bugtraq - Re: IPv4 fragmentation --> The Rose Attack

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date: Thu, 8 Apr 2004 17:24:57 +0200 (CEST)
From: Paul Starzetz <ihaquer@...c.pl>
To: bugtraq@...urityfocus.com
Cc: gandalf@...ital.net
Subject: Re: IPv4 fragmentation  --> The Rose Attack

gandalf@...ital.net wrote:

>The attack is simple.  Two parts of a fragmented packet are sent to the
>machine being attacked.  The first fragment (payload 32 bytes long) is the
>initial offset zero fragment of a SYN packet.  The final (second) fragment
>of the SYN packet is also 32 bytes in size, but is set to an offset of 64800
>bytes into the datagram.

There is a similar fragmentation attack which works pretty nice for Linux. 
From the source code of ip_fragment.c follows that the worst case is if 
you send small fragments of a datagram beginning from 0 to lets say 60000 
in pieces of 8 bytes each. This will cause the defragmentation code to 
build a linear list of socket buffers. If you now continue to send the last 
fragment, the kernel will cycle over that list over and over, for every 
packet and finally kfree the last fragment and replace it by the new one.

That causes a really nice load... It killed at least a 2.4.25 running on 
an Athlon 850.

-- 
Paul Starzetz
iSEC Security Research
http://isec.pl/