[llvm] r339533 - [SelectionDAG] In PromoteFloatRes_BITCAST, insert a bitcast before the fp16_to_fp in case the input type isn't an i16.

Craig Topper via llvm-commits llvm-commits at lists.llvm.org
Sun Aug 12 22:26:50 PDT 2018


Author: ctopper
Date: Sun Aug 12 22:26:49 2018
New Revision: 339533

URL: http://llvm.org/viewvc/llvm-project?rev=339533&view=rev
Log:
[SelectionDAG] In PromoteFloatRes_BITCAST, insert a bitcast before the fp16_to_fp in case the input type isn't an i16.

The bitcast can be further legalized as needed.

Fixes PR38533.

Added:
    llvm/trunk/test/CodeGen/X86/pr38533.ll
Modified:
    llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp

Modified: llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp?rev=339533&r1=339532&r2=339533&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp Sun Aug 12 22:26:49 2018
@@ -1941,8 +1941,10 @@ void DAGTypeLegalizer::PromoteFloatResul
 SDValue DAGTypeLegalizer::PromoteFloatRes_BITCAST(SDNode *N) {
   EVT VT = N->getValueType(0);
   EVT NVT = TLI.getTypeToTransformTo(*DAG.getContext(), VT);
-  return DAG.getNode(GetPromotionOpcode(VT, NVT), SDLoc(N), NVT,
-                     N->getOperand(0));
+  // Input type isn't guaranteed to be i16 so bitcast if not. The bitcast
+  // will be legalized further if necessary.
+  SDValue Cast = DAG.getBitcast(MVT::i16, N->getOperand(0));
+  return DAG.getNode(GetPromotionOpcode(VT, NVT), SDLoc(N), NVT, Cast);
 }
 
 SDValue DAGTypeLegalizer::PromoteFloatRes_ConstantFP(SDNode *N) {

Added: llvm/trunk/test/CodeGen/X86/pr38533.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/pr38533.ll?rev=339533&view=auto
==============================================================================
--- llvm/trunk/test/CodeGen/X86/pr38533.ll (added)
+++ llvm/trunk/test/CodeGen/X86/pr38533.ll Sun Aug 12 22:26:49 2018
@@ -0,0 +1,11 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown | FileCheck %s
+
+define void @constant_fold_vector_to_half() {
+; CHECK-LABEL: constant_fold_vector_to_half:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movw $16384, (%rax) # imm = 0x4000
+; CHECK-NEXT:    retq
+  store volatile half bitcast (<4 x i4> <i4 0, i4 0, i4 0, i4 4> to half), half* undef
+  ret void
+}




More information about the llvm-commits mailing list