- Devel - libvirt List Archives

[PATCH] qemuMigrationSrcBeginPhase: Don't call 'qemuBlockNodesEnsureActive' with offline VM
by Peter Krempa 18 Mar '26

18 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> Commits 7b5566ce67b18a and f879d5f40385358 ( v11.8.0-92-gf879d5f403 ) moved around code for re-activating block backends after migration. While previously it was done when migration failed now we do it when we need qemu to do some block operations. Commit 7b5566ce67b18a introduced a bug where 'qemuBlockNodesEnsureActive' is called also when the VM is offline (for migration with 'VIR_MIGRATE_OFFLINE'). That doesn't work though as attempting to enter the monitor of an inactive VM returns failure. Address it by calling 'qemuBlockNodesEnsureActive' only on active VMs during migration. Fixes: 7b5566ce67b18a2bebe68fdb07e046f25185f8d3 Resolves: https://redhat.atlassian.net/browse/RHEL-156800 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- src/qemu/qemu_migration.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index fec808ccfb..3141018a78 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -2870,7 +2870,8 @@ qemuMigrationSrcBeginPhase(virQEMUDriver *driver, vm->newDef && !qemuDomainVcpuHotplugIsInOrder(vm->newDef))) cookieFlags |= QEMU_MIGRATION_COOKIE_CPU_HOTPLUG; - if (qemuBlockNodesEnsureActive(vm, vm->job->asyncJob) < 0) + if (virDomainObjIsActive(vm) && + qemuBlockNodesEnsureActive(vm, vm->job->asyncJob) < 0) return NULL; return qemuMigrationSrcBeginXML(vm, xmlin, -- 2.53.0

2 1

Plans for 12.2.0 release (freeze on 2026-03-26)
by Jiri Denemark 18 Mar '26

18 Mar '26

We are getting close to 12.2.0 release of libvirt. To aim for the release on Wednesday 01 Apr I suggest entering the freeze on Thursday 26 Mar and tagging RC2 on Monday 30 Mar. I hope this works for everyone. Jirka

1 0

[PATCH] virt-aa-helper : grant access to unix socket for qgsd
by Hector Cao 18 Mar '26

18 Mar '26

For quote generation and remote attestation, Intel TDX VM communicates with the QGSD (Quote Generation Service) on the host via unix socket. The unix socket can be specified via: <launchSecurity type='tdx'> ... <quoteGenerationService path='/var/run/tdx-qgs/qgs.socket'/> </launchSecurity> in case the path field is omitted, a default path is used. This commit generates the needed rule in the dynamic AA profile for the VM. Signed-off-by: Hector Cao <hector.cao(a)canonical.com> --- src/security/virt-aa-helper.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index e932e79dab..53059ad5fa 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -1397,6 +1397,19 @@ get_files(vahControl * ctl) break; case VIR_DOMAIN_LAUNCH_SECURITY_PV: case VIR_DOMAIN_LAUNCH_SECURITY_TDX: + /* To communicate with the QGSD daemon running on the host, if we use the unix socket, + * allow the VM to access to the unix socket. + */ + if (ctl->def->sec->data.tdx.haveQGS) { + const char *qgsd_sock_file = QGS_UNIX_SOCKET_FILE; + if (ctl->def->sec->data.tdx.qgs_unix_path) { + qgsd_sock_file = ctl->def->sec->data.tdx.qgs_unix_path; + } + if (vah_add_file(&buf, qgsd_sock_file, "rw") != 0) { + return -1; + } + } + break; case VIR_DOMAIN_LAUNCH_SECURITY_NONE: case VIR_DOMAIN_LAUNCH_SECURITY_LAST: break; -- 2.43.0

3 3

[PATCH v3 0/1] virt-aa-helper : grant access to unix socket for qgsd
by Hector Cao 18 Mar '26

18 Mar '26

Thanks Daniel for the review, I fixed the indent in the previous mail but realized afterwards that I need to add an other improvement to the patch, even though functionally, the previous patch works. The change is about moving the VIR_DOMAIN_LAUNCH_SECURITY_PV case after VIR_DOMAIN_LAUNCH_SECURITY_TDX block. Hector Cao (1): virt-aa-helper : grant access to unix socket for qgsd src/security/virt-aa-helper.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) -- 2.43.0

1 1

[PATCH 0/5] qemuhotplug: Introduce interface-network-hostdev
by Michal Privoznik 17 Mar '26

17 Mar '26

*** BLURB HERE *** Michal Prívozník (5): tests: Drop WITH_QEMU from qemu specific tests binaries/libraries qemuxmlconftest: Separate fake drivers into a separate file virnetworkportxml2xmldata: Use different PCI address in plug-hostdev-pci.xml qemuhotplugtest: Use fake drivers qemuhotplug: Introduce interface-network-hostdev tests/meson.build | 5 +- tests/qemudomaincheckpointxml2xmltest.c | 31 +- tests/qemudomainsnapshotxml2xmltest.c | 29 +- tests/qemufakedrivers.c | 368 +++++++++++++++ tests/qemufakedrivers.h | 19 + tests/qemuhotplugtest.c | 20 + .../qemuhotplug-interface-network-hostdev.xml | 4 + ...ug-base-live+interface-network-hostdev.xml | 63 +++ tests/qemumemlocktest.c | 25 +- ...-virtio-teaming-network.x86_64-latest.args | 2 +- tests/qemuxmlconftest.c | 443 +++--------------- tests/testutilsqemu.c | 43 +- tests/testutilsqemu.h | 23 +- .../plug-hostdev-pci.xml | 2 +- 14 files changed, 588 insertions(+), 489 deletions(-) create mode 100644 tests/qemufakedrivers.c create mode 100644 tests/qemufakedrivers.h create mode 100644 tests/qemuhotplugtestdevices/qemuhotplug-interface-network-hostdev.xml create mode 100644 tests/qemuhotplugtestdomains/qemuhotplug-base-live+interface-network-hostdev.xml -- 2.52.0

2 7

[PATCH] network: Don't enable ip_forward for VIR_NETWORK_FORWARD_OPEN
by Michal Privoznik 17 Mar '26

17 Mar '26

From: Michal Privoznik <mprivozn(a)redhat.com> For a network that's <forward mode="open"/> there are no firewall rules added. We should not assume that users will configure NAT, and if they do it should be their responsibility to enable IP forwarding too. Resolves: https://gitlab.com/libvirt/libvirt/-/work_items/863 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/network/bridge_driver.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index d50d42c98c..ecfce5d9a4 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -2080,8 +2080,9 @@ networkStartNetworkVirtual(virNetworkDriverState *driver, } } - /* If forward.type != NONE, turn on global IP forwarding */ - if (def->forward.type != VIR_NETWORK_FORWARD_NONE) { + /* If forward.type != NONE and != OPEN, turn on global IP forwarding */ + if (def->forward.type != VIR_NETWORK_FORWARD_NONE && + def->forward.type != VIR_NETWORK_FORWARD_OPEN) { if (v6present && !virNetDevIPCheckIPv6Forwarding()) goto error; /* Precise error message already provided */ -- 2.52.0

2 1

A few questions about libvirt-java
by Takashi Kajinami 17 Mar '26

17 Mar '26

Hi, I've been working in a downstream project using libvirt-java to manage resources in libvirt from our application written in Java. The library has been working well so far, but I'm interested in contributing further improvements of the library, such as; - Replace finalize by more modern mechanisms such as (Auto)Closeable (finalize is deprecated in recent OpenJDK versions) - Complete migration to maven (to get rid of maven-ant-tasks which was already retired) However some of the planned contributions (especially the ones listed above) have relatively large impact and I'd like to ask a few questions so that so that we can drive the effort more smoothly. 1. Do you think that these are good/reasonable improvements ? 2. May I request a new release ? The last release (0.5.3) was released 4 years ago and several changes such as new features have been merged since then. Creating a release allows us to mave backword incompatible changes with less user impact. 3. Can we bump minimum JDK version ? Current minimum JSK version supported by the library is 1.8.0 but this is quite old ELS. I wonder if we can bump it to a newer version such as OpenJDK 17 or even 21. This is required to replace "finalize". Is there requirement about the platform the library should support ? Any feedback or guidance is very much appreciated ! Thank you, Takashi -- Takashi Kajinami irc: tkajinam github: https://github.com/kajinamit launchpad: https://launchpad.net/~kajinamit

2 3

[PATCH 0/3] Random improvements
by Michal Privoznik 17 Mar '26

17 Mar '26

I've noticed an old branch. I've accumulated these while working on the networkxmlconftest. Michal Prívozník (3): src: Drop NULL check before calling virBufferEscapeString() network: Format <ip/> element using virXMLFormatElement() networkxmlconftest: s/fail/cleanup/ src/conf/domain_conf.c | 17 ++-- src/conf/network_conf.c | 87 ++++++++----------- src/conf/node_device_conf.c | 4 +- src/qemu/qemu_domain.c | 16 ++-- .../nat-network-dns-hosts-modified.xml | 3 +- .../nat-network-dns-more-hosts.xml | 3 +- .../nat-network-dns-srv-record.xml | 12 +-- .../nat-network-dns-srv-records.xml | 12 +-- .../nat-network-dns-txt-modify-ok.xml | 12 +-- .../nat-network-dns-txt-none.xml | 12 +-- .../nat-network-dns-txt-records.xml | 12 +-- .../nat-network-forward-ifaces.xml | 12 +-- .../nat-network-host-updated.xml | 12 +-- .../nat-network-hosts.xml | 12 +-- .../nat-network-no-forward-ifaces.xml | 12 +-- .../nat-network-no-hosts.xml | 3 +- .../nat-network-no-range.xml | 12 +-- .../nat-network-one-host.xml | 12 +-- tests/networkxml2xmlupdateout/nat-network.xml | 12 +-- .../dhcp6-nat-network.expect.xml | 6 +- .../nat-network-dns-forward-plain.expect.xml | 3 +- ...network-dns-forwarder-no-resolv.expect.xml | 3 +- .../nat-network-dns-forwarders.expect.xml | 3 +- .../nat-network-dns-hosts.expect.xml | 3 +- .../nat-network-dns-local-domain.expect.xml | 3 +- ...-network-dns-srv-record-minimal.expect.xml | 12 +-- .../nat-network-dns-srv-record.expect.xml | 12 +-- .../nat-network-dns-srv-records.expect.xml | 12 +-- .../nat-network-dns-txt-record.expect.xml | 12 +-- .../nat-network-explicit-flood.expect.xml | 12 +-- ...nat-network-forward-nat-address.expect.xml | 12 +-- .../nat-network-forward-nat-ipv6.expect.xml | 3 +- ...-network-forward-nat-no-address.expect.xml | 12 +-- .../nat-network-mtu.expect.xml | 12 +-- .../nat-network-name-with-quotes.expect.xml | 12 +-- .../networkxmlconfdata/nat-network.expect.xml | 12 +-- .../open-network.expect.xml | 3 +- .../ptr-domains-auto.expect.xml | 12 +-- .../routed-network-no-dns.expect.xml | 3 +- .../routed-network.expect.xml | 3 +- tests/networkxmlconftest.c | 18 ++-- 41 files changed, 167 insertions(+), 293 deletions(-) -- 2.52.0

2 4

[PATCH v3] qemu: Store tapfd path in domstatus XML
by Wesley Hershberger 17 Mar '26

17 Mar '26

Introduce a read-only `tapfd` element for direct interfaces (macvtap), which contains the path to the backing tapfd for that interface (e.g. `/dev/tapXX`). The element is only included when the domain is being formatted for internal consumption (VIR_DOMAIN_DEF_FORMAT_STATUS) and is not accepted in user-provided XML (!VIR_DOMAIN_DEF_PARSE_INACTIVE). This is used by the AppArmor security driver when re-generating profiles. Partial-Resolves: https://gitlab.com/libvirt/libvirt/-/issues/692 Bug-Ubuntu: https://bugs.launchpad.net/bugs/2126574 Signed-off-by: Wesley Hershberger <wesley.hershberger(a)canonical.com> --- Resending this patch as I've not recieved a response on my previous submission. Fixed the bug URL in the commit message as I missed that feedback item on my last mail. This submission is a partial revision of a previous series with a fix for the macvtap component of gitlab#692 [1][2]. I haven't had bandwidth to resolve the blockcommit component since the complexity there is somewhat higher (and is also lower priority for us). I kept the separate `tapfd` element rather than reusing the existing `backend` element (virDomainNetBackend.tap) to avoid making a user-visible change [3]. I'd be happy to use the existing field instead if you think that would make more sense. I opted not to introduce/modify a security driver API for FD+path as the patch here is sufficient to resolve the bug, but would be willing to do so if that would make this change more palatable. I've opened a MR to libvirt-tck with test cases that demonstrate the bugs that this fixes [4]. apparmor/110-macvtap.t passes with this patch applied. Thanks for the reviews and continued consideration. ~Wesley [1] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/UNNB… [2] https://gitlab.com/libvirt/libvirt/-/issues/692 [3] https://libvirt.org/formatdomain.html#setting-network-backend-specific-opti… [4] https://gitlab.com/libvirt/libvirt-tck/-/merge_requests/73 --- Changes in v3: - Fix buglink in commit message - Link to v2: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/IPEB… Changes in v2: - Drop `virt-aa-helper: Ask for no deny rule...` as it was applied - Drop `qemu: Store blockcommit permissions...` due to unresolved concerns - Pass tapfd path through netdef instead of resolving from fd - Link to v1: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/UNNB… --- src/conf/domain_conf.c | 8 ++++++++ src/conf/domain_conf.h | 1 + src/hypervisor/domain_interface.c | 2 +- src/lxc/lxc_process.c | 1 + src/qemu/qemu_interface.c | 1 + src/security/security_apparmor.c | 1 + src/security/virt-aa-helper.c | 5 +++++ src/util/virnetdevmacvlan.c | 18 +++++++++++------- src/util/virnetdevmacvlan.h | 4 +++- 9 files changed, 32 insertions(+), 9 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 03a05366e1..9714a1e141 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -2971,6 +2971,7 @@ virDomainNetDefFree(virDomainNetDef *def) g_free(def->virtio); g_free(def->coalesce); g_free(def->sourceDev); + g_free(def->tapfdpath); virNetDevIPInfoClear(&def->guestIP); virNetDevIPInfoClear(&def->hostIP); @@ -10634,6 +10635,10 @@ virDomainNetDefParseXML(virDomainXMLOption *xmlopt, return NULL; } + if (!(flags & VIR_DOMAIN_DEF_PARSE_INACTIVE)) { + def->tapfdpath = virXPathString("string(./tapfd/@path)", ctxt); + } + if (virNetworkPortOptionsParseXML(ctxt, &def->isolatedPort) < 0) return NULL; @@ -26004,6 +26009,9 @@ virDomainNetDefFormat(virBuffer *buf, if (def->mtu) virBufferAsprintf(buf, "<mtu size='%u'/>\n", def->mtu); + if (def->tapfdpath && (flags & VIR_DOMAIN_DEF_FORMAT_STATUS)) + virBufferAsprintf(buf, "<tapfd path='%s'/>\n", def->tapfdpath); + virDomainNetDefCoalesceFormatXML(buf, def->coalesce); virDomainDeviceInfoFormat(buf, &def->info, flags | VIR_DOMAIN_DEF_FORMAT_ALLOW_BOOT diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index e63230beec..3bb4de6274 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1212,6 +1212,7 @@ struct _virDomainNetDef { char *downscript; char *domain_name; /* backend domain name */ char *ifname; /* interface name on the host (<target dev='x'/>) */ + char *tapfdpath; /* Path in /dev for macvtap (<tapfd path="/dev/tapXX">) */ virTristateBool managed_tap; virNetDevIPInfo hostIP; char *ifname_guest_actual; diff --git a/src/hypervisor/domain_interface.c b/src/hypervisor/domain_interface.c index 5bc698d272..37e3d453a0 100644 --- a/src/hypervisor/domain_interface.c +++ b/src/hypervisor/domain_interface.c @@ -111,7 +111,7 @@ virDomainInterfaceEthernetConnect(virDomainDef *def, if (virNetDevMacVLanIsMacvtap(net->ifname)) { auditdev = net->ifname; - if (virNetDevMacVLanTapOpen(net->ifname, tapfd, tapfdSize) < 0) + if (virNetDevMacVLanTapOpen(net->ifname, tapfd, tapfdSize, &net->tapfdpath) < 0) goto cleanup; if (virNetDevMacVLanTapSetup(tapfd, tapfdSize, virDomainInterfaceIsVnetCompatModel(net)) < 0) { diff --git a/src/lxc/lxc_process.c b/src/lxc/lxc_process.c index 1bca9e8dae..c731b28871 100644 --- a/src/lxc/lxc_process.c +++ b/src/lxc/lxc_process.c @@ -379,6 +379,7 @@ virLXCProcessSetupInterfaceDirect(virLXCDriver *driver, VIR_NETDEV_VPORT_PROFILE_OP_CREATE, cfg->stateDir, NULL, 0, + &net->tapfdpath, macvlan_create_flags) < 0) return NULL; diff --git a/src/qemu/qemu_interface.c b/src/qemu/qemu_interface.c index 23a23d201a..edc53d53b3 100644 --- a/src/qemu/qemu_interface.c +++ b/src/qemu/qemu_interface.c @@ -81,6 +81,7 @@ qemuInterfaceDirectConnect(virDomainDef *def, &res_ifname, vmop, cfg->stateDir, tapfd, tapfdSize, + &net->tapfdpath, macvlan_create_flags) < 0) goto cleanup; diff --git a/src/security/security_apparmor.c b/src/security/security_apparmor.c index 40f13ec1a5..3ff80e1dc9 100644 --- a/src/security/security_apparmor.c +++ b/src/security/security_apparmor.c @@ -156,6 +156,7 @@ load_profile(virSecurityManager *mgr G_GNUC_UNUSED, if (virDomainDefFormatInternal(def, NULL, &buf, VIR_DOMAIN_DEF_FORMAT_SECURE | + VIR_DOMAIN_DEF_FORMAT_STATUS | VIR_DOMAIN_DEF_FORMAT_VOLUME_TRANSLATED) < 0) return -1; diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index e932e79dab..60e03c2ce8 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -1192,6 +1192,11 @@ get_files(vahControl * ctl) vhu->type) != 0) return -1; } + + if (net->tapfdpath) { + if (vah_add_file(&buf, net->tapfdpath, "rwk") != 0) + return -1; + } } for (i = 0; i < ctl->def->nmems; i++) { diff --git a/src/util/virnetdevmacvlan.c b/src/util/virnetdevmacvlan.c index cde9d70eef..fcf63e08ff 100644 --- a/src/util/virnetdevmacvlan.c +++ b/src/util/virnetdevmacvlan.c @@ -152,24 +152,24 @@ int virNetDevMacVLanDelete(const char *ifname) int virNetDevMacVLanTapOpen(const char *ifname, int *tapfd, - size_t tapfdSize) + size_t tapfdSize, + char **tapname) { int retries = 10; int ret = -1; int ifindex; size_t i = 0; - g_autofree char *tapname = NULL; if (virNetDevGetIndex(ifname, &ifindex) < 0) return -1; - tapname = g_strdup_printf("/dev/tap%d", ifindex); + *tapname = g_strdup_printf("/dev/tap%d", ifindex); for (i = 0; i < tapfdSize; i++) { int fd = -1; while (fd < 0) { - if ((fd = open(tapname, O_RDWR)) >= 0) { + if ((fd = open(*tapname, O_RDWR)) >= 0) { tapfd[i] = fd; } else if (retries-- > 0) { /* may need to wait for udev to be done */ @@ -178,7 +178,7 @@ virNetDevMacVLanTapOpen(const char *ifname, /* However, if haven't succeeded, quit. */ virReportSystemError(errno, _("cannot open macvtap tap device %1$s"), - tapname); + *tapname); goto cleanup; } } @@ -188,6 +188,7 @@ virNetDevMacVLanTapOpen(const char *ifname, cleanup: if (ret < 0) { + g_free(*tapname); while (i--) VIR_FORCE_CLOSE(tapfd[i]); } @@ -659,6 +660,7 @@ virNetDevMacVLanCreateWithVPortProfile(const char *ifnameRequested, char *stateDir, int *tapfd, size_t tapfdSize, + char **tapfdpath, unsigned int flags) { g_autofree char *ifname = NULL; @@ -729,7 +731,7 @@ virNetDevMacVLanCreateWithVPortProfile(const char *ifnameRequested, } if (flags & VIR_NETDEV_MACVLAN_CREATE_WITH_TAP) { - if (virNetDevMacVLanTapOpen(ifname, tapfd, tapfdSize) < 0) + if (virNetDevMacVLanTapOpen(ifname, tapfd, tapfdSize, tapfdpath) < 0) goto disassociate_exit; if (virNetDevMacVLanTapSetup(tapfd, tapfdSize, vnet_hdr) < 0) @@ -888,7 +890,8 @@ int virNetDevMacVLanDelete(const char *ifname G_GNUC_UNUSED) int virNetDevMacVLanTapOpen(const char *ifname G_GNUC_UNUSED, int *tapfd G_GNUC_UNUSED, - size_t tapfdSize G_GNUC_UNUSED) + size_t tapfdSize G_GNUC_UNUSED, + char **tapname G_GNUC_UNUSED) { virReportSystemError(ENOSYS, "%s", _("Cannot create macvlan devices on this platform")); @@ -917,6 +920,7 @@ int virNetDevMacVLanCreateWithVPortProfile(const char *ifname G_GNUC_UNUSED, char *stateDir G_GNUC_UNUSED, int *tapfd G_GNUC_UNUSED, size_t tapfdSize G_GNUC_UNUSED, + char **tapfdpath G_GNUC_UNUSED, unsigned int unused_flags G_GNUC_UNUSED) { virReportSystemError(ENOSYS, "%s", diff --git a/src/util/virnetdevmacvlan.h b/src/util/virnetdevmacvlan.h index 31e4804cdc..7424b87965 100644 --- a/src/util/virnetdevmacvlan.h +++ b/src/util/virnetdevmacvlan.h @@ -72,13 +72,15 @@ int virNetDevMacVLanCreateWithVPortProfile(const char *ifname, char *stateDir, int *tapfd, size_t tapfdSize, + char **tapfdpath, unsigned int flags) ATTRIBUTE_NONNULL(2) ATTRIBUTE_NONNULL(3) ATTRIBUTE_NONNULL(6) ATTRIBUTE_NONNULL(8) ATTRIBUTE_NONNULL(10) G_GNUC_WARN_UNUSED_RESULT; int virNetDevMacVLanTapOpen(const char *ifname, int *tapfd, - size_t tapfdSize) + size_t tapfdSize, + char **tapname) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2) G_GNUC_WARN_UNUSED_RESULT; --- base-commit: 54f5032e576a886cfb8232abbfa08a2f6e9f9962 change-id: 20260105-apparmor-races-d03238ee4d93 Best regards, -- Wesley Hershberger <wesley.hershberger(a)canonical.com>

1 0

[PATCH] virt-aa-helper: Prevent spurious denials for AoE disks
by PUSHKARAJ PATIL 13 Mar '26

13 Mar '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. Allow read access to /dev/etherd/e*.* in the virt-aa-helper profile to avoid these spurious denials. 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..80e9ef2b08 100644 --- a/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in +++ b/src/security/apparmor/usr.lib.libvirt.virt-aa-helper.in @@ -73,6 +73,7 @@ profile virt-aa-helper @libexecdir@/virt-aa-helper { /**.vhd r, /**.[iI][sS][oO] r, /**/disk{,.*} r, + /dev/etherd/e*.* r, include if exists <local/usr.lib.libvirt.virt-aa-helper> } -- 2.50.1 (Apple Git-155)

1 0

[PATCH 0/4] libvirt: AMD IOMMU fixes
by Xiaotian Feng 13 Mar '26

13 Mar '26

This series refines libvirt's logic for large vCPU configurations (>255) on x86/q35 machine types. Historically, the threshold was tied to Intel's EIM capability, but the real requirement is x2APIC support for APIC IDs beyond 255. AMD platforms can satisfy this via xtsup on amd-iommu. Summary of changes: 1. Fix virDomainIOMMUDefEquals to include pt and xtsup attributes for proper checks. 2. Correct error messages for passthrough (pt) and xtsup attributes to reflect their actual names instead of "dma translation". 3. Rename QEMU_MAX_VCPUS_WITHOUT_EIM to QEMU_MAX_VCPUS_WITHOUT_X2APIC and update validation logic to accept either: - intel-iommu with eim='on', or - amd-iommu with xtsup='on' for guests with more than 255 vCPUs. Error messages now mention x2APIC mode instead of extended interrupt mode. 4. Add QEMU_CAPS_AMD_IOMMU_XTSUP capability and enable xtsup by default for AMD IOMMU when a Q35 domain has >255 vCPUs, similar to Intel EIM auto-enable logic. Also ensure intremap is turned on when required. This makes the behavior vendor-neutral and improves usability for AMD EPYC guests with large vCPU counts. No ABI changes beyond stricter equality checks; this is a clarification and extension of existing logic. Xiaotian Feng (4): conf: fix virDomainIOMMUDefEquals for amd_iommu conf: fix error log for passthrough and xtsup attributes conf: support >255 vcpu w/ amd-iommu xtsup qemu: Enable AMD IOMMU XTSUP by default src/conf/domain_conf.c | 6 ++++-- src/qemu/qemu_capabilities.c | 2 ++ src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_postparse.c | 42 +++++++++++++++++++++++------------- src/qemu/qemu_validate.c | 11 +++++----- src/qemu/qemu_validate.h | 2 +- 6 files changed, 41 insertions(+), 23 deletions(-) -- 2.34.1

2 6

[PATCH] storage: mpath: fix error reporting in virStorageBackendGetMaps
by Lucas Amaral 13 Mar '26

13 Mar '26

virStorageBackendGetMaps() returns 1 on failure instead of the conventional -1, and does not call virReportError() in any of its error paths. On top of that, both callers silently discard the return value: virStorageBackendGetMaps() ignores the result of virStorageBackendCreateVols(), and virStorageBackendMpathRefreshPool() ignores the result of virStorageBackendGetMaps(). Fix all three issues: normalize error returns to -1, add virReportError() calls for each failure path in virStorageBackendGetMaps(), and check return values in both callers so that errors are properly propagated. Signed-off-by: Lucas Amaral <lucaaamaral(a)gmail.com> --- This is similar in spirit to commits e9b931d3e4 ("virpci: Report an error if virPCIGetVirtualFunctionIndex() fails") and 6c2c9e21ac ("virstorageobj: Make virStoragePoolObjAddVol() report an error on failure"), applying the same pattern to the mpath storage backend. Found while auditing storage backends for functions that return error codes without calling virReportError(). src/storage/storage_backend_mpath.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/src/storage/storage_backend_mpath.c b/src/storage/storage_backend_mpath.c index 9fc3983c52..a30153a93b 100644 --- a/src/storage/storage_backend_mpath.c +++ b/src/storage/storage_backend_mpath.c @@ -197,19 +197,25 @@ virStorageBackendGetMaps(virStoragePoolObj *pool) struct dm_names *names = NULL; if (!(dmt = dm_task_create(DM_DEVICE_LIST))) { - retval = 1; + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to create devmapper task")); + retval = -1; goto out; } dm_task_no_open_count(dmt); if (!dm_task_run(dmt)) { - retval = 1; + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to run devmapper task")); + retval = -1; goto out; } if (!(names = dm_task_get_names(dmt))) { - retval = 1; + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to get devmapper device names")); + retval = -1; goto out; } @@ -218,7 +224,10 @@ virStorageBackendGetMaps(virStoragePoolObj *pool) goto out; } - virStorageBackendCreateVols(pool, names); + if (virStorageBackendCreateVols(pool, names) < 0) { + retval = -1; + goto out; + } out: if (dmt != NULL) @@ -248,7 +257,8 @@ virStorageBackendMpathRefreshPool(virStoragePoolObj *pool) virWaitForDevices(); - virStorageBackendGetMaps(pool); + if (virStorageBackendGetMaps(pool) < 0) + return -1; return 0; } -- 2.52.0

2 1

[libvirt PATCH] viriommufd: Set IOMMU_OPTION_RLIMIT_MODE only when running privileged
by Pavel Hrdina 13 Mar '26

13 Mar '26

From: Pavel Hrdina <phrdina(a)redhat.com> If libvirt daemon is running unprivileged it will fail so we should not even try to set it. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/qemu/qemu_process.c | 2 +- src/util/viriommufd.c | 6 +++--- src/util/viriommufd.h | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index ab7cf03c0e..ecd05b4bf6 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -7732,7 +7732,7 @@ qemuProcessOpenIommuFd(virDomainObj *vm) VIR_DEBUG("Opening IOMMU FD for domain %s", vm->def->name); - if ((iommufd = virIOMMUFDOpenDevice()) < 0) + if ((iommufd = virIOMMUFDOpenDevice(priv->driver->privileged)) < 0) return -1; if (qemuSecuritySetImageFDLabel(priv->driver->securityManager, vm->def, iommufd) < 0) diff --git a/src/util/viriommufd.c b/src/util/viriommufd.c index b62d59241d..82920923a2 100644 --- a/src/util/viriommufd.c +++ b/src/util/viriommufd.c @@ -80,14 +80,14 @@ virIOMMUFDSetRLimitMode(int fd, bool processAccounting) } int -virIOMMUFDOpenDevice(void) +virIOMMUFDOpenDevice(bool privileged) { int fd = -1; if ((fd = open(VIR_IOMMU_DEV_PATH, O_RDWR | O_CLOEXEC)) < 0) virReportSystemError(errno, "%s", _("cannot open IOMMUFD device")); - if (virIOMMUFDSetRLimitMode(fd, true) < 0) { + if (privileged && virIOMMUFDSetRLimitMode(fd, true) < 0) { VIR_FORCE_CLOSE(fd); return -1; } @@ -98,7 +98,7 @@ virIOMMUFDOpenDevice(void) #else int -virIOMMUFDOpenDevice(void) +virIOMMUFDOpenDevice(bool privileged G_GNUC_UNUSED) { virReportError(VIR_ERR_NO_SUPPORT, "%s", _("IOMMUFD is not supported on this platform")); diff --git a/src/util/viriommufd.h b/src/util/viriommufd.h index 223f44eb5c..7bad5c7472 100644 --- a/src/util/viriommufd.h +++ b/src/util/viriommufd.h @@ -22,6 +22,6 @@ #define VIR_IOMMU_DEV_PATH "/dev/iommu" -int virIOMMUFDOpenDevice(void); +int virIOMMUFDOpenDevice(bool privileged); bool virIOMMUFDSupported(void); -- 2.53.0

2 1

[PATCH 0/5] hyperv: Implement basic snapshot query APIs
by Jonathon Jongsma 13 Mar '26

13 Mar '26

This is the first patch series adding snapshot support for the hyperv driver. It only implements the basic functions for querying information about snapshots and doesn't handle creating or deleting snapshots yet. Jonathon Jongsma (5): hyperv: Implement domainSnapshotLookupByName() hyperv: implement virDomainListAllSnapshots()/virDomainSnapshotNum() hyperv: implement virDomainSnapshotGetXMLDesc() hyperv: Implement virDomainSnapshotCurrent()/virDomainHasCurrentSnapshot() hyperv: Implement virDomainSnapshotGetParent() src/hyperv/hyperv_driver.c | 276 +++++++++++++++++++++++++++++++++++++ src/hyperv/hyperv_wmi.c | 20 +++ src/hyperv/hyperv_wmi.h | 7 + 3 files changed, 303 insertions(+) -- 2.53.0

2 7

virsh feature request: cold restart and/or synchronous shutdown option
by roy-orbitson＠devo.net.au 13 Mar '26

13 Mar '26

When libvirt components get updated on a hypervisor, VMs continue running with obsolete files. Programs like "needrestart" alert one to this condition with messages like: VM guests are running outdated hypervisor (qemu) binaries on this host: 'bazquux' with pid 1234 Normal restarting of a VM (from within or with `virsh reboot ...`) does not resolve it because it's 'warm'; VMs must be completely shut down then started again to utilise the updates. Afaik, to do so reliably, in a way that minimises downtime, requires scripts like this: (d=bazquux; virsh shutdown --domain "$d" && until LC_ALL=C LANG=C virsh domstate --domain "$d" | grep -Fi 'shut off' &> /dev/null; do echo ...; sleep 1; done && virsh start --domain "$d") Adding a timeout complicates that further. Options for this logic to be performed by virsh would be beneficial, as a single operation: virsh reboot --domain bazquux --cold And by allowing poweroff to wait: (d=bazquux; virsh shutdown --domain "$d" --synchronous && virsh start --domain "$d") The latter would also assist with scripting changes to a VM that can only be performed whilst it is not running (shutdown -> alter -> start). Ideally, these would accept `--timeout n` so that if the shutdown portion was not completed quickly enough, they would fail and exit with a non-zero status, and the former would not start the VM again. I'm not sure of a good way to handle this where `on_poweroff` is set to something that doesn't result in the machine being shut off.

5 8

[PATCH v1 0/1] rpc: fix virNetClient->error leak on teardown
by Leander Kohler 12 Mar '26

12 Mar '26

Hi all, this is my first contribution to libvirt upstream. I am with Cyberus Technology, where I work on cloud-hypervisor and its integration with OpenStack and libvirt. The following fixes a leak in virNetClient teardown. virNetClientMarkClose() may save the current error in client->error (via virSaveLastError()). We usually free that in virNetClientCloseLocked(), but we can also end up disposing the client without going through that cleanup. One path where this happens is: virConnectDispose() -> remoteConnectClose() -> doRemoteClose() -> virNetClientClose(priv->client) -> g_clear_pointer(&priv->client, virObjectUnref) In that case, client->error can stay allocated until finalization, and ASan reports a leak. Fix is to also clear client->error in virNetClientDispose() as a final fallback. I included an ASan trace below. Thanks, Leander --- Line numbers in this trace come from our downstream libvirt fork and may differ from upstream, but the call path/function names are the same. [ 276.247727] virtchd[4030]: ==4030==ERROR: LeakSanitizer: detected memory leaks Direct leak of 80 byte(s) in 1 object(s) allocated from: #0 0x7fcd6009e110 in calloc (libasan.so.8+0xfc110) #1 0x7fcd5dcb95bd in g_malloc0 (libglib-2.0.so.0+0x6d5bd) #2 0x7fcd5eb15f31 in virSaveLastError ../src/util/virerror.c:413 #3 0x7fcd5ee6c76c in virNetClientMarkClose ../src/rpc/virnetclient.c:776 #4 0x7fcd5ee6f782 in virNetClientCloseInternal ../src/rpc/virnetclient.c:837 #5 0x7fcd5f0150e4 in doRemoteClose ../src/remote/remote_driver.c:1414 #6 0x7fcd5f0150e4 in remoteConnectClose ../src/remote/remote_driver.c:1442 #7 0x7fcd5f1a90ef in virConnectDispose ../src/datatypes.c:174 #8 0x7fcd5ea58ebb in vir_object_finalize ../src/util/virobject.c:323 #9 0x7fcd5dc05c52 in g_object_unref (libgobject-2.0.so.0+0x1ec52) #10 0x7fcd5ea59fc7 in virObjectUnref ../src/util/virobject.c:377 #11 0x7fcd5ea59fc7 in virObjectUnref ../src/util/virobject.c:370 #12 0x7fcd5ef3a92f in glib_autoptr_clear_virConnect ../src/datatypes.h:548 #13 0x7fcd5ef3a92f in glib_autoptr_cleanup_virConnect ../src/datatypes.h:548 #14 0x7fcd5ef3a92f in virDomainInterfaceDeleteDevice ../src/hypervisor/domain_interface.c:419 #15 0x7fcd594a0cd3 in virCHProcessStopOrKill ../src/ch/ch_process.c:1308 #16 0x7fcd594a8bfa in virCHProcessStop ../src/ch/ch_process.c:1363 #17 0x7fcd5945d8c7 in chDomainDestroyFlags ../src/ch/ch_driver.c:959 #18 0x7fcd5f0f69c0 in virDomainDestroy ../src/libvirt-domain.c:514 #19 0x55c25777376a in remoteDispatchDomainDestroy src/remote/remote_daemon_dispatch_stubs.h:5485 #20 0x55c25777376a in remoteDispatchDomainDestroyHelper src/remote/remote_daemon_dispatch_stubs.h:5464 #21 0x7fcd5ee7c42c in virNetServerProgramDispatchCall ../src/rpc/virnetserverprogram.c:423 #22 0x7fcd5ee7c42c in virNetServerProgramDispatch ../src/rpc/virnetserverprogram.c:299 #23 0x7fcd5ee97c6c in virNetServerProcessMsg ../src/rpc/virnetserver.c:135 #24 0x7fcd5ee98b43 in virNetServerHandleJob ../src/rpc/virnetserver.c:155 #25 0x7fcd5eadbce2 in virThreadPoolWorker ../src/util/virthreadpool.c:164 #26 0x7fcd5eada3c9 in virThreadHelper ../src/util/virthread.c:256 #27 0x7fcd60004cd5 in asan_thread_start(void*) (libasan.so.8+0x62cd5) #28 0x7fcd5d200979 in start_thread (libc.so.6+0x9a979) #29 0x7fcd5d288d2b in __GI___clone3 (libc.so.6+0x122d2b) Objects leaked above: 0x507000064400 (80 bytes) SUMMARY: AddressSanitizer: 80 byte(s) leaked in 1 allocation(s). Leander Kohler (1): rpc: free saved close error in virNetClientDispose src/rpc/virnetclient.c | 2 ++ 1 file changed, 2 insertions(+) -- 2.51.2

2 2

[PATCH] util: implement virHostCPUGetOnlineBitmap() for FreeBSD
by Roman Bogorodskiy 12 Mar '26

12 Mar '26

Implement virHostCPUGetOnlineBitmap() for FreeBSD. As FreeBSD supports neither plugging nor taking CPUs offline, all CPUs are always online. Signed-off-by: Roman Bogorodskiy <bogorodskiy(a)gmail.com> --- src/util/virhostcpu.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/src/util/virhostcpu.c b/src/util/virhostcpu.c index 7cecb4b2c8..8688b6ec67 100644 --- a/src/util/virhostcpu.c +++ b/src/util/virhostcpu.c @@ -1120,6 +1120,19 @@ virHostCPUGetOnlineBitmap(void) virFileReadValueBitmap(&ret, "%s/cpu/online", SYSFS_SYSTEM_PATH); + return ret; +#elif defined(__FreeBSD__) + virBitmap *ret = NULL; + int ncpus = virHostCPUGetCountAppleFreeBSD(); + + if (ncpus == -1) + return ret; + + /* FreeBSD does not support CPU hotplug, so it is safe to assume + * that all available CPUs are online */ + ret = virBitmapNew(ncpus); + virBitmapSetAll(ret); + return ret; #else virReportError(VIR_ERR_NO_SUPPORT, "%s", -- 2.52.0

2 1

[RFC PATCH v2 0/2] qemu: Support vCMDQ
by Nathan Chen 11 Mar '26

11 Mar '26

Hi, This is a follow up to the previous RFC patchset [0] for supporting Tegra241 vCMDQ. This patchset is tested with Shameer's v3 qemu series [1] for vCMDQ support. The accelerated SMMUv3 series for qemu and Libvirt [2] are pre-requisites for vCMDQ support, as accelerated SMMUv3 must be enabled alongside vCMDQ. A new 'cmdqv' <iommu> driver attribute is introduced to toggle enabling vCMDQ support. For instance, specifying hostdevs associated with multiple accelerated SMMUs + CMDQV enabled, configured to be routed to pcie-expander-bus controllers in a way where VFIO device to SMMUv3 associations are matched with the host: <devices> ... <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-expander-bus'> <model name='pxb-pcie'/> <target busNr='248'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </controller> ... <controller type='pci' index='21' model='pcie-root-port'> <model name='pcie-root-port'/> <target chassis='21' port='0x0'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x0'/> </controller> <controller type='pci' index='22' model='pcie-root-port'> <model name='pcie-root-port'/> <target chassis='22' port='0xa8'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x00' function='0x0'/> </controller> ... <hostdev mode='subsystem' type='pci' managed='no'> <source> <address domain='0x0009' bus='0x01' slot='0x00' function='0x0'/> </source> <address type='pci' domain='0x0000' bus='0x15' slot='0x00' function='0x0'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='no'> <source> <address domain='0x0019' bus='0x01' slot='0x00' function='0x0'/> </source> <address type='pci' domain='0x0000' bus='0x16' slot='0x00' function='0x0'/> </hostdev> <iommu model='smmuv3'> <driver pciBus='1' accel='on' ats='on' ril='on' ssidSize='20' oas='44' cmdqv='on'/> </iommu> <iommu model='smmuv3'> <driver pciBus='2' accel='on' ats='on' ril='on' ssidSize='20' oas='44' cmdqv='on'/> </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":"pxb-pcie","bus_nr":248,"id":"pci.2","bus":"pcie.0","addr":"0x2"}' \ -device '{"driver":"pcie-root-port","port":0,"chassis":21,"id":"pci.21","bus":"pci.1","addr":"0x0"}' \ -device '{"driver":"pcie-root-port","port":168,"chassis":22,"id":"pci.22","bus":"pci.2","addr":"0x0"}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pci.1","id":"iommu0","accel":true,"ats":true,"cmdqv":"on","ril":true,"pasid":true,"oas":44}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pci.2","id":"iommu1","accel":true,"ats":true,"cmdqv":"on","ril":true,"pasid":true,"oas":44}' \ -device '{"driver":"vfio-pci","host":"0009:01:00.0","id":"hostdev0","bus":"pci.21","addr":"0x0"}' \ -device '{"driver":"vfio-pci","host":"0019:01:00.0","id":"hostdev1","bus":"pci.22","addr":"0x0"}' \ This series is on Github: https://github.com/NathanChenNVIDIA/libvirt/tree/cmdqv-03-26/ Thanks, Nathan Changes from v1: - Remove 'identifier' attribute for SMMUv3 - Change QEMU command line property from 'tegra241-cmdqv' to 'cmdqv' - Support QEMU OnOffAuto CMDQV property instead of expecting a bool [0] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/NHVK… [1] https://lore.kernel.org/all/20260226105056.897-1-skolothumtho@nvidia.com/ [2] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/5GG7… Nathan Chen (2): qemu: Support NVIDIA Tegra241 CMDQV for SMMUv3 tests: qemuxmlconfdata: provide cmdqv sample XML and CLI args docs/formatdomain.rst | 6 ++ src/conf/domain_conf.c | 15 ++++ src/conf/domain_conf.h | 1 + src/conf/domain_validate.c | 5 +- src/conf/schemas/domaincommon.rng | 5 ++ src/qemu/qemu_command.c | 1 + ...u-smmuv3-pci-bus-cmdqv.aarch64-latest.args | 45 +++++++++++ ...mu-smmuv3-pci-bus-cmdqv.aarch64-latest.xml | 76 +++++++++++++++++++ .../iommu-smmuv3-pci-bus-cmdqv.xml | 66 ++++++++++++++++ tests/qemuxmlconftest.c | 1 + 10 files changed, 220 insertions(+), 1 deletion(-) create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-cmdqv.aarch64-latest.args create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-cmdqv.aarch64-latest.xml create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-cmdqv.xml -- 2.43.0

1 2

[PATCH v7 0/2] qemu: Support accelerated vSMMUs
by Nathan Chen 11 Mar '26

11 Mar '26

Hi, This is a follow up to the sixth RFC patchset [0] for supporting multiple accelerated vSMMU instances. Support for user-creatable accelerated SMMUv3 in QEMU has been applied to target-arm.next [1]. For instance, specifying hostdevs associated with multiple accelerated SMMUs, configured to be routed to pcie-expander-bus controllers in a way where VFIO device to SMMUv3 associations are matched with the host: <devices> ... <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-expander-bus'> <model name='pxb-pcie'/> <target busNr='248'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </controller> ... <controller type='pci' index='21' model='pcie-root-port'> <model name='pcie-root-port'/> <target chassis='21' port='0x0'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x0'/> </controller> <controller type='pci' index='22' model='pcie-root-port'> <model name='pcie-root-port'/> <target chassis='22' port='0xa8'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x00' function='0x0'/> </controller> ... <hostdev mode='subsystem' type='pci' managed='no'> <source> <address domain='0x0009' bus='0x01' slot='0x00' function='0x0'/> </source> <vpasidCapOffset>0xff8</vpasidCapOffset> <address type='pci' domain='0x0000' bus='0x15' slot='0x00' function='0x0'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='no'> <source> <address domain='0x0019' bus='0x01' slot='0x00' function='0x0'/> </source> <vpasidCapOffset>0xff8</vpasidCapOffset> <address type='pci' domain='0x0000' bus='0x16' slot='0x00' function='0x0'/> </hostdev> <iommu model='smmuv3'> <driver parentIdx='1' accel='on' ats='on' ril='off' ssidSize='20' oas='44'/> </iommu> <iommu model='smmuv3'> <driver parentIdx='2' accel='on' ats='on' ril='off' ssidSize='20' oas='44'/> </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":"pxb-pcie","bus_nr":248,"id":"pci.2","bus":"pcie.0","addr":"0x2"}' \ -device '{"driver":"pcie-root-port","port":0,"chassis":21,"id":"pci.21","bus":"pci.1","addr":"0x0"}' \ -device '{"driver":"pcie-root-port","port":168,"chassis":22,"id":"pci.22","bus":"pci.2","addr":"0x0"}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pci.1","id":"iommu0","accel":true,"ats":true,"ril":false,"ssidsize":20,"oas":44}' \ -device '{"driver":"arm-smmuv3","primary-bus":"pci.2","id":"iommu1","accel":true,"ats":true,"ril":false,"ssidsize":20,"oas":44}' \ -device '{"driver":"vfio-pci","host":"0009:01:00.0","id":"hostdev0","x-vpasid-cap-offset":4088,"bus":"pci.21","addr":"0x0"}' \ -device '{"driver":"vfio-pci","host":"0019:01:00.0","id":"hostdev1","x-vpasid-cap-offset":4088,"bus":"pci.22","addr":"0x0"}' \ This series is on Github: https://github.com/NathanChenNVIDIA/libvirt/tree/smmuv3-accel-01-26-v7/ Thanks, Nathan Changes from v6: - Remove smmuv3 'pasid' attribute - Add smmuv3 'ssidsize' attribute - Add support for 'vpasidCapOffset' element - Modify logic for building accel smmuv3 qemu command line to omit ATS, RIL, accel flags when not needed (enabled/disabled by default in QEMU SMMUv3) [0] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/VRSS… [1] https://lore.kernel.org/qemu-devel/20260126104342.253965-1-skolothumtho@nvi… Nathan Chen (2): qemu: Add support for HW-accelerated nested SMMUv3 tests: qemuxmlconfdata: provide HW-accel smmuv3 sample XML and CLI args docs/formatdomain.rst | 36 +++++++ src/conf/domain_conf.c | 101 +++++++++++++++++- src/conf/domain_conf.h | 6 ++ src/conf/domain_validate.c | 36 ++++++- src/conf/schemas/domaincommon.rng | 32 ++++++ src/qemu/qemu_command.c | 23 ++++ src/util/virpci.h | 4 + ...u-smmuv3-pci-bus-accel.aarch64-latest.args | 38 +++++++ ...mu-smmuv3-pci-bus-accel.aarch64-latest.xml | 64 +++++++++++ .../iommu-smmuv3-pci-bus-accel.xml | 54 ++++++++++ tests/qemuxmlconftest.c | 1 + 11 files changed, 390 insertions(+), 5 deletions(-) create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-accel.aarch64-latest.args create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-accel.aarch64-latest.xml create mode 100644 tests/qemuxmlconfdata/iommu-smmuv3-pci-bus-accel.xml -- 2.43.0

2 9

[PATCH 0/4] Add support for expanding CPU features in domain capabilities
by Jiri Denemark 11 Mar '26

11 Mar '26

Consistently with all other CPU definitions, domain capabilities report host-model CPU as a base model and additional list of features. Getting all features supported on a host is possible by passing the host-model CPU definition to hypervisor-cpu-baseline with --features option. But while this is very easy in virsh (domain capabilities can by directly piped to hypervisor-cpu-baseline), doing this via API is more complicated. The relevant part of the domain XML needs to be copied into a new XML document, formatted and passed to virConnectBaselineHypervisorCPU. This series adds the ability to expand CPU features directly in the domain capabilities XML without having to call another API. Jiri Denemark (4): Introduce EXPAND_CPU_FEATURES flag for domain capabilities qemu: Implement VIR_CONNECT_GET_DOMAIN_CAPABILITIES_EXPAND_CPU_FEATURES virsh: Add --expand-cpu-features option for domcapabilities docs: Clarify host-model description in domain capabilities docs/formatdomaincaps.rst | 4 ++++ docs/manpages/virsh.rst | 5 +++++ include/libvirt/libvirt-domain.h | 2 ++ src/libvirt-domain.c | 5 +++++ src/qemu/qemu_driver.c | 9 ++++++++- tools/virsh-host.c | 7 +++++++ 6 files changed, 31 insertions(+), 1 deletion(-) -- 2.53.0

2 5

[PATCH] tests: Create fake root dirs later
by Michal Privoznik 11 Mar '26

11 Mar '26

From: Michal Privoznik <mprivozn(a)redhat.com> In one of previous commits the virTestMain() function was changed to actually create fake HOME, XDG_RUNTIME_DIR, ... directories instead of setting spoofed values in the environment. But alongside with this, the call to virTestFakeRootDirInit() was moved (to location where environment was poisoned). And this would not matter if it wasn't for mocking. Because what we ended up with is virTestFakeRootDirInit() is called and then (optionally) the process re-execs itself (with mocks loaded). This means that previously created root dirs are never cleaned up and just pollute builddir. Therefore, restore original location from which the function was called. Fixes: 79d97d2b4f0b55ea80f8330144953e2b93927e25 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- tests/testutils.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tests/testutils.c b/tests/testutils.c index c984f66d0e..14e5c56fcf 100644 --- a/tests/testutils.c +++ b/tests/testutils.c @@ -859,9 +859,6 @@ int virTestMain(int argc, preloads[npreloads] = NULL; } - if (!(fakerootdir = virTestFakeRootDirInit())) - return EXIT_FAILURE; - va_start(ap, func); while ((lib = va_arg(ap, const char *))) { g_autofree char *abs_lib_path = g_strdup_printf("%s/%s", abs_builddir, lib); @@ -934,6 +931,9 @@ int virTestMain(int argc, failedTests = virBitmapNew(1); + if (!(fakerootdir = virTestFakeRootDirInit())) + return EXIT_FAILURE; + ret = (func)(); virResetLastError(); -- 2.52.0

2 1

[PATCH v2] scripts: dtrace2systemtap: Fix argument name extraction regex to avoid '*' in names
by Peter Krempa 11 Mar '26

11 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> When commit d249170bf609d2c modified the arguments of 'virNetTLSContextNew' which has a systemtap probe defined the 'dtrace2systemtap' script was no longer to correctly generate the definitions for it. First problem which stemmed from mis-detecting the string array argumment as string, which would use 'user_string' to extract it was fixed in cb33103c4afbce681. After that the generated probe is still not correct because it doesn't strip all the '*' from pointers and thus for double pointers it generates the following invalid definition: probe libvirt.rpc.tls_context_new = process("/usr/lib64/libvirt.so").mark("rpc_tls_context_new") { ctxt = $arg1; cacert = user_string($arg2); cacrl = user_string($arg3); *cert = $arg4; *keys = $arg5; sanityCheckCert = $arg6; requireValidCert = $arg7; isServer = $arg8; } Leading to the following failure: # stap -ve 'probe oneshot {exit()}' parse error: expected literal string or number saw: operator '*' at /usr/share/systemtap/tapset/libvirt_probes-64.stp:204:3 source: *cert = $arg4; ^ 1 parse error. To address the issue the regex extracting the arguments needs to optionally match any number of '*' instead of just one. Resolves: https://issues.redhat.com/browse/RHEL-153832 Fixes: cb33103c4afbce68134be112ecc5d0251e542650 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- scripts/dtrace2systemtap.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/dtrace2systemtap.py b/scripts/dtrace2systemtap.py index 361dab1075..12fc390bcd 100755 --- a/scripts/dtrace2systemtap.py +++ b/scripts/dtrace2systemtap.py @@ -127,7 +127,7 @@ for file in filelist: if re.match(r'''.*char\s\*[^\*].*''', arg) is not None: isstr = True - m = re.search(r'''^.*\s\*?(\S+)$''', arg) + m = re.search(r'''^.*\s\**(\S+)$''', arg) if m is not None: arg = m.group(1) else: -- 2.53.0

2 1

Failed to find parent device on Arrow Lake processor
by Bertrand Leconte 10 Mar '26

10 Mar '26

Hello, I am using libvirt within Qubes OS (https://www.qubes-or.org) My laptop is a Lenovo P16 Gen 3 with an Intel Core Ultra 9 275HX, which is in the Arrow Lake family. The associated PCH is an Intel 800 Series WM880. The particularity of this platform is that the PCH includes a dedicated PCI Root Complex in addition to the CPU one. That means that I have devices connected to `0000:00` bus (addresses like 0000:00:xx.x) and devices connected to `0000:80` bus (addresses like 0000:80:xx.x). The consequence of that architecture is that there is no PCI Root Port serving the `0000:80` bus. When I want to PCI passthrough my USB controller device `0000:80:14.0`, the function `virPCIDeviceIsBehindSwitchLackingACS()` says that `Failed to find parent device for 0000:80:14.0`. When I look into the source code, it looks like it enumerates all PCI devices, searching for the Root Port managing the bus `0000:80`. As I said before, it cannot find any as it's A Root Complex that manage the `0000:80` bus and not a Root Port (a Root Complex it's not a PCI/PCIe device). However, for any device connected to PCI bus `0000:00`, it cannot find any PCI Root Port either but there is an exception in the code for it. @Bloged (one of Qubes OS users) proposed in https://forum.qubes-os.org/t/aistone-x6ar57ty-aka-tongfang-x6-16-intel-high… to change the code to diff --git a/src/util/virpci.c b/src/util/virpci.c index 3816369..fbed8f5 100644 --- a/src/util/virpci.c +++ b/src/util/virpci.c @@ -2555,12 +2555,20 @@ virPCIDeviceIsBehindSwitchLackingACS(virPCIDevice *dev) if (virPCIDeviceGetParent(dev, &parent) < 0) return -1; if (!parent) { - /* if we have no parent, and this is the root bus, ACS doesn't come - * into play since devices on the root bus can't P2P without going - * through the root IOMMU. + /* if we have no parent, and this is a root-like bus (0 or 0x80), + * ACS doesn't come into play since devices on these buses can't P2P + * without going through the root IOMMU. + * + * Arrow Lake (Intel Core Ultra 200 series) uses bus 0x80 for integrated + * devices like USB controllers, with a complex PCH topology that confuses + * parent device detection. Treat bus 0x80 like bus 0 for ACS purposes. */ if (dev->address.bus == 0) { return 0; + } else if (dev->address.bus == 0x80) { + VIR_DEBUG("%s %s: Arrow Lake bus 0x80 device, treating as root bus for ACS check", + dev->id, dev->name); + return 0; } else { virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to find parent device for %1$s"), -- This patch works, but it does not look very generic. In addition to that, a thread in 2017 (this mailing list) seems to indicate that this part of the code may not be necessary: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/QRI4…. My question: can someone that knows this part of the code have a look to see what is the best way to correct it: either propose a patch to `virPCIDeviceIsBehindSwitchLackingACS` that is more generic or to change the rest of the code so it's no more used? (By the way, internationalization of error messages in syslog/dmesg/journald does not look like a good idea. It is very hard to understand if the message is given in French in this mailing list…) Thanks very much for your help Bertrand

6 17

[PATCH] systemtap: Fix probe definition for 'rpc_tls_context_new'
by Peter Krempa 10 Mar '26

10 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> Commit d249170bf609d2c modified the arguments of 'virNetTLSContextNew' which has a systemtap probe point defined. This in turn meant that the probe point needed to be modified too. Unfortunately the systemtap generator doesn't handle double pointers correctly and in fact systemtap doesn't even seem to have a possibility to fetch a list of strings from userspace natively. This meant that an invalid probe definition was generated: probe libvirt.rpc.tls_context_new = process("/usr/lib64/libvirt.so").mark("rpc_tls_context_new") { ctxt = $arg1; cacert = user_string($arg2); cacrl = user_string($arg3); *cert = $arg4; *keys = $arg5; sanityCheckCert = $arg6; requireValidCert = $arg7; isServer = $arg8; } Leading to the following failure: # stap -ve 'probe oneshot {exit()}' parse error: expected literal string or number saw: operator '*' at /usr/share/systemtap/tapset/libvirt_probes-64.stp:204:3 source: *cert = $arg4; ^ 1 parse error. To address the issue declare the 'cert' and 'keys' parameters as 'void *', we can't really do anything else as string lists aren't supported, which will make our generator generate correct code once again. Resolves: https://issues.redhat.com/browse/RHEL-153832 Fixes: d249170bf609d2cb89c36477b0f9ca0908f25985 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- src/libvirt_probes.d | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/libvirt_probes.d b/src/libvirt_probes.d index d9e75d9797..16d1decf46 100644 --- a/src/libvirt_probes.d +++ b/src/libvirt_probes.d @@ -54,7 +54,7 @@ provider libvirt { # file: src/rpc/virnettlscontext.c # prefix: rpc probe rpc_tls_context_new(void *ctxt, const char *cacert, const char *cacrl, - const char **cert, const char **keys, + void *cert, void *keys, int sanityCheckCert, int requireValidCert, int isServer); probe rpc_tls_context_dispose(void *ctxt); probe rpc_tls_context_session_allow(void *ctxt, void *sess, const char *dname); -- 2.53.0

2 4

Re: [PATCH 2/6] system/hostmem: Remove HostMemoryBackend::use_canonical_path field
by Thomas Huth 10 Mar '26

10 Mar '26

On 07/03/2026 16.41, Philippe Mathieu-Daudé wrote: > The HostMemoryBackend::use_canonical_path boolean > was only set in the hw_compat_3_1[] array, via the > 'x-use-canonical-path-for-ramblock-id=true' property. > We removed all machines using that array, but the > comment added in commit 8db0b20415c ("machine: add > missing doc for memory-backend option") requested to > not remove this property "as long as 4.0 and older > machine types exists". Previous commit removed the > last machines in this range, we can now remove that > property and all the code around it. > > Signed-off-by: Philippe Mathieu-Daudé <philmd(a)linaro.org> > --- > qapi/qom.json | 8 +------- > include/system/hostmem.h | 2 +- > backends/hostmem.c | 34 ---------------------------------- > qemu-options.hx | 10 ---------- > 4 files changed, 2 insertions(+), 52 deletions(-) > > diff --git a/qapi/qom.json b/qapi/qom.json > index c653248f85d..3e13e3fd25c 100644 > --- a/qapi/qom.json > +++ b/qapi/qom.json > @@ -660,11 +660,6 @@ > # > # @size: size of the memory region in bytes > # > -# @x-use-canonical-path-for-ramblock-id: if true, the canonical path > -# is used for ramblock-id. Disable this for 4.0 machine types or > -# older to allow migration with newer QEMU versions. > -# (default: false generally, but true for machine types <= 4.0) > -# > # .. note:: prealloc=true and reserve=false cannot be set at the same > # time. With reserve=true, the behavior depends on the operating > # system: for example, Linux will not reserve swap space for shared > @@ -683,8 +678,7 @@ > '*prealloc-context': 'str', > '*share': 'bool', > '*reserve': 'bool', > - 'size': 'size', > - '*x-use-canonical-path-for-ramblock-id': 'bool' } } > + 'size': 'size' } } > > ## > # @MemoryBackendFileProperties: > diff --git a/include/system/hostmem.h b/include/system/hostmem.h > index 88fa791ac78..cafcf9c9c74 100644 > --- a/include/system/hostmem.h > +++ b/include/system/hostmem.h > @@ -74,7 +74,7 @@ struct HostMemoryBackend { > > /* protected */ > uint64_t size; > - bool merge, dump, use_canonical_path; > + bool merge, dump; > bool prealloc, is_mapped, share, reserve; > bool guest_memfd, aligned; > uint32_t prealloc_threads; > diff --git a/backends/hostmem.c b/backends/hostmem.c > index 15d4365b697..7da7cd54e05 100644 > --- a/backends/hostmem.c > +++ b/backends/hostmem.c > @@ -40,10 +40,6 @@ QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_INTERLEAVE != MPOL_INTERLEAVE); > char * > host_memory_backend_get_name(HostMemoryBackend *backend) > { > - if (!backend->use_canonical_path) { > - return g_strdup(object_get_canonical_path_component(OBJECT(backend))); > - } I'm a little bit confused by this one here. If I've got that right, the flag was set to "true" for the old QEMU machines <= 3.1, and it is set to false by default for new machines? If so, shouldn't we rather always use the above return and remove the one below instead? > return object_get_canonical_path(OBJECT(backend)); > } > > @@ -484,23 +480,6 @@ static void host_memory_backend_set_reserve(Object *o, bool value, Error **errp) > } > #endif /* CONFIG_LINUX */ > > -static bool > -host_memory_backend_get_use_canonical_path(Object *obj, Error **errp) > -{ > - HostMemoryBackend *backend = MEMORY_BACKEND(obj); > - > - return backend->use_canonical_path; > -} > - > -static void > -host_memory_backend_set_use_canonical_path(Object *obj, bool value, > - Error **errp) > -{ > - HostMemoryBackend *backend = MEMORY_BACKEND(obj); > - > - backend->use_canonical_path = value; > -} > - > static void > host_memory_backend_class_init(ObjectClass *oc, const void *data) > { > @@ -563,19 +542,6 @@ host_memory_backend_class_init(ObjectClass *oc, const void *data) > object_class_property_set_description(oc, "reserve", > "Reserve swap space (or huge pages) if applicable"); > #endif /* CONFIG_LINUX */ > - /* > - * Do not delete/rename option. This option must be considered stable > - * (as if it didn't have the 'x-' prefix including deprecation period) as > - * long as 4.0 and older machine types exists. By the way, this option seems to be used by libvirt - we should maybe make sure that its removal does not break libvirt by accident...? Thomas

2 1

[PATCH 0/4] qemu: bump qemu capabilities and use 'flat' mode of 'query-block'
by Peter Krempa 10 Mar '26

10 Mar '26

Peter Krempa (4): qemucapabilitiestest: Update 'caps_11.0.0_x86_64' to 'v10.2.0-2062-g084a6c6e73" qemucapabilitiestest: Update 'caps_11.0.0_aarch64' to 'v10.2.0-2062-g084a6c6e73"aarch qemu: capabilities: Introduce QEMU_CAPS_QUERY_BLOCK_FLAT qemu: monitor: Use 'flat' mode of 'query-block' src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_monitor.c | 1 + src/qemu/qemu_monitor_json.c | 12 +- src/qemu/qemu_monitor_priv.h | 3 + .../caps_11.0.0_aarch64.replies | 5587 ++++++++-------- .../caps_11.0.0_aarch64.xml | 39 +- .../caps_11.0.0_x86_64.replies | 5815 +++++++++-------- .../caps_11.0.0_x86_64.xml | 287 +- 9 files changed, 6254 insertions(+), 5493 deletions(-) -- 2.53.0

3 6

[PATCH] docs: storage: Note that refreshing the storage pool is needed on outside modification
by Peter Krempa 10 Mar '26

10 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> Add a note stating that pool contents and metadata are not auto-updated when the pool is modified outside of libvirt. Closes: https://gitlab.com/libvirt/libvirt/-/work_items/749 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- docs/formatstorage.rst | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/docs/formatstorage.rst b/docs/formatstorage.rst index 9d9a4143eb..7ed57732af 100644 --- a/docs/formatstorage.rst +++ b/docs/formatstorage.rst @@ -25,6 +25,13 @@ attribute ``type``, which is one of ``dir``, ``fs``, ``netfs``, ``disk``, ``iscsi-direct`` ( :since:`since 4.7.0` ). This corresponds to the storage backend drivers listed further along in this document. + .. note:: + Any modification of the storage pool outside of libvirt or on a different + host for shared storage pools may not update the cached volume or pool + metadata. The pool metadata can be refreshed by calling + ``virsh pool-refresh POOLNAME`` or by the + `virStoragePoolRefresh <html/libvirt-libvirt-storage.html#virStoragePoolRefresh>`__ API. + Storage pool general metadata ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -- 2.53.0

2 1

[libvirt PATCH] qemu_hotplug: Fix crash when attaching network inteface with hostdev network
by Pavel Hrdina 10 Mar '26

10 Mar '26

From: Pavel Hrdina <phrdina(a)redhat.com> The following interface XML will crash libvirt when attached to running VM: <interface type='network'> <source network='hostdev_net'/> </interface> Instead of trying to figure out every single place where hostdev is used as struct directly and allocating empty private data check if iommufd is configured as that will happen only for <hostdev> device where the private data are correctly allocated. This patch is best viewed with `git show -w`. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/qemu/qemu_hotplug.c | 37 +++++++++++++++++++++---------------- 1 file changed, 21 insertions(+), 16 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index b3f2a173a8..966e746f2b 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -1630,20 +1630,23 @@ qemuDomainAttachHostPCIDevice(virQEMUDriver *driver, qemuDomainObjEnterMonitor(vm); - if (objprops) { - if ((ret = qemuFDPassDirectTransferMonitor(priv->iommufd, priv->mon)) < 0) - goto exit_monitor; - if ((ret = qemuMonitorAddObject(priv->mon, &objprops, NULL)) < 0) - goto exit_monitor; + if (virHostdevIsPCIDeviceWithIOMMUFD(hostdev)) { + if (objprops) { + if ((ret = qemuFDPassDirectTransferMonitor(priv->iommufd, priv->mon)) < 0) + goto exit_monitor; + + if ((ret = qemuMonitorAddObject(priv->mon, &objprops, NULL)) < 0) + goto exit_monitor; - priv->iommufdState = true; - removeiommufd = true; + priv->iommufdState = true; + removeiommufd = true; + } + + if ((ret = qemuFDPassDirectTransferMonitor(hostdevPriv->vfioDeviceFd, priv->mon)) < 0) + goto exit_monitor; } - if ((ret = qemuFDPassDirectTransferMonitor(hostdevPriv->vfioDeviceFd, priv->mon)) < 0) - goto exit_monitor; - if ((ret = qemuDomainAttachExtensionDevice(priv->mon, hostdev->info)) < 0) goto exit_monitor; @@ -1674,15 +1677,17 @@ qemuDomainAttachHostPCIDevice(virQEMUDriver *driver, if (teardownmemlock && qemuDomainAdjustMaxMemLock(vm) < 0) VIR_WARN("Unable to reset maximum locked memory on hotplug fail"); - qemuDomainObjEnterMonitor(vm); + if (virHostdevIsPCIDeviceWithIOMMUFD(hostdev)) { + qemuDomainObjEnterMonitor(vm); - if (removeiommufd) - ignore_value(qemuMonitorDelObject(priv->mon, "iommufd0", false)); + if (removeiommufd) + ignore_value(qemuMonitorDelObject(priv->mon, "iommufd0", false)); - qemuFDPassDirectTransferMonitorRollback(hostdevPriv->vfioDeviceFd, priv->mon); - qemuFDPassDirectTransferMonitorRollback(priv->iommufd, priv->mon); + qemuFDPassDirectTransferMonitorRollback(hostdevPriv->vfioDeviceFd, priv->mon); + qemuFDPassDirectTransferMonitorRollback(priv->iommufd, priv->mon); - qemuDomainObjExitMonitor(vm); + qemuDomainObjExitMonitor(vm); + } if (releaseaddr) qemuDomainReleaseDeviceAddress(vm, info); -- 2.53.0

2 1

[PATCH] util: Move 'virProcessLimitResourceToLabel' into same preprocessor if-block as only caller
by Peter Krempa 10 Mar '26

10 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> 'virProcessLimitResourceToLabel' is called only from 'virProcessGetLimitFromProc' but the latter has different conditions when it's compiled. In certain cases this could lead to build failures. Fixes: 90fe839f8a0 Closes: https://gitlab.com/libvirt/libvirt/-/work_items/848 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- src/util/virprocess.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/src/util/virprocess.c b/src/util/virprocess.c index e8120c1bdc..e06ed94b0d 100644 --- a/src/util/virprocess.c +++ b/src/util/virprocess.c @@ -743,36 +743,36 @@ virProcessPrLimit(pid_t pid G_GNUC_UNUSED, #endif #if WITH_GETRLIMIT +# if defined(__linux__) static const char* virProcessLimitResourceToLabel(int resource) { switch (resource) { -# if defined(RLIMIT_MEMLOCK) +# if defined(RLIMIT_MEMLOCK) case RLIMIT_MEMLOCK: return "Max locked memory"; -# endif /* defined(RLIMIT_MEMLOCK) */ +# endif /* defined(RLIMIT_MEMLOCK) */ -# if defined(RLIMIT_NPROC) +# if defined(RLIMIT_NPROC) case RLIMIT_NPROC: return "Max processes"; -# endif /* defined(RLIMIT_NPROC) */ +# endif /* defined(RLIMIT_NPROC) */ -# if defined(RLIMIT_NOFILE) +# if defined(RLIMIT_NOFILE) case RLIMIT_NOFILE: return "Max open files"; -# endif /* defined(RLIMIT_NOFILE) */ +# endif /* defined(RLIMIT_NOFILE) */ -# if defined(RLIMIT_CORE) +# if defined(RLIMIT_CORE) case RLIMIT_CORE: return "Max core file size"; -# endif /* defined(RLIMIT_CORE) */ +# endif /* defined(RLIMIT_CORE) */ default: return NULL; } } -# if defined(__linux__) static int virProcessGetLimitFromProc(pid_t pid, int resource, -- 2.53.0

2 1

[PATCH] qemuhotplugtest: Run "interface-hostdev" test cases only on Linux
by Michal Privoznik 09 Mar '26

09 Mar '26

From: Michal Privoznik <mprivozn(a)redhat.com> In one of my previous commits, I've introduced "interface-hostdev" attach and detach test cases to qemuhotplugtest. And they work flawlessly, on Linux. But on anything else they fail because our virpci.c module is basically just a bunch of stub functions that do nothing but report an error, rendering my changes to virpcimock futile. BTW: this is similar to what I had done in v12.1.0-rc1~199. Fixes: f9bb819fc4841dbdff801629bf58f9fd6d7d93eb Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Verified in my FreeBSD VM. tests/qemuhotplugtest.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/tests/qemuhotplugtest.c b/tests/qemuhotplugtest.c index 6cb616c59b..146e9e33c1 100644 --- a/tests/qemuhotplugtest.c +++ b/tests/qemuhotplugtest.c @@ -697,10 +697,18 @@ mymain(void) DO_TEST_DETACH("ppc64", "pseries-base-live", "hostdev-pci", false, false, "device_del", QMP_DEVICE_DELETED("hostdev0") QMP_OK); +#ifdef __linux__ + /* While <interface type='hostdev'/> is nearly the same as <hostdev/>, + * there are subtle differences, e.g. checking that PCI device specified in + * <interface/> is a VF. Checks like these are done by walking sysfs which + * is limited to Linux, obviously. And while our virpcimock creates + * necessary structure, on non-Linux the virpci.c is compiled with stubs + * that do nothing but report an error. */ DO_TEST_ATTACH("x86_64", "base-live", "interface-hostdev", false, true, "device_add", QMP_OK); DO_TEST_DETACH("x86_64", "base-live", "interface-hostdev", false, false, "device_del", QMP_DEVICE_DELETED("hostdev0") QMP_OK); +#endif DO_TEST_ATTACH("x86_64", "base-live", "interface-vdpa", false, true, "query-fdsets", "{\"return\":[{\"fdset-id\":99999}]}", "add-fd", "{ \"return\": { \"fdset-id\": 1, \"fd\": 95 }}", -- 2.52.0

2 1

[PATCH] util: extend virHostCPUGetInfo() for FreeBSD
by Roman Bogorodskiy 09 Mar '26

09 Mar '26

Extend virHostCPUGetInfo() to report more data on FreeBSD, such as: - NUMA domain count - CPU core count - CPU threads per core count Signed-off-by: Roman Bogorodskiy <bogorodskiy(a)gmail.com> --- src/util/virhostcpu.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/src/util/virhostcpu.c b/src/util/virhostcpu.c index 113aa6881e..7cecb4b2c8 100644 --- a/src/util/virhostcpu.c +++ b/src/util/virhostcpu.c @@ -1014,6 +1014,22 @@ virHostCPUGetInfo(virArch hostarch G_GNUC_UNUSED, *mhz = cpu_freq / 1000000; # endif +# ifdef __FreeBSD__ +# define FILL_DATA_SYSCTL(var, sysctl, error_msg) \ + do { \ + size_t _len = sizeof(*var); \ + if (sysctlbyname(sysctl, var, &_len, NULL, 0) < 0) { \ + virReportSystemError(errno, "%s", error_msg); \ + return -1; \ + } \ + } while (0) + + FILL_DATA_SYSCTL(nodes, "vm.ndomains", _("cannot obtain NUMA domain count")); + FILL_DATA_SYSCTL(cores, "kern.smp.cores", _("cannot obtain CPU core count")); + FILL_DATA_SYSCTL(threads, "kern.smp.threads_per_core", + _("cannot obtain CPU threads per core")); +# endif /* __FreeBSD__ */ + return 0; #else /* XXX Solaris will need an impl later if they port QEMU driver */ -- 2.52.0

2 1

[PATCH 0/3] docs: bhyve documentation updates
by Roman Bogorodskiy 09 Mar '26

09 Mar '26

Roman Bogorodskiy (3): docs: bhyve: document NUMA domains configuration docs: bhyve: fix typo in version number docs: bhyve: add arm64 guest example docs/drvbhyve.rst | 64 ++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 63 insertions(+), 1 deletion(-) -- 2.52.0

2 4

[PATCH] qemu: processShutdownCompletedEvent: Remove inactive VM object after shutdown
by Peter Krempa 09 Mar '26

09 Mar '26

From: Peter Krempa <pkrempa(a)redhat.com> When the qemu process can't be successfully killed (e.g. when it's stuck in a long system call) libvirt creates a watch and waits for the monitor socket to go away before cleaning up the domain. The cleanup code in 'processShutdownCompletedEvent' called 'qemuProcessStop' but didn't call also 'qemuDomainRemoveInactive' which would break if a transient VM would undergo the delayed cleanup as we'd still have it's VM object around. Fixes: e62c26a20dced58ea342d9cb8f5e9164dc3bb023 Closes: https://gitlab.com/libvirt/libvirt/-/work_items/853 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- src/qemu/qemu_driver.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 9e4d0f8640..7d33477636 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4079,6 +4079,7 @@ processShutdownCompletedEvent(virDomainObj *vm) if (virDomainObjIsActive(vm)) { qemuProcessStop(vm, VIR_DOMAIN_SHUTOFF_UNKNOWN, VIR_ASYNC_JOB_NONE, 0); + qemuDomainRemoveInactive(vm, 0, false); } qemuProcessEndStopJob(vm); -- 2.53.0

2 1

Re: [PATCH v2 14/21] hw/core/qdev-properties: allow qdev properties accept flags
by Daniel P. Berrangé 09 Mar '26

09 Mar '26

On Wed, Feb 11, 2026 at 03:30:06PM +0800, Zhao Liu wrote: > On Tue, Feb 10, 2026 at 09:56:08AM +0000, Daniel P. Berrangé wrote: > > Date: Tue, 10 Feb 2026 09:56:08 +0000 > > From: "Daniel P. Berrangé" <berrange(a)redhat.com> > > Subject: Re: [PATCH v2 14/21] hw/core/qdev-properties: allow qdev > > properties accept flags > > > > On Tue, Feb 10, 2026 at 11:23:41AM +0800, Zhao Liu wrote: > > > Update qdev property interfaces (qdev_property_add_static() and > > > qdev_class_add_property()) to accept and pass 'ObjectPropertyFlags'. > > > This enables marking qdev properties with flags such as DEPRECATED or > > > INTERNAL. > > > > > > To facilitate this at the definition level, extend the boolean and > > > uint8_t property macros (as the examples) to accept variable arguments > > > (VA_ARGS). This allows callers to optionally specify flags in the > > > property definition. > > > > > > Example: > > > > > > DEFINE_PROP_UINT8("version", IOAPICCommonState, version, IOAPIC_VER_DEF, > > > .flags = OBJECT_PROPERTY_DEPRECATED), > > > > In other places where we track deprecation in QEMU, we have not used > > a boolean flag. Instead we have used a "const char *deprecation_note" > > internally, which lets us provide a user facing message, to be printed > > out in the warn_report, informing them what to do instead (either the > > feature is entirely removed, or there is a better alternative). IMHO > > we should be following the same pattern for properties, as it is much > > more user friendly than just printing a totally generic message > > "XXXX is deprecated, stop using it" > > Yes, rich deprecation hint is better. I think this still depends on > USER_SET - distinguish internal/external or not :-(. > > Since when we mark a property as deprecated, its code remains in the > code tree, and internal calls should not trigger warnings. Deprecation > hints are intended to reminder external users. This depends on where you put the deprecation check. IIUC, all the user facing codepaths for setting properties end up calling through object_set_properties_from_qdict, but internal codepaths don't use that. That method can check & emit the deprecation warnings, without us needing any explicit tracking of "user set" - the use context is derived from the codepath With regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|

4 8

[PATCH 0/4] qemuhotplugtest: Introduce interface-hostdev test case
by Michal Privoznik 09 Mar '26

09 Mar '26

Inspired by the following patch: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/YQHQ… Michal Prívozník (4): virpcimock: Create sriov_totalvfs file and virtfnN symlinks virpcimock: Create net/ subdir for devices virnetdevmock: Mock virNetDevSaveNetConfig() and virNetDevSetNetConfig() qemuhotplugtest: Introduce interface-hostdev test case src/util/virnetdev.h | 5 +- tests/qemuhotplugtest.c | 7 +- .../qemuhotplug-interface-hostdev.xml | 6 ++ ...emuhotplug-base-live+interface-hostdev.xml | 63 +++++++++++++++ tests/virnetdevmock.c | 22 ++++++ tests/virpcimock.c | 78 +++++++++++++++++-- 6 files changed, 171 insertions(+), 10 deletions(-) create mode 100644 tests/qemuhotplugtestdevices/qemuhotplug-interface-hostdev.xml create mode 100644 tests/qemuhotplugtestdomains/qemuhotplug-base-live+interface-hostdev.xml -- 2.52.0

2 5

[PATCH 0/6] qemu: fix a passt bug, add a passt feature
by Laine Stump 09 Mar '26

09 Mar '26

The first patch cleans up a dangling socket that was hanging around after we finished using a passt process in vhostuser mode. The rest are to support specifying the default route for a passt interface with <route gateway='x.x.x.x'/> Laine Stump (6): qemu: delete passt "repair" socket when appropriate qemu: reorganize qemuValidateDomainDeviceDefNetwork() qemu: only limit IPv4 prefix for slirp util: make the route arg to virNetDevIPRouteGetPrefix const conf/util: fix non-specification of IP route destination address qemu: support setting default route for passt interfaces inside the guest src/conf/networkcommon_conf.c | 74 ++++++++------- src/conf/schemas/networkcommon.rng | 4 +- src/qemu/qemu_passt.c | 30 ++++++- src/qemu/qemu_validate.c | 140 +++++++++++++++++++---------- src/util/virnetdevip.c | 14 ++- src/util/virnetdevip.h | 2 +- 6 files changed, 170 insertions(+), 94 deletions(-) -- 2.52.0

3 17

[PULL v2 00/27] Misc patches queue
by Daniel P. Berrangé 07 Mar '26

07 Mar '26

The following changes since commit 3fb456e9a0e9eef6a71d9b49bfff596a0f0046e9: Merge tag 'pull-request-2026-03-02' of https://gitlab.com/thuth/qemu into staging (2026-03-02 14:01:46 +0000) are available in the Git repository at: https://gitlab.com/berrange/qemu tags/next-pr-pull-request for you to fetch changes up to ca61f91ef9b0d10333881fd0070303ea33cbc72e: util/oslib-posix: increase memprealloc thread count to 32 (2026-03-05 17:40:24 +0000) ---------------------------------------------------------------- * Increase pre-alloc max thread count to 32 * Fix checkpatch warning for new/removed files with --terse * Detect more GPL boilerplate in checkpatch * Fix lean of data in TLS/websock GSource cancellation * Tweak CPU docs for DiamondRapids * Unconditionally enable thread naming * Fix race setting thread naming on Win32 * Add API docs for logging APIs * Fix interleaved error/trace output * Fix missing error prefixes with warn_report * Add detailed error reporting for VNC passworrd changes * Refactoring of error_vprintf & related funcs ---------------------------------------------------------------- Daniel P. Berrangé (25): scripts: detect another GPL license boilerplate variant io: separate freeing of tasks from marking them as complete io: fix cleanup for TLS I/O source data on cancellation io: fix cleanup for websock I/O source data on cancellation docs: simplify DiamondRapids CPU docs qemu-options: remove extraneous [] around arg values include: define constant for early constructor priority monitor: initialize global data from a constructor system: unconditionally enable thread naming util: fix race setting thread name on Win32 util: expose qemu_thread_set_name audio: make jackaudio use qemu_thread_set_name util: set the name for the 'main' thread on Windows util: add API to fetch the current thread name util: introduce some API docs for logging APIs util: avoid repeated prefix on incremental qemu_log calls util/log: add missing error reporting in qemu_log_trylock_with_err ui: add proper error reporting for password changes ui: remove redundant use of error_printf_unless_qmp() monitor: remove redundant error_[v]printf_unless_qmp monitor: refactor error_vprintf() monitor: move error_vprintf back to error-report.c util: fix interleaving of error & trace output util: don't skip error prefixes when QMP is active util: fix interleaving of error prefixes Jon Kohler (1): util/oslib-posix: increase memprealloc thread count to 32 Markus Armbruster (1): scripts/checkpatch: Fix MAINTAINERS update warning with --terse audio/jackaudio.c | 28 +++++-- docs/about/deprecated.rst | 7 ++ docs/system/cpu-models-x86.rst.inc | 16 ++-- include/io/task.h | 29 +++++-- include/monitor/monitor.h | 3 - include/qemu/compiler.h | 8 ++ include/qemu/log-for-trace.h | 17 +++- include/qemu/log.h | 39 ++++++++++ include/qemu/thread.h | 3 +- include/ui/console.h | 2 +- include/ui/qemu-spice-module.h | 3 +- io/channel-tls.c | 64 ++++++++++----- io/channel-websock.c | 46 ++++++++--- io/task.c | 8 +- meson.build | 21 +++++ monitor/monitor.c | 51 +++--------- qemu-options.hx | 112 +++++++++++++-------------- rust/util/src/log.rs | 6 ++ scripts/checkpatch.pl | 11 +-- stubs/error-printf.c | 23 ------ stubs/meson.build | 1 - stubs/monitor-core.c | 19 ++++- system/vl.c | 12 +-- tests/functional/generic/test_vnc.py | 4 +- tests/unit/test-io-task.c | 26 +++++++ tests/unit/test-util-sockets.c | 1 + ui/spice-core.c | 25 ++++-- ui/spice-module.c | 7 +- ui/ui-qmp-cmds.c | 19 ++--- ui/vnc-stubs.c | 6 +- ui/vnc.c | 16 ++-- util/error-report.c | 92 +++++++++++++++++----- util/log.c | 50 +++++++----- util/oslib-posix.c | 2 +- util/qemu-thread-posix.c | 89 +++++++++++++++------ util/qemu-thread-win32.c | 87 ++++++++++++++------- 36 files changed, 631 insertions(+), 322 deletions(-) delete mode 100644 stubs/error-printf.c -- 2.53.0

3 29

[PATCH] test: Default to ROM type for loader
by Jim Fehlig 06 Mar '26

06 Mar '26

From: Jim Fehlig <jfehlig(a)suse.com> Commit 1504b7f687 moved the corresponding logic from the generic postparse code to drivers but failed to update the test driver, which causes failures in virt-manager's test suite. Fixes: 1504b7f687bdfc679377e605d076776b18533468 Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- As an alternative to this patch, we can consider reverting 1504b7f687 (and the subsequent bhyve fix) to avoid duplicating the logic. src/test/test_driver.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 1165689de7..536e291861 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -421,6 +421,22 @@ testDomainDevicesDefPostParse(virDomainDeviceDef *dev G_GNUC_UNUSED, } +static int +testDomainDefPostParse(virDomainDef *def, + unsigned int parseFlags G_GNUC_UNUSED, + void *opaque G_GNUC_UNUSED, + void *parseOpaque G_GNUC_UNUSED) +{ + if (def->os.loader && + def->os.loader->path && + !def->os.loader->type) { + def->os.loader->type = VIR_DOMAIN_LOADER_TYPE_ROM; + } + + return 0; +} + + static void testDomainObjPrivateFree(void *data) { @@ -448,6 +464,7 @@ testDriverNew(void) VIR_DOMAIN_DEF_FEATURE_FW_AUTOSELECT | VIR_DOMAIN_DEF_FEATURE_NET_MODEL_STRING, .devicesPostParseCallback = testDomainDevicesDefPostParse, + .domainPostParseCallback = testDomainDefPostParse, .defArch = VIR_ARCH_I686, }; virDomainXMLPrivateDataCallbacks privatecb = { -- 2.51.0

2 1

[PATCH 1/2] hyperv: fix flags passed to virDomainDefParseString()
by Jonathon Jongsma 06 Mar '26

06 Mar '26

This function expects flags that are bitwise-or values of the VIR_DOMAIN_DEF_PARSE_* constants, but we were passing two flags that were not part of this flag set: - 1 < VIR_DOMAIN_VIRT_HYPERV - VIR_DOMAIN_XML_INACTIVE Replace VIR_DOMAIN_XML_INACTIVE with VIR_DOMAIN_DEF_PARSE_INACTIVE (which fortunately happens to be the exact same value). Remove flag "1 < VIR_DOMAIN_VIRT_HYPERV", which turns out to the the same value as the flag VIR_DOMAIN_DEF_PARSE_ABI_UPDATE_MIGRATION, which seems unnecessary and undesirable here. Signed-off-by: Jonathon Jongsma <jjongsma(a)redhat.com> --- src/hyperv/hyperv_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hyperv/hyperv_driver.c b/src/hyperv/hyperv_driver.c index b01b4919fe..8832769c7b 100644 --- a/src/hyperv/hyperv_driver.c +++ b/src/hyperv/hyperv_driver.c @@ -2937,7 +2937,7 @@ hypervDomainDefineXML(virConnectPtr conn, const char *xml) /* parse xml */ def = virDomainDefParseString(xml, priv->xmlopt, NULL, - 1 << VIR_DOMAIN_VIRT_HYPERV | VIR_DOMAIN_XML_INACTIVE); + VIR_DOMAIN_DEF_PARSE_INACTIVE); if (!def) goto error; -- 2.53.0

2 2

[PATCH] bhyve: support NUMA configuration for domains
by Roman Bogorodskiy 06 Mar '26

06 Mar '26

Bhyve supports NUMA domains configuration using the '-n' command line argument: -n id,size,cpus[,domain_policy] Here, "id" is a numeric NUMA domain id, "size" is the total VM memory size with units format similar to the "-m" switch, "cpus" is a cpuset, and "domain_policy" is an optional domainset(9) memory allocation policy. The "domain_policy" is currently not used by the libvirt driver. This argument is repeated for every NUMA domain to be configured, e.g.: bhyve \ ... -n id=0,size=107,cpus=0-3 -n id=1,size=107,cpus=4-7 To support that: * Add a corresponding capability; it is considered supported if the bhyve binary has the '-n' command line switch. * Generate command line arguments for NUMA from <cpu><numa>..</numa></cpu> domain configuration. Additionally, validate that: * NUMA domains can be only configured with the UEFI loaders. * No more than 8 domains configured per VM as limited by bhyve. Signed-off-by: Roman Bogorodskiy <bogorodskiy(a)gmail.com> --- src/bhyve/bhyve_capabilities.c | 3 ++ src/bhyve/bhyve_capabilities.h | 1 + src/bhyve/bhyve_command.c | 26 ++++++++++++ src/bhyve/bhyve_domain.c | 19 +++++++++ .../bhyvexml2argv-numa-empty-cpuset.xml | 29 +++++++++++++ .../bhyvexml2argv-numa-too-many-domains.xml | 36 ++++++++++++++++ .../x86_64/bhyvexml2argv-numa.args | 14 +++++++ .../x86_64/bhyvexml2argv-numa.ldargs | 1 + .../x86_64/bhyvexml2argv-numa.xml | 29 +++++++++++++ tests/bhyvexml2argvtest.c | 10 ++++- .../x86_64/bhyvexml2xmlout-numa.xml | 42 +++++++++++++++++++ tests/bhyvexml2xmltest.c | 1 + 12 files changed, 210 insertions(+), 1 deletion(-) create mode 100644 tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-empty-cpuset.xml create mode 100644 tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-too-many-domains.xml create mode 100644 tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.args create mode 100644 tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.ldargs create mode 100644 tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.xml create mode 100644 tests/bhyvexml2xmloutdata/x86_64/bhyvexml2xmlout-numa.xml diff --git a/src/bhyve/bhyve_capabilities.c b/src/bhyve/bhyve_capabilities.c index d2a48ed30c..c3fb88fe9f 100644 --- a/src/bhyve/bhyve_capabilities.c +++ b/src/bhyve/bhyve_capabilities.c @@ -258,6 +258,9 @@ bhyveProbeCapsFromHelp(unsigned int *caps, char *binary) if (strstr(help, "-A:") != NULL) *caps |= BHYVE_CAP_ACPI; + if (strstr(help, "-n:") != NULL) + *caps |= BHYVE_CAP_NUMA; + return 0; } diff --git a/src/bhyve/bhyve_capabilities.h b/src/bhyve/bhyve_capabilities.h index d5346df7ba..31fd9ab86a 100644 --- a/src/bhyve/bhyve_capabilities.h +++ b/src/bhyve/bhyve_capabilities.h @@ -56,6 +56,7 @@ typedef enum { BHYVE_CAP_VIRTIO_RND = 1 << 10, BHYVE_CAP_NVME = 1 << 11, BHYVE_CAP_ACPI = 1 << 12, + BHYVE_CAP_NUMA = 1 << 13, } virBhyveCapsFlags; int virBhyveProbeGrubCaps(virBhyveGrubCapsFlags *caps); diff --git a/src/bhyve/bhyve_command.c b/src/bhyve/bhyve_command.c index 37618812bc..931d7dd551 100644 --- a/src/bhyve/bhyve_command.c +++ b/src/bhyve/bhyve_command.c @@ -905,6 +905,7 @@ virBhyveProcessBuildBhyveCmd(struct _bhyveConn *driver, virDomainDef *def, unsigned nusbcontrollers = 0; unsigned nisacontrollers = 0; unsigned nvcpus = virDomainDefGetVcpus(def); + size_t ncells = virDomainNumaGetNodeCount(def->numa); /* CPUs */ virCommandAddArg(cmd, "-c"); @@ -955,6 +956,31 @@ virBhyveProcessBuildBhyveCmd(struct _bhyveConn *driver, virDomainDef *def, } } + /* NUMA */ + if (ncells) { + if (!(bhyveDriverGetBhyveCaps(driver) & BHYVE_CAP_NUMA)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("Installed bhyve binary does not support NUMA configuration")); + return NULL; + } + + if (def->os.bootloader || !def->os.loader) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("NUMA configuration is only supported when booting using UEFI")); + return NULL; + } + + for (i = 0; i < ncells; i++) { + unsigned long long memSize = virDomainNumaGetNodeMemorySize(def->numa, i); + virBitmap *cpus = virDomainNumaGetNodeCpumask(def->numa, i); + g_autofree char *cpumask = virBitmapFormat(cpus); + + virCommandAddArg(cmd, "-n"); + virCommandAddArgFormat(cmd, "id=%zu,size=%llu,cpus=%s", i, VIR_DIV_UP(memSize, 1024), + cpumask); + } + } + /* Memory */ virCommandAddArg(cmd, "-m"); virCommandAddArgFormat(cmd, "%llu", diff --git a/src/bhyve/bhyve_domain.c b/src/bhyve/bhyve_domain.c index 85960c6e12..4594d7673f 100644 --- a/src/bhyve/bhyve_domain.c +++ b/src/bhyve/bhyve_domain.c @@ -411,6 +411,7 @@ bhyveDomainDefValidate(const virDomainDef *def, void *parseOpaque G_GNUC_UNUSED) { size_t i; + size_t ncells; virStorageSource *src = NULL; g_autoptr(GHashTable) nvme_controllers = g_hash_table_new(g_direct_hash, g_direct_equal); @@ -445,6 +446,24 @@ bhyveDomainDefValidate(const virDomainDef *def, return -1; } + ncells = virDomainNumaGetNodeCount(def->numa); + if (ncells) { + if (ncells > 8) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("Only up to 8 NUMA domains are supported")); + return -1; + } + + for (i = 0; i < ncells; i++) { + if (!virDomainNumaGetNodeCpumask(def->numa, i)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("NUMA domain id %1$zu: empty cpusets are not allowed"), + i); + return -1; + } + } + } + if (!def->os.loader) return 0; diff --git a/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-empty-cpuset.xml b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-empty-cpuset.xml new file mode 100644 index 0000000000..9a5fc282ba --- /dev/null +++ b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-empty-cpuset.xml @@ -0,0 +1,29 @@ +<domain type='bhyve'> + <name>bhyve</name> + <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> + <memory>219136</memory> + <vcpu>8</vcpu> + <os firmware='efi'> + <type>hvm</type> + </os> + <cpu> + <numa> + <cell id='0' cpus='0-3' memory='109568' unit='KiB'/> + <cell id='1' memory='109568' unit='KiB'/> + </numa> + </cpu> + <devices> + <disk type='file'> + <driver name='file' type='raw'/> + <source file='/tmp/freebsd.img'/> + <target dev='hda' bus='sata'/> + <address type='drive' controller='0' bus='0' target='2' unit='0'/> + </disk> + <interface type='bridge'> + <mac address='52:54:00:b9:94:02'/> + <model type='virtio'/> + <source bridge="virbr0"/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </interface> + </devices> +</domain> diff --git a/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-too-many-domains.xml b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-too-many-domains.xml new file mode 100644 index 0000000000..bcabe5cd85 --- /dev/null +++ b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa-too-many-domains.xml @@ -0,0 +1,36 @@ +<domain type='bhyve'> + <name>bhyve</name> + <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> + <memory>876544</memory> + <vcpu>9</vcpu> + <os firmware='efi'> + <type>hvm</type> + </os> + <cpu> + <numa> + <cell id='0' cpus='0' memory='109568' unit='KiB'/> + <cell id='1' cpus='1' memory='109568' unit='KiB'/> + <cell id='2' cpus='2' memory='109568' unit='KiB'/> + <cell id='3' cpus='3' memory='109568' unit='KiB'/> + <cell id='4' cpus='4' memory='109568' unit='KiB'/> + <cell id='5' cpus='5' memory='109568' unit='KiB'/> + <cell id='6' cpus='6' memory='109568' unit='KiB'/> + <cell id='7' cpus='7' memory='109568' unit='KiB'/> + <cell id='8' cpus='8' memory='109568' unit='KiB'/> + </numa> + </cpu> + <devices> + <disk type='file'> + <driver name='file' type='raw'/> + <source file='/tmp/freebsd.img'/> + <target dev='hda' bus='sata'/> + <address type='drive' controller='0' bus='0' target='2' unit='0'/> + </disk> + <interface type='bridge'> + <mac address='52:54:00:b9:94:02'/> + <model type='virtio'/> + <source bridge="virbr0"/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </interface> + </devices> +</domain> diff --git a/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.args b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.args new file mode 100644 index 0000000000..15efd1c357 --- /dev/null +++ b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.args @@ -0,0 +1,14 @@ +bhyve \ +-c 8 \ +-n id=0,size=107,cpus=0-3 \ +-n id=1,size=107,cpus=4-7 \ +-m 214 \ +-u \ +-H \ +-P \ +-s 0:0,hostbridge \ +-l bootrom,fakefirmwaredir/BHYVE_UEFI.fd,fakenvramdir/bhyve_VARS.fd \ +-s 1:0,lpc \ +-s 2:0,ahci,hd:/tmp/freebsd.img \ +-s 3:0,virtio-net,faketapdev,mac=52:54:00:b9:94:02 \ +bhyve diff --git a/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.ldargs b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.ldargs new file mode 100644 index 0000000000..421376db9e --- /dev/null +++ b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.ldargs @@ -0,0 +1 @@ +dummy diff --git a/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.xml b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.xml new file mode 100644 index 0000000000..8a0da7830d --- /dev/null +++ b/tests/bhyvexml2argvdata/x86_64/bhyvexml2argv-numa.xml @@ -0,0 +1,29 @@ +<domain type='bhyve'> + <name>bhyve</name> + <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> + <memory>219136</memory> + <vcpu>8</vcpu> + <os firmware='efi'> + <type>hvm</type> + </os> + <cpu> + <numa> + <cell id='0' cpus='0-3' memory='109568' unit='KiB'/> + <cell id='1' cpus='4-7' memory='109568' unit='KiB'/> + </numa> + </cpu> + <devices> + <disk type='file'> + <driver name='file' type='raw'/> + <source file='/tmp/freebsd.img'/> + <target dev='hda' bus='sata'/> + <address type='drive' controller='0' bus='0' target='2' unit='0'/> + </disk> + <interface type='bridge'> + <mac address='52:54:00:b9:94:02'/> + <model type='virtio'/> + <source bridge="virbr0"/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </interface> + </devices> +</domain> diff --git a/tests/bhyvexml2argvtest.c b/tests/bhyvexml2argvtest.c index 2330e70bbf..b7749fec6f 100644 --- a/tests/bhyvexml2argvtest.c +++ b/tests/bhyvexml2argvtest.c @@ -202,7 +202,7 @@ mymain(void) BHYVE_CAP_FBUF | BHYVE_CAP_XHCI | \ BHYVE_CAP_CPUTOPOLOGY | BHYVE_CAP_SOUND_HDA | \ BHYVE_CAP_VNC_PASSWORD | BHYVE_CAP_VIRTIO_9P | \ - BHYVE_CAP_NVME; + BHYVE_CAP_NVME | BHYVE_CAP_NUMA; DO_TEST("base"); DO_TEST("wired"); @@ -254,6 +254,11 @@ mymain(void) DO_TEST("isa-controller"); DO_TEST_FAILURE("isa-multiple-controllers"); DO_TEST("firmware-efi"); + DO_TEST("numa"); + DO_TEST_FAILURE("numa-empty-cpuset"); + DO_TEST_FAILURE("numa-too-many-domains"); + driver.bhyvecaps &= ~BHYVE_CAP_NUMA; + DO_TEST_FAILURE("numa"); fakefirmwaredir = g_steal_pointer(&driver.config->firmwareDir); driver.config->firmwareDir = g_steal_pointer(&fakefirmwareemptydir); DO_TEST_PREPARE_ERROR("firmware-efi"); @@ -345,10 +350,13 @@ mymain(void) driver.caps = virBhyveCapsBuild(); /* bhyve does not support UTC clock on ARM */ driver.bhyvecaps ^= BHYVE_CAP_RTC_UTC; + /* bhyve does not support NUMA on ARM */ + driver.bhyvecaps &= ~BHYVE_CAP_NUMA; DO_TEST("base"); DO_TEST("console"); DO_TEST("bootloader"); + DO_TEST_FAILURE("numa"); virObjectUnref(driver.caps); virObjectUnref(driver.xmlopt); diff --git a/tests/bhyvexml2xmloutdata/x86_64/bhyvexml2xmlout-numa.xml b/tests/bhyvexml2xmloutdata/x86_64/bhyvexml2xmlout-numa.xml new file mode 100644 index 0000000000..ecc147db78 --- /dev/null +++ b/tests/bhyvexml2xmloutdata/x86_64/bhyvexml2xmlout-numa.xml @@ -0,0 +1,42 @@ +<domain type='bhyve'> + <name>bhyve</name> + <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> + <memory unit='KiB'>219136</memory> + <currentMemory unit='KiB'>219136</currentMemory> + <vcpu placement='static'>8</vcpu> + <os firmware='efi'> + <type arch='x86_64'>hvm</type> + <boot dev='hd'/> + </os> + <cpu> + <numa> + <cell id='0' cpus='0-3' memory='109568' unit='KiB'/> + <cell id='1' cpus='4-7' memory='109568' unit='KiB'/> + </numa> + </cpu> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <disk type='file' device='disk'> + <driver name='file' type='raw'/> + <source file='/tmp/freebsd.img'/> + <target dev='hda' bus='sata'/> + <address type='drive' controller='0' bus='0' target='2' unit='0'/> + </disk> + <controller type='pci' index='0' model='pci-root'/> + <controller type='isa' index='0'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x0'/> + </controller> + <controller type='sata' index='0'> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </controller> + <interface type='bridge'> + <mac address='52:54:00:b9:94:02'/> + <source bridge='virbr0'/> + <model type='virtio'/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> + </interface> + </devices> +</domain> diff --git a/tests/bhyvexml2xmltest.c b/tests/bhyvexml2xmltest.c index 7f9de2bc36..950aaea672 100644 --- a/tests/bhyvexml2xmltest.c +++ b/tests/bhyvexml2xmltest.c @@ -132,6 +132,7 @@ mymain(void) DO_TEST_DIFFERENT("passthru-multiple-devs"); DO_TEST_DIFFERENT("slirp"); DO_TEST_DIFFERENT("virtio-scsi"); + DO_TEST_DIFFERENT("numa"); /* Address allocation tests */ DO_TEST_DIFFERENT("addr-single-sata-disk"); -- 2.52.0

2 1

[PATCH] util: json: fix memory leak in virJSONValueFromJsonC()
by Elizaveta Tereshkina 06 Mar '26

06 Mar '26

In the 'json_type_array' case, the error path uses 'g_free()', which doesn't release other nested memory allocations. Replace it with 'virJSONValueFree()' to properly free the entire 'virJSONValue' structure. Fixes: da66bf53b0 (util: json: check return value of virJSONValueFromJsonC) Signed-off-by: Elizaveta Tereshkina <teryoshkina.ea(a)gmail.com> --- src/util/virjson.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/util/virjson.c b/src/util/virjson.c index 454bd657be..72bd78453a 100644 --- a/src/util/virjson.c +++ b/src/util/virjson.c @@ -1438,7 +1438,7 @@ virJSONValueFromJsonC(json_object *jobj) cur = virJSONValueFromJsonC(val); if (!cur) { - g_free(ret); + virJSONValueFree(ret); return NULL; } -- 2.39.2

2 1

[PATCH 0/2] qemu: hotplug: Fix use-after-free on media change via 'virsh attach-device'
by Peter Krempa 06 Mar '26

06 Mar '26

Reported via gitlab: https://gitlab.com/libvirt/libvirt/-/issues/859 I'll also later explore if it's possible to refactor the code to not touch the device definition wrappers after hotplug so that the pointer can be cleared properly when it's stolen into the domain definition. Peter Krempa (2): qemuDomainAttachDeviceDiskLive: Remove 'disk' variable qemu: hotplug: Don't access disk definititon after it was freed after media change src/qemu/qemu_hotplug.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) -- 2.53.0

2 3

[PATCH] util: json: fix memory leak in virJSONValueFromJsonC()
by Elizaveta Tereshkina 06 Mar '26

06 Mar '26

In the 'json_type_array' case, the error path uses 'g_free()', which doesn't release other nested memory allocations. Replace it with 'virJSONValueFree()' to properly free the entire 'virJSONValue' structure. Fixes: da66bf53b0 (util: json: check return value of virJSONValueFromJsonC) Signed-off-by: Elizaveta Tereshkina <teryoshkina.ea(a)gmail.com> --- src/util/virjson.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/util/virjson.c b/src/util/virjson.c index 454bd657be..5780dbb094 100644 --- a/src/util/virjson.c +++ b/src/util/virjson.c @@ -1438,7 +1438,7 @@ virJSONValueFromJsonC(json_object *jobj) cur = virJSONValueFromJsonC(val); if (!cur) { - g_free(ret); + virJSONValue(ret); return NULL; } -- 2.39.2

1 1

[PATCH 0/3] Add 'scalable_mode' and 'first_stage' attributes to intel IOMMU
by Zhenzhong Duan 06 Mar '26

06 Mar '26

Hi, This adds 'scalable_mode' and 'first_stage' attributes to intel IOMMU so that emulated device can run with first stage translation. This is based on Nathan Chen's iommufd series v2[1], so for passthrough device, nested tranlation is enabled automatically with below config: <iommu model='intel'> <driver caching_mode='on' iotlb='on' scalable_mode='on' first_stage='on'/> </iommu> Tests: Basic net card passthrough libvirt selftest Thanks Zhenzhong [1] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/KFYU… Zhenzhong Duan (3): qemu: Introduce scalable_mode and first_stage capabilities conf: add support for 'scalable_mode' and 'first_stage' attributes to intel IOMMU qemu: Add command line support for scalable mode and first stage of virtual vtd docs/formatdomain.rst | 8 +++++ src/conf/domain_conf.c | 35 ++++++++++++++++++- src/conf/domain_conf.h | 2 ++ src/conf/domain_validate.c | 6 ++++ src/conf/schemas/domaincommon.rng | 10 ++++++ src/qemu/qemu_capabilities.c | 4 +++ src/qemu/qemu_capabilities.h | 2 ++ src/qemu/qemu_command.c | 2 ++ src/qemu/qemu_validate.c | 12 +++++++ .../caps_10.0.0_x86_64+amdsev.xml | 2 ++ .../caps_10.0.0_x86_64.xml | 2 ++ .../caps_10.1.0_x86_64+inteltdx.xml | 2 ++ .../caps_10.1.0_x86_64.xml | 2 ++ .../caps_10.2.0_x86_64.xml | 2 ++ .../caps_6.2.0_x86_64.xml | 1 + .../caps_7.0.0_x86_64.xml | 1 + .../caps_7.1.0_x86_64.xml | 1 + .../caps_7.2.0_x86_64+hvf.xml | 1 + .../caps_7.2.0_x86_64.xml | 1 + .../caps_8.0.0_x86_64.xml | 1 + .../caps_8.1.0_x86_64.xml | 1 + .../caps_8.2.0_x86_64.xml | 1 + .../caps_9.0.0_x86_64.xml | 1 + .../caps_9.1.0_x86_64.xml | 1 + .../caps_9.2.0_x86_64+amdsev.xml | 1 + .../caps_9.2.0_x86_64.xml | 1 + 26 files changed, 102 insertions(+), 1 deletion(-) -- 2.47.3

3 6

[PATCH] hw/riscv: Remove deprecated 'riscv,delegate' device-tree property
by Philippe Mathieu-Daudé 05 Mar '26

05 Mar '26

The "riscv,delegate" DT property was added in QEMU 7.0 as part of the AIA APLIC support. The property changed name during the review process in Linux and the correct name ended up being "riscv,delegation". The incorrect name was added as alias, and deprecated in v9.1 (commit 38facfa8432), so can be removed for v11.0. Signed-off-by: Philippe Mathieu-Daudé <philmd(a)linaro.org> --- Binary file pc-bios/opensbi-riscv32-generic-fw_dynamic.bin matches Binary file pc-bios/opensbi-riscv64-generic-fw_dynamic.bin matches --- docs/about/deprecated.rst | 11 ----------- docs/about/removed-features.rst | 11 +++++++++++ hw/riscv/virt.c | 9 --------- 3 files changed, 11 insertions(+), 20 deletions(-) diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst index 1d5c4f3707c..cc5502e5394 100644 --- a/docs/about/deprecated.rst +++ b/docs/about/deprecated.rst @@ -494,17 +494,6 @@ versions, aliases will point to newer CPU model versions depending on the machine type, so management software must resolve CPU model aliases before starting a virtual machine. -RISC-V "virt" board "riscv,delegate" DT property (since 9.1) -'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' - -The "riscv,delegate" DT property was added in QEMU 7.0 as part of -the AIA APLIC support. The property changed name during the review -process in Linux and the correct name ended up being -"riscv,delegation". Changing the DT property name will break all -available firmwares that are using the current (wrong) name. The -property is kept as is in 9.1, together with "riscv,delegation", to -give more time for firmware developers to change their code. - x86 "isapc" board use of modern x86 CPUs (since 10.2) ''''''''''''''''''''''''''''''''''''''''''''''''''''' diff --git a/docs/about/removed-features.rst b/docs/about/removed-features.rst index 41bec8b8e38..0a62e87ee35 100644 --- a/docs/about/removed-features.rst +++ b/docs/about/removed-features.rst @@ -1232,6 +1232,17 @@ converting to using TCG plugins they should contact the qemu-devel mailing list. +Firmware, ACPI, Device Tree +--------------------------- + +RISC-V "virt" board "riscv,delegate" DT property (removed in 11.0) +'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' + +The "riscv,delegate" DT property was added in QEMU 7.0 as part of +the AIA APLIC support. The property changed name during the review +process in Linux and the correct name ended up being "riscv,delegation". + + System emulator devices ----------------------- diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c index bbce2fb6671..a1c323e66df 100644 --- a/hw/riscv/virt.c +++ b/hw/riscv/virt.c @@ -666,15 +666,6 @@ static void create_fdt_one_aplic(RISCVVirtState *s, int socket, qemu_fdt_setprop_cells(ms->fdt, aplic_name, "riscv,delegation", aplic_child_phandle, 0x1, VIRT_IRQCHIP_NUM_SOURCES); - /* - * DEPRECATED_9.1: Compat property kept temporarily - * to allow old firmwares to work with AIA. Do *not* - * use 'riscv,delegate' in new code: use - * 'riscv,delegation' instead. - */ - qemu_fdt_setprop_cells(ms->fdt, aplic_name, "riscv,delegate", - aplic_child_phandle, 0x1, - VIRT_IRQCHIP_NUM_SOURCES); } riscv_socket_fdt_write_id(ms, aplic_name, socket); -- 2.52.0

4 4

[PATCH 1/2] hyperv: improve error message when redefining domain
by Jonathon Jongsma 04 Mar '26

04 Mar '26

The current error message results in something like the following when running `virsh define` for an existing domain: `domain Domain already exists with UUID '$UUID' exists already` Improve the error message and make it behave like the esx driver and indicate that we do not yet support redefining existing domains in hyperv. Also avoid using the public LookupByUUID() API to check for existance, which requires unnecessarily allocating and de-allocating a virDomainPtr object. Signed-off-by: Jonathon Jongsma <jjongsma(a)redhat.com> --- src/hyperv/hyperv_driver.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/src/hyperv/hyperv_driver.c b/src/hyperv/hyperv_driver.c index b01b4919fe..6e9917f92a 100644 --- a/src/hyperv/hyperv_driver.c +++ b/src/hyperv/hyperv_driver.c @@ -2933,6 +2933,8 @@ hypervDomainDefineXML(virConnectPtr conn, const char *xml) virDomainPtr domain = NULL; g_autoptr(hypervInvokeParamsList) params = NULL; g_autoptr(GHashTable) defineSystemParam = NULL; + g_autoptr(Msvm_ComputerSystem) existing = NULL; + char uuid_string[VIR_UUID_STRING_BUFLEN]; size_t i = 0; /* parse xml */ @@ -2943,13 +2945,10 @@ hypervDomainDefineXML(virConnectPtr conn, const char *xml) goto error; /* abort if a domain with this UUID already exists */ - if ((domain = hypervDomainLookupByUUID(conn, def->uuid))) { - char uuid_string[VIR_UUID_STRING_BUFLEN]; - virUUIDFormat(domain->uuid, uuid_string); - virReportError(VIR_ERR_DOM_EXIST, _("Domain already exists with UUID '%1$s'"), uuid_string); - - // Don't use the 'exit' label, since we don't want to delete the existing domain. - virObjectUnref(domain); + virUUIDFormat(def->uuid, uuid_string); + if (hypervMsvmComputerSystemFromUUID(priv, uuid_string, &existing) == 0 && existing != NULL) { + virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", + _("Domain already exists, editing existing domains is not supported yet")); return NULL; } -- 2.53.0

3 4

[libvirt PATCH 0/3] qemu: IOMMUFD security label fix
by Pavel Hrdina 03 Mar '26

03 Mar '26

Pavel Hrdina (3): security_apparmor: Use g_auto* in AppArmorSetSecurityHostdevLabel security: Cleanup hostdev label error logic qemu: Fix IOMMUFD and VFIO security labels src/qemu/qemu_hotplug.c | 2 +- src/qemu/qemu_process.c | 16 +++-- src/qemu/qemu_process.h | 3 +- src/security/security_apparmor.c | 76 +++++++++-------------- src/security/security_dac.c | 102 ++++++++++++++----------------- src/security/security_selinux.c | 90 ++++++++++++--------------- 6 files changed, 130 insertions(+), 159 deletions(-) -- 2.53.0

3 7

[RFC 0/2] Fix build with libvirtd disabled
by Jim Fehlig 03 Mar '26

03 Mar '26

I'm considering dropping the monolithic libvirtd in future versions of SUSE distros, which got me to checking the usability of -Ddriver_libvirtd=disabled. I found it's not very usable :-). This series if an RFC attempt to rectify the situation. I'm not sure if we should build the various "common" components unconditionally as this change does, or only if the monolithic libvirtd and/or one or more modular daemons are also being built. Thanks for any comments! Jim Fehlig (2): build: Fix build with libvirtd disabled spec: Add option to enable/disable libvirtd docs/manpages/meson.build | 10 +- libvirt.spec.in | 25 +++- meson.build | 209 +++++++++++++--------------- meson_options.txt | 14 +- src/access/meson.build | 3 - src/locking/meson.build | 281 +++++++++++++++++++------------------- src/logging/meson.build | 104 +++++++------- src/meson.build | 200 +++++++++++++-------------- src/remote/meson.build | 174 +++++++++++------------ src/security/meson.build | 2 +- src/util/meson.build | 24 ++-- tools/meson.build | 26 ++-- 12 files changed, 526 insertions(+), 546 deletions(-) -- 2.51.0

3 5

[PATCH] secret: install service file only if init_script is 'systemd'
by Roman Bogorodskiy 03 Mar '26

03 Mar '26

Signed-off-by: Roman Bogorodskiy <bogorodskiy(a)gmail.com> --- src/secret/meson.build | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/src/secret/meson.build b/src/secret/meson.build index cfcc861f4f..32105dad9f 100644 --- a/src/secret/meson.build +++ b/src/secret/meson.build @@ -54,13 +54,15 @@ if conf.has('WITH_SECRETS') virt_secret_init_encryption_conf.set('localstatedir', localstatedir) - configure_file( - input: 'virt-secret-init-encryption.service.in', - output: '@0@.service'.format('virt-secret-init-encryption'), - configuration: virt_secret_init_encryption_conf, - install: true, - install_dir: unitdir, - ) + if init_script == 'systemd' + configure_file( + input: 'virt-secret-init-encryption.service.in', + output: '@0@.service'.format('virt-secret-init-encryption'), + configuration: virt_secret_init_encryption_conf, + install: true, + install_dir: unitdir, + ) + endif virt_daemon_units += { 'service': 'virtsecretd', -- 2.52.0

2 1

Release of libvirt-12.1.0
by Jiri Denemark 02 Mar '26

02 Mar '26

The 12.1.0 release of both libvirt and libvirt-python is tagged and signed tarballs are available at https://download.libvirt.org/ https://download.libvirt.org/python/ Thanks everybody who helped with this release by sending patches, reviewing, testing, or providing feedback. Your work is greatly appreciated. * New features * qemu: Advertise firmware features in domain capabilities XML The contents of the ``<firmwareFeatures/>`` element can be used to determine ahead of time whether a firmware matching certain characteristics, for example Secure Boot support, is available for the selected architecture and machine type. * qemu: Add support for uefi-vars device and firmware builds using it This is particularly noteworthy for people running aarch64 VMs with the 'virt' machine type, as it makes it finally possible to use Secure Boot with that combination. In most cases, no special steps are needed to take advantage of this: assuming that you have installed a recent version of QEMU, as well as a build of edk2 that includes the necessary binaries, you can just `enable Secure Boot <kbase/secureboot.html>`__ as you normally would. To explicitly request that the uefi-vars device is used even for scenarios where that would normally not be the case, it's enough to add an empty ``<varstore/>`` element in the domain XML. More details are available in the `guest firmware configuration <formatdomain.html#guest-firmware>`__ section of the documentation. * hyperv: improve API coverage for the hyperv driver The `virDomainInterfaceAddresses()` and `virDomainGetBlockInfo()` APIs are now supported by the hyperv driver. In addition, the domain xml for hyperv domains will indicate via firmware features whether secure boot is enabled. It also honors these firmware features when creating new domains. * bhyve: Add support for vCPU pinning configuration Bhyve guests can now have vCPU pinning configured:: <cputune> <vcpupin vcpu="0" cpuset="1,2,3"/> </cputune> Additionally, the ``domainGetVcpuPinInfo`` API is implemented for querying vCPU pinning information. * qemu: Support block operation latency histograms Libvirt now allows configuring qemu's block latency histogram collection as well as returns them via the bulk stats API. * Improvements * Introduce granule attribute for virtio-iommu In case when guest page size doesn't match the host page size (typically aarch64) the ``virtio-iommu`` needs to know the guest page size so it can allocate memory aligned to guest page size. * Parse hyperv features even for host-model Two releases ago, in v11.9.0 new ``host-model`` mode for Hyper-V enlightenments was introduced. Starting with this release, users can additionally override the defaults that are picked when domain is started and features are expanded. * bhyve: Improve loader configuration for arm64 guests If loader is not explicitly configured, use the loader from the ``sysutils/u-boot-bhyve-arm64`` port/package for the arm64 guests. * Bug fixes * Fix build with remote driver disabled Some parts of code were wrongly annotated as depended on remote driver. But they were used even from client side drivers. This is now fixed and libvirt builds properly even with remote driver disabled. * Various fixes to libvirt-guests.sh Firstly, the exit code of various commands was ignored (which may lead the script to wrongly determine persistent/transient domain state, for instance). Secondly, due to logical error, the script might have incorrectly asses state a domain is in. * AppArmor: Ask for no deny rule for readonly disk elements For read only disks, libvirt created an AppArmor profile which disallowed any future write rules. But when doing a blockcommit, libvirt needs to allow hypervisor to write to even readonly disks. The rule in the profile was changed so that future write rules can be added, temporarily. * esx: Allow connecting to IPv6 server Due to a bug in our code, if an IPv6 address was provided in connection URI, libvirt would fail to connect to VMWare server. This is now fixed. * qemu: Use device alias if interface has no name The ``virDomainInterfaceAddresses()`` API (or ``virsh domifaddr``) returns an array interfaces among with their addresses. But some interface names might be unknown, for instance if the API is told to parse host's ARP table then PCI assigned NICs or slirp/passt lack interface name. If that's the case, let the API return domain's ``<interface/>`` alias. * bhyve: hyperv: Various memory leak fixes * qemu: Fix failures when restoring save/managed-save images with upcoming qemu versions Current git version of qemu would return an error when attempting to load an existing (managed) save image as we relied on deprecated features that were now removed. Enjoy. Jirka

1 0