Introduce a separate event for CodeDisableOpt

Introduce a separate event for CodeDeopt

The reuse of CodeCreateEvent for deopt events caused a CodeCreateEvent
fired twice for a code object. When the event was processed for the first
time it seized the no-fp-ranges from code object, so the second event
had no ranges info leaving code entry without them.
As a result when a cpu profile sample falls into the region it missed the
2nd stack frame.

LOG=N
BUG=
R=bmeurer@chromium.org, loislo@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=21418

[yurys, 2014-05-20 07:59:01]

Don't you need to update tickprocessor.js and profviz code?

https://codereview.chromium.org/290093005/diff/1/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/290093005/diff/1/src/log.cc#newcode1454
src/log.cc:1454: void Logger::CodeDeoptEvent(Code* code,
How does it correlate to existing Logger::CodeDeoptEvent which will produce
"code-deopt" event in the log file?

[loislo, 2014-05-20 08:14:04]

https://codereview.chromium.org/290093005/diff/1/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/290093005/diff/1/src/objects.cc#newcode10690
src/objects.cc:10690: PROFILE(GetIsolate(), CodeDeoptEvent(code(), this));
I'd say that the event named incorrectly.
It is not a CodeDeoptEvent. It is CodeOptimizationDisabledEvent.

[loislo, 2014-05-20 08:16:17]

https://codereview.chromium.org/290093005/diff/1/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/290093005/diff/1/src/log.cc#newcode1454
src/log.cc:1454: void Logger::CodeDeoptEvent(Code* code,
On 2014/05/20 07:59:02, yurys wrote:
> How does it correlate to existing Logger::CodeDeoptEvent which will produce
> "code-deopt" event in the log file?

We need to rename LogExistingFunction into CodeOptimizationDisabledEvent

and use CodeDeoptEvent for creating deopt statistics for js files

[alph, 2014-05-21 13:55:26]

On 2014/05/20 07:59:01, yurys wrote:
> Don't you need to update tickprocessor.js and profviz code?

profviz seems to skip unknown commands.
https://code.google.com/p/chromium/codesearch#chromium/src/v8/tools/logreader...

> https://codereview.chromium.org/290093005/diff/1/src/log.cc
> File src/log.cc (right):
> 
> https://codereview.chromium.org/290093005/diff/1/src/log.cc#newcode1454
> src/log.cc:1454: void Logger::CodeDeoptEvent(Code* code,
> How does it correlate to existing Logger::CodeDeoptEvent which will produce
> "code-deopt" event in the log file?

Renamed mine to CodeDisableOptEvent.

[alph, 2014-05-21 13:55:32]

https://codereview.chromium.org/290093005/diff/1/src/objects.cc
File src/objects.cc (right):

https://codereview.chromium.org/290093005/diff/1/src/objects.cc#newcode10690
src/objects.cc:10690: PROFILE(GetIsolate(), CodeDeoptEvent(code(), this));
On 2014/05/20 08:14:05, loislo wrote:
> I'd say that the event named incorrectly.
> It is not a CodeDeoptEvent. It is CodeOptimizationDisabledEvent.

Done.

[yurys, 2014-05-21 14:13:30]

https://codereview.chromium.org/290093005/diff/20001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/290093005/diff/20001/src/log.cc#newcode1465
src/log.cc:1465: msg.Append("\"%s\"\n", name.get());
This looks odd, you are logging just function name without any header of the
event. Also you might need to update tests for tickprocessor.js and profviz.

[alph, 2014-05-21 15:06:47]

Benedict, could you please take a look.

https://codereview.chromium.org/290093005/diff/20001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/290093005/diff/20001/src/log.cc#newcode1465
src/log.cc:1465: msg.Append("\"%s\"\n", name.get());
On 2014/05/21 14:13:31, yurys wrote:
> This looks odd, you are logging just function name without any header of the
> event. Also you might need to update tests for tickprocessor.js and profviz.

Done.

[loislo, 2014-05-21 19:04:36]

lgtm

[loislo, 2014-05-21 19:04:38]

lgtm

lgtm

[Benedikt Meurer, 2014-05-22 05:27:13]

LGTM

[alph, 2014-05-22 05:36:38]

Committed patchset #4 manually as r21418 (presubmit successful).