[all-commits] [llvm/llvm-project] a6a72d: [AMDGPU][GlobalISel] Avoid selecting S_PACK with c...

Wed Jan 20 02:56:45 PST 2021

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a6a72dfdf2e132d64ea73ddbbc0d3431b6483724
      https://github.com/llvm/llvm-project/commit/a6a72dfdf2e132d64ea73ddbbc0d3431b6483724
  Author: Mirko Brkusanin <Mirko.Brkusanin at amd.com>
  Date:   2021-01-20 (Wed, 20 Jan 2021)

  Changed paths:
    M llvm/include/llvm/CodeGen/GlobalISel/Utils.h
    M llvm/lib/CodeGen/GlobalISel/Utils.cpp
    M llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
    M llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-build-vector-trunc.v2s16.mir
    M llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.fdot2.ll

  Log Message:
  -----------
  [AMDGPU][GlobalISel] Avoid selecting S_PACK with constants

If constants are hidden behind G_ANYEXT we can treat them same way as G_SEXT.
For that purpose we extend getConstantVRegValWithLookThrough with option
to handle G_ANYEXT same way as G_SEXT.

Differential Revision: https://reviews.llvm.org/D92219