I was talking to my friend Toby earlier today, bitching to him about binary flags. I was stuck on an issue with them and was havng a hell of a time getting it sorted out, since I wasn't sure what the flags needed to be.

Anyways, I was having a hard time explaining to him how it works as he is not a developer, but a network guru. So, I put it in his world, since I could demonstrate to him the same process using IP addresses and subnet masks. Another example is the umask in *nix operatin systems. After trying to explain it, I thought I'd put together some sample code showing how this works in the networking world.

Here is a C example showing how one would get the "valid" subnet mask for an IP address:

Toggle C Source for subnet_lookup.c:

// File: subnet_lookup.c // Build: $ gcc subnet_lookup.c -o subnet_lookup // #include  #include  #include  #include  #include  #include    int main(int argc, char **argv) {   struct sockaddr_in adr_inet;   unsigned msb;   char class;   char *netmask;   char *ip_addr;    if (argc < 2) {        printf("Usage: %s \n", argv[0]);        exit(1);   }   ip_addr = argv[1];    memset(&adr_inet, 0, sizeof adr_inet);   adr_inet.sin_family = AF_INET;   adr_inet.sin_port = htons(9000);   adr_inet.sin_addr.sin_addr = inet_addr(ip_addr);   msb = *(unsigned char *) &adr_inet.sin_addr.s_addr;    if ((msb & 0x80) == 0x00) {       class = 'A';       netmask = "255.0.0.0";   }   else if ((msb & 0xC0) == 0x80) {       class = 'B';       netmask = "255.255.0.0";   }   else if ((msb & 0xE0) == 0xC0) {       class = 'C';       netmask = "255.255.255.0";   }   else if ((msb & 0xF0) == 0xE0) {       class = 'D';       netmask = "255.255.255.255";   }   else {       class = 'E';       netmask = "255.255.255.255";   }    printf("IP Address %s is a class %c address with a netmask of %s\n",       ip_addr, class, netmask);    exit(0); } 

The important part of this starts with:

  msb = *(unsigned char *) &adr_inet.sin_addr.s_addr;    if ((msb & 0x80) == 0x00) {       class = 'A';       netmask = "255.0.0.0";   } 

This returns the IP you specified on the command line in the correct format. We then combine that with a hex value to see if it is equal to another hex value. Since IP addresses and subnet masks are 32bit binary, when &'d together, it's like doing binary addition. And how does that work you might ask? Here is an example:

  1 + 1 = 10 (which is 0, you would carry the 1)   0 + 1 = 1   0 + 0 = 0   1 + 0 + 0 = 1   1 + 1 + 0 = 10 (again this is 0, you would carry the 1)   1 + 1 + 1 = 1 (this would be 11, you would carry the 1 here too) 

Okay, confused yet? Good! No, seriously though, it's pretty easy since you can only use 1's or 0's. Here is a more complete example of how binary values would be added together:

01101 + 10111 = 100100  This is why:    0 1 1 0 1 +  1 0 1 1 1 ------------- 1 0 0 1 0 0 

Not so bad huh? Starting from the right, we take 1 + 1, which REALLY is 10. So, it becomes 0, and you "carry" the 1. With me still? great! So, now we take 1 + 0 + 1, which again is 10. This too becomes 0 and you "carry" the 1 again. This time, we take 1 + 1 + 1, which is 11. If you've been following along, you would make it 1, then "carry" the 1.

Anyways, I've rambled way more then I intended to. If you haven't gotten it figured out by now, you either A) have no hope at all, B) consult the Oracle.