[llvm] [AMDGPU] Remove unnecessary add instructions in ctlz.i8 (PR #77615)

[Matt Arsenault via llvm-commits]

================
@@ -3062,6 +3070,26 @@ static bool isCttzOpc(unsigned Opc) {
   return Opc == ISD::CTTZ || Opc == ISD::CTTZ_ZERO_UNDEF;
 }
 
+SDValue AMDGPUTargetLowering::lowerCTLZResults(SDValue Op,
+                                               SelectionDAG &DAG) const {
+  auto SL = SDLoc(Op);
+  auto Arg = Op.getOperand(0u);
+  auto ResultVT = Op.getValueType();
+
+  if (ResultVT != MVT::i8 && ResultVT != MVT::i16)
+    return {};
+
+  assert(isCtlzOpc(Op.getOpcode()));
+  assert(ResultVT == Arg.getValueType());
+
+  auto const LeadingZeroes = 32u - ResultVT.getFixedSizeInBits();
+  auto NewOp = DAG.getNode(ISD::ZERO_EXTEND, SL, MVT::i32, Arg);
+  auto ShiftVal = DAG.getConstant(LeadingZeroes, SL, MVT::i32);
+  NewOp = DAG.getNode(ISD::SHL, SL, MVT::i32, NewOp, ShiftVal);
+  NewOp = DAG.getNode(Op.getOpcode(), SL, MVT::i32, NewOp);
+  return DAG.getNode(ISD::TRUNCATE, SL, ResultVT, NewOp);
----------------
arsenm wrote:

Unfortunate duplication of generic legalization logic 

https://github.com/llvm/llvm-project/pull/77615