提交 d4b7648d 编写于 作者: C Christian König 提交者: Alex Deucher

drm/amdgpu: fix amdgpu_sync_resv v2

Fixes a bug introduced by AMDGPU_GEM_CREATE_EXPLICIT_SYNC. We still need
to wait for pipelined moves in the shared fences list.

v2: fix typo
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAndres Rodriguez <andresx7@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
上级 fdb1a223
......@@ -194,9 +194,6 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
f = reservation_object_get_excl(resv);
r = amdgpu_sync_fence(adev, sync, f, false);
if (explicit_sync)
return r;
flist = reservation_object_get_list(resv);
if (!flist || r)
return r;
......@@ -215,11 +212,11 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
(fence_owner == AMDGPU_FENCE_OWNER_VM)))
continue;
/* Ignore fence from the same owner as
/* Ignore fence from the same owner and explicit one as
* long as it isn't undefined.
*/
if (owner != AMDGPU_FENCE_OWNER_UNDEFINED &&
fence_owner == owner)
(fence_owner == owner || explicit_sync))
continue;
}
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册