[PATCH] Enable vmla.f32 generation on Cortex-A15

[Renato Golin]

Hi Silviu,

Patch looks good to me, though this if could be fused together:

-  if (!Subtarget->isCortexA8() && !Subtarget->isLikeA9() &&
+  if (Subtarget->isCortexA15())
+    return true;
+
+  if (!Subtarget->isCortexA8() && !Subtarget->isCortexA9() &&
       !Subtarget->isSwift())
     return true;

Also, it seems that the way "Darwin" is selected here is via Swift, which
is probably not the best, but I think that will do for now.

cheers,
--renato



On 25 July 2013 12:07, Silviu Baranga <Silviu.Baranga at arm.com> wrote:

> Hi,
>
> The attached patch enables the generation of vmla.f32 instructions
> for Cortex-A15 CPUs.
>
> To patch disables the MLxExpansion pass and modifies the
> "hasNoVMLxHazardUse" function (this was preventing the vmla.f32
> instructions from being selected).
>
> This also adds the VFPv4 feature for Cortex-A15 and tweaks the
> DontUseFusedMAC predicate (the previous logic disabled vmla.f32
> generation on all CPUs with VFPv4 on non-Darwin targets).
>
> Please review!
>
> Thanks,
> Silviu
>
> -- IMPORTANT NOTICE: The contents of this email and any attachments are
> confidential and may also be privileged. If you are not the intended
> recipient, please notify the sender immediately and do not disclose the
> contents to any other person, use it for any purpose, or store or copy the
> information in any medium.  Thank you.
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130726/e6209ca5/attachment.html>