We only want to mark the transition from unfenced GPU access by an
execbuffer, so that we are forced to flush any pending writes through
the fence before updating the register.
In applying this fix for a corruption bug, we do lose the ability to
detect the earliest end of GPU fenced access, thus disabling the inherent
optimization.
Signed-off-by: Chris Wilson <ch...@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vet...@ffwll.ch>
---
drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 20a4cc5..a07911f 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -911,7 +911,7 @@ i915_gem_execbuffer_move_to_active(struct list_head
*objects,
obj->base.read_domains = obj->base.pending_read_domains;
obj->base.write_domain = obj->base.pending_write_domain;
- obj->fenced_gpu_access = obj->pending_fenced_gpu_access;
+ obj->fenced_gpu_access |= obj->pending_fenced_gpu_access;
i915_gem_object_move_to_active(obj, ring, seqno);
if (obj->base.write_domain) {
--
1.7.4.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
