================
@@ -2563,9 +2563,15 @@ EmitAsmStores(CodeGenFunction &CGF, const AsmStmt &S,
if ((i < ResultRegIsFlagReg.size()) && ResultRegIsFlagReg[i]) {
// Target must guarantee the Value `Tmp` here is lowered to a boolean
// value.
- llvm::Constant *Two = llvm::ConstantInt::get(Tmp->getType(), 2);
+ unsigned CCUpperBound = 2;
+ if (CGF.getTarget().getTriple().getArch() == llvm::Triple::systemz) {
----------------
anoopkg6 wrote:
Removing llvm.assume intrinsic will cause performance hit. With CC range known,
intrinsic guides optimizer to generate more optimized code. Verified it for
SystemZ.
ResultRegIsFlagReg[i] will be true only if AsmStmt S has its ith
OutputConstraint starts with "{@cc".
I have tried to abstract out Triple check. It adds to one virtual function call
though. Any suggestions for abstracting it in a cleaner way?
https://github.com/llvm/llvm-project/pull/125970
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
