[llvm] [AMDGPU] Remove AMDGPUISD::FFBH_I32 and add ISD::CTLS lowering (PR #187694)

[Matt Arsenault via llvm-commits]

================
@@ -3418,6 +3420,20 @@ SDValue AMDGPUTargetLowering::LowerCTLZ_CTTZ(SDValue Op, SelectionDAG &DAG) cons
   return DAG.getNode(ISD::ZERO_EXTEND, SL, MVT::i64, NewOpr);
 }
 
+SDValue AMDGPUTargetLowering::LowerCTLS(SDValue Op, SelectionDAG &DAG) const {
+  SDLoc SL(Op);
+  SDValue Src = Op.getOperand(0);
+  assert(Src.getValueType() == MVT::i32 && "LowerCTLS only supports i32");
+  unsigned BitWidth = Src.getValueType().getScalarSizeInBits();
----------------
arsenm wrote:

If you're going to assert i32, might as well hardcode 32 

https://github.com/llvm/llvm-project/pull/187694