On Tue, Sep 09, 2014 at 07:02:43AM +0100, Chris Wilson wrote:
> Before we process the final unbind on an object and move it to the
> unbound list, it is semantically cleaner if there are no more active
> references to the object. (An active reference would imply that it was
> still being accessed by the GPU after it became inaccessible.) The
> caveat is that all callsites must be prepared for the object to
> disappeared during the unbind - i.e. they must hold their own reference.
>
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Oh dear, more fuel for gem nightmares ;-)
Both patches merged to dinq, thanks a lot. Aside, git diff -b is awesome
for getting rid of indent changes in diffs. Just discovered that ...
-Daniel
> ---
> drivers/gpu/drm/i915/i915_gem.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 4ca3a6dcf10b..f5362cabd921 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -2944,6 +2944,9 @@ 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);
>
> --
> 2.1.0
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
