Re: nmap.bind to set NSE socket source address

[Tom Sellers <nmap () fadedcode net>]

David Fifield wrote:

> > Take a look at the attached .pcap file -- I did DHCPINFORM, which worked  
> > fine with the changes, then I did DHCPDISCOVER, which didn't.
> >
> > Do you know if there's any way to receive broadcast packets without a pcap?
>
> The DHCP server in my router sends back a unicast DHCP offer in response
> to DHCPDISCOVER, so it must not be every server that does that.
>
> I think what you would normally do is make a socket bound to
> 255.255.255.255:67 and try to receive on it. In Python this works:
>
> import socket
> s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, socket.IPPROTO_UDP)
> s.bind("255.255.255.255", 67)
> print s.recv()
>
> But that currently won't work in NSE, because the socket descriptor
> isn't created until a connection is attempted. There's no way to create
> a socket without also connecting it. Something about the connection
> keeps the receive from working. This is what I tried:
>
> s = nmap.new_socket()
> assert(s:bind("255.255.255.255", 67))
> assert(s:connect(host.ip, 68, "udp"))
> status, data = s:receive()
>
> So we would have to rearchitect part of Nsock to allocate a socket on
> nsi_new, which I think is pretty reasonable.
>
> David Fifield


I have never used them, but would nmap.pcap_open [1] and nmap.pcap_receive be
useful for this?  It looks like some work, but it looks like pcap_open
will take a filter expression that could limit the amount of data you have
to process.

Tom


1.  http://nmap.org/nsedoc/lib/nmap.html#pcap_open

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