-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Greetings OpenBSD Project Team,

While reviewing our vulnerability reports we were notified of a bug in the code 
used by OpenBSD to reassemble IP 
fragments.  We are tracking this issue as VU#243620.  We do not classify this 
as a vulnerability and will not be 
publishing it to our public Vulnerability Notes database at 
http://www.kb.cert.org/vuls/.  Nonetheless, we want to 
assist the reporter in whatever ways we can to help resolve this issue.

Antonios Atlasis ([email protected]) reported that he sent this issue 
to a contact at OpenBSD previously. 
Have you addressed this issue as of this time? If so is there a public 
reference to any available fixes?

A copy of the original report is included at the bottom of this message.   
Please be sure to include VU#243620 in 
the subject when replying to this email. 

If you have any questions or concerns, please let us know.

Best Regards,
Todd
- --
Vulnerability Analysis Team
__________________________________________________________
CERT(R) Coordination Center    |             [email protected]
Software Engineering Institute | Hotline : +1 412.268.7090
Carnegie Mellon University     |     FAX : +1 412.268.6989
Pittsburgh, PA 15213-3890      |      http://www.cert.org/
==========================================================
CERT and CERT Coordination Center are registered in the U.S. Patent and 
Trademark Office.

The Software Engineering Institute is sponsored by the U.S. Department of 
Defense.

- ----- BEGIN ORIGINAL VULNERABILITY REPORT -----

OpenBSD 5.2/5.3 accepts fragments even if the first fragment had been sent 60 
seconds or more 
earlier the last. Specifically:

According to RFC 2460 (section 4.5, p.21):
"If insufficient fragments are received to complete reassembly of a packet
within 60 seconds of the reception of the first-arriving fragment of that
packet, reassembly of that packet must be abandoned and all the fragments
that have been received for that packet must be discarded."

On the contrary, OpenBSD 5.2 and 5.3 (previous versions may also be affected, 
but I have not tested 
them) discards fragments only if the time interval between two consecutive 
fragments is more than 
60 secs. However, it does not discard them if the time interval between the 
first and the last 
fragment is a few minutes or more, as long as the time-interval between two 
consecutive fragments 
is less than 60 secs.

For example, I have found that OpenBSD 5.2/5.3 accept up to 28 fragments with 
30 sec intervals 
between them (this will take up to 14 minutes).

The aforementioned behaviour can have the following security consequences.

- - Easy, accurate and not-detectable fingerprinting of OpenBSD (since other 
major OS conform to the 
RFC regarding the acceptance of delayed fragments.
- - IPS/IDS evasion as long as the target is OpenBSD (since usually IDS/IPS do 
not store incoming 
fragments for up to 14 minutes).
- - Undesirable load of the memory, if this attack is used simultaneously by 
various different, real 
or spoofed, IPv6 sources.

For testing purposes, the ICMPv6 Echo Request messages were used.
I used the default installation of OpenBSD 5.2/5.3.

REFERENCES
RFC 2460, Dec 1998

- ----- END ORIGINAL VULNERABILITY REPORT -----
iQEVAwUBUgjzN+5xi1xMnAiGAQKBKQf/YjzJKkA6ydHUd2LYJSYF8dXUvVRo+6Il
4J7JzoM6G/+qRoc1t3qOWbDP1m0Sd7yKynGGcZP75yA7KZ3neKQ1/ZupkSv0Rr9P
f56nu/m6Y2WNopDmfXAXIAW2I27aOcIybaCFDruKiucvWIW0S5KScijpEBtVATRC
RvuLzXdvQgIZ/mLw8IMHO4Az8afxOCbMHRAF5cY8cSlqpFZgmjgj+wqMg/Pen2Di
ZMjxIjJcFexsoHcU4XBXKV70gvLe/JHQriGL2E/ADr7qWjzJm6d2R0+9MQxSgx3S
ng2FPFK2JV5VNsP8OX+FQO2PEazSB5CBu+GPOwbBE1H5zu/obnGxnw==
=fasp
-----END PGP SIGNATURE-----

Reply via email to