[Ksplice][Ubuntu-15.04-Updates] New updates available via Ksplice (USN-2842-1)

[Oracle Ksplice]

Synopsis: USN-2842-1 can now be patched using Ksplice
CVEs: CVE-2015-7799 CVE-2015-7884 CVE-2015-7885 CVE-2015-8104

Systems running Ubuntu 15.04 Vivid can now use Ksplice to patch
against the latest Ubuntu Security Notice, USN-2842-1.

INSTALLING THE UPDATES

We recommend that all users of Ksplice Uptrack on Ubuntu 15.04 Vivid
install these updates.

On systems that have "autoinstall = yes" in /etc/uptrack/uptrack.conf,
these updates will be installed automatically and you do not need to
take any action.

Alternatively, you can install these updates by running:

# /usr/sbin/uptrack-upgrade -y


DESCRIPTION

* Denial-of-service in ISDN PPP device opening.

Missing allocation failure checks could result in a NULL pointer
dereference when opening an ISDN PPP device.  A local user with access
to the device could use this flaw to crash the system.


* CVE-2015-7799: Denial-of-service in PPP compression slot parameters.

Missing validation of VJ compression slot parameters for a PPP device
could result in a NULL pointer dereference and kernel crash.  A local
user with access to the PPP device could use this flaw to crash the
system.


* CVE-2015-7884: Information leak in Virtual Video Test Driver ioctl.

Missing initialization of an ioctl() structure field could result in a
leak of 16 bytes of kernel stack contents to user-space.


* CVE-2015-7885: Information leak in Digi Neo and Classic PCI driver.

Missing initialization of on-stack structures could result in an
information leak of 16 bytes to user-space.


* Out of bounds memory access in the UBI driver.

A lack of input validation when parsing a UBI image could cause out of
bounds memory accesses and lead to a kernel crash.  A local user able to
mount a special handcrafted image could use this flaw to cause a
denial-of-service.


* Out of bounds memory access in get_wchan().

A logic error when checking bounds of the current stack pointer in
get_wchan() could lead to out of bounds memory accesses.  A local,
un-privileged user could use this flaw to cause a kernel panic.


* Kernel BUG when unmapping a hugetlbfs page.

A logic error in the hugetlbfs when unmapping a page that is mapped both
with MAP_SHARED and MAP_PRIVATE could trigger a BUG() assertion.  A local,
un-privileged user could use this flaw to cause a denial-of-service.


* Mark unused space between __ex_table_end and rodata as non-executable.

Unused space between the end of the __ex_table and the start of rodata had
incorrect RWX protection, which could potentially be used to inject code in
a running kernel and facilitate exploits.


* Kernel panic when reshaping a RAID5 to RAID0.

A flaw in the RAID sub-system could lead to device errors and trigger a
kernel BUG() assertion when reshaping a RAID5 to a RAID0 in certain
circumstances.  A local, privileged user could use this flaw to cause a
denial-of-service.


* Permission bypass in the tty driver.

A flaw in the tty code would allow someone with a file descriptor opened
write only to re-open the tty with different flags, allowing him to control
the terminal when this should require both read and write access to the
tty.


* NULL pointer dereference in the Multiple devices driver when switching from cleaner policy.

A flaw in the Multiple devices driver could lead to a NULL pointer
dereference in certain circumstances.  A local attack could use this flaw
to cause a denial-of-service.


* Kernel crash when using ahash driver without import/export callback.

Ahash drivers are required to provide import/export callbacks to be
registered with the ahash crypto sub-system, otherwise they could lead to a
kernel crash under certain circumstances.  A local, un-privileged user
could use this flaw to cause a denial-of-service.


* Use-after-free in Btrfs filesystem when iterating extended refs.

A flaw in the Btrfs filesystem code when iterating over extended refs leads
to a use-after-free and kernel panic.  A local, un-privileged user could
use this flaw to cause a denial-of-service.


* Remote denial-of-service when receiving socket buffers with partial checksums.

A flaw in the socket buffer code dealing with partial checksums causes out
of bounds memory accesses on the socket buffer and kernel panic.  A remote
attacker could use this flaw to cause a denial-of-service.


* NULL pointer dereference in the Point to point over ethernet protocol.

A flaw in the Point to point over ethernet driver could lead to a NULL
pointer dereference and kernel panic when flushing the device.  A local,
un-privileged user could use this flaw to cause a denial-of-service.


* Denial-of-service when allocating from offline NUMA node in Openvswitch driver.

A flaw in the Openvswitch driver when allocating from the cache for an
offline NUMA node leads to a VM_BUG_ON() assertion to trigger and kernel
panic.  An attacker could use this flaw to cause a denial-of-service.


* Information leak when getting strings from the ethtool device.

A lack of cleaning an allocated buffer that is copied to user space on
ETHTOOL_GSTRINGS requests could leak information about the running kernel.
This could help an attacker to elevate privileges.


* Use-after-free in Rados block device driver on failure to probe.

A flaw in the Rados block device driver leads to double-free if there's an
error when probing the parent device.  A local, privileged user could use
this flaw to cause a denial-of-service.


* Kernel panic when dumping eBPF filter via SO_GET_FILTER.

A logic error in the BPF subsystem can trigger a NULL pointer
dereference and kernel panic when dumping a eBPF filter via the
SO_GET_FILTER sockopt.


* Use-after-free in Infiniband Connected Mode Service ID Resolution.

Incorrect handling of Service ID Resolution requests could result in a
use-after-free condition and kernel crash.


* Denial-of-service in PCI numa_node sysfs attribute.

Missing range checks could result in an out-of-bounds access when
writing to the num_node override attribute of a PCI device triggering a
kernel crash, or possibly allowing privilege escalation.


* Memory leak in overlayfs copying to upper filesystem.

Incorrect error handling could result in a memory leak when the
overlayfs filesystem failed to copy files from the lower to upper
filesystem.


* NULL pointer dereference in Marvell 88SE64XX/88SE94XX task preparation.

A missing NULL pointer check could result in a NULL pointer dereference
and kernel crash when performing tasks on a Marvell 88SE64XX/88SE94XX
device under low memory conditions.


* Denial-of-service in software RAID5 stripe cleaning.

Incorrect locking during stripe cleaning could result in an infinite
loop and system crash.  A local, unprivileged user with write access to
a filesystem on a RAID5 device could use this flaw to crash the system.


* Memory corruption in Mellanox MLX4 slave events.

Incorrect size arguments to memcpy() calls could result in memory
corruption of MLX4 devices, causing a kernel crash.


* NULL pointer dereference in PPP over Ethernet device releasing.

An incorrect check for disconnected PPP over Ethernet devices could
result in a NULL pointer dereference and kernel crash when closing the
device.


* CVE-2015-8104: KVM host denial-of-service in debug exception.

A guest could cause a denial-of-service on a KVM host by triggering a
debug exception to fire during an existing debug exception.  This could
cause the host to get trapped in an infinite loop causing a
denial-of-service.  A privileged user in a guest could use this flaw to
crash the host.


* Privilege escalation in Silicon Labs Si2168 firmware loading.

Missing bounds checks when reading firmware could result in a stack
buffer overflow and code execution.  A privileged user with maliciously
crafted firmware could use this flaw to gain code execution in the
kernel.


* Address space layout randomization bypass in position independent executables.

A weakness in the address space layout randomization (ASLR)
implementation for position independent executables (PIE) could allow an
attacker that could leak a single address from the executable to infer
the addresses of all libraries in the process.  This flaw could then be
used with another exploit to gain reliable code execution.


* Kernel hang in Multiple devices driver when destroying a device.

Incorrect lock ordering when destroying a RAID device could lead to a
deadlock and kernel hang.

SUPPORT

Ksplice support is available at ksplice-support_ww at oracle.com.