May 2026 - Devel - libvirt List Archives

[PATCH v3 0/5] qemu: support async live vCPU unplug
by Akash Kulhalli 12 May '26

12 May '26

This series adds support for asynchronous live vCPU unplug in the QEMU driver. For live vCPU unplug libvirt currently waits only for a short synchronous completion window. If the unplug completes later there is no dedicated event reporting successful removal of the vCPU. This series adds a new domain event for successful unplug completion and extends the public APIs to allow requesting asynchronous unplug. As recommended, this changeset is split into five parts- Patch 1 adds the new event and the required public/internal plumbing. Patch 2 wires the event into the existing QEMU unplug completion path. Patch 3 adds the internal QEMU plumbing needed to skip waiting for unplug completion, but leaves it unused for now. Patches 4 and 5 then expose the new behaviour through virDomainSetVcpusFlags() and virDomainSetVcpu() together with the corresponding virsh changes. The new public flags are limited to live unplug. Rejected unplug requests continue to be reported via device-removal-failed. Most impactful change from the previous iteration is that event emission moves from 1 vcpu-removed per qemu event model to one vcpu-removed per vcpu actually removed. This change is effectively a no-op for x86 (because there is a 1:1 mapping), but is relevant for architectures where more than 1 vcpu might be grouped under a single logical qemu vcpu object. Changes from v2: - s/virDomainSetVcpuBehaviour/virDomainSetVcpuFlags/g - Reduce redundant phrasing - emit `vcpu-removed` for every vcpu removed instead of per qemu event - bump version ton 12.4.0 All test runs are still passing cleanly. Akash Kulhalli (5): API/qemu: add async unplug flag to virDomainSetVcpu API/qemu: add async unplug flag to virDomainSetVcpusFlags qemu: thread async vcpu unplug through internal helpers qemu: emit vcpu-removed event on unplug completion conf,remote: add vcpu-removed domain event Related discussion: RFC: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/2J53… v2: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/YZID…

2 10

[PATCH 1/1] virStorageSource: note that capacity/allocation/physical are stale caches
by Denis V. Lunev 12 May '26

12 May '26

These three fields are cached values that do not reflect reality unless the caller refreshes them. 'allocation' is in addition ultra-unreliable: any guest write into a previously unallocated part of a sparse image invalidates it, even right after a refresh. Document this on the struct so new callers do not trust the values. Signed-off-by: Denis V. Lunev <den(a)openvz.org> CC: Peter Krempa <pkrempa(a)redhat.com> --- src/conf/storage_source_conf.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h index 5ddcebb282..fa71a9ef39 100644 --- a/src/conf/storage_source_conf.h +++ b/src/conf/storage_source_conf.h @@ -326,6 +326,8 @@ struct _virStorageSource { virStoragePerms *perms; virStorageTimestamps *timestamps; + + /* Cached, unreliable. Refresh before use. */ unsigned long long capacity; /* in bytes, 0 if unknown */ unsigned long long allocation; /* in bytes, 0 if unknown */ unsigned long long physical; /* in bytes, 0 if unknown */ -- 2.51.0

2 1

[libvirt PATCH] util: virGetSubIDs: do not limit file size
by Ján Tomko 12 May '26

12 May '26

From: Ján Tomko <jtomko(a)redhat.com> https://redhat.atlassian.net/browse/RHEL-174491 Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- src/util/virutil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/util/virutil.c b/src/util/virutil.c index 187b8202dd..2cc7edac50 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -1223,7 +1223,7 @@ virGetSubIDs(virSubID **retval, const char *file) *retval = NULL; - if (virFileReadAll(file, BUFSIZ, &buf) < 0) + if (virFileReadAll(file, SIZE_MAX, &buf) < 0) return -1; lines = g_strsplit(buf, "\n", 0); -- 2.54.0

2 1

[libvirt PATCH v2] util: virGetSubIDs: do not limit file size
by Ján Tomko 12 May '26

12 May '26

From: Ján Tomko <jtomko(a)redhat.com> https://redhat.atlassian.net/browse/RHEL-174491 Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- v2: it's `int maxlen`, not size_t src/util/virutil.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/util/virutil.c b/src/util/virutil.c index 187b8202dd..24b0a10878 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -1223,7 +1223,7 @@ virGetSubIDs(virSubID **retval, const char *file) *retval = NULL; - if (virFileReadAll(file, BUFSIZ, &buf) < 0) + if (virFileReadAll(file, INT_MAX, &buf) < 0) return -1; lines = g_strsplit(buf, "\n", 0); -- 2.54.0

1 0

[PATCH] test: qemu: Fix error message when populating fd groups
by Peter Krempa 12 May '26

12 May '26

From: Peter Krempa <pkrempa(a)redhat.com> The 'new->fds' array is not yet initialized at the point where the check if the FD is occupied happens so the error would always report that FD '0' is in use. Use 'new->testfds' instead. Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- tests/testutilsqemu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/testutilsqemu.c b/tests/testutilsqemu.c index e9bdbdbbe7..e7a61d0c6f 100644 --- a/tests/testutilsqemu.c +++ b/tests/testutilsqemu.c @@ -727,7 +727,7 @@ testQemuInfoSetArgs(testQemuInfo *info, new->testfds[i] = va_arg(argptr, unsigned int); if (fcntl(new->testfds[i], F_GETFD) != -1) { - fprintf(stderr, "fd '%d' is already in use\n", new->fds[i]); + fprintf(stderr, "fd '%d' is already in use\n", new->testfds[i]); abort(); } -- 2.54.0

2 1

[PATCH 0/2] Prohibit newlines in VIR_WARN strings
by Peter Krempa 12 May '26

12 May '26

Make them the same as we do with error messages since they often end up in logs. Since they are not translated [1] the existing check didn't catch that. [1] IMO in some cases they are in fact used as errors, and maybe would be worth to be translated. But this is for another series Peter Krempa (2): Don't break up strings for VIR_WARN messages syntax-check: Enforce no linebreaks in VIR_WARN messages build-aux/syntax-check.mk | 8 +++++ src/bhyve/bhyve_command.c | 4 +-- src/conf/virdomainjob.c | 5 +-- src/cpu/cpu_x86.c | 5 ++- src/esx/esx_vi.c | 3 +- src/hypervisor/virhostdev.c | 12 +++---- src/interface/interface_backend_netcf.c | 12 +++---- src/libxl/libxl_conf.c | 4 +-- src/libxl/xen_xl.c | 4 +-- src/lxc/lxc_driver.c | 3 +- src/lxc/lxc_process.c | 3 +- src/network/bridge_driver.c | 15 +++------ src/nwfilter/nwfilter_dhcpsnoop.c | 3 +- src/qemu/qemu_block.c | 3 +- src/qemu/qemu_command.c | 3 +- src/qemu/qemu_conf.c | 4 +-- src/qemu/qemu_domain.c | 44 +++++++++---------------- src/qemu/qemu_domain_address.c | 10 ++---- src/qemu/qemu_driver.c | 10 +++--- src/qemu/qemu_firmware.c | 6 ++-- src/qemu/qemu_hotplug.c | 3 +- src/qemu/qemu_migration.c | 11 +++---- src/qemu/qemu_monitor_json.c | 4 +-- src/remote/remote_driver.c | 3 +- src/rpc/virnetserverclient.c | 5 ++- src/security/security_dac.c | 3 +- src/security/security_selinux.c | 7 ++-- src/security/security_stack.c | 34 +++++++------------ src/storage/storage_backend_rbd.c | 4 +-- src/storage/storage_backend_zfs.c | 3 +- src/storage/storage_util.c | 4 +-- src/storage_file/storage_file_probe.c | 3 +- src/vbox/vbox_common.c | 4 +-- src/vmx/vmx.c | 6 ++-- src/vz/vz_sdk.c | 3 +- 35 files changed, 96 insertions(+), 162 deletions(-) -- 2.54.0

2 4

[PATCH 1/1] qemu: refresh disk->src->capacity after block_resize
by Denis V. Lunev 12 May '26

12 May '26

qemuDomainBlockResize() issues block_resize via QMP, which updates the qcow2 header on disk and bs->total_sectors in qemu, but the cached disk->src->capacity in libvirt is left at its pre-resize value. All current readers re-read capacity from either QMP or the file itself before returning to the user, so the staleness is not externally observable today. Still, the cached field should reflect what we just asked qemu to write; keeping the two in sync avoids surprises for any future consumer that trusts the cached value. Signed-off-by: Denis V. Lunev <den(a)openvz.org> CC: Peter Krempa <pkrempa(a)redhat.com> CC: Michal Privoznik <mprivozn(a)redhat.com> --- src/qemu/qemu_driver.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 2d31d4aa31..a419c5cc95 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -9598,6 +9598,9 @@ qemuDomainBlockResize(virDomainPtr dom, } qemuDomainObjExitMonitor(vm); + /* Keep the cached capacity in sync with qemu representation. */ + disk->src->capacity = size; + if (persistDisk) { g_clear_pointer(&persistDisk->src->sliceStorage, virStorageSourceSliceFree); qemuDomainSaveConfig(vm); -- 2.51.0

3 3

RFC: esx_driver -- vSphere SOAP API (possibly incompatible) upgrade
by Martin Kletzander 11 May '26

11 May '26

Hello, so we found an issue in the ESX driver which is pretty easy to solve, but requires an update to the communication with the server. The newer VI API allows for the instanceUuid parameter when searching for VMs, but only since 4.0, but because we do not send any "SOAPAction:" header it behaves as if we used the oldest one. If we add the header "SOAPAction: urn:vim25/4.0" (at least 4.0) it works, but I'm not sure if there are any incompatibilities and whether there is a use case for older clusters that do not support at least this version. If there is a reasonable use case for it (well, anyone saying "here, me, I'd like to keep it") we could change the behaviour based on the version we parse when connecting to the server. But that's an extra bit of mess, although not totally big one. So I would like to know your opinion on this. Adding few people into Cc who found about the issue. Have a nice day, Martin

3 4

[PATCH v2 00/35] qemu: Add infrastructure to prevent accidental disk shrink via 'virDomainBlockResize' (part 2)
by Peter Krempa 07 May '26

07 May '26

v2: - possibility to query flags for every API at once A comment in v1 suggested that polling flags one by one is ineffective. It was also limited to virDomainBlockResize. This version adds an API 'virConnectGetIntrospection' which allows to probe for every flag of every API at once via the introspection XML which also informs users about supported typed parameters. Patches 1-34 are new in the series. Patch 35 is adapted to use the new introspection API. This is partially a RFC: - documentation and schema for the new XML will be contributed later - I've contemplated adding typed parameters as input for the new API if we'd want to extend it to e.g. allow probing supported stuff for e.g. a hypervisor version Breakdown of patches: Semi-unrelated refactors/fixes: remote_protocol-structs: fix mis-aligned 'remote_domain_set_throttle_group_args' qemu: driver: Unify coding style util: typedparam: Convert VIR_TYPED_PARAM_CHECK_TYPE into a function util: virTypedParamValidateType: Don't report unknown typed parameter type as '(null)' util: Replace open-coded internals of VIR_TYPED_PARAMS_DEBUG with 'virTypedParamDebugstr' util: typedparam: Unexport virTypedParameterTypeFromString/virTypedParameterTypeToString util: typedparam: Refactor and fix typed param validation scripts: check-symfile: Allow also symbols in 'readonly' section virHostCPUGet(Map|Stats): Remove unused 'flags' virHostMemGet(Stats|Parameters): Remove unused 'flags' virNodeSuspend: Remove unused 'flags' Changes to allow the introspection script to parse flags: qemu: backup: Move 'virCheckFlags' to top level functions qemu: attach/detach device: Move 'virCheckFlags' to top level functions qemu: dump: Move 'virCheckFlags' to top level functions qemu: checkpoint: Move 'virCheckFlags' to top level functions qemuDomainRestore(Flags|Params): Refactor flag checking qemu: snapshot: Move flag checks to top level functions qemuDomainGetLaunchSecurityInfo: Move flag check to top level qemuDomainDetachDeviceAlias: Move 'flags' validation to top level qemuDomainManagedSaveDefineXML: Add top-level flag validation qemuDomainQemuMonitorCommand: Add top-level flag validation qemuConnectDomainQemuMonitorEventRegister: Add top-level flag validation qemuDomainGetMetadata: Add top-level flag validation API addition and implementation of the introspection api supporting flag introspection: API: Introduce 'virConnectGetIntrospection' virsh: Introduce 'introspection' command which uses 'virConnectGetIntrospection' scripts: Introduce 'getintrospection' script qemu: Implement 'virConnectGetIntrospection' Cleanups needed in order to support typed parameter introspection util: typedparam: Introduce 'virTypedParamsValidateTemplate' qemu: migration: Use 'virTypedParamsValidateTemplate' for migration params qemu: Reimplement 'qemuDomainValidateBlockIoTune' using 'virTypedParamsValidateTemplate' util: hostmem: Make parameters for 'virHostMemSetParameters' introspectable qemuDomainSetSchedulerParameters: Make typed parameters introspectable qemuDomainSetIOThreadParams: Move typed parameter validation to top level Implementation of typed parameter introspection: introspection: Add introspection of input typed parameters Original patch to prevent shrink in virsh: virsh: blockresize: Use VIR_DOMAIN_BLOCK_RESIZE_EXTEND when available and introduce --allow-shrink docs/manpages/virsh.rst | 30 +- include/libvirt/libvirt-host.h | 3 + scripts/check-symfile.py | 2 +- scripts/genintrospection.py | 347 ++++++++++++++++++++ scripts/meson.build | 1 + src/bhyve/bhyve_driver.c | 20 +- src/ch/ch_driver.c | 11 +- src/conf/domain_conf.c | 3 +- src/conf/domain_conf.h | 4 + src/conf/domain_event.c | 4 +- src/conf/domain_event.h | 5 + src/driver-hypervisor.h | 5 + src/libvirt-host.c | 40 +++ src/libvirt_private.syms | 10 +- src/libvirt_public.syms | 5 + src/lxc/lxc_driver.c | 24 +- src/openvz/openvz_driver.c | 12 +- src/qemu/meson.build | 18 ++ src/qemu/qemu_backup.c | 8 +- src/qemu/qemu_backup.h | 3 +- src/qemu/qemu_checkpoint.c | 11 - src/qemu/qemu_driver.c | 543 ++++++++++++++++++++------------ src/qemu/qemu_migration.c | 29 ++ src/qemu/qemu_migration.h | 32 +- src/qemu/qemu_snapshot.c | 27 -- src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 15 +- src/remote_protocol-structs | 21 +- src/util/meson.build | 1 + src/util/virhostcpu.c | 10 +- src/util/virhostcpu.h | 6 +- src/util/virhostmem.c | 53 ++-- src/util/virhostmem.h | 11 +- src/util/virintrospection.c | 82 +++++ src/util/virintrospection.h | 20 ++ src/util/virnodesuspend.c | 5 +- src/util/virnodesuspend.h | 3 +- src/util/virtypedparam-public.c | 54 ++-- src/util/virtypedparam.c | 222 +++++++++---- src/util/virtypedparam.h | 26 +- src/vz/vz_driver.c | 12 +- tests/virhostcputest.c | 2 +- tools/virsh-domain.c | 59 ++++ tools/virsh-host.c | 43 +++ 44 files changed, 1384 insertions(+), 459 deletions(-) create mode 100644 scripts/genintrospection.py create mode 100644 src/util/virintrospection.c create mode 100644 src/util/virintrospection.h -- 2.54.0

3 44

[PATCH 0/6] tests: Detect for unused files
by Michal Privoznik 07 May '26

07 May '26

One of the functionality implemented into qemuxmlconftest is to detect unused test files. This generalizes the idea and implements it into two other tests (networkxmlconftest and cputest). But if we find this useful, it can be implemented into the rest of the tests. Now, I was contemplating an idea of turning `existingTestCases` variable into a global one (just like we have testDebug, testVerbose, etc.). It would make the changes a lot cleaner but also it's a global variable after all. What do you think? Michal Prívozník (6): testutils: Introduce unused file detection qemuxmlconftest: Switch to virTestEnumerateTestCases() networkxmlconfdata: Remove passthrough-pf.conf networkxmlconftest: Detect unused files cputestdata: Drop unused files cputest: Detect unused files tests/cputest.c | 103 ++++++++++++++---- tests/cputestdata/ppc64-guest-host-model.xml | 3 - .../ppc64-host+guest-host-model.xml | 3 - tests/cputestdata/x86_64-bogus-vendor.xml | 4 - tests/networkxmlconfdata/passthrough-pf.conf | 11 -- tests/networkxmlconftest.c | 35 +++++- tests/qemuxmlconftest.c | 82 ++++---------- tests/testutils.c | 58 ++++++++++ tests/testutils.h | 14 +++ 9 files changed, 207 insertions(+), 106 deletions(-) delete mode 100644 tests/cputestdata/ppc64-guest-host-model.xml delete mode 100644 tests/cputestdata/ppc64-host+guest-host-model.xml delete mode 100644 tests/cputestdata/x86_64-bogus-vendor.xml delete mode 100644 tests/networkxmlconfdata/passthrough-pf.conf -- 2.53.0

2 7