[all-commits] [llvm/llvm-project] 7842e1: [AMDGPU] Fix large return values with amdgpu_gfx

[Sebastian Neubauer via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 7842e1725e80863cb5462351afbc293cb3a19111
      https://github.com/llvm/llvm-project/commit/7842e1725e80863cb5462351afbc293cb3a19111
  Author: Sebastian Neubauer <sebastian.neubauer at amd.com>
  Date:   2021-04-15 (Thu, 15 Apr 2021)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUCallingConv.td
    M llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-call-return-values.ll
    A llvm/test/CodeGen/AMDGPU/gfx-callable-return-types.ll

  Log Message:
  -----------
  [AMDGPU] Fix large return values with amdgpu_gfx

Returning in memory is not supported, so fall back to sret.
Also, extend i1 and i16 to i32. Otherwise, they would be passed through
memory.

Differential Revision: https://reviews.llvm.org/D100543