cc: Update main_thread_missed_last_deadline in aborted draws.

cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/03f72e193087882c04ee5f4bff819e30279938c8
Cr-Commit-Position: refs/heads/master@{#376609}

[brianderson, 2016-02-19 19:28:04]

Description was changed from

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=582749
==========

to

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[brianderson, 2016-02-19 20:26:15]

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

[brianderson, 2016-02-19 20:26:16]

ptal, split out from https://codereview.chromium.org/1681393003/#ps100001

[brianderson, 2016-02-19 23:36:00]

Description was changed from

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[sunnyps, 2016-02-20 00:27:00]

Is there any special code for handling the aborted draw case? I only see an
update to main_thread_missed_last_deadline_ in WillDraw, is that sufficient?

https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
File cc/scheduler/scheduler_state_machine.cc (left):

https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
cc/scheduler/scheduler_state_machine.cc:616: active_tree_needs_first_draw_ =
true;
Why was this removed?

[brianderson, 2016-02-20 00:43:23]

On 2016/02/20 00:27:00, sunnyps wrote:
> Is there any special code for handling the aborted draw case? I only see an
> update to main_thread_missed_last_deadline_ in WillDraw, is that sufficient?

WillDrawInternal gets called by regular draws and also by
SchedulerStateMachine::AbortDrawAndSwap.

I think I could get this to work today by only updating
main_thread_missed_last_deadline_ from SchedulerStateMachine::AbortDrawAndSwap,
but doing it this way works for all cases where we might draw outside of the
deadline in the future. For example, if deadline wasn't a concept for WebView or
if we try to draw ASAP during reszie.

> 
>
https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
> File cc/scheduler/scheduler_state_machine.cc (left):
> 
>
https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
> cc/scheduler/scheduler_state_machine.cc:616: active_tree_needs_first_draw_ =
> true;
> Why was this removed?

Oops. This should really be a part of https://codereview.chromium.org/1681393003
instead. I'll move it there.

This line is extraneous and not needed for correct operation of forced draws.
All it does is trick the UMA stats into thinking there's a new active tree when
there isn't. It throws off expectations like "we must have seen an activation
before a draw with a new active tree".

[brianderson, 2016-02-20 00:49:46]

https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
File cc/scheduler/scheduler_state_machine.cc (left):

https://codereview.chromium.org/1715863002/diff/1/cc/scheduler/scheduler_stat...
cc/scheduler/scheduler_state_machine.cc:616: active_tree_needs_first_draw_ =
true;
On 2016/02/20 00:27:00, sunnyps wrote:
> Why was this removed?

Added it back in for this patch.

[sunnyps, 2016-02-20 01:19:12]

LGTM

[brianderson, 2016-02-20 01:25:41]

The CQ bit was checked by brianderson@chromium.org

[commit-bot: I haz the power, 2016-02-20 01:27:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1715863002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1715863002/20001

[commit-bot: I haz the power, 2016-02-20 02:38:26]

Description was changed from

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[commit-bot: I haz the power, 2016-02-20 02:38:27]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-02-20 02:40:35]

Description was changed from

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
cc: Update main_thread_missed_last_deadline in aborted draws.

This prevents the scheduler from trying to recover
latency by skipping a BeginMainFrame immediately after
CanDraw goes from false to true, which occurs
during changes in viewport size.

It also makes recording the main_thread_missed_last_deadline
match an upcoming metric that records the the frame time
delta between the main-side and impl-side updates.

BUG=588318, 582749
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/03f72e193087882c04ee5f4bff819e30279938c8
Cr-Commit-Position: refs/heads/master@{#376609}
==========

[commit-bot: I haz the power, 2016-02-20 02:40:36]

Patchset 2 (id:??) landed as
https://crrev.com/03f72e193087882c04ee5f4bff819e30279938c8
Cr-Commit-Position: refs/heads/master@{#376609}