diff options
Diffstat (limited to 'llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp')
| -rw-r--r-- | llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp | 11 |
1 files changed, 9 insertions, 2 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp index 8fcabeba5ed..37c8de93be5 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp @@ -611,7 +611,7 @@ unsigned GCNTTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index, } bool GCNTTIImpl::areInlineCompatible(const Function *Caller, - const Function *Callee) const { + const Function *Callee) const { const TargetMachine &TM = getTLI()->getTargetMachine(); const FeatureBitset &CallerBits = TM.getSubtargetImpl(*Caller)->getFeatureBits(); @@ -620,7 +620,14 @@ bool GCNTTIImpl::areInlineCompatible(const Function *Caller, FeatureBitset RealCallerBits = CallerBits & ~InlineFeatureIgnoreList; FeatureBitset RealCalleeBits = CalleeBits & ~InlineFeatureIgnoreList; - return ((RealCallerBits & RealCalleeBits) == RealCalleeBits); + if ((RealCallerBits & RealCalleeBits) != RealCalleeBits) + return false; + + // FIXME: dx10_clamp can just take the caller setting, but there seems to be + // no way to support merge for backend defined attributes. + AMDGPU::SIModeRegisterDefaults CallerMode(*Caller); + AMDGPU::SIModeRegisterDefaults CalleeMode(*Callee); + return CallerMode.isInlineCompatible(CalleeMode); } void GCNTTIImpl::getUnrollingPreferences(Loop *L, ScalarEvolution &SE, |
