Introduce gpu_trace_event for gpu performance analysis.

Introduce gpu_trace_event for gpu performance analysis.

This changelist is a lightweight version of issue 6551019, but with
the code change confined to the gpu subsystem.

The intent of this change is to enable forward progress on gpu tracing.

Work on the aformentioned issue, i.e. fusing gpu_trace_event and trace_event, will continue.

BUG=
TEST=

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=78300

[nduca, 2011-03-14 20:51:34]

OK folks, same as before, but in gpu/common/. Key notes:
- jbates has a followon changelist that adds plumbing that plumbs trace data
about:gpu
- vince will continue to work on moving this system to base/. I really think
that is the right thing to do.

Hopefully a quick LGTM for everyone, but let me know if you spot issues.

[Ken Russell (switch to Gerrit), 2011-03-14 23:57:35]

LGTM. A few tiny nits not necessary to fix.

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
File chrome/browser/ui/webui/gpu_internals_ui.cc (right):

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
chrome/browser/ui/webui/gpu_internals_ui.cc:80: void OnBeginToEndTracing(const
ListValue* list);
The name "OnBeginToEndTracing" is a little hard to parse. Perhaps
"OnEndingTracing"?

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc
File gpu/common/gpu_trace_event.cc (right):

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc#n...
gpu/common/gpu_trace_event.cc:124: // TODO(nduca) replace with a hash_map
Colon after TODO (e.g., TODO(nduca): )

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc#n...
gpu/common/gpu_trace_event.cc:139: /* enable all categories */
Use C++ style comments instead?

[greggman, 2011-03-15 00:19:50]

only a few style issues otherwise LGTM

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
File chrome/browser/ui/webui/gpu_internals_ui.cc (right):

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
chrome/browser/ui/webui/gpu_internals_ui.cc:107: bool traceEnabled_;
trace_enabled_? (style guide)

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.h
File gpu/common/gpu_trace_event.h (right):

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.h#ne...
gpu/common/gpu_trace_event.h:206: 
remove this line blank (try gcl lint cl)

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.h#ne...
gpu/common/gpu_trace_event.h:247: // TODO(nduca): switch to per-thread tace
buffers to reduce thread
tace -> trace

[nduca, 2011-03-15 01:21:40]

Thanks folks.

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
File chrome/browser/ui/webui/gpu_internals_ui.cc (right):

http://codereview.chromium.org/6691013/diff/1/chrome/browser/ui/webui/gpu_int...
chrome/browser/ui/webui/gpu_internals_ui.cc:80: void OnBeginToEndTracing(const
ListValue* list);
On 2011/03/14 23:57:35, kbr wrote:
> The name "OnBeginToEndTracing" is a little hard to parse. Perhaps
> "OnEndingTracing"?

Done.

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc
File gpu/common/gpu_trace_event.cc (right):

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc#n...
gpu/common/gpu_trace_event.cc:124: // TODO(nduca) replace with a hash_map
On 2011/03/14 23:57:35, kbr wrote:
> Colon after TODO (e.g., TODO(nduca): )

Done.

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.cc#n...
gpu/common/gpu_trace_event.cc:139: /* enable all categories */
On 2011/03/14 23:57:35, kbr wrote:
> Use C++ style comments instead?

Done.

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.h
File gpu/common/gpu_trace_event.h (right):

http://codereview.chromium.org/6691013/diff/1/gpu/common/gpu_trace_event.h#ne...
gpu/common/gpu_trace_event.h:247: // TODO(nduca): switch to per-thread tace
buffers to reduce thread
On 2011/03/15 00:19:50, greggman wrote:
> tace -> trace

Done.