Fix TSAN issue wrt assertions in the optimizing compiler thread.

Fix TSAN issue wrt assertions in the optimizing compiler thread.

R=hpayer@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=20322

[Hannes Payer (out of office), 2014-03-27 17:16:20]

https://codereview.chromium.org/212603013/diff/1/src/handles-inl.h
File src/handles-inl.h (right):

https://codereview.chromium.org/212603013/diff/1/src/handles-inl.h#newcode61
src/handles-inl.h:61: if (location_ == NULL || other.location_ == NULL) return
false;
I guess you moved that code to have the asserts in the beginning of the
function?

https://codereview.chromium.org/212603013/diff/1/src/hydrogen-environment-liv...
File src/hydrogen-environment-liveness.cc (right):

https://codereview.chromium.org/212603013/diff/1/src/hydrogen-environment-liv...
src/hydrogen-environment-liveness.cc:236: Heap::RelocationLock
for_heap_access(isolate()->heap());
what about moving that locking code to where it is needed, i.e.
ZapEnvironmentSlotsForInstruction?

https://codereview.chromium.org/212603013/diff/1/src/hydrogen-environment-liv...
src/hydrogen-environment-liveness.cc:244: Heap::RelocationLock
for_heap_access(isolate()->heap());
what about moving that locking code to where it is needed, i.e.
ZapEnvironmentSlotsInSuccessors?

[Yang, 2014-03-28 08:27:39]

https://codereview.chromium.org/212603013/diff/1/src/handles-inl.h
File src/handles-inl.h (right):

https://codereview.chromium.org/212603013/diff/1/src/handles-inl.h#newcode61
src/handles-inl.h:61: if (location_ == NULL || other.location_ == NULL) return
false;
On 2014/03/27 17:16:20, Hannes Payer wrote:
> I guess you moved that code to have the asserts in the beginning of the
> function?

Yes. This is to tighten the assertion.

https://codereview.chromium.org/212603013/diff/1/src/hydrogen-environment-liv...
File src/hydrogen-environment-liveness.cc (right):

https://codereview.chromium.org/212603013/diff/1/src/hydrogen-environment-liv...
src/hydrogen-environment-liveness.cc:236: Heap::RelocationLock
for_heap_access(isolate()->heap());
On 2014/03/27 17:16:20, Hannes Payer wrote:
> what about moving that locking code to where it is needed, i.e.
> ZapEnvironmentSlotsForInstruction?

I refactored the handle comparison.

[Hannes Payer (out of office), 2014-03-28 08:32:18]

I like it, LGTM with two nits...

https://codereview.chromium.org/212603013/diff/20001/src/hydrogen-environment...
File src/hydrogen-environment-liveness.cc (right):

https://codereview.chromium.org/212603013/diff/20001/src/hydrogen-environment...
src/hydrogen-environment-liveness.cc:88: block->last_environment()->closure()));
+4 spaces indent

https://codereview.chromium.org/212603013/diff/20001/src/hydrogen-environment...
src/hydrogen-environment-liveness.cc:230: // Acquire relocation lock to
dereference for assertion.
remove this comments

[Yang, 2014-03-28 08:50:09]

Committed patchset #3 manually as r20322 (presubmit successful).