MailboxSync: Imply fence with sync point

MailboxSync: Imply fence with sync point

Instead of implying the fences based on read and write access
to the underlying texture, have sync point insertion imply
an EGL fence which is acquired when we wait on that sync point.

BUG=399226
Committed: https://crrev.com/173a20dabc8f32182b48e999bd16575598ba4a44
Cr-Commit-Position: refs/heads/master@{#300709}

[no sievers, 2014-10-21 21:06:54]

sievers@chromium.org changed reviewers:
+ piman@chromium.org

[no sievers, 2014-10-21 21:09:13]

ptal

It doesn't handle this scenario:

ctx A, B in share group 1 on thread T1. ctx C in share group 2 on thread T2.

A: modify texture and glFlush()
B: InsertSyncPoint
C: WaitSyncPoint

But that is a known problem with EGL and we can use virtualization for all share
groups to work around that. (Currently WebView only uses virtualization for one
of the two GL threads, but we can enable it for the other thread also.)

[piman, 2014-10-21 23:28:51]

lgtm

https://codereview.chromium.org/661973003/diff/100001/gpu/command_buffer/serv...
File gpu/command_buffer/service/in_process_command_buffer.cc (right):

https://codereview.chromium.org/661973003/diff/100001/gpu/command_buffer/serv...
gpu/command_buffer/service/in_process_command_buffer.cc:183: }
nit: you don't need this trivial wrapper any more, you can just fold into
InProcessCommandBuffer::WaitSyncPointOnGpuThread

[no sievers, 2014-10-22 15:59:22]

https://codereview.chromium.org/661973003/diff/100001/gpu/command_buffer/serv...
File gpu/command_buffer/service/in_process_command_buffer.cc (right):

https://codereview.chromium.org/661973003/diff/100001/gpu/command_buffer/serv...
gpu/command_buffer/service/in_process_command_buffer.cc:183: }
On 2014/10/21 23:28:51, piman (Very slow to review) wrote:
> nit: you don't need this trivial wrapper any more, you can just fold into
> InProcessCommandBuffer::WaitSyncPointOnGpuThread

Done.

[no sievers, 2014-10-22 15:59:40]

The CQ bit was checked by sievers@chromium.org

[commit-bot: I haz the power, 2014-10-22 16:00:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/661973003/140001

[no sievers, 2014-10-22 16:26:34]

The CQ bit was checked by sievers@chromium.org

[commit-bot: I haz the power, 2014-10-22 16:28:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/661973003/160001

[commit-bot: I haz the power, 2014-10-22 18:19:41]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2014-10-22 18:21:01]

Patchset 9 (id:??) landed as
https://crrev.com/173a20dabc8f32182b48e999bd16575598ba4a44
Cr-Commit-Position: refs/heads/master@{#300709}