Bug Fix Advisory qemu-kvm-rhev bug fix and enhancement update

Advisory: RHBA-2016:2673-1
Type: Bug Fix Advisory
Severity: N/A
Issued on: 2016-11-07
Last updated on: 2016-11-07
Affected Products: Red Hat Virtualization ( v.4 for RHEL 7)

Details

Updated qemu-kvm-rhev packages that fix several bugs and add various
enhancements are now available for Red Hat Virtualization Hypervisor 7.

Kernel-based Virtual Machine (KVM) is a full virtualization solution for Linux
on AMD64 and Intel 64 systems. The qemu-kvm-rhev package provides the user-space
component for running virtual machines using KVM in environments managed by Red
Hat Enterprise Virtualization Manager.

The qemu-kvm-rhev packages have been upgraded to upstream version 2.6.0, which
provides a number of bug fixes and enhancements over the previous version.
(BZ#1289417)

In addition, this update adds the following enhancements:

* Support for the XSAVES, XSAVEC, and XRSTORS instructions in Intel Xeon v5
guests has been added to the KVM hypervisor and kernel code, and to the libvirt
API. (BZ#1327589)

* This update adds support for removing virtual CPUs (vCPUs) from running KVM
guest virtual machines on AMD64 and Intel 64 hosts, also referred to as the vCPU
hot unplug. In combination with the vCPU hot plug, this enables users to
dynamically allocate resources to guests without the need to shut them down.
(BZ#1087672)

* It is now possible to hot plug vCPUs to guest virtual machines running on IBM
System z hosts. (BZ#1172917)

* This update enables the post-copy guest migration feature for KVM guests. When
used, post-copy migration immediately pauses guest vCPUs on the source host,
transfers only a minimum of memory pages, and activates the guest vCPUs on the
destination host. This significantly reduces the downtime of the migrated
virtual machine, but the migration cannot be reverted while it is in progress.
In addition, post-copy migration is guaranteed to finish regardless of how
rapidly the memory pages of the source guest change, so it is optimal for
migrating guests in heavy continuous use. (BZ#830966)

* With this update, Non-Volatile Dual In-line Memory Module (NVDIMM) persistent
memory is supported and can be assigned to guest virtual machines. (BZ#1270345)

* This update adds libvirt support for Intel e1000e network interface controller
for PCI-Express. This enables more effective use of virtual networks on recent
Intel machine models and operating systems. (BZ#1343092)

* With this update, it is possible to use TLS encryption with the Transmission
Control Protocol (TCP) back end for character devices. As a result, using the
TCP back end is more secure and efficient. (BZ#1300773)

Users of qemu-kvm-rhev are advised to upgrade to these updated packages, which
fix these bugs and add these enhancements.


Solution

Before applying this update, make sure all previously released errata
relevant to your system have been applied.

For details on how to apply this update, refer to:

https://access.redhat.com/articles/11258

Updated packages

Red Hat Virtualization ( v.4 for RHEL 7)

SRPMS:
qemu-kvm-rhev-2.6.0-27.el7.src.rpm
File outdated by:  RHSA-2017:0985
    MD5: 68f8819216de7dfc31abca02226674b6
SHA-256: 20aad33d141692928891c56232f3baffe1d66537bbe1826403b10db59a6975ac
 
x86_64:
qemu-img-rhev-2.6.0-27.el7.x86_64.rpm
File outdated by:  RHSA-2017:0985
    MD5: 2f21167afc58a1676d72ccd04d8a4dd1
SHA-256: 4d9f9fb6de67fb7c7cc4ed9d170a435322afe6db52ff18422b627c1ddefbafa8
qemu-kvm-common-rhev-2.6.0-27.el7.x86_64.rpm
File outdated by:  RHSA-2017:0985
    MD5: 583b2e68d65b2ea53fb3497c4c31c529
SHA-256: c58f599922c46931e8f3af3c2e7736c1af3f5f983ce2b428db7fc19892ea9506
qemu-kvm-rhev-2.6.0-27.el7.x86_64.rpm
File outdated by:  RHSA-2017:0985
    MD5: aca20204b6bd76fd61d76bd640dd4389
SHA-256: b4f5a77f325bf93a73980a00af141571c7ff6b4b37fd35dc3ccd1f776191c132
qemu-kvm-tools-rhev-2.6.0-27.el7.x86_64.rpm
File outdated by:  RHSA-2017:0985
    MD5: 07eb4578de7d2f3399ea9eef0abcea2f
SHA-256: d8677214a68f0d4cffc40342e0eb34c5b4adb991a97413749bbc4592b2dcc346
 
(The unlinked packages above are only available from the Red Hat Network)

Bugs fixed (see bugzilla for more information)

1055427 - [Q35] core dump when boot with 128 virtio-blk-pci data disks via PCIe-switch
1070027 - qemu crash during iofuzz test: qemu: hw/usb/core.c:707: usb_ep_get: Assertion `pid == 0x69 || pid == 0xe1' failed.
1082559 - udisksd[2008]: Error probing device: Error sending ATA command IDENTIFY PACKET DEVICE to /dev/sr0
1114962 - - drive-mirror with "mode":"existing" fails poorly if destination is not large enough
1124815 - fail to detect the new cdrom device in guest after "change" it to anther iso (scsi-cd only)
1126666 - qemu-kvm does not quit when booting guest w/ 161 vcpus and "-no-kvm"
1138511 - win2012r2 guest bsod(DRIVER_POWER_STATE_FAILURE) when doing restart guest
1147343 - qemu-img create -f qcow2 overlay -b ssh://xen/ segfaults on RHEL 7
1153951 - Qemu core dumped when guest reboot through guest-agent command [via isa-serial].
1154115 - [RFE] qemu chardev socket should support TLS and auth
1158310 - Error "sr 0:0:1:0: rejecting I/O to offline device" when eject cdrom
1160964 - can boot guest successfully inside qemu if maxcpus don't match topology in -smp
1161582 - Should improve error info when can't find backing file for snapshot
1166031 - Windows and Linux guest can not receive data from host via virtserialport via Unix socket backend
1177309 - Guest using rbd based image as disk failed to start when sandbox was enabled
1184125 - -cpu host: cache CPUID passthrough may not make sense depending on VM CPU topology
1202788 - There is no error prompt when set the io throttling parameters FOO_max without FOO and iops_size without iops
1202822 - QEMU SMRAM support for proper secure boot using OVMF
1209708 - floppy fail to be accessible(Location is not available) after change from none
1210221 - Netperf UDP_STREAM Lost most of the packets on spapr-vlan device
1212317 - floppy device can not be detected in the win8-64 guests under OVMF
1220252 - qemu-kvm aborts with -chardev udp,...
1221023 - the I/O throttling properties of cdrom will be changed to be none after change to a new cdrom
1232914 - preserve sparseness for drive mirror operations over distributed storage - qemu
1235499 - leaked_storage: Variable "local_err" going out of scope leaks the storage it points to
1235580 - RFE: Enable the intel-iommu device in QEMU
1235732 - spice-gtk shows outdated screen state after migration [qemu-kvm-rhev]
1238532 - qemu needs a better warning when the qemu-ifup script is busy
1238574 - qemu-kvm does not expose expected cpu topology to guest when wrong cpu topology is defined.
1240310 - {"execute":"query-block"} should not show the cdrom device that have ejected
1240497 - qemu-kvm-rhev: dump-guest-memory creates invalid header with format kdump-{zlib,lzo,snappy} on ppc64
1243548 - AIO is processing only 1 request, even if >1 requests in virtio
1244064 - the guest agent will always stay in 'disconnected' status after wakeup a guest which configured 2 cpus from 'pmsuspended' status
1245864 - qemu quit when rebooting guest which hotplug memory >=13 times
1247933 - RFE: qemu-kvm-rhev: support multiple volume hosts for gluster volumes
1248406 - Numa node assignation info is not consistency between hmp and in guest
1248758 - match the OEM ID and OEM Table ID fields of the FADT and the RSDT to those of the SLIC
1251379 - [RHEL-7.2] Package is 100% lost when ping from host to Win2012r2 guest with 64000 size
1251943 - "ENOSPAC" caused iscsi error message flood during install guest
1257103 - test case 033 099 of qemu-iotests failed for vhdx
1259440 - cpu: Improve error message when using an incompatible model
1261726 - virsh hotplug/unplug host usb device disconnects spice
1261797 - contents of MSR_TSC_AUX are not migrated
1262784 - quit qemu-kvm during boot with data plane enabled cause segmentation fault
1263591 - Guest network works abnormally(ping out or netperf test failed) when use multi queue of the virtio-net-pci macvtap
1263916 - boot guest with '-snapshot', base file is deleted after block-commit and live snapshot
1263986 - qemu crash while start a guest with invalid vnc socket path
1264726 - During block-commit, stop guest, qemu will be unresponsive until commit job finished
1264728 - Allow EEH on spapr-pci-host-bridge
1265179 - With dataplane, when migrate to remote NBD disk after drive-mirror, qemu core dump ( both src host and des host)
1265196 - posix_fallocate emulation on NFS fails with Bad file descriptor if fd is opened O_WRONLY
1265576 - hugepage could not be used inside guest if start the guest with NUMA supported huge pages
1270144 - qemu process exit after hot-plug a memory device which guest use 1G hugepages
1272759 - [PCIe] Make virtio devices pci-express on Q35 machines
1272939 - [PCIe] Add multi-root support for Q35
1273063 - Vlan table display repeat four times in qmp when queues=4
1273196 - Q35 assigned devices report Code 10 on Win7 & Win8.1 guests in default config
1273689 - Creating internal snapshot fails with data-plane
1273717 - Host lost usb device file if guest quit as soon as start with "usb-host" option
1273750 - Camera stops work after remote-viewer re-connection [qemu-kvm-rhev]
1275230 - memory unplug triggers BUG_ON: kernel BUG at mm/memory_hotplug.c:703!
1276032 - Crash on QMP input exceeding limits
1277705 - Fix handling of thin write correctly in active-commit and drive-mirror (including qemu-img commit)
1277922 - Aborted (core dumped) for "qemu-kvm: block.c:3332: bdrv_aligned_pwritev: Assertion `!qiov || bytes == qiov->size' failed." after 'lvextend' the guest primary image and 'cont' the installing guest for the 2nd time
1278447 - ceph.conf properties override qemu's command-line properties
1278894 - cold-plugged DIMM can't be unplugged on the first unplug request
1282833 - [PPC64LE] Guest freezes if qemu allocates smaller page table than requested
1285595 - vubr coredumpd during repeatedly load/unload virtio-net driver inside guest.
1286492 - vhost-user interface cannot get ip if host is a numa machine
1286506 - qemu-kvm core dumped when booting qemu-kvm process with queues=0 and vhost-user
1289285 - Live Migration dynamic cpu throttling for auto-convergence (qemu-kvm-rhev)
1289417 - Rebase to QEMU 2.6
1290150 - Include example kvm.conf with nested options commented out
1290683 - [abrt] qemu-img: get_block_status(): qemu-img killed by SIGABRT
1291209 - The cpu model that could be boot in command line does not match with host /proc/cpuinfo
1291979 - [ppc64le] Virtio-input support - qemu-kvm-rhev
1295422 - fix ivshmem segfault with msi=off option
1295907 - [RFE] add virtio-gpu and virtio-vga support
1296800 - -mem-prealloc option does not take effect when no huge page is allocated
1297334 - Guest should failed to boot if set iops,bps to negative number
1297748 - [RFE] Allocate guard pages after guest RAM to protect against buffer overflows
1298027 - qemu-img compare is unnessarily slow on raw files
1298886 - [abrt] qemu-img: get_block_status(): qemu-img killed by SIGABRT
1299030 - drmgr with invalid value in guest crashes qemu
1300773 - RFE: add support for native TLS encryption on chardev TCP transports
1301019 - RFE: add support for LUKS disk encryption format driver w/ RBD, iSCSI, and qcow2
1301057 - RFE: add support for securely passing passwords to QEMU block drivers
1303181 - qemu 2.1-2.4 crashes with segfault in phys_section_destroy
1316368 - qmp no response while start/cancel block job repeatly
1317977 - qemu-kvm-rhev supports a lot of CPU models
1319400 - live merge - qemu-kvm hangs in aio_bh_poll
1319673 - qemu-kvm core dump with "-M rhel6.6.0" and "-cpu Westmere"
1320473 - virtio-pci doesn't reset VQ correctly at device reset
1320476 - Failed to upgrade qemu-kvm-tools-rhev from 2.3.0 to 2.5.0
1320714 - [RFE] Allow the libgfapi logging level to be controlled.
1322713 - Guest Call Trace when system_reset on hmp launch guest with usb-storage device.
1323273 - migration: i82801b11 bridge broken (~q35)
1323294 - AVX-2 migration optimisation
1323976 - PCI: Add 64-bit MMIO support to PXB devices
1325991 - qemu NBD server rejects clients that use unknown NBD_OPT_*
1328020 - qemu-kvm-rhev gets SEGV when try to attach a disk image in a nfs pool to a vm with insufficient nfs authority
1329526 - seabios can not recognize virtio 1.0 virtual block
1329641 - [RFE]Ceph/RBD block driver for qemu-kvm : change error_setg() to error_setg_errno()
1330916 - image cannot be fully allocated via qemu-img with sparse_size specified as 0
1333282 - Disable new devices in QEMU 2.6
1333299 - RFE: Enable nvdimm device
1333318 - ivshmem-plain support in RHEL 7.3
1335705 - 'QEMU 2.5.94 monitor' is used for qemu-kvm-rhev-2.6.0-1.el7.x86_64
1335720 - watchdog action 'inject-nmi' takes no effect
1336296 - failed dependencies on SLOF
1336310 - virtio-scsi data-plane does not support block management QMP commands
1336517 - Disable hv-vpindex, hv-runtime, hv-reset, hv-synic & hv-stimer enlightenment for Windows
1336649 - [RHEL.7.3] Guest will not boot up when specify aio=native and snapshot=on together
1336659 - Core dump when re-launch guest with encrypted block device
1336705 - Drive mirror with option granularity fail
1337033 - kvm_stat AttributeError: 'ArchPPC' object has no attribute 'exit_reasons'
1337100 - redhat_drive_add should report error to qmp if it fails to initialize
1337917 - qemu-kvm-rhev: Only ship /usr/share/qemu-kvm/sgabios.bin on x86
1338043 - scsi-block fix - receive the right SCSI status on reads and writes
1338638 - Migration fails after ejecting the cdrom in the guest
1338755 - qemu-kvm-rhev doesn't reload udev rules before triggering for kvm device
1339196 - qemu-kvm (on target host) killed by SIGABRT when migrating a guest from AMD host to Intel host.
1339467 - User cannot create windows 7 virtual machine in rhevm 3.6.5.
1339498 - Core dump when do 'block-job-complete' after 'drive-mirror'
1341492 - QEMU on POWER does not support the PowerISA 2.07 compatibility mode
1341531 - qemu gets SIGSEGV when hot-plug a scsi hostdev device with duplicate target address
1342015 - Migration: Fixup machine types and HW_COMPAT (stage 1b)
1342330 - There is no error prompt when set the io throttling parameters iops_size without iops
1342999 - 'cache=x' cannot work with empty cdrom
1343021 - Core dump when quit from HMP after migration finished
1343092 - RFE: Integrate e1000e implementation in downstream QEMU
1343891 - Disable spapr-rng device in downstream qemu 2.6
1344259 - Q35: Device assignment to PCIe ports
1344269 - Migration: Fixup machine types and HW_COMPAT (stage 2a)
1346237 - win 10.x86_64 guest coredump when execute avocado test case: win_virtio_update.install_driver
1346627 - qemu discards EEH ioctl results
1346688 - [Q35] vfio read-only SR-IOV capability confuses OVMF
1346920 - vIOMMU: prevent unsupported configurations with vfio
1346976 - Regression from CVE-2016-3712: windows installer fails to start [rhel-7.3]
1347077 - vhost-user: A socket file is not deleted after VM's port is detached.
1347402 - vgabios-virtio.bin should be symlinked in qemu-kvm-rhev
1347498 - [ppc64le] Guest can't boot up with hugepage memdev
1348763 - Fix dirty marking with block discard requests
1350196 - Enable IOMMU device with -device intel-iommu
1350735 - memory locking limit for regular users is too low to launch guests through libvirt
1351409 - When hotplug memory, guest will shutdown as Insufficient free host memory pages available to allocate
1351442 - "TSC frequency mismatch" warning message after migration
1352799 - Client information from hmp doesn't vanish after client disconnect when using vnc display
1352860 - Migration is failed from host RHEL7.2.z to host RHEL7.3 with "-M pc-i440fx-rhel7.0.0 -device nec-usb-xhci"
1352865 - Boot guest with two virtio-scsi-pci devices and spice, QEMU core dump after executing '(qemu)__com.redhat_drive_add'
1353070 - Migration is failed from host RHEL7.2.z to host RHEL7.3 with "-M rhel6.6.0 -device e1000-82540em"
1353180 - 7.3->7.2 migration: qemu-kvm: usbredirparser: error unserialize caps mismatch
1353801 - The default io throttling group name is null, which makes all throttled disks with a default group name in the same group
1354090 - Boot guest with vhostuser server mode, QEMU prompt 'Segmentation fault' after executing '(qemu)system_powerdown'
1354341 - guest hang after cancel migration then migrate again
1354443 - /usr/libexec/qemu-kvm: undefined symbol: libusb_free_ss_endpoint_companion_descriptor
1355665 - Suggest to limit the burst value to be not less than the throttle value
1355683 - qemu core dump when do postcopy migration again after canceling a migration in postcopy phase
1355704 - spice: core dump when 'quit'
1355902 - vhost-user reconnect misc fixes and improvements
1355927 - qemu SIGABRT when doing inactive blockcommit with external system checkpoint snapshot
1355982 - qemu will abort after type two"system_reset" after the guest poweroff
1356376 - [Q35] Nic which passthrough from host didn't be found in guest when enable multifunction
1357202 - [Q35] S3 should be disabled by default for the pc-q35-rhel7.3.0 machine type
1358141 - Removal of the "x-" prefix for dynamic cpu throttling
1358653 - [RFE] Interrupt remapping support for Intel vIOMMUs
1359655 - Qemu crashes when connecting to a guest started with "-vnc none" by virt-viewer
1360225 - Can't extract guest memory dump from qemu core
1360612 - Memory leak on incomplete JSON parse
1360664 - [virtio] Update default virtio-1 behavior for virtio devices
1361205 - nvdimm: fix memory leak in error code path
1361443 - ppc64le: Introduce stable cpu_index for cpu hotplugging
1362019 - Crashes when using query-hotpluggable-cpus with pseries-rhel7.2.0 machine type
1363679 - RHEL guest hangs with kernel-irqchip=off and smp>1
1363812 - qemu-kvm-rhev: -cpu POWER8 no longer works
1363998 - Live migration via a compressed file causes the guest desktop to freeze
1364035 - [ppc64le][VFIO]Qemu complains:vfio_dma_map(0x10033d3a980, 0x1f34f0000, 0x10000, 0x3fff9a6d0000) = -6 (No such device or address)
1365613 - [PCI] The default MMIO range reserved by firmware for PCI bridges is not enough to hotplug virtio-1 devices
1365747 - Migrate guest(win10) after hot plug/unplug memory balloon device [Missing section footer for 0000:00:07.0/virtio-net]
1366471 - QEMU prints "usb-kbd: warning: key event queue full" when pressing keys during SLOF boot
1368032 - kernel crash after hot remove virtio-gpu device
1368153 - Please hide fw_cfg device in windows guest in order to make svvp test pass
1370005 - Fail to get network device info(eth0) in guest with virtio-net-pci/vhostforce
1370703 - [Balloon] Whql Job "Commom scenario stress with IO" failed on 2008-32/64
1371211 - Qemu 2.6 won't boot guest with 2 meg hugepages
1372763 - RHSA-2016-1756 breaks migration of instances
1373802 - Network can't recover when trigger EEH one time
1374251 - qemu-kvm-rhev core dumped when enabling virtio-scsi "data plane" and executing "eject"
1375677 - Crash when performing VNC websockets handshake
1377920 - Guest fails reboot and causes kernel-panic
1378788 - Race condition during virtio-blk dataplane stop triggers "Virtqueue size exceeded"
557939 - QMP: Full introspection support
878717 - Vncserver can't only listen on one kind of protocl (IPV4 or IPv6) at same time
887844 - AHCI does not support restarting requests (i.e. rerror=stop and werror=stop/enospc)
953062 - RFE: AHCI live migration
970551 - failed pci-serial hotplug leaves the chardev busy
995415 - passthrough (scsi-block) disk with "readonly=on" can be writed in guest



These packages are GPG signed by Red Hat for security. Our key and details on how to verify the signature are available from:
https://www.redhat.com/security/team/key/#package

The Red Hat security contact is secalert@redhat.com. More contact details at http://www.redhat.com/security/team/contact/