[clang] [clang]Implement the c23 stdc bit builtins (PR #185978)

[Erich Keane via cfe-commits]

================
@@ -3732,6 +3732,170 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID,
   case Builtin::BI_rotr64:
     return emitRotate(E, true);
 
+  case Builtin::BI__builtin_stdc_leading_zeros:
+  case Builtin::BI__builtin_stdc_leading_ones:
+  case Builtin::BI__builtin_stdc_trailing_zeros:
+  case Builtin::BI__builtin_stdc_trailing_ones:
+  case Builtin::BI__builtin_stdc_first_leading_zero:
+  case Builtin::BI__builtin_stdc_first_leading_one:
+  case Builtin::BI__builtin_stdc_first_trailing_zero:
+  case Builtin::BI__builtin_stdc_first_trailing_one:
+  case Builtin::BI__builtin_stdc_count_zeros:
+  case Builtin::BI__builtin_stdc_count_ones:
+  case Builtin::BI__builtin_stdc_has_single_bit:
+  case Builtin::BI__builtin_stdc_bit_width:
+  case Builtin::BI__builtin_stdc_bit_floor:
+  case Builtin::BI__builtin_stdc_bit_ceil: {
+    Value *ArgValue = EmitScalarExpr(E->getArg(0));
+    llvm::Type *ArgType = ArgValue->getType();
+    llvm::Type *ResultType = ConvertType(E->getType());
+    unsigned BitWidth = ArgType->getIntegerBitWidth();
+    Value *Zero = ConstantInt::get(ArgType, 0);
+    Value *One = ConstantInt::get(ArgType, 1);
+    Value *Result;
+
+    switch (BuiltinIDIfNoAsmLabel) {
----------------
erichkeane wrote:

Switching on the SAME types again is awful, we should be figuring out a way to do this better. One such way would be to do an implementation inside each of the above 'case' statements, and turn the `get arg value` stuff above, and the cleanup below into their own functions.

https://github.com/llvm/llvm-project/pull/185978