diff options
| author | Craig Topper <craig.topper@intel.com> | 2018-01-27 08:41:03 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@intel.com> | 2018-01-27 08:41:03 +0000 |
| commit | c80f0ced84ba856a86889b70701150dad1bab395 (patch) | |
| tree | ab6fbcaf64b1e07775311a2e849a6fcb707033b1 /llvm/lib/CodeGen | |
| parent | 77a5c965607757d3c0f7e83525a41f655bb5db89 (diff) | |
| download | bcm5719-llvm-c80f0ced84ba856a86889b70701150dad1bab395.tar.gz bcm5719-llvm-c80f0ced84ba856a86889b70701150dad1bab395.zip | |
[SelectionDAG] Make DAGTypeLegalizer::PromoteSetCCOperands handle SETEQ/SETNE correctly for vector types.
The code was using getValueSizeInBits and combining with the result of a call to DAG.ComputeNumSignBits. But for vector types getValueSizeInBits returns the width of the full vector while ComputeNumSignBits is going to give a number no larger than the width of a single element. So we should be using getScalarValueSizeInBits to get the element width.
llvm-svn: 323583
Diffstat (limited to 'llvm/lib/CodeGen')
| -rw-r--r-- | llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp index eaa827309b0..31e87692cbd 100644 --- a/llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp @@ -990,11 +990,11 @@ void DAGTypeLegalizer::PromoteSetCCOperands(SDValue &NewLHS,SDValue &NewRHS, // than the width of NewLHS/NewRH, we can avoid inserting real truncate // instruction, which is redudant eventually. unsigned OpLEffectiveBits = - OpL.getValueSizeInBits() - DAG.ComputeNumSignBits(OpL) + 1; + OpL.getScalarValueSizeInBits() - DAG.ComputeNumSignBits(OpL) + 1; unsigned OpREffectiveBits = - OpR.getValueSizeInBits() - DAG.ComputeNumSignBits(OpR) + 1; - if (OpLEffectiveBits <= NewLHS.getValueSizeInBits() && - OpREffectiveBits <= NewRHS.getValueSizeInBits()) { + OpR.getScalarValueSizeInBits() - DAG.ComputeNumSignBits(OpR) + 1; + if (OpLEffectiveBits <= NewLHS.getScalarValueSizeInBits() && + OpREffectiveBits <= NewRHS.getScalarValueSizeInBits()) { NewLHS = OpL; NewRHS = OpR; } else { |
