Several local and remote vulnerabilities have been discovered in the
Linux kernel that may lead to a denial of service or the execution of
arbitrary code. The Common Vulnerabilieis and Exposures project
identifies the following problems:
A race condition in the sysfs filesystem allows local users to
read kernel memory and cause a denial of service (crash).
Alexander Nyberg discovered that the ptrace() system call does not
properly verify addresses on the amd64 architecture which can be
exploited by a local attacker to crash the kernel.
A problem in the offset handling in the xattr file system code for
ext3 has been discovered that may allow users on 64-bit systems
that have access to an ext3 filesystem with extended attributes to
cause the kernel to crash.
Chris Wright discovered that the mmap() function could create
illegal memory maps that could be exploited by a local user to
crash the kernel or potentially execute arbitrary code.
A vulnerability on the IA-64 architecture can lead local attackers
to overwrite kernel memory and crash the kernel.
A vulnerability has been discovered in the ptrace() system call on
the amd64 architecture that allowas a local attacker to cause the
kernel to crash.
A buffer overflow in the ptrace system call for 64-bit
architectures allows local users to write bytes into arbitrary
Zou Nan Hai has discovered that a local user could cause the
kernel to hang on the amd64 architecture after invoking syscall()
with specially crafted arguments.
A vulnerability has been discovered in the stack segment fault
handler that could allow a local attacker to cause stack exception
that will leed the kernel to crash under certain circumstances.
Balazs Scheidler discovered that a local attacker could call
setsockopt() with an invalid xfrm_user policy message which would
cause the kernel to write beyond the boundaries of an array and
Vladimir Volovich discovered a bug in the zlib routines which are
also present in the Linux kernel and allows remote attackers to
krash the kernel
Another vulnerability has been discovered in the zlib routines
which are also present in the Linux kernel and allows remote
attackers to crash the kernel
Peter Sandstrom noticed that snmpwalk from a remote host could
cause a denial of service (kernel oops from null dereference) via
certain UDP packets that lead to a function call with the wrong
Andreas Gruenbacher discovered a bug in the ext2 and ext3 file
systems. When data areas are to be shared among two inodes not
all information were compared for equality, which could expose
wrong ACLs for files.
Chad Walstrom discovered that the ipt_recent kernel module on
64-bit processors such as AMD64 ows remote attackers to cause a
denial of service (kernel panic) via certain attacks such as SSH
The mprotect code on Itanium IA-64 Montecito processors does not
properly maintain cache coherency as required by the architecture,
which allows local users to cause a denial of service and possibly
corrupt data by modifying PTE protections.
A race condition in the thread management may allow local users to
cause a denial of service (deadlock) when threads are sharing
memory and waiting for a thread that has just performed an exec.
When one thread is tracing another thread that shares the same
memory map a local user could cause a denial of service (deadlock)
by forcing a core dump when the traced thread is in the
A bug in the ioremap() system call has been discovered on the
amd65 architecture that could allow local users to cause a
denial of service or an information leak when performing a lookup
of a non-existant memory page.
The HFS and HFS+ (hfsplus) modules allow local attackers to cause
a denial of service (oops) by using hfsplus to mount a filesystem
that is not hfsplus.
A race ondition in the ebtables netfilter module on an SMB system
running under high load may allow remote attackers to cause a
denial of service (crash).
Roland McGrath discovered exec() does not properly clear
posix-timers in multi-threaded environments, which results in a
resource leak and could allow a large number of multiple local
users to cause a denial of service by using more posix-timers than
specified by the quota for a single user.
The kernel allows remote attackers to poison the bridge forwarding
table using frames that have already been dropped by filtering,
which can cause the bridge to forward spoofed packets.
The ioctl for the packet radio ROSE protocol does not properly
verify the arguments when setting a new router, which allows
attackers to trigger out-of-bounds errors.
A race condition in on SMP systems allows local users to cause a
denial of service (null dereference) by causing a connection timer
to expire while the connection table is being flushed before the
appropriate lock is acquired.
An error in the NAT code allows remote attackers to cause a denial
of service (memory corruption) by causing two packets for the same
protocol to be NATed at the same time, which leads to memory
A missing memory cleanup in the thread handling routines before
copying data into userspace allows a user process to obtain
This update also contains a number of corrections for issues that
turned out to have no security implication afterwards.
The following matrix explains which kernel version for which architecture
fix the problems mentioned above: