[PATCH] D109446: DAG: Fix incorrect folding of fmul -1 to fneg

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: cameron.mcinally, spatel, efriedma, andrew.w.kaylor, kpn.
Herald added subscribers: ecnelises, kerbowa, hiraditya, nhaehnle, jvesely.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

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.


https://reviews.llvm.org/D109446

Files:
  llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  llvm/test/CodeGen/AArch64/arm64-fmadd.ll
  llvm/test/CodeGen/AArch64/fp16_intrinsic_scalar_3op.ll
  llvm/test/CodeGen/AMDGPU/fneg-combines.ll
  llvm/test/CodeGen/ARM/fnegs.ll
  llvm/test/CodeGen/Hexagon/opt-fneg.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D109446.371358.patch
Type: text/x-patch
Size: 7169 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210908/7d1b00f2/attachment.bin>