Telemetry: Remove 'overhead' time from thread times measurement.

Telemetry: Remove 'overhead' time from thread times measurement.

There is about 0.5ms of trace overhead per frame on the fast
path, so we might as well remove that so traces contribute
less noise.

This also merges 'browser_main' and 'browser_compositor'
threads since there isn't always both and they are small to
begin with. Best to do this before cleaning up the names on
the dashboard.

BUG=321439
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=248910

[epenner, 2014-01-23 01:17:09]

Ptal.

[epenner, 2014-01-24 05:44:50]

Rebased ptal. Also added some little clean ups.

https://codereview.chromium.org/139743018/diff/50001/tools/perf/metrics/timel...
File tools/perf/metrics/timeline.py (right):

https://codereview.chromium.org/139743018/diff/50001/tools/perf/metrics/timel...
tools/perf/metrics/timeline.py:217: 
This is the real change here.

https://codereview.chromium.org/139743018/diff/50001/tools/perf/metrics/timel...
tools/perf/metrics/timeline.py:234: 
And this.

[epenner, 2014-01-28 23:47:37]

Rebased. Ptal.

[nduca, 2014-01-29 21:05:30]

chunk mismatch error

[epennerAtGoogle, 2014-01-29 21:17:28]

On 2014/01/29 21:05:30, nduca wrote:
> chunk mismatch error

Oops! And I seem to have not uploaded my rebase.

Real rebase patch up now. Note the real change is quite small and then there is
some renaming and cleanup.

[epenner, 2014-02-04 22:52:16]

Ptal. No more chunk mismatch errors.

[ernstm, 2014-02-05 00:07:35]

On 2014/02/04 22:52:16, epenner wrote:
> Ptal. No more chunk mismatch errors.

Please add a test case (or extend an existing one) with trace overhead events.

[nduca, 2014-02-05 00:34:11]

with tests and manfreds lg then lgtm.

As a comment, I'm worried about the density of this code. Its very hard to read.
I know you understand it, but every time I return to it, I have a hard time. I
can't quite advise how to make it cleaner, but I typically find adding classes
and using object orientation helps.

https://codereview.chromium.org/139743018/diff/110001/tools/perf/metrics/time...
File tools/perf/metrics/timeline.py (right):

https://codereview.chromium.org/139743018/diff/110001/tools/perf/metrics/time...
tools/perf/metrics/timeline.py:167: OverheadTraceCategory =
"trace_event_overhead"
global vars sould be prefixed with _ when not intended for use outside the
module

https://codereview.chromium.org/139743018/diff/110001/tools/perf/metrics/time...
tools/perf/metrics/timeline.py:172: def ClockOverhead(event):
GetClockOverheadFromEvent

https://codereview.chromium.org/139743018/diff/110001/tools/perf/metrics/time...
tools/perf/metrics/timeline.py:179: def CpuOverhead(event):
GetCpuOverheadFromEvent or someting like that

[ernstm, 2014-02-05 00:50:16]

On 2014/02/05 00:34:11, nduca wrote:
> with tests and manfreds lg then lgtm.

lgtm % test

[epenner, 2014-02-05 02:21:19]

I added a unit test.

I also did some renaming and added some white space in between classes. I can't
think of any new objects to add at the moment. Perhaps I can break this up into
several files to further improve readability down the road.

[epenner, 2014-02-05 02:22:21]

The CQ bit was checked by epenner@chromium.org

[commit-bot: I haz the power, 2014-02-05 03:49:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/epenner@chromium.org/139743018/180001

[epennerAtGoogle, 2014-02-05 03:52:07]

The CQ bit was checked by epenner@google.com

[commit-bot: I haz the power, 2014-02-05 04:11:39]

Failed to trigger a try job on win_x64_rel
HTTP Error 400: Bad Request

[commit-bot: I haz the power, 2014-02-05 04:13:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/epenner@chromium.org/139743018/290001

[commit-bot: I haz the power, 2014-02-05 07:49:57]

Change committed as 248910