May 2026 - Devel - libvirt List Archives

[PATCH 0/7] qemu: add async vCPU unplug coverage
by Akash Kulhalli 10 Jun '26

10 Jun '26

This series adds qemuhotplugtest coverage for async vCPU unplug. Async vCPU unplug does not complete at the device_del reply. The guest CPU remains in a pending removal state until QEMU reports DEVICE_DELETED and the QEMU driver handles the resulting process event. The new harness support lets the tests observe that intermediate state and then complete the unplug through the same process-event path used by the driver. The setvcpu coverage exercises both a single-vCPU x86 hotplug entity and a grouped ppc64 hotplug entity. These cases check the pending live XML before completion and the final live/config XML after completion, while also validating the expected vcpu-removed domain events. The setvcpus coverage adds a separate count-based x86 downscale case. This keeps the independent count-to-entity selection path covered without duplicating the grouped-entity coverage from setvcpu. The series also includes a small domain API cleanup to use the vCPU-specific flag aliases consistently in the vCPU entry points. This keeps the implementation and diagnostics aligned with the API being handled. This is a cosmetic change only, since the underlying values are all mapped to their equivalent enums. While adding this coverage, I realised that virsh was being too restrictive in its parameter combination validation; only QEMU does not support both `--async` and `--guest`. Other drivers may wish to support this behaviour. Accordingly, I've changed the behaviour so that virsh now lets the drivers reject the combination if they're unsupported. The decision now rests with the driver, which is more appropriate. Akash Kulhalli (7): qemu: expose process event handler to tests tests: add helper for DEVICE_DELETED events qemuhotplugtest: support async vcpu unplug completion qemuhotplugtest: test setvcpu async unplug qemuhotplugtest: test setvcpus async downscale domain: use vcpu-specific flag aliases consistently virsh: allow async vcpu options to reach drivers -- 2.49.0

2 16

[PATCH 0/3] qemuxmlconftest: Error out if invalid FDs will be passed
by Peter Krempa 10 Jun '26

10 Jun '26

As promised in: [PATCH 00/30] qemuxmlconftest: Use real FDs in test cases these patches make qemuxmlconftest error out if fake FDs are passed to virCommandPassFD. Peter Krempa (3): testutils: Turn 'virTestDummyFDContext' into a proper type virTestDummyFDContext: Add fields to track errors and 'virTestDummyFDContextMarkError' qemuxmlconftest: Fail if test case tried to pass STDIO or invalid fds to 'virCommandPassFD' tests/qemuxml2argvmock.c | 3 +++ tests/qemuxmlconftest.c | 12 +++++++++- tests/testutils.c | 47 ++++++++++++++++++++++++++++++++++------ tests/testutils.h | 8 ++++++- tests/testutilsqemu.h | 4 +++- 5 files changed, 64 insertions(+), 10 deletions(-) -- 2.54.0

2 4

[PATCH] qemu: validate: treat only real backing chains as NVRAM backingStore
by Efim Shevrin 10 Jun '26

10 Jun '26

qemuDomainInitializePflashStorageSource() always attaches a non-NULL src->backingStore used as an empty virStorageSource chain terminator (type VIR_STORAGE_TYPE_NONE). qemuValidateDomainDefNvram() incorrectly interpreted every non-NULL backingStore as a genuine backing overlay and reported VIR_ERR_CONFIG_UNSUPPORTED, so legitimate UEFI/NVRAM setups were rejected. Check virStorageSourceIsBacking(src->backingStore) instead of a plain pointer test so only a real backing node is rejected. Regression introduced in commit bca731d0f562f0842f56ec2206fdbd721a468f5b. Signed-off-by: Fima Shevrin <efim.shevrin(a)virtuozzo.com> --- src/qemu/qemu_validate.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c index 642244b62e..d8347d1964 100644 --- a/src/qemu/qemu_validate.c +++ b/src/qemu/qemu_validate.c @@ -740,7 +740,14 @@ qemuValidateDomainDefNvram(const virDomainDef *def, return -1; } - if (src->backingStore) { + /* qemuDomainInitializePflashStorageSource() always sets + * src->backingStore to a fresh empty virStorageSource as a + * chain terminator, so a plain `if (src->backingStore)` check + * is always true and rejects every UEFI/NVRAM domain after + * upstream commit that introduced the terminator. Use + * virStorageSourceIsBacking() (type != VIR_STORAGE_TYPE_NONE) + * so we only reject genuine backing chains. */ + if (virStorageSourceIsBacking(src->backingStore)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("backingStore is not supported with NVRAM")); return -1; -- 2.47.1

2 3

[PATCH v2 03/16] migration/qapi: Rename MigrationStats to MigrationRAMStats
by Peter Xu 10 Jun '26

10 Jun '26

This stats is only about RAM, make it accurate. This paves way for statistics for all devices. Thanks to Markus, who pointed out that docs/devel/qapi-code-gen.rst has a section "Compatibility considerations" stated: Since type names are not visible in the Client JSON Protocol, types may be freely renamed. Even certain refactorings are invisible, such as splitting members from one type into a common base type. Hence this change is not ABI violation according to the document. While at it, touch up the lines to make it read better, correct the restriction on migration status being 'active' or 'completed': over time we grew too many new status that will also report "ram" section. Cc: Daniel P. Berrangé <berrange(a)redhat.com> Cc: devel(a)lists.libvirt.org Reviewed-by: Markus Armbruster <armbru(a)redhat.com> Reviewed-by: Juraj Marcin <jmarcin(a)redhat.com> Reviewed-by: Michal Privoznik <mprivozn(a)redhat.com> Signed-off-by: Peter Xu <peterx(a)redhat.com> --- docs/about/removed-features.rst | 2 +- qapi/migration.json | 10 +++++----- migration/migration-stats.h | 2 +- 3 files changed, 7 insertions(+), 7 deletions(-) diff --git a/docs/about/removed-features.rst b/docs/about/removed-features.rst index e75db08410..626162022a 100644 --- a/docs/about/removed-features.rst +++ b/docs/about/removed-features.rst @@ -699,7 +699,7 @@ was superseded by ``sections``. ``query-migrate`` return value member ``skipped`` (removed in 9.1) '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' -Member ``skipped`` of the ``MigrationStats`` struct hasn't been used +Member ``skipped`` of the ``MigrationRAMStats`` struct hasn't been used for more than 10 years. Removed with no replacement. ``migrate`` command option ``inc`` (removed in 9.1) diff --git a/qapi/migration.json b/qapi/migration.json index 7134d4ce47..e3ad3f0604 100644 --- a/qapi/migration.json +++ b/qapi/migration.json @@ -12,7 +12,7 @@ { 'include': 'sockets.json' } ## -# @MigrationStats: +# @MigrationRAMStats: # # Detailed migration status. # @@ -64,7 +64,7 @@ # # Since: 0.14 ## -{ 'struct': 'MigrationStats', +{ 'struct': 'MigrationRAMStats', 'data': {'transferred': 'int', 'remaining': 'int', 'total': 'int' , 'duplicate': 'int', 'normal': 'int', @@ -209,8 +209,8 @@ # If this field is not returned, no migration process has been # initiated # -# @ram: `MigrationStats` containing detailed migration status, only -# returned if status is 'active' or 'completed'(since 1.2) +# @ram: Detailed migration RAM statistics, only returned if migration +# is in progress or completed (since 1.2) # # @xbzrle-cache: `XBZRLECacheStats` containing detailed XBZRLE # migration statistics, only returned if XBZRLE feature is on and @@ -309,7 +309,7 @@ # Since: 0.14 ## { 'struct': 'MigrationInfo', - 'data': {'*status': 'MigrationStatus', '*ram': 'MigrationStats', + 'data': {'*status': 'MigrationStatus', '*ram': 'MigrationRAMStats', '*vfio': 'VfioStats', '*xbzrle-cache': 'XBZRLECacheStats', '*total-time': 'int', diff --git a/migration/migration-stats.h b/migration/migration-stats.h index c0f50144c9..1153520f7a 100644 --- a/migration/migration-stats.h +++ b/migration/migration-stats.h @@ -27,7 +27,7 @@ /* * These are the ram migration statistic counters. It is loosely - * based on MigrationStats. + * based on MigrationRAMStats. */ typedef struct { /* -- 2.53.0

4 5

Reboot events
by Marek Marczykowski-Górecki 10 Jun '26

10 Jun '26

Hi, What should be the events fired when VM requests reboot? Currently the libxl driver fires VIR_DOMAIN_EVENT_STOPPED, and if on_reboot is set to restart, it fires VIR_DOMAIN_EVENT_STARTED next to it (in fact, _STARTED gets fired before _STOPPED here, but it's another issue). What I'd like to do, is to set on_reboot to destroy, but still receive information whether VM requested just shutdown or reboot, so my application (qubesd daemon) can act on it. I see VIR_DOMAIN_EVENT_ID_REBOOT, but it's not used in libxl. What's the intended semantic of it? Should it be fired next to the lifecycle event (VIR_DOMAIN_EVENT_STOPPED/VIR_DOMAIN_EVENT_STARTED)? Or instead? Is it okay to fire it even if on_reboot=destroy? -- Best Regards, Marek Marczykowski-Górecki Invisible Things Lab

2 3

[PATCH] lxc: truncate LOOP_GET_STATUS64.lo_file_name for long loop backing paths
by Radoslaw Smigielski 10 Jun '26

10 Jun '26

LXC domains with long file-backed filesystem path fail to start when the backing image path is longer than LO_NAME_SIZE (64 bytes, 63 characters plus NUL). When long file path is passed, virFileLoopDeviceAssociate() -> virStrcpy() fails and user gets missleading error and domain fails to start. Example: <filesystem type='file' accessmode='passthrough'> <driver type='loop' format='raw'/> <source file='/root/demoaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa.raw'/> <target dir='/'/> </filesystem> To match losetup behavior we copy the path with virStrcpy() and allow truncation of lo_file_name only if needed, while still calling open() on the unchanged path. Finally log VIR_WARN when the path is expected to be truncated. But still report VIR_ERR_INTERNAL_ERROR for all other virStrcpy() failures. Fixes: https://gitlab.com/libvirt/libvirt/-/work_items/63 Signed-off-by: Radoslaw Smigielski <rsmigiel(a)redhat.com> --- src/util/virfile.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) diff --git a/src/util/virfile.c b/src/util/virfile.c index a0c6cb804862..ae33deb8d223 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c @@ -995,11 +995,18 @@ int virFileLoopDeviceAssociate(const char *file, lo.lo_flags = LO_FLAGS_AUTOCLEAR; - /* Set backing file name for LOOP_GET_STATUS64 queries */ + /* lo_file_name is loop device name, max length is LO_NAME_SIZE bytes. + * Truncate loop device name if file path is longer than LO_NAME_SIZE, + * and still use the full path to open backing file. */ if (virStrcpy((char *) lo.lo_file_name, file, LO_NAME_SIZE) < 0) { - virReportSystemError(errno, - _("Unable to set backing file %1$s"), file); - goto cleanup; + if (strlen(file) >= LO_NAME_SIZE) { + VIR_WARN("Loop backing device name %s truncated to %d bytes.", + file, LO_NAME_SIZE); + } else { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Unable to set loop lo_file_name for %1$s"), file); + goto cleanup; + } } if ((fsfd = open(file, O_RDWR)) < 0) { -- 2.54.0

3 7

[PATCH] virt-aa-helper: Prevent spurious denials for AoE disks
by PUSHKARAJ PATIL 10 Jun '26

10 Jun '26

virt-aa-helper calls virStorageSourceGetMetadata before adding a disk path to a domain's apparmor profile. This probes the device and may trigger an AppArmor denial when the disk is an AoE device under /dev/etherd/. The return value of virStorageSourceGetMetadata is not checked, so the denial has no functional impact but results in noisy dmesg logs. Explicitly deny read access to /dev/etherd/e*.* in the virt-aa-helper profile to avoid these spurious denials. Co-Authored-By: Peter Krempa <pkrempa(a)redhat.com> Signed-off-by: PUSHKARAJ PATIL <pushkaraj.patil(a)in.ibm.com> --- src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in | 1 + 1 file changed, 1 insertion(+) diff --git a/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in b/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in index e209a8bff7..1f1b80b9f9 100644 --- a/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in +++ b/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in @@ -38,6 +38,7 @@ profile virt-aa-helper @libexecdir@/virt-aa-helper { deny /dev/dm-* r, deny /dev/drbd[0-9]* r, deny /dev/dasd* r, + deny /dev/etherd/e*.* r, deny /dev/nvme* r, deny /dev/zd[0-9]* r, deny /dev/mapper/ r, -- 2.50.1 (Apple Git-155)

3 2

[PATCH 1/1] qemu: Inherit SCSI controller model from existing peers
by Denis V. Lunev 09 Jun '26

09 Jun '26

virDomainDefMaybeAddHostdevSCSIcontroller() already inherits the model of an existing SCSI controller when auto-adding peers driven by a hostdev whose drive address overflows the highest declared idx. virDomainDefAddDiskControllersForType(), called for the disk-driven analogue, does not -- it passes model=-1 unconditionally. The unspecified peers then arrive at qemuDomainDefaultSCSIControllerModel() and pick up whichever model the capability cascade settles on (typically lsilogic). A user-declared virtio-scsi controller at idx 0 ends up alongside auto-added lsilogic peers at idx 1+. Plug the same inheritance into qemuDomainDefaultSCSIControllerModel() itself so it covers both the disk-driven path and any other caller that arrives with an unresolved model. The function already prefers architectural defaults (PSeries, ARM virt, RISC-V virt, LoongArch, s390, built-in ESP) over capabilities; peer inheritance slots in just before the capability cascade. A virtio-scsi domain keeps the new peers as virtio-scsi; an lsisas1068 domain stays lsisas1068; a fresh domain with no SCSI controllers still gets the cap-driven default. controller-scsi-auto, the only existing test that flows through this helper, is unaffected (it has no peers to inherit from). controller-scsi-inherit-model covers the new behavior. Signed-off-by: Denis V. Lunev <den(a)openvz.org> --- src/qemu/qemu_domain.c | 15 +++++++ ...ller-scsi-inherit-model.x86_64-latest.args | 38 ++++++++++++++++ ...oller-scsi-inherit-model.x86_64-latest.xml | 45 +++++++++++++++++++ .../controller-scsi-inherit-model.xml | 20 +++++++++ tests/qemuxmlconftest.c | 1 + 5 files changed, 119 insertions(+) create mode 100644 tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.args create mode 100644 tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.xml create mode 100644 tests/qemuxmlconfdata/controller-scsi-inherit-model.xml diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 84c8645259..21f2641521 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -4289,6 +4289,8 @@ virDomainControllerModelSCSI qemuDomainDefaultSCSIControllerModel(const virDomainDef *def, virQEMUCaps *qemuCaps) { + size_t i; + /* For machine types with built-in SCSI controllers, the choice * of model is obvious */ if (qemuDomainHasBuiltinESP(def)) @@ -4305,6 +4307,19 @@ qemuDomainDefaultSCSIControllerModel(const virDomainDef *def, if (qemuDomainIsPSeries(def)) return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI; + /* Inherit the model from any peer SCSI controller already + * resolved on this domain. */ + for (i = 0; i < def->ncontrollers; i++) { + const virDomainControllerDef *peer = def->controllers[i]; + + if (peer->type != VIR_DOMAIN_CONTROLLER_TYPE_SCSI) + continue; + if (peer->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_DEFAULT || + peer->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_AUTO) + continue; + return peer->model; + } + /* If there is no preference, base the choice on device * availability. In this case, lsilogic is favored over * virtio-scsi for backwards compatibility reasons */ diff --git a/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.args b/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.args new file mode 100644 index 0000000000..2bcd9ed92f --- /dev/null +++ b/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.args @@ -0,0 +1,38 @@ +LC_ALL=C \ +PATH=/bin \ +HOME=/var/lib/libvirt/qemu/domain--1-guest \ +USER=test \ +LOGNAME=test \ +XDG_DATA_HOME=/var/lib/libvirt/qemu/domain--1-guest/.local/share \ +XDG_CACHE_HOME=/var/lib/libvirt/qemu/domain--1-guest/.cache \ +XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-guest/.config \ +/usr/bin/qemu-system-x86_64 \ +-name guest=guest,debug-threads=on \ +-S \ +-object '{"qom-type":"secret","id":"masterKey0","format":"raw","file":"/var/lib/libvirt/qemu/domain--1-guest/master-key.aes"}' \ +-machine pc,usb=off,dump-guest-core=off,memory-backend=pc.ram,acpi=off \ +-accel tcg \ +-cpu qemu64 \ +-m size=219136k \ +-object '{"qom-type":"memory-backend-ram","id":"pc.ram","size":224395264}' \ +-overcommit mem-lock=off \ +-smp 8,sockets=8,cores=1,threads=1 \ +-uuid c7a5fdbd-edaf-9455-926a-d65c16db1809 \ +-display none \ +-no-user-config \ +-nodefaults \ +-chardev socket,id=charmonitor,fd=@mon-fd@,server=on,wait=off \ +-mon chardev=charmonitor,id=monitor,mode=control \ +-rtc base=utc \ +-no-shutdown \ +-boot strict=on \ +-device '{"driver":"virtio-scsi-pci","id":"scsi0","bus":"pci.0","addr":"0x2"}' \ +-device '{"driver":"virtio-scsi-pci","id":"scsi1","bus":"pci.0","addr":"0x3"}' \ +-device '{"driver":"virtio-scsi-pci","id":"scsi2","bus":"pci.0","addr":"0x4"}' \ +-device '{"driver":"virtio-scsi-pci","id":"scsi3","bus":"pci.0","addr":"0x5"}' \ +-blockdev '{"driver":"file","filename":"/tmp/data.qcow2","node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}' \ +-blockdev '{"node-name":"libvirt-1-format","read-only":false,"driver":"qcow2","file":"libvirt-1-storage"}' \ +-device '{"driver":"scsi-hd","bus":"scsi3.0","channel":0,"scsi-id":0,"lun":3,"device_id":"drive-scsi3-0-0-3","drive":"libvirt-1-format","id":"scsi3-0-0-3","bootindex":1}' \ +-audiodev '{"id":"audio1","driver":"none"}' \ +-sandbox on,obsolete=deny,elevateprivileges=deny,spawn=deny,resourcecontrol=deny \ +-msg timestamp=on diff --git a/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.xml b/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.xml new file mode 100644 index 0000000000..b971c89a1b --- /dev/null +++ b/tests/qemuxmlconfdata/controller-scsi-inherit-model.x86_64-latest.xml @@ -0,0 +1,45 @@ +<domain type='qemu'> + <name>guest</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory unit='KiB'>219136</memory> + <currentMemory unit='KiB'>219136</currentMemory> + <vcpu placement='static'>8</vcpu> + <os> + <type arch='x86_64' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <cpu mode='custom' match='exact' check='none'> + <model fallback='forbid'>qemu64</model> + </cpu> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu-system-x86_64</emulator> + <disk type='file' device='disk'> + <driver name='qemu' type='qcow2'/> + <source file='/tmp/data.qcow2'/> + <target dev='sdy' bus='scsi'/> + <address type='drive' controller='3' bus='0' target='0' unit='3'/> + </disk> + <controller type='scsi' index='0' model='virtio-scsi'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </controller> + <controller type='scsi' index='1' model='virtio-scsi'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </controller> + <controller type='scsi' index='2' model='virtio-scsi'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> + </controller> + <controller type='scsi' index='3' model='virtio-scsi'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> + </controller> + <controller type='usb' index='0' model='none'/> + <controller type='pci' index='0' model='pci-root'/> + <input type='mouse' bus='ps2'/> + <input type='keyboard' bus='ps2'/> + <audio id='1' type='none'/> + <memballoon model='none'/> + </devices> +</domain> diff --git a/tests/qemuxmlconfdata/controller-scsi-inherit-model.xml b/tests/qemuxmlconfdata/controller-scsi-inherit-model.xml new file mode 100644 index 0000000000..45e96b7ea1 --- /dev/null +++ b/tests/qemuxmlconfdata/controller-scsi-inherit-model.xml @@ -0,0 +1,20 @@ +<domain type='qemu'> + <name>guest</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory unit='KiB'>219136</memory> + <vcpu placement='static'>8</vcpu> + <os> + <type arch='x86_64' machine='pc'>hvm</type> + </os> + <devices> + <emulator>/usr/bin/qemu-system-x86_64</emulator> + <controller type='scsi' index='0' model='virtio-scsi'/> + <disk type='file' device='disk'> + <driver name='qemu' type='qcow2'/> + <source file='/tmp/data.qcow2'/> + <target dev='sdy' bus='scsi'/> + </disk> + <controller type='usb' index='0' model='none'/> + <memballoon model='none'/> + </devices> +</domain> diff --git a/tests/qemuxmlconftest.c b/tests/qemuxmlconftest.c index 8e0f4bcb4f..7a441c1fa2 100644 --- a/tests/qemuxmlconftest.c +++ b/tests/qemuxmlconftest.c @@ -1690,6 +1690,7 @@ mymain(void) DO_TEST_CAPS_LATEST_PARSE_ERROR("disk-scsi-product-length"); DO_TEST_CAPS_LATEST("controller-virtio-scsi"); DO_TEST_CAPS_LATEST("controller-scsi-auto"); + DO_TEST_CAPS_LATEST("controller-scsi-inherit-model"); DO_TEST_FULL("controller-scsi-default-unavailable", ".x86_64-latest", ARG_CAPS_ARCH, "x86_64", ARG_CAPS_VER, "latest", -- 2.51.0

3 2

[PATCH v3 0/2] allow migrations with the ext transport
by Mark Cave-Ayland 09 Jun '26

09 Jun '26

[MCA: I've now adopted these patches from Sergey] Hello, I have a use case for live-migration where the socket of the destination libvirtd is proxied by an mTLS WebSocket proxy. Since there is no WebSocket support in Libvirt, I've decided to use the existing ext transport and point it to a proxy binary on the source host of live-migration. While doing that, I've discovered a couple of roadblocks: 1. There is no way of passing cmdline arguments to the ext program 2. Live-migrations using ext transport fail localhost checks This short series addresses those limitations. v2 -> v3: - fix passing of extArgs v1 -> v2: - new way of passing argv Sergey Dyasli (2): remote: allow passing argv to the ext transport remote: allow migrations with the ext transport src/libvirt-domain.c | 8 +++++--- src/remote/remote_driver.c | 30 ++++++++++++++++++++++++++---- src/util/viruri.c | 32 ++++++++++++++++++++++++++++++++ src/util/viruri.h | 2 ++ 4 files changed, 65 insertions(+), 7 deletions(-) -- 2.43.0

2 5

[PATCH 1/1] qemu: stop silently narrowing the guest CPU during live migration
by Denis V. Lunev 08 Jun '26

08 Jun '26

qemuDomainMakeCPUMigratable() strips features marked added='yes' (in src/cpu_map/x86_*.xml) from the migration cookie when the source CPU was specified as host-model. The intent was libvirt-protocol compat with older destinations; the cost is guest CPU compat, paid silently on every migration. Every Intel x86 CPU model from Westmere through Sapphire Rapids carries 60+ added='yes' features, including vmx-exit-load-perf-global-ctrl and vmx-entry-load-perf-global-ctrl that control the LOAD_IA32_PERF_GLOBAL_CTRL allowed-1 bits of MSR_IA32_VMX_{EXIT,ENTRY}_CTLS. A host-model live migration on any of these models drops those features from the destination's qemu argv. Modern qemu gates the nested VMX capability MSRs on the explicit -cpu list, so the guest's MSR view shifts. Linux distributions that load kvm_intel by default snapshot those MSRs at module-load time and validate every newly online CPU against that snapshot. An automated test that boots a guest, live-migrates it, and hot-plugs additional vCPUs reliably trips kvm_intel: Inconsistent VMCS config on CPU N kvm: enabling virtualization on CPUN failed smpboot: CPU N is now offline the guest agent's online attempt returns -EIO and the hot-plug fails. Live migration must preserve the guest's CPU specification bit for bit -- this is a hard contract, not an optimisation target. Drop the strip. If a destination libvirt does not know a feature in the cookie, its parser rejects the migration with a precise unknown-feature error: operators can upgrade or narrow the source CPU definition. Either is visible; the status quo is not. This effectively reverts 14d3517410 ("qemu: domain: Drop added features from migratable CPU") together with its follow-up aae8a5774b ("qemu: Drop vmx-* from migratable CPU model only when origCPU is set"), and removes the now-unused origCPU plumbing in qemuDomainMakeCPUMigratable() and its callers. Signed-off-by: Denis V. Lunev <den(a)openvz.org> --- src/qemu/qemu_domain.c | 46 ++------------------------------ src/qemu/qemu_domain.h | 3 +-- src/qemu/qemu_migration_cookie.c | 5 +--- 3 files changed, 4 insertions(+), 50 deletions(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 84c8645259..eccd279bbe 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -5307,36 +5307,10 @@ qemuDomainDefCopy(virQEMUDriver *driver, } -typedef struct { - const char * const *added; - GStrv keep; -} qemuDomainDropAddedCPUFeaturesData; - - -static bool -qemuDomainDropAddedCPUFeatures(const char *name, - virCPUFeaturePolicy policy G_GNUC_UNUSED, - void *opaque) -{ - qemuDomainDropAddedCPUFeaturesData *data = opaque; - - if (!g_strv_contains(data->added, name)) - return true; - - if (data->keep && g_strv_contains((const char **) data->keep, name)) - return true; - - return false; -} - - int qemuDomainMakeCPUMigratable(virArch arch, - virCPUDef *cpu, - virCPUDef *origCPU) + virCPUDef *cpu) { - qemuDomainDropAddedCPUFeaturesData data = { 0 }; - if (cpu->mode != VIR_CPU_MODE_CUSTOM || !cpu->model || !ARCH_IS_X86(arch)) @@ -5353,22 +5327,6 @@ qemuDomainMakeCPUMigratable(virArch arch, virCPUDefUpdateFeature(cpu, "pconfig", VIR_CPU_FEATURE_DISABLE); } - if (origCPU) { - if (virCPUx86GetAddedFeatures(cpu->model, &data.added) < 0) - return -1; - - /* Drop features marked as added in a cpu model, but only - * when they are not mentioned in origCPU, i.e., when they were not - * explicitly mentioned by the user. - */ - if (data.added) { - g_auto(GStrv) keep = virCPUDefListExplicitFeatures(origCPU); - data.keep = keep; - - virCPUDefFilterFeatures(cpu, qemuDomainDropAddedCPUFeatures, &data); - } - } - return 0; } @@ -5569,7 +5527,7 @@ qemuDomainDefFormatBufInternal(virQEMUDriver *driver, } if (def->cpu && - qemuDomainMakeCPUMigratable(def->os.arch, def->cpu, origCPU) < 0) + qemuDomainMakeCPUMigratable(def->os.arch, def->cpu) < 0) return -1; /* Old libvirt doesn't understand <audio> elements so diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index b321a64e96..0600af8b7c 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -1066,8 +1066,7 @@ qemuDomainValidateActualNetDef(const virDomainNetDef *net, int qemuDomainMakeCPUMigratable(virArch arch, - virCPUDef *cpu, - virCPUDef *origCPU); + virCPUDef *cpu); int qemuDomainInitializePflashStorageSource(virDomainObj *vm, diff --git a/src/qemu/qemu_migration_cookie.c b/src/qemu/qemu_migration_cookie.c index 7311a8294b..ebc2072ab0 100644 --- a/src/qemu/qemu_migration_cookie.c +++ b/src/qemu/qemu_migration_cookie.c @@ -535,15 +535,12 @@ static int qemuMigrationCookieAddCPU(qemuMigrationCookie *mig, virDomainObj *vm) { - qemuDomainObjPrivate *priv = vm->privateData; - if (mig->cpu) return 0; mig->cpu = virCPUDefCopy(vm->def->cpu); - if (qemuDomainMakeCPUMigratable(vm->def->os.arch, mig->cpu, - priv->origCPU) < 0) + if (qemuDomainMakeCPUMigratable(vm->def->os.arch, mig->cpu) < 0) return -1; mig->flags |= QEMU_MIGRATION_COOKIE_CPU; -- 2.51.0

3 2