Move CpuProfiler code create events behind Logger code api.

Move CpuProfiler code create events behind Logger code api. 

CpuProfiler has almost the same api for CodeCreate* events
but it was calling separately.

BUG=260203
R=svenpanne@chromium.org, yurys@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=15817

[yurys, 2013-07-22 07:39:41]

https://codereview.chromium.org/19916002/diff/2001/src/log.cc
File src/log.cc (left):

https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1212
src/log.cc:1212: if (!log_->IsEnabled() || !FLAG_log_code) return;
What's the reason for this rearrangement here an in other places?

https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1271
src/log.cc:1271: if (!is_logging_code_events()) return;
It checked for logging_nesting_ > 0; now there is no such check, can you explain
why we don't need it?

[loislo, 2013-07-22 09:31:42]

https://codereview.chromium.org/19916002/diff/2001/src/log.cc
File src/log.cc (left):

https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1212
src/log.cc:1212: if (!log_->IsEnabled() || !FLAG_log_code) return;
On 2013/07/22 07:39:41, Yury Semikhatsky wrote:
> What's the reason for this rearrangement here an in other places?

No special reason. I just made it consistent with the other places with the same
check.

https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1271
src/log.cc:1271: if (!is_logging_code_events()) return;
We already have this check in the macro LOG_CODE_EVENT,
so current implementation does this check twice.

[yurys, 2013-07-22 09:39:30]

https://codereview.chromium.org/19916002/diff/2001/src/log.cc
File src/log.cc (left):

https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1271
src/log.cc:1271: if (!is_logging_code_events()) return;
On 2013/07/22 09:31:42, loislo wrote:
> We already have this check in the macro LOG_CODE_EVENT,
> so current implementation does this check twice.

I disagree, now all these calls will be executed if
cpu_profiler()->is_profiling() == true no matter what is_logging_code_events()
state is.

[loislo, 2013-07-22 15:22:15]

On 2013/07/22 09:39:30, Yury Semikhatsky wrote:
> https://codereview.chromium.org/19916002/diff/2001/src/log.cc
> File src/log.cc (left):
> 
> https://codereview.chromium.org/19916002/diff/2001/src/log.cc#oldcode1271
> src/log.cc:1271: if (!is_logging_code_events()) return;
> On 2013/07/22 09:31:42, loislo wrote:
> > We already have this check in the macro LOG_CODE_EVENT,
> > so current implementation does this check twice.
> 
> I disagree, now all these calls will be executed if
> cpu_profiler()->is_profiling() == true no matter what is_logging_code_events()
> state is.

comments addressed

[yurys, 2013-07-22 16:17:52]

lgtm

[Sven Panne, 2013-07-23 08:04:20]

LGTM with a nit

https://codereview.chromium.org/19916002/diff/13001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/19916002/diff/13001/src/log.cc#newcode57
src/log.cc:57: #define PROFILER_LOG(Call)                               \
This macro is unsafe because of 2 reasons (well, actually more :-):

   * It contains a sequence of instructions, not a single one.

   * It has an unbalanced "if".

Although it doesn't matter at the current call sites, I would prefer a "do { ...
} while(false)" wrapper around it, just to be sure.

[loislo, 2013-07-23 08:12:24]

Committed patchset #4 manually as r15817 (presubmit successful).

[loislo, 2013-07-23 08:12:55]

https://codereview.chromium.org/19916002/diff/13001/src/log.cc
File src/log.cc (right):

https://codereview.chromium.org/19916002/diff/13001/src/log.cc#newcode57
src/log.cc:57: #define PROFILER_LOG(Call)                               \
On 2013/07/23 08:04:20, Sven Panne wrote:
> This macro is unsafe because of 2 reasons (well, actually more :-):
> 
>    * It contains a sequence of instructions, not a single one.
> 
>    * It has an unbalanced "if".
> 
> Although it doesn't matter at the current call sites, I would prefer a "do {
...
> } while(false)" wrapper around it, just to be sure.

Done.