[llvm] [AArch64] Fold vector select with power-of-2 bit-test to CMTST+BSP (PR #209100)

[Mugundan S via llvm-commits]

================
@@ -16644,6 +16644,28 @@ static SDValue tryLowerToBSL(SDValue N, SelectionDAG &DAG) {
                            N0->getOperand(1 - i), N1->getOperand(1 - j));
     }
 
+  // Fold: or(and(xor(AArch64ISD::CMTST(X,M), allones), A),
----------------
MGN-GIT wrote:

Yep, I have removed this extra BSP

https://github.com/llvm/llvm-project/pull/209100