[llvm] [TableGen][RISCV] Support sub-operands in CompressInstEmitter.cpp. (PR #133039)

[Craig Topper via llvm-commits]

================
@@ -207,90 +207,109 @@ void CompressInstEmitter::addDagOperandMapping(const Record *Rec,
                                                const CodeGenInstruction &Inst,
                                                IndexedMap<OpData> &OperandMap,
                                                bool IsSourceInst) {
+  unsigned NumMIOperands = 0;
+  for (const auto &Op : Inst.Operands)
+    NumMIOperands += Op.MINumOperands;
+  OperandMap.grow(NumMIOperands);
+
   // TiedCount keeps track of the number of operands skipped in Inst
   // operands list to get to the corresponding Dag operand. This is
   // necessary because the number of operands in Inst might be greater
   // than number of operands in the Dag due to how tied operands
   // are represented.
   unsigned TiedCount = 0;
-  for (unsigned I = 0, E = Inst.Operands.size(); I != E; ++I) {
-    int TiedOpIdx = Inst.Operands[I].getTiedRegister();
+  unsigned OpNo = 0;
+  for (const auto &Opnd : Inst.Operands) {
+    int TiedOpIdx = Opnd.getTiedRegister();
     if (-1 != TiedOpIdx) {
       // Set the entry in OperandMap for the tied operand we're skipping.
-      OperandMap[I].Kind = OperandMap[TiedOpIdx].Kind;
-      OperandMap[I].Data = OperandMap[TiedOpIdx].Data;
-      TiedCount++;
+      OperandMap[OpNo].Kind = OperandMap[TiedOpIdx].Kind;
+      OperandMap[OpNo].Data = OperandMap[TiedOpIdx].Data;
+      ++OpNo;
+      ++TiedCount;
       continue;
     }
-    if (const DefInit *DI = dyn_cast<DefInit>(Dag->getArg(I - TiedCount))) {
-      if (DI->getDef()->isSubClassOf("Register")) {
-        // Check if the fixed register belongs to the Register class.
-        if (!validateRegister(DI->getDef(), Inst.Operands[I].Rec))
+    for (unsigned SubOp = 0; SubOp != Opnd.MINumOperands; ++SubOp, ++OpNo) {
----------------
topperc wrote:

I'm not sure this provides much value. Opnd is already a const reference so it can't change out from under us and we're not calling a function that needs to calculate the size.

https://github.com/llvm/llvm-project/pull/133039