Merge tag 'drm-intel-fixes-2013-07-11' of git://people.freedesktop.org/~danvet/drm-intel

One feature latecomer, I've forgotten to merge the patch to reeanble the
Haswell power well feature now that the audio interaction is fixed up.
Since that was the only unfixed issue with it I've figured I could throw
it in a bit late, and it's trivial to revert in case I'm wrong.

Otherwise all bug/regression fixes:
- Fix status page reinit after gpu hangs, spotted by more paranoid igt
  checks.
- Fix object list walking fumble regression in the shrinker (only the
  counting part, the actual shrinking code was correct so no Oops
  potential), from Xiong Zhang.
- Fix DP 1.2 bw limits (Imre).
- Restore legacy forcewake on ivb, too many broken biosen out there. We
  dump a warn though that recent userspace might fall over with that
  config (Guenter Roeck).
- Patch up the gen2 cs tlb w/a.
- Improve the fence coherency w/a now that we have a better understanding
  what's going on. The removed wbinvd+ipi should make -rt folks happy. Big
  thanks to Jon Bloomfield for figuring this out, patches from Chris.
- Fix write-read race when switching ring (Chris). Spotted with code
  inspection, but now we also have an igt for it.

There's an ugly regression we're still working on introduced between
3.10-rc7 and 3.10.0. Unfortunately we can't just revert the offender since
that one fixes another regression :( I've asked Steven to include my
-fixes branch into linux-next to prevent such fallout in the future,
hopefully.

* tag 'drm-intel-fixes-2013-07-11' of git://people.freedesktop.org/~danvet/drm-intel:
  Revert "drm/i915: Workaround incoherence between fences and LLC across multiple CPUs"
  drm/i915: Fix incoherence with fence updates on Sandybridge+
  drm/i915: Fix write-read race with multiple rings
  Partially revert "drm/i915: unconditionally use mt forcewake on hsw/ivb"
  drm/i915: fix lane bandwidth capping for DP 1.2 sinks
  drm/i915: fix up ring cleanup for the i830/i845 CS tlb w/a
  drm/i915: Correct obj->mm_list link to dev_priv->dev_priv->mm.inactive_list
  drm/i915: switch disable_power_well default value to 1
  drm/i915: reinit status page registers after gpu reset

1 files changed, 33 insertions, 50 deletions

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 4200c32407ec..97afd2639fb6 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1880,6 +1880,10 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 	u32 seqno = intel_ring_get_seqno(ring);
 
 	BUG_ON(ring == NULL);
+	if (obj->ring != ring && obj->last_write_seqno) {
+		/* Keep the seqno relative to the current ring */
+		obj->last_write_seqno = seqno;
+	}
 	obj->ring = ring;
 
 	/* Add a reference if we're newly entering the active list. */
@@ -2653,7 +2657,6 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	int fence_reg;
 	int fence_pitch_shift;
-	uint64_t val;
 
 	if (INTEL_INFO(dev)->gen >= 6) {
 		fence_reg = FENCE_REG_SANDYBRIDGE_0;
@@ -2663,8 +2666,23 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 		fence_pitch_shift = I965_FENCE_PITCH_SHIFT;
 	}
 
+	fence_reg += reg * 8;
+
+	/* To w/a incoherency with non-atomic 64-bit register updates,
+	 * we split the 64-bit update into two 32-bit writes. In order
+	 * for a partial fence not to be evaluated between writes, we
+	 * precede the update with write to turn off the fence register,
+	 * and only enable the fence as the last step.
+	 *
+	 * For extra levels of paranoia, we make sure each step lands
+	 * before applying the next step.
+	 */
+	I915_WRITE(fence_reg, 0);
+	POSTING_READ(fence_reg);
+
 	if (obj) {
 		u32 size = obj->gtt_space->size;
+		uint64_t val;
 
 		val = (uint64_t)((obj->gtt_offset + size - 4096) &
 				 0xfffff000) << 32;
@@ -2673,12 +2691,16 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 		if (obj->tiling_mode == I915_TILING_Y)
 			val |= 1 << I965_FENCE_TILING_Y_SHIFT;
 		val |= I965_FENCE_REG_VALID;
-	} else
-		val = 0;
 
-	fence_reg += reg * 8;
-	I915_WRITE64(fence_reg, val);
-	POSTING_READ(fence_reg);
+		I915_WRITE(fence_reg + 4, val >> 32);
+		POSTING_READ(fence_reg + 4);
+
+		I915_WRITE(fence_reg + 0, val);
+		POSTING_READ(fence_reg);
+	} else {
+		I915_WRITE(fence_reg + 4, 0);
+		POSTING_READ(fence_reg + 4);
+	}
 }
 
 static void i915_write_fence_reg(struct drm_device *dev, int reg,
@@ -2796,56 +2818,17 @@ static inline int fence_number(struct drm_i915_private *dev_priv,
 	return fence - dev_priv->fence_regs;
 }
 
-struct write_fence {
-	struct drm_device *dev;
-	struct drm_i915_gem_object *obj;
-	int fence;
-};
-
-static void i915_gem_write_fence__ipi(void *data)
-{
-	struct write_fence *args = data;
-
-	/* Required for SNB+ with LLC */
-	wbinvd();
-
-	/* Required for VLV */
-	i915_gem_write_fence(args->dev, args->fence, args->obj);
-}
-
 static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 					 struct drm_i915_fence_reg *fence,
 					 bool enable)
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
-	struct write_fence args = {
-		.dev = obj->base.dev,
-		.fence = fence_number(dev_priv, fence),
-		.obj = enable ? obj : NULL,
-	};
-
-	/* In order to fully serialize access to the fenced region and
-	 * the update to the fence register we need to take extreme
-	 * measures on SNB+. In theory, the write to the fence register
-	 * flushes all memory transactions before, and coupled with the
-	 * mb() placed around the register write we serialise all memory
-	 * operations with respect to the changes in the tiler. Yet, on
-	 * SNB+ we need to take a step further and emit an explicit wbinvd()
-	 * on each processor in order to manually flush all memory
-	 * transactions before updating the fence register.
-	 *
-	 * However, Valleyview complicates matter. There the wbinvd is
-	 * insufficient and unlike SNB/IVB requires the serialising
-	 * register write. (Note that that register write by itself is
-	 * conversely not sufficient for SNB+.) To compromise, we do both.
-	 */
-	if (INTEL_INFO(args.dev)->gen >= 6)
-		on_each_cpu(i915_gem_write_fence__ipi, &args, 1);
-	else
-		i915_gem_write_fence(args.dev, args.fence, args.obj);
+	int reg = fence_number(dev_priv, fence);
+
+	i915_gem_write_fence(obj->base.dev, reg, enable ? obj : NULL);
 
 	if (enable) {
-		obj->fence_reg = args.fence;
+		obj->fence_reg = reg;
 		fence->obj = obj;
 		list_move_tail(&fence->lru_list, &dev_priv->mm.fence_list);
 	} else {
@@ -4611,7 +4594,7 @@ i915_gem_inactive_shrink(struct shrinker *shrinker, struct shrink_control *sc)
 	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)
 		if (obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;
-	list_for_each_entry(obj, &dev_priv->mm.inactive_list, global_list)
+	list_for_each_entry(obj, &dev_priv->mm.inactive_list, mm_list)
 		if (obj->pin_count == 0 && obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;