Synopsis: Moderate: kernel security and bug fix update
Advisory ID: SLSA-2017:0307-1
Issue Date: 2017-02-23
CVE Numbers: CVE-2016-6136
* When creating audit records for parameters to executed children
processes, an attacker can convince the Linux kernel audit subsystem can
create corrupt records which may allow an attacker to misrepresent or
evade logging of executing commands. (CVE-2016-6136, Moderate)
* A flaw was found in the Linux kernel’s implementation of the SCTP
protocol. A remote attacker could trigger an out-of-bounds read with an
offset of up to 64kB potentially causing the system to crash.
* The qlnic driver previously attempted to fetch pending transmission
descriptors before all writes were complete, which lead to firmware hangs.
With this update, the qlcnic driver has been fixed to complete all writes
before the hardware fetches any pending transmission descriptors. As a
result, the firmware no longer hangs with the qlcnic driver.
* Previously, when a NFS share was mounted, the file-system (FS) cache was
incorrectly enabled even when the “-o fsc” option was not used in the
mount command. Consequently, the cachefilesd service stored files in the
NFS share even when not instructed to by the user. With this update, NFS
does not use the FS cache if not instructed by the “-o fsc” option. As a
result, NFS no longer enables caching if the “-o fsc” option is not used.
* Previously, an NFS client and NFS server got into a NFS4 protocol loop
involving a WRITE action and a NFS4ERR_EXPIRED response when the
current_fileid counter got to the wraparound point by overflowing the
value of 32 bits. This update fixes the NFS server to handle the
current_fileid wraparound. As a result, the described NFS4 protocol loop
no longer occurs.
* Previously, certain configurations of the Hewlett Packard Smart Array
(HPSA) devices caused hardware to be set offline incorrectly when the HPSA
driver was expected to wait for existing I/O operations to complete.
Consequently, a kernel panic occurred. This update prevents the described
problem. As a result, the kernel panic no longer occurs.
* Previously, memory corruption by copying data into the wrong memory
locations sometimes occurred, because the __copy_tofrom_user() function
was returning incorrect values. This update fixes the __copy_tofrom_user()
function so that it no longer returns larger values than the number of
bytes it was asked to copy. As a result, memory corruption no longer
occurs in he described scenario.
* Previously, guest virtual machines (VMs) on a Hyper-V server cluster got
in some cases rebooted during the graceful node failover test, because the
host kept sending heartbeat packets independently of guests responding to
them. This update fixes the bug by properly responding to all the
heartbeat messages in the queue, even if they are pending. As a result,
guest VMs no longer get rebooted under the described circumstances.
* When the “punching hole” feature of the fallocate utility was used on an
ext4 file system inode with extent depth of 1, the extent tree of the
inode sometimes became corrupted. With this update, the underlying source
code has been fixed, and extent tree corruption no longer occurs in the
– Scientific Linux Development Team