[OpenAFS-announce] OpenAFS Security Advisory 2003-002: Rx connection hijacking

Nickolai Zeldovich kolya@MIT.EDU
Fri, 18 Apr 2003 16:06:11 -0400 (EDT)

Hash: SHA1

		OpenAFS Security Advisory 2003-002

Topic: Rx connection hijacking vulnerability in OpenAFS clients and servers

Issued:		18-Apr-2003
Last Update:	18-Apr-2003
Affected:	OpenAFS 1.0 - 1.2.7, OpenAFS 1.3.0 - 1.3.2

A remote user can hijack Rx connections between unpatched AFS clients and
servers.  This allows the attacker to observe and modify the data stream,
unless rxkad mode crypt ("fs setcrypt on") is used.


There is a bug in the Rx RPC protocol, used by AFS, which can be exploited
by an attacker to hijack arbitrary Rx connections.  This allows the
attacker to mount a denial of service attack by breaking arbitrary Rx
connections.  Additionally, unless encryption is used, such as rxkad mode
crypt ("fs setcrypt on") and the user accessing files is authenticated
(has valid tokens), the attacker can observe and modify the data being

The AFS cache manager and other AFS administrative clients (such as pts,
fs, vos, etc) are vulnerable to these attacks.  Vulnerable AFS servers
allow connections from AFS cache managers to be hijacked, but not
connections from the other AFS administrative clients (such as pts, fs,
vos, etc).

There are no known publicly-available exploits for this vulnerability at
this time.


A remote attacker can cause communication failure between any AFS server
and client, by hijacking the Rx connection between them. In cases when
rxkad mode crypt is not turned on, or users are accessing AFS without
tokens, a remote attacker can observe and modify the data being
transferred.  This allows the attacker to read and modify the contents of
any files accessible to users of vulnerable AFS clients.

If executable files are ran from AFS on a vulnerable client, the attacker
can inject malicious code of his choice, possibly causing a compromise of
the client machine.  The attacker can also inject data to make an
executable file in AFS appear to be setuid root to the client, causing
their code to execute as root.


All releases of OpenAFS 1.0.x and 1.1.x.
All releases of OpenAFS 1.2.x, up to and including OpenAFS 1.2.7.
All releases of OpenAFS 1.3.x, up to and including OpenAFS 1.3.2.


The OpenAFS project recommends that all users upgrade to OpenAFS version
1.2.9 or newer.  Note that both clients and servers must be upgraded in
order to fix this vulnerability.  The latest stable OpenAFS release is
always available from http://www.openafs.org/release/latest.html.

No update is presently available for the OpenAFS-unstable series.

For those who are unable to upgrade, apply the patch referenced below to
correct the Rx hijacking vulnerability, and rebuild your tree.

This patch may be found at:


The associated detached PGP signature is at:


It was generated against OpenAFS 1.2.7, but should apply to earlier
releases, possibly with some offset.

This announcement and code patches related to it may be found on the
OpenAFS security advisory page at:


The main OpenAFS web page is at:



Thanks to Jeffrey Hutzelman <jhutz@cs.cmu.edu> for his help in analyzing
and fixing this vulnerability.


The Rx RPC protocol, used by AFS to communicate between clients and
servers, supports multi-homed hosts by, in some cases, re-routing Rx
connections to new IP addresses.  This feature can be exploited by an
attacker to re-route arbitrary Rx connections to a host of their choice,
by sending a special Rx packet that causes the recipient to believe that
the peer has changed its IP address.

The Rx protocol will allow connections to be re-routed in all cases except
for connections from administrative AFS clients, such as pts or vos, on
the server.  That is, the server will not re-route such an Rx connection
(although the client will).  This means the attacker can only observe or
modify the traffic sent by the client to the server, in such a scenario.

Version: GnuPG v1.2.1 (SunOS)