[all-commits] [llvm/llvm-project] dcf485: MC: Centralize X86 PC-relative fixup adjustment in...

[Fangrui Song via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: dcf485609c5c4fd01045876e630dbd5ba220543e
      https://github.com/llvm/llvm-project/commit/dcf485609c5c4fd01045876e630dbd5ba220543e
  Author: Fangrui Song <i at maskray.me>
  Date:   2025-07-08 (Tue, 08 Jul 2025)

  Changed paths:
    M bolt/lib/Target/X86/X86MCPlusBuilder.cpp
    M llvm/include/llvm/MC/MCValue.h
    M llvm/lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp
    M llvm/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp
    M llvm/test/CodeGen/X86/AMX/amx-lower-tile-copy.ll
    M llvm/test/CodeGen/X86/AMX/amx-spill-merge.ll
    M llvm/test/CodeGen/X86/absolute-cmp.ll
    M llvm/test/CodeGen/X86/apx/and.ll
    M llvm/test/CodeGen/X86/apx/ccmp.ll
    M llvm/test/CodeGen/X86/apx/kmov-postrapseudos.ll
    M llvm/test/CodeGen/X86/apx/or.ll
    M llvm/test/CodeGen/X86/apx/sub.ll
    M llvm/test/CodeGen/X86/apx/xor.ll
    M llvm/test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll
    M llvm/test/CodeGen/X86/avx2-intrinsics-x86.ll
    M llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
    M llvm/test/CodeGen/X86/avx512-vec-cmp.ll
    M llvm/test/CodeGen/X86/avx512bf16-vl-intrinsics.ll
    M llvm/test/CodeGen/X86/avx512bw-intrinsics.ll
    M llvm/test/CodeGen/X86/avx512bwvl-intrinsics.ll
    M llvm/test/CodeGen/X86/avx512fp16-fmaxnum.ll
    M llvm/test/CodeGen/X86/avx512fp16-fminnum.ll
    M llvm/test/CodeGen/X86/avx512vl-arith.ll
    M llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
    M llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
    M llvm/test/CodeGen/X86/bmi.ll
    M llvm/test/CodeGen/X86/cmp.ll
    M llvm/test/CodeGen/X86/conditional-tailcall-pgso.ll
    M llvm/test/CodeGen/X86/conditional-tailcall.ll
    M llvm/test/CodeGen/X86/fma-scalar-combine.ll
    M llvm/test/CodeGen/X86/fma.ll
    M llvm/test/CodeGen/X86/fold-rmw-ops.ll
    M llvm/test/CodeGen/X86/pr38865.ll
    M llvm/test/CodeGen/X86/pr61384.ll
    M llvm/test/CodeGen/X86/sse2-intrinsics-x86-upgrade.ll
    M llvm/test/CodeGen/X86/sse2-intrinsics-x86.ll
    M llvm/test/CodeGen/X86/sse41-intrinsics-x86.ll
    M llvm/test/CodeGen/X86/sse41.ll
    M llvm/test/CodeGen/X86/vec_fpext.ll
    M llvm/test/CodeGen/X86/x86-interrupt_cc.ll
    M llvm/test/MC/ELF/mc-dump.s

  Log Message:
  -----------
  MC: Centralize X86 PC-relative fixup adjustment in MCAssembler

Move the X86 PC-relative fixup adjustment from
X86MCCodeEmitter::emitImmediate to MCAssembler, leveraging a generalized
evaluateFixup. This saves a MCBinaryExpr. For `call foo`, the fixup
expression is now `foo` instead of `foo-4`. There is no change in
generated relocations.

In bolt/lib/Target/X86/X86MCPlusBuilder.cpp, createRelocation needs to
decrease the addend.

Both max-rss and instructions:u show a minor decrease.
https://llvm-compile-time-tracker.com/compare.php?from=ea600576a6f94d6f28925c4b99962cc26b463c29&to=016e8fd4ddf851e5555f606c6394241d68f1a7bb&stat=max-rss&linkStats=on

Next: Update targets that use FKF_IsAlignedDownTo32Bits to define
`evaluateFixup` and remove FKF_IsAlignedDownTo32Bits from the generic
code.

Pull Request: https://github.com/llvm/llvm-project/pull/147113



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