3 <!-- libvirt release notes
5 This file will be processed to produce both HTML and plain text versions
8 Keep the style consistent with existing entries as much as possible.
10 Each change should be documented by a short, one-sentence summary, which
11 should fit in a single line and should not contain any formatting tags.
13 You can optionally add a description if you feel like the summary alone is
14 not enough to document the change accurately. The description may contain
15 a <code> tag for switching to non-proportional font. No other tags are
18 Lines should be kept under 80 columns, and should not exceed 100 columns.
20 This file is validated against docs/news.rng schema.
22 Use the following template to add a new release section:
24 <release version="FIXME" date="unreleased">
25 <section title="New features">
27 <section title="Improvements">
29 <section title="Bug fixes">
33 If relevant for a given release, add a section for removed
36 <section title="Removed features">
42 <release version="v5.7.0" date="unreleased">
43 <section title="New features">
46 qemu: Support Direct Mode for Hyper-V Synthetic timers
49 The QEMU driver now supports Direct Mode for Hyper-V Synthetic timers
54 <section title="Improvements">
56 <section title="Bug fixes">
59 <release version="v5.6.0" date="2019-08-05">
60 <section title="New features">
63 qemu: Introduce a new video model of type 'bochs'
66 Introduce a new video model type that supports the
67 <code>bochs-display</code> device that was added in qemu version 3.0.
72 api: new virDomainCheckpoint APIs
75 Introduce several new APIs for creating and managing
76 checkpoints in the test and qemu drivers (the latter
77 requires qcow2 images). Checkpoints serve as a way to tell
78 which portions of a disk have changed since a point in time.
83 qemu: Add support for overriding max threads per process limit
86 systemd-based systems impose a limit on the number of threads a
87 process can spawn, which in some cases can be exceeded by QEMU
88 processes running VMs. Add a <code>max_threads_per_process</code>
89 option to qemu.conf to override the system default.
94 Remember original owners and SELinux labels of files
97 When a domain is starting up libvirt changes DAC and
98 SELinux labels so that domain can access it. However,
99 it never remembered the original labels and therefore
100 the file was returned back to <code>root:root</code>.
101 With this release, the original labels are remembered
102 and restored properly.
107 network: Allow passing arbitrary options to dnsmasq
110 This works similarly to the existing support for passing arbitary
111 options to QEMU, and just like that feature it comes with no
116 <section title="Removed features">
119 xen: Remove sxpr config support
122 Remove the sxpr style config parser and formatter a year after the
123 xend driver was removed.
127 <section title="Improvements">
130 qemu: Allow XML validation for snapshot creation
133 Add flag <code>VIR_DOMAIN_SNAPSHOT_CREATE_VALIDATE</code> to validate
134 snapshot input XML. For virsh, users can use it as <code>virsh
135 snapshot-create --validate</code>.
140 Support encrypted soft TPM
143 A soft TPM backend could be encrypted with passphrase. Now libvirt
144 supports using a <code>secret</code> object to hold the passphrase,
145 and referring to it via the <code>encryption</code> element of the
151 test driver: Expand API coverage
154 Additional APIs have been implemented in the test driver.
159 Implement per-driver locking
162 Drivers now acquire a lock when they're loaded, ensuring that there
163 can never be two instances of the same driver active at a time.
168 nss: Report newer addresses first
171 In some cases, a guest might be assigned a new IP address by DHCP
172 before the previous lease has expired, in which case the NSS plugin
173 will correctly report both addresses; many applications, however,
174 ignore all addresses but the first, and may thus end up trying to
175 connect using a stale address. To prevent that from happening, the
176 NSS plugin will now always report the newest address first.
181 util: Optimize mass closing of FDs when spawning child processes
184 When the limit on the number of FDs is very high, closing all
185 unwanted FDs after calling <code>fork()</code> can take a lot of
186 time and delay the start of the child process. libvirt will now
187 use an optimized algorithm that minimizes such delays.
191 <section title="Bug fixes">
194 logging: Ensure virtlogd rollover takes priority over logrotate
197 virtlogd implements its own rollover mechanism, but until now
198 logrotate could end up acting on the logs before virtlogd had a
199 chance to do so itself.
204 <release version="v5.5.0" date="2019-07-02">
205 <section title="Security">
208 api: Prevent access to several APIs over read-only connections
211 Certain APIs give root-equivalent access to the host, and as such
212 should be limited to privileged users. CVE-2019-10161,
213 CVE-2019-10166, CVE-2019-10167, CVE-2019-10168.
217 <section title="New features">
220 qemu: Support SMMUv3 IOMMU
223 SMMUv3 is an IOMMU implementation for ARM virt guests.
228 network: Introduce the network port API
231 This new public API can be used by virtualization drivers to
232 manage network resources associated with guests, and is a further
233 step towards splitting libvirtd into multiple daemons.
237 <section title="Removed features">
240 qemu: Remove support for virDomainQemuAttach and virConnectDomainXMLFromNative APIs
243 The qemu implementations for the APIs mentioned above were removed
244 and the APIs now return an error. The implementation was stale for a
245 long time and did not work with modern QEMU command lines, generated
246 from libvirt or otherwise.
251 Stop supporting migration of config files from pre-XDG layout
254 The new layout was introduced with libvirt 0.9.13 (Jul 2012).
259 Remove Avahi mDNS support
262 This feature was never used outside of virt-manager, which has
263 itself stopped using it a while ago.
267 <section title="Improvements">
270 sysinfo: Report SMBIOS information on aarch64
273 While SMBIOS support has historically been limited to x86_64, modern
274 aarch64 machines often offer access to the same information as well,
275 and libvirt now exposes it to the user when that's the case.
280 test driver: Expand API coverage
283 Even more APIs that were missing from the test driver have now been
289 virt-xml-validate: Allow input to be read from stdin
294 qemu: Validate spapr-vio addresses as 32-bit
297 libvirt has always considered these addresses (used for pSeries
298 guests) as 64-bit, but the sPAPR specification says that they're
303 <section title="Bug fixes">
306 qemu: Set process affinity correctly when using <numatune>
309 libvirt would mistakenly interpret the <code>nodeset</code>
310 attribute as a list of CPUs instead of as a list of NUMA node, and
311 the process affinity would be set incorrectly as a result; this has
317 <release version="v5.4.0" date="2019-06-03">
318 <section title="Security">
321 cpu: Introduce support for the md-clear CPUID bit
324 This bit is set when microcode provides the mechanism to invoke a
325 flush of various exploitable CPU buffers by invoking the x86
326 <code>VERW</code> instruction. CVE-2018-12126, CVE-2018-12127,
327 CVE-2018-12130, CVE-2019-11091.
332 Restrict user access to virt-admin, virtlogd and virtlockd
335 The intended users for these facilities are the <code>root</code>
336 user and the <code>libvirtd</code> service respectively, but these
337 restrictions were not enforced correctly. CVE-2019-10132.
341 <section title="Improvements">
344 test driver: Expand API coverage
347 Several APIs that were missing from the test driver have now been
353 Avoid unnecessary static linking
356 Most binaries shipped as part of libvirt, for example
357 <code>virtlogd</code> and <code>libvirt_iohelper</code>, were
358 embedding parts of the library even though they also linked against
359 the <code>libvirt.so</code> dynamic library. This is no longer the
360 case, which results in both the disk and memory footprint being
366 qemu: Report stat-htlb-pgalloc and stat-htlb-pgfail balloon stats
369 These stats have been introduced in QEMU 3.0.
373 <section title="Bug fixes">
376 qemu: Fix emulator scheduler support
379 Setting the scheduler for QEMU's main thread before QEMU had a chance
380 to start up other threads was misleading as it would affect other
381 threads (vCPU and I/O) as well. In some particular situations this
382 could also lead to an error when the thread for vCPU #0 was being
383 moved to its cpu,cpuacct cgroup. This was fixed so that the
384 scheduler for the main thread is set after QEMU starts.
389 apparmor: Allow hotplug of vhost-scsi devices
394 <release version="v5.3.0" date="2019-05-04">
395 <section title="New features">
398 qemu: Add support for setting the emulator scheduler parameters
401 I/O threads and vCPU threads already support setting schedulers, but
402 until now it was impossible to do so for the main QEMU thread
403 (emulator thread in the libvirt naming). This is, however, requested
404 for some very specific scenarios, for example when vCPU threads are
405 running at such priority that could starve the main thread.
409 <section title="Removed features">
412 vbox: Drop support for VirtualBox 4.x releases
415 Support for all the 4.x releases was ended by
416 VirtualBox maintainers in December 2015. Therefore,
417 libvirt support for these releases is dropped.
421 <section title="Improvements">
424 qemu: Use PCI by default for RISC-V guests
427 PCI support for RISC-V guests was already available in libvirt
428 5.1.0, but it required the user to opt-in by manually assigning
429 PCI addresses: with this release, RISC-V guests will use PCI
430 automatically when running against a recent enough (4.0.0+) QEMU
436 qemu: Advertise firmware autoselection in domain capabilities
439 The firmware autoselection feature is now exposed in
440 domain capabilities and management applications can
441 query for accepted values, i.e. values that are accepted
442 and for which libvirt found firmware descriptor files.
443 Firmware Secure Boot support is also advertised.
451 YAJL 2 is widely adopted and maintaining side by side
452 support for two versions is unnecessary.
456 <section title="Bug fixes">
459 rpc: cleanup in virNetTLSContextNew
462 Failed new gnutls context allocations in virNetTLSContextNew
463 function results in double free and segfault. Occasional memory
464 leaks may also occur.
469 virsh: various completers fixes
472 There were some possible crashers, memory leaks, etc.
478 qemu: Make hugepages work with memfd backend
481 Due to a bug in command line generation libvirt did not
482 honor hugepages setting with memfd backend.
487 Enforce ACL write permission for getting guest time & hostname
490 Getting the guest time and hostname both require use of
491 guest agent commands. These must not be allowed for
492 read-only users, so the permissions check must validate
493 "write" permission not "read".
498 <release version="v5.2.0" date="2019-04-03">
499 <section title="New features">
502 Add Storage Pool Capabilities output
505 Add support to list an enumerated list of supported Storage
506 Pools via the virConnectGetCapabilities API when connected
507 via a Storage Driver. Add support to get a more detailed
508 list XML output Storage Pool Capabilities vis the
509 virConnectGetStoragePoolCapabilites API.
514 qemu: Support virtio-{non-}transitional device models
517 <code>virtio-transitional</code> and
518 <code>virtio-non-transitional</code> <code>model</code> values
519 were added to the QEMU driver for the following devices:
520 <code>disk</code>, <code>interface</code>, <code>filesystem</code>,
521 <code>rng</code>, <code>vsock</code>, <code>memballoon</code>,
522 <code>controller</code> type <code>scsi</code>,
523 <code>controller</code> type <code>virtio-serial</code>,
524 <code>input</code> bus <code>virtio</code>
525 type <code>passthrough</code>,
526 <code>hostdev</code> type <code>scsi_host</code>. These new
527 models can be used to give fine grained control over what
528 virtio device version is presented to the guest.
533 qemu: Enable firmware autoselection
536 Libvirt allows users to provide loader path for some time now.
537 However, this puts some burden on users because they need to
538 know what firmware meets their requirements. Now that QEMU
539 ships firmware description files this burden can be moved onto
540 libvirt. It is as easy as setting the <code>firmware</code>
541 attribute in the <code>os</code> element (accepted values are
542 <code>bios</code> and <code>efi</code>). Moreover, libvirt
543 automatically enables domain features needed for firmware it
549 snapshots: Add support for topological listings
552 A new flag VIR_DOMAIN_SNAPSHOT_LIST_TOPOLOGICAL is available
553 for the various snapshot listing APIs such as
554 virDomainListAllSnapshots(). For drivers that support the
555 flag, the listed snapshots are guaranteed to be sorted such
556 that parents occur before children.
561 Xen: Add support for max grant frames setting
564 Add support for Xen's max_grant_frames setting by adding a
565 new xenbus controller type with a maxGrantFrames attribute.
566 E.g. <code><controller type='xenbus' maxGrantFrames='64'/></code>
571 qemu: Add support for parallel migration
574 With QEMU 4.0.0 libvirt can enable parallel migration which causes
575 the memory pages to be processed in parallel by several threads and
576 sent to the destination host using several connections at the same
577 time. This may increase migration speed in case a single thread is
578 unable to saturate the network link.
582 <section title="Removed features">
585 Drop support for Upstart and "Red Hat" init scripts
588 Not a single one of the platforms we target still uses Upstart,
589 and the Upstart project itself has been abandoned for several years
590 now; the same is true for the "Red Hat" (really System V) init
591 scripts, since RHEL 7 and later releases use systemd.
595 <section title="Improvements">
598 Report class information for PCI node device capability.
603 Split setup of IPv4 and IPv6 top level chain
606 The requirement resulting from private chains improvement done
607 in <code>v5.1.0</code> was refined so that only tables from
608 corresponding IP version are required. This means that if a
609 network doesn't have <code>IPv6</code> enabled then those
610 tables are not required.
615 Don't default to building the QEMU driver
618 Historically, the QEMU driver has been special in that it was
619 enabled by default, with the option to explicitly opt-out of it;
620 starting now, we're enabling it opportunistically if we detect that
621 all requirements are available, just like we do with other drivers.
625 <section title="Bug fixes">
628 virt-host-validate: Fix IOMMU check on s390x
633 qemu: Allow creating pSeries guests with graphics and no USB mouse
636 It's now possible to prevent libvirt from automatically adding a
637 USB mouse to pSeries guests by including a USB tablet in the input
638 XML: doing so is desiderable as using a tablet results in a much
639 better user experience when working with GUIs.
644 qemu: Set $HOME and XGD variables for qemu:///system guests
647 This avoids files being accidentally created under <code>/</code> or
648 the guests not being able to start because they lack the necessary
649 permissions to write to that location.
654 <release version="v5.1.0" date="2019-03-04">
655 <section title="New features">
658 bhyve: Add support for additional command-line arguments
661 The bhyve driver now supports passing additional command-line
662 arguments to the bhyve process using the new
663 <code><bhyve:commandline></code> element in domain
669 network: Support setting a firewalld "zone" for virtual network bridges
672 All libvirt virtual networks with bridges managed by libvirt
673 (i.e. those with forward mode of "nat", "route", "open", or
674 no forward mode) will now be placed in a special firewalld
675 zone called "libvirt" by default. The zone of any network
676 bridge can be changed using the <code>zone</code> attribute
677 of the network's <code>bridge</code> element.
682 bhyve: Support for ignoring unknown MSRs reads and writes
685 A new <features> element <msrs unknown='ignore'/> was
686 introduced and the bhyve driver supports it to control unknown
687 Model Specific Registers (MSRs) reads and writes.
692 qemu: Add support for encrypted VNC TLS keys
695 Use the password stored in the secret driver under the uuid
696 specified by the <code>vnc_tls_x509_secret_uuid</code> option
702 Add storage pool namespace options
705 Allow for adjustment of RBD configuration options via Storage
706 Pool XML Namespace adjustments.
711 qemu: Add support for setting post-copy migration bandwidth
714 Users can now limit the bandwidth of post-copy migration, e.g.
715 via <code>virsh migrate --postcopy-bandwidth</code>.
719 <section title="Improvements">
722 Create private chains for virtual network firewall rules
725 Historically firewall rules for virtual networks were added
726 straight into the base chains. This works but has a number of
727 bugs and design limitations. To address them, libvirt now puts
728 firewall rules into its own chains. Note that with this change the
729 <code>filter</code>, <code>nat</code> and <code>mangle</code> tables
730 are required for both <code>IPv4</code> and <code>IPv6</code>.
735 Detect CEPH and GPFS as shared FS
738 When starting a migration libvirt performs some sanity checks
739 to make sure domain will be able to run on the destination.
740 One of the requirements is that the disk has to either be
741 migrated too or be accessible from a network filesystem. CEPH
742 and GPFS weren't detected as a network filesystem.
747 Advertise network MTU via DHCP when specified
750 If network MTU is set and the network has DHCP enabled,
751 advertise the MTU in DHCP transaction too so that clients can
752 adjust their link accordingly.
757 qemu: Allocate memory at the configured NUMA nodes from start
760 Libvirt used to just start QEMU, let it allocate memory for
761 the guest, and then use CGroups to move the memory to
762 configured NUMA nodes. This is suboptimal as huge chunks of
763 memory have to be moved. Moreover, this relies on ability to
764 move memory later which is not always true. A change was made
765 to set process affinity correctly from the start so that memory
766 is allocated on the configured nodes from the beginning.
771 Support for newer Wireshark
774 Adapt libvirt to use the more recent release requiring a
775 source build configuration of libvirt
776 <code>--with-wireshark</code> to upgrade to the more recent
782 Batch mode virsh and virt-admin parsing improvements
785 When parsing a single-argument command_string in batch mode,
786 virsh and virt-admin now permit newlines in addition to
787 semicolons for splitting commands, and backslash-newline for
788 splitting long lines, to be more like shell parsing.
792 <section title="Bug fixes">
795 qemu: Use CAP_DAC_OVERRIDE during QEMU capabilities probing
798 By default, libvirt runs the QEMU process as <code>qemu:qemu</code>
799 which could cause issues during probing as some features like AMD SEV
800 might be inaccessible to QEMU because of file system permissions.
801 Therefore, <code>CAP_DAC_OVERRIDE</code> is granted to overcome these
802 for the purposes of probing.
807 storage: Add default mount options for fs/netfs storage pools
810 Altered the command line generation for fs/netfs storage pools to
811 add some default options. For Linux based systems, the options
812 added are "nodev, nosuid, noexec". For FreeBSD based systems,
813 the options added are "nosuid, noexec".
818 qemu: Allow use of PCI for RISC-V guests
821 This works with QEMU 4.0.0+ only and is opt-in at the moment, since
822 it requires users to manually assign PCI addresses, but is otherwise
828 network: Fix virtual networks on systems using firewalld+nftables
831 Because of the transitional state of firewalld's new support
832 for nftables, not all iptables features required by libvirt
833 are yet available, so libvirt must continue to use iptables
834 for its own packet filtering rules even when the firewalld
835 backend is set to use nftables. However, due to the way
836 iptables support is implemented in kernels using nftables
837 (iptables rules are converted to nftables rules and
838 processed in a separate hook from the native nftables
839 rules), guest networking was broken on hosts with firewalld
840 configured to use nftables as the backend. This has been
841 fixed by putting libvirt-managed bridges in their own
842 firewalld zone, so that guest traffic can be forwarded
843 beyond the host and host services can be exposed to guests
844 on the virtual network without opening up those same
845 services to the rest of the physical network. This means
846 that host access from virtual machines is no longer
847 controlled by the firewalld default zone (usually "public"),
848 but rather by the new firewalld zone called "libvirt"
849 (unless configured otherwise using the new zone
850 attribute of the network bridge element).
855 qemu: Fix i6300esb watchdog hotplug on Q35
858 Ensure that libvirt allocates a PCI address for the device so
859 that QEMU did not default to an address that would not allow
865 lxc: Don't reboot host on virDomainReboot
868 If the container is really a simple one (init is just bash and
869 the whole root is passed through) then virDomainReboot and
870 virDomainShutdown would reboot or shutdown the host. The
871 solution is to use different method to reboot or shutdown the
872 container in that case (e.g. signal).
877 rpc: Various stream fixes
880 One particular race was fixed, one locking problem and error
881 reporting from streams was made better.
886 qemu: Fix guestfwd hotplug/hotunplug
889 Fixed the generation of the guestfwd hotplug/unplug command
890 sent to QEMU to match the syntax used when creating the
891 initial command line.
896 qemu: Forbid CDROMs on virtio bus
899 Attempting to create an empty virtio-blk drive or attempting
900 to eject it results into an error. Forbid configurations
901 where users would attempt to use CDROMs in virtio bus.
906 qemu: Use 'raw' for 'volume' disks without format
909 Storage pools might want to specify format of the image when
910 translating the volume thus libvirt can't add any default
911 format when parsing the XML. Add an explicit format when
912 starting the VM and format is not present neither by user
913 specifying it nor by the storage pool translation function.
918 qemu: Assume 'raw' default storage format also for network storage
921 Post parse callback adds the 'raw' type only for local files.
922 Remote files can also have backing store (even local) so we
923 should do this also for network backed storage.
928 qemu: Fix block job progress reporting and advocate for READY event
931 In some cases QEMU can get to 100% and still not reach the
932 synchronised phase. Initiating a pivot in that case will fail.
933 Therefore it is strongly advised to wait for
934 <code>VIR_DOMAIN_BLOCK_JOB_READY</code> event which does not
935 suffer from this problem.
940 qemu: Don't format image properties for empty drive
943 If a <code>-drive</code> has no image, then formatting
944 attributes such as cache, readonly, etc. would cause errors to
945 be reported from QEMU. This was fixed by not supplying the
946 attributes for devices without an image.
951 External snapshot metadata redefinition is fixed
954 Attempting to use VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE to
955 reinstate the metadata describing an external snapshot
956 created earlier for an offline domain no longer fails.
961 <release version="v5.0.0" date="2019-01-15">
962 <section title="New features">
965 Xen: Add support for openvswitch
968 The libxl driver now supports virtual interfaces that connect to
969 an openvswitch bridge, including interfaces with VLAN tagging and
970 trunking configuration.
975 qemu: Report whether KVM nesting is available
978 Running nested KVM guests requires specific configuration steps to
979 be performed on the host; libvirt will now report in the host
980 capabilities whether KVM nesting support is available.
984 <section title="Removed features">
990 The UML driver was unmaintained and not tested for
991 quite some time now. Worse, there is a bug that causes
992 it to deadlock on some very basic operations (e.g.
993 dumping domain XML). These facts make us believe no one
998 <section title="Improvements">
1001 qemu: Add support for ARMv6l guests
1006 Support more NVDIMM configuration options
1009 Introduce more configuration options. For the source element, add
1010 the 'alignsize' and 'pmem' subelements. For the target element, add
1011 the 'readonly' subelement.
1016 cpu: Add support for "stibp" x86_64 feature
1019 Add cpu flag stibp (Single Thread Indirect Branch Predictors) to
1020 prevent indirect branch predictions from being controlled by the
1021 sibling Hyperthread.
1026 libxl: Handle external domain destroy
1029 Historically, if a domain was destroyed using <code>xl</code>
1030 rather than through libvirt APIs, libvirt would not be aware of
1031 the fact and keep considering it as running. This is no longer the
1037 Start selecting the first available DRI device for OpenGL operations
1040 If OpenGL support is needed (either with SPICE gl enabled or with
1041 egl-headless), libvirt is now able to pick the first available DRI
1042 device for the job. At the same time, this improvement is also a
1043 bugfix as it prevents permission-related issues with regards to our
1044 mount namespaces and the default DRI render node's permissions which
1045 would normally prevent QEMU from accessing such a device.
1050 qemu: Add support for postcopy-requests migration statistics
1053 The <code>virDomainJobInfo</code> can get number page requests
1054 received from the destination host during post-copy migration.
1058 <section title="Bug fixes">
1061 lxc: Don't forbid interfaces with type=direct
1064 Such interfaces are supported by lxc and should be allowed.
1069 qemu: Fully clean up RNG devices on detach
1072 Some RNG device types, such as those using EGD, might need extra
1073 clean up on the host in addition to removing the guest-side device.
1078 <release version="v4.10.0" date="2018-12-03">
1079 <section title="New features">
1082 qemu: Add Hyper-V PV IPI and Enlightened VMCS support
1085 The QEMU driver now has support for Hyper-V PV IPI and Enlightened VMCS
1086 for Windows and Hyper-V guests.
1091 qemu: Added support for PCI devices on S390
1094 PCI addresses can now include the new zpci element which contains
1095 uid (user-defined identifier) and fid (PCI function identifier)
1096 attributes and makes the corresponding devices usable by S390
1102 Support changing IOThread polling parameters for a live guest
1105 Introduced virDomainSetIOThreadParams which allows dynamically
1106 setting the IOThread polling parameters used by QEMU to manage
1107 the thread polling interval and the algorithm for growth or
1108 shrink of the polling time. The values only affect a running
1109 guest with IOThreads. The guest's IOThread polling values can
1110 be viewed via the domain statistics.
1115 Xen: Add support for PVH
1118 The libxl driver now supports Xen's PVH virtual machine type.
1119 PVH machines are enabled with the new "xenpvh" OS type, e.g.
1120 <code><os><type>xenpvh</type></os></code>
1125 qemu: Added support for CMT (Cache Monitoring Technology)
1128 Introduced cache monitoring using the <code>monitor</code>
1129 element in <code>cachetune</code> for vCPU threads. Added
1130 interfaces to get and display the cache utilization statistics
1131 through the command 'virsh domstats' via the
1132 virConnectGetAllDomainStats API.
1137 qemu: Add support for nested HV for pSeries guests
1140 Nested HV support makes it possible to run nested (L2) guests
1141 with minimal performance penalty when compared to regular (L1)
1142 guests on ppc64 hardware.
1146 <section title="Bug fixes">
1149 Xen: Handle soft reset shutdown event
1152 The pvops Linux kernel uses soft reset to handle the crash
1153 machine operation. The libxl driver now supports the soft
1154 reset shutdown event, allowing proper crash handling of
1155 pvops-based HVM domains.
1160 <release version="v4.9.0" date="2018-11-04">
1161 <section title="New features">
1164 util: Add cgroup v2 support
1167 cgroup v2 support has been implemented in libvirt, with both
1168 "unified" (v2 only) and "hybrid" (v2 + v1) setups being usable;
1169 existing "legacy" (v1 only) setups will keep working.
1174 qemu: Add vfio AP support
1177 The QEMU driver now has support to passthrough adjunct processors
1178 into QEMU guests on S390.
1182 <section title="Improvements">
1185 rpc: Make 'genprotocol' output reproducible
1188 This is another step towards making libvirt builds fully
1193 <section title="Bug fixes">
1196 security: Fix permissions for UNIX sockets
1199 Since 4.5.0, libvirt is using FD passing to hand sockets over to
1200 QEMU, which in theory removes the need for them to be accessible by
1201 the user under which the QEMU process is running; however, other
1202 processes such as vdsm need to access the sockets as well, which
1203 means adjusting permissions is still necessary.
1208 cpu_map: Add Icelake model definitions
1211 These CPU models will be available in the upcoming 3.1.0 QEMU
1217 util: Properly parse URIs with missing trailing slash
1220 Some storage URIs were not parsed correctly, in which case libvirt
1221 ended up emitting XML that it would then refuse to parse back.
1226 <release version="v4.8.0" date="2018-10-01">
1227 <section title="New features">
1230 Xen: Support PM Suspend and Wakeup
1233 The libxl driver now supports the virDomainPMSuspendForDuration
1234 and virDomainPMWakeup APIs.
1238 <section title="Removed features">
1241 Xen: Drop support for Xen 4.4 and 4.5
1244 Xen 4.4 and 4.5 are no longer supported by the Xen community.
1245 Drop support for these older versions and require Xen >= 4.6.
1250 nwfilter: Disallow binding creation in session mode
1253 Ensure that a filter binding creation is not attempted in session
1254 mode and generates a proper error message.
1258 <section title="Improvements">
1261 qemu: Retrieve guest hostname through QEMU Guest Agent command
1264 QEMU is now able to retrieve the guest hostname using a new QEMU-GA
1265 command called 'guest-get-host-name'. Virsh users can execute
1266 'domhostname' for QEMU driver for domains configured to use the
1272 virsh: Implement vsh-table in virsh and virsh-admin
1275 The new API fixes problems with table-alignment, making the tables
1276 more readable and deals with unicode.
1280 <section title="Bug fixes">
1283 storage: Allow inputvol to be encrypted
1286 When creating a storage volume based on another volume, the base
1287 input volume is allowed to be encrypted.
1292 virsh: Require explicit --domain for domxml-to-native
1295 The --domain option for domxml-to-native virsh command has always
1296 been documented as required, but commit v4.3.0-127-gd86531daf2
1297 accidentally made it optional.
1302 lxc_monitor: Avoid AB / BA lock race
1305 A deadlock situation could occur when autostarting a LXC domain
1306 'guest' due to two threads attempting to take opposing locks while
1307 holding opposing locks (AB BA problem).
1312 <release version="v4.7.0" date="2018-09-03">
1313 <section title="New features">
1316 storage: add storage pool iscsi-direct
1319 Introduce a new storage pool backend that uses libiscsi instead of
1320 iscsiadm. It support basic pool operations: checkPool and refreshPool.
1325 Add support for MBA (Memory Bandwidth Allocation technology)
1328 Domain vCPU threads can now have allocated some parts of host memory
1329 bandwidth by using the <code>memorytune</code> element in <code>cputune</code>.
1334 qemu: Add support for RISC-V guests
1337 riscv32 and riscv64 guest architectures are now supported.
1341 <section title="Improvements">
1344 qemu: Add ccw support for vhost-vsock
1347 Support the vhost-vsock-ccw device on S390.
1352 qemu: Make default machine type independent of QEMU
1355 We can't control whether or not QEMU will change its default
1356 machine type in the future, or whether downstream distributions
1357 will decide to compile out some machine types, so our only option
1358 to provide a predictable behavior is taking care of the default
1359 ourselves; management applications and users are encouraged to
1360 explicitly pick a machine type when creating new guests.
1365 apparmor: Various improvements
1368 Rules have been added to deal with a number of scenarios that
1369 didn't work correctly.
1373 <section title="Bug fixes">
1376 esx: Truncate CPU model name
1379 Some CPU model names are too long to be stored into the
1380 corresponding property, and should be explicitly truncated
1381 to avoid unexpected behavior in users of the
1382 <code>virNodeGetInfo()</code> API such as
1383 <code>virsh nodeinfo</code>.
1388 utils: Remove arbitrary limit on socket_id/core_id
1391 Both values were assumed to be smaller than 4096, but in fact
1392 they are entirely hardware-dependent and there have been reports
1393 of machines presenting much bigger values, preventing libvirt from
1394 working correctly; all such limits have now been removed.
1399 <release version="v4.6.0" date="2018-08-06">
1400 <section title="New features">
1403 qemu: Implement the HTM pSeries feature
1406 Users can now decide whether HTM (Hardware Transactional Memory)
1407 support should be available to the guest.
1412 qemu: Enable VNC console for mediated devices
1415 Host devices now support a new atribute 'display' which can be used
1416 to turn on frame buffer rendering on a vgpu mediated device instead of
1417 on an emulated GPU, like QXL.
1421 <section title="Improvements">
1424 qemu: Introduce a new video model of type 'none'
1427 Introduce a new video model type that disables the automatic addition
1428 of a video device to domains with 'graphics' specified in their XML.
1429 This can be useful with GPU mediated devices which can serve as the
1430 only rendering devices within the guest.
1435 virsh: Add --alias to attach-disk and attach-interface commands
1438 Add option --alias to set customized device alias name when
1439 using attach-disk or attach-interface commands.
1444 virsh: Support usb and sata address to attach-disk
1447 Usb or sata address could be used when attach-disk with --address.
1448 For example, use usb address as usb:<bus>.<port>, use
1449 sata address as <controller>.<bus>.<unit>.
1454 <release version="v4.5.0" date="2018-07-02">
1455 <section title="New features">
1458 qemu: Provide TPM emulator support
1461 Support QEMU's TPM emulator based on swtpm. Each QEMU guest gets
1462 its own virtual TPM.
1467 bhyve: Support specifying guest CPU topology
1470 Bhyve's guest CPU topology could be specified using the
1471 <code><cpu><topology ../></cpu></code> element.
1476 qemu: Add support for extended TSEG size
1479 Support specifying extended TSEG size for SMM in QEMU.
1484 qemu: Add support for SEV guests
1487 SEV (Secure Encrypted Virtualization) is a feature available on AMD
1488 CPUs that encrypts the guest memory and makes it inaccessible even
1493 <section title="Removed features">
1496 Remove support for qcow/default encrypted volumes
1499 Disallow using a qcow encrypted volume for the guest and
1500 disallow creation of the qcow or default encrypted volume
1501 from the storage driver. Support for qcow encrypted volumes
1502 has been phasing out since QEMU 2.3 and by QEMU 2.9 creation
1503 of a qcow encrypted volume via qemu-img required usage of
1504 secret objects, but that support was never added to libvirt.
1509 Make GnuTLS mandatory
1512 Building without GnuTLS is no longer possible.
1517 qemu: Remove allow_disk_format_probing configuration option
1520 The option represented a security risk when used with malicious
1521 disk images, so users were recommended against enabling it; with
1522 this release, it's been removed altogether.
1526 <section title="Improvements">
1529 capabilities: Provide info about host IOMMU support
1532 Capabilities XML now provide information about host IOMMU support.
1537 virsh: Add --all to domblkinfo command
1540 Alter the <code>domblkinfo</code> command to add the option
1541 --all in order to display the size details of each domain
1542 block device from one command in a output table.
1547 qemu: Allow concurrent access to monitor and guest agent
1550 Historically libvirt prevented concurrent accesses to
1551 the qemu monitor and the guest agent. Therefore two
1552 independent calls (one querying the monitor and the
1553 other querying guest agent) would serialize which hurts
1554 performance. The code was reworked to allow two
1555 independent calls run at the same time.
1560 qemu: Allow configuring the page size for HPT pSeries guests
1563 For HPT pSeries guests, the size of the host pages used to back guest
1564 memory and the usable guest page sizes are connected; the new setting
1565 can be used to request that a certain page size is available in the
1571 Add support to use an raw input volume for encryption
1574 It is now possible to provide a raw input volume as input for
1575 to generate a luks encrypted volume via either virsh vol-create-from
1576 or virStorageVolCreateXMLFrom.
1581 qemu: Add support for vsock hot (un)plug and cold (un)plug
1586 qemu: Add support for NBD over TLS
1589 NBD volumes can now be accessed securely.
1594 qemu: Implement FD passing for Unix sockets
1597 Instead of having QEMU open the socket and then connecting to it,
1598 which is inherently racy, starting with QEMU 2.12 we can open the
1599 socket ourselves and pass it to QEMU, avoiding race conditions.
1604 virsh: Introduce --nowait option for domstat command
1607 When this option is specified, virsh will try to fetch the guest
1608 stats but abort instead of stalling if they can't be retrieved right
1613 <section title="Bug fixes">
1616 qemu: Fix a potential libvirtd crash on VM reconnect
1619 Initialization of the driver worker pool needs to come before libvirtd
1620 trying to reconnect to all machines, since one of the QEMU processes
1621 migh have already emitted events which need to be handled prior to
1622 us getting to the worker pool initialization.
1627 qemu: Fix domain resume after failed migration
1630 Recent versions of QEMU activate block devices before the guest CPU
1631 has been started, which makes it impossible to roll back a failed
1632 migration. Use the <code>late-block-activate</code> migration
1633 capability if supported to avoid the issue.
1638 vmx: Permit guests to have an odd number of vCPUs
1641 An odd number of vCPUs greater than 1 was forbidden in the past,
1642 but current versions of ESXi have lifted that restriction.
1647 <release version="v4.4.0" date="2018-06-04">
1648 <section title="New features">
1651 bhyve: Support locking guest memory
1654 Bhyve's guest memory may be wired using the
1655 <code><memoryBacking><locked/></memoryBacking></code>
1661 qemu: Provide VFIO channel I/O passthrough support
1664 Support passthrough devices that use channel I/O based mechanism in
1665 a QEMU virtual machine.
1670 qemu: Add support for migration of VMs with non-shared storage over TLS
1673 It's now possible to use the VIR_MIGRATE_TLS flag together with
1674 VIR_MIGRATE_NON_SHARED_DISK. The connection is then secured using the
1675 TLS environment which is setup for the migration connection.
1680 Add support for VM Generation ID
1683 The VM Generatation ID exposes a 128-bit, cryptographically
1684 random, integer value identifier, referred to as a Globally
1685 Unique Identifier (GUID) to the guest in order to notify the
1686 guest operating system when the virtual machine is executed
1687 with a different configuration. Add a new domain XML processing
1688 and a domain capabilities feature.
1693 Introduce virDomainDetachDeviceAlias
1696 This new API enables users to detach device using only its alias.
1701 Introduce new virConnectCompareHypervisorCPU and virConnectBaselineHypervisorCPU APIs
1704 Unlike the old virConnectCompareCPU and virConnectBaselineCPU APIs,
1705 both new APIs consider capabilities of a specific hypervisor.
1710 Introduce SCSI persistent reservations support
1713 The QEMU driver gained support for qemu-pr-helper which enables
1714 guests to issue SCSI commands for persistent reservation.
1719 qemu: Implement multiple screen support for virDomainScreenshot
1722 While the virDomainScreenshot API supported multihead video cards,
1723 the implementation was missing. But now that QEMU implemented it
1724 libvirt has done as well.
1729 qemu: add support for vhost-vsock-device
1732 A new vsock device was introduced, allowing communication between
1733 the guest and the host via the AF_VSOCK family.
1737 <section title="Improvements">
1740 qemu: Add suport for OpenGL rendering with SDL
1743 Domains using SDL as a graphics backend will now be able to use
1744 OpenGL accelerated rendering.
1749 qemu: Add support for 'output' audio codec
1752 Support QEMU's 'hda-output' codec advertising only a line-out for ich6
1753 and ich9 sound devices.
1758 virsh: Enhance event name completion
1761 Implement event name completion for some commands (e.g. event,
1762 secret-event, pool-event and nodedev-event)
1767 <release version="v4.3.0" date="2018-05-02">
1768 <section title="New features">
1771 qemu: Add support for the pcie-to-pci-bridge controller
1774 Pure PCIe guests such as x86_64/q35 and aarch64/virt will now
1775 add this controller when traditional PCI devices are in use.
1780 Xen: Support setting CPU features for host-passthrough model
1783 The CPU model presented to Xen HVM domains is equivalent to libvirt's
1784 host-passthrough model, although individual features can be enabled
1785 and disabled via the cpuid setting. The libvirt libxl driver now
1786 supports enabling and disabling individual features of the
1787 host-passthrough CPU model.
1791 <section title="Removed features">
1794 Xen: Drop the legacy xend-based driver
1797 The xm/xend toolstack was deprecated in Xen 4.2 and removed
1798 from the Xen sources in the 4.5 development cycle. The libvirt
1799 driver based on xend is now removed from the libvirt sources.
1803 <section title="Improvements">
1806 qemu: Support hot plug and hot unplug of mediated devices
1809 Libvirt now allows mediated devices to be hot plugged and hot
1810 unplugged from a guest rather than reporting an error that this isn't
1811 supported. In fact, kernel has been supporting this since 4.10.
1815 <section title="Bug fixes">
1818 Improve handling of device mapper targets
1821 When starting a domain with a disk backed by a device
1822 mapper volume libvirt also needs to allow the storage
1823 backing the device mapper in CGroups. In the past
1824 kernel did not care, but starting from 4.16 CGroups are
1825 consulted on each access to the device mapper target.
1830 <release version="v4.2.0" date="2018-04-01">
1831 <section title="New features">
1834 Support building with Python 3
1837 Python is required to build libvirt, and up until now only Python 2
1838 could be used as an interpreter. All scripts used during build have
1839 now been made compatible with Python 3, which means both major
1840 releases of the language are fully supported.
1845 qemu: Provide ccw address support for graphics and input devices
1848 Support the virtio-gpu-ccw device as a video device and
1849 virtio-{keyboard, mouse, tablet}-ccw devices as input devices
1854 <section title="Improvements">
1857 qemu: Add logging of guest crash information on S390
1860 On S390, when the guest crashes and QEMU exposes the guest crash
1861 information, log the relevant data to the domain log file.
1866 qemu: use arp table of host to get the IP address of guests
1869 Find IP address of a VM by arp table on hosts.
1870 If someone customizing IP address inside VM, it will be helpful.
1875 Xen: Remove hard-coded scheduler weight
1878 The libxl driver was accidentally hard-coding the per-domain
1879 scheduler weight to 1000, silently ignoring any user-provided
1880 <code><shares></code> in <code><cputune></code>. The
1881 driver now honors <code><shares></code>, and defers setting
1882 a default value to Xen. Note that the Xen default is 256, so any
1883 domains started after this improvement will have one fourth the
1884 shares of previously started domains. If all domains must have
1885 equal CPU shares, administrators must manually set the weight of
1886 previously started domains to 256, or restart them.
1890 <section title="Bug fixes">
1893 qemu: TLS migration now enforces use of TLS for the NBD connection
1896 When the VIR_MIGRATE_TLS flag was used with the migration API libvirt
1897 did not ensure that the NBD connection was using TLS as well. The code
1898 now rejects such migration as the TLS transport for NBD is not ready
1899 yet, but prevents a false sense of security that TLS would be used.
1900 The support TLS for NBD will be added soon.
1905 <release version="v4.1.0" date="2018-03-05">
1906 <section title="New features">
1909 Added support for CAT (Cache allocation Technology)
1912 Domain vCPU threads can now have allocated some parts of host cache
1913 using the <code>cachetune</code> element in <code>cputune</code>.
1918 Allow opening secondary drivers
1921 Up until now it was possible to connect to only hypervisor drivers
1922 (e.g. qemu:///system, lxc:///, vbox:///system, and so on). The
1923 internal drivers (like network driver, node device driver, etc.) were
1924 hidden from users and users could use them only indirectly. Starting
1925 with this release new connection URIs are accepted. For instance
1926 network:///system, storage:///system and so on.
1931 virtlogd, virtlockd: Add support for admin protocol
1934 These two daemons now support admin protocol through which some admin
1935 info can be gathered or some configuration tweaked on the fly.
1939 <section title="Improvements">
1942 virsh: Enhance bash completion
1945 Implement more bash completions so that basic libvirt
1946 objects can be auto-completed (e.g. networks,
1947 interfaces, NWFilters, and so on).
1952 qemu: Use VIR_ERR_DEVICE_MISSING for various hotplug/detach messages
1957 qemu: Allow showing the dump progress for memory only dump
1960 Alter the QEMU dump-guest-memory command processing to check
1961 for and allow asynchronous completion which then allows for
1962 the virsh dump --memory-only --verbose command to display percent
1968 conf: add support for setting Chassis SMBIOS data fields
1973 libxl: add support for setting clock offset and adjustment
1978 Make port allocator global
1981 Up until now each driver had their own port allocator module. This
1982 meant that info on port usage was not shared. Starting with this
1983 release, the port allocator module is made global and therefore
1984 drivers allocate ports from global pool.
1989 Fixed some compiler warnings that appear with GCC 8
1993 <section title="Bug fixes">
1996 qemu: Check for unsafe migration more thoroughly
1999 If a domain disk is stored on local filesystem (e.g. ext4) but is
2000 not being migrated it is very likely that domain is not able to
2001 run on destination. Regardless of share/cache mode.
2006 qemu: Fix updating device with boot order
2009 Starting with 3.7.0 release updating any device with boot order would
2010 fail with 'boot order X is already used by another device' while in
2011 fact it was the very same device.
2016 virlog: determine the hostname on startup CVE-2018-6764
2019 At later point it might not be possible or even safe to use
2020 getaddrinfo(). It can in turn result in a load of NSS module which
2021 can even be loaded from unsage guest filesystem and thus escape the
2022 confinment of its container.
2027 qemu: Rework vCPU statistics fetching
2030 Fetching vCPU statistics was very expensive because it lead to waking
2031 up vCPU threads in QEMU and thus it degraded performance. The code
2032 was reworked so that fetching statistics does not wake up halted
2038 qemu: unlink memory backing file on domain shutdown
2041 Depending on the filesystem where domain memory is stored, some files
2042 might have been left behind. This is not a problem on hugetlbfs, but
2043 it is a problem on regular filesystems like ext4.
2048 qemu: Fix shutting down domains in parallel
2051 If multiple domains were being shut down in parallel, libvirtd might
2057 nodedev: Update PCI mdev capabilities dynamically
2060 PCI devices may have other nested capabilities, like SRIOV and mdev
2061 which depend on the device being plugged into the native vendor
2062 driver. However, in case such a device is directly assigned to a guest
2063 using VFIO driver, the device will naturally lose these capabilities
2064 and libvirt needs to reflect that.
2069 <release version="v4.0.0" date="2018-01-19">
2070 <section title="New features">
2073 tools: Provide bash completion support
2076 Both <code>virsh</code> and <code>virt-admin</code> now implement
2077 basic bash completion support.
2082 qemu: Refresh capabilities on host microcode update
2085 A microcode update can cause the CPUID bits to change; therefore,
2086 the capabilities cache should be rebuilt when such an update is
2087 detected on the host.
2092 lxc: Set hostname based on container name
2096 <section title="Improvements">
2099 CPU frequency reporting improvements
2102 The CPU frequency will now be reported by <code>virsh nodeinfo</code>
2103 and other tools for s390 hosts; at the same time; CPU frequency has
2104 been disabled on aarch64 hosts because there's no way to detect it
2110 libxl: Mark domain0 as persistent
2115 Xen: Add support for multiple IP addresses on interface devices
2120 qemu: Add support for hot unplugging redirdev device
2124 <section title="Bug fixes">
2127 qemu: Enforce vCPU hotplug granularity constraints
2130 QEMU 2.7 and newer don't allow guests to start unless the initial
2131 vCPUs count is a multiple of the vCPU hotplug granularity, so
2132 validate it and report an error if needed.
2137 <release version="v3.10.0" date="2017-12-04">
2138 <section title="New features">
2141 conf: Support defining distances between virtual NUMA cells
2144 A NUMA hardware architecture supports the notion of distances
2145 between NUMA cells. This can now be specified using the
2146 <code><distances></code> element within the NUMA cell
2147 configuration. Drivers which support this include Xen and QEMU.
2152 Xen: Support defining vNUMA topology
2155 Xen now supports defining a virtual NUMA topology for VMs,
2156 including specifying distances between NUMA cells.
2161 qemu: Add the ability to configure HPT resizing for pSeries guests
2164 The user can now decide whether HPT (Hash Page Table) resizing
2165 should be enabled, disabled or required instead of leaving it up to
2166 hypervisor defaults and negotiation between the guest and the host.
2171 qemu: Add vmcoreinfo feature
2174 Starting with QEMU 2.11, the guest can save kernel debug
2175 details when this feature is enabled and the kernel supports
2176 it. It is useful to process kernel dump with KASLR enabled,
2177 and also provides various kernel details to crash tools.
2182 conf: Move the auth and encryption definitions to disk source
2185 Allow parsing and formatting of the <code>auth</code> and
2186 <code>encryption</code> sub-elements to be a child of the
2187 <code>source</code> element. This will allow adding an
2188 <code>auth</code> sub-element to a <code>backingStore</code>
2189 or <code>mirror</code> elements as a means to track specific
2190 authentication and/or encryption needs.
2194 <section title="Improvements">
2197 vbox: Add VirtualBox 5.2 support
2202 vbox: Add support for configuring storage controllers
2205 The VirtualBox driver now supports the <code><controller></code>
2206 element in the domain XML for configuring storage controllers in VBOX
2207 VMs. Additionally, libvirt's domain XML schema was updated to allow
2208 optional <code>model</code> attribute for <code><controller
2209 type='ide'></code> which is used by the VBOX driver to set the
2210 IDE controller model to be one of 'piix4', 'piix4' (default), or
2211 'ich6'. Finally, with this change <code>dumpxml</code> generates
2212 <code><controller></code> elements that correspond to current
2213 VBOX VM storage controller configuration.
2218 vbox: Add support for attaching empty removable disks
2221 The VirutalBox driver now supports adding CD-ROM and floppy disk
2222 devices that do not have the disk source specified. Previously such
2223 devices were silently ignored.
2228 vbox: Add support for attaching SAS storage controllers
2231 In VirtualBox, SCSI and SAS are distinct controller types whereas
2232 libvirt does not make such distinction. Therefore, the VBOX driver was
2233 updated to allow attaching SAS controllers via <code><controller
2234 type='scsi' model='lsisas1068'></code> element. If there are
2235 both SCSI and SAS controllers present in the VBOX VM, the domain XML
2236 can associate the disk device using the <code><address></code>
2237 element with the <code>controller</code> attribute, and optionally,
2238 set the port via <code>unit</code> attribute.
2243 qemu: Generate predictable paths for qemu memory backends
2246 In some cases management applications need to know
2247 paths passed to memory-backend-file objects upfront.
2248 Libvirt now generates predictable paths so applications
2249 can prepare the files if they need to do so.
2254 Shareable disks work properly with recent qemu
2257 Recent qemu versions added image locking to avoid potential corruption
2258 of disk images. This broke shareable disks with libvirt since the
2259 feature was turned on by default in qemu. Libvirt now enables sharing
2260 of those disks in qemu so that the image locking is not applied in
2261 that case. Additionally libvirt now checks that shareable disks have
2262 supported format (raw) to avoid metadata corruption.
2267 Improve serial console behavior on non-x86 architectures
2270 ppc64, aarch64 and s390x guests were treating the <serial>
2271 and <console> elements differently from x86, in some cases
2272 presenting misleading information to the user. The behavior is now
2273 consistent across all architectures and the information reported
2278 <section title="Bug fixes">
2281 vbox: Do not ignore failures to attach disk devices when defining
2284 The <code>define</code> now fails and reports an error if any of the
2285 <code>controller</code> or <code>disk</code> devices specified in the
2286 domain XML fail to attach to the VirtualBox VM.
2291 vbox: Fix dumpxml to always output disk devices
2294 The VirtualBox driver was ignoring any disk devices in
2295 <code>dumpxml</code> output if there was a SAS storage controller
2301 vbox: Fix dumpxml to always generate valid domain XML
2304 When a VirtualBox VM has multiple disks attached, each to a different
2305 storage controller that uses 'sd' prefix for block device names e.g.
2306 one disk attached to SATA and one to SCSI controller, it no longer
2307 generates XML where both would have 'sda' device name assigned.
2308 Instead it properly assigns 'sda' and 'sdb' to those disks in the
2309 order of appearance.
2314 Securely pass iSCSI authentication data
2317 Rather than supplying the authentication data as part of the
2318 iSCSI URL for a disk or host device, utilize the encrypted
2319 secret object to securely pass the authentication data.
2324 <release version="v3.9.0" date="2017-11-02">
2325 <section title="New features">
2328 Add capability to allow hot (un)plug of a domain watchdog device
2333 Allow users to set device aliases
2336 Users can set aliases to domain devices and thus identify them
2342 qemu: Support multiqueue for virtio-blk
2345 Multiqueue support for <code>virtio-blk</code> has been available
2346 in QEMU ever since 2.7.0, and now libvirt guests can enable it.
2351 Add virDomainSetLifecycleAction API
2354 Provided a new API to allow dynamic guest lifecycle control for
2355 guest reactions to poweroff, restart, or crash type events related
2356 to the domain XML <code>on_poweroff</code>, <code>on_reboot</code>,
2357 and <code>on_crash</code> elements. The
2358 <code>virsh set-lifecycle-action</code> command was created to
2359 control the actions.
2364 qemu: Allow cold(un)plugging and hot(un)plugging input devices
2369 net: Implement QoS for vhostuser
2373 <section title="Improvements">
2376 Allow a logical volume to be create using LUKS
2379 A logical volume may be created using an <code>encryption</code>
2380 element using "luks" format. This does require a previously created
2381 <code>secret</code> to store the passphrase used to encrypt the
2382 volume Adding the volume to a domain can then either provide the
2383 secret or allow the consumer in the guest to provide the passphrase
2384 in order to decrypt the volume.
2389 net: Ignore auto-generated MAC address when detaching an interface
2392 If the MAC address has not been specified by the user, libvirt will
2393 try and fill in the gaps by generating one; however, for some error
2394 paths that led to some confusing error messages, so when an
2395 auto-generated MAC address is specified the error message will not
2396 include the auto-generated MAC.
2401 net: Enable MAC address lookup for virDomainInterfaceStats
2406 apparmor: Several improvements
2409 Changes include permitting access to data about USB devices and
2410 <code>dnsmasq</code> instances, allowing spaces in guest names and
2416 cpu: Use CPU information obtained from QEMU when possible
2419 Recent QEMU versions can expose information about which CPU models
2420 are available and usable on the host; libvirt will now make use of
2421 such information whenever possible.
2426 hyperv: Various improvements
2429 The error reported when clients can't connect to Hyper-V has been
2430 made more descriptive, and memory limits for guests are now mapped
2431 to more appropriate libvirt equivalents.
2436 qemu: Report QEMU error on failed migration
2439 Instead of reporting a generic error, ask QEMU for a more detailed
2440 and thus hopefully more helpful one.
2445 vbox: Implement autoport for RDP
2448 libvirt will now obtain the (dynamically allocated) RDP port number
2449 from VirtualBox itself, avoiding conflicts between multiple guests
2450 wanting to use RDP at the same time.
2455 qemu: Allow rotation of small logs
2458 On a host where numerous unique instances are executed per day, it's
2459 quite possible that, even though each of the single log files are
2460 fairly small, collectively the quantity and volume may add tens of
2461 thousands of log files to the <code>/var/log/libvirt/qemu/</code>
2462 directory. Removing the constraints that log have to be bigger than
2463 100 KiB before they can be rotated solves the issue.
2467 <section title="Bug fixes">
2470 Fix swapped interface statistics and QoS
2473 Due to internal implementation, reported statistics for
2474 some types of interfaces were swapped (RX appeared in
2475 TX and vice versa). Similarly, QoS was set in reversed
2481 Properly resize local LUKS encrypted volume
2484 Resizing of a local LUKS encrypted volume will now use qemu-img
2485 to resize the volume. This will require configuring a secret for
2486 the LUKS encrypted volume.
2491 qemu: Reserve PCI addresses for implicit i440fx devices
2494 Failing to do so causes the addresses to be considered usable by
2495 libvirt, which means they could be assigned to more than one device
2496 resulting in the guest failing to start.
2501 spec: Restart libvirtd only at the end of the upgrade process
2504 Use <code>%posttrans</code> to make sure <code>libvirtd</code>
2505 is not restarted before all other components, such as the library
2506 itself and storage / hypervisor drivers, have already been upgraded.
2510 <section title="Security">
2513 qemu: Ensure TLS clients always verify the server certificate
2516 While it's reasonable to turn off client certificate validation,
2517 as setting it up can be non-trivial, clients should always verify
2518 the server certificate to avoid MITM attacks. However, libvirt was
2519 using the same knob to control both checks, leading to
2520 CVE-2017-1000256 / LSN-2017-0002.
2525 <release version="v3.8.0" date="2017-10-04">
2526 <section title="New features">
2529 qemu: Added support for cold-(un)plug of watchdog devices
2534 qemu: Added support for setting IP address os usernet interfaces
2539 qemu: Added support for Veritas Hyperscale (VxHS) block devices
2544 storage: Added new events for pool-build and pool-delete
2548 <section title="Improvements">
2551 qemu: Set DAC permissions properly for spice rendernode
2554 When a <code>rendernode</code> path is set for SPICE GL on
2555 <code>qemu:///system</code>, we now correctly set DAC permissions
2556 on the device at VM startup. This is the last remaining hurdle to
2557 let SPICE GL work for <code>qemu:///system</code> without any
2558 external host changes.
2563 nodedev: Add switchdev offload query to NIC capabilities
2566 Allow querying the NIC interface capabilities for the
2567 availability of switchdev offloading (also known as
2568 kernel-forward-plane-offload).
2573 New CPU models for AMD and Intel
2576 AMD EPYC and Intel Skylake-Server CPU models were added together with
2582 Improve long waiting when saving a domain
2585 While waiting for a write to disk to be finished, e.g. during save,
2586 even simple operations like <code>virsh list</code> would be blocking
2587 due to domain lock. This is now resolved by unlocking the domain
2588 in places where it is not needed.
2592 <section title="Bug fixes">
2595 Proper units are now used in virsh manpage for dom(mem)stats
2598 Previously the documentation used multiples of 1000, but now it is
2599 fixed to use multiples of 1024.
2604 qemu: Fix error reporting when disk attachment fails
2607 There was a possibility for the actual error to be overridden or
2608 cleared during the rollback.
2613 qemu: Fix assignment of graphics ports after daemon restart
2616 This could be seen with newer kernels that have bug regarding
2617 SO_REUSEADDR. After libvirtd was restarted it could assign already
2618 used address to new guests which would make them fail to start. This
2619 is fixed by marking used ports unavailable when reconnecting to
2620 running QEMU domains.
2625 Fix message decoding which was causing a very strange bug
2628 When parsing an RPC message with file descriptors was interrupted and
2629 had to restart, the offset of the payload was calculated badly causing
2630 strange issues like not being able to find a domain that was not
2636 <release version="v3.7.0" date="2017-09-04">
2637 <section title="New features">
2640 qemu: Add managedsave-edit commands
2643 Using managedsave-dumpxml, managedsave-define and managedsave-edit
2644 commands, now we can dump and edit the XML configuration of domain
2645 which has managedsave image.
2650 qemu: Add migrate-getmaxdowntime command
2653 Currently, the maximum tolerable downtime for a domain being migrated
2654 is write-only from libvirt, via migrate-setmaxdowntime. This
2655 implements a complementary migrate-getmaxdowntime command
2660 bhyve: Support autoport for VNC ports
2663 It's no longer necessary to explicitly specify VNC port for the bhyve
2664 guests. With the autoport feature it will be allocated automatically.
2665 Please refer to the bhyve driver documentation for examples.
2670 qemu: Added support for setting heads of virtio GPU
2675 qemu: Added support to configure reconnect timeout for chardev devices
2678 When you have a TCP or UNIX chardev device and it's connected somewhere
2679 you can configure reconnect timeout if the connection is closed.
2683 <section title="Improvements">
2686 qemu: Report a clear error when dropping a VM during startup
2689 "Failed to load config for domain 'DOMNAME'" is now reported if a VM
2690 config can't be parsed for some reason, and thus provides a clear
2691 indication for users (and devs).
2696 apparmor: Update for QEMU 2.10 compatibility
2699 Starting with QEMU 2.10, disk images and NVRAM files get
2700 automatically locked to prevent them from being corrupted; however,
2701 file locking needs to be explicitly allowed through
2702 <code>virt-aa-helper</code> or AppArmor will reject the requests and
2703 the guest will not be able to run.
2708 virsh: List Unix sockets in 'domdisplay' output
2711 VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets
2712 as connection endpoints, but such a configuration was not handled
2713 correctly by <code>virsh domdisplay</code>, causing the respective
2714 endpoints to be missing from the output.
2719 qemu: Don't check whether offline migration is safe
2722 Since offline migration only copies the guest definition to the
2723 destination host, data corruption is not a concern and the operation
2724 can always be performed safely.
2729 virt-host-validate: Fix IOMMU detection on ppc64
2733 <section title="Bug fixes">
2736 qemu: Better support for international domain names (with wide characters)
2739 There were some issues with multi-byte domains getting lost
2740 on daemon restart due to truncation, so the code now handles
2741 multi-byte names a bit better.
2746 qemu: Support long domain names with namespaces
2749 Domains with extremely long names would fail to start due to
2750 temporary namespace paths being created with the whole name.
2751 The path is now generated with shortened name instead.
2756 qemu: Tolerate missing emulator binary during libvirtd restart
2759 For some time libvirt required qemu capabilities being present when
2760 parsing VM configs during startup. As a side effect VM configs would
2761 fail to parse and thus vanish, if the emulator binary would be
2762 uninstalled or broken. Libvirt now tolerates when capabilities
2763 are missing during startup.
2768 qemu: Prevent pSeries guests from disappearing in some situations
2771 pSeries guest would disappear if any of the host devices they were
2772 configured to use was not available during libvirtd startup, which
2773 could easily happen for SR-IOV Virtual Functions. This scenario is
2774 now handled correctly.
2779 qemu: Honor <on_reboot/> setting
2782 The setting was accepted by the parser, but not actually implemented.
2787 Fix --verbose option for all daemons
2790 Since v3.0.0, the option had been ignored by all libvirt daemons
2791 (<code>libvirtd</code>, <code>virtlogd</code> and
2792 <code>virtlockd</code>); it's now working as intended once again.
2797 <release version="v3.6.0" date="2017-08-02">
2798 <section title="New features">
2801 hyperv: Implement virDomainSetMemory and virDomainSendKey APIs
2806 qemu: Support multiple PHBs for pSeries guests
2809 pSeries guests can now have multiple PHBs (PCI Host Bridges), which
2810 show up as separate PCI domains in the guest. To create additional
2811 PHBs, simply add PCI controllers with model <code>pci-root</code>
2812 to the guest configuration.
2817 qemu: Isolate hostdevs on pSeries guests
2820 To enable better error reporting and recovery, unrelated hostdevs
2821 will now be automatically isolated on pSeries guests by placing them
2822 on separate PHBs (PCI Host Bridges).
2826 <section title="Improvements">
2829 qemu: platform serial devices can now use chardev features
2832 QEMU VMs that depend on platform serial devices can now use
2833 QEMU's <code>-chardev</code> option, which enables access to
2834 advanced features like log file configuration. This applies
2835 to the default serial devices for arm, aarch64, and some ppc
2841 Require use of GCC 4.4 or Clang compilers
2844 We only ever test libvirt with GCC or Clang (which
2845 provides a GCC compatible compilation environment).
2846 Between them, these compilers cover every supported
2847 operating system platform, including Windows.
2852 qemu: shared disks with directsync cache should be safe for migration
2855 At present shared disks can be migrated with either readonly or
2856 <code>cache=none</code>. But <code>cache=directsync</code> should be
2857 safe for migration, because both <code>cache=directsync</code> and
2858 <code>cache=none</code> don't use the host page cache, and
2859 <code>cache=direct</code> write through qemu block layer cache.
2864 Handle hotplug change on VLAN configuration using OVS
2867 Libvirt was taught to handle VLAN change for running OVS interface.
2871 <section title="Bug fixes">
2874 qemu: Use vCPU 'node-id' property and pass it back to qemu
2877 vCPU properties gathered from query-hotpluggable-cpus need to be
2878 passed back to QEMU. As QEMU did not use the node-id property until
2879 now and libvirt forgot to pass it back properly (it was parsed but
2880 not passed around) we did not honor this.
2885 Miscellaneous stream fixes
2888 After introducing sparse stream features there were still some known
2889 bugs left. Those are fixed in this release.
2894 qemu: Miscellaneous domain NS fixes
2897 Libvirt starts qemu domains in separate Linux namespaces for a while
2898 now. However, there were still some bugs lingering. For instance
2899 libvirt did not know how to handle file based bind mounts.
2904 Various CPU driver improvements
2907 There were some minor bugs when using 'host-model' CPU.
2912 <release version="v3.5.0" date="2017-07-04">
2913 <section title="New features">
2916 qemu: Add support for loadparm for a boot device
2919 Add an optional boot parameter 'loadparm' for a boot device.
2920 Loadparm is an 8 byte parameter that, when present, is queried by
2921 S390 guests via sclp or diag 308. Linux guests on S390 use it to
2922 select a boot entry.
2927 Support controlling how video devices are exposed to the bhyve guests
2930 The <code>vgaconf</code> attribute was added to <code>video</code>'s
2931 <code>driver</code> element. Possible values are: <code>on</code>,
2932 <code>off</code>, and <code>io</code>. It controls the way how
2933 bhyve exposes video devices to its guests; various guest OSes might
2934 require different settings to boot properly.
2939 qemu: Add support for live updates of coalesce settings
2942 Users can now use <code>virsh update-device</code> to change
2943 the coalesce settings of an interfaces while the domain is
2949 qemu: Allow VirtIO devices to use vIOMMU
2952 It is now possible to turn on IOTBL for the vIOMMU and have VirtIO
2953 devices use it, provided they have been configured appropriately.
2957 <section title="Improvements">
2960 qemu: block copy job can be used with persistent domains
2963 Until now it was not possible to use block copy with persistent VMs.
2964 In use cases where it's not required to recover the job after VM
2965 shutdown, it's possible to specify VIR_DOMAIN_BLOCK_COPY_TRANSIENT_JOB
2966 flag to start the copy job.
2971 JSON pseudo-protocol backing store parser supports new format of qemu 2.9
2974 QEMU 2.9 modified a few structures corresponding to the JSON format
2975 of specifying a backing store for a disk image. Libvirt now implements
2981 Capabilities now include info about host's CAT settings
2984 Various information about resource control from the host is
2985 gathered and presented in capabilities if available.
2990 apparmor: Several improvements
2993 Allow access to Ceph config, EFI firmware on both x86_64 and
2994 aarch64, device tree on ppc64 and more.
2999 qemu: Support host-model on POWER9 machines
3003 <section title="Bug fixes">
3006 qemu: snapshot: retrieve image metadata from user provided files
3009 Disk images of an external snapshot created with
3010 VIR_DOMAIN_SNAPSHOT_CREATE_REUSE_EXT flag specified would not be
3011 scanned for metadata after recent changes.
3013 The metadata is necessary to allow keeping relative paths between
3014 images when doing a block-commit.
3019 Parse decimal numbers in a locale-independent way
3022 Some locales, such as <code>de_DE</code> and <code>pt_BR</code>,
3023 use comma rather than dot to separate the integer part from the
3024 fractional part of a decimal number; however, several data sources
3025 such as the kernel use a locale-independent representation and need
3026 to be treated accordingly.
3031 Support compilation with newer compiler and libc versions
3034 Several fixes have been included to make compilation with Clang
3035 4.0.0, GCC 7.1 and glibc >= 2.25.90 possible.
3040 qemu: Query name for vhost-user interfaces at runtime
3043 This makes it possible to use <code>virsh</code> subcommands such
3044 as <code>domiflist</code> and <code>domifstat</code> on vhost-user
3050 qemu: Set MTU for hotplugged interfaces correctly
3053 When hotplugging a network interface, the MTU was only set on the
3054 guest side. Set it on the host side as well.
3059 qemu: Forbid updating MTU for interfaces of running guests
3062 The MTU setting can't be modified while the guest is running, so any
3063 attempt to alter it at runtime will now result in an error rather
3064 than being silently ignored.
3069 qemu: Fix specifying QXL heads with older QEMU releases
3072 Specifying the number of QXL heads was not working correctly for
3073 QEMU releases older than 1.6.
3078 qemu: Fix migration to older libvirt/QEMU versions
3081 When the guest is started, libvirt updates the CPU definition to
3082 reflect the actual CPU features to enforce ABI. We need to send
3083 original and updated CPU definition in order to support migration
3084 to older libvirt/QEMU versions. Only the updated CPU definition
3085 was sent to destination.
3090 <release version="v3.4.0" date="2017-06-02">
3091 <section title="New features">
3094 Improved streams to efficiently transfer sparseness
3097 New extension to virStream was implemented so that
3098 virStorageVolDownload and virStorageVolUpload can preserve file
3104 I/O APIC type can be specified for QEMU/KVM domains
3107 The <code>ioapic</code> tag was added to domain
3108 <code>features</code>, so the type of the I/O APIC can now
3109 be specified (e.g. putting it in userspace for KVM domains).
3114 The reason for VM shutdown is reported, if known
3117 QEMU 2.10 will be able to report the reason for shutting
3118 down (whether that was caused by the guest or not), and
3119 libvirt is prepared for that and reports that information in
3120 its shutdown event as well, if it is known.
3124 <section title="Improvements">
3127 Repository now has new README.md file
3130 The new file uses markdown syntax, so it looks better on
3131 github and possibly other web pages, but it has also more
3132 useful information. The old README is now symlink to the
3138 qemu: Use GICv2 by default for aarch64/virt TCG guests
3141 The emulated GICv3 has some limitations that make it unusable as a
3142 default; use GICv2 until they're sorted out. This change makes it
3143 once again possible to run aarch64/virt guests on a x86_64 host
3144 without having to tweak their configuration.
3149 Additional capabilities for the node_device module
3152 Introduce two new capabilities to the node_device module. The first
3153 is for CCW devices, most common on the S390 architecture. The second
3154 is for fibre channel-backed SCSI devices and exposes the
3155 fc_remote_port sub-capability to SCSI target devices.
3160 Node devices now report Mediated device capabilities
3163 Endpoint devices support new <code>mdev</code> capability
3164 and their parents now report the supported types in new
3165 <code>mdev_types</code> capability.
3170 Capabilities now report information about host caches
3173 If supported in the kernel, host capabilities will now list
3174 L3 caches. The code for other levels was added as well, but
3175 only L3 caches are reported currently.
3180 POWER9 CPU model was added
3183 It is now properly reported in host capabilities.
3188 libxl: NUMA sibling distances are now reported in host capabilities
3193 VMDK version 3 files are now properly detected
3198 Interrupt remapping and Extended interrupt mode for IOMMU devices
3201 These two new features can now be controlled with new
3202 <code><driver intremap='on/off' eim='on/off'/></code>
3203 tag for <code>iommu</code> devices.
3208 Graphics in libxl domains now have default addresses
3211 Even though there were default addresses before this change,
3212 they were not saved in the XML. It is now possible to see
3213 and control the listen addresses properly.
3218 Default USB controllers are now added for devices in libxl domains
3221 Even though they were added automatically when USB device
3222 was attached, they could've been missing in some other
3223 cases. The logic is now fixed so there are always USB
3224 controllers, even if there was none of them in the specified
3230 Limits for RPC messages were increased
3233 Hitting the RPC limits we have is easier every day, so they
3234 were increased once again and some guessing logic was
3235 improved as well. It is now possible to get more stats than
3236 ever using the <code>virConnectGetAllDomainStats()</code>
3237 call and push through even bigger requests and replies for
3242 <section title="Bug fixes">
3245 qemu: Create memory_backing_dir on startup
3248 Libvirt's policy is that directories are created on startup if
3249 they don't exist. We've missed this one.
3254 PCIe 4.0 cards now report proper link speeds
3257 It could happen that the link speed for PCIe devices was not
3258 properly reported or the nodedev-dumpxml just failed. That
3259 was due to mistake in the field width, but should now work
3265 qemu: Do not report errors on shutdown
3268 For some users, in some rare cases, it could happen that
3269 there was an error message "internal error: End of file from
3270 qemu monitor" in the logs even though no problem happened.
3271 The detection of these false positives was improved and such
3272 errors should not show any more.
3277 User-specified UNIX socket paths for virtio channels should not be reset
3280 It could happen, in some cases, that libvirt would mistake a
3281 user-specified path for its own generated one and thus
3282 remove it from the XML. The detection of such addresses was
3288 Fix address reservation during RNG hot-plug
3291 When error occurred in a specific point in time during the
3292 hot-plug of an RNG device, it could happen that an address
3293 was released even though another device was already using
3294 it, making it possible to hot-plug another device with that
3295 address, effectively having duplicated addresses in the XML.
3300 <release version="v3.3.0" date="2017-05-05">
3301 <section title="New features">
3304 net: Add support for coalesce settings
3307 Enabling data batching through these settings can improve network
3308 performance for guests.
3313 qemu: Add support for guest CPU cache specification
3316 This features allows fine-grained control of the cache behavior
3322 qemu: Add support for the qemu-xhci USB controller
3326 <section title="Improvements">
3329 hyperv: Support Hyper-V 2012 and newer
3332 Starting with Hyper-V 2012 the API has changed causing the existing
3333 driver to be unable to send and process requests properly. This has
3334 been resolved by adding abstractions to handle the differences and
3335 ease handling such breaks if they happen in the future.
3340 libxl: Add support for nested HVM domains
3343 Xen has supported nested HVM domains since version 4.4. The
3344 libvirt libxl driver now supports nested HVM domains by
3345 specifying the host-passthrough CPU mode when defining a
3351 qemu: Implement ACPI support for aarch64 guests
3354 Up until this point, ACPI support was only advertised for x86_64
3355 guests and disabling it for aarch64 guests was not possible at all.
3360 vz: Add support for changing the number of vCPUs
3365 qemu: Automatically choose the best USB controller for guests
3368 The recently introduced qemu-xhci USB controller is the best choice
3369 for both ppc64 and aarch64 guests, so use it by default on those
3370 architectures if available.
3375 daemon: Increase default task limit for libvirtd
3378 The default number of tasks for the pids cgroup controller is 512,
3379 which libvirtd can quickly bump into when starting lots of guests.
3380 Raise the limit to a more reasonable 32768.
3385 docs: Include man pages describing key code names and values
3390 virsh: Report initialization errors
3393 Sometimes virsh might be unable to start: when that happens, report
3394 useful diagnostics instead of failing silently.
3398 <section title="Bug fixes">
3401 nss: Don't require a network restart for libvirt_guest
3404 Previously, the libvirt_guest NSS module would only work properly
3405 after the corresponding network had been restarted; now newly
3406 started guests will be reported correctly right away.
3411 storage: Remove unavailable transient pools after restart
3414 Solve an issue where transient storage pools would be stuck in an
3415 unmanageable state if the source disappeared and libvirtd was
3416 subsequently restarted.
3421 storage: Fix capacity value for LUKS encrypted volumes
3424 The 'capacity' value (e.g. guest logical size) for a LUKS volume is
3425 smaller than the 'physical' value of the file in the file system, so
3426 we need to account for that.
3431 qemu: Fix regression when hyperv/vendor_id feature is used
3434 Guests using the feature would not be started at all; it is now
3435 possible to start them as expected.
3440 qemu: Do not crash on USB address with no port and invalid bus
3445 crypto: Always pad data before encrypting it
3448 If this step is not performed, when the data length matches the
3449 chunk size the decryption routines will misinterpret the last byte
3450 of data as the padding length and fail to decode it correctly.
3455 <release version="v3.2.0" date="2017-04-02">
3456 <section title="New features">
3459 The virt-host-validate tool now supports bhyve hypervisor
3464 Introduce NVDIMM memory model
3467 NVDIMM is new type of memory introduced into QEMU 2.6. The idea is
3468 that we have a non-volatile memory module that keeps the data
3469 persistent across domain reboots and offers much faster data
3470 accesses. However, due to a bug in QEMU, this feature is not enabled
3471 for QEMUs older than 2.9.0.
3476 qemu: Introduce support for generic PCIe Root Ports
3479 For new controllers, a generic device (pcie-root-port) will be used
3480 by default instead of the Intel-specific device (ioh3420), provided
3481 the QEMU binary supports it.
3486 qemu: Add support for checking guest CPU ABI compatibility
3489 When migrating a domain to a different host, restoring a domain from
3490 a file or reverting a snapshot libvirt will make sure the guest CPU
3491 QEMU presents to the guest OS exactly matches the one provided on
3492 the source host (or before the domain's state was saved). This
3493 enhanced check may also be requested when starting a new domain to
3494 ensure the virtual CPU exactly matches the one specified in the XML.
3499 qemu: Add support to migrate using TLS
3502 Add the ability to migrate QEMU guests using TLS via a new flag
3503 VIR_MIGRATE_TLS or virsh migrate '--tls' option. Requires using at
3504 least QEMU 2.9.0 in order to work properly.
3509 qemu: add mediated devices framework support
3512 Recent kernel version introduced new mediated device framework, so
3513 provide an initial support of this framework for libvirt, mainly by
3514 introducing a new host device type in the XML.
3519 qemu: Add support for setting TSC frequency
3522 Setting TSC frequency is required to enable migration for domains
3523 with 'invtsc' CPU feature turned on.
3528 Add support for block device threshold event
3531 When using thin provisioning, management tools need to resize the
3532 disk in certain cases. To avoid having them to poll disk usage
3533 this version introduces an event which will be fired when a given
3534 offset of the storage is written by the hypervisor. Together with the
3535 API it allows registering thresholds for given storage backing
3536 volumes and this event will then notify management if the threshold
3537 is exceeded. Currently only the qemu driver supports this.
3542 bhyve: Add support for UEFI boot ROM, VNC, and USB tablet
3545 The bhyve driver now supports booting using the UEFI boot ROM,
3546 so non-FreeBSD guests that support UEFI could be booted without
3547 using an external boot loader like grub-bhyve. Video is also
3548 supported now, allowing to connect to guests via VNC and use
3549 an USB tablet as an input device. Please refer to
3550 the driver page for domain XML examples.
3554 <section title="Improvements">
3557 qemu: Detect host CPU model by asking QEMU on x86_64
3560 Previously, libvirt detected the host CPU model using CPUID
3561 instruction, which caused libvirt to detect a lot of CPU features
3562 that are not supported by QEMU/KVM. Asking QEMU makes sure we
3563 don't start it with unsupported features.
3568 perf: Add more perf statistics
3571 Add support to get the count of cpu clock time, task clock time,
3572 page faults, context switches, cpu migrations, minor page faults,
3573 major page faults, alignment faults, emulation faults by
3574 applications running on the platform.
3579 Write hyperv crash information into vm log
3582 qemu's implementation of the hyperv panic notifier now reports
3583 information about the crash from the guest os. Starting with this
3584 version, libvirt logs the information to the vm log file for possible
3589 <section title="Bug fixes">
3592 QEMU: Use adaptive timeout for connecting to monitor
3595 When starting qemu, libvirt waits for qemu to create the monitor
3596 socket which libvirt connects to. Historically, there was sharp 30
3597 second timeout after which the qemu process was killed. This
3598 approach is suboptimal as in some scenarios with huge amounts of
3599 guest RAM it can take a minute or more for kernel to allocate and
3600 zero out pages for qemu. The timeout is now flexible and computed by
3601 libvirt at domain startup.
3606 Overwrite (clear) 2 KB instead of just 512 bytes when initializing logical device
3611 Describe the logical backend requirements better for pool-create-as
3616 <release version="v3.1.0" date="2017-03-03">
3617 <section title="New features">
3620 storage: Add Virtuozzo storage backend storage pool
3623 Add new storage backend to support pool and volume management
3624 within the Virtuozzo Storage environment. Virtuozzo Storage is
3625 a highly available distributed software defined storage with
3626 built-in replication and disaster recovery.
3631 qemu: Add support for memory backing with file source
3634 Add support in numa topology for file source inside memory backing
3635 (hugepages are not needed) Three new elements
3636 <source/>,<access/> and <allocation/> were added
3637 to <memoryBacking/> element. Also new configuration parameter
3638 <code>memory_backing_dir</code> was added to qemu.conf.
3643 network: make openvswitch call timeout configurable
3646 Adding the ability to specify the timeout value in seconds for
3647 openvswitch calls in the libvirtd configuration file.
3652 bhyve: add e1000 NIC support
3655 Add support for e1000 NIC. Previously, the only available option
3656 was <code>virtio-net</code>.
3661 libxl: add tunneled migration support
3664 Add tunneled migration to libxl driver, which is always capable of
3665 strong encryption and doesn't require any extra network connection
3666 other than what's required for remote access of libvirtd.
3671 qemu: add rendernode argument
3674 Add a new attribute 'rendernode' to <gl> spice element.
3679 nodedev: add drm capability
3682 Add a new 'drm' capability for Direct Rendering Manager (DRM)
3683 devices, providing device type information.
3688 Add API for individual/specific vCPU hotplug
3691 The new API allows selecting specific vCPUs to be added/removed from
3692 the VM. The existing APIs allowed only adding/removing from the end
3693 which did not play well with NUMA.
3697 <section title="Improvements">
3700 virsh: pool-list: allow both --uuid and --name in one command
3703 Adjust the virsh-pool command to support the --uuid and/or --name
3704 options in order to print just the --name and/or --uuid of pools.
3709 Introduce MTU to domain <interface/> and <network>
3712 Allow setting MTU size for some types of domain interface
3718 libxl: improve support for <timer> configurations
3721 Add support for multiple timers. Extend the tsc timer to
3722 support the emulate mode. Improve conversion of timer XML
3728 storage: modularize the storage driver
3731 Split up the storage driver backends into loadable modules so that
3732 binary distributions don't have to compromise on shipping the storage
3733 driver with all backends which may pull in too many dependencies.
3737 <section title="Bug fixes">
3740 nodedev: Fabric name must not be required for fc_host capability
3743 fabric_name is one of many fc_host attributes in Linux that is
3744 optional and left to the low-level driver to decide if it is
3745 implemented. For example the zfcp device driver does not provide a
3746 fabric name for an fcp host. The requirement for the existence of
3747 a fabric name has been removed by making it optional.
3752 bhyve: change address allocation schema for SATA disks
3755 Previously, the bhyve driver assigned PCI addresses to SATA disks
3756 directly rather than assigning that to a controller and
3757 using SATA addresses for disks. It was implemented this way
3758 because bhyve has no notion of an explicit SATA controller.
3759 However, as this doesn't match libvirt's understanding of
3760 disk addresses, the bhyve driver was changed to follow
3761 the common schema and have PCI addresses for SATA controllers
3762 and SATA addresses for disks. If you're having issues
3763 because of this, it's recommended to edit the domain's XML
3764 and remove <address type='pci'> from the <disk>
3765 elements with <target bus='sata'/> and let libvirt
3766 regenerate it properly.
3771 libxl: maximum memory fixes
3774 Fix reporting of domain maximum memory. Fix setting dom0
3780 libxl: fix disk detach when <driver> not specified
3785 libxl: fix dom0 autoballooning with Xen 4.8
3790 qemu: Allow empty script path to <interface/>
3793 Historically, this was always allowed. Unfortunately, due to some
3794 rework done for 1.3.2 release a bug was dragged in which suddenly
3795 stop allowing domain with such configuration to start.
3800 <release version="v3.0.0" date="2017-01-17">
3801 <section title="New features">
3804 Domain events for metadata content changes
3807 The domain events framework has a new event ID that can
3808 be used to get notifications when domain metadata content
3814 Event notifications for the secret object
3817 The secret object now supports event notifications, covering
3818 lifcycle changes and secret value changes.
3823 New localPtr attribute for "ip" element in network XML
3828 qemu: Support QEMU group I/O throttling
3831 Add the capability to allow group I/O throttling via a new
3832 domain <disk> <iotune> subelement "group_name"
3833 to allow sharing I/O throttling quota between multiple drives.
3838 nss: Introduce libvirt_guest
3841 New <code>libvirt_guest</code> nss module that translates libvirt
3842 guest names into IP addresses.
3847 daemon: Add support for runtime logging settings adjustment
3850 Logging-related settings like log outputs and filters can now be
3851 adjusted during runtime using the admin interface without the
3852 necessity of the daemon's restart.
3857 storage: Add virStorageVolInfoFlags API
3860 Add the API to support using the VIR_STORAGE_VOL_GET_PHYSICAL
3861 flag in order to return the host physical size in bytes
3862 of the image container in the allocation field of the
3863 _virStorageVolInfo structure. The --physical flag has been
3864 added to the virsh vol-info command to access the data.
3869 libxl: Implement virDomainGetMaxVcpus API
3874 storage: Add overwrite flag checking for logical pool
3877 Add support for the OVERWRITE flags for the logical storage
3878 backend including checking for existing data on the target
3879 volumes when building a new logical pool on target volume(s).
3884 qemu: Add support for guest CPU configuration on s390(x)
3888 <section title="Improvements">
3891 perf: Add more perf statistics
3894 Add support to get the count of branch instructions
3895 executed, branch misses, bus cycles, stalled frontend
3896 cpu cycles, stalled backend cpu cycles, and ref cpu
3897 cycles by applications running on the platform.
3902 conf: Display <physical> for volume xml
3905 Add a display of the <physical> size of a disk
3906 volume in the output of the volume XML.
3911 qemu: Use virtio-pci by default for aarch64 mach-virt guests
3914 virtio-pci provides several advantages over virtio-mmio, such
3915 as the ability to hotplug devices and improved performance.
3916 While opting in to virtio-pci has been possible for a while,
3917 newly-defined guests will now use it automatically.
3922 vbox: remove support for VirtualBox 3.x and older
3925 Those old VirtualBox versions have been unsupported by
3926 upstream for a long time and the API of 4.0 and newer has
3927 diverged enough to require code abstractions to handle differences.
3928 Removing support for those old versions drops lots of code from
3929 the driver and simplifies the logic to ease implementation of new
3930 features going forward.
3935 virsh: pool-info: introduce option --bytes
3938 Add option --bytes to virsh pool-info in order ti allow display
3939 of units in bytes rather than default of human readable output.
3944 scsi: Add parent wwnn/wwpn or fabric capability for createVport
3947 Improve the algorithm searching for the parent scsi_host device
3948 for vHBA/NPIV scsi_host creation. Rather than supplying the
3949 "parent" by name, it's now possible to define the parent by
3950 it's wwnn/wwpn or fabric_wwn in the node device create XML or
3951 the storage pool XML.
3956 qemu: aggregate pcie-root-ports onto multiple functions of a slot
3959 When pcie-root-ports are added to pcie-root in order to
3960 provide a place to connect PCI Express endpoint devices,
3961 libvirt now aggregates multiple root ports together onto the
3962 same slot (up to 8 per slot) in order to conserve slots.
3963 Using this method, it's possible to connect more than 200
3964 endpoint devices to a guest that uses PCIe without requiring
3965 setup of any PCIe switches.
3969 <section title="Bug fixes">
3972 lxc: fix accidental killing of containers during libvirtd restart
3975 The libvirt_lxc process was previously not moved into the
3976 container scope. As a result, if systemd reloads its config
3977 after a container is started, when libvirtd is later restarted
3978 it will accidentally kill the containers.
3983 qemu: Correct GetBlockInfo values
3986 For an active domain, correct the physical value provided for
3987 a raw sparse file backed storage and the allocation value provided
3988 for a qcow2 file backed storage that hasn't yet been opened on
3994 qemu: Make virtio console usable on ppc64 guests
3997 The chardev detection code has been improved and can now handle this
3998 configuration properly.
4003 qemu: Enable mount namespace
4006 To avoid funny races with udev relabelling devices under our hands and
4007 to enhance security, libvirt now spawns each qemu process with its own
4013 storage: Fix implementation of no-overwrite for file system backend
4016 Fix file system storage backend implementation of the OVERWRITE
4017 flags to be consistent between code and documentation. Add checks
4018 to ensure that when building a new file system on a target volume
4019 that there is not something already on the disk in a format that
4020 libvirt can recognize.
4025 qemu: Create hugepage path on per domain basis
4028 Historically, all hugepage enabled domains shared the same path under
4029 hugetlbfs. This left libvirt unable to correctly set security labels
4030 on it. With this release, however, each domain is put into a
4031 separate path which is also correctly labeled.
4036 conf: Reject domains with duplicate drive addresses
4039 Reject duplicate drive addresses for disks and hostdevs at
4045 libxl: reverse defaults on HVM net device attach
4048 Fixes network interface attach for HVM domains when no model is
4049 specified. Emulated hotplug isn't yet supported and hence we should
4050 default to the general working scenario.
4055 libxl: always enable pae for x86_64 HVM
4058 By default pae is disabled in libxl. Without an explicit <pae/>
4059 setting in the domain <features> configuration, an x86_64 HVM
4060 domain would be get an i686 environment. pae should always be enabled
4061 for x86_64 HVM domains.
4066 qemu: Fix XML dump of autogenerated websocket
4069 As a result autogenerated websocket port is regenerated on domain restore,
4070 migration and such as it should be.
4075 <release version="v2.5.0" date="2016-12-04">
4076 <section title="New features">
4079 shmem: Add support for additional models
4082 The shmem device can now utilize QEMU's ivshmem-plain and
4083 ivshmem-doorbell, more modern versions of ivshmem.
4088 vbox: Add VirtualBox 5.1 support
4093 libssh: New transport
4096 The new libssh transport allows one to connect to a running
4097 libvirtd via SSH, using the libssh library; for example:
4098 <code>qemu+libssh://server/system</code>.
4103 vhost-scsi: Add support scsi_host hostdev passthrough
4106 Add the capability to pass through a scsi_host HBA and the
4107 associated LUNs to the guest.
4112 Allow debugging of gluster volumes in qemu
4115 Users can now enable debug logging for native gluster
4116 volumes in qemu using the "gluster_debug_level" option in qemu.conf
4121 Pre-allocate memory slots for memory hotplug
4124 Slot numbers for memory devices are now automatically allocated and
4125 thus persistent. In addition slot numbers can be specified without
4126 providing a base address, which simplifies user configuration
4131 qemu: Express devices will be placed on PCIe bus by default
4134 For machine types that use a PCI Express root bus
4135 (e.g. x86_64/Q35 and aarch64/virt), any unaddressed PCI
4136 device that is an Express device (all virtio-1.0 devices,
4137 e1000e, nec-xhci, vfio assigned devices) will be placed on
4138 an Express controller (i.e. a pcie-root-port) instead of a
4139 legacy PCI controller (i.e. pci-bridge) with the root ports
4144 <section title="Improvements">
4147 docs: Better documentation for migration APIs and flags
4152 vbox: Address thread safety issues
4157 virsh: Add support for passing an alternative persistent XML to migrate command
4162 vhostuser: Allow hotplug of multiqueue devices
4167 NEWS: Switch to an improved format
4170 List user-visible changes instead of single commits for a better
4171 high-level overview of differences between libvirt releases.
4176 website: Modernize layout and branding
4179 The libvirt website looked very cluttered and outdated; it has now
4180 been completely overhauled, resulting in a design that's better
4181 organized and more pleasant to look at.
4185 <section title="Bug fixes">
4188 vz: Fix migration in P2P mode
4193 Forbid newline character in names of some libvirt objects
4198 Fix compilation on macOS