[Webkit-unassigned] [Bug 130638] [Win64] ASM LLINT is not enabled.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=130638





--- Comment #60 from peavo at outlook.com  2014-06-25 08:56:17 PST ---
(In reply to comment #58)
> (From update of attachment 233709 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=233709&action=review
>

Thanks for reviewing :)

> I would r+ this patch but first, I'd like you to try undoing the .opPutByIdSlow changes I suggested.  If they are not needed anymore, let's undo them.  Let me know if they are still or not.  Thanks.
> 

They are not needed anymore, undone in latest patch.

> > Source/JavaScriptCore/assembler/MacroAssemblerX86_64.h:166
> > +        load64(Address(X86Registers::esp, 32), scratchRegister);
> > +        store64(scratchRegister, Address(X86Registers::esp, -32));
> > +
> > +        load64(Address(X86Registers::esp, 40), scratchRegister);
> > +        store64(scratchRegister, Address(X86Registers::esp, -24));
> 
> To answer your question about using a POKE_ARGUMENT_OFFSET of -4 to achieve this, I think you made a good point there.  I think it will work (and one might want to go with that solution if performance is an issue here).  However, it's not obvious from reading the code as to how things work since it differs from how all other ports expect it to work.  So, if you were to go with that solution, I would like to see a comment here to describe why the 5th and 6th arguments are already in place and no further work is needed.  I would also want a comment in the definition POKE_ARGUMENT_OFFSET to explain why it has a negative offset.  However, if performance is not an issue, then let's stick with this more obvious solution for now.
> 
Sounds good, I added a comment regarding the copying of argument 5 and 6.

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