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="v5.5.0" date="unreleased">
43 <section title="New features">
46 qemu: Support SMMUv3 IOMMU
49 SMMUv3 is an IOMMU implementation for ARM virt guests.
53 <section title="Improvements">
55 <section title="Bug fixes">
58 <release version="v5.4.0" date="2019-06-03">
59 <section title="Security">
62 cpu: Introduce support for the md-clear CPUID bit
65 This bit is set when microcode provides the mechanism to invoke a
66 flush of various exploitable CPU buffers by invoking the x86
67 <code>VERW</code> instruction. CVE-2018-12126, CVE-2018-12127,
68 CVE-2018-12130, CVE-2019-11091.
73 Restrict user access to virt-admin, virtlogd and virtlockd
76 The intended users for these facilities are the <code>root</code>
77 user and the <code>libvirtd</code> service respectively, but these
78 restrictions were not enforced correctly. CVE-2019-10132.
82 <section title="Improvements">
85 test driver: Expand API coverage
88 Several APIs that were missing from the test driver have now been
94 Avoid unnecessary static linking
97 Most binaries shipped as part of libvirt, for example
98 <code>virtlogd</code> and <code>libvirt_iohelper</code>, were
99 embedding parts of the library even though they also linked against
100 the <code>libvirt.so</code> dynamic library. This is no longer the
101 case, which results in both the disk and memory footprint being
107 qemu: Report stat-htlb-pgalloc and stat-htlb-pgfail balloon stats
110 These stats have been introduced in QEMU 3.0.
114 <section title="Bug fixes">
117 qemu: Fix emulator scheduler support
120 Setting the scheduler for QEMU's main thread before QEMU had a chance
121 to start up other threads was misleading as it would affect other
122 threads (vCPU and I/O) as well. In some particular situations this
123 could also lead to an error when the thread for vCPU #0 was being
124 moved to its cpu,cpuacct cgroup. This was fixed so that the
125 scheduler for the main thread is set after QEMU starts.
130 apparmor: Allow hotplug of vhost-scsi devices
135 <release version="v5.3.0" date="2019-05-04">
136 <section title="New features">
139 qemu: Add support for setting the emulator scheduler parameters
142 I/O threads and vCPU threads already support setting schedulers, but
143 until now it was impossible to do so for the main QEMU thread
144 (emulator thread in the libvirt naming). This is, however, requested
145 for some very specific scenarios, for example when vCPU threads are
146 running at such priority that could starve the main thread.
150 <section title="Removed features">
153 vbox: Drop support for VirtualBox 4.x releases
156 Support for all the 4.x releases was ended by
157 VirtualBox maintainers in December 2015. Therefore,
158 libvirt support for these releases is dropped.
162 <section title="Improvements">
165 qemu: Use PCI by default for RISC-V guests
168 PCI support for RISC-V guests was already available in libvirt
169 5.1.0, but it required the user to opt-in by manually assigning
170 PCI addresses: with this release, RISC-V guests will use PCI
171 automatically when running against a recent enough (4.0.0+) QEMU
177 qemu: Advertise firmware autoselection in domain capabilities
180 The firmware autoselection feature is now exposed in
181 domain capabilities and management applications can
182 query for accepted values, i.e. values that are accepted
183 and for which libvirt found firmware descriptor files.
184 Firmware Secure Boot support is also advertised.
192 YAJL 2 is widely adopted and maintaining side by side
193 support for two versions is unnecessary.
197 <section title="Bug fixes">
200 rpc: cleanup in virNetTLSContextNew
203 Failed new gnutls context allocations in virNetTLSContextNew
204 function results in double free and segfault. Occasional memory
205 leaks may also occur.
210 virsh: various completers fixes
213 There were some possible crashers, memory leaks, etc.
219 qemu: Make hugepages work with memfd backend
222 Due to a bug in command line generation libvirt did not
223 honor hugepages setting with memfd backend.
228 Enforce ACL write permission for getting guest time & hostname
231 Getting the guest time and hostname both require use of
232 guest agent commands. These must not be allowed for
233 read-only users, so the permissions check must validate
234 "write" permission not "read".
239 <release version="v5.2.0" date="2019-04-03">
240 <section title="New features">
243 Add Storage Pool Capabilities output
246 Add support to list an enumerated list of supported Storage
247 Pools via the virConnectGetCapabilities API when connected
248 via a Storage Driver. Add support to get a more detailed
249 list XML output Storage Pool Capabilities vis the
250 virConnectGetStoragePoolCapabilites API.
255 qemu: Support virtio-{non-}transitional device models
258 <code>virtio-transitional</code> and
259 <code>virtio-non-transitional</code> <code>model</code> values
260 were added to the QEMU driver for the following devices:
261 <code>disk</code>, <code>interface</code>, <code>filesystem</code>,
262 <code>rng</code>, <code>vsock</code>, <code>memballoon</code>,
263 <code>controller</code> type <code>scsi</code>,
264 <code>controller</code> type <code>virtio-serial</code>,
265 <code>input</code> bus <code>virtio</code>
266 type <code>passthrough</code>,
267 <code>hostdev</code> type <code>scsi_host</code>. These new
268 models can be used to give fine grained control over what
269 virtio device version is presented to the guest.
274 qemu: Enable firmware autoselection
277 Libvirt allows users to provide loader path for some time now.
278 However, this puts some burden on users because they need to
279 know what firmware meets their requirements. Now that QEMU
280 ships firmware description files this burden can be moved onto
281 libvirt. It is as easy as setting the <code>firmware</code>
282 attribute in the <code>os</code> element (accepted values are
283 <code>bios</code> and <code>efi</code>). Moreover, libvirt
284 automatically enables domain features needed for firmware it
290 snapshots: Add support for topological listings
293 A new flag VIR_DOMAIN_SNAPSHOT_LIST_TOPOLOGICAL is available
294 for the various snapshot listing APIs such as
295 virDomainListAllSnapshots(). For drivers that support the
296 flag, the listed snapshots are guaranteed to be sorted such
297 that parents occur before children.
302 Xen: Add support for max grant frames setting
305 Add support for Xen's max_grant_frames setting by adding a
306 new xenbus controller type with a maxGrantFrames attribute.
307 E.g. <code><controller type='xenbus' maxGrantFrames='64'/></code>
312 qemu: Add support for parallel migration
315 With QEMU 4.0.0 libvirt can enable parallel migration which causes
316 the memory pages to be processed in parallel by several threads and
317 sent to the destination host using several connections at the same
318 time. This may increase migration speed in case a single thread is
319 unable to saturate the network link.
323 <section title="Removed features">
326 Drop support for Upstart and "Red Hat" init scripts
329 Not a single one of the platforms we target still uses Upstart,
330 and the Upstart project itself has been abandoned for several years
331 now; the same is true for the "Red Hat" (really System V) init
332 scripts, since RHEL 7 and later releases use systemd.
336 <section title="Improvements">
339 Report class information for PCI node device capability.
344 Split setup of IPv4 and IPv6 top level chain
347 The requirement resulting from private chains improvement done
348 in <code>v5.1.0</code> was refined so that only tables from
349 corresponding IP version are required. This means that if a
350 network doesn't have <code>IPv6</code> enabled then those
351 tables are not required.
356 Don't default to building the QEMU driver
359 Historically, the QEMU driver has been special in that it was
360 enabled by default, with the option to explicitly opt-out of it;
361 starting now, we're enabling it opportunistically if we detect that
362 all requirements are available, just like we do with other drivers.
366 <section title="Bug fixes">
369 virt-host-validate: Fix IOMMU check on s390x
374 qemu: Allow creating pSeries guests with graphics and no USB mouse
377 It's now possible to prevent libvirt from automatically adding a
378 USB mouse to pSeries guests by including a USB tablet in the input
379 XML: doing so is desiderable as using a tablet results in a much
380 better user experience when working with GUIs.
385 qemu: Set $HOME and XGD variables for qemu:///system guests
388 This avoids files being accidentally created under <code>/</code> or
389 the guests not being able to start because they lack the necessary
390 permissions to write to that location.
395 <release version="v5.1.0" date="2019-03-04">
396 <section title="New features">
399 bhyve: Add support for additional command-line arguments
402 The bhyve driver now supports passing additional command-line
403 arguments to the bhyve process using the new
404 <code><bhyve:commandline></code> element in domain
410 network: Support setting a firewalld "zone" for virtual network bridges
413 All libvirt virtual networks with bridges managed by libvirt
414 (i.e. those with forward mode of "nat", "route", "open", or
415 no forward mode) will now be placed in a special firewalld
416 zone called "libvirt" by default. The zone of any network
417 bridge can be changed using the <code>zone</code> attribute
418 of the network's <code>bridge</code> element.
423 bhyve: Support for ignoring unknown MSRs reads and writes
426 A new <features> element <msrs unknown='ignore'/> was
427 introduced and the bhyve driver supports it to control unknown
428 Model Specific Registers (MSRs) reads and writes.
433 qemu: Add support for encrypted VNC TLS keys
436 Use the password stored in the secret driver under the uuid
437 specified by the <code>vnc_tls_x509_secret_uuid</code> option
443 Add storage pool namespace options
446 Allow for adjustment of RBD configuration options via Storage
447 Pool XML Namespace adjustments.
452 qemu: Add support for setting post-copy migration bandwidth
455 Users can now limit the bandwidth of post-copy migration, e.g.
456 via <code>virsh migrate --postcopy-bandwidth</code>.
460 <section title="Improvements">
463 Create private chains for virtual network firewall rules
466 Historically firewall rules for virtual networks were added
467 straight into the base chains. This works but has a number of
468 bugs and design limitations. To address them, libvirt now puts
469 firewall rules into its own chains. Note that with this change the
470 <code>filter</code>, <code>nat</code> and <code>mangle</code> tables
471 are required for both <code>IPv4</code> and <code>IPv6</code>.
476 Detect CEPH and GPFS as shared FS
479 When starting a migration libvirt performs some sanity checks
480 to make sure domain will be able to run on the destination.
481 One of the requirements is that the disk has to either be
482 migrated too or be accessible from a network filesystem. CEPH
483 and GPFS weren't detected as a network filesystem.
488 Advertise network MTU via DHCP when specified
491 If network MTU is set and the network has DHCP enabled,
492 advertise the MTU in DHCP transaction too so that clients can
493 adjust their link accordingly.
498 qemu: Allocate memory at the configured NUMA nodes from start
501 Libvirt used to just start QEMU, let it allocate memory for
502 the guest, and then use CGroups to move the memory to
503 configured NUMA nodes. This is suboptimal as huge chunks of
504 memory have to be moved. Moreover, this relies on ability to
505 move memory later which is not always true. A change was made
506 to set process affinity correctly from the start so that memory
507 is allocated on the configured nodes from the beginning.
512 Support for newer Wireshark
515 Adapt libvirt to use the more recent release requiring a
516 source build configuration of libvirt
517 <code>--with-wireshark</code> to upgrade to the more recent
523 Batch mode virsh and virt-admin parsing improvements
526 When parsing a single-argument command_string in batch mode,
527 virsh and virt-admin now permit newlines in addition to
528 semicolons for splitting commands, and backslash-newline for
529 splitting long lines, to be more like shell parsing.
533 <section title="Bug fixes">
536 qemu: Use CAP_DAC_OVERRIDE during QEMU capabilities probing
539 By default, libvirt runs the QEMU process as <code>qemu:qemu</code>
540 which could cause issues during probing as some features like AMD SEV
541 might be inaccessible to QEMU because of file system permissions.
542 Therefore, <code>CAP_DAC_OVERRIDE</code> is granted to overcome these
543 for the purposes of probing.
548 storage: Add default mount options for fs/netfs storage pools
551 Altered the command line generation for fs/netfs storage pools to
552 add some default options. For Linux based systems, the options
553 added are "nodev, nosuid, noexec". For FreeBSD based systems,
554 the options added are "nosuid, noexec".
559 qemu: Allow use of PCI for RISC-V guests
562 This works with QEMU 4.0.0+ only and is opt-in at the moment, since
563 it requires users to manually assign PCI addresses, but is otherwise
569 network: Fix virtual networks on systems using firewalld+nftables
572 Because of the transitional state of firewalld's new support
573 for nftables, not all iptables features required by libvirt
574 are yet available, so libvirt must continue to use iptables
575 for its own packet filtering rules even when the firewalld
576 backend is set to use nftables. However, due to the way
577 iptables support is implemented in kernels using nftables
578 (iptables rules are converted to nftables rules and
579 processed in a separate hook from the native nftables
580 rules), guest networking was broken on hosts with firewalld
581 configured to use nftables as the backend. This has been
582 fixed by putting libvirt-managed bridges in their own
583 firewalld zone, so that guest traffic can be forwarded
584 beyond the host and host services can be exposed to guests
585 on the virtual network without opening up those same
586 services to the rest of the physical network. This means
587 that host access from virtual machines is no longer
588 controlled by the firewalld default zone (usually "public"),
589 but rather by the new firewalld zone called "libvirt"
590 (unless configured otherwise using the new zone
591 attribute of the network bridge element).
596 qemu: Fix i6300esb watchdog hotplug on Q35
599 Ensure that libvirt allocates a PCI address for the device so
600 that QEMU did not default to an address that would not allow
606 lxc: Don't reboot host on virDomainReboot
609 If the container is really a simple one (init is just bash and
610 the whole root is passed through) then virDomainReboot and
611 virDomainShutdown would reboot or shutdown the host. The
612 solution is to use different method to reboot or shutdown the
613 container in that case (e.g. signal).
618 rpc: Various stream fixes
621 One particular race was fixed, one locking problem and error
622 reporting from streams was made better.
627 qemu: Fix guestfwd hotplug/hotunplug
630 Fixed the generation of the guestfwd hotplug/unplug command
631 sent to QEMU to match the syntax used when creating the
632 initial command line.
637 qemu: Forbid CDROMs on virtio bus
640 Attempting to create an empty virtio-blk drive or attempting
641 to eject it results into an error. Forbid configurations
642 where users would attempt to use CDROMs in virtio bus.
647 qemu: Use 'raw' for 'volume' disks without format
650 Storage pools might want to specify format of the image when
651 translating the volume thus libvirt can't add any default
652 format when parsing the XML. Add an explicit format when
653 starting the VM and format is not present neither by user
654 specifying it nor by the storage pool translation function.
659 qemu: Assume 'raw' default storage format also for network storage
662 Post parse callback adds the 'raw' type only for local files.
663 Remote files can also have backing store (even local) so we
664 should do this also for network backed storage.
669 qemu: Fix block job progress reporting and advocate for READY event
672 In some cases QEMU can get to 100% and still not reach the
673 synchronised phase. Initiating a pivot in that case will fail.
674 Therefore it is strongly advised to wait for
675 <code>VIR_DOMAIN_BLOCK_JOB_READY</code> event which does not
676 suffer from this problem.
681 qemu: Don't format image properties for empty drive
684 If a <code>-drive</code> has no image, then formatting
685 attributes such as cache, readonly, etc. would cause errors to
686 be reported from QEMU. This was fixed by not supplying the
687 attributes for devices without an image.
692 External snapshot metadata redefinition is fixed
695 Attempting to use VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE to
696 reinstate the metadata describing an external snapshot
697 created earlier for an offline domain no longer fails.
702 <release version="v5.0.0" date="2019-01-15">
703 <section title="New features">
706 Xen: Add support for openvswitch
709 The libxl driver now supports virtual interfaces that connect to
710 an openvswitch bridge, including interfaces with VLAN tagging and
711 trunking configuration.
716 qemu: Report whether KVM nesting is available
719 Running nested KVM guests requires specific configuration steps to
720 be performed on the host; libvirt will now report in the host
721 capabilities whether KVM nesting support is available.
725 <section title="Removed features">
731 The UML driver was unmaintained and not tested for
732 quite some time now. Worse, there is a bug that causes
733 it to deadlock on some very basic operations (e.g.
734 dumping domain XML). These facts make us believe no one
739 <section title="Improvements">
742 qemu: Add support for ARMv6l guests
747 Support more NVDIMM configuration options
750 Introduce more configuration options. For the source element, add
751 the 'alignsize' and 'pmem' subelements. For the target element, add
752 the 'readonly' subelement.
757 cpu: Add support for "stibp" x86_64 feature
760 Add cpu flag stibp (Single Thread Indirect Branch Predictors) to
761 prevent indirect branch predictions from being controlled by the
767 libxl: Handle external domain destroy
770 Historically, if a domain was destroyed using <code>xl</code>
771 rather than through libvirt APIs, libvirt would not be aware of
772 the fact and keep considering it as running. This is no longer the
778 Start selecting the first available DRI device for OpenGL operations
781 If OpenGL support is needed (either with SPICE gl enabled or with
782 egl-headless), libvirt is now able to pick the first available DRI
783 device for the job. At the same time, this improvement is also a
784 bugfix as it prevents permission-related issues with regards to our
785 mount namespaces and the default DRI render node's permissions which
786 would normally prevent QEMU from accessing such a device.
791 qemu: Add support for postcopy-requests migration statistics
794 The <code>virDomainJobInfo</code> can get number page requests
795 received from the destination host during post-copy migration.
799 <section title="Bug fixes">
802 lxc: Don't forbid interfaces with type=direct
805 Such interfaces are supported by lxc and should be allowed.
810 qemu: Fully clean up RNG devices on detach
813 Some RNG device types, such as those using EGD, might need extra
814 clean up on the host in addition to removing the guest-side device.
819 <release version="v4.10.0" date="2018-12-03">
820 <section title="New features">
823 qemu: Add Hyper-V PV IPI and Enlightened VMCS support
826 The QEMU driver now has support for Hyper-V PV IPI and Enlightened VMCS
827 for Windows and Hyper-V guests.
832 qemu: Added support for PCI devices on S390
835 PCI addresses can now include the new zpci element which contains
836 uid (user-defined identifier) and fid (PCI function identifier)
837 attributes and makes the corresponding devices usable by S390
843 Support changing IOThread polling parameters for a live guest
846 Introduced virDomainSetIOThreadParams which allows dynamically
847 setting the IOThread polling parameters used by QEMU to manage
848 the thread polling interval and the algorithm for growth or
849 shrink of the polling time. The values only affect a running
850 guest with IOThreads. The guest's IOThread polling values can
851 be viewed via the domain statistics.
856 Xen: Add support for PVH
859 The libxl driver now supports Xen's PVH virtual machine type.
860 PVH machines are enabled with the new "xenpvh" OS type, e.g.
861 <code><os><type>xenpvh</type></os></code>
866 qemu: Added support for CMT (Cache Monitoring Technology)
869 Introduced cache monitoring using the <code>monitor</code>
870 element in <code>cachetune</code> for vCPU threads. Added
871 interfaces to get and display the cache utilization statistics
872 through the command 'virsh domstats' via the
873 virConnectGetAllDomainStats API.
878 qemu: Add support for nested HV for pSeries guests
881 Nested HV support makes it possible to run nested (L2) guests
882 with minimal performance penalty when compared to regular (L1)
883 guests on ppc64 hardware.
887 <section title="Bug fixes">
890 Xen: Handle soft reset shutdown event
893 The pvops Linux kernel uses soft reset to handle the crash
894 machine operation. The libxl driver now supports the soft
895 reset shutdown event, allowing proper crash handling of
896 pvops-based HVM domains.
901 <release version="v4.9.0" date="2018-11-04">
902 <section title="New features">
905 util: Add cgroup v2 support
908 cgroup v2 support has been implemented in libvirt, with both
909 "unified" (v2 only) and "hybrid" (v2 + v1) setups being usable;
910 existing "legacy" (v1 only) setups will keep working.
915 qemu: Add vfio AP support
918 The QEMU driver now has support to passthrough adjunct processors
919 into QEMU guests on S390.
923 <section title="Improvements">
926 rpc: Make 'genprotocol' output reproducible
929 This is another step towards making libvirt builds fully
934 <section title="Bug fixes">
937 security: Fix permissions for UNIX sockets
940 Since 4.5.0, libvirt is using FD passing to hand sockets over to
941 QEMU, which in theory removes the need for them to be accessible by
942 the user under which the QEMU process is running; however, other
943 processes such as vdsm need to access the sockets as well, which
944 means adjusting permissions is still necessary.
949 cpu_map: Add Icelake model definitions
952 These CPU models will be available in the upcoming 3.1.0 QEMU
958 util: Properly parse URIs with missing trailing slash
961 Some storage URIs were not parsed correctly, in which case libvirt
962 ended up emitting XML that it would then refuse to parse back.
967 <release version="v4.8.0" date="2018-10-01">
968 <section title="New features">
971 Xen: Support PM Suspend and Wakeup
974 The libxl driver now supports the virDomainPMSuspendForDuration
975 and virDomainPMWakeup APIs.
979 <section title="Removed features">
982 Xen: Drop support for Xen 4.4 and 4.5
985 Xen 4.4 and 4.5 are no longer supported by the Xen community.
986 Drop support for these older versions and require Xen >= 4.6.
991 nwfilter: Disallow binding creation in session mode
994 Ensure that a filter binding creation is not attempted in session
995 mode and generates a proper error message.
999 <section title="Improvements">
1002 qemu: Retrieve guest hostname through QEMU Guest Agent command
1005 QEMU is now able to retrieve the guest hostname using a new QEMU-GA
1006 command called 'guest-get-host-name'. Virsh users can execute
1007 'domhostname' for QEMU driver for domains configured to use the
1013 virsh: Implement vsh-table in virsh and virsh-admin
1016 The new API fixes problems with table-alignment, making the tables
1017 more readable and deals with unicode.
1021 <section title="Bug fixes">
1024 storage: Allow inputvol to be encrypted
1027 When creating a storage volume based on another volume, the base
1028 input volume is allowed to be encrypted.
1033 virsh: Require explicit --domain for domxml-to-native
1036 The --domain option for domxml-to-native virsh command has always
1037 been documented as required, but commit v4.3.0-127-gd86531daf2
1038 accidentally made it optional.
1043 lxc_monitor: Avoid AB / BA lock race
1046 A deadlock situation could occur when autostarting a LXC domain
1047 'guest' due to two threads attempting to take opposing locks while
1048 holding opposing locks (AB BA problem).
1053 <release version="v4.7.0" date="2018-09-03">
1054 <section title="New features">
1057 storage: add storage pool iscsi-direct
1060 Introduce a new storage pool backend that uses libiscsi instead of
1061 iscsiadm. It support basic pool operations: checkPool and refreshPool.
1066 Add support for MBA (Memory Bandwidth Allocation technology)
1069 Domain vCPU threads can now have allocated some parts of host memory
1070 bandwidth by using the <code>memorytune</code> element in <code>cputune</code>.
1075 qemu: Add support for RISC-V guests
1078 riscv32 and riscv64 guest architectures are now supported.
1082 <section title="Improvements">
1085 qemu: Add ccw support for vhost-vsock
1088 Support the vhost-vsock-ccw device on S390.
1093 qemu: Make default machine type independent of QEMU
1096 We can't control whether or not QEMU will change its default
1097 machine type in the future, or whether downstream distributions
1098 will decide to compile out some machine types, so our only option
1099 to provide a predictable behavior is taking care of the default
1100 ourselves; management applications and users are encouraged to
1101 explicitly pick a machine type when creating new guests.
1106 apparmor: Various improvements
1109 Rules have been added to deal with a number of scenarios that
1110 didn't work correctly.
1114 <section title="Bug fixes">
1117 esx: Truncate CPU model name
1120 Some CPU model names are too long to be stored into the
1121 corresponding property, and should be explicitly truncated
1122 to avoid unexpected behavior in users of the
1123 <code>virNodeGetInfo()</code> API such as
1124 <code>virsh nodeinfo</code>.
1129 utils: Remove arbitrary limit on socket_id/core_id
1132 Both values were assumed to be smaller than 4096, but in fact
1133 they are entirely hardware-dependent and there have been reports
1134 of machines presenting much bigger values, preventing libvirt from
1135 working correctly; all such limits have now been removed.
1140 <release version="v4.6.0" date="2018-08-06">
1141 <section title="New features">
1144 qemu: Implement the HTM pSeries feature
1147 Users can now decide whether HTM (Hardware Transactional Memory)
1148 support should be available to the guest.
1153 qemu: Enable VNC console for mediated devices
1156 Host devices now support a new atribute 'display' which can be used
1157 to turn on frame buffer rendering on a vgpu mediated device instead of
1158 on an emulated GPU, like QXL.
1162 <section title="Improvements">
1165 qemu: Introduce a new video model of type 'none'
1168 Introduce a new video model type that disables the automatic addition
1169 of a video device to domains with 'graphics' specified in their XML.
1170 This can be useful with GPU mediated devices which can serve as the
1171 only rendering devices within the guest.
1176 virsh: Add --alias to attach-disk and attach-interface commands
1179 Add option --alias to set customized device alias name when
1180 using attach-disk or attach-interface commands.
1185 virsh: Support usb and sata address to attach-disk
1188 Usb or sata address could be used when attach-disk with --address.
1189 For example, use usb address as usb:<bus>.<port>, use
1190 sata address as <controller>.<bus>.<unit>.
1195 <release version="v4.5.0" date="2018-07-02">
1196 <section title="New features">
1199 qemu: Provide TPM emulator support
1202 Support QEMU's TPM emulator based on swtpm. Each QEMU guest gets
1203 its own virtual TPM.
1208 bhyve: Support specifying guest CPU topology
1211 Bhyve's guest CPU topology could be specified using the
1212 <code><cpu><topology ../></cpu></code> element.
1217 qemu: Add support for extended TSEG size
1220 Support specifying extended TSEG size for SMM in QEMU.
1225 qemu: Add support for SEV guests
1228 SEV (Secure Encrypted Virtualization) is a feature available on AMD
1229 CPUs that encrypts the guest memory and makes it inaccessible even
1234 <section title="Removed features">
1237 Remove support for qcow/default encrypted volumes
1240 Disallow using a qcow encrypted volume for the guest and
1241 disallow creation of the qcow or default encrypted volume
1242 from the storage driver. Support for qcow encrypted volumes
1243 has been phasing out since QEMU 2.3 and by QEMU 2.9 creation
1244 of a qcow encrypted volume via qemu-img required usage of
1245 secret objects, but that support was never added to libvirt.
1250 Make GnuTLS mandatory
1253 Building without GnuTLS is no longer possible.
1258 qemu: Remove allow_disk_format_probing configuration option
1261 The option represented a security risk when used with malicious
1262 disk images, so users were recommended against enabling it; with
1263 this release, it's been removed altogether.
1267 <section title="Improvements">
1270 capabilities: Provide info about host IOMMU support
1273 Capabilities XML now provide information about host IOMMU support.
1278 virsh: Add --all to domblkinfo command
1281 Alter the <code>domblkinfo</code> command to add the option
1282 --all in order to display the size details of each domain
1283 block device from one command in a output table.
1288 qemu: Allow concurrent access to monitor and guest agent
1291 Historically libvirt prevented concurrent accesses to
1292 the qemu monitor and the guest agent. Therefore two
1293 independent calls (one querying the monitor and the
1294 other querying guest agent) would serialize which hurts
1295 performance. The code was reworked to allow two
1296 independent calls run at the same time.
1301 qemu: Allow configuring the page size for HPT pSeries guests
1304 For HPT pSeries guests, the size of the host pages used to back guest
1305 memory and the usable guest page sizes are connected; the new setting
1306 can be used to request that a certain page size is available in the
1312 Add support to use an raw input volume for encryption
1315 It is now possible to provide a raw input volume as input for
1316 to generate a luks encrypted volume via either virsh vol-create-from
1317 or virStorageVolCreateXMLFrom.
1322 qemu: Add support for vsock hot (un)plug and cold (un)plug
1327 qemu: Add support for NBD over TLS
1330 NBD volumes can now be accessed securely.
1335 qemu: Implement FD passing for Unix sockets
1338 Instead of having QEMU open the socket and then connecting to it,
1339 which is inherently racy, starting with QEMU 2.12 we can open the
1340 socket ourselves and pass it to QEMU, avoiding race conditions.
1345 virsh: Introduce --nowait option for domstat command
1348 When this option is specified, virsh will try to fetch the guest
1349 stats but abort instead of stalling if they can't be retrieved right
1354 <section title="Bug fixes">
1357 qemu: Fix a potential libvirtd crash on VM reconnect
1360 Initialization of the driver worker pool needs to come before libvirtd
1361 trying to reconnect to all machines, since one of the QEMU processes
1362 migh have already emitted events which need to be handled prior to
1363 us getting to the worker pool initialization.
1368 qemu: Fix domain resume after failed migration
1371 Recent versions of QEMU activate block devices before the guest CPU
1372 has been started, which makes it impossible to roll back a failed
1373 migration. Use the <code>late-block-activate</code> migration
1374 capability if supported to avoid the issue.
1379 vmx: Permit guests to have an odd number of vCPUs
1382 An odd number of vCPUs greater than 1 was forbidden in the past,
1383 but current versions of ESXi have lifted that restriction.
1388 <release version="v4.4.0" date="2018-06-04">
1389 <section title="New features">
1392 bhyve: Support locking guest memory
1395 Bhyve's guest memory may be wired using the
1396 <code><memoryBacking><locked/></memoryBacking></code>
1402 qemu: Provide VFIO channel I/O passthrough support
1405 Support passthrough devices that use channel I/O based mechanism in
1406 a QEMU virtual machine.
1411 qemu: Add support for migration of VMs with non-shared storage over TLS
1414 It's now possible to use the VIR_MIGRATE_TLS flag together with
1415 VIR_MIGRATE_NON_SHARED_DISK. The connection is then secured using the
1416 TLS environment which is setup for the migration connection.
1421 Add support for VM Generation ID
1424 The VM Generatation ID exposes a 128-bit, cryptographically
1425 random, integer value identifier, referred to as a Globally
1426 Unique Identifier (GUID) to the guest in order to notify the
1427 guest operating system when the virtual machine is executed
1428 with a different configuration. Add a new domain XML processing
1429 and a domain capabilities feature.
1434 Introduce virDomainDetachDeviceAlias
1437 This new API enables users to detach device using only its alias.
1442 Introduce new virConnectCompareHypervisorCPU and virConnectBaselineHypervisorCPU APIs
1445 Unlike the old virConnectCompareCPU and virConnectBaselineCPU APIs,
1446 both new APIs consider capabilities of a specific hypervisor.
1451 Introduce SCSI persistent reservations support
1454 The QEMU driver gained support for qemu-pr-helper which enables
1455 guests to issue SCSI commands for persistent reservation.
1460 qemu: Implement multiple screen support for virDomainScreenshot
1463 While the virDomainScreenshot API supported multihead video cards,
1464 the implementation was missing. But now that QEMU implemented it
1465 libvirt has done as well.
1470 qemu: add support for vhost-vsock-device
1473 A new vsock device was introduced, allowing communication between
1474 the guest and the host via the AF_VSOCK family.
1478 <section title="Improvements">
1481 qemu: Add suport for OpenGL rendering with SDL
1484 Domains using SDL as a graphics backend will now be able to use
1485 OpenGL accelerated rendering.
1490 qemu: Add support for 'output' audio codec
1493 Support QEMU's 'hda-output' codec advertising only a line-out for ich6
1494 and ich9 sound devices.
1499 virsh: Enhance event name completion
1502 Implement event name completion for some commands (e.g. event,
1503 secret-event, pool-event and nodedev-event)
1508 <release version="v4.3.0" date="2018-05-02">
1509 <section title="New features">
1512 qemu: Add support for the pcie-to-pci-bridge controller
1515 Pure PCIe guests such as x86_64/q35 and aarch64/virt will now
1516 add this controller when traditional PCI devices are in use.
1521 Xen: Support setting CPU features for host-passthrough model
1524 The CPU model presented to Xen HVM domains is equivalent to libvirt's
1525 host-passthrough model, although individual features can be enabled
1526 and disabled via the cpuid setting. The libvirt libxl driver now
1527 supports enabling and disabling individual features of the
1528 host-passthrough CPU model.
1532 <section title="Removed features">
1535 Xen: Drop the legacy xend-based driver
1538 The xm/xend toolstack was deprecated in Xen 4.2 and removed
1539 from the Xen sources in the 4.5 development cycle. The libvirt
1540 driver based on xend is now removed from the libvirt sources.
1544 <section title="Improvements">
1547 qemu: Support hot plug and hot unplug of mediated devices
1550 Libvirt now allows mediated devices to be hot plugged and hot
1551 unplugged from a guest rather than reporting an error that this isn't
1552 supported. In fact, kernel has been supporting this since 4.10.
1556 <section title="Bug fixes">
1559 Improve handling of device mapper targets
1562 When starting a domain with a disk backed by a device
1563 mapper volume libvirt also needs to allow the storage
1564 backing the device mapper in CGroups. In the past
1565 kernel did not care, but starting from 4.16 CGroups are
1566 consulted on each access to the device mapper target.
1571 <release version="v4.2.0" date="2018-04-01">
1572 <section title="New features">
1575 Support building with Python 3
1578 Python is required to build libvirt, and up until now only Python 2
1579 could be used as an interpreter. All scripts used during build have
1580 now been made compatible with Python 3, which means both major
1581 releases of the language are fully supported.
1586 qemu: Provide ccw address support for graphics and input devices
1589 Support the virtio-gpu-ccw device as a video device and
1590 virtio-{keyboard, mouse, tablet}-ccw devices as input devices
1595 <section title="Improvements">
1598 qemu: Add logging of guest crash information on S390
1601 On S390, when the guest crashes and QEMU exposes the guest crash
1602 information, log the relevant data to the domain log file.
1607 qemu: use arp table of host to get the IP address of guests
1610 Find IP address of a VM by arp table on hosts.
1611 If someone customizing IP address inside VM, it will be helpful.
1616 Xen: Remove hard-coded scheduler weight
1619 The libxl driver was accidentally hard-coding the per-domain
1620 scheduler weight to 1000, silently ignoring any user-provided
1621 <code><shares></code> in <code><cputune></code>. The
1622 driver now honors <code><shares></code>, and defers setting
1623 a default value to Xen. Note that the Xen default is 256, so any
1624 domains started after this improvement will have one fourth the
1625 shares of previously started domains. If all domains must have
1626 equal CPU shares, administrators must manually set the weight of
1627 previously started domains to 256, or restart them.
1631 <section title="Bug fixes">
1634 qemu: TLS migration now enforces use of TLS for the NBD connection
1637 When the VIR_MIGRATE_TLS flag was used with the migration API libvirt
1638 did not ensure that the NBD connection was using TLS as well. The code
1639 now rejects such migration as the TLS transport for NBD is not ready
1640 yet, but prevents a false sense of security that TLS would be used.
1641 The support TLS for NBD will be added soon.
1646 <release version="v4.1.0" date="2018-03-05">
1647 <section title="New features">
1650 Added support for CAT (Cache allocation Technology)
1653 Domain vCPU threads can now have allocated some parts of host cache
1654 using the <code>cachetune</code> element in <code>cputune</code>.
1659 Allow opening secondary drivers
1662 Up until now it was possible to connect to only hypervisor drivers
1663 (e.g. qemu:///system, lxc:///, vbox:///system, and so on). The
1664 internal drivers (like network driver, node device driver, etc.) were
1665 hidden from users and users could use them only indirectly. Starting
1666 with this release new connection URIs are accepted. For instance
1667 network:///system, storage:///system and so on.
1672 virtlogd, virtlockd: Add support for admin protocol
1675 These two daemons now support admin protocol through which some admin
1676 info can be gathered or some configuration tweaked on the fly.
1680 <section title="Improvements">
1683 virsh: Enhance bash completion
1686 Implement more bash completions so that basic libvirt
1687 objects can be auto-completed (e.g. networks,
1688 interfaces, NWFilters, and so on).
1693 qemu: Use VIR_ERR_DEVICE_MISSING for various hotplug/detach messages
1698 qemu: Allow showing the dump progress for memory only dump
1701 Alter the QEMU dump-guest-memory command processing to check
1702 for and allow asynchronous completion which then allows for
1703 the virsh dump --memory-only --verbose command to display percent
1709 conf: add support for setting Chassis SMBIOS data fields
1714 libxl: add support for setting clock offset and adjustment
1719 Make port allocator global
1722 Up until now each driver had their own port allocator module. This
1723 meant that info on port usage was not shared. Starting with this
1724 release, the port allocator module is made global and therefore
1725 drivers allocate ports from global pool.
1730 Fixed some compiler warnings that appear with GCC 8
1734 <section title="Bug fixes">
1737 qemu: Check for unsafe migration more thoroughly
1740 If a domain disk is stored on local filesystem (e.g. ext4) but is
1741 not being migrated it is very likely that domain is not able to
1742 run on destination. Regardless of share/cache mode.
1747 qemu: Fix updating device with boot order
1750 Starting with 3.7.0 release updating any device with boot order would
1751 fail with 'boot order X is already used by another device' while in
1752 fact it was the very same device.
1757 virlog: determine the hostname on startup CVE-2018-6764
1760 At later point it might not be possible or even safe to use
1761 getaddrinfo(). It can in turn result in a load of NSS module which
1762 can even be loaded from unsage guest filesystem and thus escape the
1763 confinment of its container.
1768 qemu: Rework vCPU statistics fetching
1771 Fetching vCPU statistics was very expensive because it lead to waking
1772 up vCPU threads in QEMU and thus it degraded performance. The code
1773 was reworked so that fetching statistics does not wake up halted
1779 qemu: unlink memory backing file on domain shutdown
1782 Depending on the filesystem where domain memory is stored, some files
1783 might have been left behind. This is not a problem on hugetlbfs, but
1784 it is a problem on regular filesystems like ext4.
1789 qemu: Fix shutting down domains in parallel
1792 If multiple domains were being shut down in parallel, libvirtd might
1798 nodedev: Update PCI mdev capabilities dynamically
1801 PCI devices may have other nested capabilities, like SRIOV and mdev
1802 which depend on the device being plugged into the native vendor
1803 driver. However, in case such a device is directly assigned to a guest
1804 using VFIO driver, the device will naturally lose these capabilities
1805 and libvirt needs to reflect that.
1810 <release version="v4.0.0" date="2018-01-19">
1811 <section title="New features">
1814 tools: Provide bash completion support
1817 Both <code>virsh</code> and <code>virt-admin</code> now implement
1818 basic bash completion support.
1823 qemu: Refresh capabilities on host microcode update
1826 A microcode update can cause the CPUID bits to change; therefore,
1827 the capabilities cache should be rebuilt when such an update is
1828 detected on the host.
1833 lxc: Set hostname based on container name
1837 <section title="Improvements">
1840 CPU frequency reporting improvements
1843 The CPU frequency will now be reported by <code>virsh nodeinfo</code>
1844 and other tools for s390 hosts; at the same time; CPU frequency has
1845 been disabled on aarch64 hosts because there's no way to detect it
1851 libxl: Mark domain0 as persistent
1856 Xen: Add support for multiple IP addresses on interface devices
1861 qemu: Add support for hot unplugging redirdev device
1865 <section title="Bug fixes">
1868 qemu: Enforce vCPU hotplug granularity constraints
1871 QEMU 2.7 and newer don't allow guests to start unless the initial
1872 vCPUs count is a multiple of the vCPU hotplug granularity, so
1873 validate it and report an error if needed.
1878 <release version="v3.10.0" date="2017-12-04">
1879 <section title="New features">
1882 conf: Support defining distances between virtual NUMA cells
1885 A NUMA hardware architecture supports the notion of distances
1886 between NUMA cells. This can now be specified using the
1887 <code><distances></code> element within the NUMA cell
1888 configuration. Drivers which support this include Xen and QEMU.
1893 Xen: Support defining vNUMA topology
1896 Xen now supports defining a virtual NUMA topology for VMs,
1897 including specifying distances between NUMA cells.
1902 qemu: Add the ability to configure HPT resizing for pSeries guests
1905 The user can now decide whether HPT (Hash Page Table) resizing
1906 should be enabled, disabled or required instead of leaving it up to
1907 hypervisor defaults and negotiation between the guest and the host.
1912 qemu: Add vmcoreinfo feature
1915 Starting with QEMU 2.11, the guest can save kernel debug
1916 details when this feature is enabled and the kernel supports
1917 it. It is useful to process kernel dump with KASLR enabled,
1918 and also provides various kernel details to crash tools.
1923 conf: Move the auth and encryption definitions to disk source
1926 Allow parsing and formatting of the <code>auth</code> and
1927 <code>encryption</code> sub-elements to be a child of the
1928 <code>source</code> element. This will allow adding an
1929 <code>auth</code> sub-element to a <code>backingStore</code>
1930 or <code>mirror</code> elements as a means to track specific
1931 authentication and/or encryption needs.
1935 <section title="Improvements">
1938 vbox: Add VirtualBox 5.2 support
1943 vbox: Add support for configuring storage controllers
1946 The VirtualBox driver now supports the <code><controller></code>
1947 element in the domain XML for configuring storage controllers in VBOX
1948 VMs. Additionally, libvirt's domain XML schema was updated to allow
1949 optional <code>model</code> attribute for <code><controller
1950 type='ide'></code> which is used by the VBOX driver to set the
1951 IDE controller model to be one of 'piix4', 'piix4' (default), or
1952 'ich6'. Finally, with this change <code>dumpxml</code> generates
1953 <code><controller></code> elements that correspond to current
1954 VBOX VM storage controller configuration.
1959 vbox: Add support for attaching empty removable disks
1962 The VirutalBox driver now supports adding CD-ROM and floppy disk
1963 devices that do not have the disk source specified. Previously such
1964 devices were silently ignored.
1969 vbox: Add support for attaching SAS storage controllers
1972 In VirtualBox, SCSI and SAS are distinct controller types whereas
1973 libvirt does not make such distinction. Therefore, the VBOX driver was
1974 updated to allow attaching SAS controllers via <code><controller
1975 type='scsi' model='lsisas1068'></code> element. If there are
1976 both SCSI and SAS controllers present in the VBOX VM, the domain XML
1977 can associate the disk device using the <code><address></code>
1978 element with the <code>controller</code> attribute, and optionally,
1979 set the port via <code>unit</code> attribute.
1984 qemu: Generate predictable paths for qemu memory backends
1987 In some cases management applications need to know
1988 paths passed to memory-backend-file objects upfront.
1989 Libvirt now generates predictable paths so applications
1990 can prepare the files if they need to do so.
1995 Shareable disks work properly with recent qemu
1998 Recent qemu versions added image locking to avoid potential corruption
1999 of disk images. This broke shareable disks with libvirt since the
2000 feature was turned on by default in qemu. Libvirt now enables sharing
2001 of those disks in qemu so that the image locking is not applied in
2002 that case. Additionally libvirt now checks that shareable disks have
2003 supported format (raw) to avoid metadata corruption.
2008 Improve serial console behavior on non-x86 architectures
2011 ppc64, aarch64 and s390x guests were treating the <serial>
2012 and <console> elements differently from x86, in some cases
2013 presenting misleading information to the user. The behavior is now
2014 consistent across all architectures and the information reported
2019 <section title="Bug fixes">
2022 vbox: Do not ignore failures to attach disk devices when defining
2025 The <code>define</code> now fails and reports an error if any of the
2026 <code>controller</code> or <code>disk</code> devices specified in the
2027 domain XML fail to attach to the VirtualBox VM.
2032 vbox: Fix dumpxml to always output disk devices
2035 The VirtualBox driver was ignoring any disk devices in
2036 <code>dumpxml</code> output if there was a SAS storage controller
2042 vbox: Fix dumpxml to always generate valid domain XML
2045 When a VirtualBox VM has multiple disks attached, each to a different
2046 storage controller that uses 'sd' prefix for block device names e.g.
2047 one disk attached to SATA and one to SCSI controller, it no longer
2048 generates XML where both would have 'sda' device name assigned.
2049 Instead it properly assigns 'sda' and 'sdb' to those disks in the
2050 order of appearance.
2055 Securely pass iSCSI authentication data
2058 Rather than supplying the authentication data as part of the
2059 iSCSI URL for a disk or host device, utilize the encrypted
2060 secret object to securely pass the authentication data.
2065 <release version="v3.9.0" date="2017-11-02">
2066 <section title="New features">
2069 Add capability to allow hot (un)plug of a domain watchdog device
2074 Allow users to set device aliases
2077 Users can set aliases to domain devices and thus identify them
2083 qemu: Support multiqueue for virtio-blk
2086 Multiqueue support for <code>virtio-blk</code> has been available
2087 in QEMU ever since 2.7.0, and now libvirt guests can enable it.
2092 Add virDomainSetLifecycleAction API
2095 Provided a new API to allow dynamic guest lifecycle control for
2096 guest reactions to poweroff, restart, or crash type events related
2097 to the domain XML <code>on_poweroff</code>, <code>on_reboot</code>,
2098 and <code>on_crash</code> elements. The
2099 <code>virsh set-lifecycle-action</code> command was created to
2100 control the actions.
2105 qemu: Allow cold(un)plugging and hot(un)plugging input devices
2110 net: Implement QoS for vhostuser
2114 <section title="Improvements">
2117 Allow a logical volume to be create using LUKS
2120 A logical volume may be created using an <code>encryption</code>
2121 element using "luks" format. This does require a previously created
2122 <code>secret</code> to store the passphrase used to encrypt the
2123 volume Adding the volume to a domain can then either provide the
2124 secret or allow the consumer in the guest to provide the passphrase
2125 in order to decrypt the volume.
2130 net: Ignore auto-generated MAC address when detaching an interface
2133 If the MAC address has not been specified by the user, libvirt will
2134 try and fill in the gaps by generating one; however, for some error
2135 paths that led to some confusing error messages, so when an
2136 auto-generated MAC address is specified the error message will not
2137 include the auto-generated MAC.
2142 net: Enable MAC address lookup for virDomainInterfaceStats
2147 apparmor: Several improvements
2150 Changes include permitting access to data about USB devices and
2151 <code>dnsmasq</code> instances, allowing spaces in guest names and
2157 cpu: Use CPU information obtained from QEMU when possible
2160 Recent QEMU versions can expose information about which CPU models
2161 are available and usable on the host; libvirt will now make use of
2162 such information whenever possible.
2167 hyperv: Various improvements
2170 The error reported when clients can't connect to Hyper-V has been
2171 made more descriptive, and memory limits for guests are now mapped
2172 to more appropriate libvirt equivalents.
2177 qemu: Report QEMU error on failed migration
2180 Instead of reporting a generic error, ask QEMU for a more detailed
2181 and thus hopefully more helpful one.
2186 vbox: Implement autoport for RDP
2189 libvirt will now obtain the (dynamically allocated) RDP port number
2190 from VirtualBox itself, avoiding conflicts between multiple guests
2191 wanting to use RDP at the same time.
2196 qemu: Allow rotation of small logs
2199 On a host where numerous unique instances are executed per day, it's
2200 quite possible that, even though each of the single log files are
2201 fairly small, collectively the quantity and volume may add tens of
2202 thousands of log files to the <code>/var/log/libvirt/qemu/</code>
2203 directory. Removing the constraints that log have to be bigger than
2204 100 KiB before they can be rotated solves the issue.
2208 <section title="Bug fixes">
2211 Fix swapped interface statistics and QoS
2214 Due to internal implementation, reported statistics for
2215 some types of interfaces were swapped (RX appeared in
2216 TX and vice versa). Similarly, QoS was set in reversed
2222 Properly resize local LUKS encrypted volume
2225 Resizing of a local LUKS encrypted volume will now use qemu-img
2226 to resize the volume. This will require configuring a secret for
2227 the LUKS encrypted volume.
2232 qemu: Reserve PCI addresses for implicit i440fx devices
2235 Failing to do so causes the addresses to be considered usable by
2236 libvirt, which means they could be assigned to more than one device
2237 resulting in the guest failing to start.
2242 spec: Restart libvirtd only at the end of the upgrade process
2245 Use <code>%posttrans</code> to make sure <code>libvirtd</code>
2246 is not restarted before all other components, such as the library
2247 itself and storage / hypervisor drivers, have already been upgraded.
2251 <section title="Security">
2254 qemu: Ensure TLS clients always verify the server certificate
2257 While it's reasonable to turn off client certificate validation,
2258 as setting it up can be non-trivial, clients should always verify
2259 the server certificate to avoid MITM attacks. However, libvirt was
2260 using the same knob to control both checks, leading to
2261 CVE-2017-1000256 / LSN-2017-0002.
2266 <release version="v3.8.0" date="2017-10-04">
2267 <section title="New features">
2270 qemu: Added support for cold-(un)plug of watchdog devices
2275 qemu: Added support for setting IP address os usernet interfaces
2280 qemu: Added support for Veritas Hyperscale (VxHS) block devices
2285 storage: Added new events for pool-build and pool-delete
2289 <section title="Improvements">
2292 qemu: Set DAC permissions properly for spice rendernode
2295 When a <code>rendernode</code> path is set for SPICE GL on
2296 <code>qemu:///system</code>, we now correctly set DAC permissions
2297 on the device at VM startup. This is the last remaining hurdle to
2298 let SPICE GL work for <code>qemu:///system</code> without any
2299 external host changes.
2304 nodedev: Add switchdev offload query to NIC capabilities
2307 Allow querying the NIC interface capabilities for the
2308 availability of switchdev offloading (also known as
2309 kernel-forward-plane-offload).
2314 New CPU models for AMD and Intel
2317 AMD EPYC and Intel Skylake-Server CPU models were added together with
2323 Improve long waiting when saving a domain
2326 While waiting for a write to disk to be finished, e.g. during save,
2327 even simple operations like <code>virsh list</code> would be blocking
2328 due to domain lock. This is now resolved by unlocking the domain
2329 in places where it is not needed.
2333 <section title="Bug fixes">
2336 Proper units are now used in virsh manpage for dom(mem)stats
2339 Previously the documentation used multiples of 1000, but now it is
2340 fixed to use multiples of 1024.
2345 qemu: Fix error reporting when disk attachment fails
2348 There was a possibility for the actual error to be overridden or
2349 cleared during the rollback.
2354 qemu: Fix assignment of graphics ports after daemon restart
2357 This could be seen with newer kernels that have bug regarding
2358 SO_REUSEADDR. After libvirtd was restarted it could assign already
2359 used address to new guests which would make them fail to start. This
2360 is fixed by marking used ports unavailable when reconnecting to
2361 running QEMU domains.
2366 Fix message decoding which was causing a very strange bug
2369 When parsing an RPC message with file descriptors was interrupted and
2370 had to restart, the offset of the payload was calculated badly causing
2371 strange issues like not being able to find a domain that was not
2377 <release version="v3.7.0" date="2017-09-04">
2378 <section title="New features">
2381 qemu: Add managedsave-edit commands
2384 Using managedsave-dumpxml, managedsave-define and managedsave-edit
2385 commands, now we can dump and edit the XML configuration of domain
2386 which has managedsave image.
2391 qemu: Add migrate-getmaxdowntime command
2394 Currently, the maximum tolerable downtime for a domain being migrated
2395 is write-only from libvirt, via migrate-setmaxdowntime. This
2396 implements a complementary migrate-getmaxdowntime command
2401 bhyve: Support autoport for VNC ports
2404 It's no longer necessary to explicitly specify VNC port for the bhyve
2405 guests. With the autoport feature it will be allocated automatically.
2406 Please refer to the bhyve driver documentation for examples.
2411 qemu: Added support for setting heads of virtio GPU
2416 qemu: Added support to configure reconnect timeout for chardev devices
2419 When you have a TCP or UNIX chardev device and it's connected somewhere
2420 you can configure reconnect timeout if the connection is closed.
2424 <section title="Improvements">
2427 qemu: Report a clear error when dropping a VM during startup
2430 "Failed to load config for domain 'DOMNAME'" is now reported if a VM
2431 config can't be parsed for some reason, and thus provides a clear
2432 indication for users (and devs).
2437 apparmor: Update for QEMU 2.10 compatibility
2440 Starting with QEMU 2.10, disk images and NVRAM files get
2441 automatically locked to prevent them from being corrupted; however,
2442 file locking needs to be explicitly allowed through
2443 <code>virt-aa-helper</code> or AppArmor will reject the requests and
2444 the guest will not be able to run.
2449 virsh: List Unix sockets in 'domdisplay' output
2452 VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets
2453 as connection endpoints, but such a configuration was not handled
2454 correctly by <code>virsh domdisplay</code>, causing the respective
2455 endpoints to be missing from the output.
2460 qemu: Don't check whether offline migration is safe
2463 Since offline migration only copies the guest definition to the
2464 destination host, data corruption is not a concern and the operation
2465 can always be performed safely.
2470 virt-host-validate: Fix IOMMU detection on ppc64
2474 <section title="Bug fixes">
2477 qemu: Better support for international domain names (with wide characters)
2480 There were some issues with multi-byte domains getting lost
2481 on daemon restart due to truncation, so the code now handles
2482 multi-byte names a bit better.
2487 qemu: Support long domain names with namespaces
2490 Domains with extremely long names would fail to start due to
2491 temporary namespace paths being created with the whole name.
2492 The path is now generated with shortened name instead.
2497 qemu: Tolerate missing emulator binary during libvirtd restart
2500 For some time libvirt required qemu capabilities being present when
2501 parsing VM configs during startup. As a side effect VM configs would
2502 fail to parse and thus vanish, if the emulator binary would be
2503 uninstalled or broken. Libvirt now tolerates when capabilities
2504 are missing during startup.
2509 qemu: Prevent pSeries guests from disappearing in some situations
2512 pSeries guest would disappear if any of the host devices they were
2513 configured to use was not available during libvirtd startup, which
2514 could easily happen for SR-IOV Virtual Functions. This scenario is
2515 now handled correctly.
2520 qemu: Honor <on_reboot/> setting
2523 The setting was accepted by the parser, but not actually implemented.
2528 Fix --verbose option for all daemons
2531 Since v3.0.0, the option had been ignored by all libvirt daemons
2532 (<code>libvirtd</code>, <code>virtlogd</code> and
2533 <code>virtlockd</code>); it's now working as intended once again.
2538 <release version="v3.6.0" date="2017-08-02">
2539 <section title="New features">
2542 hyperv: Implement virDomainSetMemory and virDomainSendKey APIs
2547 qemu: Support multiple PHBs for pSeries guests
2550 pSeries guests can now have multiple PHBs (PCI Host Bridges), which
2551 show up as separate PCI domains in the guest. To create additional
2552 PHBs, simply add PCI controllers with model <code>pci-root</code>
2553 to the guest configuration.
2558 qemu: Isolate hostdevs on pSeries guests
2561 To enable better error reporting and recovery, unrelated hostdevs
2562 will now be automatically isolated on pSeries guests by placing them
2563 on separate PHBs (PCI Host Bridges).
2567 <section title="Improvements">
2570 qemu: platform serial devices can now use chardev features
2573 QEMU VMs that depend on platform serial devices can now use
2574 QEMU's <code>-chardev</code> option, which enables access to
2575 advanced features like log file configuration. This applies
2576 to the default serial devices for arm, aarch64, and some ppc
2582 Require use of GCC 4.4 or Clang compilers
2585 We only ever test libvirt with GCC or Clang (which
2586 provides a GCC compatible compilation environment).
2587 Between them, these compilers cover every supported
2588 operating system platform, including Windows.
2593 qemu: shared disks with directsync cache should be safe for migration
2596 At present shared disks can be migrated with either readonly or
2597 <code>cache=none</code>. But <code>cache=directsync</code> should be
2598 safe for migration, because both <code>cache=directsync</code> and
2599 <code>cache=none</code> don't use the host page cache, and
2600 <code>cache=direct</code> write through qemu block layer cache.
2605 Handle hotplug change on VLAN configuration using OVS
2608 Libvirt was taught to handle VLAN change for running OVS interface.
2612 <section title="Bug fixes">
2615 qemu: Use vCPU 'node-id' property and pass it back to qemu
2618 vCPU properties gathered from query-hotpluggable-cpus need to be
2619 passed back to QEMU. As QEMU did not use the node-id property until
2620 now and libvirt forgot to pass it back properly (it was parsed but
2621 not passed around) we did not honor this.
2626 Miscellaneous stream fixes
2629 After introducing sparse stream features there were still some known
2630 bugs left. Those are fixed in this release.
2635 qemu: Miscellaneous domain NS fixes
2638 Libvirt starts qemu domains in separate Linux namespaces for a while
2639 now. However, there were still some bugs lingering. For instance
2640 libvirt did not know how to handle file based bind mounts.
2645 Various CPU driver improvements
2648 There were some minor bugs when using 'host-model' CPU.
2653 <release version="v3.5.0" date="2017-07-04">
2654 <section title="New features">
2657 qemu: Add support for loadparm for a boot device
2660 Add an optional boot parameter 'loadparm' for a boot device.
2661 Loadparm is an 8 byte parameter that, when present, is queried by
2662 S390 guests via sclp or diag 308. Linux guests on S390 use it to
2663 select a boot entry.
2668 Support controlling how video devices are exposed to the bhyve guests
2671 The <code>vgaconf</code> attribute was added to <code>video</code>'s
2672 <code>driver</code> element. Possible values are: <code>on</code>,
2673 <code>off</code>, and <code>io</code>. It controls the way how
2674 bhyve exposes video devices to its guests; various guest OSes might
2675 require different settings to boot properly.
2680 qemu: Add support for live updates of coalesce settings
2683 Users can now use <code>virsh update-device</code> to change
2684 the coalesce settings of an interfaces while the domain is
2690 qemu: Allow VirtIO devices to use vIOMMU
2693 It is now possible to turn on IOTBL for the vIOMMU and have VirtIO
2694 devices use it, provided they have been configured appropriately.
2698 <section title="Improvements">
2701 qemu: block copy job can be used with persistent domains
2704 Until now it was not possible to use block copy with persistent VMs.
2705 In use cases where it's not required to recover the job after VM
2706 shutdown, it's possible to specify VIR_DOMAIN_BLOCK_COPY_TRANSIENT_JOB
2707 flag to start the copy job.
2712 JSON pseudo-protocol backing store parser supports new format of qemu 2.9
2715 QEMU 2.9 modified a few structures corresponding to the JSON format
2716 of specifying a backing store for a disk image. Libvirt now implements
2722 Capabilities now include info about host's CAT settings
2725 Various information about resource control from the host is
2726 gathered and presented in capabilities if available.
2731 apparmor: Several improvements
2734 Allow access to Ceph config, EFI firmware on both x86_64 and
2735 aarch64, device tree on ppc64 and more.
2740 qemu: Support host-model on POWER9 machines
2744 <section title="Bug fixes">
2747 qemu: snapshot: retrieve image metadata from user provided files
2750 Disk images of an external snapshot created with
2751 VIR_DOMAIN_SNAPSHOT_CREATE_REUSE_EXT flag specified would not be
2752 scanned for metadata after recent changes.
2754 The metadata is necessary to allow keeping relative paths between
2755 images when doing a block-commit.
2760 Parse decimal numbers in a locale-independent way
2763 Some locales, such as <code>de_DE</code> and <code>pt_BR</code>,
2764 use comma rather than dot to separate the integer part from the
2765 fractional part of a decimal number; however, several data sources
2766 such as the kernel use a locale-independent representation and need
2767 to be treated accordingly.
2772 Support compilation with newer compiler and libc versions
2775 Several fixes have been included to make compilation with Clang
2776 4.0.0, GCC 7.1 and glibc >= 2.25.90 possible.
2781 qemu: Query name for vhost-user interfaces at runtime
2784 This makes it possible to use <code>virsh</code> subcommands such
2785 as <code>domiflist</code> and <code>domifstat</code> on vhost-user
2791 qemu: Set MTU for hotplugged interfaces correctly
2794 When hotplugging a network interface, the MTU was only set on the
2795 guest side. Set it on the host side as well.
2800 qemu: Forbid updating MTU for interfaces of running guests
2803 The MTU setting can't be modified while the guest is running, so any
2804 attempt to alter it at runtime will now result in an error rather
2805 than being silently ignored.
2810 qemu: Fix specifying QXL heads with older QEMU releases
2813 Specifying the number of QXL heads was not working correctly for
2814 QEMU releases older than 1.6.
2819 qemu: Fix migration to older libvirt/QEMU versions
2822 When the guest is started, libvirt updates the CPU definition to
2823 reflect the actual CPU features to enforce ABI. We need to send
2824 original and updated CPU definition in order to support migration
2825 to older libvirt/QEMU versions. Only the updated CPU definition
2826 was sent to destination.
2831 <release version="v3.4.0" date="2017-06-02">
2832 <section title="New features">
2835 Improved streams to efficiently transfer sparseness
2838 New extension to virStream was implemented so that
2839 virStorageVolDownload and virStorageVolUpload can preserve file
2845 I/O APIC type can be specified for QEMU/KVM domains
2848 The <code>ioapic</code> tag was added to domain
2849 <code>features</code>, so the type of the I/O APIC can now
2850 be specified (e.g. putting it in userspace for KVM domains).
2855 The reason for VM shutdown is reported, if known
2858 QEMU 2.10 will be able to report the reason for shutting
2859 down (whether that was caused by the guest or not), and
2860 libvirt is prepared for that and reports that information in
2861 its shutdown event as well, if it is known.
2865 <section title="Improvements">
2868 Repository now has new README.md file
2871 The new file uses markdown syntax, so it looks better on
2872 github and possibly other web pages, but it has also more
2873 useful information. The old README is now symlink to the
2879 qemu: Use GICv2 by default for aarch64/virt TCG guests
2882 The emulated GICv3 has some limitations that make it unusable as a
2883 default; use GICv2 until they're sorted out. This change makes it
2884 once again possible to run aarch64/virt guests on a x86_64 host
2885 without having to tweak their configuration.
2890 Additional capabilities for the node_device module
2893 Introduce two new capabilities to the node_device module. The first
2894 is for CCW devices, most common on the S390 architecture. The second
2895 is for fibre channel-backed SCSI devices and exposes the
2896 fc_remote_port sub-capability to SCSI target devices.
2901 Node devices now report Mediated device capabilities
2904 Endpoint devices support new <code>mdev</code> capability
2905 and their parents now report the supported types in new
2906 <code>mdev_types</code> capability.
2911 Capabilities now report information about host caches
2914 If supported in the kernel, host capabilities will now list
2915 L3 caches. The code for other levels was added as well, but
2916 only L3 caches are reported currently.
2921 POWER9 CPU model was added
2924 It is now properly reported in host capabilities.
2929 libxl: NUMA sibling distances are now reported in host capabilities
2934 VMDK version 3 files are now properly detected
2939 Interrupt remapping and Extended interrupt mode for IOMMU devices
2942 These two new features can now be controlled with new
2943 <code><driver intremap='on/off' eim='on/off'/></code>
2944 tag for <code>iommu</code> devices.
2949 Graphics in libxl domains now have default addresses
2952 Even though there were default addresses before this change,
2953 they were not saved in the XML. It is now possible to see
2954 and control the listen addresses properly.
2959 Default USB controllers are now added for devices in libxl domains
2962 Even though they were added automatically when USB device
2963 was attached, they could've been missing in some other
2964 cases. The logic is now fixed so there are always USB
2965 controllers, even if there was none of them in the specified
2971 Limits for RPC messages were increased
2974 Hitting the RPC limits we have is easier every day, so they
2975 were increased once again and some guessing logic was
2976 improved as well. It is now possible to get more stats than
2977 ever using the <code>virConnectGetAllDomainStats()</code>
2978 call and push through even bigger requests and replies for
2983 <section title="Bug fixes">
2986 qemu: Create memory_backing_dir on startup
2989 Libvirt's policy is that directories are created on startup if
2990 they don't exist. We've missed this one.
2995 PCIe 4.0 cards now report proper link speeds
2998 It could happen that the link speed for PCIe devices was not
2999 properly reported or the nodedev-dumpxml just failed. That
3000 was due to mistake in the field width, but should now work
3006 qemu: Do not report errors on shutdown
3009 For some users, in some rare cases, it could happen that
3010 there was an error message "internal error: End of file from
3011 qemu monitor" in the logs even though no problem happened.
3012 The detection of these false positives was improved and such
3013 errors should not show any more.
3018 User-specified UNIX socket paths for virtio channels should not be reset
3021 It could happen, in some cases, that libvirt would mistake a
3022 user-specified path for its own generated one and thus
3023 remove it from the XML. The detection of such addresses was
3029 Fix address reservation during RNG hot-plug
3032 When error occurred in a specific point in time during the
3033 hot-plug of an RNG device, it could happen that an address
3034 was released even though another device was already using
3035 it, making it possible to hot-plug another device with that
3036 address, effectively having duplicated addresses in the XML.
3041 <release version="v3.3.0" date="2017-05-05">
3042 <section title="New features">
3045 net: Add support for coalesce settings
3048 Enabling data batching through these settings can improve network
3049 performance for guests.
3054 qemu: Add support for guest CPU cache specification
3057 This features allows fine-grained control of the cache behavior
3063 qemu: Add support for the qemu-xhci USB controller
3067 <section title="Improvements">
3070 hyperv: Support Hyper-V 2012 and newer
3073 Starting with Hyper-V 2012 the API has changed causing the existing
3074 driver to be unable to send and process requests properly. This has
3075 been resolved by adding abstractions to handle the differences and
3076 ease handling such breaks if they happen in the future.
3081 libxl: Add support for nested HVM domains
3084 Xen has supported nested HVM domains since version 4.4. The
3085 libvirt libxl driver now supports nested HVM domains by
3086 specifying the host-passthrough CPU mode when defining a
3092 qemu: Implement ACPI support for aarch64 guests
3095 Up until this point, ACPI support was only advertised for x86_64
3096 guests and disabling it for aarch64 guests was not possible at all.
3101 vz: Add support for changing the number of vCPUs
3106 qemu: Automatically choose the best USB controller for guests
3109 The recently introduced qemu-xhci USB controller is the best choice
3110 for both ppc64 and aarch64 guests, so use it by default on those
3111 architectures if available.
3116 daemon: Increase default task limit for libvirtd
3119 The default number of tasks for the pids cgroup controller is 512,
3120 which libvirtd can quickly bump into when starting lots of guests.
3121 Raise the limit to a more reasonable 32768.
3126 docs: Include man pages describing key code names and values
3131 virsh: Report initialization errors
3134 Sometimes virsh might be unable to start: when that happens, report
3135 useful diagnostics instead of failing silently.
3139 <section title="Bug fixes">
3142 nss: Don't require a network restart for libvirt_guest
3145 Previously, the libvirt_guest NSS module would only work properly
3146 after the corresponding network had been restarted; now newly
3147 started guests will be reported correctly right away.
3152 storage: Remove unavailable transient pools after restart
3155 Solve an issue where transient storage pools would be stuck in an
3156 unmanageable state if the source disappeared and libvirtd was
3157 subsequently restarted.
3162 storage: Fix capacity value for LUKS encrypted volumes
3165 The 'capacity' value (e.g. guest logical size) for a LUKS volume is
3166 smaller than the 'physical' value of the file in the file system, so
3167 we need to account for that.
3172 qemu: Fix regression when hyperv/vendor_id feature is used
3175 Guests using the feature would not be started at all; it is now
3176 possible to start them as expected.
3181 qemu: Do not crash on USB address with no port and invalid bus
3186 crypto: Always pad data before encrypting it
3189 If this step is not performed, when the data length matches the
3190 chunk size the decryption routines will misinterpret the last byte
3191 of data as the padding length and fail to decode it correctly.
3196 <release version="v3.2.0" date="2017-04-02">
3197 <section title="New features">
3200 The virt-host-validate tool now supports bhyve hypervisor
3205 Introduce NVDIMM memory model
3208 NVDIMM is new type of memory introduced into QEMU 2.6. The idea is
3209 that we have a non-volatile memory module that keeps the data
3210 persistent across domain reboots and offers much faster data
3211 accesses. However, due to a bug in QEMU, this feature is not enabled
3212 for QEMUs older than 2.9.0.
3217 qemu: Introduce support for generic PCIe Root Ports
3220 For new controllers, a generic device (pcie-root-port) will be used
3221 by default instead of the Intel-specific device (ioh3420), provided
3222 the QEMU binary supports it.
3227 qemu: Add support for checking guest CPU ABI compatibility
3230 When migrating a domain to a different host, restoring a domain from
3231 a file or reverting a snapshot libvirt will make sure the guest CPU
3232 QEMU presents to the guest OS exactly matches the one provided on
3233 the source host (or before the domain's state was saved). This
3234 enhanced check may also be requested when starting a new domain to
3235 ensure the virtual CPU exactly matches the one specified in the XML.
3240 qemu: Add support to migrate using TLS
3243 Add the ability to migrate QEMU guests using TLS via a new flag
3244 VIR_MIGRATE_TLS or virsh migrate '--tls' option. Requires using at
3245 least QEMU 2.9.0 in order to work properly.
3250 qemu: add mediated devices framework support
3253 Recent kernel version introduced new mediated device framework, so
3254 provide an initial support of this framework for libvirt, mainly by
3255 introducing a new host device type in the XML.
3260 qemu: Add support for setting TSC frequency
3263 Setting TSC frequency is required to enable migration for domains
3264 with 'invtsc' CPU feature turned on.
3269 Add support for block device threshold event
3272 When using thin provisioning, management tools need to resize the
3273 disk in certain cases. To avoid having them to poll disk usage
3274 this version introduces an event which will be fired when a given
3275 offset of the storage is written by the hypervisor. Together with the
3276 API it allows registering thresholds for given storage backing
3277 volumes and this event will then notify management if the threshold
3278 is exceeded. Currently only the qemu driver supports this.
3283 bhyve: Add support for UEFI boot ROM, VNC, and USB tablet
3286 The bhyve driver now supports booting using the UEFI boot ROM,
3287 so non-FreeBSD guests that support UEFI could be booted without
3288 using an external boot loader like grub-bhyve. Video is also
3289 supported now, allowing to connect to guests via VNC and use
3290 an USB tablet as an input device. Please refer to
3291 the driver page for domain XML examples.
3295 <section title="Improvements">
3298 qemu: Detect host CPU model by asking QEMU on x86_64
3301 Previously, libvirt detected the host CPU model using CPUID
3302 instruction, which caused libvirt to detect a lot of CPU features
3303 that are not supported by QEMU/KVM. Asking QEMU makes sure we
3304 don't start it with unsupported features.
3309 perf: Add more perf statistics
3312 Add support to get the count of cpu clock time, task clock time,
3313 page faults, context switches, cpu migrations, minor page faults,
3314 major page faults, alignment faults, emulation faults by
3315 applications running on the platform.
3320 Write hyperv crash information into vm log
3323 qemu's implementation of the hyperv panic notifier now reports
3324 information about the crash from the guest os. Starting with this
3325 version, libvirt logs the information to the vm log file for possible
3330 <section title="Bug fixes">
3333 QEMU: Use adaptive timeout for connecting to monitor
3336 When starting qemu, libvirt waits for qemu to create the monitor
3337 socket which libvirt connects to. Historically, there was sharp 30
3338 second timeout after which the qemu process was killed. This
3339 approach is suboptimal as in some scenarios with huge amounts of
3340 guest RAM it can take a minute or more for kernel to allocate and
3341 zero out pages for qemu. The timeout is now flexible and computed by
3342 libvirt at domain startup.
3347 Overwrite (clear) 2 KB instead of just 512 bytes when initializing logical device
3352 Describe the logical backend requirements better for pool-create-as
3357 <release version="v3.1.0" date="2017-03-03">
3358 <section title="New features">
3361 storage: Add Virtuozzo storage backend storage pool
3364 Add new storage backend to support pool and volume management
3365 within the Virtuozzo Storage environment. Virtuozzo Storage is
3366 a highly available distributed software defined storage with
3367 built-in replication and disaster recovery.
3372 qemu: Add support for memory backing with file source
3375 Add support in numa topology for file source inside memory backing
3376 (hugepages are not needed) Three new elements
3377 <source/>,<access/> and <allocation/> were added
3378 to <memoryBacking/> element. Also new configuration parameter
3379 <code>memory_backing_dir</code> was added to qemu.conf.
3384 network: make openvswitch call timeout configurable
3387 Adding the ability to specify the timeout value in seconds for
3388 openvswitch calls in the libvirtd configuration file.
3393 bhyve: add e1000 NIC support
3396 Add support for e1000 NIC. Previously, the only available option
3397 was <code>virtio-net</code>.
3402 libxl: add tunneled migration support
3405 Add tunneled migration to libxl driver, which is always capable of
3406 strong encryption and doesn't require any extra network connection
3407 other than what's required for remote access of libvirtd.
3412 qemu: add rendernode argument
3415 Add a new attribute 'rendernode' to <gl> spice element.
3420 nodedev: add drm capability
3423 Add a new 'drm' capability for Direct Rendering Manager (DRM)
3424 devices, providing device type information.
3429 Add API for individual/specific vCPU hotplug
3432 The new API allows selecting specific vCPUs to be added/removed from
3433 the VM. The existing APIs allowed only adding/removing from the end
3434 which did not play well with NUMA.
3438 <section title="Improvements">
3441 virsh: pool-list: allow both --uuid and --name in one command
3444 Adjust the virsh-pool command to support the --uuid and/or --name
3445 options in order to print just the --name and/or --uuid of pools.
3450 Introduce MTU to domain <interface/> and <network>
3453 Allow setting MTU size for some types of domain interface
3459 libxl: improve support for <timer> configurations
3462 Add support for multiple timers. Extend the tsc timer to
3463 support the emulate mode. Improve conversion of timer XML
3469 storage: modularize the storage driver
3472 Split up the storage driver backends into loadable modules so that
3473 binary distributions don't have to compromise on shipping the storage
3474 driver with all backends which may pull in too many dependencies.
3478 <section title="Bug fixes">
3481 nodedev: Fabric name must not be required for fc_host capability
3484 fabric_name is one of many fc_host attributes in Linux that is
3485 optional and left to the low-level driver to decide if it is
3486 implemented. For example the zfcp device driver does not provide a
3487 fabric name for an fcp host. The requirement for the existence of
3488 a fabric name has been removed by making it optional.
3493 bhyve: change address allocation schema for SATA disks
3496 Previously, the bhyve driver assigned PCI addresses to SATA disks
3497 directly rather than assigning that to a controller and
3498 using SATA addresses for disks. It was implemented this way
3499 because bhyve has no notion of an explicit SATA controller.
3500 However, as this doesn't match libvirt's understanding of
3501 disk addresses, the bhyve driver was changed to follow
3502 the common schema and have PCI addresses for SATA controllers
3503 and SATA addresses for disks. If you're having issues
3504 because of this, it's recommended to edit the domain's XML
3505 and remove <address type='pci'> from the <disk>
3506 elements with <target bus='sata'/> and let libvirt
3507 regenerate it properly.
3512 libxl: maximum memory fixes
3515 Fix reporting of domain maximum memory. Fix setting dom0
3521 libxl: fix disk detach when <driver> not specified
3526 libxl: fix dom0 autoballooning with Xen 4.8
3531 qemu: Allow empty script path to <interface/>
3534 Historically, this was always allowed. Unfortunately, due to some
3535 rework done for 1.3.2 release a bug was dragged in which suddenly
3536 stop allowing domain with such configuration to start.
3541 <release version="v3.0.0" date="2017-01-17">
3542 <section title="New features">
3545 Domain events for metadata content changes
3548 The domain events framework has a new event ID that can
3549 be used to get notifications when domain metadata content
3555 Event notifications for the secret object
3558 The secret object now supports event notifications, covering
3559 lifcycle changes and secret value changes.
3564 New localPtr attribute for "ip" element in network XML
3569 qemu: Support QEMU group I/O throttling
3572 Add the capability to allow group I/O throttling via a new
3573 domain <disk> <iotune> subelement "group_name"
3574 to allow sharing I/O throttling quota between multiple drives.
3579 nss: Introduce libvirt_guest
3582 New <code>libvirt_guest</code> nss module that translates libvirt
3583 guest names into IP addresses.
3588 daemon: Add support for runtime logging settings adjustment
3591 Logging-related settings like log outputs and filters can now be
3592 adjusted during runtime using the admin interface without the
3593 necessity of the daemon's restart.
3598 storage: Add virStorageVolInfoFlags API
3601 Add the API to support using the VIR_STORAGE_VOL_GET_PHYSICAL
3602 flag in order to return the host physical size in bytes
3603 of the image container in the allocation field of the
3604 _virStorageVolInfo structure. The --physical flag has been
3605 added to the virsh vol-info command to access the data.
3610 libxl: Implement virDomainGetMaxVcpus API
3615 storage: Add overwrite flag checking for logical pool
3618 Add support for the OVERWRITE flags for the logical storage
3619 backend including checking for existing data on the target
3620 volumes when building a new logical pool on target volume(s).
3625 qemu: Add support for guest CPU configuration on s390(x)
3629 <section title="Improvements">
3632 perf: Add more perf statistics
3635 Add support to get the count of branch instructions
3636 executed, branch misses, bus cycles, stalled frontend
3637 cpu cycles, stalled backend cpu cycles, and ref cpu
3638 cycles by applications running on the platform.
3643 conf: Display <physical> for volume xml
3646 Add a display of the <physical> size of a disk
3647 volume in the output of the volume XML.
3652 qemu: Use virtio-pci by default for aarch64 mach-virt guests
3655 virtio-pci provides several advantages over virtio-mmio, such
3656 as the ability to hotplug devices and improved performance.
3657 While opting in to virtio-pci has been possible for a while,
3658 newly-defined guests will now use it automatically.
3663 vbox: remove support for VirtualBox 3.x and older
3666 Those old VirtualBox versions have been unsupported by
3667 upstream for a long time and the API of 4.0 and newer has
3668 diverged enough to require code abstractions to handle differences.
3669 Removing support for those old versions drops lots of code from
3670 the driver and simplifies the logic to ease implementation of new
3671 features going forward.
3676 virsh: pool-info: introduce option --bytes
3679 Add option --bytes to virsh pool-info in order ti allow display
3680 of units in bytes rather than default of human readable output.
3685 scsi: Add parent wwnn/wwpn or fabric capability for createVport
3688 Improve the algorithm searching for the parent scsi_host device
3689 for vHBA/NPIV scsi_host creation. Rather than supplying the
3690 "parent" by name, it's now possible to define the parent by
3691 it's wwnn/wwpn or fabric_wwn in the node device create XML or
3692 the storage pool XML.
3697 qemu: aggregate pcie-root-ports onto multiple functions of a slot
3700 When pcie-root-ports are added to pcie-root in order to
3701 provide a place to connect PCI Express endpoint devices,
3702 libvirt now aggregates multiple root ports together onto the
3703 same slot (up to 8 per slot) in order to conserve slots.
3704 Using this method, it's possible to connect more than 200
3705 endpoint devices to a guest that uses PCIe without requiring
3706 setup of any PCIe switches.
3710 <section title="Bug fixes">
3713 lxc: fix accidental killing of containers during libvirtd restart
3716 The libvirt_lxc process was previously not moved into the
3717 container scope. As a result, if systemd reloads its config
3718 after a container is started, when libvirtd is later restarted
3719 it will accidentally kill the containers.
3724 qemu: Correct GetBlockInfo values
3727 For an active domain, correct the physical value provided for
3728 a raw sparse file backed storage and the allocation value provided
3729 for a qcow2 file backed storage that hasn't yet been opened on
3735 qemu: Make virtio console usable on ppc64 guests
3738 The chardev detection code has been improved and can now handle this
3739 configuration properly.
3744 qemu: Enable mount namespace
3747 To avoid funny races with udev relabelling devices under our hands and
3748 to enhance security, libvirt now spawns each qemu process with its own
3754 storage: Fix implementation of no-overwrite for file system backend
3757 Fix file system storage backend implementation of the OVERWRITE
3758 flags to be consistent between code and documentation. Add checks
3759 to ensure that when building a new file system on a target volume
3760 that there is not something already on the disk in a format that
3761 libvirt can recognize.
3766 qemu: Create hugepage path on per domain basis
3769 Historically, all hugepage enabled domains shared the same path under
3770 hugetlbfs. This left libvirt unable to correctly set security labels
3771 on it. With this release, however, each domain is put into a
3772 separate path which is also correctly labeled.
3777 conf: Reject domains with duplicate drive addresses
3780 Reject duplicate drive addresses for disks and hostdevs at
3786 libxl: reverse defaults on HVM net device attach
3789 Fixes network interface attach for HVM domains when no model is
3790 specified. Emulated hotplug isn't yet supported and hence we should
3791 default to the general working scenario.
3796 libxl: always enable pae for x86_64 HVM
3799 By default pae is disabled in libxl. Without an explicit <pae/>
3800 setting in the domain <features> configuration, an x86_64 HVM
3801 domain would be get an i686 environment. pae should always be enabled
3802 for x86_64 HVM domains.
3807 qemu: Fix XML dump of autogenerated websocket
3810 As a result autogenerated websocket port is regenerated on domain restore,
3811 migration and such as it should be.
3816 <release version="v2.5.0" date="2016-12-04">
3817 <section title="New features">
3820 shmem: Add support for additional models
3823 The shmem device can now utilize QEMU's ivshmem-plain and
3824 ivshmem-doorbell, more modern versions of ivshmem.
3829 vbox: Add VirtualBox 5.1 support
3834 libssh: New transport
3837 The new libssh transport allows one to connect to a running
3838 libvirtd via SSH, using the libssh library; for example:
3839 <code>qemu+libssh://server/system</code>.
3844 vhost-scsi: Add support scsi_host hostdev passthrough
3847 Add the capability to pass through a scsi_host HBA and the
3848 associated LUNs to the guest.
3853 Allow debugging of gluster volumes in qemu
3856 Users can now enable debug logging for native gluster
3857 volumes in qemu using the "gluster_debug_level" option in qemu.conf
3862 Pre-allocate memory slots for memory hotplug
3865 Slot numbers for memory devices are now automatically allocated and
3866 thus persistent. In addition slot numbers can be specified without
3867 providing a base address, which simplifies user configuration
3872 qemu: Express devices will be placed on PCIe bus by default
3875 For machine types that use a PCI Express root bus
3876 (e.g. x86_64/Q35 and aarch64/virt), any unaddressed PCI
3877 device that is an Express device (all virtio-1.0 devices,
3878 e1000e, nec-xhci, vfio assigned devices) will be placed on
3879 an Express controller (i.e. a pcie-root-port) instead of a
3880 legacy PCI controller (i.e. pci-bridge) with the root ports
3885 <section title="Improvements">
3888 docs: Better documentation for migration APIs and flags
3893 vbox: Address thread safety issues
3898 virsh: Add support for passing an alternative persistent XML to migrate command
3903 vhostuser: Allow hotplug of multiqueue devices
3908 NEWS: Switch to an improved format
3911 List user-visible changes instead of single commits for a better
3912 high-level overview of differences between libvirt releases.
3917 website: Modernize layout and branding
3920 The libvirt website looked very cluttered and outdated; it has now
3921 been completely overhauled, resulting in a design that's better
3922 organized and more pleasant to look at.
3926 <section title="Bug fixes">
3929 vz: Fix migration in P2P mode
3934 Forbid newline character in names of some libvirt objects
3939 Fix compilation on macOS