[all-commits] [llvm/llvm-project] 4778b8: [AMDGPU][MC] Avoid creating lit64() operands unles...

[Ivan Kosarev via All-commits]

  Branch: refs/heads/users/kosarev/drop-implicit-lit64-mods
  Home:   https://github.com/llvm/llvm-project
  Commit: 4778b8c12d7154836deb65b3122608530dde3e25
      https://github.com/llvm/llvm-project/commit/4778b8c12d7154836deb65b3122608530dde3e25
  Author: Ivan Kosarev <ivan.kosarev at amd.com>
  Date:   2025-09-29 (Mon, 29 Sep 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
    M llvm/lib/Target/AMDGPU/Disassembler/AMDGPUDisassembler.cpp
    M llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUMCCodeEmitter.cpp
    M llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
    M llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.h
    M llvm/test/MC/AMDGPU/gfx1250_asm_salu_lit64.s
    M llvm/test/MC/AMDGPU/gfx1250_asm_sop1.s
    M llvm/test/MC/AMDGPU/gfx1250_asm_valu_lit64.s
    M llvm/test/MC/AMDGPU/gfx1250_asm_vop1-fake16.s
    M llvm/test/MC/AMDGPU/gfx1250_asm_vop1.s
    M llvm/test/MC/AMDGPU/gfx1250_asm_vop2.s
    M llvm/test/MC/AMDGPU/gfx12_asm_sop1.s
    M llvm/test/MC/AMDGPU/gfx12_asm_sop2.s
    M llvm/test/MC/AMDGPU/gfx12_asm_sopc.s
    M llvm/test/MC/AMDGPU/literals.s
    M llvm/test/MC/AMDGPU/vop3-literal.s
    M llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_salu_lit64.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_sop1.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_valu_lit64.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_vop1.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_vop2.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx12_dasm_sop1.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx12_dasm_sop2.txt
    M llvm/test/MC/Disassembler/AMDGPU/gfx12_dasm_sopc.txt

  Log Message:
  -----------
  [AMDGPU][MC] Avoid creating lit64() operands unless asked or needed.

There should normally be no need to generate implicit lit64()
modifiers on the assembler side. It's the encoder's responsibility
to recognise literals that are implicitly 64 bits wide.

The exceptions are where we rewrite floating-point operand values
as integer ones, which would not be assembled back to the original
values unless wrapped into lit64().

Respect explicit lit() modifiers for non-inline values as
necessary to avoid regressions in MC tests. This change still
doesn't prevent use of inline constants where lit()/lit64 is
specified; subject to a separate patch.

On disassembling, only create lit64() operands where necessary for
correct round-tripping.

Add round-tripping tests where useful and feasible.



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