MITKRB5-SA-2010-004 [CVE-2010-1320] double free in KDC

[Tom Yu]

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

MITKRB5-SA-2010-004

MIT krb5 Security Advisory 2010-004
Original release: 2010-04-20

Topic: double free in KDC

CVE-2010-1320

CVSSv2 Vector: AV:N/AC:L/Au:S/C:C/I:C/A:C/E:POC/RL:OF/RC:C

CVSSv2 Base Score:      9

Access Vector:          Network
Access Complexity:      Low
Authentication:         Single
Confidentiality Impact: Complete
Integrity Impact:       Complete
Availability Impact:    Complete

CVSSv2 Temporal Score:  7

Exploitability:         Proof-of-Concept
Remediation Level:      Official Fix
Report Confidence:      Confirmed

SUMMARY
=======

A double free vulnerability exists in the KDC in MIT krb5 releases
krb5-1.7 and later.

This is an implementation vulnerability in MIT krb5, and not a
vulnerability in the Kerberos protocol.

IMPACT
======

An authenticated remote attacker can crash the KDC by inducing the KDC
to perform a double free.  Under some circumstances on some platforms,
this could also allow malicious code execution.  Successfully inducing
code execution by exploiting a double free is believed to be
difficult, and no such exploits are known to exist for this
vulnerability.

AFFECTED SOFTWARE
=================

* KDC in krb5-1.7 and later

FIXES
=====

* The upcoming krb5-1.8.2 release, as well as an upcoming krb5-1.7
  series release, will contain a fix.

* Apply the following patch:

diff --git a/src/kdc/do_tgs_req.c b/src/kdc/do_tgs_req.c
index b2f0655..76ca94a 100644
- --- a/src/kdc/do_tgs_req.c
+++ b/src/kdc/do_tgs_req.c
@@ -543,6 +543,7 @@ tgt_again:
            to the caller */
         ticket_reply = *(header_ticket);
         enc_tkt_reply = *(header_ticket->enc_part2);
+        enc_tkt_reply.authorization_data = NULL;
         clear(enc_tkt_reply.flags, TKT_FLG_INVALID);
     }
 
@@ -554,6 +555,7 @@ tgt_again:
            to the caller */
         ticket_reply = *(header_ticket);
         enc_tkt_reply = *(header_ticket->enc_part2);
+        enc_tkt_reply.authorization_data = NULL;
 
         old_life = enc_tkt_reply.times.endtime - enc_tkt_reply.times.starttime;
 


  This patch is also available at

  http://web.mit.edu/kerberos/advisories/2010-004-patch.txt

  A PGP-signed patch is available at

  http://web.mit.edu/kerberos/advisories/2010-004-patch.txt.asc

REFERENCES
==========

This announcement is posted at:

  http://web.mit.edu/kerberos/advisories/MITKRB5-SA-2010-004.txt

This announcement and related security advisories may be found on the
MIT Kerberos security advisory page at:

        http://web.mit.edu/kerberos/advisories/index.html

The main MIT Kerberos web page is at:

        http://web.mit.edu/kerberos/index.html

CVSSv2:

    http://www.first.org/cvss/cvss-guide.html
    http://nvd.nist.gov/cvss.cfm?calculator&adv&version=2

CVE: CVE-2010-1320
http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-1320

Debian bug #577490
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=577490

ACKNOWLEDGMENTS
===============

Joel Johnson reported this bug to Debian.  Brian Almeida provided a
report that this bug affects krb5-1.7 releases.

CONTACT
=======

The MIT Kerberos Team security contact address is
<krbcore-security at mit.edu>.  When sending sensitive information,
please PGP-encrypt it using the following key:

pub   2048R/8B8DF501 2010-01-15 [expires: 2011-02-01]
uid     MIT Kerberos Team Security Contact <krbcore-security at mit.edu>

DETAILS
=======

When process_tgs_req() handles renewal or validation of existing
tickets, it copies header_ticket->enc_part2 (from the ticket that is
being validated or renewed) to enc_tkt_reply (the new ticket being
generated for the reply).  This causes
enc_tkt_reply.authorization_data to be an alias for memory that
belongs to the request.  As process_tgs_req() exits, it frees both
header_ticket and enc_tkt_reply, which causes the aliased memory to be
freed twice.

In the krb5-1.8 releases, unlike prior MIT krb5 releases, the TGS
request processing normally adds a "signedpath" authorization data
element, which causes merge_authdata() to run.  Inside
merge_authdata() is a call to realloc() that can potentially cause the
authorization data from header_ticket to be freed (if realloc() needs
to relocate the memory) an additional time before the cleanup code at
the end of process_tgs_req().  The krb5-1.7 releases can still
encounter this condition under less common circumstances, such as when
a client provides authorization data that it wants the KDC include in
the new ticket.

REVISION HISTORY
================

2010-04-20      original release

Copyright (C) 2010 Massachusetts Institute of Technology
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (SunOS)

iEYEARECAAYFAkvN7+QACgkQSO8fWy4vZo5+dQCfV+qgkScspXVbQ9rgZ/L8m1Rp
9mcAn31H+mi9pTcmEyhMzRuXD47VVKv3
=W2dB
-----END PGP SIGNATURE-----