[PATCH] SelectionDAG: Fix incorrect condition checks in some cases of folding FADD/FMUL combinations

[Stephen Lin]

Oops, forgot to rebase after an update. Here’s the rebased version:



On Jun 14, 2013, at 9:17 AM, Nadav Rotem <nrotem at apple.com> wrote:

> The FMA stuff looks okay (I did not review the order of operands carefully).  Your patch has some unrelated changes in it (windows/signals.inc .. )
> 
> Nadav
> 
> On Jun 14, 2013, at 9:12 AM, Stephen Lin <swlin at apple.com> wrote:
> 
>> Hi,
>> 
>> This is a fix for a bug noticed while in the process of implementing a different patch...a case was duplicated and order of operands was swapped, but some variable names were missed.
>> 
>> Fairly obvious fix, but it would be nice to have another set of eyes look over it to make sure that I didn’t miss anything else.
>> 
>> Also, some comments were slightly inaccurate so I fixed the comments to match what the code was doing. Alternatively, I could have fixed the code to match the comments, but it didn’t seem like there would be any benefit to doing so in this case.
>> 
>> <fix-unsafe-combine-fadd.patch>
>> 
>> Thanks,
>> Stephen

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130614/16f897c9/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fix-unsafe-combine-fadd.patch
Type: application/octet-stream
Size: 6566 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130614/16f897c9/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130614/16f897c9/attachment-0001.html>