[llvm] [ARM][TableGen][MC] Change the ARM mnemonic operands to be optional for ASM parsing (PR #83436)

[Simon Tatham via llvm-commits]

================
@@ -4054,6 +4099,65 @@ static MCRegister MatchRegisterName(StringRef Name);
 
 /// }
 
+static bool isDataTypeToken(StringRef Tok) {
+  static const DenseSet<StringRef> DataTypes{
+      ".8",  ".16",  ".32",  ".64",  ".i8", ".i16", ".i32", ".i64",
+      ".u8", ".u16", ".u32", ".u64", ".s8", ".s16", ".s32", ".s64",
+      ".p8", ".p16", ".f32", ".f64", ".f",  ".d"};
+  return DataTypes.contains(Tok);
+}
+
+static unsigned getMnemonicOpsEndInd(const OperandVector &Operands) {
+  unsigned MnemonicOpsEndInd = 1;
+  // Special case for CPS which has a Mnemonic side token for possibly storing
+  // ie/id variant
+  if (Operands[0]->isToken() &&
+      static_cast<ARMOperand &>(*Operands[0]).getToken() == "cps") {
+    if (Operands.size() > 1 && Operands[1]->isImm() &&
+        static_cast<ARMOperand &>(*Operands[1]).getImm()->getKind() ==
+            llvm::MCExpr::Constant &&
+        (dyn_cast<MCConstantExpr>(
+             static_cast<ARMOperand &>(*Operands[1]).getImm())
+                 ->getValue() == ARM_PROC::IE ||
+         dyn_cast<MCConstantExpr>(
+             static_cast<ARMOperand &>(*Operands[1]).getImm())
+                 ->getValue() == ARM_PROC::ID))
+      ++MnemonicOpsEndInd;
+  }
+
+  // In some circumstances the code code moves to the right
+  bool RHSCondCode = false;
+  while (MnemonicOpsEndInd < Operands.size()) {
+    auto Op = static_cast<ARMOperand &>(*Operands[MnemonicOpsEndInd]);
+    // Special case for it instructions which have a condition code on the RHS
+    if (Op.isITMask()) {
+      RHSCondCode = true;
+      MnemonicOpsEndInd++;
+    }
+    // Special case for it instructions which have a condition code on the RHS
+    else if (Op.isToken() &&
+             (
+                 // There are several special cases not covered by
+                 // isDataTypeToken
+                 Op.getToken() == ".w" || Op.getToken() == ".bf16" ||
+                 Op.getToken() == ".p64" || Op.getToken() == ".f16" ||
+                 isDataTypeToken(Op.getToken()))) {
+      // In the mnemonic operators the cond code must always precede the data
+      // type. So we can now safely assume any subsequent cond code is on the
+      // RHS. As is the cdase for VCMP and VPT.
----------------
statham-arm wrote:

(typo, while I'm here anyway: cdase → case)

https://github.com/llvm/llvm-project/pull/83436