``, ``<description>`` or ``<metadata>``. * qemu: Add support for vDPA block devices With a new enough version of qemu, libvirt will allow you to assign vDPA block devices to a domain. This is configured with:: <disk type='vhostvdpa'> <source dev='/dev/vhost-vdpa-0'> ... * **Improvements** * qemu: add nbdkit backend for network disks Up until now, libvirt supported network disks (http, ftp, ssh) by passing the URL to qemu and having the appropriate qemu block drivers handle the disk I/O. However, by handling the network I/O outside of the qemu process, we get several advantages, such as reduced attack surface and improved stability of qemu. Therefore, when available, libvirt will use nbdkit as a backend for these network disks and export an NBD disk to qemu. * virnetdevopenvswitch: Propagate OVS error messages When configuring OVS interfaces/bridges libvirt used to report its own error messages instead of passing (more accurate) error messages from `ovs-vsctl`. This is now changed. * Various virtio-mem/virtio-pmem fixes Now libvirt validates more values of virtio-mem and virtio-pmem devices, e.g. overlapping memory addresses or alignment. v9.7.0 (2023-09-01) =================== * **New features** * qemu: basic support for use of "VFIO variant" drivers A VFIO variant driver is a device-specific driver that can be used in place of the generic vfio-pci driver, and provides extra functionality to support things like live migration of guests with vfio-assigned devices. It can currently be used by: 1) setting ``managed='no'`` in the XML configuration for the device 2) pre-binding the variant driver using the ``--driver`` option of ``virsh nodedev-detach``. * network: Support for ``<title>`` and ``<description>`` fields in Network XML The network object adds two more user defined metadata fields ``<title>`` and ``<description>``. Two new APIs ``virNetworkGetMetadata()`` and ``virNetworkSetMetadata()`` can be used to view and modify the above including the existing ``<metadata>`` field. virsh adds two new commands ``net-desc`` and ``net-metadata`` to view/modify the same. ``net-list`` adds a new option ``--title`` that prints the content of ``<title>`` in an extra column within the default ``--table`` output. * **Bug fixes** * qemu: Various fixes to firmware selection The changes made to firmware selection in libvirt 9.2.0 have unfortunately introduced a number of regressions. All known issues in this area have now been resolved. v9.6.0 (2023-08-01) =================== * **Security** * ``CVE-2023-3750``: Fix race condition in storage driver leading to a crash In **libvirt-8.3** a bug was introduced which in rare cases could cause ``libvirtd`` or ``virtstoraged`` to crash if multiple clients attempted to look up a storage volume by key, path or target path, while other clients attempted to access something from the same storage pool. * **Improvements** * apparmor: All profiles and abstractions now support local overrides This has long been the case for the ``virt-aa-helper`` profile, but has now been extended to all other profiles and abstractions. The mechanism used is the standard AppArmor 3.x one, where the contents of ``foo`` and ``abstractions/foo`` can be overridden by creating ``local/foo`` and ``abstractions/foo.d`` respectively. * qemu: Support ``removable`` attribute for scsi disk Now the scsi disk device (``/disk@device='disk'`` and ``/disk/target@bus='scsi'``) supports the ``removable`` attribute at ``/disk/target@removable```. * qemu: Add NUMA node automatically for memory hotplug Users no longer need to specify guest NUMA node in the domain XML when enabling memory hotplug, libvirt automatically adds one when it is missing. * qemu: Consider ``BeeGFS`` as a shared filesystem Allow migration with non-shared storage for VMs accessing storage via ``BeeGFS``. * **Bug fixes** * qemu: Adapt to new way of specifying PC speaker PC speaker is now usable again with newer QEMU since the change of how it is specified on the command line. * qemu_tpm: Try harder to create emulator state Libvirt no longer considers empty directory valid SWTPM state and setup is now run properly in such case. v9.5.0 (2023-07-03) =================== * **New features** * qemu: Allow configuring the ``discard-no-unref`` feature of ``qcow2`` driver The new ``discard_no_unref`` attribute of the ``disk`` ``driver`` element controls whether the ``qcow2`` driver in qemu unrefs clusters inside the image on discard requests. Disabling cluster unrefing decreases fragmentation of the image. * **Improvements** * qemu: Include maximum physical address size in baseline CPU When computing a baseline CPU definition for a set of hosts, we need to include maximum physical address size in the result to make sure it is compatible with all hosts even if their supported physical address sizes differ. * conf: Properly handle slots for non-DIMM ``<memory>`` devices Memory devices such as ``virtio-mem`` don't need a memory slot as they are PCI devices. ``libvirt`` now properly accounts the memory slots for such devices as well as specifying the ``slots`` attribute of the ``<maxMemory>`` element is no longer needed unless DIMM-like devices are to be used. * ``passt`` log and port forwarding improvements Libvirt now ensures that the ``passt`` helper process can access the configured log file even when it's placed in a directory without permissions. The ``<portForward>`` element of a passt-backed interface can now omit the ``address`` attribute as it's enough to specify a ``dev``. * **Bug fixes** * lxc: Allow seeking in ``/proc/meminfo`` to resove failure with new ``procps`` package New version of the ``free`` command from ``procps`` package seeks into the ``/proc/meminfo`` file, which was not supported by the instance of the file exposed via LXC causing a failure. * qemu: Fix rare race-condition when detaching a device The device removal handler callback function didn't re-check the state of the unplug operation after a timeout, which could rarely cause that the device was removed from the VM but not the definition. * qemu: Fix NUMA memory allocation logic QEMU allocates memory via the emulator thread thus that has to be allowed to access all configured NUMA nodes of the VM rather than just the one where it's supposed to be pinned. * qemu: Fix setup of ``hostdev`` backed ``<interface>`` The proper steps to initialize the host device were skipped for interfaces due to a logic bug preventing start of VM which used them. v9.4.0 (2023-06-01) =================== * **New features** * qemu: Support compression for parallel migration QEMU supports parallel migration to be compressed using either zstd or zlib. * cpu_map: Add SapphireRapids cpu model This model is introduced since QEMU 8.0. * **Improvements** * Adapt to musl-1.2.4 The latest version of musl stopped declaring some symbols that libvirt's test suite used (for redirecting ``stat()`` family of functions), leaving the tests broken. This is now fixed and the test suite works even with the latest version of musl. * conf: Introduce ``<address/>`` for virtio-mem and virtio-pmem To ensure guest ABI stability, libvirt persists address for memory devices, now including ``virtio-mem`` and ``virtio-pmem``. The address can be also specified by user. * **Bug fixes** * qemu: Account for NVMe disks when calculating memlock limit on hotplug When no ``<hard_limit/>`` is set, libvirt still tries to guess a sensible limit for memlock for domains. But this limit was not calculated properly on a hotplug of ``<disk type='nvme'/>``. * numa: Deny other memory modes than ``restrictive``` if a memnode is ``restrictive`` Due to a missing check it was possible to define a domain with incorrect ``<numatune/>``. For instance it was possible to have a ``<memnode mode="restrictive"/>`` and ``<memory/>`` of a different mode. This is now forbidden and if either all ``<memnode/>``-s and ``<memory/>`` have to have ``restrictive`` mode, or none. * qemu: Start emulator thread with more generous ``cpuset.mems`` To ensure memory is allocated only from configured NUMA nodes, libvirt sets up cpuset CGgroup controller, even before QEMU is executed. But this may prevent QEMU from setting affinity of threads that allocate memory. Since these threads are spawned from the emulator thread, the initial set up must be more generous and include union of all host NUMA nodes that are allowed in the domain definition. Once QEMU has allocated all its memory, the emulator thread is restricted further, as it otherwise would be. v9.3.0 (2023-05-02) =================== * **New features** * qemu: Introduce support for ``igb`` network interface model ``igb`` is a successor to the ``e1000e`` network device using PCIe interface. It was introduced in QEMU 8.0 * qemu: Improve handling of maximum physical address configuration * **Improvements** * qemu: Change default machine type for ARM and RISC-V ARM and RISC-V architectures now use the ``virt`` machine type by default. The previous defaults were nearly unusable and had to be overridden in most cases. * Improve translatable strings format substitutions All translatable error messages with substitution strings were converted to use positional modifiers to allow translators to shuffle around words in the translation. The translations in Weblate were also updated to match. * qemu: Improve validation of ``watchdog`` devices Certain invalid configurations of ``watchdog`` device are now properly detected: - hotplug of always-present platform watchdogs is forbidden - ``iTCO`` watchdog can be configured only once - ``ib700`` watchdog is allowed only on ``i440fx`` machines * Improved output of ``virt-host-validate`` on ARM Our validation tool now parses the ``IORT`` data on ARM to properly detect presence of SMMU and other features. * **Bug fixes** * qemu: Fix inactive internal snapshots of VM with UEFI firmware Recent changes to UEFI firmware handling resulted into breaking support for inactive internal snapshots of VMs with UEFI which historically worked. (Although the intention was to disallow them together with active ones, but the check did not work properly.) Preserve existing functionality by allowing such snapshots explicitly. * qemu: Properly configure locked memory limit for VMs with ``<disk type='nvme'`` The NVMe driver in qemu requires some memory to be locked. This was not taken into account in the code which calculates the memory limits based on devices present in the configuration * Fix native build on win32 Various improvements to the build system now allow users to build the client library of libvirt on win32 natively. * qemu: Properly detect tray of hotplugged CD-ROM devices Media in a CD-ROM device which was hotplugged could not be changed as the presence of the tray was not detected properly on hotplug. v9.2.0 (2023-04-01) =================== * **New features** * qemu: Add support for QCOW2 formatted firmware This type of firmware can be picked up either automatically, if the corresponding JSON descriptor has the highest priority, or manually by using ``<loader format='qcow2'/>`` in the domain XML. * qemu: Implement QEMU NBD reconnect delay attribute Support the nbd reconnect-delay of QEMU. It will set the delay time for reconnect after an unexpected disconnect or a serious error. * **Improvements** * qemu: Make firmware selection persistent Up until now, firmware autoselection has been performed at domain startup time: as a result, changes to the JSON firmware descriptors present on the system could have translated to a different firmware being chosen for subsequent startups of the same domain, potentially rendering it unbootable or lowering the security guarantees. Firmware selection now happens once, when the domain is defined, and its results are stored in the domain XML to be reused, unchanged, for all subsequent boots. * qemu: passt now works when SELinux/AppArmor is enabled In the case of SELinux, this requires passt-specific support code to be present in the host policy, so it might only work with upcoming operating systems and not with existing ones. * xen: Support custom UEFI firmware paths The Xen libxl driver now supports specifying a custom UEFI firmware path. Previously the Xen default was used in all cases. * **Bug fixes** * qemu: Fix validation of the HPET timer Due to a logic bug introduced in libvirt 9.0.0, VM configurations explicitly enabling the HPET timer were rejected. * qemu: Fix thread-context .host-nodes generation With new enough QEMU, libvirt instructs QEMU to set affinity of memory allocation threads. But this may have resulted in QEMU being unable to do so, as affinity to NUMA nodes inaccessible to emulator thread might have been requested. * rpc: fix typo in admin code generation Fix the bug in the remote ``virt-admin`` code generator, that resulted in a crash. Introduced in libvirt 9.1.0. * qemu: relax shared memory check for vhostuser daemons Fix hotplug of virtiofs ``filesystem`` after restarting libvirtd. Before, libvirtd would incorrectly complain about missing shared memory. v9.1.0 (2023-03-01) =================== * **Removed features** * vbox: removed support for version 5.2 and 6.0 APIs Libvirt no longer supports use of VirtualBox 5.2 and 6.0 since these versions reached their end of life on 2020/07. * **New features** * vbox: added support for version 7.0 API Libvirt can now support use of the VirtualBox 7.0, This is compile tested only, so we are looking for feedback from users on how well it works in practice. * qemu: Support crypto device Support crypto device(virtio crypto only), also add support for QEMU with backend ``builtin`` and ``lkcf``. * qemu: added support for pvpanic-pci device A pvpanic device can be now defined as a PCI device (the original is an ISA device) with ``<panic model='pvpanic'/>``. * qemu: support automatic restart of inadvertently terminated passt process If the passt process that is serving as the backend of a -netdev stream is terminated unexpectedly, libvirt now listens to QEMU's notification of this, and starts up a new passt instance, thus preserving network connectivity. * **Improvements** * RPM packaging changes The ``libvirt-daemon`` subpackage is split into several new subpackages, allowing installation of a modular daemon configuration without the traditional monolithic libvirtd. * **Bug fixes** * QEMU: iTCO watchdog made operational The watchdog was always included when q35 machine type was used, but needed an extra bit of configuration in order to be operational. This is now done by default when running a QEMU domain with q35 machine type. This is not a change in the guest ABI, but it is a guest visible behavior change since the watchdog that did not fire before will now fire once used. To switch to the previous behavior the watchdog action must be set to ``none``. * QEMU: fix deleting memory snapshot when deleting external snapshots When external snapshot deletion was introduced it did not remove memory snapshot when it existed. In addition when external memory only snapshot was created libvirt failed without producing any error. * QEMU: properly report passt startup errors Due to how the child passt process was started, the initial support for passt (added in 9.0.0) would not see errors encountered during startup, so libvirt would continue to setup and start the guest; this led to a running guest with no network connectivity. (NB: On systems that use them, it is still necessary to disable SELinux/AppArmor to start passt. This is a temporary limitation, and use of the feature in production is strongly discouraged until it has been lifted.) * qemu: Fix error when attempting to change media in a CDROM drive Due to a logic bug introduced in libvirt-9.0 attempts to change media in a CDROM would previously fail with an error stating that the tray isn't open. * qemu: Properly handle block job transitions Starting with libvirt-9.0 the block job state machine improperly handled some job transitions, which resulted into some block jobs not being properly terminated. This could cause problems such as errors when detaching a disk after snapshot. * virsh: Make domif-setlink work more than once There was a bug introduced in the previous release which made ``virsh domif-setlink`` work exactly once over given domain. The bug was fixed and now the command can be run multiple times. * qemu: Make domain startup fail if NIC already exists When starting a domain with an ``<interface/>`` that's supposed to be managed by libvirt (``managed='yes'``) but corresponding TAP device already exists, report an error and make the startup process fail. * qemu: Deal with nested mounts when umount()-ing /dev When setting up private ``/dev`` for a domain (also known as ``namespaces`` in ``qemu.conf``), libvirt preserves mount points nested under ``/dev`` (e.g. ``/dev/shm``, ``/dev/pts`` and so on). But there was a bug which resulted in inability to construct the namespace when there were two or more filesystems mounted on the same path. This is common scenario with containers and thus the bug was fixed. * remote: Pass ``mode`` and ``socket`` URI parameters to virt-ssh-helper When connecting to a remote host using SSH transport, ``?mode=`` and ``?socket=`` URI parameters were ignored. This prevented users from connecting to a monolithic daemon running on a remote host. * qemu: Various ``swtpm`` related fixes There are more cleanups and small bug fixes with regards to emulated ``<tpm/>``. For instance with migration when the ``swtpm`` state is on a shared volume, or seclabel setting/restoring. v9.0.0 (2023-01-16) =================== * **New features** * QEMU: implement external snapshot deletion External snapshot deletion is now possible using the existing API ``virDomainSnapshotDelete()``. Flags that allow deleting children or children only are not supported. * QEMU: support passt (https://passt.top) passt can be used to connect an emulated network device to the host's network without requiring libvirt to have any sort of elevated privileges. This is configured with:: <interface type='user'> <backend type='passt'> ... * QEMU: add external backend for swtpm Connecting the VM to a swtpm daemon started outside of libvirt is now possible. * QEMU: Support for passing FDs instead of opening files for `<disk>` A new API `virDomainFDAssociate` gives the users the option to pass FDs to libvirt and then use them when starting a VM. Currently the FDs can be used instead of directly opening files as `<disk>` backend. * **Improvements** * qemu: Prefer PNG for domain screenshots With sufficiently new QEMU (v7.1.0) screenshots change format from PPM to PNG. * tools: Fix install_mode for some scripts Scripts from the following list were installed with group write bit set: virt-xml-validate, virt-pki-validate, virt-sanlock-cleanup, libvirt-guests.sh. This was changed so that only the owner is able to write them. * qemu: Allow multiple nodes for preferred policy Due to restrictions of old kernels and libnuma APIs, the preferred NUMA policy accepted just a single host NUMA node. With recent enough kernel (v5.15.0) and libnuma (v2.0.15) it's possible to set multiple nodes. * secret: Inhibit shutdown of daemon for ephemeral secrets When an ephemeral secret is defined then automatic shutdown of virtsecretd is inhibited. This is to avoid ephemeral secrets disappearing shortly before their use. * qemu: Report Hyper-V Enlightenments in domcapabilities The supported Hyper-V Enlightenments are now reported in domain capabilities XML. * **Bug fixes** * Fix NULL-pointer dereference `virXMLPropStringRequired` Fix a bug where when parsing a XML property which is required to be present by using `virXMLPropStringRequired` the parser will crash instead of reporting an error. * qemu: Init ext devices paths on reconnect Paths for external devices are not stored in the status XML. Therefore, when the daemon restarted and was reconnecting to a running domain, these paths were left blank which led to the daemon crash. * qemu: Validate arguments passed to `virConnectGetDomainCapabilities` There was a code path in which insufficient validation of input arguments of `virConnectGetDomainCapabilities` API was possible which led to the daemon crash. This path is now fixed. v8.10.0 (2022-12-01) ==================== * **New features** * Tool for validating SEV firmware boot measurement of QEMU VMs The ``virt-qemu-sev-validate`` program will compare a reported SEV/SEV-ES domain launch measurement, to a computed launch measurement. This determines whether the domain has been tampered with during launch. * Support for SGX EPC (enclave page cache) Users can add a ``<memory model='sgx-epc'>`` device to launch a VM with ``Intel Software Guard Extensions``. * Support migration of vTPM state of QEMU vms on shared storage Pass ``--migration`` option if appropriate in order for ``swtpm`` to properly migrate on shared storage. * **Improvements** * Mark close callback (un-)register API as high priority High priority APIs use a separate thread pool thus can help in eliminating problems with stuck VMs. Marking the close callback API as high priority allows ``virsh`` to properly connect to the daemon in case the normal priority workers are stuck allowing other high priority API usage. * Updated x86 CPU features The following features for the x86 platform were added: ``v-vmsave-vmload``, ``vgif``, ``avx512-vp2intersect``, ``avx512-fp16``, ``serialize``, ``tsx-ldtrk``, ``arch-lbr``, ``xfd``, ``intel-pt-lip``, ``avic``, ``sgx``, ``sgxlc``, ``sgx-exinfo``, ``sgx1``, ``sgx2``, ``sgx-debug``, ``sgx-mode64``, ``sgx-provisionkey``, ``sgx-tokenkey``, ``sgx-kss``, ``bus-lock-detect``, ``pks``, ``amx``. * Add support for ``hv-avic`` Hyper-V enlightenment ``qemu-6.2`` introduced support for the ``hv-avic`` enlightenment which allows to use Hyper-V SynIC with hardware APICv/AVIC enabled. * qemu: Run memory preallocation with numa-pinned threads Run the thread allocating memory in the proper NUMA node to reduce overhead. * RPM packaging changes - add optional dependency of ``libvirt-daemon`` on ``libvirt-client`` The ``libvirt-guests.`` tool requires the ``virsh`` client to work properly, but we don't want to require the installation of the daemon if the tool is not used. - relax required ``python3-libvirt`` version for ``libvirt-client-qemu`` The ``virt-qemu-qmp-proxy`` tool requires python but doesn't strictly need the newest version. Remove the strict versioning requirement in order to prevent cyclic dependency when building. * **Bug fixes** * Skip initialization of ``cache`` capabilities if host doesn't support them Hypervisor drivers would fail to initialize on ``aarch64`` hosts with following error :: virStateInitialize:657 : Initialisation of cloud-hypervisor state driver failed: no error which prevented the startup of the daemon. * Allow incoming connections to guests on routed networks w/firewalld A change in handling of implicit rules in ``firewalld 1.0.0`` broke incoming connections to VMs when using ``routed`` network. This is fixed by adding a new ``libvirt-routed`` zone configured to once again allow incoming sessions to guests on routed networks. * Fix infinite loop in nodedev driver Certain udev entries might be of a size that makes libudev emit EINVAL which caused a busy loop burning CPU. Fix it by ignoring the return code. v8.9.0 (2022-11-01) =================== * **New features** * Add ``virt-qemu-qmp-proxy`` for emulating a QMP socket for libvirt managed VMs ``virt-qemu-qmp-proxy`` tool provides a way to expose an emulated QMP server socket for a VM managed by libvirt. This allows existing QMP-only clients to work with libvirt managed VMs. **Note:** libvirt is not interpreting the communication between the tool using the proxy and qemu itself, so any state-changing commands may desynchronize libvirt. Use at your own risk. * qemu: Core Scheduling support To avoid side channel attacks, the Linux kernel allows creating groups of processes that trust each other and thus can be scheduled to run on hyperthreads of a CPU core at the same time. This is now implemented for QEMU domains too (see ``sched_core`` knob in qemu.conf), although not enabled by default, just yet. * **Improvements** * qemu: Add hypervisor-specific statistics to ``virConnectGetAllDomainStats`` The new stats group ``VIR_DOMAIN_STATS_VM`` of ``virConnectGetAllDomainStats``, also exposed as ``virsh domstats --vm``, returns hypervisor-specific stats fields for given VM. * Add ``vendor`` attribute for CPU models in domain capabilities Users can now see the vendor of each CPU model in domain capabilities and use it, e.g., for filtering usable CPU models based on host CPU vendor. * virsh: Add ``--model`` option for ``hypervisor-cpu-baseline`` This is a shortcut for calling ``hypervisor-cpu-baseline`` with a single CPU model and no additional features. It can be used for determining which features block a particular CPU model from being usable. * Improved documentation of CPU ``usable`` attribute in domain capabilities * Report ``channel`` and ``redirdev`` devices in domain capabilities The channel and redirect devices supported by the hypervisor are now reported in domain capabilities. * meson: Bump minimal required meson version Newer meson versions deprecate some functions used. These were replaced with their newer counterparts and the minimal required mesion version was bumped to 0.56.0. * qemu: Add flags to keep or remove TPM state for ``virDomainUndefineFlags`` ``VIR_DOMAIN_UNDEFINE_TPM`` and ``VIR_DOMAIN_UNDEFINE_KEEP_TPM`` specify accordingly to delete or keep a TPM's persistent state directory structure and files when undefining a domain. In virsh the flags are exposed as ``--tpm`` and ``--keep-tpm`` for the sub-command ``undefine``. * **Bug fixes** * qemu: Disable all blocker features in CPU baseline Three years ago QEMU renamed some CPU features (mostly those containing an underscore). When such renamed feature was reported by QEMU as blocking usability of a CPU model, we would fail to explicitly disable it when creating a baseline CPU definition using this model. This bug did not have any functional impact when the default ``check='partial'`` attribute was used for guest CPU definition in domain XML, but it could have caused failures to start a domain with ``check='full'`` in some cases. * qemu: Do not crash after restart with active migration In 8.8.0 release libvirt daemon would crash after it was restarted during an active outgoing migration. * qemu: Refresh state after restore from a save image When a domain is restored from a saved image, libvirt now queries QEMU for those parts of runtime information that were not part of the save image. For instance: MAC address of a macvtap NICs, tray state of CD-ROMs, allocated size of virtio-mem, and others. v8.8.0 (2022-10-03) =================== * **Removed features** * storage: Remove 'sheepdog' storage driver backend The 'sheepdog' project is no longer maintained and upstream bug reports are unaddressed. Libvirt thus removed the support for the sheepdog storage driver backend, following qemu's removal of sheepdog support in qemu-6.1. * **Improvements** * qemu: Implement VIR_DOMAIN_STATS_CPU_TOTAL for qemu:///session Users can now query VIR_DOMAIN_STATS_CPU_TOTAL (also known as cpu.time) statistics for session domains. * **Bug fixes** * qemu: Fix non-shared storage migration setup This release fixes a bug in setup of a migration with non-shared storage ( ``virsh migrate --copy-storage-all``) which was broken by a refactor of the code in libvirt-8.7. * selinux: Don't ignore NVMe disks when setting image label Libvirt did not set any SELinux label on NVMe disks and relied only on the default SELinux policy. This turned out to cause problem when using namespace or altered policy and thus is fixed now. * qemu: Fix a deadlock when setting up namespace When starting a domain, libvirt creates a mount namespace and manages private /dev with only a handful nodes exposed. But when creating those a deadlock inside glib might have occurred. The code was changed so that libvirt does not tickle the glib bug. * qemu: Don't build memory paths on daemon restart When the daemon is restarted it tried to create domain private paths for each mounted hugetlbfs. When this failed, the corresponding domain was killed. This operation is now performed during domain startup and memory hotplug and no longer leads to sudden kill of the domain. v8.7.0 (2022-09-01) =================== * **Removed features** * qemu: Remove support for QEMU < 4.2 In accordance with our platform support policy, the oldest supported QEMU version is now bumped from 3.1 to 4.2. * **New features** * qemu: Add support for specifying vCPU physical address size in bits Users can now specify the number of vCPU physical address bits with the `<maxphysaddr>` subelement of the `<cpu>` element. * **Improvements** * esx: Domain XMLs can now be dumped for VMs with two new interface types One is when the interface is not connected anywhere `type='null'` and one when it is connected to VMWare Distributed Switch `type='vds'`. * **Bug fixes** * qemu: increase memlock limit for a domain with multiple vfio/vdpa devices When multiple vfio or vdpa devices are assigned to a domain, the locked memory limit could be too low to map memory for all devices. The memlock limit has been increased to be proportional to the number of vdpa/vfio devices. v8.6.0 (2022-08-01) =================== * **Improvements** * conf: Improved firmware autoselection The firmware autoselection feature now behaves more intuitively, reports better error messages on failure and comes with high-level documentation. v8.5.0 (2022-07-01) =================== * **New features** * qemu: Introduce support for network backed NVRAM Users can now use remote store NVRAM image by specifying newly introduced attribute `type='network'` with `<nvram>` element. * qemu: Add support for post-copy migration recovery A new ``VIR_MIGRATE_POSTCOPY_RESUME`` flag (``virsh migrate --postcopy-resume``) was introduced for recovering from a failed post-copy migration. * qemu: Add support for zero-copy migration With QEMU 7.1.0, libvirt can enable zerocopy for parallel migration. This is implemented by adding a new ``VIR_MIGRATE_ZEROCOPY`` flag(``virsh migrate --zerocopy``). * Introduce thread_pool_min and thread_pool_max attributes to IOThread New attributes ``thread_pool_min`` and ``thread_pool_max`` were introduced to ``<iothread/>`` as well as new ``<defaultiothread/>`` element with the same attributes. This way it's possible to instruct QEMU to spawn enough worker threads for an IOThread upfront, resulting in predictable time needed to process an I/O request. * **Improvements** * Define a TFTP server without a DHCP server in network configuration It's now possible to define a network with no DHCP server but with a TFTP server. This may be useful when DHCP service is provided by other entity on the network than libvirt spawned dnsmasq. * **Bug fixes** * qemu: Restore label to temp file in qemuDomainScreenshot() When virDomainScreenshot() is called, libvirt instructs QEMU to save the screenshot into a temporary file. This file needs to be labelled correctly, so that QEMU can access it. And since the file is temporary (it's deleted after the screenshot was taken) the corresponding label restore was missing. This proven to be problematic for profile based models, like AppArmor, where the temporary files were added into the profile but never removed, which resulted in longer profile recalculation times. * qemuBuildInterfaceConnect: Initialize @tapfd array Due to an uninitialized array, unsuccessful attempt to start a guest with an ``<interface/>`` might have resulted in closing of a random FD and thus sudden disconnect of a client or other random failures. * qemu: Fix hotplug of network interfaces A logic bug introduced in a recent refactor was fixed. The bug caused a problem when hot-adding a network interface, which failed with the following error:: error: internal error: unable to execute QEMU command 'netdev_add': File descriptor named '(null)' has not been found * Fix ``startupPolicy`` validation for ``block`` disks Setting of ``startupPolicy`` for a block disk would result in an error due to a logic bug in a recent refactor. * qemu: Fix crash when overriding device properties via ``<qemu:override>`` element Adding an override for a device property would result in a crash of the qemu driver. v8.4.0 (2022-06-01) =================== * **New features** * qemu: D-Bus display Libvirt is now able to setup a D-Bus display export, either with a private bus or in p2p mode. This display is available in QEMU 7.0.0. * qemu: ppc64 Power10 processor support Support for the recently released IBM Power10 processor was added. * qemu: Introduce ``absolute`` clock offset The ``absolute`` clock offset type allows to set the guest clock to an arbitrary epoch timestamp at each start. This is useful if some VM needs to be kept set to an arbitrary time for e.g. testing or working around broken software. * qemu: add qemu-vdagent channel This paravirtualized qemu vdagent channel can enable copy and paste between a guest and a VNC client. It is available in QEMU 6.1.0. * api: Add new APIs ``virDomainSaveParams`` and ``virDomainRestoreParams`` * ``virDomainSaveParams``: An alternative domain saving API, extends ``virDomainSaveFlags`` by adding parameters. * ``virDomainRestoreParams``: An alternative domain restoring API, extends ``virDomainRestoreFlags`` by adding parameters. * **Bug fixes** * Improve heuristics for computing baseline CPU models Both ``virConnectBaselineHypervisorCPU`` and ``virConnectBaselineCPU`` were in some cases computing the result using a CPU model which was newer than some of the input models. For example, ``Cascadelake-Server`` was used as a baseline for ``Skylake-Server-IBRS`` and ``Cascadelake-Server``. The CPU model selection heuristics was improved to choose a more appropriate model. v8.3.0 (2022-05-02) =================== * **Removed features** * qemu: Remove support for QEMU < 3.1 In accordance with our platform support policy, the oldest supported QEMU version is now bumped from 2.11 to 3.1. * **New features** * qemu: Introduce support for virtio-iommu This IOMMU device can be used with both Q35 and ARM virt guests. * qemu: Introduce attributes rss and rss_hash_report for net interface They can enable in-qemu/ebpf RSS and in-qemu RSS hash report for virtio NIC. Require QEMU >= 5.1. v8.2.0 (2022-04-01) =================== * **New features** * qemu: Introduce ``manual`` disk snapshot mode This new mode allows users to synchronize libvirt snapshots with snapshots which need to be done outside of libvirt e.g. when 'vhost-user-blk' is used to back the disk. * Introduce memory allocation threads When starting a QEMU guest, libvirt can now instruct QEMU to allocate guest's memory in parallel. This may be handy when guest has large amounts of memory. * **Improvements** * qemu: ``VIR_MIGRATE_PARAM_TLS_DESTINATION`` now works with non-shared storage migration The setting now also applies to the NBD connections for non-shared storage migration allowing migration to proceed even when the user expects certificate name not to match. * qemu: Allow overrides of device properties via the qemu namespace Users wishing to override or modify properties of devices configured by libvirt can use the ``<qemu:deviceOverride>`` QEMU namespace element to specify the overrides instead of relying on the argv passthrough of the ``-set`` qemu commandline option which no longer works with new qemu. * qemu: Allow passing file descriptors to ``virsh qemu-monitor-command`` Passing FDs allows users wanting to experiment with qemu driven by libvirt use commands like ``add-fd`` properly. * libxl: Turn on user aliases Users can now use so called user aliases for XEN domains. * Implement support for FUSE3 The LXC driver uses fuse to overwrite some lines in ``/proc/meminfo`` inside containers so that they see correct amount of memory given to them. The code was changed so that both ``fuse`` and ``fuse3`` are supported. * Improve domain save/restore throughput Code that's handling save or restore of QEMU domains was changed resulting in better performance of I/O and thus shortening time needed for the operation. * **Bug fixes** * Both build and tests should now pass on Alpine Linux or any other distribution with musl libc. * virsh: Fix integer overflow in allocpages On hosts which support hugepages larger than 1GiB ``virsh allocpages`` failed to accept them because of an integer overflow. This is now fixed. * qemu: Fix segmentation fault in virDomainUndefineFlags When a domain without any ``<loader/>`` was being undefined, libvirt has crashed. This is now fixed. * lxc: Fix unaligned reads of /proc/meminfo within a container When /proc/meminfo was read in chunks smaller than the entire file, libvirt would produce mangled output. While porting the code to FUSE3 this area was reworked and the file can now be read with any granularity. * qemu: Be less aggressive around cgroup_device_acl A basic set of devices common to every domain can be set in ``qemu.conf`` via cgroup_device_acl knob. Devices from this set are allowed in CGroup and created in domain private namespace for every domain. However, upon device hotunplug it may have had happened that libvirt mistakenly denied a device from this set and/or removed it from the namespace. For instance, /dev/urandom was removed and denied in CGroup on RNG hotunplug. * nodedev: trigger mdev device definition update on udev add and remove When nodedev objects are added and removed mdev device definitions are updated to report correct associated parent. v8.1.0 (2022-03-01) =================== * **New features** * qemu: Add hvf domain type for Hypervisor.framework It works on Intel machines as well as recent machines powered by Apple Silicon. QEMU 6.2.0 is needed for Apple Silicon support. * qemu: Support mode option for dirtyrate calculation Introduce ``virDomainDirtyRateCalcFlags`` as parameter of ``virDomainStartDirtyRateCalc``, which is used to specify the mode of dirty page rate calculation. Add ``--mode`` option to ``virsh domdirtyrate-calc``, which can be either of the following 3 options: ``page-sampling, dirty-bitmap, dirty-ring``. Add ``calc_mode`` field for dirtyrate statistics returned by ``virsh domstats --dirtyrate``, also add ``vCPU dirtyrate`` if ``dirty-ring`` mode was used in last measurement. * **Improvements** * packaging: sysconfig files no longer installed libvirt used to provide defaults in various /etc/sysconfig/ files, such as /etc/sysconfig/libvirtd. Since these files are owned by the admin, this made it difficult to change built-in defaults in case such file was modified by the admin. The built-in defaults are now part of the provided systemd unit files, such as libvirtd.service. These unit files continue to parse sysconfig files, in case they are created by the admin and filled with the desired key=value pairs. * virnetdev: Ignore EPERM on implicit clearing of VF VLAN ID Libvirt will now ignore EPERM errors on attempts to implicitly clear a VLAN ID (when a VLAN is not explicitly provided via an interface XML using a 0 or a non-zero value) as SmartNIC DPUs do not expose VLAN programming capabilities to the hypervisor host. This allows Libvirt clients to avoid specifying a VLAN and expect VF configuration to work since Libvirt tries to clear a VLAN in the same operation as setting a MAC address for VIR_DOMAIN_NET_TYPE_HOSTDEV devices which is now split into two distinct operations. EPERM errors received while trying to program a non-zero VLAN ID or explicitly program a VLAN ID 0 will still cause errors as before so there is no change in behavior in those cases. * **Bug fixes** * Remove unix sockets from filesystem when disabling a '.socket' systemd unit The presence of the socket files is used by our remote driver to determine which service to access. Since neither systemd nor the daemons clean up the socket file clients were running into problems when a modular deployment was switched to monolithic ``libvirtd``. * qemu: Fixes of fd passing during hotplug and hotunplug of chardevs FDs used as chardev backing are now properly removed when hot-unplugging a chardev from qemu and hotplugged chardevs now properly use ``virtlogd`` to handle the input and output from qemu. * RPM: Run pre/post-install steps on ``daemon-driver-storage-core`` Previously the pre/post-install code was part of the meta-package which installed all storage driver sub-packages thus a minimalistic install of the storage driver didn't behave correctly. v8.0.0 (2022-01-14) =================== * **Security** * libxl: Fix potential deadlock and crash (CVE-2021-4147) A rogue guest could continuously reboot itself and cause libvirtd on the host to deadlock or crash, resulting in a denial of service condition. * **Removed features** * qemu: Explicitly forbid live changing nodeset for strict numatune For ``strict`` mode of <numatune/> it can't be guaranteed that memory is moved completely onto new set of nodes (e.g. QEMU might have locked pieces of its memory) thus breaking the strict promise. If live migration of QEMU memory between NUMA nodes is desired, users are advised to use ``restrictive`` mode instead. * **New features** * qemu: Synchronous write mode for disk copy operations The ``blockdev-mirror`` block job supports a mode where writes from the VM are synchronously propagated to the destination of the copy. This ensures that the job will converge under heavy I/O. Implement the mode for the copy blockjob as ``VIR_DOMAIN_BLOCK_COPY_SYNCHRONOUS_WRITES`` flag exposed via ``virsh blockcopy --synchronous-writes`` and for non-shared storage migration as ``VIR_MIGRATE_NON_SHARED_SYNCHRONOUS_WRITES`` exposed via ``virsh migrate --copy-storage-synchronous-writes``. * Introduce TCG domain features Libvirt is now able to set the size of translation block cache size (tb-size) for TCG domains. * qemu: Add new API to inject a launch secret in a domain New API ``virDomainSetLaunchSecurityState()`` and virsh command ``domsetlaunchsecstate`` are added to support injecting a launch secret in a domain's memory. * **Improvements** * libxl: Implement the virDomainGetMessages API * qemu: Preserve qcow2 sub-cluster allocation state after external snapshots and block-copy The new image which is installed as an overlay on top of the current chain when taking an external snapshot, or the target of a block copy operation now enables sub-cluster allocation (``extended_l2``) if the original image has the option enabled. * **Bug fixes** * qemu: Fix device hot-unplug with ``libvirt-7.9`` or ``libvirt-7.10`` used with ``qemu-6.2`` An internal change to the configuration format used by the above libvirt versions triggers a bug in ``qemu-6.2`` where qemu no longer emits the event notifying that the device was unplugged successfully and thus libvirt never removes the device from the definition. This impacts only devices which were present at startup of the VM, hotplugged devices behave correctly. This is fixed in ``libvirt-8.0`` by reverting to the old configuration approach until qemu is fixed. As a workaround for ``libvirt-7.9`` and ``libvirt-7.10`` the old configuration approach can be forced by: Option 1, global ``qemu.conf``:: capability_filters = [ "device.json" ] Option 2, per VM XML override:: <domain type='kvm' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'> [...] <qemu:capabilities> <qemu:del capability='device.json'/> </qemu:capabilities> </domain> * Fix sparse streams with split daemon In split daemon scenario, a client connected to a hypervisor driver and using sparse streams (e.g. ``virsh vol-download --sparse``) would make the hypervisor daemon enter an infinite loop without any data transfer. This is now fixed. * Build no longer requires RPC library Code and its cross dependencies were fixed so that build without remote driver and thus an RPC library (like ``tirpc``) fails no more. * virnetdevopenvswitch: Fix 'burst' value passed to ovs-vsctl When a ``<bandwidth/>`` was defined for a TAP device that's plugged into an OvS bridge values passed to the OvS were incorrectly recalculated resulting in slightly different limits being applied. v7.10.0 (2021-12-01) ==================== * **New features** * Added virt-pki-query-dn binary This binary helps users figure out the format of Distinguished Name from a certificate file the way that libvirt expects it in tls_allowed_dn_list option of libvirtd.conf configuration file * **Improvements** * qemu: Report guest interface information in ``virDomainGetGuestInfo`` Libvirt is now able to report interface information from the guest's perspective (using guest agent). * qemu: detect guest side errors during device removal Libvirt is now able to detect guest side errors during device removal by using the DEVICE_UNPLUG_GUEST_ERROR event, available in QEMU 6.2.0. * Minimum SSF setting The libvirtd.conf option tcp_min_ssf can be used to override the minimum permitted security strength factor for non-TLS remote connections. The current hardcoded minimum is 56 (single-DES) and will be raised to 112 in the future. Setting a minimum lower than 112 is not supported. * qemu: Report stats also for block copy destination and backup job scratch images The statistics are available via the bulk domain stats API. * **Bug fixes** * qemu: Don't format 'ramfb' attribute when disabled Fix a regression caused by the conversion to JSON -device arguments where 'ramfb' would be put on the commandline of 'vfio-pci' which doesn't have it, thus breaking VMs with a mediated host device. * qemu: Fix block copy and backup to encrypted storage An oversight in last release lead to a spurious error being reported when encrypted storage was requested for disk images which are not directly backing the disk, which is now fixed. v7.9.0 (2021-11-01) =================== * **New features** * Introduce virtio-mem ``<memory/>`` model New virtio-mem model is introduced for ``<memory/>`` device which is a paravirtualized mechanism of adding/removing memory to/from a VM. Use ``virDomainUpdateDeviceFlags()`` API to adjust amount of memory or ``virsh update-memory-device`` for convenience. * qemu: support disabling hotplug of devices on the pci-root controller the <target hotplug='on|off'/> option is now supported for the pci-root controller on i440fx-based (x86 "pc") machinetypes. This can be used to disable hotplug/unplug of devices from this controller. The default behavior is unchanged (hotplug is allowed). * Support hotplug and hotunplug for virtiofs Filesystems backed by virtiofsd can now be hotplugged and hotunplugged. * virpcivpd: Add a PCI VPD parser A parser for the standard PCI/PCIe VPD ("I.3. VPD Definitions" in PCI 2.2+ and an equivalent definition in "6.28.1 VPD Format" PCIe 4.0) was added along with relevant types to represent PCI VPD in memory. This functionality got added for Linux only at this point (kernels above v2.6.26 have support for exposing VPD via sysfs). * virpci: Add PCI VPD-related helper functions to virpci In order to utilize the PCI VPD parser, a couple of helper functions got introduced to check for the presence of a VPD file in the sysfs tree and to invoke the PCI VPD parser to get a list of resources representing PCI VPD contents in memory. * nodedev: Add PCI VPD capability support Support for serializing and deserializing PCI VPD data structures is added following the addition of the PCI VPD parser. A new PCI device capability called "vpd" is introduced holding string resources and keyword resources found in PCI VPD. * qemu: Support page_per_vq for driver element This optional virtio attribute ``page_per_vq`` controls the layout of the notification capabilities exposed to the guest. It is recommended for the vDPA devices. * qemu: Support librbd encryption Add an encryption engine ``librbd``. It will provides the image-level encryption of librbd. It requires QEMU >= 6.1.0 and librbd >= 16.1.0. * **Improvements** * Use of JSON syntax with ``-device`` with upcoming QEMU-6.2 Libvirt started using JSON directly with the ``-device`` commandline parameter as it's considered the preferred stable syntax for further QEMU releases. If any problems with the conversion are encountered please report them as soon as possible. * **Bug fixes** * qemu: Fix problems on ``virsh domstats`` with qemu <5.2.0 Libvirt v7.2.0 and later called query-dirty-rate, which was introduced in qemu-5.2.0, regardless of qemu version and failed in qemu-5.1.0. This release fixes the bug. * Don't enter endless loop when unable to accept new clients If libvirtd (or any other daemon) hit the ulimit for maximum number of open files but there are still client connections pending then libvirtd (or corresponding split daemon) would enter an endless loop from which it would never recover. This behaviour is now fixed. * qemu: Run secondary driver hooks in split daemon mode Because of a bug in implementation it may happen that hooks from secondary drivers were not called in all cases, for instance a network hook wasn't called upon removal of interface after domain shut off itself. With this release the bug is fixed. v7.8.0 (2021-10-01) =================== * **New features** * nodedev: Add ability to automatically start mediated devices The autostart status of a persistent mediated devices can be managed with the new APIs ``virNodeDeviceSetAutostart()`` and ``virNodeDeviceGetAutostart()``. The corresponding virsh command is ``nodedev-autostart``. In addition, two new APIs were added to get additional information about node devices: ``virNodeDeviceIsPersistent()`` checks whether the device is persistently defined, and ``virNodeDeviceIsActive()`` checks whether the node device is currently active. This information can also be retrieved with the new virsh command ``nodedev-info``. * qemu: Add attribute ``queue_size`` for virtio-blk devices * **Improvements** * api: Add XML validation for creating of: networkport, nwfilter-binding, network * Add flag ``VIR_NETWORK_PORT_CREATE_VALIDATE`` to validate network port input xml of network-port creating. * Add flag ``VIR_NETWORK_CREATE_VALIDATE`` to validate network input xml of network creating. * Add flag ``VIR_NWFILTER_BINDING_CREATE_VALIDATE`` to validate nwfilter-binding input xml of nwfilter-binding creating. v7.7.0 (2021-09-01) =================== * **New features** * Add support for Fibre Channel VMID New VM element ``<fibrechannel appid=''/>`` was added to allow users to set their ``appid`` for each VM which will be used by kernel to create Fibre Channel VMID. This allows various QoS levels, access control or collecting telemetry data per VM. * **Improvements** * virsh: Allow XML validation for define of: storage pool, network, secret, nwfilter, interface * Add flag ``VIR_STORAGE_POOL_DEFINE_VALIDATE`` to validate storage pool input xml. For virsh, users can use it as ``virsh pool-define --validate``. * Add flag ``VIR_NETWORK_DEFINE_VALIDATE`` to validate network input xml. For virsh, users can use it as ``net-define --validate``. * Add flag ``VIR_SECRET_DEFINE_VALIDATE`` to validate secret input xml. For virsh, users can use it as ``secret-define --validate``. * Add flag ``VIR_NWFILTER_DEFINE_VALIDATE`` to validate nwfilter input xml. For virsh, users can use it as ``nwfilter-define --validate``. * Add flag ``VIR_INTERFACE_DEFINE_VALIDATE`` to validate interface input xml. For virsh, users can use it as ``iface-define --validate``. * Add SecurityManager APIs for labeling network devices New ``virSecurityManagerSetNetdevLabel`` and ``virSecurityManagerSetNetdevLabel`` APIs are introduced and implemented in the Apparmor security driver. The qemu driver uses the APIs to label vhostuser ports on hotplug and restore labeling on unplug. * vmx: Parse vm.genid and support super wide SCSI bus The genid attribute is now reported for VMX guests. Libvirt can now properly process super wide SCSI bus (64 units). * qemu: Lifecycle action (``on_poweroff``/``on_reboot``) handling improvements The handling of lifecycle actions was fixed and improved in multiple ways: - ``restart-rename`` action was forbidden The action was never properly implemented in the qemu driver and didn't actually result in a restart of the VM but rather termination. The qemu driver now rejects such configurations. - ``preserve`` action was forbidden Similarly to the previous case this never worked as the intended semantics of the actions dictate. It's better to not allow it at all until there's a proper implementation - ``reboot`` action of ``on_poweroff`` now actually works The guest OS is now rebooted instead of terminating the VM when the ``reboot`` action is used and the guest OS powers down. Note that it's incompatible with ``on_reboot`` set to ``destroy``. - Changes in action action of ``on_reboot`` are now updated with qemu Libvirtd can now properly update the ``on_reboot`` action in qemu which allows proper handling when changing between ``reboot`` and ``destroy`` actions. In addition, switching from ``reboot`` to ``destroy`` was forbidden for older qemus which don't support the update API as the guest could still reboot and execute some instructions until it was terminated. * virsh: Support vhostuser in attach-interface * **Bug fixes** * qemu: Open chardev logfile on behalf of QEMU Guests with a logfile configured for their chardevs are now able to start even when no virtlogd is configured. * virhostmem: Handle numactl-less build in hugepages allocation/reporting Some architectures don't have notion of NUMA (e.g. s390x) but do support hugepages. Libvirt silently ignored requests to allocate/report hugepage pool when built without numactl. This is now fixed and the pool can be allocated/reported on properly. * qemu: Record proper ``backing`` format for overlays of qcow2+luks images Libvirt would record ``luks`` instead of ``qcow2`` into the metadata. In practice this is a problem only when inspecting images manually via ``qemu-img`` as with libvirt users must use full specification of the backing chain in the domain XML which supersedes information recorded in the image metadata. v7.6.0 (2021-08-02) =================== * **Security** * storage: Unlock pool objects on ACL check failures in ``storagePoolLookupByTargetPath`` (CVE-2021-3667) A logic bug in ``storagePoolLookupByTargetPath`` where the storage pool object was left locked after a failure of the ACL check could potentially deprive legitimate users access to a storage pool object by users who don't have access. * **New features** * qemu: Incremental backup support via ``virDomainBackupBegin`` libvirt-7.6 along with the unreleased qemu-6.1 will fully support the change block tracking features (block-dirty-bitmaps) to be able to do incremental backups and management of the checkpoint states via the appropriate APIs. * qemu: Add support for launch security type s390-pv Specifying s390-pv as launch security type in an s390 domain prepares for running the guest in protected virtualization secure mode, also known as IBM Secure Execution. This simplifies the definition and reduces the risk of an incorrect definition, e.g. by forgetting to specify ``iommu=on`` on all virtio devices. * domstats: Add haltpolling time statistic interface Domstats now provide the data of cpu haltpolling time. This feature relies on statistics available after kernel version 5.8. This will allow the user to get more accurate CPU usage information if needed. * **Bug fixes** * qemu: Fix migration with ``VIR_MIGRATE_NON_SHARED_INC`` libvirt 7.3.0 introduced a bug where ``VIR_MIGRATE_NON_SHARED_INC`` would not actually migrate the contents of the disk due to broken logic and at the same time could trigger migration of storage when ``VIR_MIGRATE_TUNNELLED`` is requested. This release fixes the bug. * qemu: Don't emit ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` twice when registered with index When registering the threshold event with the index notation (e.g. ``vda[3]``) libvirt would emit the event also for ``vda`` if the image is in the top layer. The intention was to emit two events only when the original registration was done without the index. * qemu: Pass discard requests for disks with ``copy_on_read='on'`` When a disk using the ``copy_on_read='on'`` option is configured also with ``discard='unmap'`` the discard requests will now be passed to the underlying image freeing up the space. v7.5.0 (2021-07-01) =================== * **Security** * svirt: fix MCS label generation (CVE-2021-3631) A flaw in the way MCS labels were generated could result in a VM's resource not being fully protected from access by another VM were it to be compromised. https://gitlab.com/libvirt/libvirt/-/issues/153 * **Removed features** * xen: Remove support for Xen < 4.9 In accordance with our platform support policy, the oldest supported Xen version is now bumped from 4.6 to 4.9. * **Improvements** * docs: Document disk serial truncation status quo Disk ``<serial>`` is being truncated by QEMU before passed to the guest. Since it's impossible to fix it without running into further regressions the documentation was improved to document the intricacies. * **Bug fixes** * qemu: Fixed validation of disk ``iothread`` configuration The validation of ``iothread`` config was previously moved to a place where it caused bogus errors when address wasn't allocated when hotplugging a disk. The check is now removed as it wasn't actually necessary at all. v7.4.0 (2021-06-01) =================== * **Removed features** * qemu: Remove support for QEMU < 2.11 In accordance with our platform support policy, the oldest supported QEMU version is now bumped from 1.5 to 2.11. * **New features** * qemu: Add support for hotplugging ``<transient/>`` disks The disk hotplug code in the qemu driver now can handle hotplug of disks with automatically added overlay. * qemu: Add support for sharing base image of ``<transient/>`` disks Users can use ``<transient shareBacking='yes'/>`` to tell the qemu driver to never open the base image in write mode thus multiple VMs can share the same image. Note that the disk will be hotplugged during startup. * **Improvements** * Add win-dmp crashdump format New ``win-dmp`` format for ``virDomainCoreDumpWithFormat`` API and/or virsh ``dump --format`` was introduced. * **Bug fixes** * Allow 0 offset in XML schema for ``<slice type='storage' offset='0' size='321'/>`` Having a 0 offset so that the size of the image can be limited is a valid configuration so it was allowed in the XML schema. v7.3.0 (2021-05-03) =================== * **New features** * xen: Support domains with more than 4TB The xen driver now supports domains with more than 4TB of memory with xen >= 4.13. * qemu: add socket for virtiofs filesystems Libvirt now supports ``filesystem`` devices that connect to a ``virtiofsd`` daemon launched outside of libvirtd, via the ``socket`` attribute of the ``source`` element. * nodedev: Add ability to manage persistent mediated devices Persistent mediated devices can now be managed with libvirt. ``virNodeDeviceDefineXML()`` defines a new device, ``virNodeDeviceUndefine()`` removes an existing definition, and ``virNodeDeviceCreate()`` starts a device definition that is currently inactive. Corresponding virsh commands ``nodedev-define``, ``nodedev-undefine``, and ``nodedev-start`` were also added. ``nodedev-list`` only lists active devices by default. Inactive device definitions can be shown with the new ``--inactive`` and ``--all`` flags. * qemu: Allow use of qemu's ``-compat`` option Curious developers or testers now can enable certain ``-compat`` modes which allow to notice use of deprecated commands and options as qemu will use the selected method to notify the user. The new behaviour can be requested using either the ``deprecation_behavior`` option in ``qemu.conf`` for all VMs or using ``<qemu:deprecation behavior='...'/>`` in the VM XML. * **Improvements** * virsh: Improve errors with ``virsh snapshot-create-as`` The XML document constructed by virsh was forced through XML schema validation which yielded unintelligible error messages in cases such as when the path to the new image did not start with a slash. XML documents are no longer validated as the XML parser actually has better error messages which allow users to figure the problem out quickly. * qemu: Terminate backing store when doing a full-chain block pull When pulling everything into the overlay image the chain can be terminated since we know that it won't depend on any backing image and thus can prevent attempts to probe the backing chain. * qemu: Expose disk serial in virDomainGetGuestInfo() The ``virDomainGetGuestInfo()`` reports disk serial number among with other disk information. * **Bug fixes** * qemu: Fix crash of libvirt on full block pull of a disk When the persistent definition contains a compatible disk (meaning the definition of the running and persistent config match) a block pull job would leave a dangling pointer in the config definition which resulted in a crash. * qemu: Use proper job cancelling command Libvirt's API contract for aborting a block copy job in 'ready' state declares that the destination image of the copy will contain a consistent image of the disk from the time when the block job was aborted. This requires that libvirt uses the proper cancelling qemu command to ensure that the data is consistent which was not the case. * qemu: Don't attempt storage migration when there are no migratable disks Due to a logic bug introduced in the previous release libvirt would attempt to migrate disks in case when no disks are selected/eligible for migration. * qemu: Fix very rare race when two block job 'ready' events are delivered In certain high-load scenarios, qemu might deliver the 'ready' event twice and if it's delivered when pivoting to the destination during a block copy job, libvirt would get confused and execute the code as if the job were aborted. * lxc: Fix container destroy with CGroupsV2 When an LXC container was started and the host used CGroupsV2 it might have had created nested controllers under the container's scope. Libvirt was unaware and thus destroying the container failed with a cryptic error: ``failed to get cgroup backend for 'pathOfController'``. The CGroup removal code was reworked and is now capable of dealing with such scenario. * bash-completion: Fix argument passing to $1 Due to a bug in bash completion script, the auto completion did not work properly when a connection URI or read only flag were specified on ``virsh`` or ``virt-admin`` command line. v7.2.0 (2021-04-01) =================== * **New features** * qemu: Implement domain memory dirty rate calculation API New API ``virDomainStartDirtyRateCalc()`` and virsh command ``domdirtyrate-calc`` are added to start calculating a live domain's memory dirty rate. * qemu: Support reporting memory dirty rate stats The memory dirty rate stats can be obtained through ``virsh domstats --dirtyrate`` via the virConnectGetAllDomainStats API. * qemu: Full disk backups via ``virDomainBackupBegin`` The qemu hypervisor driver now allows taking full disk backups via the ``virDomainBackupBegin`` API and the corresponding virsh wrapper. In future releases the feature will be extended to also support incremental backups (where only the difference since the last backup is copied) when qemu adds the required functionality. * Add support for audio backend specific settings With this release a new ``<audio/>`` element is introduced that allows users to configure audio output for their guests. * **Improvements** * qemu: Compatibility with QEMU 6.0 for certain hot-(un)-plug operations Libvirt 7.2.0 is required for compatibility with the upcoming QEMU 6.0 release for hotplug and hotunplug of certain devices and helpers, such as iothreads, chardevs, RNG devices, disks with secret, ... * qemu: Various improvements to embedded mode Embedded mode for the QEMU driver, as well as the ``virt-qemu-run`` tool saw improvements in handling of domain life cycle, temporary directories creation (important when using disk secrets) and other minor fixes. * Documentation of split daemon related config files Split daemons read configuration files upon their start. These were never documented though. * **Bug fixes** * Check host CPU for forbidden features CPU feature policy did not work as expected with ``host-passthrough`` and features supported by physical host. CPU features were not filtered out when ``@check`` was set to ``full``. * Fix virNetworkUpdate() to work with split daemons Due to a bug in our code, virNetworkUpdate() did not work with split daemon unless management application connected to virtnetworkd directly. * qemu: increase locked memory limit when a vDPA device is present Just like VFIO devices, vDPA devices may need to have all guest memory pages locked/pinned in order to operate properly. These devices are now included when calculating the limit for memory lock. * Don't log error if SRIOV PF has no associated netdev Some SRIOV PFs don't have a netdev associated with them in which case libvirtd reported an error and refused to start. This is now fixed. * qemu: Only raise memlock limit if necessary Attempting to set the memlock limit might fail if we're running in a containerized environment where ``CAP_SYS_RESOURCE`` is not available, and if the limit is already high enough there's no point in trying to raise it anyway. * Restore security context of swtpm.log If a guest with emulated TPM was started and the daemon was restarted afterwards, the security context of the per-domain ``swtpm.log`` file was not restored on domain shutdown leaving it unable to be started again. * virtlogd|virtlockd: Fixed crash when upgrading the daemons in-place A bug preventing the in-place upgrade of ``virtlogd`` and ``virtlockd`` daemons was fixed, so they can again be upgraded without dropping the log file descriptors or locks on files. v7.1.0 (2021-03-01) =================== * **Portability** * Implement Apple Silicon support libvirt now runs on the ARM-based Apple Silicon Macs. * **New features** * Introduce virtio-pmem ``<memory/>`` model The virtio-pmem is a virtio variant of NVDIMM and just like NVDIMM virtio-pmem also allows accessing host pages bypassing guest page cache. * Introduce ``<boot order/>`` for ``<filesystem>`` Booting is possible from virtiofs filesystems. Introduce an option to control the boot order, like we do for other bootable devices. * hyperv: implement new APIs The ``virDomainUndefine()``, ``virDomainUndefineFlags()``, ``virDomainDefineXML()``, ``virDomainAttachDevice()``, and ``virDomainAttachDeviceFlags()``, ``virConnectListAllNetworks()``, ``virConnectNumOfNetworks()``, ``virNetworkLookupByName()``, ``virNetworkLookupByUUID()``, ``virConnectNumOfDefinedNetworks()``, ``virConnectListDefinedNetworks()``, ``virNetworkGetAutostart()``, ``virNetworkIsActive()``, ``virNetworkIsPersistent()``, ``virNetworkGetXMLDesc()``, and ``virDomainScreenshot()``, APIs have been implemented in the Hyper-V driver. * Support <teaming> element in plain <hostdev> devices This is useful when libvirt doesn't have the privileges necessary to set the hostdev device's MAC address (which is a necessary part of the alternate <interface type='hostdev'>). * Introduce ``<disk type='vhostuser'>`` support Introduces support for QEMU vhost-user-blk device that can be used to access storage exported via the vhost-user protocol by daemons such as the ``qemu-storage-daemon``. * **Bug fixes** * qemu: Fix disk quiescing rollback when creating external snapshots If the qemu guest agent call to freeze filesystems failed when creating an external snapshot with ``VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE`` flag the filesystems would be unconditionally thawed. This could cause problems when the filesystems were frozen by an explicit call to ``virDomainFSFreeze`` since the guest agent then rejects any further freeze attempts once are filesystems frozen, an explicit freeze followed by a quiesced snapshot would fail and thaw filesystems. Users are also encouraged to use ``virDomainFSFreeze/Thaw`` manually instead of relying on ``VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE`` if they need finer grained control. * cgroups: Fix how we setup and configure cgroups on hosts with systemd When libvirt is running on host with systemd we register every VM with machined which creates the VM root cgroup for us as well. Before this fix we were directly modifying files in the VM root cgroup which was incorrect because all the files are managed by systemd. The implication was that any change done by libvirt to cgroup attributes supported by systemd could be removed which happens for example by running ``systemctl daemon-reload``. To fix the issue libvirt now uses DBus calls for some of the cgroup attributes that distribute the resources proportionally to the cgroup siblings and for the rest we have a new sub-cgroup that libvirt can managed directly. For more details why this is necessary see `systemd cgroup <https://systemd.io/CGROUP_DELEGATION/>`_ documentation. * qemu: Fix swtpm device with aarch64 The TPM TIS device name for x86 is ``tpm-tis``, whereas for aarch64 it is ``tpm-tis-device``. Fix the use of TPM TIS device with aarch64 by using the proper device name when building the QEMU command line. * libxl: Fix domain shutdown Commit fa30ee04a2 introduced the possibility of a race between the shutdown and death threads used to process domain shutdown and death events from libxl. On normal domain shutdown the shutdown thread handles all aspects of shutting down and cleaning up the domain. The death thread is only used to handle out-of-band domain destruction and is inhibited when domain shutdown is under libvirt's control. The race is avoided by also inhibiting the death thread when libvirt starts the shutdown thread. v7.0.0 (2021-01-15) =================== * **Project governance** * Formal handover of release tarball signing Starting from libvirt-6.6 the release tarballs are signed by Jiří Denemark. Releases starting with 7.0 contain a note from the previous maintainer Daniel Veillard officially handing over the signing of packages so that the transition can be verified. * **New features** * nodedev: Add node device driver support for AP devices Add support for detecting and listing Adjunct Processor(AP) cards, AP queues and AP matrix devices (which are capable of MDEV) of a KVM host system in libvirt node device driver with correct object relationships. * qemu: Allow control of ``qcow2`` metadata cache In specific usecases such as when massive storage images are used it's possible to achieve better performance by increasing the metadata cache size. The new knob allows advanced users setting the size according to qemu's documentation to suit their image. * conf: Add support for keeping TPM emulator state Currently, swtpm TPM state file is removed when a transient domain is powered off or undefined. Add per-TPM emulator option ``persistent_state`` for keeping TPM state. * cpu_map: Add Snowridge CPU model It's supported in QEMU 4.1 and newer. * qemu: Add support for NFS disk protocol Implement support for the 'nfs' native protocol driver in the qemu driver. * **Improvements** * qemu: Discourage users from polling ``virDomainGetBlockJobInfo`` for block job completion Document that waiting for events is a more robust solution. * secret: Relax XML schema for the ``usage`` name of a ``secret`` Various bits of documentation of how to use libvirt with RBD volumes used an usage name which would not pass the XML validation. Relax the requirement to make such XMLs valid. * virnetdevopenvswitch: Various improvements The code that handles ``<interface type='vhostuser'/>`` was given various improvements. So far, libvirt assumed vhostuser interfaces are handled exclusively by OpenVSwitch and refused to start a guest if it was not so. Now a guest can be started successfully even if the interface is created by some other tool (e.g. ``dpdk-testpmd``). Also, the code that detects the interface name was adapted to new versions of OpenVSwitch and thus can detect name more reliably. * qemu: Report guest disks information in ``virDomainGetGuestInfo`` Libvirt is now able to report disks and filesystems from the guest's perspective (using guest agent). And with sufficiently new guest agent (5.3.0 or newer) the API also handles disks on CCW bus. * **Bug fixes** * qemu: Fix logic bug in inactive snapshot deletion This release fixes a bug introduced in libvirt-6.9 where libvirt's snapshot metadata would not be deleted on successful snapshot deletion. * qemu: Fix VMs with ``<iotune>`` on an empty cdrom Specifying ``<iotune>`` for an empty cdrom would prevent the VM from starting as qemu doesn't accept the tuning for an empty drive. We now postpone setting the parameters until a new media is inserted. * Avoid taking extra host memory when launching pSeries guests Under certain conditions, pSeries guests were being launched with more RAM than it was specified in the domain XML by the user. New pSeries domains created with libvirt 7.0.0 will always launch with the right amount of initial memory. Existing guests that migrate from an older libvirt version to 7.0.0 will not be affected by this change. * qemu: Don't cache NUMA caps ``virsh capabilities`` contains ``<topology/>`` section which reports NUMA topology among with amount of free hugepages per each NUMA node. However, these amounts were not updated between calls. * networkGetDHCPLeases: Handle leases with infinite expiry time Since libvirt-6.3.0 it is possible to configure expiry time for DHCP leases. If the expiry time was infinite then ``virsh net-dhcp-leases`` and NSS plugins refused to work. * qemu: Don't prealloc mem for real NVDIMMs If a real life NVDIMM is assigned to a guest via ``<memory model='nvdimm'/>`` then QEMU is no longer instructed to preallocate memory for it. This prevents unnecessary wear on the NVDIMM. * network: Introduce mutex for bridge name generation When new libvirt network is defined or created and the input XML does not contain any bridge name, libvirt generates one. However, it might have happened that the same name would be generated for different networks if two or more networks were defined/created at once. v6.10.0 (2020-12-01) ==================== * **Security** * qemu: Enable client TLS certificate validation by default for ``chardev``, ``migration``, and ``backup`` servers. The default value if qemu.conf options ``chardev_tls_x509_verify``, ``migrate_tls_x509_verify``, or ``backup_tls_x509_verify`` are not specified explicitly in the config file and also the ``default_tls_x509_verify`` config option is missing are now '1'. This ensures that only legitimate clients access servers, which don't have any additional form of authentication. * qemu: Introduce "migrate_tls_force" qemu.conf option The ``migrate_tls_force`` configuration option allows administrators to always force connections used for migration to be TLS secured as if the ``VIR_MIGRATE_TLS`` flag had been used. * **New features** * qemu: Implement OpenSSH authorized key file management APIs New APIs (``virDomainAuthorizedSSHKeysGet()`` and ``virDomainAuthorizedSSHKeysSet()``) and virsh commands (``get-user-sshkeys`` and ``set-user-sshkeys``) are added to manage authorized_keys SSH file for user. * hyperv: implement new APIs The ``virDomainGetMaxMemory()``, ``virDomainSetMaxMemory()``, ``virDomainGetSchedulerType()``, ``virDomainGetSchedulerParameters()``, ``virDomainGetSchedulerParametersFlags()``, ``virDomainGetVcpus()``, ``virDomainGetVcpusFlags()``, ``virDomainGetMaxVcpus()``, ``virDomainSetVcpus()``, and ``virDomainSetVcpusFlags()`` APIs have been implemented in the Hyper-V driver. * qemu: Add 'fmode' and 'dmode' options for 9pfs Expose QEMU's 9pfs 'fmode' and 'dmode' options via attributes on the 'filesystem' node in the domain XML. These options control the creation mode of files and directories, respectively, when using accessmode=mapped. It requires QEMU 2.10 or above. * qemu: support kvm-poll-control performance hint Implement the new KVM feature 'poll-control' to set this performance hint for KVM guests. It requires QEMU 4.2 or above. * **Improvements** * virsh: Support network disks in ``virsh attach-disk`` The ``virsh attach-disk`` helper command which simplifies attaching of disks without the need for the user to formulate the disk XML manually now supports network-backed images. Users can specify the protocol and host specification with new command line arguments. Please refer to the man page of virsh for further information. * **Bug fixes** * remote: fixed performance regression in SSH tunnelling The ``virt-ssh-helper`` binary introduced in 6.8.0 had very poor scalability which impacted libvirt tunnelled migration and storage volume upload/download in particular. It has been updated and now has performance on par with netcat. * **Removed features** * hyperv: removed support for the Hyper-V V1 WMI API This drops support for Windows Server 2008R2 and 2012. The earliest supported version is now Windows 2012R2. v6.9.0 (2020-11-02) =================== * **New features** * nodedev: Add support for channel subsystem (CSS) devices on S390 A CSS device is represented as a parent device of a CCW device. This support allows to create vfio-ccw mediated devices with ``virNodeDeviceCreateXML()``. * qemu: Implement memory failure event New event is implemented that is emitted whenever a guest encounters a memory failure. * qemu: Implement support for ``<transient/>`` disks VMs based on the QEMU hypervisor now can use ``<transient/>`` option for local file-backed disks to configure a disk which discards changes made to it while the VM was active. * hyperv: implement new APIs The ``virConnectGetCapabilities()``, ``virConnectGetMaxVcpus()``, ``virConnectGetVersion()``, ``virDomainGetAutostart()``, ``virDomainSetAutostart()``, ``virNodeGetFreeMemory()``, ``virDomainReboot()``, ``virDomainReset()``, ``virDomainShutdown()``, and ``virDomainShutdownFlags()`` APIs have been implemented in the Hyper-V driver. * bhyve: implement virtio-9p filesystem support Implement virito-9p shared filesystem using the ``<filesystem/>`` element. * qemu: Add support for vDPA network devices. VMs using the QEMU hypervisor can now specify vDPA network devices using ``<interface type='vdpa'>``. The node device APIs also now list and provide XML descriptions for vDPA devices. * cpu_map: Add EPYC-Rome CPU model It's supported in QEMU 5.0.0 and newer. * cpu: Add a flag for XML validation in CPU comparison The ``virConnectCompareCPU`` and ``virConnectCompareHypervisorCPU`` API now support the ``VIR_CONNECT_COMPARE_CPU_VALIDATE_XML`` flag, which enables XML validation. For virsh, this feature is enabled by passing the ``--validate`` option to the ``cpu-compare`` and ``hypervisor-cpu-compare`` subcommands. * qemu: Introduce virtio-balloon free page reporting feature Introduce the optional attribute ``free-page-reporting`` for virtio memballoon device. It enables/disables the ability of the QEMU virtio memory balloon to return unused pages back to the hypervisor. QEMU 5.1 and newer support this feature. * **Improvements** * qemu: Make 'cbitpos' & 'reducedPhysBits' attrs optional Libvirt probes the underlying platform in order to fill in these SEV attributes automatically before launching a guest. * util: support device stats collection for SR-IOV VF hostdev For SR-IOV VF hostdevs, libvirt now supports retrieving device traffic stats via the ``virDomainInterfaceStats`` API and ``virsh domifstat``. * logging: Allow disabling log rollover Set ``max_len=0`` in ``virtlogd.conf`` to disable log rollover. * qemu: Set noqueue qdisc for TAP devices Set ``noqueue`` instead of the former ``pfifo_fast`` queue discipline for TAP devices. It will avoid needless cost of host CPU cycles and thus improve performance. * qemu: virtiofs can be used without NUMA nodes Virtiofs is supported for the VM without NUMA nodes but configured with shared memory. * **Bug fixes** * hyperv: ensure WQL queries work in all locales Relying on the "Description" field caused queries to fail on non-"en-US" systems. The queries have been updated to avoid using localized strings. * rpc: Fix ``virt-ssh-helper`` detection libvirt 6.8.0 failed to correctly detect the availability of the new ``virt-ssh-helper`` command on the remote host, and thus always used the fallback instead; this has now been fixed. v6.8.0 (2020-10-01) =================== * **Security** * qemu: double free in qemuAgentGetInterfaces() in qemu_agent.c Clients connecting to the read-write socket with limited ACL permissions may be able to crash the libvirt daemon, resulting in a denial of service, or potentially escalate their privileges on the system. CVE-2020-25637. * **New features** * xen: Add ``writeFiltering`` attribute for PCI devices By default Xen filters guest writes to the PCI configuration space of a PCI hostdev, which may cause problems for some devices. The ``writeFiltering`` attribute of the device's ``<source>`` element can be used to disable the filtering and allow all guest writes to the configuration space. * bhyve: Support setting the framebuffer resolution Libvirt can now set the framebuffer's "w" and "h" parameters using the ``resolution`` element. * bhyve: Support VNC password authentication Libvirt can now probe whether the bhyve binary supports VNC password authentication. In case it does, a VNC password can now be passed using the ``passwd`` attribute on the ``<graphics>`` element. * remote: ``virt-ssh-helper`` replaces ``nc`` for SSH tunnelling Libvirt now provides a ``virt-ssh-helper`` binary on the server side. The libvirt remote client will use this binary for setting up an SSH tunnelled connection to hosts. If not present, it will transparently fallback to the traditional ``nc`` tunnel. The new binary makes it possible for libvirt to transparently connect across hosts even if libvirt is built with a different installation prefix on the client vs server. It also enables remote access to the unprivileged per-user libvirt daemons (e.g. using a URI such as ``qemu+ssh://hostname/session``). The only requirement is that ``virt-ssh-helper`` is present in ``$PATH`` of the remote host. * esx: implement few APIs The ``virConnectListAllNetworks()``, ``virDomainGetHostname()``, and ``virDomainInterfaceAddresses()`` (only for ``VIR_DOMAIN_INTERFACE_ADDRESSES_SRC_AGENT`` source) APIs were implemented in the esx driver. * **Improvements** * qemu: Allow migration over UNIX sockets QEMU migration can now be performed completely over UNIX sockets. This is useful for containerised scenarios and can be used in both peer2peer and direct migrations. * dbus: Use GLib implementation instead of libdbus Adopting GLib DBus implementation simplifies our code as libdbus provides low-level APIs where we had to have a lot of helper functions. With this change we also remove dependency on libdbus and possibly fix all the DBus related libvirtd crashes seen over the time. * Re-introduce NVDIMM auto-alignment for pSeries Guests The auto-alignment logic was removed in v6.7.0 in favor of requiring the size provided by the user to be already aligned; however, this had the unintended consequence of breaking some existing guests. v6.8.0 restores the previous behavior with an improvement: it also reflects the auto-aligned value in the domain XML. * qemu: Preserve qcow2 cluster size after external snapshots The new overlay image which is installed on top of the current chain when taking an external snapshot now preserves the cluster size of the original top image to preserve any performance tuning done on the original image. * **Bug fixes** * qemu: Various (i)SCSI backed hostdev fixes (i)SCSI backed hostdevs now work again with an arbitrarily long user-specified device alias and also honor the 'readonly' property after a recent rewrite. * **Removed features** * node_device: Remove HAL node device backend HAL is deprecated on all supported OS so there is no need to keep it in libvirt. udev backend is used on Linux OSes and devd can be eventually implemented as replacement for FreeBSD. v6.7.0 (2020-09-01) =================== * **Packaging changes** * Libvirt switch to Meson build system Libvirt abandoned autotools and switched to Meson build system. * **New features** * qemu: Add support for initiator IQN configuration for iSCSI hostdevs Similarly to iSCSI ``<disk>`` users can use an ``<initiator>`` element inside ``<hostdev>`` with the same format to configure the ``IQN`` value used by the qemu initiator when connecting to an iSCSI target. * xen: Add support for device model command-line passthrough Xen supports passing arbitrary arguments to the QEMU device model using the ``device_model_args`` setting in xl.cfg(5). The libvirt xen driver now supports this using ``<xen:commandline/>`` XML extensions. * shmem: Add support for shmem-{plain, doorbell} ``role`` option The ``role`` attribute controls how the domain behaves on migration. With ``role=master``, the guest will copy the shared memory on migration to the destination host. With ``role=peer``, the migration is disabled. * bhyve: Sound device support This feature allows to configure guest sound device using the ``<sound>`` element, and map it to the host sound device using the ``<audio>`` element. * **Improvements** * Allow sparse streams for block devices Sparse streams (e.g. ``virsh vol-download --sparse`` or ``virsh vol-upload --sparse``) now handle if one of the stream ends is a block device. * Remove NVDIMM auto-alignment for pSeries Guests This feature was introduced in libvirt v6.2.0 as part of the overall NVDIMM support for pSeries guests. The idea was to relieve the user from knowing ppc64 alignment details, but the end result is that we ended up with inconsistencies between domain XML and actual NVDIMM size the guest is using. To promote consistency between domain XML and the guest, unaligned NVDIMM sizes for pSeries guests will now be forbidden and no size auto-alignment will be made. Instead, libvirt will suggest an aligned round up size for the user. * apparmor: Several improvements Add support for virtiofs filesystem and allow QEMU to load old shared objects after upgrade. * **Bug fixes** * virdevmapper: Deal with kernels without DM support In the previous release libvirt dropped libdevmapper in favor of its own implementation. However, it failed to deal correctly with kernels that either don't have device mapper enabled or where the dm-mod module is not loaded yet. This is now fixed. * resctrl: Use exclusive lock for /sys/fs/resctrl When two or more domains were attempted to start at once, due to a bug in implementation, resctrl was not locked properly and thus threads did not mutually exclude with each other resulting in not setting requested limitations. * mdev: Fix daemon crash when reattaching mdevs on assignment conflict If there's a list of mdevs to be assigned to a domain, but one of them (NOT the first) is already assigned to a different domain then libvirtd would crash. This is now fixed. * Fix logic in setting COW flag on btrfs When COW is not explicitly requested to be disabled or enabled, then libvirt should do nothing on non-BTRFS file systems. * Avoid crash due to race in glib event loop code Libvirt switched to glib event loop in 6.1.0 but it was also tickling a bug in glib code leading to the daemon crash. Libvirt way of calling glib was changed so the daemon crashes no more. * virdevmapper: Handle kernel without device-mapper support In the previous release, Libvirt dropped libdevmapper in favor of its own implementation. But the implementation did not handle kernels without device-mapper support. This is now fixed. * remove autogenerated macvtap names from migration XML Autogenerated macvtap device names were being left in the migration XML, which could result in libvirt erroneously deleting the macvtap device of a different guest in the aftermath of failing to restart the guest on the destination host. Removing the autogenerated names avoids this. v6.6.0 (2020-08-02) =================== * **New features** * Allow configuring of ACPI NUMA HMAT Libvirt allows configuring ACPI Heterogeneous Memory Attribute Table to hint software running inside the guest on optimization. * esx: Add a ``type`` attribute for mac addresses. This attribute allows (when set to ``static``) ignoring VMWare checks of the MAC addresses that would generate a new one if they were in its OUI (00:0c:29). * conf: add control over COW for storage pool directories The storage pool code now attempts to disable COW by default on btrfs, but management applications may wish to override this behaviour. This is now possible via new ``cow`` element. * **Improvements** * esx: Change the NIC limit for recent virtualHW versions Specifying a virtualHW version greater or equal to 7 (ESXi 4.0) will allow you to use up to 10 NICs instead of 4 as it was previously. * qemu: Support encrypted TLS keys for NBD disks The secret key used for disks can now be encrypted similarly to TLS keys used for migration, chardev and others. * qemu: ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` can now be registered for ``<mirror>`` The event can now be used also for block copy destinations by using the index of the ``<mirror>`` image. * qemu: consider available CPUs in ``vcpupin/emulatorpin`` output This patch changes the default bitmap of ``vcpupin`` and ``emulatorpin``, in the case of domains with static vcpu placement, all available CPUs instead of all possible CPUs are returned making these APIs consistent with the behavior of ``vcpuinfo``. * **Bug fixes** * virdevmapper: Don't use libdevmapper to obtain dependencies When building domain's private ``/dev`` in a namespace, libdevmapper was consulted for getting full dependency tree of domain's disks. However, this meant that libdevmapper opened ``/dev/mapper/control`` which wasn't closed and was leaked to QEMU. CVE-2020-14339 * qemu: Report correct ``index`` in ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` Starting from libvirt 5.10 with QEMU 4.2 the ``VIR_DOMAIN_EVENT_ID_BLOCK_THRESHOLD`` event would report incorrect device ``index`` when reported for an image from the backing chain of a disk. * qemu: Don't fail active layer block commit or block copy in certain cases Starting from libvirt-6.5 an active layer block commit or a block copy could fail if the same destination was used more than once. * qemu: Don't change ownership of restore file When restoring a domain from a file, Libvirt no longer changes its ownership. * qemu: Set SPAPR TPM default to 2.0 and prevent 1.2 choice The firmware (SLOF) on QEMU for ppc64 does not support TPM 1.2, so prevent the choice of TPM 1.2 when the SPAPR device model is chosen and use a default of '2.0' (TPM 2) for the backend. * qemu: Do not set ``//cpu/@migratable`` for running domains Libvirt release of 6.4.0 started to fill the default value for ``//cpu/@migratable`` attribute according to QEMU support. However, active domains either have the migratable attribute already set or they were started with older Libvirt which doesn't support the attribute. v6.5.0 (2020-07-03) =================== * **New features** * Allow firmware blobs configuration QEMU offers a way to tweak how firmware configures itself and/or provide new configuration blobs. New ``<sysinfo/>`` type is introduced that will hold these new blobs. It's possible to either specify new value as a string or provide a filename which contents then serve as the value. * nodedev: Add ability to create mediated devices Mediated devices can now be created with ``virNodeDeviceCreateXML()``. This functionality requires the ``mdevctl`` utility to be installed. The XML schema for node devices was expanded to support attributes for mediated devices. * QEMU: add TPM Proxy device support libvirt can now create guests using a new device type called "TPM Proxy". The TPM Proxy connects to a TPM Resource Manager present in the host, enabling the guest to run in secure virtual machine mode with the help of an Ultravisor. Adding a TPM Proxy to a pSeries guest brings no security benefits unless the guest is running on a PPC64 host that has Ultravisor and TPM Resource Manager support. Only one TPM Proxy is allowed per guest. A guest using a TPM Proxy device can instantiate another TPM device at the same time. This device is supported only for pSeries guests via the new 'spapr-tpm-proxy' model of the TPM 'passthrough' backend. * virhook: Support hooks placed in several files Running all scripts from directory /etc/libvirt/hooks/<driver>.d in alphabetical order. Hook script in old place will be executed as first for backward compatibility. * qemu: Add support for migratable host-passthrough CPU QEMU 2.12 made it possible for guests to use a migration-friendly version of the host-passthrough CPU. This feature is now exposed by libvirt. * **Improvements** * network: Support NAT with IPv6 It's now possible to use ``<nat ipv6="yes"/>`` in a libvirt network. * qemu: Auto-fill NUMA information for incomplete topologies If the NUMA topology is not fully described in the guest XML, libvirt will complete it by putting all unspecified CPUs in the first NUMA node. This is only done in the QEMU binary itself supports disjointed CPU ranges for NUMA nodes. * qemu: Assign hostdev-backed interfaces to PCIe slots All SR-IOV capable devices are PCIe, so when their VFs are assigned to guests they should end up in PCIe slots rather than conventional PCI ones. * **Bug fixes** * qemu: fixed crash in ``qemuDomainBlockCommit`` This release fixes a regression which was introduced in libvirt v6.4.0 where libvirtd always crashes when a block commit of a disk is requested. * qemu: fixed zPCI address auto generation on s390 Removes the correlation between the zPCI address attributes uid and fid. Fixes the validation and autogeneration of zPCI address attributes. * qemu: Skip pre-creation of NVMe disks during migration libvirt has no way to create NVMe devices on the target host, so it now just makes sure they exist and let the migration proceed in that case. v6.4.0 (2020-06-02) =================== * **New features** * qemu: Add support for pvscsi controllers pvscsi is the VMware paravirtualized SCSI controller, which has been supported in QEMU for a number of years. * cpu: Report model information for ARM CPUs ``virsh capabilities`` will now include information about the host CPU when run on ARM machines. * qemu: support network interface downscript QEMU has the ability to run a script when a NIC is brought up and down. Libvirt only enables use of the up script. Now add support for postscript when NIC is down/detached. * **Improvements** * qemu: stricter validation for disk type='lun' The 'lun' type is meant for SCSI command passthrough, which can't be achieved if qemu's block layer features are used. Disk type='lun' is now allowed only when the format is 'raw' and no other block layer features are requested. * qemu: auto-fill of incomplete NUMA topologies Domains with incomplete NUMA topologies, where the sum of vCPUs in all NUMA cells is less than the total of vCPUs, will get their first NUMA cell to be auto-filled with the remaining vCPUs. This behavior reproduces what QEMU already does in these cases. Users are encouraged to provide complete NUMA topologies to avoid unexpected changes in the domain XML. * Cooperlake x86 CPU model is added * **Bug fixes** * qemu: fixed regression in network device hotplug with new qemu versions Starting from QEMU-5.0 it's required to conform to strict schema when hotplugging network devices. Libvirt didn't conform to the schema so in versions prior to 6.4.0 network device hotplug fails in certain cases. This version fixes it and adds stricter testing to prevent further issues. * remote: Look up libxl driver correctly This makes ``xen://`` connection URIs usable in split daemon mode. * systemd: Start libvirtd after firewalld/iptables services This solves an issue where iptables rules and chains created by libvirtd would get removed by a service started after it. * network: Re-create iptables chains on firewalld restart firewalld resets all iptables rules and chains on restart, and this includes deleting those created by libvirt. * qemu: reject readonly attribute for virtiofs virtiofs does not yet support read-only shares. v6.3.0 (2020-05-05) =================== * **New features** * qemu: support disabling hotplug/unplug of PCIe devices libvirt can now set the "hotplug" option for pcie-root-ports and pcie-switch-downstream-ports, which can be used to disable hotplug/unplug of devices from these ports (default behavior is for these controllers to accept all hotplug/unplug attempts, but this is often undesirable). * vbox: added support for version 6.0 and 6.1 APIs libvirt can now support use of the VirtualBox 6.0 and 6.1 APIs. This is compile tested only, so we are looking for feedback from users on how well it works in practice. * xen: Add support for 'e820_host' hypervisor feature ``e820_host`` is a Xen-specific option only available for PV guests. When enabled it provides the guest with a virtual e820 memory map based on the host one. It must be enabled to allow hotplugging PCI devices to PV guests, particularly when memory ballooning is enabled. * xen: Add support for 'passthrough' hypervisor feature ``passthrough`` is a Xen-specific option new to Xen 4.13 that enables PCI passthrough for guests. It must be enabled to allow hotplugging PCI devices. * qemu: support async IO mode 'io_uring' Linux 5.1 introduces a fast and efficient async IO interface io_uring, then qemu implements that in version 5.0. It could be used in file, host device and host cdrom backend by ``io='io_uring'`` of disk XML. * Lease time option included for network DHCP settings Users can now configure expiry time for leases for networks where libvirt manages DHCP. The time can be specified for whole range and/or fine tuned per individual host. * qemu: Implement pSeries Spectre mitigation features Users can now setup the following capabilities of pSeries guests: CFPC (Cache Flush on Privilege Change), SBBC (Speculation Barrier Bounds Checking) and IBS (Indirect Branch Speculation). * qemu: Add support for virtio packed option The ``packed`` attribute controls if QEMU should try to use packed virtqueues. Possible values are ``on`` or ``off``. * **Improvements** * qemu: Allow checkpoint redefine for offline VMs Skip the liveness and capability checks when redefining checkpoints as we don't need QEMU interactions to update the metadata. * daemons: Improve timeout handling Daemons now support ``--timeout 0`` which suppresses daemon killing after given time of inactivity. * qemu: Add support for 'multidevs' option This option prevents misbehaviours on guest if a QEMU 9pfs export contains multiple devices, due to the potential file ID collisions this otherwise may cause. * **Bug fixes** * qemu: Various embed driver fixes When using shared resources from embed driver (e.g. hugepages, machined, etc.) libvirt now generates an unique handler that is not conflicting with other embed drivers or system or session daemons. * cpu: Distinguish Cascadelake-Server from Skylake-Server Libvirt now properly detects Cascadelake-Server and Skylake-Server processors which differ only in stepping. * qemu: Fix domain restore from a block device When using namespaces, libvirt was unable to restore a domain from a block device because libvirt tried to relabel the device inside the namespace while QEMU was given FD to the block device in the host. * node_device_udev: Handle move events Libvirt now handles ``move`` event which is emitted on a NIC rename. * qemu: Fix capabilities probing with TCG Libvirt no long assumes TCG is always available. It now detects whether QEMU supports TCG and reports it accordingly. * **Removed features** * vbox: removed support for version 5.0 and 5.1 APIs libvirt no longer supports use of VirtualBox 5.0 and 5.1 since these versions reached their end of life on 2017/05 and 2018/04 respectively. v6.2.0 (2020-04-02) =================== * **New features** * qemu: NVDIMM support for pSeries guests QEMU 5.0 implements NVDIMM memory support for pSeries guests. This is done by adding an 'uuid' element in the memory XML, which can either be provided in the XML or, if omitted, generated automatically. * qemu: Add virtiofs support This feature, introduced in QEMU 4.2, is a more modern alternative to virtio-9p, which is exposed through the same ``<filesystem/>`` element. * admin: Support reloading TLS certificates After renewing TLS certificates, it was usually necessary to restart libvirtd for the new ones to be loaded: now the same result can be obtained without restarting the daemon by using ``virt-admin server-update-tls`` . * **Removed features** * Removed support for INI style of comments With switching of our internal code to GLib, parsing of client authentication config files is handed over to GLib which does not support ``INI`` style of comments starting with a semicolon ( ``;`` ). Use number sign ( ``#`` ) instead. * **Improvements** * qemu: Don't compare local and remote hostnames on migration This check was introduced to prevent same-host migration, but did not work as expected when multiple libvirtd instances were running on the same host but in different containers. With this release, the host UUID (which should be unique to the container) is checked instead. * qemu: Use per-VM event loops Instead of using a single even loop to process communication with the QEMU monitor and guest agent, create a separate one for each VM. This helps with scalability and prevents scenarios where a single malfunctioning VM could affect all those running on the same host. * qemu: Support migration with SLIRP helper interface With QEMU 5.0, a new D-Bus backend allows migration of external processes. When needed, libvirt will start a per-vm D-Bus bus, and migrate the slirp-helper along with QEMU. * **Bug fixes** * qemu: Open backing chain late for shallow block copy reusing external images With introduction of -blockdev for QEMU storage configuration in libvirt-5.10 we've started opening the backing chain of the destination/mirror of a virDomainBlockcopy started with VIR_DOMAIN_BLOCK_COPY_REUSE_EXT | VIR_DOMAIN_BLOCK_COPY_SHALLOW flags when starting the job rather than when virDomainBlockJobAbort with VIR_DOMAIN_BLOCK_JOB_ABORT_PIVOT is issued. For users depending on this undocumented quirky pre-blockdev behaviour this caused a regression as the backing chain could not be modified while the copy of the top image was progressing due to QEMU image locking. Note that this fix also requires qemu-5.0 while -blockdev is used starting from QEMU-4.2. * Don't generate machine names containing dots Even though the guest name containing dots is not a problem for libvirt itself, we need to strip them out when registering with machined because of the latter's requirements. v6.1.0 (2020-03-03) =================== * **New features** * qemu: new rng backend type: builtin It implements qemu builtin rng backend. That uses getrandom syscall to generate random, no external rng source needed. Available since QEMU 4.2. * support for virtio+hostdev NIC <teaming> QEMU 4.2.0 and later, combined with a sufficiently recent guest virtio-net driver (e.g. the driver included in Linux kernel 4.18 and later), supports setting up a simple network bond device comprised of one virtio emulated NIC and one hostdev NIC (which must be an SRIOV VF). (in QEMU, this is known as the "virtio failover" feature). The allure of this setup is that the bond will always favor the hostdev device, providing better performance, until the guest is migrated - at that time QEMU will automatically unplug the hostdev NIC and the bond will send all traffic via the virtio NIC until migration is completed, then QEMU on the destination side will hotplug a new hostdev NIC and the bond will switch back to using the hostdev for network traffic. The result is that guests desiring the extra performance of a hostdev NIC are now migratable without network downtime (performance is just degraded during migration) and without requiring a complicated bonding configuration in the guest OS network config and complicated unplug/replug logic in the management application on the host - it can instead all be accomplished in libvirt with the interface <teaming> subelement "type" and "persistent" attributes. * support BR_ISOLATED flag for guest interfaces attached to a Linux host bridge Since Linux kernel 4.18, the Linux host bridge has had a flag BR_ISOLATED that can be applied to individual ports. When this flag is set for a port, traffic is blocked between that port and any other port that also has the BR_ISOLATED flag set. libvirt domain interface config now supports setting this flag via the <port isolated='yes'/> setting. It can also be set for all connections to a particular libvirt network by setting the same option in the network config - since the port for the host itself does not have BR_ISOLATED set, the guests can communicate with the host and the outside world, but guests on that network can't communicate with each other. This feature works for QEMU and LXC guests with interfaces attached to a Linux host bridge. * qemu: Introduce the 'armvtimer' timer type QEMU 5.0 introduces the ability to control the behavior of the virtual timer for KVM ARM/virt guests, and this new timer type exposes the same capability to libvirt users. * qemu: Storage configuration improvements Libvirt now accepts ``<backingStore type='volume'>`` and allows specifying the offset and size of the image format container inside the storage source via the ``<slices>`` subelement. * qemu: Introduce the 'tpm-spapr' TPM model This device, available starting from QEMU 5.0, is limited to pSeries guests. * qemu: support Panic Crashloaded event handling The pvpanic device now supports a 'crashloaded' event, which is emitted when a guest panic has occurred but has already been handled by the guest itself. * qemu: Implement virDomainGetHostnameFlags The ``--source`` argument to ``virsh domhostname`` can be used to specify what data source to use for the domain hostnames. Currently, in addition to the 'agent', libvirt can also use 'lease' information from dnsmasq to get the hostname. * **Improvements** * qemu: Image format probing is allowed in certain cases To resolve regressions when users didn't specify the backing image format in the overlay, libvirt now probes the format in certain secure scenarios which fixes a few common existing cases. Additionally the knowledge base was extended to provide more information on how to rectify the problem. * qemu: Support "dies" in CPU topology This CPU topology concept, new in QEMU 4.1.0, sits between the existing "socket" and "core". * libxl: Add support for Credit2 scheduler parameters * lxc: Add support LXC 3 network configuration format * **Bug fixes** * conf: Do not generate machine names ending with a dash Recent systemd versions do not allow them. * **Packaging changes** * use of gnulib has been completely eliminated Historically libvirt has embedded gnulib to provide fixes for various platform portability problems. This usage has now been eliminated and alternative approaches for platform portability problems adopted where required. This has been validated on the set of platforms covered by automated CI build testing. Other modern Linux distros using glibc are expected to work. Linux distros using non-glibc packages, and other non-Linux platforms may encounter regressions when building this release. Please report any build problems encountered back to the project maintainers for evaluation. v6.0.0 (2020-01-15) =================== * **Packaging changes** * support for python2 is removed Libvirt is no longer able to be built using the Python 2 binary. Python 3 must be used instead. * docs: the python docutils toolset is now required The use of rst2html has been introduced for the website build process since docs are now being written in the RST as an alternative to HTML. * **New features** * new PCI hostdev address type: unassigned A new PCI hostdev address type 'unassigned' is introduced. An unassigned PCI hostdev behaves like any regular PCI hostdev inside Libvirt, but it is not usable by the guest. This gives the user a new option to manage the binding of PCI devices via Libvirt, declaring PCI hostdevs in the domain XML but allowing just a subset of them to be assigned to the guest. * Provide init scripts for sub-deaemons So far libvirt shipped systemd unit files for sub-daemons. With this release, init scripts are available too. Package maintainers can choose which one to install via ``--with-init-script`` configure option. * qemu: Support cold-unplug of sound devices * qemu: Implement VIR_MIGRATE_PARAM_TLS_DESTINATION This flag, which can be enabled using ``virsh`` 's ``--tls-destination`` option, allows migration to succeed in situations where there is a mismatch between the destination's hostname and the information stored in its TLS certificate. * qemu: Support reporting memory bandwidth usage stats Implement Intel RDT-MBM in libvirt. The stats can be obtained via ``virsh domstats --memory`` . * qemu: Allow accessing NVMe disks directly Before this release there were two ways to configure a NVMe disk for a domain. The first was using <disk/> with the <source/> pointing to the ``/dev/nvmeXXXX`` . The other was using PCI assignment via <hostdev/> element. Both have their disadvantages: the former adds latency of file system and block layers of the host kernel, the latter prohibits domain migration. In this release the third way of configuring NVMe disk is added which combines the advantages and drops disadvantages of the previous two ways. It's accessible via <disk type='nvme'/>. * **Removed features** * 'phyp' Power Hypervisor driver removed The 'phyp' Power Hypervisor driver has not seen active development since 2011 and does not seem to have any real world usage. It has now been removed. * **Improvements** * qemu: xz save image compression is faster When using the xz format to compressed virtual machine saved state images, the "-3" compression level preset is now used. This results in slightly larger files, but with a massively reduced time to compress. The xz format offers the best compression level for saved state images, albeit still with the slowest running time. For the fastest possible running time, at cost of the larest compressed size, lzop should be used. * domain: Improve job stat handling It is now possible to retrieve stats for completed and failed jobs. * qemu: Don't hold monitor and agent job at the same time Before this change, a malicious (or buggy) ``qemu-guest-agent`` running in the guest could make other libvirt APIs unavailable for an unbounded amount of time. * **Bug fixes** * qemu: Report error if backing image format is not specified explicitly For a long time libvirt was assuming that a backing file is RAW when the format was not specified. This didn't pose a problem until blockdev support was enabled in last release. Libvirt now requires that the format is specified in the image metadata or domain XML and the VM will refuse to start otherwise. Additionally the error message now links to the knowledge base which summarizes how to fix the images. * qemu: Fix non-shared storage migration over NBD * qemu: Generate a single MAC address for hotplugged network devices Since libvirt 4.6.0, when hotplugging a network device that didn't have a MAC address already assigned by the user, two separate addresses would be generated: one for the live configuration, which would show up immediately, and one for the inactive configuration, which would show up after the first reboot. This situation was clearly undesirable, so a single MAC address is now generated and used both for the live configuration and the inactive one. v5.10.0 (2019-12-02) ==================== * **New features** * qemu: Introduce support for ARM CPU features The only features supported at the moment are SVE vector lengths, which were introduced in QEMU 4.2.0. * qemu: Support boot display for GPU mediated devices Until now, GPU mediated devices generally did not show any output until the guest OS had initialized the vGPU. By specifying the ``ramfb`` attribute, QEMU can be configured to use ramfb as a boot display for the device: this allows for display of firmware messages, boot loader menu, and other output before the guest OS has initialized the vGPU. * Add API to change the response timeout for guest agent commands By default, when a command is sent to the guest agent, libvirt waits forever for a response from the guest agent. If the guest is unresponsive for any reason, this can block the calling thread indefinitely. By setting a custom timeout using ``virDomainAgentSetResponseTimeout()`` , API users can change this behavior. * **Improvements** * Devices CGroup v2 support Libvirt supported all controllers of CGroup v2 but the devices controller which is implemented in this release. * Cold plug of sound device The QEMU driver now can handle cold plug of ``<sound/>`` devices. * Probe for default CPU types With QEMU 4.2.0 we can probe for the default CPU model used by QEMU for a particular machine type and store it in the domain XML. This way the chosen CPU model is more visible to users and libvirt will make sure the guest will see the exact same CPU after migration. * Adaptation to qemu's blockdev QEMU introduced a new way of specifying disks on the command line which enables fine-grained control over the block stack. Libvirt has adapted to this. * **Refactors** * More GLib integration More patches were merged that replace our internal functions with GLib ones. Also some effort was invested in replacing gnulib modules with GLib functions. * Rewrite of Perl scripts into Python Libvirt used Perl scripts to check for coding style, generate some code and things like that. To bring the number of languages used down, these scripts were rewritten into Python. * **Bug fixes** * Warn verbosely if using old loader:nvram pairs Some distributions still use ``--with-loader-nvram`` or ``nvram`` variable in qemu.conf. This is now discouraged in favour of FW descriptors. However, instead of silently ignoring user's config, libvirt warns if outdated config is detected. * Drop pconfig from Icelake-Server CPU model The pconfig feature was enabled in QEMU by accident in 3.1.0. All other newer versions do not support it and it was removed from the Icelake-Server CPU model in QEMU. * Wait longer for device removal confirmation on PPC64 After sending device hot unplug request to QEMU, libvirt waits up to 5 seconds for qemu to confirm the device removal. On some architectures (like PPC64) this can take longer time and libvirt now reflects that. * Forcibly create nodes in domain's namespace The QEMU driver starts a domain in a namespace with private ``/dev`` and creates only those nodes there which the domain is configured to have. However, it may have happened that if a node changed its minor number this change wasn't propagated to the namespace. * Various AppArmor bugfixes The AppArmor driver now knows how to handle ``<shmem/>`` devices and also snapshotting more disks at once. * Improved video model autoselection If a graphics device was added to XML that had no video device, libvirt automatically added a video device which was always of type 'cirrus' on x86_64, even if the underlying qemu didn't support cirrus. Libvirt now bases the decision on qemu's capabilities. v5.9.0 (2019-11-05) =================== * **Packaging changes** * Start linking against GLib and using its features Up until now, libvirt has been dealing with platform portability and the lack of certain features in libc by using gnulib and implementing its own functions and data structures respectively; going forward, it will prefer the facilities offered by GLib instead. * Stop distributing generated documentation Most downstreams already patch the libvirt source to some extent, so this change will probably not affect them. * Rewrite several Perl scripts in Python Phasing out Perl usage is part of the project strategy. * **New features** * qemu: Introduce a new video model of type 'ramfb' Introduce a new video model type to the domain XML that supports the ``ramfb`` standalone device in qemu. * qemu: Implement the ccf-assist pSeries feature Users can now decide whether ccf-assist (Count Cache Flush Assist) support should be available to pSeries guests. * Xen: Support specifying ACPI firmware path The libxl driver now supports specifying an ACPI firmware path using the ``acpi`` element. * qemu: Support specifying resolution for video devices * **Removed features** * logging: Drop support for including stack traces This feature was intended to aid debugging, but in practice it resulted in logs that were too verbose to be useful and also resulted in a significant performance penalty. * **Improvements** * qemu: Implement CPU comparison/baseline on s390x This functionality has been historically limited to x86_64, but it's now available on s390x too. * **Bug fixes** * lib: autostart objects exactly once If libvirtd or any of the sub-daemons is started with socket activation then objects might be autostarted more than once. For instance, if a domain under ``qemu:///session`` URI is mark as autostarted and the session daemon is started then the domain is started with it. If user shuts the domain down and the session daemon is started again, the user's wish to keep the domain shut off is ignored and the domain is autostarted again. This is now fixed. * qemu: Properly advertise bochs-display availability Support for ``bochs-display`` was introduced in libvirt 5.6.0, but until now the model was not listed in the domain capabilities. * security: Don't remember labels for TPM devices Due to the way they're implemented in the kernel, trying to remember labels for TPM devices makes it impossible to use them. * security: Properly rollback after failure in a stacked driver When multiple security drivers are involved, failure in one of them would result in only the corresponding changes being rolled back, leaving the ones performed by drivers that had been activated earlier in place. All changes are rolled back now. * Fix build with musl libc * Improve compatibility with non-bash shells v5.8.0 (2019-10-05) =================== * **New features** * qemu: Support use of precreated tap/macvtap devices by unprivileged libvirtd It is now possible for an unprivileged libvirtd to make use of tap and macvtap devices that were previously created by some other entity. This is done by setting ``managed='no'`` along with the device name in the ``target`` subelement of ``<interface type='ethernet'>`` . * qemu: Support vhost-user-gpu Support for running virtio GPUs in separate processes with vhost-user backend. It requires QEMU newer than 4.1. * Introduce virConnectSetIdentity API When split daemons are in use, this API is used to forward uid, gid and SELinux info from ``virproxyd`` to other driver daemons such as ``virtqemud`` . * **Improvements** * qemu: Support running SLIRP networking in a separate process User can configure the slirp-helper path in ``qemu.conf`` . It will start a slirp-helper process to provide SLIRP networking when the VM is started with network interface "user". That will allow stricter security policies for QEMU SLIRP network. * **Removed features** * Remove xenapi driver The xenapi driver is removed since it has not received any significant development since its initial contribution nine years ago and has no known user base. v5.7.0 (2019-09-03) =================== * **New features** * qemu: Support Direct Mode for Hyper-V Synthetic timers The QEMU driver now supports Direct Mode for Hyper-V Synthetic timers for Hyper-V guests. * lib: Add virDomainGetGuestInfo() This API is intended to aggregate several guest agent information queries and is inspired by stats API ``virDomainListGetStats()`` . It is anticipated that this information will be provided by a guest agent running within the domain. It's exposed as ``virsh guestinfo`` . * Experimental split of libvirtd into separate daemons The big monolithic libvirtd daemon can now be replaced by smaller per-driver daemons. The new split daemons are considered experimental at this time and distributions are encouraged to continue using the traditional libvirtd by default. * qemu: Support kvm-hint-dedicated performance hint With ``<hint-dedicated state='on'/>`` and ``<cpu mode='host-passthrough'/>`` , it allows a guest to enable optimizations when running on dedicated vCPUs. QEMU newer than 2.12.0 and kernel newer than 4.17 are required. * **Removed features** * Remove KVM assignment support The KVM style of PCI device assignment was removed from the kernel in version 4.12.0 after being deprecated since 4.2.0. Libvirt defaults to VFIO for a long time. Remove support for KVM device assignment from libvirt too. * libxml: min required libxml is now 2.9.1 Support for building with libxml versions older than 2.9.1 has been dropped. * **Improvements** * virsh: Support setting bandwidth in migrate subcommand In addition to postcopy bandwidth, the ``virsh migrate`` subcommand now supports specifying precopy bandwidth with the ``--bandwidth`` parameter. * libxl: Implement domain metadata getter/setter The libxl driver now supports ``virDomainGetMetadata()`` and ``virDomainSetMetadata()`` APIs. * test driver: Expand API coverage Additional APIs have been implemented in the test driver. * Report RNG device in domain capabilities XML Libvirt now reports if RNG devices are supported by the underlying hypervisor in the domain capabilities XML. * Stop linking NSS plugins with libvirt.so This reduces the amount of code and 3rd party libraries are that loaded into all processes. * Split the setuid virt-login-shell binary into two pieces The setuid virt-login-shell binary is now a tiny shim that sanitizes the process execution environment variables and arguments, before launching the trusted virt-login-shell-helper binary. * qemu: Allow migration with disk cache on When QEMU supports flushing caches at the end of migration, we can safely allow migration even if ``disk/driver/@cache`` is neither ``none`` nor ``directsync`` . * **Bug fixes** * Various security label remembering fixes In the previous release libvirt introduced remembering of original owners and SELinux labels on files. However, the feature did not work properly with snapshots, on migrations or on network filesystems. This is now fixed. * Allow greater PCI domain numbers Libvirt used to require PCI domain number to be not greater than 0xFFFF. The code was changed to allow 32 bits long numbers. * Various D-Bus fixes When D-Bus is not available, libvirt was reporting random errors. These are now gone. * Prefer read-only opening of PCI config files When enumerating PCI bus, libvirt opens config files under ``sysfs`` mount and parses them to learn various aspects of the device (e.g. its capabilities). Only in a very limited number of cases it is actually writing into the file. However, it used to open the file also for writing even if it was only reading from it. * Fix AppArmor profile Since the ``5.6.0`` release, libvirt uses ``procfs`` to learn the list of opened file descriptors when spawning a command. However, our AppArmor profile was not allowing such access. * Don't block storage driver when starting or building a pool Starting or building a storage pool can take a long time to finish. During this time the storage driver was blocked and thus no other API involving the storage driver could run. This is now fixed. v5.6.0 (2019-08-05) =================== * **New features** * qemu: Introduce a new video model of type 'bochs' Introduce a new video model type that supports the ``bochs-display`` device that was added in qemu version 3.0. * api: new virDomainCheckpoint APIs Introduce several new APIs for creating and managing checkpoints in the test and qemu drivers (the latter requires qcow2 images). Checkpoints serve as a way to tell which portions of a disk have changed since a point in time. * qemu: Add support for overriding max threads per process limit systemd-based systems impose a limit on the number of threads a process can spawn, which in some cases can be exceeded by QEMU processes running VMs. Add a ``max_threads_per_process`` option to qemu.conf to override the system default. * Remember original owners and SELinux labels of files When a domain is starting up libvirt changes DAC and SELinux labels so that domain can access it. However, it never remembered the original labels and therefore the file was returned back to ``root:root`` . With this release, the original labels are remembered and restored properly. * network: Allow passing arbitrary options to dnsmasq This works similarly to the existing support for passing arbitrary options to QEMU, and just like that feature it comes with no support guarantees. * **Removed features** * xen: Remove sxpr config support Remove the sxpr style config parser and formatter a year after the xend driver was removed. * **Improvements** * qemu: Allow XML validation for snapshot creation Add flag ``VIR_DOMAIN_SNAPSHOT_CREATE_VALIDATE`` to validate snapshot input XML. For virsh, users can use it as ``virsh snapshot-create --validate`` . * Support encrypted soft TPM A soft TPM backend could be encrypted with passphrase. Now libvirt supports using a ``secret`` object to hold the passphrase, and referring to it via the ``encryption`` element of the TPM device. * test driver: Expand API coverage Additional APIs have been implemented in the test driver. * Implement per-driver locking Drivers now acquire a lock when they're loaded, ensuring that there can never be two instances of the same driver active at a time. * nss: Report newer addresses first In some cases, a guest might be assigned a new IP address by DHCP before the previous lease has expired, in which case the NSS plugin will correctly report both addresses; many applications, however, ignore all addresses but the first, and may thus end up trying to connect using a stale address. To prevent that from happening, the NSS plugin will now always report the newest address first. * util: Optimize mass closing of FDs when spawning child processes When the limit on the number of FDs is very high, closing all unwanted FDs after calling ``fork()`` can take a lot of time and delay the start of the child process. libvirt will now use an optimized algorithm that minimizes such delays. * **Bug fixes** * logging: Ensure virtlogd rollover takes priority over logrotate virtlogd implements its own rollover mechanism, but until now logrotate could end up acting on the logs before virtlogd had a chance to do so itself. v5.5.0 (2019-07-02) =================== * **Security** * api: Prevent access to several APIs over read-only connections Certain APIs give root-equivalent access to the host, and as such should be limited to privileged users. CVE-2019-10161, CVE-2019-10166, CVE-2019-10167, CVE-2019-10168. * **New features** * qemu: Support SMMUv3 IOMMU SMMUv3 is an IOMMU implementation for ARM virt guests. * network: Introduce the network port API This new public API can be used by virtualization drivers to manage network resources associated with guests, and is a further step towards splitting libvirtd into multiple daemons. * **Removed features** * qemu: Remove support for virDomainQemuAttach and virConnectDomainXMLFromNative APIs The qemu implementations for the APIs mentioned above were removed and the APIs now return an error. The implementation was stale for a long time and did not work with modern QEMU command lines, generated from libvirt or otherwise. * Stop supporting migration of config files from pre-XDG layout The new layout was introduced with libvirt 0.9.13 (Jul 2012). * Remove Avahi mDNS support This feature was never used outside of virt-manager, which has itself stopped using it a while ago. * **Improvements** * sysinfo: Report SMBIOS information on aarch64 While SMBIOS support has historically been limited to x86_64, modern aarch64 machines often offer access to the same information as well, and libvirt now exposes it to the user when that's the case. * test driver: Expand API coverage Even more APIs that were missing from the test driver have now been implemented. * virt-xml-validate: Allow input to be read from stdin * qemu: Validate spapr-vio addresses as 32-bit libvirt has always considered these addresses (used for pSeries guests) as 64-bit, but the sPAPR specification says that they're 32-bit instead. * **Bug fixes** * qemu: Set process affinity correctly when using <numatune> libvirt would mistakenly interpret the ``nodeset`` attribute as a list of CPUs instead of as a list of NUMA node, and the process affinity would be set incorrectly as a result; this has now been fixed. v5.4.0 (2019-06-03) =================== * **Security** * cpu: Introduce support for the md-clear CPUID bit This bit is set when microcode provides the mechanism to invoke a flush of various exploitable CPU buffers by invoking the x86 ``VERW`` instruction. CVE-2018-12126, CVE-2018-12127, CVE-2018-12130, CVE-2019-11091. * Restrict user access to virt-admin, virtlogd and virtlockd The intended users for these facilities are the ``root`` user and the ``libvirtd`` service respectively, but these restrictions were not enforced correctly. CVE-2019-10132. * **Improvements** * test driver: Expand API coverage Several APIs that were missing from the test driver have now been implemented. * Avoid unnecessary static linking Most binaries shipped as part of libvirt, for example ``virtlogd`` and ``libvirt_iohelper`` , were embedding parts of the library even though they also linked against the ``libvirt.so`` dynamic library. This is no longer the case, which results in both the disk and memory footprint being reduced. * qemu: Report stat-htlb-pgalloc and stat-htlb-pgfail balloon stats These stats have been introduced in QEMU 3.0. * **Bug fixes** * qemu: Fix emulator scheduler support Setting the scheduler for QEMU's main thread before QEMU had a chance to start up other threads was misleading as it would affect other threads (vCPU and I/O) as well. In some particular situations this could also lead to an error when the thread for vCPU #0 was being moved to its cpu,cpuacct cgroup. This was fixed so that the scheduler for the main thread is set after QEMU starts. * apparmor: Allow hotplug of vhost-scsi devices v5.3.0 (2019-05-04) =================== * **New features** * qemu: Add support for setting the emulator scheduler parameters I/O threads and vCPU threads already support setting schedulers, but until now it was impossible to do so for the main QEMU thread (emulator thread in the libvirt naming). This is, however, requested for some very specific scenarios, for example when vCPU threads are running at such priority that could starve the main thread. * **Removed features** * vbox: Drop support for VirtualBox 4.x releases Support for all the 4.x releases was ended by VirtualBox maintainers in December 2015. Therefore, libvirt support for these releases is dropped. * **Improvements** * qemu: Use PCI by default for RISC-V guests PCI support for RISC-V guests was already available in libvirt 5.1.0, but it required the user to opt-in by manually assigning PCI addresses: with this release, RISC-V guests will use PCI automatically when running against a recent enough (4.0.0+) QEMU release. * qemu: Advertise firmware autoselection in domain capabilities The firmware autoselection feature is now exposed in domain capabilities and management applications can query for accepted values, i.e. values that are accepted and for which libvirt found firmware descriptor files. Firmware Secure Boot support is also advertised. * Drop YAJL 1 support YAJL 2 is widely adopted and maintaining side by side support for two versions is unnecessary. * **Bug fixes** * rpc: cleanup in virNetTLSContextNew Failed new gnutls context allocations in virNetTLSContextNew function results in double free and segfault. Occasional memory leaks may also occur. * virsh: various completers fixes There were some possible crashers, memory leaks, etc. which are now fixed. * qemu: Make hugepages work with memfd backend Due to a bug in command line generation libvirt did not honor hugepages setting with memfd backend. * Enforce ACL write permission for getting guest time & hostname Getting the guest time and hostname both require use of guest agent commands. These must not be allowed for read-only users, so the permissions check must validate "write" permission not "read". v5.2.0 (2019-04-03) =================== * **New features** * Add Storage Pool Capabilities output Add support to list an enumerated list of supported Storage Pools via the virConnectGetCapabilities API when connected via a Storage Driver. Add support to get a more detailed list XML output Storage Pool Capabilities vis the virConnectGetStoragePoolCapabilites API. * qemu: Support virtio-{non-}transitional device models ``virtio-transitional`` and ``virtio-non-transitional`` ``model`` values were added to the QEMU driver for the following devices: ``disk`` , ``interface`` , ``filesystem`` , ``rng`` , ``vsock`` , ``memballoon`` , ``controller`` type ``scsi`` , ``controller`` type ``virtio-serial`` , ``input`` bus ``virtio`` type ``passthrough`` , ``hostdev`` type ``scsi_host`` . These new models can be used to give fine grained control over what virtio device version is presented to the guest. * qemu: Enable firmware autoselection Libvirt allows users to provide loader path for some time now. However, this puts some burden on users because they need to know what firmware meets their requirements. Now that QEMU ships firmware description files this burden can be moved onto libvirt. It is as easy as setting the ``firmware`` attribute in the ``os`` element (accepted values are ``bios`` and ``efi`` ). Moreover, libvirt automatically enables domain features needed for firmware it chooses. * snapshots: Add support for topological listings A new flag VIR_DOMAIN_SNAPSHOT_LIST_TOPOLOGICAL is available for the various snapshot listing APIs such as virDomainListAllSnapshots(). For drivers that support the flag, the listed snapshots are guaranteed to be sorted such that parents occur before children. * Xen: Add support for max grant frames setting Add support for Xen's max_grant_frames setting by adding a new xenbus controller type with a maxGrantFrames attribute. E.g. ``<controller type='xenbus' maxGrantFrames='64'/>`` * qemu: Add support for parallel migration With QEMU 4.0.0 libvirt can enable parallel migration which causes the memory pages to be processed in parallel by several threads and sent to the destination host using several connections at the same time. This may increase migration speed in case a single thread is unable to saturate the network link. * **Removed features** * Drop support for Upstart and "Red Hat" init scripts Not a single one of the platforms we target still uses Upstart, and the Upstart project itself has been abandoned for several years now; the same is true for the "Red Hat" (really System V) init scripts, since RHEL 7 and later releases use systemd. * **Improvements** * Report class information for PCI node device capability. * Split setup of IPv4 and IPv6 top level chain The requirement resulting from private chains improvement done in ``v5.1.0`` was refined so that only tables from corresponding IP version are required. This means that if a network doesn't have ``IPv6`` enabled then those tables are not required. * Don't default to building the QEMU driver Historically, the QEMU driver has been special in that it was enabled by default, with the option to explicitly opt-out of it; starting now, we're enabling it opportunistically if we detect that all requirements are available, just like we do with other drivers. * **Bug fixes** * virt-host-validate: Fix IOMMU check on s390x * qemu: Allow creating pSeries guests with graphics and no USB mouse It's now possible to prevent libvirt from automatically adding a USB mouse to pSeries guests by including a USB tablet in the input XML: doing so is desiderable as using a tablet results in a much better user experience when working with GUIs. * qemu: Set $HOME and XGD variables for qemu:///system guests This avoids files being accidentally created under ``/`` or the guests not being able to start because they lack the necessary permissions to write to that location. v5.1.0 (2019-03-04) =================== * **New features** * bhyve: Add support for additional command-line arguments The bhyve driver now supports passing additional command-line arguments to the bhyve process using the new ``<bhyve:commandline>`` element in domain configuration. * network: Support setting a firewalld "zone" for virtual network bridges All libvirt virtual networks with bridges managed by libvirt (i.e. those with forward mode of "nat", "route", "open", or no forward mode) will now be placed in a special firewalld zone called "libvirt" by default. The zone of any network bridge can be changed using the ``zone`` attribute of the network's ``bridge`` element. * bhyve: Support for ignoring unknown MSRs reads and writes A new <features> element <msrs unknown='ignore'/> was introduced and the bhyve driver supports it to control unknown Model Specific Registers (MSRs) reads and writes. * qemu: Add support for encrypted VNC TLS keys Use the password stored in the secret driver under the uuid specified by the ``vnc_tls_x509_secret_uuid`` option in qemu.conf. * Add storage pool namespace options Allow for adjustment of RBD configuration options via Storage Pool XML Namespace adjustments. * qemu: Add support for setting post-copy migration bandwidth Users can now limit the bandwidth of post-copy migration, e.g. via ``virsh migrate --postcopy-bandwidth`` . * **Improvements** * Create private chains for virtual network firewall rules Historically firewall rules for virtual networks were added straight into the base chains. This works but has a number of bugs and design limitations. To address them, libvirt now puts firewall rules into its own chains. Note that with this change the ``filter`` , ``nat`` and ``mangle`` tables are required for both ``IPv4`` and ``IPv6`` . * Detect CEPH and GPFS as shared FS When starting a migration libvirt performs some sanity checks to make sure domain will be able to run on the destination. One of the requirements is that the disk has to either be migrated too or be accessible from a network filesystem. CEPH and GPFS weren't detected as a network filesystem. * Advertise network MTU via DHCP when specified If network MTU is set and the network has DHCP enabled, advertise the MTU in DHCP transaction too so that clients can adjust their link accordingly. * qemu: Allocate memory at the configured NUMA nodes from start Libvirt used to just start QEMU, let it allocate memory for the guest, and then use CGroups to move the memory to configured NUMA nodes. This is suboptimal as huge chunks of memory have to be moved. Moreover, this relies on ability to move memory later which is not always true. A change was made to set process affinity correctly from the start so that memory is allocated on the configured nodes from the beginning. * Support for newer Wireshark Adapt libvirt to use the more recent release requiring a source build configuration of libvirt ``--with-wireshark`` to upgrade to the more recent version. * Batch mode virsh and virt-admin parsing improvements When parsing a single-argument command_string in batch mode, virsh and virt-admin now permit newlines in addition to semicolons for splitting commands, and backslash-newline for splitting long lines, to be more like shell parsing. * **Bug fixes** * qemu: Use CAP_DAC_OVERRIDE during QEMU capabilities probing By default, libvirt runs the QEMU process as ``qemu:qemu`` which could cause issues during probing as some features like AMD SEV might be inaccessible to QEMU because of file system permissions. Therefore, ``CAP_DAC_OVERRIDE`` is granted to overcome these for the purposes of probing. * storage: Add default mount options for fs/netfs storage pools Altered the command line generation for fs/netfs storage pools to add some default options. For Linux based systems, the options added are "nodev, nosuid, noexec". For FreeBSD based systems, the options added are "nosuid, noexec". * qemu: Allow use of PCI for RISC-V guests This works with QEMU 4.0.0+ only and is opt-in at the moment, since it requires users to manually assign PCI addresses, but is otherwise fully functional. * network: Fix virtual networks on systems using firewalld+nftables Because of the transitional state of firewalld's new support for nftables, not all iptables features required by libvirt are yet available, so libvirt must continue to use iptables for its own packet filtering rules even when the firewalld backend is set to use nftables. However, due to the way iptables support is implemented in kernels using nftables (iptables rules are converted to nftables rules and processed in a separate hook from the native nftables rules), guest networking was broken on hosts with firewalld configured to use nftables as the backend. This has been fixed by putting libvirt-managed bridges in their own firewalld zone, so that guest traffic can be forwarded beyond the host and host services can be exposed to guests on the virtual network without opening up those same services to the rest of the physical network. This means that host access from virtual machines is no longer controlled by the firewalld default zone (usually "public"), but rather by the new firewalld zone called "libvirt" (unless configured otherwise using the new zone attribute of the network bridge element). * qemu: Fix i6300esb watchdog hotplug on Q35 Ensure that libvirt allocates a PCI address for the device so that QEMU did not default to an address that would not allow for device hotplug. * lxc: Don't reboot host on virDomainReboot If the container is really a simple one (init is just bash and the whole root is passed through) then virDomainReboot and virDomainShutdown would reboot or shutdown the host. The solution is to use different method to reboot or shutdown the container in that case (e.g. signal). * rpc: Various stream fixes One particular race was fixed, one locking problem and error reporting from streams was made better. * qemu: Fix guestfwd hotplug/hotunplug Fixed the generation of the guestfwd hotplug/unplug command sent to QEMU to match the syntax used when creating the initial command line. * qemu: Forbid CDROMs on virtio bus Attempting to create an empty virtio-blk drive or attempting to eject it results into an error. Forbid configurations where users would attempt to use CDROMs in virtio bus. * qemu: Use 'raw' for 'volume' disks without format Storage pools might want to specify format of the image when translating the volume thus libvirt can't add any default format when parsing the XML. Add an explicit format when starting the VM and format is not present neither by user specifying it nor by the storage pool translation function. * qemu: Assume 'raw' default storage format also for network storage Post parse callback adds the 'raw' type only for local files. Remote files can also have backing store (even local) so we should do this also for network backed storage. * qemu: Fix block job progress reporting and advocate for READY event In some cases QEMU can get to 100% and still not reach the synchronised phase. Initiating a pivot in that case will fail. Therefore it is strongly advised to wait for ``VIR_DOMAIN_BLOCK_JOB_READY`` event which does not suffer from this problem. * qemu: Don't format image properties for empty drive If a ``-drive`` has no image, then formatting attributes such as cache, readonly, etc. would cause errors to be reported from QEMU. This was fixed by not supplying the attributes for devices without an image. * External snapshot metadata redefinition is fixed Attempting to use VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE to reinstate the metadata describing an external snapshot created earlier for an offline domain no longer fails. v5.0.0 (2019-01-15) =================== * **New features** * Xen: Add support for openvswitch The libxl driver now supports virtual interfaces that connect to an openvswitch bridge, including interfaces with VLAN tagging and trunking configuration. * qemu: Report whether KVM nesting is available Running nested KVM guests requires specific configuration steps to be performed on the host; libvirt will now report in the host capabilities whether KVM nesting support is available. * **Removed features** * Drop UML driver The UML driver was unmaintained and not tested for quite some time now. Worse, there is a bug that causes it to deadlock on some very basic operations (e.g. dumping domain XML). These facts make us believe no one uses it. * **Improvements** * qemu: Add support for ARMv6l guests * Support more NVDIMM configuration options Introduce more configuration options. For the source element, add the 'alignsize' and 'pmem' subelements. For the target element, add the 'readonly' subelement. * cpu: Add support for "stibp" x86_64 feature Add cpu flag stibp (Single Thread Indirect Branch Predictors) to prevent indirect branch predictions from being controlled by the sibling Hyperthread. * libxl: Handle external domain destroy Historically, if a domain was destroyed using ``xl`` rather than through libvirt APIs, libvirt would not be aware of the fact and keep considering it as running. This is no longer the case. * Start selecting the first available DRI device for OpenGL operations If OpenGL support is needed (either with SPICE gl enabled or with egl-headless), libvirt is now able to pick the first available DRI device for the job. At the same time, this improvement is also a bugfix as it prevents permission-related issues with regards to our mount namespaces and the default DRI render node's permissions which would normally prevent QEMU from accessing such a device. * qemu: Add support for postcopy-requests migration statistics The ``virDomainJobInfo`` can get number page requests received from the destination host during post-copy migration. * **Bug fixes** * lxc: Don't forbid interfaces with type=direct Such interfaces are supported by lxc and should be allowed. * qemu: Fully clean up RNG devices on detach Some RNG device types, such as those using EGD, might need extra clean up on the host in addition to removing the guest-side device. v4.10.0 (2018-12-03) ==================== * **New features** * qemu: Add Hyper-V PV IPI and Enlightened VMCS support The QEMU driver now has support for Hyper-V PV IPI and Enlightened VMCS for Windows and Hyper-V guests. * qemu: Added support for PCI devices on S390 PCI addresses can now include the new zpci element which contains uid (user-defined identifier) and fid (PCI function identifier) attributes and makes the corresponding devices usable by S390 guests. * Support changing IOThread polling parameters for a live guest Introduced virDomainSetIOThreadParams which allows dynamically setting the IOThread polling parameters used by QEMU to manage the thread polling interval and the algorithm for growth or shrink of the polling time. The values only affect a running guest with IOThreads. The guest's IOThread polling values can be viewed via the domain statistics. * Xen: Add support for PVH The libxl driver now supports Xen's PVH virtual machine type. PVH machines are enabled with the new "xenpvh" OS type, e.g. ``<os><type>xenpvh</type></os>`` * qemu: Added support for CMT (Cache Monitoring Technology) Introduced cache monitoring using the ``monitor`` element in ``cachetune`` for vCPU threads. Added interfaces to get and display the cache utilization statistics through the command 'virsh domstats' via the virConnectGetAllDomainStats API. * qemu: Add support for nested HV for pSeries guests Nested HV support makes it possible to run nested (L2) guests with minimal performance penalty when compared to regular (L1) guests on ppc64 hardware. * **Bug fixes** * Xen: Handle soft reset shutdown event The pvops Linux kernel uses soft reset to handle the crash machine operation. The libxl driver now supports the soft reset shutdown event, allowing proper crash handling of pvops-based HVM domains. v4.9.0 (2018-11-04) =================== * **New features** * util: Add cgroup v2 support cgroup v2 support has been implemented in libvirt, with both "unified" (v2 only) and "hybrid" (v2 + v1) setups being usable; existing "legacy" (v1 only) setups will keep working. * qemu: Add vfio AP support The QEMU driver now has support to passthrough adjunct processors into QEMU guests on S390. * **Improvements** * rpc: Make 'genprotocol' output reproducible This is another step towards making libvirt builds fully reproducible. * **Bug fixes** * security: Fix permissions for UNIX sockets Since 4.5.0, libvirt is using FD passing to hand sockets over to QEMU, which in theory removes the need for them to be accessible by the user under which the QEMU process is running; however, other processes such as vdsm need to access the sockets as well, which means adjusting permissions is still necessary. * cpu_map: Add Icelake model definitions These CPU models will be available in the upcoming 3.1.0 QEMU release. * util: Properly parse URIs with missing trailing slash Some storage URIs were not parsed correctly, in which case libvirt ended up emitting XML that it would then refuse to parse back. v4.8.0 (2018-10-01) =================== * **New features** * Xen: Support PM Suspend and Wakeup The libxl driver now supports the virDomainPMSuspendForDuration and virDomainPMWakeup APIs. * **Removed features** * Xen: Drop support for Xen 4.4 and 4.5 Xen 4.4 and 4.5 are no longer supported by the Xen community. Drop support for these older versions and require Xen >= 4.6. * nwfilter: Disallow binding creation in session mode Ensure that a filter binding creation is not attempted in session mode and generates a proper error message. * **Improvements** * qemu: Retrieve guest hostname through QEMU Guest Agent command QEMU is now able to retrieve the guest hostname using a new QEMU-GA command called 'guest-get-host-name'. Virsh users can execute 'domhostname' for QEMU driver for domains configured to use the Guest Agent. * virsh: Implement vsh-table in virsh and virsh-admin The new API fixes problems with table-alignment, making the tables more readable and deals with unicode. * **Bug fixes** * storage: Allow inputvol to be encrypted When creating a storage volume based on another volume, the base input volume is allowed to be encrypted. * virsh: Require explicit --domain for domxml-to-native The --domain option for domxml-to-native virsh command has always been documented as required, but commit v4.3.0-127-gd86531daf2 accidentally made it optional. * lxc_monitor: Avoid AB / BA lock race A deadlock situation could occur when autostarting a LXC domain 'guest' due to two threads attempting to take opposing locks while holding opposing locks (AB BA problem). v4.7.0 (2018-09-03) =================== * **New features** * storage: add storage pool iscsi-direct Introduce a new storage pool backend that uses libiscsi instead of iscsiadm. It support basic pool operations: checkPool and refreshPool. * Add support for MBA (Memory Bandwidth Allocation technology) Domain vCPU threads can now have allocated some parts of host memory bandwidth by using the ``memorytune`` element in ``cputune`` . * qemu: Add support for RISC-V guests riscv32 and riscv64 guest architectures are now supported. * **Improvements** * qemu: Add ccw support for vhost-vsock Support the vhost-vsock-ccw device on S390. * qemu: Make default machine type independent of QEMU We can't control whether or not QEMU will change its default machine type in the future, or whether downstream distributions will decide to compile out some machine types, so our only option to provide a predictable behavior is taking care of the default ourselves; management applications and users are encouraged to explicitly pick a machine type when creating new guests. * apparmor: Various improvements Rules have been added to deal with a number of scenarios that didn't work correctly. * **Bug fixes** * esx: Truncate CPU model name Some CPU model names are too long to be stored into the corresponding property, and should be explicitly truncated to avoid unexpected behavior in users of the ``virNodeGetInfo()`` API such as ``virsh nodeinfo`` . * utils: Remove arbitrary limit on socket_id/core_id Both values were assumed to be smaller than 4096, but in fact they are entirely hardware-dependent and there have been reports of machines presenting much bigger values, preventing libvirt from working correctly; all such limits have now been removed. v4.6.0 (2018-08-06) =================== * **New features** * qemu: Implement the HTM pSeries feature Users can now decide whether HTM (Hardware Transactional Memory) support should be available to the guest. * qemu: Enable VNC console for mediated devices Host devices now support a new attribute 'display' which can be used to turn on frame buffer rendering on a vgpu mediated device instead of on an emulated GPU, like QXL. * **Improvements** * qemu: Introduce a new video model of type 'none' Introduce a new video model type that disables the automatic addition of a video device to domains with 'graphics' specified in their XML. This can be useful with GPU mediated devices which can serve as the only rendering devices within the guest. * virsh: Add --alias to attach-disk and attach-interface commands Add option --alias to set customized device alias name when using attach-disk or attach-interface commands. * virsh: Support usb and sata address to attach-disk Usb or sata address could be used when attach-disk with --address. For example, use usb address as usb:<bus>.<port>, use sata address as <controller>.<bus>.<unit>. v4.5.0 (2018-07-02) =================== * **New features** * qemu: Provide TPM emulator support Support QEMU's TPM emulator based on swtpm. Each QEMU guest gets its own virtual TPM. * bhyve: Support specifying guest CPU topology Bhyve's guest CPU topology could be specified using the ``<cpu><topology ../></cpu>`` element. * qemu: Add support for extended TSEG size Support specifying extended TSEG size for SMM in QEMU. * qemu: Add support for SEV guests SEV (Secure Encrypted Virtualization) is a feature available on AMD CPUs that encrypts the guest memory and makes it inaccessible even to the host OS. * **Removed features** * Remove support for qcow/default encrypted volumes Disallow using a qcow encrypted volume for the guest and disallow creation of the qcow or default encrypted volume from the storage driver. Support for qcow encrypted volumes has been phasing out since QEMU 2.3 and by QEMU 2.9 creation of a qcow encrypted volume via qemu-img required usage of secret objects, but that support was never added to libvirt. * Make GnuTLS mandatory Building without GnuTLS is no longer possible. * qemu: Remove allow_disk_format_probing configuration option The option represented a security risk when used with malicious disk images, so users were recommended against enabling it; with this release, it's been removed altogether. * **Improvements** * capabilities: Provide info about host IOMMU support Capabilities XML now provide information about host IOMMU support. * virsh: Add --all to domblkinfo command Alter the ``domblkinfo`` command to add the option --all in order to display the size details of each domain block device from one command in a output table. * qemu: Allow concurrent access to monitor and guest agent Historically libvirt prevented concurrent accesses to the qemu monitor and the guest agent. Therefore two independent calls (one querying the monitor and the other querying guest agent) would serialize which hurts performance. The code was reworked to allow two independent calls run at the same time. * qemu: Allow configuring the page size for HPT pSeries guests For HPT pSeries guests, the size of the host pages used to back guest memory and the usable guest page sizes are connected; the new setting can be used to request that a certain page size is available in the guest. * Add support to use an raw input volume for encryption It is now possible to provide a raw input volume as input for to generate a luks encrypted volume via either virsh vol-create-from or virStorageVolCreateXMLFrom. * qemu: Add support for vsock hot (un)plug and cold (un)plug * qemu: Add support for NBD over TLS NBD volumes can now be accessed securely. * qemu: Implement FD passing for Unix sockets Instead of having QEMU open the socket and then connecting to it, which is inherently racy, starting with QEMU 2.12 we can open the socket ourselves and pass it to QEMU, avoiding race conditions. * virsh: Introduce --nowait option for domstat command When this option is specified, virsh will try to fetch the guest stats but abort instead of stalling if they can't be retrieved right away. * **Bug fixes** * qemu: Fix a potential libvirtd crash on VM reconnect Initialization of the driver worker pool needs to come before libvirtd trying to reconnect to all machines, since one of the QEMU processes migh have already emitted events which need to be handled prior to us getting to the worker pool initialization. * qemu: Fix domain resume after failed migration Recent versions of QEMU activate block devices before the guest CPU has been started, which makes it impossible to roll back a failed migration. Use the ``late-block-activate`` migration capability if supported to avoid the issue. * vmx: Permit guests to have an odd number of vCPUs An odd number of vCPUs greater than 1 was forbidden in the past, but current versions of ESXi have lifted that restriction. v4.4.0 (2018-06-04) =================== * **New features** * bhyve: Support locking guest memory Bhyve's guest memory may be wired using the ``<memoryBacking><locked/></memoryBacking>`` element. * qemu: Provide VFIO channel I/O passthrough support Support passthrough devices that use channel I/O based mechanism in a QEMU virtual machine. * qemu: Add support for migration of VMs with non-shared storage over TLS It's now possible to use the VIR_MIGRATE_TLS flag together with VIR_MIGRATE_NON_SHARED_DISK. The connection is then secured using the TLS environment which is setup for the migration connection. * Add support for VM Generation ID The VM Generatation ID exposes a 128-bit, cryptographically random, integer value identifier, referred to as a Globally Unique Identifier (GUID) to the guest in order to notify the guest operating system when the virtual machine is executed with a different configuration. Add a new domain XML processing and a domain capabilities feature. * Introduce virDomainDetachDeviceAlias This new API enables users to detach device using only its alias. * Introduce new virConnectCompareHypervisorCPU and virConnectBaselineHypervisorCPU APIs Unlike the old virConnectCompareCPU and virConnectBaselineCPU APIs, both new APIs consider capabilities of a specific hypervisor. * Introduce SCSI persistent reservations support The QEMU driver gained support for qemu-pr-helper which enables guests to issue SCSI commands for persistent reservation. * qemu: Implement multiple screen support for virDomainScreenshot While the virDomainScreenshot API supported multihead video cards, the implementation was missing. But now that QEMU implemented it libvirt has done as well. * qemu: add support for vhost-vsock-device A new vsock device was introduced, allowing communication between the guest and the host via the AF_VSOCK family. * **Improvements** * qemu: Add support for OpenGL rendering with SDL Domains using SDL as a graphics backend will now be able to use OpenGL accelerated rendering. * qemu: Add support for 'output' audio codec Support QEMU's 'hda-output' codec advertising only a line-out for ich6 and ich9 sound devices. * virsh: Enhance event name completion Implement event name completion for some commands (e.g. event, secret-event, pool-event and nodedev-event) v4.3.0 (2018-05-02) =================== * **New features** * qemu: Add support for the pcie-to-pci-bridge controller Pure PCIe guests such as x86_64/q35 and aarch64/virt will now add this controller when traditional PCI devices are in use. * Xen: Support setting CPU features for host-passthrough model The CPU model presented to Xen HVM domains is equivalent to libvirt's host-passthrough model, although individual features can be enabled and disabled via the cpuid setting. The libvirt libxl driver now supports enabling and disabling individual features of the host-passthrough CPU model. * **Removed features** * Xen: Drop the legacy xend-based driver The xm/xend toolstack was deprecated in Xen 4.2 and removed from the Xen sources in the 4.5 development cycle. The libvirt driver based on xend is now removed from the libvirt sources. * **Improvements** * qemu: Support hot plug and hot unplug of mediated devices Libvirt now allows mediated devices to be hot plugged and hot unplugged from a guest rather than reporting an error that this isn't supported. In fact, kernel has been supporting this since 4.10. * **Bug fixes** * Improve handling of device mapper targets When starting a domain with a disk backed by a device mapper volume libvirt also needs to allow the storage backing the device mapper in CGroups. In the past kernel did not care, but starting from 4.16 CGroups are consulted on each access to the device mapper target. v4.2.0 (2018-04-01) =================== * **New features** * Support building with Python 3 Python is required to build libvirt, and up until now only Python 2 could be used as an interpreter. All scripts used during build have now been made compatible with Python 3, which means both major releases of the language are fully supported. * qemu: Provide ccw address support for graphics and input devices Support the virtio-gpu-ccw device as a video device and virtio-{keyboard, mouse, tablet}-ccw devices as input devices on S390. * **Improvements** * qemu: Add logging of guest crash information on S390 On S390, when the guest crashes and QEMU exposes the guest crash information, log the relevant data to the domain log file. * qemu: use arp table of host to get the IP address of guests Find IP address of a VM by arp table on hosts. If someone customizing IP address inside VM, it will be helpful. * Xen: Remove hard-coded scheduler weight The libxl driver was accidentally hard-coding the per-domain scheduler weight to 1000, silently ignoring any user-provided ``<shares>`` in ``<cputune>`` . The driver now honors ``<shares>`` , and defers setting a default value to Xen. Note that the Xen default is 256, so any domains started after this improvement will have one fourth the shares of previously started domains. If all domains must have equal CPU shares, administrators must manually set the weight of previously started domains to 256, or restart them. * **Bug fixes** * qemu: TLS migration now enforces use of TLS for the NBD connection When the VIR_MIGRATE_TLS flag was used with the migration API libvirt did not ensure that the NBD connection was using TLS as well. The code now rejects such migration as the TLS transport for NBD is not ready yet, but prevents a false sense of security that TLS would be used. The support TLS for NBD will be added soon. v4.1.0 (2018-03-05) =================== * **New features** * Added support for CAT (Cache allocation Technology) Domain vCPU threads can now have allocated some parts of host cache using the ``cachetune`` element in ``cputune`` . * Allow opening secondary drivers Up until now it was possible to connect to only hypervisor drivers (e.g. qemu:///system, lxc:///, vbox:///system, and so on). The internal drivers (like network driver, node device driver, etc.) were hidden from users and users could use them only indirectly. Starting with this release new connection URIs are accepted. For instance network:///system, storage:///system and so on. * virtlogd, virtlockd: Add support for admin protocol These two daemons now support admin protocol through which some admin info can be gathered or some configuration tweaked on the fly. * **Improvements** * virsh: Enhance bash completion Implement more bash completions so that basic libvirt objects can be auto-completed (e.g. networks, interfaces, NWFilters, and so on). * qemu: Use VIR_ERR_DEVICE_MISSING for various hotplug/detach messages * qemu: Allow showing the dump progress for memory only dump Alter the QEMU dump-guest-memory command processing to check for and allow asynchronous completion which then allows for the virsh dump --memory-only --verbose command to display percent completion data. * conf: add support for setting Chassis SMBIOS data fields * libxl: add support for setting clock offset and adjustment * Make port allocator global Up until now each driver had their own port allocator module. This meant that info on port usage was not shared. Starting with this release, the port allocator module is made global and therefore drivers allocate ports from global pool. * Fixed some compiler warnings that appear with GCC 8 * **Bug fixes** * qemu: Check for unsafe migration more thoroughly If a domain disk is stored on local filesystem (e.g. ext4) but is not being migrated it is very likely that domain is not able to run on destination. Regardless of share/cache mode. * qemu: Fix updating device with boot order Starting with 3.7.0 release updating any device with boot order would fail with 'boot order X is already used by another device' while in fact it was the very same device. * virlog: determine the hostname on startup CVE-2018-6764 At later point it might not be possible or even safe to use getaddrinfo(). It can in turn result in a load of NSS module which can even be loaded from unsage guest filesystem and thus escape the confinment of its container. * qemu: Rework vCPU statistics fetching Fetching vCPU statistics was very expensive because it lead to waking up vCPU threads in QEMU and thus it degraded performance. The code was reworked so that fetching statistics does not wake up halted vCPUs. * qemu: unlink memory backing file on domain shutdown Depending on the filesystem where domain memory is stored, some files might have been left behind. This is not a problem on hugetlbfs, but it is a problem on regular filesystems like ext4. * qemu: Fix shutting down domains in parallel If multiple domains were being shut down in parallel, libvirtd might have deadlocked. * nodedev: Update PCI mdev capabilities dynamically PCI devices may have other nested capabilities, like SRIOV and mdev which depend on the device being plugged into the native vendor driver. However, in case such a device is directly assigned to a guest using VFIO driver, the device will naturally lose these capabilities and libvirt needs to reflect that. v4.0.0 (2018-01-19) =================== * **New features** * tools: Provide bash completion support Both ``virsh`` and ``virt-admin`` now implement basic bash completion support. * qemu: Refresh capabilities on host microcode update A microcode update can cause the CPUID bits to change; therefore, the capabilities cache should be rebuilt when such an update is detected on the host. * lxc: Set hostname based on container name * **Improvements** * CPU frequency reporting improvements The CPU frequency will now be reported by ``virsh nodeinfo`` and other tools for s390 hosts; at the same time; CPU frequency has been disabled on aarch64 hosts because there's no way to detect it reliably. * libxl: Mark domain0 as persistent * Xen: Add support for multiple IP addresses on interface devices * qemu: Add support for hot unplugging redirdev device * **Bug fixes** * qemu: Enforce vCPU hotplug granularity constraints QEMU 2.7 and newer don't allow guests to start unless the initial vCPUs count is a multiple of the vCPU hotplug granularity, so validate it and report an error if needed. v3.10.0 (2017-12-04) ==================== * **New features** * conf: Support defining distances between virtual NUMA cells A NUMA hardware architecture supports the notion of distances between NUMA cells. This can now be specified using the ``<distances>`` element within the NUMA cell configuration. Drivers which support this include Xen and QEMU. * Xen: Support defining vNUMA topology Xen now supports defining a virtual NUMA topology for VMs, including specifying distances between NUMA cells. * qemu: Add the ability to configure HPT resizing for pSeries guests The user can now decide whether HPT (Hash Page Table) resizing should be enabled, disabled or required instead of leaving it up to hypervisor defaults and negotiation between the guest and the host. * qemu: Add vmcoreinfo feature Starting with QEMU 2.11, the guest can save kernel debug details when this feature is enabled and the kernel supports it. It is useful to process kernel dump with KASLR enabled, and also provides various kernel details to crash tools. * conf: Move the auth and encryption definitions to disk source Allow parsing and formatting of the ``auth`` and ``encryption`` sub-elements to be a child of the ``source`` element. This will allow adding an ``auth`` sub-element to a ``backingStore`` or ``mirror`` elements as a means to track specific authentication and/or encryption needs. * **Improvements** * vbox: Add VirtualBox 5.2 support * vbox: Add support for configuring storage controllers The VirtualBox driver now supports the ``<controller>`` element in the domain XML for configuring storage controllers in VBOX VMs. Additionally, libvirt's domain XML schema was updated to allow optional ``model`` attribute for ``<controller type='ide'>`` which is used by the VBOX driver to set the IDE controller model to be one of 'piix4', 'piix4' (default), or 'ich6'. Finally, with this change ``dumpxml`` generates ``<controller>`` elements that correspond to current VBOX VM storage controller configuration. * vbox: Add support for attaching empty removable disks The VirutalBox driver now supports adding CD-ROM and floppy disk devices that do not have the disk source specified. Previously such devices were silently ignored. * vbox: Add support for attaching SAS storage controllers In VirtualBox, SCSI and SAS are distinct controller types whereas libvirt does not make such distinction. Therefore, the VBOX driver was updated to allow attaching SAS controllers via ``<controller type='scsi' model='lsisas1068'>`` element. If there are both SCSI and SAS controllers present in the VBOX VM, the domain XML can associate the disk device using the ``<address>`` element with the ``controller`` attribute, and optionally, set the port via ``unit`` attribute. * qemu: Generate predictable paths for qemu memory backends In some cases management applications need to know paths passed to memory-backend-file objects upfront. Libvirt now generates predictable paths so applications can prepare the files if they need to do so. * Shareable disks work properly with recent qemu Recent qemu versions added image locking to avoid potential corruption of disk images. This broke shareable disks with libvirt since the feature was turned on by default in qemu. Libvirt now enables sharing of those disks in qemu so that the image locking is not applied in that case. Additionally libvirt now checks that shareable disks have supported format (raw) to avoid metadata corruption. * Improve serial console behavior on non-x86 architectures ppc64, aarch64 and s390x guests were treating the <serial> and <console> elements differently from x86, in some cases presenting misleading information to the user. The behavior is now consistent across all architectures and the information reported is always accurate. * **Bug fixes** * vbox: Do not ignore failures to attach disk devices when defining The ``define`` now fails and reports an error if any of the ``controller`` or ``disk`` devices specified in the domain XML fail to attach to the VirtualBox VM. * vbox: Fix dumpxml to always output disk devices The VirtualBox driver was ignoring any disk devices in ``dumpxml`` output if there was a SAS storage controller attached to the VM. * vbox: Fix dumpxml to always generate valid domain XML When a VirtualBox VM has multiple disks attached, each to a different storage controller that uses 'sd' prefix for block device names e.g. one disk attached to SATA and one to SCSI controller, it no longer generates XML where both would have 'sda' device name assigned. Instead it properly assigns 'sda' and 'sdb' to those disks in the order of appearance. * Securely pass iSCSI authentication data Rather than supplying the authentication data as part of the iSCSI URL for a disk or host device, utilize the encrypted secret object to securely pass the authentication data. v3.9.0 (2017-11-02) =================== * **New features** * Add capability to allow hot (un)plug of a domain watchdog device * Allow users to set device aliases Users can set aliases to domain devices and thus identify them easily. * qemu: Support multiqueue for virtio-blk Multiqueue support for ``virtio-blk`` has been available in QEMU ever since 2.7.0, and now libvirt guests can enable it. * Add virDomainSetLifecycleAction API Provided a new API to allow dynamic guest lifecycle control for guest reactions to poweroff, restart, or crash type events related to the domain XML ``on_poweroff`` , ``on_reboot`` , and ``on_crash`` elements. The ``virsh set-lifecycle-action`` command was created to control the actions. * qemu: Allow cold(un)plugging and hot(un)plugging input devices * net: Implement QoS for vhostuser * **Improvements** * Allow a logical volume to be create using LUKS A logical volume may be created using an ``encryption`` element using "luks" format. This does require a previously created ``secret`` to store the passphrase used to encrypt the volume Adding the volume to a domain can then either provide the secret or allow the consumer in the guest to provide the passphrase in order to decrypt the volume. * net: Ignore auto-generated MAC address when detaching an interface If the MAC address has not been specified by the user, libvirt will try and fill in the gaps by generating one; however, for some error paths that led to some confusing error messages, so when an auto-generated MAC address is specified the error message will not include the auto-generated MAC. * net: Enable MAC address lookup for virDomainInterfaceStats * apparmor: Several improvements Changes include permitting access to data about USB devices and ``dnsmasq`` instances, allowing spaces in guest names and many more. * cpu: Use CPU information obtained from QEMU when possible Recent QEMU versions can expose information about which CPU models are available and usable on the host; libvirt will now make use of such information whenever possible. * hyperv: Various improvements The error reported when clients can't connect to Hyper-V has been made more descriptive, and memory limits for guests are now mapped to more appropriate libvirt equivalents. * qemu: Report QEMU error on failed migration Instead of reporting a generic error, ask QEMU for a more detailed and thus hopefully more helpful one. * vbox: Implement autoport for RDP libvirt will now obtain the (dynamically allocated) RDP port number from VirtualBox itself, avoiding conflicts between multiple guests wanting to use RDP at the same time. * qemu: Allow rotation of small logs On a host where numerous unique instances are executed per day, it's quite possible that, even though each of the single log files are fairly small, collectively the quantity and volume may add tens of thousands of log files to the ``/var/log/libvirt/qemu/`` directory. Removing the constraints that log have to be bigger than 100 KiB before they can be rotated solves the issue. * **Bug fixes** * Fix swapped interface statistics and QoS Due to internal implementation, reported statistics for some types of interfaces were swapped (RX appeared in TX and vice versa). Similarly, QoS was set in reversed way. * Properly resize local LUKS encrypted volume Resizing of a local LUKS encrypted volume will now use qemu-img to resize the volume. This will require configuring a secret for the LUKS encrypted volume. * qemu: Reserve PCI addresses for implicit i440fx devices Failing to do so causes the addresses to be considered usable by libvirt, which means they could be assigned to more than one device resulting in the guest failing to start. * spec: Restart libvirtd only at the end of the upgrade process Use ``%posttrans`` to make sure ``libvirtd`` is not restarted before all other components, such as the library itself and storage / hypervisor drivers, have already been upgraded. * **Security** * qemu: Ensure TLS clients always verify the server certificate While it's reasonable to turn off client certificate validation, as setting it up can be non-trivial, clients should always verify the server certificate to avoid MITM attacks. However, libvirt was using the same knob to control both checks, leading to CVE-2017-1000256 / LSN-2017-0002. v3.8.0 (2017-10-04) =================== * **New features** * qemu: Added support for cold-(un)plug of watchdog devices * qemu: Added support for setting IP address os usernet interfaces * qemu: Added support for Veritas Hyperscale (VxHS) block devices * storage: Added new events for pool-build and pool-delete * **Improvements** * qemu: Set DAC permissions properly for spice rendernode When a ``rendernode`` path is set for SPICE GL on ``qemu:///system`` , we now correctly set DAC permissions on the device at VM startup. This is the last remaining hurdle to let SPICE GL work for ``qemu:///system`` without any external host changes. * nodedev: Add switchdev offload query to NIC capabilities Allow querying the NIC interface capabilities for the availability of switchdev offloading (also known as kernel-forward-plane-offload). * New CPU models for AMD and Intel AMD EPYC and Intel Skylake-Server CPU models were added together with their features * Improve long waiting when saving a domain While waiting for a write to disk to be finished, e.g. during save, even simple operations like ``virsh list`` would be blocking due to domain lock. This is now resolved by unlocking the domain in places where it is not needed. * **Bug fixes** * Proper units are now used in virsh manpage for dom(mem)stats Previously the documentation used multiples of 1000, but now it is fixed to use multiples of 1024. * qemu: Fix error reporting when disk attachment fails There was a possibility for the actual error to be overridden or cleared during the rollback. * qemu: Fix assignment of graphics ports after daemon restart This could be seen with newer kernels that have bug regarding SO_REUSEADDR. After libvirtd was restarted it could assign already used address to new guests which would make them fail to start. This is fixed by marking used ports unavailable when reconnecting to running QEMU domains. * Fix message decoding which was causing a very strange bug When parsing an RPC message with file descriptors was interrupted and had to restart, the offset of the payload was calculated badly causing strange issues like not being able to find a domain that was not requested. v3.7.0 (2017-09-04) =================== * **New features** * qemu: Add managedsave-edit commands Using managedsave-dumpxml, managedsave-define and managedsave-edit commands, now we can dump and edit the XML configuration of domain which has managedsave image. * qemu: Add migrate-getmaxdowntime command Currently, the maximum tolerable downtime for a domain being migrated is write-only from libvirt, via migrate-setmaxdowntime. This implements a complementary migrate-getmaxdowntime command * bhyve: Support autoport for VNC ports It's no longer necessary to explicitly specify VNC port for the bhyve guests. With the autoport feature it will be allocated automatically. Please refer to the bhyve driver documentation for examples. * qemu: Added support for setting heads of virtio GPU * qemu: Added support to configure reconnect timeout for chardev devices When you have a TCP or UNIX chardev device and it's connected somewhere you can configure reconnect timeout if the connection is closed. * **Improvements** * qemu: Report a clear error when dropping a VM during startup "Failed to load config for domain 'DOMNAME'" is now reported if a VM config can't be parsed for some reason, and thus provides a clear indication for users (and devs). * apparmor: Update for QEMU 2.10 compatibility Starting with QEMU 2.10, disk images and NVRAM files get automatically locked to prevent them from being corrupted; however, file locking needs to be explicitly allowed through ``virt-aa-helper`` or AppArmor will reject the requests and the guest will not be able to run. * virsh: List Unix sockets in 'domdisplay' output VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets as connection endpoints, but such a configuration was not handled correctly by ``virsh domdisplay`` , causing the respective endpoints to be missing from the output. * qemu: Don't check whether offline migration is safe Since offline migration only copies the guest definition to the destination host, data corruption is not a concern and the operation can always be performed safely. * virt-host-validate: Fix IOMMU detection on ppc64 * **Bug fixes** * qemu: Better support for international domain names (with wide characters) There were some issues with multi-byte domains getting lost on daemon restart due to truncation, so the code now handles multi-byte names a bit better. * qemu: Support long domain names with namespaces Domains with extremely long names would fail to start due to temporary namespace paths being created with the whole name. The path is now generated with shortened name instead. * qemu: Tolerate missing emulator binary during libvirtd restart For some time libvirt required qemu capabilities being present when parsing VM configs during startup. As a side effect VM configs would fail to parse and thus vanish, if the emulator binary would be uninstalled or broken. Libvirt now tolerates when capabilities are missing during startup. * qemu: Prevent pSeries guests from disappearing in some situations pSeries guest would disappear if any of the host devices they were configured to use was not available during libvirtd startup, which could easily happen for SR-IOV Virtual Functions. This scenario is now handled correctly. * qemu: Honor <on_reboot/> setting The setting was accepted by the parser, but not actually implemented. * Fix --verbose option for all daemons Since v3.0.0, the option had been ignored by all libvirt daemons ( ``libvirtd`` , ``virtlogd`` and ``virtlockd`` ); it's now working as intended once again. v3.6.0 (2017-08-02) =================== * **New features** * hyperv: Implement virDomainSetMemory and virDomainSendKey APIs * qemu: Support multiple PHBs for pSeries guests pSeries guests can now have multiple PHBs (PCI Host Bridges), which show up as separate PCI domains in the guest. To create additional PHBs, simply add PCI controllers with model ``pci-root`` to the guest configuration. * qemu: Isolate hostdevs on pSeries guests To enable better error reporting and recovery, unrelated hostdevs will now be automatically isolated on pSeries guests by placing them on separate PHBs (PCI Host Bridges). * **Improvements** * qemu: platform serial devices can now use chardev features QEMU VMs that depend on platform serial devices can now use QEMU's ``-chardev`` option, which enables access to advanced features like log file configuration. This applies to the default serial devices for arm, aarch64, and some ppc configurations. * Require use of GCC 4.4 or Clang compilers We only ever test libvirt with GCC or Clang (which provides a GCC compatible compilation environment). Between them, these compilers cover every supported operating system platform, including Windows. * qemu: shared disks with directsync cache should be safe for migration At present shared disks can be migrated with either readonly or ``cache=none`` . But ``cache=directsync`` should be safe for migration, because both ``cache=directsync`` and ``cache=none`` don't use the host page cache, and ``cache=direct`` write through qemu block layer cache. * Handle hotplug change on VLAN configuration using OVS Libvirt was taught to handle VLAN change for running OVS interface. * **Bug fixes** * qemu: Use vCPU 'node-id' property and pass it back to qemu vCPU properties gathered from query-hotpluggable-cpus need to be passed back to QEMU. As QEMU did not use the node-id property until now and libvirt forgot to pass it back properly (it was parsed but not passed around) we did not honor this. * Miscellaneous stream fixes After introducing sparse stream features there were still some known bugs left. Those are fixed in this release. * qemu: Miscellaneous domain NS fixes Libvirt starts qemu domains in separate Linux namespaces for a while now. However, there were still some bugs lingering. For instance libvirt did not know how to handle file based bind mounts. * Various CPU driver improvements There were some minor bugs when using 'host-model' CPU. v3.5.0 (2017-07-04) =================== * **New features** * qemu: Add support for loadparm for a boot device Add an optional boot parameter 'loadparm' for a boot device. Loadparm is an 8 byte parameter that, when present, is queried by S390 guests via sclp or diag 308. Linux guests on S390 use it to select a boot entry. * Support controlling how video devices are exposed to the bhyve guests The ``vgaconf`` attribute was added to ``video`` 's ``driver`` element. Possible values are: ``on`` , ``off`` , and ``io`` . It controls the way how bhyve exposes video devices to its guests; various guest OSes might require different settings to boot properly. * qemu: Add support for live updates of coalesce settings Users can now use ``virsh update-device`` to change the coalesce settings of an interfaces while the domain is running. * qemu: Allow VirtIO devices to use vIOMMU It is now possible to turn on IOTBL for the vIOMMU and have VirtIO devices use it, provided they have been configured appropriately. * **Improvements** * qemu: block copy job can be used with persistent domains Until now it was not possible to use block copy with persistent VMs. In use cases where it's not required to recover the job after VM shutdown, it's possible to specify VIR_DOMAIN_BLOCK_COPY_TRANSIENT_JOB flag to start the copy job. * JSON pseudo-protocol backing store parser supports new format of qemu 2.9 QEMU 2.9 modified a few structures corresponding to the JSON format of specifying a backing store for a disk image. Libvirt now implements the new format. * Capabilities now include info about host's CAT settings Various information about resource control from the host is gathered and presented in capabilities if available. * apparmor: Several improvements Allow access to Ceph config, EFI firmware on both x86_64 and aarch64, device tree on ppc64 and more. * qemu: Support host-model on POWER9 machines * **Bug fixes** * qemu: snapshot: retrieve image metadata from user provided files Disk images of an external snapshot created with VIR_DOMAIN_SNAPSHOT_CREATE_REUSE_EXT flag specified would not be scanned for metadata after recent changes. The metadata is necessary to allow keeping relative paths between images when doing a block-commit. * Parse decimal numbers in a locale-independent way Some locales, such as ``de_DE`` and ``pt_BR`` , use comma rather than dot to separate the integer part from the fractional part of a decimal number; however, several data sources such as the kernel use a locale-independent representation and need to be treated accordingly. * Support compilation with newer compiler and libc versions Several fixes have been included to make compilation with Clang 4.0.0, GCC 7.1 and glibc >= 2.25.90 possible. * qemu: Query name for vhost-user interfaces at runtime This makes it possible to use ``virsh`` subcommands such as ``domiflist`` and ``domifstat`` on vhost-user interfaces. * qemu: Set MTU for hotplugged interfaces correctly When hotplugging a network interface, the MTU was only set on the guest side. Set it on the host side as well. * qemu: Forbid updating MTU for interfaces of running guests The MTU setting can't be modified while the guest is running, so any attempt to alter it at runtime will now result in an error rather than being silently ignored. * qemu: Fix specifying QXL heads with older QEMU releases Specifying the number of QXL heads was not working correctly for QEMU releases older than 1.6. * qemu: Fix migration to older libvirt/QEMU versions When the guest is started, libvirt updates the CPU definition to reflect the actual CPU features to enforce ABI. We need to send original and updated CPU definition in order to support migration to older libvirt/QEMU versions. Only the updated CPU definition was sent to destination. v3.4.0 (2017-06-02) =================== * **New features** * Improved streams to efficiently transfer sparseness New extension to virStream was implemented so that virStorageVolDownload and virStorageVolUpload can preserve file sparseness. * I/O APIC type can be specified for QEMU/KVM domains The ``ioapic`` tag was added to domain ``features`` , so the type of the I/O APIC can now be specified (e.g. putting it in userspace for KVM domains). * The reason for VM shutdown is reported, if known QEMU 2.10 will be able to report the reason for shutting down (whether that was caused by the guest or not), and libvirt is prepared for that and reports that information in its shutdown event as well, if it is known. * **Improvements** * Repository now has new README.md file The new file uses markdown syntax, so it looks better on github and possibly other web pages, but it has also more useful information. The old README is now symlink to the new file. * qemu: Use GICv2 by default for aarch64/virt TCG guests The emulated GICv3 has some limitations that make it unusable as a default; use GICv2 until they're sorted out. This change makes it once again possible to run aarch64/virt guests on a x86_64 host without having to tweak their configuration. * Additional capabilities for the node_device module Introduce two new capabilities to the node_device module. The first is for CCW devices, most common on the S390 architecture. The second is for fibre channel-backed SCSI devices and exposes the fc_remote_port sub-capability to SCSI target devices. * Node devices now report Mediated device capabilities Endpoint devices support new ``mdev`` capability and their parents now report the supported types in new ``mdev_types`` capability. * Capabilities now report information about host caches If supported in the kernel, host capabilities will now list L3 caches. The code for other levels was added as well, but only L3 caches are reported currently. * POWER9 CPU model was added It is now properly reported in host capabilities. * libxl: NUMA sibling distances are now reported in host capabilities * VMDK version 3 files are now properly detected * Interrupt remapping and Extended interrupt mode for IOMMU devices These two new features can now be controlled with new ``<driver intremap='on/off' eim='on/off'/>`` tag for ``iommu`` devices. * Graphics in libxl domains now have default addresses Even though there were default addresses before this change, they were not saved in the XML. It is now possible to see and control the listen addresses properly. * Default USB controllers are now added for devices in libxl domains Even though they were added automatically when USB device was attached, they could've been missing in some other cases. The logic is now fixed so there are always USB controllers, even if there was none of them in the specified XML. * Limits for RPC messages were increased Hitting the RPC limits we have is easier every day, so they were increased once again and some guessing logic was improved as well. It is now possible to get more stats than ever using the ``virConnectGetAllDomainStats()`` call and push through even bigger requests and replies for all APIs. * **Bug fixes** * qemu: Create memory_backing_dir on startup Libvirt's policy is that directories are created on startup if they don't exist. We've missed this one. * PCIe 4.0 cards now report proper link speeds It could happen that the link speed for PCIe devices was not properly reported or the nodedev-dumpxml just failed. That was due to mistake in the field width, but should now work properly. * qemu: Do not report errors on shutdown For some users, in some rare cases, it could happen that there was an error message "internal error: End of file from qemu monitor" in the logs even though no problem happened. The detection of these false positives was improved and such errors should not show any more. * User-specified UNIX socket paths for virtio channels should not be reset It could happen, in some cases, that libvirt would mistake a user-specified path for its own generated one and thus remove it from the XML. The detection of such addresses was improved now. * Fix address reservation during RNG hot-plug When error occurred in a specific point in time during the hot-plug of an RNG device, it could happen that an address was released even though another device was already using it, making it possible to hot-plug another device with that address, effectively having duplicated addresses in the XML. v3.3.0 (2017-05-05) =================== * **New features** * net: Add support for coalesce settings Enabling data batching through these settings can improve network performance for guests. * qemu: Add support for guest CPU cache specification This features allows fine-grained control of the cache behavior of the guest CPU. * qemu: Add support for the qemu-xhci USB controller * **Improvements** * hyperv: Support Hyper-V 2012 and newer Starting with Hyper-V 2012 the API has changed causing the existing driver to be unable to send and process requests properly. This has been resolved by adding abstractions to handle the differences and ease handling such breaks if they happen in the future. * libxl: Add support for nested HVM domains Xen has supported nested HVM domains since version 4.4. The libvirt libxl driver now supports nested HVM domains by specifying the host-passthrough CPU mode when defining a domain. * qemu: Implement ACPI support for aarch64 guests Up until this point, ACPI support was only advertised for x86_64 guests and disabling it for aarch64 guests was not possible at all. * vz: Add support for changing the number of vCPUs * qemu: Automatically choose the best USB controller for guests The recently introduced qemu-xhci USB controller is the best choice for both ppc64 and aarch64 guests, so use it by default on those architectures if available. * daemon: Increase default task limit for libvirtd The default number of tasks for the pids cgroup controller is 512, which libvirtd can quickly bump into when starting lots of guests. Raise the limit to a more reasonable 32768. * docs: Include man pages describing key code names and values * virsh: Report initialization errors Sometimes virsh might be unable to start: when that happens, report useful diagnostics instead of failing silently. * **Bug fixes** * nss: Don't require a network restart for libvirt_guest Previously, the libvirt_guest NSS module would only work properly after the corresponding network had been restarted; now newly started guests will be reported correctly right away. * storage: Remove unavailable transient pools after restart Solve an issue where transient storage pools would be stuck in an unmanageable state if the source disappeared and libvirtd was subsequently restarted. * storage: Fix capacity value for LUKS encrypted volumes The 'capacity' value (e.g. guest logical size) for a LUKS volume is smaller than the 'physical' value of the file in the file system, so we need to account for that. * qemu: Fix regression when hyperv/vendor_id feature is used Guests using the feature would not be started at all; it is now possible to start them as expected. * qemu: Do not crash on USB address with no port and invalid bus * crypto: Always pad data before encrypting it If this step is not performed, when the data length matches the chunk size the decryption routines will misinterpret the last byte of data as the padding length and fail to decode it correctly. v3.2.0 (2017-04-02) =================== * **New features** * The virt-host-validate tool now supports bhyve hypervisor * Introduce NVDIMM memory model NVDIMM is new type of memory introduced into QEMU 2.6. The idea is that we have a non-volatile memory module that keeps the data persistent across domain reboots and offers much faster data accesses. However, due to a bug in QEMU, this feature is not enabled for QEMUs older than 2.9.0. * qemu: Introduce support for generic PCIe Root Ports For new controllers, a generic device (pcie-root-port) will be used by default instead of the Intel-specific device (ioh3420), provided the QEMU binary supports it. * qemu: Add support for checking guest CPU ABI compatibility When migrating a domain to a different host, restoring a domain from a file or reverting a snapshot libvirt will make sure the guest CPU QEMU presents to the guest OS exactly matches the one provided on the source host (or before the domain's state was saved). This enhanced check may also be requested when starting a new domain to ensure the virtual CPU exactly matches the one specified in the XML. * qemu: Add support to migrate using TLS Add the ability to migrate QEMU guests using TLS via a new flag VIR_MIGRATE_TLS or virsh migrate '--tls' option. Requires using at least QEMU 2.9.0 in order to work properly. * qemu: add mediated devices framework support Recent kernel version introduced new mediated device framework, so provide an initial support of this framework for libvirt, mainly by introducing a new host device type in the XML. * qemu: Add support for setting TSC frequency Setting TSC frequency is required to enable migration for domains with 'invtsc' CPU feature turned on. * Add support for block device threshold event When using thin provisioning, management tools need to resize the disk in certain cases. To avoid having them to poll disk usage this version introduces an event which will be fired when a given offset of the storage is written by the hypervisor. Together with the API it allows registering thresholds for given storage backing volumes and this event will then notify management if the threshold is exceeded. Currently only the qemu driver supports this. * bhyve: Add support for UEFI boot ROM, VNC, and USB tablet The bhyve driver now supports booting using the UEFI boot ROM, so non-FreeBSD guests that support UEFI could be booted without using an external boot loader like grub-bhyve. Video is also supported now, allowing to connect to guests via VNC and use an USB tablet as an input device. Please refer to the driver page for domain XML examples. * **Improvements** * qemu: Detect host CPU model by asking QEMU on x86_64 Previously, libvirt detected the host CPU model using CPUID instruction, which caused libvirt to detect a lot of CPU features that are not supported by QEMU/KVM. Asking QEMU makes sure we don't start it with unsupported features. * perf: Add more perf statistics Add support to get the count of cpu clock time, task clock time, page faults, context switches, cpu migrations, minor page faults, major page faults, alignment faults, emulation faults by applications running on the platform. * Write hyperv crash information into vm log qemu's implementation of the hyperv panic notifier now reports information about the crash from the guest os. Starting with this version, libvirt logs the information to the vm log file for possible debugging. * **Bug fixes** * QEMU: Use adaptive timeout for connecting to monitor When starting qemu, libvirt waits for qemu to create the monitor socket which libvirt connects to. Historically, there was sharp 30 second timeout after which the qemu process was killed. This approach is suboptimal as in some scenarios with huge amounts of guest RAM it can take a minute or more for kernel to allocate and zero out pages for qemu. The timeout is now flexible and computed by libvirt at domain startup. * Overwrite (clear) 2 KB instead of just 512 bytes when initializing logical device * Describe the logical backend requirements better for pool-create-as v3.1.0 (2017-03-03) =================== * **New features** * storage: Add Virtuozzo storage backend storage pool Add new storage backend to support pool and volume management within the Virtuozzo Storage environment. Virtuozzo Storage is a highly available distributed software defined storage with built-in replication and disaster recovery. * qemu: Add support for memory backing with file source Add support in numa topology for file source inside memory backing (hugepages are not needed) Three new elements <source/>,<access/> and <allocation/> were added to <memoryBacking/> element. Also new configuration parameter ``memory_backing_dir`` was added to qemu.conf. * network: make openvswitch call timeout configurable Adding the ability to specify the timeout value in seconds for openvswitch calls in the libvirtd configuration file. * bhyve: add e1000 NIC support Add support for e1000 NIC. Previously, the only available option was ``virtio-net`` . * libxl: add tunneled migration support Add tunneled migration to libxl driver, which is always capable of strong encryption and doesn't require any extra network connection other than what's required for remote access of libvirtd. * qemu: add rendernode argument Add a new attribute 'rendernode' to <gl> spice element. * nodedev: add drm capability Add a new 'drm' capability for Direct Rendering Manager (DRM) devices, providing device type information. * Add API for individual/specific vCPU hotplug The new API allows selecting specific vCPUs to be added/removed from the VM. The existing APIs allowed only adding/removing from the end which did not play well with NUMA. * **Improvements** * virsh: pool-list: allow both --uuid and --name in one command Adjust the virsh-pool command to support the --uuid and/or --name options in order to print just the --name and/or --uuid of pools. * Introduce MTU to domain <interface/> and <network> Allow setting MTU size for some types of domain interface and network. * libxl: improve support for <timer> configurations Add support for multiple timers. Extend the tsc timer to support the emulate mode. Improve conversion of timer XML to/from xl.cfg. * storage: modularize the storage driver Split up the storage driver backends into loadable modules so that binary distributions don't have to compromise on shipping the storage driver with all backends which may pull in too many dependencies. * **Bug fixes** * nodedev: Fabric name must not be required for fc_host capability fabric_name is one of many fc_host attributes in Linux that is optional and left to the low-level driver to decide if it is implemented. For example the zfcp device driver does not provide a fabric name for an fcp host. The requirement for the existence of a fabric name has been removed by making it optional. * bhyve: change address allocation schema for SATA disks Previously, the bhyve driver assigned PCI addresses to SATA disks directly rather than assigning that to a controller and using SATA addresses for disks. It was implemented this way because bhyve has no notion of an explicit SATA controller. However, as this doesn't match libvirt's understanding of disk addresses, the bhyve driver was changed to follow the common schema and have PCI addresses for SATA controllers and SATA addresses for disks. If you're having issues because of this, it's recommended to edit the domain's XML and remove <address type='pci'> from the <disk> elements with <target bus='sata'/> and let libvirt regenerate it properly. * libxl: maximum memory fixes Fix reporting of domain maximum memory. Fix setting dom0 maximum memory. * libxl: fix disk detach when <driver> not specified * libxl: fix dom0 autoballooning with Xen 4.8 * qemu: Allow empty script path to <interface/> Historically, this was always allowed. Unfortunately, due to some rework done for 1.3.2 release a bug was dragged in which suddenly stop allowing domain with such configuration to start. v3.0.0 (2017-01-17) =================== * **New features** * Domain events for metadata content changes The domain events framework has a new event ID that can be used to get notifications when domain metadata content changes. * Event notifications for the secret object The secret object now supports event notifications, covering lifcycle changes and secret value changes. * New localPtr attribute for "ip" element in network XML * qemu: Support QEMU group I/O throttling Add the capability to allow group I/O throttling via a new domain <disk> <iotune> subelement "group_name" to allow sharing I/O throttling quota between multiple drives. * nss: Introduce libvirt_guest New ``libvirt_guest`` nss module that translates libvirt guest names into IP addresses. * daemon: Add support for runtime logging settings adjustment Logging-related settings like log outputs and filters can now be adjusted during runtime using the admin interface without the necessity of the daemon's restart. * storage: Add virStorageVolInfoFlags API Add the API to support using the VIR_STORAGE_VOL_GET_PHYSICAL flag in order to return the host physical size in bytes of the image container in the allocation field of the _virStorageVolInfo structure. The --physical flag has been added to the virsh vol-info command to access the data. * libxl: Implement virDomainGetMaxVcpus API * storage: Add overwrite flag checking for logical pool Add support for the OVERWRITE flags for the logical storage backend including checking for existing data on the target volumes when building a new logical pool on target volume(s). * qemu: Add support for guest CPU configuration on s390(x) * **Improvements** * perf: Add more perf statistics Add support to get the count of branch instructions executed, branch misses, bus cycles, stalled frontend cpu cycles, stalled backend cpu cycles, and ref cpu cycles by applications running on the platform. * conf: Display <physical> for volume xml Add a display of the <physical> size of a disk volume in the output of the volume XML. * qemu: Use virtio-pci by default for aarch64 mach-virt guests virtio-pci provides several advantages over virtio-mmio, such as the ability to hotplug devices and improved performance. While opting in to virtio-pci has been possible for a while, newly-defined guests will now use it automatically. * vbox: remove support for VirtualBox 3.x and older Those old VirtualBox versions have been unsupported by upstream for a long time and the API of 4.0 and newer has diverged enough to require code abstractions to handle differences. Removing support for those old versions drops lots of code from the driver and simplifies the logic to ease implementation of new features going forward. * virsh: pool-info: introduce option --bytes Add option --bytes to virsh pool-info in order ti allow display of units in bytes rather than default of human readable output. * scsi: Add parent wwnn/wwpn or fabric capability for createVport Improve the algorithm searching for the parent scsi_host device for vHBA/NPIV scsi_host creation. Rather than supplying the "parent" by name, it's now possible to define the parent by it's wwnn/wwpn or fabric_wwn in the node device create XML or the storage pool XML. * qemu: aggregate pcie-root-ports onto multiple functions of a slot When pcie-root-ports are added to pcie-root in order to provide a place to connect PCI Express endpoint devices, libvirt now aggregates multiple root ports together onto the same slot (up to 8 per slot) in order to conserve slots. Using this method, it's possible to connect more than 200 endpoint devices to a guest that uses PCIe without requiring setup of any PCIe switches. * **Bug fixes** * lxc: fix accidental killing of containers during libvirtd restart The libvirt_lxc process was previously not moved into the container scope. As a result, if systemd reloads its config after a container is started, when libvirtd is later restarted it will accidentally kill the containers. * qemu: Correct GetBlockInfo values For an active domain, correct the physical value provided for a raw sparse file backed storage and the allocation value provided for a qcow2 file backed storage that hasn't yet been opened on the domain. * qemu: Make virtio console usable on ppc64 guests The chardev detection code has been improved and can now handle this configuration properly. * qemu: Enable mount namespace To avoid funny races with udev relabelling devices under our hands and to enhance security, libvirt now spawns each qemu process with its own ``/dev`` . * storage: Fix implementation of no-overwrite for file system backend Fix file system storage backend implementation of the OVERWRITE flags to be consistent between code and documentation. Add checks to ensure that when building a new file system on a target volume that there is not something already on the disk in a format that libvirt can recognize. * qemu: Create hugepage path on per domain basis Historically, all hugepage enabled domains shared the same path under hugetlbfs. This left libvirt unable to correctly set security labels on it. With this release, however, each domain is put into a separate path which is also correctly labeled. * conf: Reject domains with duplicate drive addresses Reject duplicate drive addresses for disks and hostdevs at domain definition. * libxl: reverse defaults on HVM net device attach Fixes network interface attach for HVM domains when no model is specified. Emulated hotplug isn't yet supported and hence we should default to the general working scenario. * libxl: always enable pae for x86_64 HVM By default pae is disabled in libxl. Without an explicit <pae/> setting in the domain <features> configuration, an x86_64 HVM domain would be get an i686 environment. pae should always be enabled for x86_64 HVM domains. * qemu: Fix XML dump of autogenerated websocket As a result autogenerated websocket port is regenerated on domain restore, migration and such as it should be. v2.5.0 (2016-12-04) =================== * **New features** * shmem: Add support for additional models The shmem device can now utilize QEMU's ivshmem-plain and ivshmem-doorbell, more modern versions of ivshmem. * vbox: Add VirtualBox 5.1 support * libssh: New transport The new libssh transport allows one to connect to a running libvirtd via SSH, using the libssh library; for example: ``qemu+libssh://server/system`` . * vhost-scsi: Add support scsi_host hostdev passthrough Add the capability to pass through a scsi_host HBA and the associated LUNs to the guest. * Allow debugging of gluster volumes in qemu Users can now enable debug logging for native gluster volumes in qemu using the "gluster_debug_level" option in qemu.conf * Pre-allocate memory slots for memory hotplug Slot numbers for memory devices are now automatically allocated and thus persistent. In addition slot numbers can be specified without providing a base address, which simplifies user configuration * qemu: Express devices will be placed on PCIe bus by default For machine types that use a PCI Express root bus (e.g. x86_64/Q35 and aarch64/virt), any unaddressed PCI device that is an Express device (all virtio-1.0 devices, e1000e, nec-xhci, vfio assigned devices) will be placed on an Express controller (i.e. a pcie-root-port) instead of a legacy PCI controller (i.e. pci-bridge) with the root ports added as needed. * **Improvements** * docs: Better documentation for migration APIs and flags * vbox: Address thread safety issues * virsh: Add support for passing an alternative persistent XML to migrate command * vhostuser: Allow hotplug of multiqueue devices * NEWS: Switch to an improved format List user-visible changes instead of single commits for a better high-level overview of differences between libvirt releases. * website: Modernize layout and branding The libvirt website looked very cluttered and outdated; it has now been completely overhauled, resulting in a design that's better organized and more pleasant to look at. * **Bug fixes** * vz: Fix migration in P2P mode * Forbid newline character in names of some libvirt objects * Fix compilation on macOS v2.4.0 (2016-11-01) =================== No release notes. v2.3.0 (2016-10-04) =================== No release notes. v2.2.0 (2016-09-02) =================== No release notes. v2.1.0 (2016-08-02) =================== No release notes. v2.0.0 (2016-07-01) =================== No release notes. v1.3.5 (2016-06-04) =================== No release notes. v1.3.4 (2016-05-01) =================== No release notes. v1.3.3 (2016-04-06) =================== No release notes. v1.3.2 (2016-03-01) =================== No release notes. v1.3.1 (2016-01-17) =================== No release notes. v1.3.0 (2015-12-09) =================== No release notes. v1.2.21 (2015-11-04) ==================== No release notes. v1.2.20 (2015-10-02) ==================== No release notes. v1.2.19 (2015-09-02) ==================== No release notes. v1.2.18 (2015-08-03) ==================== No release notes. v1.2.17 (2015-07-02) ==================== No release notes. v1.2.16 (2015-06-01) ==================== No release notes. v1.2.15 (2015-05-04) ==================== No release notes. v1.2.14 (2015-04-02) ==================== No release notes. v1.2.13 (2015-03-02) ==================== No release notes. v1.2.12 (2015-01-27) ==================== No release notes. v1.2.11 (2014-12-13) ==================== No release notes. v1.2.10 (2014-11-03) ==================== No release notes. v1.2.9 (2014-10-01) =================== No release notes. v1.2.8 (2014-09-02) =================== No release notes. v1.2.7 (2014-08-02) =================== No release notes. v1.2.6 (2014-07-02) =================== No release notes. v1.2.5 (2014-06-02) =================== No release notes. v1.2.4 (2014-05-04) =================== No release notes. v1.2.3 (2014-04-01) =================== No release notes. v1.2.2 (2014-03-02) =================== No release notes. v1.2.1 (2014-01-16) =================== No release notes. v1.2.0 (2013-12-02) =================== No release notes. v1.1.4 (2013-11-04) =================== No release notes. v1.1.3 (2013-10-01) =================== No release notes. v1.1.2 (2013-09-02) =================== No release notes. v1.1.1 (2013-07-30) =================== No release notes. v1.1.0 (2013-07-01) =================== No release notes. v1.0.6 (2013-06-03) =================== No release notes. v1.0.5 (2013-05-02) =================== No release notes. v1.0.4 (2013-04-01) =================== No release notes. v1.0.3 (2013-03-05) =================== No release notes. v1.0.2 (2013-01-30) =================== No release notes. v1.0.1 (2012-12-17) =================== No release notes. v1.0.0 (2012-11-02) =================== No release notes. v0.10.2 (2012-09-24) ==================== No release notes. v0.10.1 (2012-08-31) ==================== No release notes. v0.10.0 (2012-08-29) ==================== No release notes. v0.9.13 (2012-08-02) ==================== No release notes. v0.9.12 (2012-05-14) ==================== No release notes. v0.9.11 (2012-04-03) ==================== No release notes. v0.9.10 (2012-02-13) ==================== No release notes. v0.9.9 (2012-01-07) =================== No release notes. v0.9.8 (2011-12-08) =================== No release notes. v0.9.7 (2011-11-08) =================== No release notes. v0.9.6 (2011-09-22) =================== No release notes. v0.9.5 (2011-09-20) =================== No release notes. v0.9.4 (2011-08-03) =================== No release notes. v0.9.3 (2011-07-04) =================== No release notes. v0.9.2 (2011-06-06) =================== No release notes. v0.9.1 (2011-05-05) =================== No release notes. v0.9.0 (2011-04-04) =================== No release notes. v0.8.8 (2011-02-17) =================== No release notes. v0.8.7 (2011-01-04) =================== No release notes. v0.8.6 (2010-11-30) =================== No release notes. v0.8.5 (2010-10-29) =================== No release notes. v0.8.4 (2010-09-10) =================== No release notes. v0.8.3 (2010-08-04) =================== No release notes. v0.8.2 (2010-07-05) =================== No release notes. v0.8.1 (2010-04-30) =================== No release notes. v0.8.0 (2010-04-12) =================== No release notes. v0.7.7 (2010-03-05) =================== No release notes. v0.7.6 (2010-02-03) =================== No release notes. v0.7.5 (2009-12-23) =================== No release notes. v0.7.4 (2009-11-20) =================== No release notes. v0.7.3 (2009-11-20) =================== No release notes. v0.7.2 (2009-10-14) =================== No release notes. v0.7.1 (2009-09-15) =================== No release notes. v0.7.0 (2009-08-05) =================== No release notes. v0.6.5 (2009-07-03) =================== No release notes. v0.6.4 (2009-05-29) =================== No release notes. v0.6.3 (2009-04-24) =================== No release notes. v0.6.2 (2009-04-03) =================== No release notes. v0.6.1 (2009-03-03) =================== No release notes. v0.6.0 (2009-01-31) =================== No release notes. v0.5.1 (2008-12-04) =================== No release notes. v0.5.0 (2008-11-25) =================== No release notes. v0.4.6 (2008-09-23) =================== No release notes. v0.4.5 (2008-09-08) =================== No release notes. v0.4.4 (2008-06-25) =================== No release notes. v0.4.3 (2008-06-12) =================== No release notes. v0.4.2 (2008-04-08) =================== No release notes. v0.4.1 (2008-03-03) =================== No release notes. v0.4.0 (2007-12-18) =================== No release notes. v0.3.3 (2007-09-30) =================== No release notes. v0.3.2 (2007-08-21) =================== No release notes. v0.3.1 (2007-07-24) =================== No release notes. v0.3.0 (2007-07-09) =================== No release notes. v0.2.3 (2007-06-08) =================== No release notes. v0.2.2 (2007-04-17) =================== No release notes. v0.2.1 (2007-03-16) =================== No release notes. v0.2.0 (2007-02-14) =================== No release notes. v0.1.11 (2007-01-22) ==================== No release notes. v0.1.10 (2006-12-20) ==================== No release notes. v0.1.9 (2006-11-29) =================== No release notes. v0.1.8 (2006-10-16) =================== No release notes. v0.1.7 (2006-09-29) =================== No release notes. v0.1.6 (2006-09-22) =================== No release notes. v0.1.5 (2006-09-05) =================== No release notes. v0.1.4 (2006-08-16) =================== No release notes. v0.1.3 (2006-07-11) =================== No release notes. v0.1.2 (2006-07-03) =================== No release notes. v0.1.1 (2006-06-21) =================== No release notes. v0.1.0 (2006-04-10) =================== No release notes. v0.0.6 (2006-02-28) =================== No release notes. v0.0.5 (2006-02-23) =================== No release notes. v0.0.4 (2006-02-10) =================== No release notes. v0.0.3 (2006-02-09) =================== No release notes. v0.0.2 (2006-01-29) =================== No release notes. v0.0.1 (2005-12-19) =================== No release notes. .. _git log: https://gitlab.com/libvirt/libvirt/-/commits/master

================ libvirt releases ================ This is the list of official releases for libvirt, along with an overview of the changes introduced by each of them. For a more fine-grained view, use the `git log`_. v10.4.0 (unreleased) ==================== * **Security** * **Removed features** * **New features** * qemu: Support for ras feature for virt machine type It is now possible to set on/off ``ras`` feature in the domain XML for virt (Arm) machine type as ````. * SSH proxy for VM Libvirt now installs a binary helper that allows connecting to QEMU domains via SSH using the following scheme: ``ssh user@qemu/virtualMachine``. * qemu: Support for ``virtio`` sound model Sound devices can now be configured to use the virtio model with ````. This model is available from QEMU 8.2.0 onwards. * **Improvements** * **Bug fixes** * virsh/virt-admin: Fix ``--help`` option for all commands A bug introduced in `v10.3.0 (2024-05-02)`_ caused that the attempt to print help for any command by using the ``--help`` option in ``virsh`` and ``virt-admin`` would print:: $ virsh list --help error: command 'list' doesn't support option --help instead of the help output. A workaround for the affected version is to use the help command:: $ virsh help list v10.3.0 (2024-05-02) ==================== * **New features** * qemu: Proper support for USB network device USB address is now automatically assigned to USB network devices thus they can be used without manual configuration. * conf: Introduce memReserve attribute to Some PCI devices have large non-prefetchable memory. This can be a problem in case when such device needs to be hotplugged as the firmware can't foresee such situation. The user thus can override the value calculated at start to accomodate for such devices. * **Improvements** * Improve validation of USB devices Certain USB device types ('sound', 'fs', 'chr', 'ccid' and 'net') were not properly handled in the check whether the VM config supports USB and thus would result in poor error messages. * virsh: Fix behaviour of ``--name`` and ``--parent`` used together when listing checkpoint and snapshots The ``checkpoint-list`` and ``snapshot-list`` commands would ignore the ``--name`` option to print only the name when used with ``--parent``. * Extend libvirt-guests to shutdown only persistent VMs Users can now choose to shutdown only persistent VMs when the host is being shut down. * **Bug fixes** * qemu: Fix migration with custom XML Libvirt 10.2.0 would sometimes complain about incompatible CPU definition when trying to migrate or save a domain and passing a custom XML even though such XML was properly generated as migratable. Hitting this bug depends on the guest CPU definition and the host on which a particular domain was running. * qemu: Fix TLS hostname verification failure in certain non-shared storage migration scenarios In certain scenarios (parallel migration, newly also post-copy migration) libvirt would wrongly pass an empty hostname to QEMU to be used for TLS certificate hostname validation, which would result into failure of the non-shared storage migration step:: error: internal error: unable to execute QEMU command 'blockdev-add': Certificate does not match the hostname * Create OVS ports as transient Libvirt now creates OVS ports as transient which prevents them from reappearing or going stale on sudden reboots. * Clear OVS QoS settings when domain shuts down Libvirt now clears QoS settings on domain shutdown, so they no longer pile up in OVS database. v10.2.0 (2024-04-02) ==================== * **New features** * ch: Basic save and restore support for ch driver The ch driver now supports basic save and restore operations. This is functional on domains without any network, host device config defined. The ``path`` parameter for save and restore should be a directory. * qemu: Support for driver type ``mtp`` in ```` devices The ``mtp`` driver type exposes the ``usb-mtp`` device in QEMU. The guest can access files on this driver through the Media Transfer Protocol (MTP). * qemu: Added support for the loongarch64 architecture It is now possible for libvirt to run loongarch64 guests, including on other architectures via TCG. For the best results, it is recommended to use the upcoming QEMU 9.0.0 release together with the development version of edk2. * qemu: Introduce virDomainGraphicsReload API Reloading the graphics display is now supported for QEMU guests using VNC. This is useful to make QEMU reload the TLS certificates without restarting the guest. Available via the ``virDomainGraphicsReload`` API and the ``domdisplay-reload`` virsh command. * **Bug fixes** * qemu: Fix migration from libvirt older than 9.10.0 when vmx is enabled A domain with vmx feature enabled (which may be even done automatically with ``mode='host-model'``) started by libvirt 9.9.0 or older cannot be migrated to libvirt 9.10.0, 10.0.0, and 10.1.0 as the target host would complain about a lot of extra ``vmx-*`` features. Migration of similar domains started by the affected releases to libvirt 9.9.0 and older does not work either. Since libvirt 10.2.0 migration works again with libvirt 9.9.0 and older in both directions. Migration from the affected releases to 10.2.0 works as well, but the other direction remains broken unless the fix is backported. * node_device: Don't report spurious errors from PCI VPD parsing In last release the PCI Vital Product Data parser was enhanced to report errors but that effort failed as some kernels have the file but don't allow reading it causing logs to be spammed with:: libvirtd[21055]: operation failed: failed to read the PCI VPD data Since the data is used only in the node device XML and errors are ignored if the parsing failed, this release removes all the error reporting. * qemu: set correct SELinux label for unprivileged virtiofsd It is now possible to use virtiofsd-based ```` shares even if the guest is confined using SELinux. * qemu: fix a crash on unprivileged virtiofsd hotplug Hotplugging virtiofsd-based filesystems works now. * virt-admin: Fix segfault when libvirtd dies ``virt-admin`` no longer crashes when ``libvirtd`` unexpectedly closes the connection. v10.1.0 (2024-03-01) ==================== * **Security** * ``CVE-2024-1441``: Fix off-by-one error leading to a crash In **libvirt-1.0.0** there were couple of interface listing APIs introduced which had an off-by-one error. That error could lead to a very rare crash if an array was passed to those functions which did not fit all the interfaces. In **libvirt-5.10** a check for non-NULL arrays has been adjusted to allow for NULL arrays with size 0 instead of rejecting all NULL arrays. However that made the above issue significantly worse since that off-by-one error now did not write beyond an array, but dereferenced said NULL pointer making the crash certain in a specific scenario in which a NULL array of size 0 was passed to the aforementioned functions. * **New features** * nodedev: Support updating mdevs The node device driver has been extended to allow updating mediated node devices. Options are available to target the update against the persistent, active or both configurations of a mediated device. **Note:** The support is only available with at least mdevctl v1.3.0 installed. * qemu: Add support for /dev/userfaultfd On hosts with new enough kernel which supports /dev/userfaultfd libvirt will now automatically grant QEMU access to this device. It's no longer needed to set vm.unprivileged_userfaultfd sysctl. * qemu: Support clusters in CPU topology It is now possible to configure the guest CPU topology to use clusters. Additionally, if CPU clusters are present in the host topology, they will be reported as part of the capabilities XML. * network: Make virtual domains resolvable from the host When starting a virtual network with a new ``register='yes'`` attribute in the ```` element, libvirt will configure ``systemd-resolved`` to resolve names of the connected guests using the name server started for this network. * qemu: Introduce dynamicMemslots attribute for virtio-mem QEMU now allows setting ``.dynamic-memslots`` attribute for virtio-mem-pci devices. When turned on, it allows memory exposed to guest to be split into multiple memory slots and thus smaller memory footprint (see the original commit for detailed explanation). * **Improvements** * nodedev: Add ability to update persistent mediated devices by defining them Existing persistent mediated devices can now also be updated by ``virNodeDeviceDefineXML()`` as long as parent and UUID remain unchanged. * ch: Enable ``ethernet`` interface mode support ```` can now be used for CH domains. * viraccessdriverpolkit: Add missing vtpm case Secrets with ```` were left unable to be checked for in the access driver, i.e. in ACL rules. Missing code was provided. * virt-admin: Notify users to use explicit URI if connection fails ``virt-admin`` doesn't try to guess the URI of the daemon to manage so a failure to connect may be confusing for users if modular daemons are used. Add a hint to use the URI of the dameon to manage. * **Bug fixes** * qemu_process: Skip over non-virtio non-TAP NIC models when refreshing rx-filter If ``trustGuestRxFilters`` is enabled for a vNIC that doesn't support it, libvirt may throw an error when such domain is being started, loaded from a saved state, migrated, etc. These errors are now silenced, but make sure to fix such configurations (after previous release it is even possible to change ``trustGuestRxFilters`` value on live domains via ``virDomainUpdateDeviceFlags()`` or ``virsh device-update``). * domain: Fix check for overlapping ```` devices A bug was identified which caused libvirt to report two NVDIMMs as overlapping even though they weren't. This now fixed. * vmx: Accept empty fileName for cdrom-image Turns out, ``fileName`` attribute (which contains path to CDROM image) can be set to an empty string (``""``) to denote a state in which the CDROM has no medium in it. Libvirt used to reject such configuration file, but not anymore. * qemu_hotplug: Don't lose 'created' flag in qemuDomainChangeNet() When starting a domain, libvirt tracks what resources it created for it and which were pre-existing and uses this information to preserve pre-existing resources when cleaning up after said domain is shut off. But for macvtaps this information was lost after the macvtap device was changed (e.g. via ``virsh update-device``). * Fix virStream hole handling When a client sent multiple holes into a virStream it may have caused daemon hangup as the daemon stopped processing RPC from the client temporarily. This is now fixed. * nodedev: Don't generate broken XML with certain hardware A broken node device XML would be generated in a rare case when a hardware device had certain characters in the VPD fields. * qemu: Fix reservation of manually specified port for disk migration A manually specified port would not be relased after disk migration making it impossible to use it again. v10.0.0 (2024-01-15) ==================== * **New features** * qemu: Enable ``postcopy-preempt`` migration capability Post-copy migrations are now started with ``postcopy-preempt`` capability enabled as long as it is supported by both sides of migration. This should enable faster migration of memory pages that the destination tries to read before they are migrated from the source. * qemu: Add support for mapping iothreads to virtqueues of ``virtio-blk`` devices QEMU added the possibility to map multiple ``iothreads`` to a single ``virtio-blk`` device and map them even to specific virtqueues. Libvirt adds a ```` subelement of the `` `` element that users can use to configure the mapping. * qemu: Allow automatic resize of block-device-backed disk to full size of the device The new flag ``VIR_DOMAIN_BLOCK_RESIZE_CAPACITY`` for ``virDomainBlockResize`` allows resizing a block-device backed ``raw`` disk of a VM without the need to specify the full size of the block device. * qemu: automatic selection/binding of VFIO variant drivers When a device is assigned to a guest using VFIO with ````, libvirt will now search the running kernel's modules.alias file for the most specific match to that device for a VFIO driver, and bind that driver to the device rather than vfio-pci. A specific driver can also be forced, using the ```` attribute. * qemu: add runtime configuration option for nbdkit Since the new nbdkit support requires a recent selinux policy that is not widely available yet, it is now possible to build libvirt with nbdkit support for remote disks but disabled at runtime. This behavior is controlled via the storage_use_nbdkit option of the qemu driver configuration file. The option will default to being disabled, but this may change in a future release and can be customized with the nbdkit_config_default build option. * qemu: add ID mapping support for virtiofsd New ```` element was added for virtiofsd-based ```` devices. It can be used to set up UID and GID mapping between host and guest, making running virtiofsd unprivileged much more useful. * **Improvements** * qemu: Improve migration XML use when persisting VM on destination When migrating a VM with a custom migration XML, use it as a base for persisting it on the destination as users could have changed non-ABI breaking facts which would prevent subsequent start if the old XML were used. * qemu: Simplify non-shared storage migration to ``raw`` block devices The phase of copying storage during migration without shared storage requires that both the source and destination image are identical in size. This may not be possible if the destination is backed by a block device and the source image size is not a multiple of the block device block size. Libvirt aleviates this by automatically adding a ```` to match the size of the source image rather than failing the migration. * test driver: Support for hotplug/hotunplug of PCI devices The test driver now supports basic hotplug and hotunplug of PCI devices. * qemu: allow virtiofsd to run unprivileged Nowadays virtiofsd no longer requires to run with root privileges, so the restriction to always run as root is now removed from libvirt too. * **Bug fixes** * qemu: Various migration bug fixes and debuggability improvement This release fixes multiple bugs in virsh and libvirt in handling of migration arguments and XMLs and modifies error reporting for better debugging. * conf: Restore setting default bus for input devices Because of a regression, starting from 9.3.0 libvirt did not autofill bus for input devices. With this release the regression was identified and fixed. * qemu: Relax check for memory device coldplug Because of a check that was too aggressive, a virtio-mem memory device could not be cold plugged. This is now fixed. * qemu: Be less aggressive when dropping channel source paths Another regression is resolved, (introduced in 9.7.0) when libvirt was too aggressive when dropping parsed paths for sources * qemuDomainChangeNet: Reflect trustGuestRxFilters change On device-update, when a user requested change of trustGuestRxFilters for a domain's libvirt did nothing. It did not throw an error nor did it reflect the change. Starting with this release, the change is reflected. v9.10.0 (2023-12-01) ==================== * **New features** * Introduce pipewire audio backend The QEMU hypervisor driver now allows setting ``pipewire`` backend for ``