[llvm] r247074 - AMDGPU/SI: Fix input vcc operand for VOP2b instructions

[Matt Arsenault via llvm-commits]

> On Sep 9, 2015, at 12:20 AM, Michel Dänzer via llvm-commits <llvm-commits at lists.llvm.org> wrote:
> 
> 
> Hi Matt,
> 
> 
> On 09.09.2015 06:15, Matt Arsenault via llvm-commits wrote:
>> Author: arsenm
>> Date: Tue Sep  8 16:15:00 2015
>> New Revision: 247074
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=247074&view=rev
>> Log:
>> AMDGPU/SI: Fix input vcc operand for VOP2b instructions
>> 
>> Adds vcc to output string input for e32. Allows option
>> of using e64 encoding with assembler.
>> 
>> Also fixes these instructions not implicitly reading exec.
> 
> This change broke a bunch of fp64 related piglit (graphics) tests on my
> Kaveri. I'm attaching an example of good and bad LLVM IR.
> 
> 
> -- 
> Earthling Michel Dänzer               |               http://www.amd.com
> Libre software enthusiast             |             Mesa and X developer
> <fs-inverse-dmat2.bad><fs-inverse-dmat2.good>_______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits

r247123 should fix it