Nmap Development mailing list archives

Re: Adding new NSE discovered targets to Nmap


From: Patrik Karlsson <patrik () cqure net>
Date: Mon, 16 Aug 2010 10:20:55 +0200

Hi Fyodor,

On 16 aug 2010, at 10.09, Fyodor wrote:

On Mon, Aug 16, 2010 at 08:06:28AM +0200, Patrik Karlsson wrote:

On 16 aug 2010, at 04.17, Ron wrote:

On Sun, 15 Aug 2010 19:00:25 -0700 Fyodor <fyodor () insecure org> wrote:

So if setting this enables a useful script to work, and we
don't identify any problems with using the flag, then we might just
set it in all cases rather than provide it as an optional function.

I suspect that, with the new prerule scripts, broadcasting is
going to become rather common.

I agree, I can think of a handful scripts that could use this.
Unfortunately, like I wrote in [1] I don't think changing this alone
will allow the scripts to work.

I agree as well.  We need to figure out how to make this work for a
useful script, then we can check the broadcast change in along with
the new script.  As Patrik notes, we still have a problem with
receiving responses in some cases.  Maybe someone can look at what API
calls the real dhclient and "ping -b" applications use.

This is what the comments to the nsock_connect_udp function say:

"The advantages to having a connected UDP socket (as
   opposed to just specifying an address with sendto() are that we can
   now use a consistent set of write/read calls for TCP/UDP, received
   packets from the non-partner are automatically dropped by the OS,
   and the OS can provide asynchronous errors (see Unix Network
   Programming pp224)."

Wouldn't that suggest that there would be a problem receiving responses in all cases where broadcast is used?


Cheers,
Fyodor



//Patrik
--
Patrik Karlsson
http://www.cqure.net
http://www.twitter.com/nevdull77





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


Current thread: