diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2020-02-19 13:23:11 -0500 |
|---|---|---|
| committer | Matt Arsenault <arsenm2@gmail.com> | 2020-03-09 22:51:54 -0400 |
| commit | ed72bcae3467f6a3874ddad13c9212ebce818150 (patch) | |
| tree | 0e2116d91eacbf6cb1fc2d857899d285c033bf6f /llvm/lib/CodeGen | |
| parent | 3f981cdde90fada6a6aa0f6ebe2f53105b80d523 (diff) | |
| download | llvm-ed72bcae3467f6a3874ddad13c9212ebce818150.zip llvm-ed72bcae3467f6a3874ddad13c9212ebce818150.tar.gz llvm-ed72bcae3467f6a3874ddad13c9212ebce818150.tar.bz2 | |
AMDGPU/GlobalISel: Fix mishandling SGPR v2s16 add/sub/mul
We weren't considering the packed case correctly, and this was passing
through to the selector. The selector only checked the size, so this
would incorrectly compile to a single 32-bit scalar add.
As usual, the LegalizerHelper is somewhat awkward to use from
applyMappingImpl. I think this is the first place we've needed
multi-step legalization here though.
Diffstat (limited to 'llvm/lib/CodeGen')
0 files changed, 0 insertions, 0 deletions