Various upstream fixes and a security fix high severity issue that allows remote DOS:
This laptop needed to be updated before the install, nvidia driver and
Removed several old kernels as well.
Installed the tmb kernel and rebooted to Mate.
System: Host: markab Kernel: 4.14.62-tmb-desktop-1.mga6 x86_64
CPU: Quad core Intel Core i7-5700HQ (-HT-MCP-) speed/max: 2761/3500 MHz
Machine: Device: laptop System: GIGABYTE product: X5
Graphics: Card-1: NVIDIA GM204M [GeForce GTX 965M]
Card-2: NVIDIA GM204M [GeForce GTX 965M]
GLX Version: 4.6.0 NVIDIA 390.77
Network: Card-1: Qualcomm Atheros Killer E220x Gigabit Ethernet Controller
Card-2: Intel Wireless 7265 driver: iwlwifi
RAM: 16 GB
Desktop functional, resumes after closing and opening the lid. NFS share
mounted under the MCC.
USB automounted when plugged in and applications on it could be run without issues.
Running on one GPU and wifi. Stress tests OK.
Tried various desktops:
Plasma seemed to be running normally.
Cinnamon was difficult to use because most applications launched at fullscreen with no
way to reduce them except by corner dragging. firefox occupied twice the screen depth.
This needs to be checked in other scenarios. Later.
GNOME Classic came up OK but desktop icons were oversized, all but the weather applet.
No way to log out until I clicked the charger icon on the toolbar.
Tried GNOME on Xorg. That was functional - no desktop icons and the toolbar showed a
GNOME (Wayland?) worked as expected.
LXDE login produced a serviceable desktop.
LXQt came up fine with one oddity, two Home icons on the desktop. One used
Dolphin as file manager and the other something else.
So, other desktops work fine within their limitations. The quibbles mentioned
doubtless have nothing to do with the kernel.
Flushing this out as more is coming...
Advisory, added to svn
subject: Updated kernel-tmb packages fix security vulnerabilities
This kernel-tmb update is based on the upstream 4.14.62 and fixes atleast
the following security issues:
kernel KVM before versions kernel 4.16, kernel 4.16-rc7, kernel 4.17-rc1,
kernel 4.17-rc2 and kernel 4.17-rc3 is vulnerable to a flaw in the way the
Linux kernel's KVM hypervisor handled exceptions delivered after a stack
switch operation via Mov SS or Pop SS instructions. During the stack switch
operation, the processor did not deliver interrupts and exceptions, rather
they are delivered once the first instruction after the stack switch is
executed. An unprivileged KVM guest user could use this flaw to crash the
guest or, potentially, escalate their privileges in the guest
Linux kernel vhost since version 4.8 does not properly initialize memory in
messages passed between virtual guests and the host operating system in the
vhost/vhost.c:vhost_new_msg() function. This can allow local privileged
users to read some kernel memory contents when reading from the
/dev/vhost-net device file (CVE-2018-1118).
Security researchers from FICORA have identified a remote denial of
service attack against the Linux kernel caused by inefficient
implementation of TCP segment reassembly, named "SegmentSmack".
A remote attacker could consume a lot of CPU resources in the kernel
networking stack with just a low bandwidth or single host attack by
using lots of small TCP segments packets. Usually large botnets are
needed for similar effect. The rate needed for this denial of service
attack to be effective is several magnitudes lower than the usual
packet processing capability of the machine, as the attack exploits
worst case behaviour of existing algorithms (CVE-2018-5390).
In the function sbusfb_ioctl_helper() in drivers/video/fbdev/sbuslib.c
in the Linux kernel through 4.15, an integer signedness error allows
arbitrary information leakage for the FBIOPUTCMAP_SPARC and
FBIOGETCMAP_SPARC commands (CVE-2018-6412).
In some circumstances, some operating systems or hypervisors may not expect
or properly handle an Intel architecture hardware debug exception. The error
appears to be due to developer interpretation of existing documentation for
certain Intel architecture interrupt/exception instructions, namely MOV SS
and POP SS. An authenticated attacker may be able to read sensitive data in
memory or control low-level operating system functions (CVE-2018-8897).
Linux kernel is vulnerable to a heap-based buffer overflow in the
fs/ext4/xattr.c:ext4_xattr_set_entry() function. An attacker could exploit
this by operating on a mounted crafted ext4 image (CVE-2018-10840).
The kvm functions that were used in the emulation of fxrstor, fxsave,
sgdt and sidt were originally meant for task switching, and as such they
did not check privilege levels. This allowed guest userspace to guest
kernel write (CVE-2018-10853).
A flaw was found in Linux kernel ext4 File System. A use-after-free in
ext4_ext_remove_space() when mounting and operating a crafted ext4 image
Linux kernel ext4 filesystem is vulnerable to an out-of-bound access in the
ext4_ext_drop_refs() function when operating on a crafted ext4 filesystem
A flaw was found in Linux kernel ext4 filesystem. A local user can cause a
use-after-free in ext4_xattr_set_entry function and so a denial of service
or possibly unspecified other impact by when renaming a file in a crafted
ext4 filesystem image (CVE-2018-10879).
A flaw was found in Linux kernel ext4 filesystem code. A stack-out-of-bounds
write in ext4_update_inline_data() is possible when mounting and writing to
a crafted ext4 image. An attacker could use this to cause a system crash
and a denial of service (CVE-2018-10880).
A flaw was found in Linux kernel ext4 filesystem. A local user can cause an
out-of-bound access in ext4_get_group_info function and so a denial of
service and a system crash by mounting and operating on a crafted ext4
filesystem image (CVE-2018-10881).
A flaw was found in Linux kernel ext4 File System. An out-of-bound write
when unmounting a crafted ext4 image in fs/jbd2/transaction.c. An attacker
could use this to cause a denial of service (system crash) (CVE-2018-10882).
A flaw was found in Linux kernel ext4 File System. An out-of-bound write in
jbd2_journal_dirty_metadata() that kernel cannot handle when mounting and
operating a crafted ext4 image. An attacker could use this to cause a
denial of service (system crash) (CVE-2018-10883).
In the Linux kernel 4.13 through 4.16.11, ext4_read_inline_data() in
fs/ext4/inline.c performs a memcpy with an untrusted length value in certain
circumstances involving a crafted filesystem that stores the system.data
extended attribute value in a dedicated inode (CVE-2018-11412).
In arch/x86/kvm/vmx.c in the Linux kernel before 4.17.2, when nested
virtualization is used, local attackers could cause L1 KVM guests to
VMEXIT, potentially allowing privilege escalations and denial of service
attacks due to lack of checking of CPL (CVE-2018-12904).
The inode_init_owner function in fs/inode.c in the Linux kernel through
4.17.4 allows local users to create files with an unintended group
ownership, in a scenario where a directory is SGID to a certain group and
is writable by a user who is not a member of that group. Here, the
non-member can trigger creation of a plain file whose group ownership is
that group. The intended behavior was that the non-member can trigger
creation of a directory (but not a plain file) whose group ownership is
that group. The non-member can escalate privileges by making the plain
file executable and SGID (CVE-2018-13405).
An issue was discovered in the Linux kernel through 4.17.11, as used
in Xen through 4.11.x. The xen_failsafe_callback entry point in
arch/x86/entry/entry_64.S does not properly maintain RBX, which allows
local users to cause a denial of service (uninitialized memory usage
and system crash). Within Xen, 64-bit x86 PV Linux guest OS users can
trigger a guest OS crash or possibly gain privileges (CVE-2018-14678).
Other changes in this update:
* WireGuard has been updated to 0.0.20180802
* enable Mellanox5 support (mga#23263)
* enable SMARTPQI support (mga#23305)
* ext4: check for allocation block validity with block group locked,
fixes possible data corruption under heavy load
* Add PCI ID for Cannon Lake PCH-LP and Ice Lake LP AHCI
For other upstream fixes in this update, see the referenced changelogs.
An update for this issue has been pushed to the Mageia Updates repository.