Use ThreadData::Now() for TrackingInfo::time_posted initialization

Use ThreadData::Now() for TrackingInfo::time_posted initialization

TrackingInfo::time_posted is only used for task profiling, however, it
is always initialized with TimeTicks::Now(). As TimeTicks::Now() can be
relatively expensive, particularly on Android, instead use
ThreadData::Now(), effectively a no-op when task profiling is disabled.

With task profiling disabled by default on Android, this saves ~1-2us
per task, or ~40% of the fixed cost of task posting.

BUG=417156, 405348
Committed: https://crrev.com/71fc9b28eff22bfe923f1ba08bbe0e62a82f7139
Cr-Commit-Position: refs/heads/master@{#310853}

[jdduke (slow), 2015-01-09 16:40:33]

jdduke@chromium.org changed reviewers:
+ danakj@chromium.org, jamesr@chromium.org

[jdduke (slow), 2015-01-09 16:40:34]

jamesr@, danakj@: PTAL. As far as I can tell, |time_posted| can safely be null
when task profiling is disabled.

[jdduke (slow), 2015-01-09 16:45:46]

jdduke@chromium.org changed reviewers:
+ skyostil@chromium.org

[jdduke (slow), 2015-01-09 16:45:46]

+skyostil@ for content/renderer/scheduler.

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
File content/renderer/scheduler/task_queue_manager.cc (right):

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
content/renderer/scheduler/task_queue_manager.cc:235: (task.time_posted -
tracked_objects::TrackedTime()).InMilliseconds());
Sami, should we consider removing this entry entirely?

[Sami, 2015-01-09 16:48:32]

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
File content/renderer/scheduler/task_queue_manager.cc (right):

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
content/renderer/scheduler/task_queue_manager.cc:235: (task.time_posted -
tracked_objects::TrackedTime()).InMilliseconds());
On 2015/01/09 16:45:46, jdduke wrote:
> Sami, should we consider removing this entry entirely?

Yes, kill it. We can reconstruct it from trace events anyway.

[jdduke (slow), 2015-01-09 16:54:22]

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
File content/renderer/scheduler/task_queue_manager.cc (right):

https://codereview.chromium.org/846663002/diff/1/content/renderer/scheduler/t...
content/renderer/scheduler/task_queue_manager.cc:235: (task.time_posted -
tracked_objects::TrackedTime()).InMilliseconds());
On 2015/01/09 16:48:32, Sami wrote:
> On 2015/01/09 16:45:46, jdduke wrote:
> > Sami, should we consider removing this entry entirely?
> 
> Yes, kill it. We can reconstruct it from trace events anyway.

Done.

[danakj, 2015-01-09 17:16:22]

LGTM % jamesr

[Sami, 2015-01-09 17:17:29]

content/renderer/scheduler/task_queue_manager.cc lgtm.

[jamesr, 2015-01-09 19:17:29]

lgtm

nice!!

[jdduke (slow), 2015-01-09 19:53:44]

The CQ bit was checked by jdduke@chromium.org

[commit-bot: I haz the power, 2015-01-09 19:54:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/846663002/40001

[commit-bot: I haz the power, 2015-01-09 21:02:08]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-01-09 21:03:00]

Patchset 3 (id:??) landed as
https://crrev.com/71fc9b28eff22bfe923f1ba08bbe0e62a82f7139
Cr-Commit-Position: refs/heads/master@{#310853}