Local Root Exploit in Udev
The udev subsystem allows the Linux kernel, together with a userland program, to manage device nodes dynamically, adding and removing them at will. It has now been revealed that the communication channel between the kernel and program fails to authenticate, so that users can assume root privileges.
The udev subsystem and the udevd daemon communicate in userspace over the netlink interface. Sending KOBJECT_UEVENT messages unfortunately doesn't verify who sent them. The result is that normal users can assume read privileges for a random device. If the device has a major and minor number of the root block device, invasive code can be applied to alter the system. A root exploit could then be quite simple for an attacker.
The root exploit was discovered by Sebastian Krahmer of the SUSE Security Team, which had become apprised of the CVE-2009-1185 spoofing exposure. The udev “trickery" exploit also points to another vulnerability in the CVE-2009-1186 stack buffer overflow exposure. Unfortunately, MITRE has not released further details on these vulnerabilities. All larger Linux distros have openly declared to be affected by them, so the distros are now providing updated packages.