[Webkit-unassigned] [Bug 142924] New: FTL (and maybe DFG) should have GC deferred indefinitely in the slow paths of any node that doesn't clobber world

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=142924

            Bug ID: 142924
           Summary: FTL (and maybe DFG) should have GC deferred
                    indefinitely in the slow paths of any node that
                    doesn't clobber world
    Classification: Unclassified
           Product: WebKit
           Version: 528+ (Nightly build)
          Hardware: All
                OS: All
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: JavaScriptCore
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: fpizlo at apple.com

And we should insert fast "should GC?" checks on loop backedges and function exits, of any code that could GC but not clobber world.  Those insertion points should be computed before SSA conversion, so that SSA could create loops without worrying about creating new GC points.

This can be viewed as an alternate solution to most of the issues that https://bugs.webkit.org/show_bug.cgi?id=142921 would fix.

-- 
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/20150320/b93210b4/attachment-0002.html>