[all-commits] [llvm/llvm-project] ca93d4: AMDGPU/GlobalISel: Handle atomic sextload and zext...

[Matt Arsenault via All-commits]

  Branch: refs/heads/users/arsenm/amdgpu-global-isel-fix-atomic-extloads
  Home:   https://github.com/llvm/llvm-project
  Commit: ca93d439010444c48ac6533e6ce032e768bf5141
      https://github.com/llvm/llvm-project/commit/ca93d439010444c48ac6533e6ce032e768bf5141
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-10-31 (Thu, 31 Oct 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUGISel.td
    M llvm/lib/Target/AMDGPU/AMDGPUInstructions.td
    M llvm/lib/Target/AMDGPU/BUFInstructions.td
    M llvm/lib/Target/AMDGPU/DSInstructions.td
    M llvm/lib/Target/AMDGPU/FLATInstructions.td
    M llvm/lib/Target/AMDGPU/SIInstrInfo.td
    A llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_flat.ll
    A llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_global.ll
    A llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_local_2.ll

  Log Message:
  -----------
  AMDGPU/GlobalISel: Handle atomic sextload and zextload

Atomic loads are handled differently from the DAG, and have separate opcodes
and explicit control over the extensions, like ordinary loads. Add
new patterns for these.

There's room for cleanup and improvement. d16 cases aren't handled.

Fixes #111645


  Commit: cfcc3ff2ade09c2cb78b150e43ce212fef1c832d
      https://github.com/llvm/llvm-project/commit/cfcc3ff2ade09c2cb78b150e43ce212fef1c832d
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-10-31 (Thu, 31 Oct 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/BUFInstructions.td

  Log Message:
  -----------
  Remove i16 result mubuf types.

This would only be used if we started using direct i16
values on gfx6/gfx7, but they should legalize all to i32


  Commit: 3b0f3462247c8945a3b7e859fe26e48497df5801
      https://github.com/llvm/llvm-project/commit/3b0f3462247c8945a3b7e859fe26e48497df5801
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-10-31 (Thu, 31 Oct 2024)

  Changed paths:
    M llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_global.ll

  Log Message:
  -----------
  Add missing gfx6 run line


  Commit: 2548945e35124f89297f2bce3b7ea54e4fddefb3
      https://github.com/llvm/llvm-project/commit/2548945e35124f89297f2bce3b7ea54e4fddefb3
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-10-31 (Thu, 31 Oct 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUInstructions.td
    M llvm/lib/Target/AMDGPU/FLATInstructions.td
    M llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_flat.ll
    M llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_global.ll
    M llvm/test/CodeGen/AMDGPU/GlobalISel/atomic_load_local_2.ll

  Log Message:
  -----------
  Fix memory predicates


Compare: https://github.com/llvm/llvm-project/compare/5c0c2905a400...2548945e3512

To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications