This reverts commit eb1711bb94991e93669c5a1b5f84f11be2d51ea1.
This patch has been reported by two people to cause memory corruption
and application crashes. It exists solely to fix a problem with the
VT-d workarounds for Ironlake. I would back those out as well as
without this change, that work-around causes an infinite recursion for
anyone foolish enough to actually try VT-d for graphics.
So, without this patch, and when using VT-d for graphics on Ironlake,
people will experience infinite recursion crashes. With this patch, we
get memory corruption on all systems.
Cc: dri-devel at lists.freedesktop.org
Cc: Daniel Vetter <daniel at ffwll.ch>
Reported-by: Alex Villac??s Lasso <a_villacis at palosanto.com>
Reported-by: Dirk Hohndel <dirk at hohndel.org>
Signed-off-by: Keith Packard <keithp at keithp.com>
---
drivers/gpu/drm/i915/i915_gem.c | 7 +------
1 files changed, 1 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 60ff1b6..8359dc7 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2026,13 +2026,8 @@ i915_wait_request(struct intel_ring_buffer *ring,
* to handle this, the waiter on a request often wants an associated
* buffer to have made it to the inactive list, and we would need
* a separate wait queue to handle that.
- *
- * To avoid a recursion with the ilk VT-d workaround (that calls
- * gpu_idle when unbinding objects with interruptible==false) don't
- * retire requests in that case (because it might call unbind if the
- * active list holds the last reference to the object).
*/
- if (ret == 0 && dev_priv->mm.interruptible)
+ if (ret == 0)
i915_gem_retire_requests_ring(ring);
return ret;
--
1.7.7.3
