[X86] Change how we prefer shift by immediate over folding a load into a shift.
BMI2 added new shift by register instructions that have the ability to fold a load.
Normally without doing anything special isel would prefer folding a load over folding an immediate because the load folding pattern has higher "complexity". This would require an instruction to move the immediate into a register. We would rather fold the immediate instead and have a separate instruction for the load.
We used to enforce this priority by artificially lowering the complexity of the load pattern.
This patch changes this to instead reject the load fold in isProfitableToFoldLoad if there is an immediate. This is more consistent with other binops and feels less hacky.
llvm-svn: 335804
diff --git a/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp b/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
index 3382ee8..c833075 100644
--- a/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
+++ b/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
@@ -568,7 +568,20 @@
if (Val.getOpcode() == ISD::TargetGlobalTLSAddress)
return false;
}
+
+ break;
}
+ case ISD::SHL:
+ case ISD::SRA:
+ case ISD::SRL:
+ // Don't fold a load into a shift by immediate. The BMI2 instructions
+ // support folding a load, but not an immediate. The legacy instructions
+ // support folding an immediate, but can't fold a load. Folding an
+ // immediate is preferable to folding a load.
+ if (isa<ConstantSDNode>(U->getOperand(1)))
+ return false;
+
+ break;
}
}