On programming without malloc

Re:On programming without malloc by cpeterso on Thursday December 05, 2013 @12:14AM (#6852820)

djb's qmail does almost exactly this! It works surprisingly well for short-lived forked processes.

  typedef union { char irrelevant[ALIGNMENT]; double d; } aligned;
  static aligned realspace[SPACE / ALIGNMENT];
  #define space ((char *) realspace)
  static unsigned int avail = SPACE; /* multiple of ALIGNMENT; 0<=avail<=SPACE */

  char *alloc(n)
  unsigned int n;
  {
    n = ALIGNMENT + n - (n & (ALIGNMENT - 1)); /* XXX: could overflow */
    if (n <= avail) {
      avail -= n;
      return space + avail;
    }
    return malloc(n);
  }

  void alloc_free(x)
  char *x;
  {
    if (x >= space && x < space + SPACE) {
      return; /* XXX: assuming that pointers are flat */
    }
    free(x);
  }

[ Parent | on HN ]

Re:On programming without malloc by a-priori on Thursday December 05, 2013 @10:02AM (#6854674)

Cool, didn't know that! It's basically a degenerate pooled memory allocator. That one is a bit more clever in that it falls through to a proper malloc when it runs out of space. That's assuming, of course, that you have a proper malloc to fall through too.

Mine will eventually wrap around and clobber whatever's at 0x0. It assumes you have a very liberal page allocator in the background to handle page faults for whatever addresses you decide to point to and automatically wire them up to RAM.

[ Parent | on HN ]