diff options
author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-10-28 17:38:44 -0700 |
---|---|---|
committer | Matt Arsenault <arsenm2@gmail.com> | 2019-11-19 19:25:26 +0530 |
commit | b696b9dba7b66307c225c2612def8bbb328805b7 (patch) | |
tree | de7c4d11b305bf3541bc334af8fdd13d7ceae60b /llvm/lib/Target/ARM/ARMISelLowering.cpp | |
parent | 714445e406f043e861024f394fd58eb26c15e51a (diff) | |
download | bcm5719-llvm-b696b9dba7b66307c225c2612def8bbb328805b7.tar.gz bcm5719-llvm-b696b9dba7b66307c225c2612def8bbb328805b7.zip |
DAG: Add function context to isFMAFasterThanFMulAndFAdd
AMDGPU needs to know the FP mode for the function to answer this
correctly when this is removed from the subtarget.
AArch64 had to make this more complicated by using this from an IR
hook, so add an IR typed overload.
Diffstat (limited to 'llvm/lib/Target/ARM/ARMISelLowering.cpp')
-rw-r--r-- | llvm/lib/Target/ARM/ARMISelLowering.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/llvm/lib/Target/ARM/ARMISelLowering.cpp b/llvm/lib/Target/ARM/ARMISelLowering.cpp index c9314007c0a..a33535ecd17 100644 --- a/llvm/lib/Target/ARM/ARMISelLowering.cpp +++ b/llvm/lib/Target/ARM/ARMISelLowering.cpp @@ -14826,7 +14826,8 @@ int ARMTargetLowering::getScalingFactorCost(const DataLayout &DL, /// /// For MVE, we set this to true as it helps simplify the need for some /// patterns (and we don't have the non-fused floating point instruction). -bool ARMTargetLowering::isFMAFasterThanFMulAndFAdd(EVT VT) const { +bool ARMTargetLowering::isFMAFasterThanFMulAndFAdd(const MachineFunction &MF, + EVT VT) const { if (!Subtarget->hasMVEFloatOps()) return false; |