[cc]: Add a BeginMainFrameNotExpectedSoon signal and route it to the RendererScheduler.

[cc]: Add a BeginMainFrameNotExpectedSoon signal and route it to the RendererScheduler.

In order to implement long idle times, the RendererScheduler needs to
know when the compositor has stoped drawing frames. To enable this, add
a BeginMainFrameNotExpectedSoon signal which is sent when the cc
scheduler stops requesting BeginFrames from the Browser.

Design Doc: https://docs.google.com/a/chromium.org/document/d/1yBlUdYW8VTIfB-DqhvQqUeP0kf-Ap1W4cao2yQq58Do/edit

BUG=455713
TBR=danakj@chromium.org, torne@chromium.org, aa@chromium.org

Committed: https://crrev.com/0a19362adbaa53af4fb7ccb5819f482b30ca75af
Cr-Commit-Position: refs/heads/master@{#316804}

[rmcilroy, 2015-02-05 14:37:15]

rmcilroy@chromium.org changed reviewers:
+ brianderson@chromium.org, skyostil@chromium.org

[rmcilroy, 2015-02-05 14:37:16]

brianderson@ / skyostil@ please take a look.  

I originally involved the state machine in the decision about whether to send
the BeginFrameNoLongerRequired signal, however it turned out that doing it when
we stop requesting BeginFrames from the BeginFrameSource was a much simpler
method of doing the same thing.  PTAL, thanks.

[brianderson, 2015-02-11 01:58:11]

This looks really good.

We can add optimizations later for the corner cases where the impl thread is
active and we don't expect BeginMainFrames, but this will already cover most of
the important use cases.

Do we want to bikeshed on the name? BeginMainFrameNoLongerRequired seems so
final. How about BeginMainFrameNotExpectedSoon?

In any case, lgtm.

https://codereview.chromium.org/904633004/diff/1/cc/trees/single_thread_proxy.h
File cc/trees/single_thread_proxy.h (right):

https://codereview.chromium.org/904633004/diff/1/cc/trees/single_thread_proxy...
cc/trees/single_thread_proxy.h:127: void BeginMainFrameNoLongerRequired();
Doesn't look like this is actually implemented.

[rmcilroy, 2015-02-11 12:01:18]

rmcilroy@chromium.org changed reviewers:
+ danakj@chromium.org, jochen@chromium.org

[rmcilroy, 2015-02-11 12:01:19]

Adding reviewers for OWNERS:

  danakj@: Please review changes in ui/compositor
  jochen@: Please review changes in content/renderer/gpu

Sami: could you also please task a look at the changes in
content/renderer/scheduler.

> Do we want to bikeshed on the name? 
> BeginMainFrameNoLongerRequired seems so
> final. How about BeginMainFrameNotExpectedSoon?

Good call - renamed to BeginMainFrameNotExpectedSoon.

https://codereview.chromium.org/904633004/diff/1/cc/trees/single_thread_proxy.h
File cc/trees/single_thread_proxy.h (right):

https://codereview.chromium.org/904633004/diff/1/cc/trees/single_thread_proxy...
cc/trees/single_thread_proxy.h:127: void BeginMainFrameNoLongerRequired();
On 2015/02/11 01:58:11, brianderson wrote:
> Doesn't look like this is actually implemented.

Good catch, thanks. Done.

[rmcilroy, 2015-02-11 12:03:25]

New patchsets have been uploaded after l-g-t-m from brianderson@chromium.org

[Sami, 2015-02-11 15:28:25]

Overall looks great. I've added a couple suggestions.

https://codereview.chromium.org/904633004/diff/20001/cc/scheduler/scheduler.h
File cc/scheduler/scheduler.h (right):

https://codereview.chromium.org/904633004/diff/20001/cc/scheduler/scheduler.h...
cc/scheduler/scheduler.h:52: virtual void SendBeginMainFrameNotExpectedSoon() =
0;
Could you add a scheduler unit test for this? Just a basic one for now -- we can
add the tricky corner cases later as we run into them :)

https://codereview.chromium.org/904633004/diff/20001/cc/trees/thread_proxy.cc
File cc/trees/thread_proxy.cc (right):

https://codereview.chromium.org/904633004/diff/20001/cc/trees/thread_proxy.cc...
cc/trees/thread_proxy.cc:865: TRACE_EVENT0("cc",
"ThreadProxy::BeginMainFrameNotExpectedSoon");
DCHECK(IsMainThread());

https://codereview.chromium.org/904633004/diff/20001/content/renderer/schedul...
File content/renderer/scheduler/renderer_scheduler.h (right):

https://codereview.chromium.org/904633004/diff/20001/content/renderer/schedul...
content/renderer/scheduler/renderer_scheduler.h:40: // need to be drawn.
Could you add a note saying this needs to be called from the main thread (and a
DCHECK on the .cc side)?

https://codereview.chromium.org/904633004/diff/20001/content/renderer/schedul...
content/renderer/scheduler/renderer_scheduler.h:40: // need to be drawn.
Could you add a note saying this needs to be called from the main thread (and a
DCHECK on the .cc side)?

[jochen (gone - plz use gerrit), 2015-02-12 08:33:08]

content/renderer/gpu lgtm

[rmcilroy, 2015-02-13 16:30:43]

New patchsets have been uploaded after l-g-t-m from jochen@chromium.org

[rmcilroy, 2015-02-13 16:32:46]

Comments addressed from Sami - thanks for the review.

https://codereview.chromium.org/904633004/diff/20001/cc/scheduler/scheduler.h
File cc/scheduler/scheduler.h (right):

https://codereview.chromium.org/904633004/diff/20001/cc/scheduler/scheduler.h...
cc/scheduler/scheduler.h:52: virtual void SendBeginMainFrameNotExpectedSoon() =
0;
On 2015/02/11 15:28:24, Sami wrote:
> Could you add a scheduler unit test for this? Just a basic one for now -- we
can
> add the tricky corner cases later as we run into them :)

Done.  Also update all the other tests to expect these signals where
appropriate.

https://codereview.chromium.org/904633004/diff/20001/cc/trees/thread_proxy.cc
File cc/trees/thread_proxy.cc (right):

https://codereview.chromium.org/904633004/diff/20001/cc/trees/thread_proxy.cc...
cc/trees/thread_proxy.cc:865: TRACE_EVENT0("cc",
"ThreadProxy::BeginMainFrameNotExpectedSoon");
On 2015/02/11 15:28:25, Sami wrote:
> DCHECK(IsMainThread());

Done.

https://codereview.chromium.org/904633004/diff/20001/content/renderer/schedul...
File content/renderer/scheduler/renderer_scheduler.h (right):

https://codereview.chromium.org/904633004/diff/20001/content/renderer/schedul...
content/renderer/scheduler/renderer_scheduler.h:40: // need to be drawn.
On 2015/02/11 15:28:25, Sami wrote:
> Could you add a note saying this needs to be called from the main thread (and
a
> DCHECK on the .cc side)?

Done.

[rmcilroy, 2015-02-17 16:16:44]

Ping?  skyostil@ / danakj@ any other comments?

[Sami, 2015-02-17 16:33:16]

Sorry, missed this one. cc/ and content/renderer/scheduler lgtm.

[rmcilroy, 2015-02-18 09:34:55]

The CQ bit was checked by rmcilroy@chromium.org

[commit-bot: I haz the power, 2015-02-18 09:35:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/904633004/60001

[commit-bot: I haz the power, 2015-02-18 09:48:05]

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

[commit-bot: I haz the power, 2015-02-18 09:48:06]

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

[rmcilroy, 2015-02-18 10:07:52]

New patchsets have been uploaded after l-g-t-m from skyostil@chromium.org

[rmcilroy, 2015-02-18 10:12:13]

rmcilroy@chromium.org changed reviewers:
+ aa@chromium.org, torne@chromium.org

[rmcilroy, 2015-02-18 10:12:14]

TBRing Aaron, Torne and Dana for nop implementations of
BeginMainFrameNotExpectedSoon on Mojo HTML Viewer, Webview and compositor
layer_tree_host_client implementations.

[rmcilroy, 2015-02-18 10:12:50]

The CQ bit was checked by rmcilroy@chromium.org

[commit-bot: I haz the power, 2015-02-18 10:13:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/904633004/80001

[Torne, 2015-02-18 10:15:36]

android_webview LGTM

[commit-bot: I haz the power, 2015-02-18 12:34:38]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-02-18 12:35:19]

Patchset 5 (id:??) landed as
https://crrev.com/0a19362adbaa53af4fb7ccb5819f482b30ca75af
Cr-Commit-Position: refs/heads/master@{#316804}