Bug 126313 - Stop using ThreadCondition in BlockAllocator
Summary: Stop using ThreadCondition in BlockAllocator
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Anders Carlsson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-12-30 16:40 PST by Anders Carlsson
Modified: 2014-01-20 11:27 PST (History)
0 users

See Also:

Attachments
Patch (5.37 KB, patch)
2013-12-30 16:41 PST, Anders Carlsson 		no flags Details | Formatted Diff | Diff
Patch (5.37 KB, patch)
2013-12-30 17:14 PST, Anders Carlsson 		sam: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Anders Carlsson 2013-12-30 16:40:13 PST 
Stop using ThreadCondition in BlockAllocator
Comment 1 Anders Carlsson 2013-12-30 16:41:36 PST 
Created attachment 220136 [details]
Patch
Comment 2 Anders Carlsson 2013-12-30 17:14:57 PST 
Created attachment 220138 [details]
Patch
Comment 3 Anders Carlsson 2013-12-30 17:18:38 PST 
Committed r161161: <http://trac.webkit.org/changeset/161161>
Comment 4 Alexey Proskuryakov 2013-12-30 22:53:40 PST 
Rolled out with the rest of today's threading fixes, which broke WebKit2 on Mountain Lion.

I don't know if this particular change caused any trouble - there was so much churn today that I could't tell which patch caused which badness. Perhaps we should simply re-land if we can get EWS green.

I think that r161163 was a follow-up fix to this.
Comment 5 Filip Pizlo 2014-01-01 09:52:07 PST 
Comment on attachment 220138 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=220138&action=review

> Source/JavaScriptCore/heap/BlockAllocator.cpp:106
> +    std::unique_lock<std::mutex> lock(m_emptyRegionConditionMutex);

Why is this a unique_lock instead of lock_guard?
Comment 6 Anders Carlsson 2014-01-06 11:12:55 PST 
(In reply to comment #5)
> (From update of attachment 220138 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=220138&action=review
> 
> > Source/JavaScriptCore/heap/BlockAllocator.cpp:106
> > +    std::unique_lock<std::mutex> lock(m_emptyRegionConditionMutex);
> 
> Why is this a unique_lock instead of lock_guard?

Unique locks are used together with std::condition_variable.
Comment 7 Anders Carlsson 2014-01-20 11:27:19 PST 
Committed r162362: <http://trac.webkit.org/changeset/162362>