Add profiling_controller that manages profilers.

Add profiling_controller that manages profilers.
Remove StartProfiling & StopProfiling methods from browser.py.

profiling_controller is owned by platform, so this will allow decoupling lifetime of profilers from browser. 
This cannot be achieved yet in this patch since we have some
profilers whose lifetime dependent on the browser, e.g: v8, tracing... To do
this, profilers' APIs should be modified for case browser_backend is optional if
they are not dependent on the browsers (e.g: monsoon, tcp_dump...).
For now, StartProfiling will explode if the number of running browsers is not 1.

BUG=

Committed: https://crrev.com/54ab6c560f9738070d152bdd5246a3e32ee459dd
Cr-Commit-Position: refs/heads/master@{#293618}

[nednguyen, 2014-08-26 23:12:29]

nednguyen@google.com changed reviewers:
+ tonyg@chromium.org

[nednguyen, 2014-09-04 16:05:00]

nednguyen@google.com changed reviewers:
+ nduca@chromium.org

[tonyg, 2014-09-04 17:10:19]

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
File tools/perf/measurements/loading_profile.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
tools/perf/measurements/loading_profile.py:37: profile_files =
tab.browser.StopProfiling()
How does this work?

Note that we could simply remove this Measurement if it's in the way. I'm not
aware of anyone using it and I've always thought it is kind of strange for a
Measurement to try to do something with a Profiler.

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
File tools/perf/measurements/memory.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
tools/perf/measurements/memory.py:37: if
tab.browser.platform.profiling_controller.is_profiler_active(
This also seems like the wrong place for this. The Profiler itself should be
running this. Is that possible under the API these days?

https://codereview.chromium.org/511653002/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/profiling_controller.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiling_controller.py:10: def
StartProfiling(self, profiler_name, base_output_file):
I'd prefer the more terse names Start() and Stop() and IsActive() since they're
on a class already named ProfilingController.

[nednguyen, 2014-09-04 17:31:46]

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
File tools/perf/measurements/loading_profile.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
tools/perf/measurements/loading_profile.py:37: profile_files =
tab.browser.StopProfiling()
On 2014/09/04 17:10:19, tonyg wrote:
> How does this work?
> 
> Note that we could simply remove this Measurement if it's in the way. I'm not
> aware of anyone using it and I've always thought it is kind of strange for a
> Measurement to try to do something with a Profiler.

Removing patch in https://codereview.chromium.org/540483004/

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
File tools/perf/measurements/memory.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
tools/perf/measurements/memory.py:37: if
tab.browser.platform.profiling_controller.is_profiler_active(
On 2014/09/04 17:10:19, tonyg wrote:
> This also seems like the wrong place for this. The Profiler itself should be
> running this. Is that possible under the API these days?

Yeah, we leak the profiler API through the handles in browser object. Maybe this
optimization can be moved to tcmalloc_heap_profiler.WillCloseBrowser method,
wdyt?

[nednguyen, 2014-09-05 17:08:34]

On 2014/09/04 17:31:46, nednguyen wrote:
>
https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
> File tools/perf/measurements/loading_profile.py (right):
> 
>
https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
> tools/perf/measurements/loading_profile.py:37: profile_files =
> tab.browser.StopProfiling()
> On 2014/09/04 17:10:19, tonyg wrote:
> > How does this work?
> > 
> > Note that we could simply remove this Measurement if it's in the way. I'm
not
> > aware of anyone using it and I've always thought it is kind of strange for a
> > Measurement to try to do something with a Profiler.
> 
> Removing patch in https://codereview.chromium.org/540483004/
> 
>
https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
> File tools/perf/measurements/memory.py (right):
> 
>
https://codereview.chromium.org/511653002/diff/40001/tools/perf/measurements/...
> tools/perf/measurements/memory.py:37: if
> tab.browser.platform.profiling_controller.is_profiler_active(
> On 2014/09/04 17:10:19, tonyg wrote:
> > This also seems like the wrong place for this. The Profiler itself should be
> > running this. Is that possible under the API these days?
> 
> Yeah, we leak the profiler API through the handles in browser object. Maybe
this
> optimization can be moved to tcmalloc_heap_profiler.WillCloseBrowser method,
> wdyt?

Rebase. PTAL

[nednguyen, 2014-09-05 17:08:43]

https://codereview.chromium.org/511653002/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/profiling_controller.py (right):

https://codereview.chromium.org/511653002/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiling_controller.py:10: def
StartProfiling(self, profiler_name, base_output_file):
On 2014/09/04 17:10:19, tonyg wrote:
> I'd prefer the more terse names Start() and Stop() and IsActive() since
they're
> on a class already named ProfilingController.

Done. I also remove is_profiler_active since it's no longer used.

[tonyg, 2014-09-05 17:13:39]

lgtm

[nednguyen, 2014-09-05 17:19:13]

The CQ bit was checked by nednguyen@google.com

[commit-bot: I haz the power, 2014-09-05 17:20:09]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nednguyen@google.com/511653002/80001

[commit-bot: I haz the power, 2014-09-05 23:37:32]

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

[commit-bot: I haz the power, 2014-09-05 23:37:33]

Exceeded time limit waiting for builds to trigger.

[nednguyen, 2014-09-05 23:58:31]

The CQ bit was checked by nednguyen@google.com

[commit-bot: I haz the power, 2014-09-06 00:00:09]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nednguyen@google.com/511653002/80001

[commit-bot: I haz the power, 2014-09-06 06:02:56]

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

[commit-bot: I haz the power, 2014-09-06 06:02:57]

Exceeded time limit waiting for builds to trigger.

[nednguyen, 2014-09-06 14:33:19]

The CQ bit was checked by nednguyen@google.com

[commit-bot: I haz the power, 2014-09-06 14:34:26]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nednguyen@google.com/511653002/80001

[commit-bot: I haz the power, 2014-09-06 14:35:51]

Committed patchset #5 (id:80001) as bc2b2dd0bfa664b9874e71effbede3dded57e469

[commit-bot: I haz the power, 2014-09-10 03:42:38]

Patchset 5 (id:??) landed as
https://crrev.com/54ab6c560f9738070d152bdd5246a3e32ee459dd
Cr-Commit-Position: refs/heads/master@{#293618}