Caldera OpenLinux StartKDE Script LD_LIBRARY_PATH Vulnerability
BugTraq ID: 4400
Remote: No
Date Published: Apr 01 2002 12:00A
Relevant URL:
http://www.securityfocus.com/bid/4400
Summary:

OpenLinux is a freely available, open source implementation of the Linux
operating system.  It is maintained and distributed by Caldera.

A problem with the OpenLinux startkde script could lead to arbitrary
library attacks.  The problem is in the initialization of an environment
variable.

The startkde script insecurely initializes the LD_LIBRARY_PATH environment
variable.  When the script is executed, it by default searches the current
working directory.  Any libraries needed by KDE that are found in the
current working directory will be loaded.

This vulnerability requires that a user start KDE with the startkde script
outside of his or her home directory.  Additionally, it requires that the
directory the script is executed in be write accessible to other system
users.

OpenBSD PF TTL Fingerprinting Vulnerability
BugTraq ID: 4401
Remote: Yes
Date Published: Mar 31 2002 12:00A
Relevant URL:
http://www.securityfocus.com/bid/4401
Summary:

PF is a packet filter implementation, available for OpenBSD.  It is
developed and maintained by the OpenBSD Project.

A problem with PF could allow remote users to gain information about open
ports on a system.  The problem is in the response to some types of
traffic.

Under some circumstances, PF sends responses that can allow an attacker to
gain information about the firewall ruleset.  When an attempt is made to
connect to a system via TCP on a port that is filtered by PF, and PF
returns a RST, it is possible to differentiate between filtered and
unfiltered ports.  A port that is filtered will return a RST with a TTL
field set to 128, whereas the operating system returns a value of 64 by
default.

In the event that a system is filtering certain ports, analysis of the TTL
values returned by the system could lead to accurately predicting ports
that are open but have no services running on them, as would be indicated
by a TTL of 64.  For ports filtered by PF, the returned TTL of 128 would
give information about a port that can't be accessed due to PF.

This vulnerability can only be exploited in rulesets which return RST
values for unauthorized connection attempts.  Rulesets that do not return
RSTs but simply drop the incoming packet are not affected.

IPFilter TTL Fingerprinting Vulnerability
BugTraq ID: 4403
Remote: Yes
Date Published: Mar 31 2002 12:00A
Relevant URL:
http://www.securityfocus.com/bid/4403
Summary:

IPFilter is a freely available, open source firewall package written by
Darren Reed.  It is available for multiple platforms, including Unix and
Linux operating systems.

A problem with IPFilter could allow remote users to gain information about
open ports on a system.  The problem is in the response to some types of
traffic.

Under some circumstances, IPFilter sends responses that can allow an
attacker to gain information about the firewall ruleset.  When an attempt
is made to connect to a system via TCP on a port that is filtered by
IPFilter, and IPFilter returns a RST, it is possible to differentiate
between filtered and unfiltered ports.  A port that is filtered by
IPFilter will return a RST with a TTL field set to 60, whereas the
operating system will return it's default TTL value for a RST.

In the event that a system is filtering certain ports, analysis of the TTL
values returned by the system could lead to accurately predicting ports
that are open but have no services running on them, as would be indicated
by a TTL characteristic of the default operating system.  For ports
filtered by IPFilter, the returned TTL of 60 would give information about
a port that can't be accessed due to filtering by IPFilter.

This vulnerability can only be exploited in rulesets which return RST
values for unauthorized connection attempts.  Rulesets that do not return
RSTs but simply drop the incoming packet are not affected.  Firewalls that
filter all ports by default will not exhibit this behavior.

Cyrus SASL LDAP+MySQL Authentication Patch SQL Command Execution Vulnerability
BugTraq ID: 4409
Remote: Yes
Date Published: Apr 02 2002 12:00A
Relevant URL:
http://www.securityfocus.com/bid/4409
Summary:

The Cyrus SASL LDAP+MySQL patch is a freely available, open source
enhancement patch.  It is designed for use on the Unix and Linux operating
systems.

A problem with the patch could make it possible for remote users to gain
access to the mail account of any user.  The problem is in the handling of
user input.

The Cyrus SASL LDAP+MySQL patch is designed to integrate LDAP and MySQL
authentication with Cyrus SASL.  This makes it possible to centralize
authentication data.

Due to a design problem in the patch, users may gain access to the mail
accounts of others.  By passing a specially crafted SQL command to the
password challenge, it is possible to provoke a successful authentication
response from the MySQL server.  This would give access to the mail of the
user specified in the login challenge.

Exploitation of this vulnerability may offer intermitted success through
the use of a string such as ') OR 1=1 HAVING FLOOR(RAND()*100)=1 AND
('1'='1.  If the attacker has knowledge of database layout via another
vulnerability that allows SQL command stuffing, the probability of
exploitation increases significantly.

This problem may allow a remote user to gain access to the mail spool of
the desired user.

[ des trucs avec Oracle et Lotus aussi ]


-
Pour poster une annonce: [EMAIL PROTECTED]

Répondre à