[llvm] [AMDGPU] Promote uniform ops to I32 in ISel (PR #106383)

[Matt Arsenault via llvm-commits]

================
@@ -6746,6 +6748,122 @@ SDValue SITargetLowering::lowerFLDEXP(SDValue Op, SelectionDAG &DAG) const {
   return DAG.getNode(ISD::FLDEXP, DL, VT, Op.getOperand(0), TruncExp);
 }
 
+static unsigned getExtOpcodeForPromotedOp(SDValue Op) {
----------------
arsenm wrote:

I would hope there would be a generic helper for this somewhere 

https://github.com/llvm/llvm-project/pull/106383