Merge tag 'drm-next-2022-12-13' of git://anongit.freedesktop.org/drm/drm

Pull drm updates from Dave Airlie:
 "The biggest highlight is that the accel subsystem framework is merged.
  Hopefully for 6.3 we will be able to line up a driver to use it.

  In drivers land, i915 enables DG2 support by default now, and nouveau
  has a big stability refactoring and initial ampere support, AMD
  includes new hw IP support and should build on ARM again. There is
  also an ofdrm driver to take over offb on platforms it's used.

  Stuff outside my tree, the dma-buf patches hit a few places, the vc4
  firmware changes also do, and i915 has some interactions with MEI for
  discrete GPUs. I think all of those should have been acked/reviewed by
  relevant parties.

  New driver:
   - ofdrm - replacement for offb

  fbdev:
   - add support for nomodeset

  fourcc:
   - add Vivante tiled modifier

  core:
   - atomic-helpers: CRTC primary plane test fixes, fb access hooks
   - connector: TV API consistency, cmdline parser improvements
   - send connector hotplug on cleanup
   - sort makefile objects

  tests:
   - sort kunit tests
   - improve DP-MST tests
   - add kunit helpers to create a device

  sched:
   - module param for scheduling policy
   - refcounting fix

  buddy:
   - add back random seed log

  ttm:
   - convert ttm_resource to size_t
   - optimize pool allocations

  edid:
   - HFVSDB parsing support fixes
   - logging/debug improvements
   - DSC quirks

  dma-buf:
   - Add unlocked vmap and attachment mapping
   - move drivers to common locking convention
   - locking improvements

  firmware:
   - new API for rPI firmware and vc4

  xilinx:
   - zynqmp: displayport bridge support
   - dpsub fix

  bridge:
   - adv7533: Remove dynamic lane switching
   - it6505: Runtime PM support, sync improvements
   - ps8640: Handle AUX defer messages
   - tc358775: Drop soft-reset over I2C

  panel:
   - panel-edp: Add INX N116BGE-EA2 C2 and C4 support.
   - Jadard JD9365DA-H3
   - NewVision NV3051D

  amdgpu:
   - DCN support on ARM
   - DCN 2.1 secure display
   - Sienna Cichlid mode2 reset fixes
   - new GC 11.x firmware versions
   - drop AMD specific DSC workarounds in favour of drm code
   - clang warning fixes
   - scheduler rework
   - SR-IOV fixes
   - GPUVM locking fixes
   - fix memory leak in CS IOCTL error path
   - flexible array updates
   - enable new GC/PSP/SMU/NBIO IP
   - GFX preemption support for gfx9

  amdkfd:
   - cache size fixes
   - userptr fixes
   - enable cooperative launch on gfx 10.3
   - enable GC 11.0.4 KFD support

  radeon:
   - replace kmap with kmap_local_page
   - ACPI ref count fix
   - HDA audio notifier support

  i915:
   - DG2 enabled by default
   - MTL enablement work
   - hotplug refactoring
   - VBT improvements
   - Display and watermark refactoring
   - ADL-P workaround
   - temp disable runtime_pm for discrete-
   - fix for A380 as a secondary GPU
   - Wa_18017747507 for DG2
   - CS timestamp support fixes for gen5 and earlier
   - never purge busy TTM objects
   - use i915_sg_dma_sizes for all backends
   - demote GuC kernel contexts to normal priority
   - gvt: refactor for new MDEV interface
   - enable DC power states on eDP ports
   - fix gen 2/3 workarounds

  nouveau:
   - fix page fault handling
   - Ampere acceleration support
   - driver stability improvements
   - nva3 backlight support

  msm:
   - MSM_INFO_GET_FLAGS support
   - DPU: XR30 and P010 image formats
   - Qualcomm SM6115 support
   - DSI PHY support for QCM2290
   - HDMI: refactored dev init path
   - remove exclusive-fence hack
   - fix speed-bin detection
   - enable clamp to idle on 7c3
   - improved hangcheck detection

  vmwgfx:
   - fb and cursor refactoring
   - convert to generic hashtable
   - cursor improvements

  etnaviv:
   - hw workarounds
   - softpin MMU fixes

  ast:
   - atomic gamma LUT support
   - convert to SHMEM

  lcdif:
   - support YUV planes
   - Increase DMA burst size
   - FIFO threshold tuning

  meson:
   - fix return type of cvbs mode_valid

  mgag200:
   - fix PLL setup on some revisions

  sun4i:
   - A100 and D1 support

  udl:
   - modesetting improvements
   - hot unplug support

  vc4:
   - support PAL-M
   - fix regression preventing 4K @ 60Hz
   - fix NULL ptr deref

  v3d:
   - switch to drm managed resources

  renesas:
   - RZ/G2L DSI support
   - DU Kconfig cleanup

  mediatek:
   - fixup dpi and hdmi
   - MT8188 dpi support
   - MT8195 AFBC support

  tegra:
   - NVDEC hardware on Tegra234 SoC

  hdlcd:
   - switch to drm managed resources

  ingenic:
   - fix registration error path

  hisilicon:
   - convert to drm_mode_init

  maildp:
   - use managed resources

  mtk:
   - use drm_mode_init

  rockchip:
   - use drm_mode_copy"

* tag 'drm-next-2022-12-13' of git://anongit.freedesktop.org/drm/drm: (1397 commits)
  drm/amdgpu: fix mmhub register base coding error
  drm/amdgpu: add tmz support for GC IP v11.0.4
  drm/amdgpu: enable GFX Clock Gating control for GC IP v11.0.4
  drm/amdgpu: enable GFX Power Gating for GC IP v11.0.4
  drm/amdgpu: enable GFX IP v11.0.4 CG support
  drm/amdgpu: Make amdgpu_ring_mux functions as static
  drm/amdgpu: generally allow over-commit during BO allocation
  drm/amd/display: fix array index out of bound error in DCN32 DML
  drm/amd/display: 3.2.215
  drm/amd/display: set optimized required for comp buf changes
  drm/amd/display: Add debug option to skip PSR CRTC disable
  drm/amd/display: correct DML calc error of UrgentLatency
  drm/amd/display: correct static_screen_event_mask
  drm/amd/display: Ensure commit_streams returns the DC return code
  drm/amd/display: read invalid ddc pin status cause engine busy
  drm/amd/display: Bypass DET swath fill check for max clocks
  drm/amd/display: Disable uclk pstate for subvp pipes
  drm/amd/display: Fix DCN2.1 default DSC clocks
  drm/amd/display: Enable dp_hdmi21_pcon support
  drm/amd/display: prevent seamless boot on displays that don't have the preferred dig
  ...

1 files changed, 205 insertions, 267 deletions

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 8f86f56e7ca4..73c88b1c9545 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -30,6 +30,8 @@
 #include "display/skl_watermark.h"
 
 #include "gt/intel_engine_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
 #include "gt/intel_gt_regs.h"
 
 #include "i915_drv.h"
@@ -58,25 +60,20 @@ static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
 		 * Must match Sampler, Pixel Back End, and Media. See
 		 * WaCompressedResourceSamplerPbeMediaNewHashMode.
 		 */
-		intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
-			   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) |
-			   SKL_DE_COMPRESSED_HASH_MODE);
+		intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PAR1_1, 0, SKL_DE_COMPRESSED_HASH_MODE);
 	}
 
 	/* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl,cfl */
-	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
-		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) | SKL_EDP_PSR_FIX_RDWRAP);
+	intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PAR1_1, 0, SKL_EDP_PSR_FIX_RDWRAP);
 
 	/* WaEnableChickenDCPR:skl,bxt,kbl,glk,cfl */
-	intel_uncore_write(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1,
-		   intel_uncore_read(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1) | MASK_WAKEMEM);
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1, 0, MASK_WAKEMEM);
 
 	/*
 	 * WaFbcWakeMemOn:skl,bxt,kbl,glk,cfl
 	 * Display WA #0859: skl,bxt,kbl,glk,cfl
 	 */
-	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
-		   DISP_FBC_MEMORY_WAKE);
+	intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, 0, DISP_FBC_MEMORY_WAKE);
 }
 
 static void bxt_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -84,15 +81,13 @@ static void bxt_init_clock_gating(struct drm_i915_private *dev_priv)
 	gen9_init_clock_gating(dev_priv);
 
 	/* WaDisableSDEUnitClockGating:bxt */
-	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
-		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_UCGCTL6, 0, GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
 	/*
 	 * FIXME:
 	 * GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ applies on 3x6 GT SKUs only.
 	 */
-	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
-		   GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ);
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_UCGCTL6, 0, GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ);
 
 	/*
 	 * Wa: Backlight PWM may stop in the asserted state, causing backlight
@@ -113,16 +108,13 @@ static void bxt_init_clock_gating(struct drm_i915_private *dev_priv)
 	 * WaFbcTurnOffFbcWatermark:bxt
 	 * Display WA #0562: bxt
 	 */
-	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
-		   DISP_FBC_WM_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, 0, DISP_FBC_WM_DIS);
 
 	/*
 	 * WaFbcHighMemBwCorruptionAvoidance:bxt
 	 * Display WA #0883: bxt
 	 */
-	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
-			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
-			   DPFC_DISABLE_DUMMY0);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A), 0, DPFC_DISABLE_DUMMY0);
 }
 
 static void glk_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -895,19 +887,14 @@ static void pnv_update_wm(struct drm_i915_private *dev_priv)
 		wm = intel_calculate_wm(pixel_rate, &pnv_cursor_wm,
 					pnv_display_wm.fifo_size,
 					4, latency->cursor_sr);
-		reg = intel_uncore_read(&dev_priv->uncore, DSPFW3);
-		reg &= ~DSPFW_CURSOR_SR_MASK;
-		reg |= FW_WM(wm, CURSOR_SR);
-		intel_uncore_write(&dev_priv->uncore, DSPFW3, reg);
+		intel_uncore_rmw(&dev_priv->uncore, DSPFW3, DSPFW_CURSOR_SR_MASK,
+				 FW_WM(wm, CURSOR_SR));
 
 		/* Display HPLL off SR */
 		wm = intel_calculate_wm(pixel_rate, &pnv_display_hplloff_wm,
 					pnv_display_hplloff_wm.fifo_size,
 					cpp, latency->display_hpll_disable);
-		reg = intel_uncore_read(&dev_priv->uncore, DSPFW3);
-		reg &= ~DSPFW_HPLL_SR_MASK;
-		reg |= FW_WM(wm, HPLL_SR);
-		intel_uncore_write(&dev_priv->uncore, DSPFW3, reg);
+		intel_uncore_rmw(&dev_priv->uncore, DSPFW3, DSPFW_HPLL_SR_MASK, FW_WM(wm, HPLL_SR));
 
 		/* cursor HPLL off SR */
 		wm = intel_calculate_wm(pixel_rate, &pnv_cursor_hplloff_wm,
@@ -1337,34 +1324,14 @@ static bool g4x_compute_fbc_en(const struct g4x_wm_state *wm_state,
 	return true;
 }
 
-static int g4x_compute_pipe_wm(struct intel_atomic_state *state,
-			       struct intel_crtc *crtc)
+static int _g4x_compute_pipe_wm(struct intel_crtc_state *crtc_state)
 {
-	struct intel_crtc_state *crtc_state =
-		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
 	struct g4x_wm_state *wm_state = &crtc_state->wm.g4x.optimal;
 	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
 	const struct g4x_pipe_wm *raw;
-	const struct intel_plane_state *old_plane_state;
-	const struct intel_plane_state *new_plane_state;
-	struct intel_plane *plane;
 	enum plane_id plane_id;
-	int i, level;
-	unsigned int dirty = 0;
-
-	for_each_oldnew_intel_plane_in_state(state, plane,
-					     old_plane_state,
-					     new_plane_state, i) {
-		if (new_plane_state->hw.crtc != &crtc->base &&
-		    old_plane_state->hw.crtc != &crtc->base)
-			continue;
-
-		if (g4x_raw_plane_wm_compute(crtc_state, new_plane_state))
-			dirty |= BIT(plane->id);
-	}
-
-	if (!dirty)
-		return 0;
+	int level;
 
 	level = G4X_WM_LEVEL_NORMAL;
 	if (!g4x_raw_crtc_wm_is_valid(crtc_state, level))
@@ -1417,6 +1384,34 @@ static int g4x_compute_pipe_wm(struct intel_atomic_state *state,
 	return 0;
 }
 
+static int g4x_compute_pipe_wm(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *old_plane_state;
+	const struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	unsigned int dirty = 0;
+	int i;
+
+	for_each_oldnew_intel_plane_in_state(state, plane,
+					     old_plane_state,
+					     new_plane_state, i) {
+		if (new_plane_state->hw.crtc != &crtc->base &&
+		    old_plane_state->hw.crtc != &crtc->base)
+			continue;
+
+		if (g4x_raw_plane_wm_compute(crtc_state, new_plane_state))
+			dirty |= BIT(plane->id);
+	}
+
+	if (!dirty)
+		return 0;
+
+	return _g4x_compute_pipe_wm(crtc_state);
+}
+
 static int g4x_compute_intermediate_wm(struct intel_atomic_state *state,
 				       struct intel_crtc *crtc)
 {
@@ -1431,7 +1426,7 @@ static int g4x_compute_intermediate_wm(struct intel_atomic_state *state,
 	enum plane_id plane_id;
 
 	if (!new_crtc_state->hw.active ||
-	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi)) {
+	    intel_crtc_needs_modeset(new_crtc_state)) {
 		*intermediate = *optimal;
 
 		intermediate->cxsr = false;
@@ -1857,64 +1852,17 @@ static bool vlv_raw_crtc_wm_is_valid(const struct intel_crtc_state *crtc_state,
 		vlv_raw_plane_wm_is_valid(crtc_state, PLANE_CURSOR, level);
 }
 
-static int vlv_compute_pipe_wm(struct intel_atomic_state *state,
-			       struct intel_crtc *crtc)
+static int _vlv_compute_pipe_wm(struct intel_crtc_state *crtc_state)
 {
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	struct intel_crtc_state *crtc_state =
-		intel_atomic_get_new_crtc_state(state, crtc);
 	struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
 	const struct vlv_fifo_state *fifo_state =
 		&crtc_state->wm.vlv.fifo_state;
 	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
 	int num_active_planes = hweight8(active_planes);
-	bool needs_modeset = drm_atomic_crtc_needs_modeset(&crtc_state->uapi);
-	const struct intel_plane_state *old_plane_state;
-	const struct intel_plane_state *new_plane_state;
-	struct intel_plane *plane;
 	enum plane_id plane_id;
-	int level, ret, i;
-	unsigned int dirty = 0;
-
-	for_each_oldnew_intel_plane_in_state(state, plane,
-					     old_plane_state,
-					     new_plane_state, i) {
-		if (new_plane_state->hw.crtc != &crtc->base &&
-		    old_plane_state->hw.crtc != &crtc->base)
-			continue;
-
-		if (vlv_raw_plane_wm_compute(crtc_state, new_plane_state))
-			dirty |= BIT(plane->id);
-	}
-
-	/*
-	 * DSPARB registers may have been reset due to the
-	 * power well being turned off. Make sure we restore
-	 * them to a consistent state even if no primary/sprite
-	 * planes are initially active.
-	 */
-	if (needs_modeset)
-		crtc_state->fifo_changed = true;
-
-	if (!dirty)
-		return 0;
-
-	/* cursor changes don't warrant a FIFO recompute */
-	if (dirty & ~BIT(PLANE_CURSOR)) {
-		const struct intel_crtc_state *old_crtc_state =
-			intel_atomic_get_old_crtc_state(state, crtc);
-		const struct vlv_fifo_state *old_fifo_state =
-			&old_crtc_state->wm.vlv.fifo_state;
-
-		ret = vlv_compute_fifo(crtc_state);
-		if (ret)
-			return ret;
-
-		if (needs_modeset ||
-		    memcmp(old_fifo_state, fifo_state,
-			   sizeof(*fifo_state)) != 0)
-			crtc_state->fifo_changed = true;
-	}
+	int level;
 
 	/* initially allow all levels */
 	wm_state->num_levels = intel_wm_num_levels(dev_priv);
@@ -1961,6 +1909,66 @@ static int vlv_compute_pipe_wm(struct intel_atomic_state *state,
 	return 0;
 }
 
+static int vlv_compute_pipe_wm(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *old_plane_state;
+	const struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	unsigned int dirty = 0;
+	int i;
+
+	for_each_oldnew_intel_plane_in_state(state, plane,
+					     old_plane_state,
+					     new_plane_state, i) {
+		if (new_plane_state->hw.crtc != &crtc->base &&
+		    old_plane_state->hw.crtc != &crtc->base)
+			continue;
+
+		if (vlv_raw_plane_wm_compute(crtc_state, new_plane_state))
+			dirty |= BIT(plane->id);
+	}
+
+	/*
+	 * DSPARB registers may have been reset due to the
+	 * power well being turned off. Make sure we restore
+	 * them to a consistent state even if no primary/sprite
+	 * planes are initially active. We also force a FIFO
+	 * recomputation so that we are sure to sanitize the
+	 * FIFO setting we took over from the BIOS even if there
+	 * are no active planes on the crtc.
+	 */
+	if (intel_crtc_needs_modeset(crtc_state))
+		dirty = ~0;
+
+	if (!dirty)
+		return 0;
+
+	/* cursor changes don't warrant a FIFO recompute */
+	if (dirty & ~BIT(PLANE_CURSOR)) {
+		const struct intel_crtc_state *old_crtc_state =
+			intel_atomic_get_old_crtc_state(state, crtc);
+		const struct vlv_fifo_state *old_fifo_state =
+			&old_crtc_state->wm.vlv.fifo_state;
+		const struct vlv_fifo_state *new_fifo_state =
+			&crtc_state->wm.vlv.fifo_state;
+		int ret;
+
+		ret = vlv_compute_fifo(crtc_state);
+		if (ret)
+			return ret;
+
+		if (intel_crtc_needs_modeset(crtc_state) ||
+		    memcmp(old_fifo_state, new_fifo_state,
+			   sizeof(*new_fifo_state)) != 0)
+			crtc_state->fifo_changed = true;
+	}
+
+	return _vlv_compute_pipe_wm(crtc_state);
+}
+
 #define VLV_FIFO(plane, value) \
 	(((value) << DSPARB_ ## plane ## _SHIFT_VLV) & DSPARB_ ## plane ## _MASK_VLV)
 
@@ -2075,7 +2083,7 @@ static int vlv_compute_intermediate_wm(struct intel_atomic_state *state,
 	int level;
 
 	if (!new_crtc_state->hw.active ||
-	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi)) {
+	    intel_crtc_needs_modeset(new_crtc_state)) {
 		*intermediate = *optimal;
 
 		intermediate->cxsr = false;
@@ -3133,7 +3141,7 @@ static int ilk_compute_intermediate_wm(struct intel_atomic_state *state,
 	 */
 	*a = new_crtc_state->wm.ilk.optimal;
 	if (!new_crtc_state->hw.active ||
-	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi) ||
+	    intel_crtc_needs_modeset(new_crtc_state) ||
 	    state->skip_intermediate_wm)
 		return 0;
 
@@ -3458,7 +3466,6 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 {
 	struct ilk_wm_values *previous = &dev_priv->display.wm.hw;
 	unsigned int dirty;
-	u32 val;
 
 	dirty = ilk_compute_wm_dirty(dev_priv, previous, results);
 	if (!dirty)
@@ -3474,32 +3481,20 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
 		intel_uncore_write(&dev_priv->uncore, WM0_PIPE_ILK(PIPE_C), results->wm_pipe[2]);
 
 	if (dirty & WM_DIRTY_DDB) {
-		if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
-			val = intel_uncore_read(&dev_priv->uncore, WM_MISC);
-			if (results->partitioning == INTEL_DDB_PART_1_2)
-				val &= ~WM_MISC_DATA_PARTITION_5_6;
-			else
-				val |= WM_MISC_DATA_PARTITION_5_6;
-			intel_uncore_write(&dev_priv->uncore, WM_MISC, val);
-		} else {
-			val = intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL2);
-			if (results->partitioning == INTEL_DDB_PART_1_2)
-				val &= ~DISP_DATA_PARTITION_5_6;
-			else
-				val |= DISP_DATA_PARTITION_5_6;
-			intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL2, val);
-		}
-	}
-
-	if (dirty & WM_DIRTY_FBC) {
-		val = intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL);
-		if (results->enable_fbc_wm)
-			val &= ~DISP_FBC_WM_DIS;
+		if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+			intel_uncore_rmw(&dev_priv->uncore, WM_MISC, WM_MISC_DATA_PARTITION_5_6,
+					 results->partitioning == INTEL_DDB_PART_1_2 ? 0 :
+					 WM_MISC_DATA_PARTITION_5_6);
 		else
-			val |= DISP_FBC_WM_DIS;
-		intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, val);
+			intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL2, DISP_DATA_PARTITION_5_6,
+					 results->partitioning == INTEL_DDB_PART_1_2 ? 0 :
+					 DISP_DATA_PARTITION_5_6);
 	}
 
+	if (dirty & WM_DIRTY_FBC)
+		intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, DISP_FBC_WM_DIS,
+				 results->enable_fbc_wm ? 0 : DISP_FBC_WM_DIS);
+
 	if (dirty & WM_DIRTY_LP(1) &&
 	    previous->wm_lp_spr[0] != results->wm_lp_spr[0])
 		intel_uncore_write(&dev_priv->uncore, WM1S_LP_ILK, results->wm_lp_spr[0]);
@@ -3824,6 +3819,8 @@ void g4x_wm_get_hw_state(struct drm_i915_private *dev_priv)
 					     plane_id, USHRT_MAX);
 		g4x_raw_fbc_wm_set(crtc_state, level, USHRT_MAX);
 
+		g4x_invalidate_wms(crtc, active, level);
+
 		crtc_state->wm.g4x.optimal = *active;
 		crtc_state->wm.g4x.intermediate = *active;
 
@@ -3860,37 +3857,30 @@ void g4x_wm_sanitize(struct drm_i915_private *dev_priv)
 			to_intel_crtc_state(crtc->base.state);
 		struct intel_plane_state *plane_state =
 			to_intel_plane_state(plane->base.state);
-		struct g4x_wm_state *wm_state = &crtc_state->wm.g4x.optimal;
 		enum plane_id plane_id = plane->id;
-		int level;
+		int level, num_levels = intel_wm_num_levels(dev_priv);
 
 		if (plane_state->uapi.visible)
 			continue;
 
-		for (level = 0; level < 3; level++) {
+		for (level = 0; level < num_levels; level++) {
 			struct g4x_pipe_wm *raw =
 				&crtc_state->wm.g4x.raw[level];
 
 			raw->plane[plane_id] = 0;
-			wm_state->wm.plane[plane_id] = 0;
-		}
 
-		if (plane_id == PLANE_PRIMARY) {
-			for (level = 0; level < 3; level++) {
-				struct g4x_pipe_wm *raw =
-					&crtc_state->wm.g4x.raw[level];
+			if (plane_id == PLANE_PRIMARY)
 				raw->fbc = 0;
-			}
-
-			wm_state->sr.fbc = 0;
-			wm_state->hpll.fbc = 0;
-			wm_state->fbc_en = false;
 		}
 	}
 
 	for_each_intel_crtc(&dev_priv->drm, crtc) {
 		struct intel_crtc_state *crtc_state =
 			to_intel_crtc_state(crtc->base.state);
+		int ret;
+
+		ret = _g4x_compute_pipe_wm(crtc_state);
+		drm_WARN_ON(&dev_priv->drm, ret);
 
 		crtc_state->wm.g4x.intermediate =
 			crtc_state->wm.g4x.optimal;
@@ -4016,30 +4006,27 @@ void vlv_wm_sanitize(struct drm_i915_private *dev_priv)
 			to_intel_crtc_state(crtc->base.state);
 		struct intel_plane_state *plane_state =
 			to_intel_plane_state(plane->base.state);
-		struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
-		const struct vlv_fifo_state *fifo_state =
-			&crtc_state->wm.vlv.fifo_state;
 		enum plane_id plane_id = plane->id;
-		int level;
+		int level, num_levels = intel_wm_num_levels(dev_priv);
 
 		if (plane_state->uapi.visible)
 			continue;
 
-		for (level = 0; level < wm_state->num_levels; level++) {
+		for (level = 0; level < num_levels; level++) {
 			struct g4x_pipe_wm *raw =
 				&crtc_state->wm.vlv.raw[level];
 
 			raw->plane[plane_id] = 0;
-
-			wm_state->wm[level].plane[plane_id] =
-				vlv_invert_wm_value(raw->plane[plane_id],
-						    fifo_state->plane[plane_id]);
 		}
 	}
 
 	for_each_intel_crtc(&dev_priv->drm, crtc) {
 		struct intel_crtc_state *crtc_state =
 			to_intel_crtc_state(crtc->base.state);
+		int ret;
+
+		ret = _vlv_compute_pipe_wm(crtc_state);
+		drm_WARN_ON(&dev_priv->drm, ret);
 
 		crtc_state->wm.vlv.intermediate =
 			crtc_state->wm.vlv.optimal;
@@ -4057,9 +4044,9 @@ void vlv_wm_sanitize(struct drm_i915_private *dev_priv)
  */
 static void ilk_init_lp_watermarks(struct drm_i915_private *dev_priv)
 {
-	intel_uncore_write(&dev_priv->uncore, WM3_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM3_LP_ILK) & ~WM_LP_ENABLE);
-	intel_uncore_write(&dev_priv->uncore, WM2_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM2_LP_ILK) & ~WM_LP_ENABLE);
-	intel_uncore_write(&dev_priv->uncore, WM1_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM1_LP_ILK) & ~WM_LP_ENABLE);
+	intel_uncore_rmw(&dev_priv->uncore, WM3_LP_ILK, WM_LP_ENABLE, 0);
+	intel_uncore_rmw(&dev_priv->uncore, WM2_LP_ILK, WM_LP_ENABLE, 0);
+	intel_uncore_rmw(&dev_priv->uncore, WM1_LP_ILK, WM_LP_ENABLE, 0);
 
 	/*
 	 * Don't touch WM_LP_SPRITE_ENABLE here.
@@ -4113,11 +4100,9 @@ static void g4x_disable_trickle_feed(struct drm_i915_private *dev_priv)
 	enum pipe pipe;
 
 	for_each_pipe(dev_priv, pipe) {
-		intel_uncore_write(&dev_priv->uncore, DSPCNTR(pipe),
-			   intel_uncore_read(&dev_priv->uncore, DSPCNTR(pipe)) |
-			   DISP_TRICKLE_FEED_DISABLE);
+		intel_uncore_rmw(&dev_priv->uncore, DSPCNTR(pipe), 0, DISP_TRICKLE_FEED_DISABLE);
 
-		intel_uncore_write(&dev_priv->uncore, DSPSURF(pipe), intel_uncore_read(&dev_priv->uncore, DSPSURF(pipe)));
+		intel_uncore_rmw(&dev_priv->uncore, DSPSURF(pipe), 0, 0);
 		intel_uncore_posting_read(&dev_priv->uncore, DSPSURF(pipe));
 	}
 }
@@ -4164,19 +4149,13 @@ static void ilk_init_clock_gating(struct drm_i915_private *dev_priv)
 	 */
 	if (IS_IRONLAKE_M(dev_priv)) {
 		/* WaFbcAsynchFlipDisableFbcQueue:ilk */
-		intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1,
-			   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1) |
-			   ILK_FBCQ_DIS);
-		intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
-			   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
-			   ILK_DPARB_GATE);
+		intel_uncore_rmw(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1, 0, ILK_FBCQ_DIS);
+		intel_uncore_rmw(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2, 0, ILK_DPARB_GATE);
 	}
 
 	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, dspclk_gate);
 
-	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
-		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
-		   ILK_ELPIN_409_SELECT);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2, 0, ILK_ELPIN_409_SELECT);
 
 	g4x_disable_trickle_feed(dev_priv);
 
@@ -4196,8 +4175,7 @@ static void cpt_init_clock_gating(struct drm_i915_private *dev_priv)
 	intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE |
 		   PCH_DPLUNIT_CLOCK_GATE_DISABLE |
 		   PCH_CPUNIT_CLOCK_GATE_DISABLE);
-	intel_uncore_write(&dev_priv->uncore, SOUTH_CHICKEN2, intel_uncore_read(&dev_priv->uncore, SOUTH_CHICKEN2) |
-		   DPLS_EDP_PPS_FIX_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, SOUTH_CHICKEN2, 0, DPLS_EDP_PPS_FIX_DIS);
 	/* The below fixes the weird display corruption, a few pixels shifted
 	 * downward, on (only) LVDS of some HP laptops with IVY.
 	 */
@@ -4235,9 +4213,7 @@ static void gen6_init_clock_gating(struct drm_i915_private *dev_priv)
 
 	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, dspclk_gate);
 
-	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
-		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
-		   ILK_ELPIN_409_SELECT);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2, 0, ILK_ELPIN_409_SELECT);
 
 	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1,
 		   intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
@@ -4297,14 +4273,12 @@ static void lpt_init_clock_gating(struct drm_i915_private *dev_priv)
 	 * disabled when not needed anymore in order to save power.
 	 */
 	if (HAS_PCH_LPT_LP(dev_priv))
-		intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D,
-			   intel_uncore_read(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D) |
-			   PCH_LP_PARTITION_LEVEL_DISABLE);
+		intel_uncore_rmw(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D,
+				 0, PCH_LP_PARTITION_LEVEL_DISABLE);
 
 	/* WADPOClockGatingDisable:hsw */
-	intel_uncore_write(&dev_priv->uncore, TRANS_CHICKEN1(PIPE_A),
-		   intel_uncore_read(&dev_priv->uncore, TRANS_CHICKEN1(PIPE_A)) |
-		   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, TRANS_CHICKEN1(PIPE_A),
+			 0, TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
 }
 
 static void lpt_suspend_hw(struct drm_i915_private *dev_priv)
@@ -4325,22 +4299,22 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv,
 	u32 val;
 
 	/* WaTempDisableDOPClkGating:bdw */
-	misccpctl = intel_uncore_read(&dev_priv->uncore, GEN7_MISCCPCTL);
-	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl & ~GEN7_DOP_CLOCK_GATE_ENABLE);
+	misccpctl = intel_gt_mcr_multicast_rmw(to_gt(dev_priv), GEN8_MISCCPCTL,
+					       GEN8_DOP_CLOCK_GATE_ENABLE, 0);
 
-	val = intel_uncore_read(&dev_priv->uncore, GEN8_L3SQCREG1);
+	val = intel_gt_mcr_read_any(to_gt(dev_priv), GEN8_L3SQCREG1);
 	val &= ~L3_PRIO_CREDITS_MASK;
 	val |= L3_GENERAL_PRIO_CREDITS(general_prio_credits);
 	val |= L3_HIGH_PRIO_CREDITS(high_prio_credits);
-	intel_uncore_write(&dev_priv->uncore, GEN8_L3SQCREG1, val);
+	intel_gt_mcr_multicast_write(to_gt(dev_priv), GEN8_L3SQCREG1, val);
 
 	/*
 	 * Wait at least 100 clocks before re-enabling clock gating.
 	 * See the definition of L3SQCREG1 in BSpec.
 	 */
-	intel_uncore_posting_read(&dev_priv->uncore, GEN8_L3SQCREG1);
+	intel_gt_mcr_read_any(to_gt(dev_priv), GEN8_L3SQCREG1);
 	udelay(1);
-	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl);
+	intel_gt_mcr_multicast_write(to_gt(dev_priv), GEN8_MISCCPCTL, misccpctl);
 }
 
 static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4363,8 +4337,7 @@ static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
 
 	/* Wa_1409825376:tgl (pre-prod)*/
 	if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_C0))
-		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
-			   TGL_VRH_GATING_DIS);
+		intel_uncore_rmw(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, 0, TGL_VRH_GATING_DIS);
 
 	/* Wa_14013723622:tgl,rkl,dg1,adl-s */
 	if (DISPLAY_VER(dev_priv) == 12)
@@ -4389,8 +4362,7 @@ static void dg1_init_clock_gating(struct drm_i915_private *dev_priv)
 
 	/* Wa_1409836686:dg1[a0] */
 	if (IS_DG1_GRAPHICS_STEP(dev_priv, STEP_A0, STEP_B0))
-		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
-			   DPT_GATING_DIS);
+		intel_uncore_rmw(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, 0, DPT_GATING_DIS);
 }
 
 static void xehpsdv_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4432,8 +4404,7 @@ static void cnp_init_clock_gating(struct drm_i915_private *dev_priv)
 		return;
 
 	/* Display WA #1181 WaSouthDisplayDisablePWMCGEGating: cnp */
-	intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, intel_uncore_read(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D) |
-		   CNP_PWM_CGE_GATING_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, 0, CNP_PWM_CGE_GATING_DISABLE);
 }
 
 static void cfl_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4442,23 +4413,20 @@ static void cfl_init_clock_gating(struct drm_i915_private *dev_priv)
 	gen9_init_clock_gating(dev_priv);
 
 	/* WAC6entrylatency:cfl */
-	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
-		   FBC_LLC_FULLY_OPEN);
+	intel_uncore_rmw(&dev_priv->uncore, FBC_LLC_READ_CTRL, 0, FBC_LLC_FULLY_OPEN);
 
 	/*
 	 * WaFbcTurnOffFbcWatermark:cfl
 	 * Display WA #0562: cfl
 	 */
-	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
-		   DISP_FBC_WM_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, 0, DISP_FBC_WM_DIS);
 
 	/*
 	 * WaFbcNukeOnHostModify:cfl
 	 * Display WA #0873: cfl
 	 */
-	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
-			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
-			   DPFC_NUKE_ON_ANY_MODIFICATION);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			 0, DPFC_NUKE_ON_ANY_MODIFICATION);
 }
 
 static void kbl_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4466,33 +4434,30 @@ static void kbl_init_clock_gating(struct drm_i915_private *dev_priv)
 	gen9_init_clock_gating(dev_priv);
 
 	/* WAC6entrylatency:kbl */
-	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
-		   FBC_LLC_FULLY_OPEN);
+	intel_uncore_rmw(&dev_priv->uncore, FBC_LLC_READ_CTRL, 0, FBC_LLC_FULLY_OPEN);
 
 	/* WaDisableSDEUnitClockGating:kbl */
 	if (IS_KBL_GRAPHICS_STEP(dev_priv, 0, STEP_C0))
-		intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
-			   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+		intel_uncore_rmw(&dev_priv->uncore, GEN8_UCGCTL6,
+				 0, GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
 	/* WaDisableGamClockGating:kbl */
 	if (IS_KBL_GRAPHICS_STEP(dev_priv, 0, STEP_C0))
-		intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1, intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
-			   GEN6_GAMUNIT_CLOCK_GATE_DISABLE);
+		intel_uncore_rmw(&dev_priv->uncore, GEN6_UCGCTL1,
+				 0, GEN6_GAMUNIT_CLOCK_GATE_DISABLE);
 
 	/*
 	 * WaFbcTurnOffFbcWatermark:kbl
 	 * Display WA #0562: kbl
 	 */
-	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
-		   DISP_FBC_WM_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, 0, DISP_FBC_WM_DIS);
 
 	/*
 	 * WaFbcNukeOnHostModify:kbl
 	 * Display WA #0873: kbl
 	 */
-	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
-			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
-			   DPFC_NUKE_ON_ANY_MODIFICATION);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			 0, DPFC_NUKE_ON_ANY_MODIFICATION);
 }
 
 static void skl_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4500,35 +4465,30 @@ static void skl_init_clock_gating(struct drm_i915_private *dev_priv)
 	gen9_init_clock_gating(dev_priv);
 
 	/* WaDisableDopClockGating:skl */
-	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, intel_uncore_read(&dev_priv->uncore, GEN7_MISCCPCTL) &
-		   ~GEN7_DOP_CLOCK_GATE_ENABLE);
+	intel_gt_mcr_multicast_rmw(to_gt(dev_priv), GEN8_MISCCPCTL,
+				   GEN8_DOP_CLOCK_GATE_ENABLE, 0);
 
 	/* WAC6entrylatency:skl */
-	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
-		   FBC_LLC_FULLY_OPEN);
+	intel_uncore_rmw(&dev_priv->uncore, FBC_LLC_READ_CTRL, 0, FBC_LLC_FULLY_OPEN);
 
 	/*
 	 * WaFbcTurnOffFbcWatermark:skl
 	 * Display WA #0562: skl
 	 */
-	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
-		   DISP_FBC_WM_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, DISP_ARB_CTL, 0, DISP_FBC_WM_DIS);
 
 	/*
 	 * WaFbcNukeOnHostModify:skl
 	 * Display WA #0873: skl
 	 */
-	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
-			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
-			   DPFC_NUKE_ON_ANY_MODIFICATION);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			 0, DPFC_NUKE_ON_ANY_MODIFICATION);
 
 	/*
 	 * WaFbcHighMemBwCorruptionAvoidance:skl
 	 * Display WA #0883: skl
 	 */
-	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
-			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
-			   DPFC_DISABLE_DUMMY0);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A), 0, DPFC_DISABLE_DUMMY0);
 }
 
 static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -4536,43 +4496,37 @@ static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)
 	enum pipe pipe;
 
 	/* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
-	intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A),
-		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A)) |
-		   HSW_FBCQ_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A), 0, HSW_FBCQ_DIS);
 
 	/* WaSwitchSolVfFArbitrationPriority:bdw */
-	intel_uncore_write(&dev_priv->uncore, GAM_ECOCHK, intel_uncore_read(&dev_priv->uncore, GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
+	intel_uncore_rmw(&dev_priv->uncore, GAM_ECOCHK, 0, HSW_ECOCHK_ARB_PRIO_SOL);
 
 	/* WaPsrDPAMaskVBlankInSRD:bdw */
-	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
-		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) | DPA_MASK_VBLANK_SRD);
+	intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PAR1_1, 0, DPA_MASK_VBLANK_SRD);
 
 	for_each_pipe(dev_priv, pipe) {
 		/* WaPsrDPRSUnmaskVBlankInSRD:bdw */
-		intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
-			   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe)) |
-			   BDW_DPRS_MASK_VBLANK_SRD);
+		intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
+				 0, BDW_DPRS_MASK_VBLANK_SRD);
 	}
 
 	/* WaVSRefCountFullforceMissDisable:bdw */
 	/* WaDSRefCountFullforceMissDisable:bdw */
-	intel_uncore_write(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
-		   intel_uncore_read(&dev_priv->uncore, GEN7_FF_THREAD_MODE) &
-		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
+			 GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME, 0);
 
 	intel_uncore_write(&dev_priv->uncore, RING_PSMI_CTL(RENDER_RING_BASE),
 		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
 
 	/* WaDisableSDEUnitClockGating:bdw */
-	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
-		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_UCGCTL6, 0, GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
 	/* WaProgramL3SqcReg1Default:bdw */
 	gen8_set_l3sqc_credits(dev_priv, 30, 2);
 
 	/* WaKVMNotificationOnConfigChange:bdw */
-	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR2_1, intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR2_1)
-		   | KVM_CONFIG_CHANGE_NOTIFICATION_SELECT);
+	intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PAR2_1,
+			 0, KVM_CONFIG_CHANGE_NOTIFICATION_SELECT);
 
 	lpt_init_clock_gating(dev_priv);
 
@@ -4581,38 +4535,30 @@ static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)
 	 * Also see the CHICKEN2 write in bdw_init_workarounds() to disable DOP
 	 * clock gating.
 	 */
-	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1,
-		   intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) | GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN6_UCGCTL1, 0, GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
 }
 
 static void hsw_init_clock_gating(struct drm_i915_private *dev_priv)
 {
 	/* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
-	intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A),
-		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A)) |
-		   HSW_FBCQ_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A), 0, HSW_FBCQ_DIS);
 
 	/* This is required by WaCatErrorRejectionIssue:hsw */
-	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
-		   intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
-		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+			 0, GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
 	/* WaSwitchSolVfFArbitrationPriority:hsw */
-	intel_uncore_write(&dev_priv->uncore, GAM_ECOCHK, intel_uncore_read(&dev_priv->uncore, GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
+	intel_uncore_rmw(&dev_priv->uncore, GAM_ECOCHK, 0, HSW_ECOCHK_ARB_PRIO_SOL);
 
 	lpt_init_clock_gating(dev_priv);
 }
 
 static void ivb_init_clock_gating(struct drm_i915_private *dev_priv)
 {
-	u32 snpcr;
-
 	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
 
 	/* WaFbcAsynchFlipDisableFbcQueue:ivb */
-	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1,
-		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1) |
-		   ILK_FBCQ_DIS);
+	intel_uncore_rmw(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1, 0, ILK_FBCQ_DIS);
 
 	/* WaDisableBackToBackFlipFix:ivb */
 	intel_uncore_write(&dev_priv->uncore, IVB_CHICKEN3,
@@ -4638,16 +4584,13 @@ static void ivb_init_clock_gating(struct drm_i915_private *dev_priv)
 		   GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
 
 	/* This is required by WaCatErrorRejectionIssue:ivb */
-	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
-			intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
-			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+			 0, GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
 	g4x_disable_trickle_feed(dev_priv);
 
-	snpcr = intel_uncore_read(&dev_priv->uncore, GEN6_MBCUNIT_SNPCR);
-	snpcr &= ~GEN6_MBC_SNPCR_MASK;
-	snpcr |= GEN6_MBC_SNPCR_MED;
-	intel_uncore_write(&dev_priv->uncore, GEN6_MBCUNIT_SNPCR, snpcr);
+	intel_uncore_rmw(&dev_priv->uncore, GEN6_MBCUNIT_SNPCR, GEN6_MBC_SNPCR_MASK,
+			 GEN6_MBC_SNPCR_MED);
 
 	if (!HAS_PCH_NOP(dev_priv))
 		cpt_init_clock_gating(dev_priv);
@@ -4667,9 +4610,8 @@ static void vlv_init_clock_gating(struct drm_i915_private *dev_priv)
 		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
 	/* This is required by WaCatErrorRejectionIssue:vlv */
-	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
-		   intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
-		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+			 0, GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
 	/*
 	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
@@ -4681,8 +4623,7 @@ static void vlv_init_clock_gating(struct drm_i915_private *dev_priv)
 	/* WaDisableL3Bank2xClockGate:vlv
 	 * Disabling L3 clock gating- MMIO 940c[25] = 1
 	 * Set bit 25, to disable L3_BANK_2x_CLK_GATING */
-	intel_uncore_write(&dev_priv->uncore, GEN7_UCGCTL4,
-		   intel_uncore_read(&dev_priv->uncore, GEN7_UCGCTL4) | GEN7_L3BANK2X_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_UCGCTL4, 0, GEN7_L3BANK2X_CLOCK_GATE_DISABLE);
 
 	/*
 	 * WaDisableVLVClockGating_VBIIssue:vlv
@@ -4696,21 +4637,18 @@ static void chv_init_clock_gating(struct drm_i915_private *dev_priv)
 {
 	/* WaVSRefCountFullforceMissDisable:chv */
 	/* WaDSRefCountFullforceMissDisable:chv */
-	intel_uncore_write(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
-		   intel_uncore_read(&dev_priv->uncore, GEN7_FF_THREAD_MODE) &
-		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
+	intel_uncore_rmw(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
+			 GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME, 0);
 
 	/* WaDisableSemaphoreAndSyncFlipWait:chv */
 	intel_uncore_write(&dev_priv->uncore, RING_PSMI_CTL(RENDER_RING_BASE),
 		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
 
 	/* WaDisableCSUnitClockGating:chv */
-	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1, intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
-		   GEN6_CSUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN6_UCGCTL1, 0, GEN6_CSUNIT_CLOCK_GATE_DISABLE);
 
 	/* WaDisableSDEUnitClockGating:chv */
-	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
-		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_UCGCTL6, 0, GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
 
 	/*
 	 * WaProgramL3SqcReg1Default:chv