[all-commits] [llvm/llvm-project] ed72bc: AMDGPU/GlobalISel: Fix mishandling SGPR v2s16 add/...

[Matt Arsenault via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: ed72bcae3467f6a3874ddad13c9212ebce818150
      https://github.com/llvm/llvm-project/commit/ed72bcae3467f6a3874ddad13c9212ebce818150
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2020-03-09 (Mon, 09 Mar 2020)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
    M llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp
    A llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-scalar-packed.xfail.mir
    R llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-add.mir
    A llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-add.s16.mir
    A llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-add.s32.mir
    A llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-add.v2s16.mir

  Log Message:
  -----------
  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.