Replace upvalue-based shell dispatch with OP_LUSH opcode

Add a dedicated OP_LUSH A B instruction that loads shell C functions
directly from the registry, eliminating the need for upvalue allocation
in mainfunc(), upvalue population in lua_load(), and the fragile
LUSH_NUM_UPVALS heuristic. Every chunk no longer carries 4 extra
upvalues.

lopcodes.h

@@ -338,6 +338,8 @@ OP_CLOSURE,/*	A Bx	R[A] := closure(KPROTO[Bx])			*/
 
 OP_VARARG,/*	A C	R[A], R[A+1], ..., R[A+C-2] = vararg		*/
 
+OP_LUSH,/*	A B	R[A] := lush_func[B]				*/
+
 OP_GETVARG, /* A B C	R[A] := R[B][R[C]], R[B] is vararg parameter    */
 
 OP_ERRNNIL,/*	A Bx	raise error if R[A] ~= nil (K[Bx] is global name)*/