[llvm] [AMDGPU][SDAG] Legalise v2i32 or/xor/and instructions to make use of 64-bit wide instructions (PR #140694)

[Matt Arsenault via llvm-commits]

================
@@ -1986,6 +1994,13 @@ bool SITargetLowering::shouldConvertConstantLoadToIntImm(const APInt &Imm,
   return true;
 }
 
+bool SITargetLowering::shouldFoldSelectWithIdentityConstant(unsigned BinOpcode,
+                                                            EVT VT) const {
+  return (BinOpcode == ISD::AND || BinOpcode == ISD::OR ||
+          BinOpcode == ISD::XOR) &&
+         VT.getScalarType() == MVT::i64;
+}
----------------
arsenm wrote:

Can this be done separately 

https://github.com/llvm/llvm-project/pull/140694