Re: [nmap-svn] r21154 - nbase

[David Fifield <david () bamsoftware com>]

On Sat, Nov 20, 2010 at 11:19:35PM -0800, commit-mailer () insecure org wrote:
> Author: kris
> Date: Sat Nov 20 23:19:34 2010
> New Revision: 21154
>
> Log:
> Fix a bug in sockaddr_storage_cmp() which could cause it to return the wrong
> value when the address families are different, but typically only with the
> wrong sign for 1.  Practically, it should never return 0 when the sockaddrs
> differ in family since it would also require the address data for one to match
> the corresponding data in the other.
>
> I found this via code review and upon a quick inspection it doesn't appear to
> affect any current uses of this function.  On most (all?) systems it would only
> occur if the first sockaddr was inet6 and the second was inet (I would imagine
> that AF_INET6 > AF_INET is always true, even though these values vary between
> systems).
>
> Doesn't get too much more obscure than that ;)
>
>
>
> Modified:
>    nbase/nbase_misc.c
>
> Modified: nbase/nbase_misc.c
> ==============================================================================
> --- nbase/nbase_misc.c        (original)
> +++ nbase/nbase_misc.c        Sat Nov 20 23:19:34 2010
> @@ -158,7 +158,7 @@
>    const struct sockaddr_storage *b) {
>    if (a->ss_family < b->ss_family)
>      return -1;
> -  else if (a->ss_family < b->ss_family)
> +  else if (a->ss_family > b->ss_family)
>      return 1;
>    if (a->ss_family == AF_INET) {
>      struct sockaddr_in *sin_a = (struct sockaddr_in *) a;

Nice find.

David Fifield
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/