[Webkit-unassigned] [Bug 99706] MIPS LLInt implementation.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=99706





--- Comment #10 from Oliver Hunt <oliver at apple.com>  2012-10-18 10:53:34 PST ---
(In reply to comment #9)
> (In reply to comment #8)
> > (From update of attachment 169392 [details] [details])
> > View in context: https://bugs.webkit.org/attachment.cgi?id=169392&action=review
> > 
> > r=me, but it would be nice if you could respond to my comments :D
> > 
> > > Source/JavaScriptCore/llint/LLIntOfflineAsmConfig.h:126
> > > +    ".set noreorder\n" \
> > 
> > Quick question, what does this do?  Disable instruction reordering, or disable function reordering - LLInt should be safe against function reordering so if it's the latter this is probably unnecessary.  If it's not then just ignore me :D
> 
> This macro turns off gas' optimization of instruction reordering and must be used around .cpload macro which is necessary for PIC compatibility.

Righto

> 
> > 
> > > Source/JavaScriptCore/llint/LowLevelInterpreter.asm:163
> > > +        move sourceRegister, lr
> > 
> > I feel that we should just have 
> > if C_LOOP or ARMv7 or MIPS
> >     move sourceRegister, lr
> > elsif X86 or X86_64
> >      ...
> > else
> >     error
> > end
> Could be. Should I make a new patch with this modification?

Nah, but maybe make the change as a followup?

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