- Devel - libvirt List Archives

[PATCH v3 0/3] qemu: Implement support for multiple device-pluggable SMMUv3s
by Nathan Chen 20 Nov '25

20 Nov '25

Hi, This is a follow up to the second patchset [0] for supporting multiple device-pluggable vSMMU instances. (re-sending this series for review) This patchset implements support for specifying multiple <iommu> devices within the VM definition when smmuv3 IOMMU model is specified with a pciBus driver attribute, and is tested with Shameer's recently accepted qemu series for user-creatable vSMMU devices [1] For instance, specifying the associated emulated device in a VM definition with multiple IOMMUs, configured to be routed to pcie-expander-bus or pcie-root controllers: <devices> <controller type='pci' index='0' model='pcie-root'/> <controller type='pci' index='1' model='pcie-expander-bus'> <model name='pxb-pcie'/> <target busNr='252'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x0'/> </controller> <controller type='pci' index='2' model='pcie-root-port'> <model name='pcie-root-port'/> <target chassis='2' port='0x0'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x0'/> </controller> ... <rng model='virtio'> <backend model='random'>/dev/urandom</backend> <address type='pci' domain='0x0000' bus='0x02' slot='0x00' function='0x0'/> </rng> <rng model='virtio'> <backend model='random'>/dev/urandom</backend> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </rng> <iommu model='smmuv3'> <driver pciBus='0'/> </iommu> <iommu model='smmuv3'> <driver pciBus='1'/> </iommu> </devices> This would get translated to a qemu command line with the arguments below. -device '{"driver":"pxb-pcie","bus_nr":252,"id":"pci.1","bus":"pcie.0","addr":"0x1"}' \ -device '{"driver":"pcie-root-port","port":0,"chassis":2,"id":"pci.2","bus":"pci.1","addr":"0x0"}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pcie.0","id":"smmuv3.0"}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pci.1","id":"smmuv3.1"}' \ ... -object '{"qom-type":"rng-random","id":"objrng0","filename":"/dev/urandom"}' \ -device '{"driver":"virtio-rng-pci","rng":"objrng0","id":"rng0","bus":"pci.2","addr":"0x0"}' \ -object '{"qom-type":"rng-random","id":"objrng1","filename":"/dev/urandom"}' \ -device '{"driver":"virtio-rng-pci","rng":"objrng1","id":"rng1","bus":"pcie.0","addr":"0x4"}' \ Changes from v2: - Make multi-iommu adoption the first commit - Consolidate smmuv3Dev into smmuv3 model - Change 'parentIdx' attribute name to 'pciBus' - Change naming of virDomainDef 'iommu' attribute to 'iommus' - Move multi-iommu check to validation code - Move qemu dev props function to device-pluggable smmuv3 commit - Fix conditional statements per feedback Changes from v1: - Moved parentIdx attribute under <driver> - Revised qemuxmlconfdata test cases to attach emulated devices This series is on Github: https://github.com/NathanChenNVIDIA/libvirt/tree/nestedsmmuv3-no-accel-10-2… Thanks, Nathan [0] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/IF6O… [1] https://lore.kernel.org/qemu-devel/20250829082543.7680-1-skolothumtho@nvidi… Nathan Chen (3): conf: Support multiple device-pluggable smmuv3 IOMMUs qemu: Implement pluggable-device smmuv3 tests: qemuxmlconfdata: provide device-pluggable smmuv3 sample XML and CLI args docs/formatdomain.rst | 4 + src/conf/domain_conf.c | 100 +++++++-- src/conf/domain_conf.h | 10 +- src/conf/domain_validate.c | 61 ++++-- src/conf/schemas/domaincommon.rng | 9 +- src/libvirt_private.syms | 2 + src/qemu/qemu_alias.c | 15 +- src/qemu/qemu_command.c | 197 ++++++++++++------ src/qemu/qemu_domain.c | 2 +- src/qemu/qemu_domain_address.c | 5 +- src/qemu/qemu_driver.c | 8 +- src/qemu/qemu_postparse.c | 11 +- src/qemu/qemu_validate.c | 2 +- .../iommu-smmuv3-pci-bus.aarch64-latest.args | 41 ++++ .../iommu-smmuv3-pci-bus.aarch64-latest.xml | 62 ++++++ .../qemuxmlconfdata/iommu-smmuv3-pci-bus.xml | 49 +++++ tests/qemuxmlconftest.c | 1 + 17 files changed, 455 insertions(+), 124 deletions(-) create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus.aarch64-latest.args create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus.aarch64-latest.xml create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus.xml -- 2.43.0

2 7

[PATCH v3 0/7] USB hostdev: allow addressing by port
by Maximilian Martin 20 Nov '25

20 Nov '25

Currently, only vendor/product and bus/device matching are supported for USB host devices. Neither of these provide a stable and persistent way of assigning a guest a specific host device. Vendor/product can be ambiguous. Device numbers change on every enumeration. This patch adds a bus/port matching, which allows a specific port on the host to be specified using the dotted notation found in Linux's "devpath" sysfs attribute. The path series is based on the previous work of Thomas Hebb: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/message/7U3… This resubmission includes Daniel's patch which extends the USB hostdev test logic. Maximilian Martin (7): tests: validate an XML config with USB vendor/product set virusb test data: add devpath files for port addressing domain_conf, virhostdev, virusb, virusb test: add bus/port matching schema: add USB port attribute tests: validate an XML config with USB bus/port set nodedev: add USB port to nodedev XML docs: add description for USB port matching docs/formatdomain.rst | 29 ++-- src/conf/domain_conf.c | 58 ++++++- src/conf/domain_conf.h | 1 + src/conf/node_device_conf.c | 5 + src/conf/node_device_conf.h | 1 + src/conf/schemas/basictypes.rng | 31 ++++ src/conf/schemas/domaincommon.rng | 41 +---- src/conf/schemas/nodedev.rng | 3 + src/hypervisor/virhostdev.c | 131 ++++++++------ src/libvirt_private.syms | 2 - src/node_device/node_device_udev.c | 4 + src/util/virusb.c | 160 ++++++------------ src/util/virusb.h | 22 +-- .../usb_device_1d6b_1_0000_00_1d_0.xml | 1 + ...ostdev-usb-address-port.x86_64-latest.args | 36 ++++ ...hostdev-usb-address-port.x86_64-latest.xml | 45 +++++ .../hostdev-usb-address-port.xml | 28 +++ ...tdev-usb-vendor-product.x86_64-latest.args | 35 ++++ ...stdev-usb-vendor-product.x86_64-latest.xml | 44 +++++ .../hostdev-usb-vendor-product.xml | 36 ++++ tests/qemuxmlconftest.c | 25 +++ tests/virusbtest.c | 149 +++++++++++----- .../sys_bus_usb/devices/1-1.5.3.1/devpath | 1 + .../sys_bus_usb/devices/1-1.5.3.3/devpath | 1 + .../sys_bus_usb/devices/1-1.5.3/devpath | 1 + .../sys_bus_usb/devices/1-1.5.4/devpath | 1 + .../sys_bus_usb/devices/1-1.5.5/devpath | 1 + .../sys_bus_usb/devices/1-1.5.6/devpath | 1 + .../sys_bus_usb/devices/1-1.5/devpath | 1 + .../sys_bus_usb/devices/1-1.6/devpath | 1 + .../sys_bus_usb/devices/1-1/devpath | 1 + .../sys_bus_usb/devices/2-1.2/devpath | 1 + .../sys_bus_usb/devices/2-1/devpath | 1 + .../sys_bus_usb/devices/usb1/devpath | 1 + .../sys_bus_usb/devices/usb2/devpath | 1 + .../sys_bus_usb/devices/usb3/devpath | 1 + .../sys_bus_usb/devices/usb4/devpath | 1 + 37 files changed, 631 insertions(+), 271 deletions(-) create mode 100644 tests/qemuxmlconfdata/hostdev-usb-address-port.x86_64-latest.args create mode 100644 tests/qemuxmlconfdata/hostdev-usb-address-port.x86_64-latest.xml create mode 100644 tests/qemuxmlconfdata/hostdev-usb-address-port.xml create mode 100644 tests/qemuxmlconfdata/hostdev-usb-vendor-product.x86_64-latest.args create mode 100644 tests/qemuxmlconfdata/hostdev-usb-vendor-product.x86_64-latest.xml create mode 100644 tests/qemuxmlconfdata/hostdev-usb-vendor-product.xml create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.3.1/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.3.3/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.3/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.4/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.5/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5.6/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.5/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1.6/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/1-1/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/2-1.2/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/2-1/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/usb1/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/usb2/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/usb3/devpath create mode 100644 tests/virusbtestdata/sys_bus_usb/devices/usb4/devpath -- 2.39.5

2 11

RFC Gitlab #692
by Wesley Hershberger 20 Nov '25

20 Nov '25

Hi all, I'm working on Gitlab #692 [1] and would appreciate a quick discussion/review/sanity check of the direction I'm planning to take before I invest many days in a solution that misses some major consideration. I'll provide a summary of the issue here and then lay out the two approaches I'm considering. The AppArmor sVirt driver creates & loads a dynamic AppArmor profile by passing a domain's XML description to a separate program `virt-aa-helper` (see [2] for more details). When a device is hotplugged to a domain, the `virt-aa-helper` is called with a flag (`-F`) which appends a rule for that device's backing resource to the profile. The profile is reloaded from the XML description whenever a device is hotunplugged. However, the XML description for some devices does not contain enough information to create the corresponding AppArmor rule. When a profile containing a dynamic rule is reloaded, the dynamic rule is lost, and the device fails (this has been observed with macvtap devices). A similar situation occurs when temporary write access to snapshot layers is needed during block-commit; the `virt-aa-helper` is called with `-F` to add the temporary access and the profile is reloaded at the end of the operation to remove it again. However, this can have the unintended side-effect of wiping out other temporary accesses (for concurrent blockcommits) or removing access for devices added with `-F` (macvtap). A solution must: - Be persistent across libvirtd restarts - Correctly synchronize concurrent profile modifications We've conceived of two broad approaches to this issue: 1. Add fields to the domain XML schema (and persist them) so that the profile can be correctly generated using only the domain's XML description. 2. Create a separate XML file `/etc/apparmor.d/libvirt/libvirt-UUID-dynamic.xml` with only the information needed to produce the dynamic rules; access would be synchronized by a lock in libvirtd. `virt-aa-helper` would consume this file every time it runs; the file would be removed on Domain shutdown. Option 1 (Add fields to Domain schema) Concerns: ================================================ - User-visible schema changes - Do additional fields make sense in the domain schema? e.g. - a macvtap device path `/dev/tapX` - permissions info in `backingStore` elements - It may be non-trivial to chase down every device that could be affected by this issue and plumb through the additional fields (i.e. a complete solution may not be possible) Option 2 (Separate AppArmor XML) Concerns: ========================================== - User-visible additional file - Not generalizable for (future) sVirt drivers that depend on synchronizing a global resource - May require adding remove/restore hooks for the following sVirt interfaces: - virSecurityDomainSetImageFDLabel domainSetSecurityImageFDLabel; - virSecurityDomainSetTapFDLabel domainSetSecurityTapFDLabel; I'd appreciate any input you'd be able to provide regarding the above trade-offs; specifically looking for guidance on what you would consider merging (in contrast to previous approaches [3]). Thanks very much for taking the time to review this, ~Wesley Hershberger Canonical Support [1] https://gitlab.com/libvirt/libvirt/-/issues/692 [2] https://gitlab.com/apparmor/apparmor/-/wikis/Libvirt [3] https://www.mail-archive.com/devel@lists.libvirt.org/msg07703.html

2 1

[PATCH 0/8] qemu: Add option to preserve shutdown VM during backup job and job/reconnection fixes
by Peter Krempa 19 Nov '25

19 Nov '25

This series first fixes few job handling, reconnection and backup job bugs and then implements an option to keep the VM process around when the guest OS shuts down during backup, so that the backup doesn't need to be restarted. Peter Krempa (8): virDomainNestedJobAllowed: Allow VIR_JOB_MODIFY_MIGRATION_SAFE if VIR_JOB_MODIFY is allowed qemuProcessReconnect: Continue reconnection if VM untergoes fake-reboot qemu: backup: Don't attempt to stop the NBD server twice qemuBlockJobProcessEventConcludedBackup: Notify the backup job later lib: Introduce VIR_DOMAIN_EVENT_SUSPENDED_GUEST_SHUTDOWN event reason lib: Introduce VIR_DOMAIN_BACKUP_BEGIN_PRESERVE_SHUTDOWN_DOMAIN flag qemu: backup: Add support for VIR_DOMAIN_BACKUP_BEGIN_PRESERVE_SHUTDOWN_DOMAIN kbase: Add note about preserving VM on shutdown to backup article docs/kbase/live_full_disk_backup.rst | 18 +++++ docs/manpages/virsh.rst | 6 ++ examples/c/misc/event-test.c | 3 + include/libvirt/libvirt-domain.h | 7 +- src/conf/backup_conf.h | 4 ++ src/conf/virdomainjob.c | 1 + src/libvirt-domain.c | 5 ++ src/qemu/qemu_backup.c | 61 +++++++++++++--- src/qemu/qemu_backup.h | 4 ++ src/qemu/qemu_blockjob.c | 7 +- src/qemu/qemu_driver.c | 2 +- src/qemu/qemu_process.c | 101 +++++++++++++++++++++++++-- src/qemu/qemu_process.h | 3 +- tools/virsh-backup.c | 7 ++ tools/virsh-domain-event.c | 3 +- 15 files changed, 210 insertions(+), 22 deletions(-) -- 2.51.1

2 11

[PATCH] docs: domain: add watchdog notification support
by Massimiliano Minella via B4 Relay 19 Nov '25

19 Nov '25

From: Massimiliano Minella <massimiliano.minella(a)se.com> Since v0.8.0 a watchdog notification is available under event ID VIR_DOMAIN_EVENT_ID_WATCHDOG, update the documentation to remove the previous limitation. Signed-off-by: Massimiliano Minella <massimiliano.minella(a)se.com> --- Remove limitation on notification support for watchdog. --- docs/formatdomain.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git docs/formatdomain.rst docs/formatdomain.rst index f50dce477f..d70022a239 100644 --- docs/formatdomain.rst +++ docs/formatdomain.rst @@ -8350,8 +8350,8 @@ The watchdog device requires an additional driver and management daemon in the guest. Just enabling the watchdog in the libvirt configuration does not do anything useful on its own. -Currently libvirt does not support notification when the watchdog fires. This -feature is planned for a future version of libvirt. +:since:`Since 0.8.0`, a notification is available when the watchdog fires, using +the event ID ``VIR_DOMAIN_EVENT_ID_WATCHDOG``. Having multiple watchdogs is usually not something very common, but be aware that this might happen, for example, when an implicit watchdog device is added --- base-commit: eb99bbaf2e5b1e91f4999c7a1ccc90d67dfbe471 change-id: 20251003-update_wd_doc-5aa9b43be09a Best regards, -- Massimiliano Minella <massimiliano.minella(a)se.com>

2 1

RE: [PATCH] docs: domain: add watchdog notification support
by Massimiliano MINELLA 19 Nov '25

19 Nov '25

Hello, do you have any feedback on this patch? Best regards, Massimiliano General -----Original Message----- From: Massimiliano Minella via B4 Relay <devnull+massimiliano.minella.se.com(a)kernel.org> Sent: vendredi 3 octobre 2025 10:45 To: devel(a)lists.libvirt.org Cc: Massimiliano MINELLA <massimiliano.minella(a)se.com> Subject: [PATCH] docs: domain: add watchdog notification support [External email: Use caution with links and attachments] ________________________________ From: Massimiliano Minella <massimiliano.minella(a)se.com> Since v0.8.0 a watchdog notification is available under event ID VIR_DOMAIN_EVENT_ID_WATCHDOG, update the documentation to remove the previous limitation. Signed-off-by: Massimiliano Minella <massimiliano.minella(a)se.com> --- Remove limitation on notification support for watchdog. --- docs/formatdomain.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git docs/formatdomain.rst docs/formatdomain.rst index f50dce477f..c99560a3bd 100644 --- docs/formatdomain.rst +++ docs/formatdomain.rst @@ -8350,8 +8350,8 @@ The watchdog device requires an additional driver and management daemon in the guest. Just enabling the watchdog in the libvirt configuration does not do anything useful on its own. -Currently libvirt does not support notification when the watchdog fires. This -feature is planned for a future version of libvirt. +:since:`Since 0.8.0` a notification when the watchdog fires is +available using the event ID ``VIR_DOMAIN_EVENT_ID_WATCHDOG``. Having multiple watchdogs is usually not something very common, but be aware that this might happen, for example, when an implicit watchdog device is added --- base-commit: eb99bbaf2e5b1e91f4999c7a1ccc90d67dfbe471 change-id: 20251003-update_wd_doc-5aa9b43be09a Best regards, -- Massimiliano Minella <massimiliano.minella(a)se.com>

1 0

[PATCH 0/2] Fix detect_zeroes setup in qemu and improve tests
by Peter Krempa 19 Nov '25

19 Nov '25

See 2/2 for explanation. Peter Krempa (2): qemuxmlconftest: Improve coverage of 'disk-detect-zeroes' test case qemuDomainPrepareDiskSourceData: Setup 'detect_zeroes' for all layers src/qemu/qemu_domain.c | 6 +--- .../xml2json/file-backing_basic-detect.json | 3 ++ .../file-backing_basic-unmap-detect.json | 3 ++ .../file-backing_basic-unmap-ignore.json | 3 ++ .../disk-detect-zeroes.x86_64-latest.args | 20 +++++++++--- tests/qemuxmlconfdata/disk-detect-zeroes.xml | 31 +++++++++++++++++-- 6 files changed, 55 insertions(+), 11 deletions(-) -- 2.51.1

3 4

Plans for 11.10.0 release (freeze on 2025-11-25)
by Jiri Denemark 18 Nov '25

18 Nov '25

We are getting close to 11.10.0 release of libvirt. To aim for the release on Monday 01 Dec I suggest entering the freeze on Tuesday 25 Nov and tagging RC2 on Thursday 27 Nov. I hope this works for everyone. Jirka

1 0

[PATCH-for-10.2 v3 0/2] docs: Update around MIPS/PPC host removal
by Philippe Mathieu-Daudé 18 Nov '25

18 Nov '25

Since v2: - Generic 32-bit host removal message (danpb) Since v1: - 11.0 -> 10.2 v0 Cover: - Clarify MIPS host removal - Mention PPC host was removed Philippe Mathieu-Daudé (2): docs: Correct release of MIPS deprecations / removals docs: Mention 32-bit PPC host as removed docs/about/deprecated.rst | 2 +- docs/about/removed-features.rst | 7 ++++++- 2 files changed, 7 insertions(+), 2 deletions(-) -- 2.51.0

4 6

libvirt serial console on non-x86 platforms
by Leigh Brown 18 Nov '25

18 Nov '25

Hi Daniel, On a whim, I tried installing libvirt on a Raspberry Pi 5. I was pleasantly surprised to find that it all worked, apart from one small wrinkle. I typically use virt-install and set up a serial console, but this failed with the error: ERROR unsupported configuration: BIOS serial console only supported on x86 architectures I created the VM with a graphical console then manually adjusted the config afterwards to use a serial console and it all worked fine. After a bit of digging I saw an email[1] from yourself talking about how graphics=off has semantic effects beyond just controlling whether the firmware prints to the serial or not. After a bit more digging, I found out that Qemu 8.0[2] removed the sga device due to SeaBIOS 1.11.0 and newer supporting this feature natively when QEMU is started with the option -M graphics=off. I made the following small patch against the Debian package: --- libvirt-11.3.0.orig/src/qemu/qemu_validate.c +++ libvirt-11.3.0/src/qemu/qemu_validate.c @@ -1304,15 +1304,7 @@ qemuValidateDomainDef(const virDomainDef /* On x86 -machine graphics=off toggles the use of the * serial console in SeaBIOS (and theoretically other * firmwares). - * On non-x86, it has also sorts of other effects - * on QEMU device models created and so we don't - * want to allow its use. */ - if (!ARCH_IS_X86(def->os.arch)) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", - _("BIOS serial console only supported on x86 architectures")); - return -1; - } if (!def->nserials) { virReportError(VIR_ERR_XML_ERROR, "%s", _("need at least one serial port to use BIOS serial output")); With that change, I can now create VMs with a serial console through virt-install on the Raspberry PI 5. Have I got this right? If so, I'm happy to submit the patch. Regards, Leigh. -- [1] https://mail.gnu.org/archive/html/qemu-devel/2021-09/msg02417.html [2] https://wiki.qemu.org/ChangeLog/8.0

2 4