[PATCH] D71353: Fix denormal-fp-math flag and attribute interaction

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: echristo, spatel.
Herald added subscribers: hiraditya, tpr, wdng, jholewinski.
Herald added a project: LLVM.
arsenm added parent revisions: D69878: Consoldiate internal denormal flushing controls, D71258: Set some fast math attributes in setFunctionAttributes, D69978: Separately track input and output denormal mode.

Make this behave similar to unsafe-fp-math and co.

      

AMDGPU has a variety of tests with denormals enabled/disabled that
would require a painful level of test duplication without a flag. This
doesn't expose setting the separate input/output modes, or add a flag
for the f32 version yet.

      

Tests will be included in future patch.


https://reviews.llvm.org/D71353

Files:
  llvm/include/llvm/CodeGen/CommandFlags.inc
  llvm/include/llvm/CodeGen/MachineFunction.h
  llvm/include/llvm/CodeGen/SelectionDAG.h
  llvm/include/llvm/Target/TargetOptions.h
  llvm/lib/CodeGen/MachineFunction.cpp
  llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
  llvm/lib/CodeGen/TargetOptionsImpl.cpp
  llvm/lib/Target/ARM/ARMAsmPrinter.cpp
  llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
  llvm/lib/Target/TargetMachine.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D71353.233348.patch
Type: text/x-patch
Size: 10636 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20191211/85b42b71/attachment.bin>