[Webkit-unassigned] [Bug 129429] [Win32][LLINT] Crash when running JSC stress tests.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=129429





--- Comment #13 from Michael Saboff <msaboff at apple.com>  2014-02-27 11:09:16 PST ---
(In reply to comment #12)
> Thanks for your comments :)
> 
> I have probably been doing some unnecessary work here.
> As suggested, I created a test program in C with a simple function which had a
> parameter of size > 4K.
> The assembly then shows a call to _chkstk, which probably does the job needed.
> 
> See
> 
> http://msdn.microsoft.com/en-us/library/ms648426(VS.85).aspx 
> 
> or e.g.
> 
> http://www.codeguru.com/cpp/v-s/debug/article.php/c19241/Adventures-with-chkstk.htm
> 
> Maybe we should be using this function instead?
> 
> I also believe the guard page is put between committed and uncommitted memory,
> see e.g. the comment in WebKit\Source\WTF\wtf\StackBounds.cpp in the Windows implementation.

There is other support on the Web for the notion that Windows moves the guard page - yuck!

Old Microsoft post: http://support.microsoft.com/kb/100775
Someone's reverse engineering: http://j00ru.vexillium.org/?p=1594

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