SamplingProfiler should hold API lock before reporting results

[Tadeu Zagallo]

...

[Tadeu Zagallo]

Created attachment 378882 [details]
Patch

[Yusuke Suzuki]

Comment on attachment 378882 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=378882&action=review

r=me with nit.

> Source/JavaScriptCore/runtime/SamplingProfiler.cpp:1029
> +        JSLockHolder holder(m_vm);

Can you ensure that this locking order (registerationLock -> APILock) is correct? (maybe, in practice, it does not matter).

[Tadeu Zagallo]

Thanks for the review!

(In reply to Yusuke Suzuki from comment #2)
> Can you ensure that this locking order (registerationLock -> APILock) is
> correct? (maybe, in practice, it does not matter).

The registrationLock is only guarding the append to the HashSet, `reportDataToOptionFile` is actually called later, from within the `atexit` lambda.

[WebKit Commit Bot]

Comment on attachment 378882 [details]
Patch

Clearing flags on attachment: 378882

Committed r249915: <https://trac.webkit.org/changeset/249915>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/55412906>