[LegalizeTypes] Expand FNEG to bitwise op for IEEE FP types

Summary:
Except for custom floating point types x86_fp80 and ppc_fp128,
expand Y = FNEG(X) to Y = X ^ sign mask to avoid library call.
 Using bitwise operation can improve code size and performance.

Reviewers: efriedma

Reviewed By: efriedma

Subscribers: efriedma, kpn, arsenm, eli.friedman, javed.absar, rbar, johnrusso, simoncook, sabuasal, niosHD, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, asb, llvm-commits

Differential Revision: https://reviews.llvm.org/D57875

llvm-svn: 353757

1 files changed, 9 insertions, 0 deletions

diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp
index 17dee8d3ed4..ae00d69ad0e 100644
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp
@@ -440,6 +440,15 @@ SDValue DAGTypeLegalizer::SoftenFloatRes_FNEG(SDNode *N, unsigned ResNo) {
     return SDValue(N, ResNo);
   EVT NVT = TLI.getTypeToTransformTo(*DAG.getContext(), N->getValueType(0));
   SDLoc dl(N);
+
+  EVT FloatVT = N->getValueType(ResNo);
+  if (FloatVT == MVT::f32 || FloatVT == MVT::f64 || FloatVT == MVT::f128) {
+    // Expand Y = FNEG(X) -> Y = X ^ sign mask
+    APInt SignMask = APInt::getSignMask(NVT.getSizeInBits());
+    return DAG.getNode(ISD::XOR, dl, NVT, GetSoftenedFloat(N->getOperand(0)),
+                       DAG.getConstant(SignMask, dl, NVT));
+  }
+
   // Expand Y = FNEG(X) -> Y = SUB -0.0, X
   SDValue Ops[2] = { DAG.getConstantFP(-0.0, dl, N->getValueType(0)),
                      GetSoftenedFloat(N->getOperand(0)) };