[Webkit-unassigned] [Bug 91354] Machine stack marker should not gather values already piled up on the stack before JSC working

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=91354





--- Comment #3 from hojong.han at samsung.com  2012-08-12 23:39:58 PST ---
(In reply to comment #2)
> (From update of attachment 152470 [details])
> I don't think this change is valid.  The StackBounds will be captured on the first call into WebCore, when the thread data is initialized.  This may occur at a deep stack depth.  
Setting top occurs at a deep stack depth because it's followed after initializingThread(), but not deep as much as origin which is the bottom of the stack.

> Subsequent calls into WebCore may occur at a much shallower stack depth.  
I think subsequent calls at a shallower stack depth do not matter of JSC working because in most cased it's deep enough from the perspective of JSC.

> It certainly is the case that there may be some range of the stack that we currently mark, which one could safely avoid marking – but I don't think this patch safely determines what that line is (also, I'm not sure how one would safely define this high water mark).
Do you think it's hardly defined to set high water mark?
Plz give me any feedback. thanks

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