iDefense Security Advisory 10.10.06: FreeBSD ptrace PT_LWPINFO Denial of Service Vulnerability

[Bill Moran]

This report seems pretty vague.  I'm unsure as to whether the alleged
"bug" gives the user any more permissions than he'd already have?  Anyone
know any details?

FreeBSD ptrace PT_LWPINFO Denial of Service Vulnerability

iDefense Security Advisory 10.10.06
http://www.idefense.com/intelligence/vulnerabilities/
Oct 10, 2006

I. BACKGROUND

FreeBSD is a modern operating system for x86, amd64, Alpha, IA-64, PC-98
and SPARC architectures. It's based on the UNIX operating system, BSD,
which was created at the University of California, Berkeley.  More
information can be obtained from the FreeBSD Project web site at
http://www.FreeBSD.org/

II. DESCRIPTION

The PT_LWPINFO ptrace command allows a tracer to get information on a
running thread.

Due to the use of signed integers and a lack of proper input validation,
a situation can occur in the kernel where a panic will cause DoS. The
affected code follows.

953         case PT_LWPINFO:
954                 if (data == 0 || data > sizeof(*pl))
955                         return (EINVAL);

Since the "data" variable is a signed integer, the check on line 954 can
easily be bypassed. Eventually, the negative value is passed to
copyout(), which will result in a kernel panic or corruption of the user
space memory.

III. ANALYSIS

Exploitation of this vulnerability would result in a denial of service
condition on the affected host. In some cases exploitation resulted in a
hard lock up of the machine, where as other times a kernel panic was
caused leading to reboot.

iDefense considers this a LOW severity vulnerability due to the local
access requirement.

IV. DETECTION

iDefense has confirmed the existence of this problem in FreeBSD version
6.0-RELEASE. FreeBSD 6.1-RELEASE is not affected. It is suspected that
other versions are also affected.

V. WORKAROUND

iDefense is not aware of any workaround for this issue.

VI. VENDOR RESPONSE

"The policy of the FreeBSD Security Team is that local denial of service
bugs
not be treated as security issues; it is possible that this problem will be
corrected in a future Erratum."

VII. CVE INFORMATION

The Common Vulnerabilities and Exposures (CVE) project has assigned the
name CVE-2006-4516 to this issue. This is a candidate for inclusion in
the CVE list (http://cve.mitre.org), which standardizes names for
security problems.

VIII. DISCLOSURE TIMELINE

08/18/2006  Initial vendor notification
10/06/2006  Initial vendor response
10/10/2006  Public disclosure


-- 
Bill Moran

Sometimes I think I'm stupid.  The rest of the time I'm sure of it.