FreeBSD System Call Signed Integer Buffer Overflow Vulnerability
BugTraq ID: 5493
Remote: No
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5493
Summary:

A vulnerability has been reported for the FreeBSD system. Reportedly, a
few system calls are vulnerable to signed integer buffer overflow
conditions.

The vulnerability is the result of system calls assuming that some
arguments were given as positive integers while, in actuality, the
arguments were handled as signed integers. If a negative value was
supplied for the argument, the boundary checking code would fail.

This results in the kernel returning a large portion of kernel memory
which may contain sensitive information including passwords. An attacker
may be able to use the information obtained to elevate privileges.

This vulnerability has been reported to affect the accept(2),
getsockname(2), and getpeername(2) system calls, and the vesa(4)
FBIO_GETPALETTE ioctl(2). The associated files in the kernel source are:

src/sys/i386/isa/vesa.c src/sys/kern/uipc_syscalls.c
src/sys/conf/newvers.sh
Welcome to the SecurityFocus.com 'week in review' newsletter issue

FreeBSD has reported that all versions of FreeBSD, up to and including
4.6.1-RELEASE-p10, are vulnerable to this issue.

PostgreSQL cash_words Function Buffer Overflow Vulnerability
BugTraq ID: 5497
Remote: No
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5497
Summary:

PostgreSQL is a freely distributed Object-Relational DBMS.

A buffer overflow vulnerability has been reported for PostgreSQL.
Reportedly, PostgreSQL doesn't properly handle overly long queries when
selecting the cash_words() function.

It is possible to cause the database server process to crash when issuing
a cash_words() function as follows:

psql> select cash_words('-700000000000000000000000000000');

It is believed that an attacker could potentially exploit this condition
to overwrite stack variables with attacker-supplied values.  It is highly
possible that exploitation could result in execution of malicious
attacker-supplied code as the database server process.

This vulnerability has been reported for PostgreSQL versions 7.2 and
earlier.

nCipher PKCS#11 Symmetric Message Signature Verification Vulnerability
BugTraq ID: 5498
Remote: Unknown
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5498
Summary:

nCipher produces a range of hardware and software security products which
support a range of cryptographic operations. A vulnerability has been
reported in the nCipher cryptographic library, related to the checking of
some message signatures.

The RSA PKCS#11 specification allows the signing of messages with a
symmetric key. Verification of these signatures is supported by the
nCipher cryptographic library. However, an error in the library
implementation may result in incorrect results being returned when
signatures are verified.

Under some conditions, the vulnerable function C_Verify may return the
'CKR_OK' message when an invalid signature is verified. The
'CKR_SIGNATURE_INVALID' message would normally be expected under this
condition. As a result, products and processes which rely on this library
function may make erroneous trust decisions regarding messages with
invalid signatures.

The consequences of exploitation will be highly dependent on the nature of
the application using the vulnerable library. It is likely that
exploitation will allow an attacker to inject or modify encrypted
information which is normally protected by a signature. Impersonation of
trusted parties may be possible.

Reportedly, the vulnerable signature mechanism is used by a number of
common protocols, including SSLv2, SSH and IPSEC.

This issue exists in versions 1.2.0 and later of the nCipher cryptographic
library.

Lynx Command Line URL CRLF Injection Vulnerability
BugTraq ID: 5499
Remote: Yes
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5499
Summary:

Lynx is a freely distributable, text-based WWW client. It is available for
use on various operating systems and platforms including Linux and Unix
variant and Microsoft Windows operating environments.

A CRLF injection vulnerability has been reported for Lynx that may allow
an attacker to include extra HTTP headers when viewing web pages. If Lynx
is called from the command line, carriage return and line feed (CRLF)
characters may be included in the specified URL. These characters are not
escaped when the input is used to construct a HTTP request.

As CRLF is used as a delimiter between headers under the HTTP protocol,
exploitation of this vulnerability will result in additional headers being
included in the HTTP request.

Injection of a 'Host' header may cause the request to be serviced as if
made to a different domain, if the server in question supports multiple
hosts. It may also be possible to inject arbitrary cookie data.

It is still possible for attackers to exploit this vulnerability even if
the '-realm' and '-restrictions=useragen' options are used. Reportedly, it
is also possible for an attacker to contact other type of servers,
including POP3 servers and MTAs (Mail Transfer Agents).

This vulnerability has been reported for Lynx versions 2.8.4rel.1,
2.8.5dev.8, 2.8.3rel.1 and 2.8.2rel.1. It is not known whether other
versions are affected.

*** Links 0.9.6 and ELinks have also been reported as being vulnerable.
Some versions of Links and ELinks URL encode space characters so an
attacker needs to use tab characters, instead of spaces, to exploit the
issue on these browsers.

W3C Jigsaw Proxy Server Cross-Site Scripting Vulnerability
BugTraq ID: 5506
Remote: Yes
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5506
Summary:

The W3C Jigsaw project includes a HTTP proxy server written in Java.

When the proxy server cannot successfully resolve a fully qualified domain
name, an error page is served to the client.  The requested URL is
included in the content of this page without being adequately sanitized.
Consequently, embedded script code may execute within the context of the
requested URL (and it's domain).  Exploitation may result in theft of
cookie information or impersonation of websites associated with the
domain.

MySQL Null Root Password Weak Default Configuration Vulnerability
BugTraq ID: 5503
Remote: Yes
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5503
Summary:

MySQL is is an open source relational database project, and is available
for a number of operating systems, including Microsoft Windows.

A weak default configuration problem has been reported in some versions of
MySQL. Reportedly, the root user of the database is defined with no
password, and granted login privileges from any host.

Users unaware of this may fail to define a strong password for the root
user. While the MySQL security documentation does suggest verifying that
the root user has a password defined, an inexperienced administrator may
overlook this step.

Exploitation of this issue can allow a remote attacker to connect to the
database with full privileges. Exploitation may result in access to
sensitive information, or allow denial of service attacks through the
destruction of data.

This issue has been reported in the Windows binary release of MySQL. Other
versions may share this default configuration, this has not however been
confirmed.

MySQL Bind Address Not Enabled Weak Default Configuration Vulnerability
BugTraq ID: 5511
Remote: Yes
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5511
Summary:

MySQL is is an open source relational database project, and is available
for a number of operating systems, including Microsoft Windows.

MySQL supports the 'bind-address' configuration directive. This restricts
database access to the defined address. If remote administration is not
required, this variable may be set to the loopback address 127.0.0.1,
preventing access from any remote system.

This option is not enabled by default, possibly allowing remote access to
default installations of the server. The MySQL security documentation
does, however, suggest restricting remote access to the server to only
required hosts.

This issue has been reported in the Windows binary release of MySQL. Other
versions may share this default configuration, this has not however been
confirmed.

MySQL Logging Not Enabled Weak Default Configuration Vulnerability
BugTraq ID: 5513
Remote: Yes
Date Published: Aug 19 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5513
Summary:

MySQL is is an open source relational database project, and is available
for a number of operating systems, including Microsoft Windows.

Reportedly, most logging is disabled by default in MySQL. If not
explicitely enabled, an administrator may not detect malicious actions or
attacks against the database.

Logging of errors may, however, be enabled by default.

This issue has been reported in the Windows binary release of MySQL. Other
versions may share this default configuration, this has not however been
confirmed.

Mozilla Bonsai Multiple Cross Site Scripting Vulnerabilities
BugTraq ID: 5516
Remote: Yes
Date Published: Aug 20 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5516
Summary:

Mozilla Bonsai is a tool that allows a user to perform queries on the
contents of a CVS archive.

Multiple cross site scripting vulnerabilities have been reported for the
Bonsai tool.

An attacker may exploit this vulnerability by causing a victim user to
follow a malicious link. Attacker-supplied code may execute within the
context of the site hosting the vulnerable software when the malicious
link is visited.

This type of vulnerability may be used to steal cookies or perform other
web-based attacks. It may be possible to take actions as an user of the
Bonsai system.

This vulnerability has been reported for Mozilla Bonsai 1.3 (including all
current and CVS versions).

Mozilla Bonsai Path Disclosure Vulnerability
BugTraq ID: 5517
Remote: Yes
Date Published: Aug 20 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5517
Summary:

Mozilla Bonsai is a tool that allows a user to perform queries on the
contents of a CVS archive.

A path disclosure vulnerability has been reported in Mozilla Bonsai. This
issue is reported to affect all current and CVS versions of the utility.

An attacker can exploit this vulnerability by making a malformed request
to Bonsai. This causes Bonsai to return an error page to the requesting
user. This error page will contain the absolute path information about the
requested file.

Information disclosed in this manner may be used by remote attackers in
intelligence gathering and may aid in further attacks against the
vulnerable host.

SCPOnly SSH Environment Shell Escaping Vulnerability
BugTraq ID: 5526
Remote: No
Date Published: Aug 20 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5526
Summary:

scponly is a freely available, open source restricted secure copy client.
It is available for Unix and Linux operating systems.

A problem with scponly could make it possible for a user to gain
unintended access to a system running a vulnerable version.

The default installation of scponly does not place sufficient access
controls on the .ssh subdirectory.  Due to this oversight, it is possible
for a remote user to upload files which may allow command execution.
This could lead to unintended command execution, and regular shell access
to a vulnerable host.

The problem is in the environment file contained within the .ssh
subdirectory.  If this file is installed with permissions that allow user
modification to the file (which is the default behavior), a user would be
able to upload a new version of this file.  This file could contain
malicious commands, such as changing the user's shell, and would be
executed by the user upon the next log-in.

PostgreSQL Repeat Function Buffer Overflow Vulnerability
BugTraq ID: 5527
Remote: Yes
Date Published: Aug 20 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5527
Summary:

PostgreSQL is a freely distributed Object-Relational DBMS.

An overflow vulnerability has been reported in some versions of
PostgreSQL. The issue lies in the handling of large integer arguments by
the repeat() function, used to create a text string with a specified
number of copies of the original argument.

Reportedly, if this function is called with an extremely large integer
argument, memory allocated on the heap will be overflowed. An attacker
able to call this function may corrupt adjacent data. It may be possible
to corrupt control structures used by some heap implementations and force
the database process to execute arbitrary, attacker-supplied code.

PostgreSQL String Pad Function Buffer Overflow Vulnerability
BugTraq ID: 5528
Remote: No
Date Published: Aug 20 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5528
Summary:

PostgreSQL is a freely distributed Object-Relational DBMS.

A buffer overflow vulnerability has been reported for PostgreSQL.
Reportedly, PostgreSQL doesn't properly handle overly large integer
arguments given to the lpad() and rpad() funtions. The functions are
lpad() and rpad() found in the file,
src/backend/utils/adt/oracle_compat.c, and serve to pad an existing text
string with another up to a given length.

This vulnerability only affects data bases that were created using special
international encodings. For example, databases that were created using a
'UNICODE' encoding are vulnerable to this issue.

It is possible to cause the database server process to crash when issuing the lpad() 
or rpad() function as follows:

my_db=# select lpad('xxxxx',1431655765,'yyyyyyyyyyyyyyyy'); my_db=# select
rpad('xxxxx',1431655765,'yyyyyyyyyyyyyyyy');

This will cause PostgreSQL to improperly allocate space on the system
stack. Thus, it is believed that an attacker could potentially exploit
this condition to overwrite stack variables with malicious
attacker-supplied values. It is highly possible that exploitation could
result in execution of malicious attacker-supplied code as the database
server process.

Reportedly, databases created with EUC_JP, EUC_CN, EUC_KR, EUC_TW,
UNICODE, or MULE_INTERNAL encodings are vulnerable to this issue.


Linux Kernel 2.4.18 Security Issues
BugTraq ID: 5539
Remote: Unknown
Date Published: Aug 21 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5539
Summary:

Red Hat has issued an advisory reporting the correction of several
vulnerabilities in version 2.4.18 of the Linux kernel.  Some of the
security issues are related to the following device drivers:

stradis rio500 se401 usbvideo apm

Furhermore, vulnerabilities reportedly exist in components of the procfs
virtual filesystem that may cause kernel memory to be exposed.  It should
be assumed that at the very least, local attackers may exploit these
vulnerabilities to elevate privileges.

SecurityFocus is currently completing analysis of the reported
vulnerabilities and will issue individual alerts for each.

Apache Tomcat 4.1 JSP Request Cross Site Scripting Vulnerability
BugTraq ID: 5542
Remote: Yes
Date Published: Aug 21 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5542
Summary:

Jakarta Tomcat is a Java Servlet and JSP server produced by the Apache
Software Foundation. Tomcat is available for Microsoft Windows, Linux, and
other Unix based operating systems.

A cross site scripting vulnerability has been reported in some versions of
Tomcat. Reportedly, if a HTTP request is made for a JSP, malicious script
code embedded in the URI may be included in a page generated by Tomcat.

An attacker may generate a link to a vulnerable site, and include
arbitrary malicious script code. If a user is enticed into following this
link, the supplied code will be returned by the server, and execute within
the context of the vulnerable site.

Exploitation may result in the disclosure of sensitive cookie data, or the
ability to take actions as an authenticated user of the vulnerable site.
The consequences of exploitation will be highly dependant on the details
of the vulnerable site.

This may be related to the issues discussed in BID 2982. This has not,
however, been confirmed.

Multiple Vendor IPv4-IPv6 Transition Address Spoofing Vulnerability
BugTraq ID: 5545
Remote: Yes
Date Published: Aug 22 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5545
Summary:

IPv6 is a protocol designed to replace IPv4. IPv6 allows for the
encapsulation of IPv4 addresses, in order to facilitate transition between
the two standards, and allow the usage of IPv4 legacy applications under
IPv6 networking.

Additionally, many systems are expected to support both IPv4 and IPv6
traffic, in order to allow a transition period between the two standards.

Malicious parties may be able to abuse this feature in order to spoof IPv4
addresses. Under some circumstances, IPv4 addresses may be extracted from
IPv6 traffic and passed to applications. These applications will not be
able to distinguish between legitimate IPv4 traffic and that embedded in
IPv6 traffic.

If trust decisions are made based on this information, an attacker may be
able to bypass some security measures. For example, certain applications
may restrict access to a limited range of IPv4 addresses, only allow
access from the loopback address 127.0.0.1, or perform reverse lookup
checks on IPv4 addresses.

The details and consequences of exploitation will be highly dependant on
the specifics of deployed applications. It may be possible to gain
unauthorized access to systems, or to generate malicious network traffic
through the usage of the loopback address or broadcast addresses.

D-Link Remote Administration Arbitrary DHCP Address Release Vulnerability
BugTraq ID: 5544
Remote: Yes
Date Published: Aug 22 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5544
Summary:

The DI-804 is a hardware gateway and firewall solution distributed and
maintained by D-Link.

A problem with the DI-804 could make it possible for a remote user to deny
service to legitimate users of the device.

It has been reported that a problem with the remote administration
interface could allow for the release of DHCP allocated addresses.  When
remote administration is enabled, insufficient access control is allegedly
placed on the /release.html page.  This page is used to manipulate DHCP
allocated addresses, and could be used to revoke leases on assigned
addresses.

This problem makes it possible for a remote user to access the DHCP
address release page, and release arbitrary DHCP assigned addresses.  It
should be noted that this vulnerability is only capable of being exploited
when the web administration interface is enabled.

D-Link Remote Administration Information Leakage Vulnerability
BugTraq ID: 5553
Remote: Yes
Date Published: Aug 22 2002 12:00AM
Relevant URL:
http://www.securityfocus.com/bid/5553
Summary:

The DI-804 is a hardware gateway and firewall solution distributed and
maintained by D-Link.

A problem with the DI-804 could make it possible for a remote user to gain
sensitive information about the device.

It has been reported that a problem with the remote administration
interface could allow users to gain sensitive information.  It is possible
to access to the Device information and Device status pages.  These pages
contain information such as the WAN IP, netmask, name server information,
DHCP log, and MAC address to IP address mappings.

This could allow an attacker to gain sensitive information, and result in
an organized attack on network resources.


-
Pour poster une annonce: [EMAIL PROTECTED]

Répondre à