[Webkit-unassigned] [Bug 136436] REGRESSION(r173031): crashes during run-layout-jsc on x86/Linux

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=136436





--- Comment #11 from Julien Brianceau <jbriance at cisco.com>  2014-09-02 15:44:46 PST ---
(In reply to comment #10)
> Created an attachment (id=237519)
 --> (https://bugs.webkit.org/attachment.cgi?id=237519&action=review) [details]
> Patch
> 
> Well, we think alike.  I was working on this one and testing when you posted.  I prefer keying this off of argument registers and not CPU type.
> 
> BTW, X86-32 on Mac works with the old code on debug because the arguments are moved to local stack variables and for release because the args are moved into registers.

Looks good to me.

As the stack must remain aligned (otherwise it will crash in an alignment check in the LLINT), do you think it's worth to assert something like this too ?

   ASSERT(!((sizeof(Register)*4) % stackAlignmentBytes()));

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.