[llvm] r214011 - [x86] Fix PR20355 (for real). There are many layers to this bug.

[Quentin Colombet]

On Aug 13, 2014, at 4:51 PM, Chandler Carruth <chandlerc at google.com> wrote:

> 
> On Wed, Aug 13, 2014 at 4:50 PM, Quentin Colombet <qcolombet at apple.com> wrote:
> Hi Chandler, hi Ben,
> 
> I haven’t been able to come up with a test case that uses the low part of MUL_LOHI, even with Ben’s suggestion.
> 
> The questions now are:
> 1. Do we leave the code as is and wait for it to fail?
> or
> 2. Do we fix the mask as I think Ben, You, and I agree that this is wrong?
> 
> I would prefer #2, though #1 would give us a test case when it eventually fails!
> 
> No, we should do #2... I'm just really frustrated that we have tons of logic here and no way to test it. =[ That is kind of terrible in general.


Tell me about it :(.

Anyway, thanks for the follow-up on the whole problem.

Fixed the mask in revision 215604.

-Quentin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140813/dde6671b/attachment.html>