RHSA-2018:2384 Important: kernel security and bug fix update

Updated 14 Aug 2018

The kernel packages contain the Linux kernel, the core of any Linux operating system.

Red Hat Product Security has rated this update as having Important security impact. Common Vulnerability Scoring System (CVSS) base scores, which give detailed severity ratings, are available for each vulnerability from the CVE links in the References section.

Security Fixes are described in RHSA-2018:2384.

This update fixes the following bugs :

A change in the host_busy counter of the SCSI kernel driver introduced a race condition when comparing the host_busy value against the value of the host_failed counter. As a consequence, certain system configurations sometimes missed a wakeup of the SCSI error handler. With this update, the Advanced Technology Attachment (ATA) code is changed to use block layer tagging. As a result, the SCSI error handler no longer misses a wakeup in the described scenario. (This content is not included.BZ#1581728)
Previously, when a user performed a CPU hotplug, the /proc/cpuinfo virtual file displayed an incorrect number of cores. With this update, the number of cores is initialized when a CPU hotplug occurs. As a result, /proc/cpuinfo displays the correct number of cores in the described scenario. (This content is not included.BZ#1588563)
An update to the qla2xxx driver inadvertently caused that the login_retry_count parameter was set to a fixed value instead of using a value configured in the Non-volatile random-access memory (NVRAM). As a consequence, under some conditions, the Host Bus Adapter (HBA) failed to login to the remote Fibre Channel (FC) port through a switch. This update sets login_retry_count to the value configured in NVRAM. As a result, the HBA is consistently now able to login to the switch port in the described scenario. (This content is not included.BZ#1588937)
Previously, inconsistencies in the Local Advanced Programmable Interrupt Controller (LAPIC) periodic timer implementation occasionally led to accumulated timing errors. As a consequence, the guest machines that used the period timer were likely to become unresponsive or experienced pauses, particularly after migration. This update synchronizes timers on each tick, preventing the errors from accumulating. As a result, the guest machines no longer hang or experience pauses in the described scenario. (This content is not included.BZ#1594292)
When creating VLAN over bond on QLogic cards, the references to network devices were not decreased properly. Consequently, the operating system became unresponsive on shutdown/reboot with the following message:
```
  "unregister_netdevice: waiting for p2p2 to become free. Usage count = 2"
```

With this update, the Remote Direct Memory Access (RDMA) devices are removed before unregistering the network device. As a result, all references to network devices are decreased and shutdown/reboot is successful in the described scenario. (This content is not included.BZ#1594700)

On the VMware hypervisor with Fault Tolerance and ballooning enabled, the page pointers were not properly initialized during a balloon reset. As a consequence, the following messages repeated in the kernel log:
```
  WARNING: at mm/vmalloc.c:1491 __vunmap+0xd3/0x100()
  Trying to vfree() nonexistent vm area (ffffc90000697000)
```

With this update, the page pointers are properly initialized. As a result, the kernel no longer logs the above messages in the described scenario. (This content is not included.BZ#1595601)

Previously, the code responsible for processing a list of unused dentry objects received an arbitrary number of objects. As a consequence, a list with multiple dentry objects occasionally caused a soft lockup. With this update, another task is allowed to run and the current task from the run queue can be re-selected to run later. As a result, the soft lockup no longer occurs in the described scenario. (This content is not included.BZ#1596184)

Article Type

General