dma-fence: Allow wait_any_timeout for all fences
authorDaniel Vetter <daniel.vetter@ffwll.ch>
Thu, 3 May 2018 14:25:51 +0000 (16:25 +0200)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Mon, 2 Jul 2018 08:20:53 +0000 (10:20 +0200)
commit796422f227eef662256dcf74f609abd48cfd1cc9
tree8cb0876de40a565ea0d78372005d3319148daf34
parentc701317a3eb8c012364a8d468f20eabf6df1ad77
dma-fence: Allow wait_any_timeout for all fences

When this was introduced in

commit a519435a96597d8cd96123246fea4ae5a6c90b02
Author: Christian König <christian.koenig@amd.com>
Date:   Tue Oct 20 16:34:16 2015 +0200

    dma-buf/fence: add fence_wait_any_timeout function v2

there was a restriction added that this only works if the dma-fence
uses the dma_fence_default_wait hook. Which works for amdgpu, which is
the only caller. Well, until you share some buffers with e.g. i915,
then you get an -EINVAL.

But there's really no reason for this, because all drivers must
support callbacks. The special ->wait hook is only as an optimization;
if the driver needs to create a worker thread for an active callback,
then it can avoid to do that if it knows that there's a process
context available already. So ->wait is just an optimization, just
using the logic in dma_fence_default_wait() should work for all
drivers.

Let's remove this restriction.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Gustavo Padovan <gustavo@padovan.org>
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180503142603.28513-4-daniel.vetter@ffwll.ch
drivers/dma-buf/dma-fence.c