5 This is the list of official releases for libvirt, along with an overview of
6 the changes introduced by each of them.
8 For a more fine-grained view, use the `git log`_.
16 * **Removed features**
30 * qemu: Support for ras feature for virt machine type
32 It is now possible to set on/off ``ras`` feature in the domain XML for virt
33 (Arm) machine type as ``<ras state='on'/>``.
37 Libvirt now installs a binary helper that allows connecting to QEMU domains
38 via SSH using the following scheme: ``ssh user@qemu/virtualMachine``.
40 * qemu: Support for ``virtio`` sound model
42 Sound devices can now be configured to use the virtio model with
43 ``<sound model='virtio'/>``. This model is available from QEMU 8.2.0
46 * network: use nftables to setup virtual network firewall rules
48 The network driver can now use nftables rules for the virtual
49 network firewalls, rather than iptables. With the standard build
50 options, nftables is preferred over iptables (with fallback to
51 iptables if nftables isn't installed), but this can be modified at
52 build time, or at runtime via the firewall_backend setting in
53 network.conf. (NB: the nwfilter driver still uses
58 * qemu: add zstd to supported compression formats
60 Extend the list of supported formats of QEMU save image by adding zstd
63 * qemu: Implement support for hotplugging evdev input devices
65 As of this release, hotplug and hotunplug of evdev ``<input/>`` devices is
70 * virsh/virt-admin: Fix ``--help`` option for all commands
72 A bug introduced in `v10.3.0 (2024-05-02)`_ caused that the attempt to print
73 help for any command by using the ``--help`` option in ``virsh`` and
74 ``virt-admin`` would print::
77 error: command 'list' doesn't support option --help
79 instead of the help output. A workaround for the affected version is to use
84 * qemu: Fix ``virsh save`` and migration when storage in question is root_squashed NFS
86 Attempting to save a VM to a root_squash NFS mount or migrating with disks
87 hosted on such mount could, in some scenarios, result in error stating::
91 The bug was introduced in `v10.1.0 (2024-03-01)`_.
93 * qemu: Don't set affinity for isolcpus unless explicitly requested
95 When starting a domain, by default libvirt sets affinity of QEMU process to
96 all online CPUs. This also included isolated CPUs (``isolcpus=``) which is
97 wrong. As of this release, isolated CPUs are left untouched, unless
98 explicitly configured in domain XML.
100 * qemu_hotplug: Properly assign USB address to hotplugged usb-net device
102 Previously, the network device hotplug logic would try to ensure only CCW
103 or PCI addresses. With recent support for the usb-net model, USB addresses
104 for usb-net network devices are assigned automatically.
106 * qemu: Fix hotplug of ``virtiofs`` filesystem device with ``<boot order=`` set
108 The bug was introduced in `v10.3.0 (2024-05-02)`_ when attempting to reject
109 unsupported configurations. During hotplug the addresses are
110 assigned after validation and thus errorneously reject valid configs.
118 * qemu: Proper support for USB network device
120 USB address is now automatically assigned to USB network devices thus they
121 can be used without manual configuration.
123 * conf: Introduce memReserve attribute to <controller/>
125 Some PCI devices have large non-prefetchable memory. This can be a problem
126 in case when such device needs to be hotplugged as the firmware can't
127 foresee such situation. The user thus can override the value calculated at
128 start to accomodate for such devices.
132 * Improve validation of USB devices
134 Certain USB device types ('sound', 'fs', 'chr', 'ccid' and 'net') were not
135 properly handled in the check whether the VM config supports USB and thus
136 would result in poor error messages.
138 * virsh: Fix behaviour of ``--name`` and ``--parent`` used together when listing checkpoint and snapshots
140 The ``checkpoint-list`` and ``snapshot-list`` commands would ignore the
141 ``--name`` option to print only the name when used with ``--parent``.
143 * Extend libvirt-guests to shutdown only persistent VMs
145 Users can now choose to shutdown only persistent VMs when the host is being
150 * qemu: Fix migration with custom XML
152 Libvirt 10.2.0 would sometimes complain about incompatible CPU definition
153 when trying to migrate or save a domain and passing a custom XML even
154 though such XML was properly generated as migratable. Hitting this bug
155 depends on the guest CPU definition and the host on which a particular
158 * qemu: Fix TLS hostname verification failure in certain non-shared storage migration scenarios
160 In certain scenarios (parallel migration, newly also post-copy migration)
161 libvirt would wrongly pass an empty hostname to QEMU to be used for TLS
162 certificate hostname validation, which would result into failure of the
163 non-shared storage migration step::
165 error: internal error: unable to execute QEMU command 'blockdev-add': Certificate does not match the hostname
167 * Create OVS ports as transient
169 Libvirt now creates OVS ports as transient which prevents them from
170 reappearing or going stale on sudden reboots.
172 * Clear OVS QoS settings when domain shuts down
174 Libvirt now clears QoS settings on domain shutdown, so they no longer pile
183 * ch: Basic save and restore support for ch driver
185 The ch driver now supports basic save and restore operations. This is
186 functional on domains without any network, host device config defined.
187 The ``path`` parameter for save and restore should be a directory.
189 * qemu: Support for driver type ``mtp`` in ``<filesystem/>`` devices
191 The ``mtp`` driver type exposes the ``usb-mtp`` device in QEMU. The
192 guest can access files on this driver through the Media Transfer
195 * qemu: Added support for the loongarch64 architecture
197 It is now possible for libvirt to run loongarch64 guests, including on
198 other architectures via TCG. For the best results, it is recommended to
199 use the upcoming QEMU 9.0.0 release together with the development version
202 * qemu: Introduce virDomainGraphicsReload API
204 Reloading the graphics display is now supported for QEMU guests using
205 VNC. This is useful to make QEMU reload the TLS certificates without
206 restarting the guest. Available via the ``virDomainGraphicsReload`` API
207 and the ``domdisplay-reload`` virsh command.
211 * qemu: Fix migration from libvirt older than 9.10.0 when vmx is enabled
213 A domain with vmx feature enabled (which may be even done automatically
214 with ``mode='host-model'``) started by libvirt 9.9.0 or older cannot be
215 migrated to libvirt 9.10.0, 10.0.0, and 10.1.0 as the target host would
216 complain about a lot of extra ``vmx-*`` features. Migration of similar
217 domains started by the affected releases to libvirt 9.9.0 and older
218 does not work either. Since libvirt 10.2.0 migration works again with
219 libvirt 9.9.0 and older in both directions. Migration from the affected
220 releases to 10.2.0 works as well, but the other direction remains broken
221 unless the fix is backported.
223 * node_device: Don't report spurious errors from PCI VPD parsing
225 In last release the PCI Vital Product Data parser was enhanced to report
226 errors but that effort failed as some kernels have the file but don't allow
227 reading it causing logs to be spammed with::
229 libvirtd[21055]: operation failed: failed to read the PCI VPD data
231 Since the data is used only in the node device XML and errors are ignored if
232 the parsing failed, this release removes all the error reporting.
234 * qemu: set correct SELinux label for unprivileged virtiofsd
236 It is now possible to use virtiofsd-based ``<filesystem>`` shares even
237 if the guest is confined using SELinux.
239 * qemu: fix a crash on unprivileged virtiofsd hotplug
241 Hotplugging virtiofsd-based filesystems works now.
243 * virt-admin: Fix segfault when libvirtd dies
245 ``virt-admin`` no longer crashes when ``libvirtd`` unexpectedly closes
254 * ``CVE-2024-1441``: Fix off-by-one error leading to a crash
256 In **libvirt-1.0.0** there were couple of interface listing APIs
257 introduced which had an off-by-one error. That error could lead to a
258 very rare crash if an array was passed to those functions which did
259 not fit all the interfaces.
261 In **libvirt-5.10** a check for non-NULL arrays has been adjusted to
262 allow for NULL arrays with size 0 instead of rejecting all NULL
263 arrays. However that made the above issue significantly worse since
264 that off-by-one error now did not write beyond an array, but
265 dereferenced said NULL pointer making the crash certain in a
266 specific scenario in which a NULL array of size 0 was passed to the
267 aforementioned functions.
271 * nodedev: Support updating mdevs
273 The node device driver has been extended to allow updating mediated node
274 devices. Options are available to target the update against the persistent,
275 active or both configurations of a mediated device.
276 **Note:** The support is only available with at least mdevctl v1.3.0 installed.
278 * qemu: Add support for /dev/userfaultfd
280 On hosts with new enough kernel which supports /dev/userfaultfd libvirt will
281 now automatically grant QEMU access to this device. It's no longer needed to
282 set vm.unprivileged_userfaultfd sysctl.
284 * qemu: Support clusters in CPU topology
286 It is now possible to configure the guest CPU topology to use clusters.
287 Additionally, if CPU clusters are present in the host topology, they will
288 be reported as part of the capabilities XML.
290 * network: Make virtual domains resolvable from the host
292 When starting a virtual network with a new ``register='yes'`` attribute
293 in the ``<domain>`` element, libvirt will configure ``systemd-resolved``
294 to resolve names of the connected guests using the name server started
297 * qemu: Introduce dynamicMemslots attribute for virtio-mem
299 QEMU now allows setting ``.dynamic-memslots`` attribute for virtio-mem-pci
300 devices. When turned on, it allows memory exposed to guest to be split into
301 multiple memory slots and thus smaller memory footprint (see the original
302 commit for detailed explanation).
306 * nodedev: Add ability to update persistent mediated devices by defining them
308 Existing persistent mediated devices can now also be updated by
309 ``virNodeDeviceDefineXML()`` as long as parent and UUID remain unchanged.
311 * ch: Enable ``ethernet`` interface mode support
313 ``<interface type='ethernet'/>`` can now be used for CH domains.
315 * viraccessdriverpolkit: Add missing vtpm case
317 Secrets with ``<usage type='vtpm'>`` were left unable to be checked for in
318 the access driver, i.e. in ACL rules. Missing code was provided.
320 * virt-admin: Notify users to use explicit URI if connection fails
322 ``virt-admin`` doesn't try to guess the URI of the daemon to manage so a
323 failure to connect may be confusing for users if modular daemons are used.
324 Add a hint to use the URI of the dameon to manage.
328 * qemu_process: Skip over non-virtio non-TAP NIC models when refreshing rx-filter
330 If ``trustGuestRxFilters`` is enabled for a vNIC that doesn't support it,
331 libvirt may throw an error when such domain is being started, loaded from a
332 saved state, migrated, etc. These errors are now silenced, but make sure to
333 fix such configurations (after previous release it is even possible to
334 change ``trustGuestRxFilters`` value on live domains via
335 ``virDomainUpdateDeviceFlags()`` or ``virsh device-update``).
337 * domain: Fix check for overlapping ``<memory/>`` devices
339 A bug was identified which caused libvirt to report two NVDIMMs as
340 overlapping even though they weren't. This now fixed.
342 * vmx: Accept empty fileName for cdrom-image
344 Turns out, ``fileName`` attribute (which contains path to CDROM image) can
345 be set to an empty string (``""``) to denote a state in which the CDROM has
346 no medium in it. Libvirt used to reject such configuration file, but not
349 * qemu_hotplug: Don't lose 'created' flag in qemuDomainChangeNet()
351 When starting a domain, libvirt tracks what resources it created for it and
352 which were pre-existing and uses this information to preserve pre-existing
353 resources when cleaning up after said domain is shut off. But for macvtaps
354 this information was lost after the macvtap device was changed (e.g. via
355 ``virsh update-device``).
357 * Fix virStream hole handling
359 When a client sent multiple holes into a virStream it may have caused
360 daemon hangup as the daemon stopped processing RPC from the client
361 temporarily. This is now fixed.
363 * nodedev: Don't generate broken XML with certain hardware
365 A broken node device XML would be generated in a rare case when a hardware
366 device had certain characters in the VPD fields.
368 * qemu: Fix reservation of manually specified port for disk migration
370 A manually specified port would not be relased after disk migration making
371 it impossible to use it again.
379 * qemu: Enable ``postcopy-preempt`` migration capability
381 Post-copy migrations are now started with ``postcopy-preempt``
382 capability enabled as long as it is supported by both sides of migration.
383 This should enable faster migration of memory pages that the destination
384 tries to read before they are migrated from the source.
386 * qemu: Add support for mapping iothreads to virtqueues of ``virtio-blk`` devices
388 QEMU added the possibility to map multiple ``iothreads`` to a single
389 ``virtio-blk`` device and map them even to specific virtqueues. Libvirt
390 adds a ``<iothreads>`` subelement of the ``<disk> <driver>`` element that
391 users can use to configure the mapping.
393 * qemu: Allow automatic resize of block-device-backed disk to full size of the device
395 The new flag ``VIR_DOMAIN_BLOCK_RESIZE_CAPACITY`` for
396 ``virDomainBlockResize`` allows resizing a block-device backed ``raw`` disk
397 of a VM without the need to specify the full size of the block device.
399 * qemu: automatic selection/binding of VFIO variant drivers
401 When a device is assigned to a guest using VFIO with ``<hostdev
402 managed='yes'>``, libvirt will now search the running kernel's
403 modules.alias file for the most specific match to that device for
404 a VFIO driver, and bind that driver to the device rather than
405 vfio-pci. A specific driver can also be forced, using the
406 ``<driver model='plugh'/>`` attribute.
408 * qemu: add runtime configuration option for nbdkit
410 Since the new nbdkit support requires a recent selinux policy that is not
411 widely available yet, it is now possible to build libvirt with nbdkit
412 support for remote disks but disabled at runtime. This behavior is
413 controlled via the storage_use_nbdkit option of the qemu driver
414 configuration file. The option will default to being disabled, but this may
415 change in a future release and can be customized with the
416 nbdkit_config_default build option.
418 * qemu: add ID mapping support for virtiofsd
420 New ``<idmap>`` element was added for virtiofsd-based ``<filesystem>``
421 devices. It can be used to set up UID and GID mapping between host
422 and guest, making running virtiofsd unprivileged much more useful.
426 * qemu: Improve migration XML use when persisting VM on destination
428 When migrating a VM with a custom migration XML, use it as a base for
429 persisting it on the destination as users could have changed non-ABI
430 breaking facts which would prevent subsequent start if the old XML were used.
432 * qemu: Simplify non-shared storage migration to ``raw`` block devices
434 The phase of copying storage during migration without shared storage
435 requires that both the source and destination image are identical in size.
436 This may not be possible if the destination is backed by a block device
437 and the source image size is not a multiple of the block device block size.
439 Libvirt aleviates this by automatically adding a ``<slice>`` to match the
440 size of the source image rather than failing the migration.
442 * test driver: Support for hotplug/hotunplug of PCI devices
444 The test driver now supports basic hotplug and hotunplug of PCI devices.
446 * qemu: allow virtiofsd to run unprivileged
448 Nowadays virtiofsd no longer requires to run with root privileges, so the
449 restriction to always run as root is now removed from libvirt too.
453 * qemu: Various migration bug fixes and debuggability improvement
455 This release fixes multiple bugs in virsh and libvirt in handling of
456 migration arguments and XMLs and modifies error reporting for better
459 * conf: Restore setting default bus for input devices
461 Because of a regression, starting from 9.3.0 libvirt did not autofill bus
462 for input devices. With this release the regression was identified and
465 * qemu: Relax check for memory device coldplug
467 Because of a check that was too aggressive, a virtio-mem memory device
468 could not be cold plugged. This is now fixed.
470 * qemu: Be less aggressive when dropping channel source paths
472 Another regression is resolved, (introduced in 9.7.0) when libvirt was too
473 aggressive when dropping parsed paths for <channel/> sources
475 * qemuDomainChangeNet: Reflect trustGuestRxFilters change
477 On device-update, when a user requested change of trustGuestRxFilters for a
478 domain's <interface/> libvirt did nothing. It did not throw an error nor
479 did it reflect the change. Starting with this release, the change is
488 * Introduce pipewire audio backend
490 The QEMU hypervisor driver now allows setting ``pipewire`` backend for
495 * Adapt to qemu's use of protocol drivers in QCOW2 'backing file format' field
497 QEMU allows creating images where the 'backing file format' is actually a
498 protocol name such as 'file'/'host_device'/'nbd'/etc.. Adapt libvirt to
499 properly handle such images and don't assume automatic format probing is
500 necessary, which is in many cases forbidden due to security implications.
504 * qemu: Fix setup of images on hotplug of disk
506 Internal image metadata was not setup correctly which could cause some disk
507 hotplug configurations (namely those including backing images) to fail.
509 * qemu: Fix qemu crash when reverting an internal snapshot
511 Libvirt attempted to start qemu with wrong arguments when attempting to
512 revert to an internal snapshot causing qemu to crash.
514 * qemu: Fix hotplug of empty cdrom
516 Empty cdrom drive couldn't be hotplugged as libvirt wanted to setup the
517 storage backing it unconditionally.
525 * QEMU: implement reverting external snapshots
527 Reverting external snapshots is now possible using the existing API
528 ``virDomainSnapshotRevert()``. Management application can check host
529 capabilities for ``<externalSnapshot/>`` element within the list of
530 guest features to see if the current libvirt supports both deleting
531 and reverting external snapshots.
533 * virsh: add ``console --resume`` support
535 The ``virsh console`` subcommand now accepts a ``--resume`` option. This
536 will resume a paused guest after connecting to the console.
540 * virsh: Improve ``virsh start --console`` behavior
542 The ``virsh start --console`` now tries to connect to the guest console
543 before starting the vCPUs.
545 * virsh: Improve ``virsh create --console`` behavior
547 The ``virsh create --console`` now tries to connect to the guest console
548 before starting the vCPUs.
556 * network: New metadata change event
558 The network object now has a new event ID ``VIR_NETWORK_EVENT_ID_METADATA_CHANGE``
559 that can be used to get notifications upon changes in any of ``<title>``,
560 ``<description>`` or ``<metadata>``.
562 * qemu: Add support for vDPA block devices
564 With a new enough version of qemu, libvirt will allow you to assign vDPA block
565 devices to a domain. This is configured with::
567 <disk type='vhostvdpa'>
568 <source dev='/dev/vhost-vdpa-0'>
573 * qemu: add nbdkit backend for network disks
575 Up until now, libvirt supported network disks (http, ftp, ssh) by passing
576 the URL to qemu and having the appropriate qemu block drivers handle the
577 disk I/O. However, by handling the network I/O outside of the qemu process,
578 we get several advantages, such as reduced attack surface and improved
579 stability of qemu. Therefore, when available, libvirt will use nbdkit as a
580 backend for these network disks and export an NBD disk to qemu.
582 * virnetdevopenvswitch: Propagate OVS error messages
584 When configuring OVS interfaces/bridges libvirt used to report its own
585 error messages instead of passing (more accurate) error messages from
586 `ovs-vsctl`. This is now changed.
588 * Various virtio-mem/virtio-pmem fixes
590 Now libvirt validates more values of virtio-mem and virtio-pmem devices,
591 e.g. overlapping memory addresses or alignment.
599 * qemu: basic support for use of "VFIO variant" drivers
601 A VFIO variant driver is a device-specific driver that can
602 be used in place of the generic vfio-pci driver, and provides
603 extra functionality to support things like live migration of
604 guests with vfio-assigned devices. It can currently be used by:
606 1) setting ``managed='no'`` in the XML configuration for the device
607 2) pre-binding the variant driver using the ``--driver`` option of
608 ``virsh nodedev-detach``.
610 * network: Support for ``<title>`` and ``<description>`` fields in Network XML
612 The network object adds two more user defined metadata fields ``<title>``
613 and ``<description>``.
614 Two new APIs ``virNetworkGetMetadata()`` and ``virNetworkSetMetadata()`` can be
615 used to view and modify the above including the existing ``<metadata>`` field.
617 virsh adds two new commands ``net-desc`` and ``net-metadata`` to view/modify the same.
618 ``net-list`` adds a new option ``--title`` that prints the content of ``<title>``
619 in an extra column within the default ``--table`` output.
623 * qemu: Various fixes to firmware selection
625 The changes made to firmware selection in libvirt 9.2.0 have unfortunately
626 introduced a number of regressions. All known issues in this area have now
635 * ``CVE-2023-3750``: Fix race condition in storage driver leading to a crash
637 In **libvirt-8.3** a bug was introduced which in rare cases could cause
638 ``libvirtd`` or ``virtstoraged`` to crash if multiple clients attempted to
639 look up a storage volume by key, path or target path, while other clients
640 attempted to access something from the same storage pool.
644 * apparmor: All profiles and abstractions now support local overrides
646 This has long been the case for the ``virt-aa-helper`` profile, but has
647 now been extended to all other profiles and abstractions. The mechanism
648 used is the standard AppArmor 3.x one, where the contents of ``foo`` and
649 ``abstractions/foo`` can be overridden by creating ``local/foo`` and
650 ``abstractions/foo.d`` respectively.
652 * qemu: Support ``removable`` attribute for scsi disk
654 Now the scsi disk device (``/disk@device='disk'`` and
655 ``/disk/target@bus='scsi'``) supports the ``removable`` attribute at
656 ``/disk/target@removable```.
658 * qemu: Add NUMA node automatically for memory hotplug
660 Users no longer need to specify guest NUMA node in the domain XML when
661 enabling memory hotplug, libvirt automatically adds one when it is missing.
663 * qemu: Consider ``BeeGFS`` as a shared filesystem
665 Allow migration with non-shared storage for VMs accessing storage via
670 * qemu: Adapt to new way of specifying PC speaker
672 PC speaker is now usable again with newer QEMU since the change of how it
673 is specified on the command line.
675 * qemu_tpm: Try harder to create emulator state
677 Libvirt no longer considers empty directory valid SWTPM state and setup is
678 now run properly in such case.
686 * qemu: Allow configuring the ``discard-no-unref`` feature of ``qcow2`` driver
688 The new ``discard_no_unref`` attribute of the ``disk`` ``driver`` element
689 controls whether the ``qcow2`` driver in qemu unrefs clusters inside the
690 image on discard requests. Disabling cluster unrefing decreases fragmentation
695 * qemu: Include maximum physical address size in baseline CPU
697 When computing a baseline CPU definition for a set of hosts, we need to
698 include maximum physical address size in the result to make sure it is
699 compatible with all hosts even if their supported physical address sizes
702 * conf: Properly handle slots for non-DIMM ``<memory>`` devices
704 Memory devices such as ``virtio-mem`` don't need a memory slot as they are
705 PCI devices. ``libvirt`` now properly accounts the memory slots for such
706 devices as well as specifying the ``slots`` attribute of the ``<maxMemory>``
707 element is no longer needed unless DIMM-like devices are to be used.
709 * ``passt`` log and port forwarding improvements
711 Libvirt now ensures that the ``passt`` helper process can access the
712 configured log file even when it's placed in a directory without permissions.
714 The ``<portForward>`` element of a passt-backed interface can now omit the
715 ``address`` attribute as it's enough to specify a ``dev``.
719 * lxc: Allow seeking in ``/proc/meminfo`` to resove failure with new ``procps`` package
721 New version of the ``free`` command from ``procps`` package seeks into the
722 ``/proc/meminfo`` file, which was not supported by the instance of the file
723 exposed via LXC causing a failure.
725 * qemu: Fix rare race-condition when detaching a device
727 The device removal handler callback function didn't re-check the state of
728 the unplug operation after a timeout, which could rarely cause that the
729 device was removed from the VM but not the definition.
731 * qemu: Fix NUMA memory allocation logic
733 QEMU allocates memory via the emulator thread thus that has to be allowed
734 to access all configured NUMA nodes of the VM rather than just the one where
735 it's supposed to be pinned.
737 * qemu: Fix setup of ``hostdev`` backed ``<interface>``
739 The proper steps to initialize the host device were skipped for interfaces
740 due to a logic bug preventing start of VM which used them.
748 * qemu: Support compression for parallel migration
750 QEMU supports parallel migration to be compressed using either zstd or zlib.
752 * cpu_map: Add SapphireRapids cpu model
754 This model is introduced since QEMU 8.0.
758 * Adapt to musl-1.2.4
760 The latest version of musl stopped declaring some symbols that libvirt's
761 test suite used (for redirecting ``stat()`` family of functions), leaving
762 the tests broken. This is now fixed and the test suite works even with the
763 latest version of musl.
765 * conf: Introduce ``<address/>`` for virtio-mem and virtio-pmem
767 To ensure guest ABI stability, libvirt persists address for memory devices,
768 now including ``virtio-mem`` and ``virtio-pmem``. The address can be also
773 * qemu: Account for NVMe disks when calculating memlock limit on hotplug
775 When no ``<hard_limit/>`` is set, libvirt still tries to guess a sensible
776 limit for memlock for domains. But this limit was not calculated properly
777 on a hotplug of ``<disk type='nvme'/>``.
779 * numa: Deny other memory modes than ``restrictive``` if a memnode is ``restrictive``
781 Due to a missing check it was possible to define a domain with incorrect
782 ``<numatune/>``. For instance it was possible to have a ``<memnode
783 mode="restrictive"/>`` and ``<memory/>`` of a different mode. This is now
784 forbidden and if either all ``<memnode/>``-s and ``<memory/>`` have to have
785 ``restrictive`` mode, or none.
787 * qemu: Start emulator thread with more generous ``cpuset.mems``
789 To ensure memory is allocated only from configured NUMA nodes, libvirt sets
790 up cpuset CGgroup controller, even before QEMU is executed. But this may
791 prevent QEMU from setting affinity of threads that allocate memory. Since
792 these threads are spawned from the emulator thread, the initial set up must
793 be more generous and include union of all host NUMA nodes that are allowed
794 in the domain definition. Once QEMU has allocated all its memory, the
795 emulator thread is restricted further, as it otherwise would be.
803 * qemu: Introduce support for ``igb`` network interface model
805 ``igb`` is a successor to the ``e1000e`` network device using PCIe interface.
806 It was introduced in QEMU 8.0
808 * qemu: Improve handling of maximum physical address configuration
812 * qemu: Change default machine type for ARM and RISC-V
814 ARM and RISC-V architectures now use the ``virt`` machine type by default.
815 The previous defaults were nearly unusable and had to be overridden in most
818 * Improve translatable strings format substitutions
820 All translatable error messages with substitution strings were converted to
821 use positional modifiers to allow translators to shuffle around words in
822 the translation. The translations in Weblate were also updated to match.
824 * qemu: Improve validation of ``watchdog`` devices
826 Certain invalid configurations of ``watchdog`` device are now properly
829 - hotplug of always-present platform watchdogs is forbidden
830 - ``iTCO`` watchdog can be configured only once
831 - ``ib700`` watchdog is allowed only on ``i440fx`` machines
833 * Improved output of ``virt-host-validate`` on ARM
835 Our validation tool now parses the ``IORT`` data on ARM to properly detect
836 presence of SMMU and other features.
840 * qemu: Fix inactive internal snapshots of VM with UEFI firmware
842 Recent changes to UEFI firmware handling resulted into breaking support
843 for inactive internal snapshots of VMs with UEFI which historically worked.
844 (Although the intention was to disallow them together with active ones, but
845 the check did not work properly.)
847 Preserve existing functionality by allowing such snapshots explicitly.
849 * qemu: Properly configure locked memory limit for VMs with ``<disk type='nvme'``
851 The NVMe driver in qemu requires some memory to be locked. This was not
852 taken into account in the code which calculates the memory limits based
853 on devices present in the configuration
855 * Fix native build on win32
857 Various improvements to the build system now allow users to build the client
858 library of libvirt on win32 natively.
860 * qemu: Properly detect tray of hotplugged CD-ROM devices
862 Media in a CD-ROM device which was hotplugged could not be changed as the
863 presence of the tray was not detected properly on hotplug.
871 * qemu: Add support for QCOW2 formatted firmware
873 This type of firmware can be picked up either automatically, if the
874 corresponding JSON descriptor has the highest priority, or manually by
875 using ``<loader format='qcow2'/>`` in the domain XML.
877 * qemu: Implement QEMU NBD reconnect delay attribute
879 Support the nbd reconnect-delay of QEMU. It will set the delay time for
880 reconnect after an unexpected disconnect or a serious error.
884 * qemu: Make firmware selection persistent
886 Up until now, firmware autoselection has been performed at domain startup
887 time: as a result, changes to the JSON firmware descriptors present on the
888 system could have translated to a different firmware being chosen for
889 subsequent startups of the same domain, potentially rendering it unbootable
890 or lowering the security guarantees. Firmware selection now happens once,
891 when the domain is defined, and its results are stored in the domain XML
892 to be reused, unchanged, for all subsequent boots.
894 * qemu: passt now works when SELinux/AppArmor is enabled
896 In the case of SELinux, this requires passt-specific support code to be
897 present in the host policy, so it might only work with upcoming operating
898 systems and not with existing ones.
900 * xen: Support custom UEFI firmware paths
902 The Xen libxl driver now supports specifying a custom UEFI firmware path.
903 Previously the Xen default was used in all cases.
907 * qemu: Fix validation of the HPET timer
909 Due to a logic bug introduced in libvirt 9.0.0, VM configurations
910 explicitly enabling the HPET timer were rejected.
912 * qemu: Fix thread-context .host-nodes generation
914 With new enough QEMU, libvirt instructs QEMU to set affinity of memory
915 allocation threads. But this may have resulted in QEMU being unable to do
916 so, as affinity to NUMA nodes inaccessible to emulator thread might have
919 * rpc: fix typo in admin code generation
921 Fix the bug in the remote ``virt-admin`` code generator, that resulted
922 in a crash. Introduced in libvirt 9.1.0.
924 * qemu: relax shared memory check for vhostuser daemons
926 Fix hotplug of virtiofs ``filesystem`` after restarting libvirtd.
927 Before, libvirtd would incorrectly complain about missing shared
934 * **Removed features**
936 * vbox: removed support for version 5.2 and 6.0 APIs
938 Libvirt no longer supports use of VirtualBox 5.2 and 6.0 since these
939 versions reached their end of life on 2020/07.
943 * vbox: added support for version 7.0 API
945 Libvirt can now support use of the VirtualBox 7.0, This is compile tested
946 only, so we are looking for feedback from users on how well it works in
949 * qemu: Support crypto device
951 Support crypto device(virtio crypto only), also add support for QEMU with
952 backend ``builtin`` and ``lkcf``.
954 * qemu: added support for pvpanic-pci device
956 A pvpanic device can be now defined as a PCI device (the original is an ISA
957 device) with ``<panic model='pvpanic'/>``.
959 * qemu: support automatic restart of inadvertently terminated passt process
961 If the passt process that is serving as the backend of a -netdev
962 stream is terminated unexpectedly, libvirt now listens to QEMU's
963 notification of this, and starts up a new passt instance, thus
964 preserving network connectivity.
968 * RPM packaging changes
970 The ``libvirt-daemon`` subpackage is split into several new subpackages,
971 allowing installation of a modular daemon configuration without the
972 traditional monolithic libvirtd.
976 * QEMU: iTCO watchdog made operational
978 The watchdog was always included when q35 machine type was used, but needed
979 an extra bit of configuration in order to be operational. This is now done
980 by default when running a QEMU domain with q35 machine type. This is not a
981 change in the guest ABI, but it is a guest visible behavior change since the
982 watchdog that did not fire before will now fire once used. To switch to the
983 previous behavior the watchdog action must be set to ``none``.
985 * QEMU: fix deleting memory snapshot when deleting external snapshots
987 When external snapshot deletion was introduced it did not remove memory
988 snapshot when it existed. In addition when external memory only snapshot
989 was created libvirt failed without producing any error.
991 * QEMU: properly report passt startup errors
993 Due to how the child passt process was started, the initial
994 support for passt (added in 9.0.0) would not see errors
995 encountered during startup, so libvirt would continue to setup and
996 start the guest; this led to a running guest with no network
999 (NB: On systems that use them, it is still necessary to disable
1000 SELinux/AppArmor to start passt. This is a temporary limitation,
1001 and use of the feature in production is strongly discouraged
1002 until it has been lifted.)
1004 * qemu: Fix error when attempting to change media in a CDROM drive
1006 Due to a logic bug introduced in libvirt-9.0 attempts to change media in a
1007 CDROM would previously fail with an error stating that the tray isn't open.
1009 * qemu: Properly handle block job transitions
1011 Starting with libvirt-9.0 the block job state machine improperly handled
1012 some job transitions, which resulted into some block jobs not being
1013 properly terminated. This could cause problems such as errors when
1014 detaching a disk after snapshot.
1016 * virsh: Make domif-setlink work more than once
1018 There was a bug introduced in the previous release which made ``virsh
1019 domif-setlink`` work exactly once over given domain. The bug was fixed and
1020 now the command can be run multiple times.
1022 * qemu: Make domain startup fail if NIC already exists
1024 When starting a domain with an ``<interface/>`` that's supposed to be
1025 managed by libvirt (``managed='yes'``) but corresponding TAP device already
1026 exists, report an error and make the startup process fail.
1028 * qemu: Deal with nested mounts when umount()-ing /dev
1030 When setting up private ``/dev`` for a domain (also known as ``namespaces``
1031 in ``qemu.conf``), libvirt preserves mount points nested under ``/dev``
1032 (e.g. ``/dev/shm``, ``/dev/pts`` and so on). But there was a bug which
1033 resulted in inability to construct the namespace when there were two or
1034 more filesystems mounted on the same path. This is common scenario with
1035 containers and thus the bug was fixed.
1037 * remote: Pass ``mode`` and ``socket`` URI parameters to virt-ssh-helper
1039 When connecting to a remote host using SSH transport, ``?mode=`` and
1040 ``?socket=`` URI parameters were ignored. This prevented users from
1041 connecting to a monolithic daemon running on a remote host.
1043 * qemu: Various ``swtpm`` related fixes
1045 There are more cleanups and small bug fixes with regards to emulated
1046 ``<tpm/>``. For instance with migration when the ``swtpm`` state is on a
1047 shared volume, or seclabel setting/restoring.
1055 * QEMU: implement external snapshot deletion
1057 External snapshot deletion is now possible using the existing API
1058 ``virDomainSnapshotDelete()``. Flags that allow deleting children
1059 or children only are not supported.
1061 * QEMU: support passt (https://passt.top)
1063 passt can be used to connect an emulated network device to the
1064 host's network without requiring libvirt to have any sort of
1065 elevated privileges. This is configured with::
1067 <interface type='user'>
1068 <backend type='passt'>
1071 * QEMU: add external backend for swtpm
1073 Connecting the VM to a swtpm daemon started outside of libvirt
1076 * QEMU: Support for passing FDs instead of opening files for `<disk>`
1078 A new API `virDomainFDAssociate` gives the users the option to pass FDs
1079 to libvirt and then use them when starting a VM. Currently the FDs can
1080 be used instead of directly opening files as `<disk>` backend.
1084 * qemu: Prefer PNG for domain screenshots
1086 With sufficiently new QEMU (v7.1.0) screenshots change format from PPM to PNG.
1088 * tools: Fix install_mode for some scripts
1090 Scripts from the following list were installed with group write bit set:
1091 virt-xml-validate, virt-pki-validate, virt-sanlock-cleanup,
1092 libvirt-guests.sh. This was changed so that only the owner is able to write
1095 * qemu: Allow multiple nodes for preferred policy
1097 Due to restrictions of old kernels and libnuma APIs, the preferred NUMA
1098 policy accepted just a single host NUMA node. With recent enough kernel
1099 (v5.15.0) and libnuma (v2.0.15) it's possible to set multiple nodes.
1101 * secret: Inhibit shutdown of daemon for ephemeral secrets
1103 When an ephemeral secret is defined then automatic shutdown of virtsecretd
1104 is inhibited. This is to avoid ephemeral secrets disappearing shortly
1107 * qemu: Report Hyper-V Enlightenments in domcapabilities
1109 The supported Hyper-V Enlightenments are now reported in domain
1114 * Fix NULL-pointer dereference `virXMLPropStringRequired`
1116 Fix a bug where when parsing a XML property which is required to be present
1117 by using `virXMLPropStringRequired` the parser will crash instead of
1120 * qemu: Init ext devices paths on reconnect
1122 Paths for external devices are not stored in the status XML. Therefore,
1123 when the daemon restarted and was reconnecting to a running domain, these
1124 paths were left blank which led to the daemon crash.
1126 * qemu: Validate arguments passed to `virConnectGetDomainCapabilities`
1128 There was a code path in which insufficient validation of input arguments
1129 of `virConnectGetDomainCapabilities` API was possible which led to the
1130 daemon crash. This path is now fixed.
1133 v8.10.0 (2022-12-01)
1134 ====================
1138 * Tool for validating SEV firmware boot measurement of QEMU VMs
1140 The ``virt-qemu-sev-validate`` program will compare a reported SEV/SEV-ES
1141 domain launch measurement, to a computed launch measurement. This
1142 determines whether the domain has been tampered with during launch.
1144 * Support for SGX EPC (enclave page cache)
1146 Users can add a ``<memory model='sgx-epc'>`` device to launch a VM with
1147 ``Intel Software Guard Extensions``.
1149 * Support migration of vTPM state of QEMU vms on shared storage
1151 Pass ``--migration`` option if appropriate in order for ``swtpm`` to
1152 properly migrate on shared storage.
1156 * Mark close callback (un-)register API as high priority
1158 High priority APIs use a separate thread pool thus can help in eliminating
1159 problems with stuck VMs. Marking the close callback API as high priority
1160 allows ``virsh`` to properly connect to the daemon in case the normal
1161 priority workers are stuck allowing other high priority API usage.
1163 * Updated x86 CPU features
1165 The following features for the x86 platform were added:
1166 ``v-vmsave-vmload``, ``vgif``, ``avx512-vp2intersect``, ``avx512-fp16``,
1167 ``serialize``, ``tsx-ldtrk``, ``arch-lbr``, ``xfd``, ``intel-pt-lip``,
1168 ``avic``, ``sgx``, ``sgxlc``, ``sgx-exinfo``, ``sgx1``, ``sgx2``,
1169 ``sgx-debug``, ``sgx-mode64``, ``sgx-provisionkey``, ``sgx-tokenkey``,
1170 ``sgx-kss``, ``bus-lock-detect``, ``pks``, ``amx``.
1172 * Add support for ``hv-avic`` Hyper-V enlightenment
1174 ``qemu-6.2`` introduced support for the ``hv-avic`` enlightenment which
1175 allows to use Hyper-V SynIC with hardware APICv/AVIC enabled.
1177 * qemu: Run memory preallocation with numa-pinned threads
1179 Run the thread allocating memory in the proper NUMA node to reduce overhead.
1181 * RPM packaging changes
1183 - add optional dependency of ``libvirt-daemon`` on ``libvirt-client``
1185 The ``libvirt-guests.`` tool requires the ``virsh`` client to work
1186 properly, but we don't want to require the installation of the daemon
1187 if the tool is not used.
1189 - relax required ``python3-libvirt`` version for ``libvirt-client-qemu``
1191 The ``virt-qemu-qmp-proxy`` tool requires python but doesn't strictly
1192 need the newest version. Remove the strict versioning requirement in
1193 order to prevent cyclic dependency when building.
1197 * Skip initialization of ``cache`` capabilities if host doesn't support them
1199 Hypervisor drivers would fail to initialize on ``aarch64`` hosts with
1202 virStateInitialize:657 : Initialisation of cloud-hypervisor state driver failed: no error
1204 which prevented the startup of the daemon.
1206 * Allow incoming connections to guests on routed networks w/firewalld
1208 A change in handling of implicit rules in ``firewalld 1.0.0`` broke
1209 incoming connections to VMs when using ``routed`` network. This is fixed
1210 by adding a new ``libvirt-routed`` zone configured to once again allow
1211 incoming sessions to guests on routed networks.
1213 * Fix infinite loop in nodedev driver
1215 Certain udev entries might be of a size that makes libudev emit EINVAL
1216 which caused a busy loop burning CPU. Fix it by ignoring the return code.
1224 * Add ``virt-qemu-qmp-proxy`` for emulating a QMP socket for libvirt managed VMs
1226 ``virt-qemu-qmp-proxy`` tool provides a way to expose an emulated QMP server
1227 socket for a VM managed by libvirt. This allows existing QMP-only clients
1228 to work with libvirt managed VMs.
1230 **Note:** libvirt is not interpreting the communication between the tool
1231 using the proxy and qemu itself, so any state-changing commands may
1232 desynchronize libvirt. Use at your own risk.
1234 * qemu: Core Scheduling support
1236 To avoid side channel attacks, the Linux kernel allows creating groups of
1237 processes that trust each other and thus can be scheduled to run on
1238 hyperthreads of a CPU core at the same time. This is now implemented for
1239 QEMU domains too (see ``sched_core`` knob in qemu.conf), although not
1240 enabled by default, just yet.
1244 * qemu: Add hypervisor-specific statistics to ``virConnectGetAllDomainStats``
1246 The new stats group ``VIR_DOMAIN_STATS_VM`` of
1247 ``virConnectGetAllDomainStats``, also exposed as ``virsh domstats --vm``,
1248 returns hypervisor-specific stats fields for given VM.
1250 * Add ``vendor`` attribute for CPU models in domain capabilities
1252 Users can now see the vendor of each CPU model in domain capabilities and
1253 use it, e.g., for filtering usable CPU models based on host CPU vendor.
1255 * virsh: Add ``--model`` option for ``hypervisor-cpu-baseline``
1257 This is a shortcut for calling ``hypervisor-cpu-baseline`` with a single
1258 CPU model and no additional features. It can be used for determining which
1259 features block a particular CPU model from being usable.
1261 * Improved documentation of CPU ``usable`` attribute in domain capabilities
1263 * Report ``channel`` and ``redirdev`` devices in domain capabilities
1265 The channel and redirect devices supported by the hypervisor are now
1266 reported in domain capabilities.
1268 * meson: Bump minimal required meson version
1270 Newer meson versions deprecate some functions used. These were replaced
1271 with their newer counterparts and the minimal required mesion version was
1274 * qemu: Add flags to keep or remove TPM state for ``virDomainUndefineFlags``
1276 ``VIR_DOMAIN_UNDEFINE_TPM`` and ``VIR_DOMAIN_UNDEFINE_KEEP_TPM`` specify
1277 accordingly to delete or keep a TPM's persistent state directory structure
1278 and files when undefining a domain. In virsh the flags are exposed as
1279 ``--tpm`` and ``--keep-tpm`` for the sub-command ``undefine``.
1283 * qemu: Disable all blocker features in CPU baseline
1285 Three years ago QEMU renamed some CPU features (mostly those containing
1286 an underscore). When such renamed feature was reported by QEMU as blocking
1287 usability of a CPU model, we would fail to explicitly disable it when
1288 creating a baseline CPU definition using this model. This bug did not have
1289 any functional impact when the default ``check='partial'`` attribute was
1290 used for guest CPU definition in domain XML, but it could have caused
1291 failures to start a domain with ``check='full'`` in some cases.
1293 * qemu: Do not crash after restart with active migration
1295 In 8.8.0 release libvirt daemon would crash after it was restarted during
1296 an active outgoing migration.
1298 * qemu: Refresh state after restore from a save image
1300 When a domain is restored from a saved image, libvirt now queries QEMU for
1301 those parts of runtime information that were not part of the save image.
1302 For instance: MAC address of a macvtap NICs, tray state of CD-ROMs,
1303 allocated size of virtio-mem, and others.
1309 * **Removed features**
1311 * storage: Remove 'sheepdog' storage driver backend
1313 The 'sheepdog' project is no longer maintained and upstream bug reports
1314 are unaddressed. Libvirt thus removed the support for the sheepdog storage
1315 driver backend, following qemu's removal of sheepdog support in qemu-6.1.
1319 * qemu: Implement VIR_DOMAIN_STATS_CPU_TOTAL for qemu:///session
1321 Users can now query VIR_DOMAIN_STATS_CPU_TOTAL (also known as cpu.time)
1322 statistics for session domains.
1326 * qemu: Fix non-shared storage migration setup
1328 This release fixes a bug in setup of a migration with non-shared storage
1329 ( ``virsh migrate --copy-storage-all``) which was broken by a refactor of
1330 the code in libvirt-8.7.
1332 * selinux: Don't ignore NVMe disks when setting image label
1334 Libvirt did not set any SELinux label on NVMe disks and relied only on the
1335 default SELinux policy. This turned out to cause problem when using
1336 namespace or altered policy and thus is fixed now.
1338 * qemu: Fix a deadlock when setting up namespace
1340 When starting a domain, libvirt creates a mount namespace and manages
1341 private /dev with only a handful nodes exposed. But when creating those a
1342 deadlock inside glib might have occurred. The code was changed so that
1343 libvirt does not tickle the glib bug.
1345 * qemu: Don't build memory paths on daemon restart
1347 When the daemon is restarted it tried to create domain private paths for
1348 each mounted hugetlbfs. When this failed, the corresponding domain was
1349 killed. This operation is now performed during domain startup and memory
1350 hotplug and no longer leads to sudden kill of the domain.
1356 * **Removed features**
1358 * qemu: Remove support for QEMU < 4.2
1360 In accordance with our platform support policy, the oldest supported QEMU
1361 version is now bumped from 3.1 to 4.2.
1365 * qemu: Add support for specifying vCPU physical address size in bits
1367 Users can now specify the number of vCPU physical address bits with
1368 the `<maxphysaddr>` subelement of the `<cpu>` element.
1372 * esx: Domain XMLs can now be dumped for VMs with two new interface types
1374 One is when the interface is not connected anywhere `type='null'` and one
1375 when it is connected to VMWare Distributed Switch `type='vds'`.
1379 * qemu: increase memlock limit for a domain with multiple vfio/vdpa devices
1381 When multiple vfio or vdpa devices are assigned to a domain, the locked
1382 memory limit could be too low to map memory for all devices. The memlock
1383 limit has been increased to be proportional to the number of vdpa/vfio
1392 * conf: Improved firmware autoselection
1394 The firmware autoselection feature now behaves more intuitively, reports
1395 better error messages on failure and comes with high-level documentation.
1403 * qemu: Introduce support for network backed NVRAM
1405 Users can now use remote store NVRAM image by specifying newly introduced
1406 attribute `type='network'` with `<nvram>` element.
1408 * qemu: Add support for post-copy migration recovery
1410 A new ``VIR_MIGRATE_POSTCOPY_RESUME`` flag (``virsh migrate --postcopy-resume``)
1411 was introduced for recovering from a failed post-copy migration.
1413 * qemu: Add support for zero-copy migration
1415 With QEMU 7.1.0, libvirt can enable zerocopy for parallel migration. This
1416 is implemented by adding a new ``VIR_MIGRATE_ZEROCOPY`` flag(``virsh migrate
1419 * Introduce thread_pool_min and thread_pool_max attributes to IOThread
1421 New attributes ``thread_pool_min`` and ``thread_pool_max`` were introduced
1422 to ``<iothread/>`` as well as new ``<defaultiothread/>`` element with the
1423 same attributes. This way it's possible to instruct QEMU to spawn enough
1424 worker threads for an IOThread upfront, resulting in predictable time
1425 needed to process an I/O request.
1429 * Define a TFTP server without a DHCP server in network configuration
1431 It's now possible to define a network with no DHCP server but with a TFTP
1432 server. This may be useful when DHCP service is provided by other entity on
1433 the network than libvirt spawned dnsmasq.
1437 * qemu: Restore label to temp file in qemuDomainScreenshot()
1439 When virDomainScreenshot() is called, libvirt instructs QEMU to save the
1440 screenshot into a temporary file. This file needs to be labelled correctly,
1441 so that QEMU can access it. And since the file is temporary (it's deleted
1442 after the screenshot was taken) the corresponding label restore was
1443 missing. This proven to be problematic for profile based models, like
1444 AppArmor, where the temporary files were added into the profile but never
1445 removed, which resulted in longer profile recalculation times.
1447 * qemuBuildInterfaceConnect: Initialize @tapfd array
1449 Due to an uninitialized array, unsuccessful attempt to start a guest with
1450 an ``<interface/>`` might have resulted in closing of a random FD and thus
1451 sudden disconnect of a client or other random failures.
1453 * qemu: Fix hotplug of network interfaces
1455 A logic bug introduced in a recent refactor was fixed. The bug caused a
1456 problem when hot-adding a network interface, which failed with the
1459 error: internal error: unable to execute QEMU command 'netdev_add': File descriptor named '(null)' has not been found
1461 * Fix ``startupPolicy`` validation for ``block`` disks
1463 Setting of ``startupPolicy`` for a block disk would result in an error due
1464 to a logic bug in a recent refactor.
1466 * qemu: Fix crash when overriding device properties via ``<qemu:override>`` element
1468 Adding an override for a device property would result in a crash of the qemu
1477 * qemu: D-Bus display
1479 Libvirt is now able to setup a D-Bus display export, either with a private
1480 bus or in p2p mode. This display is available in QEMU 7.0.0.
1482 * qemu: ppc64 Power10 processor support
1484 Support for the recently released IBM Power10 processor was added.
1486 * qemu: Introduce ``absolute`` clock offset
1488 The ``absolute`` clock offset type allows to set the guest clock to an
1489 arbitrary epoch timestamp at each start. This is useful if some VM needs
1490 to be kept set to an arbitrary time for e.g. testing or working around
1493 * qemu: add qemu-vdagent channel
1495 This paravirtualized qemu vdagent channel can enable copy and paste between
1496 a guest and a VNC client. It is available in QEMU 6.1.0.
1498 * api: Add new APIs ``virDomainSaveParams`` and ``virDomainRestoreParams``
1500 * ``virDomainSaveParams``: An alternative domain saving API, extends
1501 ``virDomainSaveFlags`` by adding parameters.
1502 * ``virDomainRestoreParams``: An alternative domain restoring API, extends
1503 ``virDomainRestoreFlags`` by adding parameters.
1507 * Improve heuristics for computing baseline CPU models
1509 Both ``virConnectBaselineHypervisorCPU`` and ``virConnectBaselineCPU`` were
1510 in some cases computing the result using a CPU model which was newer than
1511 some of the input models. For example, ``Cascadelake-Server`` was used as a
1512 baseline for ``Skylake-Server-IBRS`` and ``Cascadelake-Server``. The CPU
1513 model selection heuristics was improved to choose a more appropriate model.
1519 * **Removed features**
1521 * qemu: Remove support for QEMU < 3.1
1523 In accordance with our platform support policy, the oldest supported QEMU
1524 version is now bumped from 2.11 to 3.1.
1528 * qemu: Introduce support for virtio-iommu
1530 This IOMMU device can be used with both Q35 and ARM virt guests.
1532 * qemu: Introduce attributes rss and rss_hash_report for net interface
1534 They can enable in-qemu/ebpf RSS and in-qemu RSS hash report for virtio NIC.
1535 Require QEMU >= 5.1.
1543 * qemu: Introduce ``manual`` disk snapshot mode
1545 This new mode allows users to synchronize libvirt snapshots with snapshots
1546 which need to be done outside of libvirt e.g. when 'vhost-user-blk' is used
1549 * Introduce memory allocation threads
1551 When starting a QEMU guest, libvirt can now instruct QEMU to allocate
1552 guest's memory in parallel. This may be handy when guest has large amounts
1557 * qemu: ``VIR_MIGRATE_PARAM_TLS_DESTINATION`` now works with non-shared storage migration
1559 The setting now also applies to the NBD connections for non-shared storage
1560 migration allowing migration to proceed even when the user expects certificate
1563 * qemu: Allow overrides of device properties via the qemu namespace
1565 Users wishing to override or modify properties of devices configured by
1566 libvirt can use the ``<qemu:deviceOverride>`` QEMU namespace element to
1567 specify the overrides instead of relying on the argv passthrough of the
1568 ``-set`` qemu commandline option which no longer works with new qemu.
1570 * qemu: Allow passing file descriptors to ``virsh qemu-monitor-command``
1572 Passing FDs allows users wanting to experiment with qemu driven by libvirt
1573 use commands like ``add-fd`` properly.
1575 * libxl: Turn on user aliases
1577 Users can now use so called user aliases for XEN domains.
1579 * Implement support for FUSE3
1581 The LXC driver uses fuse to overwrite some lines in ``/proc/meminfo``
1582 inside containers so that they see correct amount of memory given to them.
1583 The code was changed so that both ``fuse`` and ``fuse3`` are supported.
1585 * Improve domain save/restore throughput
1587 Code that's handling save or restore of QEMU domains was changed resulting
1588 in better performance of I/O and thus shortening time needed for the operation.
1592 * Both build and tests should now pass on Alpine Linux or any other
1593 distribution with musl libc.
1595 * virsh: Fix integer overflow in allocpages
1597 On hosts which support hugepages larger than 1GiB ``virsh allocpages``
1598 failed to accept them because of an integer overflow. This is now fixed.
1600 * qemu: Fix segmentation fault in virDomainUndefineFlags
1602 When a domain without any ``<loader/>`` was being undefined, libvirt has
1603 crashed. This is now fixed.
1605 * lxc: Fix unaligned reads of /proc/meminfo within a container
1607 When /proc/meminfo was read in chunks smaller than the entire file, libvirt
1608 would produce mangled output. While porting the code to FUSE3 this area was
1609 reworked and the file can now be read with any granularity.
1611 * qemu: Be less aggressive around cgroup_device_acl
1613 A basic set of devices common to every domain can be set in ``qemu.conf``
1614 via cgroup_device_acl knob. Devices from this set are allowed in CGroup and
1615 created in domain private namespace for every domain. However, upon device
1616 hotunplug it may have had happened that libvirt mistakenly denied a device
1617 from this set and/or removed it from the namespace. For instance,
1618 /dev/urandom was removed and denied in CGroup on RNG hotunplug.
1620 * nodedev: trigger mdev device definition update on udev add and remove
1622 When nodedev objects are added and removed mdev device definitions are
1623 updated to report correct associated parent.
1631 * qemu: Add hvf domain type for Hypervisor.framework
1633 It works on Intel machines as well as recent machines powered by Apple
1634 Silicon. QEMU 6.2.0 is needed for Apple Silicon support.
1636 * qemu: Support mode option for dirtyrate calculation
1638 Introduce ``virDomainDirtyRateCalcFlags`` as parameter of
1639 ``virDomainStartDirtyRateCalc``, which is used to specify the mode of
1640 dirty page rate calculation.
1642 Add ``--mode`` option to ``virsh domdirtyrate-calc``, which can be
1643 either of the following 3 options:
1644 ``page-sampling, dirty-bitmap, dirty-ring``.
1646 Add ``calc_mode`` field for dirtyrate statistics returned by
1647 ``virsh domstats --dirtyrate``, also add ``vCPU dirtyrate`` if
1648 ``dirty-ring`` mode was used in last measurement.
1652 * packaging: sysconfig files no longer installed
1654 libvirt used to provide defaults in various /etc/sysconfig/ files, such
1655 as /etc/sysconfig/libvirtd. Since these files are owned by the admin, this
1656 made it difficult to change built-in defaults in case such file was
1657 modified by the admin. The built-in defaults are now part of the provided
1658 systemd unit files, such as libvirtd.service. These unit files continue
1659 to parse sysconfig files, in case they are created by the admin and filled
1660 with the desired key=value pairs.
1662 * virnetdev: Ignore EPERM on implicit clearing of VF VLAN ID
1664 Libvirt will now ignore EPERM errors on attempts to implicitly clear a
1665 VLAN ID (when a VLAN is not explicitly provided via an interface XML
1666 using a 0 or a non-zero value) as SmartNIC DPUs do not expose VLAN
1667 programming capabilities to the hypervisor host. This allows Libvirt
1668 clients to avoid specifying a VLAN and expect VF configuration to work
1669 since Libvirt tries to clear a VLAN in the same operation
1670 as setting a MAC address for VIR_DOMAIN_NET_TYPE_HOSTDEV devices which
1671 is now split into two distinct operations. EPERM errors received while
1672 trying to program a non-zero VLAN ID or explicitly program a VLAN ID 0
1673 will still cause errors as before so there is no change in behavior
1678 * Remove unix sockets from filesystem when disabling a '.socket' systemd unit
1680 The presence of the socket files is used by our remote driver to determine
1681 which service to access. Since neither systemd nor the daemons clean up the
1682 socket file clients were running into problems when a modular deployment was
1683 switched to monolithic ``libvirtd``.
1685 * qemu: Fixes of fd passing during hotplug and hotunplug of chardevs
1687 FDs used as chardev backing are now properly removed when hot-unplugging
1688 a chardev from qemu and hotplugged chardevs now properly use ``virtlogd``
1689 to handle the input and output from qemu.
1691 * RPM: Run pre/post-install steps on ``daemon-driver-storage-core``
1693 Previously the pre/post-install code was part of the meta-package which
1694 installed all storage driver sub-packages thus a minimalistic install
1695 of the storage driver didn't behave correctly.
1703 * libxl: Fix potential deadlock and crash (CVE-2021-4147)
1705 A rogue guest could continuously reboot itself and cause libvirtd on the
1706 host to deadlock or crash, resulting in a denial of service condition.
1708 * **Removed features**
1710 * qemu: Explicitly forbid live changing nodeset for strict numatune
1712 For ``strict`` mode of <numatune/> it can't be guaranteed that memory is
1713 moved completely onto new set of nodes (e.g. QEMU might have locked pieces
1714 of its memory) thus breaking the strict promise. If live migration of QEMU
1715 memory between NUMA nodes is desired, users are advised to use
1716 ``restrictive`` mode instead.
1720 * qemu: Synchronous write mode for disk copy operations
1722 The ``blockdev-mirror`` block job supports a mode where writes from the VM
1723 are synchronously propagated to the destination of the copy. This ensures
1724 that the job will converge under heavy I/O.
1726 Implement the mode for the copy blockjob as
1727 ``VIR_DOMAIN_BLOCK_COPY_SYNCHRONOUS_WRITES`` flag exposed via
1728 ``virsh blockcopy --synchronous-writes`` and for non-shared storage migration
1729 as ``VIR_MIGRATE_NON_SHARED_SYNCHRONOUS_WRITES`` exposed via
1730 ``virsh migrate --copy-storage-synchronous-writes``.
1732 * Introduce TCG domain features
1734 Libvirt is now able to set the size of translation block cache size
1735 (tb-size) for TCG domains.
1737 * qemu: Add new API to inject a launch secret in a domain
1739 New API ``virDomainSetLaunchSecurityState()`` and virsh command
1740 ``domsetlaunchsecstate`` are added to support injecting a launch secret
1741 in a domain's memory.
1745 * libxl: Implement the virDomainGetMessages API
1747 * qemu: Preserve qcow2 sub-cluster allocation state after external snapshots and block-copy
1749 The new image which is installed as an overlay on top of the current chain
1750 when taking an external snapshot, or the target of a block copy operation
1751 now enables sub-cluster allocation (``extended_l2``) if the original
1752 image has the option enabled.
1756 * qemu: Fix device hot-unplug with ``libvirt-7.9`` or ``libvirt-7.10`` used with ``qemu-6.2``
1758 An internal change to the configuration format used by the above libvirt
1759 versions triggers a bug in ``qemu-6.2`` where qemu no longer emits the
1760 event notifying that the device was unplugged successfully and thus libvirt
1761 never removes the device from the definition.
1763 This impacts only devices which were present at startup of the VM, hotplugged
1764 devices behave correctly.
1766 This is fixed in ``libvirt-8.0`` by reverting to the old configuration
1767 approach until qemu is fixed.
1769 As a workaround for ``libvirt-7.9`` and ``libvirt-7.10`` the old configuration
1770 approach can be forced by:
1772 Option 1, global ``qemu.conf``::
1774 capability_filters = [ "device.json" ]
1776 Option 2, per VM XML override::
1778 <domain type='kvm' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'>
1783 <qemu:del capability='device.json'/>
1784 </qemu:capabilities>
1787 * Fix sparse streams with split daemon
1789 In split daemon scenario, a client connected to a hypervisor driver and
1790 using sparse streams (e.g. ``virsh vol-download --sparse``) would make the
1791 hypervisor daemon enter an infinite loop without any data transfer. This is
1794 * Build no longer requires RPC library
1796 Code and its cross dependencies were fixed so that build without remote
1797 driver and thus an RPC library (like ``tirpc``) fails no more.
1799 * virnetdevopenvswitch: Fix 'burst' value passed to ovs-vsctl
1801 When a ``<bandwidth/>`` was defined for a TAP device that's plugged into an
1802 OvS bridge values passed to the OvS were incorrectly recalculated resulting
1803 in slightly different limits being applied.
1806 v7.10.0 (2021-12-01)
1807 ====================
1811 * Added virt-pki-query-dn binary
1813 This binary helps users figure out the format of Distinguished Name
1814 from a certificate file the way that libvirt expects it in
1815 tls_allowed_dn_list option of libvirtd.conf configuration file
1819 * qemu: Report guest interface information in ``virDomainGetGuestInfo``
1821 Libvirt is now able to report interface information from the guest's
1822 perspective (using guest agent).
1824 * qemu: detect guest side errors during device removal
1826 Libvirt is now able to detect guest side errors during device removal by
1827 using the DEVICE_UNPLUG_GUEST_ERROR event, available in QEMU 6.2.0.
1829 * Minimum SSF setting
1831 The libvirtd.conf option tcp_min_ssf can be used to override the minimum
1832 permitted security strength factor for non-TLS remote connections.
1833 The current hardcoded minimum is 56 (single-DES) and will be raised to 112
1834 in the future. Setting a minimum lower than 112 is not supported.
1836 * qemu: Report stats also for block copy destination and backup job scratch
1839 The statistics are available via the bulk domain stats API.
1843 * qemu: Don't format 'ramfb' attribute when disabled
1845 Fix a regression caused by the conversion to JSON -device arguments where
1846 'ramfb' would be put on the commandline of 'vfio-pci' which doesn't have it,
1847 thus breaking VMs with a mediated host device.
1849 * qemu: Fix block copy and backup to encrypted storage
1851 An oversight in last release lead to a spurious error being reported when
1852 encrypted storage was requested for disk images which are not directly
1853 backing the disk, which is now fixed.
1861 * Introduce virtio-mem ``<memory/>`` model
1863 New virtio-mem model is introduced for ``<memory/>`` device which is a
1864 paravirtualized mechanism of adding/removing memory to/from a VM. Use
1865 ``virDomainUpdateDeviceFlags()`` API to adjust amount of memory or ``virsh
1866 update-memory-device`` for convenience.
1868 * qemu: support disabling hotplug of devices on the pci-root controller
1870 the <target hotplug='on|off'/> option is now supported for the
1871 pci-root controller on i440fx-based (x86 "pc") machinetypes. This
1872 can be used to disable hotplug/unplug of devices from this
1873 controller. The default behavior is unchanged (hotplug is
1876 * Support hotplug and hotunplug for virtiofs
1878 Filesystems backed by virtiofsd can now be hotplugged and hotunplugged.
1880 * virpcivpd: Add a PCI VPD parser
1882 A parser for the standard PCI/PCIe VPD ("I.3. VPD Definitions" in PCI 2.2+
1883 and an equivalent definition in "6.28.1 VPD Format" PCIe 4.0) was added
1884 along with relevant types to represent PCI VPD in memory. This
1885 functionality got added for Linux only at this point (kernels above
1886 v2.6.26 have support for exposing VPD via sysfs).
1888 * virpci: Add PCI VPD-related helper functions to virpci
1890 In order to utilize the PCI VPD parser, a couple of helper functions got
1891 introduced to check for the presence of a VPD file in the sysfs tree and
1892 to invoke the PCI VPD parser to get a list of resources representing PCI
1893 VPD contents in memory.
1895 * nodedev: Add PCI VPD capability support
1897 Support for serializing and deserializing PCI VPD data structures is added
1898 following the addition of the PCI VPD parser. A new PCI device capability
1899 called "vpd" is introduced holding string resources and keyword resources
1902 * qemu: Support page_per_vq for driver element
1904 This optional virtio attribute ``page_per_vq`` controls the layout of the
1905 notification capabilities exposed to the guest. It is recommended for the
1908 * qemu: Support librbd encryption
1910 Add an encryption engine ``librbd``. It will provides the image-level
1911 encryption of librbd. It requires QEMU >= 6.1.0 and librbd >= 16.1.0.
1915 * Use of JSON syntax with ``-device`` with upcoming QEMU-6.2
1917 Libvirt started using JSON directly with the ``-device`` commandline
1918 parameter as it's considered the preferred stable syntax for further QEMU
1919 releases. If any problems with the conversion are encountered please
1920 report them as soon as possible.
1924 * qemu: Fix problems on ``virsh domstats`` with qemu <5.2.0
1926 Libvirt v7.2.0 and later called query-dirty-rate, which was introduced in
1927 qemu-5.2.0, regardless of qemu version and failed in qemu-5.1.0. This
1928 release fixes the bug.
1930 * Don't enter endless loop when unable to accept new clients
1932 If libvirtd (or any other daemon) hit the ulimit for maximum number of open
1933 files but there are still client connections pending then libvirtd (or
1934 corresponding split daemon) would enter an endless loop from which it would
1935 never recover. This behaviour is now fixed.
1937 * qemu: Run secondary driver hooks in split daemon mode
1939 Because of a bug in implementation it may happen that hooks from secondary
1940 drivers were not called in all cases, for instance a network hook wasn't
1941 called upon removal of interface after domain shut off itself. With this
1942 release the bug is fixed.
1950 * nodedev: Add ability to automatically start mediated devices
1952 The autostart status of a persistent mediated devices can be managed with
1953 the new APIs ``virNodeDeviceSetAutostart()`` and
1954 ``virNodeDeviceGetAutostart()``. The corresponding virsh command is
1955 ``nodedev-autostart``. In addition, two new APIs were added to get
1956 additional information about node devices: ``virNodeDeviceIsPersistent()``
1957 checks whether the device is persistently defined, and
1958 ``virNodeDeviceIsActive()`` checks whether the node device is currently
1959 active. This information can also be retrieved with the new virsh command
1962 * qemu: Add attribute ``queue_size`` for virtio-blk devices
1966 * api: Add XML validation for creating of: networkport, nwfilter-binding,
1969 * Add flag ``VIR_NETWORK_PORT_CREATE_VALIDATE`` to validate network port
1970 input xml of network-port creating.
1971 * Add flag ``VIR_NETWORK_CREATE_VALIDATE`` to validate network input xml of
1973 * Add flag ``VIR_NWFILTER_BINDING_CREATE_VALIDATE`` to validate
1974 nwfilter-binding input xml of nwfilter-binding creating.
1982 * Add support for Fibre Channel VMID
1984 New VM element ``<fibrechannel appid=''/>`` was added to allow users to set
1985 their ``appid`` for each VM which will be used by kernel to create Fibre
1986 Channel VMID. This allows various QoS levels, access control or collecting
1987 telemetry data per VM.
1991 * virsh: Allow XML validation for define of: storage pool, network, secret,
1994 * Add flag ``VIR_STORAGE_POOL_DEFINE_VALIDATE`` to validate storage pool
1995 input xml. For virsh, users can use it as ``virsh pool-define --validate``.
1996 * Add flag ``VIR_NETWORK_DEFINE_VALIDATE`` to validate network input xml. For
1997 virsh, users can use it as ``net-define --validate``.
1998 * Add flag ``VIR_SECRET_DEFINE_VALIDATE`` to validate secret input xml. For
1999 virsh, users can use it as ``secret-define --validate``.
2000 * Add flag ``VIR_NWFILTER_DEFINE_VALIDATE`` to validate nwfilter input xml.
2001 For virsh, users can use it as ``nwfilter-define --validate``.
2002 * Add flag ``VIR_INTERFACE_DEFINE_VALIDATE`` to validate interface input xml.
2003 For virsh, users can use it as ``iface-define --validate``.
2005 * Add SecurityManager APIs for labeling network devices
2007 New ``virSecurityManagerSetNetdevLabel`` and ``virSecurityManagerSetNetdevLabel``
2008 APIs are introduced and implemented in the Apparmor security driver.
2009 The qemu driver uses the APIs to label vhostuser ports on hotplug and
2010 restore labeling on unplug.
2012 * vmx: Parse vm.genid and support super wide SCSI bus
2014 The genid attribute is now reported for VMX guests. Libvirt can now
2015 properly process super wide SCSI bus (64 units).
2017 * qemu: Lifecycle action (``on_poweroff``/``on_reboot``) handling improvements
2019 The handling of lifecycle actions was fixed and improved in multiple ways:
2021 - ``restart-rename`` action was forbidden
2023 The action was never properly implemented in the qemu driver and didn't
2024 actually result in a restart of the VM but rather termination. The qemu
2025 driver now rejects such configurations.
2027 - ``preserve`` action was forbidden
2029 Similarly to the previous case this never worked as the intended semantics
2030 of the actions dictate. It's better to not allow it at all until there's a
2031 proper implementation
2033 - ``reboot`` action of ``on_poweroff`` now actually works
2035 The guest OS is now rebooted instead of terminating the VM when the
2036 ``reboot`` action is used and the guest OS powers down. Note that it's
2037 incompatible with ``on_reboot`` set to ``destroy``.
2039 - Changes in action action of ``on_reboot`` are now updated with qemu
2041 Libvirtd can now properly update the ``on_reboot`` action in qemu which
2042 allows proper handling when changing between ``reboot`` and ``destroy``
2043 actions. In addition, switching from ``reboot`` to ``destroy`` was
2044 forbidden for older qemus which don't support the update API as the guest
2045 could still reboot and execute some instructions until it was terminated.
2047 * virsh: Support vhostuser in attach-interface
2051 * qemu: Open chardev logfile on behalf of QEMU
2053 Guests with a logfile configured for their chardevs are now able to start
2054 even when no virtlogd is configured.
2056 * virhostmem: Handle numactl-less build in hugepages allocation/reporting
2058 Some architectures don't have notion of NUMA (e.g. s390x) but do support
2059 hugepages. Libvirt silently ignored requests to allocate/report hugepage
2060 pool when built without numactl. This is now fixed and the pool can be
2061 allocated/reported on properly.
2063 * qemu: Record proper ``backing`` format for overlays of qcow2+luks images
2065 Libvirt would record ``luks`` instead of ``qcow2`` into the metadata. In
2066 practice this is a problem only when inspecting images manually via
2067 ``qemu-img`` as with libvirt users must use full specification of the
2068 backing chain in the domain XML which supersedes information recorded in
2077 * storage: Unlock pool objects on ACL check failures in ``storagePoolLookupByTargetPath`` (CVE-2021-3667)
2079 A logic bug in ``storagePoolLookupByTargetPath`` where the storage pool
2080 object was left locked after a failure of the ACL check could potentially
2081 deprive legitimate users access to a storage pool object by users who don't
2086 * qemu: Incremental backup support via ``virDomainBackupBegin``
2088 libvirt-7.6 along with the unreleased qemu-6.1 will fully support the change
2089 block tracking features (block-dirty-bitmaps) to be able to do incremental
2090 backups and management of the checkpoint states via the appropriate APIs.
2092 * qemu: Add support for launch security type s390-pv
2094 Specifying s390-pv as launch security type in an s390 domain prepares for
2095 running the guest in protected virtualization secure mode, also known as
2096 IBM Secure Execution. This simplifies the definition and reduces the risk
2097 of an incorrect definition, e.g. by forgetting to specify ``iommu=on`` on
2100 * domstats: Add haltpolling time statistic interface
2102 Domstats now provide the data of cpu haltpolling time. This feature relies
2103 on statistics available after kernel version 5.8. This will allow the user
2104 to get more accurate CPU usage information if needed.
2108 * qemu: Fix migration with ``VIR_MIGRATE_NON_SHARED_INC``
2110 libvirt 7.3.0 introduced a bug where ``VIR_MIGRATE_NON_SHARED_INC`` would
2111 not actually migrate the contents of the disk due to broken logic and at
2112 the same time could trigger migration of storage when
2113 ``VIR_MIGRATE_TUNNELLED`` is requested. This release fixes the bug.
2115 * qemu: Don't emit ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` twice when registered with index
2117 When registering the threshold event with the index notation (e.g.
2118 ``vda[3]``) libvirt would emit the event also for ``vda`` if the image is
2119 in the top layer. The intention was to emit two events only when the
2120 original registration was done without the index.
2122 * qemu: Pass discard requests for disks with ``copy_on_read='on'``
2124 When a disk using the ``copy_on_read='on'`` option is configured also with
2125 ``discard='unmap'`` the discard requests will now be passed to the
2126 underlying image freeing up the space.
2134 * svirt: fix MCS label generation (CVE-2021-3631)
2136 A flaw in the way MCS labels were generated could result in a VM's
2137 resource not being fully protected from access by another VM were
2138 it to be compromised. https://gitlab.com/libvirt/libvirt/-/issues/153
2140 * **Removed features**
2142 * xen: Remove support for Xen < 4.9
2144 In accordance with our platform support policy, the oldest supported Xen
2145 version is now bumped from 4.6 to 4.9.
2149 * docs: Document disk serial truncation status quo
2151 Disk ``<serial>`` is being truncated by QEMU before passed to the guest.
2152 Since it's impossible to fix it without running into further regressions
2153 the documentation was improved to document the intricacies.
2157 * qemu: Fixed validation of disk ``iothread`` configuration
2159 The validation of ``iothread`` config was previously moved to a place where
2160 it caused bogus errors when address wasn't allocated when hotplugging a
2161 disk. The check is now removed as it wasn't actually necessary at all.
2167 * **Removed features**
2169 * qemu: Remove support for QEMU < 2.11
2171 In accordance with our platform support policy, the oldest supported QEMU
2172 version is now bumped from 1.5 to 2.11.
2176 * qemu: Add support for hotplugging ``<transient/>`` disks
2178 The disk hotplug code in the qemu driver now can handle hotplug of disks
2179 with automatically added overlay.
2181 * qemu: Add support for sharing base image of ``<transient/>`` disks
2183 Users can use ``<transient shareBacking='yes'/>`` to tell the qemu driver to
2184 never open the base image in write mode thus multiple VMs can share the
2185 same image. Note that the disk will be hotplugged during startup.
2189 * Add win-dmp crashdump format
2191 New ``win-dmp`` format for ``virDomainCoreDumpWithFormat`` API and/or virsh
2192 ``dump --format`` was introduced.
2196 * Allow 0 offset in XML schema for ``<slice type='storage' offset='0' size='321'/>``
2198 Having a 0 offset so that the size of the image can be limited is a
2199 valid configuration so it was allowed in the XML schema.
2207 * xen: Support domains with more than 4TB
2209 The xen driver now supports domains with more than 4TB of memory with
2212 * qemu: add socket for virtiofs filesystems
2214 Libvirt now supports ``filesystem`` devices that connect to
2215 a ``virtiofsd`` daemon launched outside of libvirtd, via the
2216 ``socket`` attribute of the ``source`` element.
2218 * nodedev: Add ability to manage persistent mediated devices
2220 Persistent mediated devices can now be managed with libvirt.
2221 ``virNodeDeviceDefineXML()`` defines a new device,
2222 ``virNodeDeviceUndefine()`` removes an existing definition, and
2223 ``virNodeDeviceCreate()`` starts a device definition that is currently
2224 inactive. Corresponding virsh commands ``nodedev-define``,
2225 ``nodedev-undefine``, and ``nodedev-start`` were also added.
2226 ``nodedev-list`` only lists active devices by default. Inactive device
2227 definitions can be shown with the new ``--inactive`` and ``--all`` flags.
2229 * qemu: Allow use of qemu's ``-compat`` option
2231 Curious developers or testers now can enable certain ``-compat`` modes which
2232 allow to notice use of deprecated commands and options as qemu will use the
2233 selected method to notify the user. The new behaviour can be requested using
2234 either the ``deprecation_behavior`` option in ``qemu.conf`` for all VMs or
2235 using ``<qemu:deprecation behavior='...'/>`` in the VM XML.
2239 * virsh: Improve errors with ``virsh snapshot-create-as``
2241 The XML document constructed by virsh was forced through XML schema
2242 validation which yielded unintelligible error messages in cases such as
2243 when the path to the new image did not start with a slash. XML documents
2244 are no longer validated as the XML parser actually has better error
2245 messages which allow users to figure the problem out quickly.
2247 * qemu: Terminate backing store when doing a full-chain block pull
2249 When pulling everything into the overlay image the chain can be terminated
2250 since we know that it won't depend on any backing image and thus can prevent
2251 attempts to probe the backing chain.
2253 * qemu: Expose disk serial in virDomainGetGuestInfo()
2255 The ``virDomainGetGuestInfo()`` reports disk serial number among with other
2260 * qemu: Fix crash of libvirt on full block pull of a disk
2262 When the persistent definition contains a compatible disk (meaning the
2263 definition of the running and persistent config match) a block pull job
2264 would leave a dangling pointer in the config definition which resulted
2267 * qemu: Use proper job cancelling command
2269 Libvirt's API contract for aborting a block copy job in 'ready' state
2270 declares that the destination image of the copy will contain a consistent
2271 image of the disk from the time when the block job was aborted. This
2272 requires that libvirt uses the proper cancelling qemu command to ensure
2273 that the data is consistent which was not the case.
2275 * qemu: Don't attempt storage migration when there are no migratable disks
2277 Due to a logic bug introduced in the previous release libvirt would attempt
2278 to migrate disks in case when no disks are selected/eligible for migration.
2280 * qemu: Fix very rare race when two block job 'ready' events are delivered
2282 In certain high-load scenarios, qemu might deliver the 'ready' event twice
2283 and if it's delivered when pivoting to the destination during a block copy
2284 job, libvirt would get confused and execute the code as if the job were
2287 * lxc: Fix container destroy with CGroupsV2
2289 When an LXC container was started and the host used CGroupsV2 it might have
2290 had created nested controllers under the container's scope. Libvirt was
2291 unaware and thus destroying the container failed with a cryptic error:
2292 ``failed to get cgroup backend for 'pathOfController'``. The CGroup removal
2293 code was reworked and is now capable of dealing with such scenario.
2295 * bash-completion: Fix argument passing to $1
2297 Due to a bug in bash completion script, the auto completion did not work
2298 properly when a connection URI or read only flag were specified on
2299 ``virsh`` or ``virt-admin`` command line.
2307 * qemu: Implement domain memory dirty rate calculation API
2309 New API ``virDomainStartDirtyRateCalc()`` and virsh command
2310 ``domdirtyrate-calc`` are added to start calculating a live domain's
2313 * qemu: Support reporting memory dirty rate stats
2315 The memory dirty rate stats can be obtained through ``virsh domstats
2316 --dirtyrate`` via the virConnectGetAllDomainStats API.
2318 * qemu: Full disk backups via ``virDomainBackupBegin``
2320 The qemu hypervisor driver now allows taking full disk backups via the
2321 ``virDomainBackupBegin`` API and the corresponding virsh wrapper.
2323 In future releases the feature will be extended to also support incremental
2324 backups (where only the difference since the last backup is copied) when
2325 qemu adds the required functionality.
2327 * Add support for audio backend specific settings
2329 With this release a new ``<audio/>`` element is introduced that allows
2330 users to configure audio output for their guests.
2334 * qemu: Compatibility with QEMU 6.0 for certain hot-(un)-plug operations
2336 Libvirt 7.2.0 is required for compatibility with the upcoming QEMU 6.0
2337 release for hotplug and hotunplug of certain devices and helpers, such as
2338 iothreads, chardevs, RNG devices, disks with secret, ...
2340 * qemu: Various improvements to embedded mode
2342 Embedded mode for the QEMU driver, as well as the ``virt-qemu-run`` tool
2343 saw improvements in handling of domain life cycle, temporary directories
2344 creation (important when using disk secrets) and other minor fixes.
2346 * Documentation of split daemon related config files
2348 Split daemons read configuration files upon their start. These were never
2353 * Check host CPU for forbidden features
2355 CPU feature policy did not work as expected with ``host-passthrough`` and
2356 features supported by physical host. CPU features were not filtered out
2357 when ``@check`` was set to ``full``.
2359 * Fix virNetworkUpdate() to work with split daemons
2361 Due to a bug in our code, virNetworkUpdate() did not work with split daemon
2362 unless management application connected to virtnetworkd directly.
2364 * qemu: increase locked memory limit when a vDPA device is present
2366 Just like VFIO devices, vDPA devices may need to have all guest memory
2367 pages locked/pinned in order to operate properly. These devices are now
2368 included when calculating the limit for memory lock.
2370 * Don't log error if SRIOV PF has no associated netdev
2372 Some SRIOV PFs don't have a netdev associated with them in which case
2373 libvirtd reported an error and refused to start. This is now fixed.
2375 * qemu: Only raise memlock limit if necessary
2377 Attempting to set the memlock limit might fail if we're running
2378 in a containerized environment where ``CAP_SYS_RESOURCE`` is not
2379 available, and if the limit is already high enough there's no
2380 point in trying to raise it anyway.
2382 * Restore security context of swtpm.log
2384 If a guest with emulated TPM was started and the daemon was restarted
2385 afterwards, the security context of the per-domain ``swtpm.log`` file was
2386 not restored on domain shutdown leaving it unable to be started again.
2388 * virtlogd|virtlockd: Fixed crash when upgrading the daemons in-place
2390 A bug preventing the in-place upgrade of ``virtlogd`` and ``virtlockd``
2391 daemons was fixed, so they can again be upgraded without dropping the log
2392 file descriptors or locks on files.
2400 * Implement Apple Silicon support
2402 libvirt now runs on the ARM-based Apple Silicon Macs.
2406 * Introduce virtio-pmem ``<memory/>`` model
2408 The virtio-pmem is a virtio variant of NVDIMM and just like NVDIMM
2409 virtio-pmem also allows accessing host pages bypassing guest page cache.
2411 * Introduce ``<boot order/>`` for ``<filesystem>``
2413 Booting is possible from virtiofs filesystems. Introduce an option
2414 to control the boot order, like we do for other bootable devices.
2416 * hyperv: implement new APIs
2418 The ``virDomainUndefine()``, ``virDomainUndefineFlags()``,
2419 ``virDomainDefineXML()``, ``virDomainAttachDevice()``, and
2420 ``virDomainAttachDeviceFlags()``, ``virConnectListAllNetworks()``,
2421 ``virConnectNumOfNetworks()``, ``virNetworkLookupByName()``,
2422 ``virNetworkLookupByUUID()``, ``virConnectNumOfDefinedNetworks()``,
2423 ``virConnectListDefinedNetworks()``, ``virNetworkGetAutostart()``,
2424 ``virNetworkIsActive()``, ``virNetworkIsPersistent()``,
2425 ``virNetworkGetXMLDesc()``, and ``virDomainScreenshot()``, APIs have been
2426 implemented in the Hyper-V driver.
2428 * Support <teaming> element in plain <hostdev> devices
2430 This is useful when libvirt doesn't have the privileges necessary
2431 to set the hostdev device's MAC address (which is a necessary
2432 part of the alternate <interface type='hostdev'>).
2434 * Introduce ``<disk type='vhostuser'>`` support
2436 Introduces support for QEMU vhost-user-blk device that can be used
2437 to access storage exported via the vhost-user protocol by daemons such
2438 as the ``qemu-storage-daemon``.
2442 * qemu: Fix disk quiescing rollback when creating external snapshots
2444 If the qemu guest agent call to freeze filesystems failed when creating
2445 an external snapshot with ``VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE`` flag the
2446 filesystems would be unconditionally thawed. This could cause problems when
2447 the filesystems were frozen by an explicit call to ``virDomainFSFreeze``
2448 since the guest agent then rejects any further freeze attempts once are
2449 filesystems frozen, an explicit freeze followed by a quiesced snapshot
2450 would fail and thaw filesystems.
2452 Users are also encouraged to use ``virDomainFSFreeze/Thaw`` manually instead
2453 of relying on ``VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE`` if they need finer
2456 * cgroups: Fix how we setup and configure cgroups on hosts with systemd
2458 When libvirt is running on host with systemd we register every VM with
2459 machined which creates the VM root cgroup for us as well. Before this fix
2460 we were directly modifying files in the VM root cgroup which was incorrect
2461 because all the files are managed by systemd. The implication was that any
2462 change done by libvirt to cgroup attributes supported by systemd could be
2463 removed which happens for example by running ``systemctl daemon-reload``.
2465 To fix the issue libvirt now uses DBus calls for some of the cgroup
2466 attributes that distribute the resources proportionally to the cgroup
2467 siblings and for the rest we have a new sub-cgroup that libvirt can
2470 For more details why this is necessary see
2471 `systemd cgroup <https://systemd.io/CGROUP_DELEGATION/>`_ documentation.
2473 * qemu: Fix swtpm device with aarch64
2475 The TPM TIS device name for x86 is ``tpm-tis``, whereas for aarch64 it is
2476 ``tpm-tis-device``. Fix the use of TPM TIS device with aarch64 by using
2477 the proper device name when building the QEMU command line.
2479 * libxl: Fix domain shutdown
2481 Commit fa30ee04a2 introduced the possibility of a race between the
2482 shutdown and death threads used to process domain shutdown and death
2483 events from libxl. On normal domain shutdown the shutdown thread handles
2484 all aspects of shutting down and cleaning up the domain. The death
2485 thread is only used to handle out-of-band domain destruction and is
2486 inhibited when domain shutdown is under libvirt's control. The race is
2487 avoided by also inhibiting the death thread when libvirt starts the
2494 * **Project governance**
2496 * Formal handover of release tarball signing
2498 Starting from libvirt-6.6 the release tarballs are signed by Jiřà Denemark.
2499 Releases starting with 7.0 contain a note from the previous maintainer
2500 Daniel Veillard officially handing over the signing of packages so that the
2501 transition can be verified.
2505 * nodedev: Add node device driver support for AP devices
2507 Add support for detecting and listing Adjunct Processor(AP) cards, AP
2508 queues and AP matrix devices (which are capable of MDEV) of a KVM host
2509 system in libvirt node device driver with correct object relationships.
2511 * qemu: Allow control of ``qcow2`` metadata cache
2513 In specific usecases such as when massive storage images are used it's
2514 possible to achieve better performance by increasing the metadata cache
2515 size. The new knob allows advanced users setting the size according to
2516 qemu's documentation to suit their image.
2518 * conf: Add support for keeping TPM emulator state
2520 Currently, swtpm TPM state file is removed when a transient domain is
2521 powered off or undefined. Add per-TPM emulator option ``persistent_state``
2522 for keeping TPM state.
2524 * cpu_map: Add Snowridge CPU model
2526 It's supported in QEMU 4.1 and newer.
2528 * qemu: Add support for NFS disk protocol
2530 Implement support for the 'nfs' native protocol driver in the qemu driver.
2534 * qemu: Discourage users from polling ``virDomainGetBlockJobInfo`` for block
2537 Document that waiting for events is a more robust solution.
2539 * secret: Relax XML schema for the ``usage`` name of a ``secret``
2541 Various bits of documentation of how to use libvirt with RBD volumes used
2542 an usage name which would not pass the XML validation. Relax the requirement
2543 to make such XMLs valid.
2545 * virnetdevopenvswitch: Various improvements
2547 The code that handles ``<interface type='vhostuser'/>`` was given various
2548 improvements. So far, libvirt assumed vhostuser interfaces are handled
2549 exclusively by OpenVSwitch and refused to start a guest if it was not so.
2550 Now a guest can be started successfully even if the interface is created by
2551 some other tool (e.g. ``dpdk-testpmd``). Also, the code that detects the
2552 interface name was adapted to new versions of OpenVSwitch and thus can
2553 detect name more reliably.
2555 * qemu: Report guest disks information in ``virDomainGetGuestInfo``
2557 Libvirt is now able to report disks and filesystems from the guest's
2558 perspective (using guest agent). And with sufficiently new guest agent
2559 (5.3.0 or newer) the API also handles disks on CCW bus.
2563 * qemu: Fix logic bug in inactive snapshot deletion
2565 This release fixes a bug introduced in libvirt-6.9 where libvirt's
2566 snapshot metadata would not be deleted on successful snapshot deletion.
2568 * qemu: Fix VMs with ``<iotune>`` on an empty cdrom
2570 Specifying ``<iotune>`` for an empty cdrom would prevent the VM from
2571 starting as qemu doesn't accept the tuning for an empty drive. We now
2572 postpone setting the parameters until a new media is inserted.
2574 * Avoid taking extra host memory when launching pSeries guests
2576 Under certain conditions, pSeries guests were being launched with more
2577 RAM than it was specified in the domain XML by the user. New pSeries
2578 domains created with libvirt 7.0.0 will always launch with the right
2579 amount of initial memory. Existing guests that migrate from an older
2580 libvirt version to 7.0.0 will not be affected by this change.
2582 * qemu: Don't cache NUMA caps
2584 ``virsh capabilities`` contains ``<topology/>`` section which reports NUMA
2585 topology among with amount of free hugepages per each NUMA node. However,
2586 these amounts were not updated between calls.
2588 * networkGetDHCPLeases: Handle leases with infinite expiry time
2590 Since libvirt-6.3.0 it is possible to configure expiry time for DHCP
2591 leases. If the expiry time was infinite then ``virsh net-dhcp-leases``
2592 and NSS plugins refused to work.
2594 * qemu: Don't prealloc mem for real NVDIMMs
2596 If a real life NVDIMM is assigned to a guest via ``<memory model='nvdimm'/>``
2597 then QEMU is no longer instructed to preallocate memory
2598 for it. This prevents unnecessary wear on the NVDIMM.
2600 * network: Introduce mutex for bridge name generation
2602 When new libvirt network is defined or created and the input XML does not
2603 contain any bridge name, libvirt generates one. However, it might have
2604 happened that the same name would be generated for different networks if
2605 two or more networks were defined/created at once.
2608 v6.10.0 (2020-12-01)
2609 ====================
2613 * qemu: Enable client TLS certificate validation by default for ``chardev``,
2614 ``migration``, and ``backup`` servers.
2616 The default value if qemu.conf options ``chardev_tls_x509_verify``,
2617 ``migrate_tls_x509_verify``, or ``backup_tls_x509_verify`` are not
2618 specified explicitly in the config file and also the
2619 ``default_tls_x509_verify`` config option is missing are now '1'. This
2620 ensures that only legitimate clients access servers, which don't have any
2621 additional form of authentication.
2623 * qemu: Introduce "migrate_tls_force" qemu.conf option
2625 The ``migrate_tls_force`` configuration option allows administrators to
2626 always force connections used for migration to be TLS secured as if the
2627 ``VIR_MIGRATE_TLS`` flag had been used.
2631 * qemu: Implement OpenSSH authorized key file management APIs
2633 New APIs (``virDomainAuthorizedSSHKeysGet()`` and
2634 ``virDomainAuthorizedSSHKeysSet()``) and virsh commands
2635 (``get-user-sshkeys`` and ``set-user-sshkeys``) are added to manage
2636 authorized_keys SSH file for user.
2638 * hyperv: implement new APIs
2640 The ``virDomainGetMaxMemory()``, ``virDomainSetMaxMemory()``,
2641 ``virDomainGetSchedulerType()``, ``virDomainGetSchedulerParameters()``,
2642 ``virDomainGetSchedulerParametersFlags()``, ``virDomainGetVcpus()``,
2643 ``virDomainGetVcpusFlags()``, ``virDomainGetMaxVcpus()``,
2644 ``virDomainSetVcpus()``, and ``virDomainSetVcpusFlags()`` APIs have been
2645 implemented in the Hyper-V driver.
2647 * qemu: Add 'fmode' and 'dmode' options for 9pfs
2649 Expose QEMU's 9pfs 'fmode' and 'dmode' options via attributes on the
2650 'filesystem' node in the domain XML. These options control the creation
2651 mode of files and directories, respectively, when using accessmode=mapped.
2652 It requires QEMU 2.10 or above.
2654 * qemu: support kvm-poll-control performance hint
2656 Implement the new KVM feature 'poll-control' to set this performance hint
2657 for KVM guests. It requires QEMU 4.2 or above.
2661 * virsh: Support network disks in ``virsh attach-disk``
2663 The ``virsh attach-disk`` helper command which simplifies attaching of disks
2664 without the need for the user to formulate the disk XML manually now
2665 supports network-backed images. Users can specify the protocol and host
2666 specification with new command line arguments. Please refer to the man
2667 page of virsh for further information.
2671 * remote: fixed performance regression in SSH tunnelling
2673 The ``virt-ssh-helper`` binary introduced in 6.8.0 had very
2674 poor scalability which impacted libvirt tunnelled migration
2675 and storage volume upload/download in particular. It has been
2676 updated and now has performance on par with netcat.
2678 * **Removed features**
2680 * hyperv: removed support for the Hyper-V V1 WMI API
2682 This drops support for Windows Server 2008R2 and 2012.
2683 The earliest supported version is now Windows 2012R2.
2691 * nodedev: Add support for channel subsystem (CSS) devices on S390
2693 A CSS device is represented as a parent device of a CCW device.
2694 This support allows to create vfio-ccw mediated devices with
2695 ``virNodeDeviceCreateXML()``.
2697 * qemu: Implement memory failure event
2699 New event is implemented that is emitted whenever a guest encounters a
2702 * qemu: Implement support for ``<transient/>`` disks
2704 VMs based on the QEMU hypervisor now can use ``<transient/>`` option for
2705 local file-backed disks to configure a disk which discards changes made to
2706 it while the VM was active.
2708 * hyperv: implement new APIs
2710 The ``virConnectGetCapabilities()``, ``virConnectGetMaxVcpus()``,
2711 ``virConnectGetVersion()``, ``virDomainGetAutostart()``,
2712 ``virDomainSetAutostart()``, ``virNodeGetFreeMemory()``,
2713 ``virDomainReboot()``, ``virDomainReset()``, ``virDomainShutdown()``, and
2714 ``virDomainShutdownFlags()`` APIs have been implemented in the Hyper-V
2717 * bhyve: implement virtio-9p filesystem support
2719 Implement virito-9p shared filesystem using the ``<filesystem/>`` element.
2721 * qemu: Add support for vDPA network devices.
2723 VMs using the QEMU hypervisor can now specify vDPA network devices
2724 using ``<interface type='vdpa'>``. The node device APIs also now
2725 list and provide XML descriptions for vDPA devices.
2727 * cpu_map: Add EPYC-Rome CPU model
2729 It's supported in QEMU 5.0.0 and newer.
2731 * cpu: Add a flag for XML validation in CPU comparison
2733 The ``virConnectCompareCPU`` and ``virConnectCompareHypervisorCPU`` API
2734 now support the ``VIR_CONNECT_COMPARE_CPU_VALIDATE_XML`` flag, which
2735 enables XML validation. For virsh, this feature is enabled by passing
2736 the ``--validate`` option to the ``cpu-compare`` and
2737 ``hypervisor-cpu-compare`` subcommands.
2739 * qemu: Introduce virtio-balloon free page reporting feature
2741 Introduce the optional attribute ``free-page-reporting`` for virtio
2742 memballoon device. It enables/disables the ability of the QEMU virtio
2743 memory balloon to return unused pages back to the hypervisor. QEMU 5.1
2744 and newer support this feature.
2748 * qemu: Make 'cbitpos' & 'reducedPhysBits' attrs optional
2750 Libvirt probes the underlying platform in order to fill in these SEV
2751 attributes automatically before launching a guest.
2753 * util: support device stats collection for SR-IOV VF hostdev
2755 For SR-IOV VF hostdevs, libvirt now supports retrieving device traffic
2756 stats via the ``virDomainInterfaceStats`` API and ``virsh domifstat``.
2758 * logging: Allow disabling log rollover
2760 Set ``max_len=0`` in ``virtlogd.conf`` to disable log rollover.
2762 * qemu: Set noqueue qdisc for TAP devices
2764 Set ``noqueue`` instead of the former ``pfifo_fast`` queue discipline
2765 for TAP devices. It will avoid needless cost of host CPU cycles and
2766 thus improve performance.
2768 * qemu: virtiofs can be used without NUMA nodes
2770 Virtiofs is supported for the VM without NUMA nodes but configured with
2775 * hyperv: ensure WQL queries work in all locales
2777 Relying on the "Description" field caused queries to fail on non-"en-US"
2778 systems. The queries have been updated to avoid using localized strings.
2780 * rpc: Fix ``virt-ssh-helper`` detection
2782 libvirt 6.8.0 failed to correctly detect the availability of the new
2783 ``virt-ssh-helper`` command on the remote host, and thus always used the
2784 fallback instead; this has now been fixed.
2792 * qemu: double free in qemuAgentGetInterfaces() in qemu_agent.c
2794 Clients connecting to the read-write socket with limited ACL permissions
2795 may be able to crash the libvirt daemon, resulting in a denial of service,
2796 or potentially escalate their privileges on the system. CVE-2020-25637.
2800 * xen: Add ``writeFiltering`` attribute for PCI devices
2802 By default Xen filters guest writes to the PCI configuration space of a
2803 PCI hostdev, which may cause problems for some devices. The ``writeFiltering``
2804 attribute of the device's ``<source>`` element can be used to disable the
2805 filtering and allow all guest writes to the configuration space.
2807 * bhyve: Support setting the framebuffer resolution
2809 Libvirt can now set the framebuffer's "w" and "h" parameters
2810 using the ``resolution`` element.
2812 * bhyve: Support VNC password authentication
2814 Libvirt can now probe whether the bhyve binary supports
2815 VNC password authentication. In case it does, a VNC password
2816 can now be passed using the ``passwd`` attribute on
2817 the ``<graphics>`` element.
2819 * remote: ``virt-ssh-helper`` replaces ``nc`` for SSH tunnelling
2821 Libvirt now provides a ``virt-ssh-helper`` binary on the server
2822 side. The libvirt remote client will use this binary for setting
2823 up an SSH tunnelled connection to hosts. If not present, it will
2824 transparently fallback to the traditional ``nc`` tunnel. The new
2825 binary makes it possible for libvirt to transparently connect
2826 across hosts even if libvirt is built with a different installation
2827 prefix on the client vs server. It also enables remote access to
2828 the unprivileged per-user libvirt daemons (e.g. using a URI such as
2829 ``qemu+ssh://hostname/session``). The only requirement is that
2830 ``virt-ssh-helper`` is present in ``$PATH`` of the remote host.
2832 * esx: implement few APIs
2834 The ``virConnectListAllNetworks()``, ``virDomainGetHostname()``, and
2835 ``virDomainInterfaceAddresses()`` (only for
2836 ``VIR_DOMAIN_INTERFACE_ADDRESSES_SRC_AGENT`` source) APIs were implemented
2841 * qemu: Allow migration over UNIX sockets
2843 QEMU migration can now be performed completely over UNIX sockets. This is
2844 useful for containerised scenarios and can be used in both peer2peer and
2847 * dbus: Use GLib implementation instead of libdbus
2849 Adopting GLib DBus implementation simplifies our code as libdbus provides
2850 low-level APIs where we had to have a lot of helper functions. With this
2851 change we also remove dependency on libdbus and possibly fix all the DBus
2852 related libvirtd crashes seen over the time.
2854 * Re-introduce NVDIMM auto-alignment for pSeries Guests
2856 The auto-alignment logic was removed in v6.7.0 in favor of requiring the
2857 size provided by the user to be already aligned; however, this had the
2858 unintended consequence of breaking some existing guests. v6.8.0 restores
2859 the previous behavior with an improvement: it also reflects the auto-aligned
2860 value in the domain XML.
2862 * qemu: Preserve qcow2 cluster size after external snapshots
2864 The new overlay image which is installed on top of the current chain when
2865 taking an external snapshot now preserves the cluster size of the original
2866 top image to preserve any performance tuning done on the original image.
2870 * qemu: Various (i)SCSI backed hostdev fixes
2872 (i)SCSI backed hostdevs now work again with an arbitrarily long
2873 user-specified device alias and also honor the 'readonly' property after a
2876 * **Removed features**
2878 * node_device: Remove HAL node device backend
2880 HAL is deprecated on all supported OS so there is no need to keep it
2881 in libvirt. udev backend is used on Linux OSes and devd can be eventually
2882 implemented as replacement for FreeBSD.
2888 * **Packaging changes**
2890 * Libvirt switch to Meson build system
2892 Libvirt abandoned autotools and switched to Meson build system.
2896 * qemu: Add support for initiator IQN configuration for iSCSI hostdevs
2898 Similarly to iSCSI ``<disk>`` users can use an ``<initiator>`` element
2899 inside ``<hostdev>`` with the same format to configure the ``IQN`` value
2900 used by the qemu initiator when connecting to an iSCSI target.
2902 * xen: Add support for device model command-line passthrough
2904 Xen supports passing arbitrary arguments to the QEMU device model using
2905 the ``device_model_args`` setting in xl.cfg(5). The libvirt xen driver now
2906 supports this using ``<xen:commandline/>`` XML extensions.
2908 * shmem: Add support for shmem-{plain, doorbell} ``role`` option
2910 The ``role`` attribute controls how the domain behaves on migration. With
2911 ``role=master``, the guest will copy the shared memory on migration to
2912 the destination host. With ``role=peer``, the migration is disabled.
2914 * bhyve: Sound device support
2916 This feature allows to configure guest sound device using
2917 the ``<sound>`` element, and map it to the host sound device using
2918 the ``<audio>`` element.
2922 * Allow sparse streams for block devices
2924 Sparse streams (e.g. ``virsh vol-download --sparse`` or ``virsh vol-upload
2925 --sparse``) now handle if one of the stream ends is a block device.
2927 * Remove NVDIMM auto-alignment for pSeries Guests
2929 This feature was introduced in libvirt v6.2.0 as part of the overall
2930 NVDIMM support for pSeries guests. The idea was to relieve the user
2931 from knowing ppc64 alignment details, but the end result is that we
2932 ended up with inconsistencies between domain XML and actual NVDIMM
2933 size the guest is using. To promote consistency between domain XML
2934 and the guest, unaligned NVDIMM sizes for pSeries guests will now be
2935 forbidden and no size auto-alignment will be made. Instead, libvirt will
2936 suggest an aligned round up size for the user.
2938 * apparmor: Several improvements
2940 Add support for virtiofs filesystem and allow QEMU to load old
2941 shared objects after upgrade.
2945 * virdevmapper: Deal with kernels without DM support
2947 In the previous release libvirt dropped libdevmapper in favor of its own
2948 implementation. However, it failed to deal correctly with kernels that
2949 either don't have device mapper enabled or where the dm-mod module is not
2950 loaded yet. This is now fixed.
2952 * resctrl: Use exclusive lock for /sys/fs/resctrl
2954 When two or more domains were attempted to start at once, due to a bug in
2955 implementation, resctrl was not locked properly and thus threads did not
2956 mutually exclude with each other resulting in not setting requested
2959 * mdev: Fix daemon crash when reattaching mdevs on assignment conflict
2961 If there's a list of mdevs to be assigned to a domain, but one of them (NOT
2962 the first) is already assigned to a different domain then libvirtd would
2963 crash. This is now fixed.
2965 * Fix logic in setting COW flag on btrfs
2967 When COW is not explicitly requested to be disabled or enabled, then
2968 libvirt should do nothing on non-BTRFS file systems.
2970 * Avoid crash due to race in glib event loop code
2972 Libvirt switched to glib event loop in 6.1.0 but it was also tickling a bug
2973 in glib code leading to the daemon crash. Libvirt way of calling glib was
2974 changed so the daemon crashes no more.
2976 * virdevmapper: Handle kernel without device-mapper support
2978 In the previous release, Libvirt dropped libdevmapper in favor of its own
2979 implementation. But the implementation did not handle kernels without
2980 device-mapper support. This is now fixed.
2982 * remove autogenerated macvtap names from migration XML
2984 Autogenerated macvtap device names were being left in the
2985 migration XML, which could result in libvirt erroneously deleting
2986 the macvtap device of a different guest in the aftermath of
2987 failing to restart the guest on the destination host. Removing the
2988 autogenerated names avoids this.
2996 * Allow configuring of ACPI NUMA HMAT
2998 Libvirt allows configuring ACPI Heterogeneous Memory Attribute Table to
2999 hint software running inside the guest on optimization.
3001 * esx: Add a ``type`` attribute for mac addresses.
3003 This attribute allows (when set to ``static``) ignoring VMWare checks of the
3004 MAC addresses that would generate a new one if they were in its OUI
3007 * conf: add control over COW for storage pool directories
3009 The storage pool code now attempts to disable COW by default on btrfs, but
3010 management applications may wish to override this behaviour. This is now
3011 possible via new ``cow`` element.
3015 * esx: Change the NIC limit for recent virtualHW versions
3017 Specifying a virtualHW version greater or equal to 7 (ESXi 4.0) will allow
3018 you to use up to 10 NICs instead of 4 as it was previously.
3020 * qemu: Support encrypted TLS keys for NBD disks
3022 The secret key used for disks can now be encrypted similarly to TLS keys
3023 used for migration, chardev and others.
3025 * qemu: ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` can now be registered for ``<mirror>``
3027 The event can now be used also for block copy destinations by using the
3028 index of the ``<mirror>`` image.
3030 * qemu: consider available CPUs in ``vcpupin/emulatorpin`` output
3032 This patch changes the default bitmap of ``vcpupin`` and ``emulatorpin``,
3033 in the case of domains with static vcpu placement, all available CPUs
3034 instead of all possible CPUs are returned making these APIs consistent with
3035 the behavior of ``vcpuinfo``.
3039 * virdevmapper: Don't use libdevmapper to obtain dependencies
3041 When building domain's private ``/dev`` in a namespace, libdevmapper was
3042 consulted for getting full dependency tree of domain's disks. However, this
3043 meant that libdevmapper opened ``/dev/mapper/control`` which wasn't closed
3044 and was leaked to QEMU. CVE-2020-14339
3046 * qemu: Report correct ``index`` in ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD``
3048 Starting from libvirt 5.10 with QEMU 4.2 the
3049 ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` event would report incorrect device
3050 ``index`` when reported for an image from the backing chain of a disk.
3052 * qemu: Don't fail active layer block commit or block copy in certain cases
3054 Starting from libvirt-6.5 an active layer block commit or a block copy could
3055 fail if the same destination was used more than once.
3057 * qemu: Don't change ownership of restore file
3059 When restoring a domain from a file, Libvirt no longer changes its ownership.
3061 * qemu: Set SPAPR TPM default to 2.0 and prevent 1.2 choice
3063 The firmware (SLOF) on QEMU for ppc64 does not support TPM 1.2, so prevent
3064 the choice of TPM 1.2 when the SPAPR device model is chosen and use a
3065 default of '2.0' (TPM 2) for the backend.
3067 * qemu: Do not set ``//cpu/@migratable`` for running domains
3069 Libvirt release of 6.4.0 started to fill the default value for
3070 ``//cpu/@migratable`` attribute according to QEMU support. However, active
3071 domains either have the migratable attribute already set or they were
3072 started with older Libvirt which doesn't support the attribute.
3080 * Allow firmware blobs configuration
3082 QEMU offers a way to tweak how firmware configures itself
3083 and/or provide new configuration blobs. New ``<sysinfo/>``
3084 type is introduced that will hold these new blobs.
3085 It's possible to either specify new value as a string or
3086 provide a filename which contents then serve as the value.
3088 * nodedev: Add ability to create mediated devices
3090 Mediated devices can now be created with ``virNodeDeviceCreateXML()``. This
3091 functionality requires the ``mdevctl`` utility to be installed. The XML
3092 schema for node devices was expanded to support attributes for mediated
3095 * QEMU: add TPM Proxy device support
3097 libvirt can now create guests using a new device type called
3098 "TPM Proxy". The TPM Proxy connects to a TPM Resource Manager
3099 present in the host, enabling the guest to run in secure virtual
3100 machine mode with the help of an Ultravisor. Adding a TPM Proxy to
3101 a pSeries guest brings no security benefits unless the guest is
3102 running on a PPC64 host that has Ultravisor and TPM Resource Manager
3103 support. Only one TPM Proxy is allowed per guest. A guest using
3104 a TPM Proxy device can instantiate another TPM device at the same
3105 time. This device is supported only for pSeries guests via the new
3106 'spapr-tpm-proxy' model of the TPM 'passthrough' backend.
3108 * virhook: Support hooks placed in several files
3110 Running all scripts from directory /etc/libvirt/hooks/<driver>.d in
3111 alphabetical order. Hook script in old place will be executed
3112 as first for backward compatibility.
3114 * qemu: Add support for migratable host-passthrough CPU
3116 QEMU 2.12 made it possible for guests to use a migration-friendly
3117 version of the host-passthrough CPU. This feature is now exposed by
3122 * network: Support NAT with IPv6
3124 It's now possible to use ``<nat ipv6="yes"/>`` in a libvirt network.
3126 * qemu: Auto-fill NUMA information for incomplete topologies
3128 If the NUMA topology is not fully described in the guest XML, libvirt
3129 will complete it by putting all unspecified CPUs in the first NUMA node.
3130 This is only done in the QEMU binary itself supports disjointed CPU
3131 ranges for NUMA nodes.
3133 * qemu: Assign hostdev-backed interfaces to PCIe slots
3135 All SR-IOV capable devices are PCIe, so when their VFs are assigned to
3136 guests they should end up in PCIe slots rather than conventional PCI ones.
3140 * qemu: fixed crash in ``qemuDomainBlockCommit``
3142 This release fixes a regression which was introduced in libvirt v6.4.0
3143 where libvirtd always crashes when a block commit of a disk is requested.
3145 * qemu: fixed zPCI address auto generation on s390
3147 Removes the correlation between the zPCI address attributes uid and fid.
3148 Fixes the validation and autogeneration of zPCI address attributes.
3150 * qemu: Skip pre-creation of NVMe disks during migration
3152 libvirt has no way to create NVMe devices on the target host, so it now
3153 just makes sure they exist and let the migration proceed in that case.
3161 * qemu: Add support for pvscsi controllers
3163 pvscsi is the VMware paravirtualized SCSI controller, which has been
3164 supported in QEMU for a number of years.
3166 * cpu: Report model information for ARM CPUs
3168 ``virsh capabilities`` will now include information about the host CPU when
3169 run on ARM machines.
3171 * qemu: support network interface downscript
3173 QEMU has the ability to run a script when a NIC is brought up and down.
3174 Libvirt only enables use of the up script. Now add support for postscript
3175 when NIC is down/detached.
3179 * qemu: stricter validation for disk type='lun'
3181 The 'lun' type is meant for SCSI command passthrough, which can't be
3182 achieved if qemu's block layer features are used. Disk type='lun' is now
3183 allowed only when the format is 'raw' and no other block layer features are
3186 * qemu: auto-fill of incomplete NUMA topologies
3188 Domains with incomplete NUMA topologies, where the sum of vCPUs in all NUMA
3189 cells is less than the total of vCPUs, will get their first NUMA cell to
3190 be auto-filled with the remaining vCPUs. This behavior reproduces what QEMU
3191 already does in these cases. Users are encouraged to provide complete NUMA
3192 topologies to avoid unexpected changes in the domain XML.
3194 * Cooperlake x86 CPU model is added
3198 * qemu: fixed regression in network device hotplug with new qemu versions
3200 Starting from QEMU-5.0 it's required to conform to strict schema when
3201 hotplugging network devices. Libvirt didn't conform to the schema so in
3202 versions prior to 6.4.0 network device hotplug fails in certain cases. This
3203 version fixes it and adds stricter testing to prevent further issues.
3205 * remote: Look up libxl driver correctly
3207 This makes ``xen://`` connection URIs usable in split daemon mode.
3209 * systemd: Start libvirtd after firewalld/iptables services
3211 This solves an issue where iptables rules and chains created by libvirtd
3212 would get removed by a service started after it.
3214 * network: Re-create iptables chains on firewalld restart
3216 firewalld resets all iptables rules and chains on restart, and this
3217 includes deleting those created by libvirt.
3219 * qemu: reject readonly attribute for virtiofs
3221 virtiofs does not yet support read-only shares.
3229 * qemu: support disabling hotplug/unplug of PCIe devices
3231 libvirt can now set the "hotplug" option for pcie-root-ports and
3232 pcie-switch-downstream-ports, which can be used to disable hotplug/unplug
3233 of devices from these ports (default behavior is for these controllers to
3234 accept all hotplug/unplug attempts, but this is often undesirable).
3236 * vbox: added support for version 6.0 and 6.1 APIs
3238 libvirt can now support use of the VirtualBox 6.0 and 6.1 APIs. This is
3239 compile tested only, so we are looking for feedback from users on how well
3240 it works in practice.
3242 * xen: Add support for 'e820_host' hypervisor feature
3244 ``e820_host`` is a Xen-specific option only available for PV guests. When
3245 enabled it provides the guest with a virtual e820 memory map based on the
3246 host one. It must be enabled to allow hotplugging PCI devices to PV guests,
3247 particularly when memory ballooning is enabled.
3249 * xen: Add support for 'passthrough' hypervisor feature
3251 ``passthrough`` is a Xen-specific option new to Xen 4.13 that enables PCI
3252 passthrough for guests. It must be enabled to allow hotplugging PCI
3255 * qemu: support async IO mode 'io_uring'
3257 Linux 5.1 introduces a fast and efficient async IO interface io_uring, then
3258 qemu implements that in version 5.0. It could be used in file, host device
3259 and host cdrom backend by ``io='io_uring'`` of disk XML.
3261 * Lease time option included for network DHCP settings
3263 Users can now configure expiry time for leases for networks where libvirt
3264 manages DHCP. The time can be specified for whole range and/or fine tuned
3265 per individual host.
3267 * qemu: Implement pSeries Spectre mitigation features
3269 Users can now setup the following capabilities of pSeries guests: CFPC
3270 (Cache Flush on Privilege Change), SBBC (Speculation Barrier Bounds
3271 Checking) and IBS (Indirect Branch Speculation).
3273 * qemu: Add support for virtio packed option
3275 The ``packed`` attribute controls if QEMU should try to use packed
3276 virtqueues. Possible values are ``on`` or ``off``.
3280 * qemu: Allow checkpoint redefine for offline VMs
3282 Skip the liveness and capability checks when redefining checkpoints as we
3283 don't need QEMU interactions to update the metadata.
3285 * daemons: Improve timeout handling
3287 Daemons now support ``--timeout 0`` which suppresses daemon killing after
3288 given time of inactivity.
3290 * qemu: Add support for 'multidevs' option
3292 This option prevents misbehaviours on guest if a QEMU 9pfs export contains
3293 multiple devices, due to the potential file ID collisions this otherwise
3298 * qemu: Various embed driver fixes
3300 When using shared resources from embed driver (e.g. hugepages, machined,
3301 etc.) libvirt now generates an unique handler that is not conflicting with
3302 other embed drivers or system or session daemons.
3304 * cpu: Distinguish Cascadelake-Server from Skylake-Server
3306 Libvirt now properly detects Cascadelake-Server and Skylake-Server
3307 processors which differ only in stepping.
3309 * qemu: Fix domain restore from a block device
3311 When using namespaces, libvirt was unable to restore a domain from a block
3312 device because libvirt tried to relabel the device inside the namespace
3313 while QEMU was given FD to the block device in the host.
3315 * node_device_udev: Handle move events
3317 Libvirt now handles ``move`` event which is emitted on a NIC rename.
3319 * qemu: Fix capabilities probing with TCG
3321 Libvirt no long assumes TCG is always available. It now detects whether
3322 QEMU supports TCG and reports it accordingly.
3324 * **Removed features**
3326 * vbox: removed support for version 5.0 and 5.1 APIs
3328 libvirt no longer supports use of VirtualBox 5.0 and 5.1 since these
3329 versions reached their end of life on 2017/05 and 2018/04 respectively.
3337 * qemu: NVDIMM support for pSeries guests
3339 QEMU 5.0 implements NVDIMM memory support for pSeries guests. This is done
3340 by adding an 'uuid' element in the memory XML, which can either be provided
3341 in the XML or, if omitted, generated automatically.
3343 * qemu: Add virtiofs support
3345 This feature, introduced in QEMU 4.2, is a more modern alternative to
3346 virtio-9p, which is exposed through the same ``<filesystem/>`` element.
3348 * admin: Support reloading TLS certificates
3350 After renewing TLS certificates, it was usually necessary to restart
3351 libvirtd for the new ones to be loaded: now the same result can be obtained
3352 without restarting the daemon by using ``virt-admin server-update-tls`` .
3354 * **Removed features**
3356 * Removed support for INI style of comments
3358 With switching of our internal code to GLib, parsing of client
3359 authentication config files is handed over to GLib which does not support
3360 ``INI`` style of comments starting with a semicolon ( ``;`` ). Use number
3361 sign ( ``#`` ) instead.
3365 * qemu: Don't compare local and remote hostnames on migration
3367 This check was introduced to prevent same-host migration, but did not work
3368 as expected when multiple libvirtd instances were running on the same host
3369 but in different containers. With this release, the host UUID (which should
3370 be unique to the container) is checked instead.
3372 * qemu: Use per-VM event loops
3374 Instead of using a single even loop to process communication with the QEMU
3375 monitor and guest agent, create a separate one for each VM. This helps with
3376 scalability and prevents scenarios where a single malfunctioning VM could
3377 affect all those running on the same host.
3379 * qemu: Support migration with SLIRP helper interface
3381 With QEMU 5.0, a new D-Bus backend allows migration of external processes.
3382 When needed, libvirt will start a per-vm D-Bus bus, and migrate the
3383 slirp-helper along with QEMU.
3387 * qemu: Open backing chain late for shallow block copy reusing external
3390 With introduction of -blockdev for QEMU storage configuration in
3391 libvirt-5.10 we've started opening the backing chain of the
3392 destination/mirror of a virDomainBlockcopy started with
3393 VIR_DOMAIN_BLOCK_COPY_REUSE_EXT | VIR_DOMAIN_BLOCK_COPY_SHALLOW flags when
3394 starting the job rather than when virDomainBlockJobAbort with
3395 VIR_DOMAIN_BLOCK_JOB_ABORT_PIVOT is issued. For users depending on this
3396 undocumented quirky pre-blockdev behaviour this caused a regression as the
3397 backing chain could not be modified while the copy of the top image was
3398 progressing due to QEMU image locking. Note that this fix also requires
3399 qemu-5.0 while -blockdev is used starting from QEMU-4.2.
3401 * Don't generate machine names containing dots
3403 Even though the guest name containing dots is not a problem for libvirt
3404 itself, we need to strip them out when registering with machined because of
3405 the latter's requirements.
3413 * qemu: new rng backend type: builtin
3415 It implements qemu builtin rng backend. That uses getrandom syscall to
3416 generate random, no external rng source needed. Available since QEMU 4.2.
3418 * support for virtio+hostdev NIC <teaming>
3420 QEMU 4.2.0 and later, combined with a sufficiently recent guest virtio-net
3421 driver (e.g. the driver included in Linux kernel 4.18 and later), supports
3422 setting up a simple network bond device comprised of one virtio emulated
3423 NIC and one hostdev NIC (which must be an SRIOV VF). (in QEMU, this is
3424 known as the "virtio failover" feature). The allure of this setup is that
3425 the bond will always favor the hostdev device, providing better
3426 performance, until the guest is migrated - at that time QEMU will
3427 automatically unplug the hostdev NIC and the bond will send all traffic via
3428 the virtio NIC until migration is completed, then QEMU on the destination
3429 side will hotplug a new hostdev NIC and the bond will switch back to using
3430 the hostdev for network traffic. The result is that guests desiring the
3431 extra performance of a hostdev NIC are now migratable without network
3432 downtime (performance is just degraded during migration) and without
3433 requiring a complicated bonding configuration in the guest OS network
3434 config and complicated unplug/replug logic in the management application on
3435 the host - it can instead all be accomplished in libvirt with the interface
3436 <teaming> subelement "type" and "persistent" attributes.
3438 * support BR_ISOLATED flag for guest interfaces attached to a Linux host
3441 Since Linux kernel 4.18, the Linux host bridge has had a flag BR_ISOLATED
3442 that can be applied to individual ports. When this flag is set for a port,
3443 traffic is blocked between that port and any other port that also has the
3444 BR_ISOLATED flag set. libvirt domain interface config now supports setting
3445 this flag via the <port isolated='yes'/> setting. It can also be set for
3446 all connections to a particular libvirt network by setting the same option
3447 in the network config - since the port for the host itself does not have
3448 BR_ISOLATED set, the guests can communicate with the host and the outside
3449 world, but guests on that network can't communicate with each other. This
3450 feature works for QEMU and LXC guests with interfaces attached to a Linux
3453 * qemu: Introduce the 'armvtimer' timer type
3455 QEMU 5.0 introduces the ability to control the behavior of the virtual
3456 timer for KVM ARM/virt guests, and this new timer type exposes the same
3457 capability to libvirt users.
3459 * qemu: Storage configuration improvements
3461 Libvirt now accepts ``<backingStore type='volume'>`` and allows specifying
3462 the offset and size of the image format container inside the storage source
3463 via the ``<slices>`` subelement.
3465 * qemu: Introduce the 'tpm-spapr' TPM model
3467 This device, available starting from QEMU 5.0, is limited to pSeries
3470 * qemu: support Panic Crashloaded event handling
3472 The pvpanic device now supports a 'crashloaded' event, which is emitted
3473 when a guest panic has occurred but has already been handled by the guest
3476 * qemu: Implement virDomainGetHostnameFlags
3478 The ``--source`` argument to ``virsh domhostname`` can be used to specify
3479 what data source to use for the domain hostnames. Currently, in addition
3480 to the 'agent', libvirt can also use 'lease' information from dnsmasq to
3485 * qemu: Image format probing is allowed in certain cases
3487 To resolve regressions when users didn't specify the backing image format
3488 in the overlay, libvirt now probes the format in certain secure scenarios
3489 which fixes a few common existing cases. Additionally the knowledge base
3490 was extended to provide more information on how to rectify the problem.
3492 * qemu: Support "dies" in CPU topology
3494 This CPU topology concept, new in QEMU 4.1.0, sits between the existing
3495 "socket" and "core".
3497 * libxl: Add support for Credit2 scheduler parameters
3499 * lxc: Add support LXC 3 network configuration format
3503 * conf: Do not generate machine names ending with a dash
3505 Recent systemd versions do not allow them.
3507 * **Packaging changes**
3509 * use of gnulib has been completely eliminated
3511 Historically libvirt has embedded gnulib to provide fixes for various
3512 platform portability problems. This usage has now been eliminated and
3513 alternative approaches for platform portability problems adopted where
3514 required. This has been validated on the set of platforms covered by
3515 automated CI build testing. Other modern Linux distros using glibc are
3516 expected to work. Linux distros using non-glibc packages, and other
3517 non-Linux platforms may encounter regressions when building this release.
3518 Please report any build problems encountered back to the project
3519 maintainers for evaluation.
3525 * **Packaging changes**
3527 * support for python2 is removed
3529 Libvirt is no longer able to be built using the Python 2 binary. Python 3
3530 must be used instead.
3532 * docs: the python docutils toolset is now required
3534 The use of rst2html has been introduced for the website build process since
3535 docs are now being written in the RST as an alternative to HTML.
3539 * new PCI hostdev address type: unassigned
3541 A new PCI hostdev address type 'unassigned' is introduced. An unassigned
3542 PCI hostdev behaves like any regular PCI hostdev inside Libvirt, but it is
3543 not usable by the guest. This gives the user a new option to manage the
3544 binding of PCI devices via Libvirt, declaring PCI hostdevs in the domain
3545 XML but allowing just a subset of them to be assigned to the guest.
3547 * Provide init scripts for sub-deaemons
3549 So far libvirt shipped systemd unit files for sub-daemons. With this
3550 release, init scripts are available too. Package maintainers can choose
3551 which one to install via ``--with-init-script`` configure option.
3553 * qemu: Support cold-unplug of sound devices
3555 * qemu: Implement VIR_MIGRATE_PARAM_TLS_DESTINATION
3557 This flag, which can be enabled using ``virsh`` 's ``--tls-destination``
3558 option, allows migration to succeed in situations where there is a mismatch
3559 between the destination's hostname and the information stored in its TLS
3562 * qemu: Support reporting memory bandwidth usage stats
3564 Implement Intel RDT-MBM in libvirt. The stats can be obtained via ``virsh
3565 domstats --memory`` .
3567 * qemu: Allow accessing NVMe disks directly
3569 Before this release there were two ways to configure a NVMe disk for a
3570 domain. The first was using <disk/> with the <source/> pointing to the
3571 ``/dev/nvmeXXXX`` . The other was using PCI assignment via <hostdev/>
3572 element. Both have their disadvantages: the former adds latency of file
3573 system and block layers of the host kernel, the latter prohibits domain
3574 migration. In this release the third way of configuring NVMe disk is added
3575 which combines the advantages and drops disadvantages of the previous two
3576 ways. It's accessible via <disk type='nvme'/>.
3578 * **Removed features**
3580 * 'phyp' Power Hypervisor driver removed
3582 The 'phyp' Power Hypervisor driver has not seen active development since
3583 2011 and does not seem to have any real world usage. It has now been
3588 * qemu: xz save image compression is faster
3590 When using the xz format to compressed virtual machine saved state images,
3591 the "-3" compression level preset is now used. This results in slightly
3592 larger files, but with a massively reduced time to compress. The xz format
3593 offers the best compression level for saved state images, albeit still with
3594 the slowest running time. For the fastest possible running time, at cost of
3595 the larest compressed size, lzop should be used.
3597 * domain: Improve job stat handling
3599 It is now possible to retrieve stats for completed and failed jobs.
3601 * qemu: Don't hold monitor and agent job at the same time
3603 Before this change, a malicious (or buggy) ``qemu-guest-agent`` running in
3604 the guest could make other libvirt APIs unavailable for an unbounded amount
3609 * qemu: Report error if backing image format is not specified explicitly
3611 For a long time libvirt was assuming that a backing file is RAW when the
3612 format was not specified. This didn't pose a problem until blockdev support
3613 was enabled in last release. Libvirt now requires that the format is
3614 specified in the image metadata or domain XML and the VM will refuse to
3615 start otherwise. Additionally the error message now links to the knowledge
3616 base which summarizes how to fix the images.
3618 * qemu: Fix non-shared storage migration over NBD
3620 * qemu: Generate a single MAC address for hotplugged network devices
3622 Since libvirt 4.6.0, when hotplugging a network device that didn't have a
3623 MAC address already assigned by the user, two separate addresses would be
3624 generated: one for the live configuration, which would show up immediately,
3625 and one for the inactive configuration, which would show up after the first
3626 reboot. This situation was clearly undesirable, so a single MAC address is
3627 now generated and used both for the live configuration and the inactive
3631 v5.10.0 (2019-12-02)
3632 ====================
3636 * qemu: Introduce support for ARM CPU features
3638 The only features supported at the moment are SVE vector lengths, which
3639 were introduced in QEMU 4.2.0.
3641 * qemu: Support boot display for GPU mediated devices
3643 Until now, GPU mediated devices generally did not show any output until the
3644 guest OS had initialized the vGPU. By specifying the ``ramfb`` attribute,
3645 QEMU can be configured to use ramfb as a boot display for the device: this
3646 allows for display of firmware messages, boot loader menu, and other output
3647 before the guest OS has initialized the vGPU.
3649 * Add API to change the response timeout for guest agent commands
3651 By default, when a command is sent to the guest agent, libvirt waits
3652 forever for a response from the guest agent. If the guest is unresponsive
3653 for any reason, this can block the calling thread indefinitely. By setting
3654 a custom timeout using ``virDomainAgentSetResponseTimeout()`` , API users
3655 can change this behavior.
3659 * Devices CGroup v2 support
3661 Libvirt supported all controllers of CGroup v2 but the devices controller
3662 which is implemented in this release.
3664 * Cold plug of sound device
3666 The QEMU driver now can handle cold plug of ``<sound/>`` devices.
3668 * Probe for default CPU types
3670 With QEMU 4.2.0 we can probe for the default CPU model used by QEMU for a
3671 particular machine type and store it in the domain XML. This way the chosen
3672 CPU model is more visible to users and libvirt will make sure the guest
3673 will see the exact same CPU after migration.
3675 * Adaptation to qemu's blockdev
3677 QEMU introduced a new way of specifying disks on the command line which
3678 enables fine-grained control over the block stack. Libvirt has adapted to
3683 * More GLib integration
3685 More patches were merged that replace our internal functions with GLib
3686 ones. Also some effort was invested in replacing gnulib modules with GLib
3689 * Rewrite of Perl scripts into Python
3691 Libvirt used Perl scripts to check for coding style, generate some code and
3692 things like that. To bring the number of languages used down, these scripts
3693 were rewritten into Python.
3697 * Warn verbosely if using old loader:nvram pairs
3699 Some distributions still use ``--with-loader-nvram`` or ``nvram`` variable
3700 in qemu.conf. This is now discouraged in favour of FW descriptors. However,
3701 instead of silently ignoring user's config, libvirt warns if outdated
3704 * Drop pconfig from Icelake-Server CPU model
3706 The pconfig feature was enabled in QEMU by accident in 3.1.0. All other
3707 newer versions do not support it and it was removed from the Icelake-Server
3710 * Wait longer for device removal confirmation on PPC64
3712 After sending device hot unplug request to QEMU, libvirt waits up to 5
3713 seconds for qemu to confirm the device removal. On some architectures (like
3714 PPC64) this can take longer time and libvirt now reflects that.
3716 * Forcibly create nodes in domain's namespace
3718 The QEMU driver starts a domain in a namespace with private ``/dev`` and
3719 creates only those nodes there which the domain is configured to have.
3720 However, it may have happened that if a node changed its minor number this
3721 change wasn't propagated to the namespace.
3723 * Various AppArmor bugfixes
3725 The AppArmor driver now knows how to handle ``<shmem/>`` devices and also
3726 snapshotting more disks at once.
3728 * Improved video model autoselection
3730 If a graphics device was added to XML that had no video device, libvirt
3731 automatically added a video device which was always of type 'cirrus' on
3732 x86_64, even if the underlying qemu didn't support cirrus. Libvirt now
3733 bases the decision on qemu's capabilities.
3739 * **Packaging changes**
3741 * Start linking against GLib and using its features
3743 Up until now, libvirt has been dealing with platform portability and the
3744 lack of certain features in libc by using gnulib and implementing its own
3745 functions and data structures respectively; going forward, it will prefer
3746 the facilities offered by GLib instead.
3748 * Stop distributing generated documentation
3750 Most downstreams already patch the libvirt source to some extent, so this
3751 change will probably not affect them.
3753 * Rewrite several Perl scripts in Python
3755 Phasing out Perl usage is part of the project strategy.
3759 * qemu: Introduce a new video model of type 'ramfb'
3761 Introduce a new video model type to the domain XML that supports the
3762 ``ramfb`` standalone device in qemu.
3764 * qemu: Implement the ccf-assist pSeries feature
3766 Users can now decide whether ccf-assist (Count Cache Flush Assist) support
3767 should be available to pSeries guests.
3769 * Xen: Support specifying ACPI firmware path
3771 The libxl driver now supports specifying an ACPI firmware path using the
3774 * qemu: Support specifying resolution for video devices
3776 * **Removed features**
3778 * logging: Drop support for including stack traces
3780 This feature was intended to aid debugging, but in practice it resulted in
3781 logs that were too verbose to be useful and also resulted in a significant
3782 performance penalty.
3786 * qemu: Implement CPU comparison/baseline on s390x
3788 This functionality has been historically limited to x86_64, but it's now
3789 available on s390x too.
3793 * lib: autostart objects exactly once
3795 If libvirtd or any of the sub-daemons is started with socket activation
3796 then objects might be autostarted more than once. For instance, if a domain
3797 under ``qemu:///session`` URI is mark as autostarted and the session daemon
3798 is started then the domain is started with it. If user shuts the domain
3799 down and the session daemon is started again, the user's wish to keep the
3800 domain shut off is ignored and the domain is autostarted again. This is now
3803 * qemu: Properly advertise bochs-display availability
3805 Support for ``bochs-display`` was introduced in libvirt 5.6.0, but until
3806 now the model was not listed in the domain capabilities.
3808 * security: Don't remember labels for TPM devices
3810 Due to the way they're implemented in the kernel, trying to remember labels
3811 for TPM devices makes it impossible to use them.
3813 * security: Properly rollback after failure in a stacked driver
3815 When multiple security drivers are involved, failure in one of them would
3816 result in only the corresponding changes being rolled back, leaving the
3817 ones performed by drivers that had been activated earlier in place. All
3818 changes are rolled back now.
3820 * Fix build with musl libc
3822 * Improve compatibility with non-bash shells
3830 * qemu: Support use of precreated tap/macvtap devices by unprivileged
3833 It is now possible for an unprivileged libvirtd to make use of tap and
3834 macvtap devices that were previously created by some other entity. This is
3835 done by setting ``managed='no'`` along with the device name in the
3836 ``target`` subelement of ``<interface type='ethernet'>`` .
3838 * qemu: Support vhost-user-gpu
3840 Support for running virtio GPUs in separate processes with vhost-user
3841 backend. It requires QEMU newer than 4.1.
3843 * Introduce virConnectSetIdentity API
3845 When split daemons are in use, this API is used to forward uid, gid and
3846 SELinux info from ``virproxyd`` to other driver daemons such as
3851 * qemu: Support running SLIRP networking in a separate process
3853 User can configure the slirp-helper path in ``qemu.conf`` . It will start a
3854 slirp-helper process to provide SLIRP networking when the VM is started
3855 with network interface "user". That will allow stricter security policies
3856 for QEMU SLIRP network.
3858 * **Removed features**
3860 * Remove xenapi driver
3862 The xenapi driver is removed since it has not received any significant
3863 development since its initial contribution nine years ago and has no known
3872 * qemu: Support Direct Mode for Hyper-V Synthetic timers
3874 The QEMU driver now supports Direct Mode for Hyper-V Synthetic timers for
3877 * lib: Add virDomainGetGuestInfo()
3879 This API is intended to aggregate several guest agent information queries
3880 and is inspired by stats API ``virDomainListGetStats()`` . It is
3881 anticipated that this information will be provided by a guest agent running
3882 within the domain. It's exposed as ``virsh guestinfo`` .
3884 * Experimental split of libvirtd into separate daemons
3886 The big monolithic libvirtd daemon can now be replaced by smaller
3887 per-driver daemons. The new split daemons are considered experimental at
3888 this time and distributions are encouraged to continue using the
3889 traditional libvirtd by default.
3891 * qemu: Support kvm-hint-dedicated performance hint
3893 With ``<hint-dedicated state='on'/>`` and ``<cpu
3894 mode='host-passthrough'/>`` , it allows a guest to enable optimizations
3895 when running on dedicated vCPUs. QEMU newer than 2.12.0 and kernel newer
3896 than 4.17 are required.
3898 * **Removed features**
3900 * Remove KVM assignment support
3902 The KVM style of PCI device assignment was removed from the kernel in
3903 version 4.12.0 after being deprecated since 4.2.0. Libvirt defaults to VFIO
3904 for a long time. Remove support for KVM device assignment from libvirt too.
3906 * libxml: min required libxml is now 2.9.1
3908 Support for building with libxml versions older than 2.9.1 has been
3913 * virsh: Support setting bandwidth in migrate subcommand
3915 In addition to postcopy bandwidth, the ``virsh migrate`` subcommand now
3916 supports specifying precopy bandwidth with the ``--bandwidth`` parameter.
3918 * libxl: Implement domain metadata getter/setter
3920 The libxl driver now supports ``virDomainGetMetadata()`` and
3921 ``virDomainSetMetadata()`` APIs.
3923 * test driver: Expand API coverage
3925 Additional APIs have been implemented in the test driver.
3927 * Report RNG device in domain capabilities XML
3929 Libvirt now reports if RNG devices are supported by the underlying
3930 hypervisor in the domain capabilities XML.
3932 * Stop linking NSS plugins with libvirt.so
3934 This reduces the amount of code and 3rd party libraries are that loaded
3937 * Split the setuid virt-login-shell binary into two pieces
3939 The setuid virt-login-shell binary is now a tiny shim that sanitizes the
3940 process execution environment variables and arguments, before launching the
3941 trusted virt-login-shell-helper binary.
3943 * qemu: Allow migration with disk cache on
3945 When QEMU supports flushing caches at the end of migration, we can safely
3946 allow migration even if ``disk/driver/@cache`` is neither ``none`` nor
3951 * Various security label remembering fixes
3953 In the previous release libvirt introduced remembering of original owners
3954 and SELinux labels on files. However, the feature did not work properly
3955 with snapshots, on migrations or on network filesystems. This is now fixed.
3957 * Allow greater PCI domain numbers
3959 Libvirt used to require PCI domain number to be not greater than 0xFFFF.
3960 The code was changed to allow 32 bits long numbers.
3962 * Various D-Bus fixes
3964 When D-Bus is not available, libvirt was reporting random errors. These are
3967 * Prefer read-only opening of PCI config files
3969 When enumerating PCI bus, libvirt opens config files under ``sysfs`` mount
3970 and parses them to learn various aspects of the device (e.g. its
3971 capabilities). Only in a very limited number of cases it is actually
3972 writing into the file. However, it used to open the file also for writing
3973 even if it was only reading from it.
3975 * Fix AppArmor profile
3977 Since the ``5.6.0`` release, libvirt uses ``procfs`` to learn the list of
3978 opened file descriptors when spawning a command. However, our AppArmor
3979 profile was not allowing such access.
3981 * Don't block storage driver when starting or building a pool
3983 Starting or building a storage pool can take a long time to finish. During
3984 this time the storage driver was blocked and thus no other API involving
3985 the storage driver could run. This is now fixed.
3993 * qemu: Introduce a new video model of type 'bochs'
3995 Introduce a new video model type that supports the ``bochs-display`` device
3996 that was added in qemu version 3.0.
3998 * api: new virDomainCheckpoint APIs
4000 Introduce several new APIs for creating and managing checkpoints in the
4001 test and qemu drivers (the latter requires qcow2 images). Checkpoints serve
4002 as a way to tell which portions of a disk have changed since a point in
4005 * qemu: Add support for overriding max threads per process limit
4007 systemd-based systems impose a limit on the number of threads a process can
4008 spawn, which in some cases can be exceeded by QEMU processes running VMs.
4009 Add a ``max_threads_per_process`` option to qemu.conf to override the
4012 * Remember original owners and SELinux labels of files
4014 When a domain is starting up libvirt changes DAC and SELinux labels so that
4015 domain can access it. However, it never remembered the original labels and
4016 therefore the file was returned back to ``root:root`` . With this release,
4017 the original labels are remembered and restored properly.
4019 * network: Allow passing arbitrary options to dnsmasq
4021 This works similarly to the existing support for passing arbitrary options
4022 to QEMU, and just like that feature it comes with no support guarantees.
4024 * **Removed features**
4026 * xen: Remove sxpr config support
4028 Remove the sxpr style config parser and formatter a year after the xend
4033 * qemu: Allow XML validation for snapshot creation
4035 Add flag ``VIR_DOMAIN_SNAPSHOT_CREATE_VALIDATE`` to validate snapshot input
4036 XML. For virsh, users can use it as ``virsh snapshot-create --validate`` .
4038 * Support encrypted soft TPM
4040 A soft TPM backend could be encrypted with passphrase. Now libvirt supports
4041 using a ``secret`` object to hold the passphrase, and referring to it via
4042 the ``encryption`` element of the TPM device.
4044 * test driver: Expand API coverage
4046 Additional APIs have been implemented in the test driver.
4048 * Implement per-driver locking
4050 Drivers now acquire a lock when they're loaded, ensuring that there can
4051 never be two instances of the same driver active at a time.
4053 * nss: Report newer addresses first
4055 In some cases, a guest might be assigned a new IP address by DHCP before
4056 the previous lease has expired, in which case the NSS plugin will correctly
4057 report both addresses; many applications, however, ignore all addresses but
4058 the first, and may thus end up trying to connect using a stale address. To
4059 prevent that from happening, the NSS plugin will now always report the
4060 newest address first.
4062 * util: Optimize mass closing of FDs when spawning child processes
4064 When the limit on the number of FDs is very high, closing all unwanted FDs
4065 after calling ``fork()`` can take a lot of time and delay the start of the
4066 child process. libvirt will now use an optimized algorithm that minimizes
4071 * logging: Ensure virtlogd rollover takes priority over logrotate
4073 virtlogd implements its own rollover mechanism, but until now logrotate
4074 could end up acting on the logs before virtlogd had a chance to do so
4083 * api: Prevent access to several APIs over read-only connections
4085 Certain APIs give root-equivalent access to the host, and as such should be
4086 limited to privileged users. CVE-2019-10161, CVE-2019-10166,
4087 CVE-2019-10167, CVE-2019-10168.
4091 * qemu: Support SMMUv3 IOMMU
4093 SMMUv3 is an IOMMU implementation for ARM virt guests.
4095 * network: Introduce the network port API
4097 This new public API can be used by virtualization drivers to manage network
4098 resources associated with guests, and is a further step towards splitting
4099 libvirtd into multiple daemons.
4101 * **Removed features**
4103 * qemu: Remove support for virDomainQemuAttach and
4104 virConnectDomainXMLFromNative APIs
4106 The qemu implementations for the APIs mentioned above were removed and the
4107 APIs now return an error. The implementation was stale for a long time and
4108 did not work with modern QEMU command lines, generated from libvirt or
4111 * Stop supporting migration of config files from pre-XDG layout
4113 The new layout was introduced with libvirt 0.9.13 (Jul 2012).
4115 * Remove Avahi mDNS support
4117 This feature was never used outside of virt-manager, which has itself
4118 stopped using it a while ago.
4122 * sysinfo: Report SMBIOS information on aarch64
4124 While SMBIOS support has historically been limited to x86_64, modern
4125 aarch64 machines often offer access to the same information as well, and
4126 libvirt now exposes it to the user when that's the case.
4128 * test driver: Expand API coverage
4130 Even more APIs that were missing from the test driver have now been
4133 * virt-xml-validate: Allow input to be read from stdin
4135 * qemu: Validate spapr-vio addresses as 32-bit
4137 libvirt has always considered these addresses (used for pSeries guests) as
4138 64-bit, but the sPAPR specification says that they're 32-bit instead.
4142 * qemu: Set process affinity correctly when using <numatune>
4144 libvirt would mistakenly interpret the ``nodeset`` attribute as a list of
4145 CPUs instead of as a list of NUMA node, and the process affinity would be
4146 set incorrectly as a result; this has now been fixed.
4154 * cpu: Introduce support for the md-clear CPUID bit
4156 This bit is set when microcode provides the mechanism to invoke a flush of
4157 various exploitable CPU buffers by invoking the x86 ``VERW`` instruction.
4158 CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, CVE-2019-11091.
4160 * Restrict user access to virt-admin, virtlogd and virtlockd
4162 The intended users for these facilities are the ``root`` user and the
4163 ``libvirtd`` service respectively, but these restrictions were not enforced
4164 correctly. CVE-2019-10132.
4168 * test driver: Expand API coverage
4170 Several APIs that were missing from the test driver have now been
4173 * Avoid unnecessary static linking
4175 Most binaries shipped as part of libvirt, for example ``virtlogd`` and
4176 ``libvirt_iohelper`` , were embedding parts of the library even though they
4177 also linked against the ``libvirt.so`` dynamic library. This is no longer
4178 the case, which results in both the disk and memory footprint being
4181 * qemu: Report stat-htlb-pgalloc and stat-htlb-pgfail balloon stats
4183 These stats have been introduced in QEMU 3.0.
4187 * qemu: Fix emulator scheduler support
4189 Setting the scheduler for QEMU's main thread before QEMU had a chance to
4190 start up other threads was misleading as it would affect other threads
4191 (vCPU and I/O) as well. In some particular situations this could also lead
4192 to an error when the thread for vCPU #0 was being moved to its cpu,cpuacct
4193 cgroup. This was fixed so that the scheduler for the main thread is set
4196 * apparmor: Allow hotplug of vhost-scsi devices
4204 * qemu: Add support for setting the emulator scheduler parameters
4206 I/O threads and vCPU threads already support setting schedulers, but until
4207 now it was impossible to do so for the main QEMU thread (emulator thread in
4208 the libvirt naming). This is, however, requested for some very specific
4209 scenarios, for example when vCPU threads are running at such priority that
4210 could starve the main thread.
4212 * **Removed features**
4214 * vbox: Drop support for VirtualBox 4.x releases
4216 Support for all the 4.x releases was ended by VirtualBox maintainers in
4217 December 2015. Therefore, libvirt support for these releases is dropped.
4221 * qemu: Use PCI by default for RISC-V guests
4223 PCI support for RISC-V guests was already available in libvirt 5.1.0, but
4224 it required the user to opt-in by manually assigning PCI addresses: with
4225 this release, RISC-V guests will use PCI automatically when running against
4226 a recent enough (4.0.0+) QEMU release.
4228 * qemu: Advertise firmware autoselection in domain capabilities
4230 The firmware autoselection feature is now exposed in domain capabilities
4231 and management applications can query for accepted values, i.e. values that
4232 are accepted and for which libvirt found firmware descriptor files.
4233 Firmware Secure Boot support is also advertised.
4235 * Drop YAJL 1 support
4237 YAJL 2 is widely adopted and maintaining side by side support for two
4238 versions is unnecessary.
4242 * rpc: cleanup in virNetTLSContextNew
4244 Failed new gnutls context allocations in virNetTLSContextNew function
4245 results in double free and segfault. Occasional memory leaks may also
4248 * virsh: various completers fixes
4250 There were some possible crashers, memory leaks, etc. which are now fixed.
4252 * qemu: Make hugepages work with memfd backend
4254 Due to a bug in command line generation libvirt did not honor hugepages
4255 setting with memfd backend.
4257 * Enforce ACL write permission for getting guest time & hostname
4259 Getting the guest time and hostname both require use of guest agent
4260 commands. These must not be allowed for read-only users, so the permissions
4261 check must validate "write" permission not "read".
4269 * Add Storage Pool Capabilities output
4271 Add support to list an enumerated list of supported Storage Pools via the
4272 virConnectGetCapabilities API when connected via a Storage Driver. Add
4273 support to get a more detailed list XML output Storage Pool Capabilities
4274 vis the virConnectGetStoragePoolCapabilites API.
4276 * qemu: Support virtio-{non-}transitional device models
4278 ``virtio-transitional`` and ``virtio-non-transitional`` ``model`` values
4279 were added to the QEMU driver for the following devices: ``disk`` ,
4280 ``interface`` , ``filesystem`` , ``rng`` , ``vsock`` , ``memballoon`` ,
4281 ``controller`` type ``scsi`` , ``controller`` type ``virtio-serial`` ,
4282 ``input`` bus ``virtio`` type ``passthrough`` , ``hostdev`` type
4283 ``scsi_host`` . These new models can be used to give fine grained control
4284 over what virtio device version is presented to the guest.
4286 * qemu: Enable firmware autoselection
4288 Libvirt allows users to provide loader path for some time now. However,
4289 this puts some burden on users because they need to know what firmware
4290 meets their requirements. Now that QEMU ships firmware description files
4291 this burden can be moved onto libvirt. It is as easy as setting the
4292 ``firmware`` attribute in the ``os`` element (accepted values are ``bios``
4293 and ``efi`` ). Moreover, libvirt automatically enables domain features
4294 needed for firmware it chooses.
4296 * snapshots: Add support for topological listings
4298 A new flag VIR_DOMAIN_SNAPSHOT_LIST_TOPOLOGICAL is available for the
4299 various snapshot listing APIs such as virDomainListAllSnapshots(). For
4300 drivers that support the flag, the listed snapshots are guaranteed to be
4301 sorted such that parents occur before children.
4303 * Xen: Add support for max grant frames setting
4305 Add support for Xen's max_grant_frames setting by adding a new xenbus
4306 controller type with a maxGrantFrames attribute. E.g. ``<controller
4307 type='xenbus' maxGrantFrames='64'/>``
4309 * qemu: Add support for parallel migration
4311 With QEMU 4.0.0 libvirt can enable parallel migration which causes the
4312 memory pages to be processed in parallel by several threads and sent to the
4313 destination host using several connections at the same time. This may
4314 increase migration speed in case a single thread is unable to saturate the
4317 * **Removed features**
4319 * Drop support for Upstart and "Red Hat" init scripts
4321 Not a single one of the platforms we target still uses Upstart, and the
4322 Upstart project itself has been abandoned for several years now; the same
4323 is true for the "Red Hat" (really System V) init scripts, since RHEL 7 and
4324 later releases use systemd.
4328 * Report class information for PCI node device capability.
4330 * Split setup of IPv4 and IPv6 top level chain
4332 The requirement resulting from private chains improvement done in
4333 ``v5.1.0`` was refined so that only tables from corresponding IP version
4334 are required. This means that if a network doesn't have ``IPv6`` enabled
4335 then those tables are not required.
4337 * Don't default to building the QEMU driver
4339 Historically, the QEMU driver has been special in that it was enabled by
4340 default, with the option to explicitly opt-out of it; starting now, we're
4341 enabling it opportunistically if we detect that all requirements are
4342 available, just like we do with other drivers.
4346 * virt-host-validate: Fix IOMMU check on s390x
4348 * qemu: Allow creating pSeries guests with graphics and no USB mouse
4350 It's now possible to prevent libvirt from automatically adding a USB mouse
4351 to pSeries guests by including a USB tablet in the input XML: doing so is
4352 desiderable as using a tablet results in a much better user experience when
4355 * qemu: Set $HOME and XGD variables for qemu:///system guests
4357 This avoids files being accidentally created under ``/`` or the guests not
4358 being able to start because they lack the necessary permissions to write to
4367 * bhyve: Add support for additional command-line arguments
4369 The bhyve driver now supports passing additional command-line arguments to
4370 the bhyve process using the new ``<bhyve:commandline>`` element in domain
4373 * network: Support setting a firewalld "zone" for virtual network bridges
4375 All libvirt virtual networks with bridges managed by libvirt (i.e. those
4376 with forward mode of "nat", "route", "open", or no forward mode) will now
4377 be placed in a special firewalld zone called "libvirt" by default. The zone
4378 of any network bridge can be changed using the ``zone`` attribute of the
4379 network's ``bridge`` element.
4381 * bhyve: Support for ignoring unknown MSRs reads and writes
4383 A new <features> element <msrs unknown='ignore'/> was introduced and the
4384 bhyve driver supports it to control unknown Model Specific Registers (MSRs)
4387 * qemu: Add support for encrypted VNC TLS keys
4389 Use the password stored in the secret driver under the uuid specified by
4390 the ``vnc_tls_x509_secret_uuid`` option in qemu.conf.
4392 * Add storage pool namespace options
4394 Allow for adjustment of RBD configuration options via Storage Pool XML
4395 Namespace adjustments.
4397 * qemu: Add support for setting post-copy migration bandwidth
4399 Users can now limit the bandwidth of post-copy migration, e.g. via ``virsh
4400 migrate --postcopy-bandwidth`` .
4404 * Create private chains for virtual network firewall rules
4406 Historically firewall rules for virtual networks were added straight into
4407 the base chains. This works but has a number of bugs and design
4408 limitations. To address them, libvirt now puts firewall rules into its own
4409 chains. Note that with this change the ``filter`` , ``nat`` and ``mangle``
4410 tables are required for both ``IPv4`` and ``IPv6`` .
4412 * Detect CEPH and GPFS as shared FS
4414 When starting a migration libvirt performs some sanity checks to make sure
4415 domain will be able to run on the destination. One of the requirements is
4416 that the disk has to either be migrated too or be accessible from a network
4417 filesystem. CEPH and GPFS weren't detected as a network filesystem.
4419 * Advertise network MTU via DHCP when specified
4421 If network MTU is set and the network has DHCP enabled, advertise the MTU
4422 in DHCP transaction too so that clients can adjust their link accordingly.
4424 * qemu: Allocate memory at the configured NUMA nodes from start
4426 Libvirt used to just start QEMU, let it allocate memory for the guest, and
4427 then use CGroups to move the memory to configured NUMA nodes. This is
4428 suboptimal as huge chunks of memory have to be moved. Moreover, this relies
4429 on ability to move memory later which is not always true. A change was made
4430 to set process affinity correctly from the start so that memory is
4431 allocated on the configured nodes from the beginning.
4433 * Support for newer Wireshark
4435 Adapt libvirt to use the more recent release requiring a source build
4436 configuration of libvirt ``--with-wireshark`` to upgrade to the more recent
4439 * Batch mode virsh and virt-admin parsing improvements
4441 When parsing a single-argument command_string in batch mode, virsh and
4442 virt-admin now permit newlines in addition to semicolons for splitting
4443 commands, and backslash-newline for splitting long lines, to be more like
4448 * qemu: Use CAP_DAC_OVERRIDE during QEMU capabilities probing
4450 By default, libvirt runs the QEMU process as ``qemu:qemu`` which could
4451 cause issues during probing as some features like AMD SEV might be
4452 inaccessible to QEMU because of file system permissions. Therefore,
4453 ``CAP_DAC_OVERRIDE`` is granted to overcome these for the purposes of
4456 * storage: Add default mount options for fs/netfs storage pools
4458 Altered the command line generation for fs/netfs storage pools to add some
4459 default options. For Linux based systems, the options added are "nodev,
4460 nosuid, noexec". For FreeBSD based systems, the options added are "nosuid,
4463 * qemu: Allow use of PCI for RISC-V guests
4465 This works with QEMU 4.0.0+ only and is opt-in at the moment, since it
4466 requires users to manually assign PCI addresses, but is otherwise fully
4469 * network: Fix virtual networks on systems using firewalld+nftables
4471 Because of the transitional state of firewalld's new support for nftables,
4472 not all iptables features required by libvirt are yet available, so libvirt
4473 must continue to use iptables for its own packet filtering rules even when
4474 the firewalld backend is set to use nftables. However, due to the way
4475 iptables support is implemented in kernels using nftables (iptables rules
4476 are converted to nftables rules and processed in a separate hook from the
4477 native nftables rules), guest networking was broken on hosts with firewalld
4478 configured to use nftables as the backend. This has been fixed by putting
4479 libvirt-managed bridges in their own firewalld zone, so that guest traffic
4480 can be forwarded beyond the host and host services can be exposed to guests
4481 on the virtual network without opening up those same services to the rest
4482 of the physical network. This means that host access from virtual machines
4483 is no longer controlled by the firewalld default zone (usually "public"),
4484 but rather by the new firewalld zone called "libvirt" (unless configured
4485 otherwise using the new zone attribute of the network bridge element).
4487 * qemu: Fix i6300esb watchdog hotplug on Q35
4489 Ensure that libvirt allocates a PCI address for the device so that QEMU did
4490 not default to an address that would not allow for device hotplug.
4492 * lxc: Don't reboot host on virDomainReboot
4494 If the container is really a simple one (init is just bash and the whole
4495 root is passed through) then virDomainReboot and virDomainShutdown would
4496 reboot or shutdown the host. The solution is to use different method to
4497 reboot or shutdown the container in that case (e.g. signal).
4499 * rpc: Various stream fixes
4501 One particular race was fixed, one locking problem and error reporting from
4502 streams was made better.
4504 * qemu: Fix guestfwd hotplug/hotunplug
4506 Fixed the generation of the guestfwd hotplug/unplug command sent to QEMU to
4507 match the syntax used when creating the initial command line.
4509 * qemu: Forbid CDROMs on virtio bus
4511 Attempting to create an empty virtio-blk drive or attempting to eject it
4512 results into an error. Forbid configurations where users would attempt to
4513 use CDROMs in virtio bus.
4515 * qemu: Use 'raw' for 'volume' disks without format
4517 Storage pools might want to specify format of the image when translating
4518 the volume thus libvirt can't add any default format when parsing the XML.
4519 Add an explicit format when starting the VM and format is not present
4520 neither by user specifying it nor by the storage pool translation function.
4522 * qemu: Assume 'raw' default storage format also for network storage
4524 Post parse callback adds the 'raw' type only for local files. Remote files
4525 can also have backing store (even local) so we should do this also for
4526 network backed storage.
4528 * qemu: Fix block job progress reporting and advocate for READY event
4530 In some cases QEMU can get to 100% and still not reach the synchronised
4531 phase. Initiating a pivot in that case will fail. Therefore it is strongly
4532 advised to wait for ``VIR_DOMAIN_BLOCK_JOB_READY`` event which does not
4533 suffer from this problem.
4535 * qemu: Don't format image properties for empty drive
4537 If a ``-drive`` has no image, then formatting attributes such as cache,
4538 readonly, etc. would cause errors to be reported from QEMU. This was fixed
4539 by not supplying the attributes for devices without an image.
4541 * External snapshot metadata redefinition is fixed
4543 Attempting to use VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE to reinstate the
4544 metadata describing an external snapshot created earlier for an offline
4545 domain no longer fails.
4553 * Xen: Add support for openvswitch
4555 The libxl driver now supports virtual interfaces that connect to an
4556 openvswitch bridge, including interfaces with VLAN tagging and trunking
4559 * qemu: Report whether KVM nesting is available
4561 Running nested KVM guests requires specific configuration steps to be
4562 performed on the host; libvirt will now report in the host capabilities
4563 whether KVM nesting support is available.
4565 * **Removed features**
4569 The UML driver was unmaintained and not tested for quite some time now.
4570 Worse, there is a bug that causes it to deadlock on some very basic
4571 operations (e.g. dumping domain XML). These facts make us believe no one
4576 * qemu: Add support for ARMv6l guests
4578 * Support more NVDIMM configuration options
4580 Introduce more configuration options. For the source element, add the
4581 'alignsize' and 'pmem' subelements. For the target element, add the
4582 'readonly' subelement.
4584 * cpu: Add support for "stibp" x86_64 feature
4586 Add cpu flag stibp (Single Thread Indirect Branch Predictors) to prevent
4587 indirect branch predictions from being controlled by the sibling
4590 * libxl: Handle external domain destroy
4592 Historically, if a domain was destroyed using ``xl`` rather than through
4593 libvirt APIs, libvirt would not be aware of the fact and keep considering
4594 it as running. This is no longer the case.
4596 * Start selecting the first available DRI device for OpenGL operations
4598 If OpenGL support is needed (either with SPICE gl enabled or with
4599 egl-headless), libvirt is now able to pick the first available DRI device
4600 for the job. At the same time, this improvement is also a bugfix as it
4601 prevents permission-related issues with regards to our mount namespaces and
4602 the default DRI render node's permissions which would normally prevent QEMU
4603 from accessing such a device.
4605 * qemu: Add support for postcopy-requests migration statistics
4607 The ``virDomainJobInfo`` can get number page requests received from the
4608 destination host during post-copy migration.
4612 * lxc: Don't forbid interfaces with type=direct
4614 Such interfaces are supported by lxc and should be allowed.
4616 * qemu: Fully clean up RNG devices on detach
4618 Some RNG device types, such as those using EGD, might need extra clean up
4619 on the host in addition to removing the guest-side device.
4622 v4.10.0 (2018-12-03)
4623 ====================
4627 * qemu: Add Hyper-V PV IPI and Enlightened VMCS support
4629 The QEMU driver now has support for Hyper-V PV IPI and Enlightened VMCS for
4630 Windows and Hyper-V guests.
4632 * qemu: Added support for PCI devices on S390
4634 PCI addresses can now include the new zpci element which contains uid
4635 (user-defined identifier) and fid (PCI function identifier) attributes and
4636 makes the corresponding devices usable by S390 guests.
4638 * Support changing IOThread polling parameters for a live guest
4640 Introduced virDomainSetIOThreadParams which allows dynamically setting the
4641 IOThread polling parameters used by QEMU to manage the thread polling
4642 interval and the algorithm for growth or shrink of the polling time. The
4643 values only affect a running guest with IOThreads. The guest's IOThread
4644 polling values can be viewed via the domain statistics.
4646 * Xen: Add support for PVH
4648 The libxl driver now supports Xen's PVH virtual machine type. PVH machines
4649 are enabled with the new "xenpvh" OS type, e.g.
4650 ``<os><type>xenpvh</type></os>``
4652 * qemu: Added support for CMT (Cache Monitoring Technology)
4654 Introduced cache monitoring using the ``monitor`` element in ``cachetune``
4655 for vCPU threads. Added interfaces to get and display the cache utilization
4656 statistics through the command 'virsh domstats' via the
4657 virConnectGetAllDomainStats API.
4659 * qemu: Add support for nested HV for pSeries guests
4661 Nested HV support makes it possible to run nested (L2) guests with minimal
4662 performance penalty when compared to regular (L1) guests on ppc64 hardware.
4666 * Xen: Handle soft reset shutdown event
4668 The pvops Linux kernel uses soft reset to handle the crash machine
4669 operation. The libxl driver now supports the soft reset shutdown event,
4670 allowing proper crash handling of pvops-based HVM domains.
4678 * util: Add cgroup v2 support
4680 cgroup v2 support has been implemented in libvirt, with both "unified" (v2
4681 only) and "hybrid" (v2 + v1) setups being usable; existing "legacy" (v1
4682 only) setups will keep working.
4684 * qemu: Add vfio AP support
4686 The QEMU driver now has support to passthrough adjunct processors into QEMU
4691 * rpc: Make 'genprotocol' output reproducible
4693 This is another step towards making libvirt builds fully reproducible.
4697 * security: Fix permissions for UNIX sockets
4699 Since 4.5.0, libvirt is using FD passing to hand sockets over to QEMU,
4700 which in theory removes the need for them to be accessible by the user
4701 under which the QEMU process is running; however, other processes such as
4702 vdsm need to access the sockets as well, which means adjusting permissions
4705 * cpu_map: Add Icelake model definitions
4707 These CPU models will be available in the upcoming 3.1.0 QEMU release.
4709 * util: Properly parse URIs with missing trailing slash
4711 Some storage URIs were not parsed correctly, in which case libvirt ended up
4712 emitting XML that it would then refuse to parse back.
4720 * Xen: Support PM Suspend and Wakeup
4722 The libxl driver now supports the virDomainPMSuspendForDuration and
4723 virDomainPMWakeup APIs.
4725 * **Removed features**
4727 * Xen: Drop support for Xen 4.4 and 4.5
4729 Xen 4.4 and 4.5 are no longer supported by the Xen community. Drop support
4730 for these older versions and require Xen >= 4.6.
4732 * nwfilter: Disallow binding creation in session mode
4734 Ensure that a filter binding creation is not attempted in session mode and
4735 generates a proper error message.
4739 * qemu: Retrieve guest hostname through QEMU Guest Agent command
4741 QEMU is now able to retrieve the guest hostname using a new QEMU-GA command
4742 called 'guest-get-host-name'. Virsh users can execute 'domhostname' for
4743 QEMU driver for domains configured to use the Guest Agent.
4745 * virsh: Implement vsh-table in virsh and virsh-admin
4747 The new API fixes problems with table-alignment, making the tables more
4748 readable and deals with unicode.
4752 * storage: Allow inputvol to be encrypted
4754 When creating a storage volume based on another volume, the base input
4755 volume is allowed to be encrypted.
4757 * virsh: Require explicit --domain for domxml-to-native
4759 The --domain option for domxml-to-native virsh command has always been
4760 documented as required, but commit v4.3.0-127-gd86531daf2 accidentally made
4763 * lxc_monitor: Avoid AB / BA lock race
4765 A deadlock situation could occur when autostarting a LXC domain 'guest' due
4766 to two threads attempting to take opposing locks while holding opposing
4767 locks (AB BA problem).
4775 * storage: add storage pool iscsi-direct
4777 Introduce a new storage pool backend that uses libiscsi instead of
4778 iscsiadm. It support basic pool operations: checkPool and refreshPool.
4780 * Add support for MBA (Memory Bandwidth Allocation technology)
4782 Domain vCPU threads can now have allocated some parts of host memory
4783 bandwidth by using the ``memorytune`` element in ``cputune`` .
4785 * qemu: Add support for RISC-V guests
4787 riscv32 and riscv64 guest architectures are now supported.
4791 * qemu: Add ccw support for vhost-vsock
4793 Support the vhost-vsock-ccw device on S390.
4795 * qemu: Make default machine type independent of QEMU
4797 We can't control whether or not QEMU will change its default machine type
4798 in the future, or whether downstream distributions will decide to compile
4799 out some machine types, so our only option to provide a predictable
4800 behavior is taking care of the default ourselves; management applications
4801 and users are encouraged to explicitly pick a machine type when creating
4804 * apparmor: Various improvements
4806 Rules have been added to deal with a number of scenarios that didn't work
4811 * esx: Truncate CPU model name
4813 Some CPU model names are too long to be stored into the corresponding
4814 property, and should be explicitly truncated to avoid unexpected behavior
4815 in users of the ``virNodeGetInfo()`` API such as ``virsh nodeinfo`` .
4817 * utils: Remove arbitrary limit on socket_id/core_id
4819 Both values were assumed to be smaller than 4096, but in fact they are
4820 entirely hardware-dependent and there have been reports of machines
4821 presenting much bigger values, preventing libvirt from working correctly;
4822 all such limits have now been removed.
4830 * qemu: Implement the HTM pSeries feature
4832 Users can now decide whether HTM (Hardware Transactional Memory) support
4833 should be available to the guest.
4835 * qemu: Enable VNC console for mediated devices
4837 Host devices now support a new attribute 'display' which can be used to
4838 turn on frame buffer rendering on a vgpu mediated device instead of on an
4839 emulated GPU, like QXL.
4843 * qemu: Introduce a new video model of type 'none'
4845 Introduce a new video model type that disables the automatic addition of a
4846 video device to domains with 'graphics' specified in their XML. This can be
4847 useful with GPU mediated devices which can serve as the only rendering
4848 devices within the guest.
4850 * virsh: Add --alias to attach-disk and attach-interface commands
4852 Add option --alias to set customized device alias name when using
4853 attach-disk or attach-interface commands.
4855 * virsh: Support usb and sata address to attach-disk
4857 Usb or sata address could be used when attach-disk with --address. For
4858 example, use usb address as usb:<bus>.<port>, use sata address as
4859 <controller>.<bus>.<unit>.
4867 * qemu: Provide TPM emulator support
4869 Support QEMU's TPM emulator based on swtpm. Each QEMU guest gets its own
4872 * bhyve: Support specifying guest CPU topology
4874 Bhyve's guest CPU topology could be specified using the ``<cpu><topology
4875 ../></cpu>`` element.
4877 * qemu: Add support for extended TSEG size
4879 Support specifying extended TSEG size for SMM in QEMU.
4881 * qemu: Add support for SEV guests
4883 SEV (Secure Encrypted Virtualization) is a feature available on AMD CPUs
4884 that encrypts the guest memory and makes it inaccessible even to the host
4887 * **Removed features**
4889 * Remove support for qcow/default encrypted volumes
4891 Disallow using a qcow encrypted volume for the guest and disallow creation
4892 of the qcow or default encrypted volume from the storage driver. Support
4893 for qcow encrypted volumes has been phasing out since QEMU 2.3 and by QEMU
4894 2.9 creation of a qcow encrypted volume via qemu-img required usage of
4895 secret objects, but that support was never added to libvirt.
4897 * Make GnuTLS mandatory
4899 Building without GnuTLS is no longer possible.
4901 * qemu: Remove allow_disk_format_probing configuration option
4903 The option represented a security risk when used with malicious disk
4904 images, so users were recommended against enabling it; with this release,
4905 it's been removed altogether.
4909 * capabilities: Provide info about host IOMMU support
4911 Capabilities XML now provide information about host IOMMU support.
4913 * virsh: Add --all to domblkinfo command
4915 Alter the ``domblkinfo`` command to add the option --all in order to
4916 display the size details of each domain block device from one command in a
4919 * qemu: Allow concurrent access to monitor and guest agent
4921 Historically libvirt prevented concurrent accesses to the qemu monitor and
4922 the guest agent. Therefore two independent calls (one querying the monitor
4923 and the other querying guest agent) would serialize which hurts
4924 performance. The code was reworked to allow two independent calls run at
4927 * qemu: Allow configuring the page size for HPT pSeries guests
4929 For HPT pSeries guests, the size of the host pages used to back guest
4930 memory and the usable guest page sizes are connected; the new setting can
4931 be used to request that a certain page size is available in the guest.
4933 * Add support to use an raw input volume for encryption
4935 It is now possible to provide a raw input volume as input for to generate a
4936 luks encrypted volume via either virsh vol-create-from or
4937 virStorageVolCreateXMLFrom.
4939 * qemu: Add support for vsock hot (un)plug and cold (un)plug
4941 * qemu: Add support for NBD over TLS
4943 NBD volumes can now be accessed securely.
4945 * qemu: Implement FD passing for Unix sockets
4947 Instead of having QEMU open the socket and then connecting to it, which is
4948 inherently racy, starting with QEMU 2.12 we can open the socket ourselves
4949 and pass it to QEMU, avoiding race conditions.
4951 * virsh: Introduce --nowait option for domstat command
4953 When this option is specified, virsh will try to fetch the guest stats but
4954 abort instead of stalling if they can't be retrieved right away.
4958 * qemu: Fix a potential libvirtd crash on VM reconnect
4960 Initialization of the driver worker pool needs to come before libvirtd
4961 trying to reconnect to all machines, since one of the QEMU processes migh
4962 have already emitted events which need to be handled prior to us getting to
4963 the worker pool initialization.
4965 * qemu: Fix domain resume after failed migration
4967 Recent versions of QEMU activate block devices before the guest CPU has
4968 been started, which makes it impossible to roll back a failed migration.
4969 Use the ``late-block-activate`` migration capability if supported to avoid
4972 * vmx: Permit guests to have an odd number of vCPUs
4974 An odd number of vCPUs greater than 1 was forbidden in the past, but
4975 current versions of ESXi have lifted that restriction.
4983 * bhyve: Support locking guest memory
4985 Bhyve's guest memory may be wired using the
4986 ``<memoryBacking><locked/></memoryBacking>`` element.
4988 * qemu: Provide VFIO channel I/O passthrough support
4990 Support passthrough devices that use channel I/O based mechanism in a QEMU
4993 * qemu: Add support for migration of VMs with non-shared storage over TLS
4995 It's now possible to use the VIR_MIGRATE_TLS flag together with
4996 VIR_MIGRATE_NON_SHARED_DISK. The connection is then secured using the TLS
4997 environment which is setup for the migration connection.
4999 * Add support for VM Generation ID
5001 The VM Generatation ID exposes a 128-bit, cryptographically random, integer
5002 value identifier, referred to as a Globally Unique Identifier (GUID) to the
5003 guest in order to notify the guest operating system when the virtual
5004 machine is executed with a different configuration. Add a new domain XML
5005 processing and a domain capabilities feature.
5007 * Introduce virDomainDetachDeviceAlias
5009 This new API enables users to detach device using only its alias.
5011 * Introduce new virConnectCompareHypervisorCPU and
5012 virConnectBaselineHypervisorCPU APIs
5014 Unlike the old virConnectCompareCPU and virConnectBaselineCPU APIs, both
5015 new APIs consider capabilities of a specific hypervisor.
5017 * Introduce SCSI persistent reservations support
5019 The QEMU driver gained support for qemu-pr-helper which enables guests to
5020 issue SCSI commands for persistent reservation.
5022 * qemu: Implement multiple screen support for virDomainScreenshot
5024 While the virDomainScreenshot API supported multihead video cards, the
5025 implementation was missing. But now that QEMU implemented it libvirt has
5028 * qemu: add support for vhost-vsock-device
5030 A new vsock device was introduced, allowing communication between the guest
5031 and the host via the AF_VSOCK family.
5035 * qemu: Add support for OpenGL rendering with SDL
5037 Domains using SDL as a graphics backend will now be able to use OpenGL
5038 accelerated rendering.
5040 * qemu: Add support for 'output' audio codec
5042 Support QEMU's 'hda-output' codec advertising only a line-out for ich6 and
5045 * virsh: Enhance event name completion
5047 Implement event name completion for some commands (e.g. event,
5048 secret-event, pool-event and nodedev-event)
5056 * qemu: Add support for the pcie-to-pci-bridge controller
5058 Pure PCIe guests such as x86_64/q35 and aarch64/virt will now add this
5059 controller when traditional PCI devices are in use.
5061 * Xen: Support setting CPU features for host-passthrough model
5063 The CPU model presented to Xen HVM domains is equivalent to libvirt's
5064 host-passthrough model, although individual features can be enabled and
5065 disabled via the cpuid setting. The libvirt libxl driver now supports
5066 enabling and disabling individual features of the host-passthrough CPU
5069 * **Removed features**
5071 * Xen: Drop the legacy xend-based driver
5073 The xm/xend toolstack was deprecated in Xen 4.2 and removed from the Xen
5074 sources in the 4.5 development cycle. The libvirt driver based on xend is
5075 now removed from the libvirt sources.
5079 * qemu: Support hot plug and hot unplug of mediated devices
5081 Libvirt now allows mediated devices to be hot plugged and hot unplugged
5082 from a guest rather than reporting an error that this isn't supported. In
5083 fact, kernel has been supporting this since 4.10.
5087 * Improve handling of device mapper targets
5089 When starting a domain with a disk backed by a device mapper volume libvirt
5090 also needs to allow the storage backing the device mapper in CGroups. In
5091 the past kernel did not care, but starting from 4.16 CGroups are consulted
5092 on each access to the device mapper target.
5100 * Support building with Python 3
5102 Python is required to build libvirt, and up until now only Python 2 could
5103 be used as an interpreter. All scripts used during build have now been made
5104 compatible with Python 3, which means both major releases of the language
5105 are fully supported.
5107 * qemu: Provide ccw address support for graphics and input devices
5109 Support the virtio-gpu-ccw device as a video device and virtio-{keyboard,
5110 mouse, tablet}-ccw devices as input devices on S390.
5114 * qemu: Add logging of guest crash information on S390
5116 On S390, when the guest crashes and QEMU exposes the guest crash
5117 information, log the relevant data to the domain log file.
5119 * qemu: use arp table of host to get the IP address of guests
5121 Find IP address of a VM by arp table on hosts. If someone customizing IP
5122 address inside VM, it will be helpful.
5124 * Xen: Remove hard-coded scheduler weight
5126 The libxl driver was accidentally hard-coding the per-domain scheduler
5127 weight to 1000, silently ignoring any user-provided ``<shares>`` in
5128 ``<cputune>`` . The driver now honors ``<shares>`` , and defers setting a
5129 default value to Xen. Note that the Xen default is 256, so any domains
5130 started after this improvement will have one fourth the shares of
5131 previously started domains. If all domains must have equal CPU shares,
5132 administrators must manually set the weight of previously started domains
5133 to 256, or restart them.
5137 * qemu: TLS migration now enforces use of TLS for the NBD connection
5139 When the VIR_MIGRATE_TLS flag was used with the migration API libvirt did
5140 not ensure that the NBD connection was using TLS as well. The code now
5141 rejects such migration as the TLS transport for NBD is not ready yet, but
5142 prevents a false sense of security that TLS would be used. The support TLS
5143 for NBD will be added soon.
5151 * Added support for CAT (Cache allocation Technology)
5153 Domain vCPU threads can now have allocated some parts of host cache using
5154 the ``cachetune`` element in ``cputune`` .
5156 * Allow opening secondary drivers
5158 Up until now it was possible to connect to only hypervisor drivers (e.g.
5159 qemu:///system, lxc:///, vbox:///system, and so on). The internal drivers
5160 (like network driver, node device driver, etc.) were hidden from users and
5161 users could use them only indirectly. Starting with this release new
5162 connection URIs are accepted. For instance network:///system,
5163 storage:///system and so on.
5165 * virtlogd, virtlockd: Add support for admin protocol
5167 These two daemons now support admin protocol through which some admin info
5168 can be gathered or some configuration tweaked on the fly.
5172 * virsh: Enhance bash completion
5174 Implement more bash completions so that basic libvirt objects can be
5175 auto-completed (e.g. networks, interfaces, NWFilters, and so on).
5177 * qemu: Use VIR_ERR_DEVICE_MISSING for various hotplug/detach messages
5179 * qemu: Allow showing the dump progress for memory only dump
5181 Alter the QEMU dump-guest-memory command processing to check for and allow
5182 asynchronous completion which then allows for the virsh dump --memory-only
5183 --verbose command to display percent completion data.
5185 * conf: add support for setting Chassis SMBIOS data fields
5187 * libxl: add support for setting clock offset and adjustment
5189 * Make port allocator global
5191 Up until now each driver had their own port allocator module. This meant
5192 that info on port usage was not shared. Starting with this release, the
5193 port allocator module is made global and therefore drivers allocate ports
5196 * Fixed some compiler warnings that appear with GCC 8
5200 * qemu: Check for unsafe migration more thoroughly
5202 If a domain disk is stored on local filesystem (e.g. ext4) but is not being
5203 migrated it is very likely that domain is not able to run on destination.
5204 Regardless of share/cache mode.
5206 * qemu: Fix updating device with boot order
5208 Starting with 3.7.0 release updating any device with boot order would fail
5209 with 'boot order X is already used by another device' while in fact it was
5210 the very same device.
5212 * virlog: determine the hostname on startup CVE-2018-6764
5214 At later point it might not be possible or even safe to use getaddrinfo().
5215 It can in turn result in a load of NSS module which can even be loaded from
5216 unsage guest filesystem and thus escape the confinment of its container.
5218 * qemu: Rework vCPU statistics fetching
5220 Fetching vCPU statistics was very expensive because it lead to waking up
5221 vCPU threads in QEMU and thus it degraded performance. The code was
5222 reworked so that fetching statistics does not wake up halted vCPUs.
5224 * qemu: unlink memory backing file on domain shutdown
5226 Depending on the filesystem where domain memory is stored, some files might
5227 have been left behind. This is not a problem on hugetlbfs, but it is a
5228 problem on regular filesystems like ext4.
5230 * qemu: Fix shutting down domains in parallel
5232 If multiple domains were being shut down in parallel, libvirtd might have
5235 * nodedev: Update PCI mdev capabilities dynamically
5237 PCI devices may have other nested capabilities, like SRIOV and mdev which
5238 depend on the device being plugged into the native vendor driver. However,
5239 in case such a device is directly assigned to a guest using VFIO driver,
5240 the device will naturally lose these capabilities and libvirt needs to
5249 * tools: Provide bash completion support
5251 Both ``virsh`` and ``virt-admin`` now implement basic bash completion
5254 * qemu: Refresh capabilities on host microcode update
5256 A microcode update can cause the CPUID bits to change; therefore, the
5257 capabilities cache should be rebuilt when such an update is detected on the
5260 * lxc: Set hostname based on container name
5264 * CPU frequency reporting improvements
5266 The CPU frequency will now be reported by ``virsh nodeinfo`` and other
5267 tools for s390 hosts; at the same time; CPU frequency has been disabled on
5268 aarch64 hosts because there's no way to detect it reliably.
5270 * libxl: Mark domain0 as persistent
5272 * Xen: Add support for multiple IP addresses on interface devices
5274 * qemu: Add support for hot unplugging redirdev device
5278 * qemu: Enforce vCPU hotplug granularity constraints
5280 QEMU 2.7 and newer don't allow guests to start unless the initial vCPUs
5281 count is a multiple of the vCPU hotplug granularity, so validate it and
5282 report an error if needed.
5285 v3.10.0 (2017-12-04)
5286 ====================
5290 * conf: Support defining distances between virtual NUMA cells
5292 A NUMA hardware architecture supports the notion of distances between NUMA
5293 cells. This can now be specified using the ``<distances>`` element within
5294 the NUMA cell configuration. Drivers which support this include Xen and
5297 * Xen: Support defining vNUMA topology
5299 Xen now supports defining a virtual NUMA topology for VMs, including
5300 specifying distances between NUMA cells.
5302 * qemu: Add the ability to configure HPT resizing for pSeries guests
5304 The user can now decide whether HPT (Hash Page Table) resizing should be
5305 enabled, disabled or required instead of leaving it up to hypervisor
5306 defaults and negotiation between the guest and the host.
5308 * qemu: Add vmcoreinfo feature
5310 Starting with QEMU 2.11, the guest can save kernel debug details when this
5311 feature is enabled and the kernel supports it. It is useful to process
5312 kernel dump with KASLR enabled, and also provides various kernel details to
5315 * conf: Move the auth and encryption definitions to disk source
5317 Allow parsing and formatting of the ``auth`` and ``encryption``
5318 sub-elements to be a child of the ``source`` element. This will allow
5319 adding an ``auth`` sub-element to a ``backingStore`` or ``mirror`` elements
5320 as a means to track specific authentication and/or encryption needs.
5324 * vbox: Add VirtualBox 5.2 support
5326 * vbox: Add support for configuring storage controllers
5328 The VirtualBox driver now supports the ``<controller>`` element in the
5329 domain XML for configuring storage controllers in VBOX VMs. Additionally,
5330 libvirt's domain XML schema was updated to allow optional ``model``
5331 attribute for ``<controller type='ide'>`` which is used by the VBOX driver
5332 to set the IDE controller model to be one of 'piix4', 'piix4' (default), or
5333 'ich6'. Finally, with this change ``dumpxml`` generates ``<controller>``
5334 elements that correspond to current VBOX VM storage controller
5337 * vbox: Add support for attaching empty removable disks
5339 The VirutalBox driver now supports adding CD-ROM and floppy disk devices
5340 that do not have the disk source specified. Previously such devices were
5343 * vbox: Add support for attaching SAS storage controllers
5345 In VirtualBox, SCSI and SAS are distinct controller types whereas libvirt
5346 does not make such distinction. Therefore, the VBOX driver was updated to
5347 allow attaching SAS controllers via ``<controller type='scsi'
5348 model='lsisas1068'>`` element. If there are both SCSI and SAS controllers
5349 present in the VBOX VM, the domain XML can associate the disk device using
5350 the ``<address>`` element with the ``controller`` attribute, and
5351 optionally, set the port via ``unit`` attribute.
5353 * qemu: Generate predictable paths for qemu memory backends
5355 In some cases management applications need to know paths passed to
5356 memory-backend-file objects upfront. Libvirt now generates predictable
5357 paths so applications can prepare the files if they need to do so.
5359 * Shareable disks work properly with recent qemu
5361 Recent qemu versions added image locking to avoid potential corruption of
5362 disk images. This broke shareable disks with libvirt since the feature was
5363 turned on by default in qemu. Libvirt now enables sharing of those disks in
5364 qemu so that the image locking is not applied in that case. Additionally
5365 libvirt now checks that shareable disks have supported format (raw) to
5366 avoid metadata corruption.
5368 * Improve serial console behavior on non-x86 architectures
5370 ppc64, aarch64 and s390x guests were treating the <serial> and <console>
5371 elements differently from x86, in some cases presenting misleading
5372 information to the user. The behavior is now consistent across all
5373 architectures and the information reported is always accurate.
5377 * vbox: Do not ignore failures to attach disk devices when defining
5379 The ``define`` now fails and reports an error if any of the ``controller``
5380 or ``disk`` devices specified in the domain XML fail to attach to the
5383 * vbox: Fix dumpxml to always output disk devices
5385 The VirtualBox driver was ignoring any disk devices in ``dumpxml`` output
5386 if there was a SAS storage controller attached to the VM.
5388 * vbox: Fix dumpxml to always generate valid domain XML
5390 When a VirtualBox VM has multiple disks attached, each to a different
5391 storage controller that uses 'sd' prefix for block device names e.g. one
5392 disk attached to SATA and one to SCSI controller, it no longer generates
5393 XML where both would have 'sda' device name assigned. Instead it properly
5394 assigns 'sda' and 'sdb' to those disks in the order of appearance.
5396 * Securely pass iSCSI authentication data
5398 Rather than supplying the authentication data as part of the iSCSI URL for
5399 a disk or host device, utilize the encrypted secret object to securely pass
5400 the authentication data.
5408 * Add capability to allow hot (un)plug of a domain watchdog device
5410 * Allow users to set device aliases
5412 Users can set aliases to domain devices and thus identify them easily.
5414 * qemu: Support multiqueue for virtio-blk
5416 Multiqueue support for ``virtio-blk`` has been available in QEMU ever since
5417 2.7.0, and now libvirt guests can enable it.
5419 * Add virDomainSetLifecycleAction API
5421 Provided a new API to allow dynamic guest lifecycle control for guest
5422 reactions to poweroff, restart, or crash type events related to the domain
5423 XML ``on_poweroff`` , ``on_reboot`` , and ``on_crash`` elements. The
5424 ``virsh set-lifecycle-action`` command was created to control the actions.
5426 * qemu: Allow cold(un)plugging and hot(un)plugging input devices
5428 * net: Implement QoS for vhostuser
5432 * Allow a logical volume to be create using LUKS
5434 A logical volume may be created using an ``encryption`` element using
5435 "luks" format. This does require a previously created ``secret`` to store
5436 the passphrase used to encrypt the volume Adding the volume to a domain can
5437 then either provide the secret or allow the consumer in the guest to
5438 provide the passphrase in order to decrypt the volume.
5440 * net: Ignore auto-generated MAC address when detaching an interface
5442 If the MAC address has not been specified by the user, libvirt will try and
5443 fill in the gaps by generating one; however, for some error paths that led
5444 to some confusing error messages, so when an auto-generated MAC address is
5445 specified the error message will not include the auto-generated MAC.
5447 * net: Enable MAC address lookup for virDomainInterfaceStats
5449 * apparmor: Several improvements
5451 Changes include permitting access to data about USB devices and ``dnsmasq``
5452 instances, allowing spaces in guest names and many more.
5454 * cpu: Use CPU information obtained from QEMU when possible
5456 Recent QEMU versions can expose information about which CPU models are
5457 available and usable on the host; libvirt will now make use of such
5458 information whenever possible.
5460 * hyperv: Various improvements
5462 The error reported when clients can't connect to Hyper-V has been made more
5463 descriptive, and memory limits for guests are now mapped to more
5464 appropriate libvirt equivalents.
5466 * qemu: Report QEMU error on failed migration
5468 Instead of reporting a generic error, ask QEMU for a more detailed and thus
5469 hopefully more helpful one.
5471 * vbox: Implement autoport for RDP
5473 libvirt will now obtain the (dynamically allocated) RDP port number from
5474 VirtualBox itself, avoiding conflicts between multiple guests wanting to
5475 use RDP at the same time.
5477 * qemu: Allow rotation of small logs
5479 On a host where numerous unique instances are executed per day, it's quite
5480 possible that, even though each of the single log files are fairly small,
5481 collectively the quantity and volume may add tens of thousands of log files
5482 to the ``/var/log/libvirt/qemu/`` directory. Removing the constraints that
5483 log have to be bigger than 100 KiB before they can be rotated solves the
5488 * Fix swapped interface statistics and QoS
5490 Due to internal implementation, reported statistics for some types of
5491 interfaces were swapped (RX appeared in TX and vice versa). Similarly, QoS
5492 was set in reversed way.
5494 * Properly resize local LUKS encrypted volume
5496 Resizing of a local LUKS encrypted volume will now use qemu-img to resize
5497 the volume. This will require configuring a secret for the LUKS encrypted
5500 * qemu: Reserve PCI addresses for implicit i440fx devices
5502 Failing to do so causes the addresses to be considered usable by libvirt,
5503 which means they could be assigned to more than one device resulting in the
5504 guest failing to start.
5506 * spec: Restart libvirtd only at the end of the upgrade process
5508 Use ``%posttrans`` to make sure ``libvirtd`` is not restarted before all
5509 other components, such as the library itself and storage / hypervisor
5510 drivers, have already been upgraded.
5514 * qemu: Ensure TLS clients always verify the server certificate
5516 While it's reasonable to turn off client certificate validation, as setting
5517 it up can be non-trivial, clients should always verify the server
5518 certificate to avoid MITM attacks. However, libvirt was using the same knob
5519 to control both checks, leading to CVE-2017-1000256 / LSN-2017-0002.
5527 * qemu: Added support for cold-(un)plug of watchdog devices
5529 * qemu: Added support for setting IP address os usernet interfaces
5531 * qemu: Added support for Veritas Hyperscale (VxHS) block devices
5533 * storage: Added new events for pool-build and pool-delete
5537 * qemu: Set DAC permissions properly for spice rendernode
5539 When a ``rendernode`` path is set for SPICE GL on ``qemu:///system`` , we
5540 now correctly set DAC permissions on the device at VM startup. This is the
5541 last remaining hurdle to let SPICE GL work for ``qemu:///system`` without
5542 any external host changes.
5544 * nodedev: Add switchdev offload query to NIC capabilities
5546 Allow querying the NIC interface capabilities for the availability of
5547 switchdev offloading (also known as kernel-forward-plane-offload).
5549 * New CPU models for AMD and Intel
5551 AMD EPYC and Intel Skylake-Server CPU models were added together with their
5554 * Improve long waiting when saving a domain
5556 While waiting for a write to disk to be finished, e.g. during save, even
5557 simple operations like ``virsh list`` would be blocking due to domain lock.
5558 This is now resolved by unlocking the domain in places where it is not
5563 * Proper units are now used in virsh manpage for dom(mem)stats
5565 Previously the documentation used multiples of 1000, but now it is fixed to
5566 use multiples of 1024.
5568 * qemu: Fix error reporting when disk attachment fails
5570 There was a possibility for the actual error to be overridden or cleared
5571 during the rollback.
5573 * qemu: Fix assignment of graphics ports after daemon restart
5575 This could be seen with newer kernels that have bug regarding SO_REUSEADDR.
5576 After libvirtd was restarted it could assign already used address to new
5577 guests which would make them fail to start. This is fixed by marking used
5578 ports unavailable when reconnecting to running QEMU domains.
5580 * Fix message decoding which was causing a very strange bug
5582 When parsing an RPC message with file descriptors was interrupted and had
5583 to restart, the offset of the payload was calculated badly causing strange
5584 issues like not being able to find a domain that was not requested.
5592 * qemu: Add managedsave-edit commands
5594 Using managedsave-dumpxml, managedsave-define and managedsave-edit
5595 commands, now we can dump and edit the XML configuration of domain which
5596 has managedsave image.
5598 * qemu: Add migrate-getmaxdowntime command
5600 Currently, the maximum tolerable downtime for a domain being migrated is
5601 write-only from libvirt, via migrate-setmaxdowntime. This implements a
5602 complementary migrate-getmaxdowntime command
5604 * bhyve: Support autoport for VNC ports
5606 It's no longer necessary to explicitly specify VNC port for the bhyve
5607 guests. With the autoport feature it will be allocated automatically.
5608 Please refer to the bhyve driver documentation for examples.
5610 * qemu: Added support for setting heads of virtio GPU
5612 * qemu: Added support to configure reconnect timeout for chardev devices
5614 When you have a TCP or UNIX chardev device and it's connected somewhere you
5615 can configure reconnect timeout if the connection is closed.
5619 * qemu: Report a clear error when dropping a VM during startup
5621 "Failed to load config for domain 'DOMNAME'" is now reported if a VM config
5622 can't be parsed for some reason, and thus provides a clear indication for
5625 * apparmor: Update for QEMU 2.10 compatibility
5627 Starting with QEMU 2.10, disk images and NVRAM files get automatically
5628 locked to prevent them from being corrupted; however, file locking needs to
5629 be explicitly allowed through ``virt-aa-helper`` or AppArmor will reject
5630 the requests and the guest will not be able to run.
5632 * virsh: List Unix sockets in 'domdisplay' output
5634 VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets as
5635 connection endpoints, but such a configuration was not handled correctly by
5636 ``virsh domdisplay`` , causing the respective endpoints to be missing from
5639 * qemu: Don't check whether offline migration is safe
5641 Since offline migration only copies the guest definition to the destination
5642 host, data corruption is not a concern and the operation can always be
5645 * virt-host-validate: Fix IOMMU detection on ppc64
5649 * qemu: Better support for international domain names (with wide characters)
5651 There were some issues with multi-byte domains getting lost on daemon
5652 restart due to truncation, so the code now handles multi-byte names a bit
5655 * qemu: Support long domain names with namespaces
5657 Domains with extremely long names would fail to start due to temporary
5658 namespace paths being created with the whole name. The path is now
5659 generated with shortened name instead.
5661 * qemu: Tolerate missing emulator binary during libvirtd restart
5663 For some time libvirt required qemu capabilities being present when parsing
5664 VM configs during startup. As a side effect VM configs would fail to parse
5665 and thus vanish, if the emulator binary would be uninstalled or broken.
5666 Libvirt now tolerates when capabilities are missing during startup.
5668 * qemu: Prevent pSeries guests from disappearing in some situations
5670 pSeries guest would disappear if any of the host devices they were
5671 configured to use was not available during libvirtd startup, which could
5672 easily happen for SR-IOV Virtual Functions. This scenario is now handled
5675 * qemu: Honor <on_reboot/> setting
5677 The setting was accepted by the parser, but not actually implemented.
5679 * Fix --verbose option for all daemons
5681 Since v3.0.0, the option had been ignored by all libvirt daemons (
5682 ``libvirtd`` , ``virtlogd`` and ``virtlockd`` ); it's now working as
5683 intended once again.
5691 * hyperv: Implement virDomainSetMemory and virDomainSendKey APIs
5693 * qemu: Support multiple PHBs for pSeries guests
5695 pSeries guests can now have multiple PHBs (PCI Host Bridges), which show up
5696 as separate PCI domains in the guest. To create additional PHBs, simply add
5697 PCI controllers with model ``pci-root`` to the guest configuration.
5699 * qemu: Isolate hostdevs on pSeries guests
5701 To enable better error reporting and recovery, unrelated hostdevs will now
5702 be automatically isolated on pSeries guests by placing them on separate
5703 PHBs (PCI Host Bridges).
5707 * qemu: platform serial devices can now use chardev features
5709 QEMU VMs that depend on platform serial devices can now use QEMU's
5710 ``-chardev`` option, which enables access to advanced features like log
5711 file configuration. This applies to the default serial devices for arm,
5712 aarch64, and some ppc configurations.
5714 * Require use of GCC 4.4 or Clang compilers
5716 We only ever test libvirt with GCC or Clang (which provides a GCC
5717 compatible compilation environment). Between them, these compilers cover
5718 every supported operating system platform, including Windows.
5720 * qemu: shared disks with directsync cache should be safe for migration
5722 At present shared disks can be migrated with either readonly or
5723 ``cache=none`` . But ``cache=directsync`` should be safe for migration,
5724 because both ``cache=directsync`` and ``cache=none`` don't use the host
5725 page cache, and ``cache=direct`` write through qemu block layer cache.
5727 * Handle hotplug change on VLAN configuration using OVS
5729 Libvirt was taught to handle VLAN change for running OVS interface.
5733 * qemu: Use vCPU 'node-id' property and pass it back to qemu
5735 vCPU properties gathered from query-hotpluggable-cpus need to be passed
5736 back to QEMU. As QEMU did not use the node-id property until now and
5737 libvirt forgot to pass it back properly (it was parsed but not passed
5738 around) we did not honor this.
5740 * Miscellaneous stream fixes
5742 After introducing sparse stream features there were still some known bugs
5743 left. Those are fixed in this release.
5745 * qemu: Miscellaneous domain NS fixes
5747 Libvirt starts qemu domains in separate Linux namespaces for a while now.
5748 However, there were still some bugs lingering. For instance libvirt did not
5749 know how to handle file based bind mounts.
5751 * Various CPU driver improvements
5753 There were some minor bugs when using 'host-model' CPU.
5761 * qemu: Add support for loadparm for a boot device
5763 Add an optional boot parameter 'loadparm' for a boot device. Loadparm is an
5764 8 byte parameter that, when present, is queried by S390 guests via sclp or
5765 diag 308. Linux guests on S390 use it to select a boot entry.
5767 * Support controlling how video devices are exposed to the bhyve guests
5769 The ``vgaconf`` attribute was added to ``video`` 's ``driver`` element.
5770 Possible values are: ``on`` , ``off`` , and ``io`` . It controls the way
5771 how bhyve exposes video devices to its guests; various guest OSes might
5772 require different settings to boot properly.
5774 * qemu: Add support for live updates of coalesce settings
5776 Users can now use ``virsh update-device`` to change the coalesce settings
5777 of an interfaces while the domain is running.
5779 * qemu: Allow VirtIO devices to use vIOMMU
5781 It is now possible to turn on IOTBL for the vIOMMU and have VirtIO devices
5782 use it, provided they have been configured appropriately.
5786 * qemu: block copy job can be used with persistent domains
5788 Until now it was not possible to use block copy with persistent VMs. In use
5789 cases where it's not required to recover the job after VM shutdown, it's
5790 possible to specify VIR_DOMAIN_BLOCK_COPY_TRANSIENT_JOB flag to start the
5793 * JSON pseudo-protocol backing store parser supports new format of qemu 2.9
5795 QEMU 2.9 modified a few structures corresponding to the JSON format of
5796 specifying a backing store for a disk image. Libvirt now implements the new
5799 * Capabilities now include info about host's CAT settings
5801 Various information about resource control from the host is gathered and
5802 presented in capabilities if available.
5804 * apparmor: Several improvements
5806 Allow access to Ceph config, EFI firmware on both x86_64 and aarch64,
5807 device tree on ppc64 and more.
5809 * qemu: Support host-model on POWER9 machines
5813 * qemu: snapshot: retrieve image metadata from user provided files
5815 Disk images of an external snapshot created with
5816 VIR_DOMAIN_SNAPSHOT_CREATE_REUSE_EXT flag specified would not be scanned
5817 for metadata after recent changes. The metadata is necessary to allow
5818 keeping relative paths between images when doing a block-commit.
5820 * Parse decimal numbers in a locale-independent way
5822 Some locales, such as ``de_DE`` and ``pt_BR`` , use comma rather than dot
5823 to separate the integer part from the fractional part of a decimal number;
5824 however, several data sources such as the kernel use a locale-independent
5825 representation and need to be treated accordingly.
5827 * Support compilation with newer compiler and libc versions
5829 Several fixes have been included to make compilation with Clang 4.0.0, GCC
5830 7.1 and glibc >= 2.25.90 possible.
5832 * qemu: Query name for vhost-user interfaces at runtime
5834 This makes it possible to use ``virsh`` subcommands such as ``domiflist``
5835 and ``domifstat`` on vhost-user interfaces.
5837 * qemu: Set MTU for hotplugged interfaces correctly
5839 When hotplugging a network interface, the MTU was only set on the guest
5840 side. Set it on the host side as well.
5842 * qemu: Forbid updating MTU for interfaces of running guests
5844 The MTU setting can't be modified while the guest is running, so any
5845 attempt to alter it at runtime will now result in an error rather than
5846 being silently ignored.
5848 * qemu: Fix specifying QXL heads with older QEMU releases
5850 Specifying the number of QXL heads was not working correctly for QEMU
5851 releases older than 1.6.
5853 * qemu: Fix migration to older libvirt/QEMU versions
5855 When the guest is started, libvirt updates the CPU definition to reflect
5856 the actual CPU features to enforce ABI. We need to send original and
5857 updated CPU definition in order to support migration to older libvirt/QEMU
5858 versions. Only the updated CPU definition was sent to destination.
5866 * Improved streams to efficiently transfer sparseness
5868 New extension to virStream was implemented so that virStorageVolDownload
5869 and virStorageVolUpload can preserve file sparseness.
5871 * I/O APIC type can be specified for QEMU/KVM domains
5873 The ``ioapic`` tag was added to domain ``features`` , so the type of the
5874 I/O APIC can now be specified (e.g. putting it in userspace for KVM
5877 * The reason for VM shutdown is reported, if known
5879 QEMU 2.10 will be able to report the reason for shutting down (whether that
5880 was caused by the guest or not), and libvirt is prepared for that and
5881 reports that information in its shutdown event as well, if it is known.
5885 * Repository now has new README.md file
5887 The new file uses markdown syntax, so it looks better on github and
5888 possibly other web pages, but it has also more useful information. The old
5889 README is now symlink to the new file.
5891 * qemu: Use GICv2 by default for aarch64/virt TCG guests
5893 The emulated GICv3 has some limitations that make it unusable as a default;
5894 use GICv2 until they're sorted out. This change makes it once again
5895 possible to run aarch64/virt guests on a x86_64 host without having to
5896 tweak their configuration.
5898 * Additional capabilities for the node_device module
5900 Introduce two new capabilities to the node_device module. The first is for
5901 CCW devices, most common on the S390 architecture. The second is for fibre
5902 channel-backed SCSI devices and exposes the fc_remote_port sub-capability
5903 to SCSI target devices.
5905 * Node devices now report Mediated device capabilities
5907 Endpoint devices support new ``mdev`` capability and their parents now
5908 report the supported types in new ``mdev_types`` capability.
5910 * Capabilities now report information about host caches
5912 If supported in the kernel, host capabilities will now list L3 caches. The
5913 code for other levels was added as well, but only L3 caches are reported
5916 * POWER9 CPU model was added
5918 It is now properly reported in host capabilities.
5920 * libxl: NUMA sibling distances are now reported in host capabilities
5922 * VMDK version 3 files are now properly detected
5924 * Interrupt remapping and Extended interrupt mode for IOMMU devices
5926 These two new features can now be controlled with new ``<driver
5927 intremap='on/off' eim='on/off'/>`` tag for ``iommu`` devices.
5929 * Graphics in libxl domains now have default addresses
5931 Even though there were default addresses before this change, they were not
5932 saved in the XML. It is now possible to see and control the listen
5935 * Default USB controllers are now added for devices in libxl domains
5937 Even though they were added automatically when USB device was attached,
5938 they could've been missing in some other cases. The logic is now fixed so
5939 there are always USB controllers, even if there was none of them in the
5942 * Limits for RPC messages were increased
5944 Hitting the RPC limits we have is easier every day, so they were increased
5945 once again and some guessing logic was improved as well. It is now possible
5946 to get more stats than ever using the ``virConnectGetAllDomainStats()``
5947 call and push through even bigger requests and replies for all APIs.
5951 * qemu: Create memory_backing_dir on startup
5953 Libvirt's policy is that directories are created on startup if they don't
5954 exist. We've missed this one.
5956 * PCIe 4.0 cards now report proper link speeds
5958 It could happen that the link speed for PCIe devices was not properly
5959 reported or the nodedev-dumpxml just failed. That was due to mistake in the
5960 field width, but should now work properly.
5962 * qemu: Do not report errors on shutdown
5964 For some users, in some rare cases, it could happen that there was an error
5965 message "internal error: End of file from qemu monitor" in the logs even
5966 though no problem happened. The detection of these false positives was
5967 improved and such errors should not show any more.
5969 * User-specified UNIX socket paths for virtio channels should not be reset
5971 It could happen, in some cases, that libvirt would mistake a user-specified
5972 path for its own generated one and thus remove it from the XML. The
5973 detection of such addresses was improved now.
5975 * Fix address reservation during RNG hot-plug
5977 When error occurred in a specific point in time during the hot-plug of an
5978 RNG device, it could happen that an address was released even though
5979 another device was already using it, making it possible to hot-plug another
5980 device with that address, effectively having duplicated addresses in the
5989 * net: Add support for coalesce settings
5991 Enabling data batching through these settings can improve network
5992 performance for guests.
5994 * qemu: Add support for guest CPU cache specification
5996 This features allows fine-grained control of the cache behavior of the
5999 * qemu: Add support for the qemu-xhci USB controller
6003 * hyperv: Support Hyper-V 2012 and newer
6005 Starting with Hyper-V 2012 the API has changed causing the existing driver
6006 to be unable to send and process requests properly. This has been resolved
6007 by adding abstractions to handle the differences and ease handling such
6008 breaks if they happen in the future.
6010 * libxl: Add support for nested HVM domains
6012 Xen has supported nested HVM domains since version 4.4. The libvirt libxl
6013 driver now supports nested HVM domains by specifying the host-passthrough
6014 CPU mode when defining a domain.
6016 * qemu: Implement ACPI support for aarch64 guests
6018 Up until this point, ACPI support was only advertised for x86_64 guests and
6019 disabling it for aarch64 guests was not possible at all.
6021 * vz: Add support for changing the number of vCPUs
6023 * qemu: Automatically choose the best USB controller for guests
6025 The recently introduced qemu-xhci USB controller is the best choice for
6026 both ppc64 and aarch64 guests, so use it by default on those architectures
6029 * daemon: Increase default task limit for libvirtd
6031 The default number of tasks for the pids cgroup controller is 512, which
6032 libvirtd can quickly bump into when starting lots of guests. Raise the
6033 limit to a more reasonable 32768.
6035 * docs: Include man pages describing key code names and values
6037 * virsh: Report initialization errors
6039 Sometimes virsh might be unable to start: when that happens, report useful
6040 diagnostics instead of failing silently.
6044 * nss: Don't require a network restart for libvirt_guest
6046 Previously, the libvirt_guest NSS module would only work properly after the
6047 corresponding network had been restarted; now newly started guests will be
6048 reported correctly right away.
6050 * storage: Remove unavailable transient pools after restart
6052 Solve an issue where transient storage pools would be stuck in an
6053 unmanageable state if the source disappeared and libvirtd was subsequently
6056 * storage: Fix capacity value for LUKS encrypted volumes
6058 The 'capacity' value (e.g. guest logical size) for a LUKS volume is smaller
6059 than the 'physical' value of the file in the file system, so we need to
6062 * qemu: Fix regression when hyperv/vendor_id feature is used
6064 Guests using the feature would not be started at all; it is now possible to
6065 start them as expected.
6067 * qemu: Do not crash on USB address with no port and invalid bus
6069 * crypto: Always pad data before encrypting it
6071 If this step is not performed, when the data length matches the chunk size
6072 the decryption routines will misinterpret the last byte of data as the
6073 padding length and fail to decode it correctly.
6081 * The virt-host-validate tool now supports bhyve hypervisor
6083 * Introduce NVDIMM memory model
6085 NVDIMM is new type of memory introduced into QEMU 2.6. The idea is that we
6086 have a non-volatile memory module that keeps the data persistent across
6087 domain reboots and offers much faster data accesses. However, due to a bug
6088 in QEMU, this feature is not enabled for QEMUs older than 2.9.0.
6090 * qemu: Introduce support for generic PCIe Root Ports
6092 For new controllers, a generic device (pcie-root-port) will be used by
6093 default instead of the Intel-specific device (ioh3420), provided the QEMU
6096 * qemu: Add support for checking guest CPU ABI compatibility
6098 When migrating a domain to a different host, restoring a domain from a file
6099 or reverting a snapshot libvirt will make sure the guest CPU QEMU presents
6100 to the guest OS exactly matches the one provided on the source host (or
6101 before the domain's state was saved). This enhanced check may also be
6102 requested when starting a new domain to ensure the virtual CPU exactly
6103 matches the one specified in the XML.
6105 * qemu: Add support to migrate using TLS
6107 Add the ability to migrate QEMU guests using TLS via a new flag
6108 VIR_MIGRATE_TLS or virsh migrate '--tls' option. Requires using at least
6109 QEMU 2.9.0 in order to work properly.
6111 * qemu: add mediated devices framework support
6113 Recent kernel version introduced new mediated device framework, so provide
6114 an initial support of this framework for libvirt, mainly by introducing a
6115 new host device type in the XML.
6117 * qemu: Add support for setting TSC frequency
6119 Setting TSC frequency is required to enable migration for domains with
6120 'invtsc' CPU feature turned on.
6122 * Add support for block device threshold event
6124 When using thin provisioning, management tools need to resize the disk in
6125 certain cases. To avoid having them to poll disk usage this version
6126 introduces an event which will be fired when a given offset of the storage
6127 is written by the hypervisor. Together with the API it allows registering
6128 thresholds for given storage backing volumes and this event will then
6129 notify management if the threshold is exceeded. Currently only the qemu
6130 driver supports this.
6132 * bhyve: Add support for UEFI boot ROM, VNC, and USB tablet
6134 The bhyve driver now supports booting using the UEFI boot ROM, so
6135 non-FreeBSD guests that support UEFI could be booted without using an
6136 external boot loader like grub-bhyve. Video is also supported now, allowing
6137 to connect to guests via VNC and use an USB tablet as an input device.
6138 Please refer to the driver page for domain XML examples.
6142 * qemu: Detect host CPU model by asking QEMU on x86_64
6144 Previously, libvirt detected the host CPU model using CPUID instruction,
6145 which caused libvirt to detect a lot of CPU features that are not supported
6146 by QEMU/KVM. Asking QEMU makes sure we don't start it with unsupported
6149 * perf: Add more perf statistics
6151 Add support to get the count of cpu clock time, task clock time, page
6152 faults, context switches, cpu migrations, minor page faults, major page
6153 faults, alignment faults, emulation faults by applications running on the
6156 * Write hyperv crash information into vm log
6158 qemu's implementation of the hyperv panic notifier now reports information
6159 about the crash from the guest os. Starting with this version, libvirt logs
6160 the information to the vm log file for possible debugging.
6164 * QEMU: Use adaptive timeout for connecting to monitor
6166 When starting qemu, libvirt waits for qemu to create the monitor socket
6167 which libvirt connects to. Historically, there was sharp 30 second timeout
6168 after which the qemu process was killed. This approach is suboptimal as in
6169 some scenarios with huge amounts of guest RAM it can take a minute or more
6170 for kernel to allocate and zero out pages for qemu. The timeout is now
6171 flexible and computed by libvirt at domain startup.
6173 * Overwrite (clear) 2 KB instead of just 512 bytes when initializing logical
6176 * Describe the logical backend requirements better for pool-create-as
6184 * storage: Add Virtuozzo storage backend storage pool
6186 Add new storage backend to support pool and volume management within the
6187 Virtuozzo Storage environment. Virtuozzo Storage is a highly available
6188 distributed software defined storage with built-in replication and disaster
6191 * qemu: Add support for memory backing with file source
6193 Add support in numa topology for file source inside memory backing
6194 (hugepages are not needed) Three new elements <source/>,<access/> and
6195 <allocation/> were added to <memoryBacking/> element. Also new
6196 configuration parameter ``memory_backing_dir`` was added to qemu.conf.
6198 * network: make openvswitch call timeout configurable
6200 Adding the ability to specify the timeout value in seconds for openvswitch
6201 calls in the libvirtd configuration file.
6203 * bhyve: add e1000 NIC support
6205 Add support for e1000 NIC. Previously, the only available option was
6208 * libxl: add tunneled migration support
6210 Add tunneled migration to libxl driver, which is always capable of strong
6211 encryption and doesn't require any extra network connection other than
6212 what's required for remote access of libvirtd.
6214 * qemu: add rendernode argument
6216 Add a new attribute 'rendernode' to <gl> spice element.
6218 * nodedev: add drm capability
6220 Add a new 'drm' capability for Direct Rendering Manager (DRM) devices,
6221 providing device type information.
6223 * Add API for individual/specific vCPU hotplug
6225 The new API allows selecting specific vCPUs to be added/removed from the
6226 VM. The existing APIs allowed only adding/removing from the end which did
6227 not play well with NUMA.
6231 * virsh: pool-list: allow both --uuid and --name in one command
6233 Adjust the virsh-pool command to support the --uuid and/or --name options
6234 in order to print just the --name and/or --uuid of pools.
6236 * Introduce MTU to domain <interface/> and <network>
6238 Allow setting MTU size for some types of domain interface and network.
6240 * libxl: improve support for <timer> configurations
6242 Add support for multiple timers. Extend the tsc timer to support the
6243 emulate mode. Improve conversion of timer XML to/from xl.cfg.
6245 * storage: modularize the storage driver
6247 Split up the storage driver backends into loadable modules so that binary
6248 distributions don't have to compromise on shipping the storage driver with
6249 all backends which may pull in too many dependencies.
6253 * nodedev: Fabric name must not be required for fc_host capability
6255 fabric_name is one of many fc_host attributes in Linux that is optional and
6256 left to the low-level driver to decide if it is implemented. For example
6257 the zfcp device driver does not provide a fabric name for an fcp host. The
6258 requirement for the existence of a fabric name has been removed by making
6261 * bhyve: change address allocation schema for SATA disks
6263 Previously, the bhyve driver assigned PCI addresses to SATA disks directly
6264 rather than assigning that to a controller and using SATA addresses for
6265 disks. It was implemented this way because bhyve has no notion of an
6266 explicit SATA controller. However, as this doesn't match libvirt's
6267 understanding of disk addresses, the bhyve driver was changed to follow the
6268 common schema and have PCI addresses for SATA controllers and SATA
6269 addresses for disks. If you're having issues because of this, it's
6270 recommended to edit the domain's XML and remove <address type='pci'> from
6271 the <disk> elements with <target bus='sata'/> and let libvirt regenerate it
6274 * libxl: maximum memory fixes
6276 Fix reporting of domain maximum memory. Fix setting dom0 maximum memory.
6278 * libxl: fix disk detach when <driver> not specified
6280 * libxl: fix dom0 autoballooning with Xen 4.8
6282 * qemu: Allow empty script path to <interface/>
6284 Historically, this was always allowed. Unfortunately, due to some rework
6285 done for 1.3.2 release a bug was dragged in which suddenly stop allowing
6286 domain with such configuration to start.
6294 * Domain events for metadata content changes
6296 The domain events framework has a new event ID that can be used to get
6297 notifications when domain metadata content changes.
6299 * Event notifications for the secret object
6301 The secret object now supports event notifications, covering lifcycle
6302 changes and secret value changes.
6304 * New localPtr attribute for "ip" element in network XML
6306 * qemu: Support QEMU group I/O throttling
6308 Add the capability to allow group I/O throttling via a new domain <disk>
6309 <iotune> subelement "group_name" to allow sharing I/O throttling quota
6310 between multiple drives.
6312 * nss: Introduce libvirt_guest
6314 New ``libvirt_guest`` nss module that translates libvirt guest names into
6317 * daemon: Add support for runtime logging settings adjustment
6319 Logging-related settings like log outputs and filters can now be adjusted
6320 during runtime using the admin interface without the necessity of the
6323 * storage: Add virStorageVolInfoFlags API
6325 Add the API to support using the VIR_STORAGE_VOL_GET_PHYSICAL flag in order
6326 to return the host physical size in bytes of the image container in the
6327 allocation field of the _virStorageVolInfo structure. The --physical flag
6328 has been added to the virsh vol-info command to access the data.
6330 * libxl: Implement virDomainGetMaxVcpus API
6332 * storage: Add overwrite flag checking for logical pool
6334 Add support for the OVERWRITE flags for the logical storage backend
6335 including checking for existing data on the target volumes when building a
6336 new logical pool on target volume(s).
6338 * qemu: Add support for guest CPU configuration on s390(x)
6342 * perf: Add more perf statistics
6344 Add support to get the count of branch instructions executed, branch
6345 misses, bus cycles, stalled frontend cpu cycles, stalled backend cpu
6346 cycles, and ref cpu cycles by applications running on the platform.
6348 * conf: Display <physical> for volume xml
6350 Add a display of the <physical> size of a disk volume in the output of the
6353 * qemu: Use virtio-pci by default for aarch64 mach-virt guests
6355 virtio-pci provides several advantages over virtio-mmio, such as the
6356 ability to hotplug devices and improved performance. While opting in to
6357 virtio-pci has been possible for a while, newly-defined guests will now use
6360 * vbox: remove support for VirtualBox 3.x and older
6362 Those old VirtualBox versions have been unsupported by upstream for a long
6363 time and the API of 4.0 and newer has diverged enough to require code
6364 abstractions to handle differences. Removing support for those old versions
6365 drops lots of code from the driver and simplifies the logic to ease
6366 implementation of new features going forward.
6368 * virsh: pool-info: introduce option --bytes
6370 Add option --bytes to virsh pool-info in order ti allow display of units in
6371 bytes rather than default of human readable output.
6373 * scsi: Add parent wwnn/wwpn or fabric capability for createVport
6375 Improve the algorithm searching for the parent scsi_host device for
6376 vHBA/NPIV scsi_host creation. Rather than supplying the "parent" by name,
6377 it's now possible to define the parent by it's wwnn/wwpn or fabric_wwn in
6378 the node device create XML or the storage pool XML.
6380 * qemu: aggregate pcie-root-ports onto multiple functions of a slot
6382 When pcie-root-ports are added to pcie-root in order to provide a place to
6383 connect PCI Express endpoint devices, libvirt now aggregates multiple root
6384 ports together onto the same slot (up to 8 per slot) in order to conserve
6385 slots. Using this method, it's possible to connect more than 200 endpoint
6386 devices to a guest that uses PCIe without requiring setup of any PCIe
6391 * lxc: fix accidental killing of containers during libvirtd restart
6393 The libvirt_lxc process was previously not moved into the container scope.
6394 As a result, if systemd reloads its config after a container is started,
6395 when libvirtd is later restarted it will accidentally kill the containers.
6397 * qemu: Correct GetBlockInfo values
6399 For an active domain, correct the physical value provided for a raw sparse
6400 file backed storage and the allocation value provided for a qcow2 file
6401 backed storage that hasn't yet been opened on the domain.
6403 * qemu: Make virtio console usable on ppc64 guests
6405 The chardev detection code has been improved and can now handle this
6406 configuration properly.
6408 * qemu: Enable mount namespace
6410 To avoid funny races with udev relabelling devices under our hands and to
6411 enhance security, libvirt now spawns each qemu process with its own
6414 * storage: Fix implementation of no-overwrite for file system backend
6416 Fix file system storage backend implementation of the OVERWRITE flags to be
6417 consistent between code and documentation. Add checks to ensure that when
6418 building a new file system on a target volume that there is not something
6419 already on the disk in a format that libvirt can recognize.
6421 * qemu: Create hugepage path on per domain basis
6423 Historically, all hugepage enabled domains shared the same path under
6424 hugetlbfs. This left libvirt unable to correctly set security labels on it.
6425 With this release, however, each domain is put into a separate path which
6426 is also correctly labeled.
6428 * conf: Reject domains with duplicate drive addresses
6430 Reject duplicate drive addresses for disks and hostdevs at domain
6433 * libxl: reverse defaults on HVM net device attach
6435 Fixes network interface attach for HVM domains when no model is specified.
6436 Emulated hotplug isn't yet supported and hence we should default to the
6437 general working scenario.
6439 * libxl: always enable pae for x86_64 HVM
6441 By default pae is disabled in libxl. Without an explicit <pae/> setting in
6442 the domain <features> configuration, an x86_64 HVM domain would be get an
6443 i686 environment. pae should always be enabled for x86_64 HVM domains.
6445 * qemu: Fix XML dump of autogenerated websocket
6447 As a result autogenerated websocket port is regenerated on domain restore,
6448 migration and such as it should be.
6456 * shmem: Add support for additional models
6458 The shmem device can now utilize QEMU's ivshmem-plain and ivshmem-doorbell,
6459 more modern versions of ivshmem.
6461 * vbox: Add VirtualBox 5.1 support
6463 * libssh: New transport
6465 The new libssh transport allows one to connect to a running libvirtd via
6466 SSH, using the libssh library; for example: ``qemu+libssh://server/system``
6469 * vhost-scsi: Add support scsi_host hostdev passthrough
6471 Add the capability to pass through a scsi_host HBA and the associated LUNs
6474 * Allow debugging of gluster volumes in qemu
6476 Users can now enable debug logging for native gluster volumes in qemu using
6477 the "gluster_debug_level" option in qemu.conf
6479 * Pre-allocate memory slots for memory hotplug
6481 Slot numbers for memory devices are now automatically allocated and thus
6482 persistent. In addition slot numbers can be specified without providing a
6483 base address, which simplifies user configuration
6485 * qemu: Express devices will be placed on PCIe bus by default
6487 For machine types that use a PCI Express root bus (e.g. x86_64/Q35 and
6488 aarch64/virt), any unaddressed PCI device that is an Express device (all
6489 virtio-1.0 devices, e1000e, nec-xhci, vfio assigned devices) will be placed
6490 on an Express controller (i.e. a pcie-root-port) instead of a legacy PCI
6491 controller (i.e. pci-bridge) with the root ports added as needed.
6495 * docs: Better documentation for migration APIs and flags
6497 * vbox: Address thread safety issues
6499 * virsh: Add support for passing an alternative persistent XML to migrate
6502 * vhostuser: Allow hotplug of multiqueue devices
6504 * NEWS: Switch to an improved format
6506 List user-visible changes instead of single commits for a better high-level
6507 overview of differences between libvirt releases.
6509 * website: Modernize layout and branding
6511 The libvirt website looked very cluttered and outdated; it has now been
6512 completely overhauled, resulting in a design that's better organized and
6513 more pleasant to look at.
6517 * vz: Fix migration in P2P mode
6519 * Forbid newline character in names of some libvirt objects
6521 * Fix compilation on macOS
6590 v1.2.21 (2015-11-04)
6591 ====================
6596 v1.2.20 (2015-10-02)
6597 ====================
6602 v1.2.19 (2015-09-02)
6603 ====================
6608 v1.2.18 (2015-08-03)
6609 ====================
6614 v1.2.17 (2015-07-02)
6615 ====================
6620 v1.2.16 (2015-06-01)
6621 ====================
6626 v1.2.15 (2015-05-04)
6627 ====================
6632 v1.2.14 (2015-04-02)
6633 ====================
6638 v1.2.13 (2015-03-02)
6639 ====================
6644 v1.2.12 (2015-01-27)
6645 ====================
6650 v1.2.11 (2014-12-13)
6651 ====================
6656 v1.2.10 (2014-11-03)
6657 ====================
6794 v0.10.2 (2012-09-24)
6795 ====================
6800 v0.10.1 (2012-08-31)
6801 ====================
6806 v0.10.0 (2012-08-29)
6807 ====================
6812 v0.9.13 (2012-08-02)
6813 ====================
6818 v0.9.12 (2012-05-14)
6819 ====================
6824 v0.9.11 (2012-04-03)
6825 ====================
6830 v0.9.10 (2012-02-13)
6831 ====================
7136 v0.1.11 (2007-01-22)
7137 ====================
7142 v0.1.10 (2006-12-20)
7143 ====================
7244 .. _git log: https://gitlab.com/libvirt/libvirt/-/commits/master