Bug 116630 - fourthTier: It should be possible to use more than one compiler thread
Summary: fourthTier: It should be possible to use more than one compiler thread
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Filip Pizlo
URL:
Keywords:
Depends on:
Blocks: 112836
  Show dependency treegraph
 
Reported: 2013-05-22 12:42 PDT by Filip Pizlo
Modified: 2013-05-22 13:42 PDT (History)
7 users (show)

See Also:


Attachments
the patch (11.08 KB, patch)
2013-05-22 12:45 PDT, Filip Pizlo
mhahnenberg: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Filip Pizlo 2013-05-22 12:42:28 PDT
Patch forthcoming.
Comment 1 Filip Pizlo 2013-05-22 12:45:31 PDT
Created attachment 202614 [details]
the patch
Comment 2 Filip Pizlo 2013-05-22 12:46:39 PDT
Comment on attachment 202614 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=202614&action=review

> Source/JavaScriptCore/runtime/SymbolTable.h:371
>      {
> -        ConcurrentJITLocker locker(m_lock);
> -        return get(locker, key);
> +        SymbolTableEntry result;
> +        {
> +            ConcurrentJITLocker locker(m_lock);
> +            result = get(locker, key);
> +        }
> +        return result;
>      }
>      

I'll investigate if this is necessary.
Comment 3 Mark Hahnenberg 2013-05-22 12:56:26 PDT
Comment on attachment 202614 [details]
the patch

r=me as long as reverting the symbol table stuff is okay to revert.
Comment 4 Filip Pizlo 2013-05-22 13:41:48 PDT
Landed in http://trac.webkit.org/changeset/150536
Comment 5 Filip Pizlo 2013-05-22 13:42:20 PDT
(In reply to comment #2)
> (From update of attachment 202614 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=202614&action=review
> 
> > Source/JavaScriptCore/runtime/SymbolTable.h:371
> >      {
> > -        ConcurrentJITLocker locker(m_lock);
> > -        return get(locker, key);
> > +        SymbolTableEntry result;
> > +        {
> > +            ConcurrentJITLocker locker(m_lock);
> > +            result = get(locker, key);
> > +        }
> > +        return result;
> >      }
> >      
> 
> I'll investigate if this is necessary.

And it wasn't necessary.  I backed this out before landing.