Nmap Development mailing list archives

Re: [NSE script] IPv6 RA flood


From: David Fifield <david () bamsoftware com>
Date: Fri, 14 Sep 2012 23:40:21 -0700

On Sat, May 05, 2012 at 05:49:52PM +0200, Adam Števko wrote:
Hi guys,

IPv6 deployment is on the rise and there are some protocol
vulnerabilities. One of them is flooding network with Router
Advertisments causing machines to recompute route table entries
leading to 100% CPU utilization.  Based on my testing, I found these
platforms vulnerable: Windows (was unusable), Solaris (was usable, but
the console lagged a bit). Linux and FreeBSD were unaffected (issue
was fixed few days later it was announced). This work is inspired by
THC IPv6 suite. 

In the future I would like to add support for packet fragmentation and
giving ability to bypass RA Guard, work on more NSE scripts inspired
by tools from THC IPv6 suite and commit them to nmap script library.

Script URL: https://bitbucket.org/xenol/nse-scripts/src/0c9b7397daeb/ipv6-ra-flood.nse

Hi Adam. I'm sorry this script has been ignored for so long.

I changed the script's imports to run under Lua 5.2 and tried it. It
pegged a Windows 7 CPU at 100% almost immediately, and continued to do
so after I killed the Nmap process. It works as advertised.

I'd like to add this script if you'll make just a few simple changes.
Update the code to Lua 5.2; this may be as simple as changing the
require statements. A guide is here: https://secwiki.org/w/Nmap/Lua_5.2.
Don't just choose the first interface from the list if no argument was
given; abort the script. See the prerule of url-snarf for an example.

Since the effect, at least on Windows 7, appears to be persistent, do
you think the script should have a default time limit? It wouldn't be
the first script that runs forever, but I think people are still
surprised when a script does that.

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

Current thread: