diff options
| author | Ben Skeggs <bskeggs@redhat.com> | 2012-03-21 13:51:03 +1000 |
|---|---|---|
| committer | Ben Skeggs <bskeggs@redhat.com> | 2012-03-22 17:17:36 +1000 |
| commit | b08abd4e9a11d637d3c2ff52b2ebbc1b3f686d06 (patch) | |
| tree | b965122d55743e4094cccdb1c3884299693ee7b7 /drivers/gpu | |
| parent | b5b2e5988bd18a2f6e3f192adf7439599de00d3f (diff) | |
| download | linux-stable-b08abd4e9a11d637d3c2ff52b2ebbc1b3f686d06.tar.gz linux-stable-b08abd4e9a11d637d3c2ff52b2ebbc1b3f686d06.tar.bz2 linux-stable-b08abd4e9a11d637d3c2ff52b2ebbc1b3f686d06.zip | |
drm/nouveau: move fence sequence check to start of loop
I want to be able to use REF_CNT from other places in the kernel without
pushing a fence object onto the list of emitted fences.
The current code makes an assumption that every time the acked sequence is
bumped that there's at least one fence on the list that'll be signalled.
This will no longer be true in the near future.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Diffstat (limited to 'drivers/gpu')
| -rw-r--r-- | drivers/gpu/drm/nouveau/nouveau_fence.c | 11 |
1 files changed, 5 insertions, 6 deletions
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index f676ecd3fd3c..c1dc20f6cb85 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -93,18 +93,17 @@ nouveau_fence_update(struct nouveau_channel *chan) } list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) { - sequence = fence->sequence; + if (fence->sequence > chan->fence.sequence_ack) + break; + fence->signalled = true; list_del(&fence->entry); - - if (unlikely(fence->work)) + if (fence->work) fence->work(fence->priv, true); kref_put(&fence->refcount, nouveau_fence_del); - - if (sequence == chan->fence.sequence_ack) - break; } + out: spin_unlock(&chan->fence.lock); } |