OOPIF: Set timer for per-frame UpdateState.

OOPIF: Set timer for per-frame UpdateState.

We continue to use a per-page timer on RenderView, but it keeps track of
which frames have pending state updates to send.  In most cases, only a
single frame will need to send an update.  If there are multiple such
frames, each one will send an UpdateState IPC when the timer fires.

BUG=545219
TEST=Scroll and form state restored in OOPIFs after close
     (once OOPIF session restore is fixed)

Committed: https://crrev.com/69b184a47779e5edb02437d77398c7d4cf35ec2f
Cr-Commit-Position: refs/heads/master@{#361821}

[Charlie Reis, 2015-11-25 20:22:11]

creis@chromium.org changed reviewers:
+ avi@chromium.org

[Charlie Reis, 2015-11-25 20:22:12]

Avi, can you take a look?  This approach has the nice property that it won't
send UpdateState IPCs for frames that haven't changed, but we can still keep a
single timer for the page.

https://codereview.chromium.org/1449353002/diff/60001/content/renderer/render...
File content/renderer/render_view_browsertest.cc (left):

https://codereview.chromium.org/1449353002/diff/60001/content/renderer/render...
content/renderer/render_view_browsertest.cc:455: TEST_F(RenderViewImplTest,
DISABLED_OnNavStateChanged) {
I'm thinking I'll re-enable the test first in
https://codereview.chromium.org/1474003002/ (in case it's flaky) then land this
CL with updates to it.

[Avi (use Gerrit), 2015-11-25 20:37:11]

lgtm

This looks like a nice simple step forward. Is there a reason why you didn't
want to go all-out and give the timer to the frame?

[Charlie Reis, 2015-11-25 21:21:17]

On 2015/11/25 20:37:11, Avi wrote:
> lgtm
> 
> This looks like a nice simple step forward. Is there a reason why you didn't
> want to go all-out and give the timer to the frame?

Having a single timer per page is nice because there's fewer wake-ups (as
opposed to one timer for each frame), which tends to be better for battery, etc.
 In other words, if multiple frames need to send updates, it's better to do the
IPCs in the same event.

I'm thinking it will make sense to move the timer to a per-page class once we
introduce one, allowing us to kill the timer in RenderView.  That will be easier
once we turn on UseSubframeNavigationEntries by default and RenderView no longer
needs its timer for default mode.

[Charlie Reis, 2015-11-26 00:49:42]

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

[commit-bot: I haz the power, 2015-11-26 00:52:01]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1449353002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1449353002/80001

[commit-bot: I haz the power, 2015-11-26 02:55:18]

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

[commit-bot: I haz the power, 2015-11-26 02:55:19]

Dry run: Try jobs failed on following builders:
  ios_dbg_simulator_ninja on tryserver.chromium.mac (JOB_TIMED_OUT, no build
URL)
  ios_rel_device_ninja on tryserver.chromium.mac (JOB_TIMED_OUT, no build URL)

[Charlie Reis, 2015-11-26 05:24:54]

The CQ bit was checked by creis@chromium.org

[Charlie Reis, 2015-11-26 05:24:55]

The patchset sent to the CQ was uploaded after l-g-t-m from avi@chromium.org
Link to the patchset: https://codereview.chromium.org/1449353002/#ps80001
(title: "Rebase")

[commit-bot: I haz the power, 2015-11-26 05:25:10]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1449353002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1449353002/80001

[commit-bot: I haz the power, 2015-11-26 05:30:02]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-11-26 05:31:10]

Description was changed from

==========
OOPIF: Set timer for per-frame UpdateState.

We continue to use a per-page timer on RenderView, but it keeps track of
which frames have pending state updates to send.  In most cases, only a
single frame will need to send an update.  If there are multiple such
frames, each one will send an UpdateState IPC when the timer fires.

BUG=545219
TEST=Scroll and form state restored in OOPIFs after close
     (once OOPIF session restore is fixed)
==========

to

==========
OOPIF: Set timer for per-frame UpdateState.

We continue to use a per-page timer on RenderView, but it keeps track of
which frames have pending state updates to send.  In most cases, only a
single frame will need to send an update.  If there are multiple such
frames, each one will send an UpdateState IPC when the timer fires.

BUG=545219
TEST=Scroll and form state restored in OOPIFs after close
     (once OOPIF session restore is fixed)

Committed: https://crrev.com/69b184a47779e5edb02437d77398c7d4cf35ec2f
Cr-Commit-Position: refs/heads/master@{#361821}
==========

[commit-bot: I haz the power, 2015-11-26 05:31:11]

Patchset 5 (id:??) landed as
https://crrev.com/69b184a47779e5edb02437d77398c7d4cf35ec2f
Cr-Commit-Position: refs/heads/master@{#361821}