Showing all newswire headlines

The zlib compression library is being used by many applications to provide data compression/decompression routines. An error in a decompression routine can corrupt the internal data structures of malloc by a double call to the free() function. If the data processed by the compression library is provided from an untrusted source, it may be possible for an attacker to interfere with the process using the zlib routines. The attack scenario includes a denial of service attack and memory/data disclosure, but it may also be possible to insert arbitrary code into the running program and to execute this code. This update fixes the known problems in the libz/zlib as a permanent fix. There exists no temporary workaround that can efficiently remedy the problem.

The zlib library provides in-memory compression/decompression functions. The library is widely used throughout Linux and other operating systems. While performing tests on the gdk-pixbuf library, Matthias Clasen created an invalid PNG image that caused libpng to crash. Upon further investigation, this turned out to be a bug in zlib 1.1.3 where certain types of input will cause zlib to free the same area of memory twice (called a "double free"). This bug can be used to crash any program that takes untrusted compressed input. Web browsers or email programs that display image attachments or other programs that uncompress data are particularly affected. This vulnerability makes it easy to perform various denial-of-service attacks against such programs. It is also possible that an attacker could manage a more significant exploit, since the result of a double free is the corruption of the malloc() implementation's data structures. This could include running arbitrary code on local or remote systems. Most packages in Red Hat Linux use the shared zlib library and can be protected against vulnerability by updating to the errata zlib package. However, we have identified a number of packages in Red Hat Linux that either statically link to zlib or contain an internal version of zlib code. Although no exploits for this issue or these packages are currently known to exist, this is a serious vulnerability which could be locally or remotely exploited. All users should upgrade affected packages immediately. Additionally, if you have any programs that you have compiled yourself, you should check to see if they use zlib. If they link to the shared zlib library then they will not be vulnerable once the shared zlib library is updated to the errata package. However, if any programs that decompress arbitrary data statically link to zlib or use their own version of the zlib code internally, then they need to be patched or recompiled.

The zlib compression library provides in-memory compression and decompression functions. It is widely used throughout Linux and other operating systems. While performing tests on the gdk-pixbuf library, Matthias Clasen created an invalid PNG image that caused libpng to crash. Upon further investigation, this turned out to be a bug in zlib 1.1.3. Certain input will cause zlib to free an area of memory twice (also called a "double free"). This bug can be used to crash any program that takes untrusted compressed input. Web browsers or email programs that display image attachments or other programs that uncompress data are particularly affected. This vulnerability makes it easy to perform various denial-of-service attacks against such programs. However, since the result of a double free is the corruption of the malloc implementation's data structures, it is possible that an attacker could manage a more significant exploit, such as running arbitrary code on the affected system.

Several security related problems have been found in the xtell package, a simple messaging client and server. In detail, these problems contain several buffer overflows, a problem in connection with symbolic links, unauthorized directory traversal when the path contains "..". These problems could lead into an attacker being able to execute arbitrary code on the server machine. The server runs with nobody privileges by default, so this would be the account to be exploited.

Ed Moyle recently found a buffer overflow in Apache-SSL and mod_ssl. With session caching enabled, mod_ssl will serialize SSL session variables to store them for later use. These variables were stored in a buffer of a fixed size without proper boundary checks.

Joost Pol <joost@pine.nl> reports that OpenSSH versions 2.0 through 3.0.2 have an off-by-one bug in the channel allocation code. This vulnerability can be exploited by authenticated users to gain root privilege or by a malicious server exploiting a client with this bug.

Updated openssh packages are now available for Red Hat Linux 7, 7.1, and 7.2 which close a remotely-exploitable vulnerability in sshd.

A problem was found in versions of improved mod_frontpage prior to 1.6.1 regarding a lack of boundary checks in fpexec.c. This means that the suid root binary is exploitable for buffer overflows. This could be exploited by remote attackers to execute arbitrary code on the server with superuser privileges. Although there are no known exploits available, if you use mod_frontpage you are strongly encouraged to upgrade. This update for Mandrake Linux has been completely reworked and is easier to configure and use, as well as supporting the new FrontPage 2002 extensions.

Ed Moyle discovered a buffer overflow in mod_ssl's session caching mechanisms that use shared memory and dbm. This could potentially be triggered by sending a very long client certificate to the server.

Joost Pol found a bug in the channel code of all versions of OpenSSH from 2.0 to 3.0.2. This bug can allow authenticated users with an existing account on the vulnerable system to obtain root privilege or by a malicious server attacking a vulnerable client. OpenSSH 3.1 is not vulnerable to this problem. The provided packages fix this vulnerability.

New openssh packages are available to fix security problems.

Updated mod_ssl packages for Red Hat Linux 7, 7.1, and 7.2 are available which close a buffer overflow in mod_ssl.

Joost Pol discovered an off-by-one bug in a routine in the openssh code for checking channel IDs. This bug can be exploited on the remote side by an already authenticated user, qualifying this bug as a local security vulnerability, and on the local side if a malicious server attacks the connected client, qualifying this bug as a remote vulnerability. If the error is being exploited, it leads to arbitrary code execution in the process under attack (either a local ssh client, attacking the userID of the client user, or a remote secure shell daemon that has an authenticated user session running, attacking the root account of the remote system). Please note that the possible attack scenario is different from the usual attack scheme because "local vulnerability" refers to the remote side and vice versa.

Tim Waugh found several insecure uses of temporary files in the xsane program, which is used for scanning. This was fixed for Debian/stable by moving those files into a securely created directory within the /tmp directory.

Kim Nielsen recently found an internal problem with the CVS server and reported it to the vuln-dev mailing list. The problem is triggered by an improperly initialized global variable. A user exploiting this can crash the CVS server, which may be accessed through the pserver service and running under a remote user id. It is not yet clear if the remote account can be exposed, through.

Updated radiusd-cistron packages, which fix various security issues, are now available.

The widely used proxy-server squid contains a heap overflow in one of its URL constructing functions. Incorrect length-calculations for the user and passwd fields in ftp-URLs turned out to be the origin of the problem. Only users from hosts listed in squids ACL-files could trigger the overflow. The ftp-URL problem is not present in the 6.4, 7.0 and 7.1 distributions, but other security releated bugs have been fixed there. A complete history can be found at

This fixes several security problems in the POST handling code used for uploading files through forms. All sites using PHP are urged to upgrade as soon as possible.

Zorgon found several buffer overflows in cfsd, a daemon that pushes encryption services into the Unix(tm) file system. We are not yet sure if these overflows can successfully be exploited to gain root access to the machine running the CFS daemon. However, since cfsd can easily be forced to die, a malicious user can easily perform a denial of service attack to it.

Stefan Esser, who is also a member of the PHP team, found several flaws in the way PHP handles multipart/form-data POST requests (as described in RFC1867) known as POST fileuploads. Each of the flaws could allow an attacker to execute arbitrary code on the victim's system.