_priv->dma_type == SAVAGE_DMA_AGP)
-		drm_core_ioremapfree(dev_priv->cmd_dma, dev);
+		drm_legacy_ioremapfree(dev_priv->cmd_dma, dev);
 
 	if (dev_priv->dma_type == SAVAGE_DMA_AGP &&
 	    dev->agp_buffer_map && dev->agp_buffer_map->handle) {
-		drm_core_ioremapfree(dev->agp_buffer_map, dev);
+		drm_legacy_ioremapfree(dev->agp_buffer_map, dev);
 		/* make sure the next instance (which may be running
 		 * in PCI mode) doesn't try to use an old
 		 * agp_buffer_map. */
@@ -1050,7 +1051,7 @@ void savage_reclaim_buffers(struct drm_device *dev, struct drm_file *file_priv)
 		return;
 
 	if (file_priv->master && file_priv->master->lock.hw_lock) {
-		drm_idlelock_take(&file_priv->master->lock);
+		drm_legacy_idlelock_take(&file_priv->master->lock);
 		release_idlelock = 1;
 	}
 
@@ -1069,7 +1070,7 @@ void savage_reclaim_buffers(struct drm_device *dev, struct drm_file *file_priv)
 	}
 
 	if (release_idlelock)
-		drm_idlelock_release(&file_priv->master->lock);
+		drm_legacy_idlelock_release(&file_priv->master->lock);
 }
 
 const struct drm_ioctl_desc savage_ioctls[] = {
diff --git a/drivers/gpu/drm/savage/savage_drv.c b/drivers/gpu/drm/savage/savage_drv.c
index 3c030216e888..21aed1febeb4 100644
--- a/drivers/gpu/drm/savage/savage_drv.c
+++ b/drivers/gpu/drm/savage/savage_drv.c
@@ -40,7 +40,7 @@ static const struct file_operations savage_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -57,6 +57,7 @@ static struct drm_driver driver = {
 	.preclose = savage_reclaim_buffers,
 	.lastclose = savage_driver_lastclose,
 	.unload = savage_driver_unload,
+	.set_busid = drm_pci_set_busid,
 	.ioctls = savage_ioctls,
 	.dma_ioctl = savage_bci_buffers,
 	.fops = &savage_driver_fops,
diff --git a/drivers/gpu/drm/savage/savage_drv.h b/drivers/gpu/drm/savage/savage_drv.h
index 335f8fcf1041..37b699571ad0 100644
--- a/drivers/gpu/drm/savage/savage_drv.h
+++ b/drivers/gpu/drm/savage/savage_drv.h
@@ -26,6 +26,8 @@
 #ifndef __SAVAGE_DRV_H__
 #define __SAVAGE_DRV_H__
 
+#include <drm/drm_legacy.h>
+
 #define DRIVER_AUTHOR	"Felix Kuehling"
 
 #define DRIVER_NAME	"savage"
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_backlight.c b/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
index 463aee18f774..33dd41afea0e 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_backlight.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_backlight.c  --  SH Mobile DRM Backlight
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_backlight.h b/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
index 9477595d2ff3..bac719ecc301 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_backlight.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_backlight.h  --  SH Mobile DRM Backlight
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
index faf176b2daf9..0ddce4d046d9 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_crtc.c  --  SH Mobile DRM CRTCs
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -692,7 +692,7 @@ static void shmob_drm_connector_destroy(struct drm_connector *connector)
 	struct shmob_drm_connector *scon = to_shmob_connector(connector);
 
 	shmob_drm_backlight_exit(scon);
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 }
 
@@ -726,7 +726,7 @@ int shmob_drm_connector_create(struct shmob_drm_device *sdev,
 		return ret;
 
 	drm_connector_helper_add(connector, &connector_helper_funcs);
-	ret = drm_sysfs_connector_add(connector);
+	ret = drm_connector_register(connector);
 	if (ret < 0)
 		goto err_cleanup;
 
@@ -749,7 +749,7 @@ int shmob_drm_connector_create(struct shmob_drm_device *sdev,
 err_backlight:
 	shmob_drm_backlight_exit(&sdev->connector);
 err_sysfs:
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 err_cleanup:
 	drm_connector_cleanup(connector);
 	return ret;
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
index e5bd109c4c38..eddad6dcc88a 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_crtc.h  --  SH Mobile DRM CRTCs
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
index 82c84c7fd4f6..e62cbde81e50 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_drv.c  --  SH Mobile DRM driver
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
@@ -267,6 +267,7 @@ static struct drm_driver shmob_drm_driver = {
 	.load			= shmob_drm_load,
 	.unload			= shmob_drm_unload,
 	.preclose		= shmob_drm_preclose,
+	.set_busid		= drm_platform_set_busid,
 	.irq_handler		= shmob_drm_irq,
 	.get_vblank_counter	= drm_vblank_count,
 	.enable_vblank		= shmob_drm_enable_vblank,
@@ -297,7 +298,7 @@ static struct drm_driver shmob_drm_driver = {
  * Power management
  */
 
-#if CONFIG_PM_SLEEP
+#ifdef CONFIG_PM_SLEEP
 static int shmob_drm_pm_suspend(struct device *dev)
 {
 	struct shmob_drm_device *sdev = dev_get_drvdata(dev);
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.h b/drivers/gpu/drm/shmobile/shmob_drm_drv.h
index 4d46b811b5a7..02ea315ba69a 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm.h  --  SH Mobile DRM driver
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.c b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
index fc0ef0ca7d04..aaf98ace4a90 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_kms.c  --  SH Mobile DRM Mode Setting
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.h b/drivers/gpu/drm/shmobile/shmob_drm_kms.h
index 9495c9111308..06d5b7caa026 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_kms.h  --  SH Mobile DRM Mode Setting
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.c b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
index 060ae03e5f9b..1805bb23b113 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_plane.c  --  SH Mobile DRM Planes
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.h b/drivers/gpu/drm/shmobile/shmob_drm_plane.h
index 99623d05e3b0..a58cc1fc3240 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_plane.h  --  SH Mobile DRM Planes
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_regs.h b/drivers/gpu/drm/shmobile/shmob_drm_regs.h
index 7923cdd6368e..ea17d4415b9e 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_regs.h
+++ b/drivers/gpu/drm/shmobile/shmob_drm_regs.h
@@ -1,7 +1,7 @@
 /*
  * shmob_drm_regs.h  --  SH Mobile DRM registers
  *
- * Copyright (C) 2012 Renesas Corporation
+ * Copyright (C) 2012 Renesas Electronics Corporation
  *
  * Laurent Pinchart (laurent.pinchart@ideasonboard.com)
  *
diff --git a/drivers/gpu/drm/sis/sis_drv.c b/drivers/gpu/drm/sis/sis_drv.c
index 756f787b7143..79bce76cb8f7 100644
--- a/drivers/gpu/drm/sis/sis_drv.c
+++ b/drivers/gpu/drm/sis/sis_drv.c
@@ -70,7 +70,7 @@ static const struct file_operations sis_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -108,6 +108,7 @@ static struct drm_driver driver = {
 	.open = sis_driver_open,
 	.preclose = sis_reclaim_buffers_locked,
 	.postclose = sis_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 	.dma_quiescent = sis_idle,
 	.lastclose = sis_lastclose,
 	.ioctls = sis_ioctls,
diff --git a/drivers/gpu/drm/sis/sis_drv.h b/drivers/gpu/drm/sis/sis_drv.h
index c31c0253054d..16f972b2a76a 100644
--- a/drivers/gpu/drm/sis/sis_drv.h
+++ b/drivers/gpu/drm/sis/sis_drv.h
@@ -28,6 +28,8 @@
 #ifndef _SIS_DRV_H_
 #define _SIS_DRV_H_
 
+#include <drm/drm_legacy.h>
+
 /* General customization:
  */
 
diff --git a/drivers/gpu/drm/sis/sis_mm.c b/drivers/gpu/drm/sis/sis_mm.c
index 77f288e4a0a6..93ad8a5704d1 100644
--- a/drivers/gpu/drm/sis/sis_mm.c
+++ b/drivers/gpu/drm/sis/sis_mm.c
@@ -319,12 +319,12 @@ void sis_reclaim_buffers_locked(struct drm_device *dev,
 	if (!(file->minor->master && file->master->lock.hw_lock))
 		return;
 
-	drm_idlelock_take(&file->master->lock);
+	drm_legacy_idlelock_take(&file->master->lock);
 
 	mutex_lock(&dev->struct_mutex);
 	if (list_empty(&file_priv->obj_list)) {
 		mutex_unlock(&dev->struct_mutex);
-		drm_idlelock_release(&file->master->lock);
+		drm_legacy_idlelock_release(&file->master->lock);
 
 		return;
 	}
@@ -345,7 +345,7 @@ void sis_reclaim_buffers_locked(struct drm_device *dev,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	drm_idlelock_release(&file->master->lock);
+	drm_legacy_idlelock_release(&file->master->lock);
 
 	return;
 }
diff --git a/drivers/gpu/drm/sti/Kconfig b/drivers/gpu/drm/sti/Kconfig
new file mode 100644
index 000000000000..ae8850f3e63b
--- /dev/null
+++ b/drivers/gpu/drm/sti/Kconfig
@@ -0,0 +1,15 @@
+config DRM_STI
+	tristate "DRM Support for STMicroelectronics SoC stiH41x Series"
+	depends on DRM && (SOC_STIH415 || SOC_STIH416 || ARCH_MULTIPLATFORM)
+	select RESET_CONTROLLER
+	select DRM_KMS_HELPER
+	select DRM_GEM_CMA_HELPER
+	select DRM_KMS_CMA_HELPER
+	help
+	  Choose this option to enable DRM on STM stiH41x chipset
+
+config DRM_STI_FBDEV
+	bool "DRM frame buffer device for STMicroelectronics SoC stiH41x Serie"
+	depends on DRM_STI
+	help
+	  Choose this option to enable FBDEV on top of DRM for STM stiH41x chipset
diff --git a/drivers/gpu/drm/sti/Makefile b/drivers/gpu/drm/sti/Makefile
new file mode 100644
index 000000000000..04ac2ceef27f
--- /dev/null
+++ b/drivers/gpu/drm/sti/Makefile
@@ -0,0 +1,21 @@
+sticompositor-y := \
+	sti_layer.o \
+	sti_mixer.o \
+	sti_gdp.o \
+	sti_vid.o \
+	sti_compositor.o \
+	sti_drm_crtc.o \
+	sti_drm_plane.o
+
+stihdmi-y := sti_hdmi.o \
+	sti_hdmi_tx3g0c55phy.o \
+	sti_hdmi_tx3g4c28phy.o \
+
+obj-$(CONFIG_DRM_STI) = \
+	sti_vtg.o \
+	sti_vtac.o \
+	stihdmi.o \
+	sti_hda.o \
+	sti_tvout.o \
+	sticompositor.o \
+	sti_drm_drv.o
\ No newline at end of file
diff --git a/drivers/gpu/drm/sti/NOTES b/drivers/gpu/drm/sti/NOTES
new file mode 100644
index 000000000000..57e257969198
--- /dev/null
+++ b/drivers/gpu/drm/sti/NOTES
@@ -0,0 +1,58 @@
+1. stiH display hardware IP
+---------------------------
+The STMicroelectronics stiH SoCs use a common chain of HW display IP blocks:
+- The High Quality Video Display Processor (HQVDP) gets video frames from a
+  video decoder and does high quality video processing, including scaling.
+
+- The Compositor is a multiplane, dual-mixer (Main & Aux) digital processor. It
+  has several inputs:
+  - The graphics planes are internally processed by the Generic Display
+    Pipeline (GDP).
+  - The video plug (VID) connects to the HQVDP output.
+  - The cursor handles ... a cursor.
+- The TV OUT pre-formats (convert, clip, round) the compositor output data
+- The HDMI / DVO / HD Analog / SD analog IP builds the video signals
+  - DVO (Digital Video Output) handles a 24bits parallel signal
+  - The HD analog signal is typically driven by a YCbCr cable, supporting up to
+    1080i mode.
+  - The SD analog signal is typically used for legacy TV
+- The VTG (Video Timing Generators) build Vsync signals used by the other HW IP
+Note that some stiH drivers support only a subset of thee HW IP.
+
+                  .-------------.   .-----------.   .-----------.
+GPU >-------------+GDP     Main |   |           +---+    HDMI   +--> HDMI
+GPU >-------------+GDP     mixer+---+           |   :===========:
+GPU >-------------+Cursor       |   |           +---+    DVO    +--> 24b//
+        -------   |  COMPOSITOR |   |  TV OUT   |   :===========:
+       |       |  |             |   |           +---+ HD analog +--> YCbCr
+Vid >--+ HQVDP +--+VID     Aux  +---+           |   :===========:
+dec    |       |  |        mixer|   |           +---+ SD analog +--> CVBS
+       '-------'  '-------------'   '-----------'   '-----------'
+                   .-----------.
+                   |       main+--> Vsync
+                   | VTG       |
+                   |        aux+--> Vsync
+                   '-----------'
+
+2. DRM / HW mapping
+-------------------
+These IP are mapped to the DRM objects as following:
+- The CRTCs are mapped to the Compositor Main and Aux Mixers
+- The Framebuffers and planes are mapped to the Compositor GDP (non video
+  buffers) and to HQVDP+VID (video buffers)
+- The Cursor is mapped to the Compositor Cursor
+- The Encoders are mapped to the TVOut
+- The Bridges/Connectors are mapped to the HDMI / DVO / HD Analog / SD analog
+
+FB & planes         Cursor      CRTC     Encoders    Bridges/Connectors
+   |                   |          |         |                       |
+   |                   |          |         |                       |
+   |              .-------------. | .-----------.   .-----------.   |
+   +------------> |GDP |   Main | | |       +-> |   |    HDMI   | <-+
+   +------------> |GDP v   mixer|<+ |       |   |   :===========:   |
+   |              |Cursor       | | |       +-> |   |    DVO    | <-+
+   |    -------   |  COMPOSITOR | | |TV OUT |   |   :===========:   |
+   |   |       |  |             | | |       +-> |   | HD analog | <-+
+   +-> | HQVDP |  |VID     Aux  |<+ |       |   |   :===========:   |
+       |       |  |        mixer|   |       +-> |   | SD analog | <-+
+       '-------'  '-------------'   '-----------'   '-----------'
diff --git a/drivers/gpu/drm/sti/sti_compositor.c b/drivers/gpu/drm/sti/sti_compositor.c
new file mode 100644
index 000000000000..390d93e9a06c
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_compositor.c
@@ -0,0 +1,281 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+
+#include <drm/drmP.h>
+
+#include "sti_compositor.h"
+#include "sti_drm_crtc.h"
+#include "sti_drm_drv.h"
+#include "sti_drm_plane.h"
+#include "sti_gdp.h"
+#include "sti_vtg.h"
+
+/*
+ * stiH407 compositor properties
+ */
+struct sti_compositor_data stih407_compositor_data = {
+	.nb_subdev = 6,
+	.subdev_desc = {
+			{STI_GPD_SUBDEV, (int)STI_GDP_0, 0x100},
+			{STI_GPD_SUBDEV, (int)STI_GDP_1, 0x200},
+			{STI_GPD_SUBDEV, (int)STI_GDP_2, 0x300},
+			{STI_GPD_SUBDEV, (int)STI_GDP_3, 0x400},
+			{STI_VID_SUBDEV, (int)STI_VID_0, 0x700},
+			{STI_MIXER_MAIN_SUBDEV, STI_MIXER_MAIN, 0xC00}
+	},
+};
+
+/*
+ * stiH416 compositor properties
+ * Note:
+ * on stih416 MIXER_AUX has a different base address from MIXER_MAIN
+ * Moreover, GDPx is different for Main and Aux Mixer. So this subdev map does
+ * not fit for stiH416 if we want to enable the MIXER_AUX.
+ */
+struct sti_compositor_data stih416_compositor_data = {
+	.nb_subdev = 3,
+	.subdev_desc = {
+			{STI_GPD_SUBDEV, (int)STI_GDP_0, 0x100},
+			{STI_GPD_SUBDEV, (int)STI_GDP_1, 0x200},
+			{STI_MIXER_MAIN_SUBDEV, STI_MIXER_MAIN, 0xC00}
+	},
+};
+
+static int sti_compositor_init_subdev(struct sti_compositor *compo,
+		struct sti_compositor_subdev_descriptor *desc,
+		unsigned int array_size)
+{
+	unsigned int i, mixer_id = 0, layer_id = 0;
+
+	for (i = 0; i < array_size; i++) {
+		switch (desc[i].type) {
+		case STI_MIXER_MAIN_SUBDEV:
+		case STI_MIXER_AUX_SUBDEV:
+			compo->mixer[mixer_id++] =
+			    sti_mixer_create(compo->dev, desc[i].id,
+					     compo->regs + desc[i].offset);
+			break;
+		case STI_GPD_SUBDEV:
+		case STI_VID_SUBDEV:
+			compo->layer[layer_id++] =
+			    sti_layer_create(compo->dev, desc[i].id,
+					     compo->regs + desc[i].offset);
+			break;
+			/* case STI_CURSOR_SUBDEV : TODO */
+		default:
+			DRM_ERROR("Unknow subdev compoment type\n");
+			return 1;
+		}
+
+	}
+	compo->nb_mixers = mixer_id;
+	compo->nb_layers = layer_id;
+
+	return 0;
+}
+
+static int sti_compositor_bind(struct device *dev, struct device *master,
+	void *data)
+{
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
+	unsigned int i, crtc = 0, plane = 0;
+	struct sti_drm_private *dev_priv = drm_dev->dev_private;
+	struct drm_plane *cursor = NULL;
+	struct drm_plane *primary = NULL;
+
+	dev_priv->compo = compo;
+
+	for (i = 0; i < compo->nb_layers; i++) {
+		if (compo->layer[i]) {
+			enum sti_layer_desc desc = compo->layer[i]->desc;
+			enum sti_layer_type type = desc & STI_LAYER_TYPE_MASK;
+			enum drm_plane_type plane_type = DRM_PLANE_TYPE_OVERLAY;
+
+			if (compo->mixer[crtc])
+				plane_type = DRM_PLANE_TYPE_PRIMARY;
+
+			switch (type) {
+			case STI_CUR:
+				cursor = sti_drm_plane_init(drm_dev,
+						compo->layer[i],
+						(1 << crtc) - 1,
+						DRM_PLANE_TYPE_CURSOR);
+				break;
+			case STI_GDP:
+			case STI_VID:
+				primary = sti_drm_plane_init(drm_dev,
+						compo->layer[i],
+						(1 << crtc) - 1, plane_type);
+				plane++;
+				break;
+			case STI_BCK:
+				break;
+			}
+
+			/* The first planes are reserved for primary planes*/
+			if (compo->mixer[crtc]) {
+				sti_drm_crtc_init(drm_dev, compo->mixer[crtc],
+						primary, cursor);
+				crtc++;
+				cursor = NULL;
+			}
+		}
+	}
+
+	drm_vblank_init(drm_dev, crtc);
+	/* Allow usage of vblank without having to call drm_irq_install */
+	drm_dev->irq_enabled = 1;
+
+	DRM_DEBUG_DRIVER("Initialized %d DRM CRTC(s) and %d DRM plane(s)\n",
+			 crtc, plane);
+	DRM_DEBUG_DRIVER("DRM plane(s) for VID/VDP not created yet\n");
+
+	return 0;
+}
+
+static void sti_compositor_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	/* do nothing */
+}
+
+static const struct component_ops sti_compositor_ops = {
+	.bind	= sti_compositor_bind,
+	.unbind	= sti_compositor_unbind,
+};
+
+static const struct of_device_id compositor_of_match[] = {
+	{
+		.compatible = "st,stih416-compositor",
+		.data = &stih416_compositor_data,
+	}, {
+		.compatible = "st,stih407-compositor",
+		.data = &stih407_compositor_data,
+	}, {
+		/* end node */
+	}
+};
+MODULE_DEVICE_TABLE(of, compositor_of_match);
+
+static int sti_compositor_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
+	struct device_node *vtg_np;
+	struct sti_compositor *compo;
+	struct resource *res;
+	int err;
+
+	compo = devm_kzalloc(dev, sizeof(*compo), GFP_KERNEL);
+	if (!compo) {
+		DRM_ERROR("Failed to allocate compositor context\n");
+		return -ENOMEM;
+	}
+	compo->dev = dev;
+	compo->vtg_vblank_nb.notifier_call = sti_drm_crtc_vblank_cb;
+
+	/* populate data structure depending on compatibility */
+	BUG_ON(!of_match_node(compositor_of_match, np)->data);
+
+	memcpy(&compo->data, of_match_node(compositor_of_match, np)->data,
+	       sizeof(struct sti_compositor_data));
+
+	/* Get Memory ressources */
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (res == NULL) {
+		DRM_ERROR("Get memory resource failed\n");
+		return -ENXIO;
+	}
+	compo->regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (compo->regs == NULL) {
+		DRM_ERROR("Register mapping failed\n");
+		return -ENXIO;
+	}
+
+	/* Get clock resources */
+	compo->clk_compo_main = devm_clk_get(dev, "compo_main");
+	if (IS_ERR(compo->clk_compo_main)) {
+		DRM_ERROR("Cannot get compo_main clock\n");
+		return PTR_ERR(compo->clk_compo_main);
+	}
+
+	compo->clk_compo_aux = devm_clk_get(dev, "compo_aux");
+	if (IS_ERR(compo->clk_compo_aux)) {
+		DRM_ERROR("Cannot get compo_aux clock\n");
+		return PTR_ERR(compo->clk_compo_aux);
+	}
+
+	compo->clk_pix_main = devm_clk_get(dev, "pix_main");
+	if (IS_ERR(compo->clk_pix_main)) {
+		DRM_ERROR("Cannot get pix_main clock\n");
+		return PTR_ERR(compo->clk_pix_main);
+	}
+
+	compo->clk_pix_aux = devm_clk_get(dev, "pix_aux");
+	if (IS_ERR(compo->clk_pix_aux)) {
+		DRM_ERROR("Cannot get pix_aux clock\n");
+		return PTR_ERR(compo->clk_pix_aux);
+	}
+
+	/* Get reset resources */
+	compo->rst_main = devm_reset_control_get(dev, "compo-main");
+	/* Take compo main out of reset */
+	if (!IS_ERR(compo->rst_main))
+		reset_control_deassert(compo->rst_main);
+
+	compo->rst_aux = devm_reset_control_get(dev, "compo-aux");
+	/* Take compo aux out of reset */
+	if (!IS_ERR(compo->rst_aux))
+		reset_control_deassert(compo->rst_aux);
+
+	vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 0);
+	if (vtg_np)
+		compo->vtg_main = of_vtg_find(vtg_np);
+
+	vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 1);
+	if (vtg_np)
+		compo->vtg_aux = of_vtg_find(vtg_np);
+
+	/* Initialize compositor subdevices */
+	err = sti_compositor_init_subdev(compo, compo->data.subdev_desc,
+					 compo->data.nb_subdev);
+	if (err)
+		return err;
+
+	platform_set_drvdata(pdev, compo);
+
+	return component_add(&pdev->dev, &sti_compositor_ops);
+}
+
+static int sti_compositor_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &sti_compositor_ops);
+	return 0;
+}
+
+static struct platform_driver sti_compositor_driver = {
+	.driver = {
+		.name = "sti-compositor",
+		.owner = THIS_MODULE,
+		.of_match_table = compositor_of_match,
+	},
+	.probe = sti_compositor_probe,
+	.remove = sti_compositor_remove,
+};
+
+module_platform_driver(sti_compositor_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_compositor.h b/drivers/gpu/drm/sti/sti_compositor.h
new file mode 100644
index 000000000000..3ea19db72e0f
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_compositor.h
@@ -0,0 +1,90 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_COMPOSITOR_H_
+#define _STI_COMPOSITOR_H_
+
+#include <linux/clk.h>
+#include <linux/kernel.h>
+
+#include "sti_layer.h"
+#include "sti_mixer.h"
+
+#define WAIT_NEXT_VSYNC_MS      50 /*ms*/
+
+#define STI_MAX_LAYER 8
+#define STI_MAX_MIXER 2
+
+enum sti_compositor_subdev_type {
+	STI_MIXER_MAIN_SUBDEV,
+	STI_MIXER_AUX_SUBDEV,
+	STI_GPD_SUBDEV,
+	STI_VID_SUBDEV,
+	STI_CURSOR_SUBDEV,
+};
+
+struct sti_compositor_subdev_descriptor {
+	enum sti_compositor_subdev_type type;
+	int id;
+	unsigned int offset;
+};
+
+/**
+ * STI Compositor data structure
+ *
+ * @nb_subdev: number of subdevices supported by the compositor
+ * @subdev_desc: subdev list description
+ */
+#define MAX_SUBDEV 9
+struct sti_compositor_data {
+	unsigned int nb_subdev;
+	struct sti_compositor_subdev_descriptor subdev_desc[MAX_SUBDEV];
+};
+
+/**
+ * STI Compositor structure
+ *
+ * @dev: driver device
+ * @regs: registers (main)
+ * @data: device data
+ * @clk_compo_main: clock for main compo
+ * @clk_compo_aux: clock for aux compo
+ * @clk_pix_main: pixel clock for main path
+ * @clk_pix_aux: pixel clock for aux path
+ * @rst_main: reset control of the main path
+ * @rst_aux: reset control of the aux path
+ * @mixer: array of mixers
+ * @vtg_main: vtg for main data path
+ * @vtg_aux: vtg for auxillary data path
+ * @layer: array of layers
+ * @nb_mixers: number of mixers for this compositor
+ * @nb_layers: number of layers (GDP,VID,...) for this compositor
+ * @enable: true if compositor is enable else false
+ * @vtg_vblank_nb: callback for VTG VSYNC notification
+ */
+struct sti_compositor {
+	struct device *dev;
+	void __iomem *regs;
+	struct sti_compositor_data data;
+	struct clk *clk_compo_main;
+	struct clk *clk_compo_aux;
+	struct clk *clk_pix_main;
+	struct clk *clk_pix_aux;
+	struct reset_control *rst_main;
+	struct reset_control *rst_aux;
+	struct sti_mixer *mixer[STI_MAX_MIXER];
+	struct sti_vtg *vtg_main;
+	struct sti_vtg *vtg_aux;
+	struct sti_layer *layer[STI_MAX_LAYER];
+	int nb_mixers;
+	int nb_layers;
+	bool enable;
+	struct notifier_block vtg_vblank_nb;
+};
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_drm_crtc.c b/drivers/gpu/drm/sti/sti_drm_crtc.c
new file mode 100644
index 000000000000..d2ae0c0e13be
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_crtc.c
@@ -0,0 +1,421 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "sti_compositor.h"
+#include "sti_drm_drv.h"
+#include "sti_drm_crtc.h"
+#include "sti_vtg.h"
+
+static void sti_drm_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	DRM_DEBUG_KMS("\n");
+}
+
+static void sti_drm_crtc_prepare(struct drm_crtc *crtc)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	struct device *dev = mixer->dev;
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+
+	compo->enable = true;
+
+	/* Prepare and enable the compo IP clock */
+	if (mixer->id == STI_MIXER_MAIN) {
+		if (clk_prepare_enable(compo->clk_compo_main))
+			DRM_INFO("Failed to prepare/enable compo_main clk\n");
+	} else {
+		if (clk_prepare_enable(compo->clk_compo_aux))
+			DRM_INFO("Failed to prepare/enable compo_aux clk\n");
+	}
+}
+
+static void sti_drm_crtc_commit(struct drm_crtc *crtc)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	struct device *dev = mixer->dev;
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+	struct sti_layer *layer;
+
+	if ((!mixer || !compo)) {
+		DRM_ERROR("Can not find mixer or compositor)\n");
+		return;
+	}
+
+	/* get GDP which is reserved to the CRTC FB */
+	layer = to_sti_layer(crtc->primary);
+	if (layer)
+		sti_layer_commit(layer);
+	else
+		DRM_ERROR("Can not find CRTC dedicated plane (GDP0)\n");
+
+	/* Enable layer on mixer */
+	if (sti_mixer_set_layer_status(mixer, layer, true))
+		DRM_ERROR("Can not enable layer at mixer\n");
+}
+
+static bool sti_drm_crtc_mode_fixup(struct drm_crtc *crtc,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	/* accept the provided drm_display_mode, do not fix it up */
+	return true;
+}
+
+static int
+sti_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
+		      struct drm_display_mode *adjusted_mode, int x, int y,
+		      struct drm_framebuffer *old_fb)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	struct device *dev = mixer->dev;
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+	struct sti_layer *layer;
+	struct clk *clk;
+	int rate = mode->clock * 1000;
+	int res;
+	unsigned int w, h;
+
+	DRM_DEBUG_KMS("CRTC:%d (%s) fb:%d mode:%d (%s)\n",
+		      crtc->base.id, sti_mixer_to_str(mixer),
+		      crtc->primary->fb->base.id, mode->base.id, mode->name);
+
+	DRM_DEBUG_KMS("%d %d %d %d %d %d %d %d %d %d 0x%x 0x%x\n",
+		      mode->vrefresh, mode->clock,
+		      mode->hdisplay,
+		      mode->hsync_start, mode->hsync_end,
+		      mode->htotal,
+		      mode->vdisplay,
+		      mode->vsync_start, mode->vsync_end,
+		      mode->vtotal, mode->type, mode->flags);
+
+	/* Set rate and prepare/enable pixel clock */
+	if (mixer->id == STI_MIXER_MAIN)
+		clk = compo->clk_pix_main;
+	else
+		clk = compo->clk_pix_aux;
+
+	res = clk_set_rate(clk, rate);
+	if (res < 0) {
+		DRM_ERROR("Cannot set rate (%dHz) for pix clk\n", rate);
+		return -EINVAL;
+	}
+	if (clk_prepare_enable(clk)) {
+		DRM_ERROR("Failed to prepare/enable pix clk\n");
+		return -EINVAL;
+	}
+
+	sti_vtg_set_config(mixer->id == STI_MIXER_MAIN ?
+			compo->vtg_main : compo->vtg_aux, &crtc->mode);
+
+	/* a GDP is reserved to the CRTC FB */
+	layer = to_sti_layer(crtc->primary);
+	if (!layer) {
+		DRM_ERROR("Can not find GDP0)\n");
+		return -EINVAL;
+	}
+
+	/* copy the mode data adjusted by mode_fixup() into crtc->mode
+	 * so that hardware can be set to proper mode
+	 */
+	memcpy(&crtc->mode, adjusted_mode, sizeof(*adjusted_mode));
+
+	res = sti_mixer_set_layer_depth(mixer, layer);
+	if (res) {
+		DRM_ERROR("Can not set layer depth\n");
+		return -EINVAL;
+	}
+	res = sti_mixer_active_video_area(mixer, &crtc->mode);
+	if (res) {
+		DRM_ERROR("Can not set active video area\n");
+		return -EINVAL;
+	}
+
+	w = crtc->primary->fb->width - x;
+	h = crtc->primary->fb->height - y;
+
+	return sti_layer_prepare(layer, crtc->primary->fb, &crtc->mode,
+			mixer->id, 0, 0, w, h, x, y, w, h);
+}
+
+static int sti_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				      struct drm_framebuffer *old_fb)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	struct sti_layer *layer;
+	unsigned int w, h;
+	int ret;
+
+	DRM_DEBUG_KMS("CRTC:%d (%s) fb:%d (%d,%d)\n",
+		      crtc->base.id, sti_mixer_to_str(mixer),
+		      crtc->primary->fb->base.id, x, y);
+
+	/* GDP is reserved to the CRTC FB */
+	layer = to_sti_layer(crtc->primary);
+	if (!layer) {
+		DRM_ERROR("Can not find GDP0)\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	w = crtc->primary->fb->width - crtc->x;
+	h = crtc->primary->fb->height - crtc->y;
+
+	ret = sti_layer_prepare(layer, crtc->primary->fb, &crtc->mode,
+				mixer->id, 0, 0, w, h,
+				crtc->x, crtc->y, w, h);
+	if (ret) {
+		DRM_ERROR("Can not prepare layer\n");
+		goto out;
+	}
+
+	sti_drm_crtc_commit(crtc);
+out:
+	return ret;
+}
+
+static void sti_drm_crtc_load_lut(struct drm_crtc *crtc)
+{
+	/* do nothing */
+}
+
+static void sti_drm_crtc_disable(struct drm_crtc *crtc)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	struct device *dev = mixer->dev;
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+	struct sti_layer *layer;
+
+	if (!compo->enable)
+		return;
+
+	DRM_DEBUG_KMS("CRTC:%d (%s)\n", crtc->base.id, sti_mixer_to_str(mixer));
+
+	/* Disable Background */
+	sti_mixer_set_background_status(mixer, false);
+
+	/* Disable GDP */
+	layer = to_sti_layer(crtc->primary);
+	if (!layer) {
+		DRM_ERROR("Cannot find GDP0\n");
+		return;
+	}
+
+	/* Disable layer at mixer level */
+	if (sti_mixer_set_layer_status(mixer, layer, false))
+		DRM_ERROR("Can not disable %s layer at mixer\n",
+				sti_layer_to_str(layer));
+
+	/* Wait a while to be sure that a Vsync event is received */
+	msleep(WAIT_NEXT_VSYNC_MS);
+
+	/* Then disable layer itself */
+	sti_layer_disable(layer);
+
+	drm_vblank_off(crtc->dev, mixer->id);
+
+	/* Disable pixel clock and compo IP clocks */
+	if (mixer->id == STI_MIXER_MAIN) {
+		clk_disable_unprepare(compo->clk_pix_main);
+		clk_disable_unprepare(compo->clk_compo_main);
+	} else {
+		clk_disable_unprepare(compo->clk_pix_aux);
+		clk_disable_unprepare(compo->clk_compo_aux);
+	}
+
+	compo->enable = false;
+}
+
+static struct drm_crtc_helper_funcs sti_crtc_helper_funcs = {
+	.dpms = sti_drm_crtc_dpms,
+	.prepare = sti_drm_crtc_prepare,
+	.commit = sti_drm_crtc_commit,
+	.mode_fixup = sti_drm_crtc_mode_fixup,
+	.mode_set = sti_drm_crtc_mode_set,
+	.mode_set_base = sti_drm_crtc_mode_set_base,
+	.load_lut = sti_drm_crtc_load_lut,
+	.disable = sti_drm_crtc_disable,
+};
+
+static int sti_drm_crtc_page_flip(struct drm_crtc *crtc,
+				  struct drm_framebuffer *fb,
+				  struct drm_pending_vblank_event *event,
+				  uint32_t page_flip_flags)
+{
+	struct drm_device *drm_dev = crtc->dev;
+	struct drm_framebuffer *old_fb;
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	unsigned long flags;
+	int ret;
+
+	DRM_DEBUG_KMS("fb %d --> fb %d\n",
+			crtc->primary->fb->base.id, fb->base.id);
+
+	mutex_lock(&drm_dev->struct_mutex);
+
+	old_fb = crtc->primary->fb;
+	crtc->primary->fb = fb;
+	ret = sti_drm_crtc_mode_set_base(crtc, crtc->x, crtc->y, old_fb);
+	if (ret) {
+		DRM_ERROR("failed\n");
+		crtc->primary->fb = old_fb;
+		goto out;
+	}
+
+	if (event) {
+		event->pipe = mixer->id;
+
+		ret = drm_vblank_get(drm_dev, event->pipe);
+		if (ret) {
+			DRM_ERROR("Cannot get vblank\n");
+			goto out;
+		}
+
+		spin_lock_irqsave(&drm_dev->event_lock, flags);
+		if (mixer->pending_event) {
+			drm_vblank_put(drm_dev, event->pipe);
+			ret = -EBUSY;
+		} else {
+			mixer->pending_event = event;
+		}
+		spin_unlock_irqrestore(&drm_dev->event_lock, flags);
+	}
+out:
+	mutex_unlock(&drm_dev->struct_mutex);
+	return ret;
+}
+
+static void sti_drm_crtc_destroy(struct drm_crtc *crtc)
+{
+	DRM_DEBUG_KMS("\n");
+	drm_crtc_cleanup(crtc);
+}
+
+static int sti_drm_crtc_set_property(struct drm_crtc *crtc,
+				     struct drm_property *property,
+				     uint64_t val)
+{
+	DRM_DEBUG_KMS("\n");
+	return 0;
+}
+
+int sti_drm_crtc_vblank_cb(struct notifier_block *nb,
+			   unsigned long event, void *data)
+{
+	struct drm_device *drm_dev;
+	struct sti_compositor *compo =
+		container_of(nb, struct sti_compositor, vtg_vblank_nb);
+	int *crtc = data;
+	unsigned long flags;
+	struct sti_drm_private *priv;
+
+	drm_dev = compo->mixer[*crtc]->drm_crtc.dev;
+	priv = drm_dev->dev_private;
+
+	if ((event != VTG_TOP_FIELD_EVENT) &&
+	    (event != VTG_BOTTOM_FIELD_EVENT)) {
+		DRM_ERROR("unknown event: %lu\n", event);
+		return -EINVAL;
+	}
+
+	drm_handle_vblank(drm_dev, *crtc);
+
+	spin_lock_irqsave(&drm_dev->event_lock, flags);
+	if (compo->mixer[*crtc]->pending_event) {
+		drm_send_vblank_event(drm_dev, -1,
+				compo->mixer[*crtc]->pending_event);
+		drm_vblank_put(drm_dev, *crtc);
+		compo->mixer[*crtc]->pending_event = NULL;
+	}
+	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
+
+	return 0;
+}
+
+int sti_drm_crtc_enable_vblank(struct drm_device *dev, int crtc)
+{
+	struct sti_drm_private *dev_priv = dev->dev_private;
+	struct sti_compositor *compo = dev_priv->compo;
+	struct notifier_block *vtg_vblank_nb = &compo->vtg_vblank_nb;
+
+	if (sti_vtg_register_client(crtc == STI_MIXER_MAIN ?
+			compo->vtg_main : compo->vtg_aux,
+			vtg_vblank_nb, crtc)) {
+		DRM_ERROR("Cannot register VTG notifier\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(sti_drm_crtc_enable_vblank);
+
+void sti_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
+{
+	struct sti_drm_private *priv = dev->dev_private;
+	struct sti_compositor *compo = priv->compo;
+	struct notifier_block *vtg_vblank_nb = &compo->vtg_vblank_nb;
+	unsigned long flags;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (sti_vtg_unregister_client(crtc == STI_MIXER_MAIN ?
+			compo->vtg_main : compo->vtg_aux, vtg_vblank_nb))
+		DRM_DEBUG_DRIVER("Warning: cannot unregister VTG notifier\n");
+
+	/* free the resources of the pending requests */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (compo->mixer[crtc]->pending_event) {
+		drm_vblank_put(dev, crtc);
+		compo->mixer[crtc]->pending_event = NULL;
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+}
+EXPORT_SYMBOL(sti_drm_crtc_disable_vblank);
+
+static struct drm_crtc_funcs sti_crtc_funcs = {
+	.set_config = drm_crtc_helper_set_config,
+	.page_flip = sti_drm_crtc_page_flip,
+	.destroy = sti_drm_crtc_destroy,
+	.set_property = sti_drm_crtc_set_property,
+};
+
+bool sti_drm_crtc_is_main(struct drm_crtc *crtc)
+{
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+
+	if (mixer->id == STI_MIXER_MAIN)
+		return true;
+
+	return false;
+}
+
+int sti_drm_crtc_init(struct drm_device *drm_dev, struct sti_mixer *mixer,
+		struct drm_plane *primary, struct drm_plane *cursor)
+{
+	struct drm_crtc *crtc = &mixer->drm_crtc;
+	int res;
+
+	res = drm_crtc_init_with_planes(drm_dev, crtc, primary, cursor,
+			&sti_crtc_funcs);
+	if (res) {
+		DRM_ERROR("Can not initialze CRTC\n");
+		return -EINVAL;
+	}
+
+	drm_crtc_helper_add(crtc, &sti_crtc_helper_funcs);
+
+	DRM_DEBUG_DRIVER("drm CRTC:%d mapped to %s\n",
+			 crtc->base.id, sti_mixer_to_str(mixer));
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/sti/sti_drm_crtc.h b/drivers/gpu/drm/sti/sti_drm_crtc.h
new file mode 100644
index 000000000000..caca8b14f017
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_crtc.h
@@ -0,0 +1,22 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_DRM_CRTC_H_
+#define _STI_DRM_CRTC_H_
+
+#include <drm/drmP.h>
+
+struct sti_mixer;
+
+int sti_drm_crtc_init(struct drm_device *drm_dev, struct sti_mixer *mixer,
+		struct drm_plane *primary, struct drm_plane *cursor);
+int sti_drm_crtc_enable_vblank(struct drm_device *dev, int crtc);
+void sti_drm_crtc_disable_vblank(struct drm_device *dev, int crtc);
+int sti_drm_crtc_vblank_cb(struct notifier_block *nb,
+		unsigned long event, void *data);
+bool sti_drm_crtc_is_main(struct drm_crtc *drm_crtc);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_drm_drv.c b/drivers/gpu/drm/sti/sti_drm_drv.c
new file mode 100644
index 000000000000..223d93c3a05d
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_drv.c
@@ -0,0 +1,241 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <drm/drmP.h>
+
+#include <linux/component.h>
+#include <linux/debugfs.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+
+#include "sti_drm_drv.h"
+#include "sti_drm_crtc.h"
+
+#define DRIVER_NAME	"sti"
+#define DRIVER_DESC	"STMicroelectronics SoC DRM"
+#define DRIVER_DATE	"20140601"
+#define DRIVER_MAJOR	1
+#define DRIVER_MINOR	0
+
+#define STI_MAX_FB_HEIGHT	4096
+#define STI_MAX_FB_WIDTH	4096
+
+static struct drm_mode_config_funcs sti_drm_mode_config_funcs = {
+	.fb_create = drm_fb_cma_create,
+};
+
+static void sti_drm_mode_config_init(struct drm_device *dev)
+{
+	dev->mode_config.min_width = 0;
+	dev->mode_config.min_height = 0;
+
+	/*
+	 * set max width and height as default value.
+	 * this value would be used to check framebuffer size limitation
+	 * at drm_mode_addfb().
+	 */
+	dev->mode_config.max_width = STI_MAX_FB_HEIGHT;
+	dev->mode_config.max_height = STI_MAX_FB_WIDTH;
+
+	dev->mode_config.funcs = &sti_drm_mode_config_funcs;
+}
+
+static int sti_drm_load(struct drm_device *dev, unsigned long flags)
+{
+	struct sti_drm_private *private;
+	int ret;
+
+	private = kzalloc(sizeof(struct sti_drm_private), GFP_KERNEL);
+	if (!private) {
+		DRM_ERROR("Failed to allocate private\n");
+		return -ENOMEM;
+	}
+	dev->dev_private = (void *)private;
+	private->drm_dev = dev;
+
+	drm_mode_config_init(dev);
+	drm_kms_helper_poll_init(dev);
+
+	sti_drm_mode_config_init(dev);
+
+	ret = component_bind_all(dev->dev, dev);
+	if (ret)
+		return ret;
+
+	drm_helper_disable_unused_functions(dev);
+
+#ifdef CONFIG_DRM_STI_FBDEV
+	drm_fbdev_cma_init(dev, 32,
+		   dev->mode_config.num_crtc,
+		   dev->mode_config.num_connector);
+#endif
+	return 0;
+}
+
+static const struct file_operations sti_drm_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.mmap = drm_gem_cma_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.unlocked_ioctl = drm_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl = drm_compat_ioctl,
+#endif
+	.release = drm_release,
+};
+
+static struct dma_buf *sti_drm_gem_prime_export(struct drm_device *dev,
+						struct drm_gem_object *obj,
+						int flags)
+{
+	/* we want to be able to write in mmapped buffer */
+	flags |= O_RDWR;
+	return drm_gem_prime_export(dev, obj, flags);
+}
+
+static struct drm_driver sti_drm_driver = {
+	.driver_features = DRIVER_HAVE_IRQ | DRIVER_MODESET |
+	    DRIVER_GEM | DRIVER_PRIME,
+	.load = sti_drm_load,
+	.gem_free_object = drm_gem_cma_free_object,
+	.gem_vm_ops = &drm_gem_cma_vm_ops,
+	.dumb_create = drm_gem_cma_dumb_create,
+	.dumb_map_offset = drm_gem_cma_dumb_map_offset,
+	.dumb_destroy = drm_gem_dumb_destroy,
+	.fops = &sti_drm_driver_fops,
+
+	.get_vblank_counter = drm_vblank_count,
+	.enable_vblank = sti_drm_crtc_enable_vblank,
+	.disable_vblank = sti_drm_crtc_disable_vblank,
+
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_export = sti_drm_gem_prime_export,
+	.gem_prime_import = drm_gem_prime_import,
+	.gem_prime_get_sg_table = drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap = drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap = drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap = drm_gem_cma_prime_mmap,
+
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+};
+
+static int compare_of(struct device *dev, void *data)
+{
+	return dev->of_node == data;
+}
+
+static int sti_drm_bind(struct device *dev)
+{
+	return drm_platform_init(&sti_drm_driver, to_platform_device(dev));
+}
+
+static void sti_drm_unbind(struct device *dev)
+{
+	drm_put_dev(dev_get_drvdata(dev));
+}
+
+static const struct component_master_ops sti_drm_ops = {
+	.bind = sti_drm_bind,
+	.unbind = sti_drm_unbind,
+};
+
+static int sti_drm_master_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *node = dev->parent->of_node;
+	struct device_node *child_np;
+	struct component_match *match = NULL;
+
+	dma_set_coherent_mask(dev, DMA_BIT_MASK(32));
+
+	child_np = of_get_next_available_child(node, NULL);
+
+	while (child_np) {
+		component_match_add(dev, &match, compare_of, child_np);
+		of_node_put(child_np);
+		child_np = of_get_next_available_child(node, child_np);
+	}
+
+	return component_master_add_with_match(dev, &sti_drm_ops, match);
+}
+
+static int sti_drm_master_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &sti_drm_ops);
+	return 0;
+}
+
+static struct platform_driver sti_drm_master_driver = {
+	.probe = sti_drm_master_probe,
+	.remove = sti_drm_master_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = DRIVER_NAME "__master",
+	},
+};
+
+static int sti_drm_platform_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *node = dev->of_node;
+	struct platform_device *master;
+
+	of_platform_populate(node, NULL, NULL, dev);
+
+	platform_driver_register(&sti_drm_master_driver);
+	master = platform_device_register_resndata(dev,
+			DRIVER_NAME "__master", -1,
+			NULL, 0, NULL, 0);
+	if (IS_ERR(master))
+               return PTR_ERR(master);
+
+	platform_set_drvdata(pdev, master);
+	return 0;
+}
+
+static int sti_drm_platform_remove(struct platform_device *pdev)
+{
+	struct platform_device *master = platform_get_drvdata(pdev);
+
+	of_platform_depopulate(&pdev->dev);
+	platform_device_unregister(master);
+	platform_driver_unregister(&sti_drm_master_driver);
+	return 0;
+}
+
+static const struct of_device_id sti_drm_dt_ids[] = {
+	{ .compatible = "st,sti-display-subsystem", },
+	{ /* end node */ },
+};
+MODULE_DEVICE_TABLE(of, sti_drm_dt_ids);
+
+static struct platform_driver sti_drm_platform_driver = {
+	.probe = sti_drm_platform_probe,
+	.remove = sti_drm_platform_remove,
+	.driver = {
+		.owner = THIS_MODULE,
+		.name = DRIVER_NAME,
+		.of_match_table = sti_drm_dt_ids,
+	},
+};
+
+module_platform_driver(sti_drm_platform_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_drm_drv.h b/drivers/gpu/drm/sti/sti_drm_drv.h
new file mode 100644
index 000000000000..ec5e2eb8dff9
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_drv.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_DRM_DRV_H_
+#define _STI_DRM_DRV_H_
+
+#include <drm/drmP.h>
+
+struct sti_compositor;
+struct sti_tvout;
+
+/**
+ * STI drm private structure
+ * This structure is stored as private in the drm_device
+ *
+ * @compo:                 compositor
+ * @plane_zorder_property: z-order property for CRTC planes
+ * @drm_dev:               drm device
+ */
+struct sti_drm_private {
+	struct sti_compositor *compo;
+	struct drm_property *plane_zorder_property;
+	struct drm_device *drm_dev;
+};
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_drm_plane.c b/drivers/gpu/drm/sti/sti_drm_plane.c
new file mode 100644
index 000000000000..f4118d4cac22
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_plane.c
@@ -0,0 +1,195 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include "sti_compositor.h"
+#include "sti_drm_drv.h"
+#include "sti_drm_plane.h"
+#include "sti_vtg.h"
+
+enum sti_layer_desc sti_layer_default_zorder[] = {
+	STI_GDP_0,
+	STI_VID_0,
+	STI_GDP_1,
+	STI_VID_1,
+	STI_GDP_2,
+	STI_GDP_3,
+};
+
+/* (Background) < GDP0 < VID0 < GDP1 < VID1 < GDP2 < GDP3 < (ForeGround) */
+
+static int
+sti_drm_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
+		     struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		     unsigned int crtc_w, unsigned int crtc_h,
+		     uint32_t src_x, uint32_t src_y,
+		     uint32_t src_w, uint32_t src_h)
+{
+	struct sti_layer *layer = to_sti_layer(plane);
+	struct sti_mixer *mixer = to_sti_mixer(crtc);
+	int res;
+
+	DRM_DEBUG_KMS("CRTC:%d (%s) drm plane:%d (%s) drm fb:%d\n",
+		      crtc->base.id, sti_mixer_to_str(mixer),
+		      plane->base.id, sti_layer_to_str(layer), fb->base.id);
+	DRM_DEBUG_KMS("(%dx%d)@(%d,%d)\n", crtc_w, crtc_h, crtc_x, crtc_y);
+
+	res = sti_mixer_set_layer_depth(mixer, layer);
+	if (res) {
+		DRM_ERROR("Can not set layer depth\n");
+		return res;
+	}
+
+	/* src_x are in 16.16 format. */
+	res = sti_layer_prepare(layer, fb, &crtc->mode, mixer->id,
+			crtc_x, crtc_y, crtc_w, crtc_h,
+			src_x >> 16, src_y >> 16,
+			src_w >> 16, src_h >> 16);
+	if (res) {
+		DRM_ERROR("Layer prepare failed\n");
+		return res;
+	}
+
+	res = sti_layer_commit(layer);
+	if (res) {
+		DRM_ERROR("Layer commit failed\n");
+		return res;
+	}
+
+	res = sti_mixer_set_layer_status(mixer, layer, true);
+	if (res) {
+		DRM_ERROR("Can not enable layer at mixer\n");
+		return res;
+	}
+
+	return 0;
+}
+
+static int sti_drm_disable_plane(struct drm_plane *plane)
+{
+	struct sti_layer *layer;
+	struct sti_mixer *mixer;
+	int lay_res, mix_res;
+
+	if (!plane->crtc) {
+		DRM_DEBUG_DRIVER("drm plane:%d not enabled\n", plane->base.id);
+		return 0;
+	}
+	layer = to_sti_layer(plane);
+	mixer = to_sti_mixer(plane->crtc);
+
+	DRM_DEBUG_DRIVER("CRTC:%d (%s) drm plane:%d (%s)\n",
+			plane->crtc->base.id, sti_mixer_to_str(mixer),
+			plane->base.id, sti_layer_to_str(layer));
+
+	/* Disable layer at mixer level */
+	mix_res = sti_mixer_set_layer_status(mixer, layer, false);
+	if (mix_res)
+		DRM_ERROR("Can not disable layer at mixer\n");
+
+	/* Wait a while to be sure that a Vsync event is received */
+	msleep(WAIT_NEXT_VSYNC_MS);
+
+	/* Then disable layer itself */
+	lay_res = sti_layer_disable(layer);
+	if (lay_res)
+		DRM_ERROR("Layer disable failed\n");
+
+	if (lay_res || mix_res)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void sti_drm_plane_destroy(struct drm_plane *plane)
+{
+	DRM_DEBUG_DRIVER("\n");
+
+	sti_drm_disable_plane(plane);
+	drm_plane_cleanup(plane);
+}
+
+static int sti_drm_plane_set_property(struct drm_plane *plane,
+				      struct drm_property *property,
+				      uint64_t val)
+{
+	struct drm_device *dev = plane->dev;
+	struct sti_drm_private *private = dev->dev_private;
+	struct sti_layer *layer = to_sti_layer(plane);
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (property == private->plane_zorder_property) {
+		layer->zorder = val;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static struct drm_plane_funcs sti_drm_plane_funcs = {
+	.update_plane = sti_drm_update_plane,
+	.disable_plane = sti_drm_disable_plane,
+	.destroy = sti_drm_plane_destroy,
+	.set_property = sti_drm_plane_set_property,
+};
+
+static void sti_drm_plane_attach_zorder_property(struct drm_plane *plane,
+						 uint64_t default_val)
+{
+	struct drm_device *dev = plane->dev;
+	struct sti_drm_private *private = dev->dev_private;
+	struct drm_property *prop;
+	struct sti_layer *layer = to_sti_layer(plane);
+
+	prop = private->plane_zorder_property;
+	if (!prop) {
+		prop = drm_property_create_range(dev, 0, "zpos", 0,
+						 GAM_MIXER_NB_DEPTH_LEVEL - 1);
+		if (!prop)
+			return;
+
+		private->plane_zorder_property = prop;
+	}
+
+	drm_object_attach_property(&plane->base, prop, default_val);
+	layer->zorder = default_val;
+}
+
+struct drm_plane *sti_drm_plane_init(struct drm_device *dev,
+				     struct sti_layer *layer,
+				     unsigned int possible_crtcs,
+				     enum drm_plane_type type)
+{
+	int err, i;
+	uint64_t default_zorder = 0;
+
+	err = drm_universal_plane_init(dev, &layer->plane, possible_crtcs,
+			     &sti_drm_plane_funcs,
+			     sti_layer_get_formats(layer),
+			     sti_layer_get_nb_formats(layer), type);
+	if (err) {
+		DRM_ERROR("Failed to initialize plane\n");
+		return NULL;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(sti_layer_default_zorder); i++)
+		if (sti_layer_default_zorder[i] == layer->desc)
+			break;
+
+	default_zorder = i;
+
+	if (type == DRM_PLANE_TYPE_OVERLAY)
+		sti_drm_plane_attach_zorder_property(&layer->plane,
+				default_zorder);
+
+	DRM_DEBUG_DRIVER("drm plane:%d mapped to %s with zorder:%llu\n",
+			 layer->plane.base.id,
+			 sti_layer_to_str(layer), default_zorder);
+
+	return &layer->plane;
+}
diff --git a/drivers/gpu/drm/sti/sti_drm_plane.h b/drivers/gpu/drm/sti/sti_drm_plane.h
new file mode 100644
index 000000000000..4f191839f2a7
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_drm_plane.h
@@ -0,0 +1,18 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_DRM_PLANE_H_
+#define _STI_DRM_PLANE_H_
+
+#include <drm/drmP.h>
+
+struct sti_layer;
+
+struct drm_plane *sti_drm_plane_init(struct drm_device *dev,
+		struct sti_layer *layer,
+		unsigned int possible_crtcs,
+		enum drm_plane_type type);
+#endif
diff --git a/drivers/gpu/drm/sti/sti_gdp.c b/drivers/gpu/drm/sti/sti_gdp.c
new file mode 100644
index 000000000000..4e30b74559f5
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_gdp.c
@@ -0,0 +1,549 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/dma-mapping.h>
+
+#include "sti_compositor.h"
+#include "sti_gdp.h"
+#include "sti_layer.h"
+#include "sti_vtg.h"
+
+#define ENA_COLOR_FILL  BIT(8)
+#define WAIT_NEXT_VSYNC BIT(31)
+
+/* GDP color formats */
+#define GDP_RGB565      0x00
+#define GDP_RGB888      0x01
+#define GDP_RGB888_32   0x02
+#define GDP_ARGB8565    0x04
+#define GDP_ARGB8888    0x05
+#define GDP_ARGB1555    0x06
+#define GDP_ARGB4444    0x07
+#define GDP_CLUT8       0x0B
+#define GDP_YCBR888     0x10
+#define GDP_YCBR422R    0x12
+#define GDP_AYCBR8888   0x15
+
+#define GAM_GDP_CTL_OFFSET      0x00
+#define GAM_GDP_AGC_OFFSET      0x04
+#define GAM_GDP_VPO_OFFSET      0x0C
+#define GAM_GDP_VPS_OFFSET      0x10
+#define GAM_GDP_PML_OFFSET      0x14
+#define GAM_GDP_PMP_OFFSET      0x18
+#define GAM_GDP_SIZE_OFFSET     0x1C
+#define GAM_GDP_NVN_OFFSET      0x24
+#define GAM_GDP_KEY1_OFFSET     0x28
+#define GAM_GDP_KEY2_OFFSET     0x2C
+#define GAM_GDP_PPT_OFFSET      0x34
+#define GAM_GDP_CML_OFFSET      0x3C
+#define GAM_GDP_MST_OFFSET      0x68
+
+#define GAM_GDP_ALPHARANGE_255  BIT(5)
+#define GAM_GDP_AGC_FULL_RANGE  0x00808080
+#define GAM_GDP_PPT_IGNORE      (BIT(1) | BIT(0))
+#define GAM_GDP_SIZE_MAX        0x7FF
+
+#define GDP_NODE_NB_BANK	2
+#define GDP_NODE_PER_FIELD	2
+
+struct sti_gdp_node {
+	u32 gam_gdp_ctl;
+	u32 gam_gdp_agc;
+	u32 reserved1;
+	u32 gam_gdp_vpo;
+	u32 gam_gdp_vps;
+	u32 gam_gdp_pml;
+	u32 gam_gdp_pmp;
+	u32 gam_gdp_size;
+	u32 reserved2;
+	u32 gam_gdp_nvn;
+	u32 gam_gdp_key1;
+	u32 gam_gdp_key2;
+	u32 reserved3;
+	u32 gam_gdp_ppt;
+	u32 reserved4;
+	u32 gam_gdp_cml;
+};
+
+struct sti_gdp_node_list {
+	struct sti_gdp_node *top_field;
+	struct sti_gdp_node *btm_field;
+};
+
+/**
+ * STI GDP structure
+ *
+ * @layer:		layer structure
+ * @clk_pix:            pixel clock for the current gdp
+ * @vtg_field_nb:       callback for VTG FIELD (top or bottom) notification
+ * @is_curr_top:        true if the current node processed is the top field
+ * @node_list:		array of node list
+ */
+struct sti_gdp {
+	struct sti_layer layer;
+	struct clk *clk_pix;
+	struct notifier_block vtg_field_nb;
+	bool is_curr_top;
+	struct sti_gdp_node_list node_list[GDP_NODE_NB_BANK];
+};
+
+#define to_sti_gdp(x) container_of(x, struct sti_gdp, layer)
+
+static const uint32_t gdp_supported_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGB888,
+	DRM_FORMAT_AYUV,
+	DRM_FORMAT_YUV444,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_C8,
+};
+
+static const uint32_t *sti_gdp_get_formats(struct sti_layer *layer)
+{
+	return gdp_supported_formats;
+}
+
+static unsigned int sti_gdp_get_nb_formats(struct sti_layer *layer)
+{
+	return ARRAY_SIZE(gdp_supported_formats);
+}
+
+static int sti_gdp_fourcc2format(int fourcc)
+{
+	switch (fourcc) {
+	case DRM_FORMAT_XRGB8888:
+		return GDP_RGB888_32;
+	case DRM_FORMAT_ARGB8888:
+		return GDP_ARGB8888;
+	case DRM_FORMAT_ARGB4444:
+		return GDP_ARGB4444;
+	case DRM_FORMAT_ARGB1555:
+		return GDP_ARGB1555;
+	case DRM_FORMAT_RGB565:
+		return GDP_RGB565;
+	case DRM_FORMAT_RGB888:
+		return GDP_RGB888;
+	case DRM_FORMAT_AYUV:
+		return GDP_AYCBR8888;
+	case DRM_FORMAT_YUV444:
+		return GDP_YCBR888;
+	case DRM_FORMAT_VYUY:
+		return GDP_YCBR422R;
+	case DRM_FORMAT_C8:
+		return GDP_CLUT8;
+	}
+	return -1;
+}
+
+static int sti_gdp_get_alpharange(int format)
+{
+	switch (format) {
+	case GDP_ARGB8565:
+	case GDP_ARGB8888:
+	case GDP_AYCBR8888:
+		return GAM_GDP_ALPHARANGE_255;
+	}
+	return 0;
+}
+
+/**
+ * sti_gdp_get_free_nodes
+ * @layer: gdp layer
+ *
+ * Look for a GDP node list that is not currently read by the HW.
+ *
+ * RETURNS:
+ * Pointer to the free GDP node list
+ */
+static struct sti_gdp_node_list *sti_gdp_get_free_nodes(struct sti_layer *layer)
+{
+	int hw_nvn;
+	void *virt_nvn;
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	unsigned int i;
+
+	hw_nvn = readl(layer->regs + GAM_GDP_NVN_OFFSET);
+	if (!hw_nvn)
+		goto end;
+
+	virt_nvn = dma_to_virt(layer->dev, (dma_addr_t) hw_nvn);
+
+	for (i = 0; i < GDP_NODE_NB_BANK; i++)
+		if ((virt_nvn != gdp->node_list[i].btm_field) &&
+		    (virt_nvn != gdp->node_list[i].top_field))
+			return &gdp->node_list[i];
+
+	/* in hazardious cases restart with the first node */
+	DRM_ERROR("inconsistent NVN for %s: 0x%08X\n",
+			sti_layer_to_str(layer), hw_nvn);
+
+end:
+	return &gdp->node_list[0];
+}
+
+/**
+ * sti_gdp_get_current_nodes
+ * @layer: GDP layer
+ *
+ * Look for GDP nodes that are currently read by the HW.
+ *
+ * RETURNS:
+ * Pointer to the current GDP node list
+ */
+static
+struct sti_gdp_node_list *sti_gdp_get_current_nodes(struct sti_layer *layer)
+{
+	int hw_nvn;
+	void *virt_nvn;
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	unsigned int i;
+
+	hw_nvn = readl(layer->regs + GAM_GDP_NVN_OFFSET);
+	if (!hw_nvn)
+		goto end;
+
+	virt_nvn = dma_to_virt(layer->dev, (dma_addr_t) hw_nvn);
+
+	for (i = 0; i < GDP_NODE_NB_BANK; i++)
+		if ((virt_nvn == gdp->node_list[i].btm_field) ||
+				(virt_nvn == gdp->node_list[i].top_field))
+			return &gdp->node_list[i];
+
+end:
+	DRM_DEBUG_DRIVER("Warning, NVN 0x%08X for %s does not match any node\n",
+				hw_nvn, sti_layer_to_str(layer));
+
+	return NULL;
+}
+
+/**
+ * sti_gdp_prepare_layer
+ * @lay: gdp layer
+ * @first_prepare: true if it is the first time this function is called
+ *
+ * Update the free GDP node list according to the layer properties.
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+static int sti_gdp_prepare_layer(struct sti_layer *layer, bool first_prepare)
+{
+	struct sti_gdp_node_list *list;
+	struct sti_gdp_node *top_field, *btm_field;
+	struct drm_display_mode *mode = layer->mode;
+	struct device *dev = layer->dev;
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	struct sti_compositor *compo = dev_get_drvdata(dev);
+	int format;
+	unsigned int depth, bpp;
+	int rate = mode->clock * 1000;
+	int res;
+	u32 ydo, xdo, yds, xds;
+
+	list = sti_gdp_get_free_nodes(layer);
+	top_field = list->top_field;
+	btm_field = list->btm_field;
+
+	dev_dbg(dev, "%s %s top_node:0x%p btm_node:0x%p\n", __func__,
+			sti_layer_to_str(layer), top_field, btm_field);
+
+	/* Build the top field from layer params */
+	top_field->gam_gdp_agc = GAM_GDP_AGC_FULL_RANGE;
+	top_field->gam_gdp_ctl = WAIT_NEXT_VSYNC;
+	format = sti_gdp_fourcc2format(layer->format);
+	if (format == -1) {
+		DRM_ERROR("Format not supported by GDP %.4s\n",
+			  (char *)&layer->format);
+		return 1;
+	}
+	top_field->gam_gdp_ctl |= format;
+	top_field->gam_gdp_ctl |= sti_gdp_get_alpharange(format);
+	top_field->gam_gdp_ppt &= ~GAM_GDP_PPT_IGNORE;
+
+	/* pixel memory location */
+	drm_fb_get_bpp_depth(layer->format, &depth, &bpp);
+	top_field->gam_gdp_pml = (u32) layer->paddr + layer->offsets[0];
+	top_field->gam_gdp_pml += layer->src_x * (bpp >> 3);
+	top_field->gam_gdp_pml += layer->src_y * layer->pitches[0];
+
+	/* input parameters */
+	top_field->gam_gdp_pmp = layer->pitches[0];
+	top_field->gam_gdp_size =
+	    clamp_val(layer->src_h, 0, GAM_GDP_SIZE_MAX) << 16 |
+	    clamp_val(layer->src_w, 0, GAM_GDP_SIZE_MAX);
+
+	/* output parameters */
+	ydo = sti_vtg_get_line_number(*mode, layer->dst_y);
+	yds = sti_vtg_get_line_number(*mode, layer->dst_y + layer->dst_h - 1);
+	xdo = sti_vtg_get_pixel_number(*mode, layer->dst_x);
+	xds = sti_vtg_get_pixel_number(*mode, layer->dst_x + layer->dst_w - 1);
+	top_field->gam_gdp_vpo = (ydo << 16) | xdo;
+	top_field->gam_gdp_vps = (yds << 16) | xds;
+
+	/* Same content and chained together */
+	memcpy(btm_field, top_field, sizeof(*btm_field));
+	top_field->gam_gdp_nvn = virt_to_dma(dev, btm_field);
+	btm_field->gam_gdp_nvn = virt_to_dma(dev, top_field);
+
+	/* Interlaced mode */
+	if (layer->mode->flags & DRM_MODE_FLAG_INTERLACE)
+		btm_field->gam_gdp_pml = top_field->gam_gdp_pml +
+		    layer->pitches[0];
+
+	if (first_prepare) {
+		/* Register gdp callback */
+		if (sti_vtg_register_client(layer->mixer_id == STI_MIXER_MAIN ?
+				compo->vtg_main : compo->vtg_aux,
+				&gdp->vtg_field_nb, layer->mixer_id)) {
+			DRM_ERROR("Cannot register VTG notifier\n");
+			return 1;
+		}
+
+		/* Set and enable gdp clock */
+		if (gdp->clk_pix) {
+			res = clk_set_rate(gdp->clk_pix, rate);
+			if (res < 0) {
+				DRM_ERROR("Cannot set rate (%dHz) for gdp\n",
+						rate);
+				return 1;
+			}
+
+			if (clk_prepare_enable(gdp->clk_pix)) {
+				DRM_ERROR("Failed to prepare/enable gdp\n");
+				return 1;
+			}
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * sti_gdp_commit_layer
+ * @lay: gdp layer
+ *
+ * Update the NVN field of the 'right' field of the current GDP node (being
+ * used by the HW) with the address of the updated ('free') top field GDP node.
+ * - In interlaced mode the 'right' field is the bottom field as we update
+ *   frames starting from their top field
+ * - In progressive mode, we update both bottom and top fields which are
+ *   equal nodes.
+ * At the next VSYNC, the updated node list will be used by the HW.
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+static int sti_gdp_commit_layer(struct sti_layer *layer)
+{
+	struct sti_gdp_node_list *updated_list = sti_gdp_get_free_nodes(layer);
+	struct sti_gdp_node *updated_top_node = updated_list->top_field;
+	struct sti_gdp_node *updated_btm_node = updated_list->btm_field;
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	u32 dma_updated_top = virt_to_dma(layer->dev, updated_top_node);
+	u32 dma_updated_btm = virt_to_dma(layer->dev, updated_btm_node);
+	struct sti_gdp_node_list *curr_list = sti_gdp_get_current_nodes(layer);
+
+	dev_dbg(layer->dev, "%s %s top/btm_node:0x%p/0x%p\n", __func__,
+			sti_layer_to_str(layer),
+			updated_top_node, updated_btm_node);
+	dev_dbg(layer->dev, "Current NVN:0x%X\n",
+		readl(layer->regs + GAM_GDP_NVN_OFFSET));
+	dev_dbg(layer->dev, "Posted buff: %lx current buff: %x\n",
+		(unsigned long)layer->paddr,
+		readl(layer->regs + GAM_GDP_PML_OFFSET));
+
+	if (curr_list == NULL) {
+		/* First update or invalid node should directly write in the
+		 * hw register */
+		DRM_DEBUG_DRIVER("%s first update (or invalid node)",
+				sti_layer_to_str(layer));
+
+		writel(gdp->is_curr_top == true ?
+				dma_updated_btm : dma_updated_top,
+				layer->regs + GAM_GDP_NVN_OFFSET);
+		return 0;
+	}
+
+	if (layer->mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		if (gdp->is_curr_top == true) {
+			/* Do not update in the middle of the frame, but
+			 * postpone the update after the bottom field has
+			 * been displayed */
+			curr_list->btm_field->gam_gdp_nvn = dma_updated_top;
+		} else {
+			/* Direct update to avoid one frame delay */
+			writel(dma_updated_top,
+				layer->regs + GAM_GDP_NVN_OFFSET);
+		}
+	} else {
+		/* Direct update for progressive to avoid one frame delay */
+		writel(dma_updated_top, layer->regs + GAM_GDP_NVN_OFFSET);
+	}
+
+	return 0;
+}
+
+/**
+ * sti_gdp_disable_layer
+ * @lay: gdp layer
+ *
+ * Disable a GDP.
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+static int sti_gdp_disable_layer(struct sti_layer *layer)
+{
+	unsigned int i;
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	struct sti_compositor *compo = dev_get_drvdata(layer->dev);
+
+	DRM_DEBUG_DRIVER("%s\n", sti_layer_to_str(layer));
+
+	/* Set the nodes as 'to be ignored on mixer' */
+	for (i = 0; i < GDP_NODE_NB_BANK; i++) {
+		gdp->node_list[i].top_field->gam_gdp_ppt |= GAM_GDP_PPT_IGNORE;
+		gdp->node_list[i].btm_field->gam_gdp_ppt |= GAM_GDP_PPT_IGNORE;
+	}
+
+	if (sti_vtg_unregister_client(layer->mixer_id == STI_MIXER_MAIN ?
+			compo->vtg_main : compo->vtg_aux, &gdp->vtg_field_nb))
+		DRM_DEBUG_DRIVER("Warning: cannot unregister VTG notifier\n");
+
+	if (gdp->clk_pix)
+		clk_disable_unprepare(gdp->clk_pix);
+
+	return 0;
+}
+
+/**
+ * sti_gdp_field_cb
+ * @nb: notifier block
+ * @event: event message
+ * @data: private data
+ *
+ * Handle VTG top field and bottom field event.
+ *
+ * RETURNS:
+ * 0 on success.
+ */
+int sti_gdp_field_cb(struct notifier_block *nb,
+		unsigned long event, void *data)
+{
+	struct sti_gdp *gdp = container_of(nb, struct sti_gdp, vtg_field_nb);
+
+	switch (event) {
+	case VTG_TOP_FIELD_EVENT:
+		gdp->is_curr_top = true;
+		break;
+	case VTG_BOTTOM_FIELD_EVENT:
+		gdp->is_curr_top = false;
+		break;
+	default:
+		DRM_ERROR("unsupported event: %lu\n", event);
+		break;
+	}
+
+	return 0;
+}
+
+static void sti_gdp_init(struct sti_layer *layer)
+{
+	struct sti_gdp *gdp = to_sti_gdp(layer);
+	struct device_node *np = layer->dev->of_node;
+	dma_addr_t dma;
+	void *base;
+	unsigned int i, size;
+
+	/* Allocate all the nodes within a single memory page */
+	size = sizeof(struct sti_gdp_node) *
+	    GDP_NODE_PER_FIELD * GDP_NODE_NB_BANK;
+
+	base = dma_alloc_writecombine(layer->dev,
+			size, &dma, GFP_KERNEL | GFP_DMA);
+	if (!base) {
+		DRM_ERROR("Failed to allocate memory for GDP node\n");
+		return;
+	}
+	memset(base, 0, size);
+
+	for (i = 0; i < GDP_NODE_NB_BANK; i++) {
+		if (virt_to_dma(layer->dev, base) & 0xF) {
+			DRM_ERROR("Mem alignment failed\n");
+			return;
+		}
+		gdp->node_list[i].top_field = base;
+		DRM_DEBUG_DRIVER("node[%d].top_field=%p\n", i, base);
+		base += sizeof(struct sti_gdp_node);
+
+		if (virt_to_dma(layer->dev, base) & 0xF) {
+			DRM_ERROR("Mem alignment failed\n");
+			return;
+		}
+		gdp->node_list[i].btm_field = base;
+		DRM_DEBUG_DRIVER("node[%d].btm_field=%p\n", i, base);
+		base += sizeof(struct sti_gdp_node);
+	}
+
+	if (of_device_is_compatible(np, "st,stih407-compositor")) {
+		/* GDP of STiH407 chip have its own pixel clock */
+		char *clk_name;
+
+		switch (layer->desc) {
+		case STI_GDP_0:
+			clk_name = "pix_gdp1";
+			break;
+		case STI_GDP_1:
+			clk_name = "pix_gdp2";
+			break;
+		case STI_GDP_2:
+			clk_name = "pix_gdp3";
+			break;
+		case STI_GDP_3:
+			clk_name = "pix_gdp4";
+			break;
+		default:
+			DRM_ERROR("GDP id not recognized\n");
+			return;
+		}
+
+		gdp->clk_pix = devm_clk_get(layer->dev, clk_name);
+		if (IS_ERR(gdp->clk_pix))
+			DRM_ERROR("Cannot get %s clock\n", clk_name);
+	}
+}
+
+static const struct sti_layer_funcs gdp_ops = {
+	.get_formats = sti_gdp_get_formats,
+	.get_nb_formats = sti_gdp_get_nb_formats,
+	.init = sti_gdp_init,
+	.prepare = sti_gdp_prepare_layer,
+	.commit = sti_gdp_commit_layer,
+	.disable = sti_gdp_disable_layer,
+};
+
+struct sti_layer *sti_gdp_create(struct device *dev, int id)
+{
+	struct sti_gdp *gdp;
+
+	gdp = devm_kzalloc(dev, sizeof(*gdp), GFP_KERNEL);
+	if (!gdp) {
+		DRM_ERROR("Failed to allocate memory for GDP\n");
+		return NULL;
+	}
+
+	gdp->layer.ops = &gdp_ops;
+	gdp->vtg_field_nb.notifier_call = sti_gdp_field_cb;
+
+	return (struct sti_layer *)gdp;
+}
diff --git a/drivers/gpu/drm/sti/sti_gdp.h b/drivers/gpu/drm/sti/sti_gdp.h
new file mode 100644
index 000000000000..1dab68274ad3
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_gdp.h
@@ -0,0 +1,16 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_GDP_H_
+#define _STI_GDP_H_
+
+#include <linux/types.h>
+
+struct sti_layer *sti_gdp_create(struct device *dev, int id);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_hda.c b/drivers/gpu/drm/sti/sti_hda.c
new file mode 100644
index 000000000000..2ae9a9b73666
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hda.c
@@ -0,0 +1,794 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+
+/* HDformatter registers */
+#define HDA_ANA_CFG                     0x0000
+#define HDA_ANA_SCALE_CTRL_Y            0x0004
+#define HDA_ANA_SCALE_CTRL_CB           0x0008
+#define HDA_ANA_SCALE_CTRL_CR           0x000C
+#define HDA_ANA_ANC_CTRL                0x0010
+#define HDA_ANA_SRC_Y_CFG               0x0014
+#define HDA_COEFF_Y_PH1_TAP123          0x0018
+#define HDA_COEFF_Y_PH1_TAP456          0x001C
+#define HDA_COEFF_Y_PH2_TAP123          0x0020
+#define HDA_COEFF_Y_PH2_TAP456          0x0024
+#define HDA_COEFF_Y_PH3_TAP123          0x0028
+#define HDA_COEFF_Y_PH3_TAP456          0x002C
+#define HDA_COEFF_Y_PH4_TAP123          0x0030
+#define HDA_COEFF_Y_PH4_TAP456          0x0034
+#define HDA_ANA_SRC_C_CFG               0x0040
+#define HDA_COEFF_C_PH1_TAP123          0x0044
+#define HDA_COEFF_C_PH1_TAP456          0x0048
+#define HDA_COEFF_C_PH2_TAP123          0x004C
+#define HDA_COEFF_C_PH2_TAP456          0x0050
+#define HDA_COEFF_C_PH3_TAP123          0x0054
+#define HDA_COEFF_C_PH3_TAP456          0x0058
+#define HDA_COEFF_C_PH4_TAP123          0x005C
+#define HDA_COEFF_C_PH4_TAP456          0x0060
+#define HDA_SYNC_AWGI                   0x0300
+
+/* HDA_ANA_CFG */
+#define CFG_AWG_ASYNC_EN                BIT(0)
+#define CFG_AWG_ASYNC_HSYNC_MTD         BIT(1)
+#define CFG_AWG_ASYNC_VSYNC_MTD         BIT(2)
+#define CFG_AWG_SYNC_DEL                BIT(3)
+#define CFG_AWG_FLTR_MODE_SHIFT         4
+#define CFG_AWG_FLTR_MODE_MASK          (0xF << CFG_AWG_FLTR_MODE_SHIFT)
+#define CFG_AWG_FLTR_MODE_SD            (0 << CFG_AWG_FLTR_MODE_SHIFT)
+#define CFG_AWG_FLTR_MODE_ED            (1 << CFG_AWG_FLTR_MODE_SHIFT)
+#define CFG_AWG_FLTR_MODE_HD            (2 << CFG_AWG_FLTR_MODE_SHIFT)
+#define CFG_SYNC_ON_PBPR_MASK           BIT(8)
+#define CFG_PREFILTER_EN_MASK           BIT(9)
+#define CFG_PBPR_SYNC_OFF_SHIFT         16
+#define CFG_PBPR_SYNC_OFF_MASK          (0x7FF << CFG_PBPR_SYNC_OFF_SHIFT)
+#define CFG_PBPR_SYNC_OFF_VAL           0x117 /* Voltage dependent. stiH416 */
+
+/* Default scaling values */
+#define SCALE_CTRL_Y_DFLT               0x00C50256
+#define SCALE_CTRL_CB_DFLT              0x00DB0249
+#define SCALE_CTRL_CR_DFLT              0x00DB0249
+
+/* Video DACs control */
+#define VIDEO_DACS_CONTROL_MASK         0x0FFF
+#define VIDEO_DACS_CONTROL_SYSCFG2535   0x085C /* for stih416 */
+#define DAC_CFG_HD_OFF_SHIFT            5
+#define DAC_CFG_HD_OFF_MASK             (0x7 << DAC_CFG_HD_OFF_SHIFT)
+#define VIDEO_DACS_CONTROL_SYSCFG5072   0x0120 /* for stih407 */
+#define DAC_CFG_HD_HZUVW_OFF_MASK       BIT(1)
+
+
+/* Upsampler values for the alternative 2X Filter */
+#define SAMPLER_COEF_NB                 8
+#define HDA_ANA_SRC_Y_CFG_ALT_2X        0x01130000
+static u32 coef_y_alt_2x[] = {
+	0x00FE83FB, 0x1F900401, 0x00000000, 0x00000000,
+	0x00F408F9, 0x055F7C25, 0x00000000, 0x00000000
+};
+
+#define HDA_ANA_SRC_C_CFG_ALT_2X        0x01750004
+static u32 coef_c_alt_2x[] = {
+	0x001305F7, 0x05274BD0, 0x00000000, 0x00000000,
+	0x0004907C, 0x09C80B9D, 0x00000000, 0x00000000
+};
+
+/* Upsampler values for the 4X Filter */
+#define HDA_ANA_SRC_Y_CFG_4X            0x01ED0005
+#define HDA_ANA_SRC_C_CFG_4X            0x01ED0004
+static u32 coef_yc_4x[] = {
+	0x00FC827F, 0x008FE20B, 0x00F684FC, 0x050F7C24,
+	0x00F4857C, 0x0A1F402E, 0x00FA027F, 0x0E076E1D
+};
+
+/* AWG instructions for some video modes */
+#define AWG_MAX_INST                    64
+
+/* 720p@50 */
+static u32 AWGi_720p_50[] = {
+	0x00000971, 0x00000C26, 0x0000013B, 0x00000CDA,
+	0x00000104, 0x00000E7E, 0x00000E7F, 0x0000013B,
+	0x00000D8E, 0x00000104, 0x00001804, 0x00000971,
+	0x00000C26, 0x0000003B, 0x00000FB4, 0x00000FB5,
+	0x00000104, 0x00001AE8
+};
+
+#define NN_720p_50 ARRAY_SIZE(AWGi_720p_50)
+
+/* 720p@60 */
+static u32 AWGi_720p_60[] = {
+	0x00000971, 0x00000C26, 0x0000013B, 0x00000CDA,
+	0x00000104, 0x00000E7E, 0x00000E7F, 0x0000013B,
+	0x00000C44, 0x00000104, 0x00001804, 0x00000971,
+	0x00000C26, 0x0000003B, 0x00000F0F, 0x00000F10,
+	0x00000104, 0x00001AE8
+};
+
+#define NN_720p_60 ARRAY_SIZE(AWGi_720p_60)
+
+/* 1080p@30 */
+static u32 AWGi_1080p_30[] = {
+	0x00000971, 0x00000C2A, 0x0000013B, 0x00000C56,
+	0x00000104, 0x00000FDC, 0x00000FDD, 0x0000013B,
+	0x00000C2A, 0x00000104, 0x00001804, 0x00000971,
+	0x00000C2A, 0x0000003B, 0x00000EBE, 0x00000EBF,
+	0x00000EBF, 0x00000104, 0x00001A2F, 0x00001C4B,
+	0x00001C52
+};
+
+#define NN_1080p_30 ARRAY_SIZE(AWGi_1080p_30)
+
+/* 1080p@25 */
+static u32 AWGi_1080p_25[] = {
+	0x00000971, 0x00000C2A, 0x0000013B, 0x00000C56,
+	0x00000104, 0x00000FDC, 0x00000FDD, 0x0000013B,
+	0x00000DE2, 0x00000104, 0x00001804, 0x00000971,
+	0x00000C2A, 0x0000003B, 0x00000F51, 0x00000F51,
+	0x00000F52, 0x00000104, 0x00001A2F, 0x00001C4B,
+	0x00001C52
+};
+
+#define NN_1080p_25 ARRAY_SIZE(AWGi_1080p_25)
+
+/* 1080p@24 */
+static u32 AWGi_1080p_24[] = {
+	0x00000971, 0x00000C2A, 0x0000013B, 0x00000C56,
+	0x00000104, 0x00000FDC, 0x00000FDD, 0x0000013B,
+	0x00000E50, 0x00000104, 0x00001804, 0x00000971,
+	0x00000C2A, 0x0000003B, 0x00000F76, 0x00000F76,
+	0x00000F76, 0x00000104, 0x00001A2F, 0x00001C4B,
+	0x00001C52
+};
+
+#define NN_1080p_24 ARRAY_SIZE(AWGi_1080p_24)
+
+/* 720x480p@60 */
+static u32 AWGi_720x480p_60[] = {
+	0x00000904, 0x00000F18, 0x0000013B, 0x00001805,
+	0x00000904, 0x00000C3D, 0x0000003B, 0x00001A06
+};
+
+#define NN_720x480p_60 ARRAY_SIZE(AWGi_720x480p_60)
+
+/* Video mode category */
+enum sti_hda_vid_cat {
+	VID_SD,
+	VID_ED,
+	VID_HD_74M,
+	VID_HD_148M
+};
+
+struct sti_hda_video_config {
+	struct drm_display_mode mode;
+	u32 *awg_instr;
+	int nb_instr;
+	enum sti_hda_vid_cat vid_cat;
+};
+
+/* HD analog supported modes
+ * Interlaced modes may be added when supported by the whole display chain
+ */
+static const struct sti_hda_video_config hda_supported_modes[] = {
+	/* 1080p30 74.250Mhz */
+	{{DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_1080p_30, NN_1080p_30, VID_HD_74M},
+	/* 1080p30 74.176Mhz */
+	{{DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74176, 1920, 2008,
+		   2052, 2200, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_1080p_30, NN_1080p_30, VID_HD_74M},
+	/* 1080p24 74.250Mhz */
+	{{DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2558,
+		   2602, 2750, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_1080p_24, NN_1080p_24, VID_HD_74M},
+	/* 1080p24 74.176Mhz */
+	{{DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74176, 1920, 2558,
+		   2602, 2750, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_1080p_24, NN_1080p_24, VID_HD_74M},
+	/* 1080p25 74.250Mhz */
+	{{DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448,
+		   2492, 2640, 0, 1080, 1084, 1089, 1125, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_1080p_25, NN_1080p_25, VID_HD_74M},
+	/* 720p60 74.250Mhz */
+	{{DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1390,
+		   1430, 1650, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_720p_60, NN_720p_60, VID_HD_74M},
+	/* 720p60 74.176Mhz */
+	{{DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74176, 1280, 1390,
+		   1430, 1650, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_720p_60, NN_720p_60, VID_HD_74M},
+	/* 720p50 74.250Mhz */
+	{{DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1720,
+		   1760, 1980, 0, 720, 725, 730, 750, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC)},
+	 AWGi_720p_50, NN_720p_50, VID_HD_74M},
+	/* 720x480p60 27.027Mhz */
+	{{DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27027, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC)},
+	 AWGi_720x480p_60, NN_720x480p_60, VID_ED},
+	/* 720x480p60 27.000Mhz */
+	{{DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736,
+		   798, 858, 0, 480, 489, 495, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC)},
+	 AWGi_720x480p_60, NN_720x480p_60, VID_ED}
+};
+
+/**
+ * STI hd analog structure
+ *
+ * @dev: driver device
+ * @drm_dev: pointer to drm device
+ * @mode: current display mode selected
+ * @regs: HD analog register
+ * @video_dacs_ctrl: video DACS control register
+ * @enabled: true if HD analog is enabled else false
+ */
+struct sti_hda {
+	struct device dev;
+	struct drm_device *drm_dev;
+	struct drm_display_mode mode;
+	void __iomem *regs;
+	void __iomem *video_dacs_ctrl;
+	struct clk *clk_pix;
+	struct clk *clk_hddac;
+	bool enabled;
+};
+
+struct sti_hda_connector {
+	struct drm_connector drm_connector;
+	struct drm_encoder *encoder;
+	struct sti_hda *hda;
+};
+
+#define to_sti_hda_connector(x) \
+	container_of(x, struct sti_hda_connector, drm_connector)
+
+static u32 hda_read(struct sti_hda *hda, int offset)
+{
+	return readl(hda->regs + offset);
+}
+
+static void hda_write(struct sti_hda *hda, u32 val, int offset)
+{
+	writel(val, hda->regs + offset);
+}
+
+/**
+ * Search for a video mode in the supported modes table
+ *
+ * @mode: mode being searched
+ * @idx: index of the found mode
+ *
+ * Return true if mode is found
+ */
+static bool hda_get_mode_idx(struct drm_display_mode mode, int *idx)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(hda_supported_modes); i++)
+		if (drm_mode_equal(&hda_supported_modes[i].mode, &mode)) {
+			*idx = i;
+			return true;
+		}
+	return false;
+}
+
+/**
+ * Enable the HD DACS
+ *
+ * @hda: pointer to HD analog structure
+ * @enable: true if HD DACS need to be enabled, else false
+ */
+static void hda_enable_hd_dacs(struct sti_hda *hda, bool enable)
+{
+	u32 mask;
+
+	if (hda->video_dacs_ctrl) {
+		u32 val;
+
+		switch ((u32)hda->video_dacs_ctrl & VIDEO_DACS_CONTROL_MASK) {
+		case VIDEO_DACS_CONTROL_SYSCFG2535:
+			mask = DAC_CFG_HD_OFF_MASK;
+			break;
+		case VIDEO_DACS_CONTROL_SYSCFG5072:
+			mask = DAC_CFG_HD_HZUVW_OFF_MASK;
+			break;
+		default:
+			DRM_INFO("Video DACS control register not supported!");
+			return;
+		}
+
+		val = readl(hda->video_dacs_ctrl);
+		if (enable)
+			val &= ~mask;
+		else
+			val |= mask;
+
+		writel(val, hda->video_dacs_ctrl);
+	}
+}
+
+/**
+ * Configure AWG, writing instructions
+ *
+ * @hda: pointer to HD analog structure
+ * @awg_instr: pointer to AWG instructions table
+ * @nb: nb of AWG instructions
+ */
+static void sti_hda_configure_awg(struct sti_hda *hda, u32 *awg_instr, int nb)
+{
+	unsigned int i;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	for (i = 0; i < nb; i++)
+		hda_write(hda, awg_instr[i], HDA_SYNC_AWGI + i * 4);
+	for (i = nb; i < AWG_MAX_INST; i++)
+		hda_write(hda, 0, HDA_SYNC_AWGI + i * 4);
+}
+
+static void sti_hda_disable(struct drm_bridge *bridge)
+{
+	struct sti_hda *hda = bridge->driver_private;
+	u32 val;
+
+	if (!hda->enabled)
+		return;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* Disable HD DAC and AWG */
+	val = hda_read(hda, HDA_ANA_CFG);
+	val &= ~CFG_AWG_ASYNC_EN;
+	hda_write(hda, val, HDA_ANA_CFG);
+	hda_write(hda, 0, HDA_ANA_ANC_CTRL);
+
+	hda_enable_hd_dacs(hda, false);
+
+	/* Disable/unprepare hda clock */
+	clk_disable_unprepare(hda->clk_hddac);
+	clk_disable_unprepare(hda->clk_pix);
+
+	hda->enabled = false;
+}
+
+static void sti_hda_pre_enable(struct drm_bridge *bridge)
+{
+	struct sti_hda *hda = bridge->driver_private;
+	u32 val, i, mode_idx;
+	u32 src_filter_y, src_filter_c;
+	u32 *coef_y, *coef_c;
+	u32 filter_mode;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (hda->enabled)
+		return;
+
+	/* Prepare/enable clocks */
+	if (clk_prepare_enable(hda->clk_pix))
+		DRM_ERROR("Failed to prepare/enable hda_pix clk\n");
+	if (clk_prepare_enable(hda->clk_hddac))
+		DRM_ERROR("Failed to prepare/enable hda_hddac clk\n");
+
+	if (!hda_get_mode_idx(hda->mode, &mode_idx)) {
+		DRM_ERROR("Undefined mode\n");
+		return;
+	}
+
+	switch (hda_supported_modes[mode_idx].vid_cat) {
+	case VID_HD_148M:
+		DRM_ERROR("Beyond HD analog capabilities\n");
+		return;
+	case VID_HD_74M:
+		/* HD use alternate 2x filter */
+		filter_mode = CFG_AWG_FLTR_MODE_HD;
+		src_filter_y = HDA_ANA_SRC_Y_CFG_ALT_2X;
+		src_filter_c = HDA_ANA_SRC_C_CFG_ALT_2X;
+		coef_y = coef_y_alt_2x;
+		coef_c = coef_c_alt_2x;
+		break;
+	case VID_ED:
+		/* ED uses 4x filter */
+		filter_mode = CFG_AWG_FLTR_MODE_ED;
+		src_filter_y = HDA_ANA_SRC_Y_CFG_4X;
+		src_filter_c = HDA_ANA_SRC_C_CFG_4X;
+		coef_y = coef_yc_4x;
+		coef_c = coef_yc_4x;
+		break;
+	case VID_SD:
+		DRM_ERROR("Not supported\n");
+		return;
+	default:
+		DRM_ERROR("Undefined resolution\n");
+		return;
+	}
+	DRM_DEBUG_DRIVER("Using HDA mode #%d\n", mode_idx);
+
+	/* Enable HD Video DACs */
+	hda_enable_hd_dacs(hda, true);
+
+	/* Configure scaler */
+	hda_write(hda, SCALE_CTRL_Y_DFLT, HDA_ANA_SCALE_CTRL_Y);
+	hda_write(hda, SCALE_CTRL_CB_DFLT, HDA_ANA_SCALE_CTRL_CB);
+	hda_write(hda, SCALE_CTRL_CR_DFLT, HDA_ANA_SCALE_CTRL_CR);
+
+	/* Configure sampler */
+	hda_write(hda , src_filter_y, HDA_ANA_SRC_Y_CFG);
+	hda_write(hda, src_filter_c,  HDA_ANA_SRC_C_CFG);
+	for (i = 0; i < SAMPLER_COEF_NB; i++) {
+		hda_write(hda, coef_y[i], HDA_COEFF_Y_PH1_TAP123 + i * 4);
+		hda_write(hda, coef_c[i], HDA_COEFF_C_PH1_TAP123 + i * 4);
+	}
+
+	/* Configure main HDFormatter */
+	val = 0;
+	val |= (hda->mode.flags & DRM_MODE_FLAG_INTERLACE) ?
+	    0 : CFG_AWG_ASYNC_VSYNC_MTD;
+	val |= (CFG_PBPR_SYNC_OFF_VAL << CFG_PBPR_SYNC_OFF_SHIFT);
+	val |= filter_mode;
+	hda_write(hda, val, HDA_ANA_CFG);
+
+	/* Configure AWG */
+	sti_hda_configure_awg(hda, hda_supported_modes[mode_idx].awg_instr,
+			      hda_supported_modes[mode_idx].nb_instr);
+
+	/* Enable AWG */
+	val = hda_read(hda, HDA_ANA_CFG);
+	val |= CFG_AWG_ASYNC_EN;
+	hda_write(hda, val, HDA_ANA_CFG);
+
+	hda->enabled = true;
+}
+
+static void sti_hda_set_mode(struct drm_bridge *bridge,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	struct sti_hda *hda = bridge->driver_private;
+	u32 mode_idx;
+	int hddac_rate;
+	int ret;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	memcpy(&hda->mode, mode, sizeof(struct drm_display_mode));
+
+	if (!hda_get_mode_idx(hda->mode, &mode_idx)) {
+		DRM_ERROR("Undefined mode\n");
+		return;
+	}
+
+	switch (hda_supported_modes[mode_idx].vid_cat) {
+	case VID_HD_74M:
+		/* HD use alternate 2x filter */
+		hddac_rate = mode->clock * 1000 * 2;
+		break;
+	case VID_ED:
+		/* ED uses 4x filter */
+		hddac_rate = mode->clock * 1000 * 4;
+		break;
+	default:
+		DRM_ERROR("Undefined mode\n");
+		return;
+	}
+
+	/* HD DAC = 148.5Mhz or 108 Mhz */
+	ret = clk_set_rate(hda->clk_hddac, hddac_rate);
+	if (ret < 0)
+		DRM_ERROR("Cannot set rate (%dHz) for hda_hddac clk\n",
+			  hddac_rate);
+
+	/* HDformatter clock = compositor clock */
+	ret = clk_set_rate(hda->clk_pix, mode->clock * 1000);
+	if (ret < 0)
+		DRM_ERROR("Cannot set rate (%dHz) for hda_pix clk\n",
+			  mode->clock * 1000);
+}
+
+static void sti_hda_bridge_nope(struct drm_bridge *bridge)
+{
+	/* do nothing */
+}
+
+static void sti_hda_brigde_destroy(struct drm_bridge *bridge)
+{
+	drm_bridge_cleanup(bridge);
+	kfree(bridge);
+}
+
+static const struct drm_bridge_funcs sti_hda_bridge_funcs = {
+	.pre_enable = sti_hda_pre_enable,
+	.enable = sti_hda_bridge_nope,
+	.disable = sti_hda_disable,
+	.post_disable = sti_hda_bridge_nope,
+	.mode_set = sti_hda_set_mode,
+	.destroy = sti_hda_brigde_destroy,
+};
+
+static int sti_hda_connector_get_modes(struct drm_connector *connector)
+{
+	unsigned int i;
+	int count = 0;
+	struct sti_hda_connector *hda_connector
+		= to_sti_hda_connector(connector);
+	struct sti_hda *hda = hda_connector->hda;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	for (i = 0; i < ARRAY_SIZE(hda_supported_modes); i++) {
+		struct drm_display_mode *mode =
+			drm_mode_duplicate(hda->drm_dev,
+					&hda_supported_modes[i].mode);
+		if (!mode)
+			continue;
+		mode->vrefresh = drm_mode_vrefresh(mode);
+
+		/* the first mode is the preferred mode */
+		if (i == 0)
+			mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+		drm_mode_probed_add(connector, mode);
+		count++;
+	}
+
+	drm_mode_sort(&connector->modes);
+
+	return count;
+}
+
+#define CLK_TOLERANCE_HZ 50
+
+static int sti_hda_connector_mode_valid(struct drm_connector *connector,
+					struct drm_display_mode *mode)
+{
+	int target = mode->clock * 1000;
+	int target_min = target - CLK_TOLERANCE_HZ;
+	int target_max = target + CLK_TOLERANCE_HZ;
+	int result;
+	int idx;
+	struct sti_hda_connector *hda_connector
+		= to_sti_hda_connector(connector);
+	struct sti_hda *hda = hda_connector->hda;
+
+	if (!hda_get_mode_idx(*mode, &idx)) {
+		return MODE_BAD;
+	} else {
+		result = clk_round_rate(hda->clk_pix, target);
+
+		DRM_DEBUG_DRIVER("target rate = %d => available rate = %d\n",
+				 target, result);
+
+		if ((result < target_min) || (result > target_max)) {
+			DRM_DEBUG_DRIVER("hda pixclk=%d not supported\n",
+					 target);
+			return MODE_BAD;
+		}
+	}
+
+	return MODE_OK;
+}
+
+struct drm_encoder *sti_hda_best_encoder(struct drm_connector *connector)
+{
+	struct sti_hda_connector *hda_connector
+		= to_sti_hda_connector(connector);
+
+	/* Best encoder is the one associated during connector creation */
+	return hda_connector->encoder;
+}
+
+static struct drm_connector_helper_funcs sti_hda_connector_helper_funcs = {
+	.get_modes = sti_hda_connector_get_modes,
+	.mode_valid = sti_hda_connector_mode_valid,
+	.best_encoder = sti_hda_best_encoder,
+};
+
+static enum drm_connector_status
+sti_hda_connector_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static void sti_hda_connector_destroy(struct drm_connector *connector)
+{
+	struct sti_hda_connector *hda_connector
+		= to_sti_hda_connector(connector);
+
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+	kfree(hda_connector);
+}
+
+static struct drm_connector_funcs sti_hda_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = sti_hda_connector_detect,
+	.destroy = sti_hda_connector_destroy,
+};
+
+static struct drm_encoder *sti_hda_find_encoder(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (encoder->encoder_type == DRM_MODE_ENCODER_DAC)
+			return encoder;
+	}
+
+	return NULL;
+}
+
+static int sti_hda_bind(struct device *dev, struct device *master, void *data)
+{
+	struct sti_hda *hda = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
+	struct drm_encoder *encoder;
+	struct sti_hda_connector *connector;
+	struct drm_connector *drm_connector;
+	struct drm_bridge *bridge;
+	int err;
+
+	/* Set the drm device handle */
+	hda->drm_dev = drm_dev;
+
+	encoder = sti_hda_find_encoder(drm_dev);
+	if (!encoder)
+		return -ENOMEM;
+
+	connector = devm_kzalloc(dev, sizeof(*connector), GFP_KERNEL);
+	if (!connector)
+		return -ENOMEM;
+
+	connector->hda = hda;
+
+		bridge = devm_kzalloc(dev, sizeof(*bridge), GFP_KERNEL);
+	if (!bridge)
+		return -ENOMEM;
+
+	bridge->driver_private = hda;
+	drm_bridge_init(drm_dev, bridge, &sti_hda_bridge_funcs);
+
+	encoder->bridge = bridge;
+	connector->encoder = encoder;
+
+	drm_connector = (struct drm_connector *)connector;
+
+	drm_connector->polled = DRM_CONNECTOR_POLL_HPD;
+
+	drm_connector_init(drm_dev, drm_connector,
+			&sti_hda_connector_funcs, DRM_MODE_CONNECTOR_Component);
+	drm_connector_helper_add(drm_connector,
+			&sti_hda_connector_helper_funcs);
+
+	err = drm_connector_register(drm_connector);
+	if (err)
+		goto err_connector;
+
+	err = drm_mode_connector_attach_encoder(drm_connector, encoder);
+	if (err) {
+		DRM_ERROR("Failed to attach a connector to a encoder\n");
+		goto err_sysfs;
+	}
+
+	return 0;
+
+err_sysfs:
+	drm_connector_unregister(drm_connector);
+err_connector:
+	drm_bridge_cleanup(bridge);
+	drm_connector_cleanup(drm_connector);
+	return -EINVAL;
+}
+
+static void sti_hda_unbind(struct device *dev,
+		struct device *master, void *data)
+{
+	/* do nothing */
+}
+
+static const struct component_ops sti_hda_ops = {
+	.bind = sti_hda_bind,
+	.unbind = sti_hda_unbind,
+};
+
+static int sti_hda_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct sti_hda *hda;
+	struct resource *res;
+
+	DRM_INFO("%s\n", __func__);
+
+	hda = devm_kzalloc(dev, sizeof(*hda), GFP_KERNEL);
+	if (!hda)
+		return -ENOMEM;
+
+	hda->dev = pdev->dev;
+
+	/* Get resources */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "hda-reg");
+	if (!res) {
+		DRM_ERROR("Invalid hda resource\n");
+		return -ENOMEM;
+	}
+	hda->regs = devm_ioremap_nocache(dev, res->start, resource_size(res));
+	if (!hda->regs)
+		return -ENOMEM;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+			"video-dacs-ctrl");
+	if (res) {
+		hda->video_dacs_ctrl = devm_ioremap_nocache(dev, res->start,
+				resource_size(res));
+		if (!hda->video_dacs_ctrl)
+			return -ENOMEM;
+	} else {
+		/* If no existing video-dacs-ctrl resource continue the probe */
+		DRM_DEBUG_DRIVER("No video-dacs-ctrl resource\n");
+		hda->video_dacs_ctrl = NULL;
+	}
+
+	/* Get clock resources */
+	hda->clk_pix = devm_clk_get(dev, "pix");
+	if (IS_ERR(hda->clk_pix)) {
+		DRM_ERROR("Cannot get hda_pix clock\n");
+		return PTR_ERR(hda->clk_pix);
+	}
+
+	hda->clk_hddac = devm_clk_get(dev, "hddac");
+	if (IS_ERR(hda->clk_hddac)) {
+		DRM_ERROR("Cannot get hda_hddac clock\n");
+		return PTR_ERR(hda->clk_hddac);
+	}
+
+	platform_set_drvdata(pdev, hda);
+
+	return component_add(&pdev->dev, &sti_hda_ops);
+}
+
+static int sti_hda_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &sti_hda_ops);
+	return 0;
+}
+
+static const struct of_device_id hda_of_match[] = {
+	{ .compatible = "st,stih416-hda", },
+	{ .compatible = "st,stih407-hda", },
+	{ /* end node */ }
+};
+MODULE_DEVICE_TABLE(of, hda_of_match);
+
+struct platform_driver sti_hda_driver = {
+	.driver = {
+		.name = "sti-hda",
+		.owner = THIS_MODULE,
+		.of_match_table = hda_of_match,
+	},
+	.probe = sti_hda_probe,
+	.remove = sti_hda_remove,
+};
+
+module_platform_driver(sti_hda_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_hdmi.c b/drivers/gpu/drm/sti/sti_hdmi.c
new file mode 100644
index 000000000000..b22968c08d1f
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi.c
@@ -0,0 +1,809 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Vincent Abriou <vincent.abriou@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/hdmi.h>
+#include <linux/module.h>
+#include <linux/of_gpio.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_edid.h>
+
+#include "sti_hdmi.h"
+#include "sti_hdmi_tx3g4c28phy.h"
+#include "sti_hdmi_tx3g0c55phy.h"
+#include "sti_vtg.h"
+
+#define HDMI_CFG                        0x0000
+#define HDMI_INT_EN                     0x0004
+#define HDMI_INT_STA                    0x0008
+#define HDMI_INT_CLR                    0x000C
+#define HDMI_STA                        0x0010
+#define HDMI_ACTIVE_VID_XMIN            0x0100
+#define HDMI_ACTIVE_VID_XMAX            0x0104
+#define HDMI_ACTIVE_VID_YMIN            0x0108
+#define HDMI_ACTIVE_VID_YMAX            0x010C
+#define HDMI_DFLT_CHL0_DAT              0x0110
+#define HDMI_DFLT_CHL1_DAT              0x0114
+#define HDMI_DFLT_CHL2_DAT              0x0118
+#define HDMI_SW_DI_1_HEAD_WORD          0x0210
+#define HDMI_SW_DI_1_PKT_WORD0          0x0214
+#define HDMI_SW_DI_1_PKT_WORD1          0x0218
+#define HDMI_SW_DI_1_PKT_WORD2          0x021C
+#define HDMI_SW_DI_1_PKT_WORD3          0x0220
+#define HDMI_SW_DI_1_PKT_WORD4          0x0224
+#define HDMI_SW_DI_1_PKT_WORD5          0x0228
+#define HDMI_SW_DI_1_PKT_WORD6          0x022C
+#define HDMI_SW_DI_CFG                  0x0230
+
+#define HDMI_IFRAME_SLOT_AVI            1
+
+#define  XCAT(prefix, x, suffix)        prefix ## x ## suffix
+#define  HDMI_SW_DI_N_HEAD_WORD(x)      XCAT(HDMI_SW_DI_, x, _HEAD_WORD)
+#define  HDMI_SW_DI_N_PKT_WORD0(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD0)
+#define  HDMI_SW_DI_N_PKT_WORD1(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD1)
+#define  HDMI_SW_DI_N_PKT_WORD2(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD2)
+#define  HDMI_SW_DI_N_PKT_WORD3(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD3)
+#define  HDMI_SW_DI_N_PKT_WORD4(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD4)
+#define  HDMI_SW_DI_N_PKT_WORD5(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD5)
+#define  HDMI_SW_DI_N_PKT_WORD6(x)      XCAT(HDMI_SW_DI_, x, _PKT_WORD6)
+
+#define HDMI_IFRAME_DISABLED            0x0
+#define HDMI_IFRAME_SINGLE_SHOT         0x1
+#define HDMI_IFRAME_FIELD               0x2
+#define HDMI_IFRAME_FRAME               0x3
+#define HDMI_IFRAME_MASK                0x3
+#define HDMI_IFRAME_CFG_DI_N(x, n)       ((x) << ((n-1)*4)) /* n from 1 to 6 */
+
+#define HDMI_CFG_DEVICE_EN              BIT(0)
+#define HDMI_CFG_HDMI_NOT_DVI           BIT(1)
+#define HDMI_CFG_HDCP_EN                BIT(2)
+#define HDMI_CFG_ESS_NOT_OESS           BIT(3)
+#define HDMI_CFG_H_SYNC_POL_NEG         BIT(4)
+#define HDMI_CFG_SINK_TERM_DET_EN       BIT(5)
+#define HDMI_CFG_V_SYNC_POL_NEG         BIT(6)
+#define HDMI_CFG_422_EN                 BIT(8)
+#define HDMI_CFG_FIFO_OVERRUN_CLR       BIT(12)
+#define HDMI_CFG_FIFO_UNDERRUN_CLR      BIT(13)
+#define HDMI_CFG_SW_RST_EN              BIT(31)
+
+#define HDMI_INT_GLOBAL                 BIT(0)
+#define HDMI_INT_SW_RST                 BIT(1)
+#define HDMI_INT_PIX_CAP                BIT(3)
+#define HDMI_INT_HOT_PLUG               BIT(4)
+#define HDMI_INT_DLL_LCK                BIT(5)
+#define HDMI_INT_NEW_FRAME              BIT(6)
+#define HDMI_INT_GENCTRL_PKT            BIT(7)
+#define HDMI_INT_SINK_TERM_PRESENT      BIT(11)
+
+#define HDMI_DEFAULT_INT (HDMI_INT_SINK_TERM_PRESENT \
+			| HDMI_INT_DLL_LCK \
+			| HDMI_INT_HOT_PLUG \
+			| HDMI_INT_GLOBAL)
+
+#define HDMI_WORKING_INT (HDMI_INT_SINK_TERM_PRESENT \
+			| HDMI_INT_GENCTRL_PKT \
+			| HDMI_INT_NEW_FRAME \
+			| HDMI_INT_DLL_LCK \
+			| HDMI_INT_HOT_PLUG \
+			| HDMI_INT_PIX_CAP \
+			| HDMI_INT_SW_RST \
+			| HDMI_INT_GLOBAL)
+
+#define HDMI_STA_SW_RST                 BIT(1)
+
+struct sti_hdmi_connector {
+	struct drm_connector drm_connector;
+	struct drm_encoder *encoder;
+	struct sti_hdmi *hdmi;
+};
+
+#define to_sti_hdmi_connector(x) \
+	container_of(x, struct sti_hdmi_connector, drm_connector)
+
+u32 hdmi_read(struct sti_hdmi *hdmi, int offset)
+{
+	return readl(hdmi->regs + offset);
+}
+
+void hdmi_write(struct sti_hdmi *hdmi, u32 val, int offset)
+{
+	writel(val, hdmi->regs + offset);
+}
+
+/**
+ * HDMI interrupt handler threaded
+ *
+ * @irq: irq number
+ * @arg: connector structure
+ */
+static irqreturn_t hdmi_irq_thread(int irq, void *arg)
+{
+	struct sti_hdmi *hdmi = arg;
+
+	/* Hot plug/unplug IRQ */
+	if (hdmi->irq_status & HDMI_INT_HOT_PLUG) {
+		/* read gpio to get the status */
+		hdmi->hpd = gpio_get_value(hdmi->hpd_gpio);
+		if (hdmi->drm_dev)
+			drm_helper_hpd_irq_event(hdmi->drm_dev);
+	}
+
+	/* Sw reset and PLL lock are exclusive so we can use the same
+	 * event to signal them
+	 */
+	if (hdmi->irq_status & (HDMI_INT_SW_RST | HDMI_INT_DLL_LCK)) {
+		hdmi->event_received = true;
+		wake_up_interruptible(&hdmi->wait_event);
+	}
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * HDMI interrupt handler
+ *
+ * @irq: irq number
+ * @arg: connector structure
+ */
+static irqreturn_t hdmi_irq(int irq, void *arg)
+{
+	struct sti_hdmi *hdmi = arg;
+
+	/* read interrupt status */
+	hdmi->irq_status = hdmi_read(hdmi, HDMI_INT_STA);
+
+	/* clear interrupt status */
+	hdmi_write(hdmi, hdmi->irq_status, HDMI_INT_CLR);
+
+	/* force sync bus write */
+	hdmi_read(hdmi, HDMI_INT_STA);
+
+	return IRQ_WAKE_THREAD;
+}
+
+/**
+ * Set hdmi active area depending on the drm display mode selected
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ */
+static void hdmi_active_area(struct sti_hdmi *hdmi)
+{
+	u32 xmin, xmax;
+	u32 ymin, ymax;
+
+	xmin = sti_vtg_get_pixel_number(hdmi->mode, 0);
+	xmax = sti_vtg_get_pixel_number(hdmi->mode, hdmi->mode.hdisplay - 1);
+	ymin = sti_vtg_get_line_number(hdmi->mode, 0);
+	ymax = sti_vtg_get_line_number(hdmi->mode, hdmi->mode.vdisplay - 1);
+
+	hdmi_write(hdmi, xmin, HDMI_ACTIVE_VID_XMIN);
+	hdmi_write(hdmi, xmax, HDMI_ACTIVE_VID_XMAX);
+	hdmi_write(hdmi, ymin, HDMI_ACTIVE_VID_YMIN);
+	hdmi_write(hdmi, ymax, HDMI_ACTIVE_VID_YMAX);
+}
+
+/**
+ * Overall hdmi configuration
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ */
+static void hdmi_config(struct sti_hdmi *hdmi)
+{
+	u32 conf;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* Clear overrun and underrun fifo */
+	conf = HDMI_CFG_FIFO_OVERRUN_CLR | HDMI_CFG_FIFO_UNDERRUN_CLR;
+
+	/* Enable HDMI mode not DVI */
+	conf |= HDMI_CFG_HDMI_NOT_DVI | HDMI_CFG_ESS_NOT_OESS;
+
+	/* Enable sink term detection */
+	conf |= HDMI_CFG_SINK_TERM_DET_EN;
+
+	/* Set Hsync polarity */
+	if (hdmi->mode.flags & DRM_MODE_FLAG_NHSYNC) {
+		DRM_DEBUG_DRIVER("H Sync Negative\n");
+		conf |= HDMI_CFG_H_SYNC_POL_NEG;
+	}
+
+	/* Set Vsync polarity */
+	if (hdmi->mode.flags & DRM_MODE_FLAG_NVSYNC) {
+		DRM_DEBUG_DRIVER("V Sync Negative\n");
+		conf |= HDMI_CFG_V_SYNC_POL_NEG;
+	}
+
+	/* Enable HDMI */
+	conf |= HDMI_CFG_DEVICE_EN;
+
+	hdmi_write(hdmi, conf, HDMI_CFG);
+}
+
+/**
+ * Prepare and configure the AVI infoframe
+ *
+ * AVI infoframe are transmitted at least once per two video field and
+ * contains information about HDMI transmission mode such as color space,
+ * colorimetry, ...
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ * Return negative value if error occurs
+ */
+static int hdmi_avi_infoframe_config(struct sti_hdmi *hdmi)
+{
+	struct drm_display_mode *mode = &hdmi->mode;
+	struct hdmi_avi_infoframe infoframe;
+	u8 buffer[HDMI_INFOFRAME_SIZE(AVI)];
+	u8 *frame = buffer + HDMI_INFOFRAME_HEADER_SIZE;
+	u32 val;
+	int ret;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	ret = drm_hdmi_avi_infoframe_from_display_mode(&infoframe, mode);
+	if (ret < 0) {
+		DRM_ERROR("failed to setup AVI infoframe: %d\n", ret);
+		return ret;
+	}
+
+	/* fixed infoframe configuration not linked to the mode */
+	infoframe.colorspace = HDMI_COLORSPACE_RGB;
+	infoframe.quantization_range = HDMI_QUANTIZATION_RANGE_DEFAULT;
+	infoframe.colorimetry = HDMI_COLORIMETRY_NONE;
+
+	ret = hdmi_avi_infoframe_pack(&infoframe, buffer, sizeof(buffer));
+	if (ret < 0) {
+		DRM_ERROR("failed to pack AVI infoframe: %d\n", ret);
+		return ret;
+	}
+
+	/* Disable transmission slot for AVI infoframe */
+	val = hdmi_read(hdmi, HDMI_SW_DI_CFG);
+	val &= ~HDMI_IFRAME_CFG_DI_N(HDMI_IFRAME_MASK, HDMI_IFRAME_SLOT_AVI);
+	hdmi_write(hdmi, val, HDMI_SW_DI_CFG);
+
+	/* Infoframe header */
+	val = buffer[0x0];
+	val |= buffer[0x1] << 8;
+	val |= buffer[0x2] << 16;
+	hdmi_write(hdmi, val, HDMI_SW_DI_N_HEAD_WORD(HDMI_IFRAME_SLOT_AVI));
+
+	/* Infoframe packet bytes */
+	val = frame[0x0];
+	val |= frame[0x1] << 8;
+	val |= frame[0x2] << 16;
+	val |= frame[0x3] << 24;
+	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD0(HDMI_IFRAME_SLOT_AVI));
+
+	val = frame[0x4];
+	val |= frame[0x5] << 8;
+	val |= frame[0x6] << 16;
+	val |= frame[0x7] << 24;
+	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD1(HDMI_IFRAME_SLOT_AVI));
+
+	val = frame[0x8];
+	val |= frame[0x9] << 8;
+	val |= frame[0xA] << 16;
+	val |= frame[0xB] << 24;
+	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD2(HDMI_IFRAME_SLOT_AVI));
+
+	val = frame[0xC];
+	hdmi_write(hdmi, val, HDMI_SW_DI_N_PKT_WORD3(HDMI_IFRAME_SLOT_AVI));
+
+	/* Enable transmission slot for AVI infoframe
+	 * According to the hdmi specification, AVI infoframe should be
+	 * transmitted at least once per two video fields
+	 */
+	val = hdmi_read(hdmi, HDMI_SW_DI_CFG);
+	val |= HDMI_IFRAME_CFG_DI_N(HDMI_IFRAME_FIELD, HDMI_IFRAME_SLOT_AVI);
+	hdmi_write(hdmi, val, HDMI_SW_DI_CFG);
+
+	return 0;
+}
+
+/**
+ * Software reset of the hdmi subsystem
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ */
+#define HDMI_TIMEOUT_SWRESET  100   /*milliseconds */
+static void hdmi_swreset(struct sti_hdmi *hdmi)
+{
+	u32 val;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* Enable hdmi_audio clock only during hdmi reset */
+	if (clk_prepare_enable(hdmi->clk_audio))
+		DRM_INFO("Failed to prepare/enable hdmi_audio clk\n");
+
+	/* Sw reset */
+	hdmi->event_received = false;
+
+	val = hdmi_read(hdmi, HDMI_CFG);
+	val |= HDMI_CFG_SW_RST_EN;
+	hdmi_write(hdmi, val, HDMI_CFG);
+
+	/* Wait reset completed */
+	wait_event_interruptible_timeout(hdmi->wait_event,
+					 hdmi->event_received == true,
+					 msecs_to_jiffies
+					 (HDMI_TIMEOUT_SWRESET));
+
+	/*
+	 * HDMI_STA_SW_RST bit is set to '1' when SW_RST bit in HDMI_CFG is
+	 * set to '1' and clk_audio is running.
+	 */
+	if ((hdmi_read(hdmi, HDMI_STA) & HDMI_STA_SW_RST) == 0)
+		DRM_DEBUG_DRIVER("Warning: HDMI sw reset timeout occurs\n");
+
+	val = hdmi_read(hdmi, HDMI_CFG);
+	val &= ~HDMI_CFG_SW_RST_EN;
+	hdmi_write(hdmi, val, HDMI_CFG);
+
+	/* Disable hdmi_audio clock. Not used anymore for drm purpose */
+	clk_disable_unprepare(hdmi->clk_audio);
+}
+
+static void sti_hdmi_disable(struct drm_bridge *bridge)
+{
+	struct sti_hdmi *hdmi = bridge->driver_private;
+
+	u32 val = hdmi_read(hdmi, HDMI_CFG);
+
+	if (!hdmi->enabled)
+		return;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* Disable HDMI */
+	val &= ~HDMI_CFG_DEVICE_EN;
+	hdmi_write(hdmi, val, HDMI_CFG);
+
+	hdmi_write(hdmi, 0xffffffff, HDMI_INT_CLR);
+
+	/* Stop the phy */
+	hdmi->phy_ops->stop(hdmi);
+
+	/* Set the default channel data to be a dark red */
+	hdmi_write(hdmi, 0x0000, HDMI_DFLT_CHL0_DAT);
+	hdmi_write(hdmi, 0x0000, HDMI_DFLT_CHL1_DAT);
+	hdmi_write(hdmi, 0x0060, HDMI_DFLT_CHL2_DAT);
+
+	/* Disable/unprepare hdmi clock */
+	clk_disable_unprepare(hdmi->clk_phy);
+	clk_disable_unprepare(hdmi->clk_tmds);
+	clk_disable_unprepare(hdmi->clk_pix);
+
+	hdmi->enabled = false;
+}
+
+static void sti_hdmi_pre_enable(struct drm_bridge *bridge)
+{
+	struct sti_hdmi *hdmi = bridge->driver_private;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (hdmi->enabled)
+		return;
+
+	/* Prepare/enable clocks */
+	if (clk_prepare_enable(hdmi->clk_pix))
+		DRM_ERROR("Failed to prepare/enable hdmi_pix clk\n");
+	if (clk_prepare_enable(hdmi->clk_tmds))
+		DRM_ERROR("Failed to prepare/enable hdmi_tmds clk\n");
+	if (clk_prepare_enable(hdmi->clk_phy))
+		DRM_ERROR("Failed to prepare/enable hdmi_rejec_pll clk\n");
+
+	hdmi->enabled = true;
+
+	/* Program hdmi serializer and start phy */
+	if (!hdmi->phy_ops->start(hdmi)) {
+		DRM_ERROR("Unable to start hdmi phy\n");
+		return;
+	}
+
+	/* Program hdmi active area */
+	hdmi_active_area(hdmi);
+
+	/* Enable working interrupts */
+	hdmi_write(hdmi, HDMI_WORKING_INT, HDMI_INT_EN);
+
+	/* Program hdmi config */
+	hdmi_config(hdmi);
+
+	/* Program AVI infoframe */
+	if (hdmi_avi_infoframe_config(hdmi))
+		DRM_ERROR("Unable to configure AVI infoframe\n");
+
+	/* Sw reset */
+	hdmi_swreset(hdmi);
+}
+
+static void sti_hdmi_set_mode(struct drm_bridge *bridge,
+		struct drm_display_mode *mode,
+		struct drm_display_mode *adjusted_mode)
+{
+	struct sti_hdmi *hdmi = bridge->driver_private;
+	int ret;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	/* Copy the drm display mode in the connector local structure */
+	memcpy(&hdmi->mode, mode, sizeof(struct drm_display_mode));
+
+	/* Update clock framerate according to the selected mode */
+	ret = clk_set_rate(hdmi->clk_pix, mode->clock * 1000);
+	if (ret < 0) {
+		DRM_ERROR("Cannot set rate (%dHz) for hdmi_pix clk\n",
+			  mode->clock * 1000);
+		return;
+	}
+	ret = clk_set_rate(hdmi->clk_phy, mode->clock * 1000);
+	if (ret < 0) {
+		DRM_ERROR("Cannot set rate (%dHz) for hdmi_rejection_pll clk\n",
+			  mode->clock * 1000);
+		return;
+	}
+}
+
+static void sti_hdmi_bridge_nope(struct drm_bridge *bridge)
+{
+	/* do nothing */
+}
+
+static void sti_hdmi_brigde_destroy(struct drm_bridge *bridge)
+{
+	drm_bridge_cleanup(bridge);
+	kfree(bridge);
+}
+
+static const struct drm_bridge_funcs sti_hdmi_bridge_funcs = {
+	.pre_enable = sti_hdmi_pre_enable,
+	.enable = sti_hdmi_bridge_nope,
+	.disable = sti_hdmi_disable,
+	.post_disable = sti_hdmi_bridge_nope,
+	.mode_set = sti_hdmi_set_mode,
+	.destroy = sti_hdmi_brigde_destroy,
+};
+
+static int sti_hdmi_connector_get_modes(struct drm_connector *connector)
+{
+	struct i2c_adapter *i2c_adap;
+	struct edid *edid;
+	int count;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	i2c_adap = i2c_get_adapter(1);
+	if (!i2c_adap)
+		goto fail;
+
+	edid = drm_get_edid(connector, i2c_adap);
+	if (!edid)
+		goto fail;
+
+	count = drm_add_edid_modes(connector, edid);
+	drm_mode_connector_update_edid_property(connector, edid);
+
+	kfree(edid);
+	return count;
+
+fail:
+	DRM_ERROR("Can not read HDMI EDID\n");
+	return 0;
+}
+
+#define CLK_TOLERANCE_HZ 50
+
+static int sti_hdmi_connector_mode_valid(struct drm_connector *connector,
+					struct drm_display_mode *mode)
+{
+	int target = mode->clock * 1000;
+	int target_min = target - CLK_TOLERANCE_HZ;
+	int target_max = target + CLK_TOLERANCE_HZ;
+	int result;
+	struct sti_hdmi_connector *hdmi_connector
+		= to_sti_hdmi_connector(connector);
+	struct sti_hdmi *hdmi = hdmi_connector->hdmi;
+
+
+	result = clk_round_rate(hdmi->clk_pix, target);
+
+	DRM_DEBUG_DRIVER("target rate = %d => available rate = %d\n",
+			 target, result);
+
+	if ((result < target_min) || (result > target_max)) {
+		DRM_DEBUG_DRIVER("hdmi pixclk=%d not supported\n", target);
+		return MODE_BAD;
+	}
+
+	return MODE_OK;
+}
+
+struct drm_encoder *sti_hdmi_best_encoder(struct drm_connector *connector)
+{
+	struct sti_hdmi_connector *hdmi_connector
+		= to_sti_hdmi_connector(connector);
+
+	/* Best encoder is the one associated during connector creation */
+	return hdmi_connector->encoder;
+}
+
+static struct drm_connector_helper_funcs sti_hdmi_connector_helper_funcs = {
+	.get_modes = sti_hdmi_connector_get_modes,
+	.mode_valid = sti_hdmi_connector_mode_valid,
+	.best_encoder = sti_hdmi_best_encoder,
+};
+
+/* get detection status of display device */
+static enum drm_connector_status
+sti_hdmi_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct sti_hdmi_connector *hdmi_connector
+		= to_sti_hdmi_connector(connector);
+	struct sti_hdmi *hdmi = hdmi_connector->hdmi;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (hdmi->hpd) {
+		DRM_DEBUG_DRIVER("hdmi cable connected\n");
+		return connector_status_connected;
+	}
+
+	DRM_DEBUG_DRIVER("hdmi cable disconnected\n");
+	return connector_status_disconnected;
+}
+
+static void sti_hdmi_connector_destroy(struct drm_connector *connector)
+{
+	struct sti_hdmi_connector *hdmi_connector
+		= to_sti_hdmi_connector(connector);
+
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+	kfree(hdmi_connector);
+}
+
+static struct drm_connector_funcs sti_hdmi_connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.detect = sti_hdmi_connector_detect,
+	.destroy = sti_hdmi_connector_destroy,
+};
+
+static struct drm_encoder *sti_hdmi_find_encoder(struct drm_device *dev)
+{
+	struct drm_encoder *encoder;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (encoder->encoder_type == DRM_MODE_ENCODER_TMDS)
+			return encoder;
+	}
+
+	return NULL;
+}
+
+static int sti_hdmi_bind(struct device *dev, struct device *master, void *data)
+{
+	struct sti_hdmi *hdmi = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
+	struct drm_encoder *encoder;
+	struct sti_hdmi_connector *connector;
+	struct drm_connector *drm_connector;
+	struct drm_bridge *bridge;
+	struct i2c_adapter *i2c_adap;
+	int err;
+
+	i2c_adap = i2c_get_adapter(1);
+	if (!i2c_adap)
+		return -EPROBE_DEFER;
+
+	/* Set the drm device handle */
+	hdmi->drm_dev = drm_dev;
+
+	encoder = sti_hdmi_find_encoder(drm_dev);
+	if (!encoder)
+		return -ENOMEM;
+
+	connector = devm_kzalloc(dev, sizeof(*connector), GFP_KERNEL);
+	if (!connector)
+		return -ENOMEM;
+
+	connector->hdmi = hdmi;
+
+	bridge = devm_kzalloc(dev, sizeof(*bridge), GFP_KERNEL);
+	if (!bridge)
+		return -ENOMEM;
+
+	bridge->driver_private = hdmi;
+	drm_bridge_init(drm_dev, bridge, &sti_hdmi_bridge_funcs);
+
+	encoder->bridge = bridge;
+	connector->encoder = encoder;
+
+	drm_connector = (struct drm_connector *)connector;
+
+	drm_connector->polled = DRM_CONNECTOR_POLL_HPD;
+
+	drm_connector_init(drm_dev, drm_connector,
+			&sti_hdmi_connector_funcs, DRM_MODE_CONNECTOR_HDMIA);
+	drm_connector_helper_add(drm_connector,
+			&sti_hdmi_connector_helper_funcs);
+
+	err = drm_connector_register(drm_connector);
+	if (err)
+		goto err_connector;
+
+	err = drm_mode_connector_attach_encoder(drm_connector, encoder);
+	if (err) {
+		DRM_ERROR("Failed to attach a connector to a encoder\n");
+		goto err_sysfs;
+	}
+
+	/* Enable default interrupts */
+	hdmi_write(hdmi, HDMI_DEFAULT_INT, HDMI_INT_EN);
+
+	return 0;
+
+err_sysfs:
+	drm_connector_unregister(drm_connector);
+err_connector:
+	drm_bridge_cleanup(bridge);
+	drm_connector_cleanup(drm_connector);
+	return -EINVAL;
+}
+
+static void sti_hdmi_unbind(struct device *dev,
+		struct device *master, void *data)
+{
+	/* do nothing */
+}
+
+static const struct component_ops sti_hdmi_ops = {
+	.bind = sti_hdmi_bind,
+	.unbind = sti_hdmi_unbind,
+};
+
+static const struct of_device_id hdmi_of_match[] = {
+	{
+		.compatible = "st,stih416-hdmi",
+		.data = &tx3g0c55phy_ops,
+	}, {
+		.compatible = "st,stih407-hdmi",
+		.data = &tx3g4c28phy_ops,
+	}, {
+		/* end node */
+	}
+};
+MODULE_DEVICE_TABLE(of, hdmi_of_match);
+
+static int sti_hdmi_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct sti_hdmi *hdmi;
+	struct device_node *np = dev->of_node;
+	struct resource *res;
+	int ret;
+
+	DRM_INFO("%s\n", __func__);
+
+	hdmi = devm_kzalloc(dev, sizeof(*hdmi), GFP_KERNEL);
+	if (!hdmi)
+		return -ENOMEM;
+
+	hdmi->dev = pdev->dev;
+
+	/* Get resources */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "hdmi-reg");
+	if (!res) {
+		DRM_ERROR("Invalid hdmi resource\n");
+		return -ENOMEM;
+	}
+	hdmi->regs = devm_ioremap_nocache(dev, res->start, resource_size(res));
+	if (!hdmi->regs)
+		return -ENOMEM;
+
+	if (of_device_is_compatible(np, "st,stih416-hdmi")) {
+		res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+						   "syscfg");
+		if (!res) {
+			DRM_ERROR("Invalid syscfg resource\n");
+			return -ENOMEM;
+		}
+		hdmi->syscfg = devm_ioremap_nocache(dev, res->start,
+						    resource_size(res));
+		if (!hdmi->syscfg)
+			return -ENOMEM;
+
+	}
+
+	hdmi->phy_ops = (struct hdmi_phy_ops *)
+		of_match_node(hdmi_of_match, np)->data;
+
+	/* Get clock resources */
+	hdmi->clk_pix = devm_clk_get(dev, "pix");
+	if (IS_ERR(hdmi->clk_pix)) {
+		DRM_ERROR("Cannot get hdmi_pix clock\n");
+		return PTR_ERR(hdmi->clk_pix);
+	}
+
+	hdmi->clk_tmds = devm_clk_get(dev, "tmds");
+	if (IS_ERR(hdmi->clk_tmds)) {
+		DRM_ERROR("Cannot get hdmi_tmds clock\n");
+		return PTR_ERR(hdmi->clk_tmds);
+	}
+
+	hdmi->clk_phy = devm_clk_get(dev, "phy");
+	if (IS_ERR(hdmi->clk_phy)) {
+		DRM_ERROR("Cannot get hdmi_phy clock\n");
+		return PTR_ERR(hdmi->clk_phy);
+	}
+
+	hdmi->clk_audio = devm_clk_get(dev, "audio");
+	if (IS_ERR(hdmi->clk_audio)) {
+		DRM_ERROR("Cannot get hdmi_audio clock\n");
+		return PTR_ERR(hdmi->clk_audio);
+	}
+
+	hdmi->hpd_gpio = of_get_named_gpio(np, "hdmi,hpd-gpio", 0);
+	if (hdmi->hpd_gpio < 0) {
+		DRM_ERROR("Failed to get hdmi hpd-gpio\n");
+		return -EIO;
+	}
+
+	hdmi->hpd = gpio_get_value(hdmi->hpd_gpio);
+
+	init_waitqueue_head(&hdmi->wait_event);
+
+	hdmi->irq = platform_get_irq_byname(pdev, "irq");
+
+	ret = devm_request_threaded_irq(dev, hdmi->irq, hdmi_irq,
+			hdmi_irq_thread, IRQF_ONESHOT, dev_name(dev), hdmi);
+	if (ret) {
+		DRM_ERROR("Failed to register HDMI interrupt\n");
+		return ret;
+	}
+
+	hdmi->reset = devm_reset_control_get(dev, "hdmi");
+	/* Take hdmi out of reset */
+	if (!IS_ERR(hdmi->reset))
+		reset_control_deassert(hdmi->reset);
+
+	platform_set_drvdata(pdev, hdmi);
+
+	return component_add(&pdev->dev, &sti_hdmi_ops);
+}
+
+static int sti_hdmi_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &sti_hdmi_ops);
+	return 0;
+}
+
+struct platform_driver sti_hdmi_driver = {
+	.driver = {
+		.name = "sti-hdmi",
+		.owner = THIS_MODULE,
+		.of_match_table = hdmi_of_match,
+	},
+	.probe = sti_hdmi_probe,
+	.remove = sti_hdmi_remove,
+};
+
+module_platform_driver(sti_hdmi_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_hdmi.h b/drivers/gpu/drm/sti/sti_hdmi.h
new file mode 100644
index 000000000000..61bec6557ceb
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi.h
@@ -0,0 +1,88 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Vincent Abriou <vincent.abriou@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_HDMI_H_
+#define _STI_HDMI_H_
+
+#include <linux/platform_device.h>
+
+#include <drm/drmP.h>
+
+#define HDMI_STA           0x0010
+#define HDMI_STA_DLL_LCK   BIT(5)
+
+struct sti_hdmi;
+
+struct hdmi_phy_ops {
+	bool (*start)(struct sti_hdmi *hdmi);
+	void (*stop)(struct sti_hdmi *hdmi);
+};
+
+/**
+ * STI hdmi structure
+ *
+ * @dev: driver device
+ * @drm_dev: pointer to drm device
+ * @mode: current display mode selected
+ * @regs: hdmi register
+ * @syscfg: syscfg register for pll rejection configuration
+ * @clk_pix: hdmi pixel clock
+ * @clk_tmds: hdmi tmds clock
+ * @clk_phy: hdmi phy clock
+ * @clk_audio: hdmi audio clock
+ * @irq: hdmi interrupt number
+ * @irq_status: interrupt status register
+ * @phy_ops: phy start/stop operations
+ * @enabled: true if hdmi is enabled else false
+ * @hpd_gpio: hdmi hot plug detect gpio number
+ * @hpd: hot plug detect status
+ * @wait_event: wait event
+ * @event_received: wait event status
+ * @reset: reset control of the hdmi phy
+ */
+struct sti_hdmi {
+	struct device dev;
+	struct drm_device *drm_dev;
+	struct drm_display_mode mode;
+	void __iomem *regs;
+	void __iomem *syscfg;
+	struct clk *clk_pix;
+	struct clk *clk_tmds;
+	struct clk *clk_phy;
+	struct clk *clk_audio;
+	int irq;
+	u32 irq_status;
+	struct hdmi_phy_ops *phy_ops;
+	bool enabled;
+	int hpd_gpio;
+	bool hpd;
+	wait_queue_head_t wait_event;
+	bool event_received;
+	struct reset_control *reset;
+};
+
+u32 hdmi_read(struct sti_hdmi *hdmi, int offset);
+void hdmi_write(struct sti_hdmi *hdmi, u32 val, int offset);
+
+/**
+ * hdmi phy config structure
+ *
+ * A pointer to an array of these structures is passed to a TMDS (HDMI) output
+ * via the control interface to provide board and SoC specific
+ * configurations of the HDMI PHY. Each entry in the array specifies a hardware
+ * specific configuration for a given TMDS clock frequency range.
+ *
+ * @min_tmds_freq: Lower bound of TMDS clock frequency this entry applies to
+ * @max_tmds_freq: Upper bound of TMDS clock frequency this entry applies to
+ * @config: SoC specific register configuration
+ */
+struct hdmi_phy_config {
+	u32 min_tmds_freq;
+	u32 max_tmds_freq;
+	u32 config[4];
+};
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.c b/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.c
new file mode 100644
index 000000000000..49ae8e44b285
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.c
@@ -0,0 +1,336 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Vincent Abriou <vincent.abriou@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include "sti_hdmi_tx3g0c55phy.h"
+
+#define HDMI_SRZ_PLL_CFG                0x0504
+#define HDMI_SRZ_TAP_1                  0x0508
+#define HDMI_SRZ_TAP_2                  0x050C
+#define HDMI_SRZ_TAP_3                  0x0510
+#define HDMI_SRZ_CTRL                   0x0514
+
+#define HDMI_SRZ_PLL_CFG_POWER_DOWN     BIT(0)
+#define HDMI_SRZ_PLL_CFG_VCOR_SHIFT     1
+#define HDMI_SRZ_PLL_CFG_VCOR_425MHZ    0
+#define HDMI_SRZ_PLL_CFG_VCOR_850MHZ    1
+#define HDMI_SRZ_PLL_CFG_VCOR_1700MHZ   2
+#define HDMI_SRZ_PLL_CFG_VCOR_3000MHZ   3
+#define HDMI_SRZ_PLL_CFG_VCOR_MASK      3
+#define HDMI_SRZ_PLL_CFG_VCOR(x)        (x << HDMI_SRZ_PLL_CFG_VCOR_SHIFT)
+#define HDMI_SRZ_PLL_CFG_NDIV_SHIFT     8
+#define HDMI_SRZ_PLL_CFG_NDIV_MASK      (0x1F << HDMI_SRZ_PLL_CFG_NDIV_SHIFT)
+#define HDMI_SRZ_PLL_CFG_MODE_SHIFT     16
+#define HDMI_SRZ_PLL_CFG_MODE_13_5_MHZ  0x1
+#define HDMI_SRZ_PLL_CFG_MODE_25_2_MHZ  0x4
+#define HDMI_SRZ_PLL_CFG_MODE_27_MHZ    0x5
+#define HDMI_SRZ_PLL_CFG_MODE_33_75_MHZ 0x6
+#define HDMI_SRZ_PLL_CFG_MODE_40_5_MHZ  0x7
+#define HDMI_SRZ_PLL_CFG_MODE_54_MHZ    0x8
+#define HDMI_SRZ_PLL_CFG_MODE_67_5_MHZ  0x9
+#define HDMI_SRZ_PLL_CFG_MODE_74_25_MHZ 0xA
+#define HDMI_SRZ_PLL_CFG_MODE_81_MHZ    0xB
+#define HDMI_SRZ_PLL_CFG_MODE_82_5_MHZ  0xC
+#define HDMI_SRZ_PLL_CFG_MODE_108_MHZ   0xD
+#define HDMI_SRZ_PLL_CFG_MODE_148_5_MHZ 0xE
+#define HDMI_SRZ_PLL_CFG_MODE_165_MHZ   0xF
+#define HDMI_SRZ_PLL_CFG_MODE_MASK      0xF
+#define HDMI_SRZ_PLL_CFG_MODE(x)        (x << HDMI_SRZ_PLL_CFG_MODE_SHIFT)
+
+#define HDMI_SRZ_CTRL_POWER_DOWN        (1 << 0)
+#define HDMI_SRZ_CTRL_EXTERNAL_DATA_EN  (1 << 1)
+
+/* sysconf registers */
+#define HDMI_REJECTION_PLL_CONFIGURATION 0x0858	/* SYSTEM_CONFIG2534 */
+#define HDMI_REJECTION_PLL_STATUS        0x0948	/* SYSTEM_CONFIG2594 */
+
+#define REJECTION_PLL_HDMI_ENABLE_SHIFT 0
+#define REJECTION_PLL_HDMI_ENABLE_MASK  (0x1 << REJECTION_PLL_HDMI_ENABLE_SHIFT)
+#define REJECTION_PLL_HDMI_PDIV_SHIFT   24
+#define REJECTION_PLL_HDMI_PDIV_MASK    (0x7 << REJECTION_PLL_HDMI_PDIV_SHIFT)
+#define REJECTION_PLL_HDMI_NDIV_SHIFT   16
+#define REJECTION_PLL_HDMI_NDIV_MASK    (0xFF << REJECTION_PLL_HDMI_NDIV_SHIFT)
+#define REJECTION_PLL_HDMI_MDIV_SHIFT   8
+#define REJECTION_PLL_HDMI_MDIV_MASK    (0xFF << REJECTION_PLL_HDMI_MDIV_SHIFT)
+
+#define REJECTION_PLL_HDMI_REJ_PLL_LOCK BIT(0)
+
+#define HDMI_TIMEOUT_PLL_LOCK  50   /*milliseconds */
+
+/**
+ * pll mode structure
+ *
+ * A pointer to an array of these structures is passed to a TMDS (HDMI) output
+ * via the control interface to provide board and SoC specific
+ * configurations of the HDMI PHY. Each entry in the array specifies a hardware
+ * specific configuration for a given TMDS clock frequency range. The array
+ * should be terminated with an entry that has all fields set to zero.
+ *
+ * @min: Lower bound of TMDS clock frequency this entry applies to
+ * @max: Upper bound of TMDS clock frequency this entry applies to
+ * @mode: SoC specific register configuration
+ */
+struct pllmode {
+	u32 min;
+	u32 max;
+	u32 mode;
+};
+
+#define NB_PLL_MODE 7
+static struct pllmode pllmodes[NB_PLL_MODE] = {
+	{13500000, 13513500, HDMI_SRZ_PLL_CFG_MODE_13_5_MHZ},
+	{25174800, 25200000, HDMI_SRZ_PLL_CFG_MODE_25_2_MHZ},
+	{27000000, 27027000, HDMI_SRZ_PLL_CFG_MODE_27_MHZ},
+	{54000000, 54054000, HDMI_SRZ_PLL_CFG_MODE_54_MHZ},
+	{72000000, 74250000, HDMI_SRZ_PLL_CFG_MODE_74_25_MHZ},
+	{108000000, 108108000, HDMI_SRZ_PLL_CFG_MODE_108_MHZ},
+	{148351648, 297000000, HDMI_SRZ_PLL_CFG_MODE_148_5_MHZ}
+};
+
+#define NB_HDMI_PHY_CONFIG 5
+static struct hdmi_phy_config hdmiphy_config[NB_HDMI_PHY_CONFIG] = {
+	{0, 40000000, {0x00101010, 0x00101010, 0x00101010, 0x02} },
+	{40000000, 140000000, {0x00111111, 0x00111111, 0x00111111, 0x02} },
+	{140000000, 160000000, {0x00131313, 0x00101010, 0x00101010, 0x02} },
+	{160000000, 250000000, {0x00131313, 0x00111111, 0x00111111, 0x03FE} },
+	{250000000, 300000000, {0x00151515, 0x00101010, 0x00101010, 0x03FE} },
+};
+
+#define PLL_CHANGE_DELAY	1 /* ms */
+
+/**
+ * Disable the pll rejection
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ * return true if the pll has been disabled
+ */
+static bool disable_pll_rejection(struct sti_hdmi *hdmi)
+{
+	u32 val;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	val = readl(hdmi->syscfg + HDMI_REJECTION_PLL_CONFIGURATION);
+	val &= ~REJECTION_PLL_HDMI_ENABLE_MASK;
+	writel(val, hdmi->syscfg + HDMI_REJECTION_PLL_CONFIGURATION);
+
+	msleep(PLL_CHANGE_DELAY);
+	val = readl(hdmi->syscfg + HDMI_REJECTION_PLL_STATUS);
+
+	return !(val & REJECTION_PLL_HDMI_REJ_PLL_LOCK);
+}
+
+/**
+ * Enable the old BCH/rejection PLL is now reused to provide the CLKPXPLL
+ * clock input to the new PHY PLL that generates the serializer clock
+ * (TMDS*10) and the TMDS clock which is now fed back into the HDMI
+ * formatter instead of the TMDS clock line from ClockGenB.
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ * return true if pll has been correctly set
+ */
+static bool enable_pll_rejection(struct sti_hdmi *hdmi)
+{
+	unsigned int inputclock;
+	u32 mdiv, ndiv, pdiv, val;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	if (!disable_pll_rejection(hdmi))
+		return false;
+
+	inputclock = hdmi->mode.clock * 1000;
+
+	DRM_DEBUG_DRIVER("hdmi rejection pll input clock = %dHz\n", inputclock);
+
+
+	/* Power up the HDMI rejection PLL
+	 * Note: On this SoC (stiH416) we are forced to have the input clock
+	 * be equal to the HDMI pixel clock.
+	 *
+	 * The values here have been suggested by validation however they are
+	 * still provisional and subject to change.
+	 *
+	 * PLLout = (Fin*Mdiv) / ((2 * Ndiv) / 2^Pdiv)
+	 */
+	if (inputclock < 50000000) {
+		/*
+		 * For slower clocks we need to multiply more to keep the
+		 * internal VCO frequency within the physical specification
+		 * of the PLL.
+		 */
+		pdiv = 4;
+		ndiv = 240;
+		mdiv = 30;
+	} else {
+		pdiv = 2;
+		ndiv = 60;
+		mdiv = 30;
+	}
+
+	val = readl(hdmi->syscfg + HDMI_REJECTION_PLL_CONFIGURATION);
+
+	val &= ~(REJECTION_PLL_HDMI_PDIV_MASK |
+		REJECTION_PLL_HDMI_NDIV_MASK |
+		REJECTION_PLL_HDMI_MDIV_MASK |
+		REJECTION_PLL_HDMI_ENABLE_MASK);
+
+	val |=	(pdiv << REJECTION_PLL_HDMI_PDIV_SHIFT) |
+		(ndiv << REJECTION_PLL_HDMI_NDIV_SHIFT) |
+		(mdiv << REJECTION_PLL_HDMI_MDIV_SHIFT) |
+		(0x1 << REJECTION_PLL_HDMI_ENABLE_SHIFT);
+
+	writel(val, hdmi->syscfg + HDMI_REJECTION_PLL_CONFIGURATION);
+
+	msleep(PLL_CHANGE_DELAY);
+	val = readl(hdmi->syscfg + HDMI_REJECTION_PLL_STATUS);
+
+	return (val & REJECTION_PLL_HDMI_REJ_PLL_LOCK);
+}
+
+/**
+ * Start hdmi phy macro cell tx3g0c55
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ * Return false if an error occur
+ */
+static bool sti_hdmi_tx3g0c55phy_start(struct sti_hdmi *hdmi)
+{
+	u32 ckpxpll = hdmi->mode.clock * 1000;
+	u32 val, tmdsck, freqvco, pllctrl = 0;
+	unsigned int i;
+
+	if (!enable_pll_rejection(hdmi))
+		return false;
+
+	DRM_DEBUG_DRIVER("ckpxpll = %dHz\n", ckpxpll);
+
+	/* Assuming no pixel repetition and 24bits color */
+	tmdsck = ckpxpll;
+	pllctrl = 2 << HDMI_SRZ_PLL_CFG_NDIV_SHIFT;
+
+	/*
+	 * Setup the PLL mode parameter based on the ckpxpll. If we haven't got
+	 * a clock frequency supported by one of the specific PLL modes then we
+	 * will end up using the generic mode (0) which only supports a 10x
+	 * multiplier, hence only 24bit color.
+	 */
+	for (i = 0; i < NB_PLL_MODE; i++) {
+		if (ckpxpll >= pllmodes[i].min && ckpxpll <= pllmodes[i].max)
+			pllctrl |= HDMI_SRZ_PLL_CFG_MODE(pllmodes[i].mode);
+	}
+
+	freqvco = tmdsck * 10;
+	if (freqvco <= 425000000UL)
+		pllctrl |= HDMI_SRZ_PLL_CFG_VCOR(HDMI_SRZ_PLL_CFG_VCOR_425MHZ);
+	else if (freqvco <= 850000000UL)
+		pllctrl |= HDMI_SRZ_PLL_CFG_VCOR(HDMI_SRZ_PLL_CFG_VCOR_850MHZ);
+	else if (freqvco <= 1700000000UL)
+		pllctrl |= HDMI_SRZ_PLL_CFG_VCOR(HDMI_SRZ_PLL_CFG_VCOR_1700MHZ);
+	else if (freqvco <= 2970000000UL)
+		pllctrl |= HDMI_SRZ_PLL_CFG_VCOR(HDMI_SRZ_PLL_CFG_VCOR_3000MHZ);
+	else {
+		DRM_ERROR("PHY serializer clock out of range\n");
+		goto err;
+	}
+
+	/*
+	 * Configure and power up the PHY PLL
+	 */
+	hdmi->event_received = false;
+	DRM_DEBUG_DRIVER("pllctrl = 0x%x\n", pllctrl);
+	hdmi_write(hdmi, pllctrl, HDMI_SRZ_PLL_CFG);
+
+	/* wait PLL interrupt */
+	wait_event_interruptible_timeout(hdmi->wait_event,
+					 hdmi->event_received == true,
+					 msecs_to_jiffies
+					 (HDMI_TIMEOUT_PLL_LOCK));
+
+	if ((hdmi_read(hdmi, HDMI_STA) & HDMI_STA_DLL_LCK) == 0) {
+		DRM_ERROR("hdmi phy pll not locked\n");
+		goto err;
+	}
+
+	DRM_DEBUG_DRIVER("got PHY PLL Lock\n");
+
+	/*
+	 * To configure the source termination and pre-emphasis appropriately
+	 * for different high speed TMDS clock frequencies a phy configuration
+	 * table must be provided, tailored to the SoC and board combination.
+	 */
+	for (i = 0; i < NB_HDMI_PHY_CONFIG; i++) {
+		if ((hdmiphy_config[i].min_tmds_freq <= tmdsck) &&
+		    (hdmiphy_config[i].max_tmds_freq >= tmdsck)) {
+			val = hdmiphy_config[i].config[0];
+			hdmi_write(hdmi, val, HDMI_SRZ_TAP_1);
+			val = hdmiphy_config[i].config[1];
+			hdmi_write(hdmi, val, HDMI_SRZ_TAP_2);
+			val = hdmiphy_config[i].config[2];
+			hdmi_write(hdmi, val, HDMI_SRZ_TAP_3);
+			val = hdmiphy_config[i].config[3];
+			val |= HDMI_SRZ_CTRL_EXTERNAL_DATA_EN;
+			val &= ~HDMI_SRZ_CTRL_POWER_DOWN;
+			hdmi_write(hdmi, val, HDMI_SRZ_CTRL);
+
+			DRM_DEBUG_DRIVER("serializer cfg 0x%x 0x%x 0x%x 0x%x\n",
+					 hdmiphy_config[i].config[0],
+					 hdmiphy_config[i].config[1],
+					 hdmiphy_config[i].config[2],
+					 hdmiphy_config[i].config[3]);
+			return true;
+		}
+	}
+
+	/*
+	 * Default, power up the serializer with no pre-emphasis or source
+	 * termination.
+	 */
+	hdmi_write(hdmi, 0x0, HDMI_SRZ_TAP_1);
+	hdmi_write(hdmi, 0x0, HDMI_SRZ_TAP_2);
+	hdmi_write(hdmi, 0x0, HDMI_SRZ_TAP_3);
+	hdmi_write(hdmi, HDMI_SRZ_CTRL_EXTERNAL_DATA_EN, HDMI_SRZ_CTRL);
+
+	return true;
+
+err:
+	disable_pll_rejection(hdmi);
+
+	return false;
+}
+
+/**
+ * Stop hdmi phy macro cell tx3g0c55
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ */
+static void sti_hdmi_tx3g0c55phy_stop(struct sti_hdmi *hdmi)
+{
+	DRM_DEBUG_DRIVER("\n");
+
+	hdmi->event_received = false;
+
+	hdmi_write(hdmi, HDMI_SRZ_CTRL_POWER_DOWN, HDMI_SRZ_CTRL);
+	hdmi_write(hdmi, HDMI_SRZ_PLL_CFG_POWER_DOWN, HDMI_SRZ_PLL_CFG);
+
+	/* wait PLL interrupt */
+	wait_event_interruptible_timeout(hdmi->wait_event,
+					 hdmi->event_received == true,
+					 msecs_to_jiffies
+					 (HDMI_TIMEOUT_PLL_LOCK));
+
+	if (hdmi_read(hdmi, HDMI_STA) & HDMI_STA_DLL_LCK)
+		DRM_ERROR("hdmi phy pll not well disabled\n");
+
+	disable_pll_rejection(hdmi);
+}
+
+struct hdmi_phy_ops tx3g0c55phy_ops = {
+	.start = sti_hdmi_tx3g0c55phy_start,
+	.stop = sti_hdmi_tx3g0c55phy_stop,
+};
diff --git a/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.h b/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.h
new file mode 100644
index 000000000000..068237b3a303
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi_tx3g0c55phy.h
@@ -0,0 +1,14 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_HDMI_TX3G0C55PHY_H_
+#define _STI_HDMI_TX3G0C55PHY_H_
+
+#include "sti_hdmi.h"
+
+extern struct hdmi_phy_ops tx3g0c55phy_ops;
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.c b/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.c
new file mode 100644
index 000000000000..8e0ceb0ced33
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.c
@@ -0,0 +1,211 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Vincent Abriou <vincent.abriou@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include "sti_hdmi_tx3g4c28phy.h"
+
+#define HDMI_SRZ_CFG                             0x504
+#define HDMI_SRZ_PLL_CFG                         0x510
+#define HDMI_SRZ_ICNTL                           0x518
+#define HDMI_SRZ_CALCODE_EXT                     0x520
+
+#define HDMI_SRZ_CFG_EN                          BIT(0)
+#define HDMI_SRZ_CFG_DISABLE_BYPASS_SINK_CURRENT BIT(1)
+#define HDMI_SRZ_CFG_EXTERNAL_DATA               BIT(16)
+#define HDMI_SRZ_CFG_RBIAS_EXT                   BIT(17)
+#define HDMI_SRZ_CFG_EN_SINK_TERM_DETECTION      BIT(18)
+#define HDMI_SRZ_CFG_EN_BIASRES_DETECTION        BIT(19)
+#define HDMI_SRZ_CFG_EN_SRC_TERMINATION          BIT(24)
+
+#define HDMI_SRZ_CFG_INTERNAL_MASK  (HDMI_SRZ_CFG_EN     | \
+		HDMI_SRZ_CFG_DISABLE_BYPASS_SINK_CURRENT | \
+		HDMI_SRZ_CFG_EXTERNAL_DATA               | \
+		HDMI_SRZ_CFG_RBIAS_EXT                   | \
+		HDMI_SRZ_CFG_EN_SINK_TERM_DETECTION      | \
+		HDMI_SRZ_CFG_EN_BIASRES_DETECTION        | \
+		HDMI_SRZ_CFG_EN_SRC_TERMINATION)
+
+#define PLL_CFG_EN                               BIT(0)
+#define PLL_CFG_NDIV_SHIFT                       (8)
+#define PLL_CFG_IDF_SHIFT                        (16)
+#define PLL_CFG_ODF_SHIFT                        (24)
+
+#define ODF_DIV_1                                (0)
+#define ODF_DIV_2                                (1)
+#define ODF_DIV_4                                (2)
+#define ODF_DIV_8                                (3)
+
+#define HDMI_TIMEOUT_PLL_LOCK  50  /*milliseconds */
+
+struct plldividers_s {
+	uint32_t min;
+	uint32_t max;
+	uint32_t idf;
+	uint32_t odf;
+};
+
+/*
+ * Functional specification recommended values
+ */
+#define NB_PLL_MODE 5
+static struct plldividers_s plldividers[NB_PLL_MODE] = {
+	{0, 20000000, 1, ODF_DIV_8},
+	{20000000, 42500000, 2, ODF_DIV_8},
+	{42500000, 85000000, 4, ODF_DIV_4},
+	{85000000, 170000000, 8, ODF_DIV_2},
+	{170000000, 340000000, 16, ODF_DIV_1}
+};
+
+#define NB_HDMI_PHY_CONFIG 2
+static struct hdmi_phy_config hdmiphy_config[NB_HDMI_PHY_CONFIG] = {
+	{0, 250000000, {0x0, 0x0, 0x0, 0x0} },
+	{250000000, 300000000, {0x1110, 0x0, 0x0, 0x0} },
+};
+
+/**
+ * Start hdmi phy macro cell tx3g4c28
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ *
+ * Return false if an error occur
+ */
+static bool sti_hdmi_tx3g4c28phy_start(struct sti_hdmi *hdmi)
+{
+	u32 ckpxpll = hdmi->mode.clock * 1000;
+	u32 val, tmdsck, idf, odf, pllctrl = 0;
+	bool foundplldivides = false;
+	int i;
+
+	DRM_DEBUG_DRIVER("ckpxpll = %dHz\n", ckpxpll);
+
+	for (i = 0; i < NB_PLL_MODE; i++) {
+		if (ckpxpll >= plldividers[i].min &&
+		    ckpxpll < plldividers[i].max) {
+			idf = plldividers[i].idf;
+			odf = plldividers[i].odf;
+			foundplldivides = true;
+			break;
+		}
+	}
+
+	if (!foundplldivides) {
+		DRM_ERROR("input TMDS clock speed (%d) not supported\n",
+			  ckpxpll);
+		goto err;
+	}
+
+	/* Assuming no pixel repetition and 24bits color */
+	tmdsck = ckpxpll;
+	pllctrl |= 40 << PLL_CFG_NDIV_SHIFT;
+
+	if (tmdsck > 340000000) {
+		DRM_ERROR("output TMDS clock (%d) out of range\n", tmdsck);
+		goto err;
+	}
+
+	pllctrl |= idf << PLL_CFG_IDF_SHIFT;
+	pllctrl |= odf << PLL_CFG_ODF_SHIFT;
+
+	/*
+	 * Configure and power up the PHY PLL
+	 */
+	hdmi->event_received = false;
+	DRM_DEBUG_DRIVER("pllctrl = 0x%x\n", pllctrl);
+	hdmi_write(hdmi, (pllctrl | PLL_CFG_EN), HDMI_SRZ_PLL_CFG);
+
+	/* wait PLL interrupt */
+	wait_event_interruptible_timeout(hdmi->wait_event,
+					 hdmi->event_received == true,
+					 msecs_to_jiffies
+					 (HDMI_TIMEOUT_PLL_LOCK));
+
+	if ((hdmi_read(hdmi, HDMI_STA) & HDMI_STA_DLL_LCK) == 0) {
+		DRM_ERROR("hdmi phy pll not locked\n");
+		goto err;
+	}
+
+	DRM_DEBUG_DRIVER("got PHY PLL Lock\n");
+
+	val = (HDMI_SRZ_CFG_EN |
+	       HDMI_SRZ_CFG_EXTERNAL_DATA |
+	       HDMI_SRZ_CFG_EN_BIASRES_DETECTION |
+	       HDMI_SRZ_CFG_EN_SINK_TERM_DETECTION);
+
+	if (tmdsck > 165000000)
+		val |= HDMI_SRZ_CFG_EN_SRC_TERMINATION;
+
+	/*
+	 * To configure the source termination and pre-emphasis appropriately
+	 * for different high speed TMDS clock frequencies a phy configuration
+	 * table must be provided, tailored to the SoC and board combination.
+	 */
+	for (i = 0; i < NB_HDMI_PHY_CONFIG; i++) {
+		if ((hdmiphy_config[i].min_tmds_freq <= tmdsck) &&
+		    (hdmiphy_config[i].max_tmds_freq >= tmdsck)) {
+			val |= (hdmiphy_config[i].config[0]
+				& ~HDMI_SRZ_CFG_INTERNAL_MASK);
+			hdmi_write(hdmi, val, HDMI_SRZ_CFG);
+
+			val = hdmiphy_config[i].config[1];
+			hdmi_write(hdmi, val, HDMI_SRZ_ICNTL);
+
+			val = hdmiphy_config[i].config[2];
+			hdmi_write(hdmi, val, HDMI_SRZ_CALCODE_EXT);
+
+			DRM_DEBUG_DRIVER("serializer cfg 0x%x 0x%x 0x%x\n",
+					 hdmiphy_config[i].config[0],
+					 hdmiphy_config[i].config[1],
+					 hdmiphy_config[i].config[2]);
+			return true;
+		}
+	}
+
+	/*
+	 * Default, power up the serializer with no pre-emphasis or
+	 * output swing correction
+	 */
+	hdmi_write(hdmi, val,  HDMI_SRZ_CFG);
+	hdmi_write(hdmi, 0x0, HDMI_SRZ_ICNTL);
+	hdmi_write(hdmi, 0x0, HDMI_SRZ_CALCODE_EXT);
+
+	return true;
+
+err:
+	return false;
+}
+
+/**
+ * Stop hdmi phy macro cell tx3g4c28
+ *
+ * @hdmi: pointer on the hdmi internal structure
+ */
+static void sti_hdmi_tx3g4c28phy_stop(struct sti_hdmi *hdmi)
+{
+	int val = 0;
+
+	DRM_DEBUG_DRIVER("\n");
+
+	hdmi->event_received = false;
+
+	val = HDMI_SRZ_CFG_EN_SINK_TERM_DETECTION;
+	val |= HDMI_SRZ_CFG_EN_BIASRES_DETECTION;
+
+	hdmi_write(hdmi, val, HDMI_SRZ_CFG);
+	hdmi_write(hdmi, 0, HDMI_SRZ_PLL_CFG);
+
+	/* wait PLL interrupt */
+	wait_event_interruptible_timeout(hdmi->wait_event,
+					 hdmi->event_received == true,
+					 msecs_to_jiffies
+					 (HDMI_TIMEOUT_PLL_LOCK));
+
+	if (hdmi_read(hdmi, HDMI_STA) & HDMI_STA_DLL_LCK)
+		DRM_ERROR("hdmi phy pll not well disabled\n");
+}
+
+struct hdmi_phy_ops tx3g4c28phy_ops = {
+	.start = sti_hdmi_tx3g4c28phy_start,
+	.stop = sti_hdmi_tx3g4c28phy_stop,
+};
diff --git a/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.h b/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.h
new file mode 100644
index 000000000000..f99a7ff281ef
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_hdmi_tx3g4c28phy.h
@@ -0,0 +1,14 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_HDMI_TX3G4C28PHY_H_
+#define _STI_HDMI_TX3G4C28PHY_H_
+
+#include "sti_hdmi.h"
+
+extern struct hdmi_phy_ops tx3g4c28phy_ops;
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_layer.c b/drivers/gpu/drm/sti/sti_layer.c
new file mode 100644
index 000000000000..06a587c4f1bb
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_layer.c
@@ -0,0 +1,197 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+
+#include "sti_compositor.h"
+#include "sti_gdp.h"
+#include "sti_layer.h"
+#include "sti_vid.h"
+
+const char *sti_layer_to_str(struct sti_layer *layer)
+{
+	switch (layer->desc) {
+	case STI_GDP_0:
+		return "GDP0";
+	case STI_GDP_1:
+		return "GDP1";
+	case STI_GDP_2:
+		return "GDP2";
+	case STI_GDP_3:
+		return "GDP3";
+	case STI_VID_0:
+		return "VID0";
+	case STI_VID_1:
+		return "VID1";
+	case STI_CURSOR:
+		return "CURSOR";
+	default:
+		return "<UNKNOWN LAYER>";
+	}
+}
+
+struct sti_layer *sti_layer_create(struct device *dev, int desc,
+				   void __iomem *baseaddr)
+{
+
+	struct sti_layer *layer = NULL;
+
+	switch (desc & STI_LAYER_TYPE_MASK) {
+	case STI_GDP:
+		layer = sti_gdp_create(dev, desc);
+		break;
+	case STI_VID:
+		layer = sti_vid_create(dev);
+		break;
+	}
+
+	if (!layer) {
+		DRM_ERROR("Failed to create layer\n");
+		return NULL;
+	}
+
+	layer->desc = desc;
+	layer->dev = dev;
+	layer->regs = baseaddr;
+
+	layer->ops->init(layer);
+
+	DRM_DEBUG_DRIVER("%s created\n", sti_layer_to_str(layer));
+
+	return layer;
+}
+
+int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
+		      struct drm_display_mode *mode, int mixer_id,
+		      int dest_x, int dest_y, int dest_w, int dest_h,
+		      int src_x, int src_y, int src_w, int src_h)
+{
+	int ret;
+	unsigned int i;
+	struct drm_gem_cma_object *cma_obj;
+
+	if (!layer || !fb || !mode) {
+		DRM_ERROR("Null fb, layer or mode\n");
+		return 1;
+	}
+
+	cma_obj = drm_fb_cma_get_gem_obj(fb, 0);
+	if (!cma_obj) {
+		DRM_ERROR("Can't get CMA GEM object for fb\n");
+		return 1;
+	}
+
+	layer->fb = fb;
+	layer->mode = mode;
+	layer->mixer_id = mixer_id;
+	layer->dst_x = dest_x;
+	layer->dst_y = dest_y;
+	layer->dst_w = clamp_val(dest_w, 0, mode->crtc_hdisplay - dest_x);
+	layer->dst_h = clamp_val(dest_h, 0, mode->crtc_vdisplay - dest_y);
+	layer->src_x = src_x;
+	layer->src_y = src_y;
+	layer->src_w = src_w;
+	layer->src_h = src_h;
+	layer->format = fb->pixel_format;
+	layer->paddr = cma_obj->paddr;
+	for (i = 0; i < 4; i++) {
+		layer->pitches[i] = fb->pitches[i];
+		layer->offsets[i] = fb->offsets[i];
+	}
+
+	DRM_DEBUG_DRIVER("%s is associated with mixer_id %d\n",
+			 sti_layer_to_str(layer),
+			 layer->mixer_id);
+	DRM_DEBUG_DRIVER("%s dst=(%dx%d)@(%d,%d) - src=(%dx%d)@(%d,%d)\n",
+			 sti_layer_to_str(layer),
+			 layer->dst_w, layer->dst_h, layer->dst_x, layer->dst_y,
+			 layer->src_w, layer->src_h, layer->src_x,
+			 layer->src_y);
+
+	DRM_DEBUG_DRIVER("drm FB:%d format:%.4s phys@:0x%lx\n", fb->base.id,
+			 (char *)&layer->format, (unsigned long)layer->paddr);
+
+	if (!layer->ops->prepare)
+		goto err_no_prepare;
+
+	ret = layer->ops->prepare(layer, !layer->enabled);
+	if (!ret)
+		layer->enabled = true;
+
+	return ret;
+
+err_no_prepare:
+	DRM_ERROR("Cannot prepare\n");
+	return 1;
+}
+
+int sti_layer_commit(struct sti_layer *layer)
+{
+	if (!layer)
+		return 1;
+
+	if (!layer->ops->commit)
+		goto err_no_commit;
+
+	return layer->ops->commit(layer);
+
+err_no_commit:
+	DRM_ERROR("Cannot commit\n");
+	return 1;
+}
+
+int sti_layer_disable(struct sti_layer *layer)
+{
+	int ret;
+
+	DRM_DEBUG_DRIVER("%s\n", sti_layer_to_str(layer));
+	if (!layer)
+		return 1;
+
+	if (!layer->enabled)
+		return 0;
+
+	if (!layer->ops->disable)
+		goto err_no_disable;
+
+	ret = layer->ops->disable(layer);
+	if (!ret)
+		layer->enabled = false;
+	else
+		DRM_ERROR("Disable failed\n");
+
+	return ret;
+
+err_no_disable:
+	DRM_ERROR("Cannot disable\n");
+	return 1;
+}
+
+const uint32_t *sti_layer_get_formats(struct sti_layer *layer)
+{
+	if (!layer)
+		return NULL;
+
+	if (!layer->ops->get_formats)
+		return NULL;
+
+	return layer->ops->get_formats(layer);
+}
+
+unsigned int sti_layer_get_nb_formats(struct sti_layer *layer)
+{
+	if (!layer)
+		return 0;
+
+	if (!layer->ops->get_nb_formats)
+		return 0;
+
+	return layer->ops->get_nb_formats(layer);
+}
diff --git a/drivers/gpu/drm/sti/sti_layer.h b/drivers/gpu/drm/sti/sti_layer.h
new file mode 100644
index 000000000000..198c3774cc12
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_layer.h
@@ -0,0 +1,123 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_LAYER_H_
+#define _STI_LAYER_H_
+
+#include <drm/drmP.h>
+
+#define to_sti_layer(x) container_of(x, struct sti_layer, plane)
+
+#define STI_LAYER_TYPE_SHIFT 8
+#define STI_LAYER_TYPE_MASK (~((1<<STI_LAYER_TYPE_SHIFT)-1))
+
+struct sti_layer;
+
+enum sti_layer_type {
+	STI_GDP = 1 << STI_LAYER_TYPE_SHIFT,
+	STI_VID = 2 << STI_LAYER_TYPE_SHIFT,
+	STI_CUR = 3 << STI_LAYER_TYPE_SHIFT,
+	STI_BCK = 4 << STI_LAYER_TYPE_SHIFT
+};
+
+enum sti_layer_id_of_type {
+	STI_ID_0 = 0,
+	STI_ID_1 = 1,
+	STI_ID_2 = 2,
+	STI_ID_3 = 3
+};
+
+enum sti_layer_desc {
+	STI_GDP_0       = STI_GDP | STI_ID_0,
+	STI_GDP_1       = STI_GDP | STI_ID_1,
+	STI_GDP_2       = STI_GDP | STI_ID_2,
+	STI_GDP_3       = STI_GDP | STI_ID_3,
+	STI_VID_0       = STI_VID | STI_ID_0,
+	STI_VID_1       = STI_VID | STI_ID_1,
+	STI_CURSOR      = STI_CUR,
+	STI_BACK        = STI_BCK
+};
+
+/**
+ * STI layer functions structure
+ *
+ * @get_formats:	get layer supported formats
+ * @get_nb_formats:	get number of format supported
+ * @init:               initialize the layer
+ * @prepare:		prepare layer before rendering
+ * @commit:		set layer for rendering
+ * @disable:		disable layer
+ */
+struct sti_layer_funcs {
+	const uint32_t* (*get_formats)(struct sti_layer *layer);
+	unsigned int (*get_nb_formats)(struct sti_layer *layer);
+	void (*init)(struct sti_layer *layer);
+	int (*prepare)(struct sti_layer *layer, bool first_prepare);
+	int (*commit)(struct sti_layer *layer);
+	int (*disable)(struct sti_layer *layer);
+};
+
+/**
+ * STI layer structure
+ *
+ * @plane:              drm plane it is bound to (if any)
+ * @fb:                 drm fb it is bound to
+ * @mode:               display mode
+ * @desc:               layer type & id
+ * @device:		driver device
+ * @regs:		layer registers
+ * @ops:                layer functions
+ * @zorder:             layer z-order
+ * @mixer_id:           id of the mixer used to display the layer
+ * @enabled:            to know if the layer is active or not
+ * @src_x src_y:        coordinates of the input (fb) area
+ * @src_w src_h:        size of the input (fb) area
+ * @dst_x dst_y:        coordinates of the output (crtc) area
+ * @dst_w dst_h:        size of the output (crtc) area
+ * @format:             format
+ * @pitches:            pitch of 'planes' (eg: Y, U, V)
+ * @offsets:            offset of 'planes'
+ * @paddr:              physical address of the input buffer
+ */
+struct sti_layer {
+	struct drm_plane plane;
+	struct drm_framebuffer *fb;
+	struct drm_display_mode *mode;
+	enum sti_layer_desc desc;
+	struct device *dev;
+	void __iomem *regs;
+	const struct sti_layer_funcs *ops;
+	int zorder;
+	int mixer_id;
+	bool enabled;
+	int src_x, src_y;
+	int src_w, src_h;
+	int dst_x, dst_y;
+	int dst_w, dst_h;
+	uint32_t format;
+	unsigned int pitches[4];
+	unsigned int offsets[4];
+	dma_addr_t paddr;
+};
+
+struct sti_layer *sti_layer_create(struct device *dev, int desc,
+			void __iomem *baseaddr);
+int sti_layer_prepare(struct sti_layer *layer, struct drm_framebuffer *fb,
+			struct drm_display_mode *mode,
+			int mixer_id,
+			int dest_x, int dest_y,
+			int dest_w, int dest_h,
+			int src_x, int src_y,
+			int src_w, int src_h);
+int sti_layer_commit(struct sti_layer *layer);
+int sti_layer_disable(struct sti_layer *layer);
+const uint32_t *sti_layer_get_formats(struct sti_layer *layer);
+unsigned int sti_layer_get_nb_formats(struct sti_layer *layer);
+const char *sti_layer_to_str(struct sti_layer *layer);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_mixer.c b/drivers/gpu/drm/sti/sti_mixer.c
new file mode 100644
index 000000000000..79f369db9fb6
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_mixer.c
@@ -0,0 +1,249 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include "sti_compositor.h"
+#include "sti_mixer.h"
+#include "sti_vtg.h"
+
+/* Identity: G=Y , B=Cb , R=Cr */
+static const u32 mixerColorSpaceMatIdentity[] = {
+	0x10000000, 0x00000000, 0x10000000, 0x00001000,
+	0x00000000, 0x00000000, 0x00000000, 0x00000000
+};
+
+/* regs offset */
+#define GAM_MIXER_CTL      0x00
+#define GAM_MIXER_BKC      0x04
+#define GAM_MIXER_BCO      0x0C
+#define GAM_MIXER_BCS      0x10
+#define GAM_MIXER_AVO      0x28
+#define GAM_MIXER_AVS      0x2C
+#define GAM_MIXER_CRB      0x34
+#define GAM_MIXER_ACT      0x38
+#define GAM_MIXER_MBP      0x3C
+#define GAM_MIXER_MX0      0x80
+
+/* id for depth of CRB reg */
+#define GAM_DEPTH_VID0_ID  1
+#define GAM_DEPTH_VID1_ID  2
+#define GAM_DEPTH_GDP0_ID  3
+#define GAM_DEPTH_GDP1_ID  4
+#define GAM_DEPTH_GDP2_ID  5
+#define GAM_DEPTH_GDP3_ID  6
+#define GAM_DEPTH_MASK_ID  7
+
+/* mask in CTL reg */
+#define GAM_CTL_BACK_MASK  BIT(0)
+#define GAM_CTL_VID0_MASK  BIT(1)
+#define GAM_CTL_VID1_MASK  BIT(2)
+#define GAM_CTL_GDP0_MASK  BIT(3)
+#define GAM_CTL_GDP1_MASK  BIT(4)
+#define GAM_CTL_GDP2_MASK  BIT(5)
+#define GAM_CTL_GDP3_MASK  BIT(6)
+
+const char *sti_mixer_to_str(struct sti_mixer *mixer)
+{
+	switch (mixer->id) {
+	case STI_MIXER_MAIN:
+		return "MAIN_MIXER";
+	case STI_MIXER_AUX:
+		return "AUX_MIXER";
+	default:
+		return "<UNKNOWN MIXER>";
+	}
+}
+
+static inline u32 sti_mixer_reg_read(struct sti_mixer *mixer, u32 reg_id)
+{
+	return readl(mixer->regs + reg_id);
+}
+
+static inline void sti_mixer_reg_write(struct sti_mixer *mixer,
+				       u32 reg_id, u32 val)
+{
+	writel(val, mixer->regs + reg_id);
+}
+
+void sti_mixer_set_background_status(struct sti_mixer *mixer, bool enable)
+{
+	u32 val = sti_mixer_reg_read(mixer, GAM_MIXER_CTL);
+
+	val &= ~GAM_CTL_BACK_MASK;
+	val |= enable;
+	sti_mixer_reg_write(mixer, GAM_MIXER_CTL, val);
+}
+
+static void sti_mixer_set_background_color(struct sti_mixer *mixer,
+					   u8 red, u8 green, u8 blue)
+{
+	u32 val = (red << 16) | (green << 8) | blue;
+
+	sti_mixer_reg_write(mixer, GAM_MIXER_BKC, val);
+}
+
+static void sti_mixer_set_background_area(struct sti_mixer *mixer,
+					  struct drm_display_mode *mode)
+{
+	u32 ydo, xdo, yds, xds;
+
+	ydo = sti_vtg_get_line_number(*mode, 0);
+	yds = sti_vtg_get_line_number(*mode, mode->vdisplay - 1);
+	xdo = sti_vtg_get_pixel_number(*mode, 0);
+	xds = sti_vtg_get_pixel_number(*mode, mode->hdisplay - 1);
+
+	sti_mixer_reg_write(mixer, GAM_MIXER_BCO, ydo << 16 | xdo);
+	sti_mixer_reg_write(mixer, GAM_MIXER_BCS, yds << 16 | xds);
+}
+
+int sti_mixer_set_layer_depth(struct sti_mixer *mixer, struct sti_layer *layer)
+{
+	int layer_id = 0, depth = layer->zorder;
+	u32 mask, val;
+
+	if (depth >= GAM_MIXER_NB_DEPTH_LEVEL)
+		return 1;
+
+	switch (layer->desc) {
+	case STI_GDP_0:
+		layer_id = GAM_DEPTH_GDP0_ID;
+		break;
+	case STI_GDP_1:
+		layer_id = GAM_DEPTH_GDP1_ID;
+		break;
+	case STI_GDP_2:
+		layer_id = GAM_DEPTH_GDP2_ID;
+		break;
+	case STI_GDP_3:
+		layer_id = GAM_DEPTH_GDP3_ID;
+		break;
+	case STI_VID_0:
+		layer_id = GAM_DEPTH_VID0_ID;
+		break;
+	case STI_VID_1:
+		layer_id = GAM_DEPTH_VID1_ID;
+		break;
+	default:
+		DRM_ERROR("Unknown layer %d\n", layer->desc);
+		return 1;
+	}
+	mask = GAM_DEPTH_MASK_ID << (3 * depth);
+	layer_id = layer_id << (3 * depth);
+
+	DRM_DEBUG_DRIVER("%s %s depth=%d\n", sti_mixer_to_str(mixer),
+			 sti_layer_to_str(layer), depth);
+	dev_dbg(mixer->dev, "GAM_MIXER_CRB val 0x%x mask 0x%x\n",
+		layer_id, mask);
+
+	val = sti_mixer_reg_read(mixer, GAM_MIXER_CRB);
+	val &= ~mask;
+	val |= layer_id;
+	sti_mixer_reg_write(mixer, GAM_MIXER_CRB, val);
+
+	dev_dbg(mixer->dev, "Read GAM_MIXER_CRB 0x%x\n",
+		sti_mixer_reg_read(mixer, GAM_MIXER_CRB));
+	return 0;
+}
+
+int sti_mixer_active_video_area(struct sti_mixer *mixer,
+				struct drm_display_mode *mode)
+{
+	u32 ydo, xdo, yds, xds;
+
+	ydo = sti_vtg_get_line_number(*mode, 0);
+	yds = sti_vtg_get_line_number(*mode, mode->vdisplay - 1);
+	xdo = sti_vtg_get_pixel_number(*mode, 0);
+	xds = sti_vtg_get_pixel_number(*mode, mode->hdisplay - 1);
+
+	DRM_DEBUG_DRIVER("%s active video area xdo:%d ydo:%d xds:%d yds:%d\n",
+			 sti_mixer_to_str(mixer), xdo, ydo, xds, yds);
+	sti_mixer_reg_write(mixer, GAM_MIXER_AVO, ydo << 16 | xdo);
+	sti_mixer_reg_write(mixer, GAM_MIXER_AVS, yds << 16 | xds);
+
+	sti_mixer_set_background_color(mixer, 0xFF, 0, 0);
+
+	sti_mixer_set_background_area(mixer, mode);
+	sti_mixer_set_background_status(mixer, true);
+	return 0;
+}
+
+static u32 sti_mixer_get_layer_mask(struct sti_layer *layer)
+{
+	switch (layer->desc) {
+	case STI_BACK:
+		return GAM_CTL_BACK_MASK;
+	case STI_GDP_0:
+		return GAM_CTL_GDP0_MASK;
+	case STI_GDP_1:
+		return GAM_CTL_GDP1_MASK;
+	case STI_GDP_2:
+		return GAM_CTL_GDP2_MASK;
+	case STI_GDP_3:
+		return GAM_CTL_GDP3_MASK;
+	case STI_VID_0:
+		return GAM_CTL_VID0_MASK;
+	case STI_VID_1:
+		return GAM_CTL_VID1_MASK;
+	default:
+		return 0;
+	}
+}
+
+int sti_mixer_set_layer_status(struct sti_mixer *mixer,
+			       struct sti_layer *layer, bool status)
+{
+	u32 mask, val;
+
+	DRM_DEBUG_DRIVER("%s %s %s\n", status ? "enable" : "disable",
+			 sti_mixer_to_str(mixer), sti_layer_to_str(layer));
+
+	mask = sti_mixer_get_layer_mask(layer);
+	if (!mask) {
+		DRM_ERROR("Can not find layer mask\n");
+		return -EINVAL;
+	}
+
+	val = sti_mixer_reg_read(mixer, GAM_MIXER_CTL);
+	val &= ~mask;
+	val |= status ? mask : 0;
+	sti_mixer_reg_write(mixer, GAM_MIXER_CTL, val);
+
+	return 0;
+}
+
+void sti_mixer_set_matrix(struct sti_mixer *mixer)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(mixerColorSpaceMatIdentity); i++)
+		sti_mixer_reg_write(mixer, GAM_MIXER_MX0 + (i * 4),
+				    mixerColorSpaceMatIdentity[i]);
+}
+
+struct sti_mixer *sti_mixer_create(struct device *dev, int id,
+				   void __iomem *baseaddr)
+{
+	struct sti_mixer *mixer = devm_kzalloc(dev, sizeof(*mixer), GFP_KERNEL);
+	struct device_node *np = dev->of_node;
+
+	dev_dbg(dev, "%s\n", __func__);
+	if (!mixer) {
+		DRM_ERROR("Failed to allocated memory for mixer\n");
+		return NULL;
+	}
+	mixer->regs = baseaddr;
+	mixer->dev = dev;
+	mixer->id = id;
+
+	if (of_device_is_compatible(np, "st,stih416-compositor"))
+		sti_mixer_set_matrix(mixer);
+
+	DRM_DEBUG_DRIVER("%s created. Regs=%p\n",
+			 sti_mixer_to_str(mixer), mixer->regs);
+
+	return mixer;
+}
diff --git a/drivers/gpu/drm/sti/sti_mixer.h b/drivers/gpu/drm/sti/sti_mixer.h
new file mode 100644
index 000000000000..874372102e52
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_mixer.h
@@ -0,0 +1,54 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_MIXER_H_
+#define _STI_MIXER_H_
+
+#include <drm/drmP.h>
+
+#include "sti_layer.h"
+
+#define to_sti_mixer(x) container_of(x, struct sti_mixer, drm_crtc)
+
+/**
+ * STI Mixer subdevice structure
+ *
+ * @dev: driver device
+ * @regs: mixer registers
+ * @id: id of the mixer
+ * @drm_crtc: crtc object link to the mixer
+ * @pending_event: set if a flip event is pending on crtc
+ */
+struct sti_mixer {
+	struct device *dev;
+	void __iomem *regs;
+	int id;
+	struct drm_crtc	drm_crtc;
+	struct drm_pending_vblank_event *pending_event;
+};
+
+const char *sti_mixer_to_str(struct sti_mixer *mixer);
+
+struct sti_mixer *sti_mixer_create(struct device *dev, int id,
+		void __iomem *baseaddr);
+
+int sti_mixer_set_layer_status(struct sti_mixer *mixer,
+		struct sti_layer *layer, bool status);
+int sti_mixer_set_layer_depth(struct sti_mixer *mixer, struct sti_layer *layer);
+int sti_mixer_active_video_area(struct sti_mixer *mixer,
+		struct drm_display_mode *mode);
+
+void sti_mixer_set_background_status(struct sti_mixer *mixer, bool enable);
+
+/* depth in Cross-bar control = z order */
+#define GAM_MIXER_NB_DEPTH_LEVEL 7
+
+#define STI_MIXER_MAIN 0
+#define STI_MIXER_AUX  1
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_tvout.c b/drivers/gpu/drm/sti/sti_tvout.c
new file mode 100644
index 000000000000..b8afe490356a
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_tvout.c
@@ -0,0 +1,648 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Vincent Abriou <vincent.abriou@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+
+/* glue registers */
+#define TVO_CSC_MAIN_M0                  0x000
+#define TVO_CSC_MAIN_M1                  0x004
+#define TVO_CSC_MAIN_M2                  0x008
+#define TVO_CSC_MAIN_M3                  0x00c
+#define TVO_CSC_MAIN_M4                  0x010
+#define TVO_CSC_MAIN_M5                  0x014
+#define TVO_CSC_MAIN_M6                  0x018
+#define TVO_CSC_MAIN_M7                  0x01c
+#define TVO_MAIN_IN_VID_FORMAT           0x030
+#define TVO_CSC_AUX_M0                   0x100
+#define TVO_CSC_AUX_M1                   0x104
+#define TVO_CSC_AUX_M2                   0x108
+#define TVO_CSC_AUX_M3                   0x10c
+#define TVO_CSC_AUX_M4                   0x110
+#define TVO_CSC_AUX_M5                   0x114
+#define TVO_CSC_AUX_M6                   0x118
+#define TVO_CSC_AUX_M7                   0x11c
+#define TVO_AUX_IN_VID_FORMAT            0x130
+#define TVO_VIP_HDF                      0x400
+#define TVO_HD_SYNC_SEL                  0x418
+#define TVO_HD_DAC_CFG_OFF               0x420
+#define TVO_VIP_HDMI                     0x500
+#define TVO_HDMI_FORCE_COLOR_0           0x504
+#define TVO_HDMI_FORCE_COLOR_1           0x508
+#define TVO_HDMI_CLIP_VALUE_B_CB         0x50c
+#define TVO_HDMI_CLIP_VALUE_Y_G          0x510
+#define TVO_HDMI_CLIP_VALUE_R_CR         0x514
+#define TVO_HDMI_SYNC_SEL                0x518
+#define TVO_HDMI_DFV_OBS                 0x540
+
+#define TVO_IN_FMT_SIGNED                BIT(0)
+#define TVO_SYNC_EXT                     BIT(4)
+
+#define TVO_VIP_REORDER_R_SHIFT          24
+#define TVO_VIP_REORDER_G_SHIFT          20
+#define TVO_VIP_REORDER_B_SHIFT          16
+#define TVO_VIP_REORDER_MASK             0x3
+#define TVO_VIP_REORDER_Y_G_SEL          0
+#define TVO_VIP_REORDER_CB_B_SEL         1
+#define TVO_VIP_REORDER_CR_R_SEL         2
+
+#define TVO_VIP_CLIP_SHIFT               8
+#define TVO_VIP_CLIP_MASK                0x7
+#define TVO_VIP_CLIP_DISABLED            0
+#define TVO_VIP_CLIP_EAV_SAV             1
+#define TVO_VIP_CLIP_LIMITED_RANGE_RGB_Y 2
+#define TVO_VIP_CLIP_LIMITED_RANGE_CB_CR 3
+#define TVO_VIP_CLIP_PROG_RANGE          4
+
+#define TVO_VIP_RND_SHIFT                4
+#define TVO_VIP_RND_MASK                 0x3
+#define TVO_VIP_RND_8BIT_ROUNDED         0
+#define TVO_VIP_RND_10BIT_ROUNDED        1
+#define TVO_VIP_RND_12BIT_ROUNDED        2
+
+#define TVO_VIP_SEL_INPUT_MASK           0xf
+#define TVO_VIP_SEL_INPUT_MAIN           0x0
+#define TVO_VIP_SEL_INPUT_AUX            0x8
+#define TVO_VIP_SEL_INPUT_FORCE_COLOR    0xf
+#define TVO_VIP_SEL_INPUT_BYPASS_MASK    0x1
+#define TVO_VIP_SEL_INPUT_BYPASSED       1
+
+#define TVO_SYNC_MAIN_VTG_SET_REF        0x00
+#define TVO_SYNC_MAIN_VTG_SET_1          0x01
+#define TVO_SYNC_MAIN_VTG_SET_2          0x02
+#define TVO_SYNC_MAIN_VTG_SET_3          0x03
+#define TVO_SYNC_MAIN_VTG_SET_4          0x04
+#define TVO_SYNC_MAIN_VTG_SET_5          0x05
+#define TVO_SYNC_MAIN_VTG_SET_6          0x06
+#define TVO_SYNC_AUX_VTG_SET_REF         0x10
+#define TVO_SYNC_AUX_VTG_SET_1           0x11
+#define TVO_SYNC_AUX_VTG_SET_2           0x12
+#define TVO_SYNC_AUX_VTG_SET_3           0x13
+#define TVO_SYNC_AUX_VTG_SET_4           0x14
+#define TVO_SYNC_AUX_VTG_SET_5           0x15
+#define TVO_SYNC_AUX_VTG_SET_6           0x16
+
+#define TVO_SYNC_HD_DCS_SHIFT            8
+
+#define ENCODER_MAIN_CRTC_MASK           BIT(0)
+
+/* enum listing the supported output data format */
+enum sti_tvout_video_out_type {
+	STI_TVOUT_VIDEO_OUT_RGB,
+	STI_TVOUT_VIDEO_OUT_YUV,
+};
+
+struct sti_tvout {
+	struct device *dev;
+	struct drm_device *drm_dev;
+	void __iomem *regs;
+	struct reset_control *reset;
+	struct drm_encoder *hdmi;
+	struct drm_encoder *hda;
+};
+
+struct sti_tvout_encoder {
+	struct drm_encoder encoder;
+	struct sti_tvout *tvout;
+};
+
+#define to_sti_tvout_encoder(x) \
+	container_of(x, struct sti_tvout_encoder, encoder)
+
+#define to_sti_tvout(x) to_sti_tvout_encoder(x)->tvout
+
+/* preformatter conversion matrix */
+static const u32 rgb_to_ycbcr_601[8] = {
+	0xF927082E, 0x04C9FEAB, 0x01D30964, 0xFA95FD3D,
+	0x0000082E, 0x00002000, 0x00002000, 0x00000000
+};
+
+/* 709 RGB to YCbCr */
+static const u32 rgb_to_ycbcr_709[8] = {
+	0xF891082F, 0x0367FF40, 0x01280B71, 0xF9B1FE20,
+	0x0000082F, 0x00002000, 0x00002000, 0x00000000
+};
+
+static u32 tvout_read(struct sti_tvout *tvout, int offset)
+{
+	return readl(tvout->regs + offset);
+}
+
+static void tvout_write(struct sti_tvout *tvout, u32 val, int offset)
+{
+	writel(val, tvout->regs + offset);
+}
+
+/**
+ * Set the clipping mode of a VIP
+ *
+ * @tvout: tvout structure
+ * @cr_r:
+ * @y_g:
+ * @cb_b:
+ */
+static void tvout_vip_set_color_order(struct sti_tvout *tvout,
+				      u32 cr_r, u32 y_g, u32 cb_b)
+{
+	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+
+	val &= ~(TVO_VIP_REORDER_MASK << TVO_VIP_REORDER_R_SHIFT);
+	val &= ~(TVO_VIP_REORDER_MASK << TVO_VIP_REORDER_G_SHIFT);
+	val &= ~(TVO_VIP_REORDER_MASK << TVO_VIP_REORDER_B_SHIFT);
+	val |= cr_r << TVO_VIP_REORDER_R_SHIFT;
+	val |= y_g << TVO_VIP_REORDER_G_SHIFT;
+	val |= cb_b << TVO_VIP_REORDER_B_SHIFT;
+
+	tvout_write(tvout, val, TVO_VIP_HDMI);
+}
+
+/**
+ * Set the clipping mode of a VIP
+ *
+ * @tvout: tvout structure
+ * @range: clipping range
+ */
+static void tvout_vip_set_clip_mode(struct sti_tvout *tvout, u32 range)
+{
+	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+
+	val &= ~(TVO_VIP_CLIP_MASK << TVO_VIP_CLIP_SHIFT);
+	val |= range << TVO_VIP_CLIP_SHIFT;
+	tvout_write(tvout, val, TVO_VIP_HDMI);
+}
+
+/**
+ * Set the rounded value of a VIP
+ *
+ * @tvout: tvout structure
+ * @rnd: rounded val per component
+ */
+static void tvout_vip_set_rnd(struct sti_tvout *tvout, u32 rnd)
+{
+	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+
+	val &= ~(TVO_VIP_RND_MASK << TVO_VIP_RND_SHIFT);
+	val |= rnd << TVO_VIP_RND_SHIFT;
+	tvout_write(tvout, val, TVO_VIP_HDMI);
+}
+
+/**
+ * Select the VIP input
+ *
+ * @tvout: tvout structure
+ * @sel_input: selected_input (main/aux + conv)
+ */
+static void tvout_vip_set_sel_input(struct sti_tvout *tvout,
+				    bool main_path,
+				    bool sel_input_logic_inverted,
+				    enum sti_tvout_video_out_type video_out)
+{
+	u32 sel_input;
+	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+
+	if (main_path)
+		sel_input = TVO_VIP_SEL_INPUT_MAIN;
+	else
+		sel_input = TVO_VIP_SEL_INPUT_AUX;
+
+	switch (video_out) {
+	case STI_TVOUT_VIDEO_OUT_RGB:
+		sel_input |= TVO_VIP_SEL_INPUT_BYPASSED;
+		break;
+	case STI_TVOUT_VIDEO_OUT_YUV:
+		sel_input &= ~TVO_VIP_SEL_INPUT_BYPASSED;
+		break;
+	}
+
+	/* on stih407 chip the sel_input bypass mode logic is inverted */
+	if (sel_input_logic_inverted)
+		sel_input = sel_input ^ TVO_VIP_SEL_INPUT_BYPASS_MASK;
+
+	val &= ~TVO_VIP_SEL_INPUT_MASK;
+	val |= sel_input;
+	tvout_write(tvout, val, TVO_VIP_HDMI);
+}
+
+/**
+ * Select the input video signed or unsigned
+ *
+ * @tvout: tvout structure
+ * @in_vid_signed: used video input format
+ */
+static void tvout_vip_set_in_vid_fmt(struct sti_tvout *tvout, u32 in_vid_fmt)
+{
+	u32 val = tvout_read(tvout, TVO_VIP_HDMI);
+
+	val &= ~TVO_IN_FMT_SIGNED;
+	val |= in_vid_fmt;
+	tvout_write(tvout, val, TVO_MAIN_IN_VID_FORMAT);
+}
+
+/**
+ * Start VIP block for HDMI output
+ *
+ * @tvout: pointer on tvout structure
+ * @main_path: true if main path has to be used in the vip configuration
+ *	  else aux path is used.
+ */
+static void tvout_hdmi_start(struct sti_tvout *tvout, bool main_path)
+{
+	struct device_node *node = tvout->dev->of_node;
+	bool sel_input_logic_inverted = false;
+
+	dev_dbg(tvout->dev, "%s\n", __func__);
+
+	if (main_path) {
+		DRM_DEBUG_DRIVER("main vip for hdmi\n");
+		/* select the input sync for hdmi = VTG set 1 */
+		tvout_write(tvout, TVO_SYNC_MAIN_VTG_SET_1, TVO_HDMI_SYNC_SEL);
+	} else {
+		DRM_DEBUG_DRIVER("aux vip for hdmi\n");
+		/* select the input sync for hdmi = VTG set 1 */
+		tvout_write(tvout, TVO_SYNC_AUX_VTG_SET_1, TVO_HDMI_SYNC_SEL);
+	}
+
+	/* set color channel order */
+	tvout_vip_set_color_order(tvout,
+				  TVO_VIP_REORDER_CR_R_SEL,
+				  TVO_VIP_REORDER_Y_G_SEL,
+				  TVO_VIP_REORDER_CB_B_SEL);
+
+	/* set clipping mode (Limited range RGB/Y) */
+	tvout_vip_set_clip_mode(tvout, TVO_VIP_CLIP_LIMITED_RANGE_RGB_Y);
+
+	/* set round mode (rounded to 8-bit per component) */
+	tvout_vip_set_rnd(tvout, TVO_VIP_RND_8BIT_ROUNDED);
+
+	if (of_device_is_compatible(node, "st,stih407-tvout")) {
+		/* set input video format */
+		tvout_vip_set_in_vid_fmt(tvout->regs + TVO_MAIN_IN_VID_FORMAT,
+					 TVO_IN_FMT_SIGNED);
+		sel_input_logic_inverted = true;
+	}
+
+	/* input selection */
+	tvout_vip_set_sel_input(tvout, main_path,
+			sel_input_logic_inverted, STI_TVOUT_VIDEO_OUT_RGB);
+}
+
+/**
+ * Start HDF VIP and HD DAC
+ *
+ * @tvout: pointer on tvout structure
+ * @main_path: true if main path has to be used in the vip configuration
+ *	  else aux path is used.
+ */
+static void tvout_hda_start(struct sti_tvout *tvout, bool main_path)
+{
+	struct device_node *node = tvout->dev->of_node;
+	bool sel_input_logic_inverted = false;
+
+	dev_dbg(tvout->dev, "%s\n", __func__);
+
+	if (!main_path) {
+		DRM_ERROR("HD Analog on aux not implemented\n");
+		return;
+	}
+
+	DRM_DEBUG_DRIVER("main vip for HDF\n");
+
+	/* set color channel order */
+	tvout_vip_set_color_order(tvout->regs + TVO_VIP_HDF,
+				  TVO_VIP_REORDER_CR_R_SEL,
+				  TVO_VIP_REORDER_Y_G_SEL,
+				  TVO_VIP_REORDER_CB_B_SEL);
+
+	/* set clipping mode (Limited range RGB/Y) */
+	tvout_vip_set_clip_mode(tvout->regs + TVO_VIP_HDF,
+				TVO_VIP_CLIP_LIMITED_RANGE_CB_CR);
+
+	/* set round mode (rounded to 10-bit per component) */
+	tvout_vip_set_rnd(tvout->regs + TVO_VIP_HDF, TVO_VIP_RND_10BIT_ROUNDED);
+
+	if (of_device_is_compatible(node, "st,stih407-tvout")) {
+		/* set input video format */
+		tvout_vip_set_in_vid_fmt(tvout, TVO_IN_FMT_SIGNED);
+		sel_input_logic_inverted = true;
+	}
+
+	/* Input selection */
+	tvout_vip_set_sel_input(tvout->regs + TVO_VIP_HDF,
+				main_path,
+				sel_input_logic_inverted,
+				STI_TVOUT_VIDEO_OUT_YUV);
+
+	/* select the input sync for HD analog = VTG set 3
+	 * and HD DCS = VTG set 2 */
+	tvout_write(tvout,
+		(TVO_SYNC_MAIN_VTG_SET_2 << TVO_SYNC_HD_DCS_SHIFT)
+		| TVO_SYNC_MAIN_VTG_SET_3,
+		TVO_HD_SYNC_SEL);
+
+	/* power up HD DAC */
+	tvout_write(tvout, 0, TVO_HD_DAC_CFG_OFF);
+}
+
+static void sti_tvout_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool sti_tvout_encoder_mode_fixup(struct drm_encoder *encoder,
+				       const struct drm_display_mode *mode,
+				       struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static void sti_tvout_encoder_mode_set(struct drm_encoder *encoder,
+				       struct drm_display_mode *mode,
+				       struct drm_display_mode *adjusted_mode)
+{
+}
+
+static void sti_tvout_encoder_prepare(struct drm_encoder *encoder)
+{
+}
+
+static void sti_tvout_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct sti_tvout_encoder *sti_encoder = to_sti_tvout_encoder(encoder);
+
+	drm_encoder_cleanup(encoder);
+	kfree(sti_encoder);
+}
+
+static const struct drm_encoder_funcs sti_tvout_encoder_funcs = {
+	.destroy = sti_tvout_encoder_destroy,
+};
+
+static void sti_hda_encoder_commit(struct drm_encoder *encoder)
+{
+	struct sti_tvout *tvout = to_sti_tvout(encoder);
+
+	tvout_hda_start(tvout, true);
+}
+
+static void sti_hda_encoder_disable(struct drm_encoder *encoder)
+{
+	struct sti_tvout *tvout = to_sti_tvout(encoder);
+
+	/* reset VIP register */
+	tvout_write(tvout, 0x0, TVO_VIP_HDF);
+
+	/* power down HD DAC */
+	tvout_write(tvout, 1, TVO_HD_DAC_CFG_OFF);
+}
+
+static const struct drm_encoder_helper_funcs sti_hda_encoder_helper_funcs = {
+	.dpms = sti_tvout_encoder_dpms,
+	.mode_fixup = sti_tvout_encoder_mode_fixup,
+	.mode_set = sti_tvout_encoder_mode_set,
+	.prepare = sti_tvout_encoder_prepare,
+	.commit = sti_hda_encoder_commit,
+	.disable = sti_hda_encoder_disable,
+};
+
+static struct drm_encoder *sti_tvout_create_hda_encoder(struct drm_device *dev,
+		struct sti_tvout *tvout)
+{
+	struct sti_tvout_encoder *encoder;
+	struct drm_encoder *drm_encoder;
+
+	encoder = devm_kzalloc(tvout->dev, sizeof(*encoder), GFP_KERNEL);
+	if (!encoder)
+		return NULL;
+
+	encoder->tvout = tvout;
+
+	drm_encoder = (struct drm_encoder *) encoder;
+
+	drm_encoder->possible_crtcs = ENCODER_MAIN_CRTC_MASK;
+	drm_encoder->possible_clones = 1 << 0;
+
+	drm_encoder_init(dev, drm_encoder,
+			&sti_tvout_encoder_funcs, DRM_MODE_ENCODER_DAC);
+
+	drm_encoder_helper_add(drm_encoder, &sti_hda_encoder_helper_funcs);
+
+	return drm_encoder;
+}
+
+static void sti_hdmi_encoder_commit(struct drm_encoder *encoder)
+{
+	struct sti_tvout *tvout = to_sti_tvout(encoder);
+
+	tvout_hdmi_start(tvout, true);
+}
+
+static void sti_hdmi_encoder_disable(struct drm_encoder *encoder)
+{
+	struct sti_tvout *tvout = to_sti_tvout(encoder);
+
+	/* reset VIP register */
+	tvout_write(tvout, 0x0, TVO_VIP_HDMI);
+}
+
+static const struct drm_encoder_helper_funcs sti_hdmi_encoder_helper_funcs = {
+	.dpms = sti_tvout_encoder_dpms,
+	.mode_fixup = sti_tvout_encoder_mode_fixup,
+	.mode_set = sti_tvout_encoder_mode_set,
+	.prepare = sti_tvout_encoder_prepare,
+	.commit = sti_hdmi_encoder_commit,
+	.disable = sti_hdmi_encoder_disable,
+};
+
+static struct drm_encoder *sti_tvout_create_hdmi_encoder(struct drm_device *dev,
+		struct sti_tvout *tvout)
+{
+	struct sti_tvout_encoder *encoder;
+	struct drm_encoder *drm_encoder;
+
+	encoder = devm_kzalloc(tvout->dev, sizeof(*encoder), GFP_KERNEL);
+	if (!encoder)
+		return NULL;
+
+	encoder->tvout = tvout;
+
+	drm_encoder = (struct drm_encoder *) encoder;
+
+	drm_encoder->possible_crtcs = ENCODER_MAIN_CRTC_MASK;
+	drm_encoder->possible_clones = 1 << 1;
+
+	drm_encoder_init(dev, drm_encoder,
+			&sti_tvout_encoder_funcs, DRM_MODE_ENCODER_TMDS);
+
+	drm_encoder_helper_add(drm_encoder, &sti_hdmi_encoder_helper_funcs);
+
+	return drm_encoder;
+}
+
+static void sti_tvout_create_encoders(struct drm_device *dev,
+		struct sti_tvout *tvout)
+{
+	tvout->hdmi = sti_tvout_create_hdmi_encoder(dev, tvout);
+	tvout->hda = sti_tvout_create_hda_encoder(dev, tvout);
+}
+
+static void sti_tvout_destroy_encoders(struct sti_tvout *tvout)
+{
+	if (tvout->hdmi)
+		drm_encoder_cleanup(tvout->hdmi);
+	tvout->hdmi = NULL;
+
+	if (tvout->hda)
+		drm_encoder_cleanup(tvout->hda);
+	tvout->hda = NULL;
+}
+
+static int sti_tvout_bind(struct device *dev, struct device *master, void *data)
+{
+	struct sti_tvout *tvout = dev_get_drvdata(dev);
+	struct drm_device *drm_dev = data;
+	unsigned int i;
+	int ret;
+
+	tvout->drm_dev = drm_dev;
+
+	/* set preformatter matrix */
+	for (i = 0; i < 8; i++) {
+		tvout_write(tvout, rgb_to_ycbcr_601[i],
+			TVO_CSC_MAIN_M0 + (i * 4));
+		tvout_write(tvout, rgb_to_ycbcr_601[i],
+			TVO_CSC_AUX_M0 + (i * 4));
+	}
+
+	sti_tvout_create_encoders(drm_dev, tvout);
+
+	ret = component_bind_all(dev, drm_dev);
+	if (ret)
+		sti_tvout_destroy_encoders(tvout);
+
+	return ret;
+}
+
+static void sti_tvout_unbind(struct device *dev, struct device *master,
+	void *data)
+{
+	/* do nothing */
+}
+
+static const struct component_ops sti_tvout_ops = {
+	.bind	= sti_tvout_bind,
+	.unbind	= sti_tvout_unbind,
+};
+
+static int compare_of(struct device *dev, void *data)
+{
+	return dev->of_node == data;
+}
+
+static int sti_tvout_master_bind(struct device *dev)
+{
+	return 0;
+}
+
+static void sti_tvout_master_unbind(struct device *dev)
+{
+	/* do nothing */
+}
+
+static const struct component_master_ops sti_tvout_master_ops = {
+	.bind = sti_tvout_master_bind,
+	.unbind = sti_tvout_master_unbind,
+};
+
+static int sti_tvout_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *node = dev->of_node;
+	struct sti_tvout *tvout;
+	struct resource *res;
+	struct device_node *child_np;
+	struct component_match *match = NULL;
+
+	DRM_INFO("%s\n", __func__);
+
+	if (!node)
+		return -ENODEV;
+
+	tvout = devm_kzalloc(dev, sizeof(*tvout), GFP_KERNEL);
+	if (!tvout)
+		return -ENOMEM;
+
+	tvout->dev = dev;
+
+	/* get Memory ressources */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "tvout-reg");
+	if (!res) {
+		DRM_ERROR("Invalid glue resource\n");
+		return -ENOMEM;
+	}
+	tvout->regs = devm_ioremap_nocache(dev, res->start, resource_size(res));
+	if (!tvout->regs)
+		return -ENOMEM;
+
+	/* get reset resources */
+	tvout->reset = devm_reset_control_get(dev, "tvout");
+	/* take tvout out of reset */
+	if (!IS_ERR(tvout->reset))
+		reset_control_deassert(tvout->reset);
+
+	platform_set_drvdata(pdev, tvout);
+
+	of_platform_populate(node, NULL, NULL, dev);
+
+	child_np = of_get_next_available_child(node, NULL);
+
+	while (child_np) {
+		component_match_add(dev, &match, compare_of, child_np);
+		of_node_put(child_np);
+		child_np = of_get_next_available_child(node, child_np);
+	}
+
+	component_master_add_with_match(dev, &sti_tvout_master_ops, match);
+
+	return component_add(dev, &sti_tvout_ops);
+}
+
+static int sti_tvout_remove(struct platform_device *pdev)
+{
+	component_master_del(&pdev->dev, &sti_tvout_master_ops);
+	component_del(&pdev->dev, &sti_tvout_ops);
+	return 0;
+}
+
+static const struct of_device_id tvout_of_match[] = {
+	{ .compatible = "st,stih416-tvout", },
+	{ .compatible = "st,stih407-tvout", },
+	{ /* end node */ }
+};
+MODULE_DEVICE_TABLE(of, tvout_of_match);
+
+struct platform_driver sti_tvout_driver = {
+	.driver = {
+		.name = "sti-tvout",
+		.owner = THIS_MODULE,
+		.of_match_table = tvout_of_match,
+	},
+	.probe = sti_tvout_probe,
+	.remove = sti_tvout_remove,
+};
+
+module_platform_driver(sti_tvout_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_vid.c b/drivers/gpu/drm/sti/sti_vid.c
new file mode 100644
index 000000000000..10ced6a479f4
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_vid.c
@@ -0,0 +1,138 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <drm/drmP.h>
+
+#include "sti_layer.h"
+#include "sti_vid.h"
+#include "sti_vtg.h"
+
+/* Registers */
+#define VID_CTL                 0x00
+#define VID_ALP                 0x04
+#define VID_CLF                 0x08
+#define VID_VPO                 0x0C
+#define VID_VPS                 0x10
+#define VID_KEY1                0x28
+#define VID_KEY2                0x2C
+#define VID_MPR0                0x30
+#define VID_MPR1                0x34
+#define VID_MPR2                0x38
+#define VID_MPR3                0x3C
+#define VID_MST                 0x68
+#define VID_BC                  0x70
+#define VID_TINT                0x74
+#define VID_CSAT                0x78
+
+/* Registers values */
+#define VID_CTL_IGNORE          (BIT(31) | BIT(30))
+#define VID_CTL_PSI_ENABLE      (BIT(2) | BIT(1) | BIT(0))
+#define VID_ALP_OPAQUE          0x00000080
+#define VID_BC_DFLT             0x00008000
+#define VID_TINT_DFLT           0x00000000
+#define VID_CSAT_DFLT           0x00000080
+/* YCbCr to RGB BT709:
+ * R = Y+1.5391Cr
+ * G = Y-0.4590Cr-0.1826Cb
+ * B = Y+1.8125Cb */
+#define VID_MPR0_BT709          0x0A800000
+#define VID_MPR1_BT709          0x0AC50000
+#define VID_MPR2_BT709          0x07150545
+#define VID_MPR3_BT709          0x00000AE8
+
+static int sti_vid_prepare_layer(struct sti_layer *vid, bool first_prepare)
+{
+	u32 val;
+
+	/* Unmask */
+	val = readl(vid->regs + VID_CTL);
+	val &= ~VID_CTL_IGNORE;
+	writel(val, vid->regs + VID_CTL);
+
+	return 0;
+}
+
+static int sti_vid_commit_layer(struct sti_layer *vid)
+{
+	struct drm_display_mode *mode = vid->mode;
+	u32 ydo, xdo, yds, xds;
+
+	ydo = sti_vtg_get_line_number(*mode, vid->dst_y);
+	yds = sti_vtg_get_line_number(*mode, vid->dst_y + vid->dst_h - 1);
+	xdo = sti_vtg_get_pixel_number(*mode, vid->dst_x);
+	xds = sti_vtg_get_pixel_number(*mode, vid->dst_x + vid->dst_w - 1);
+
+	writel((ydo << 16) | xdo, vid->regs + VID_VPO);
+	writel((yds << 16) | xds, vid->regs + VID_VPS);
+
+	return 0;
+}
+
+static int sti_vid_disable_layer(struct sti_layer *vid)
+{
+	u32 val;
+
+	/* Mask */
+	val = readl(vid->regs + VID_CTL);
+	val |= VID_CTL_IGNORE;
+	writel(val, vid->regs + VID_CTL);
+
+	return 0;
+}
+
+static const uint32_t *sti_vid_get_formats(struct sti_layer *layer)
+{
+	return NULL;
+}
+
+static unsigned int sti_vid_get_nb_formats(struct sti_layer *layer)
+{
+	return 0;
+}
+
+static void sti_vid_init(struct sti_layer *vid)
+{
+	/* Enable PSI, Mask layer */
+	writel(VID_CTL_PSI_ENABLE | VID_CTL_IGNORE, vid->regs + VID_CTL);
+
+	/* Opaque */
+	writel(VID_ALP_OPAQUE, vid->regs + VID_ALP);
+
+	/* Color conversion parameters */
+	writel(VID_MPR0_BT709, vid->regs + VID_MPR0);
+	writel(VID_MPR1_BT709, vid->regs + VID_MPR1);
+	writel(VID_MPR2_BT709, vid->regs + VID_MPR2);
+	writel(VID_MPR3_BT709, vid->regs + VID_MPR3);
+
+	/* Brightness, contrast, tint, saturation */
+	writel(VID_BC_DFLT, vid->regs + VID_BC);
+	writel(VID_TINT_DFLT, vid->regs + VID_TINT);
+	writel(VID_CSAT_DFLT, vid->regs + VID_CSAT);
+}
+
+static const struct sti_layer_funcs vid_ops = {
+	.get_formats = sti_vid_get_formats,
+	.get_nb_formats = sti_vid_get_nb_formats,
+	.init = sti_vid_init,
+	.prepare = sti_vid_prepare_layer,
+	.commit = sti_vid_commit_layer,
+	.disable = sti_vid_disable_layer,
+};
+
+struct sti_layer *sti_vid_create(struct device *dev)
+{
+	struct sti_layer *vid;
+
+	vid = devm_kzalloc(dev, sizeof(*vid), GFP_KERNEL);
+	if (!vid) {
+		DRM_ERROR("Failed to allocate memory for VID\n");
+		return NULL;
+	}
+
+	vid->ops = &vid_ops;
+
+	return vid;
+}
diff --git a/drivers/gpu/drm/sti/sti_vid.h b/drivers/gpu/drm/sti/sti_vid.h
new file mode 100644
index 000000000000..2c0aecd63294
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_vid.h
@@ -0,0 +1,12 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Fabien Dessenne <fabien.dessenne@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_VID_H_
+#define _STI_VID_H_
+
+struct sti_layer *sti_vid_create(struct device *dev);
+
+#endif
diff --git a/drivers/gpu/drm/sti/sti_vtac.c b/drivers/gpu/drm/sti/sti_vtac.c
new file mode 100644
index 000000000000..97bcdac23ae1
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_vtac.c
@@ -0,0 +1,223 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+
+#include <drm/drmP.h>
+
+/* registers offset */
+#define VTAC_CONFIG                     0x00
+#define VTAC_RX_FIFO_CONFIG             0x04
+#define VTAC_FIFO_CONFIG_VAL            0x04
+
+#define VTAC_SYS_CFG8521                0x824
+#define VTAC_SYS_CFG8522                0x828
+
+/* Number of phyts per pixel */
+#define VTAC_2_5_PPP                    0x0005
+#define VTAC_3_PPP                      0x0006
+#define VTAC_4_PPP                      0x0008
+#define VTAC_5_PPP                      0x000A
+#define VTAC_6_PPP                      0x000C
+#define VTAC_13_PPP                     0x001A
+#define VTAC_14_PPP                     0x001C
+#define VTAC_15_PPP                     0x001E
+#define VTAC_16_PPP                     0x0020
+#define VTAC_17_PPP                     0x0022
+#define VTAC_18_PPP                     0x0024
+
+/* enable bits */
+#define VTAC_ENABLE                     0x3003
+
+#define VTAC_TX_PHY_ENABLE_CLK_PHY      BIT(0)
+#define VTAC_TX_PHY_ENABLE_CLK_DLL      BIT(1)
+#define VTAC_TX_PHY_PLL_NOT_OSC_MODE    BIT(3)
+#define VTAC_TX_PHY_RST_N_DLL_SWITCH    BIT(4)
+#define VTAC_TX_PHY_PROG_N3             BIT(9)
+
+
+/**
+ * VTAC mode structure
+ *
+ * @vid_in_width: Video Data Resolution
+ * @phyts_width: Width of phyt buses(phyt low and phyt high).
+ * @phyts_per_pixel: Number of phyts sent per pixel
+ */
+struct sti_vtac_mode {
+	u32 vid_in_width;
+	u32 phyts_width;
+	u32 phyts_per_pixel;
+};
+
+static const struct sti_vtac_mode vtac_mode_main = {
+	.vid_in_width = 0x2,
+	.phyts_width = 0x2,
+	.phyts_per_pixel = VTAC_5_PPP,
+};
+static const struct sti_vtac_mode vtac_mode_aux = {
+	.vid_in_width = 0x1,
+	.phyts_width = 0x0,
+	.phyts_per_pixel = VTAC_17_PPP,
+};
+
+/**
+ * VTAC structure
+ *
+ * @dev: pointer to device structure
+ * @regs: ioremapped registers for RX and TX devices
+ * @phy_regs: phy registers for TX device
+ * @clk: clock
+ * @mode: main or auxillary configuration mode
+ */
+struct sti_vtac {
+	struct device *dev;
+	void __iomem *regs;
+	void __iomem *phy_regs;
+	struct clk *clk;
+	const struct sti_vtac_mode *mode;
+};
+
+static void sti_vtac_rx_set_config(struct sti_vtac *vtac)
+{
+	u32 config;
+
+	/* Enable VTAC clock */
+	if (clk_prepare_enable(vtac->clk))
+		DRM_ERROR("Failed to prepare/enable vtac_rx clock.\n");
+
+	writel(VTAC_FIFO_CONFIG_VAL, vtac->regs + VTAC_RX_FIFO_CONFIG);
+
+	config = VTAC_ENABLE;
+	config |= vtac->mode->vid_in_width << 4;
+	config |= vtac->mode->phyts_width << 16;
+	config |= vtac->mode->phyts_per_pixel << 23;
+	writel(config, vtac->regs + VTAC_CONFIG);
+}
+
+static void sti_vtac_tx_set_config(struct sti_vtac *vtac)
+{
+	u32 phy_config;
+	u32 config;
+
+	/* Enable VTAC clock */
+	if (clk_prepare_enable(vtac->clk))
+		DRM_ERROR("Failed to prepare/enable vtac_tx clock.\n");
+
+	/* Configure vtac phy */
+	phy_config = 0x00000000;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8522);
+	phy_config = VTAC_TX_PHY_ENABLE_CLK_PHY;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config |= VTAC_TX_PHY_PROG_N3;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config |= VTAC_TX_PHY_ENABLE_CLK_DLL;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config |= VTAC_TX_PHY_RST_N_DLL_SWITCH;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config = readl(vtac->phy_regs + VTAC_SYS_CFG8521);
+	phy_config |= VTAC_TX_PHY_PLL_NOT_OSC_MODE;
+	writel(phy_config, vtac->phy_regs + VTAC_SYS_CFG8521);
+
+	/* Configure vtac tx */
+	config = VTAC_ENABLE;
+	config |= vtac->mode->vid_in_width << 4;
+	config |= vtac->mode->phyts_width << 16;
+	config |= vtac->mode->phyts_per_pixel << 23;
+	writel(config, vtac->regs + VTAC_CONFIG);
+}
+
+static const struct of_device_id vtac_of_match[] = {
+	{
+		.compatible = "st,vtac-main",
+		.data = &vtac_mode_main,
+	}, {
+		.compatible = "st,vtac-aux",
+		.data = &vtac_mode_aux,
+	}, {
+		/* end node */
+	}
+};
+MODULE_DEVICE_TABLE(of, vtac_of_match);
+
+static int sti_vtac_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
+	const struct of_device_id *id;
+	struct sti_vtac *vtac;
+	struct resource *res;
+
+	vtac = devm_kzalloc(dev, sizeof(*vtac), GFP_KERNEL);
+	if (!vtac)
+		return -ENOMEM;
+
+	vtac->dev = dev;
+
+	id = of_match_node(vtac_of_match, np);
+	if (!id)
+		return -ENOMEM;
+
+	vtac->mode = id->data;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		DRM_ERROR("Invalid resource\n");
+		return -ENOMEM;
+	}
+	vtac->regs = devm_ioremap_resource(dev, res);
+	if (IS_ERR(vtac->regs))
+		return PTR_ERR(vtac->regs);
+
+
+	vtac->clk = devm_clk_get(dev, "vtac");
+	if (IS_ERR(vtac->clk)) {
+		DRM_ERROR("Cannot get vtac clock\n");
+		return PTR_ERR(vtac->clk);
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (res) {
+		vtac->phy_regs = devm_ioremap_nocache(dev, res->start,
+						 resource_size(res));
+		sti_vtac_tx_set_config(vtac);
+	} else {
+
+		sti_vtac_rx_set_config(vtac);
+	}
+
+	platform_set_drvdata(pdev, vtac);
+	DRM_INFO("%s %s\n", __func__, dev_name(vtac->dev));
+
+	return 0;
+}
+
+static int sti_vtac_remove(struct platform_device *pdev)
+{
+	return 0;
+}
+
+struct platform_driver sti_vtac_driver = {
+	.driver = {
+		.name = "sti-vtac",
+		.owner = THIS_MODULE,
+		.of_match_table = vtac_of_match,
+	},
+	.probe = sti_vtac_probe,
+	.remove = sti_vtac_remove,
+};
+
+module_platform_driver(sti_vtac_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c
new file mode 100644
index 000000000000..740d6e347a62
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_vtg.c
@@ -0,0 +1,366 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Authors: Benjamin Gaignard <benjamin.gaignard@st.com>
+ *          Fabien Dessenne <fabien.dessenne@st.com>
+ *          Vincent Abriou <vincent.abriou@st.com>
+ *          for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#include <linux/module.h>
+#include <linux/notifier.h>
+#include <linux/platform_device.h>
+
+#include <drm/drmP.h>
+
+#include "sti_vtg.h"
+
+#define VTG_TYPE_MASTER         0
+#define VTG_TYPE_SLAVE_BY_EXT0  1
+
+/* registers offset */
+#define VTG_MODE            0x0000
+#define VTG_CLKLN           0x0008
+#define VTG_HLFLN           0x000C
+#define VTG_DRST_AUTOC      0x0010
+#define VTG_VID_TFO         0x0040
+#define VTG_VID_TFS         0x0044
+#define VTG_VID_BFO         0x0048
+#define VTG_VID_BFS         0x004C
+
+#define VTG_HOST_ITS        0x0078
+#define VTG_HOST_ITS_BCLR   0x007C
+#define VTG_HOST_ITM_BCLR   0x0088
+#define VTG_HOST_ITM_BSET   0x008C
+
+#define VTG_H_HD_1          0x00C0
+#define VTG_TOP_V_VD_1      0x00C4
+#define VTG_BOT_V_VD_1      0x00C8
+#define VTG_TOP_V_HD_1      0x00CC
+#define VTG_BOT_V_HD_1      0x00D0
+
+#define VTG_H_HD_2          0x00E0
+#define VTG_TOP_V_VD_2      0x00E4
+#define VTG_BOT_V_VD_2      0x00E8
+#define VTG_TOP_V_HD_2      0x00EC
+#define VTG_BOT_V_HD_2      0x00F0
+
+#define VTG_H_HD_3          0x0100
+#define VTG_TOP_V_VD_3      0x0104
+#define VTG_BOT_V_VD_3      0x0108
+#define VTG_TOP_V_HD_3      0x010C
+#define VTG_BOT_V_HD_3      0x0110
+
+#define VTG_IRQ_BOTTOM      BIT(0)
+#define VTG_IRQ_TOP         BIT(1)
+#define VTG_IRQ_MASK        (VTG_IRQ_TOP | VTG_IRQ_BOTTOM)
+
+/* delay introduced by the Arbitrary Waveform Generator in nb of pixels */
+#define AWG_DELAY_HD        (-9)
+#define AWG_DELAY_ED        (-8)
+#define AWG_DELAY_SD        (-7)
+
+LIST_HEAD(vtg_lookup);
+
+/**
+ * STI VTG structure
+ *
+ * @dev: pointer to device driver
+ * @data: data associated to the device
+ * @irq: VTG irq
+ * @type: VTG type (main or aux)
+ * @notifier_list: notifier callback
+ * @crtc_id: the crtc id for vblank event
+ * @slave: slave vtg
+ * @link: List node to link the structure in lookup list
+ */
+struct sti_vtg {
+	struct device *dev;
+	struct device_node *np;
+	void __iomem *regs;
+	int irq;
+	u32 irq_status;
+	struct raw_notifier_head notifier_list;
+	int crtc_id;
+	struct sti_vtg *slave;
+	struct list_head link;
+};
+
+static void vtg_register(struct sti_vtg *vtg)
+{
+	list_add_tail(&vtg->link, &vtg_lookup);
+}
+
+struct sti_vtg *of_vtg_find(struct device_node *np)
+{
+	struct sti_vtg *vtg;
+
+	list_for_each_entry(vtg, &vtg_lookup, link) {
+		if (vtg->np == np)
+			return vtg;
+	}
+	return NULL;
+}
+EXPORT_SYMBOL(of_vtg_find);
+
+static void vtg_reset(struct sti_vtg *vtg)
+{
+	/* reset slave and then master */
+	if (vtg->slave)
+		vtg_reset(vtg->slave);
+
+	writel(1, vtg->regs + VTG_DRST_AUTOC);
+}
+
+static void vtg_set_mode(struct sti_vtg *vtg,
+			 int type, const struct drm_display_mode *mode)
+{
+	u32 tmp;
+
+	if (vtg->slave)
+		vtg_set_mode(vtg->slave, VTG_TYPE_SLAVE_BY_EXT0, mode);
+
+	writel(mode->htotal, vtg->regs + VTG_CLKLN);
+	writel(mode->vtotal * 2, vtg->regs + VTG_HLFLN);
+
+	tmp = (mode->vtotal - mode->vsync_start + 1) << 16;
+	tmp |= mode->htotal - mode->hsync_start;
+	writel(tmp, vtg->regs + VTG_VID_TFO);
+	writel(tmp, vtg->regs + VTG_VID_BFO);
+
+	tmp = (mode->vdisplay + mode->vtotal - mode->vsync_start + 1) << 16;
+	tmp |= mode->hdisplay + mode->htotal - mode->hsync_start;
+	writel(tmp, vtg->regs + VTG_VID_TFS);
+	writel(tmp, vtg->regs + VTG_VID_BFS);
+
+	/* prepare VTG set 1 and 2 for HDMI and VTG set 3 for HD DAC */
+	tmp = (mode->hsync_end - mode->hsync_start) << 16;
+	writel(tmp, vtg->regs + VTG_H_HD_1);
+	writel(tmp, vtg->regs + VTG_H_HD_2);
+
+	tmp = (mode->vsync_end - mode->vsync_start + 1) << 16;
+	tmp |= 1;
+	writel(tmp, vtg->regs + VTG_TOP_V_VD_1);
+	writel(tmp, vtg->regs + VTG_BOT_V_VD_1);
+	writel(0, vtg->regs + VTG_TOP_V_HD_1);
+	writel(0, vtg->regs + VTG_BOT_V_HD_1);
+
+	/* prepare VTG set 2 for for HD DCS */
+	writel(tmp, vtg->regs + VTG_TOP_V_VD_2);
+	writel(tmp, vtg->regs + VTG_BOT_V_VD_2);
+	writel(0, vtg->regs + VTG_TOP_V_HD_2);
+	writel(0, vtg->regs + VTG_BOT_V_HD_2);
+
+	/* prepare VTG set 3 for HD Analog in HD mode */
+	tmp = (mode->hsync_end - mode->hsync_start + AWG_DELAY_HD) << 16;
+	tmp |= mode->htotal + AWG_DELAY_HD;
+	writel(tmp, vtg->regs + VTG_H_HD_3);
+
+	tmp = (mode->vsync_end - mode->vsync_start) << 16;
+	tmp |= mode->vtotal;
+	writel(tmp, vtg->regs + VTG_TOP_V_VD_3);
+	writel(tmp, vtg->regs + VTG_BOT_V_VD_3);
+
+	tmp = (mode->htotal + AWG_DELAY_HD) << 16;
+	tmp |= mode->htotal + AWG_DELAY_HD;
+	writel(tmp, vtg->regs + VTG_TOP_V_HD_3);
+	writel(tmp, vtg->regs + VTG_BOT_V_HD_3);
+
+	/* mode */
+	writel(type, vtg->regs + VTG_MODE);
+}
+
+static void vtg_enable_irq(struct sti_vtg *vtg)
+{
+	/* clear interrupt status and mask */
+	writel(0xFFFF, vtg->regs + VTG_HOST_ITS_BCLR);
+	writel(0xFFFF, vtg->regs + VTG_HOST_ITM_BCLR);
+	writel(VTG_IRQ_MASK, vtg->regs + VTG_HOST_ITM_BSET);
+}
+
+void sti_vtg_set_config(struct sti_vtg *vtg,
+		const struct drm_display_mode *mode)
+{
+	/* write configuration */
+	vtg_set_mode(vtg, VTG_TYPE_MASTER, mode);
+
+	vtg_reset(vtg);
+
+	/* enable irq for the vtg vblank synchro */
+	if (vtg->slave)
+		vtg_enable_irq(vtg->slave);
+	else
+		vtg_enable_irq(vtg);
+}
+EXPORT_SYMBOL(sti_vtg_set_config);
+
+/**
+ * sti_vtg_get_line_number
+ *
+ * @mode: display mode to be used
+ * @y:    line
+ *
+ * Return the line number according to the display mode taking
+ * into account the Sync and Back Porch information.
+ * Video frame line numbers start at 1, y starts at 0.
+ * In interlaced modes the start line is the field line number of the odd
+ * field, but y is still defined as a progressive frame.
+ */
+u32 sti_vtg_get_line_number(struct drm_display_mode mode, int y)
+{
+	u32 start_line = mode.vtotal - mode.vsync_start + 1;
+
+	if (mode.flags & DRM_MODE_FLAG_INTERLACE)
+		start_line *= 2;
+
+	return start_line + y;
+}
+EXPORT_SYMBOL(sti_vtg_get_line_number);
+
+/**
+ * sti_vtg_get_pixel_number
+ *
+ * @mode: display mode to be used
+ * @x:    row
+ *
+ * Return the pixel number according to the display mode taking
+ * into account the Sync and Back Porch information.
+ * Pixels are counted from 0.
+ */
+u32 sti_vtg_get_pixel_number(struct drm_display_mode mode, int x)
+{
+	return mode.htotal - mode.hsync_start + x;
+}
+EXPORT_SYMBOL(sti_vtg_get_pixel_number);
+
+int sti_vtg_register_client(struct sti_vtg *vtg,
+		struct notifier_block *nb, int crtc_id)
+{
+	if (vtg->slave)
+		return sti_vtg_register_client(vtg->slave, nb, crtc_id);
+
+	vtg->crtc_id = crtc_id;
+	return raw_notifier_chain_register(&vtg->notifier_list, nb);
+}
+EXPORT_SYMBOL(sti_vtg_register_client);
+
+int sti_vtg_unregister_client(struct sti_vtg *vtg, struct notifier_block *nb)
+{
+	if (vtg->slave)
+		return sti_vtg_unregister_client(vtg->slave, nb);
+
+	return raw_notifier_chain_unregister(&vtg->notifier_list, nb);
+}
+EXPORT_SYMBOL(sti_vtg_unregister_client);
+
+static irqreturn_t vtg_irq_thread(int irq, void *arg)
+{
+	struct sti_vtg *vtg = arg;
+	u32 event;
+
+	event = (vtg->irq_status & VTG_IRQ_TOP) ?
+		VTG_TOP_FIELD_EVENT : VTG_BOTTOM_FIELD_EVENT;
+
+	raw_notifier_call_chain(&vtg->notifier_list, event, &vtg->crtc_id);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t vtg_irq(int irq, void *arg)
+{
+	struct sti_vtg *vtg = arg;
+
+	vtg->irq_status = readl(vtg->regs + VTG_HOST_ITS);
+
+	writel(vtg->irq_status, vtg->regs + VTG_HOST_ITS_BCLR);
+
+	/* force sync bus write */
+	readl(vtg->regs + VTG_HOST_ITS);
+
+	return IRQ_WAKE_THREAD;
+}
+
+static int vtg_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *np;
+	struct sti_vtg *vtg;
+	struct resource *res;
+	char irq_name[32];
+	int ret;
+
+	vtg = devm_kzalloc(dev, sizeof(*vtg), GFP_KERNEL);
+	if (!vtg)
+		return -ENOMEM;
+
+	vtg->dev = dev;
+	vtg->np = pdev->dev.of_node;
+
+	/* Get Memory ressources */
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		DRM_ERROR("Get memory resource failed\n");
+		return -ENOMEM;
+	}
+	vtg->regs = devm_ioremap_nocache(dev, res->start, resource_size(res));
+
+	np = of_parse_phandle(pdev->dev.of_node, "st,slave", 0);
+	if (np) {
+		vtg->slave = of_vtg_find(np);
+
+		if (!vtg->slave)
+			return -EPROBE_DEFER;
+	} else {
+		vtg->irq = platform_get_irq(pdev, 0);
+		if (IS_ERR_VALUE(vtg->irq)) {
+			DRM_ERROR("Failed to get VTG interrupt\n");
+			return vtg->irq;
+		}
+
+		snprintf(irq_name, sizeof(irq_name), "vsync-%s",
+				dev_name(vtg->dev));
+
+		RAW_INIT_NOTIFIER_HEAD(&vtg->notifier_list);
+
+		ret = devm_request_threaded_irq(dev, vtg->irq, vtg_irq,
+				vtg_irq_thread, IRQF_ONESHOT, irq_name, vtg);
+		if (IS_ERR_VALUE(ret)) {
+			DRM_ERROR("Failed to register VTG interrupt\n");
+			return ret;
+		}
+	}
+
+	vtg_register(vtg);
+	platform_set_drvdata(pdev, vtg);
+
+	DRM_INFO("%s %s\n", __func__, dev_name(vtg->dev));
+
+	return 0;
+}
+
+static int vtg_remove(struct platform_device *pdev)
+{
+	return 0;
+}
+
+static const struct of_device_id vtg_of_match[] = {
+	{ .compatible = "st,vtg", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, vtg_of_match);
+
+struct platform_driver sti_vtg_driver = {
+	.driver = {
+		.name = "sti-vtg",
+		.owner = THIS_MODULE,
+		.of_match_table = vtg_of_match,
+	},
+	.probe	= vtg_probe,
+	.remove = vtg_remove,
+};
+
+module_platform_driver(sti_vtg_driver);
+
+MODULE_AUTHOR("Benjamin Gaignard <benjamin.gaignard@st.com>");
+MODULE_DESCRIPTION("STMicroelectronics SoC DRM driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/sti/sti_vtg.h b/drivers/gpu/drm/sti/sti_vtg.h
new file mode 100644
index 000000000000..e84d23f1f57f
--- /dev/null
+++ b/drivers/gpu/drm/sti/sti_vtg.h
@@ -0,0 +1,28 @@
+/*
+ * Copyright (C) STMicroelectronics SA 2014
+ * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics.
+ * License terms:  GNU General Public License (GPL), version 2
+ */
+
+#ifndef _STI_VTG_H_
+#define _STI_VTG_H_
+
+#define VTG_TOP_FIELD_EVENT     1
+#define VTG_BOTTOM_FIELD_EVENT  2
+
+struct sti_vtg;
+struct drm_display_mode;
+struct notifier_block;
+
+struct sti_vtg *of_vtg_find(struct device_node *np);
+void sti_vtg_set_config(struct sti_vtg *vtg,
+		const struct drm_display_mode *mode);
+int sti_vtg_register_client(struct sti_vtg *vtg,
+		struct notifier_block *nb, int crtc_id);
+int sti_vtg_unregister_client(struct sti_vtg *vtg,
+		struct notifier_block *nb);
+
+u32 sti_vtg_get_line_number(struct drm_display_mode mode, int y);
+u32 sti_vtg_get_pixel_number(struct drm_display_mode mode, int x);
+
+#endif
diff --git a/drivers/gpu/drm/tdfx/tdfx_drv.c b/drivers/gpu/drm/tdfx/tdfx_drv.c
index 3492ca5c46d3..fab5ebcb0fef 100644
--- a/drivers/gpu/drm/tdfx/tdfx_drv.c
+++ b/drivers/gpu/drm/tdfx/tdfx_drv.c
@@ -36,6 +36,7 @@
 #include "tdfx_drv.h"
 
 #include <drm/drm_pciids.h>
+#include <drm/drm_legacy.h>
 
 static struct pci_device_id pciidlist[] = {
 	tdfx_PCI_IDS
@@ -46,7 +47,7 @@ static const struct file_operations tdfx_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -55,6 +56,7 @@ static const struct file_operations tdfx_driver_fops = {
 };
 
 static struct drm_driver driver = {
+	.set_busid = drm_pci_set_busid,
 	.fops = &tdfx_driver_fops,
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
index ef40381f3909..054a79f143ae 100644
--- a/drivers/gpu/drm/tegra/dc.c
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -18,6 +18,8 @@
 struct tegra_dc_soc_info {
 	bool supports_interlacing;
 	bool supports_cursor;
+	bool supports_block_linear;
+	unsigned int pitch_align;
 };
 
 struct tegra_plane {
@@ -212,15 +214,44 @@ static int tegra_dc_setup_window(struct tegra_dc *dc, unsigned int index,
 	tegra_dc_writel(dc, h_offset, DC_WINBUF_ADDR_H_OFFSET);
 	tegra_dc_writel(dc, v_offset, DC_WINBUF_ADDR_V_OFFSET);
 
-	if (window->tiled) {
-		value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
-			DC_WIN_BUFFER_ADDR_MODE_TILE;
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = window->tiling.value;
+
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_dc_writel(dc, value, DC_WINBUF_SURFACE_KIND);
 	} else {
-		value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
-			DC_WIN_BUFFER_ADDR_MODE_LINEAR;
-	}
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
+				DC_WIN_BUFFER_ADDR_MODE_LINEAR;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
+				DC_WIN_BUFFER_ADDR_MODE_TILE;
+			break;
 
-	tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			DRM_ERROR("hardware doesn't support block linear mode\n");
+			return -EINVAL;
+		}
+
+		tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+	}
 
 	value = WIN_ENABLE;
 
@@ -288,6 +319,7 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	struct tegra_dc *dc = to_tegra_dc(crtc);
 	struct tegra_dc_window window;
 	unsigned int i;
+	int err;
 
 	memset(&window, 0, sizeof(window));
 	window.src.x = src_x >> 16;
@@ -301,7 +333,10 @@ static int tegra_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
 	window.format = tegra_dc_format(fb->pixel_format, &window.swap);
 	window.bits_per_pixel = fb->bits_per_pixel;
 	window.bottom_up = tegra_fb_is_bottom_up(fb);
-	window.tiled = tegra_fb_is_tiled(fb);
+
+	err = tegra_fb_get_tiling(fb, &window.tiling);
+	if (err < 0)
+		return err;
 
 	for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) {
 		struct tegra_bo *bo = tegra_fb_get_plane(fb, i);
@@ -402,8 +437,14 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 {
 	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned int h_offset = 0, v_offset = 0;
+	struct tegra_bo_tiling tiling;
 	unsigned int format, swap;
 	unsigned long value;
+	int err;
+
+	err = tegra_fb_get_tiling(fb, &tiling);
+	if (err < 0)
+		return err;
 
 	tegra_dc_writel(dc, WINDOW_A_SELECT, DC_CMD_DISPLAY_WINDOW_HEADER);
 
@@ -417,15 +458,44 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 	tegra_dc_writel(dc, format, DC_WIN_COLOR_DEPTH);
 	tegra_dc_writel(dc, swap, DC_WIN_BYTE_SWAP);
 
-	if (tegra_fb_is_tiled(fb)) {
-		value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
-			DC_WIN_BUFFER_ADDR_MODE_TILE;
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = tiling.value;
+
+		switch (tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_dc_writel(dc, value, DC_WINBUF_SURFACE_KIND);
 	} else {
-		value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
-			DC_WIN_BUFFER_ADDR_MODE_LINEAR;
-	}
+		switch (tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
+				DC_WIN_BUFFER_ADDR_MODE_LINEAR;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
+				DC_WIN_BUFFER_ADDR_MODE_TILE;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			DRM_ERROR("hardware doesn't support block linear mode\n");
+			return -EINVAL;
+		}
 
-	tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+		tegra_dc_writel(dc, value, DC_WIN_BUFFER_ADDR_MODE);
+	}
 
 	/* make sure bottom-up buffers are properly displayed */
 	if (tegra_fb_is_bottom_up(fb)) {
@@ -666,7 +736,6 @@ static const struct drm_crtc_funcs tegra_crtc_funcs = {
 
 static void tegra_crtc_disable(struct drm_crtc *crtc)
 {
-	struct tegra_dc *dc = to_tegra_dc(crtc);
 	struct drm_device *drm = crtc->dev;
 	struct drm_plane *plane;
 
@@ -682,7 +751,7 @@ static void tegra_crtc_disable(struct drm_crtc *crtc)
 		}
 	}
 
-	drm_vblank_off(drm, dc->pipe);
+	drm_crtc_vblank_off(crtc);
 }
 
 static bool tegra_crtc_mode_fixup(struct drm_crtc *crtc,
@@ -771,8 +840,6 @@ static int tegra_crtc_mode_set(struct drm_crtc *crtc,
 	u32 value;
 	int err;
 
-	drm_vblank_pre_modeset(crtc->dev, dc->pipe);
-
 	err = tegra_crtc_setup_clk(crtc, mode);
 	if (err) {
 		dev_err(dc->dev, "failed to setup clock for CRTC: %d\n", err);
@@ -826,6 +893,8 @@ static void tegra_crtc_prepare(struct drm_crtc *crtc)
 	unsigned int syncpt;
 	unsigned long value;
 
+	drm_crtc_vblank_off(crtc);
+
 	/* hardware initialization */
 	reset_control_deassert(dc->rst);
 	usleep_range(10000, 20000);
@@ -873,7 +942,7 @@ static void tegra_crtc_commit(struct drm_crtc *crtc)
 	value = GENERAL_ACT_REQ | WIN_A_ACT_REQ;
 	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
 
-	drm_vblank_post_modeset(crtc->dev, dc->pipe);
+	drm_crtc_vblank_on(crtc);
 }
 
 static void tegra_crtc_load_lut(struct drm_crtc *crtc)
@@ -1214,12 +1283,20 @@ static int tegra_dc_init(struct host1x_client *client)
 {
 	struct drm_device *drm = dev_get_drvdata(client->parent);
 	struct tegra_dc *dc = host1x_client_to_dc(client);
+	struct tegra_drm *tegra = drm->dev_private;
 	int err;
 
 	drm_crtc_init(drm, &dc->base, &tegra_crtc_funcs);
 	drm_mode_crtc_set_gamma_size(&dc->base, 256);
 	drm_crtc_helper_add(&dc->base, &tegra_crtc_helper_funcs);
 
+	/*
+	 * Keep track of the minimum pitch alignment across all display
+	 * controllers.
+	 */
+	if (dc->soc->pitch_align > tegra->pitch_align)
+		tegra->pitch_align = dc->soc->pitch_align;
+
 	err = tegra_dc_rgb_init(drm, dc);
 	if (err < 0 && err != -ENODEV) {
 		dev_err(dc->dev, "failed to initialize RGB output: %d\n", err);
@@ -1277,16 +1354,29 @@ static const struct host1x_client_ops dc_client_ops = {
 static const struct tegra_dc_soc_info tegra20_dc_soc_info = {
 	.supports_interlacing = false,
 	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 8,
 };
 
 static const struct tegra_dc_soc_info tegra30_dc_soc_info = {
 	.supports_interlacing = false,
 	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 8,
+};
+
+static const struct tegra_dc_soc_info tegra114_dc_soc_info = {
+	.supports_interlacing = false,
+	.supports_cursor = false,
+	.supports_block_linear = false,
+	.pitch_align = 64,
 };
 
 static const struct tegra_dc_soc_info tegra124_dc_soc_info = {
 	.supports_interlacing = true,
 	.supports_cursor = true,
+	.supports_block_linear = true,
+	.pitch_align = 64,
 };
 
 static const struct of_device_id tegra_dc_of_match[] = {
@@ -1303,6 +1393,7 @@ static const struct of_device_id tegra_dc_of_match[] = {
 		/* sentinel */
 	}
 };
+MODULE_DEVICE_TABLE(of, tegra_dc_of_match);
 
 static int tegra_dc_parse_dt(struct tegra_dc *dc)
 {
@@ -1430,6 +1521,7 @@ static int tegra_dc_remove(struct platform_device *pdev)
 		return err;
 	}
 
+	reset_control_assert(dc->rst);
 	clk_disable_unprepare(dc->clk);
 
 	return 0;
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h
index 78c5feff95d2..705c93b00794 100644
--- a/drivers/gpu/drm/tegra/dc.h
+++ b/drivers/gpu/drm/tegra/dc.h
@@ -428,6 +428,11 @@
 #define DC_WINBUF_ADDR_V_OFFSET_NS		0x809
 
 #define DC_WINBUF_UFLOW_STATUS			0x80a
+#define DC_WINBUF_SURFACE_KIND			0x80b
+#define DC_WINBUF_SURFACE_KIND_PITCH	(0 << 0)
+#define DC_WINBUF_SURFACE_KIND_TILED	(1 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK	(2 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(x) (((x) & 0x7) << 4)
 
 #define DC_WINBUF_AD_UFLOW_STATUS		0xbca
 #define DC_WINBUF_BD_UFLOW_STATUS		0xdca
diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c
index 3f132e356e9c..d6b55e3e3716 100644
--- a/drivers/gpu/drm/tegra/dpaux.c
+++ b/drivers/gpu/drm/tegra/dpaux.c
@@ -382,6 +382,7 @@ static const struct of_device_id tegra_dpaux_of_match[] = {
 	{ .compatible = "nvidia,tegra124-dpaux", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_dpaux_of_match);
 
 struct platform_driver tegra_dpaux_driver = {
 	.driver = {
@@ -532,9 +533,9 @@ int tegra_dpaux_train(struct tegra_dpaux *dpaux, struct drm_dp_link *link,
 
 	for (i = 0; i < link->num_lanes; i++)
 		values[i] = DP_TRAIN_MAX_PRE_EMPHASIS_REACHED |
-			    DP_TRAIN_PRE_EMPHASIS_0 |
+			    DP_TRAIN_PRE_EMPH_LEVEL_0 |
 			    DP_TRAIN_MAX_SWING_REACHED |
-			    DP_TRAIN_VOLTAGE_SWING_400;
+			    DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
 
 	err = drm_dp_dpcd_write(&dpaux->aux, DP_TRAINING_LANE0_SET, values,
 				link->num_lanes);
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
index 3396f9f6a9f7..59736bb810cd 100644
--- a/drivers/gpu/drm/tegra/drm.c
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -40,6 +40,12 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 
 	drm_mode_config_init(drm);
 
+	err = tegra_drm_fb_prepare(drm);
+	if (err < 0)
+		return err;
+
+	drm_kms_helper_poll_init(drm);
+
 	err = host1x_device_init(device);
 	if (err < 0)
 		return err;
@@ -59,8 +65,6 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 	if (err < 0)
 		return err;
 
-	drm_kms_helper_poll_init(drm);
-
 	return 0;
 }
 
@@ -128,6 +132,45 @@ host1x_bo_lookup(struct drm_device *drm, struct drm_file *file, u32 handle)
 	return &bo->base;
 }
 
+static int host1x_reloc_copy_from_user(struct host1x_reloc *dest,
+				       struct drm_tegra_reloc __user *src,
+				       struct drm_device *drm,
+				       struct drm_file *file)
+{
+	u32 cmdbuf, target;
+	int err;
+
+	err = get_user(cmdbuf, &src->cmdbuf.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->cmdbuf.offset, &src->cmdbuf.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(target, &src->target.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->target.offset, &src->cmdbuf.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->shift, &src->shift);
+	if (err < 0)
+		return err;
+
+	dest->cmdbuf.bo = host1x_bo_lookup(drm, file, cmdbuf);
+	if (!dest->cmdbuf.bo)
+		return -ENOENT;
+
+	dest->target.bo = host1x_bo_lookup(drm, file, target);
+	if (!dest->target.bo)
+		return -ENOENT;
+
+	return 0;
+}
+
 int tegra_drm_submit(struct tegra_drm_context *context,
 		     struct drm_tegra_submit *args, struct drm_device *drm,
 		     struct drm_file *file)
@@ -180,26 +223,13 @@ int tegra_drm_submit(struct tegra_drm_context *context,
 		cmdbufs++;
 	}
 
-	if (copy_from_user(job->relocarray, relocs,
-			   sizeof(*relocs) * num_relocs)) {
-		err = -EFAULT;
-		goto fail;
-	}
-
+	/* copy and resolve relocations from submit */
 	while (num_relocs--) {
-		struct host1x_reloc *reloc = &job->relocarray[num_relocs];
-		struct host1x_bo *cmdbuf, *target;
-
-		cmdbuf = host1x_bo_lookup(drm, file, (u32)reloc->cmdbuf);
-		target = host1x_bo_lookup(drm, file, (u32)reloc->target);
-
-		reloc->cmdbuf = cmdbuf;
-		reloc->target = target;
-
-		if (!reloc->target || !reloc->cmdbuf) {
-			err = -ENOENT;
+		err = host1x_reloc_copy_from_user(&job->relocarray[num_relocs],
+						  &relocs[num_relocs], drm,
+						  file);
+		if (err < 0)
 			goto fail;
-		}
 	}
 
 	if (copy_from_user(job->waitchk, waitchks,
@@ -451,11 +481,151 @@ static int tegra_get_syncpt_base(struct drm_device *drm, void *data,
 
 	return 0;
 }
+
+static int tegra_gem_set_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_set_tiling *args = data;
+	enum tegra_bo_tiling_mode mode;
+	struct drm_gem_object *gem;
+	unsigned long value = 0;
+	struct tegra_bo *bo;
+
+	switch (args->mode) {
+	case DRM_TEGRA_GEM_TILING_MODE_PITCH:
+		mode = TEGRA_BO_TILING_MODE_PITCH;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_TILED:
+		mode = TEGRA_BO_TILING_MODE_TILED;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_BLOCK:
+		mode = TEGRA_BO_TILING_MODE_BLOCK;
+
+		if (args->value > 5)
+			return -EINVAL;
+
+		value = args->value;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	bo->tiling.mode = mode;
+	bo->tiling.value = value;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_get_tiling *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+	int err = 0;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	switch (bo->tiling.mode) {
+	case TEGRA_BO_TILING_MODE_PITCH:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_PITCH;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_TILED:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_TILED;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_BLOCK:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_BLOCK;
+		args->value = bo->tiling.value;
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	drm_gem_object_unreference(gem);
+
+	return err;
+}
+
+static int tegra_gem_set_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_set_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	if (args->flags & ~DRM_TEGRA_GEM_FLAGS)
+		return -EINVAL;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	bo->flags = 0;
+
+	if (args->flags & DRM_TEGRA_GEM_BOTTOM_UP)
+		bo->flags |= TEGRA_BO_BOTTOM_UP;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_get_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(drm, file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	args->flags = 0;
+
+	if (bo->flags & TEGRA_BO_BOTTOM_UP)
+		args->flags |= DRM_TEGRA_GEM_BOTTOM_UP;
+
+	drm_gem_object_unreference(gem);
+
+	return 0;
+}
 #endif
 
 static const struct drm_ioctl_desc tegra_drm_ioctls[] = {
 #ifdef CONFIG_DRM_TEGRA_STAGING
-	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_UNLOCKED | DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_READ, tegra_syncpt_read, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_INCR, tegra_syncpt_incr, DRM_UNLOCKED),
@@ -465,6 +635,10 @@ static const struct drm_ioctl_desc tegra_drm_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT, tegra_get_syncpt, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_SUBMIT, tegra_submit, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT_BASE, tegra_get_syncpt_base, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_TILING, tegra_gem_set_tiling, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_TILING, tegra_gem_get_tiling, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_FLAGS, tegra_gem_set_flags, DRM_UNLOCKED),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_FLAGS, tegra_gem_get_flags, DRM_UNLOCKED),
 #endif
 };
 
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
index 6b8fe9d86ed4..e89c70fa82d5 100644
--- a/drivers/gpu/drm/tegra/drm.h
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -19,6 +19,8 @@
 #include <drm/drm_fb_helper.h>
 #include <drm/drm_fixed.h>
 
+#include "gem.h"
+
 struct reset_control;
 
 struct tegra_fb {
@@ -43,6 +45,8 @@ struct tegra_drm {
 #ifdef CONFIG_DRM_TEGRA_FBDEV
 	struct tegra_fbdev *fbdev;
 #endif
+
+	unsigned int pitch_align;
 };
 
 struct tegra_drm_client;
@@ -160,7 +164,8 @@ struct tegra_dc_window {
 	unsigned int stride[2];
 	unsigned long base[3];
 	bool bottom_up;
-	bool tiled;
+
+	struct tegra_bo_tiling tiling;
 };
 
 /* from dc.c */
@@ -279,7 +284,9 @@ int tegra_dpaux_train(struct tegra_dpaux *dpaux, struct drm_dp_link *link,
 struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
 				    unsigned int index);
 bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer);
-bool tegra_fb_is_tiled(struct drm_framebuffer *framebuffer);
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling);
+int tegra_drm_fb_prepare(struct drm_device *drm);
 int tegra_drm_fb_init(struct drm_device *drm);
 void tegra_drm_fb_exit(struct drm_device *drm);
 #ifdef CONFIG_DRM_TEGRA_FBDEV
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
index bd56f2affa78..f7874458926a 100644
--- a/drivers/gpu/drm/tegra/dsi.c
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -474,7 +474,8 @@ static int tegra_output_dsi_enable(struct tegra_output *output)
 	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
 
 	value = tegra_dsi_readl(dsi, DSI_CONTROL);
-	value |= DSI_CONTROL_HS_CLK_CTRL;
+	if (dsi->flags & MIPI_DSI_CLOCK_NON_CONTINUOUS)
+		value |= DSI_CONTROL_HS_CLK_CTRL;
 	value &= ~DSI_CONTROL_TX_TRIG(3);
 	value &= ~DSI_CONTROL_DCS_ENABLE;
 	value |= DSI_CONTROL_VIDEO_ENABLE;
@@ -982,6 +983,7 @@ static const struct of_device_id tegra_dsi_of_match[] = {
 	{ .compatible = "nvidia,tegra114-dsi", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_dsi_of_match);
 
 struct platform_driver tegra_dsi_driver = {
 	.driver = {
diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
index 9798a7080322..3513d12d5aa1 100644
--- a/drivers/gpu/drm/tegra/fb.c
+++ b/drivers/gpu/drm/tegra/fb.c
@@ -46,14 +46,15 @@ bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer)
 	return false;
 }
 
-bool tegra_fb_is_tiled(struct drm_framebuffer *framebuffer)
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling)
 {
 	struct tegra_fb *fb = to_tegra_fb(framebuffer);
 
-	if (fb->planes[0]->flags & TEGRA_BO_TILED)
-		return true;
+	/* TODO: handle YUV formats? */
+	*tiling = fb->planes[0]->tiling;
 
-	return false;
+	return 0;
 }
 
 static void tegra_fb_destroy(struct drm_framebuffer *framebuffer)
@@ -193,6 +194,7 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 			     struct drm_fb_helper_surface_size *sizes)
 {
 	struct tegra_fbdev *fbdev = to_tegra_fbdev(helper);
+	struct tegra_drm *tegra = helper->dev->dev_private;
 	struct drm_device *drm = helper->dev;
 	struct drm_mode_fb_cmd2 cmd = { 0 };
 	unsigned int bytes_per_pixel;
@@ -207,7 +209,8 @@ static int tegra_fbdev_probe(struct drm_fb_helper *helper,
 
 	cmd.width = sizes->surface_width;
 	cmd.height = sizes->surface_height;
-	cmd.pitches[0] = sizes->surface_width * bytes_per_pixel;
+	cmd.pitches[0] = round_up(sizes->surface_width * bytes_per_pixel,
+				  tegra->pitch_align);
 	cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
 						     sizes->surface_depth);
 
@@ -267,18 +270,13 @@ release:
 	return err;
 }
 
-static struct drm_fb_helper_funcs tegra_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs tegra_fb_helper_funcs = {
 	.fb_probe = tegra_fbdev_probe,
 };
 
-static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm,
-					      unsigned int preferred_bpp,
-					      unsigned int num_crtc,
-					      unsigned int max_connectors)
+static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm)
 {
-	struct drm_fb_helper *helper;
 	struct tegra_fbdev *fbdev;
-	int err;
 
 	fbdev = kzalloc(sizeof(*fbdev), GFP_KERNEL);
 	if (!fbdev) {
@@ -286,13 +284,23 @@ static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	fbdev->base.funcs = &tegra_fb_helper_funcs;
-	helper = &fbdev->base;
+	drm_fb_helper_prepare(drm, &fbdev->base, &tegra_fb_helper_funcs);
+
+	return fbdev;
+}
+
+static int tegra_fbdev_init(struct tegra_fbdev *fbdev,
+			    unsigned int preferred_bpp,
+			    unsigned int num_crtc,
+			    unsigned int max_connectors)
+{
+	struct drm_device *drm = fbdev->base.dev;
+	int err;
 
 	err = drm_fb_helper_init(drm, &fbdev->base, num_crtc, max_connectors);
 	if (err < 0) {
 		dev_err(drm->dev, "failed to initialize DRM FB helper\n");
-		goto free;
+		return err;
 	}
 
 	err = drm_fb_helper_single_add_all_connectors(&fbdev->base);
@@ -301,21 +309,17 @@ static struct tegra_fbdev *tegra_fbdev_create(struct drm_device *drm,
 		goto fini;
 	}
 
-	drm_helper_disable_unused_functions(drm);
-
 	err = drm_fb_helper_initial_config(&fbdev->base, preferred_bpp);
 	if (err < 0) {
 		dev_err(drm->dev, "failed to set initial configuration\n");
 		goto fini;
 	}
 
-	return fbdev;
+	return 0;
 
 fini:
 	drm_fb_helper_fini(&fbdev->base);
-free:
-	kfree(fbdev);
-	return ERR_PTR(err);
+	return err;
 }
 
 static void tegra_fbdev_free(struct tegra_fbdev *fbdev)
@@ -366,7 +370,7 @@ static const struct drm_mode_config_funcs tegra_drm_mode_funcs = {
 #endif
 };
 
-int tegra_drm_fb_init(struct drm_device *drm)
+int tegra_drm_fb_prepare(struct drm_device *drm)
 {
 #ifdef CONFIG_DRM_TEGRA_FBDEV
 	struct tegra_drm *tegra = drm->dev_private;
@@ -381,8 +385,7 @@ int tegra_drm_fb_init(struct drm_device *drm)
 	drm->mode_config.funcs = &tegra_drm_mode_funcs;
 
 #ifdef CONFIG_DRM_TEGRA_FBDEV
-	tegra->fbdev = tegra_fbdev_create(drm, 32, drm->mode_config.num_crtc,
-					  drm->mode_config.num_connector);
+	tegra->fbdev = tegra_fbdev_create(drm);
 	if (IS_ERR(tegra->fbdev))
 		return PTR_ERR(tegra->fbdev);
 #endif
@@ -390,6 +393,21 @@ int tegra_drm_fb_init(struct drm_device *drm)
 	return 0;
 }
 
+int tegra_drm_fb_init(struct drm_device *drm)
+{
+#ifdef CONFIG_DRM_TEGRA_FBDEV
+	struct tegra_drm *tegra = drm->dev_private;
+	int err;
+
+	err = tegra_fbdev_init(tegra->fbdev, 32, drm->mode_config.num_crtc,
+			       drm->mode_config.num_connector);
+	if (err < 0)
+		return err;
+#endif
+
+	return 0;
+}
+
 void tegra_drm_fb_exit(struct drm_device *drm)
 {
 #ifdef CONFIG_DRM_TEGRA_FBDEV
diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c
index 78cc8143760a..ce023fa3e8ae 100644
--- a/drivers/gpu/drm/tegra/gem.c
+++ b/drivers/gpu/drm/tegra/gem.c
@@ -16,6 +16,7 @@
 #include <linux/dma-buf.h>
 #include <drm/tegra_drm.h>
 
+#include "drm.h"
 #include "gem.h"
 
 static inline struct tegra_bo *host1x_to_tegra_bo(struct host1x_bo *bo)
@@ -126,7 +127,7 @@ struct tegra_bo *tegra_bo_create(struct drm_device *drm, unsigned int size,
 		goto err_mmap;
 
 	if (flags & DRM_TEGRA_GEM_CREATE_TILED)
-		bo->flags |= TEGRA_BO_TILED;
+		bo->tiling.mode = TEGRA_BO_TILING_MODE_TILED;
 
 	if (flags & DRM_TEGRA_GEM_CREATE_BOTTOM_UP)
 		bo->flags |= TEGRA_BO_BOTTOM_UP;
@@ -259,8 +260,10 @@ int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
 			 struct drm_mode_create_dumb *args)
 {
 	int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	struct tegra_drm *tegra = drm->dev_private;
 	struct tegra_bo *bo;
 
+	min_pitch = round_up(min_pitch, tegra->pitch_align);
 	if (args->pitch < min_pitch)
 		args->pitch = min_pitch;
 
diff --git a/drivers/gpu/drm/tegra/gem.h b/drivers/gpu/drm/tegra/gem.h
index 2f3fe96c5154..6538b56780c2 100644
--- a/drivers/gpu/drm/tegra/gem.h
+++ b/drivers/gpu/drm/tegra/gem.h
@@ -15,9 +15,20 @@
 
 #include <drm/drm.h>
 #include <drm/drmP.h>
+#include <drm/drm_gem.h>
 
-#define TEGRA_BO_TILED     (1 << 0)
-#define TEGRA_BO_BOTTOM_UP (1 << 1)
+#define TEGRA_BO_BOTTOM_UP (1 << 0)
+
+enum tegra_bo_tiling_mode {
+	TEGRA_BO_TILING_MODE_PITCH,
+	TEGRA_BO_TILING_MODE_TILED,
+	TEGRA_BO_TILING_MODE_BLOCK,
+};
+
+struct tegra_bo_tiling {
+	enum tegra_bo_tiling_mode mode;
+	unsigned long value;
+};
 
 struct tegra_bo {
 	struct drm_gem_object gem;
@@ -26,6 +37,8 @@ struct tegra_bo {
 	struct sg_table *sgt;
 	dma_addr_t paddr;
 	void *vaddr;
+
+	struct tegra_bo_tiling tiling;
 };
 
 static inline struct tegra_bo *to_tegra_bo(struct drm_gem_object *gem)
diff --git a/drivers/gpu/drm/tegra/gr2d.c b/drivers/gpu/drm/tegra/gr2d.c
index 7c53941f2a9e..02cd3e37a6ec 100644
--- a/drivers/gpu/drm/tegra/gr2d.c
+++ b/drivers/gpu/drm/tegra/gr2d.c
@@ -121,6 +121,7 @@ static const struct of_device_id gr2d_match[] = {
 	{ .compatible = "nvidia,tegra20-gr2d" },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, gr2d_match);
 
 static const u32 gr2d_addr_regs[] = {
 	GR2D_UA_BASE_ADDR,
diff --git a/drivers/gpu/drm/tegra/gr3d.c b/drivers/gpu/drm/tegra/gr3d.c
index 30f5ba9bd6d0..0b3f2b977ba0 100644
--- a/drivers/gpu/drm/tegra/gr3d.c
+++ b/drivers/gpu/drm/tegra/gr3d.c
@@ -12,7 +12,8 @@
 #include <linux/module.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
-#include <linux/tegra-powergate.h>
+
+#include <soc/tegra/pmc.h>
 
 #include "drm.h"
 #include "gem.h"
@@ -130,6 +131,7 @@ static const struct of_device_id tegra_gr3d_match[] = {
 	{ .compatible = "nvidia,tegra20-gr3d" },
 	{ }
 };
+MODULE_DEVICE_TABLE(of, tegra_gr3d_match);
 
 static const u32 gr3d_addr_regs[] = {
 	GR3D_IDX_ATTRIBUTE( 0),
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
index ba067bb767e3..ffe26547328d 100644
--- a/drivers/gpu/drm/tegra/hdmi.c
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -1450,6 +1450,7 @@ static const struct of_device_id tegra_hdmi_of_match[] = {
 	{ .compatible = "nvidia,tegra20-hdmi", .data = &tegra20_hdmi_config },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match);
 
 static int tegra_hdmi_probe(struct platform_device *pdev)
 {
diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c
index a3e4f1eca6f7..0c67d7eebc94 100644
--- a/drivers/gpu/drm/tegra/output.c
+++ b/drivers/gpu/drm/tegra/output.c
@@ -105,7 +105,7 @@ static void drm_connector_clear(struct drm_connector *connector)
 
 static void tegra_connector_destroy(struct drm_connector *connector)
 {
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	drm_connector_clear(connector);
 }
@@ -140,7 +140,9 @@ static void tegra_encoder_dpms(struct drm_encoder *encoder, int mode)
 	if (mode != DRM_MODE_DPMS_ON) {
 		drm_panel_disable(panel);
 		tegra_output_disable(output);
+		drm_panel_unprepare(panel);
 	} else {
+		drm_panel_prepare(panel);
 		tegra_output_enable(output);
 		drm_panel_enable(panel);
 	}
@@ -318,7 +320,7 @@ int tegra_output_init(struct drm_device *drm, struct tegra_output *output)
 	drm_encoder_helper_add(&output->encoder, &encoder_helper_funcs);
 
 	drm_mode_connector_attach_encoder(&output->connector, &output->encoder);
-	drm_sysfs_connector_add(&output->connector);
+	drm_connector_register(&output->connector);
 
 	output->encoder.possible_crtcs = 0x3;
 
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
index 27c979b50111..7829e81f065d 100644
--- a/drivers/gpu/drm/tegra/sor.c
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -11,7 +11,8 @@
 #include <linux/io.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
-#include <linux/tegra-powergate.h>
+
+#include <soc/tegra/pmc.h>
 
 #include <drm/drm_dp_helper.h>
 
@@ -516,7 +517,7 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 		if (err < 0) {
 			dev_err(sor->dev, "failed to probe eDP link: %d\n",
 				err);
-			return err;
+			goto unlock;
 		}
 	}
 
@@ -525,7 +526,7 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
 
 	memset(&config, 0, sizeof(config));
-	config.bits_per_pixel = 24; /* XXX: don't hardcode? */
+	config.bits_per_pixel = output->connector.display_info.bpc * 3;
 
 	err = tegra_sor_calc_config(sor, mode, &config, &link);
 	if (err < 0)
@@ -815,12 +816,22 @@ static int tegra_output_sor_enable(struct tegra_output *output)
 	 * configure panel (24bpp, vsync-, hsync-, DP-A protocol, complete
 	 * raster, associate with display controller)
 	 */
-	value = SOR_STATE_ASY_VSYNCPOL |
-		SOR_STATE_ASY_HSYNCPOL |
-		SOR_STATE_ASY_PROTOCOL_DP_A |
+	value = SOR_STATE_ASY_PROTOCOL_DP_A |
 		SOR_STATE_ASY_CRC_MODE_COMPLETE |
 		SOR_STATE_ASY_OWNER(dc->pipe + 1);
 
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		value &= ~SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		value &= ~SOR_STATE_ASY_VSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL;
+
 	switch (config.bits_per_pixel) {
 	case 24:
 		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
@@ -1455,6 +1466,7 @@ static const struct of_device_id tegra_sor_of_match[] = {
 	{ .compatible = "nvidia,tegra124-sor", },
 	{ },
 };
+MODULE_DEVICE_TABLE(of, tegra_sor_of_match);
 
 struct platform_driver tegra_sor_driver = {
 	.driver = {
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index b20b69488dc9..79a34cbd29f5 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -84,6 +84,7 @@ static int modeset_init(struct drm_device *dev)
 	if ((priv->num_encoders == 0) || (priv->num_connectors == 0)) {
 		/* oh nos! */
 		dev_err(dev->dev, "no encoders/connectors found\n");
+		drm_mode_config_cleanup(dev);
 		return -ENXIO;
 	}
 
@@ -120,8 +121,8 @@ static int cpufreq_transition(struct notifier_block *nb,
 static int tilcdc_unload(struct drm_device *dev)
 {
 	struct tilcdc_drm_private *priv = dev->dev_private;
-	struct tilcdc_module *mod, *cur;
 
+	drm_fbdev_cma_fini(priv->fbdev);
 	drm_kms_helper_poll_fini(dev);
 	drm_mode_config_cleanup(dev);
 	drm_vblank_cleanup(dev);
@@ -148,11 +149,6 @@ static int tilcdc_unload(struct drm_device *dev)
 
 	pm_runtime_disable(dev->dev);
 
-	list_for_each_entry_safe(mod, cur, &module_list, list) {
-		DBG("destroying module: %s", mod->name);
-		mod->funcs->destroy(mod);
-	}
-
 	kfree(priv);
 
 	return 0;
@@ -177,33 +173,37 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	dev->dev_private = priv;
 
 	priv->wq = alloc_ordered_workqueue("tilcdc", 0);
+	if (!priv->wq) {
+		ret = -ENOMEM;
+		goto fail_free_priv;
+	}
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!res) {
 		dev_err(dev->dev, "failed to get memory resource\n");
 		ret = -EINVAL;
-		goto fail;
+		goto fail_free_wq;
 	}
 
 	priv->mmio = ioremap_nocache(res->start, resource_size(res));
 	if (!priv->mmio) {
 		dev_err(dev->dev, "failed to ioremap\n");
 		ret = -ENOMEM;
-		goto fail;
+		goto fail_free_wq;
 	}
 
 	priv->clk = clk_get(dev->dev, "fck");
 	if (IS_ERR(priv->clk)) {
 		dev_err(dev->dev, "failed to get functional clock\n");
 		ret = -ENODEV;
-		goto fail;
+		goto fail_iounmap;
 	}
 
 	priv->disp_clk = clk_get(dev->dev, "dpll_disp_ck");
 	if (IS_ERR(priv->clk)) {
 		dev_err(dev->dev, "failed to get display clock\n");
 		ret = -ENODEV;
-		goto fail;
+		goto fail_put_clk;
 	}
 
 #ifdef CONFIG_CPU_FREQ
@@ -213,7 +213,7 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 			CPUFREQ_TRANSITION_NOTIFIER);
 	if (ret) {
 		dev_err(dev->dev, "failed to register cpufreq notifier\n");
-		goto fail;
+		goto fail_put_disp_clk;
 	}
 #endif
 
@@ -258,13 +258,13 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	ret = modeset_init(dev);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize mode setting\n");
-		goto fail;
+		goto fail_cpufreq_unregister;
 	}
 
 	ret = drm_vblank_init(dev, 1);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to initialize vblank\n");
-		goto fail;
+		goto fail_mode_config_cleanup;
 	}
 
 	pm_runtime_get_sync(dev->dev);
@@ -272,7 +272,7 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	pm_runtime_put_sync(dev->dev);
 	if (ret < 0) {
 		dev_err(dev->dev, "failed to install IRQ handler\n");
-		goto fail;
+		goto fail_vblank_cleanup;
 	}
 
 	platform_set_drvdata(pdev, dev);
@@ -288,13 +288,48 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	priv->fbdev = drm_fbdev_cma_init(dev, bpp,
 			dev->mode_config.num_crtc,
 			dev->mode_config.num_connector);
+	if (IS_ERR(priv->fbdev)) {
+		ret = PTR_ERR(priv->fbdev);
+		goto fail_irq_uninstall;
+	}
 
 	drm_kms_helper_poll_init(dev);
 
 	return 0;
 
-fail:
-	tilcdc_unload(dev);
+fail_irq_uninstall:
+	pm_runtime_get_sync(dev->dev);
+	drm_irq_uninstall(dev);
+	pm_runtime_put_sync(dev->dev);
+
+fail_vblank_cleanup:
+	drm_vblank_cleanup(dev);
+
+fail_mode_config_cleanup:
+	drm_mode_config_cleanup(dev);
+
+fail_cpufreq_unregister:
+	pm_runtime_disable(dev->dev);
+#ifdef CONFIG_CPU_FREQ
+	cpufreq_unregister_notifier(&priv->freq_transition,
+			CPUFREQ_TRANSITION_NOTIFIER);
+fail_put_disp_clk:
+	clk_put(priv->disp_clk);
+#endif
+
+fail_put_clk:
+	clk_put(priv->clk);
+
+fail_iounmap:
+	iounmap(priv->mmio);
+
+fail_free_wq:
+	flush_workqueue(priv->wq);
+	destroy_workqueue(priv->wq);
+
+fail_free_priv:
+	dev->dev_private = NULL;
+	kfree(priv);
 	return ret;
 }
 
@@ -507,6 +542,7 @@ static struct drm_driver tilcdc_driver = {
 	.unload             = tilcdc_unload,
 	.preclose           = tilcdc_preclose,
 	.lastclose          = tilcdc_lastclose,
+	.set_busid          = drm_platform_set_busid,
 	.irq_handler        = tilcdc_irq,
 	.irq_preinstall     = tilcdc_irq_preinstall,
 	.irq_postinstall    = tilcdc_irq_postinstall,
@@ -628,13 +664,13 @@ static int __init tilcdc_drm_init(void)
 static void __exit tilcdc_drm_fini(void)
 {
 	DBG("fini");
-	tilcdc_tfp410_fini();
-	tilcdc_slave_fini();
-	tilcdc_panel_fini();
 	platform_driver_unregister(&tilcdc_platform_driver);
+	tilcdc_panel_fini();
+	tilcdc_slave_fini();
+	tilcdc_tfp410_fini();
 }
 
-late_initcall(tilcdc_drm_init);
+module_init(tilcdc_drm_init);
 module_exit(tilcdc_drm_fini);
 
 MODULE_AUTHOR("Rob Clark <robdclark@gmail.com");
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
index 093803683b25..7596c144a9fb 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
@@ -98,7 +98,6 @@ struct tilcdc_module;
 struct tilcdc_module_ops {
 	/* create appropriate encoders/connectors: */
 	int (*modeset_init)(struct tilcdc_module *mod, struct drm_device *dev);
-	void (*destroy)(struct tilcdc_module *mod);
 #ifdef CONFIG_DEBUG_FS
 	/* create debugfs nodes (can be NULL): */
 	int (*debugfs_init)(struct tilcdc_module *mod, struct drm_minor *minor);
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
index 86c67329b605..7a0315855e90 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -18,6 +18,7 @@
 #include <linux/pinctrl/pinmux.h>
 #include <linux/pinctrl/consumer.h>
 #include <linux/backlight.h>
+#include <linux/gpio/consumer.h>
 #include <video/display_timing.h>
 #include <video/of_display_timing.h>
 #include <video/videomode.h>
@@ -29,6 +30,7 @@ struct panel_module {
 	struct tilcdc_panel_info *info;
 	struct display_timings *timings;
 	struct backlight_device *backlight;
+	struct gpio_desc *enable_gpio;
 };
 #define to_panel_module(x) container_of(x, struct panel_module, base)
 
@@ -55,13 +57,17 @@ static void panel_encoder_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct panel_encoder *panel_encoder = to_panel_encoder(encoder);
 	struct backlight_device *backlight = panel_encoder->mod->backlight;
+	struct gpio_desc *gpio = panel_encoder->mod->enable_gpio;
 
-	if (!backlight)
-		return;
+	if (backlight) {
+		backlight->props.power = mode == DRM_MODE_DPMS_ON ?
+					 FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN;
+		backlight_update_status(backlight);
+	}
 
-	backlight->props.power = mode == DRM_MODE_DPMS_ON
-				     ? FB_BLANK_UNBLANK : FB_BLANK_POWERDOWN;
-	backlight_update_status(backlight);
+	if (gpio)
+		gpiod_set_value_cansleep(gpio,
+					 mode == DRM_MODE_DPMS_ON ? 1 : 0);
 }
 
 static bool panel_encoder_mode_fixup(struct drm_encoder *encoder,
@@ -151,6 +157,7 @@ struct panel_connector {
 static void panel_connector_destroy(struct drm_connector *connector)
 {
 	struct panel_connector *panel_connector = to_panel_connector(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(panel_connector);
 }
@@ -247,7 +254,7 @@ static struct drm_connector *panel_connector_create(struct drm_device *dev,
 	if (ret)
 		goto fail;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	return connector;
 
@@ -281,23 +288,8 @@ static int panel_modeset_init(struct tilcdc_module *mod, struct drm_device *dev)
 	return 0;
 }
 
-static void panel_destroy(struct tilcdc_module *mod)
-{
-	struct panel_module *panel_mod = to_panel_module(mod);
-
-	if (panel_mod->timings) {
-		display_timings_release(panel_mod->timings);
-		kfree(panel_mod->timings);
-	}
-
-	tilcdc_module_cleanup(mod);
-	kfree(panel_mod->info);
-	kfree(panel_mod);
-}
-
 static const struct tilcdc_module_ops panel_module_ops = {
 		.modeset_init = panel_modeset_init,
-		.destroy = panel_destroy,
 };
 
 /*
@@ -325,6 +317,7 @@ static struct tilcdc_panel_info *of_get_panel_info(struct device_node *np)
 	info = kzalloc(sizeof(*info), GFP_KERNEL);
 	if (!info) {
 		pr_err("%s: allocation failed\n", __func__);
+		of_node_put(info_np);
 		return NULL;
 	}
 
@@ -345,22 +338,21 @@ static struct tilcdc_panel_info *of_get_panel_info(struct device_node *np)
 	if (ret) {
 		pr_err("%s: error reading panel-info properties\n", __func__);
 		kfree(info);
+		of_node_put(info_np);
 		return NULL;
 	}
+	of_node_put(info_np);
 
 	return info;
 }
 
-static struct of_device_id panel_of_match[];
-
 static int panel_probe(struct platform_device *pdev)
 {
-	struct device_node *node = pdev->dev.of_node;
+	struct device_node *bl_node, *node = pdev->dev.of_node;
 	struct panel_module *panel_mod;
 	struct tilcdc_module *mod;
 	struct pinctrl *pinctrl;
-	int ret = -EINVAL;
-
+	int ret;
 
 	/* bail out early if no DT data: */
 	if (!node) {
@@ -368,11 +360,42 @@ static int panel_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	panel_mod = kzalloc(sizeof(*panel_mod), GFP_KERNEL);
+	panel_mod = devm_kzalloc(&pdev->dev, sizeof(*panel_mod), GFP_KERNEL);
 	if (!panel_mod)
 		return -ENOMEM;
 
+	bl_node = of_parse_phandle(node, "backlight", 0);
+	if (bl_node) {
+		panel_mod->backlight = of_find_backlight_by_node(bl_node);
+		of_node_put(bl_node);
+
+		if (!panel_mod->backlight)
+			return -EPROBE_DEFER;
+
+		dev_info(&pdev->dev, "found backlight\n");
+	}
+
+	panel_mod->enable_gpio = devm_gpiod_get(&pdev->dev, "enable");
+	if (IS_ERR(panel_mod->enable_gpio)) {
+		ret = PTR_ERR(panel_mod->enable_gpio);
+		if (ret != -ENOENT) {
+			dev_err(&pdev->dev, "failed to request enable GPIO\n");
+			goto fail_backlight;
+		}
+
+		/* Optional GPIO is not here, continue silently. */
+		panel_mod->enable_gpio = NULL;
+	} else {
+		ret = gpiod_direction_output(panel_mod->enable_gpio, 0);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "failed to setup GPIO\n");
+			goto fail_backlight;
+		}
+		dev_info(&pdev->dev, "found enable GPIO\n");
+	}
+
 	mod = &panel_mod->base;
+	pdev->dev.platform_data = mod;
 
 	tilcdc_module_init(mod, "panel", &panel_module_ops);
 
@@ -380,34 +403,50 @@ static int panel_probe(struct platform_device *pdev)
 	if (IS_ERR(pinctrl))
 		dev_warn(&pdev->dev, "pins are not configured\n");
 
-
 	panel_mod->timings = of_get_display_timings(node);
 	if (!panel_mod->timings) {
 		dev_err(&pdev->dev, "could not get panel timings\n");
-		goto fail;
+		ret = -EINVAL;
+		goto fail_free;
 	}
 
 	panel_mod->info = of_get_panel_info(node);
 	if (!panel_mod->info) {
 		dev_err(&pdev->dev, "could not get panel info\n");
-		goto fail;
+		ret = -EINVAL;
+		goto fail_timings;
 	}
 
 	mod->preferred_bpp = panel_mod->info->bpp;
 
-	panel_mod->backlight = of_find_backlight_by_node(node);
-	if (panel_mod->backlight)
-		dev_info(&pdev->dev, "found backlight\n");
-
 	return 0;
 
-fail:
-	panel_destroy(mod);
+fail_timings:
+	display_timings_release(panel_mod->timings);
+
+fail_free:
+	tilcdc_module_cleanup(mod);
+
+fail_backlight:
+	if (panel_mod->backlight)
+		put_device(&panel_mod->backlight->dev);
 	return ret;
 }
 
 static int panel_remove(struct platform_device *pdev)
 {
+	struct tilcdc_module *mod = dev_get_platdata(&pdev->dev);
+	struct panel_module *panel_mod = to_panel_module(mod);
+	struct backlight_device *backlight = panel_mod->backlight;
+
+	if (backlight)
+		put_device(&backlight->dev);
+
+	display_timings_release(panel_mod->timings);
+
+	tilcdc_module_cleanup(mod);
+	kfree(panel_mod->info);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.c b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
index 595068ba2d5e..3775fd49dac4 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_slave.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
@@ -166,6 +166,7 @@ struct slave_connector {
 static void slave_connector_destroy(struct drm_connector *connector)
 {
 	struct slave_connector *slave_connector = to_slave_connector(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(slave_connector);
 }
@@ -261,7 +262,7 @@ static struct drm_connector *slave_connector_create(struct drm_device *dev,
 	if (ret)
 		goto fail;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	return connector;
 
@@ -295,17 +296,8 @@ static int slave_modeset_init(struct tilcdc_module *mod, struct drm_device *dev)
 	return 0;
 }
 
-static void slave_destroy(struct tilcdc_module *mod)
-{
-	struct slave_module *slave_mod = to_slave_module(mod);
-
-	tilcdc_module_cleanup(mod);
-	kfree(slave_mod);
-}
-
 static const struct tilcdc_module_ops slave_module_ops = {
 		.modeset_init = slave_modeset_init,
-		.destroy = slave_destroy,
 };
 
 /*
@@ -355,10 +347,13 @@ static int slave_probe(struct platform_device *pdev)
 	}
 
 	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
-	if (!slave_mod)
-		return -ENOMEM;
+	if (!slave_mod) {
+		ret = -ENOMEM;
+		goto fail_adapter;
+	}
 
 	mod = &slave_mod->base;
+	pdev->dev.platform_data = mod;
 
 	mod->preferred_bpp = slave_info.bpp;
 
@@ -373,10 +368,20 @@ static int slave_probe(struct platform_device *pdev)
 	tilcdc_slave_probedefer(false);
 
 	return 0;
+
+fail_adapter:
+	i2c_put_adapter(slavei2c);
+	return ret;
 }
 
 static int slave_remove(struct platform_device *pdev)
 {
+	struct tilcdc_module *mod = dev_get_platdata(&pdev->dev);
+	struct slave_module *slave_mod = to_slave_module(mod);
+
+	tilcdc_module_cleanup(mod);
+	kfree(slave_mod);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
index c38b56b268ac..354c47ca6374 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
@@ -167,6 +167,7 @@ struct tfp410_connector {
 static void tfp410_connector_destroy(struct drm_connector *connector)
 {
 	struct tfp410_connector *tfp410_connector = to_tfp410_connector(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(tfp410_connector);
 }
@@ -261,7 +262,7 @@ static struct drm_connector *tfp410_connector_create(struct drm_device *dev,
 	if (ret)
 		goto fail;
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 
 	return connector;
 
@@ -295,23 +296,8 @@ static int tfp410_modeset_init(struct tilcdc_module *mod, struct drm_device *dev
 	return 0;
 }
 
-static void tfp410_destroy(struct tilcdc_module *mod)
-{
-	struct tfp410_module *tfp410_mod = to_tfp410_module(mod);
-
-	if (tfp410_mod->i2c)
-		i2c_put_adapter(tfp410_mod->i2c);
-
-	if (!IS_ERR_VALUE(tfp410_mod->gpio))
-		gpio_free(tfp410_mod->gpio);
-
-	tilcdc_module_cleanup(mod);
-	kfree(tfp410_mod);
-}
-
 static const struct tilcdc_module_ops tfp410_module_ops = {
 		.modeset_init = tfp410_modeset_init,
-		.destroy = tfp410_destroy,
 };
 
 /*
@@ -341,6 +327,7 @@ static int tfp410_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	mod = &tfp410_mod->base;
+	pdev->dev.platform_data = mod;
 
 	tilcdc_module_init(mod, "tfp410", &tfp410_module_ops);
 
@@ -364,6 +351,7 @@ static int tfp410_probe(struct platform_device *pdev)
 	tfp410_mod->i2c = of_find_i2c_adapter_by_node(i2c_node);
 	if (!tfp410_mod->i2c) {
 		dev_err(&pdev->dev, "could not get i2c\n");
+		of_node_put(i2c_node);
 		goto fail;
 	}
 
@@ -377,19 +365,32 @@ static int tfp410_probe(struct platform_device *pdev)
 		ret = gpio_request(tfp410_mod->gpio, "DVI_PDn");
 		if (ret) {
 			dev_err(&pdev->dev, "could not get DVI_PDn gpio\n");
-			goto fail;
+			goto fail_adapter;
 		}
 	}
 
 	return 0;
 
+fail_adapter:
+	i2c_put_adapter(tfp410_mod->i2c);
+
 fail:
-	tfp410_destroy(mod);
+	kfree(tfp410_mod);
+	tilcdc_module_cleanup(mod);
 	return ret;
 }
 
 static int tfp410_remove(struct platform_device *pdev)
 {
+	struct tilcdc_module *mod = dev_get_platdata(&pdev->dev);
+	struct tfp410_module *tfp410_mod = to_tfp410_module(mod);
+
+	i2c_put_adapter(tfp410_mod->i2c);
+	gpio_free(tfp410_mod->gpio);
+
+	tilcdc_module_cleanup(mod);
+	kfree(tfp410_mod);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 4ab9f7171c4f..d395b0bef73b 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -40,6 +40,7 @@
 #include <linux/file.h>
 #include <linux/module.h>
 #include <linux/atomic.h>
+#include <linux/reservation.h>
 
 #define TTM_ASSERT_LOCKED(param)
 #define TTM_DEBUG(fmt, arg...)
@@ -53,12 +54,13 @@ static struct attribute ttm_bo_count = {
 	.mode = S_IRUGO
 };
 
-static inline int ttm_mem_type_from_flags(uint32_t flags, uint32_t *mem_type)
+static inline int ttm_mem_type_from_place(const struct ttm_place *place,
+					  uint32_t *mem_type)
 {
 	int i;
 
 	for (i = 0; i <= TTM_PL_PRIV5; i++)
-		if (flags & (1 << i)) {
+		if (place->flags & (1 << i)) {
 			*mem_type = i;
 			return 0;
 		}
@@ -89,12 +91,12 @@ static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
 	       bo, bo->mem.num_pages, bo->mem.size >> 10,
 	       bo->mem.size >> 20);
 	for (i = 0; i < placement->num_placement; i++) {
-		ret = ttm_mem_type_from_flags(placement->placement[i],
+		ret = ttm_mem_type_from_place(&placement->placement[i],
 						&mem_type);
 		if (ret)
 			return;
 		pr_err("  placement[%d]=0x%08X (%d)\n",
-		       i, placement->placement[i], mem_type);
+		       i, placement->placement[i].flags, mem_type);
 		ttm_mem_type_debug(bo->bdev, mem_type);
 	}
 }
@@ -141,7 +143,6 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	BUG_ON(atomic_read(&bo->list_kref.refcount));
 	BUG_ON(atomic_read(&bo->kref.refcount));
 	BUG_ON(atomic_read(&bo->cpu_writers));
-	BUG_ON(bo->sync_obj != NULL);
 	BUG_ON(bo->mem.mm_node != NULL);
 	BUG_ON(!list_empty(&bo->lru));
 	BUG_ON(!list_empty(&bo->ddestroy));
@@ -402,36 +403,48 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	ww_mutex_unlock (&bo->resv->lock);
 }
 
+static void ttm_bo_flush_all_fences(struct ttm_buffer_object *bo)
+{
+	struct reservation_object_list *fobj;
+	struct fence *fence;
+	int i;
+
+	fobj = reservation_object_get_list(bo->resv);
+	fence = reservation_object_get_excl(bo->resv);
+	if (fence && !fence->ops->signaled)
+		fence_enable_sw_signaling(fence);
+
+	for (i = 0; fobj && i < fobj->shared_count; ++i) {
+		fence = rcu_dereference_protected(fobj->shared[i],
+					reservation_object_held(bo->resv));
+
+		if (!fence->ops->signaled)
+			fence_enable_sw_signaling(fence);
+	}
+}
+
 static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_global *glob = bo->glob;
-	struct ttm_bo_driver *driver = bdev->driver;
-	void *sync_obj = NULL;
 	int put_count;
 	int ret;
 
 	spin_lock(&glob->lru_lock);
-	ret = __ttm_bo_reserve(bo, false, true, false, 0);
-
-	spin_lock(&bdev->fence_lock);
-	(void) ttm_bo_wait(bo, false, false, true);
-	if (!ret && !bo->sync_obj) {
-		spin_unlock(&bdev->fence_lock);
-		put_count = ttm_bo_del_from_lru(bo);
+	ret = __ttm_bo_reserve(bo, false, true, false, NULL);
 
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_cleanup_memtype_use(bo);
+	if (!ret) {
+		if (!ttm_bo_wait(bo, false, false, true)) {
+			put_count = ttm_bo_del_from_lru(bo);
 
-		ttm_bo_list_ref_sub(bo, put_count, true);
+			spin_unlock(&glob->lru_lock);
+			ttm_bo_cleanup_memtype_use(bo);
 
-		return;
-	}
-	if (bo->sync_obj)
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-	spin_unlock(&bdev->fence_lock);
+			ttm_bo_list_ref_sub(bo, put_count, true);
 
-	if (!ret) {
+			return;
+		} else
+			ttm_bo_flush_all_fences(bo);
 
 		/*
 		 * Make NO_EVICT bos immediately available to
@@ -450,10 +463,6 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
 	spin_unlock(&glob->lru_lock);
 
-	if (sync_obj) {
-		driver->sync_obj_flush(sync_obj);
-		driver->sync_obj_unref(&sync_obj);
-	}
 	schedule_delayed_work(&bdev->wq,
 			      ((HZ / 100) < 1) ? 1 : HZ / 100);
 }
@@ -474,47 +483,29 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 					  bool interruptible,
 					  bool no_wait_gpu)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	struct ttm_bo_global *glob = bo->glob;
 	int put_count;
 	int ret;
 
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, false, true);
 
 	if (ret && !no_wait_gpu) {
-		void *sync_obj;
-
-		/*
-		 * Take a reference to the fence and unreserve,
-		 * at this point the buffer should be dead, so
-		 * no new sync objects can be attached.
-		 */
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-		spin_unlock(&bdev->fence_lock);
-
-		__ttm_bo_unreserve(bo);
+		long lret;
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 
-		ret = driver->sync_obj_wait(sync_obj, false, interruptible);
-		driver->sync_obj_unref(&sync_obj);
-		if (ret)
-			return ret;
+		lret = reservation_object_wait_timeout_rcu(bo->resv,
+							   true,
+							   interruptible,
+							   30 * HZ);
 
-		/*
-		 * remove sync_obj with ttm_bo_wait, the wait should be
-		 * finished, and no new wait object should have been added.
-		 */
-		spin_lock(&bdev->fence_lock);
-		ret = ttm_bo_wait(bo, false, false, true);
-		WARN_ON(ret);
-		spin_unlock(&bdev->fence_lock);
-		if (ret)
-			return ret;
+		if (lret < 0)
+			return lret;
+		else if (lret == 0)
+			return -EBUSY;
 
 		spin_lock(&glob->lru_lock);
-		ret = __ttm_bo_reserve(bo, false, true, false, 0);
+		ret = __ttm_bo_reserve(bo, false, true, false, NULL);
 
 		/*
 		 * We raced, and lost, someone else holds the reservation now,
@@ -528,8 +519,14 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 			spin_unlock(&glob->lru_lock);
 			return 0;
 		}
-	} else
-		spin_unlock(&bdev->fence_lock);
+
+		/*
+		 * remove sync_obj with ttm_bo_wait, the wait should be
+		 * finished, and no new wait object should have been added.
+		 */
+		ret = ttm_bo_wait(bo, false, false, true);
+		WARN_ON(ret);
+	}
 
 	if (ret || unlikely(list_empty(&bo->ddestroy))) {
 		__ttm_bo_unreserve(bo);
@@ -577,11 +574,11 @@ static int ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all)
 			kref_get(&nentry->list_kref);
 		}
 
-		ret = __ttm_bo_reserve(entry, false, true, false, 0);
+		ret = __ttm_bo_reserve(entry, false, true, false, NULL);
 		if (remove_all && ret) {
 			spin_unlock(&glob->lru_lock);
 			ret = __ttm_bo_reserve(entry, false, false,
-					       false, 0);
+					       false, NULL);
 			spin_lock(&glob->lru_lock);
 		}
 
@@ -667,9 +664,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	struct ttm_placement placement;
 	int ret = 0;
 
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bdev->fence_lock);
 
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS) {
@@ -685,8 +680,6 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	evict_mem.bus.io_reserved_vm = false;
 	evict_mem.bus.io_reserved_count = 0;
 
-	placement.fpfn = 0;
-	placement.lpfn = 0;
 	placement.num_placement = 0;
 	placement.num_busy_placement = 0;
 	bdev->driver->evict_flags(bo, &placement);
@@ -716,6 +709,7 @@ out:
 
 static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 				uint32_t mem_type,
+				const struct ttm_place *place,
 				bool interruptible,
 				bool no_wait_gpu)
 {
@@ -726,9 +720,22 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
 
 	spin_lock(&glob->lru_lock);
 	list_for_each_entry(bo, &man->lru, lru) {
-		ret = __ttm_bo_reserve(bo, false, true, false, 0);
-		if (!ret)
+		ret = __ttm_bo_reserve(bo, false, true, false, NULL);
+		if (!ret) {
+			if (place && (place->fpfn || place->lpfn)) {
+				/* Don't evict this BO if it's outside of the
+				 * requested placement range
+				 */
+				if (place->fpfn >= (bo->mem.start + bo->mem.size) ||
+				    (place->lpfn && place->lpfn <= bo->mem.start)) {
+					__ttm_bo_unreserve(bo);
+					ret = -EBUSY;
+					continue;
+				}
+			}
+
 			break;
+		}
 	}
 
 	if (ret) {
@@ -774,7 +781,7 @@ EXPORT_SYMBOL(ttm_bo_mem_put);
  */
 static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 					uint32_t mem_type,
-					struct ttm_placement *placement,
+					const struct ttm_place *place,
 					struct ttm_mem_reg *mem,
 					bool interruptible,
 					bool no_wait_gpu)
@@ -784,12 +791,12 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 	int ret;
 
 	do {
-		ret = (*man->func->get_node)(man, bo, placement, mem);
+		ret = (*man->func->get_node)(man, bo, place, mem);
 		if (unlikely(ret != 0))
 			return ret;
 		if (mem->mm_node)
 			break;
-		ret = ttm_mem_evict_first(bdev, mem_type,
+		ret = ttm_mem_evict_first(bdev, mem_type, place,
 					  interruptible, no_wait_gpu);
 		if (unlikely(ret != 0))
 			return ret;
@@ -827,18 +834,18 @@ static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man,
 
 static bool ttm_bo_mt_compatible(struct ttm_mem_type_manager *man,
 				 uint32_t mem_type,
-				 uint32_t proposed_placement,
+				 const struct ttm_place *place,
 				 uint32_t *masked_placement)
 {
 	uint32_t cur_flags = ttm_bo_type_flags(mem_type);
 
-	if ((cur_flags & proposed_placement & TTM_PL_MASK_MEM) == 0)
+	if ((cur_flags & place->flags & TTM_PL_MASK_MEM) == 0)
 		return false;
 
-	if ((proposed_placement & man->available_caching) == 0)
+	if ((place->flags & man->available_caching) == 0)
 		return false;
 
-	cur_flags |= (proposed_placement & man->available_caching);
+	cur_flags |= (place->flags & man->available_caching);
 
 	*masked_placement = cur_flags;
 	return true;
@@ -869,15 +876,14 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 
 	mem->mm_node = NULL;
 	for (i = 0; i < placement->num_placement; ++i) {
-		ret = ttm_mem_type_from_flags(placement->placement[i],
-						&mem_type);
+		const struct ttm_place *place = &placement->placement[i];
+
+		ret = ttm_mem_type_from_place(place, &mem_type);
 		if (ret)
 			return ret;
 		man = &bdev->man[mem_type];
 
-		type_ok = ttm_bo_mt_compatible(man,
-						mem_type,
-						placement->placement[i],
+		type_ok = ttm_bo_mt_compatible(man, mem_type, place,
 						&cur_flags);
 
 		if (!type_ok)
@@ -889,7 +895,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		 * Use the access and other non-mapping-related flag bits from
 		 * the memory placement flags to the current flags
 		 */
-		ttm_flag_masked(&cur_flags, placement->placement[i],
+		ttm_flag_masked(&cur_flags, place->flags,
 				~TTM_PL_MASK_MEMTYPE);
 
 		if (mem_type == TTM_PL_SYSTEM)
@@ -897,7 +903,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 
 		if (man->has_type && man->use_type) {
 			type_found = true;
-			ret = (*man->func->get_node)(man, bo, placement, mem);
+			ret = (*man->func->get_node)(man, bo, place, mem);
 			if (unlikely(ret))
 				return ret;
 		}
@@ -915,17 +921,15 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		return -EINVAL;
 
 	for (i = 0; i < placement->num_busy_placement; ++i) {
-		ret = ttm_mem_type_from_flags(placement->busy_placement[i],
-						&mem_type);
+		const struct ttm_place *place = &placement->busy_placement[i];
+
+		ret = ttm_mem_type_from_place(place, &mem_type);
 		if (ret)
 			return ret;
 		man = &bdev->man[mem_type];
 		if (!man->has_type)
 			continue;
-		if (!ttm_bo_mt_compatible(man,
-						mem_type,
-						placement->busy_placement[i],
-						&cur_flags))
+		if (!ttm_bo_mt_compatible(man, mem_type, place, &cur_flags))
 			continue;
 
 		cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
@@ -934,10 +938,9 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 		 * Use the access and other non-mapping-related flag bits from
 		 * the memory placement flags to the current flags
 		 */
-		ttm_flag_masked(&cur_flags, placement->busy_placement[i],
+		ttm_flag_masked(&cur_flags, place->flags,
 				~TTM_PL_MASK_MEMTYPE);
 
-
 		if (mem_type == TTM_PL_SYSTEM) {
 			mem->mem_type = mem_type;
 			mem->placement = cur_flags;
@@ -945,7 +948,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 			return 0;
 		}
 
-		ret = ttm_bo_mem_force_space(bo, mem_type, placement, mem,
+		ret = ttm_bo_mem_force_space(bo, mem_type, place, mem,
 						interruptible, no_wait_gpu);
 		if (ret == 0 && mem->mm_node) {
 			mem->placement = cur_flags;
@@ -966,7 +969,6 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 {
 	int ret = 0;
 	struct ttm_mem_reg mem;
-	struct ttm_bo_device *bdev = bo->bdev;
 
 	lockdep_assert_held(&bo->resv->lock.base);
 
@@ -975,9 +977,7 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	 * Have the driver move function wait for idle when necessary,
 	 * instead of doing it here.
 	 */
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bdev->fence_lock);
 	if (ret)
 		return ret;
 	mem.num_pages = bo->num_pages;
@@ -1006,20 +1006,27 @@ static bool ttm_bo_mem_compat(struct ttm_placement *placement,
 {
 	int i;
 
-	if (mem->mm_node && placement->lpfn != 0 &&
-	    (mem->start < placement->fpfn ||
-	     mem->start + mem->num_pages > placement->lpfn))
-		return false;
-
 	for (i = 0; i < placement->num_placement; i++) {
-		*new_flags = placement->placement[i];
+		const struct ttm_place *heap = &placement->placement[i];
+		if (mem->mm_node &&
+		    (mem->start < heap->fpfn ||
+		     (heap->lpfn != 0 && (mem->start + mem->num_pages) > heap->lpfn)))
+			continue;
+
+		*new_flags = heap->flags;
 		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
 		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
 			return true;
 	}
 
 	for (i = 0; i < placement->num_busy_placement; i++) {
-		*new_flags = placement->busy_placement[i];
+		const struct ttm_place *heap = &placement->busy_placement[i];
+		if (mem->mm_node &&
+		    (mem->start < heap->fpfn ||
+		     (heap->lpfn != 0 && (mem->start + mem->num_pages) > heap->lpfn)))
+			continue;
+
+		*new_flags = heap->flags;
 		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
 		    (*new_flags & mem->placement & TTM_PL_MASK_MEM))
 			return true;
@@ -1037,11 +1044,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 	uint32_t new_flags;
 
 	lockdep_assert_held(&bo->resv->lock.base);
-	/* Check that range is valid */
-	if (placement->lpfn || placement->fpfn)
-		if (placement->fpfn > placement->lpfn ||
-			(placement->lpfn - placement->fpfn) < bo->num_pages)
-			return -EINVAL;
 	/*
 	 * Check whether we need to move buffer.
 	 */
@@ -1070,15 +1072,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_validate);
 
-int ttm_bo_check_placement(struct ttm_buffer_object *bo,
-				struct ttm_placement *placement)
-{
-	BUG_ON((placement->fpfn || placement->lpfn) &&
-	       (bo->mem.num_pages > (placement->lpfn - placement->fpfn)));
-
-	return 0;
-}
-
 int ttm_bo_init(struct ttm_bo_device *bdev,
 		struct ttm_buffer_object *bo,
 		unsigned long size,
@@ -1089,6 +1082,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 		struct file *persistent_swap_storage,
 		size_t acc_size,
 		struct sg_table *sg,
+		struct reservation_object *resv,
 		void (*destroy) (struct ttm_buffer_object *))
 {
 	int ret = 0;
@@ -1142,30 +1136,38 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->persistent_swap_storage = persistent_swap_storage;
 	bo->acc_size = acc_size;
 	bo->sg = sg;
-	bo->resv = &bo->ttm_resv;
-	reservation_object_init(bo->resv);
+	if (resv) {
+		bo->resv = resv;
+		lockdep_assert_held(&bo->resv->lock.base);
+	} else {
+		bo->resv = &bo->ttm_resv;
+		reservation_object_init(&bo->ttm_resv);
+	}
 	atomic_inc(&bo->glob->bo_count);
 	drm_vma_node_reset(&bo->vma_node);
 
-	ret = ttm_bo_check_placement(bo, placement);
-
 	/*
 	 * For ttm_bo_type_device buffers, allocate
 	 * address space from the device.
 	 */
-	if (likely(!ret) &&
-	    (bo->type == ttm_bo_type_device ||
-	     bo->type == ttm_bo_type_sg))
+	if (bo->type == ttm_bo_type_device ||
+	    bo->type == ttm_bo_type_sg)
 		ret = drm_vma_offset_add(&bdev->vma_manager, &bo->vma_node,
 					 bo->mem.num_pages);
 
-	locked = ww_mutex_trylock(&bo->resv->lock);
-	WARN_ON(!locked);
+	/* passed reservation objects should already be locked,
+	 * since otherwise lockdep will be angered in radeon.
+	 */
+	if (!resv) {
+		locked = ww_mutex_trylock(&bo->resv->lock);
+		WARN_ON(!locked);
+	}
 
 	if (likely(!ret))
 		ret = ttm_bo_validate(bo, placement, interruptible, false);
 
-	ttm_bo_unreserve(bo);
+	if (!resv)
+		ttm_bo_unreserve(bo);
 
 	if (unlikely(ret))
 		ttm_bo_unref(&bo);
@@ -1223,7 +1225,7 @@ int ttm_bo_create(struct ttm_bo_device *bdev,
 	acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct ttm_buffer_object));
 	ret = ttm_bo_init(bdev, bo, size, type, placement, page_alignment,
 			  interruptible, persistent_swap_storage, acc_size,
-			  NULL, NULL);
+			  NULL, NULL, NULL);
 	if (likely(ret == 0))
 		*p_bo = bo;
 
@@ -1245,7 +1247,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 	spin_lock(&glob->lru_lock);
 	while (!list_empty(&man->lru)) {
 		spin_unlock(&glob->lru_lock);
-		ret = ttm_mem_evict_first(bdev, mem_type, false, false);
+		ret = ttm_mem_evict_first(bdev, mem_type, NULL, false, false);
 		if (ret) {
 			if (allow_errors) {
 				return ret;
@@ -1477,7 +1479,6 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	bdev->glob = glob;
 	bdev->need_dma32 = need_dma32;
 	bdev->val_seq = 0;
-	spin_lock_init(&bdev->fence_lock);
 	mutex_lock(&glob->device_list_mutex);
 	list_add_tail(&bdev->device_list, &glob->device_list);
 	mutex_unlock(&glob->device_list_mutex);
@@ -1530,77 +1531,66 @@ void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
 
 EXPORT_SYMBOL(ttm_bo_unmap_virtual);
 
-
 int ttm_bo_wait(struct ttm_buffer_object *bo,
 		bool lazy, bool interruptible, bool no_wait)
 {
-	struct ttm_bo_driver *driver = bo->bdev->driver;
-	struct ttm_bo_device *bdev = bo->bdev;
-	void *sync_obj;
-	int ret = 0;
-
-	if (likely(bo->sync_obj == NULL))
-		return 0;
+	struct reservation_object_list *fobj;
+	struct reservation_object *resv;
+	struct fence *excl;
+	long timeout = 15 * HZ;
+	int i;
 
-	while (bo->sync_obj) {
+	resv = bo->resv;
+	fobj = reservation_object_get_list(resv);
+	excl = reservation_object_get_excl(resv);
+	if (excl) {
+		if (!fence_is_signaled(excl)) {
+			if (no_wait)
+				return -EBUSY;
 
-		if (driver->sync_obj_signaled(bo->sync_obj)) {
-			void *tmp_obj = bo->sync_obj;
-			bo->sync_obj = NULL;
-			clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bdev->fence_lock);
-			continue;
+			timeout = fence_wait_timeout(excl,
+						     interruptible, timeout);
 		}
+	}
 
-		if (no_wait)
-			return -EBUSY;
+	for (i = 0; fobj && timeout > 0 && i < fobj->shared_count; ++i) {
+		struct fence *fence;
+		fence = rcu_dereference_protected(fobj->shared[i],
+						reservation_object_held(resv));
 
-		sync_obj = driver->sync_obj_ref(bo->sync_obj);
-		spin_unlock(&bdev->fence_lock);
-		ret = driver->sync_obj_wait(sync_obj,
-					    lazy, interruptible);
-		if (unlikely(ret != 0)) {
-			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bdev->fence_lock);
-			return ret;
-		}
-		spin_lock(&bdev->fence_lock);
-		if (likely(bo->sync_obj == sync_obj)) {
-			void *tmp_obj = bo->sync_obj;
-			bo->sync_obj = NULL;
-			clear_bit(TTM_BO_PRIV_FLAG_MOVING,
-				  &bo->priv_flags);
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&sync_obj);
-			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bdev->fence_lock);
-		} else {
-			spin_unlock(&bdev->fence_lock);
-			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bdev->fence_lock);
+		if (!fence_is_signaled(fence)) {
+			if (no_wait)
+				return -EBUSY;
+
+			timeout = fence_wait_timeout(fence,
+						     interruptible, timeout);
 		}
 	}
+
+	if (timeout < 0)
+		return timeout;
+
+	if (timeout == 0)
+		return -EBUSY;
+
+	reservation_object_add_excl_fence(resv, NULL);
+	clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
 	return 0;
 }
 EXPORT_SYMBOL(ttm_bo_wait);
 
 int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
 	int ret = 0;
 
 	/*
 	 * Using ttm_bo_reserve makes sure the lru lists are updated.
 	 */
 
-	ret = ttm_bo_reserve(bo, true, no_wait, false, 0);
+	ret = ttm_bo_reserve(bo, true, no_wait, false, NULL);
 	if (unlikely(ret != 0))
 		return ret;
-	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, true, no_wait);
-	spin_unlock(&bdev->fence_lock);
 	if (likely(ret == 0))
 		atomic_inc(&bo->cpu_writers);
 	ttm_bo_unreserve(bo);
@@ -1630,7 +1620,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 
 	spin_lock(&glob->lru_lock);
 	list_for_each_entry(bo, &glob->swap_lru, swap) {
-		ret = __ttm_bo_reserve(bo, false, true, false, 0);
+		ret = __ttm_bo_reserve(bo, false, true, false, NULL);
 		if (!ret)
 			break;
 	}
@@ -1657,9 +1647,7 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 	 * Wait for GPU, then move to system cached.
 	 */
 
-	spin_lock(&bo->bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bo->bdev->fence_lock);
 
 	if (unlikely(ret != 0))
 		goto out;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_manager.c b/drivers/gpu/drm/ttm/ttm_bo_manager.c
index bd850c9f4bca..964387fc5c8f 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_manager.c
@@ -49,7 +49,7 @@ struct ttm_range_manager {
 
 static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
 			       struct ttm_buffer_object *bo,
-			       struct ttm_placement *placement,
+			       const struct ttm_place *place,
 			       struct ttm_mem_reg *mem)
 {
 	struct ttm_range_manager *rman = (struct ttm_range_manager *) man->priv;
@@ -59,7 +59,7 @@ static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
 	unsigned long lpfn;
 	int ret;
 
-	lpfn = placement->lpfn;
+	lpfn = place->lpfn;
 	if (!lpfn)
 		lpfn = man->size;
 
@@ -67,13 +67,13 @@ static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
 	if (!node)
 		return -ENOMEM;
 
-	if (bo->mem.placement & TTM_PL_FLAG_TOPDOWN)
+	if (place->flags & TTM_PL_FLAG_TOPDOWN)
 		aflags = DRM_MM_CREATE_TOP;
 
 	spin_lock(&rman->lock);
 	ret = drm_mm_insert_node_in_range_generic(mm, node, mem->num_pages,
 					  mem->page_alignment, 0,
-					  placement->fpfn, lpfn,
+					  place->fpfn, lpfn,
 					  DRM_MM_SEARCH_BEST,
 					  aflags);
 	spin_unlock(&rman->lock);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 1df856f78568..882cccdad272 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -37,6 +37,7 @@
 #include <linux/slab.h>
 #include <linux/vmalloc.h>
 #include <linux/module.h>
+#include <linux/reservation.h>
 
 void ttm_bo_free_old_node(struct ttm_buffer_object *bo)
 {
@@ -444,8 +445,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 				      struct ttm_buffer_object **new_obj)
 {
 	struct ttm_buffer_object *fbo;
-	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
@@ -466,12 +465,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	drm_vma_node_reset(&fbo->vma_node);
 	atomic_set(&fbo->cpu_writers, 0);
 
-	spin_lock(&bdev->fence_lock);
-	if (bo->sync_obj)
-		fbo->sync_obj = driver->sync_obj_ref(bo->sync_obj);
-	else
-		fbo->sync_obj = NULL;
-	spin_unlock(&bdev->fence_lock);
 	kref_init(&fbo->list_kref);
 	kref_init(&fbo->kref);
 	fbo->destroy = &ttm_transfered_destroy;
@@ -487,28 +480,24 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 
 pgprot_t ttm_io_prot(uint32_t caching_flags, pgprot_t tmp)
 {
+	/* Cached mappings need no adjustment */
+	if (caching_flags & TTM_PL_FLAG_CACHED)
+		return tmp;
+
 #if defined(__i386__) || defined(__x86_64__)
 	if (caching_flags & TTM_PL_FLAG_WC)
 		tmp = pgprot_writecombine(tmp);
 	else if (boot_cpu_data.x86 > 3)
 		tmp = pgprot_noncached(tmp);
-
-#elif defined(__powerpc__)
-	if (!(caching_flags & TTM_PL_FLAG_CACHED)) {
-		pgprot_val(tmp) |= _PAGE_NO_CACHE;
-		if (caching_flags & TTM_PL_FLAG_UNCACHED)
-			pgprot_val(tmp) |= _PAGE_GUARDED;
-	}
 #endif
-#if defined(__ia64__)
+#if defined(__ia64__) || defined(__arm__) || defined(__powerpc__)
 	if (caching_flags & TTM_PL_FLAG_WC)
 		tmp = pgprot_writecombine(tmp);
 	else
 		tmp = pgprot_noncached(tmp);
 #endif
 #if defined(__sparc__) || defined(__mips__)
-	if (!(caching_flags & TTM_PL_FLAG_CACHED))
-		tmp = pgprot_noncached(tmp);
+	tmp = pgprot_noncached(tmp);
 #endif
 	return tmp;
 }
@@ -567,9 +556,7 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo,
 		 * We need to use vmap to get the desired page protection
 		 * or to make the buffer object look contiguous.
 		 */
-		prot = (mem->placement & TTM_PL_FLAG_CACHED) ?
-			PAGE_KERNEL :
-			ttm_io_prot(mem->placement, PAGE_KERNEL);
+		prot = ttm_io_prot(mem->placement, PAGE_KERNEL);
 		map->bo_kmap_type = ttm_bo_map_vmap;
 		map->virtual = vmap(ttm->pages + start_page, num_pages,
 				    0, prot);
@@ -644,30 +631,20 @@ void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 EXPORT_SYMBOL(ttm_bo_kunmap);
 
 int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
-			      void *sync_obj,
+			      struct fence *fence,
 			      bool evict,
 			      bool no_wait_gpu,
 			      struct ttm_mem_reg *new_mem)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
 	struct ttm_mem_type_manager *man = &bdev->man[new_mem->mem_type];
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	int ret;
 	struct ttm_buffer_object *ghost_obj;
-	void *tmp_obj = NULL;
 
-	spin_lock(&bdev->fence_lock);
-	if (bo->sync_obj) {
-		tmp_obj = bo->sync_obj;
-		bo->sync_obj = NULL;
-	}
-	bo->sync_obj = driver->sync_obj_ref(sync_obj);
+	reservation_object_add_excl_fence(bo->resv, fence);
 	if (evict) {
 		ret = ttm_bo_wait(bo, false, false, false);
-		spin_unlock(&bdev->fence_lock);
-		if (tmp_obj)
-			driver->sync_obj_unref(&tmp_obj);
 		if (ret)
 			return ret;
 
@@ -688,14 +665,13 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 		 */
 
 		set_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-		spin_unlock(&bdev->fence_lock);
-		if (tmp_obj)
-			driver->sync_obj_unref(&tmp_obj);
 
 		ret = ttm_buffer_object_transfer(bo, &ghost_obj);
 		if (ret)
 			return ret;
 
+		reservation_object_add_excl_fence(ghost_obj->resv, fence);
+
 		/**
 		 * If we're not moving to fixed memory, the TTM object
 		 * needs to stay alive. Otherwhise hang it on the ghost
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 0ce48e5a9cb4..8fb7213277cc 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -45,10 +45,8 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 				struct vm_area_struct *vma,
 				struct vm_fault *vmf)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
 	int ret = 0;
 
-	spin_lock(&bdev->fence_lock);
 	if (likely(!test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)))
 		goto out_unlock;
 
@@ -82,7 +80,6 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo,
 			VM_FAULT_NOPAGE;
 
 out_unlock:
-	spin_unlock(&bdev->fence_lock);
 	return ret;
 }
 
@@ -200,9 +197,8 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 						cvma.vm_page_prot);
 	} else {
 		ttm = bo->ttm;
-		if (!(bo->mem.placement & TTM_PL_FLAG_CACHED))
-			cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
-							cvma.vm_page_prot);
+		cvma.vm_page_prot = ttm_io_prot(bo->mem.placement,
+						cvma.vm_page_prot);
 
 		/* Allocate all page at once, most common usage */
 		if (ttm->bdev->driver->ttm_tt_populate(ttm)) {
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index e8dac8758528..8ce508e76208 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -32,20 +32,12 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 
-static void ttm_eu_backoff_reservation_locked(struct list_head *list)
+static void ttm_eu_backoff_reservation_reverse(struct list_head *list,
+					      struct ttm_validate_buffer *entry)
 {
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
+	list_for_each_entry_continue_reverse(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
-		if (!entry->reserved)
-			continue;
 
-		entry->reserved = false;
-		if (entry->removed) {
-			ttm_bo_add_to_lru(bo);
-			entry->removed = false;
-		}
 		__ttm_bo_unreserve(bo);
 	}
 }
@@ -56,27 +48,9 @@ static void ttm_eu_del_from_lru_locked(struct list_head *list)
 
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
-		if (!entry->reserved)
-			continue;
-
-		if (!entry->removed) {
-			entry->put_count = ttm_bo_del_from_lru(bo);
-			entry->removed = true;
-		}
-	}
-}
-
-static void ttm_eu_list_ref_sub(struct list_head *list)
-{
-	struct ttm_validate_buffer *entry;
-
-	list_for_each_entry(entry, list, head) {
-		struct ttm_buffer_object *bo = entry->bo;
+		unsigned put_count = ttm_bo_del_from_lru(bo);
 
-		if (entry->put_count) {
-			ttm_bo_list_ref_sub(bo, entry->put_count, true);
-			entry->put_count = 0;
-		}
+		ttm_bo_list_ref_sub(bo, put_count, true);
 	}
 }
 
@@ -91,11 +65,18 @@ void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
 
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
+
 	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
+	list_for_each_entry(entry, list, head) {
+		struct ttm_buffer_object *bo = entry->bo;
+
+		ttm_bo_add_to_lru(bo);
+		__ttm_bo_unreserve(bo);
+	}
+	spin_unlock(&glob->lru_lock);
+
 	if (ticket)
 		ww_acquire_fini(ticket);
-	spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_eu_backoff_reservation);
 
@@ -112,7 +93,7 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  */
 
 int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
-			   struct list_head *list)
+			   struct list_head *list, bool intr)
 {
 	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
@@ -121,60 +102,64 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
 	if (list_empty(list))
 		return 0;
 
-	list_for_each_entry(entry, list, head) {
-		entry->reserved = false;
-		entry->put_count = 0;
-		entry->removed = false;
-	}
-
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 
 	if (ticket)
 		ww_acquire_init(ticket, &reservation_ww_class);
-retry:
+
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
 
-		/* already slowpath reserved? */
-		if (entry->reserved)
-			continue;
-
-		ret = __ttm_bo_reserve(bo, true, (ticket == NULL), true,
+		ret = __ttm_bo_reserve(bo, intr, (ticket == NULL), true,
 				       ticket);
+		if (!ret && unlikely(atomic_read(&bo->cpu_writers) > 0)) {
+			__ttm_bo_unreserve(bo);
+
+			ret = -EBUSY;
+		}
 
-		if (ret == -EDEADLK) {
-			/* uh oh, we lost out, drop every reservation and try
-			 * to only reserve this buffer, then start over if
-			 * this succeeds.
-			 */
-			BUG_ON(ticket == NULL);
-			spin_lock(&glob->lru_lock);
-			ttm_eu_backoff_reservation_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ttm_eu_list_ref_sub(list);
+		if (!ret) {
+			if (!entry->shared)
+				continue;
+
+			ret = reservation_object_reserve_shared(bo->resv);
+			if (!ret)
+				continue;
+		}
+
+		/* uh oh, we lost out, drop every reservation and try
+		 * to only reserve this buffer, then start over if
+		 * this succeeds.
+		 */
+		ttm_eu_backoff_reservation_reverse(list, entry);
+
+		if (ret == -EDEADLK && intr) {
 			ret = ww_mutex_lock_slow_interruptible(&bo->resv->lock,
 							       ticket);
-			if (unlikely(ret != 0)) {
-				if (ret == -EINTR)
-					ret = -ERESTARTSYS;
-				goto err_fini;
-			}
+		} else if (ret == -EDEADLK) {
+			ww_mutex_lock_slow(&bo->resv->lock, ticket);
+			ret = 0;
+		}
 
-			entry->reserved = true;
-			if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-				ret = -EBUSY;
-				goto err;
-			}
-			goto retry;
-		} else if (ret)
-			goto err;
+		if (!ret && entry->shared)
+			ret = reservation_object_reserve_shared(bo->resv);
 
-		entry->reserved = true;
-		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-			ret = -EBUSY;
-			goto err;
+		if (unlikely(ret != 0)) {
+			if (ret == -EINTR)
+				ret = -ERESTARTSYS;
+			if (ticket) {
+				ww_acquire_done(ticket);
+				ww_acquire_fini(ticket);
+			}
+			return ret;
 		}
+
+		/* move this item to the front of the list,
+		 * forces correct iteration of the loop without keeping track
+		 */
+		list_del(&entry->head);
+		list_add(&entry->head, list);
 	}
 
 	if (ticket)
@@ -182,25 +167,12 @@ retry:
 	spin_lock(&glob->lru_lock);
 	ttm_eu_del_from_lru_locked(list);
 	spin_unlock(&glob->lru_lock);
-	ttm_eu_list_ref_sub(list);
 	return 0;
-
-err:
-	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
-	spin_unlock(&glob->lru_lock);
-	ttm_eu_list_ref_sub(list);
-err_fini:
-	if (ticket) {
-		ww_acquire_done(ticket);
-		ww_acquire_fini(ticket);
-	}
-	return ret;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 
 void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
-				 struct list_head *list, void *sync_obj)
+				 struct list_head *list, struct fence *fence)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
@@ -217,24 +189,18 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
 	glob = bo->glob;
 
 	spin_lock(&glob->lru_lock);
-	spin_lock(&bdev->fence_lock);
 
 	list_for_each_entry(entry, list, head) {
 		bo = entry->bo;
-		entry->old_sync_obj = bo->sync_obj;
-		bo->sync_obj = driver->sync_obj_ref(sync_obj);
+		if (entry->shared)
+			reservation_object_add_shared_fence(bo->resv, fence);
+		else
+			reservation_object_add_excl_fence(bo->resv, fence);
 		ttm_bo_add_to_lru(bo);
 		__ttm_bo_unreserve(bo);
-		entry->reserved = false;
 	}
-	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
 	if (ticket)
 		ww_acquire_fini(ticket);
-
-	list_for_each_entry(entry, list, head) {
-		if (entry->old_sync_obj)
-			driver->sync_obj_unref(&entry->old_sync_obj);
-	}
 }
 EXPORT_SYMBOL(ttm_eu_fence_buffer_objects);
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index dbc2def887cd..a1803fbcc898 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -300,7 +300,8 @@ static int ttm_mem_init_highmem_zone(struct ttm_mem_global *glob,
 	zone->glob = glob;
 	glob->zone_highmem = zone;
 	ret = kobject_init_and_add(
-		&zone->kobj, &ttm_mem_zone_kobj_type, &glob->kobj, zone->name);
+		&zone->kobj, &ttm_mem_zone_kobj_type, &glob->kobj, "%s",
+		zone->name);
 	if (unlikely(ret != 0)) {
 		kobject_put(&zone->kobj);
 		return ret;
diff --git a/drivers/gpu/drm/ttm/ttm_module.c b/drivers/gpu/drm/ttm/ttm_module.c
index d7f92fe9d904..66fc6395eb54 100644
--- a/drivers/gpu/drm/ttm/ttm_module.c
+++ b/drivers/gpu/drm/ttm/ttm_module.c
@@ -35,7 +35,7 @@
 #include <drm/drm_sysfs.h>
 
 static DECLARE_WAIT_QUEUE_HEAD(exit_q);
-atomic_t device_released;
+static atomic_t device_released;
 
 static struct device_type ttm_drm_class_type = {
 	.name = "ttm",
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 863bef9f9234..09874d695188 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -297,8 +297,10 @@ static void ttm_pool_update_free_locked(struct ttm_page_pool *pool,
  *
  * @pool: to free the pages from
  * @free_all: If set to true will free all pages in pool
+ * @gfp: GFP flags.
  **/
-static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free)
+static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free,
+			      gfp_t gfp)
 {
 	unsigned long irq_flags;
 	struct page *p;
@@ -309,8 +311,7 @@ static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free)
 	if (NUM_PAGES_TO_ALLOC < nr_free)
 		npages_to_free = NUM_PAGES_TO_ALLOC;
 
-	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
-			GFP_KERNEL);
+	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
 	if (!pages_to_free) {
 		pr_err("Failed to allocate memory for pool free operation\n");
 		return 0;
@@ -382,32 +383,35 @@ out:
  *
  * XXX: (dchinner) Deadlock warning!
  *
- * ttm_page_pool_free() does memory allocation using GFP_KERNEL.  that means
- * this can deadlock when called a sc->gfp_mask that is not equal to
- * GFP_KERNEL.
+ * We need to pass sc->gfp_mask to ttm_page_pool_free().
  *
  * This code is crying out for a shrinker per pool....
  */
 static unsigned long
 ttm_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 {
-	static atomic_t start_pool = ATOMIC_INIT(0);
+	static DEFINE_MUTEX(lock);
+	static unsigned start_pool;
 	unsigned i;
-	unsigned pool_offset = atomic_add_return(1, &start_pool);
+	unsigned pool_offset;
 	struct ttm_page_pool *pool;
 	int shrink_pages = sc->nr_to_scan;
 	unsigned long freed = 0;
 
-	pool_offset = pool_offset % NUM_POOLS;
+	if (!mutex_trylock(&lock))
+		return SHRINK_STOP;
+	pool_offset = ++start_pool % NUM_POOLS;
 	/* select start pool in round robin fashion */
 	for (i = 0; i < NUM_POOLS; ++i) {
 		unsigned nr_free = shrink_pages;
 		if (shrink_pages == 0)
 			break;
 		pool = &_manager->pools[(i + pool_offset)%NUM_POOLS];
-		shrink_pages = ttm_page_pool_free(pool, nr_free);
+		shrink_pages = ttm_page_pool_free(pool, nr_free,
+						  sc->gfp_mask);
 		freed += nr_free - shrink_pages;
 	}
+	mutex_unlock(&lock);
 	return freed;
 }
 
@@ -706,7 +710,7 @@ static void ttm_put_pages(struct page **pages, unsigned npages, int flags,
 	}
 	spin_unlock_irqrestore(&pool->lock, irq_flags);
 	if (npages)
-		ttm_page_pool_free(pool, npages);
+		ttm_page_pool_free(pool, npages, GFP_KERNEL);
 }
 
 /*
@@ -790,7 +794,7 @@ static int ttm_get_pages(struct page **pages, unsigned npages, int flags,
 	return 0;
 }
 
-static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags,
+static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, gfp_t flags,
 		char *name)
 {
 	spin_lock_init(&pool->lock);
@@ -846,7 +850,8 @@ void ttm_page_alloc_fini(void)
 	ttm_pool_mm_shrink_fini(_manager);
 
 	for (i = 0; i < NUM_POOLS; ++i)
-		ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES);
+		ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES,
+				   GFP_KERNEL);
 
 	kobject_put(&_manager->kobj);
 	_manager = NULL;
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
index fb8259f69839..c96db433f8af 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c
@@ -411,8 +411,10 @@ static void ttm_dma_page_put(struct dma_pool *pool, struct dma_page *d_page)
  *
  * @pool: to free the pages from
  * @nr_free: If set to true will free all pages in pool
+ * @gfp: GFP flags.
  **/
-static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free)
+static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free,
+				       gfp_t gfp)
 {
 	unsigned long irq_flags;
 	struct dma_page *dma_p, *tmp;
@@ -430,8 +432,7 @@ static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free)
 			 npages_to_free, nr_free);
 	}
 #endif
-	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *),
-			GFP_KERNEL);
+	pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), gfp);
 
 	if (!pages_to_free) {
 		pr_err("%s: Failed to allocate memory for pool free operation\n",
@@ -530,7 +531,7 @@ static void ttm_dma_free_pool(struct device *dev, enum pool_type type)
 		if (pool->type != type)
 			continue;
 		/* Takes a spinlock.. */
-		ttm_dma_page_pool_free(pool, FREE_ALL_PAGES);
+		ttm_dma_page_pool_free(pool, FREE_ALL_PAGES, GFP_KERNEL);
 		WARN_ON(((pool->npages_in_use + pool->npages_free) != 0));
 		/* This code path is called after _all_ references to the
 		 * struct device has been dropped - so nobody should be
@@ -847,6 +848,7 @@ static int ttm_dma_pool_get_pages(struct dma_pool *pool,
 	if (count) {
 		d_page = list_first_entry(&pool->free_list, struct dma_page, page_list);
 		ttm->pages[index] = d_page->p;
+		ttm_dma->cpu_address[index] = d_page->vaddr;
 		ttm_dma->dma_address[index] = d_page->dma;
 		list_move_tail(&d_page->page_list, &ttm_dma->pages_list);
 		r = 0;
@@ -978,12 +980,13 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	for (i = 0; i < ttm->num_pages; i++) {
 		ttm->pages[i] = NULL;
+		ttm_dma->cpu_address[i] = 0;
 		ttm_dma->dma_address[i] = 0;
 	}
 
 	/* shrink pool if necessary (only on !is_cached pools)*/
 	if (npages)
-		ttm_dma_page_pool_free(pool, npages);
+		ttm_dma_page_pool_free(pool, npages, GFP_KERNEL);
 	ttm->state = tt_unpopulated;
 }
 EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
@@ -993,10 +996,7 @@ EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
  *
  * XXX: (dchinner) Deadlock warning!
  *
- * ttm_dma_page_pool_free() does GFP_KERNEL memory allocation, and so attention
- * needs to be paid to sc->gfp_mask to determine if this can be done or not.
- * GFP_KERNEL memory allocation in a GFP_ATOMIC reclaim context woul dbe really
- * bad.
+ * We need to pass sc->gfp_mask to ttm_dma_page_pool_free().
  *
  * I'm getting sadder as I hear more pathetical whimpers about needing per-pool
  * shrinkers
@@ -1004,9 +1004,9 @@ EXPORT_SYMBOL_GPL(ttm_dma_unpopulate);
 static unsigned long
 ttm_dma_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 {
-	static atomic_t start_pool = ATOMIC_INIT(0);
+	static unsigned start_pool;
 	unsigned idx = 0;
-	unsigned pool_offset = atomic_add_return(1, &start_pool);
+	unsigned pool_offset;
 	unsigned shrink_pages = sc->nr_to_scan;
 	struct device_pools *p;
 	unsigned long freed = 0;
@@ -1014,8 +1014,11 @@ ttm_dma_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 	if (list_empty(&_manager->pools))
 		return SHRINK_STOP;
 
-	mutex_lock(&_manager->lock);
-	pool_offset = pool_offset % _manager->npools;
+	if (!mutex_trylock(&_manager->lock))
+		return SHRINK_STOP;
+	if (!_manager->npools)
+		goto out;
+	pool_offset = ++start_pool % _manager->npools;
 	list_for_each_entry(p, &_manager->pools, pools) {
 		unsigned nr_free;
 
@@ -1027,13 +1030,15 @@ ttm_dma_pool_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 		if (++idx < pool_offset)
 			continue;
 		nr_free = shrink_pages;
-		shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free);
+		shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free,
+						      sc->gfp_mask);
 		freed += nr_free - shrink_pages;
 
 		pr_debug("%s: (%s:%d) Asked to shrink %d, have %d more to go\n",
 			 p->pool->dev_name, p->pool->name, current->pid,
 			 nr_free, shrink_pages);
 	}
+out:
 	mutex_unlock(&_manager->lock);
 	return freed;
 }
@@ -1044,7 +1049,8 @@ ttm_dma_pool_shrink_count(struct shrinker *shrink, struct shrink_control *sc)
 	struct device_pools *p;
 	unsigned long count = 0;
 
-	mutex_lock(&_manager->lock);
+	if (!mutex_trylock(&_manager->lock))
+		return 0;
 	list_for_each_entry(p, &_manager->pools, pools)
 		count += p->pool->npages_free;
 	mutex_unlock(&_manager->lock);
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index 75f319090043..bf080abc86d1 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -55,9 +55,12 @@ static void ttm_tt_alloc_page_directory(struct ttm_tt *ttm)
 
 static void ttm_dma_tt_alloc_page_directory(struct ttm_dma_tt *ttm)
 {
-	ttm->ttm.pages = drm_calloc_large(ttm->ttm.num_pages, sizeof(void*));
-	ttm->dma_address = drm_calloc_large(ttm->ttm.num_pages,
-					    sizeof(*ttm->dma_address));
+	ttm->ttm.pages = drm_calloc_large(ttm->ttm.num_pages,
+					  sizeof(*ttm->ttm.pages) +
+					  sizeof(*ttm->dma_address) +
+					  sizeof(*ttm->cpu_address));
+	ttm->cpu_address = (void *) (ttm->ttm.pages + ttm->ttm.num_pages);
+	ttm->dma_address = (void *) (ttm->cpu_address + ttm->ttm.num_pages);
 }
 
 #ifdef CONFIG_X86
@@ -228,7 +231,7 @@ int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_bo_device *bdev,
 
 	INIT_LIST_HEAD(&ttm_dma->pages_list);
 	ttm_dma_tt_alloc_page_directory(ttm_dma);
-	if (!ttm->pages || !ttm_dma->dma_address) {
+	if (!ttm->pages) {
 		ttm_tt_destroy(ttm);
 		pr_err("Failed allocating page table\n");
 		return -ENOMEM;
@@ -243,7 +246,7 @@ void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma)
 
 	drm_free_large(ttm->pages);
 	ttm->pages = NULL;
-	drm_free_large(ttm_dma->dma_address);
+	ttm_dma->cpu_address = NULL;
 	ttm_dma->dma_address = NULL;
 }
 EXPORT_SYMBOL(ttm_dma_tt_fini);
diff --git a/drivers/gpu/drm/udl/Kconfig b/drivers/gpu/drm/udl/Kconfig
index f02528686cd5..613ab0622d6e 100644
--- a/drivers/gpu/drm/udl/Kconfig
+++ b/drivers/gpu/drm/udl/Kconfig
@@ -1,8 +1,9 @@
 config DRM_UDL
 	tristate "DisplayLink"
 	depends on DRM
+	depends on USB_SUPPORT
 	depends on USB_ARCH_HAS_HCD
-	select DRM_USB
+	select USB
 	select FB_SYS_FILLRECT
 	select FB_SYS_COPYAREA
 	select FB_SYS_IMAGEBLIT
diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c
index b44d548c56f8..0110d95522f3 100644
--- a/drivers/gpu/drm/udl/udl_connector.c
+++ b/drivers/gpu/drm/udl/udl_connector.c
@@ -34,8 +34,8 @@ static u8 *udl_get_edid(struct udl_device *udl)
 		goto error;
 
 	for (i = 0; i < EDID_LENGTH; i++) {
-		ret = usb_control_msg(udl->ddev->usbdev,
-				      usb_rcvctrlpipe(udl->ddev->usbdev, 0), (0x02),
+		ret = usb_control_msg(udl->udev,
+				      usb_rcvctrlpipe(udl->udev, 0), (0x02),
 				      (0x80 | (0x02 << 5)), i << 8, 0xA1, rbuf, 2,
 				      HZ);
 		if (ret < 1) {
@@ -105,14 +105,7 @@ static struct drm_encoder*
 udl_best_single_encoder(struct drm_connector *connector)
 {
 	int enc_id = connector->encoder_ids[0];
-	struct drm_mode_object *obj;
-	struct drm_encoder *encoder;
-
-	obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER);
-	if (!obj)
-		return NULL;
-	encoder = obj_to_encoder(obj);
-	return encoder;
+	return drm_encoder_find(connector->dev, enc_id);
 }
 
 static int udl_connector_set_property(struct drm_connector *connector,
@@ -124,7 +117,7 @@ static int udl_connector_set_property(struct drm_connector *connector,
 
 static void udl_connector_destroy(struct drm_connector *connector)
 {
-	drm_sysfs_connector_remove(connector);
+	drm_connector_unregister(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
 }
@@ -154,7 +147,7 @@ int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder)
 	drm_connector_init(dev, connector, &udl_connector_funcs, DRM_MODE_CONNECTOR_DVII);
 	drm_connector_helper_add(connector, &udl_connector_helper_funcs);
 
-	drm_sysfs_connector_add(connector);
+	drm_connector_register(connector);
 	drm_mode_connector_attach_encoder(connector, encoder);
 
 	drm_object_attach_property(&connector->base,
diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c
index 3ddd6cd98ac1..8607e9e513db 100644
--- a/drivers/gpu/drm/udl/udl_drv.c
+++ b/drivers/gpu/drm/udl/udl_drv.c
@@ -7,48 +7,13 @@
  */
 
 #include <linux/module.h>
-#include <drm/drm_usb.h>
+#include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include "udl_drv.h"
 
-static struct drm_driver driver;
-
-/*
- * There are many DisplayLink-based graphics products, all with unique PIDs.
- * So we match on DisplayLink's VID + Vendor-Defined Interface Class (0xff)
- * We also require a match on SubClass (0x00) and Protocol (0x00),
- * which is compatible with all known USB 2.0 era graphics chips and firmware,
- * but allows DisplayLink to increment those for any future incompatible chips
- */
-static struct usb_device_id id_table[] = {
-	{.idVendor = 0x17e9, .bInterfaceClass = 0xff,
-	 .bInterfaceSubClass = 0x00,
-	 .bInterfaceProtocol = 0x00,
-	 .match_flags = USB_DEVICE_ID_MATCH_VENDOR |
-			USB_DEVICE_ID_MATCH_INT_CLASS |
-			USB_DEVICE_ID_MATCH_INT_SUBCLASS |
-			USB_DEVICE_ID_MATCH_INT_PROTOCOL,},
-	{},
-};
-MODULE_DEVICE_TABLE(usb, id_table);
-
-MODULE_LICENSE("GPL");
-
-static int udl_usb_probe(struct usb_interface *interface,
-			 const struct usb_device_id *id)
+static int udl_driver_set_busid(struct drm_device *d, struct drm_master *m)
 {
-	return drm_get_usb_dev(interface, id, &driver);
-}
-
-static void udl_usb_disconnect(struct usb_interface *interface)
-{
-	struct drm_device *dev = usb_get_intfdata(interface);
-
-	drm_kms_helper_poll_disable(dev);
-	drm_connector_unplug_all(dev);
-	udl_fbdev_unplug(dev);
-	udl_drop_usb(dev);
-	drm_unplug_dev(dev);
+	return 0;
 }
 
 static const struct vm_operations_struct udl_gem_vm_ops = {
@@ -75,6 +40,7 @@ static struct drm_driver driver = {
 	.driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME,
 	.load = udl_driver_load,
 	.unload = udl_driver_unload,
+	.set_busid = udl_driver_set_busid,
 
 	/* gem hooks */
 	.gem_free_object = udl_gem_free_object,
@@ -96,6 +62,61 @@ static struct drm_driver driver = {
 	.patchlevel = DRIVER_PATCHLEVEL,
 };
 
+static int udl_usb_probe(struct usb_interface *interface,
+			 const struct usb_device_id *id)
+{
+	struct usb_device *udev = interface_to_usbdev(interface);
+	struct drm_device *dev;
+	int r;
+
+	dev = drm_dev_alloc(&driver, &interface->dev);
+	if (!dev)
+		return -ENOMEM;
+
+	r = drm_dev_register(dev, (unsigned long)udev);
+	if (r)
+		goto err_free;
+
+	usb_set_intfdata(interface, dev);
+	DRM_INFO("Initialized udl on minor %d\n", dev->primary->index);
+
+	return 0;
+
+err_free:
+	drm_dev_unref(dev);
+	return r;
+}
+
+static void udl_usb_disconnect(struct usb_interface *interface)
+{
+	struct drm_device *dev = usb_get_intfdata(interface);
+
+	drm_kms_helper_poll_disable(dev);
+	drm_connector_unplug_all(dev);
+	udl_fbdev_unplug(dev);
+	udl_drop_usb(dev);
+	drm_unplug_dev(dev);
+}
+
+/*
+ * There are many DisplayLink-based graphics products, all with unique PIDs.
+ * So we match on DisplayLink's VID + Vendor-Defined Interface Class (0xff)
+ * We also require a match on SubClass (0x00) and Protocol (0x00),
+ * which is compatible with all known USB 2.0 era graphics chips and firmware,
+ * but allows DisplayLink to increment those for any future incompatible chips
+ */
+static struct usb_device_id id_table[] = {
+	{.idVendor = 0x17e9, .bInterfaceClass = 0xff,
+	 .bInterfaceSubClass = 0x00,
+	 .bInterfaceProtocol = 0x00,
+	 .match_flags = USB_DEVICE_ID_MATCH_VENDOR |
+			USB_DEVICE_ID_MATCH_INT_CLASS |
+			USB_DEVICE_ID_MATCH_INT_SUBCLASS |
+			USB_DEVICE_ID_MATCH_INT_PROTOCOL,},
+	{},
+};
+MODULE_DEVICE_TABLE(usb, id_table);
+
 static struct usb_driver udl_driver = {
 	.name = "udl",
 	.probe = udl_usb_probe,
@@ -105,13 +126,14 @@ static struct usb_driver udl_driver = {
 
 static int __init udl_init(void)
 {
-	return drm_usb_init(&driver, &udl_driver);
+	return usb_register(&udl_driver);
 }
 
 static void __exit udl_exit(void)
 {
-	drm_usb_exit(&driver, &udl_driver);
+	usb_deregister(&udl_driver);
 }
 
 module_init(udl_init);
 module_exit(udl_exit);
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h
index 1fbf7b357f16..c7490a2489a7 100644
--- a/drivers/gpu/drm/udl/udl_drv.h
+++ b/drivers/gpu/drm/udl/udl_drv.h
@@ -15,6 +15,7 @@
 #define UDL_DRV_H
 
 #include <linux/usb.h>
+#include <drm/drm_gem.h>
 
 #define DRIVER_NAME		"udl"
 #define DRIVER_DESC		"DisplayLink"
@@ -47,6 +48,7 @@ struct udl_fbdev;
 struct udl_device {
 	struct device *dev;
 	struct drm_device *ddev;
+	struct usb_device *udev;
 
 	int sku_pixel_limit;
 
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index 377176372da8..8cbcb4589bd3 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -472,7 +472,8 @@ udl_framebuffer_init(struct drm_device *dev,
 static int udlfb_create(struct drm_fb_helper *helper,
 			struct drm_fb_helper_surface_size *sizes)
 {
-	struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper;
+	struct udl_fbdev *ufbdev =
+		container_of(helper, struct udl_fbdev, helper);
 	struct drm_device *dev = ufbdev->helper.dev;
 	struct fb_info *info;
 	struct device *device = dev->dev;
@@ -550,7 +551,7 @@ out:
 	return ret;
 }
 
-static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
+static const struct drm_fb_helper_funcs udl_fb_helper_funcs = {
 	.fb_probe = udlfb_create,
 };
 
@@ -583,7 +584,8 @@ int udl_fbdev_init(struct drm_device *dev)
 		return -ENOMEM;
 
 	udl->fbdev = ufbdev;
-	ufbdev->helper.funcs = &udl_fb_helper_funcs;
+
+	drm_fb_helper_prepare(dev, &ufbdev->helper, &udl_fb_helper_funcs);
 
 	ret = drm_fb_helper_init(dev, &ufbdev->helper,
 				 1, 1);
diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c
index c041cd73f399..8044f5fb7c49 100644
--- a/drivers/gpu/drm/udl/udl_gem.c
+++ b/drivers/gpu/drm/udl/udl_gem.c
@@ -107,14 +107,14 @@ int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	}
 }
 
-static int udl_gem_get_pages(struct udl_gem_object *obj, gfp_t gfpmask)
+static int udl_gem_get_pages(struct udl_gem_object *obj)
 {
 	struct page **pages;
 
 	if (obj->pages)
 		return 0;
 
-	pages = drm_gem_get_pages(&obj->base, gfpmask);
+	pages = drm_gem_get_pages(&obj->base);
 	if (IS_ERR(pages))
 		return PTR_ERR(pages);
 
@@ -147,7 +147,7 @@ int udl_gem_vmap(struct udl_gem_object *obj)
 		return 0;
 	}
 		
-	ret = udl_gem_get_pages(obj, GFP_KERNEL);
+	ret = udl_gem_get_pages(obj);
 	if (ret)
 		return ret;
 
@@ -205,7 +205,7 @@ int udl_gem_mmap(struct drm_file *file, struct drm_device *dev,
 	}
 	gobj = to_udl_bo(obj);
 
-	ret = udl_gem_get_pages(gobj, GFP_KERNEL);
+	ret = udl_gem_get_pages(gobj);
 	if (ret)
 		goto out;
 	ret = drm_gem_create_mmap_offset(obj);
diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c
index 7094b92d1ec7..33dbfb2c4748 100644
--- a/drivers/gpu/drm/udl/udl_main.c
+++ b/drivers/gpu/drm/udl/udl_main.c
@@ -202,7 +202,7 @@ static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
 		}
 		unode->urb = urb;
 
-		buf = usb_alloc_coherent(udl->ddev->usbdev, MAX_TRANSFER, GFP_KERNEL,
+		buf = usb_alloc_coherent(udl->udev, MAX_TRANSFER, GFP_KERNEL,
 					 &urb->transfer_dma);
 		if (!buf) {
 			kfree(unode);
@@ -211,7 +211,7 @@ static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size)
 		}
 
 		/* urb->transfer_buffer_length set to actual before submit */
-		usb_fill_bulk_urb(urb, udl->ddev->usbdev, usb_sndbulkpipe(udl->ddev->usbdev, 1),
+		usb_fill_bulk_urb(urb, udl->udev, usb_sndbulkpipe(udl->udev, 1),
 			buf, size, udl_urb_completion, unode);
 		urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
 
@@ -282,6 +282,7 @@ int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len)
 
 int udl_driver_load(struct drm_device *dev, unsigned long flags)
 {
+	struct usb_device *udev = (void*)flags;
 	struct udl_device *udl;
 	int ret = -ENOMEM;
 
@@ -290,10 +291,11 @@ int udl_driver_load(struct drm_device *dev, unsigned long flags)
 	if (!udl)
 		return -ENOMEM;
 
+	udl->udev = udev;
 	udl->ddev = dev;
 	dev->dev_private = udl;
 
-	if (!udl_parse_vendor_descriptor(dev, dev->usbdev)) {
+	if (!udl_parse_vendor_descriptor(dev, udl->udev)) {
 		ret = -ENODEV;
 		DRM_ERROR("firmware not recognized. Assume incompatible device\n");
 		goto err;
@@ -306,10 +308,23 @@ int udl_driver_load(struct drm_device *dev, unsigned long flags)
 
 	DRM_DEBUG("\n");
 	ret = udl_modeset_init(dev);
+	if (ret)
+		goto err;
 
 	ret = udl_fbdev_init(dev);
+	if (ret)
+		goto err;
+
+	ret = drm_vblank_init(dev, 1);
+	if (ret)
+		goto err_fb;
+
 	return 0;
+err_fb:
+	udl_fbdev_cleanup(dev);
 err:
+	if (udl->urbs.count)
+		udl_free_urb_list(dev);
 	kfree(udl);
 	DRM_ERROR("%d\n", ret);
 	return ret;
@@ -325,6 +340,8 @@ int udl_driver_unload(struct drm_device *dev)
 {
 	struct udl_device *udl = dev->dev_private;
 
+	drm_vblank_cleanup(dev);
+
 	if (udl->urbs.count)
 		udl_free_urb_list(dev);
 
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
index cddc4fcf35cf..dc145d320b25 100644
--- a/drivers/gpu/drm/udl/udl_modeset.c
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -363,6 +363,26 @@ static void udl_crtc_destroy(struct drm_crtc *crtc)
 	kfree(crtc);
 }
 
+static int udl_crtc_page_flip(struct drm_crtc *crtc,
+			      struct drm_framebuffer *fb,
+			      struct drm_pending_vblank_event *event,
+			      uint32_t page_flip_flags)
+{
+	struct udl_framebuffer *ufb = to_udl_fb(fb);
+	struct drm_device *dev = crtc->dev;
+	unsigned long flags;
+
+	udl_handle_damage(ufb, 0, 0, fb->width, fb->height);
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (event)
+		drm_send_vblank_event(dev, 0, event);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+	crtc->primary->fb = fb;
+
+	return 0;
+}
+
 static void udl_crtc_prepare(struct drm_crtc *crtc)
 {
 }
@@ -384,6 +404,7 @@ static struct drm_crtc_helper_funcs udl_helper_funcs = {
 static const struct drm_crtc_funcs udl_crtc_funcs = {
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = udl_crtc_destroy,
+	.page_flip = udl_crtc_page_flip,
 };
 
 static int udl_crtc_init(struct drm_device *dev)
diff --git a/drivers/gpu/drm/via/via_dma.c b/drivers/gpu/drm/via/via_dma.c
index 6fc0648dd37f..d17d8f245c1a 100644
--- a/drivers/gpu/drm/via/via_dma.c
+++ b/drivers/gpu/drm/via/via_dma.c
@@ -161,7 +161,7 @@ int via_dma_cleanup(struct drm_device *dev)
 		if (dev_priv->ring.virtual_start) {
 			via_cmdbuf_reset(dev_priv);
 
-			drm_core_ioremapfree(&dev_priv->ring.map, dev);
+			drm_legacy_ioremapfree(&dev_priv->ring.map, dev);
 			dev_priv->ring.virtual_start = NULL;
 		}
 
@@ -200,7 +200,7 @@ static int via_initialize(struct drm_device *dev,
 	dev_priv->ring.map.flags = 0;
 	dev_priv->ring.map.mtrr = 0;
 
-	drm_core_ioremap(&dev_priv->ring.map, dev);
+	drm_legacy_ioremap(&dev_priv->ring.map, dev);
 
 	if (dev_priv->ring.map.handle == NULL) {
 		via_dma_cleanup(dev);
diff --git a/drivers/gpu/drm/via/via_drv.c b/drivers/gpu/drm/via/via_drv.c
index 50abc2adfaee..ed8aa8ff861a 100644
--- a/drivers/gpu/drm/via/via_drv.c
+++ b/drivers/gpu/drm/via/via_drv.c
@@ -62,7 +62,7 @@ static const struct file_operations via_driver_fops = {
 	.open = drm_open,
 	.release = drm_release,
 	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
+	.mmap = drm_legacy_mmap,
 	.poll = drm_poll,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl = drm_compat_ioctl,
@@ -79,6 +79,7 @@ static struct drm_driver driver = {
 	.open = via_driver_open,
 	.preclose = via_reclaim_buffers_locked,
 	.postclose = via_driver_postclose,
+	.set_busid = drm_pci_set_busid,
 	.context_dtor = via_final_context,
 	.get_vblank_counter = via_get_vblank_counter,
 	.enable_vblank = via_enable_vblank,
diff --git a/drivers/gpu/drm/via/via_drv.h b/drivers/gpu/drm/via/via_drv.h
index ad0273256beb..ef8c500b4a00 100644
--- a/drivers/gpu/drm/via/via_drv.h
+++ b/drivers/gpu/drm/via/via_drv.h
@@ -25,6 +25,8 @@
 #define _VIA_DRV_H_
 
 #include <drm/drm_mm.h>
+#include <drm/drm_legacy.h>
+
 #define DRIVER_AUTHOR	"Various"
 
 #define DRIVER_NAME		"via"
diff --git a/drivers/gpu/drm/via/via_map.c b/drivers/gpu/drm/via/via_map.c
index d0ab3fb32acd..0b3522dba6e8 100644
--- a/drivers/gpu/drm/via/via_map.c
+++ b/drivers/gpu/drm/via/via_map.c
@@ -31,7 +31,7 @@ static int via_do_init_map(struct drm_device *dev, drm_via_init_t *init)
 
 	DRM_DEBUG("\n");
 
-	dev_priv->sarea = drm_getsarea(dev);
+	dev_priv->sarea = drm_legacy_getsarea(dev);
 	if (!dev_priv->sarea) {
 		DRM_ERROR("could not find sarea!\n");
 		dev->dev_private = (void *)dev_priv;
@@ -39,14 +39,14 @@ static int via_do_init_map(struct drm_device *dev, drm_via_init_t *init)
 		return -EINVAL;
 	}
 
-	dev_priv->fb = drm_core_findmap(dev, init->fb_offset);
+	dev_priv->fb = drm_legacy_findmap(dev, init->fb_offset);
 	if (!dev_priv->fb) {
 		DRM_ERROR("could not find framebuffer!\n");
 		dev->dev_private = (void *)dev_priv;
 		via_do_cleanup_map(dev);
 		return -EINVAL;
 	}
-	dev_priv->mmio = drm_core_findmap(dev, init->mmio_offset);
+	dev_priv->mmio = drm_legacy_findmap(dev, init->mmio_offset);
 	if (!dev_priv->mmio) {
 		DRM_ERROR("could not find mmio region!\n");
 		dev->dev_private = (void *)dev_priv;
diff --git a/drivers/gpu/drm/via/via_mm.c b/drivers/gpu/drm/via/via_mm.c
index d70b1e1544bf..4f20742e7788 100644
--- a/drivers/gpu/drm/via/via_mm.c
+++ b/drivers/gpu/drm/via/via_mm.c
@@ -211,12 +211,12 @@ void via_reclaim_buffers_locked(struct drm_device *dev,
 	if (!(file->minor->master && file->master->lock.hw_lock))
 		return;
 
-	drm_idlelock_take(&file->master->lock);
+	drm_legacy_idlelock_take(&file->master->lock);
 
 	mutex_lock(&dev->struct_mutex);
 	if (list_empty(&file_priv->obj_list)) {
 		mutex_unlock(&dev->struct_mutex);
-		drm_idlelock_release(&file->master->lock);
+		drm_legacy_idlelock_release(&file->master->lock);
 
 		return;
 	}
@@ -231,7 +231,7 @@ void via_reclaim_buffers_locked(struct drm_device *dev,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	drm_idlelock_release(&file->master->lock);
+	drm_legacy_idlelock_release(&file->master->lock);
 
 	return;
 }
diff --git a/drivers/gpu/drm/via/via_verifier.c b/drivers/gpu/drm/via/via_verifier.c
index 9dbc92bd1512..0677bbf4ec7e 100644
--- a/drivers/gpu/drm/via/via_verifier.c
+++ b/drivers/gpu/drm/via/via_verifier.c
@@ -31,6 +31,7 @@
 #include "via_3d_reg.h"
 #include <drm/drmP.h>
 #include <drm/via_drm.h>
+#include <drm/drm_legacy.h>
 #include "via_verifier.h"
 #include "via_drv.h"
 
diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile
index 458cdf6d81e8..ce0ab951f507 100644
--- a/drivers/gpu/drm/vmwgfx/Makefile
+++ b/drivers/gpu/drm/vmwgfx/Makefile
@@ -6,6 +6,7 @@ vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \
 	    vmwgfx_fifo.o vmwgfx_irq.o vmwgfx_ldu.o vmwgfx_ttm_glue.o \
 	    vmwgfx_overlay.o vmwgfx_marker.o vmwgfx_gmrid_manager.o \
 	    vmwgfx_fence.o vmwgfx_dmabuf.o vmwgfx_scrn.o vmwgfx_context.o \
-	    vmwgfx_surface.o vmwgfx_prime.o vmwgfx_mob.o vmwgfx_shader.o
+	    vmwgfx_surface.o vmwgfx_prime.o vmwgfx_mob.o vmwgfx_shader.o \
+	    vmwgfx_cmdbuf_res.o \
 
 obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o
diff --git a/drivers/gpu/drm/vmwgfx/svga_reg.h b/drivers/gpu/drm/vmwgfx/svga_reg.h
index 11323dd5196f..e4259c2c1acc 100644
--- a/drivers/gpu/drm/vmwgfx/svga_reg.h
+++ b/drivers/gpu/drm/vmwgfx/svga_reg.h
@@ -35,7 +35,6 @@
 /*
  * PCI device IDs.
  */
-#define PCI_VENDOR_ID_VMWARE            0x15AD
 #define PCI_DEVICE_ID_VMWARE_SVGA2      0x0405
 
 /*
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
index 6327cfc36805..cff2bf9db9d2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
@@ -30,66 +30,101 @@
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_page_alloc.h>
 
-static uint32_t vram_placement_flags = TTM_PL_FLAG_VRAM |
-	TTM_PL_FLAG_CACHED;
-
-static uint32_t vram_ne_placement_flags = TTM_PL_FLAG_VRAM |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place vram_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t sys_placement_flags = TTM_PL_FLAG_SYSTEM |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place vram_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-static uint32_t sys_ne_placement_flags = TTM_PL_FLAG_SYSTEM |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place sys_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t gmr_placement_flags = VMW_PL_FLAG_GMR |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place sys_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-static uint32_t gmr_ne_placement_flags = VMW_PL_FLAG_GMR |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
+static struct ttm_place gmr_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+};
 
-static uint32_t mob_placement_flags = VMW_PL_FLAG_MOB |
-	TTM_PL_FLAG_CACHED;
+static struct ttm_place gmr_ne_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+};
 
-struct ttm_placement vmw_vram_placement = {
+static struct ttm_place mob_placement_flags = {
 	.fpfn = 0,
 	.lpfn = 0,
+	.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+};
+
+struct ttm_placement vmw_vram_placement = {
 	.num_placement = 1,
 	.placement = &vram_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &vram_placement_flags
 };
 
-static uint32_t vram_gmr_placement_flags[] = {
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+static struct ttm_place vram_gmr_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}
 };
 
-static uint32_t gmr_vram_placement_flags[] = {
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+static struct ttm_place gmr_vram_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}
 };
 
 struct ttm_placement vmw_vram_gmr_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 2,
 	.placement = vram_gmr_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &gmr_placement_flags
 };
 
-static uint32_t vram_gmr_ne_placement_flags[] = {
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
+static struct ttm_place vram_gmr_ne_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED |
+			 TTM_PL_FLAG_NO_EVICT
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED |
+			 TTM_PL_FLAG_NO_EVICT
+	}
 };
 
 struct ttm_placement vmw_vram_gmr_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 2,
 	.placement = vram_gmr_ne_placement_flags,
 	.num_busy_placement = 1,
@@ -97,8 +132,6 @@ struct ttm_placement vmw_vram_gmr_ne_placement = {
 };
 
 struct ttm_placement vmw_vram_sys_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &vram_placement_flags,
 	.num_busy_placement = 1,
@@ -106,8 +139,6 @@ struct ttm_placement vmw_vram_sys_placement = {
 };
 
 struct ttm_placement vmw_vram_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &vram_ne_placement_flags,
 	.num_busy_placement = 1,
@@ -115,8 +146,6 @@ struct ttm_placement vmw_vram_ne_placement = {
 };
 
 struct ttm_placement vmw_sys_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &sys_placement_flags,
 	.num_busy_placement = 1,
@@ -124,24 +153,33 @@ struct ttm_placement vmw_sys_placement = {
 };
 
 struct ttm_placement vmw_sys_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.placement = &sys_ne_placement_flags,
 	.num_busy_placement = 1,
 	.busy_placement = &sys_ne_placement_flags
 };
 
-static uint32_t evictable_placement_flags[] = {
-	TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+static struct ttm_place evictable_placement_flags[] = {
+	{
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+	}, {
+		.fpfn = 0,
+		.lpfn = 0,
+		.flags = VMW_PL_FLAG_MOB | TTM_PL_FLAG_CACHED
+	}
 };
 
 struct ttm_placement vmw_evictable_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 4,
 	.placement = evictable_placement_flags,
 	.num_busy_placement = 1,
@@ -149,8 +187,6 @@ struct ttm_placement vmw_evictable_placement = {
 };
 
 struct ttm_placement vmw_srf_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.num_busy_placement = 2,
 	.placement = &gmr_placement_flags,
@@ -158,8 +194,6 @@ struct ttm_placement vmw_srf_placement = {
 };
 
 struct ttm_placement vmw_mob_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
 	.num_placement = 1,
 	.num_busy_placement = 1,
 	.placement = &mob_placement_flags,
@@ -768,44 +802,6 @@ static int vmw_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 }
 
 /**
- * FIXME: We're using the old vmware polling method to sync.
- * Do this with fences instead.
- */
-
-static void *vmw_sync_obj_ref(void *sync_obj)
-{
-
-	return (void *)
-		vmw_fence_obj_reference((struct vmw_fence_obj *) sync_obj);
-}
-
-static void vmw_sync_obj_unref(void **sync_obj)
-{
-	vmw_fence_obj_unreference((struct vmw_fence_obj **) sync_obj);
-}
-
-static int vmw_sync_obj_flush(void *sync_obj)
-{
-	vmw_fence_obj_flush((struct vmw_fence_obj *) sync_obj);
-	return 0;
-}
-
-static bool vmw_sync_obj_signaled(void *sync_obj)
-{
-	return	vmw_fence_obj_signaled((struct vmw_fence_obj *) sync_obj,
-				       DRM_VMW_FENCE_FLAG_EXEC);
-
-}
-
-static int vmw_sync_obj_wait(void *sync_obj, bool lazy, bool interruptible)
-{
-	return vmw_fence_obj_wait((struct vmw_fence_obj *) sync_obj,
-				  DRM_VMW_FENCE_FLAG_EXEC,
-				  lazy, interruptible,
-				  VMW_FENCE_WAIT_TIMEOUT);
-}
-
-/**
  * vmw_move_notify - TTM move_notify_callback
  *
  * @bo:             The TTM buffer object about to move.
@@ -829,11 +825,7 @@ static void vmw_move_notify(struct ttm_buffer_object *bo,
  */
 static void vmw_swap_notify(struct ttm_buffer_object *bo)
 {
-	struct ttm_bo_device *bdev = bo->bdev;
-
-	spin_lock(&bdev->fence_lock);
 	ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bdev->fence_lock);
 }
 
 
@@ -846,11 +838,6 @@ struct ttm_bo_driver vmw_bo_driver = {
 	.evict_flags = vmw_evict_flags,
 	.move = NULL,
 	.verify_access = vmw_verify_access,
-	.sync_obj_signaled = vmw_sync_obj_signaled,
-	.sync_obj_wait = vmw_sync_obj_wait,
-	.sync_obj_flush = vmw_sync_obj_flush,
-	.sync_obj_unref = vmw_sync_obj_unref,
-	.sync_obj_ref = vmw_sync_obj_ref,
 	.move_notify = vmw_move_notify,
 	.swap_notify = vmw_swap_notify,
 	.fault_reserve_notify = &vmw_ttm_fault_reserve_notify,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c
new file mode 100644
index 000000000000..21e9b7f8dad0
--- /dev/null
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c
@@ -0,0 +1,342 @@
+/**************************************************************************
+ *
+ * Copyright © 2014 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#include "vmwgfx_drv.h"
+
+#define VMW_CMDBUF_RES_MAN_HT_ORDER 12
+
+enum vmw_cmdbuf_res_state {
+	VMW_CMDBUF_RES_COMMITED,
+	VMW_CMDBUF_RES_ADD,
+	VMW_CMDBUF_RES_DEL
+};
+
+/**
+ * struct vmw_cmdbuf_res - Command buffer managed resource entry.
+ *
+ * @res: Refcounted pointer to a struct vmw_resource.
+ * @hash: Hash entry for the manager hash table.
+ * @head: List head used either by the staging list or the manager list
+ * of commited resources.
+ * @state: Staging state of this resource entry.
+ * @man: Pointer to a resource manager for this entry.
+ */
+struct vmw_cmdbuf_res {
+	struct vmw_resource *res;
+	struct drm_hash_item hash;
+	struct list_head head;
+	enum vmw_cmdbuf_res_state state;
+	struct vmw_cmdbuf_res_manager *man;
+};
+
+/**
+ * struct vmw_cmdbuf_res_manager - Command buffer resource manager.
+ *
+ * @resources: Hash table containing staged and commited command buffer
+ * resources
+ * @list: List of commited command buffer resources.
+ * @dev_priv: Pointer to a device private structure.
+ *
+ * @resources and @list are protected by the cmdbuf mutex for now.
+ */
+struct vmw_cmdbuf_res_manager {
+	struct drm_open_hash resources;
+	struct list_head list;
+	struct vmw_private *dev_priv;
+};
+
+
+/**
+ * vmw_cmdbuf_res_lookup - Look up a command buffer resource
+ *
+ * @man: Pointer to the command buffer resource manager
+ * @resource_type: The resource type, that combined with the user key
+ * identifies the resource.
+ * @user_key: The user key.
+ *
+ * Returns a valid refcounted struct vmw_resource pointer on success,
+ * an error pointer on failure.
+ */
+struct vmw_resource *
+vmw_cmdbuf_res_lookup(struct vmw_cmdbuf_res_manager *man,
+		      enum vmw_cmdbuf_res_type res_type,
+		      u32 user_key)
+{
+	struct drm_hash_item *hash;
+	int ret;
+	unsigned long key = user_key | (res_type << 24);
+
+	ret = drm_ht_find_item(&man->resources, key, &hash);
+	if (unlikely(ret != 0))
+		return ERR_PTR(ret);
+
+	return vmw_resource_reference
+		(drm_hash_entry(hash, struct vmw_cmdbuf_res, hash)->res);
+}
+
+/**
+ * vmw_cmdbuf_res_free - Free a command buffer resource.
+ *
+ * @man: Pointer to the command buffer resource manager
+ * @entry: Pointer to a struct vmw_cmdbuf_res.
+ *
+ * Frees a struct vmw_cmdbuf_res entry and drops its reference to the
+ * struct vmw_resource.
+ */
+static void vmw_cmdbuf_res_free(struct vmw_cmdbuf_res_manager *man,
+				struct vmw_cmdbuf_res *entry)
+{
+	list_del(&entry->head);
+	WARN_ON(drm_ht_remove_item(&man->resources, &entry->hash));
+	vmw_resource_unreference(&entry->res);
+	kfree(entry);
+}
+
+/**
+ * vmw_cmdbuf_res_commit - Commit a list of command buffer resource actions
+ *
+ * @list: Caller's list of command buffer resource actions.
+ *
+ * This function commits a list of command buffer resource
+ * additions or removals.
+ * It is typically called when the execbuf ioctl call triggering these
+ * actions has commited the fifo contents to the device.
+ */
+void vmw_cmdbuf_res_commit(struct list_head *list)
+{
+	struct vmw_cmdbuf_res *entry, *next;
+
+	list_for_each_entry_safe(entry, next, list, head) {
+		list_del(&entry->head);
+		switch (entry->state) {
+		case VMW_CMDBUF_RES_ADD:
+			entry->state = VMW_CMDBUF_RES_COMMITED;
+			list_add_tail(&entry->head, &entry->man->list);
+			break;
+		case VMW_CMDBUF_RES_DEL:
+			vmw_resource_unreference(&entry->res);
+			kfree(entry);
+			break;
+		default:
+			BUG();
+			break;
+		}
+	}
+}
+
+/**
+ * vmw_cmdbuf_res_revert - Revert a list of command buffer resource actions
+ *
+ * @man: Pointer to the command buffer resource manager
+ * @list: Caller's list of command buffer resource action
+ *
+ * This function reverts a list of command buffer resource
+ * additions or removals.
+ * It is typically called when the execbuf ioctl call triggering these
+ * actions failed for some reason, and the command stream was never
+ * submitted.
+ */
+void vmw_cmdbuf_res_revert(struct list_head *list)
+{
+	struct vmw_cmdbuf_res *entry, *next;
+	int ret;
+
+	list_for_each_entry_safe(entry, next, list, head) {
+		switch (entry->state) {
+		case VMW_CMDBUF_RES_ADD:
+			vmw_cmdbuf_res_free(entry->man, entry);
+			break;
+		case VMW_CMDBUF_RES_DEL:
+			ret = drm_ht_insert_item(&entry->man->resources,
+						 &entry->hash);
+			list_del(&entry->head);
+			list_add_tail(&entry->head, &entry->man->list);
+			entry->state = VMW_CMDBUF_RES_COMMITED;
+			break;
+		default:
+			BUG();
+			break;
+		}
+	}
+}
+
+/**
+ * vmw_cmdbuf_res_add - Stage a command buffer managed resource for addition.
+ *
+ * @man: Pointer to the command buffer resource manager.
+ * @res_type: The resource type.
+ * @user_key: The user-space id of the resource.
+ * @res: Valid (refcount != 0) pointer to a struct vmw_resource.
+ * @list: The staging list.
+ *
+ * This function allocates a struct vmw_cmdbuf_res entry and adds the
+ * resource to the hash table of the manager identified by @man. The
+ * entry is then put on the staging list identified by @list.
+ */
+int vmw_cmdbuf_res_add(struct vmw_cmdbuf_res_manager *man,
+		       enum vmw_cmdbuf_res_type res_type,
+		       u32 user_key,
+		       struct vmw_resource *res,
+		       struct list_head *list)
+{
+	struct vmw_cmdbuf_res *cres;
+	int ret;
+
+	cres = kzalloc(sizeof(*cres), GFP_KERNEL);
+	if (unlikely(cres == NULL))
+		return -ENOMEM;
+
+	cres->hash.key = user_key | (res_type << 24);
+	ret = drm_ht_insert_item(&man->resources, &cres->hash);
+	if (unlikely(ret != 0))
+		goto out_invalid_key;
+
+	cres->state = VMW_CMDBUF_RES_ADD;
+	cres->res = vmw_resource_reference(res);
+	cres->man = man;
+	list_add_tail(&cres->head, list);
+
+out_invalid_key:
+	return ret;
+}
+
+/**
+ * vmw_cmdbuf_res_remove - Stage a command buffer managed resource for removal.
+ *
+ * @man: Pointer to the command buffer resource manager.
+ * @res_type: The resource type.
+ * @user_key: The user-space id of the resource.
+ * @list: The staging list.
+ *
+ * This function looks up the struct vmw_cmdbuf_res entry from the manager
+ * hash table and, if it exists, removes it. Depending on its current staging
+ * state it then either removes the entry from the staging list or adds it
+ * to it with a staging state of removal.
+ */
+int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
+			  enum vmw_cmdbuf_res_type res_type,
+			  u32 user_key,
+			  struct list_head *list)
+{
+	struct vmw_cmdbuf_res *entry;
+	struct drm_hash_item *hash;
+	int ret;
+
+	ret = drm_ht_find_item(&man->resources, user_key | (res_type << 24),
+			       &hash);
+	if (likely(ret != 0))
+		return -EINVAL;
+
+	entry = drm_hash_entry(hash, struct vmw_cmdbuf_res, hash);
+
+	switch (entry->state) {
+	case VMW_CMDBUF_RES_ADD:
+		vmw_cmdbuf_res_free(man, entry);
+		break;
+	case VMW_CMDBUF_RES_COMMITED:
+		(void) drm_ht_remove_item(&man->resources, &entry->hash);
+		list_del(&entry->head);
+		entry->state = VMW_CMDBUF_RES_DEL;
+		list_add_tail(&entry->head, list);
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	return 0;
+}
+
+/**
+ * vmw_cmdbuf_res_man_create - Allocate a command buffer managed resource
+ * manager.
+ *
+ * @dev_priv: Pointer to a struct vmw_private
+ *
+ * Allocates and initializes a command buffer managed resource manager. Returns
+ * an error pointer on failure.
+ */
+struct vmw_cmdbuf_res_manager *
+vmw_cmdbuf_res_man_create(struct vmw_private *dev_priv)
+{
+	struct vmw_cmdbuf_res_manager *man;
+	int ret;
+
+	man = kzalloc(sizeof(*man), GFP_KERNEL);
+	if (man == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	man->dev_priv = dev_priv;
+	INIT_LIST_HEAD(&man->list);
+	ret = drm_ht_create(&man->resources, VMW_CMDBUF_RES_MAN_HT_ORDER);
+	if (ret == 0)
+		return man;
+
+	kfree(man);
+	return ERR_PTR(ret);
+}
+
+/**
+ * vmw_cmdbuf_res_man_destroy - Destroy a command buffer managed resource
+ * manager.
+ *
+ * @man: Pointer to the  manager to destroy.
+ *
+ * This function destroys a command buffer managed resource manager and
+ * unreferences / frees all command buffer managed resources and -entries
+ * associated with it.
+ */
+void vmw_cmdbuf_res_man_destroy(struct vmw_cmdbuf_res_manager *man)
+{
+	struct vmw_cmdbuf_res *entry, *next;
+
+	list_for_each_entry_safe(entry, next, &man->list, head)
+		vmw_cmdbuf_res_free(man, entry);
+
+	kfree(man);
+}
+
+/**
+ *
+ * vmw_cmdbuf_res_man_size - Return the size of a command buffer managed
+ * resource manager
+ *
+ * Returns the approximate allocation size of a command buffer managed
+ * resource manager.
+ */
+size_t vmw_cmdbuf_res_man_size(void)
+{
+	static size_t res_man_size;
+
+	if (unlikely(res_man_size == 0))
+		res_man_size =
+			ttm_round_pot(sizeof(struct vmw_cmdbuf_res_manager)) +
+			ttm_round_pot(sizeof(struct hlist_head) <<
+				      VMW_CMDBUF_RES_MAN_HT_ORDER);
+
+	return res_man_size;
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_context.c b/drivers/gpu/drm/vmwgfx/vmwgfx_context.c
index 8bb26dcd9eae..5ac92874404d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_context.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_context.c
@@ -33,6 +33,7 @@ struct vmw_user_context {
 	struct ttm_base_object base;
 	struct vmw_resource res;
 	struct vmw_ctx_binding_state cbs;
+	struct vmw_cmdbuf_res_manager *man;
 };
 
 
@@ -103,7 +104,8 @@ static const vmw_scrub_func vmw_scrub_funcs[vmw_ctx_binding_max] = {
 
 static void vmw_hw_context_destroy(struct vmw_resource *res)
 {
-
+	struct vmw_user_context *uctx =
+		container_of(res, struct vmw_user_context, res);
 	struct vmw_private *dev_priv = res->dev_priv;
 	struct {
 		SVGA3dCmdHeader header;
@@ -113,9 +115,9 @@ static void vmw_hw_context_destroy(struct vmw_resource *res)
 
 	if (res->func->destroy == vmw_gb_context_destroy) {
 		mutex_lock(&dev_priv->cmdbuf_mutex);
+		vmw_cmdbuf_res_man_destroy(uctx->man);
 		mutex_lock(&dev_priv->binding_mutex);
-		(void) vmw_context_binding_state_kill
-			(&container_of(res, struct vmw_user_context, res)->cbs);
+		(void) vmw_context_binding_state_kill(&uctx->cbs);
 		(void) vmw_gb_context_destroy(res);
 		mutex_unlock(&dev_priv->binding_mutex);
 		if (dev_priv->pinned_bo != NULL &&
@@ -152,13 +154,16 @@ static int vmw_gb_context_init(struct vmw_private *dev_priv,
 	ret = vmw_resource_init(dev_priv, res, true,
 				res_free, &vmw_gb_context_func);
 	res->backup_size = SVGA3D_CONTEXT_DATA_SIZE;
+	if (unlikely(ret != 0))
+		goto out_err;
 
-	if (unlikely(ret != 0)) {
-		if (res_free)
-			res_free(res);
-		else
-			kfree(res);
-		return ret;
+	if (dev_priv->has_mob) {
+		uctx->man = vmw_cmdbuf_res_man_create(dev_priv);
+		if (unlikely(IS_ERR(uctx->man))) {
+			ret = PTR_ERR(uctx->man);
+			uctx->man = NULL;
+			goto out_err;
+		}
 	}
 
 	memset(&uctx->cbs, 0, sizeof(uctx->cbs));
@@ -166,6 +171,13 @@ static int vmw_gb_context_init(struct vmw_private *dev_priv,
 
 	vmw_resource_activate(res, vmw_hw_context_destroy);
 	return 0;
+
+out_err:
+	if (res_free)
+		res_free(res);
+	else
+		kfree(res);
+	return ret;
 }
 
 static int vmw_context_init(struct vmw_private *dev_priv,
@@ -471,7 +483,8 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 	 */
 
 	if (unlikely(vmw_user_context_size == 0))
-		vmw_user_context_size = ttm_round_pot(sizeof(*ctx)) + 128;
+		vmw_user_context_size = ttm_round_pot(sizeof(*ctx)) + 128 +
+		  ((dev_priv->has_mob) ? vmw_cmdbuf_res_man_size() : 0);
 
 	ret = ttm_read_lock(&dev_priv->reservation_sem, true);
 	if (unlikely(ret != 0))
@@ -901,3 +914,8 @@ struct list_head *vmw_context_binding_list(struct vmw_resource *ctx)
 {
 	return &(container_of(ctx, struct vmw_user_context, res)->cbs.list);
 }
+
+struct vmw_cmdbuf_res_manager *vmw_context_res_man(struct vmw_resource *ctx)
+{
+	return container_of(ctx, struct vmw_user_context, res)->man;
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
index 70ddce8358b0..914b375763dc 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
@@ -61,7 +61,7 @@ int vmw_dmabuf_to_placement(struct vmw_private *dev_priv,
 
 	vmw_execbuf_release_pinned_bo(dev_priv);
 
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
+	ret = ttm_bo_reserve(bo, interruptible, false, false, NULL);
 	if (unlikely(ret != 0))
 		goto err;
 
@@ -105,7 +105,7 @@ int vmw_dmabuf_to_vram_or_gmr(struct vmw_private *dev_priv,
 	if (pin)
 		vmw_execbuf_release_pinned_bo(dev_priv);
 
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
+	ret = ttm_bo_reserve(bo, interruptible, false, false, NULL);
 	if (unlikely(ret != 0))
 		goto err;
 
@@ -198,13 +198,19 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *dev_priv,
 {
 	struct ttm_buffer_object *bo = &buf->base;
 	struct ttm_placement placement;
+	struct ttm_place place;
 	int ret = 0;
 
 	if (pin)
-		placement = vmw_vram_ne_placement;
+		place = vmw_vram_ne_placement.placement[0];
 	else
-		placement = vmw_vram_placement;
-	placement.lpfn = bo->num_pages;
+		place = vmw_vram_placement.placement[0];
+	place.lpfn = bo->num_pages;
+
+	placement.num_placement = 1;
+	placement.placement = &place;
+	placement.num_busy_placement = 1;
+	placement.busy_placement = &place;
 
 	ret = ttm_write_lock(&dev_priv->reservation_sem, interruptible);
 	if (unlikely(ret != 0))
@@ -212,7 +218,7 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *dev_priv,
 
 	if (pin)
 		vmw_execbuf_release_pinned_bo(dev_priv);
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
+	ret = ttm_bo_reserve(bo, interruptible, false, false, NULL);
 	if (unlikely(ret != 0))
 		goto err_unlock;
 
@@ -293,21 +299,23 @@ void vmw_bo_get_guest_ptr(const struct ttm_buffer_object *bo,
  */
 void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin)
 {
-	uint32_t pl_flags;
+	struct ttm_place pl;
 	struct ttm_placement placement;
 	uint32_t old_mem_type = bo->mem.mem_type;
 	int ret;
 
 	lockdep_assert_held(&bo->resv->lock.base);
 
-	pl_flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | VMW_PL_FLAG_MOB
+	pl.fpfn = 0;
+	pl.lpfn = 0;
+	pl.flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | VMW_PL_FLAG_MOB
 		| TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
 	if (pin)
-		pl_flags |= TTM_PL_FLAG_NO_EVICT;
+		pl.flags |= TTM_PL_FLAG_NO_EVICT;
 
 	memset(&placement, 0, sizeof(placement));
 	placement.num_placement = 1;
-	placement.placement = &pl_flags;
+	placement.placement = &pl;
 
 	ret = ttm_bo_validate(bo, &placement, false, true);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 246a62bab378..25f3c250fd98 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -316,7 +316,7 @@ static int vmw_dummy_query_bo_create(struct vmw_private *dev_priv)
 	if (unlikely(ret != 0))
 		return ret;
 
-	ret = ttm_bo_reserve(bo, false, true, false, 0);
+	ret = ttm_bo_reserve(bo, false, true, false, NULL);
 	BUG_ON(ret != 0);
 
 	ret = ttm_bo_kmap(bo, 0, 1, &map);
@@ -688,7 +688,11 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		goto out_err0;
 	}
 
-	if (unlikely(dev_priv->prim_bb_mem < dev_priv->vram_size))
+	/*
+	 * Limit back buffer size to VRAM size.  Remove this once
+	 * screen targets are implemented.
+	 */
+	if (dev_priv->prim_bb_mem > dev_priv->vram_size)
 		dev_priv->prim_bb_mem = dev_priv->vram_size;
 
 	mutex_unlock(&dev_priv->hw_mutex);
@@ -946,7 +950,6 @@ static void vmw_postclose(struct drm_device *dev,
 		drm_master_put(&vmw_fp->locked_master);
 	}
 
-	vmw_compat_shader_man_destroy(vmw_fp->shman);
 	ttm_object_file_release(&vmw_fp->tfile);
 	kfree(vmw_fp);
 }
@@ -966,16 +969,10 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv)
 	if (unlikely(vmw_fp->tfile == NULL))
 		goto out_no_tfile;
 
-	vmw_fp->shman = vmw_compat_shader_man_create(dev_priv);
-	if (IS_ERR(vmw_fp->shman))
-		goto out_no_shman;
-
 	file_priv->driver_priv = vmw_fp;
 
 	return 0;
 
-out_no_shman:
-	ttm_object_file_release(&vmw_fp->tfile);
 out_no_tfile:
 	kfree(vmw_fp);
 	return ret;
@@ -1425,6 +1422,7 @@ static struct drm_driver driver = {
 	.open = vmw_driver_open,
 	.preclose = vmw_preclose,
 	.postclose = vmw_postclose,
+	.set_busid = drm_pci_set_busid,
 
 	.dumb_create = vmw_dumb_create,
 	.dumb_map_offset = vmw_dumb_map_offset,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index c886c024c637..4ee799b43d5d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -40,10 +40,10 @@
 #include <drm/ttm/ttm_module.h>
 #include "vmwgfx_fence.h"
 
-#define VMWGFX_DRIVER_DATE "20140325"
+#define VMWGFX_DRIVER_DATE "20140704"
 #define VMWGFX_DRIVER_MAJOR 2
 #define VMWGFX_DRIVER_MINOR 6
-#define VMWGFX_DRIVER_PATCHLEVEL 0
+#define VMWGFX_DRIVER_PATCHLEVEL 1
 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000
 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024)
 #define VMWGFX_MAX_RELOCATIONS 2048
@@ -75,14 +75,11 @@
 #define VMW_RES_FENCE ttm_driver_type3
 #define VMW_RES_SHADER ttm_driver_type4
 
-struct vmw_compat_shader_manager;
-
 struct vmw_fpriv {
 	struct drm_master *locked_master;
 	struct ttm_object_file *tfile;
 	struct list_head fence_events;
 	bool gb_aware;
-	struct vmw_compat_shader_manager *shman;
 };
 
 struct vmw_dma_buffer {
@@ -124,6 +121,10 @@ struct vmw_resource {
 	void (*hw_destroy) (struct vmw_resource *res);
 };
 
+
+/*
+ * Resources that are managed using ioctls.
+ */
 enum vmw_res_type {
 	vmw_res_context,
 	vmw_res_surface,
@@ -132,6 +133,15 @@ enum vmw_res_type {
 	vmw_res_max
 };
 
+/*
+ * Resources that are managed using command streams.
+ */
+enum vmw_cmdbuf_res_type {
+	vmw_cmdbuf_res_compat_shader
+};
+
+struct vmw_cmdbuf_res_manager;
+
 struct vmw_cursor_snooper {
 	struct drm_crtc *crtc;
 	size_t age;
@@ -332,7 +342,6 @@ struct vmw_sw_context{
 	uint32_t *cmd_bounce;
 	uint32_t cmd_bounce_size;
 	struct list_head resource_list;
-	uint32_t fence_flags;
 	struct ttm_buffer_object *cur_query_bo;
 	struct list_head res_relocations;
 	uint32_t *buf_start;
@@ -341,7 +350,7 @@ struct vmw_sw_context{
 	bool needs_post_query_barrier;
 	struct vmw_resource *error_resource;
 	struct vmw_ctx_binding_state staged_bindings;
-	struct list_head staged_shaders;
+	struct list_head staged_cmd_res;
 };
 
 struct vmw_legacy_display;
@@ -694,6 +703,7 @@ extern void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes);
 extern void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes);
 extern int vmw_fifo_send_fence(struct vmw_private *dev_priv,
 			       uint32_t *seqno);
+extern void vmw_fifo_ping_host_locked(struct vmw_private *, uint32_t reason);
 extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason);
 extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv);
 extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv);
@@ -974,7 +984,8 @@ extern void vmw_context_binding_res_list_kill(struct list_head *head);
 extern void vmw_context_binding_res_list_scrub(struct list_head *head);
 extern int vmw_context_rebind_all(struct vmw_resource *ctx);
 extern struct list_head *vmw_context_binding_list(struct vmw_resource *ctx);
-
+extern struct vmw_cmdbuf_res_manager *
+vmw_context_res_man(struct vmw_resource *ctx);
 /*
  * Surface management - vmwgfx_surface.c
  */
@@ -1008,27 +1019,42 @@ extern int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 				   struct drm_file *file_priv);
 extern int vmw_shader_destroy_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv);
-extern int vmw_compat_shader_lookup(struct vmw_compat_shader_manager *man,
-				    SVGA3dShaderType shader_type,
-				    u32 *user_key);
-extern void vmw_compat_shaders_commit(struct vmw_compat_shader_manager *man,
-				      struct list_head *list);
-extern void vmw_compat_shaders_revert(struct vmw_compat_shader_manager *man,
-				      struct list_head *list);
-extern int vmw_compat_shader_remove(struct vmw_compat_shader_manager *man,
-				    u32 user_key,
-				    SVGA3dShaderType shader_type,
-				    struct list_head *list);
-extern int vmw_compat_shader_add(struct vmw_compat_shader_manager *man,
+extern int vmw_compat_shader_add(struct vmw_private *dev_priv,
+				 struct vmw_cmdbuf_res_manager *man,
 				 u32 user_key, const void *bytecode,
 				 SVGA3dShaderType shader_type,
 				 size_t size,
-				 struct ttm_object_file *tfile,
 				 struct list_head *list);
-extern struct vmw_compat_shader_manager *
-vmw_compat_shader_man_create(struct vmw_private *dev_priv);
-extern void
-vmw_compat_shader_man_destroy(struct vmw_compat_shader_manager *man);
+extern int vmw_compat_shader_remove(struct vmw_cmdbuf_res_manager *man,
+				    u32 user_key, SVGA3dShaderType shader_type,
+				    struct list_head *list);
+extern struct vmw_resource *
+vmw_compat_shader_lookup(struct vmw_cmdbuf_res_manager *man,
+			 u32 user_key, SVGA3dShaderType shader_type);
+
+/*
+ * Command buffer managed resources - vmwgfx_cmdbuf_res.c
+ */
+
+extern struct vmw_cmdbuf_res_manager *
+vmw_cmdbuf_res_man_create(struct vmw_private *dev_priv);
+extern void vmw_cmdbuf_res_man_destroy(struct vmw_cmdbuf_res_manager *man);
+extern size_t vmw_cmdbuf_res_man_size(void);
+extern struct vmw_resource *
+vmw_cmdbuf_res_lookup(struct vmw_cmdbuf_res_manager *man,
+		      enum vmw_cmdbuf_res_type res_type,
+		      u32 user_key);
+extern void vmw_cmdbuf_res_revert(struct list_head *list);
+extern void vmw_cmdbuf_res_commit(struct list_head *list);
+extern int vmw_cmdbuf_res_add(struct vmw_cmdbuf_res_manager *man,
+			      enum vmw_cmdbuf_res_type res_type,
+			      u32 user_key,
+			      struct vmw_resource *res,
+			      struct list_head *list);
+extern int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
+				 enum vmw_cmdbuf_res_type res_type,
+				 u32 user_key,
+				 struct list_head *list);
 
 
 /**
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 87df0b3674fd..596cd6dafd33 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -346,13 +346,11 @@ static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context,
 		++sw_context->cur_val_buf;
 		val_buf = &vval_buf->base;
 		val_buf->bo = ttm_bo_reference(bo);
-		val_buf->reserved = false;
+		val_buf->shared = false;
 		list_add_tail(&val_buf->head, &sw_context->validate_nodes);
 		vval_buf->validate_as_mob = validate_as_mob;
 	}
 
-	sw_context->fence_flags |= DRM_VMW_FENCE_FLAG_EXEC;
-
 	if (p_val_node)
 		*p_val_node = val_node;
 
@@ -422,28 +420,90 @@ static int vmw_resources_validate(struct vmw_sw_context *sw_context)
 	return 0;
 }
 
+
+/**
+ * vmw_cmd_res_reloc_add - Add a resource to a software context's
+ * relocation- and validation lists.
+ *
+ * @dev_priv: Pointer to a struct vmw_private identifying the device.
+ * @sw_context: Pointer to the software context.
+ * @res_type: Resource type.
+ * @id_loc: Pointer to where the id that needs translation is located.
+ * @res: Valid pointer to a struct vmw_resource.
+ * @p_val: If non null, a pointer to the struct vmw_resource_validate_node
+ * used for this resource is returned here.
+ */
+static int vmw_cmd_res_reloc_add(struct vmw_private *dev_priv,
+				 struct vmw_sw_context *sw_context,
+				 enum vmw_res_type res_type,
+				 uint32_t *id_loc,
+				 struct vmw_resource *res,
+				 struct vmw_resource_val_node **p_val)
+{
+	int ret;
+	struct vmw_resource_val_node *node;
+
+	*p_val = NULL;
+	ret = vmw_resource_relocation_add(&sw_context->res_relocations,
+					  res,
+					  id_loc - sw_context->buf_start);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = vmw_resource_val_add(sw_context, res, &node);
+	if (unlikely(ret != 0))
+		return ret;
+
+	if (res_type == vmw_res_context && dev_priv->has_mob &&
+	    node->first_usage) {
+
+		/*
+		 * Put contexts first on the list to be able to exit
+		 * list traversal for contexts early.
+		 */
+		list_del(&node->head);
+		list_add(&node->head, &sw_context->resource_list);
+
+		ret = vmw_resource_context_res_add(dev_priv, sw_context, res);
+		if (unlikely(ret != 0))
+			return ret;
+		node->staged_bindings =
+			kzalloc(sizeof(*node->staged_bindings), GFP_KERNEL);
+		if (node->staged_bindings == NULL) {
+			DRM_ERROR("Failed to allocate context binding "
+				  "information.\n");
+			return -ENOMEM;
+		}
+		INIT_LIST_HEAD(&node->staged_bindings->list);
+	}
+
+	if (p_val)
+		*p_val = node;
+
+	return 0;
+}
+
+
 /**
- * vmw_cmd_compat_res_check - Check that a resource is present and if so, put it
+ * vmw_cmd_res_check - Check that a resource is present and if so, put it
  * on the resource validate list unless it's already there.
  *
  * @dev_priv: Pointer to a device private structure.
  * @sw_context: Pointer to the software context.
  * @res_type: Resource type.
  * @converter: User-space visisble type specific information.
- * @id: user-space resource id handle.
  * @id_loc: Pointer to the location in the command buffer currently being
  * parsed from where the user-space resource id handle is located.
  * @p_val: Pointer to pointer to resource validalidation node. Populated
  * on exit.
  */
 static int
-vmw_cmd_compat_res_check(struct vmw_private *dev_priv,
-			 struct vmw_sw_context *sw_context,
-			 enum vmw_res_type res_type,
-			 const struct vmw_user_resource_conv *converter,
-			 uint32_t id,
-			 uint32_t *id_loc,
-			 struct vmw_resource_val_node **p_val)
+vmw_cmd_res_check(struct vmw_private *dev_priv,
+		  struct vmw_sw_context *sw_context,
+		  enum vmw_res_type res_type,
+		  const struct vmw_user_resource_conv *converter,
+		  uint32_t *id_loc,
+		  struct vmw_resource_val_node **p_val)
 {
 	struct vmw_res_cache_entry *rcache =
 		&sw_context->res_cache[res_type];
@@ -451,7 +511,7 @@ vmw_cmd_compat_res_check(struct vmw_private *dev_priv,
 	struct vmw_resource_val_node *node;
 	int ret;
 
-	if (id == SVGA3D_INVALID_ID) {
+	if (*id_loc == SVGA3D_INVALID_ID) {
 		if (p_val)
 			*p_val = NULL;
 		if (res_type == vmw_res_context) {
@@ -466,7 +526,7 @@ vmw_cmd_compat_res_check(struct vmw_private *dev_priv,
 	 * resource
 	 */
 
-	if (likely(rcache->valid && id == rcache->handle)) {
+	if (likely(rcache->valid && *id_loc == rcache->handle)) {
 		const struct vmw_resource *res = rcache->res;
 
 		rcache->node->first_usage = false;
@@ -480,49 +540,28 @@ vmw_cmd_compat_res_check(struct vmw_private *dev_priv,
 
 	ret = vmw_user_resource_lookup_handle(dev_priv,
 					      sw_context->fp->tfile,
-					      id,
+					      *id_loc,
 					      converter,
 					      &res);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("Could not find or use resource 0x%08x.\n",
-			  (unsigned) id);
+			  (unsigned) *id_loc);
 		dump_stack();
 		return ret;
 	}
 
 	rcache->valid = true;
 	rcache->res = res;
-	rcache->handle = id;
-
-	ret = vmw_resource_relocation_add(&sw_context->res_relocations,
-					  res,
-					  id_loc - sw_context->buf_start);
-	if (unlikely(ret != 0))
-		goto out_no_reloc;
+	rcache->handle = *id_loc;
 
-	ret = vmw_resource_val_add(sw_context, res, &node);
+	ret = vmw_cmd_res_reloc_add(dev_priv, sw_context, res_type, id_loc,
+				    res, &node);
 	if (unlikely(ret != 0))
 		goto out_no_reloc;
 
 	rcache->node = node;
 	if (p_val)
 		*p_val = node;
-
-	if (dev_priv->has_mob && node->first_usage &&
-	    res_type == vmw_res_context) {
-		ret = vmw_resource_context_res_add(dev_priv, sw_context, res);
-		if (unlikely(ret != 0))
-			goto out_no_reloc;
-		node->staged_bindings =
-			kzalloc(sizeof(*node->staged_bindings), GFP_KERNEL);
-		if (node->staged_bindings == NULL) {
-			DRM_ERROR("Failed to allocate context binding "
-				  "information.\n");
-			goto out_no_reloc;
-		}
-		INIT_LIST_HEAD(&node->staged_bindings->list);
-	}
-
 	vmw_resource_unreference(&res);
 	return 0;
 
@@ -534,31 +573,6 @@ out_no_reloc:
 }
 
 /**
- * vmw_cmd_res_check - Check that a resource is present and if so, put it
- * on the resource validate list unless it's already there.
- *
- * @dev_priv: Pointer to a device private structure.
- * @sw_context: Pointer to the software context.
- * @res_type: Resource type.
- * @converter: User-space visisble type specific information.
- * @id_loc: Pointer to the location in the command buffer currently being
- * parsed from where the user-space resource id handle is located.
- * @p_val: Pointer to pointer to resource validalidation node. Populated
- * on exit.
- */
-static int
-vmw_cmd_res_check(struct vmw_private *dev_priv,
-		  struct vmw_sw_context *sw_context,
-		  enum vmw_res_type res_type,
-		  const struct vmw_user_resource_conv *converter,
-		  uint32_t *id_loc,
-		  struct vmw_resource_val_node **p_val)
-{
-	return vmw_cmd_compat_res_check(dev_priv, sw_context, res_type,
-					converter, *id_loc, id_loc, p_val);
-}
-
-/**
  * vmw_rebind_contexts - Rebind all resources previously bound to
  * referenced contexts.
  *
@@ -572,8 +586,8 @@ static int vmw_rebind_contexts(struct vmw_sw_context *sw_context)
 	int ret;
 
 	list_for_each_entry(val, &sw_context->resource_list, head) {
-		if (likely(!val->staged_bindings))
-			continue;
+		if (unlikely(!val->staged_bindings))
+			break;
 
 		ret = vmw_context_rebind_all(val->res);
 		if (unlikely(ret != 0)) {
@@ -1626,13 +1640,14 @@ static int vmw_cmd_shader_define(struct vmw_private *dev_priv,
 	} *cmd;
 	int ret;
 	size_t size;
+	struct vmw_resource_val_node *val;
 
 	cmd = container_of(header, struct vmw_shader_define_cmd,
 			   header);
 
 	ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context,
 				user_context_converter, &cmd->body.cid,
-				NULL);
+				&val);
 	if (unlikely(ret != 0))
 		return ret;
 
@@ -1640,11 +1655,11 @@ static int vmw_cmd_shader_define(struct vmw_private *dev_priv,
 		return 0;
 
 	size = cmd->header.size - sizeof(cmd->body);
-	ret = vmw_compat_shader_add(sw_context->fp->shman,
+	ret = vmw_compat_shader_add(dev_priv,
+				    vmw_context_res_man(val->res),
 				    cmd->body.shid, cmd + 1,
 				    cmd->body.type, size,
-				    sw_context->fp->tfile,
-				    &sw_context->staged_shaders);
+				    &sw_context->staged_cmd_res);
 	if (unlikely(ret != 0))
 		return ret;
 
@@ -1672,23 +1687,24 @@ static int vmw_cmd_shader_destroy(struct vmw_private *dev_priv,
 		SVGA3dCmdDestroyShader body;
 	} *cmd;
 	int ret;
+	struct vmw_resource_val_node *val;
 
 	cmd = container_of(header, struct vmw_shader_destroy_cmd,
 			   header);
 
 	ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_context,
 				user_context_converter, &cmd->body.cid,
-				NULL);
+				&val);
 	if (unlikely(ret != 0))
 		return ret;
 
 	if (unlikely(!dev_priv->has_mob))
 		return 0;
 
-	ret = vmw_compat_shader_remove(sw_context->fp->shman,
+	ret = vmw_compat_shader_remove(vmw_context_res_man(val->res),
 				       cmd->body.shid,
 				       cmd->body.type,
-				       &sw_context->staged_shaders);
+				       &sw_context->staged_cmd_res);
 	if (unlikely(ret != 0))
 		return ret;
 
@@ -1715,7 +1731,9 @@ static int vmw_cmd_set_shader(struct vmw_private *dev_priv,
 		SVGA3dCmdHeader header;
 		SVGA3dCmdSetShader body;
 	} *cmd;
-	struct vmw_resource_val_node *ctx_node;
+	struct vmw_resource_val_node *ctx_node, *res_node = NULL;
+	struct vmw_ctx_bindinfo bi;
+	struct vmw_resource *res = NULL;
 	int ret;
 
 	cmd = container_of(header, struct vmw_set_shader_cmd,
@@ -1727,32 +1745,40 @@ static int vmw_cmd_set_shader(struct vmw_private *dev_priv,
 	if (unlikely(ret != 0))
 		return ret;
 
-	if (dev_priv->has_mob) {
-		struct vmw_ctx_bindinfo bi;
-		struct vmw_resource_val_node *res_node;
-		u32 shid = cmd->body.shid;
-
-		if (shid != SVGA3D_INVALID_ID)
-			(void) vmw_compat_shader_lookup(sw_context->fp->shman,
-							cmd->body.type,
-							&shid);
-
-		ret = vmw_cmd_compat_res_check(dev_priv, sw_context,
-					       vmw_res_shader,
-					       user_shader_converter,
-					       shid,
-					       &cmd->body.shid, &res_node);
+	if (!dev_priv->has_mob)
+		return 0;
+
+	if (cmd->body.shid != SVGA3D_INVALID_ID) {
+		res = vmw_compat_shader_lookup
+			(vmw_context_res_man(ctx_node->res),
+			 cmd->body.shid,
+			 cmd->body.type);
+
+		if (!IS_ERR(res)) {
+			ret = vmw_cmd_res_reloc_add(dev_priv, sw_context,
+						    vmw_res_shader,
+						    &cmd->body.shid, res,
+						    &res_node);
+			vmw_resource_unreference(&res);
+			if (unlikely(ret != 0))
+				return ret;
+		}
+	}
+
+	if (!res_node) {
+		ret = vmw_cmd_res_check(dev_priv, sw_context,
+					vmw_res_shader,
+					user_shader_converter,
+					&cmd->body.shid, &res_node);
 		if (unlikely(ret != 0))
 			return ret;
-
-		bi.ctx = ctx_node->res;
-		bi.res = res_node ? res_node->res : NULL;
-		bi.bt = vmw_ctx_binding_shader;
-		bi.i1.shader_type = cmd->body.type;
-		return vmw_context_binding_add(ctx_node->staged_bindings, &bi);
 	}
 
-	return 0;
+	bi.ctx = ctx_node->res;
+	bi.res = res_node ? res_node->res : NULL;
+	bi.bt = vmw_ctx_binding_shader;
+	bi.i1.shader_type = cmd->body.type;
+	return vmw_context_binding_add(ctx_node->staged_bindings, &bi);
 }
 
 /**
@@ -2309,13 +2335,9 @@ int vmw_execbuf_fence_commands(struct drm_file *file_priv,
 
 	if (p_handle != NULL)
 		ret = vmw_user_fence_create(file_priv, dev_priv->fman,
-					    sequence,
-					    DRM_VMW_FENCE_FLAG_EXEC,
-					    p_fence, p_handle);
+					    sequence, p_fence, p_handle);
 	else
-		ret = vmw_fence_create(dev_priv->fman, sequence,
-				       DRM_VMW_FENCE_FLAG_EXEC,
-				       p_fence);
+		ret = vmw_fence_create(dev_priv->fman, sequence, p_fence);
 
 	if (unlikely(ret != 0 && !synced)) {
 		(void) vmw_fallback_wait(dev_priv, false, false,
@@ -2367,7 +2389,7 @@ vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
 		BUG_ON(fence == NULL);
 
 		fence_rep.handle = fence_handle;
-		fence_rep.seqno = fence->seqno;
+		fence_rep.seqno = fence->base.seqno;
 		vmw_update_seqno(dev_priv, &dev_priv->fifo);
 		fence_rep.passed_seqno = dev_priv->last_read_seqno;
 	}
@@ -2388,12 +2410,13 @@ vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
 		ttm_ref_object_base_unref(vmw_fp->tfile,
 					  fence_handle, TTM_REF_USAGE);
 		DRM_ERROR("Fence copy error. Syncing.\n");
-		(void) vmw_fence_obj_wait(fence, fence->signal_mask,
-					  false, false,
+		(void) vmw_fence_obj_wait(fence, false, false,
 					  VMW_FENCE_WAIT_TIMEOUT);
 	}
 }
 
+
+
 int vmw_execbuf_process(struct drm_file *file_priv,
 			struct vmw_private *dev_priv,
 			void __user *user_commands,
@@ -2439,7 +2462,6 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	sw_context->fp = vmw_fpriv(file_priv);
 	sw_context->cur_reloc = 0;
 	sw_context->cur_val_buf = 0;
-	sw_context->fence_flags = 0;
 	INIT_LIST_HEAD(&sw_context->resource_list);
 	sw_context->cur_query_bo = dev_priv->pinned_bo;
 	sw_context->last_query_ctx = NULL;
@@ -2453,7 +2475,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 			goto out_unlock;
 		sw_context->res_ht_initialized = true;
 	}
-	INIT_LIST_HEAD(&sw_context->staged_shaders);
+	INIT_LIST_HEAD(&sw_context->staged_cmd_res);
 
 	INIT_LIST_HEAD(&resource_list);
 	ret = vmw_cmd_check_all(dev_priv, sw_context, kernel_commands,
@@ -2465,7 +2487,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	if (unlikely(ret != 0))
 		goto out_err_nores;
 
-	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes);
+	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes, true);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -2548,8 +2570,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	}
 
 	list_splice_init(&sw_context->resource_list, &resource_list);
-	vmw_compat_shaders_commit(sw_context->fp->shman,
-				  &sw_context->staged_shaders);
+	vmw_cmdbuf_res_commit(&sw_context->staged_cmd_res);
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
 
 	/*
@@ -2576,8 +2597,7 @@ out_unlock:
 	list_splice_init(&sw_context->resource_list, &resource_list);
 	error_resource = sw_context->error_resource;
 	sw_context->error_resource = NULL;
-	vmw_compat_shaders_revert(sw_context->fp->shman,
-				  &sw_context->staged_shaders);
+	vmw_cmdbuf_res_revert(&sw_context->staged_cmd_res);
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
 
 	/*
@@ -2650,15 +2670,14 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	INIT_LIST_HEAD(&validate_list);
 
 	pinned_val.bo = ttm_bo_reference(dev_priv->pinned_bo);
+	pinned_val.shared = false;
 	list_add_tail(&pinned_val.head, &validate_list);
 
 	query_val.bo = ttm_bo_reference(dev_priv->dummy_query_bo);
+	query_val.shared = false;
 	list_add_tail(&query_val.head, &validate_list);
 
-	do {
-		ret = ttm_eu_reserve_buffers(&ticket, &validate_list);
-	} while (ret == -ERESTARTSYS);
-
+	ret = ttm_eu_reserve_buffers(&ticket, &validate_list, false);
 	if (unlikely(ret != 0)) {
 		vmw_execbuf_unpin_panic(dev_priv);
 		goto out_no_reserve;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index b031b48dbb3c..0a474f391fad 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -374,10 +374,16 @@ static int vmw_fb_create_bo(struct vmw_private *vmw_priv,
 			    size_t size, struct vmw_dma_buffer **out)
 {
 	struct vmw_dma_buffer *vmw_bo;
-	struct ttm_placement ne_placement = vmw_vram_ne_placement;
+	struct ttm_place ne_place = vmw_vram_ne_placement.placement[0];
+	struct ttm_placement ne_placement;
 	int ret;
 
-	ne_placement.lpfn = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	ne_placement.num_placement = 1;
+	ne_placement.placement = &ne_place;
+	ne_placement.num_busy_placement = 1;
+	ne_placement.busy_placement = &ne_place;
+
+	ne_place.lpfn = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
 	(void) ttm_write_lock(&vmw_priv->reservation_sem, false);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index 436b013b4231..197164fd7803 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -35,7 +35,7 @@ struct vmw_fence_manager {
 	struct vmw_private *dev_priv;
 	spinlock_t lock;
 	struct list_head fence_list;
-	struct work_struct work;
+	struct work_struct work, ping_work;
 	u32 user_fence_size;
 	u32 fence_size;
 	u32 event_fence_action_size;
@@ -46,6 +46,7 @@ struct vmw_fence_manager {
 	bool goal_irq_on; /* Protected by @goal_irq_mutex */
 	bool seqno_valid; /* Protected by @lock, and may not be set to true
 			     without the @goal_irq_mutex held. */
+	unsigned ctx;
 };
 
 struct vmw_user_fence {
@@ -80,6 +81,12 @@ struct vmw_event_fence_action {
 	uint32_t *tv_usec;
 };
 
+static struct vmw_fence_manager *
+fman_from_fence(struct vmw_fence_obj *fence)
+{
+	return container_of(fence->base.lock, struct vmw_fence_manager, lock);
+}
+
 /**
  * Note on fencing subsystem usage of irqs:
  * Typically the vmw_fences_update function is called
@@ -102,25 +109,143 @@ struct vmw_event_fence_action {
  * objects with actions attached to them.
  */
 
-static void vmw_fence_obj_destroy_locked(struct kref *kref)
+static void vmw_fence_obj_destroy(struct fence *f)
 {
 	struct vmw_fence_obj *fence =
-		container_of(kref, struct vmw_fence_obj, kref);
+		container_of(f, struct vmw_fence_obj, base);
 
-	struct vmw_fence_manager *fman = fence->fman;
-	unsigned int num_fences;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	unsigned long irq_flags;
 
+	spin_lock_irqsave(&fman->lock, irq_flags);
 	list_del_init(&fence->head);
-	num_fences = --fman->num_fence_objects;
-	spin_unlock_irq(&fman->lock);
-	if (fence->destroy)
-		fence->destroy(fence);
-	else
-		kfree(fence);
+	--fman->num_fence_objects;
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	fence->destroy(fence);
+}
 
-	spin_lock_irq(&fman->lock);
+static const char *vmw_fence_get_driver_name(struct fence *f)
+{
+	return "vmwgfx";
+}
+
+static const char *vmw_fence_get_timeline_name(struct fence *f)
+{
+	return "svga";
+}
+
+static void vmw_fence_ping_func(struct work_struct *work)
+{
+	struct vmw_fence_manager *fman =
+		container_of(work, struct vmw_fence_manager, ping_work);
+
+	vmw_fifo_ping_host(fman->dev_priv, SVGA_SYNC_GENERIC);
+}
+
+static bool vmw_fence_enable_signaling(struct fence *f)
+{
+	struct vmw_fence_obj *fence =
+		container_of(f, struct vmw_fence_obj, base);
+
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct vmw_private *dev_priv = fman->dev_priv;
+
+	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+	u32 seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+	if (seqno - fence->base.seqno < VMW_FENCE_WRAP)
+		return false;
+
+	if (mutex_trylock(&dev_priv->hw_mutex)) {
+		vmw_fifo_ping_host_locked(dev_priv, SVGA_SYNC_GENERIC);
+		mutex_unlock(&dev_priv->hw_mutex);
+	} else
+		schedule_work(&fman->ping_work);
+
+	return true;
+}
+
+struct vmwgfx_wait_cb {
+	struct fence_cb base;
+	struct task_struct *task;
+};
+
+static void
+vmwgfx_wait_cb(struct fence *fence, struct fence_cb *cb)
+{
+	struct vmwgfx_wait_cb *wait =
+		container_of(cb, struct vmwgfx_wait_cb, base);
+
+	wake_up_process(wait->task);
+}
+
+static void __vmw_fences_update(struct vmw_fence_manager *fman);
+
+static long vmw_fence_wait(struct fence *f, bool intr, signed long timeout)
+{
+	struct vmw_fence_obj *fence =
+		container_of(f, struct vmw_fence_obj, base);
+
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct vmw_private *dev_priv = fman->dev_priv;
+	struct vmwgfx_wait_cb cb;
+	long ret = timeout;
+	unsigned long irq_flags;
+
+	if (likely(vmw_fence_obj_signaled(fence)))
+		return timeout;
+
+	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
+	vmw_seqno_waiter_add(dev_priv);
+
+	spin_lock_irqsave(f->lock, irq_flags);
+
+	if (intr && signal_pending(current)) {
+		ret = -ERESTARTSYS;
+		goto out;
+	}
+
+	cb.base.func = vmwgfx_wait_cb;
+	cb.task = current;
+	list_add(&cb.base.node, &f->cb_list);
+
+	while (ret > 0) {
+		__vmw_fences_update(fman);
+		if (test_bit(FENCE_FLAG_SIGNALED_BIT, &f->flags))
+			break;
+
+		if (intr)
+			__set_current_state(TASK_INTERRUPTIBLE);
+		else
+			__set_current_state(TASK_UNINTERRUPTIBLE);
+		spin_unlock_irqrestore(f->lock, irq_flags);
+
+		ret = schedule_timeout(ret);
+
+		spin_lock_irqsave(f->lock, irq_flags);
+		if (ret > 0 && intr && signal_pending(current))
+			ret = -ERESTARTSYS;
+	}
+
+	if (!list_empty(&cb.base.node))
+		list_del(&cb.base.node);
+	__set_current_state(TASK_RUNNING);
+
+out:
+	spin_unlock_irqrestore(f->lock, irq_flags);
+
+	vmw_seqno_waiter_remove(dev_priv);
+
+	return ret;
 }
 
+static struct fence_ops vmw_fence_ops = {
+	.get_driver_name = vmw_fence_get_driver_name,
+	.get_timeline_name = vmw_fence_get_timeline_name,
+	.enable_signaling = vmw_fence_enable_signaling,
+	.wait = vmw_fence_wait,
+	.release = vmw_fence_obj_destroy,
+};
+
 
 /**
  * Execute signal actions on fences recently signaled.
@@ -180,12 +305,14 @@ struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv)
 	INIT_LIST_HEAD(&fman->fence_list);
 	INIT_LIST_HEAD(&fman->cleanup_list);
 	INIT_WORK(&fman->work, &vmw_fence_work_func);
+	INIT_WORK(&fman->ping_work, &vmw_fence_ping_func);
 	fman->fifo_down = true;
 	fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence));
 	fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj));
 	fman->event_fence_action_size =
 		ttm_round_pot(sizeof(struct vmw_event_fence_action));
 	mutex_init(&fman->goal_irq_mutex);
+	fman->ctx = fence_context_alloc(1);
 
 	return fman;
 }
@@ -196,6 +323,7 @@ void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
 	bool lists_empty;
 
 	(void) cancel_work_sync(&fman->work);
+	(void) cancel_work_sync(&fman->ping_work);
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
 	lists_empty = list_empty(&fman->fence_list) &&
@@ -207,23 +335,16 @@ void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
 }
 
 static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
-			      struct vmw_fence_obj *fence,
-			      u32 seqno,
-			      uint32_t mask,
+			      struct vmw_fence_obj *fence, u32 seqno,
 			      void (*destroy) (struct vmw_fence_obj *fence))
 {
 	unsigned long irq_flags;
-	unsigned int num_fences;
 	int ret = 0;
 
-	fence->seqno = seqno;
+	fence_init(&fence->base, &vmw_fence_ops, &fman->lock,
+		   fman->ctx, seqno);
 	INIT_LIST_HEAD(&fence->seq_passed_actions);
-	fence->fman = fman;
-	fence->signaled = 0;
-	fence->signal_mask = mask;
-	kref_init(&fence->kref);
 	fence->destroy = destroy;
-	init_waitqueue_head(&fence->queue);
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
 	if (unlikely(fman->fifo_down)) {
@@ -231,7 +352,7 @@ static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
 		goto out_unlock;
 	}
 	list_add_tail(&fence->head, &fman->fence_list);
-	num_fences = ++fman->num_fence_objects;
+	++fman->num_fence_objects;
 
 out_unlock:
 	spin_unlock_irqrestore(&fman->lock, irq_flags);
@@ -239,38 +360,6 @@ out_unlock:
 
 }
 
-struct vmw_fence_obj *vmw_fence_obj_reference(struct vmw_fence_obj *fence)
-{
-	if (unlikely(fence == NULL))
-		return NULL;
-
-	kref_get(&fence->kref);
-	return fence;
-}
-
-/**
- * vmw_fence_obj_unreference
- *
- * Note that this function may not be entered with disabled irqs since
- * it may re-enable them in the destroy function.
- *
- */
-void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
-{
-	struct vmw_fence_obj *fence = *fence_p;
-	struct vmw_fence_manager *fman;
-
-	if (unlikely(fence == NULL))
-		return;
-
-	fman = fence->fman;
-	*fence_p = NULL;
-	spin_lock_irq(&fman->lock);
-	BUG_ON(atomic_read(&fence->kref.refcount) == 0);
-	kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
-	spin_unlock_irq(&fman->lock);
-}
-
 static void vmw_fences_perform_actions(struct vmw_fence_manager *fman,
 				struct list_head *list)
 {
@@ -326,7 +415,7 @@ static bool vmw_fence_goal_new_locked(struct vmw_fence_manager *fman,
 	list_for_each_entry(fence, &fman->fence_list, head) {
 		if (!list_empty(&fence->seq_passed_actions)) {
 			fman->seqno_valid = true;
-			iowrite32(fence->seqno,
+			iowrite32(fence->base.seqno,
 				  fifo_mem + SVGA_FIFO_FENCE_GOAL);
 			break;
 		}
@@ -353,27 +442,27 @@ static bool vmw_fence_goal_new_locked(struct vmw_fence_manager *fman,
  */
 static bool vmw_fence_goal_check_locked(struct vmw_fence_obj *fence)
 {
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	u32 goal_seqno;
 	__le32 __iomem *fifo_mem;
 
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC)
+	if (fence_is_signaled_locked(&fence->base))
 		return false;
 
-	fifo_mem = fence->fman->dev_priv->mmio_virt;
+	fifo_mem = fman->dev_priv->mmio_virt;
 	goal_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	if (likely(fence->fman->seqno_valid &&
-		   goal_seqno - fence->seqno < VMW_FENCE_WRAP))
+	if (likely(fman->seqno_valid &&
+		   goal_seqno - fence->base.seqno < VMW_FENCE_WRAP))
 		return false;
 
-	iowrite32(fence->seqno, fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	fence->fman->seqno_valid = true;
+	iowrite32(fence->base.seqno, fifo_mem + SVGA_FIFO_FENCE_GOAL);
+	fman->seqno_valid = true;
 
 	return true;
 }
 
-void vmw_fences_update(struct vmw_fence_manager *fman)
+static void __vmw_fences_update(struct vmw_fence_manager *fman)
 {
-	unsigned long flags;
 	struct vmw_fence_obj *fence, *next_fence;
 	struct list_head action_list;
 	bool needs_rerun;
@@ -382,32 +471,25 @@ void vmw_fences_update(struct vmw_fence_manager *fman)
 
 	seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 rerun:
-	spin_lock_irqsave(&fman->lock, flags);
 	list_for_each_entry_safe(fence, next_fence, &fman->fence_list, head) {
-		if (seqno - fence->seqno < VMW_FENCE_WRAP) {
+		if (seqno - fence->base.seqno < VMW_FENCE_WRAP) {
 			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			fence_signal_locked(&fence->base);
 			INIT_LIST_HEAD(&action_list);
 			list_splice_init(&fence->seq_passed_actions,
 					 &action_list);
 			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
 		} else
 			break;
 	}
 
-	needs_rerun = vmw_fence_goal_new_locked(fman, seqno);
-
-	if (!list_empty(&fman->cleanup_list))
-		(void) schedule_work(&fman->work);
-	spin_unlock_irqrestore(&fman->lock, flags);
-
 	/*
 	 * Rerun if the fence goal seqno was updated, and the
 	 * hardware might have raced with that update, so that
 	 * we missed a fence_goal irq.
 	 */
 
+	needs_rerun = vmw_fence_goal_new_locked(fman, seqno);
 	if (unlikely(needs_rerun)) {
 		new_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 		if (new_seqno != seqno) {
@@ -415,79 +497,58 @@ rerun:
 			goto rerun;
 		}
 	}
+
+	if (!list_empty(&fman->cleanup_list))
+		(void) schedule_work(&fman->work);
 }
 
-bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-			    uint32_t flags)
+void vmw_fences_update(struct vmw_fence_manager *fman)
 {
-	struct vmw_fence_manager *fman = fence->fman;
 	unsigned long irq_flags;
-	uint32_t signaled;
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
+	__vmw_fences_update(fman);
 	spin_unlock_irqrestore(&fman->lock, irq_flags);
+}
 
-	flags &= fence->signal_mask;
-	if ((signaled & flags) == flags)
-		return 1;
+bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence)
+{
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 
-	if ((signaled & DRM_VMW_FENCE_FLAG_EXEC) == 0)
-		vmw_fences_update(fman);
+	if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+		return 1;
 
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	vmw_fences_update(fman);
 
-	return ((signaled & flags) == flags);
+	return fence_is_signaled(&fence->base);
 }
 
-int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
-		       uint32_t flags, bool lazy,
+int vmw_fence_obj_wait(struct vmw_fence_obj *fence, bool lazy,
 		       bool interruptible, unsigned long timeout)
 {
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
-	long ret;
+	long ret = fence_wait_timeout(&fence->base, interruptible, timeout);
 
-	if (likely(vmw_fence_obj_signaled(fence, flags)))
+	if (likely(ret > 0))
 		return 0;
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-	vmw_seqno_waiter_add(dev_priv);
-
-	if (interruptible)
-		ret = wait_event_interruptible_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
+	else if (ret == 0)
+		return -EBUSY;
 	else
-		ret = wait_event_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
-
-	vmw_seqno_waiter_remove(dev_priv);
-
-	if (unlikely(ret == 0))
-		ret = -EBUSY;
-	else if (likely(ret > 0))
-		ret = 0;
-
-	return ret;
+		return ret;
 }
 
 void vmw_fence_obj_flush(struct vmw_fence_obj *fence)
 {
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
+	struct vmw_private *dev_priv = fman_from_fence(fence)->dev_priv;
 
 	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
 }
 
 static void vmw_fence_destroy(struct vmw_fence_obj *fence)
 {
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+
+	fence_free(&fence->base);
 
-	kfree(fence);
 	/*
 	 * Free kernel space accounting.
 	 */
@@ -497,7 +558,6 @@ static void vmw_fence_destroy(struct vmw_fence_obj *fence)
 
 int vmw_fence_create(struct vmw_fence_manager *fman,
 		     uint32_t seqno,
-		     uint32_t mask,
 		     struct vmw_fence_obj **p_fence)
 {
 	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
@@ -515,7 +575,7 @@ int vmw_fence_create(struct vmw_fence_manager *fman,
 		goto out_no_object;
 	}
 
-	ret = vmw_fence_obj_init(fman, fence, seqno, mask,
+	ret = vmw_fence_obj_init(fman, fence, seqno,
 				 vmw_fence_destroy);
 	if (unlikely(ret != 0))
 		goto out_err_init;
@@ -535,7 +595,7 @@ static void vmw_user_fence_destroy(struct vmw_fence_obj *fence)
 {
 	struct vmw_user_fence *ufence =
 		container_of(fence, struct vmw_user_fence, fence);
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 
 	ttm_base_object_kfree(ufence, base);
 	/*
@@ -559,7 +619,6 @@ static void vmw_user_fence_base_release(struct ttm_base_object **p_base)
 int vmw_user_fence_create(struct drm_file *file_priv,
 			  struct vmw_fence_manager *fman,
 			  uint32_t seqno,
-			  uint32_t mask,
 			  struct vmw_fence_obj **p_fence,
 			  uint32_t *p_handle)
 {
@@ -586,7 +645,7 @@ int vmw_user_fence_create(struct drm_file *file_priv,
 	}
 
 	ret = vmw_fence_obj_init(fman, &ufence->fence, seqno,
-				 mask, vmw_user_fence_destroy);
+				 vmw_user_fence_destroy);
 	if (unlikely(ret != 0)) {
 		kfree(ufence);
 		goto out_no_object;
@@ -629,7 +688,6 @@ out_no_object:
 
 void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
 {
-	unsigned long irq_flags;
 	struct list_head action_list;
 	int ret;
 
@@ -638,35 +696,32 @@ void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
 	 * restart when we've released the fman->lock.
 	 */
 
-	spin_lock_irqsave(&fman->lock, irq_flags);
+	spin_lock_irq(&fman->lock);
 	fman->fifo_down = true;
 	while (!list_empty(&fman->fence_list)) {
 		struct vmw_fence_obj *fence =
 			list_entry(fman->fence_list.prev, struct vmw_fence_obj,
 				   head);
-		kref_get(&fence->kref);
+		fence_get(&fence->base);
 		spin_unlock_irq(&fman->lock);
 
-		ret = vmw_fence_obj_wait(fence, fence->signal_mask,
-					 false, false,
+		ret = vmw_fence_obj_wait(fence, false, false,
 					 VMW_FENCE_WAIT_TIMEOUT);
 
 		if (unlikely(ret != 0)) {
 			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			fence_signal(&fence->base);
 			INIT_LIST_HEAD(&action_list);
 			list_splice_init(&fence->seq_passed_actions,
 					 &action_list);
 			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
 		}
 
-		spin_lock_irq(&fman->lock);
-
 		BUG_ON(!list_empty(&fence->head));
-		kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
+		fence_put(&fence->base);
+		spin_lock_irq(&fman->lock);
 	}
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	spin_unlock_irq(&fman->lock);
 }
 
 void vmw_fence_fifo_up(struct vmw_fence_manager *fman)
@@ -716,14 +771,14 @@ int vmw_fence_obj_wait_ioctl(struct drm_device *dev, void *data,
 
 	timeout = jiffies;
 	if (time_after_eq(timeout, (unsigned long)arg->kernel_cookie)) {
-		ret = ((vmw_fence_obj_signaled(fence, arg->flags)) ?
+		ret = ((vmw_fence_obj_signaled(fence)) ?
 		       0 : -EBUSY);
 		goto out;
 	}
 
 	timeout = (unsigned long)arg->kernel_cookie - timeout;
 
-	ret = vmw_fence_obj_wait(fence, arg->flags, arg->lazy, true, timeout);
+	ret = vmw_fence_obj_wait(fence, arg->lazy, true, timeout);
 
 out:
 	ttm_base_object_unref(&base);
@@ -758,12 +813,12 @@ int vmw_fence_obj_signaled_ioctl(struct drm_device *dev, void *data,
 	}
 
 	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
-	fman = fence->fman;
+	fman = fman_from_fence(fence);
 
-	arg->signaled = vmw_fence_obj_signaled(fence, arg->flags);
-	spin_lock_irq(&fman->lock);
+	arg->signaled = vmw_fence_obj_signaled(fence);
 
-	arg->signaled_flags = fence->signaled;
+	arg->signaled_flags = arg->flags;
+	spin_lock_irq(&fman->lock);
 	arg->passed_seqno = dev_priv->last_read_seqno;
 	spin_unlock_irq(&fman->lock);
 
@@ -876,7 +931,7 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
 {
 	struct vmw_event_fence_action *eaction =
 		container_of(action, struct vmw_event_fence_action, action);
-	struct vmw_fence_manager *fman = eaction->fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(eaction->fence);
 	unsigned long irq_flags;
 
 	spin_lock_irqsave(&fman->lock, irq_flags);
@@ -900,7 +955,7 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
 static void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
 			      struct vmw_fence_action *action)
 {
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	unsigned long irq_flags;
 	bool run_update = false;
 
@@ -908,7 +963,7 @@ static void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
 	spin_lock_irqsave(&fman->lock, irq_flags);
 
 	fman->pending_actions[action->type]++;
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC) {
+	if (fence_is_signaled_locked(&fence->base)) {
 		struct list_head action_list;
 
 		INIT_LIST_HEAD(&action_list);
@@ -960,7 +1015,7 @@ int vmw_event_fence_action_queue(struct drm_file *file_priv,
 				 bool interruptible)
 {
 	struct vmw_event_fence_action *eaction;
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
 	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
 	unsigned long irq_flags;
 
@@ -1000,7 +1055,8 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv,
 				  bool interruptible)
 {
 	struct vmw_event_fence_pending *event;
-	struct drm_device *dev = fence->fman->dev_priv->dev;
+	struct vmw_fence_manager *fman = fman_from_fence(fence);
+	struct drm_device *dev = fman->dev_priv->dev;
 	unsigned long irq_flags;
 	int ret;
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
index faf2e7873860..26a4add39208 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
@@ -27,6 +27,8 @@
 
 #ifndef _VMWGFX_FENCE_H_
 
+#include <linux/fence.h>
+
 #define VMW_FENCE_WAIT_TIMEOUT (5*HZ)
 
 struct vmw_private;
@@ -50,16 +52,11 @@ struct vmw_fence_action {
 };
 
 struct vmw_fence_obj {
-	struct kref kref;
-	u32 seqno;
+	struct fence base;
 
-	struct vmw_fence_manager *fman;
 	struct list_head head;
-	uint32_t signaled;
-	uint32_t signal_mask;
 	struct list_head seq_passed_actions;
 	void (*destroy)(struct vmw_fence_obj *fence);
-	wait_queue_head_t queue;
 };
 
 extern struct vmw_fence_manager *
@@ -67,17 +64,29 @@ vmw_fence_manager_init(struct vmw_private *dev_priv);
 
 extern void vmw_fence_manager_takedown(struct vmw_fence_manager *fman);
 
-extern void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p);
+static inline void
+vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
+{
+	struct vmw_fence_obj *fence = *fence_p;
+
+	*fence_p = NULL;
+	if (fence)
+		fence_put(&fence->base);
+}
 
-extern struct vmw_fence_obj *
-vmw_fence_obj_reference(struct vmw_fence_obj *fence);
+static inline struct vmw_fence_obj *
+vmw_fence_obj_reference(struct vmw_fence_obj *fence)
+{
+	if (fence)
+		fence_get(&fence->base);
+	return fence;
+}
 
 extern void vmw_fences_update(struct vmw_fence_manager *fman);
 
-extern bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-				   uint32_t flags);
+extern bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence);
 
-extern int vmw_fence_obj_wait(struct vmw_fence_obj *fence, uint32_t flags,
+extern int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
 			      bool lazy,
 			      bool interruptible, unsigned long timeout);
 
@@ -85,13 +94,11 @@ extern void vmw_fence_obj_flush(struct vmw_fence_obj *fence);
 
 extern int vmw_fence_create(struct vmw_fence_manager *fman,
 			    uint32_t seqno,
-			    uint32_t mask,
 			    struct vmw_fence_obj **p_fence);
 
 extern int vmw_user_fence_create(struct drm_file *file_priv,
 				 struct vmw_fence_manager *fman,
 				 uint32_t sequence,
-				 uint32_t mask,
 				 struct vmw_fence_obj **p_fence,
 				 uint32_t *p_handle);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index 6ccd993e26bf..09e10aefcd8e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -160,16 +160,21 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	return vmw_fifo_send_fence(dev_priv, &dummy);
 }
 
-void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
+void vmw_fifo_ping_host_locked(struct vmw_private *dev_priv, uint32_t reason)
 {
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
 
-	mutex_lock(&dev_priv->hw_mutex);
-
 	if (unlikely(ioread32(fifo_mem + SVGA_FIFO_BUSY) == 0)) {
 		iowrite32(1, fifo_mem + SVGA_FIFO_BUSY);
 		vmw_write(dev_priv, SVGA_REG_SYNC, reason);
 	}
+}
+
+void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
+{
+	mutex_lock(&dev_priv->hw_mutex);
+
+	vmw_fifo_ping_host_locked(dev_priv, reason);
 
 	mutex_unlock(&dev_priv->hw_mutex);
 }
@@ -180,8 +185,9 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 
 	mutex_lock(&dev_priv->hw_mutex);
 
+	vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);
 	while (vmw_read(dev_priv, SVGA_REG_BUSY) != 0)
-		vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);
+		;
 
 	dev_priv->last_read_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
index b1273e8e9a69..170b61be1e4e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
@@ -46,7 +46,7 @@ struct vmwgfx_gmrid_man {
 
 static int vmw_gmrid_man_get_node(struct ttm_mem_type_manager *man,
 				  struct ttm_buffer_object *bo,
-				  struct ttm_placement *placement,
+				  const struct ttm_place *place,
 				  struct ttm_mem_reg *mem)
 {
 	struct vmwgfx_gmrid_man *gman =
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 8f3edc4710f2..941a7bc0b791 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -75,7 +75,7 @@ void vmw_display_unit_cleanup(struct vmw_display_unit *du)
 		vmw_surface_unreference(&du->cursor_surface);
 	if (du->cursor_dmabuf)
 		vmw_dmabuf_unreference(&du->cursor_dmabuf);
-	drm_sysfs_connector_remove(&du->connector);
+	drm_connector_unregister(&du->connector);
 	drm_crtc_cleanup(&du->crtc);
 	drm_encoder_cleanup(&du->encoder);
 	drm_connector_cleanup(&du->connector);
@@ -136,7 +136,7 @@ int vmw_cursor_update_dmabuf(struct vmw_private *dev_priv,
 	kmap_offset = 0;
 	kmap_num = (width*height*4 + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
-	ret = ttm_bo_reserve(&dmabuf->base, true, false, false, 0);
+	ret = ttm_bo_reserve(&dmabuf->base, true, false, false, NULL);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("reserve failed\n");
 		return -EINVAL;
@@ -187,7 +187,7 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	 * can do this since the caller in the drm core doesn't check anything
 	 * which is protected by any looks.
 	 */
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 	drm_modeset_lock_all(dev_priv->dev);
 
 	/* A lot of the code assumes this */
@@ -252,7 +252,7 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	ret = 0;
 out:
 	drm_modeset_unlock_all(dev_priv->dev);
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc);
 
 	return ret;
 }
@@ -273,7 +273,7 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	 * can do this since the caller in the drm core doesn't check anything
 	 * which is protected by any looks.
 	 */
-	drm_modeset_unlock(&crtc->mutex);
+	drm_modeset_unlock_crtc(crtc);
 	drm_modeset_lock_all(dev_priv->dev);
 
 	vmw_cursor_update_position(dev_priv, shown,
@@ -281,7 +281,7 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 				   du->cursor_y + du->hotspot_y);
 
 	drm_modeset_unlock_all(dev_priv->dev);
-	drm_modeset_lock(&crtc->mutex, NULL);
+	drm_modeset_lock_crtc(crtc);
 
 	return 0;
 }
@@ -343,7 +343,7 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 	kmap_offset = cmd->dma.guest.ptr.offset >> PAGE_SHIFT;
 	kmap_num = (64*64*4) >> PAGE_SHIFT;
 
-	ret = ttm_bo_reserve(bo, true, false, false, 0);
+	ret = ttm_bo_reserve(bo, true, false, false, NULL);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("reserve failed\n");
 		return;
@@ -1501,7 +1501,6 @@ int vmw_kms_cursor_bypass_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_vmw_cursor_bypass_arg *arg = data;
 	struct vmw_display_unit *du;
-	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	int ret = 0;
 
@@ -1519,13 +1518,12 @@ int vmw_kms_cursor_bypass_ioctl(struct drm_device *dev, void *data,
 		return 0;
 	}
 
-	obj = drm_mode_object_find(dev, arg->crtc_id, DRM_MODE_OBJECT_CRTC);
-	if (!obj) {
+	crtc = drm_crtc_find(dev, arg->crtc_id);
+	if (!crtc) {
 		ret = -ENOENT;
 		goto out;
 	}
 
-	crtc = obj_to_crtc(obj);
 	du = vmw_crtc_to_du(crtc);
 
 	du->hotspot_x = arg->xhot;
@@ -1952,6 +1950,14 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC)
 	};
 	int i;
+	u32 assumed_bpp = 2;
+
+	/*
+	 * If using screen objects, then assume 32-bpp because that's what the
+	 * SVGA device is assuming
+	 */
+	if (dev_priv->sou_priv)
+		assumed_bpp = 4;
 
 	/* Add preferred mode */
 	{
@@ -1962,8 +1968,9 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		mode->vdisplay = du->pref_height;
 		vmw_guess_mode_timing(mode);
 
-		if (vmw_kms_validate_mode_vram(dev_priv, mode->hdisplay * 2,
-					       mode->vdisplay)) {
+		if (vmw_kms_validate_mode_vram(dev_priv,
+						mode->hdisplay * assumed_bpp,
+						mode->vdisplay)) {
 			drm_mode_probed_add(connector, mode);
 		} else {
 			drm_mode_destroy(dev, mode);
@@ -1985,7 +1992,8 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector,
 		    bmode->vdisplay > max_height)
 			continue;
 
-		if (!vmw_kms_validate_mode_vram(dev_priv, bmode->hdisplay * 2,
+		if (!vmw_kms_validate_mode_vram(dev_priv,
+						bmode->hdisplay * assumed_bpp,
 						bmode->vdisplay))
 			continue;
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index b2b9bd23aeee..15e185ae4c99 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -371,7 +371,7 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 	encoder->possible_crtcs = (1 << unit);
 	encoder->possible_clones = 0;
 
-	(void) drm_sysfs_connector_add(connector);
+	(void) drm_connector_register(connector);
 
 	drm_crtc_init(dev, crtc, &vmw_legacy_crtc_funcs);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index 01d68f0a69dc..026de7cea0f6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -127,12 +127,13 @@ static void vmw_resource_release(struct kref *kref)
 	if (res->backup) {
 		struct ttm_buffer_object *bo = &res->backup->base;
 
-		ttm_bo_reserve(bo, false, false, false, 0);
+		ttm_bo_reserve(bo, false, false, false, NULL);
 		if (!list_empty(&res->mob_head) &&
 		    res->func->unbind != NULL) {
 			struct ttm_validate_buffer val_buf;
 
 			val_buf.bo = bo;
+			val_buf.shared = false;
 			res->func->unbind(res, false, &val_buf);
 		}
 		res->backup_dirty = false;
@@ -429,7 +430,7 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv,
 	ret = ttm_bo_init(bdev, &vmw_bo->base, size,
 			  ttm_bo_type_device, placement,
 			  0, interruptible,
-			  NULL, acc_size, NULL, bo_free);
+			  NULL, acc_size, NULL, NULL, bo_free);
 	return ret;
 }
 
@@ -567,13 +568,18 @@ static int vmw_user_dmabuf_synccpu_grab(struct vmw_user_dma_buffer *user_bo,
 	int ret;
 
 	if (flags & drm_vmw_synccpu_allow_cs) {
-		struct ttm_bo_device *bdev = bo->bdev;
+		bool nonblock = !!(flags & drm_vmw_synccpu_dontblock);
+		long lret;
 
-		spin_lock(&bdev->fence_lock);
-		ret = ttm_bo_wait(bo, false, true,
-				  !!(flags & drm_vmw_synccpu_dontblock));
-		spin_unlock(&bdev->fence_lock);
-		return ret;
+		if (nonblock)
+			return reservation_object_test_signaled_rcu(bo->resv, true) ? 0 : -EBUSY;
+
+		lret = reservation_object_wait_timeout_rcu(bo->resv, true, true, MAX_SCHEDULE_TIMEOUT);
+		if (!lret)
+			return -EBUSY;
+		else if (lret < 0)
+			return lret;
+		return 0;
 	}
 
 	ret = ttm_bo_synccpu_write_grab
@@ -1214,8 +1220,9 @@ vmw_resource_check_buffer(struct vmw_resource *res,
 
 	INIT_LIST_HEAD(&val_list);
 	val_buf->bo = ttm_bo_reference(&res->backup->base);
+	val_buf->shared = false;
 	list_add_tail(&val_buf->head, &val_list);
-	ret = ttm_eu_reserve_buffers(NULL, &val_list);
+	ret = ttm_eu_reserve_buffers(NULL, &val_list, interruptible);
 	if (unlikely(ret != 0))
 		goto out_no_reserve;
 
@@ -1307,6 +1314,7 @@ int vmw_resource_do_evict(struct vmw_resource *res, bool interruptible)
 	BUG_ON(!func->may_evict);
 
 	val_buf.bo = NULL;
+	val_buf.shared = false;
 	ret = vmw_resource_check_buffer(res, interruptible, &val_buf);
 	if (unlikely(ret != 0))
 		return ret;
@@ -1352,6 +1360,7 @@ int vmw_resource_validate(struct vmw_resource *res)
 		return 0;
 
 	val_buf.bo = NULL;
+	val_buf.shared = false;
 	if (res->backup)
 		val_buf.bo = &res->backup->base;
 	do {
@@ -1419,25 +1428,16 @@ void vmw_fence_single_bo(struct ttm_buffer_object *bo,
 			 struct vmw_fence_obj *fence)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_driver *driver = bdev->driver;
-	struct vmw_fence_obj *old_fence_obj;
+
 	struct vmw_private *dev_priv =
 		container_of(bdev, struct vmw_private, bdev);
 
-	if (fence == NULL)
+	if (fence == NULL) {
 		vmw_execbuf_fence_commands(NULL, dev_priv, &fence, NULL);
-	else
-		driver->sync_obj_ref(fence);
-
-	spin_lock(&bdev->fence_lock);
-
-	old_fence_obj = bo->sync_obj;
-	bo->sync_obj = fence;
-
-	spin_unlock(&bdev->fence_lock);
-
-	if (old_fence_obj)
-		vmw_fence_obj_unreference(&old_fence_obj);
+		reservation_object_add_excl_fence(bo->resv, &fence->base);
+		fence_put(&fence->base);
+	} else
+		reservation_object_add_excl_fence(bo->resv, &fence->base);
 }
 
 /**
@@ -1475,10 +1475,10 @@ void vmw_resource_move_notify(struct ttm_buffer_object *bo,
 
 	if (mem->mem_type != VMW_PL_MOB) {
 		struct vmw_resource *res, *n;
-		struct ttm_bo_device *bdev = bo->bdev;
 		struct ttm_validate_buffer val_buf;
 
 		val_buf.bo = bo;
+		val_buf.shared = false;
 
 		list_for_each_entry_safe(res, n, &dma_buf->res_list, mob_head) {
 
@@ -1491,9 +1491,7 @@ void vmw_resource_move_notify(struct ttm_buffer_object *bo,
 			list_del_init(&res->mob_head);
 		}
 
-		spin_lock(&bdev->fence_lock);
 		(void) ttm_bo_wait(bo, false, false, false);
-		spin_unlock(&bdev->fence_lock);
 	}
 }
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
index a95d3a0cabe4..b295463a60b3 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
@@ -467,7 +467,7 @@ static int vmw_sou_init(struct vmw_private *dev_priv, unsigned unit)
 	encoder->possible_crtcs = (1 << unit);
 	encoder->possible_clones = 0;
 
-	(void) drm_sysfs_connector_add(connector);
+	(void) drm_connector_register(connector);
 
 	drm_crtc_init(dev, crtc, &vmw_screen_object_crtc_funcs);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
index c1559eeaffe9..8719fb3cccc9 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c
@@ -29,8 +29,6 @@
 #include "vmwgfx_resource_priv.h"
 #include "ttm/ttm_placement.h"
 
-#define VMW_COMPAT_SHADER_HT_ORDER 12
-
 struct vmw_shader {
 	struct vmw_resource res;
 	SVGA3dShaderType type;
@@ -42,49 +40,8 @@ struct vmw_user_shader {
 	struct vmw_shader shader;
 };
 
-/**
- * enum vmw_compat_shader_state - Staging state for compat shaders
- */
-enum vmw_compat_shader_state {
-	VMW_COMPAT_COMMITED,
-	VMW_COMPAT_ADD,
-	VMW_COMPAT_DEL
-};
-
-/**
- * struct vmw_compat_shader - Metadata for compat shaders.
- *
- * @handle: The TTM handle of the guest backed shader.
- * @tfile: The struct ttm_object_file the guest backed shader is registered
- * with.
- * @hash: Hash item for lookup.
- * @head: List head for staging lists or the compat shader manager list.
- * @state: Staging state.
- *
- * The structure is protected by the cmdbuf lock.
- */
-struct vmw_compat_shader {
-	u32 handle;
-	struct ttm_object_file *tfile;
-	struct drm_hash_item hash;
-	struct list_head head;
-	enum vmw_compat_shader_state state;
-};
-
-/**
- * struct vmw_compat_shader_manager - Compat shader manager.
- *
- * @shaders: Hash table containing staged and commited compat shaders
- * @list: List of commited shaders.
- * @dev_priv: Pointer to a device private structure.
- *
- * @shaders and @list are protected by the cmdbuf mutex for now.
- */
-struct vmw_compat_shader_manager {
-	struct drm_open_hash shaders;
-	struct list_head list;
-	struct vmw_private *dev_priv;
-};
+static uint64_t vmw_user_shader_size;
+static uint64_t vmw_shader_size;
 
 static void vmw_user_shader_free(struct vmw_resource *res);
 static struct vmw_resource *
@@ -98,8 +55,6 @@ static int vmw_gb_shader_unbind(struct vmw_resource *res,
 				 struct ttm_validate_buffer *val_buf);
 static int vmw_gb_shader_destroy(struct vmw_resource *res);
 
-static uint64_t vmw_user_shader_size;
-
 static const struct vmw_user_resource_conv user_shader_conv = {
 	.object_type = VMW_RES_SHADER,
 	.base_obj_to_res = vmw_user_shader_base_to_res,
@@ -347,6 +302,16 @@ static void vmw_user_shader_free(struct vmw_resource *res)
 			    vmw_user_shader_size);
 }
 
+static void vmw_shader_free(struct vmw_resource *res)
+{
+	struct vmw_shader *shader = vmw_res_to_shader(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	kfree(shader);
+	ttm_mem_global_free(vmw_mem_glob(dev_priv),
+			    vmw_shader_size);
+}
+
 /**
  * This function is called when user space has no more references on the
  * base object. It releases the base-object's reference on the resource object.
@@ -371,13 +336,13 @@ int vmw_shader_destroy_ioctl(struct drm_device *dev, void *data,
 					 TTM_REF_USAGE);
 }
 
-static int vmw_shader_alloc(struct vmw_private *dev_priv,
-			    struct vmw_dma_buffer *buffer,
-			    size_t shader_size,
-			    size_t offset,
-			    SVGA3dShaderType shader_type,
-			    struct ttm_object_file *tfile,
-			    u32 *handle)
+static int vmw_user_shader_alloc(struct vmw_private *dev_priv,
+				 struct vmw_dma_buffer *buffer,
+				 size_t shader_size,
+				 size_t offset,
+				 SVGA3dShaderType shader_type,
+				 struct ttm_object_file *tfile,
+				 u32 *handle)
 {
 	struct vmw_user_shader *ushader;
 	struct vmw_resource *res, *tmp;
@@ -442,6 +407,56 @@ out:
 }
 
 
+struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv,
+				      struct vmw_dma_buffer *buffer,
+				      size_t shader_size,
+				      size_t offset,
+				      SVGA3dShaderType shader_type)
+{
+	struct vmw_shader *shader;
+	struct vmw_resource *res;
+	int ret;
+
+	/*
+	 * Approximate idr memory usage with 128 bytes. It will be limited
+	 * by maximum number_of shaders anyway.
+	 */
+	if (unlikely(vmw_shader_size == 0))
+		vmw_shader_size =
+			ttm_round_pot(sizeof(struct vmw_shader)) + 128;
+
+	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
+				   vmw_shader_size,
+				   false, true);
+	if (unlikely(ret != 0)) {
+		if (ret != -ERESTARTSYS)
+			DRM_ERROR("Out of graphics memory for shader "
+				  "creation.\n");
+		goto out_err;
+	}
+
+	shader = kzalloc(sizeof(*shader), GFP_KERNEL);
+	if (unlikely(shader == NULL)) {
+		ttm_mem_global_free(vmw_mem_glob(dev_priv),
+				    vmw_shader_size);
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
+	res = &shader->res;
+
+	/*
+	 * From here on, the destructor takes over resource freeing.
+	 */
+	ret = vmw_gb_shader_init(dev_priv, res, shader_size,
+				 offset, shader_type, buffer,
+				 vmw_shader_free);
+
+out_err:
+	return ret ? ERR_PTR(ret) : res;
+}
+
+
 int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv)
 {
@@ -490,8 +505,8 @@ int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 	if (unlikely(ret != 0))
 		goto out_bad_arg;
 
-	ret = vmw_shader_alloc(dev_priv, buffer, arg->size, arg->offset,
-			       shader_type, tfile, &arg->shader_handle);
+	ret = vmw_user_shader_alloc(dev_priv, buffer, arg->size, arg->offset,
+				    shader_type, tfile, &arg->shader_handle);
 
 	ttm_read_unlock(&dev_priv->reservation_sem);
 out_bad_arg:
@@ -500,202 +515,83 @@ out_bad_arg:
 }
 
 /**
- * vmw_compat_shader_lookup - Look up a compat shader
- *
- * @man: Pointer to the compat shader manager.
- * @shader_type: The shader type, that combined with the user_key identifies
- * the shader.
- * @user_key: On entry, this should be a pointer to the user_key.
- * On successful exit, it will contain the guest-backed shader's TTM handle.
+ * vmw_compat_shader_id_ok - Check whether a compat shader user key and
+ * shader type are within valid bounds.
  *
- * Returns 0 on success. Non-zero on failure, in which case the value pointed
- * to by @user_key is unmodified.
- */
-int vmw_compat_shader_lookup(struct vmw_compat_shader_manager *man,
-			     SVGA3dShaderType shader_type,
-			     u32 *user_key)
-{
-	struct drm_hash_item *hash;
-	int ret;
-	unsigned long key = *user_key | (shader_type << 24);
-
-	ret = drm_ht_find_item(&man->shaders, key, &hash);
-	if (unlikely(ret != 0))
-		return ret;
-
-	*user_key = drm_hash_entry(hash, struct vmw_compat_shader,
-				   hash)->handle;
-
-	return 0;
-}
-
-/**
- * vmw_compat_shader_free - Free a compat shader.
- *
- * @man: Pointer to the compat shader manager.
- * @entry: Pointer to a struct vmw_compat_shader.
- *
- * Frees a struct vmw_compat_shder entry and drops its reference to the
- * guest backed shader.
- */
-static void vmw_compat_shader_free(struct vmw_compat_shader_manager *man,
-				   struct vmw_compat_shader *entry)
-{
-	list_del(&entry->head);
-	WARN_ON(drm_ht_remove_item(&man->shaders, &entry->hash));
-	WARN_ON(ttm_ref_object_base_unref(entry->tfile, entry->handle,
-					  TTM_REF_USAGE));
-	kfree(entry);
-}
-
-/**
- * vmw_compat_shaders_commit - Commit a list of compat shader actions.
- *
- * @man: Pointer to the compat shader manager.
- * @list: Caller's list of compat shader actions.
+ * @user_key: User space id of the shader.
+ * @shader_type: Shader type.
  *
- * This function commits a list of compat shader additions or removals.
- * It is typically called when the execbuf ioctl call triggering these
- * actions has commited the fifo contents to the device.
+ * Returns true if valid false if not.
  */
-void vmw_compat_shaders_commit(struct vmw_compat_shader_manager *man,
-			       struct list_head *list)
+static bool vmw_compat_shader_id_ok(u32 user_key, SVGA3dShaderType shader_type)
 {
-	struct vmw_compat_shader *entry, *next;
-
-	list_for_each_entry_safe(entry, next, list, head) {
-		list_del(&entry->head);
-		switch (entry->state) {
-		case VMW_COMPAT_ADD:
-			entry->state = VMW_COMPAT_COMMITED;
-			list_add_tail(&entry->head, &man->list);
-			break;
-		case VMW_COMPAT_DEL:
-			ttm_ref_object_base_unref(entry->tfile, entry->handle,
-						  TTM_REF_USAGE);
-			kfree(entry);
-			break;
-		default:
-			BUG();
-			break;
-		}
-	}
+	return user_key <= ((1 << 20) - 1) && (unsigned) shader_type < 16;
 }
 
 /**
- * vmw_compat_shaders_revert - Revert a list of compat shader actions
+ * vmw_compat_shader_key - Compute a hash key suitable for a compat shader.
  *
- * @man: Pointer to the compat shader manager.
- * @list: Caller's list of compat shader actions.
+ * @user_key: User space id of the shader.
+ * @shader_type: Shader type.
  *
- * This function reverts a list of compat shader additions or removals.
- * It is typically called when the execbuf ioctl call triggering these
- * actions failed for some reason, and the command stream was never
- * submitted.
+ * Returns a hash key suitable for a command buffer managed resource
+ * manager hash table.
  */
-void vmw_compat_shaders_revert(struct vmw_compat_shader_manager *man,
-			       struct list_head *list)
+static u32 vmw_compat_shader_key(u32 user_key, SVGA3dShaderType shader_type)
 {
-	struct vmw_compat_shader *entry, *next;
-	int ret;
-
-	list_for_each_entry_safe(entry, next, list, head) {
-		switch (entry->state) {
-		case VMW_COMPAT_ADD:
-			vmw_compat_shader_free(man, entry);
-			break;
-		case VMW_COMPAT_DEL:
-			ret = drm_ht_insert_item(&man->shaders, &entry->hash);
-			list_del(&entry->head);
-			list_add_tail(&entry->head, &man->list);
-			entry->state = VMW_COMPAT_COMMITED;
-			break;
-		default:
-			BUG();
-			break;
-		}
-	}
+	return user_key | (shader_type << 20);
 }
 
 /**
  * vmw_compat_shader_remove - Stage a compat shader for removal.
  *
- * @man: Pointer to the compat shader manager
+ * @man: Pointer to the compat shader manager identifying the shader namespace.
  * @user_key: The key that is used to identify the shader. The key is
  * unique to the shader type.
  * @shader_type: Shader type.
- * @list: Caller's list of staged shader actions.
- *
- * This function stages a compat shader for removal and removes the key from
- * the shader manager's hash table. If the shader was previously only staged
- * for addition it is completely removed (But the execbuf code may keep a
- * reference if it was bound to a context between addition and removal). If
- * it was previously commited to the manager, it is staged for removal.
+ * @list: Caller's list of staged command buffer resource actions.
  */
-int vmw_compat_shader_remove(struct vmw_compat_shader_manager *man,
+int vmw_compat_shader_remove(struct vmw_cmdbuf_res_manager *man,
 			     u32 user_key, SVGA3dShaderType shader_type,
 			     struct list_head *list)
 {
-	struct vmw_compat_shader *entry;
-	struct drm_hash_item *hash;
-	int ret;
-
-	ret = drm_ht_find_item(&man->shaders, user_key | (shader_type << 24),
-			       &hash);
-	if (likely(ret != 0))
+	if (!vmw_compat_shader_id_ok(user_key, shader_type))
 		return -EINVAL;
 
-	entry = drm_hash_entry(hash, struct vmw_compat_shader, hash);
-
-	switch (entry->state) {
-	case VMW_COMPAT_ADD:
-		vmw_compat_shader_free(man, entry);
-		break;
-	case VMW_COMPAT_COMMITED:
-		(void) drm_ht_remove_item(&man->shaders, &entry->hash);
-		list_del(&entry->head);
-		entry->state = VMW_COMPAT_DEL;
-		list_add_tail(&entry->head, list);
-		break;
-	default:
-		BUG();
-		break;
-	}
-
-	return 0;
+	return vmw_cmdbuf_res_remove(man, vmw_cmdbuf_res_compat_shader,
+				     vmw_compat_shader_key(user_key,
+							   shader_type),
+				     list);
 }
 
 /**
- * vmw_compat_shader_add - Create a compat shader and add the
- * key to the manager
+ * vmw_compat_shader_add - Create a compat shader and stage it for addition
+ * as a command buffer managed resource.
  *
- * @man: Pointer to the compat shader manager
+ * @man: Pointer to the compat shader manager identifying the shader namespace.
  * @user_key: The key that is used to identify the shader. The key is
  * unique to the shader type.
  * @bytecode: Pointer to the bytecode of the shader.
  * @shader_type: Shader type.
  * @tfile: Pointer to a struct ttm_object_file that the guest-backed shader is
  * to be created with.
- * @list: Caller's list of staged shader actions.
+ * @list: Caller's list of staged command buffer resource actions.
  *
- * Note that only the key is added to the shader manager's hash table.
- * The shader is not yet added to the shader manager's list of shaders.
  */
-int vmw_compat_shader_add(struct vmw_compat_shader_manager *man,
+int vmw_compat_shader_add(struct vmw_private *dev_priv,
+			  struct vmw_cmdbuf_res_manager *man,
 			  u32 user_key, const void *bytecode,
 			  SVGA3dShaderType shader_type,
 			  size_t size,
-			  struct ttm_object_file *tfile,
 			  struct list_head *list)
 {
 	struct vmw_dma_buffer *buf;
 	struct ttm_bo_kmap_obj map;
 	bool is_iomem;
-	struct vmw_compat_shader *compat;
-	u32 handle;
 	int ret;
+	struct vmw_resource *res;
 
-	if (user_key > ((1 << 24) - 1) || (unsigned) shader_type > 16)
+	if (!vmw_compat_shader_id_ok(user_key, shader_type))
 		return -EINVAL;
 
 	/* Allocate and pin a DMA buffer */
@@ -703,7 +599,7 @@ int vmw_compat_shader_add(struct vmw_compat_shader_manager *man,
 	if (unlikely(buf == NULL))
 		return -ENOMEM;
 
-	ret = vmw_dmabuf_init(man->dev_priv, buf, size, &vmw_sys_ne_placement,
+	ret = vmw_dmabuf_init(dev_priv, buf, size, &vmw_sys_ne_placement,
 			      true, vmw_dmabuf_bo_free);
 	if (unlikely(ret != 0))
 		goto out;
@@ -728,84 +624,40 @@ int vmw_compat_shader_add(struct vmw_compat_shader_manager *man,
 	WARN_ON(ret != 0);
 	ttm_bo_unreserve(&buf->base);
 
-	/* Create a guest-backed shader container backed by the dma buffer */
-	ret = vmw_shader_alloc(man->dev_priv, buf, size, 0, shader_type,
-			       tfile, &handle);
-	vmw_dmabuf_unreference(&buf);
+	res = vmw_shader_alloc(dev_priv, buf, size, 0, shader_type);
 	if (unlikely(ret != 0))
 		goto no_reserve;
-	/*
-	 * Create a compat shader structure and stage it for insertion
-	 * in the manager
-	 */
-	compat = kzalloc(sizeof(*compat), GFP_KERNEL);
-	if (compat == NULL)
-		goto no_compat;
-
-	compat->hash.key = user_key |  (shader_type << 24);
-	ret = drm_ht_insert_item(&man->shaders, &compat->hash);
-	if (unlikely(ret != 0))
-		goto out_invalid_key;
-
-	compat->state = VMW_COMPAT_ADD;
-	compat->handle = handle;
-	compat->tfile = tfile;
-	list_add_tail(&compat->head, list);
-
-	return 0;
 
-out_invalid_key:
-	kfree(compat);
-no_compat:
-	ttm_ref_object_base_unref(tfile, handle, TTM_REF_USAGE);
+	ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_compat_shader,
+				 vmw_compat_shader_key(user_key, shader_type),
+				 res, list);
+	vmw_resource_unreference(&res);
 no_reserve:
+	vmw_dmabuf_unreference(&buf);
 out:
 	return ret;
 }
 
 /**
- * vmw_compat_shader_man_create - Create a compat shader manager
- *
- * @dev_priv: Pointer to a device private structure.
- *
- * Typically done at file open time. If successful returns a pointer to a
- * compat shader manager. Otherwise returns an error pointer.
- */
-struct vmw_compat_shader_manager *
-vmw_compat_shader_man_create(struct vmw_private *dev_priv)
-{
-	struct vmw_compat_shader_manager *man;
-	int ret;
-
-	man = kzalloc(sizeof(*man), GFP_KERNEL);
-	if (man == NULL)
-		return ERR_PTR(-ENOMEM);
-
-	man->dev_priv = dev_priv;
-	INIT_LIST_HEAD(&man->list);
-	ret = drm_ht_create(&man->shaders, VMW_COMPAT_SHADER_HT_ORDER);
-	if (ret == 0)
-		return man;
-
-	kfree(man);
-	return ERR_PTR(ret);
-}
-
-/**
- * vmw_compat_shader_man_destroy - Destroy a compat shader manager
+ * vmw_compat_shader_lookup - Look up a compat shader
  *
- * @man: Pointer to the shader manager to destroy.
+ * @man: Pointer to the command buffer managed resource manager identifying
+ * the shader namespace.
+ * @user_key: The user space id of the shader.
+ * @shader_type: The shader type.
  *
- * Typically done at file close time.
+ * Returns a refcounted pointer to a struct vmw_resource if the shader was
+ * found. An error pointer otherwise.
  */
-void vmw_compat_shader_man_destroy(struct vmw_compat_shader_manager *man)
+struct vmw_resource *
+vmw_compat_shader_lookup(struct vmw_cmdbuf_res_manager *man,
+			 u32 user_key,
+			 SVGA3dShaderType shader_type)
 {
-	struct vmw_compat_shader *entry, *next;
-
-	mutex_lock(&man->dev_priv->cmdbuf_mutex);
-	list_for_each_entry_safe(entry, next, &man->list, head)
-		vmw_compat_shader_free(man, entry);
+	if (!vmw_compat_shader_id_ok(user_key, shader_type))
+		return ERR_PTR(-EINVAL);
 
-	mutex_unlock(&man->dev_priv->cmdbuf_mutex);
-	kfree(man);
+	return vmw_cmdbuf_res_lookup(man, vmw_cmdbuf_res_compat_shader,
+				     vmw_compat_shader_key(user_key,
+							   shader_type));
 }
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index 112f27e51bc7..63bd63f3c7df 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -185,16 +185,16 @@ static unsigned int pin_job(struct host1x_job *job)
 		struct sg_table *sgt;
 		dma_addr_t phys_addr;
 
-		reloc->target = host1x_bo_get(reloc->target);
-		if (!reloc->target)
+		reloc->target.bo = host1x_bo_get(reloc->target.bo);
+		if (!reloc->target.bo)
 			goto unpin;
 
-		phys_addr = host1x_bo_pin(reloc->target, &sgt);
+		phys_addr = host1x_bo_pin(reloc->target.bo, &sgt);
 		if (!phys_addr)
 			goto unpin;
 
 		job->addr_phys[job->num_unpins] = phys_addr;
-		job->unpins[job->num_unpins].bo = reloc->target;
+		job->unpins[job->num_unpins].bo = reloc->target.bo;
 		job->unpins[job->num_unpins].sgt = sgt;
 		job->num_unpins++;
 	}
@@ -235,21 +235,21 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	for (i = 0; i < job->num_relocs; i++) {
 		struct host1x_reloc *reloc = &job->relocarray[i];
 		u32 reloc_addr = (job->reloc_addr_phys[i] +
-			reloc->target_offset) >> reloc->shift;
+				  reloc->target.offset) >> reloc->shift;
 		u32 *target;
 
 		/* skip all other gathers */
-		if (cmdbuf != reloc->cmdbuf)
+		if (cmdbuf != reloc->cmdbuf.bo)
 			continue;
 
-		if (last_page != reloc->cmdbuf_offset >> PAGE_SHIFT) {
+		if (last_page != reloc->cmdbuf.offset >> PAGE_SHIFT) {
 			if (cmdbuf_page_addr)
 				host1x_bo_kunmap(cmdbuf, last_page,
 						 cmdbuf_page_addr);
 
 			cmdbuf_page_addr = host1x_bo_kmap(cmdbuf,
-					reloc->cmdbuf_offset >> PAGE_SHIFT);
-			last_page = reloc->cmdbuf_offset >> PAGE_SHIFT;
+					reloc->cmdbuf.offset >> PAGE_SHIFT);
+			last_page = reloc->cmdbuf.offset >> PAGE_SHIFT;
 
 			if (unlikely(!cmdbuf_page_addr)) {
 				pr_err("Could not map cmdbuf for relocation\n");
@@ -257,7 +257,7 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 			}
 		}
 
-		target = cmdbuf_page_addr + (reloc->cmdbuf_offset & ~PAGE_MASK);
+		target = cmdbuf_page_addr + (reloc->cmdbuf.offset & ~PAGE_MASK);
 		*target = reloc_addr;
 	}
 
@@ -272,7 +272,7 @@ static bool check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
 {
 	offset *= sizeof(u32);
 
-	if (reloc->cmdbuf != cmdbuf || reloc->cmdbuf_offset != offset)
+	if (reloc->cmdbuf.bo != cmdbuf || reloc->cmdbuf.offset != offset)
 		return false;
 
 	return true;
diff --git a/drivers/gpu/ipu-v3/Kconfig b/drivers/gpu/ipu-v3/Kconfig
index 2f228a2f2a48..aefdff95356d 100644
--- a/drivers/gpu/ipu-v3/Kconfig
+++ b/drivers/gpu/ipu-v3/Kconfig
@@ -1,7 +1,8 @@
 config IMX_IPUV3_CORE
 	tristate "IPUv3 core support"
-	depends on SOC_IMX5 || SOC_IMX6Q || SOC_IMX6SL || ARCH_MULTIPLATFORM
+	depends on SOC_IMX5 || SOC_IMX6Q || ARCH_MULTIPLATFORM
 	depends on RESET_CONTROLLER
+	select GENERIC_IRQ_CHIP
 	help
 	  Choose this if you have a i.MX5/6 system and want to use the Image
 	  Processing Unit. This option only enables IPU base support.
diff --git a/drivers/gpu/ipu-v3/Makefile b/drivers/gpu/ipu-v3/Makefile
index 1887972b4ac2..107ec236a4a6 100644
--- a/drivers/gpu/ipu-v3/Makefile
+++ b/drivers/gpu/ipu-v3/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_IMX_IPUV3_CORE) += imx-ipu-v3.o
 
-imx-ipu-v3-objs := ipu-common.o ipu-dc.o ipu-di.o ipu-dp.o ipu-dmfc.o ipu-smfc.o
+imx-ipu-v3-objs := ipu-common.o ipu-cpmem.o ipu-csi.o ipu-dc.o ipu-di.o \
+		ipu-dp.o ipu-dmfc.o ipu-ic.o ipu-smfc.o
diff --git a/drivers/gpu/ipu-v3/ipu-common.c b/drivers/gpu/ipu-v3/ipu-common.c
index 04e7b2eafbdd..f707d25ae78f 100644
--- a/drivers/gpu/ipu-v3/ipu-common.c
+++ b/drivers/gpu/ipu-v3/ipu-common.c
@@ -44,17 +44,6 @@ static inline void ipu_cm_write(struct ipu_soc *ipu, u32 value, unsigned offset)
 	writel(value, ipu->cm_reg + offset);
 }
 
-static inline u32 ipu_idmac_read(struct ipu_soc *ipu, unsigned offset)
-{
-	return readl(ipu->idmac_reg + offset);
-}
-
-static inline void ipu_idmac_write(struct ipu_soc *ipu, u32 value,
-		unsigned offset)
-{
-	writel(value, ipu->idmac_reg + offset);
-}
-
 void ipu_srm_dp_sync_update(struct ipu_soc *ipu)
 {
 	u32 val;
@@ -65,457 +54,184 @@ void ipu_srm_dp_sync_update(struct ipu_soc *ipu)
 }
 EXPORT_SYMBOL_GPL(ipu_srm_dp_sync_update);
 
-struct ipu_ch_param __iomem *ipu_get_cpmem(struct ipuv3_channel *channel)
-{
-	struct ipu_soc *ipu = channel->ipu;
-
-	return ipu->cpmem_base + channel->num;
-}
-EXPORT_SYMBOL_GPL(ipu_get_cpmem);
-
-void ipu_cpmem_set_high_priority(struct ipuv3_channel *channel)
-{
-	struct ipu_soc *ipu = channel->ipu;
-	struct ipu_ch_param __iomem *p = ipu_get_cpmem(channel);
-	u32 val;
-
-	if (ipu->ipu_type == IPUV3EX)
-		ipu_ch_param_write_field(p, IPU_FIELD_ID, 1);
-
-	val = ipu_idmac_read(ipu, IDMAC_CHA_PRI(channel->num));
-	val |= 1 << (channel->num % 32);
-	ipu_idmac_write(ipu, val, IDMAC_CHA_PRI(channel->num));
-};
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_high_priority);
-
-void ipu_ch_param_write_field(struct ipu_ch_param __iomem *base, u32 wbs, u32 v)
-{
-	u32 bit = (wbs >> 8) % 160;
-	u32 size = wbs & 0xff;
-	u32 word = (wbs >> 8) / 160;
-	u32 i = bit / 32;
-	u32 ofs = bit % 32;
-	u32 mask = (1 << size) - 1;
-	u32 val;
-
-	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
-
-	val = readl(&base->word[word].data[i]);
-	val &= ~(mask << ofs);
-	val |= v << ofs;
-	writel(val, &base->word[word].data[i]);
-
-	if ((bit + size - 1) / 32 > i) {
-		val = readl(&base->word[word].data[i + 1]);
-		val &= ~(mask >> (ofs ? (32 - ofs) : 0));
-		val |= v >> (ofs ? (32 - ofs) : 0);
-		writel(val, &base->word[word].data[i + 1]);
-	}
-}
-EXPORT_SYMBOL_GPL(ipu_ch_param_write_field);
-
-u32 ipu_ch_param_read_field(struct ipu_ch_param __iomem *base, u32 wbs)
-{
-	u32 bit = (wbs >> 8) % 160;
-	u32 size = wbs & 0xff;
-	u32 word = (wbs >> 8) / 160;
-	u32 i = bit / 32;
-	u32 ofs = bit % 32;
-	u32 mask = (1 << size) - 1;
-	u32 val = 0;
-
-	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
-
-	val = (readl(&base->word[word].data[i]) >> ofs) & mask;
-
-	if ((bit + size - 1) / 32 > i) {
-		u32 tmp;
-		tmp = readl(&base->word[word].data[i + 1]);
-		tmp &= mask >> (ofs ? (32 - ofs) : 0);
-		val |= tmp << (ofs ? (32 - ofs) : 0);
-	}
-
-	return val;
-}
-EXPORT_SYMBOL_GPL(ipu_ch_param_read_field);
-
-int ipu_cpmem_set_format_rgb(struct ipu_ch_param __iomem *p,
-		const struct ipu_rgb *rgb)
-{
-	int bpp = 0, npb = 0, ro, go, bo, to;
-
-	ro = rgb->bits_per_pixel - rgb->red.length - rgb->red.offset;
-	go = rgb->bits_per_pixel - rgb->green.length - rgb->green.offset;
-	bo = rgb->bits_per_pixel - rgb->blue.length - rgb->blue.offset;
-	to = rgb->bits_per_pixel - rgb->transp.length - rgb->transp.offset;
-
-	ipu_ch_param_write_field(p, IPU_FIELD_WID0, rgb->red.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS0, ro);
-	ipu_ch_param_write_field(p, IPU_FIELD_WID1, rgb->green.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS1, go);
-	ipu_ch_param_write_field(p, IPU_FIELD_WID2, rgb->blue.length - 1);
-	ipu_ch_param_write_field(p, IPU_FIELD_OFS2, bo);
-
-	if (rgb->transp.length) {
-		ipu_ch_param_write_field(p, IPU_FIELD_WID3,
-				rgb->transp.length - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_OFS3, to);
-	} else {
-		ipu_ch_param_write_field(p, IPU_FIELD_WID3, 7);
-		ipu_ch_param_write_field(p, IPU_FIELD_OFS3,
-				rgb->bits_per_pixel);
-	}
-
-	switch (rgb->bits_per_pixel) {
-	case 32:
-		bpp = 0;
-		npb = 15;
-		break;
-	case 24:
-		bpp = 1;
-		npb = 19;
-		break;
-	case 16:
-		bpp = 3;
-		npb = 31;
-		break;
-	case 8:
-		bpp = 5;
-		npb = 63;
-		break;
-	default:
-		return -EINVAL;
-	}
-	ipu_ch_param_write_field(p, IPU_FIELD_BPP, bpp);
-	ipu_ch_param_write_field(p, IPU_FIELD_NPB, npb);
-	ipu_ch_param_write_field(p, IPU_FIELD_PFS, 7); /* rgb mode */
-
-	return 0;
-}
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_rgb);
-
-int ipu_cpmem_set_format_passthrough(struct ipu_ch_param __iomem *p,
-		int width)
+enum ipu_color_space ipu_drm_fourcc_to_colorspace(u32 drm_fourcc)
 {
-	int bpp = 0, npb = 0;
-
-	switch (width) {
-	case 32:
-		bpp = 0;
-		npb = 15;
-		break;
-	case 24:
-		bpp = 1;
-		npb = 19;
-		break;
-	case 16:
-		bpp = 3;
-		npb = 31;
-		break;
-	case 8:
-		bpp = 5;
-		npb = 63;
-		break;
+	switch (drm_fourcc) {
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_BGR565:
+	case DRM_FORMAT_RGB888:
+	case DRM_FORMAT_BGR888:
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_RGBX8888:
+	case DRM_FORMAT_BGRX8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_RGBA8888:
+	case DRM_FORMAT_BGRA8888:
+		return IPUV3_COLORSPACE_RGB;
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_YUV420:
+	case DRM_FORMAT_YVU420:
+	case DRM_FORMAT_YUV422:
+	case DRM_FORMAT_YVU422:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_NV21:
+	case DRM_FORMAT_NV16:
+	case DRM_FORMAT_NV61:
+		return IPUV3_COLORSPACE_YUV;
 	default:
-		return -EINVAL;
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
-
-	ipu_ch_param_write_field(p, IPU_FIELD_BPP, bpp);
-	ipu_ch_param_write_field(p, IPU_FIELD_NPB, npb);
-	ipu_ch_param_write_field(p, IPU_FIELD_PFS, 6); /* raw mode */
-
-	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_passthrough);
+EXPORT_SYMBOL_GPL(ipu_drm_fourcc_to_colorspace);
 
-void ipu_cpmem_set_yuv_interleaved(struct ipu_ch_param __iomem *p,
-				   u32 pixel_format)
+enum ipu_color_space ipu_pixelformat_to_colorspace(u32 pixelformat)
 {
-	switch (pixel_format) {
+	switch (pixelformat) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+	case V4L2_PIX_FMT_YUV422P:
 	case V4L2_PIX_FMT_UYVY:
-		ipu_ch_param_write_field(p, IPU_FIELD_BPP, 3);    /* bits/pixel */
-		ipu_ch_param_write_field(p, IPU_FIELD_PFS, 0xA);  /* pix format */
-		ipu_ch_param_write_field(p, IPU_FIELD_NPB, 31);   /* burst size */
-		break;
 	case V4L2_PIX_FMT_YUYV:
-		ipu_ch_param_write_field(p, IPU_FIELD_BPP, 3);    /* bits/pixel */
-		ipu_ch_param_write_field(p, IPU_FIELD_PFS, 0x8);  /* pix format */
-		ipu_ch_param_write_field(p, IPU_FIELD_NPB, 31);   /* burst size */
-		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
+		return IPUV3_COLORSPACE_YUV;
+	case V4L2_PIX_FMT_RGB32:
+	case V4L2_PIX_FMT_BGR32:
+	case V4L2_PIX_FMT_RGB24:
+	case V4L2_PIX_FMT_BGR24:
+	case V4L2_PIX_FMT_RGB565:
+		return IPUV3_COLORSPACE_RGB;
+	default:
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_interleaved);
+EXPORT_SYMBOL_GPL(ipu_pixelformat_to_colorspace);
 
-void ipu_cpmem_set_yuv_planar_full(struct ipu_ch_param __iomem *p,
-		u32 pixel_format, int stride, int u_offset, int v_offset)
+bool ipu_pixelformat_is_planar(u32 pixelformat)
 {
-	switch (pixel_format) {
+	switch (pixelformat) {
 	case V4L2_PIX_FMT_YUV420:
-		ipu_ch_param_write_field(p, IPU_FIELD_SLUV, (stride / 2) - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_UBO, u_offset / 8);
-		ipu_ch_param_write_field(p, IPU_FIELD_VBO, v_offset / 8);
-		break;
 	case V4L2_PIX_FMT_YVU420:
-		ipu_ch_param_write_field(p, IPU_FIELD_SLUV, (stride / 2) - 1);
-		ipu_ch_param_write_field(p, IPU_FIELD_UBO, v_offset / 8);
-		ipu_ch_param_write_field(p, IPU_FIELD_VBO, u_offset / 8);
-		break;
+	case V4L2_PIX_FMT_YUV422P:
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
+		return true;
 	}
-}
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar_full);
-
-void ipu_cpmem_set_yuv_planar(struct ipu_ch_param __iomem *p, u32 pixel_format,
-		int stride, int height)
-{
-	int u_offset, v_offset;
-	int uv_stride = 0;
 
-	switch (pixel_format) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-		uv_stride = stride / 2;
-		u_offset = stride * height;
-		v_offset = u_offset + (uv_stride * height / 2);
-		ipu_cpmem_set_yuv_planar_full(p, pixel_format, stride,
-				u_offset, v_offset);
-		break;
-	}
+	return false;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar);
-
-static const struct ipu_rgb def_rgb_32 = {
-	.red	= { .offset = 16, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset =  0, .length = 8, },
-	.transp = { .offset = 24, .length = 8, },
-	.bits_per_pixel = 32,
-};
-
-static const struct ipu_rgb def_bgr_32 = {
-	.red	= { .offset =  0, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset = 16, .length = 8, },
-	.transp = { .offset = 24, .length = 8, },
-	.bits_per_pixel = 32,
-};
-
-static const struct ipu_rgb def_rgb_24 = {
-	.red	= { .offset = 16, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset =  0, .length = 8, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 24,
-};
-
-static const struct ipu_rgb def_bgr_24 = {
-	.red	= { .offset =  0, .length = 8, },
-	.green	= { .offset =  8, .length = 8, },
-	.blue	= { .offset = 16, .length = 8, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 24,
-};
-
-static const struct ipu_rgb def_rgb_16 = {
-	.red	= { .offset = 11, .length = 5, },
-	.green	= { .offset =  5, .length = 6, },
-	.blue	= { .offset =  0, .length = 5, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 16,
-};
-
-static const struct ipu_rgb def_bgr_16 = {
-	.red	= { .offset =  0, .length = 5, },
-	.green	= { .offset =  5, .length = 6, },
-	.blue	= { .offset = 11, .length = 5, },
-	.transp = { .offset =  0, .length = 0, },
-	.bits_per_pixel = 16,
-};
-
-#define Y_OFFSET(pix, x, y)	((x) + pix->width * (y))
-#define U_OFFSET(pix, x, y)	((pix->width * pix->height) + \
-					(pix->width * (y) / 4) + (x) / 2)
-#define V_OFFSET(pix, x, y)	((pix->width * pix->height) + \
-					(pix->width * pix->height / 4) + \
-					(pix->width * (y) / 4) + (x) / 2)
+EXPORT_SYMBOL_GPL(ipu_pixelformat_is_planar);
 
-int ipu_cpmem_set_fmt(struct ipu_ch_param __iomem *cpmem, u32 drm_fourcc)
+enum ipu_color_space ipu_mbus_code_to_colorspace(u32 mbus_code)
 {
-	switch (drm_fourcc) {
-	case DRM_FORMAT_YUV420:
-	case DRM_FORMAT_YVU420:
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 2);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 63);
-		break;
-	case DRM_FORMAT_UYVY:
-		/* bits/pixel */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_BPP, 3);
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 0xA);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 31);
-		break;
-	case DRM_FORMAT_YUYV:
-		/* bits/pixel */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_BPP, 3);
-		/* pix format */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_PFS, 0x8);
-		/* burst size */
-		ipu_ch_param_write_field(cpmem, IPU_FIELD_NPB, 31);
-		break;
-	case DRM_FORMAT_ABGR8888:
-	case DRM_FORMAT_XBGR8888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_32);
-		break;
-	case DRM_FORMAT_ARGB8888:
-	case DRM_FORMAT_XRGB8888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_32);
-		break;
-	case DRM_FORMAT_BGR888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_24);
-		break;
-	case DRM_FORMAT_RGB888:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_24);
-		break;
-	case DRM_FORMAT_RGB565:
-		ipu_cpmem_set_format_rgb(cpmem, &def_rgb_16);
-		break;
-	case DRM_FORMAT_BGR565:
-		ipu_cpmem_set_format_rgb(cpmem, &def_bgr_16);
-		break;
+	switch (mbus_code & 0xf000) {
+	case 0x1000:
+		return IPUV3_COLORSPACE_RGB;
+	case 0x2000:
+		return IPUV3_COLORSPACE_YUV;
 	default:
-		return -EINVAL;
+		return IPUV3_COLORSPACE_UNKNOWN;
 	}
-
-	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_fmt);
+EXPORT_SYMBOL_GPL(ipu_mbus_code_to_colorspace);
 
-/*
- * The V4L2 spec defines packed RGB formats in memory byte order, which from
- * point of view of the IPU corresponds to little-endian words with the first
- * component in the least significant bits.
- * The DRM pixel formats and IPU internal representation are ordered the other
- * way around, with the first named component ordered at the most significant
- * bits. Further, V4L2 formats are not well defined:
- *     http://linuxtv.org/downloads/v4l-dvb-apis/packed-rgb.html
- * We choose the interpretation which matches GStreamer behavior.
- */
-static int v4l2_pix_fmt_to_drm_fourcc(u32 pixelformat)
+int ipu_stride_to_bytes(u32 pixel_stride, u32 pixelformat)
 {
 	switch (pixelformat) {
-	case V4L2_PIX_FMT_RGB565:
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+	case V4L2_PIX_FMT_YUV422P:
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV21:
+	case V4L2_PIX_FMT_NV16:
+	case V4L2_PIX_FMT_NV61:
 		/*
-		 * Here we choose the 'corrected' interpretation of RGBP, a
-		 * little-endian 16-bit word with the red component at the most
-		 * significant bits:
-		 * g[2:0]b[4:0] r[4:0]g[5:3] <=> [16:0] R:G:B
+		 * for the planar YUV formats, the stride passed to
+		 * cpmem must be the stride in bytes of the Y plane.
+		 * And all the planar YUV formats have an 8-bit
+		 * Y component.
 		 */
-		return DRM_FORMAT_RGB565;
+		return (8 * pixel_stride) >> 3;
+	case V4L2_PIX_FMT_RGB565:
+	case V4L2_PIX_FMT_YUYV:
+	case V4L2_PIX_FMT_UYVY:
+		return (16 * pixel_stride) >> 3;
 	case V4L2_PIX_FMT_BGR24:
-		/* B G R <=> [24:0] R:G:B */
-		return DRM_FORMAT_RGB888;
 	case V4L2_PIX_FMT_RGB24:
-		/* R G B <=> [24:0] B:G:R */
-		return DRM_FORMAT_BGR888;
+		return (24 * pixel_stride) >> 3;
 	case V4L2_PIX_FMT_BGR32:
-		/* B G R A <=> [32:0] A:B:G:R */
-		return DRM_FORMAT_XRGB8888;
 	case V4L2_PIX_FMT_RGB32:
-		/* R G B A <=> [32:0] A:B:G:R */
-		return DRM_FORMAT_XBGR8888;
-	case V4L2_PIX_FMT_UYVY:
-		return DRM_FORMAT_UYVY;
-	case V4L2_PIX_FMT_YUYV:
-		return DRM_FORMAT_YUYV;
-	case V4L2_PIX_FMT_YUV420:
-		return DRM_FORMAT_YUV420;
-	case V4L2_PIX_FMT_YVU420:
-		return DRM_FORMAT_YVU420;
+		return (32 * pixel_stride) >> 3;
+	default:
+		break;
 	}
 
 	return -EINVAL;
 }
+EXPORT_SYMBOL_GPL(ipu_stride_to_bytes);
 
-enum ipu_color_space ipu_drm_fourcc_to_colorspace(u32 drm_fourcc)
+int ipu_degrees_to_rot_mode(enum ipu_rotate_mode *mode, int degrees,
+			    bool hflip, bool vflip)
 {
-	switch (drm_fourcc) {
-	case DRM_FORMAT_RGB565:
-	case DRM_FORMAT_BGR565:
-	case DRM_FORMAT_RGB888:
-	case DRM_FORMAT_BGR888:
-	case DRM_FORMAT_XRGB8888:
-	case DRM_FORMAT_XBGR8888:
-	case DRM_FORMAT_RGBX8888:
-	case DRM_FORMAT_BGRX8888:
-	case DRM_FORMAT_ARGB8888:
-	case DRM_FORMAT_ABGR8888:
-	case DRM_FORMAT_RGBA8888:
-	case DRM_FORMAT_BGRA8888:
-		return IPUV3_COLORSPACE_RGB;
-	case DRM_FORMAT_YUYV:
-	case DRM_FORMAT_UYVY:
-	case DRM_FORMAT_YUV420:
-	case DRM_FORMAT_YVU420:
-		return IPUV3_COLORSPACE_YUV;
+	u32 r90, vf, hf;
+
+	switch (degrees) {
+	case 0:
+		vf = hf = r90 = 0;
+		break;
+	case 90:
+		vf = hf = 0;
+		r90 = 1;
+		break;
+	case 180:
+		vf = hf = 1;
+		r90 = 0;
+		break;
+	case 270:
+		vf = hf = r90 = 1;
+		break;
 	default:
-		return IPUV3_COLORSPACE_UNKNOWN;
+		return -EINVAL;
 	}
-}
-EXPORT_SYMBOL_GPL(ipu_drm_fourcc_to_colorspace);
 
-int ipu_cpmem_set_image(struct ipu_ch_param __iomem *cpmem,
-		struct ipu_image *image)
-{
-	struct v4l2_pix_format *pix = &image->pix;
-	int y_offset, u_offset, v_offset;
+	hf ^= (u32)hflip;
+	vf ^= (u32)vflip;
 
-	pr_debug("%s: resolution: %dx%d stride: %d\n",
-			__func__, pix->width, pix->height,
-			pix->bytesperline);
+	*mode = (enum ipu_rotate_mode)((r90 << 2) | (hf << 1) | vf);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_degrees_to_rot_mode);
 
-	ipu_cpmem_set_resolution(cpmem, image->rect.width,
-			image->rect.height);
-	ipu_cpmem_set_stride(cpmem, pix->bytesperline);
+int ipu_rot_mode_to_degrees(int *degrees, enum ipu_rotate_mode mode,
+			    bool hflip, bool vflip)
+{
+	u32 r90, vf, hf;
 
-	ipu_cpmem_set_fmt(cpmem, v4l2_pix_fmt_to_drm_fourcc(pix->pixelformat));
+	r90 = ((u32)mode >> 2) & 0x1;
+	hf = ((u32)mode >> 1) & 0x1;
+	vf = ((u32)mode >> 0) & 0x1;
+	hf ^= (u32)hflip;
+	vf ^= (u32)vflip;
 
-	switch (pix->pixelformat) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-		y_offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
-		u_offset = U_OFFSET(pix, image->rect.left,
-				image->rect.top) - y_offset;
-		v_offset = V_OFFSET(pix, image->rect.left,
-				image->rect.top) - y_offset;
-
-		ipu_cpmem_set_yuv_planar_full(cpmem, pix->pixelformat,
-				pix->bytesperline, u_offset, v_offset);
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys + y_offset);
+	switch ((enum ipu_rotate_mode)((r90 << 2) | (hf << 1) | vf)) {
+	case IPU_ROTATE_NONE:
+		*degrees = 0;
 		break;
-	case V4L2_PIX_FMT_UYVY:
-	case V4L2_PIX_FMT_YUYV:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 2 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_90_RIGHT:
+		*degrees = 90;
 		break;
-	case V4L2_PIX_FMT_RGB32:
-	case V4L2_PIX_FMT_BGR32:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 4 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_180:
+		*degrees = 180;
 		break;
-	case V4L2_PIX_FMT_RGB565:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 2 +
-				image->rect.top * image->pix.bytesperline);
-		break;
-	case V4L2_PIX_FMT_RGB24:
-	case V4L2_PIX_FMT_BGR24:
-		ipu_cpmem_set_buffer(cpmem, 0, image->phys +
-				image->rect.left * 3 +
-				image->rect.top * image->pix.bytesperline);
+	case IPU_ROTATE_90_LEFT:
+		*degrees = 270;
 		break;
 	default:
 		return -EINVAL;
@@ -523,27 +239,7 @@ int ipu_cpmem_set_image(struct ipu_ch_param __iomem *cpmem,
 
 	return 0;
 }
-EXPORT_SYMBOL_GPL(ipu_cpmem_set_image);
-
-enum ipu_color_space ipu_pixelformat_to_colorspace(u32 pixelformat)
-{
-	switch (pixelformat) {
-	case V4L2_PIX_FMT_YUV420:
-	case V4L2_PIX_FMT_YVU420:
-	case V4L2_PIX_FMT_UYVY:
-	case V4L2_PIX_FMT_YUYV:
-		return IPUV3_COLORSPACE_YUV;
-	case V4L2_PIX_FMT_RGB32:
-	case V4L2_PIX_FMT_BGR32:
-	case V4L2_PIX_FMT_RGB24:
-	case V4L2_PIX_FMT_BGR24:
-	case V4L2_PIX_FMT_RGB565:
-		return IPUV3_COLORSPACE_RGB;
-	default:
-		return IPUV3_COLORSPACE_UNKNOWN;
-	}
-}
-EXPORT_SYMBOL_GPL(ipu_pixelformat_to_colorspace);
+EXPORT_SYMBOL_GPL(ipu_rot_mode_to_degrees);
 
 struct ipuv3_channel *ipu_idmac_get(struct ipu_soc *ipu, unsigned num)
 {
@@ -587,7 +283,26 @@ void ipu_idmac_put(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_put);
 
-#define idma_mask(ch)			(1 << (ch & 0x1f))
+#define idma_mask(ch)			(1 << ((ch) & 0x1f))
+
+/*
+ * This is an undocumented feature, a write one to a channel bit in
+ * IPU_CHA_CUR_BUF and IPU_CHA_TRIPLE_CUR_BUF will reset the channel's
+ * internal current buffer pointer so that transfers start from buffer
+ * 0 on the next channel enable (that's the theory anyway, the imx6 TRM
+ * only says these are read-only registers). This operation is required
+ * for channel linking to work correctly, for instance video capture
+ * pipelines that carry out image rotations will fail after the first
+ * streaming unless this function is called for each channel before
+ * re-enabling the channels.
+ */
+static void __ipu_idmac_reset_current_buffer(struct ipuv3_channel *channel)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned int chno = channel->num;
+
+	ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_CUR_BUF(chno));
+}
 
 void ipu_idmac_set_double_buffer(struct ipuv3_channel *channel,
 		bool doublebuffer)
@@ -605,10 +320,81 @@ void ipu_idmac_set_double_buffer(struct ipuv3_channel *channel,
 		reg &= ~idma_mask(channel->num);
 	ipu_cm_write(ipu, reg, IPU_CHA_DB_MODE_SEL(channel->num));
 
+	__ipu_idmac_reset_current_buffer(channel);
+
 	spin_unlock_irqrestore(&ipu->lock, flags);
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_set_double_buffer);
 
+static const struct {
+	int chnum;
+	u32 reg;
+	int shift;
+} idmac_lock_en_info[] = {
+	{ .chnum =  5, .reg = IDMAC_CH_LOCK_EN_1, .shift =  0, },
+	{ .chnum = 11, .reg = IDMAC_CH_LOCK_EN_1, .shift =  2, },
+	{ .chnum = 12, .reg = IDMAC_CH_LOCK_EN_1, .shift =  4, },
+	{ .chnum = 14, .reg = IDMAC_CH_LOCK_EN_1, .shift =  6, },
+	{ .chnum = 15, .reg = IDMAC_CH_LOCK_EN_1, .shift =  8, },
+	{ .chnum = 20, .reg = IDMAC_CH_LOCK_EN_1, .shift = 10, },
+	{ .chnum = 21, .reg = IDMAC_CH_LOCK_EN_1, .shift = 12, },
+	{ .chnum = 22, .reg = IDMAC_CH_LOCK_EN_1, .shift = 14, },
+	{ .chnum = 23, .reg = IDMAC_CH_LOCK_EN_1, .shift = 16, },
+	{ .chnum = 27, .reg = IDMAC_CH_LOCK_EN_1, .shift = 18, },
+	{ .chnum = 28, .reg = IDMAC_CH_LOCK_EN_1, .shift = 20, },
+	{ .chnum = 45, .reg = IDMAC_CH_LOCK_EN_2, .shift =  0, },
+	{ .chnum = 46, .reg = IDMAC_CH_LOCK_EN_2, .shift =  2, },
+	{ .chnum = 47, .reg = IDMAC_CH_LOCK_EN_2, .shift =  4, },
+	{ .chnum = 48, .reg = IDMAC_CH_LOCK_EN_2, .shift =  6, },
+	{ .chnum = 49, .reg = IDMAC_CH_LOCK_EN_2, .shift =  8, },
+	{ .chnum = 50, .reg = IDMAC_CH_LOCK_EN_2, .shift = 10, },
+};
+
+int ipu_idmac_lock_enable(struct ipuv3_channel *channel, int num_bursts)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 bursts, regval;
+	int i;
+
+	switch (num_bursts) {
+	case 0:
+	case 1:
+		bursts = 0x00; /* locking disabled */
+		break;
+	case 2:
+		bursts = 0x01;
+		break;
+	case 4:
+		bursts = 0x02;
+		break;
+	case 8:
+		bursts = 0x03;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(idmac_lock_en_info); i++) {
+		if (channel->num == idmac_lock_en_info[i].chnum)
+			break;
+	}
+	if (i >= ARRAY_SIZE(idmac_lock_en_info))
+		return -EINVAL;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	regval = ipu_idmac_read(ipu, idmac_lock_en_info[i].reg);
+	regval &= ~(0x03 << idmac_lock_en_info[i].shift);
+	regval |= (bursts << idmac_lock_en_info[i].shift);
+	ipu_idmac_write(ipu, regval, idmac_lock_en_info[i].reg);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_lock_enable);
+
 int ipu_module_enable(struct ipu_soc *ipu, u32 mask)
 {
 	unsigned long lock_flags;
@@ -661,30 +447,6 @@ int ipu_module_disable(struct ipu_soc *ipu, u32 mask)
 }
 EXPORT_SYMBOL_GPL(ipu_module_disable);
 
-int ipu_csi_enable(struct ipu_soc *ipu, int csi)
-{
-	return ipu_module_enable(ipu, csi ? IPU_CONF_CSI1_EN : IPU_CONF_CSI0_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_csi_enable);
-
-int ipu_csi_disable(struct ipu_soc *ipu, int csi)
-{
-	return ipu_module_disable(ipu, csi ? IPU_CONF_CSI1_EN : IPU_CONF_CSI0_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_csi_disable);
-
-int ipu_smfc_enable(struct ipu_soc *ipu)
-{
-	return ipu_module_enable(ipu, IPU_CONF_SMFC_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_smfc_enable);
-
-int ipu_smfc_disable(struct ipu_soc *ipu)
-{
-	return ipu_module_disable(ipu, IPU_CONF_SMFC_EN);
-}
-EXPORT_SYMBOL_GPL(ipu_smfc_disable);
-
 int ipu_idmac_get_current_buffer(struct ipuv3_channel *channel)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -694,6 +456,30 @@ int ipu_idmac_get_current_buffer(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_get_current_buffer);
 
+bool ipu_idmac_buffer_is_ready(struct ipuv3_channel *channel, u32 buf_num)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 reg = 0;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+	switch (buf_num) {
+	case 0:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF0_RDY(channel->num));
+		break;
+	case 1:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF1_RDY(channel->num));
+		break;
+	case 2:
+		reg = ipu_cm_read(ipu, IPU_CHA_BUF2_RDY(channel->num));
+		break;
+	}
+	spin_unlock_irqrestore(&ipu->lock, flags);
+
+	return ((reg & idma_mask(channel->num)) != 0);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_buffer_is_ready);
+
 void ipu_idmac_select_buffer(struct ipuv3_channel *channel, u32 buf_num)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -712,6 +498,34 @@ void ipu_idmac_select_buffer(struct ipuv3_channel *channel, u32 buf_num)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_select_buffer);
 
+void ipu_idmac_clear_buffer(struct ipuv3_channel *channel, u32 buf_num)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned int chno = channel->num;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	ipu_cm_write(ipu, 0xF0300000, IPU_GPR); /* write one to clear */
+	switch (buf_num) {
+	case 0:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF0_RDY(chno));
+		break;
+	case 1:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF1_RDY(chno));
+		break;
+	case 2:
+		ipu_cm_write(ipu, idma_mask(chno), IPU_CHA_BUF2_RDY(chno));
+		break;
+	default:
+		break;
+	}
+	ipu_cm_write(ipu, 0x0, IPU_GPR); /* write one to set */
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_clear_buffer);
+
 int ipu_idmac_enable_channel(struct ipuv3_channel *channel)
 {
 	struct ipu_soc *ipu = channel->ipu;
@@ -782,6 +596,8 @@ int ipu_idmac_disable_channel(struct ipuv3_channel *channel)
 	val &= ~idma_mask(channel->num);
 	ipu_idmac_write(ipu, val, IDMAC_CHA_EN(channel->num));
 
+	__ipu_idmac_reset_current_buffer(channel);
+
 	/* Set channel buffers NOT to be ready */
 	ipu_cm_write(ipu, 0xf0000000, IPU_GPR); /* write one to clear */
 
@@ -810,6 +626,31 @@ int ipu_idmac_disable_channel(struct ipuv3_channel *channel)
 }
 EXPORT_SYMBOL_GPL(ipu_idmac_disable_channel);
 
+/*
+ * The imx6 rev. D TRM says that enabling the WM feature will increase
+ * a channel's priority. Refer to Table 36-8 Calculated priority value.
+ * The sub-module that is the sink or source for the channel must enable
+ * watermark signal for this to take effect (SMFC_WM for instance).
+ */
+void ipu_idmac_enable_watermark(struct ipuv3_channel *channel, bool enable)
+{
+	struct ipu_soc *ipu = channel->ipu;
+	unsigned long flags;
+	u32 val;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_idmac_read(ipu, IDMAC_WM_EN(channel->num));
+	if (enable)
+		val |= 1 << (channel->num % 32);
+	else
+		val &= ~(1 << (channel->num % 32));
+	ipu_idmac_write(ipu, val, IDMAC_WM_EN(channel->num));
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_idmac_enable_watermark);
+
 static int ipu_memory_reset(struct ipu_soc *ipu)
 {
 	unsigned long timeout;
@@ -826,12 +667,66 @@ static int ipu_memory_reset(struct ipu_soc *ipu)
 	return 0;
 }
 
+/*
+ * Set the source mux for the given CSI. Selects either parallel or
+ * MIPI CSI2 sources.
+ */
+void ipu_set_csi_src_mux(struct ipu_soc *ipu, int csi_id, bool mipi_csi2)
+{
+	unsigned long flags;
+	u32 val, mask;
+
+	mask = (csi_id == 1) ? IPU_CONF_CSI1_DATA_SOURCE :
+		IPU_CONF_CSI0_DATA_SOURCE;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_cm_read(ipu, IPU_CONF);
+	if (mipi_csi2)
+		val |= mask;
+	else
+		val &= ~mask;
+	ipu_cm_write(ipu, val, IPU_CONF);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_set_csi_src_mux);
+
+/*
+ * Set the source mux for the IC. Selects either CSI[01] or the VDI.
+ */
+void ipu_set_ic_src_mux(struct ipu_soc *ipu, int csi_id, bool vdi)
+{
+	unsigned long flags;
+	u32 val;
+
+	spin_lock_irqsave(&ipu->lock, flags);
+
+	val = ipu_cm_read(ipu, IPU_CONF);
+	if (vdi) {
+		val |= IPU_CONF_IC_INPUT;
+	} else {
+		val &= ~IPU_CONF_IC_INPUT;
+		if (csi_id == 1)
+			val |= IPU_CONF_CSI_SEL;
+		else
+			val &= ~IPU_CONF_CSI_SEL;
+	}
+	ipu_cm_write(ipu, val, IPU_CONF);
+
+	spin_unlock_irqrestore(&ipu->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_set_ic_src_mux);
+
 struct ipu_devtype {
 	const char *name;
 	unsigned long cm_ofs;
 	unsigned long cpmem_ofs;
 	unsigned long srm_ofs;
 	unsigned long tpm_ofs;
+	unsigned long csi0_ofs;
+	unsigned long csi1_ofs;
+	unsigned long ic_ofs;
 	unsigned long disp0_ofs;
 	unsigned long disp1_ofs;
 	unsigned long dc_tmpl_ofs;
@@ -845,6 +740,9 @@ static struct ipu_devtype ipu_type_imx51 = {
 	.cpmem_ofs = 0x1f000000,
 	.srm_ofs = 0x1f040000,
 	.tpm_ofs = 0x1f060000,
+	.csi0_ofs = 0x1f030000,
+	.csi1_ofs = 0x1f038000,
+	.ic_ofs = 0x1f020000,
 	.disp0_ofs = 0x1e040000,
 	.disp1_ofs = 0x1e048000,
 	.dc_tmpl_ofs = 0x1f080000,
@@ -858,6 +756,9 @@ static struct ipu_devtype ipu_type_imx53 = {
 	.cpmem_ofs = 0x07000000,
 	.srm_ofs = 0x07040000,
 	.tpm_ofs = 0x07060000,
+	.csi0_ofs = 0x07030000,
+	.csi1_ofs = 0x07038000,
+	.ic_ofs = 0x07020000,
 	.disp0_ofs = 0x06040000,
 	.disp1_ofs = 0x06048000,
 	.dc_tmpl_ofs = 0x07080000,
@@ -871,6 +772,9 @@ static struct ipu_devtype ipu_type_imx6q = {
 	.cpmem_ofs = 0x00300000,
 	.srm_ofs = 0x00340000,
 	.tpm_ofs = 0x00360000,
+	.csi0_ofs = 0x00230000,
+	.csi1_ofs = 0x00238000,
+	.ic_ofs = 0x00220000,
 	.disp0_ofs = 0x00240000,
 	.disp1_ofs = 0x00248000,
 	.dc_tmpl_ofs = 0x00380000,
@@ -895,8 +799,36 @@ static int ipu_submodules_init(struct ipu_soc *ipu,
 	struct device *dev = &pdev->dev;
 	const struct ipu_devtype *devtype = ipu->devtype;
 
+	ret = ipu_cpmem_init(ipu, dev, ipu_base + devtype->cpmem_ofs);
+	if (ret) {
+		unit = "cpmem";
+		goto err_cpmem;
+	}
+
+	ret = ipu_csi_init(ipu, dev, 0, ipu_base + devtype->csi0_ofs,
+			   IPU_CONF_CSI0_EN, ipu_clk);
+	if (ret) {
+		unit = "csi0";
+		goto err_csi_0;
+	}
+
+	ret = ipu_csi_init(ipu, dev, 1, ipu_base + devtype->csi1_ofs,
+			   IPU_CONF_CSI1_EN, ipu_clk);
+	if (ret) {
+		unit = "csi1";
+		goto err_csi_1;
+	}
+
+	ret = ipu_ic_init(ipu, dev,
+			  ipu_base + devtype->ic_ofs,
+			  ipu_base + devtype->tpm_ofs);
+	if (ret) {
+		unit = "ic";
+		goto err_ic;
+	}
+
 	ret = ipu_di_init(ipu, dev, 0, ipu_base + devtype->disp0_ofs,
-			IPU_CONF_DI0_EN, ipu_clk);
+			  IPU_CONF_DI0_EN, ipu_clk);
 	if (ret) {
 		unit = "di0";
 		goto err_di_0;
@@ -949,6 +881,14 @@ err_dc:
 err_di_1:
 	ipu_di_exit(ipu, 0);
 err_di_0:
+	ipu_ic_exit(ipu);
+err_ic:
+	ipu_csi_exit(ipu, 1);
+err_csi_1:
+	ipu_csi_exit(ipu, 0);
+err_csi_0:
+	ipu_cpmem_exit(ipu);
+err_cpmem:
 	dev_err(&pdev->dev, "init %s failed with %d\n", unit, ret);
 	return ret;
 }
@@ -1025,6 +965,10 @@ static void ipu_submodules_exit(struct ipu_soc *ipu)
 	ipu_dc_exit(ipu);
 	ipu_di_exit(ipu, 1);
 	ipu_di_exit(ipu, 0);
+	ipu_ic_exit(ipu);
+	ipu_csi_exit(ipu, 1);
+	ipu_csi_exit(ipu, 0);
+	ipu_cpmem_exit(ipu);
 }
 
 static int platform_remove_devices_fn(struct device *dev, void *unused)
@@ -1116,8 +1060,10 @@ static int ipu_add_client_devices(struct ipu_soc *ipu, unsigned long ipu_base)
 				id++, &reg->pdata, sizeof(reg->pdata));
 		}
 
-		if (IS_ERR(pdev))
+		if (IS_ERR(pdev)) {
+			ret = PTR_ERR(pdev);
 			goto err_register;
+		}
 	}
 
 	return 0;
@@ -1201,6 +1147,44 @@ static void ipu_irq_exit(struct ipu_soc *ipu)
 	irq_domain_remove(ipu->domain);
 }
 
+void ipu_dump(struct ipu_soc *ipu)
+{
+	int i;
+
+	dev_dbg(ipu->dev, "IPU_CONF = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CONF));
+	dev_dbg(ipu->dev, "IDMAC_CONF = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CONF));
+	dev_dbg(ipu->dev, "IDMAC_CHA_EN1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_EN(0)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_EN2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_EN(32)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_PRI1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_PRI(0)));
+	dev_dbg(ipu->dev, "IDMAC_CHA_PRI2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_CHA_PRI(32)));
+	dev_dbg(ipu->dev, "IDMAC_BAND_EN1 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_BAND_EN(0)));
+	dev_dbg(ipu->dev, "IDMAC_BAND_EN2 = \t0x%08X\n",
+		ipu_idmac_read(ipu, IDMAC_BAND_EN(32)));
+	dev_dbg(ipu->dev, "IPU_CHA_DB_MODE_SEL0 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CHA_DB_MODE_SEL(0)));
+	dev_dbg(ipu->dev, "IPU_CHA_DB_MODE_SEL1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_CHA_DB_MODE_SEL(32)));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW1));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW2 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW2));
+	dev_dbg(ipu->dev, "IPU_FS_PROC_FLOW3 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_PROC_FLOW3));
+	dev_dbg(ipu->dev, "IPU_FS_DISP_FLOW1 = \t0x%08X\n",
+		ipu_cm_read(ipu, IPU_FS_DISP_FLOW1));
+	for (i = 0; i < 15; i++)
+		dev_dbg(ipu->dev, "IPU_INT_CTRL(%d) = \t%08X\n", i,
+			ipu_cm_read(ipu, IPU_INT_CTRL(i)));
+}
+EXPORT_SYMBOL_GPL(ipu_dump);
+
 static int ipu_probe(struct platform_device *pdev)
 {
 	const struct of_device_id *of_id =
@@ -1243,6 +1227,12 @@ static int ipu_probe(struct platform_device *pdev)
 			ipu_base + devtype->cm_ofs + IPU_CM_IDMAC_REG_OFS);
 	dev_dbg(&pdev->dev, "cpmem:    0x%08lx\n",
 			ipu_base + devtype->cpmem_ofs);
+	dev_dbg(&pdev->dev, "csi0:    0x%08lx\n",
+			ipu_base + devtype->csi0_ofs);
+	dev_dbg(&pdev->dev, "csi1:    0x%08lx\n",
+			ipu_base + devtype->csi1_ofs);
+	dev_dbg(&pdev->dev, "ic:      0x%08lx\n",
+			ipu_base + devtype->ic_ofs);
 	dev_dbg(&pdev->dev, "disp0:    0x%08lx\n",
 			ipu_base + devtype->disp0_ofs);
 	dev_dbg(&pdev->dev, "disp1:    0x%08lx\n",
@@ -1265,10 +1255,8 @@ static int ipu_probe(struct platform_device *pdev)
 	ipu->idmac_reg = devm_ioremap(&pdev->dev,
 			ipu_base + devtype->cm_ofs + IPU_CM_IDMAC_REG_OFS,
 			PAGE_SIZE);
-	ipu->cpmem_base = devm_ioremap(&pdev->dev,
-			ipu_base + devtype->cpmem_ofs, PAGE_SIZE);
 
-	if (!ipu->cm_reg || !ipu->idmac_reg || !ipu->cpmem_base)
+	if (!ipu->cm_reg || !ipu->idmac_reg)
 		return -ENOMEM;
 
 	ipu->clk = devm_clk_get(&pdev->dev, "bus");
diff --git a/drivers/gpu/ipu-v3/ipu-cpmem.c b/drivers/gpu/ipu-v3/ipu-cpmem.c
new file mode 100644
index 000000000000..3bf05bc4ab67
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-cpmem.c
@@ -0,0 +1,764 @@
+/*
+ * Copyright (C) 2012 Mentor Graphics Inc.
+ * Copyright 2005-2012 Freescale Semiconductor, Inc. All Rights Reserved.
+ *
+ * The code contained herein is licensed under the GNU General Public
+ * License. You may obtain a copy of the GNU General Public License
+ * Version 2 or later at the following locations:
+ *
+ * http://www.opensource.org/licenses/gpl-license.html
+ * http://www.gnu.org/copyleft/gpl.html
+ */
+#include <linux/types.h>
+#include <linux/bitrev.h>
+#include <linux/io.h>
+#include <drm/drm_fourcc.h>
+#include "ipu-prv.h"
+
+struct ipu_cpmem_word {
+	u32 data[5];
+	u32 res[3];
+};
+
+struct ipu_ch_param {
+	struct ipu_cpmem_word word[2];
+};
+
+struct ipu_cpmem {
+	struct ipu_ch_param __iomem *base;
+	u32 module;
+	spinlock_t lock;
+	int use_count;
+	struct ipu_soc *ipu;
+};
+
+#define IPU_CPMEM_WORD(word, ofs, size) ((((word) * 160 + (ofs)) << 8) | (size))
+
+#define IPU_FIELD_UBO		IPU_CPMEM_WORD(0, 46, 22)
+#define IPU_FIELD_VBO		IPU_CPMEM_WORD(0, 68, 22)
+#define IPU_FIELD_IOX		IPU_CPMEM_WORD(0, 90, 4)
+#define IPU_FIELD_RDRW		IPU_CPMEM_WORD(0, 94, 1)
+#define IPU_FIELD_SO		IPU_CPMEM_WORD(0, 113, 1)
+#define IPU_FIELD_SLY		IPU_CPMEM_WORD(1, 102, 14)
+#define IPU_FIELD_SLUV		IPU_CPMEM_WORD(1, 128, 14)
+
+#define IPU_FIELD_XV		IPU_CPMEM_WORD(0, 0, 10)
+#define IPU_FIELD_YV		IPU_CPMEM_WORD(0, 10, 9)
+#define IPU_FIELD_XB		IPU_CPMEM_WORD(0, 19, 13)
+#define IPU_FIELD_YB		IPU_CPMEM_WORD(0, 32, 12)
+#define IPU_FIELD_NSB_B		IPU_CPMEM_WORD(0, 44, 1)
+#define IPU_FIELD_CF		IPU_CPMEM_WORD(0, 45, 1)
+#define IPU_FIELD_SX		IPU_CPMEM_WORD(0, 46, 12)
+#define IPU_FIELD_SY		IPU_CPMEM_WORD(0, 58, 11)
+#define IPU_FIELD_NS		IPU_CPMEM_WORD(0, 69, 10)
+#define IPU_FIELD_SDX		IPU_CPMEM_WORD(0, 79, 7)
+#define IPU_FIELD_SM		IPU_CPMEM_WORD(0, 86, 10)
+#define IPU_FIELD_SCC		IPU_CPMEM_WORD(0, 96, 1)
+#define IPU_FIELD_SCE		IPU_CPMEM_WORD(0, 97, 1)
+#define IPU_FIELD_SDY		IPU_CPMEM_WORD(0, 98, 7)
+#define IPU_FIELD_SDRX		IPU_CPMEM_WORD(0, 105, 1)
+#define IPU_FIELD_SDRY		IPU_CPMEM_WORD(0, 106, 1)
+#define IPU_FIELD_BPP		IPU_CPMEM_WORD(0, 107, 3)
+#define IPU_FIELD_DEC_SEL	IPU_CPMEM_WORD(0, 110, 2)
+#define IPU_FIELD_DIM		IPU_CPMEM_WORD(0, 112, 1)
+#define IPU_FIELD_BNDM		IPU_CPMEM_WORD(0, 114, 3)
+#define IPU_FIELD_BM		IPU_CPMEM_WORD(0, 117, 2)
+#define IPU_FIELD_ROT		IPU_CPMEM_WORD(0, 119, 1)
+#define IPU_FIELD_ROT_HF_VF	IPU_CPMEM_WORD(0, 119, 3)
+#define IPU_FIELD_HF		IPU_CPMEM_WORD(0, 120, 1)
+#define IPU_FIELD_VF		IPU_CPMEM_WORD(0, 121, 1)
+#define IPU_FIELD_THE		IPU_CPMEM_WORD(0, 122, 1)
+#define IPU_FIELD_CAP		IPU_CPMEM_WORD(0, 123, 1)
+#define IPU_FIELD_CAE		IPU_CPMEM_WORD(0, 124, 1)
+#define IPU_FIELD_FW		IPU_CPMEM_WORD(0, 125, 13)
+#define IPU_FIELD_FH		IPU_CPMEM_WORD(0, 138, 12)
+#define IPU_FIELD_EBA0		IPU_CPMEM_WORD(1, 0, 29)
+#define IPU_FIELD_EBA1		IPU_CPMEM_WORD(1, 29, 29)
+#define IPU_FIELD_ILO		IPU_CPMEM_WORD(1, 58, 20)
+#define IPU_FIELD_NPB		IPU_CPMEM_WORD(1, 78, 7)
+#define IPU_FIELD_PFS		IPU_CPMEM_WORD(1, 85, 4)
+#define IPU_FIELD_ALU		IPU_CPMEM_WORD(1, 89, 1)
+#define IPU_FIELD_ALBM		IPU_CPMEM_WORD(1, 90, 3)
+#define IPU_FIELD_ID		IPU_CPMEM_WORD(1, 93, 2)
+#define IPU_FIELD_TH		IPU_CPMEM_WORD(1, 95, 7)
+#define IPU_FIELD_SL		IPU_CPMEM_WORD(1, 102, 14)
+#define IPU_FIELD_WID0		IPU_CPMEM_WORD(1, 116, 3)
+#define IPU_FIELD_WID1		IPU_CPMEM_WORD(1, 119, 3)
+#define IPU_FIELD_WID2		IPU_CPMEM_WORD(1, 122, 3)
+#define IPU_FIELD_WID3		IPU_CPMEM_WORD(1, 125, 3)
+#define IPU_FIELD_OFS0		IPU_CPMEM_WORD(1, 128, 5)
+#define IPU_FIELD_OFS1		IPU_CPMEM_WORD(1, 133, 5)
+#define IPU_FIELD_OFS2		IPU_CPMEM_WORD(1, 138, 5)
+#define IPU_FIELD_OFS3		IPU_CPMEM_WORD(1, 143, 5)
+#define IPU_FIELD_SXYS		IPU_CPMEM_WORD(1, 148, 1)
+#define IPU_FIELD_CRE		IPU_CPMEM_WORD(1, 149, 1)
+#define IPU_FIELD_DEC_SEL2	IPU_CPMEM_WORD(1, 150, 1)
+
+static inline struct ipu_ch_param __iomem *
+ipu_get_cpmem(struct ipuv3_channel *ch)
+{
+	struct ipu_cpmem *cpmem = ch->ipu->cpmem_priv;
+
+	return cpmem->base + ch->num;
+}
+
+static void ipu_ch_param_write_field(struct ipuv3_channel *ch, u32 wbs, u32 v)
+{
+	struct ipu_ch_param __iomem *base = ipu_get_cpmem(ch);
+	u32 bit = (wbs >> 8) % 160;
+	u32 size = wbs & 0xff;
+	u32 word = (wbs >> 8) / 160;
+	u32 i = bit / 32;
+	u32 ofs = bit % 32;
+	u32 mask = (1 << size) - 1;
+	u32 val;
+
+	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
+
+	val = readl(&base->word[word].data[i]);
+	val &= ~(mask << ofs);
+	val |= v << ofs;
+	writel(val, &base->word[word].data[i]);
+
+	if ((bit + size - 1) / 32 > i) {
+		val = readl(&base->word[word].data[i + 1]);
+		val &= ~(mask >> (ofs ? (32 - ofs) : 0));
+		val |= v >> (ofs ? (32 - ofs) : 0);
+		writel(val, &base->word[word].data[i + 1]);
+	}
+}
+
+static u32 ipu_ch_param_read_field(struct ipuv3_channel *ch, u32 wbs)
+{
+	struct ipu_ch_param __iomem *base = ipu_get_cpmem(ch);
+	u32 bit = (wbs >> 8) % 160;
+	u32 size = wbs & 0xff;
+	u32 word = (wbs >> 8) / 160;
+	u32 i = bit / 32;
+	u32 ofs = bit % 32;
+	u32 mask = (1 << size) - 1;
+	u32 val = 0;
+
+	pr_debug("%s %d %d %d\n", __func__, word, bit , size);
+
+	val = (readl(&base->word[word].data[i]) >> ofs) & mask;
+
+	if ((bit + size - 1) / 32 > i) {
+		u32 tmp;
+
+		tmp = readl(&base->word[word].data[i + 1]);
+		tmp &= mask >> (ofs ? (32 - ofs) : 0);
+		val |= tmp << (ofs ? (32 - ofs) : 0);
+	}
+
+	return val;
+}
+
+/*
+ * The V4L2 spec defines packed RGB formats in memory byte order, which from
+ * point of view of the IPU corresponds to little-endian words with the first
+ * component in the least significant bits.
+ * The DRM pixel formats and IPU internal representation are ordered the other
+ * way around, with the first named component ordered at the most significant
+ * bits. Further, V4L2 formats are not well defined:
+ *     http://linuxtv.org/downloads/v4l-dvb-apis/packed-rgb.html
+ * We choose the interpretation which matches GStreamer behavior.
+ */
+static int v4l2_pix_fmt_to_drm_fourcc(u32 pixelformat)
+{
+	switch (pixelformat) {
+	case V4L2_PIX_FMT_RGB565:
+		/*
+		 * Here we choose the 'corrected' interpretation of RGBP, a
+		 * little-endian 16-bit word with the red component at the most
+		 * significant bits:
+		 * g[2:0]b[4:0] r[4:0]g[5:3] <=> [16:0] R:G:B
+		 */
+		return DRM_FORMAT_RGB565;
+	case V4L2_PIX_FMT_BGR24:
+		/* B G R <=> [24:0] R:G:B */
+		return DRM_FORMAT_RGB888;
+	case V4L2_PIX_FMT_RGB24:
+		/* R G B <=> [24:0] B:G:R */
+		return DRM_FORMAT_BGR888;
+	case V4L2_PIX_FMT_BGR32:
+		/* B G R A <=> [32:0] A:B:G:R */
+		return DRM_FORMAT_XRGB8888;
+	case V4L2_PIX_FMT_RGB32:
+		/* R G B A <=> [32:0] A:B:G:R */
+		return DRM_FORMAT_XBGR8888;
+	case V4L2_PIX_FMT_UYVY:
+		return DRM_FORMAT_UYVY;
+	case V4L2_PIX_FMT_YUYV:
+		return DRM_FORMAT_YUYV;
+	case V4L2_PIX_FMT_YUV420:
+		return DRM_FORMAT_YUV420;
+	case V4L2_PIX_FMT_YUV422P:
+		return DRM_FORMAT_YUV422;
+	case V4L2_PIX_FMT_YVU420:
+		return DRM_FORMAT_YVU420;
+	case V4L2_PIX_FMT_NV12:
+		return DRM_FORMAT_NV12;
+	case V4L2_PIX_FMT_NV16:
+		return DRM_FORMAT_NV16;
+	}
+
+	return -EINVAL;
+}
+
+void ipu_cpmem_zero(struct ipuv3_channel *ch)
+{
+	struct ipu_ch_param __iomem *p = ipu_get_cpmem(ch);
+	void __iomem *base = p;
+	int i;
+
+	for (i = 0; i < sizeof(*p) / sizeof(u32); i++)
+		writel(0, base + i * sizeof(u32));
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_zero);
+
+void ipu_cpmem_set_resolution(struct ipuv3_channel *ch, int xres, int yres)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_FW, xres - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_FH, yres - 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_resolution);
+
+void ipu_cpmem_set_stride(struct ipuv3_channel *ch, int stride)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_SLY, stride - 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_stride);
+
+void ipu_cpmem_set_high_priority(struct ipuv3_channel *ch)
+{
+	struct ipu_soc *ipu = ch->ipu;
+	u32 val;
+
+	if (ipu->ipu_type == IPUV3EX)
+		ipu_ch_param_write_field(ch, IPU_FIELD_ID, 1);
+
+	val = ipu_idmac_read(ipu, IDMAC_CHA_PRI(ch->num));
+	val |= 1 << (ch->num % 32);
+	ipu_idmac_write(ipu, val, IDMAC_CHA_PRI(ch->num));
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_high_priority);
+
+void ipu_cpmem_set_buffer(struct ipuv3_channel *ch, int bufnum, dma_addr_t buf)
+{
+	if (bufnum)
+		ipu_ch_param_write_field(ch, IPU_FIELD_EBA1, buf >> 3);
+	else
+		ipu_ch_param_write_field(ch, IPU_FIELD_EBA0, buf >> 3);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_buffer);
+
+void ipu_cpmem_interlaced_scan(struct ipuv3_channel *ch, int stride)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_SO, 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_ILO, stride / 8);
+	ipu_ch_param_write_field(ch, IPU_FIELD_SLY, (stride * 2) - 1);
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_interlaced_scan);
+
+void ipu_cpmem_set_axi_id(struct ipuv3_channel *ch, u32 id)
+{
+	id &= 0x3;
+	ipu_ch_param_write_field(ch, IPU_FIELD_ID, id);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_axi_id);
+
+void ipu_cpmem_set_burstsize(struct ipuv3_channel *ch, int burstsize)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, burstsize - 1);
+};
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_burstsize);
+
+void ipu_cpmem_set_block_mode(struct ipuv3_channel *ch)
+{
+	ipu_ch_param_write_field(ch, IPU_FIELD_BM, 1);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_block_mode);
+
+void ipu_cpmem_set_rotation(struct ipuv3_channel *ch,
+			    enum ipu_rotate_mode rot)
+{
+	u32 temp_rot = bitrev8(rot) >> 5;
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_ROT_HF_VF, temp_rot);
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_rotation);
+
+int ipu_cpmem_set_format_rgb(struct ipuv3_channel *ch,
+			     const struct ipu_rgb *rgb)
+{
+	int bpp = 0, npb = 0, ro, go, bo, to;
+
+	ro = rgb->bits_per_pixel - rgb->red.length - rgb->red.offset;
+	go = rgb->bits_per_pixel - rgb->green.length - rgb->green.offset;
+	bo = rgb->bits_per_pixel - rgb->blue.length - rgb->blue.offset;
+	to = rgb->bits_per_pixel - rgb->transp.length - rgb->transp.offset;
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID0, rgb->red.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS0, ro);
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID1, rgb->green.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS1, go);
+	ipu_ch_param_write_field(ch, IPU_FIELD_WID2, rgb->blue.length - 1);
+	ipu_ch_param_write_field(ch, IPU_FIELD_OFS2, bo);
+
+	if (rgb->transp.length) {
+		ipu_ch_param_write_field(ch, IPU_FIELD_WID3,
+				rgb->transp.length - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_OFS3, to);
+	} else {
+		ipu_ch_param_write_field(ch, IPU_FIELD_WID3, 7);
+		ipu_ch_param_write_field(ch, IPU_FIELD_OFS3,
+				rgb->bits_per_pixel);
+	}
+
+	switch (rgb->bits_per_pixel) {
+	case 32:
+		bpp = 0;
+		npb = 15;
+		break;
+	case 24:
+		bpp = 1;
+		npb = 19;
+		break;
+	case 16:
+		bpp = 3;
+		npb = 31;
+		break;
+	case 8:
+		bpp = 5;
+		npb = 63;
+		break;
+	default:
+		return -EINVAL;
+	}
+	ipu_ch_param_write_field(ch, IPU_FIELD_BPP, bpp);
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, npb);
+	ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 7); /* rgb mode */
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_rgb);
+
+int ipu_cpmem_set_format_passthrough(struct ipuv3_channel *ch, int width)
+{
+	int bpp = 0, npb = 0;
+
+	switch (width) {
+	case 32:
+		bpp = 0;
+		npb = 15;
+		break;
+	case 24:
+		bpp = 1;
+		npb = 19;
+		break;
+	case 16:
+		bpp = 3;
+		npb = 31;
+		break;
+	case 8:
+		bpp = 5;
+		npb = 63;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ipu_ch_param_write_field(ch, IPU_FIELD_BPP, bpp);
+	ipu_ch_param_write_field(ch, IPU_FIELD_NPB, npb);
+	ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 6); /* raw mode */
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_format_passthrough);
+
+void ipu_cpmem_set_yuv_interleaved(struct ipuv3_channel *ch, u32 pixel_format)
+{
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_UYVY:
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3); /* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0xA);/* pix fmt */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);/* burst size */
+		break;
+	case V4L2_PIX_FMT_YUYV:
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3); /* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0x8);/* pix fmt */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);/* burst size */
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_interleaved);
+
+void ipu_cpmem_set_yuv_planar_full(struct ipuv3_channel *ch,
+				   u32 pixel_format, int stride,
+				   int u_offset, int v_offset)
+{
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YUV422P:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, (stride / 2) - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, u_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, v_offset / 8);
+		break;
+	case V4L2_PIX_FMT_YVU420:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, (stride / 2) - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, v_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, u_offset / 8);
+		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV16:
+		ipu_ch_param_write_field(ch, IPU_FIELD_SLUV, stride - 1);
+		ipu_ch_param_write_field(ch, IPU_FIELD_UBO, u_offset / 8);
+		ipu_ch_param_write_field(ch, IPU_FIELD_VBO, u_offset / 8);
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar_full);
+
+void ipu_cpmem_set_yuv_planar(struct ipuv3_channel *ch,
+			      u32 pixel_format, int stride, int height)
+{
+	int u_offset, v_offset;
+	int uv_stride = 0;
+
+	switch (pixel_format) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+		uv_stride = stride / 2;
+		u_offset = stride * height;
+		v_offset = u_offset + (uv_stride * height / 2);
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_YUV422P:
+		uv_stride = stride / 2;
+		u_offset = stride * height;
+		v_offset = u_offset + (uv_stride * height);
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV12:
+	case V4L2_PIX_FMT_NV16:
+		u_offset = stride * height;
+		ipu_cpmem_set_yuv_planar_full(ch, pixel_format, stride,
+					      u_offset, 0);
+		break;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_yuv_planar);
+
+static const struct ipu_rgb def_rgb_32 = {
+	.red	= { .offset = 16, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset =  0, .length = 8, },
+	.transp = { .offset = 24, .length = 8, },
+	.bits_per_pixel = 32,
+};
+
+static const struct ipu_rgb def_bgr_32 = {
+	.red	= { .offset =  0, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset = 16, .length = 8, },
+	.transp = { .offset = 24, .length = 8, },
+	.bits_per_pixel = 32,
+};
+
+static const struct ipu_rgb def_rgb_24 = {
+	.red	= { .offset = 16, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset =  0, .length = 8, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 24,
+};
+
+static const struct ipu_rgb def_bgr_24 = {
+	.red	= { .offset =  0, .length = 8, },
+	.green	= { .offset =  8, .length = 8, },
+	.blue	= { .offset = 16, .length = 8, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 24,
+};
+
+static const struct ipu_rgb def_rgb_16 = {
+	.red	= { .offset = 11, .length = 5, },
+	.green	= { .offset =  5, .length = 6, },
+	.blue	= { .offset =  0, .length = 5, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 16,
+};
+
+static const struct ipu_rgb def_bgr_16 = {
+	.red	= { .offset =  0, .length = 5, },
+	.green	= { .offset =  5, .length = 6, },
+	.blue	= { .offset = 11, .length = 5, },
+	.transp = { .offset =  0, .length = 0, },
+	.bits_per_pixel = 16,
+};
+
+#define Y_OFFSET(pix, x, y)	((x) + pix->width * (y))
+#define U_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * (y) / 4) + (x) / 2)
+#define V_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * pix->height / 4) +	\
+				 (pix->width * (y) / 4) + (x) / 2)
+#define U2_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * (y) / 2) + (x) / 2)
+#define V2_OFFSET(pix, x, y)	((pix->width * pix->height) +		\
+				 (pix->width * pix->height / 2) +	\
+				 (pix->width * (y) / 2) + (x) / 2)
+#define UV_OFFSET(pix, x, y)	((pix->width * pix->height) +	\
+				 (pix->width * (y) / 2) + (x))
+#define UV2_OFFSET(pix, x, y)	((pix->width * pix->height) +	\
+				 (pix->width * y) + (x))
+
+int ipu_cpmem_set_fmt(struct ipuv3_channel *ch, u32 drm_fourcc)
+{
+	switch (drm_fourcc) {
+	case DRM_FORMAT_YUV420:
+	case DRM_FORMAT_YVU420:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 2);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_YUV422:
+	case DRM_FORMAT_YVU422:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 1);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_NV12:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 4);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_NV16:
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 3);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_UYVY:
+		/* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3);
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0xA);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_YUYV:
+		/* bits/pixel */
+		ipu_ch_param_write_field(ch, IPU_FIELD_BPP, 3);
+		/* pix format */
+		ipu_ch_param_write_field(ch, IPU_FIELD_PFS, 0x8);
+		/* burst size */
+		ipu_ch_param_write_field(ch, IPU_FIELD_NPB, 31);
+		break;
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_XBGR8888:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_32);
+		break;
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_XRGB8888:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_32);
+		break;
+	case DRM_FORMAT_BGR888:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_24);
+		break;
+	case DRM_FORMAT_RGB888:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_24);
+		break;
+	case DRM_FORMAT_RGB565:
+		ipu_cpmem_set_format_rgb(ch, &def_rgb_16);
+		break;
+	case DRM_FORMAT_BGR565:
+		ipu_cpmem_set_format_rgb(ch, &def_bgr_16);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_fmt);
+
+int ipu_cpmem_set_image(struct ipuv3_channel *ch, struct ipu_image *image)
+{
+	struct v4l2_pix_format *pix = &image->pix;
+	int offset, u_offset, v_offset;
+
+	pr_debug("%s: resolution: %dx%d stride: %d\n",
+		 __func__, pix->width, pix->height,
+		 pix->bytesperline);
+
+	ipu_cpmem_set_resolution(ch, image->rect.width, image->rect.height);
+	ipu_cpmem_set_stride(ch, pix->bytesperline);
+
+	ipu_cpmem_set_fmt(ch, v4l2_pix_fmt_to_drm_fourcc(pix->pixelformat));
+
+	switch (pix->pixelformat) {
+	case V4L2_PIX_FMT_YUV420:
+	case V4L2_PIX_FMT_YVU420:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = U_OFFSET(pix, image->rect.left,
+				    image->rect.top) - offset;
+		v_offset = V_OFFSET(pix, image->rect.left,
+				    image->rect.top) - offset;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_YUV422P:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = U2_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+		v_offset = V2_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV12:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = UV_OFFSET(pix, image->rect.left,
+				     image->rect.top) - offset;
+		v_offset = 0;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_NV16:
+		offset = Y_OFFSET(pix, image->rect.left, image->rect.top);
+		u_offset = UV2_OFFSET(pix, image->rect.left,
+				      image->rect.top) - offset;
+		v_offset = 0;
+
+		ipu_cpmem_set_yuv_planar_full(ch, pix->pixelformat,
+					      pix->bytesperline,
+					      u_offset, v_offset);
+		break;
+	case V4L2_PIX_FMT_UYVY:
+	case V4L2_PIX_FMT_YUYV:
+	case V4L2_PIX_FMT_RGB565:
+		offset = image->rect.left * 2 +
+			image->rect.top * pix->bytesperline;
+		break;
+	case V4L2_PIX_FMT_RGB32:
+	case V4L2_PIX_FMT_BGR32:
+		offset = image->rect.left * 4 +
+			image->rect.top * pix->bytesperline;
+		break;
+	case V4L2_PIX_FMT_RGB24:
+	case V4L2_PIX_FMT_BGR24:
+		offset = image->rect.left * 3 +
+			image->rect.top * pix->bytesperline;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ipu_cpmem_set_buffer(ch, 0, image->phys0 + offset);
+	ipu_cpmem_set_buffer(ch, 1, image->phys1 + offset);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_set_image);
+
+void ipu_cpmem_dump(struct ipuv3_channel *ch)
+{
+	struct ipu_ch_param __iomem *p = ipu_get_cpmem(ch);
+	struct ipu_soc *ipu = ch->ipu;
+	int chno = ch->num;
+
+	dev_dbg(ipu->dev, "ch %d word 0 - %08X %08X %08X %08X %08X\n", chno,
+		readl(&p->word[0].data[0]),
+		readl(&p->word[0].data[1]),
+		readl(&p->word[0].data[2]),
+		readl(&p->word[0].data[3]),
+		readl(&p->word[0].data[4]));
+	dev_dbg(ipu->dev, "ch %d word 1 - %08X %08X %08X %08X %08X\n", chno,
+		readl(&p->word[1].data[0]),
+		readl(&p->word[1].data[1]),
+		readl(&p->word[1].data[2]),
+		readl(&p->word[1].data[3]),
+		readl(&p->word[1].data[4]));
+	dev_dbg(ipu->dev, "PFS 0x%x, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_PFS));
+	dev_dbg(ipu->dev, "BPP 0x%x, ",
+		ipu_ch_param_read_field(ch, IPU_FIELD_BPP));
+	dev_dbg(ipu->dev, "NPB 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_NPB));
+
+	dev_dbg(ipu->dev, "FW %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_FW));
+	dev_dbg(ipu->dev, "FH %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_FH));
+	dev_dbg(ipu->dev, "EBA0 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_EBA0) << 3);
+	dev_dbg(ipu->dev, "EBA1 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_EBA1) << 3);
+	dev_dbg(ipu->dev, "Stride %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SL));
+	dev_dbg(ipu->dev, "scan_order %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SO));
+	dev_dbg(ipu->dev, "uv_stride %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_SLUV));
+	dev_dbg(ipu->dev, "u_offset 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_UBO) << 3);
+	dev_dbg(ipu->dev, "v_offset 0x%x\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_VBO) << 3);
+
+	dev_dbg(ipu->dev, "Width0 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID0));
+	dev_dbg(ipu->dev, "Width1 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID1));
+	dev_dbg(ipu->dev, "Width2 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID2));
+	dev_dbg(ipu->dev, "Width3 %d+1, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_WID3));
+	dev_dbg(ipu->dev, "Offset0 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS0));
+	dev_dbg(ipu->dev, "Offset1 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS1));
+	dev_dbg(ipu->dev, "Offset2 %d, ",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS2));
+	dev_dbg(ipu->dev, "Offset3 %d\n",
+		 ipu_ch_param_read_field(ch, IPU_FIELD_OFS3));
+}
+EXPORT_SYMBOL_GPL(ipu_cpmem_dump);
+
+int ipu_cpmem_init(struct ipu_soc *ipu, struct device *dev, unsigned long base)
+{
+	struct ipu_cpmem *cpmem;
+
+	cpmem = devm_kzalloc(dev, sizeof(*cpmem), GFP_KERNEL);
+	if (!cpmem)
+		return -ENOMEM;
+
+	ipu->cpmem_priv = cpmem;
+
+	spin_lock_init(&cpmem->lock);
+	cpmem->base = devm_ioremap(dev, base, SZ_128K);
+	if (!cpmem->base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "CPMEM base: 0x%08lx remapped to %p\n",
+		base, cpmem->base);
+	cpmem->ipu = ipu;
+
+	return 0;
+}
+
+void ipu_cpmem_exit(struct ipu_soc *ipu)
+{
+}
diff --git a/drivers/gpu/ipu-v3/ipu-csi.c b/drivers/gpu/ipu-v3/ipu-csi.c
new file mode 100644
index 000000000000..d6f56471bd2a
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-csi.c
@@ -0,0 +1,741 @@
+/*
+ * Copyright (C) 2012-2014 Mentor Graphics Inc.
+ * Copyright (C) 2005-2009 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or (at your
+ * option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+ * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ * for more details.
+ */
+#include <linux/export.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/err.h>
+#include <linux/platform_device.h>
+#include <linux/videodev2.h>
+#include <uapi/linux/v4l2-mediabus.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/clkdev.h>
+
+#include "ipu-prv.h"
+
+struct ipu_csi {
+	void __iomem *base;
+	int id;
+	u32 module;
+	struct clk *clk_ipu;	/* IPU bus clock */
+	spinlock_t lock;
+	bool inuse;
+	struct ipu_soc *ipu;
+};
+
+/* CSI Register Offsets */
+#define CSI_SENS_CONF		0x0000
+#define CSI_SENS_FRM_SIZE	0x0004
+#define CSI_ACT_FRM_SIZE	0x0008
+#define CSI_OUT_FRM_CTRL	0x000c
+#define CSI_TST_CTRL		0x0010
+#define CSI_CCIR_CODE_1		0x0014
+#define CSI_CCIR_CODE_2		0x0018
+#define CSI_CCIR_CODE_3		0x001c
+#define CSI_MIPI_DI		0x0020
+#define CSI_SKIP		0x0024
+#define CSI_CPD_CTRL		0x0028
+#define CSI_CPD_RC(n)		(0x002c + ((n)*4))
+#define CSI_CPD_RS(n)		(0x004c + ((n)*4))
+#define CSI_CPD_GRC(n)		(0x005c + ((n)*4))
+#define CSI_CPD_GRS(n)		(0x007c + ((n)*4))
+#define CSI_CPD_GBC(n)		(0x008c + ((n)*4))
+#define CSI_CPD_GBS(n)		(0x00Ac + ((n)*4))
+#define CSI_CPD_BC(n)		(0x00Bc + ((n)*4))
+#define CSI_CPD_BS(n)		(0x00Dc + ((n)*4))
+#define CSI_CPD_OFFSET1		0x00ec
+#define CSI_CPD_OFFSET2		0x00f0
+
+/* CSI Register Fields */
+#define CSI_SENS_CONF_DATA_FMT_SHIFT		8
+#define CSI_SENS_CONF_DATA_FMT_MASK		0x00000700
+#define CSI_SENS_CONF_DATA_FMT_RGB_YUV444	0L
+#define CSI_SENS_CONF_DATA_FMT_YUV422_YUYV	1L
+#define CSI_SENS_CONF_DATA_FMT_YUV422_UYVY	2L
+#define CSI_SENS_CONF_DATA_FMT_BAYER		3L
+#define CSI_SENS_CONF_DATA_FMT_RGB565		4L
+#define CSI_SENS_CONF_DATA_FMT_RGB555		5L
+#define CSI_SENS_CONF_DATA_FMT_RGB444		6L
+#define CSI_SENS_CONF_DATA_FMT_JPEG		7L
+
+#define CSI_SENS_CONF_VSYNC_POL_SHIFT		0
+#define CSI_SENS_CONF_HSYNC_POL_SHIFT		1
+#define CSI_SENS_CONF_DATA_POL_SHIFT		2
+#define CSI_SENS_CONF_PIX_CLK_POL_SHIFT		3
+#define CSI_SENS_CONF_SENS_PRTCL_MASK		0x00000070
+#define CSI_SENS_CONF_SENS_PRTCL_SHIFT		4
+#define CSI_SENS_CONF_PACK_TIGHT_SHIFT		7
+#define CSI_SENS_CONF_DATA_WIDTH_SHIFT		11
+#define CSI_SENS_CONF_EXT_VSYNC_SHIFT		15
+#define CSI_SENS_CONF_DIVRATIO_SHIFT		16
+
+#define CSI_SENS_CONF_DIVRATIO_MASK		0x00ff0000
+#define CSI_SENS_CONF_DATA_DEST_SHIFT		24
+#define CSI_SENS_CONF_DATA_DEST_MASK		0x07000000
+#define CSI_SENS_CONF_JPEG8_EN_SHIFT		27
+#define CSI_SENS_CONF_JPEG_EN_SHIFT		28
+#define CSI_SENS_CONF_FORCE_EOF_SHIFT		29
+#define CSI_SENS_CONF_DATA_EN_POL_SHIFT		31
+
+#define CSI_DATA_DEST_IC			2
+#define CSI_DATA_DEST_IDMAC			4
+
+#define CSI_CCIR_ERR_DET_EN			0x01000000
+#define CSI_HORI_DOWNSIZE_EN			0x80000000
+#define CSI_VERT_DOWNSIZE_EN			0x40000000
+#define CSI_TEST_GEN_MODE_EN			0x01000000
+
+#define CSI_HSC_MASK				0x1fff0000
+#define CSI_HSC_SHIFT				16
+#define CSI_VSC_MASK				0x00000fff
+#define CSI_VSC_SHIFT				0
+
+#define CSI_TEST_GEN_R_MASK			0x000000ff
+#define CSI_TEST_GEN_R_SHIFT			0
+#define CSI_TEST_GEN_G_MASK			0x0000ff00
+#define CSI_TEST_GEN_G_SHIFT			8
+#define CSI_TEST_GEN_B_MASK			0x00ff0000
+#define CSI_TEST_GEN_B_SHIFT			16
+
+#define CSI_MAX_RATIO_SKIP_SMFC_MASK		0x00000007
+#define CSI_MAX_RATIO_SKIP_SMFC_SHIFT		0
+#define CSI_SKIP_SMFC_MASK			0x000000f8
+#define CSI_SKIP_SMFC_SHIFT			3
+#define CSI_ID_2_SKIP_MASK			0x00000300
+#define CSI_ID_2_SKIP_SHIFT			8
+
+#define CSI_COLOR_FIRST_ROW_MASK		0x00000002
+#define CSI_COLOR_FIRST_COMP_MASK		0x00000001
+
+/* MIPI CSI-2 data types */
+#define MIPI_DT_YUV420		0x18 /* YYY.../UYVY.... */
+#define MIPI_DT_YUV420_LEGACY	0x1a /* UYY.../VYY...   */
+#define MIPI_DT_YUV422		0x1e /* UYVY...         */
+#define MIPI_DT_RGB444		0x20
+#define MIPI_DT_RGB555		0x21
+#define MIPI_DT_RGB565		0x22
+#define MIPI_DT_RGB666		0x23
+#define MIPI_DT_RGB888		0x24
+#define MIPI_DT_RAW6		0x28
+#define MIPI_DT_RAW7		0x29
+#define MIPI_DT_RAW8		0x2a
+#define MIPI_DT_RAW10		0x2b
+#define MIPI_DT_RAW12		0x2c
+#define MIPI_DT_RAW14		0x2d
+
+/*
+ * Bitfield of CSI bus signal polarities and modes.
+ */
+struct ipu_csi_bus_config {
+	unsigned data_width:4;
+	unsigned clk_mode:3;
+	unsigned ext_vsync:1;
+	unsigned vsync_pol:1;
+	unsigned hsync_pol:1;
+	unsigned pixclk_pol:1;
+	unsigned data_pol:1;
+	unsigned sens_clksrc:1;
+	unsigned pack_tight:1;
+	unsigned force_eof:1;
+	unsigned data_en_pol:1;
+
+	unsigned data_fmt;
+	unsigned mipi_dt;
+};
+
+/*
+ * Enumeration of CSI data bus widths.
+ */
+enum ipu_csi_data_width {
+	IPU_CSI_DATA_WIDTH_4   = 0,
+	IPU_CSI_DATA_WIDTH_8   = 1,
+	IPU_CSI_DATA_WIDTH_10  = 3,
+	IPU_CSI_DATA_WIDTH_12  = 5,
+	IPU_CSI_DATA_WIDTH_16  = 9,
+};
+
+/*
+ * Enumeration of CSI clock modes.
+ */
+enum ipu_csi_clk_mode {
+	IPU_CSI_CLK_MODE_GATED_CLK,
+	IPU_CSI_CLK_MODE_NONGATED_CLK,
+	IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE,
+	IPU_CSI_CLK_MODE_CCIR656_INTERLACED,
+	IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR,
+	IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR,
+	IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR,
+	IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR,
+};
+
+static inline u32 ipu_csi_read(struct ipu_csi *csi, unsigned offset)
+{
+	return readl(csi->base + offset);
+}
+
+static inline void ipu_csi_write(struct ipu_csi *csi, u32 value,
+				 unsigned offset)
+{
+	writel(value, csi->base + offset);
+}
+
+/*
+ * Set mclk division ratio for generating test mode mclk. Only used
+ * for test generator.
+ */
+static int ipu_csi_set_testgen_mclk(struct ipu_csi *csi, u32 pixel_clk,
+					u32 ipu_clk)
+{
+	u32 temp;
+	u32 div_ratio;
+
+	div_ratio = (ipu_clk / pixel_clk) - 1;
+
+	if (div_ratio > 0xFF || div_ratio < 0) {
+		dev_err(csi->ipu->dev,
+			"value of pixel_clk extends normal range\n");
+		return -EINVAL;
+	}
+
+	temp = ipu_csi_read(csi, CSI_SENS_CONF);
+	temp &= ~CSI_SENS_CONF_DIVRATIO_MASK;
+	ipu_csi_write(csi, temp | (div_ratio << CSI_SENS_CONF_DIVRATIO_SHIFT),
+			  CSI_SENS_CONF);
+
+	return 0;
+}
+
+/*
+ * Find the CSI data format and data width for the given V4L2 media
+ * bus pixel format code.
+ */
+static int mbus_code_to_bus_cfg(struct ipu_csi_bus_config *cfg, u32 mbus_code)
+{
+	switch (mbus_code) {
+	case V4L2_MBUS_FMT_BGR565_2X8_BE:
+	case V4L2_MBUS_FMT_BGR565_2X8_LE:
+	case V4L2_MBUS_FMT_RGB565_2X8_BE:
+	case V4L2_MBUS_FMT_RGB565_2X8_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB565;
+		cfg->mipi_dt = MIPI_DT_RGB565;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_RGB444_2X8_PADHI_BE:
+	case V4L2_MBUS_FMT_RGB444_2X8_PADHI_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB444;
+		cfg->mipi_dt = MIPI_DT_RGB444;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_RGB555_2X8_PADHI_BE:
+	case V4L2_MBUS_FMT_RGB555_2X8_PADHI_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_RGB555;
+		cfg->mipi_dt = MIPI_DT_RGB555;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_UYVY8_2X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_UYVY;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_YUYV8_2X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_YUYV;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_UYVY8_1X16:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_UYVY;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_16;
+		break;
+	case V4L2_MBUS_FMT_YUYV8_1X16:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_YUV422_YUYV;
+		cfg->mipi_dt = MIPI_DT_YUV422;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_16;
+		break;
+	case V4L2_MBUS_FMT_SBGGR8_1X8:
+	case V4L2_MBUS_FMT_SGBRG8_1X8:
+	case V4L2_MBUS_FMT_SGRBG8_1X8:
+	case V4L2_MBUS_FMT_SRGGB8_1X8:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW8;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_SBGGR10_DPCM8_1X8:
+	case V4L2_MBUS_FMT_SGBRG10_DPCM8_1X8:
+	case V4L2_MBUS_FMT_SGRBG10_DPCM8_1X8:
+	case V4L2_MBUS_FMT_SRGGB10_DPCM8_1X8:
+	case V4L2_MBUS_FMT_SBGGR10_2X8_PADHI_BE:
+	case V4L2_MBUS_FMT_SBGGR10_2X8_PADHI_LE:
+	case V4L2_MBUS_FMT_SBGGR10_2X8_PADLO_BE:
+	case V4L2_MBUS_FMT_SBGGR10_2X8_PADLO_LE:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW10;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	case V4L2_MBUS_FMT_SBGGR10_1X10:
+	case V4L2_MBUS_FMT_SGBRG10_1X10:
+	case V4L2_MBUS_FMT_SGRBG10_1X10:
+	case V4L2_MBUS_FMT_SRGGB10_1X10:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW10;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_10;
+		break;
+	case V4L2_MBUS_FMT_SBGGR12_1X12:
+	case V4L2_MBUS_FMT_SGBRG12_1X12:
+	case V4L2_MBUS_FMT_SGRBG12_1X12:
+	case V4L2_MBUS_FMT_SRGGB12_1X12:
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_BAYER;
+		cfg->mipi_dt = MIPI_DT_RAW12;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_12;
+		break;
+	case V4L2_MBUS_FMT_JPEG_1X8:
+		/* TODO */
+		cfg->data_fmt = CSI_SENS_CONF_DATA_FMT_JPEG;
+		cfg->mipi_dt = MIPI_DT_RAW8;
+		cfg->data_width = IPU_CSI_DATA_WIDTH_8;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/*
+ * Fill a CSI bus config struct from mbus_config and mbus_framefmt.
+ */
+static void fill_csi_bus_cfg(struct ipu_csi_bus_config *csicfg,
+				 struct v4l2_mbus_config *mbus_cfg,
+				 struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	memset(csicfg, 0, sizeof(*csicfg));
+
+	mbus_code_to_bus_cfg(csicfg, mbus_fmt->code);
+
+	switch (mbus_cfg->type) {
+	case V4L2_MBUS_PARALLEL:
+		csicfg->ext_vsync = 1;
+		csicfg->vsync_pol = (mbus_cfg->flags &
+				     V4L2_MBUS_VSYNC_ACTIVE_LOW) ? 1 : 0;
+		csicfg->hsync_pol = (mbus_cfg->flags &
+				     V4L2_MBUS_HSYNC_ACTIVE_LOW) ? 1 : 0;
+		csicfg->pixclk_pol = (mbus_cfg->flags &
+				      V4L2_MBUS_PCLK_SAMPLE_FALLING) ? 1 : 0;
+		csicfg->clk_mode = IPU_CSI_CLK_MODE_GATED_CLK;
+		break;
+	case V4L2_MBUS_BT656:
+		csicfg->ext_vsync = 0;
+		if (V4L2_FIELD_HAS_BOTH(mbus_fmt->field))
+			csicfg->clk_mode = IPU_CSI_CLK_MODE_CCIR656_INTERLACED;
+		else
+			csicfg->clk_mode = IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE;
+		break;
+	case V4L2_MBUS_CSI2:
+		/*
+		 * MIPI CSI-2 requires non gated clock mode, all other
+		 * parameters are not applicable for MIPI CSI-2 bus.
+		 */
+		csicfg->clk_mode = IPU_CSI_CLK_MODE_NONGATED_CLK;
+		break;
+	default:
+		/* will never get here, keep compiler quiet */
+		break;
+	}
+}
+
+int ipu_csi_init_interface(struct ipu_csi *csi,
+			   struct v4l2_mbus_config *mbus_cfg,
+			   struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	struct ipu_csi_bus_config cfg;
+	unsigned long flags;
+	u32 data = 0;
+
+	fill_csi_bus_cfg(&cfg, mbus_cfg, mbus_fmt);
+
+	/* Set the CSI_SENS_CONF register remaining fields */
+	data |= cfg.data_width << CSI_SENS_CONF_DATA_WIDTH_SHIFT |
+		cfg.data_fmt << CSI_SENS_CONF_DATA_FMT_SHIFT |
+		cfg.data_pol << CSI_SENS_CONF_DATA_POL_SHIFT |
+		cfg.vsync_pol << CSI_SENS_CONF_VSYNC_POL_SHIFT |
+		cfg.hsync_pol << CSI_SENS_CONF_HSYNC_POL_SHIFT |
+		cfg.pixclk_pol << CSI_SENS_CONF_PIX_CLK_POL_SHIFT |
+		cfg.ext_vsync << CSI_SENS_CONF_EXT_VSYNC_SHIFT |
+		cfg.clk_mode << CSI_SENS_CONF_SENS_PRTCL_SHIFT |
+		cfg.pack_tight << CSI_SENS_CONF_PACK_TIGHT_SHIFT |
+		cfg.force_eof << CSI_SENS_CONF_FORCE_EOF_SHIFT |
+		cfg.data_en_pol << CSI_SENS_CONF_DATA_EN_POL_SHIFT;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	ipu_csi_write(csi, data, CSI_SENS_CONF);
+
+	/* Setup sensor frame size */
+	ipu_csi_write(csi,
+		      (mbus_fmt->width - 1) | ((mbus_fmt->height - 1) << 16),
+		      CSI_SENS_FRM_SIZE);
+
+	/* Set CCIR registers */
+
+	switch (cfg.clk_mode) {
+	case IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE:
+		ipu_csi_write(csi, 0x40030, CSI_CCIR_CODE_1);
+		ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		break;
+	case IPU_CSI_CLK_MODE_CCIR656_INTERLACED:
+		if (mbus_fmt->width == 720 && mbus_fmt->height == 576) {
+			/*
+			 * PAL case
+			 *
+			 * Field0BlankEnd = 0x6, Field0BlankStart = 0x2,
+			 * Field0ActiveEnd = 0x4, Field0ActiveStart = 0
+			 * Field1BlankEnd = 0x7, Field1BlankStart = 0x3,
+			 * Field1ActiveEnd = 0x5, Field1ActiveStart = 0x1
+			 */
+			ipu_csi_write(csi, 0x40596 | CSI_CCIR_ERR_DET_EN,
+					  CSI_CCIR_CODE_1);
+			ipu_csi_write(csi, 0xD07DF, CSI_CCIR_CODE_2);
+			ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+
+		} else if (mbus_fmt->width == 720 && mbus_fmt->height == 480) {
+			/*
+			 * NTSC case
+			 *
+			 * Field0BlankEnd = 0x7, Field0BlankStart = 0x3,
+			 * Field0ActiveEnd = 0x5, Field0ActiveStart = 0x1
+			 * Field1BlankEnd = 0x6, Field1BlankStart = 0x2,
+			 * Field1ActiveEnd = 0x4, Field1ActiveStart = 0
+			 */
+			ipu_csi_write(csi, 0xD07DF | CSI_CCIR_ERR_DET_EN,
+					  CSI_CCIR_CODE_1);
+			ipu_csi_write(csi, 0x40596, CSI_CCIR_CODE_2);
+			ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		} else {
+			dev_err(csi->ipu->dev,
+				"Unsupported CCIR656 interlaced video mode\n");
+			spin_unlock_irqrestore(&csi->lock, flags);
+			return -EINVAL;
+		}
+		break;
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR:
+		ipu_csi_write(csi, 0x40030 | CSI_CCIR_ERR_DET_EN,
+				   CSI_CCIR_CODE_1);
+		ipu_csi_write(csi, 0xFF0000, CSI_CCIR_CODE_3);
+		break;
+	case IPU_CSI_CLK_MODE_GATED_CLK:
+	case IPU_CSI_CLK_MODE_NONGATED_CLK:
+		ipu_csi_write(csi, 0, CSI_CCIR_CODE_1);
+		break;
+	}
+
+	dev_dbg(csi->ipu->dev, "CSI_SENS_CONF = 0x%08X\n",
+		ipu_csi_read(csi, CSI_SENS_CONF));
+	dev_dbg(csi->ipu->dev, "CSI_ACT_FRM_SIZE = 0x%08X\n",
+		ipu_csi_read(csi, CSI_ACT_FRM_SIZE));
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_init_interface);
+
+bool ipu_csi_is_interlaced(struct ipu_csi *csi)
+{
+	unsigned long flags;
+	u32 sensor_protocol;
+
+	spin_lock_irqsave(&csi->lock, flags);
+	sensor_protocol =
+		(ipu_csi_read(csi, CSI_SENS_CONF) &
+		 CSI_SENS_CONF_SENS_PRTCL_MASK) >>
+		CSI_SENS_CONF_SENS_PRTCL_SHIFT;
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	switch (sensor_protocol) {
+	case IPU_CSI_CLK_MODE_GATED_CLK:
+	case IPU_CSI_CLK_MODE_NONGATED_CLK:
+	case IPU_CSI_CLK_MODE_CCIR656_PROGRESSIVE:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_PROGRESSIVE_SDR:
+		return false;
+	case IPU_CSI_CLK_MODE_CCIR656_INTERLACED:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_DDR:
+	case IPU_CSI_CLK_MODE_CCIR1120_INTERLACED_SDR:
+		return true;
+	default:
+		dev_err(csi->ipu->dev,
+			"CSI %d sensor protocol unsupported\n", csi->id);
+		return false;
+	}
+}
+EXPORT_SYMBOL_GPL(ipu_csi_is_interlaced);
+
+void ipu_csi_get_window(struct ipu_csi *csi, struct v4l2_rect *w)
+{
+	unsigned long flags;
+	u32 reg;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	reg = ipu_csi_read(csi, CSI_ACT_FRM_SIZE);
+	w->width = (reg & 0xFFFF) + 1;
+	w->height = (reg >> 16 & 0xFFFF) + 1;
+
+	reg = ipu_csi_read(csi, CSI_OUT_FRM_CTRL);
+	w->left = (reg & CSI_HSC_MASK) >> CSI_HSC_SHIFT;
+	w->top = (reg & CSI_VSC_MASK) >> CSI_VSC_SHIFT;
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_get_window);
+
+void ipu_csi_set_window(struct ipu_csi *csi, struct v4l2_rect *w)
+{
+	unsigned long flags;
+	u32 reg;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	ipu_csi_write(csi, (w->width - 1) | ((w->height - 1) << 16),
+			  CSI_ACT_FRM_SIZE);
+
+	reg = ipu_csi_read(csi, CSI_OUT_FRM_CTRL);
+	reg &= ~(CSI_HSC_MASK | CSI_VSC_MASK);
+	reg |= ((w->top << CSI_VSC_SHIFT) | (w->left << CSI_HSC_SHIFT));
+	ipu_csi_write(csi, reg, CSI_OUT_FRM_CTRL);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_window);
+
+void ipu_csi_set_test_generator(struct ipu_csi *csi, bool active,
+				u32 r_value, u32 g_value, u32 b_value,
+				u32 pix_clk)
+{
+	unsigned long flags;
+	u32 ipu_clk = clk_get_rate(csi->clk_ipu);
+	u32 temp;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_TST_CTRL);
+
+	if (active == false) {
+		temp &= ~CSI_TEST_GEN_MODE_EN;
+		ipu_csi_write(csi, temp, CSI_TST_CTRL);
+	} else {
+		/* Set sensb_mclk div_ratio */
+		ipu_csi_set_testgen_mclk(csi, pix_clk, ipu_clk);
+
+		temp &= ~(CSI_TEST_GEN_R_MASK | CSI_TEST_GEN_G_MASK |
+			  CSI_TEST_GEN_B_MASK);
+		temp |= CSI_TEST_GEN_MODE_EN;
+		temp |= (r_value << CSI_TEST_GEN_R_SHIFT) |
+			(g_value << CSI_TEST_GEN_G_SHIFT) |
+			(b_value << CSI_TEST_GEN_B_SHIFT);
+		ipu_csi_write(csi, temp, CSI_TST_CTRL);
+	}
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_test_generator);
+
+int ipu_csi_set_mipi_datatype(struct ipu_csi *csi, u32 vc,
+			      struct v4l2_mbus_framefmt *mbus_fmt)
+{
+	struct ipu_csi_bus_config cfg;
+	unsigned long flags;
+	u32 temp;
+
+	if (vc > 3)
+		return -EINVAL;
+
+	mbus_code_to_bus_cfg(&cfg, mbus_fmt->code);
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_MIPI_DI);
+	temp &= ~(0xff << (vc * 8));
+	temp |= (cfg.mipi_dt << (vc * 8));
+	ipu_csi_write(csi, temp, CSI_MIPI_DI);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_mipi_datatype);
+
+int ipu_csi_set_skip_smfc(struct ipu_csi *csi, u32 skip,
+			  u32 max_ratio, u32 id)
+{
+	unsigned long flags;
+	u32 temp;
+
+	if (max_ratio > 5 || id > 3)
+		return -EINVAL;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	temp = ipu_csi_read(csi, CSI_SKIP);
+	temp &= ~(CSI_MAX_RATIO_SKIP_SMFC_MASK | CSI_ID_2_SKIP_MASK |
+		  CSI_SKIP_SMFC_MASK);
+	temp |= (max_ratio << CSI_MAX_RATIO_SKIP_SMFC_SHIFT) |
+		(id << CSI_ID_2_SKIP_SHIFT) |
+		(skip << CSI_SKIP_SMFC_SHIFT);
+	ipu_csi_write(csi, temp, CSI_SKIP);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_skip_smfc);
+
+int ipu_csi_set_dest(struct ipu_csi *csi, enum ipu_csi_dest csi_dest)
+{
+	unsigned long flags;
+	u32 csi_sens_conf, dest;
+
+	if (csi_dest == IPU_CSI_DEST_IDMAC)
+		dest = CSI_DATA_DEST_IDMAC;
+	else
+		dest = CSI_DATA_DEST_IC; /* IC or VDIC */
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	csi_sens_conf = ipu_csi_read(csi, CSI_SENS_CONF);
+	csi_sens_conf &= ~CSI_SENS_CONF_DATA_DEST_MASK;
+	csi_sens_conf |= (dest << CSI_SENS_CONF_DATA_DEST_SHIFT);
+	ipu_csi_write(csi, csi_sens_conf, CSI_SENS_CONF);
+
+	spin_unlock_irqrestore(&csi->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_set_dest);
+
+int ipu_csi_enable(struct ipu_csi *csi)
+{
+	ipu_module_enable(csi->ipu, csi->module);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_enable);
+
+int ipu_csi_disable(struct ipu_csi *csi)
+{
+	ipu_module_disable(csi->ipu, csi->module);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_disable);
+
+struct ipu_csi *ipu_csi_get(struct ipu_soc *ipu, int id)
+{
+	unsigned long flags;
+	struct ipu_csi *csi, *ret;
+
+	if (id > 1)
+		return ERR_PTR(-EINVAL);
+
+	csi = ipu->csi_priv[id];
+	ret = csi;
+
+	spin_lock_irqsave(&csi->lock, flags);
+
+	if (csi->inuse) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	csi->inuse = true;
+unlock:
+	spin_unlock_irqrestore(&csi->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_csi_get);
+
+void ipu_csi_put(struct ipu_csi *csi)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&csi->lock, flags);
+	csi->inuse = false;
+	spin_unlock_irqrestore(&csi->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_csi_put);
+
+int ipu_csi_init(struct ipu_soc *ipu, struct device *dev, int id,
+		 unsigned long base, u32 module, struct clk *clk_ipu)
+{
+	struct ipu_csi *csi;
+
+	if (id > 1)
+		return -ENODEV;
+
+	csi = devm_kzalloc(dev, sizeof(*csi), GFP_KERNEL);
+	if (!csi)
+		return -ENOMEM;
+
+	ipu->csi_priv[id] = csi;
+
+	spin_lock_init(&csi->lock);
+	csi->module = module;
+	csi->id = id;
+	csi->clk_ipu = clk_ipu;
+	csi->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!csi->base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "CSI%d base: 0x%08lx remapped to %p\n",
+		id, base, csi->base);
+	csi->ipu = ipu;
+
+	return 0;
+}
+
+void ipu_csi_exit(struct ipu_soc *ipu, int id)
+{
+}
+
+void ipu_csi_dump(struct ipu_csi *csi)
+{
+	dev_dbg(csi->ipu->dev, "CSI_SENS_CONF:     %08x\n",
+		ipu_csi_read(csi, CSI_SENS_CONF));
+	dev_dbg(csi->ipu->dev, "CSI_SENS_FRM_SIZE: %08x\n",
+		ipu_csi_read(csi, CSI_SENS_FRM_SIZE));
+	dev_dbg(csi->ipu->dev, "CSI_ACT_FRM_SIZE:  %08x\n",
+		ipu_csi_read(csi, CSI_ACT_FRM_SIZE));
+	dev_dbg(csi->ipu->dev, "CSI_OUT_FRM_CTRL:  %08x\n",
+		ipu_csi_read(csi, CSI_OUT_FRM_CTRL));
+	dev_dbg(csi->ipu->dev, "CSI_TST_CTRL:      %08x\n",
+		ipu_csi_read(csi, CSI_TST_CTRL));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_1:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_1));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_2:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_2));
+	dev_dbg(csi->ipu->dev, "CSI_CCIR_CODE_3:   %08x\n",
+		ipu_csi_read(csi, CSI_CCIR_CODE_3));
+	dev_dbg(csi->ipu->dev, "CSI_MIPI_DI:       %08x\n",
+		ipu_csi_read(csi, CSI_MIPI_DI));
+	dev_dbg(csi->ipu->dev, "CSI_SKIP:          %08x\n",
+		ipu_csi_read(csi, CSI_SKIP));
+}
+EXPORT_SYMBOL_GPL(ipu_csi_dump);
diff --git a/drivers/gpu/ipu-v3/ipu-ic.c b/drivers/gpu/ipu-v3/ipu-ic.c
new file mode 100644
index 000000000000..ad75588e1629
--- /dev/null
+++ b/drivers/gpu/ipu-v3/ipu-ic.c
@@ -0,0 +1,778 @@
+/*
+ * Copyright (C) 2012-2014 Mentor Graphics Inc.
+ * Copyright 2005-2012 Freescale Semiconductor, Inc. All Rights Reserved.
+ *
+ * The code contained herein is licensed under the GNU General Public
+ * License. You may obtain a copy of the GNU General Public License
+ * Version 2 or later at the following locations:
+ *
+ * http://www.opensource.org/licenses/gpl-license.html
+ * http://www.gnu.org/copyleft/gpl.html
+ */
+
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/spinlock.h>
+#include <linux/bitrev.h>
+#include <linux/io.h>
+#include <linux/err.h>
+#include "ipu-prv.h"
+
+/* IC Register Offsets */
+#define IC_CONF                 0x0000
+#define IC_PRP_ENC_RSC          0x0004
+#define IC_PRP_VF_RSC           0x0008
+#define IC_PP_RSC               0x000C
+#define IC_CMBP_1               0x0010
+#define IC_CMBP_2               0x0014
+#define IC_IDMAC_1              0x0018
+#define IC_IDMAC_2              0x001C
+#define IC_IDMAC_3              0x0020
+#define IC_IDMAC_4              0x0024
+
+/* IC Register Fields */
+#define IC_CONF_PRPENC_EN       (1 << 0)
+#define IC_CONF_PRPENC_CSC1     (1 << 1)
+#define IC_CONF_PRPENC_ROT_EN   (1 << 2)
+#define IC_CONF_PRPVF_EN        (1 << 8)
+#define IC_CONF_PRPVF_CSC1      (1 << 9)
+#define IC_CONF_PRPVF_CSC2      (1 << 10)
+#define IC_CONF_PRPVF_CMB       (1 << 11)
+#define IC_CONF_PRPVF_ROT_EN    (1 << 12)
+#define IC_CONF_PP_EN           (1 << 16)
+#define IC_CONF_PP_CSC1         (1 << 17)
+#define IC_CONF_PP_CSC2         (1 << 18)
+#define IC_CONF_PP_CMB          (1 << 19)
+#define IC_CONF_PP_ROT_EN       (1 << 20)
+#define IC_CONF_IC_GLB_LOC_A    (1 << 28)
+#define IC_CONF_KEY_COLOR_EN    (1 << 29)
+#define IC_CONF_RWS_EN          (1 << 30)
+#define IC_CONF_CSI_MEM_WR_EN   (1 << 31)
+
+#define IC_IDMAC_1_CB0_BURST_16         (1 << 0)
+#define IC_IDMAC_1_CB1_BURST_16         (1 << 1)
+#define IC_IDMAC_1_CB2_BURST_16         (1 << 2)
+#define IC_IDMAC_1_CB3_BURST_16         (1 << 3)
+#define IC_IDMAC_1_CB4_BURST_16         (1 << 4)
+#define IC_IDMAC_1_CB5_BURST_16         (1 << 5)
+#define IC_IDMAC_1_CB6_BURST_16         (1 << 6)
+#define IC_IDMAC_1_CB7_BURST_16         (1 << 7)
+#define IC_IDMAC_1_PRPENC_ROT_MASK      (0x7 << 11)
+#define IC_IDMAC_1_PRPENC_ROT_OFFSET    11
+#define IC_IDMAC_1_PRPVF_ROT_MASK       (0x7 << 14)
+#define IC_IDMAC_1_PRPVF_ROT_OFFSET     14
+#define IC_IDMAC_1_PP_ROT_MASK          (0x7 << 17)
+#define IC_IDMAC_1_PP_ROT_OFFSET        17
+#define IC_IDMAC_1_PP_FLIP_RS           (1 << 22)
+#define IC_IDMAC_1_PRPVF_FLIP_RS        (1 << 21)
+#define IC_IDMAC_1_PRPENC_FLIP_RS       (1 << 20)
+
+#define IC_IDMAC_2_PRPENC_HEIGHT_MASK   (0x3ff << 0)
+#define IC_IDMAC_2_PRPENC_HEIGHT_OFFSET 0
+#define IC_IDMAC_2_PRPVF_HEIGHT_MASK    (0x3ff << 10)
+#define IC_IDMAC_2_PRPVF_HEIGHT_OFFSET  10
+#define IC_IDMAC_2_PP_HEIGHT_MASK       (0x3ff << 20)
+#define IC_IDMAC_2_PP_HEIGHT_OFFSET     20
+
+#define IC_IDMAC_3_PRPENC_WIDTH_MASK    (0x3ff << 0)
+#define IC_IDMAC_3_PRPENC_WIDTH_OFFSET  0
+#define IC_IDMAC_3_PRPVF_WIDTH_MASK     (0x3ff << 10)
+#define IC_IDMAC_3_PRPVF_WIDTH_OFFSET   10
+#define IC_IDMAC_3_PP_WIDTH_MASK        (0x3ff << 20)
+#define IC_IDMAC_3_PP_WIDTH_OFFSET      20
+
+struct ic_task_regoffs {
+	u32 rsc;
+	u32 tpmem_csc[2];
+};
+
+struct ic_task_bitfields {
+	u32 ic_conf_en;
+	u32 ic_conf_rot_en;
+	u32 ic_conf_cmb_en;
+	u32 ic_conf_csc1_en;
+	u32 ic_conf_csc2_en;
+	u32 ic_cmb_galpha_bit;
+};
+
+static const struct ic_task_regoffs ic_task_reg[IC_NUM_TASKS] = {
+	[IC_TASK_ENCODER] = {
+		.rsc = IC_PRP_ENC_RSC,
+		.tpmem_csc = {0x2008, 0},
+	},
+	[IC_TASK_VIEWFINDER] = {
+		.rsc = IC_PRP_VF_RSC,
+		.tpmem_csc = {0x4028, 0x4040},
+	},
+	[IC_TASK_POST_PROCESSOR] = {
+		.rsc = IC_PP_RSC,
+		.tpmem_csc = {0x6060, 0x6078},
+	},
+};
+
+static const struct ic_task_bitfields ic_task_bit[IC_NUM_TASKS] = {
+	[IC_TASK_ENCODER] = {
+		.ic_conf_en = IC_CONF_PRPENC_EN,
+		.ic_conf_rot_en = IC_CONF_PRPENC_ROT_EN,
+		.ic_conf_cmb_en = 0,    /* NA */
+		.ic_conf_csc1_en = IC_CONF_PRPENC_CSC1,
+		.ic_conf_csc2_en = 0,   /* NA */
+		.ic_cmb_galpha_bit = 0, /* NA */
+	},
+	[IC_TASK_VIEWFINDER] = {
+		.ic_conf_en = IC_CONF_PRPVF_EN,
+		.ic_conf_rot_en = IC_CONF_PRPVF_ROT_EN,
+		.ic_conf_cmb_en = IC_CONF_PRPVF_CMB,
+		.ic_conf_csc1_en = IC_CONF_PRPVF_CSC1,
+		.ic_conf_csc2_en = IC_CONF_PRPVF_CSC2,
+		.ic_cmb_galpha_bit = 0,
+	},
+	[IC_TASK_POST_PROCESSOR] = {
+		.ic_conf_en = IC_CONF_PP_EN,
+		.ic_conf_rot_en = IC_CONF_PP_ROT_EN,
+		.ic_conf_cmb_en = IC_CONF_PP_CMB,
+		.ic_conf_csc1_en = IC_CONF_PP_CSC1,
+		.ic_conf_csc2_en = IC_CONF_PP_CSC2,
+		.ic_cmb_galpha_bit = 8,
+	},
+};
+
+struct ipu_ic_priv;
+
+struct ipu_ic {
+	enum ipu_ic_task task;
+	const struct ic_task_regoffs *reg;
+	const struct ic_task_bitfields *bit;
+
+	enum ipu_color_space in_cs, g_in_cs;
+	enum ipu_color_space out_cs;
+	bool graphics;
+	bool rotation;
+	bool in_use;
+
+	struct ipu_ic_priv *priv;
+};
+
+struct ipu_ic_priv {
+	void __iomem *base;
+	void __iomem *tpmem_base;
+	spinlock_t lock;
+	struct ipu_soc *ipu;
+	int use_count;
+	struct ipu_ic task[IC_NUM_TASKS];
+};
+
+static inline u32 ipu_ic_read(struct ipu_ic *ic, unsigned offset)
+{
+	return readl(ic->priv->base + offset);
+}
+
+static inline void ipu_ic_write(struct ipu_ic *ic, u32 value, unsigned offset)
+{
+	writel(value, ic->priv->base + offset);
+}
+
+struct ic_csc_params {
+	s16 coeff[3][3];	/* signed 9-bit integer coefficients */
+	s16 offset[3];		/* signed 11+2-bit fixed point offset */
+	u8 scale:2;		/* scale coefficients * 2^(scale-1) */
+	bool sat:1;		/* saturate to (16, 235(Y) / 240(U, V)) */
+};
+
+/*
+ * Y = R *  .299 + G *  .587 + B *  .114;
+ * U = R * -.169 + G * -.332 + B *  .500 + 128.;
+ * V = R *  .500 + G * -.419 + B * -.0813 + 128.;
+ */
+static const struct ic_csc_params ic_csc_rgb2ycbcr = {
+	.coeff = {
+		{ 77, 150, 29 },
+		{ 469, 427, 128 },
+		{ 128, 405, 491 },
+	},
+	.offset = { 0, 512, 512 },
+	.scale = 1,
+};
+
+/* transparent RGB->RGB matrix for graphics combining */
+static const struct ic_csc_params ic_csc_rgb2rgb = {
+	.coeff = {
+		{ 128, 0, 0 },
+		{ 0, 128, 0 },
+		{ 0, 0, 128 },
+	},
+	.scale = 2,
+};
+
+/*
+ * R = (1.164 * (Y - 16)) + (1.596 * (Cr - 128));
+ * G = (1.164 * (Y - 16)) - (0.392 * (Cb - 128)) - (0.813 * (Cr - 128));
+ * B = (1.164 * (Y - 16)) + (2.017 * (Cb - 128);
+ */
+static const struct ic_csc_params ic_csc_ycbcr2rgb = {
+	.coeff = {
+		{ 149, 0, 204 },
+		{ 149, 462, 408 },
+		{ 149, 255, 0 },
+	},
+	.offset = { -446, 266, -554 },
+	.scale = 2,
+};
+
+static int init_csc(struct ipu_ic *ic,
+		    enum ipu_color_space inf,
+		    enum ipu_color_space outf,
+		    int csc_index)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	const struct ic_csc_params *params;
+	u32 __iomem *base;
+	const u16 (*c)[3];
+	const u16 *a;
+	u32 param;
+
+	base = (u32 __iomem *)
+		(priv->tpmem_base + ic->reg->tpmem_csc[csc_index]);
+
+	if (inf == IPUV3_COLORSPACE_YUV && outf == IPUV3_COLORSPACE_RGB)
+		params = &ic_csc_ycbcr2rgb;
+	else if (inf == IPUV3_COLORSPACE_RGB && outf == IPUV3_COLORSPACE_YUV)
+		params = &ic_csc_rgb2ycbcr;
+	else if (inf == IPUV3_COLORSPACE_RGB && outf == IPUV3_COLORSPACE_RGB)
+		params = &ic_csc_rgb2rgb;
+	else {
+		dev_err(priv->ipu->dev, "Unsupported color space conversion\n");
+		return -EINVAL;
+	}
+
+	/* Cast to unsigned */
+	c = (const u16 (*)[3])params->coeff;
+	a = (const u16 *)params->offset;
+
+	param = ((a[0] & 0x1f) << 27) | ((c[0][0] & 0x1ff) << 18) |
+		((c[1][1] & 0x1ff) << 9) | (c[2][2] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[0] & 0x1fe0) >> 5) | (params->scale << 8) |
+		(params->sat << 9);
+	writel(param, base++);
+
+	param = ((a[1] & 0x1f) << 27) | ((c[0][1] & 0x1ff) << 18) |
+		((c[1][0] & 0x1ff) << 9) | (c[2][0] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[1] & 0x1fe0) >> 5);
+	writel(param, base++);
+
+	param = ((a[2] & 0x1f) << 27) | ((c[0][2] & 0x1ff) << 18) |
+		((c[1][2] & 0x1ff) << 9) | (c[2][1] & 0x1ff);
+	writel(param, base++);
+
+	param = ((a[2] & 0x1fe0) >> 5);
+	writel(param, base++);
+
+	return 0;
+}
+
+static int calc_resize_coeffs(struct ipu_ic *ic,
+			      u32 in_size, u32 out_size,
+			      u32 *resize_coeff,
+			      u32 *downsize_coeff)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+	u32 temp_size, temp_downsize;
+
+	/*
+	 * Input size cannot be more than 4096, and output size cannot
+	 * be more than 1024
+	 */
+	if (in_size > 4096) {
+		dev_err(ipu->dev, "Unsupported resize (in_size > 4096)\n");
+		return -EINVAL;
+	}
+	if (out_size > 1024) {
+		dev_err(ipu->dev, "Unsupported resize (out_size > 1024)\n");
+		return -EINVAL;
+	}
+
+	/* Cannot downsize more than 8:1 */
+	if ((out_size << 3) < in_size) {
+		dev_err(ipu->dev, "Unsupported downsize\n");
+		return -EINVAL;
+	}
+
+	/* Compute downsizing coefficient */
+	temp_downsize = 0;
+	temp_size = in_size;
+	while (((temp_size > 1024) || (temp_size >= out_size * 2)) &&
+	       (temp_downsize < 2)) {
+		temp_size >>= 1;
+		temp_downsize++;
+	}
+	*downsize_coeff = temp_downsize;
+
+	/*
+	 * compute resizing coefficient using the following equation:
+	 * resize_coeff = M * (SI - 1) / (SO - 1)
+	 * where M = 2^13, SI = input size, SO = output size
+	 */
+	*resize_coeff = (8192L * (temp_size - 1)) / (out_size - 1);
+	if (*resize_coeff >= 16384L) {
+		dev_err(ipu->dev, "Warning! Overflow on resize coeff.\n");
+		*resize_coeff = 0x3FFF;
+	}
+
+	return 0;
+}
+
+void ipu_ic_task_enable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 ic_conf;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	ic_conf |= ic->bit->ic_conf_en;
+
+	if (ic->rotation)
+		ic_conf |= ic->bit->ic_conf_rot_en;
+
+	if (ic->in_cs != ic->out_cs)
+		ic_conf |= ic->bit->ic_conf_csc1_en;
+
+	if (ic->graphics) {
+		ic_conf |= ic->bit->ic_conf_cmb_en;
+		ic_conf |= ic->bit->ic_conf_csc1_en;
+
+		if (ic->g_in_cs != ic->out_cs)
+			ic_conf |= ic->bit->ic_conf_csc2_en;
+	}
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_enable);
+
+void ipu_ic_task_disable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 ic_conf;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	ic_conf &= ~(ic->bit->ic_conf_en |
+		     ic->bit->ic_conf_csc1_en |
+		     ic->bit->ic_conf_rot_en);
+	if (ic->bit->ic_conf_csc2_en)
+		ic_conf &= ~ic->bit->ic_conf_csc2_en;
+	if (ic->bit->ic_conf_cmb_en)
+		ic_conf &= ~ic->bit->ic_conf_cmb_en;
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	ic->rotation = ic->graphics = false;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_disable);
+
+int ipu_ic_task_graphics_init(struct ipu_ic *ic,
+			      enum ipu_color_space in_g_cs,
+			      bool galpha_en, u32 galpha,
+			      bool colorkey_en, u32 colorkey)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 reg, ic_conf;
+	int ret = 0;
+
+	if (ic->task == IC_TASK_ENCODER)
+		return -EINVAL;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_conf = ipu_ic_read(ic, IC_CONF);
+
+	if (!(ic_conf & ic->bit->ic_conf_csc1_en)) {
+		/* need transparent CSC1 conversion */
+		ret = init_csc(ic, IPUV3_COLORSPACE_RGB,
+			       IPUV3_COLORSPACE_RGB, 0);
+		if (ret)
+			goto unlock;
+	}
+
+	ic->g_in_cs = in_g_cs;
+
+	if (ic->g_in_cs != ic->out_cs) {
+		ret = init_csc(ic, ic->g_in_cs, ic->out_cs, 1);
+		if (ret)
+			goto unlock;
+	}
+
+	if (galpha_en) {
+		ic_conf |= IC_CONF_IC_GLB_LOC_A;
+		reg = ipu_ic_read(ic, IC_CMBP_1);
+		reg &= ~(0xff << ic->bit->ic_cmb_galpha_bit);
+		reg |= (galpha << ic->bit->ic_cmb_galpha_bit);
+		ipu_ic_write(ic, reg, IC_CMBP_1);
+	} else
+		ic_conf &= ~IC_CONF_IC_GLB_LOC_A;
+
+	if (colorkey_en) {
+		ic_conf |= IC_CONF_KEY_COLOR_EN;
+		ipu_ic_write(ic, colorkey, IC_CMBP_2);
+	} else
+		ic_conf &= ~IC_CONF_KEY_COLOR_EN;
+
+	ipu_ic_write(ic, ic_conf, IC_CONF);
+
+	ic->graphics = true;
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_graphics_init);
+
+int ipu_ic_task_init(struct ipu_ic *ic,
+		     int in_width, int in_height,
+		     int out_width, int out_height,
+		     enum ipu_color_space in_cs,
+		     enum ipu_color_space out_cs)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	u32 reg, downsize_coeff, resize_coeff;
+	unsigned long flags;
+	int ret = 0;
+
+	/* Setup vertical resizing */
+	ret = calc_resize_coeffs(ic, in_height, out_height,
+				 &resize_coeff, &downsize_coeff);
+	if (ret)
+		return ret;
+
+	reg = (downsize_coeff << 30) | (resize_coeff << 16);
+
+	/* Setup horizontal resizing */
+	ret = calc_resize_coeffs(ic, in_width, out_width,
+				 &resize_coeff, &downsize_coeff);
+	if (ret)
+		return ret;
+
+	reg |= (downsize_coeff << 14) | resize_coeff;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ipu_ic_write(ic, reg, ic->reg->rsc);
+
+	/* Setup color space conversion */
+	ic->in_cs = in_cs;
+	ic->out_cs = out_cs;
+
+	if (ic->in_cs != ic->out_cs) {
+		ret = init_csc(ic, ic->in_cs, ic->out_cs, 0);
+		if (ret)
+			goto unlock;
+	}
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_init);
+
+int ipu_ic_task_idma_init(struct ipu_ic *ic, struct ipuv3_channel *channel,
+			  u32 width, u32 height, int burst_size,
+			  enum ipu_rotate_mode rot)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+	u32 ic_idmac_1, ic_idmac_2, ic_idmac_3;
+	u32 temp_rot = bitrev8(rot) >> 5;
+	bool need_hor_flip = false;
+	unsigned long flags;
+	int ret = 0;
+
+	if ((burst_size != 8) && (burst_size != 16)) {
+		dev_err(ipu->dev, "Illegal burst length for IC\n");
+		return -EINVAL;
+	}
+
+	width--;
+	height--;
+
+	if (temp_rot & 0x2)	/* Need horizontal flip */
+		need_hor_flip = true;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	ic_idmac_1 = ipu_ic_read(ic, IC_IDMAC_1);
+	ic_idmac_2 = ipu_ic_read(ic, IC_IDMAC_2);
+	ic_idmac_3 = ipu_ic_read(ic, IC_IDMAC_3);
+
+	switch (channel->num) {
+	case IPUV3_CHANNEL_IC_PP_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB2_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB2_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PP_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PP_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PP_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PP_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PP_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PP_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_IC_PP:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB5_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB5_BURST_16;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_PP:
+		ic_idmac_1 &= ~IC_IDMAC_1_PP_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PP_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_IC_PRP_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB6_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB6_BURST_16;
+		break;
+	case IPUV3_CHANNEL_IC_PRP_ENC_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB0_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB0_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PRPENC_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PRPENC_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PRPENC_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PRPENC_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PRPENC_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PRPENC_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_ENC:
+		ic_idmac_1 &= ~IC_IDMAC_1_PRPENC_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PRPENC_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_IC_PRP_VF_MEM:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB1_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB1_BURST_16;
+
+		if (need_hor_flip)
+			ic_idmac_1 |= IC_IDMAC_1_PRPVF_FLIP_RS;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_PRPVF_FLIP_RS;
+
+		ic_idmac_2 &= ~IC_IDMAC_2_PRPVF_HEIGHT_MASK;
+		ic_idmac_2 |= height << IC_IDMAC_2_PRPVF_HEIGHT_OFFSET;
+
+		ic_idmac_3 &= ~IC_IDMAC_3_PRPVF_WIDTH_MASK;
+		ic_idmac_3 |= width << IC_IDMAC_3_PRPVF_WIDTH_OFFSET;
+		break;
+	case IPUV3_CHANNEL_MEM_ROT_VF:
+		ic_idmac_1 &= ~IC_IDMAC_1_PRPVF_ROT_MASK;
+		ic_idmac_1 |= temp_rot << IC_IDMAC_1_PRPVF_ROT_OFFSET;
+		break;
+	case IPUV3_CHANNEL_G_MEM_IC_PRP_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB3_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB3_BURST_16;
+		break;
+	case IPUV3_CHANNEL_G_MEM_IC_PP:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB4_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB4_BURST_16;
+		break;
+	case IPUV3_CHANNEL_VDI_MEM_IC_VF:
+		if (burst_size == 16)
+			ic_idmac_1 |= IC_IDMAC_1_CB7_BURST_16;
+		else
+			ic_idmac_1 &= ~IC_IDMAC_1_CB7_BURST_16;
+		break;
+	default:
+		goto unlock;
+	}
+
+	ipu_ic_write(ic, ic_idmac_1, IC_IDMAC_1);
+	ipu_ic_write(ic, ic_idmac_2, IC_IDMAC_2);
+	ipu_ic_write(ic, ic_idmac_3, IC_IDMAC_3);
+
+	if (rot >= IPU_ROTATE_90_RIGHT)
+		ic->rotation = true;
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_task_idma_init);
+
+int ipu_ic_enable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 module = IPU_CONF_IC_EN;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (ic->rotation)
+		module |= IPU_CONF_ROT_EN;
+
+	if (!priv->use_count)
+		ipu_module_enable(priv->ipu, module);
+
+	priv->use_count++;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_enable);
+
+int ipu_ic_disable(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+	u32 module = IPU_CONF_IC_EN | IPU_CONF_ROT_EN;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	priv->use_count--;
+
+	if (!priv->use_count)
+		ipu_module_disable(priv->ipu, module);
+
+	if (priv->use_count < 0)
+		priv->use_count = 0;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_disable);
+
+struct ipu_ic *ipu_ic_get(struct ipu_soc *ipu, enum ipu_ic_task task)
+{
+	struct ipu_ic_priv *priv = ipu->ic_priv;
+	unsigned long flags;
+	struct ipu_ic *ic, *ret;
+
+	if (task >= IC_NUM_TASKS)
+		return ERR_PTR(-EINVAL);
+
+	ic = &priv->task[task];
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (ic->in_use) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	ic->in_use = true;
+	ret = ic;
+
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_ic_get);
+
+void ipu_ic_put(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	ic->in_use = false;
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_ic_put);
+
+int ipu_ic_init(struct ipu_soc *ipu, struct device *dev,
+		unsigned long base, unsigned long tpmem_base)
+{
+	struct ipu_ic_priv *priv;
+	int i;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	ipu->ic_priv = priv;
+
+	spin_lock_init(&priv->lock);
+	priv->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!priv->base)
+		return -ENOMEM;
+	priv->tpmem_base = devm_ioremap(dev, tpmem_base, SZ_64K);
+	if (!priv->tpmem_base)
+		return -ENOMEM;
+
+	dev_dbg(dev, "IC base: 0x%08lx remapped to %p\n", base, priv->base);
+
+	priv->ipu = ipu;
+
+	for (i = 0; i < IC_NUM_TASKS; i++) {
+		priv->task[i].task = i;
+		priv->task[i].priv = priv;
+		priv->task[i].reg = &ic_task_reg[i];
+		priv->task[i].bit = &ic_task_bit[i];
+	}
+
+	return 0;
+}
+
+void ipu_ic_exit(struct ipu_soc *ipu)
+{
+}
+
+void ipu_ic_dump(struct ipu_ic *ic)
+{
+	struct ipu_ic_priv *priv = ic->priv;
+	struct ipu_soc *ipu = priv->ipu;
+
+	dev_dbg(ipu->dev, "IC_CONF = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CONF));
+	dev_dbg(ipu->dev, "IC_PRP_ENC_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PRP_ENC_RSC));
+	dev_dbg(ipu->dev, "IC_PRP_VF_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PRP_VF_RSC));
+	dev_dbg(ipu->dev, "IC_PP_RSC = \t0x%08X\n",
+		ipu_ic_read(ic, IC_PP_RSC));
+	dev_dbg(ipu->dev, "IC_CMBP_1 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CMBP_1));
+	dev_dbg(ipu->dev, "IC_CMBP_2 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_CMBP_2));
+	dev_dbg(ipu->dev, "IC_IDMAC_1 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_1));
+	dev_dbg(ipu->dev, "IC_IDMAC_2 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_2));
+	dev_dbg(ipu->dev, "IC_IDMAC_3 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_3));
+	dev_dbg(ipu->dev, "IC_IDMAC_4 = \t0x%08X\n",
+		ipu_ic_read(ic, IC_IDMAC_4));
+}
+EXPORT_SYMBOL_GPL(ipu_ic_dump);
diff --git a/drivers/gpu/ipu-v3/ipu-prv.h b/drivers/gpu/ipu-v3/ipu-prv.h
index c93f50ec04f7..bfb1e8a4483f 100644
--- a/drivers/gpu/ipu-v3/ipu-prv.h
+++ b/drivers/gpu/ipu-v3/ipu-prv.h
@@ -24,23 +24,6 @@ struct ipu_soc;
 
 #include <video/imx-ipu-v3.h>
 
-#define IPUV3_CHANNEL_CSI0			 0
-#define IPUV3_CHANNEL_CSI1			 1
-#define IPUV3_CHANNEL_CSI2			 2
-#define IPUV3_CHANNEL_CSI3			 3
-#define IPUV3_CHANNEL_MEM_BG_SYNC		23
-#define IPUV3_CHANNEL_MEM_FG_SYNC		27
-#define IPUV3_CHANNEL_MEM_DC_SYNC		28
-#define IPUV3_CHANNEL_MEM_FG_SYNC_ALPHA		31
-#define IPUV3_CHANNEL_MEM_DC_ASYNC		41
-#define IPUV3_CHANNEL_ROT_ENC_MEM		45
-#define IPUV3_CHANNEL_ROT_VF_MEM		46
-#define IPUV3_CHANNEL_ROT_PP_MEM		47
-#define IPUV3_CHANNEL_ROT_ENC_MEM_OUT		48
-#define IPUV3_CHANNEL_ROT_VF_MEM_OUT		49
-#define IPUV3_CHANNEL_ROT_PP_MEM_OUT		50
-#define IPUV3_CHANNEL_MEM_BG_SYNC_ALPHA		51
-
 #define IPU_MCU_T_DEFAULT	8
 #define IPU_CM_IDMAC_REG_OFS	0x00008000
 #define IPU_CM_IC_REG_OFS	0x00020000
@@ -85,6 +68,7 @@ struct ipu_soc;
 #define IPU_DISP_TASK_STAT		IPU_CM_REG(0x0254)
 #define IPU_CHA_BUF0_RDY(ch)		IPU_CM_REG(0x0268 + 4 * ((ch) / 32))
 #define IPU_CHA_BUF1_RDY(ch)		IPU_CM_REG(0x0270 + 4 * ((ch) / 32))
+#define IPU_CHA_BUF2_RDY(ch)		IPU_CM_REG(0x0288 + 4 * ((ch) / 32))
 #define IPU_ALT_CHA_BUF0_RDY(ch)	IPU_CM_REG(0x0278 + 4 * ((ch) / 32))
 #define IPU_ALT_CHA_BUF1_RDY(ch)	IPU_CM_REG(0x0280 + 4 * ((ch) / 32))
 
@@ -148,9 +132,12 @@ struct ipuv3_channel {
 	struct ipu_soc *ipu;
 };
 
+struct ipu_cpmem;
+struct ipu_csi;
 struct ipu_dc_priv;
 struct ipu_dmfc_priv;
 struct ipu_di;
+struct ipu_ic_priv;
 struct ipu_smfc_priv;
 
 struct ipu_devtype;
@@ -164,7 +151,6 @@ struct ipu_soc {
 
 	void __iomem		*cm_reg;
 	void __iomem		*idmac_reg;
-	struct ipu_ch_param __iomem	*cpmem_base;
 
 	int			usecount;
 
@@ -176,13 +162,27 @@ struct ipu_soc {
 	int			irq_err;
 	struct irq_domain	*domain;
 
+	struct ipu_cpmem	*cpmem_priv;
 	struct ipu_dc_priv	*dc_priv;
 	struct ipu_dp_priv	*dp_priv;
 	struct ipu_dmfc_priv	*dmfc_priv;
 	struct ipu_di		*di_priv[2];
+	struct ipu_csi		*csi_priv[2];
+	struct ipu_ic_priv	*ic_priv;
 	struct ipu_smfc_priv	*smfc_priv;
 };
 
+static inline u32 ipu_idmac_read(struct ipu_soc *ipu, unsigned offset)
+{
+	return readl(ipu->idmac_reg + offset);
+}
+
+static inline void ipu_idmac_write(struct ipu_soc *ipu, u32 value,
+				   unsigned offset)
+{
+	writel(value, ipu->idmac_reg + offset);
+}
+
 void ipu_srm_dp_sync_update(struct ipu_soc *ipu);
 
 int ipu_module_enable(struct ipu_soc *ipu, u32 mask);
@@ -191,6 +191,14 @@ int ipu_module_disable(struct ipu_soc *ipu, u32 mask);
 bool ipu_idmac_channel_busy(struct ipu_soc *ipu, unsigned int chno);
 int ipu_wait_interrupt(struct ipu_soc *ipu, int irq, int ms);
 
+int ipu_csi_init(struct ipu_soc *ipu, struct device *dev, int id,
+		 unsigned long base, u32 module, struct clk *clk_ipu);
+void ipu_csi_exit(struct ipu_soc *ipu, int id);
+
+int ipu_ic_init(struct ipu_soc *ipu, struct device *dev,
+		unsigned long base, unsigned long tpmem_base);
+void ipu_ic_exit(struct ipu_soc *ipu);
+
 int ipu_di_init(struct ipu_soc *ipu, struct device *dev, int id,
 		unsigned long base, u32 module, struct clk *ipu_clk);
 void ipu_di_exit(struct ipu_soc *ipu, int id);
diff --git a/drivers/gpu/ipu-v3/ipu-smfc.c b/drivers/gpu/ipu-v3/ipu-smfc.c
index e4f85ad286fc..4ef910991413 100644
--- a/drivers/gpu/ipu-v3/ipu-smfc.c
+++ b/drivers/gpu/ipu-v3/ipu-smfc.c
@@ -8,7 +8,6 @@
  * http://www.opensource.org/licenses/gpl-license.html
  * http://www.gnu.org/copyleft/gpl.html
  */
-#define DEBUG
 #include <linux/export.h>
 #include <linux/types.h>
 #include <linux/init.h>
@@ -21,9 +20,18 @@
 
 #include "ipu-prv.h"
 
+struct ipu_smfc {
+	struct ipu_smfc_priv *priv;
+	int chno;
+	bool inuse;
+};
+
 struct ipu_smfc_priv {
 	void __iomem *base;
 	spinlock_t lock;
+	struct ipu_soc *ipu;
+	struct ipu_smfc channel[4];
+	int use_count;
 };
 
 /*SMFC Registers */
@@ -31,63 +39,166 @@ struct ipu_smfc_priv {
 #define SMFC_WMC	0x0004
 #define SMFC_BS		0x0008
 
-int ipu_smfc_set_burstsize(struct ipu_soc *ipu, int channel, int burstsize)
+int ipu_smfc_set_burstsize(struct ipu_smfc *smfc, int burstsize)
 {
-	struct ipu_smfc_priv *smfc = ipu->smfc_priv;
+	struct ipu_smfc_priv *priv = smfc->priv;
 	unsigned long flags;
 	u32 val, shift;
 
-	spin_lock_irqsave(&smfc->lock, flags);
+	spin_lock_irqsave(&priv->lock, flags);
 
-	shift = channel * 4;
-	val = readl(smfc->base + SMFC_BS);
+	shift = smfc->chno * 4;
+	val = readl(priv->base + SMFC_BS);
 	val &= ~(0xf << shift);
 	val |= burstsize << shift;
-	writel(val, smfc->base + SMFC_BS);
+	writel(val, priv->base + SMFC_BS);
 
-	spin_unlock_irqrestore(&smfc->lock, flags);
+	spin_unlock_irqrestore(&priv->lock, flags);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ipu_smfc_set_burstsize);
 
-int ipu_smfc_map_channel(struct ipu_soc *ipu, int channel, int csi_id, int mipi_id)
+int ipu_smfc_map_channel(struct ipu_smfc *smfc, int csi_id, int mipi_id)
 {
-	struct ipu_smfc_priv *smfc = ipu->smfc_priv;
+	struct ipu_smfc_priv *priv = smfc->priv;
 	unsigned long flags;
 	u32 val, shift;
 
-	spin_lock_irqsave(&smfc->lock, flags);
+	spin_lock_irqsave(&priv->lock, flags);
 
-	shift = channel * 3;
-	val = readl(smfc->base + SMFC_MAP);
+	shift = smfc->chno * 3;
+	val = readl(priv->base + SMFC_MAP);
 	val &= ~(0x7 << shift);
 	val |= ((csi_id << 2) | mipi_id) << shift;
-	writel(val, smfc->base + SMFC_MAP);
+	writel(val, priv->base + SMFC_MAP);
 
-	spin_unlock_irqrestore(&smfc->lock, flags);
+	spin_unlock_irqrestore(&priv->lock, flags);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(ipu_smfc_map_channel);
 
+int ipu_smfc_set_watermark(struct ipu_smfc *smfc, u32 set_level, u32 clr_level)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+	u32 val, shift;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	shift = smfc->chno * 6 + (smfc->chno > 1 ? 4 : 0);
+	val = readl(priv->base + SMFC_WMC);
+	val &= ~(0x3f << shift);
+	val |= ((clr_level << 3) | set_level) << shift;
+	writel(val, priv->base + SMFC_WMC);
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_set_watermark);
+
+int ipu_smfc_enable(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (!priv->use_count)
+		ipu_module_enable(priv->ipu, IPU_CONF_SMFC_EN);
+
+	priv->use_count++;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_enable);
+
+int ipu_smfc_disable(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	priv->use_count--;
+
+	if (!priv->use_count)
+		ipu_module_disable(priv->ipu, IPU_CONF_SMFC_EN);
+
+	if (priv->use_count < 0)
+		priv->use_count = 0;
+
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_disable);
+
+struct ipu_smfc *ipu_smfc_get(struct ipu_soc *ipu, unsigned int chno)
+{
+	struct ipu_smfc_priv *priv = ipu->smfc_priv;
+	struct ipu_smfc *smfc, *ret;
+	unsigned long flags;
+
+	if (chno >= 4)
+		return ERR_PTR(-EINVAL);
+
+	smfc = &priv->channel[chno];
+	ret = smfc;
+
+	spin_lock_irqsave(&priv->lock, flags);
+
+	if (smfc->inuse) {
+		ret = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	smfc->inuse = true;
+unlock:
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_get);
+
+void ipu_smfc_put(struct ipu_smfc *smfc)
+{
+	struct ipu_smfc_priv *priv = smfc->priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	smfc->inuse = false;
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+EXPORT_SYMBOL_GPL(ipu_smfc_put);
+
 int ipu_smfc_init(struct ipu_soc *ipu, struct device *dev,
 		  unsigned long base)
 {
-	struct ipu_smfc_priv *smfc;
+	struct ipu_smfc_priv *priv;
+	int i;
 
-	smfc = devm_kzalloc(dev, sizeof(*smfc), GFP_KERNEL);
-	if (!smfc)
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
 		return -ENOMEM;
 
-	ipu->smfc_priv = smfc;
-	spin_lock_init(&smfc->lock);
+	ipu->smfc_priv = priv;
+	spin_lock_init(&priv->lock);
+	priv->ipu = ipu;
 
-	smfc->base = devm_ioremap(dev, base, PAGE_SIZE);
-	if (!smfc->base)
+	priv->base = devm_ioremap(dev, base, PAGE_SIZE);
+	if (!priv->base)
 		return -ENOMEM;
 
-	pr_debug("%s: ioremap 0x%08lx -> %p\n", __func__, base, smfc->base);
+	for (i = 0; i < 4; i++) {
+		priv->channel[i].priv = priv;
+		priv->channel[i].chno = i;
+	}
+
+	pr_debug("%s: ioremap 0x%08lx -> %p\n", __func__, base, priv->base);
 
 	return 0;
 }
diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c
index 6866448083b2..37ac7b5dbd06 100644
--- a/drivers/gpu/vga/vga_switcheroo.c
+++ b/drivers/gpu/vga/vga_switcheroo.c
@@ -660,6 +660,12 @@ int vga_switcheroo_init_domain_pm_ops(struct device *dev, struct dev_pm_domain *
 }
 EXPORT_SYMBOL(vga_switcheroo_init_domain_pm_ops);
 
+void vga_switcheroo_fini_domain_pm_ops(struct device *dev)
+{
+	dev->pm_domain = NULL;
+}
+EXPORT_SYMBOL(vga_switcheroo_fini_domain_pm_ops);
+
 static int vga_switcheroo_runtime_resume_hdmi_audio(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/gpu/vga/vgaarb.c
index af0259708358..7bcbf863656e 100644
--- a/drivers/gpu/vga/vgaarb.c
+++ b/drivers/gpu/vga/vgaarb.c
@@ -41,6 +41,7 @@
 #include <linux/poll.h>
 #include <linux/miscdevice.h>
 #include <linux/slab.h>
+#include <linux/screen_info.h>
 
 #include <linux/uaccess.h>
 
@@ -112,10 +113,8 @@ both:
 	return 1;
 }
 
-#ifndef __ARCH_HAS_VGA_DEFAULT_DEVICE
 /* this is only used a cookie - it should not be dereferenced */
 static struct pci_dev *vga_default;
-#endif
 
 static void vga_arb_device_card_gone(struct pci_dev *pdev);
 
@@ -131,7 +130,6 @@ static struct vga_device *vgadev_find(struct pci_dev *pdev)
 }
 
 /* Returns the default VGA device (vgacon's babe) */
-#ifndef __ARCH_HAS_VGA_DEFAULT_DEVICE
 struct pci_dev *vga_default_device(void)
 {
 	return vga_default;
@@ -147,7 +145,6 @@ void vga_set_default_device(struct pci_dev *pdev)
 	pci_dev_put(vga_default);
 	vga_default = pci_dev_get(pdev);
 }
-#endif
 
 static inline void vga_irq_set_state(struct vga_device *vgadev, bool state)
 {
@@ -237,12 +234,10 @@ static struct vga_device *__vga_tryget(struct vga_device *vgadev,
 		if (conflict->locks & lwants)
 			return conflict;
 
-		/* Ok, now check if he owns the resource we want. We don't need
-		 * to check "decodes" since it should be impossible to own
-		 * own legacy resources you don't decode unless I have a bug
-		 * in this code...
+		/* Ok, now check if it owns the resource we want.  We can
+		 * lock resources that are not decoded, therefore a device
+		 * can own resources it doesn't decode.
 		 */
-		WARN_ON(conflict->owns & ~conflict->decodes);
 		match = lwants & conflict->owns;
 		if (!match)
 			continue;
@@ -254,13 +249,19 @@ static struct vga_device *__vga_tryget(struct vga_device *vgadev,
 		flags = 0;
 		pci_bits = 0;
 
+		/* If we can't control legacy resources via the bridge, we
+		 * also need to disable normal decoding.
+		 */
 		if (!conflict->bridge_has_one_vga) {
-			vga_irq_set_state(conflict, false);
-			flags |= PCI_VGA_STATE_CHANGE_DECODES;
-			if (match & (VGA_RSRC_LEGACY_MEM|VGA_RSRC_NORMAL_MEM))
+			if ((match & conflict->decodes) & VGA_RSRC_LEGACY_MEM)
 				pci_bits |= PCI_COMMAND_MEMORY;
-			if (match & (VGA_RSRC_LEGACY_IO|VGA_RSRC_NORMAL_IO))
+			if ((match & conflict->decodes) & VGA_RSRC_LEGACY_IO)
 				pci_bits |= PCI_COMMAND_IO;
+
+			if (pci_bits) {
+				vga_irq_set_state(conflict, false);
+				flags |= PCI_VGA_STATE_CHANGE_DECODES;
+			}
 		}
 
 		if (change_bridge)
@@ -268,18 +269,19 @@ static struct vga_device *__vga_tryget(struct vga_device *vgadev,
 
 		pci_set_vga_state(conflict->pdev, false, pci_bits, flags);
 		conflict->owns &= ~match;
-		/* If he also owned non-legacy, that is no longer the case */
-		if (match & VGA_RSRC_LEGACY_MEM)
+
+		/* If we disabled normal decoding, reflect it in owns */
+		if (pci_bits & PCI_COMMAND_MEMORY)
 			conflict->owns &= ~VGA_RSRC_NORMAL_MEM;
-		if (match & VGA_RSRC_LEGACY_IO)
+		if (pci_bits & PCI_COMMAND_IO)
 			conflict->owns &= ~VGA_RSRC_NORMAL_IO;
 	}
 
 enable_them:
 	/* ok dude, we got it, everybody conflicting has been disabled, let's
-	 * enable us. Make sure we don't mark a bit in "owns" that we don't
-	 * also have in "decodes". We can lock resources we don't decode but
-	 * not own them.
+	 * enable us.  Mark any bits in "owns" regardless of whether we
+	 * decoded them.  We can lock resources we don't decode, therefore
+	 * we must track them via "owns".
 	 */
 	flags = 0;
 	pci_bits = 0;
@@ -291,7 +293,7 @@ enable_them:
 		if (wants & (VGA_RSRC_LEGACY_IO|VGA_RSRC_NORMAL_IO))
 			pci_bits |= PCI_COMMAND_IO;
 	}
-	if (!!(wants & VGA_RSRC_LEGACY_MASK))
+	if (wants & VGA_RSRC_LEGACY_MASK)
 		flags |= PCI_VGA_STATE_CHANGE_BRIDGE;
 
 	pci_set_vga_state(vgadev->pdev, true, pci_bits, flags);
@@ -299,7 +301,7 @@ enable_them:
 	if (!vgadev->bridge_has_one_vga) {
 		vga_irq_set_state(vgadev, true);
 	}
-	vgadev->owns |= (wants & vgadev->decodes);
+	vgadev->owns |= wants;
 lock_them:
 	vgadev->locks |= (rsrc & VGA_RSRC_LEGACY_MASK);
 	if (rsrc & VGA_RSRC_LEGACY_IO)
@@ -398,7 +400,6 @@ int vga_get(struct pci_dev *pdev, unsigned int rsrc, int interruptible)
 		}
 		schedule();
 		remove_wait_queue(&vga_wait_queue, &wait);
-		set_current_state(TASK_RUNNING);
 	}
 	return rc;
 }
@@ -578,11 +579,12 @@ static bool vga_arbiter_add_pci_device(struct pci_dev *pdev)
 	/* Deal with VGA default device. Use first enabled one
 	 * by default if arch doesn't have it's own hook
 	 */
-#ifndef __ARCH_HAS_VGA_DEFAULT_DEVICE
 	if (vga_default == NULL &&
-	    ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK))
+	    ((vgadev->owns & VGA_RSRC_LEGACY_MASK) == VGA_RSRC_LEGACY_MASK)) {
+		pr_info("vgaarb: setting as boot device: PCI:%s\n",
+			pci_name(pdev));
 		vga_set_default_device(pdev);
-#endif
+	}
 
 	vga_arbiter_check_bridge_sharing(vgadev);
 
@@ -616,10 +618,8 @@ static bool vga_arbiter_del_pci_device(struct pci_dev *pdev)
 		goto bail;
 	}
 
-#ifndef __ARCH_HAS_VGA_DEFAULT_DEVICE
 	if (vga_default == pdev)
 		vga_set_default_device(NULL);
-#endif
 
 	if (vgadev->decodes & (VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM))
 		vga_decode_count--;
@@ -649,7 +649,6 @@ static inline void vga_update_device_decodes(struct vga_device *vgadev,
 	old_decodes = vgadev->decodes;
 	decodes_removed = ~new_decodes & old_decodes;
 	decodes_unlocked = vgadev->locks & decodes_removed;
-	vgadev->owns &= ~decodes_removed;
 	vgadev->decodes = new_decodes;
 
 	pr_info("vgaarb: device changed decodes: PCI:%s,olddecodes=%s,decodes=%s:owns=%s\n",
@@ -1316,6 +1315,38 @@ static int __init vga_arb_device_init(void)
 	pr_info("vgaarb: loaded\n");
 
 	list_for_each_entry(vgadev, &vga_list, list) {
+#if defined(CONFIG_X86) || defined(CONFIG_IA64)
+		/* Override I/O based detection done by vga_arbiter_add_pci_device()
+		 * as it may take the wrong device (e.g. on Apple system under EFI).
+		 *
+		 * Select the device owning the boot framebuffer if there is one.
+		 */
+		resource_size_t start, end;
+		int i;
+
+		/* Does firmware framebuffer belong to us? */
+		for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
+			if (!(pci_resource_flags(vgadev->pdev, i) & IORESOURCE_MEM))
+				continue;
+
+			start = pci_resource_start(vgadev->pdev, i);
+			end  = pci_resource_end(vgadev->pdev, i);
+
+			if (!start || !end)
+				continue;
+
+			if (screen_info.lfb_base < start ||
+			    (screen_info.lfb_base + screen_info.lfb_size) >= end)
+				continue;
+			if (!vga_default_device())
+				pr_info("vgaarb: setting as boot device: PCI:%s\n",
+					pci_name(vgadev->pdev));
+			else if (vgadev->pdev != vga_default_device())
+				pr_info("vgaarb: overriding boot device: PCI:%s\n",
+					pci_name(vgadev->pdev));
+			vga_set_default_device(vgadev->pdev);
+		}
+#endif
 		if (vgadev->bridge_has_one_vga)
 			pr_info("vgaarb: bridge control possible %s\n", pci_name(vgadev->pdev));
 		else