[webkit-reviews] review requested: [Bug 122617] transition void cti_op_* methods to JIT operations. : [Attachment 214280] Updated patch. Rebased and with fix.

[bugzilla-daemon at webkit.org]

Michael Saboff <msaboff at apple.com> has asked  for review:
Bug 122617: transition void cti_op_* methods to JIT operations.
https://bugs.webkit.org/show_bug.cgi?id=122617

Attachment 214280: Updated patch.  Rebased and with fix.
https://bugs.webkit.org/attachment.cgi?id=214280&action=review

------- Additional Comments from Michael Saboff <msaboff at apple.com>
(In reply to comment #5)
> (From update of attachment 214265 [details])
> View in context:
https://bugs.webkit.org/attachment.cgi?id=214265&action=review
> 
> > Source/JavaScriptCore/jit/CCallHelpers.h:1096
> > +	     poke(arg4, POKE_ARGUMENT_OFFSET);
> 
> Will this poke be valid when we're running on the C stack? Will each stack
frame reserve space for these pokes?

This poke is valid way to pass additional arguments.  When we move JSC over to
the C stack, we'll need to make sure there is space for calling out to C
including any calls made by the immediate callee.  My thinking is that we make
sure there is stack space for C calls beyond which JSC can use for stack
frames.
 
> > Source/JavaScriptCore/jit/JITOpcodes32_64.cpp:1185
> > +	 callOperation(operationProfileWillCall, regT1, regT0);
> 
> Wrong function = broken profiler. Did the profiler tests pass for this patch?


I made the change and ran the layout tests including the profiler tests with a
32 bit build.