cc: Get rid of PostSwapBuffersComplete.

cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973, 606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/daad1d10fb3932de760bdc3688e96c947d3c7b18
Cr-Commit-Position: refs/heads/master@{#424294}

[danakj, 2016-10-08 01:19:33]

Description was changed from

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
==========

to

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[danakj, 2016-10-08 01:20:02]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[danakj, 2016-10-08 01:20:35]

Description was changed from

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[danakj, 2016-10-08 01:20:35]

danakj@chromium.org changed reviewers:
+ boliu@chromium.org

[commit-bot: I haz the power, 2016-10-08 01:20:37]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-08 01:21:10]

enne: overall
khushal: blimp/
boliu: content/..android*

[commit-bot: I haz the power, 2016-10-08 01:23:33]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-08 01:23:35]

Dry run: Try jobs failed on following builders:
  mac_chromium_compile_dbg_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_comp...)

[danakj, 2016-10-08 01:30:15]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-08 01:30:32]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-08 01:48:00]

Patchset #2 (id:20001) has been deleted

[danakj, 2016-10-08 01:48:05]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-08 01:48:20]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-08 01:51:10]

danakj@chromium.org changed reviewers:
+ sunnyps@chromium.org

[danakj, 2016-10-08 01:51:11]

+sunnyps for cc/ stuff too as it's overlapping stuff he did recently.

[danakj, 2016-10-08 01:59:20]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-08 01:59:38]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-08 02:05:44]

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
File blimp/client/core/compositor/blimp_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
blimp/client/core/compositor/blimp_compositor.cc:253: DCHECK(surface_factory_);
I cargo-culted this, but is it value?

[commit-bot: I haz the power, 2016-10-08 02:22:54]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-08 02:22:55]

Dry run: Try jobs failed on following builders:
  cast_shell_linux on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[boliu, 2016-10-09 22:55:08]

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
File content/browser/compositor/vulkan_browser_compositor_output_surface.cc
(left):

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
content/browser/compositor/vulkan_browser_compositor_output_surface.cc:57:
PostSwapBuffersComplete();
was the previous thing calling DidSwapBuffersComplete twice per frame? by
mistake?

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
File content/browser/compositor/vulkan_browser_compositor_output_surface.h
(right):

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
content/browser/compositor/vulkan_browser_compositor_output_surface.h:44: void
SwapBuffersCallback();
not needed?

[Khushal, 2016-10-10 17:40:10]

lgtm. Thanks!

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
File blimp/client/core/compositor/blimp_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
blimp/client/core/compositor/blimp_compositor.cc:253: DCHECK(surface_factory_);
On 2016/10/08 02:05:44, danakj wrote:
> I cargo-culted this, but is it value?

The aim was to assert that these calls should come only when we are bound to a
CompositorFrameSink, which is what the lifetime of SurfaceFactory is tied to.

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/support/co...
File blimp/client/support/compositor/blimp_embedder_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/support/co...
blimp/client/support/compositor/blimp_embedder_compositor.cc:63:
task_runner_->PostTask(
Comment here that acks for submitted frames must be done asynchronously?

[danakj, 2016-10-10 20:28:20]

PTAL

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/support/co...
File blimp/client/support/compositor/blimp_embedder_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/support/co...
blimp/client/support/compositor/blimp_embedder_compositor.cc:63:
task_runner_->PostTask(
On 2016/10/10 17:40:10, Khushal wrote:
> Comment here that acks for submitted frames must be done asynchronously?

Done.

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
File content/browser/compositor/vulkan_browser_compositor_output_surface.cc
(left):

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
content/browser/compositor/vulkan_browser_compositor_output_surface.cc:57:
PostSwapBuffersComplete();
On 2016/10/09 22:55:08, boliu wrote:
> was the previous thing calling DidSwapBuffersComplete twice per frame? by
> mistake?

I think so but this was never seriously run yet, it may be that
OnGpuSwapBuffersCompleted isn't hooked up, but we'll find that out quickly
if/when we use this class again.

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
File content/browser/compositor/vulkan_browser_compositor_output_surface.h
(right):

https://codereview.chromium.org/2402173002/diff/80001/content/browser/composi...
content/browser/compositor/vulkan_browser_compositor_output_surface.h:44: void
SwapBuffersCallback();
On 2016/10/09 22:55:08, boliu wrote:
> not needed?

Done.

[boliu, 2016-10-10 20:31:28]

content/browser lgtm

[danakj, 2016-10-10 20:34:04]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-10 20:34:35]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-10 20:34:53]

danakj@chromium.org changed reviewers:
+ dtrainor@chromium.org

[danakj, 2016-10-10 20:34:54]

On 2016/10/10 17:40:10, Khushal wrote:
> lgtm. Thanks!

+dtrainor for blimp owners

[enne (OOO), 2016-10-10 20:42:19]

enne@chromium.org changed reviewers:
- dtrainor@chromium.org

[enne (OOO), 2016-10-10 20:42:20]

lgtm, this seems like nice cleanup

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
File blimp/client/core/compositor/blimp_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
blimp/client/core/compositor/blimp_compositor.cc:253: DCHECK(surface_factory_);
On 2016/10/08 at 02:05:44, danakj wrote:
> I cargo-culted this, but is it value?

I'm not sure what the point of this is.  If this method just posts a task, then
this DCHECK doesn't really say anything about the state of surface factory when
that task runs.

[danakj, 2016-10-10 20:46:36]

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
File blimp/client/core/compositor/blimp_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
blimp/client/core/compositor/blimp_compositor.cc:253: DCHECK(surface_factory_);
On 2016/10/10 20:42:20, enne wrote:
> On 2016/10/08 at 02:05:44, danakj wrote:
> > I cargo-culted this, but is it value?
> 
> I'm not sure what the point of this is.  If this method just posts a task,
then
> this DCHECK doesn't really say anything about the state of surface factory
when
> that task runs.

I think DCHECKing proxy_client_ is not null would give the same information but
make it more clear what/why it's checking. WDYT?

[danakj, 2016-10-10 20:48:13]

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
File blimp/client/core/compositor/blimp_compositor.cc (right):

https://codereview.chromium.org/2402173002/diff/80001/blimp/client/core/compo...
blimp/client/core/compositor/blimp_compositor.cc:253: DCHECK(surface_factory_);
On 2016/10/10 20:46:36, danakj wrote:
> On 2016/10/10 20:42:20, enne wrote:
> > On 2016/10/08 at 02:05:44, danakj wrote:
> > > I cargo-culted this, but is it value?
> > 
> > I'm not sure what the point of this is.  If this method just posts a task,
> then
> > this DCHECK doesn't really say anything about the state of surface factory
> when
> > that task runs.
> 
> I think DCHECKing proxy_client_ is not null would give the same information
but
> make it more clear what/why it's checking. WDYT?

No that seems incorrect actually, cuz proxy_client_ is a weak ptr so it can
become null suddenly.

[sunnyps, 2016-10-10 20:58:16]

lgtm % nits

https://codereview.chromium.org/2402173002/diff/100001/cc/output/compositor_f...
File cc/output/compositor_frame_sink.h (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/output/compositor_f...
cc/output/compositor_frame_sink.h:136: base::WeakPtrFactory<CompositorFrameSink>
weak_ptr_factory_;
nit: We don't this WeakPtrFactory any more.

https://codereview.chromium.org/2402173002/diff/100001/cc/output/output_surfa...
File cc/output/output_surface.h (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/output/output_surfa...
cc/output/output_surface.h:161: base::WeakPtrFactory<OutputSurface>
weak_ptr_factory_;
nit: Same here.

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
File cc/test/test_compositor_frame_sink.cc (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
cc/test/test_compositor_frame_sink.cc:140: // draws we don't need one.
nit: Make a note that weak ptr is not required for draw callback because the
callback is called before the surface is destroyed.

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
cc/test/test_compositor_frame_sink.cc:164: void
TestCompositorFrameSink::DidDrawCallback() {
nit: DCHECK(!display_->has_scheduler())?

[danakj, 2016-10-10 21:02:11]

https://codereview.chromium.org/2402173002/diff/100001/cc/output/compositor_f...
File cc/output/compositor_frame_sink.h (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/output/compositor_f...
cc/output/compositor_frame_sink.h:136: base::WeakPtrFactory<CompositorFrameSink>
weak_ptr_factory_;
On 2016/10/10 20:58:15, sunnyps wrote:
> nit: We don't this WeakPtrFactory any more.

Done. I just noticed this too :)

https://codereview.chromium.org/2402173002/diff/100001/cc/output/output_surfa...
File cc/output/output_surface.h (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/output/output_surfa...
cc/output/output_surface.h:159: void OnSwapBuffersComplete();
This too! Woops, deleted.

https://codereview.chromium.org/2402173002/diff/100001/cc/output/output_surfa...
cc/output/output_surface.h:161: base::WeakPtrFactory<OutputSurface>
weak_ptr_factory_;
On 2016/10/10 20:58:15, sunnyps wrote:
> nit: Same here.

Done.

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
File cc/test/test_compositor_frame_sink.cc (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
cc/test/test_compositor_frame_sink.cc:140: // draws we don't need one.
On 2016/10/10 20:58:15, sunnyps wrote:
> nit: Make a note that weak ptr is not required for draw callback because the
> callback is called before the surface is destroyed.

Done.

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
cc/test/test_compositor_frame_sink.cc:164: void
TestCompositorFrameSink::DidDrawCallback() {
On 2016/10/10 20:58:15, sunnyps wrote:
> nit: DCHECK(!display_->has_scheduler())?

Inverted but done. use callback => not synchronous => has scheduler

[danakj, 2016-10-10 21:10:10]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-10 21:10:39]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-10 21:15:04]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-10 21:15:40]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-10-10 21:32:26]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[danakj, 2016-10-10 21:32:29]

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
File cc/test/test_compositor_frame_sink.cc (right):

https://codereview.chromium.org/2402173002/diff/100001/cc/test/test_composito...
cc/test/test_compositor_frame_sink.cc:164: void
TestCompositorFrameSink::DidDrawCallback() {
On 2016/10/10 21:02:11, danakj wrote:
> On 2016/10/10 20:58:15, sunnyps wrote:
> > nit: DCHECK(!display_->has_scheduler())?
> 
> Inverted but done. use callback => not synchronous => has scheduler

Wait both paths to use this function but one posts directly and one gives it to
SubmitCompositorFrame, so can't DCHECK.

[commit-bot: I haz the power, 2016-10-10 21:32:42]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[David Trainor- moved to gerrit, 2016-10-10 22:22:29]

dtrainor@chromium.org changed reviewers:
+ dtrainor@chromium.org

[David Trainor- moved to gerrit, 2016-10-10 22:22:30]

rubberstamp lgtm

[danakj, 2016-10-10 22:24:59]

The CQ bit was unchecked by danakj@chromium.org

[danakj, 2016-10-10 22:25:06]

The CQ bit was checked by danakj@chromium.org

[danakj, 2016-10-10 22:25:06]

The patchset sent to the CQ was uploaded after l-g-t-m from
khushalsagar@chromium.org, sunnyps@chromium.org, enne@chromium.org,
boliu@chromium.org
Link to the patchset: https://codereview.chromium.org/2402173002/#ps180001
(title: "postswap: .")

[commit-bot: I haz the power, 2016-10-10 22:25:24]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-10 22:57:51]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-10 22:57:54]

Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[danakj, 2016-10-10 23:13:25]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[danakj, 2016-10-10 23:13:45]

The CQ bit was unchecked by danakj@chromium.org

[danakj, 2016-10-10 23:13:49]

The CQ bit was checked by danakj@chromium.org

[danakj, 2016-10-10 23:13:50]

The patchset sent to the CQ was uploaded after l-g-t-m from
sunnyps@chromium.org, enne@chromium.org, boliu@chromium.org,
dtrainor@chromium.org, khushalsagar@chromium.org
Link to the patchset: https://codereview.chromium.org/2402173002/#ps200001
(title: "postswap: .")

[commit-bot: I haz the power, 2016-10-10 23:13:51]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-11 00:00:57]

Description was changed from

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[commit-bot: I haz the power, 2016-10-11 00:00:59]

Committed patchset #10 (id:200001)

[commit-bot: I haz the power, 2016-10-11 00:02:39]

Description was changed from

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Get rid of PostSwapBuffersComplete.

OutputSurface and CompositorFrameSink both have this method which is
not part of the interface but used to post a call to the client's
DidSwapBuffersComplete to the current ThreadTaskRunnerHandle. This is
problematic since that is *not* the compositor task runner which can
be overridden, and it clutters these base class APIs.

This moves the calls to the client DidSwapBuffersComplete() out to each
implementation of OutputSurface or CompositorFrameSink.

Some additional changes in the process:

- BrowserCompositorOutputSurface had OnGpuSwapBuffersCompleted but it
was only used for gpu cases, so moved to GpuBrowserCompositorOutputSurface
and remove the NOTREACHED() version in
SoftwareBrowserCompositorOutputSurface.

- Blimp was not using the draw callback to provide backpressure for
submitted CompositorFrames, and just posting immediately, so hook that
up instead of doing a simple post-task dance there.

- SoftwareBrowserCompositorOutputSurface was using ThreadTaskRunnerHandle
to make another post task too but it should be using the compositor's
task runner. Since it now has it, use that.

R=enne@chromium.org, khushalsagar@chromium.org
BUG=616973,606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/daad1d10fb3932de760bdc3688e96c947d3c7b18
Cr-Commit-Position: refs/heads/master@{#424294}
==========

[commit-bot: I haz the power, 2016-10-11 00:02:40]

Patchset 10 (id:??) landed as
https://crrev.com/daad1d10fb3932de760bdc3688e96c947d3c7b18
Cr-Commit-Position: refs/heads/master@{#424294}