[PATCH] IR support for "cmpxchg weak"

[Chandler Carruth]

On Fri, Jun 13, 2014 at 2:57 PM, Tim Northover <t.p.northover at gmail.com>
wrote:

> > .. of the patch. The paragraph is confusing, as I don't see why Clang
> would
> > need to change with this patch as it doesn't set weak to true...
>
> It has to adapt to the new return type of cmpxchg instructions.
> Changing all the current icmps to extractvalues and adding a bunch
> more for when the loaded value is used.
>

Ah, yea, so its not about weak, its about result extraction. Makes sense.


>
> I'm attaching it here in case you're interested, though it's all
> pretty straightforward.
>

LGTM, but please submit the CHECK-LABEL-ification as its own patch first? =]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140613/83e09885/attachment.html>