[all-commits] [llvm/llvm-project] 6adb9a: [AMDGPU] Emit predefined macro `__AMDGCN_CUMODE__`

[Yaxun (Sam) Liu via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 6adb9a0602bcd75acd383e13c253e471edc9b189
      https://github.com/llvm/llvm-project/commit/6adb9a0602bcd75acd383e13c253e471edc9b189
  Author: Yaxun (Sam) Liu <yaxun.liu at amd.com>
  Date:   2023-05-12 (Fri, 12 May 2023)

  Changed paths:
    M clang/include/clang/Basic/DiagnosticDriverKinds.td
    M clang/lib/Basic/Targets/AMDGPU.cpp
    M clang/lib/Basic/Targets/AMDGPU.h
    M clang/lib/Driver/ToolChains/AMDGPU.cpp
    M clang/lib/Driver/ToolChains/Arch/PPC.cpp
    M clang/lib/Driver/ToolChains/Arch/RISCV.cpp
    M clang/lib/Driver/ToolChains/CommonArgs.cpp
    M clang/lib/Driver/ToolChains/CommonArgs.h
    M clang/lib/Driver/ToolChains/Hexagon.cpp
    M clang/lib/Driver/ToolChains/Hexagon.h
    A clang/test/CodeGenHIP/hip-cumode.hip
    M clang/test/Driver/amdgpu-macros.cl
    M clang/test/Driver/hip-macros.hip
    M llvm/include/llvm/TargetParser/TargetParser.h
    M llvm/lib/TargetParser/TargetParser.cpp

  Log Message:
  -----------
  [AMDGPU] Emit predefined macro `__AMDGCN_CUMODE__`

Predefine __AMDGCN_CUMODE__ as 1 or 0 when compilation assumes CU or WGP modes.

If WGP mode is not supported, ignore -mno-cumode and emit a warning.

This is needed for implementing device functions like __smid
(https://github.com/ROCm-Developer-Tools/hipamd/blob/312dff7b794337aa040be0691acc78e9f968a8d2/include/hip/amd_detail/amd_device_functions.h#L957)

Reviewed by: Matt Arsenault, Artem Belevich, Brian Sumner

Differential Revision: https://reviews.llvm.org/D145343