Nmap Development mailing list archives

Re: new Win install fails beyond localhost


From: David Fifield <david () bamsoftware com>
Date: Wed, 12 May 2010 22:32:41 -0600

On Wed, May 12, 2010 at 04:08:58PM -0500, Norris Carden wrote:
David,

Here ya go. Looks like you did a -vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv

Thanks. With this information I think I have solved it in r17542. Please
try --iflist with

http://nmap.org/dist/nmap-5.30BETA1-debug2-win32.zip

What's different in this version is that libdnet now also compares
textual interface description as well as MAC addresses when mapping dnet
names to pcap names. Using only the MAC addresses failed with the teamed
interface because there were three interfaces with the same MAC address:
the two physical NICs and the team. When eth0 was assigned to one of the
physical NICs, it could send packets on that interface, but would not
necessarily receive replies on the same interface.

What makes this difficult is we're trying to match up information coming
from two different sources. dnet gets its interface list from a Windows
function called GetIfTable while pcap gets it by reading the registry.
There are not a lot of values that are common to both data sets. MAC
address is one of these. The description returned by GetIfTable in
MIB_IFROW.bDescr is not the same as the description returned by
pcap_findalldevs in pcap_if_t.description, but the latter appears to be
the same as what you get when you call PacketRequest with an OID of
OID_GEN_FRIENDLY_NAME. That's what I've done in this patch.

David Fifield

Attachment: nmap-5.30BETA1-intf-debug2.diff
Description:

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

Current thread: