content: Make sure all CreateGpuMemoryBuffer messages are handled.

content: Make sure all CreateGpuMemoryBuffer messages are handled.

Add a startup message filter for CreateGpuMemoryBuffer messages
that is installed before establishing the channel. This guarantees
that all CreateGpuMemoryBuffer messages are handled on the IO
thread.

BUG=504121
TEST=content/test/gpu/run_gpu_test.py context_lost --extra-browser-args=--enable-native-gpu-memory-buffers --story-filter=GpuCrash.GPUProcessCrashesExactlyOnce

Committed: https://crrev.com/611d024f699b95474c26b6a3d91db13d5626d75a
Cr-Commit-Position: refs/heads/master@{#336284}

[reveman, 2015-06-24 23:26:32]

reveman@chromium.org changed reviewers:
+ rsesek@chromium.org

[reveman, 2015-06-25 04:38:51]

reveman@chromium.org changed reviewers:
+ alexst@chromium.org, piman@chromium.org
- rsesek@chromium.org

[reveman, 2015-06-25 04:38:52]

-rsesek
+piman
+alexst

[piman, 2015-06-25 20:29:00]

https://codereview.chromium.org/1210703002/diff/40001/content/gpu/gpu_child_t...
File content/gpu/gpu_child_thread.cc (right):

https://codereview.chromium.org/1210703002/diff/40001/content/gpu/gpu_child_t...
content/gpu/gpu_child_thread.cc:74: g_thread_safe_sender.Get()->Send(new
GpuHostMsg_GpuMemoryBufferCreated(
g_thread_safe_sender causes an extra thread hop (will post a task even when
you're already on the IO thread). Can we just keep a pointer to the Sender in
OnFilterAdded (and drop in OnFilterRemoved)? That's the common pattern for
filters.

[reveman, 2015-06-25 20:45:25]

ptal

https://codereview.chromium.org/1210703002/diff/40001/content/gpu/gpu_child_t...
File content/gpu/gpu_child_thread.cc (right):

https://codereview.chromium.org/1210703002/diff/40001/content/gpu/gpu_child_t...
content/gpu/gpu_child_thread.cc:74: g_thread_safe_sender.Get()->Send(new
GpuHostMsg_GpuMemoryBufferCreated(
On 2015/06/25 at 20:29:00, piman (Very slow to review) wrote:
> g_thread_safe_sender causes an extra thread hop (will post a task even when
you're already on the IO thread). Can we just keep a pointer to the Sender in
OnFilterAdded (and drop in OnFilterRemoved)? That's the common pattern for
filters.

Done.

[piman, 2015-06-25 22:09:35]

lgtm

[reveman, 2015-06-25 22:12:32]

The CQ bit was checked by reveman@chromium.org

[commit-bot: I haz the power, 2015-06-25 22:13:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1210703002/60001

[commit-bot: I haz the power, 2015-06-25 22:42:48]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-06-25 22:44:48]

Patchset 4 (id:??) landed as
https://crrev.com/611d024f699b95474c26b6a3d91db13d5626d75a
Cr-Commit-Position: refs/heads/master@{#336284}

[vignatti (out of this project), 2015-06-26 17:52:41]

reveman, this CL breaks ChildThreadImplGpuMemoryBufferBrowserTests*
content_browsertests. It blocks forever somewhere when going to
BrowserGpuMemoryBufferManager::AllocateGpuMemoryBufferForSurface().

[reveman, 2015-06-26 17:56:35]

On 2015/06/26 at 17:52:41, tiago.vignatti wrote:
> reveman, this CL breaks ChildThreadImplGpuMemoryBufferBrowserTests*
content_browsertests. It blocks forever somewhere when going to
BrowserGpuMemoryBufferManager::AllocateGpuMemoryBufferForSurface().

Thanks for the the heads up. I'll take a look.