[all-commits] [llvm/llvm-project] 0da582: GlobalISel: Handle llvm.roundeven

[Matt Arsenault via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 0da582d9b6cc8a96682628ff90b5e0e5d155358c
      https://github.com/llvm/llvm-project/commit/0da582d9b6cc8a96682628ff90b5e0e5d155358c
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2020-07-29 (Wed, 29 Jul 2020)

  Changed paths:
    M llvm/include/llvm/CodeGen/GlobalISel/LegalizerHelper.h
    M llvm/include/llvm/CodeGen/GlobalISel/LegalizerInfo.h
    M llvm/include/llvm/Support/TargetOpcodes.def
    M llvm/include/llvm/Target/GenericOpcodes.td
    M llvm/lib/CodeGen/GlobalISel/IRTranslator.cpp
    M llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
    M llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
    M llvm/lib/Target/X86/X86LegalizerInfo.cpp
    M llvm/test/CodeGen/AArch64/GlobalISel/legalizer-info-validation.mir
    A llvm/test/CodeGen/AMDGPU/GlobalISel/roundeven.ll
    A llvm/test/CodeGen/X86/GlobalISel/roundeven.ll

  Log Message:
  -----------
  GlobalISel: Handle llvm.roundeven

I still think it's highly questionable that we have two intrinsics
with identical behavior and only vary by the name of the libcall used
if it happens to be lowered that way, but try to reduce the feature
delta between SDAG and GlobalISel for recently added intrinsics. I'm
not sure which opcode should be considered the canonical one, but
lower roundeven back to round.