Re: nmap 4.11 + OpenBSD Kernel PPPOE = strange error

[kx <kxmail () gmail com>]

Honestly, no.

Quote F:

"So I think I'll wait on this until we see more reports and are
able to get a better idea what is going on. But in the meantime, it
is good that people who are annoyed by the issue can apply your patch
(hopefully after they have reported it though!)

Cheers,
-F"

As for, why does -sT work and not -sS?

-sT results in a connect() call, which lets the kernel construct and
route the packet, whereas -sS uses nmap code to construct and route
the packet.

If you would, could you at least report this:

go to  collect_dnet_routes() in tcpip.cc and add some print statements
like this:

/* Now determine which interface the route relates to */
  u32 mask;
  struct sockaddr_in *sin;
  char buf[32];
  for(i = 0; i < dcrn->numifaces; i++) {
    sin = (struct sockaddr_in *) &dcrn->ifaces[i].addr;
    mask = htonl((unsigned long) (0-1) << (32 - dcrn->ifaces[i].netmask_bits));
printf("iface     : %s\n", &dcrn->ifaces[i].devname);
inet_ntop(AF_INET, &sin->sin_addr, buf, sizeof(buf));
printf("iface addr: %s \n", buf);
printf("gw    addr: %s \n", addr_ntoa(&entry->route_gw));
inet_ntop(AF_INET, &mask, buf, sizeof(buf));
printf("mask  addr: %s \n\n", buf);
    if ((sin->sin_addr.s_addr & mask) == (entry->route_gw.addr_ip & mask)) {
      dcrn->routes[dcrn->numroutes].device = &dcrn->ifaces[i];
      break;
    }
  }

Thanks,
  kx



On 7/28/06, Rembrandt <rembrandt () jpberlin de> wrote:
> thanks and sorry for the late reply.
> Is there any roadmap to FIX that?!
>
> It`s NOT documented (at least I didn`t read it anywhere).
> The very interestign part is:
>
> -sT works (mostly?! Everytime I tested it it worked..)
> -sS does NOT work (Routing issues and foo)
>
> OpenBSD using the pppD (userland) uses the TUN interface (works
> perfectly for me). OpenBSD using the kernel pppoe uses the pppoe-Device
> (wich is like a normal NIC) and I get these strange errors.
>
> I´m not sure if it deals with the differences between tun and other
> devices but even if so: Why does -sT work then?
> If it`s a problem then why should -sT work even all others fail (with
> the routing problem. -sT option even resolves the DNS perfectly).
>
> I`ve no time (nor the right skills) to debug this but it ould be realy
> neat if somebody takes care. One free beer at the german CCC-Congress
> for the guy who patches it. :))
>
> Kind regards,
> Rembrandt


_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev