[all-commits] [llvm/llvm-project] 54d755: DAG: Fix incorrect folding of fmul -1 to fneg

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 54d755a034362814bd7a0b90f172cbba39729cf4
      https://github.com/llvm/llvm-project/commit/54d755a034362814bd7a0b90f172cbba39729cf4
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2021-09-14 (Tue, 14 Sep 2021)

  Changed paths:
    M llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
    M llvm/test/CodeGen/AArch64/arm64-fmadd.ll
    M llvm/test/CodeGen/AArch64/fp16_intrinsic_scalar_3op.ll
    M llvm/test/CodeGen/AMDGPU/fneg-combines.ll
    M llvm/test/CodeGen/ARM/fnegs.ll
    M llvm/test/CodeGen/Hexagon/opt-fneg.ll
    M llvm/test/CodeGen/PowerPC/combine-fneg.ll

  Log Message:
  -----------
  DAG: Fix incorrect folding of fmul -1 to fneg

The fmul is a canonicalizing operation, and fneg is not so this would
break denormals that need flushing and also would not quiet signaling
nans. Fold to fsub instead, which is also canonicalizing.