Product
Solaris 9 Operating System
Solaris 10 Operating System

Bug Id
6458595

Date of Workaround Release
10-OCT-2006

Date of Resolved Release
30-JAN-2007

Impact

A "use-after-free" security vulnerability in sendmail(1M) relating to the handling of long header lines may allow a local or remote unprivileged user to fill up a disk if sendmail(1M) is configured to write unique core files. The core files created by sendmail(1M) would be written to the disk partition configured with coreadm(1M). The ability to consume all available space of a disk partition (which may be the root "/" partition) is a type of denial of service (DoS).

Additional information regarding this issue is available at:

• CVE-2006-4434 at: http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2006-4434

Contributing Factors

This issue can occur in the following releases:

SPARC Platform

• Solaris 9 without patch 113575-08
• Solaris 10 without patch 125011-01

x86 Platform

• Solaris 9 without patch 114137-07
• Solaris 10 without patch 125012-01

Note 1: Sendmail versions prior to 8.12.0 are not impacted by this issue. Solaris 8 uses sendmail versions prior to 8.12.0 and thus is not impacted.

Note 2: This issue only affects systems which have sendmail(1M) versions 8.12.0 to 8.13.7 enabled and have configured sendmail(1M) to create unique core dumps.

To determine the version of sendmail(1M) running on a system, the mconnect(1) command can be used:

    $ /usr/bin/mconnect
    connecting to host localhost (127.0.0.1), port 25
    connection open
    220 an.example.com ESMTP Sendmail 8.13.5+Sun/8.13.5;
    Mon, 20 Mar 2006 17:07:57 GMT
    quit
    221 2.0.0 an.example.com closing connection

If sendmail is not running on the system the mconnect(1) command will report the following:

    $ /usr/bin/mconnect
    connecting to host localhost (127.0.0.1), port 25
    connect: Connection refused

To determine if sendmail(1M) has been configured to create unique core dumps the coreadm(1M) command can be utilized to check the global and per-process core dump settings which affect sendmail processes:

    $ coreadm | grep global
    global core file pattern: /var/cores/core.%f.%p
    global core file content: default
    global core dumps: enabled
    global setid core dumps: disabled
    global core dump logging: disabled
    # coreadm `pgrep sendmail`
    109584:  core.%f.%p      default

 

Symptoms

If the described issue occurs, the symptoms would be a decreasing amount or no free space, as displayed by the df(1M) utility, on the file system partition which sendmail(1M) is configured to write core files to using coreadm(1M) due to a large number of sendmail core dumps.

Workaround

To prevent sendmail(1M) from filling up the configured disk partition with unique core dumps, the coreadm(1M) utility can be used to configure sendmail to create statically named core files. This should be applied to both the global core file pattern (if enabled) and the per-process core file pattern.

To configure sendmail to create core files with static filenames until the sendmail service is restarted:

    # coreadm -p core.%f $(pgrep sendmail)

 

Resolution

This issue is addressed in the following releases:

SPARC Platform

• Solaris 9 with patch 113575-08 or later
• Solaris 10 with patch 125011-01 or later

x86 Platform

• Solaris 9 with patch 114137-07 or later
• Solaris 10 with patch 125012-01 or later

Modification History
Date: 27-NOV-2006

• Upated Contributing Factors and Resolution sections.

Date: 12-DEC-2006

• Updated Relief/Workaround section

Date: 30-JAN-2007

• State: Resolved
• Updated Contributing Factors and Resolution sections

References

113575-08
114137-07
125011-01
125012-01




Attachments

This solution has no attachment