[PATCH] D81779: [SelectionDAG] Add MVT::bf16 to getConstantFP()

[Aaron Smith via Phabricator via llvm-commits]

asmith created this revision.
asmith added a reviewer: stuij.
Herald added subscribers: llvm-commits, hiraditya.
Herald added a project: LLVM.

This was probably overlooked in recent bfloat patches.
Needed to handle bf16 constants in SelectionDAG.

  ConstantFP:bf16<APFloat(0)>


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D81779

Files:
  llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp


Index: llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
===================================================================
--- llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -1394,7 +1394,7 @@
   else if (EltVT == MVT::f64)
     return getConstantFP(APFloat(Val), DL, VT, isTarget);
   else if (EltVT == MVT::f80 || EltVT == MVT::f128 || EltVT == MVT::ppcf128 ||
-           EltVT == MVT::f16) {
+           EltVT == MVT::f16 || EltVT == MVT::bf16) {
     bool Ignored;
     APFloat APF = APFloat(Val);
     APF.convert(EVTToAPFloatSemantics(EltVT), APFloat::rmNearestTiesToEven,


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D81779.270552.patch
Type: text/x-patch
Size: 633 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200613/2d31b112/attachment.bin>