Fix Mac resize, delete more Mac code

Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/817a0d6803e28e21b260f7c16025bf174e52f395
Cr-Commit-Position: refs/heads/master@{#393032}

[ccameron, 2016-05-10 20:42:54]

Description was changed from

==========
Fix Mac resize, delete more Mac code

BUG=610596
==========

to

==========
Fix Mac resize, delete more Mac code

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[ccameron, 2016-05-10 20:53:55]

Description was changed from

==========
Fix Mac resize, delete more Mac code

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
==========

[ccameron, 2016-05-10 20:53:55]

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

[ccameron, 2016-05-10 20:58:52]

On today's topic of Mac resize -- the previous patch forgot to forward the new
swap completed message to the nested message loop runner.

This also continues removing now-unused code (probably should have been done in
the previous patch).

https://codereview.chromium.org/1963263002/diff/20001/content/browser/composi...
File content/browser/compositor/gpu_browser_compositor_output_surface.cc
(right):

https://codereview.chromium.org/1963263002/diff/20001/content/browser/composi...
content/browser/compositor/gpu_browser_compositor_output_surface.cc:130: if
(should_show_frames_state_ == SHOULD_SHOW_FRAMES) {
This is a change in behavior -- the previous patch incorrectly called "not
suspended for recycle" not "should draw frames".

[ccameron, 2016-05-10 22:01:28]

Description was changed from

==========
Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
==========

to

==========
Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[piman, 2016-05-10 22:24:16]

https://codereview.chromium.org/1963263002/diff/60001/gpu/ipc/client/gpu_chan...
File gpu/ipc/client/gpu_channel_host.cc (right):

https://codereview.chromium.org/1963263002/diff/60001/gpu/ipc/client/gpu_chan...
gpu/ipc/client/gpu_channel_host.cc:459: task_runner =
ui::WindowResizeHelperMac::Get()->task_runner();
This is called in the renderer - e.g. for pepper. Does this make sense there?
Should we inject the task runner instead when creating the UI context, to avoid
this extra odd dependency?

[ccameron, 2016-05-11 06:23:01]

Good point.

Moved the OS-specific logic to
GpuProcessTransportFactory::EstablishedGpuChannel. Added
ContextProviderCommandBuffer::SetDefaultTaskRunner to handle setting the default
runner. I was considering adding an argument to BindToCurrentThread, but that
seemed cumbersome, as only one caller would need the argument (adding a separate
BindToCurrentThreadWithTaskRunner was also an idea, if also touching lots of
files).

https://codereview.chromium.org/1963263002/diff/60001/gpu/ipc/client/gpu_chan...
File gpu/ipc/client/gpu_channel_host.cc (right):

https://codereview.chromium.org/1963263002/diff/60001/gpu/ipc/client/gpu_chan...
gpu/ipc/client/gpu_channel_host.cc:459: task_runner =
ui::WindowResizeHelperMac::Get()->task_runner();
On 2016/05/10 22:24:16, piman wrote:
> This is called in the renderer - e.g. for pepper. Does this make sense there?
> Should we inject the task runner instead when creating the UI context, to
avoid
> this extra odd dependency?

Good point -- moved the #if MAC parts into
GpuProcessTransportFactory::EstablishedGpuChannel (which has lots of OS-specific
goo).

[piman, 2016-05-11 17:06:15]

LGTM+nits. Thanks!

https://codereview.chromium.org/1963263002/diff/30026/content/common/gpu/clie...
File content/common/gpu/client/context_provider_command_buffer.cc (right):

https://codereview.chromium.org/1963263002/diff/30026/content/common/gpu/clie...
content/common/gpu/client/context_provider_command_buffer.cc:153: active_url_,
gpu_preference_, task_runner);
nit: std::move(task_runner)

https://codereview.chromium.org/1963263002/diff/30026/content/common/gpu/clie...
content/common/gpu/client/context_provider_command_buffer.cc:292:
default_task_runner_ = default_task_runner;
nit: std::move(default_task_runner)

https://codereview.chromium.org/1963263002/diff/30026/gpu/ipc/client/gpu_chan...
File gpu/ipc/client/gpu_channel_host.cc (right):

https://codereview.chromium.org/1963263002/diff/30026/gpu/ipc/client/gpu_chan...
gpu/ipc/client/gpu_channel_host.cc:241: AddRouteWithTaskRunner(route_id,
command_buffer->AsWeakPtr(), task_runner);
nit: std::move(task_runner)

[ccameron, 2016-05-11 17:59:47]

Thanks!

[ccameron, 2016-05-11 17:59:52]

The CQ bit was checked by ccameron@chromium.org

[ccameron, 2016-05-11 17:59:53]

The patchset sent to the CQ was uploaded after l-g-t-m from piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1963263002/#ps110001
(title: "Rebase")

[commit-bot: I haz the power, 2016-05-11 18:00:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1963263002/110001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1963263002/110001

[commit-bot: I haz the power, 2016-05-11 19:53:06]

Committed patchset #7 (id:110001)

[commit-bot: I haz the power, 2016-05-11 19:54:56]

Description was changed from

==========
Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Fix Mac resize, delete more Mac code

Delete content::RenderWidgetResizeHelper. Previously this had two
callers, both in content (one for GPU IPCs and one for renderer IPCs).
Now this has one caller in content and one in gpu/ipc/client, so
change both of these callers to talk to ui::WindowResizeHelperMac.

Update GpuChannelHost to forward the IPC
GpuCommandBufferMsg_SwapBuffersCompleted to the task runner
in ui::WindowResizeHelperMac, so that new frames can be accepted
while resizing. This was missed in the previous cleanup patch.

Remove SurfaceShouldNotShowFramesAfterSuspendForRecycle
because it is no longer needed (it can be accessed directly). Fix
its use in GpuBrowserCompositorOutputSurface's
OnGpuSwapBuffersCompleted.

This patch also accidentally pulled in an extra git cl format, so a few
whitespace fixes come in here and there as well.

BUG=610596
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/817a0d6803e28e21b260f7c16025bf174e52f395
Cr-Commit-Position: refs/heads/master@{#393032}
==========

[commit-bot: I haz the power, 2016-05-11 19:54:57]

Patchset 7 (id:??) landed as
https://crrev.com/817a0d6803e28e21b260f7c16025bf174e52f395
Cr-Commit-Position: refs/heads/master@{#393032}