Remote Buffer Overflow in Sendmail

vangogh

Original release date: March 3, 2003
Last revised: March 3, 2003 13:36:45 EST
Source: CERT/CC

A complete revision history can be found at the end of this file.



Systems Affected
Sendmail Pro (all versions)
Sendmail Switch 2.1 prior to 2.1.5
Sendmail Switch 2.2 prior to 2.2.5
Sendmail Switch 3.0 prior to 3.0.3
Sendmail for NT 2.X prior to 2.6.2
Sendmail for NT 3.0 prior to 3.0.3
Systems running open-source sendmail versions prior to 8.12.8, including UNIX and Linux systems


Overview
There is a vulnerability in sendmail that may allow remote attackers to gain the privileges of the sendmail daemon, typically root.



I. Description
Researchers at Internet Security Systems (ISS) have discovered a remotely exploitable vulnerability in sendmail. This vulnerability could allow an intruder to gain control of a vulnerable sendmail server.

Most organizations have a variety of mail transfer agents (MTAs) at various locations within their network, with at least one exposed to the Internet. Since sendmail is the most popular MTA, most medium-sized to large organizations are likely to have at least one vulnerable sendmail server. In addition, many UNIX and Linux workstations provide a sendmail implementation that is enabled and running by default.

This vulnerability is message-oriented as opposed to connection-oriented. That means that the vulnerability is triggered by the contents of a specially-crafted email message rather than by lower-level network traffic. This is important because an MTA that does not contain the vulnerability will pass the malicious message along to other MTAs that may be protected at the network level. In other words, vulnerable sendmail servers on the interior of a network are still at risk, even if the site's border MTA uses software other than sendmail. Also, messages capable of exploiting this vulnerability may pass undetected through many common packet filters or firewalls.

Sendmail has indicated to the CERT/CC that this vulnerability has been successfully exploited in a laboratory environment. We do not believe that this exploit is available to the public. However, this vulnerability is likely to draw significant attention from the intruder community, so the probability of a public exploit is high.

A successful attack against an unpatched sendmail system will not leave any messages in the system log. However, on a patched system, an attempt to exploit this vulnerability will leave the following log message:

Dropped invalid comments from header address

Although this does not represent conclusive evidence of an attack, it may be useful as an indicator.

A patched sendmail server will drop invalid headers, thus preventing downstream servers from receiving them.

The CERT/CC is tracking this issue as VU#398025. This reference number corresponds to CVE candidate CAN-2002-1337.

For more information, please see

http://www.sendmail.org
http://www.sendmail.org/8.12.8.html
http://www.sendmail.com/security/
http://www.iss.net/issEn/delivery/xforce/alertdetail.jsp?oid=21950
http://www.kb.cert.org/vuls/id/398025


II. Impact
Successful exploitation of this vulnerability may allow an attacker to gain the privileges of the sendmail daemon, typically root. Even vulnerable sendmail servers on the interior of a given network may be at risk since the vulnerability is triggered from the contents of a malicious email message.

==========see below url for detail============
http://www.cert.org/advisories/CA-2003-07.html