cc: RenderingStats overhaul

cc: RenderingStats overhaul
- Split RenderingStats into MainThreadRenderingStats and
  ImplThreadRenderingStats
- Support accumulated stats and per-frame stats in RenderingStatsInstrumentation
- Dump MainThreadRenderingStats into trace event from 
  ThreadProxy::BeginFrameOnMainThread
- Dump ImplThreadRenderingStats into trace event from
  LayerTreeHostImpl::DrawLayers
- Made naming of RenderingStats members consistent

R=nduca@chromium.org,jamesr@chromium.org
BUG=266101, 264308
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=217824

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=218149

[ernstm, 2013-08-13 17:16:26]

Nat, this CL has all the required changes in RenderingStats we need for
switching smoothness benchmark to trace events. The old API is left unchanged.
Maybe we can deprecate it once smoothness benchmark is updated?

James, could you please take a quick look at the content part of this CL?

[nduca, 2013-08-13 19:16:45]

lgtm

https://codereview.chromium.org/23088002/diff/1/cc/debug/rendering_stats_inst...
File cc/debug/rendering_stats_instrumentation.h (left):

https://codereview.chromium.org/23088002/diff/1/cc/debug/rendering_stats_inst...
cc/debug/rendering_stats_instrumentation.h:20: 
should this thing split in half too as a followup?

[jamesr, 2013-08-13 19:42:56]

content/ lgtm

[ernstm, 2013-08-13 21:10:19]

https://codereview.chromium.org/23088002/diff/1/cc/debug/rendering_stats_inst...
File cc/debug/rendering_stats_instrumentation.h (left):

https://codereview.chromium.org/23088002/diff/1/cc/debug/rendering_stats_inst...
cc/debug/rendering_stats_instrumentation.h:20: 
On 2013/08/13 19:16:46, nduca wrote:
> should this thing split in half too as a followup?

Probably yes. Filed crbug.com/272475.

[commit-bot: I haz the power, 2013-08-13 21:20:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ernstm@chromium.org/23088002/1

[commit-bot: I haz the power, 2013-08-14 00:09:52]

Retried try job too often on linux_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[ernstm, 2013-08-14 21:11:35]

On 2013/08/14 00:09:52, I haz the power (commit-bot) wrote:
> Retried try job too often on linux_rel for step(s) browser_tests
>
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

GpuFeatureTest.RafNoDamage fails, because TraceAnalyzer cannot handle Dictionary
arguments on trace events. It throws an error in trace_event_analyzer.cc:97 and
fails hard. Unless we can foresee a use case that requires TraceAnalyzer to
handle Dictionary arguments, I would suggest fixing this by ignoring the
Dictionary arguments (and issue a warning) instead of failing hard when
encountering one.

[ernstm, 2013-08-14 21:45:50]

On 2013/08/14 21:11:35, Manfred Ernst wrote:
> On 2013/08/14 00:09:52, I haz the power (commit-bot) wrote:
> > Retried try job too often on linux_rel for step(s) browser_tests
> >
>
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...
> 
> GpuFeatureTest.RafNoDamage fails, because TraceAnalyzer cannot handle
Dictionary
> arguments on trace events. It throws an error in trace_event_analyzer.cc:97
and
> fails hard. Unless we can foresee a use case that requires TraceAnalyzer to
> handle Dictionary arguments, I would suggest fixing this by ignoring the
> Dictionary arguments (and issue a warning) instead of failing hard when
> encountering one.

I've put up a separate CL for review that makes the proposed change:
https://codereview.chromium.org/23189006

[nduca, 2013-08-14 22:32:35]

What the heck is GpuFeatureTest and do we plan for it to live post-telemetry? My
understanding is that we were moving everything that is trace-event-analyzer
based to telemetry-based gpu tests. Maybe file a bug and route to kbr and
bajones about porting this test to telemetry.

[commit-bot: I haz the power, 2013-08-15 17:33:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ernstm@chromium.org/23088002/1

[commit-bot: I haz the power, 2013-08-15 18:51:36]

Change committed as 217824

[ernstm, 2013-08-16 20:11:45]

Cause of error on Mac OS 10.7:
Software compositing -> screen_frame_count recorded in main_stats instead of
impl_stats-> impl_stats overwrites main_stats' screen frame count in
EnumerateFileds -> 0

Fix:
Merge all filed enumeration into RenderingStats::EnumerateFields and properly
sum the screen frame counts of main_stats and impl_stats.

[commit-bot: I haz the power, 2013-08-16 20:26:17]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ernstm@chromium.org/23088002/35001

[nduca, 2013-08-16 20:47:51]

still lgtm :)

[commit-bot: I haz the power, 2013-08-17 02:43:22]

Change committed as 218149