[Webkit-unassigned] [Bug 177839] bmalloc mutex should be adaptive

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Oct 3 15:01:05 PDT 2017


https://bugs.webkit.org/show_bug.cgi?id=177839

--- Comment #1 from Filip Pizlo <fpizlo at apple.com> ---
@Yusuke: I was originally going to use os_unfair_lock and suggest futexes for Linux.  Unfortunately, using os_unfair_lock was a regression, but using WordLock was fine.  So, I'm about to post a patch that pulls WTF's WordLock into bmalloc.  The code is horribly duplicated, but I don't want this change to be blocked on the larger question of how to layer WTF and bmalloc.

I still think that the most performant solution for Linux is futexes.  I think we should explore making WTF::WordLock/bmalloc::Mutex use futexes on Linux.  I don't think we could do the same with WTF::Lock, since WTF::Lock has some fairness guarantees that futexes can't easily give you.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20171003/bf19b16a/attachment.html>


More information about the webkit-unassigned mailing list