Thinking about how tail call optimization might be implemented

tco.c

#include <stdio.h>

int square(int n) {
  return n * n;
}

int main(int argc, char const *argv[]) {
  int a = 3;
  int b = 2;
  square(a);
  square(b);
  return 0;
}

/* REGISTERS
 * stack pointer:       100
 * instruction pointer: 001

 * /* CODE: MAIN  ...the "CODE" segment...
 * 001  | push 3      <-- IP
 * 002  | push 2
 * 003  | push (SP-2)
 * 004  | push 120
 * 005  | pop
 * 006  | push (sp-1)
 * 007  | pop 120
 * 008  | return 0
 * /* CODE: DOUBLE2 ...
 * 120  | push
 * 121  | multiply
 * 122  | return (SP-??)
 * ...
 * /* STACK:
 * 201 | 000 // a <-- SP
 * 202 | 000 // b
 * 203 | 000
 * 204 | 000
 * 205 | 000
 * 206 | 000
 * 207 | 000
 * ...
 * ...and then we eventually find the heap...
 */