OpenBSD 2.8 errata

This is the OpenBSD 2.8 release errata & patch list:

For OpenBSD patch branch information, please refer here.

For errata on a certain release, click below:
2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.9, 3.0, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7,
3.8, 3.9, 4.0, 4.1, 4.2, 4.3, 4.4, 4.5, 4.6, 4.7, 4.8, 4.9, 5.0, 5.1, 5.2, 5.3,
5.4, 5.5.

You can also fetch a tar.gz file containing all the following patches. This file is updated once a day.

The patches below are available in CVS via the OPENBSD_2_8 patch branch.

For more detailed information on how to install patches to OpenBSD, please consult the OpenBSD FAQ.

033: SECURITY FIX: September 11, 2001 All architectures
A security hole exists in uuxqt(8) that may allow an attacker to run arbitrary commands as user uucp and use this to gain root access. The UUCP execution daemon, uuxqt(8), has a bug in its command line parsing routine may allow arbitrary commands to be run. Because some UUCP commands are run as root (and daemon) from cron it is possible to leverage compromise of the UUCP user to gain root.
A source code patch exists which remedies this problem.
032: SECURITY FIX: August 29, 2001 All architectures
A security hole exists in lpd(8) that may allow an attacker with line printer access to gain root privileges. A machine must be running lpd to be vulnerable (OpenBSD does not start lpd by default). Only machines with line printer access (ie: listed in either /etc/hosts.lpd or /etc/hosts.equiv) may be used to mount an attack.
A source code patch exists which remedies this problem.
031: SECURITY FIX: August 21, 2001 All architectures
A security hole exists in sendmail(8) that may allow an attacker on the local host to gain root privileges by specifying out-of-bounds debug parameters.
A source code patch exists which remedies this problem.
030: SECURITY FIX: June 15, 2001 All architectures
A race condition exists in the kernel execve(2) implementation that opens a small window of vulnerability for a non-privileged user to ptrace(2) attach to a suid/sgid process.
A source code patch exists which remedies this problem.
029: SECURITY FIX: May 30, 2001 All architectures
Programs using the fts(3) routines (such as rm, find, and most programs that take a -R flag) can be tricked into changing into the wrong directory if the parent dir is changed out from underneath it. This is similar to the old fts bug but happens when popping out of directories, as opposed to descending into them. A source code patch exists which remedies this problem. This is the second version of the patch.
028: SECURITY FIX: May 29, 2001 All architectures
The signal handlers in sendmail(8) contain code that is unsafe in the context of a signal handler. This leads to potentially serious race conditions. At the moment this is a theoretical attack only and can only be exploited on the local host (if at all).
A source code patch exists which remedies this problem. It updates sendmail to version 8.11.4.
027: SECURITY FIX: Apr 23, 2001 All architectures
IPF has a serious problem with fragment caching, the bug is triggered if you use the ipf(5) syntax "keep state".
A source code patch exists which remedies this problem.
026: SECURITY FIX: Apr 23, 2001 All architectures
ftpd(8) has a potential DoS related to glob(3). This patch introduces a GLOB_LIMIT, eliminating the DoS. You must have 025_glob.patch installed before installing this patch.
A source code patch exists which remedies this problem.
025: SECURITY FIX: Apr 10, 2001 All architectures
glob(3) contains multiple buffer overflows.
A source code patch exists which remedies this problem.
024: SECURITY FIX: Mar 18, 2001 All architectures
The readline library shipped with OpenBSD allows history files creation with a permissive umask(2). This can lead to the leakage of sensitive information in applications that use passwords and the like during user interaction (one such application is mysql).
A source code patch exists which remedies this problem.
023: SECURITY FIX: Mar 2, 2001 All architectures
Insufficient checks in the IPSEC AH IPv4 option handling code can lead to a buffer overrun leading to a remote DoS. This option is not on by default.
A source code patch exists which remedies this problem.
022: SECURITY FIX: Mar 2, 2001
The USER_LDT kernel option allows an attacker to gain access to privileged areas of kernel memory. This option is not on by default. A source code patch exists which remedies this problem.
021: SECURITY FIX: Feb 22, 2001 All architectures
There is an exploitable heap corruption bug in sudo.
A source code patch exists which remedies this problem.
020: IMPLEMENTATION FIX: Feb 15, 2001 All architectures
Client side ident protocol was broken in libwrap, affecting anything using libwrap including tcpd. The effect of this was that libwrap would never retrieve and log ident values from remote hosts on connections.
A source code patch exists which remedies this problem.
019: IMPLEMENTATION FIX: Jan 31, 2001 All architectures
Fix memory allocation in the PCI LANCE driver, le. A side effect of this is that OpenBSD under VMWare now works again.
A source code patch exists which remedies this problem.
018: SECURITY FIX: Jan 29, 2001 All architectures
Merge named with ISC BIND 4.9.8-REL, which fixes some buffer vulnerabilities (actually it appears that these were already impossible to exploit beforehand).
A source code patch exists which remedies this problem.
017: SECURITY FIX: Jan 22, 2001 All architectures
The rnd(4) device does not use all of its input when data is written to it.
A source code patch exists which remedies this problem.
016: RELIABILITY FIX: Jan 4, 2001 All architectures
Allow ThunderLAN cards to share interrupts nicely.
A source code patch exists which remedies this problem.
015: STABILITY FIX: Dec 22, 2000
Some machines locked up while trying to use the mouse in console mode. This patch solves that problem.
A source code patch exists which remedies this problem.
014: SECURITY FIX: Dec 22, 2000 All architectures
Improve xlock(1)'s authentication by authenticating via a pipe in an early forked process. No known vulnerability exists, this is just a precautionary patch.
A source code patch exists which remedies this problem.
In addition to a source code patch, new xlock binaries have been created for each architecture listed below. Place these binaries at /usr/X11R6/bin/xlock and chmod 4755 /usr/X11R6/bin/xlock.
- Xlock - i386
- Xlock - sparc
- Xlock - mac68k
- Xlock - powerpc
- Xlock - amiga
- Xlock - hp300
- Xlock - mvme68k
013: SECURITY FIX: Dec 18, 2000 All architectures
Procfs contained numerous overflows, which could lead an intruder to root permissions. Procfs is NOT enabled by default in OpenBSD.
A source code patch exists which remedies this problem.
012: INSTALL PROBLEM: Dec 14, 2000
The IMac DV+ (and probably some other machines) incorrectly identify their video hardware, but it is possible to work around the problem.
A source code patch exists which remedies this problem.
011: RELIABILITY FIX: Dec 13, 2000 All architectures
The crypto subsystem could incorrectly fail to run certain software ciphers, if a hardware card existed in the machine.
A source code patch exists which remedies this problem.
010: RELIABILITY FIX: Dec 11, 2000 All architectures
A crash could occur during fast routing, if IPSEC was enabled.
A source code patch exists which remedies this problem.
009: SECURITY FIX: Dec 10, 2000 All architectures
Another problem exists in the Kerberos libraries.
A source code patch exists which remedies this problem.
008: SECURITY FIX: Dec 7, 2000 All architectures
Two problems have recently been discovered in the KerberosIV code.
1. A symlink problem was discovered in the KerberosIV password checking routines /usr/bin/su and /usr/bin/login, which makes it possible for a local user to overwrite any file on the local machine.
2. It is possible to specify environment variables in telnet which will be passed over the to the remote host. This makes it possible to set environment variables on the remote side, including ones that have special meaning on the server. It is not clear at this time what the impact is, but we recommend everyone to upgrade their machines immediately.
A source code patch exists which remedies this problem.
007: INSTALL PROBLEM: Dec 4, 2000
The X packages share28.tgz and font28.tgz were not on the CD, and only available via FTP install. These packages can be added post-install by using the following command:
```
    # cd /; tar xvfpz xshare28.tgz
    # cd /; tar xvfpz xfont28.tgz
```
006: STABILITY FIX: Dec 4, 2000
On some machines, a PCIBIOS device driver interrupt allocation bug can cause a kernel hang while probing PCI devices. If you have this symptom, you can disable PCIBIOS as a workaround. To do this,
- Enter the User Kernel Configuration by booting with the option "boot -c".
- Once at the UKC> prompt, enter
```
	UKC> disable pcibios
	UKC> quit
    
```
- See FAQ 5.6 after a successful boot for instructions on how to re-write your kernel to disable PCIBIOS permanently.
005: SECURITY FIX: Dec 4, 2000 All architectures
OpenBSD 2.8's ftpd contains a one-byte overflow in the replydirname() function.
A source code patch exists which remedies this problem. You can view the OpenBSD Advisory here.
004: RELIABILITY FIX: Nov 17, 2000 All architectures
First off, AES (Rijndael) encryption and decryption were broken for IPsec and swap encryption.
Secondly, the AES code did not work properly on big endian machines.
A source code patch exists which remedies this problem. This is the second revision of the patch.
003: RELIABILITY FIX: Nov 17, 2000
Configuring a qec+qe causes a NMI panic.
A source code patch exists which remedies this problem.
002: IMPLEMENTATION FIX: Nov 10, 2000 All architectures
In ssh(1), skey support for SSH1 protocol was broken. Some people might consider that kind of important.
A source code patch exists which remedies this problem.
001: RELIABILITY FIX: Nov 10, 2000
When running a sparc with a serial console, certain types of interrupts would cause great grief.
A source code patch exists which remedies this problem.