November 2024 - Devel - libvirt List Archives

[PATCH v2 0/5] Allow Guest CPU Model with Deprecated Features Disabled
by Collin Walling 09 Dec '24

09 Dec '24

# Changelog v2 - rebased on latest master changes # Description Add support for libvirt to query and cache an array of deprecated CPU features (aka CPU properties) for the host-model. This data is queried via a full query-cpu-model-expansion and cached in the QEMU capabilities file. This model expansion will depend on the availability of the "deprecated-props" field resulting from a query-cpu-model-expansion command. Currently, only s390x supports this field. The purpose of these patches is to make it easy for users to create guests with a CPU model that will be compatible & migratable with future hardware. An updated host CPU model with deprecated features paired with the policy "disable" may be visable via an update to the virsh domcapabilities command with the --disable-deprecated-features flag. An example is shown below. Note: other CPU model queries (e.g. baseline and comparison) currently do not consider deprecated features, as their outputs do not consider feature policy. If implementation is desired, it will require a discussion on how these commands should report policies. Examples: virsh domcapabilities --disable-deprecated-features e.g. output (trimmed): <mode name='host-model' supported='yes'> <model fallback='forbid'>z14.2-base</model> <feature policy='require' name='aen'/> <feature policy='require' name='cmmnt'/> <feature policy='require' name='aefsi'/> <feature policy='require' name='diag318'/> <feature policy='require' name='mepoch'/> <feature policy='require' name='msa8'/> <feature policy='require' name='msa7'/> <feature policy='require' name='msa6'/> <feature policy='require' name='msa5'/> <feature policy='require' name='msa4'/> <feature policy='require' name='msa3'/> <feature policy='require' name='msa2'/> <feature policy='require' name='msa1'/> <feature policy='require' name='sthyi'/> <feature policy='require' name='edat'/> <feature policy='require' name='ri'/> <feature policy='require' name='edat2'/> <feature policy='require' name='etoken'/> <feature policy='require' name='vx'/> <feature policy='require' name='ipter'/> <feature policy='require' name='mepochptff'/> <feature policy='require' name='ap'/> <feature policy='require' name='vxeh'/> <feature policy='require' name='vxpd'/> <feature policy='require' name='esop'/> <feature policy='require' name='apqi'/> <feature policy='require' name='apft'/> <feature policy='require' name='els'/> <feature policy='require' name='iep'/> <feature policy='require' name='apqci'/> <feature policy='disable' name='cte'/> <feature policy='require' name='ais'/> <feature policy='disable' name='bpb'/> <feature policy='require' name='ctop'/> <feature policy='require' name='gs'/> <feature policy='require' name='ppa15'/> <feature policy='require' name='zpci'/> <feature policy='require' name='sea_esop2'/> <feature policy='disable' name='te'/> <feature policy='require' name='cmm'/> <feature policy='disable' name='csske'/> </mode> A domain may be defined with a new <cpu> XML attribute, deprecated_features='on|off': <cpu mode='host-model' check='partial' deprecated_features='off'/> e.g. after guest has started (trimmed): <cpu mode='custom' match='exact' check='partial' deprecated_features='off'> <model fallback='forbid'>z14.2-base</model> <feature policy='require' name='aen'/> <feature policy='require' name='cmmnt'/> <feature policy='require' name='aefsi'/> <feature policy='require' name='diag318'/> <feature policy='require' name='mepoch'/> <feature policy='require' name='msa8'/> <feature policy='require' name='msa7'/> <feature policy='require' name='msa6'/> <feature policy='require' name='msa5'/> <feature policy='require' name='msa4'/> <feature policy='require' name='msa3'/> <feature policy='require' name='msa2'/> <feature policy='require' name='msa1'/> <feature policy='require' name='sthyi'/> <feature policy='require' name='edat'/> <feature policy='require' name='ri'/> <feature policy='require' name='edat2'/> <feature policy='require' name='etoken'/> <feature policy='require' name='vx'/> <feature policy='require' name='ipter'/> <feature policy='require' name='mepochptff'/> <feature policy='require' name='ap'/> <feature policy='require' name='vxeh'/> <feature policy='require' name='vxpd'/> <feature policy='require' name='esop'/> <feature policy='require' name='apqi'/> <feature policy='require' name='apft'/> <feature policy='require' name='els'/> <feature policy='require' name='iep'/> <feature policy='require' name='apqci'/> <feature policy='disable' name='cte'/> <feature policy='require' name='ais'/> <feature policy='disable' name='bpb'/> <feature policy='require' name='ctop'/> <feature policy='require' name='gs'/> <feature policy='require' name='ppa15'/> <feature policy='require' name='zpci'/> <feature policy='require' name='sea_esop2'/> <feature policy='disable' name='te'/> <feature policy='require' name='cmm'/> <feature policy='disable' name='csske'/> </cpu> Collin Walling (5): qemuMonitorJSONGetCPUModelExpansion: refactor parsing functions qemu: parse deprecated-props from query-cpu-model-expansion response qemu_capabilities: query deprecated features for host-model virsh: add --disable-deprecated-features flag to domcapabilities conf: add deprecated_features attribute docs/manpages/virsh.rst | 6 + include/libvirt/libvirt-domain.h | 12 + src/conf/cpu_conf.c | 10 + src/conf/cpu_conf.h | 1 + src/conf/schemas/cputypes.rng | 12 + src/libvirt-domain.c | 2 +- src/qemu/qemu_capabilities.c | 89 +++++ src/qemu/qemu_capabilities.h | 4 + src/qemu/qemu_driver.c | 8 +- src/qemu/qemu_monitor.c | 10 + src/qemu/qemu_monitor.h | 1 + src/qemu/qemu_monitor_json.c | 64 +++- src/qemu/qemu_process.c | 4 + .../caps_9.1.0_s390x.replies | 348 +++++++++++++++++- .../qemucapabilitiesdata/caps_9.1.0_s390x.xml | 13 + .../caps_9.2.0_s390x.replies | 348 +++++++++++++++++- .../qemucapabilitiesdata/caps_9.2.0_s390x.xml | 13 + tools/virsh-host.c | 9 +- 18 files changed, 940 insertions(+), 14 deletions(-) -- 2.45.1

3 16

[PATCH 0/5] cpu_map: Add missing -v1 models
by Jiri Denemark 05 Dec '24

05 Dec '24

CPU models that do not have a list of versions attached are still advertised as aliases to corresponding -v1 variants. We should add the missing variants to the CPU map. Available on gitlab: git fetch https://gitlab.com/jirkade/libvirt.git cpu-versions Pipeline: https://gitlab.com/jirkade/libvirt/-/pipelines/1564399375 Jiri Denemark (5): cpu_map: Sort data files in meson.build sync_qemu_models_i386: Update meson.build sync_qemu_models_i386: Ignore old models sync_qemu_models_i386: Generate missing -v1 variants cpu_map: Add missing -v1 models src/cpu_map/index.xml | 2 + src/cpu_map/meson.build | 18 +++--- src/cpu_map/sync_qemu_models_i386.py | 58 +++++++++++++++++++ src/cpu_map/x86_EPYC-Genoa-v1.xml | 6 ++ src/cpu_map/x86_KnightsMill-v1.xml | 6 ++ .../domaincapsdata/qemu_5.2.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_5.2.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_5.2.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.0.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.0.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_6.0.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.1.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.1.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_6.1.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.2.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_6.2.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_6.2.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_7.0.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_7.0.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_7.0.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_7.1.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_7.1.0-tcg.x86_64.xml | 20 ++++++- tests/domaincapsdata/qemu_7.1.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_7.2.0-q35.x86_64.xml | 14 ++++- .../qemu_7.2.0-tcg.x86_64+hvf.xml | 16 ++++- .../domaincapsdata/qemu_7.2.0-tcg.x86_64.xml | 16 ++++- tests/domaincapsdata/qemu_7.2.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_8.0.0-q35.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_8.0.0-tcg.x86_64.xml | 16 ++++- tests/domaincapsdata/qemu_8.0.0.x86_64.xml | 14 ++++- .../domaincapsdata/qemu_8.1.0-q35.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_8.1.0-tcg.x86_64.xml | 56 +++++++++++++++++- tests/domaincapsdata/qemu_8.1.0.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_8.2.0-q35.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_8.2.0-tcg.x86_64.xml | 55 +++++++++++++++++- tests/domaincapsdata/qemu_8.2.0.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.0.0-q35.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.0.0-tcg.x86_64.xml | 54 ++++++++++++++++- tests/domaincapsdata/qemu_9.0.0.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.1.0-q35.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.1.0-tcg.x86_64.xml | 54 ++++++++++++++++- tests/domaincapsdata/qemu_9.1.0.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.2.0-q35.x86_64.xml | 46 ++++++++++++++- .../domaincapsdata/qemu_9.2.0-tcg.x86_64.xml | 54 ++++++++++++++++- tests/domaincapsdata/qemu_9.2.0.x86_64.xml | 46 ++++++++++++++- 45 files changed, 1152 insertions(+), 63 deletions(-) create mode 100644 src/cpu_map/x86_EPYC-Genoa-v1.xml create mode 100644 src/cpu_map/x86_KnightsMill-v1.xml -- 2.47.0

3 10

[PATCH v3 0/5] ch: handle events from cloud-hypervisor
by Purna Pavan Chandra Aekkaladevi 03 Dec '24

03 Dec '24

changes from v2->v3: * Remove patch 'utils: Implement virFileIsNamedPipe' as it is no more needed. * Remove the eventmonitorpath only if it exists * Added domain name as a prefix to logs from ch_events.c. This will make debugging easier. * Simplified event parsing logic by reserving a byte for null char. changes from v1->v2: * Rebase on latest master * Use /* */ for comments * Remove fifo file if already exists * Address other comments from Praveen Paladugu cloud-hypervisor raises various events, including VM lifecylce operations such as boot, shutdown, pause, resume, etc. Libvirt will now read these events and take the necessary actions, such as correctly updating the domain state. A FIFO file is passed to `--event-monitor` option of cloud-hypervisor. Libvirt creates a new thread that acts as the reader of the fifo file and continuously monitors for new events. Currently, shutdown events are handled by updating the domain state appropriately. Purna Pavan Chandra Aekkaladevi (5): ch: pass --event-monitor option to cloud-hypervisor ch: start a new thread for handling ch events ch: events: Read and parse cloud-hypervisor events ch: events: facilitate lifecycle events handling NEWS: Mention event handling support in ch driver NEWS.rst | 7 + po/POTFILES | 1 + src/ch/ch_events.c | 329 ++++++++++++++++++++++++++++++++++++++++++++ src/ch/ch_events.h | 54 ++++++++ src/ch/ch_monitor.c | 52 ++++++- src/ch/ch_monitor.h | 11 ++ src/ch/meson.build | 2 + 7 files changed, 449 insertions(+), 7 deletions(-) create mode 100644 src/ch/ch_events.c create mode 100644 src/ch/ch_events.h -- 2.34.1

3 14

[PATCH] qemu:qemu_snapshot: Fix a libvirtd cransh when delete snapshot
by jungle man 29 Nov '24

29 Nov '24

qemuDomainDiskByName() can return a NULL pointer on failure, but this returned value in qemuSnapshotDeleteValidate is not checked. It will make libvirtd crash. diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c index 5b3aadcbf0..52312b4a7b 100644 --- a/src/qemu/qemu_snapshot.c +++ b/src/qemu/qemu_snapshot.c @@ -4235,8 +4235,11 @@ qemuSnapshotDeleteValidate(virDomainObj *vm, virDomainDiskDef *vmdisk = NULL; virDomainDiskDef *disk = NULL; - vmdisk = qemuDomainDiskByName(vm->def, snapDisk->name); - disk = qemuDomainDiskByName(snapdef->parent.dom, snapDisk->name); + if (!(vmdisk = qemuDomainDiskByName(vm->def, snapDisk->name))) + return -1; + + if (!(disk = qemuDomainDiskByName(snapdef->parent.dom, snapDisk->name))) + return -1; if (!virStorageSourceIsSameLocation(vmdisk->src, disk->src)) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED,

2 2

[PATCH] qemu: enable QEMU core dump by default on Linux
by Daniel P. Berrangé 29 Nov '24

29 Nov '24

The Linux MADV_DONTDUMP flag was added to Linux kernels > 3.3, back in 2012, and the dump-guest-core flag was added to QEMU > 1.0 at the same time. IOW, on Linux we have long been able to assume that QEMU core dumps will exclude guest memory, unless the user has overridden the host level defaults in the domain XML. It is desirable to permit QEMU core dumps out of the box to make it easier for users to report crashes to their OS vendor without having to reconfigure and restart libvirt daemons and their running guests. While there is a risk that an admin may have set 'dump_guest_core' to true, while leaving 'max_core' to 0, on balance the benefits of easier troubleshooting outweigh the risk of changing the defaults to permit core dumps. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- src/qemu/qemu.conf.in | 8 +++++--- src/qemu/qemu_conf.c | 10 +++++++--- 2 files changed, 12 insertions(+), 6 deletions(-) diff --git a/src/qemu/qemu.conf.in b/src/qemu/qemu.conf.in index 89c9b6b913..d853136f10 100644 --- a/src/qemu/qemu.conf.in +++ b/src/qemu/qemu.conf.in @@ -709,7 +709,9 @@ # As a special case it can be set to the string "unlimited" to # to allow arbitrarily sized core dumps. # -# By default the core dump size is set to 0 disabling all dumps +# By default the core dump size is set to unlimited on +# Linux where 'dump_guest_core' defaults to false, and +# is set to 0 on other platforms, disabling all dumps # # Size is a positive integer specifying bytes or the # string "unlimited" @@ -717,8 +719,8 @@ #max_core = "unlimited" # Determine if guest RAM is included in QEMU core dumps. By -# default guest RAM will be excluded if a new enough QEMU is -# present and host kernel supports it. Setting this to '1' will +# default guest RAM will be excluded on Linux platforms, +# and included on all other patforms. Setting this to '1' will # force guest RAM to always be included in QEMU core dumps. # # This setting will be ignored if the guest XML has set the diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 7c15c521c7..8b9fe4e381 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -291,10 +291,14 @@ virQEMUDriverConfig *virQEMUDriverConfigNew(bool privileged, cfg->deprecationBehavior = g_strdup("none"); cfg->storageUseNbdkit = USE_NBDKIT_DEFAULT; -#ifndef MADV_DONTDUMP +#ifdef MADV_DONTDUMP + cfg->maxCore = ULLONG_MAX; +#else /* QEMU uses Linux extensions to madvise() (MADV_DODUMP/MADV_DONTDUMP) to - * include/exclude guest memory from core dump. These might be unavailable - * on some systems. Provide sane default. */ + * include/exclude guest memory from core dump. On non-Linux systems + * core dumps will unavoidably include all guest RAM, so toggle the + * default to reflect this and warn the admin. + */ VIR_INFO("Host kernel doesn't support MADV_DONTDUMP. Enabling dump_guest_core"); cfg->dumpGuestCore = true; #endif -- 2.46.0

2 1

[PATCH] docs: Document how to enable coredumps for session qemu
by Richard W.M. Jones 29 Nov '24

29 Nov '24

Thanks: Daniel P. Berrangé Signed-off-by: Richard W.M. Jones <rjones(a)redhat.com> --- docs/kbase/qemu-core-dump.rst | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/docs/kbase/qemu-core-dump.rst b/docs/kbase/qemu-core-dump.rst index 5c3781cf90..4bd3ff0b9d 100644 --- a/docs/kbase/qemu-core-dump.rst +++ b/docs/kbase/qemu-core-dump.rst @@ -53,6 +53,13 @@ that may get interrupted. In particular long running jobs like live migration or block device copy jobs may abort. It is thus wise to check that the host is mostly idle before restarting the daemons. +Session mode qemu +================= + +For session mode (non-root) qemu, you have to create a new file called +``$HOME/.config/libvirt/qemu.conf``, add the line(s) above, and kill +the user ``virtqemud`` instance. + Guest core dump configuration ============================= -- 2.46.0

1 0

Entering freeze for libvirt-10.10.0
by Jiri Denemark 28 Nov '24

28 Nov '24

I have just tagged v10.10.0-rc1 in the repository and pushed signed tarballs to https://download.libvirt.org/ Please give the release candidate some testing and in case you find a serious issue which should have a fix in the upcoming release, feel free to reply to this thread to make sure the issue is more visible. If you have not done so yet, please update NEWS.rst to document any significant change you made since the last release. Thanks, Jirka

1 1

[PATCH] NEWS: Document features/improvements/bug fixes I've participated in
by Michal Privoznik 28 Nov '24

28 Nov '24

There are some features/improvements/bug fixes I've either contributed or reviewed/merged. Document them for upcoming release. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- NEWS.rst | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/NEWS.rst b/NEWS.rst index 4b7f4e8244..65e21c4299 100644 --- a/NEWS.rst +++ b/NEWS.rst @@ -46,6 +46,12 @@ v10.10.0 (unreleased) required to keep data "raw" on disk, but the use case requires features of the QCOW2 format such as incremental backups. + * swtpm: Add support for profiles + + Upcoming swtpm release will have TPM profile support that allows to + restrict a TPM's provided set of crypto algorithms and commands. Users can + now select profile by using ``<profile/>`` in their TPM XML definition. + * **Improvements** * qemu: Support UEFI NVRAM images on block storage @@ -55,6 +61,21 @@ v10.10.0 (unreleased) qcow2 is used as the format, the images are now also auto-populated from the template. + * qemu: Automatically add IOMMU when needed + + When domain of 'qemu' or 'kvm' type has more than 255 vCPUs IOMMU with EIM + mode is required. Starting with this release libvirt automatically adds one + (or turns on the EIM mode if there's IOMMU without it). + + * ch: allow hostdevs in domain definition + + The Cloud Hypervisor driver (ch) now supports ``<hostdev/>``-s. + + * ch: Enable callbacks for ch domain events + + The Cloud Hypervisor driver (ch) now supports emitting events on domain + define, undefine, start, boot, stop and destroy. + * **Bug fixes** * qemu: Fix reversion and inactive deletion of internal snapshots with UEFI NVRAM @@ -63,6 +84,13 @@ v10.10.0 (unreleased) firmware was allowed, but certain operations such as reversion or inactive deletion didn't work properly as they didn't consider the NVRAM qcow2 file. + * virnetdevopenvswitch: Warn on unsupported QoS settings + + For OpenVSwitch vNICs libivrt does not set QoS directly using 'tc' but + offloads setting to OVS. But OVS is not as feature full as libvirt in this + regard and setting different 'peak' than 'average' results in vNIC always + sticking with 'peak'. Produce a warning if that's the case. + v10.9.0 (2024-11-01) ==================== -- 2.45.2

2 1

[PATCH 0/8] qemu: fixes for blockjobs with <dataStore>
by Peter Krempa 28 Nov '24

28 Nov '24

Couple fixes for the <dataStore> feature when used with blockjobs: - fixes for qemuBlockReopenAccess (code driving 'blockdev-reopen' in qemu) for bugs noticed during testing of <dataStore> - the 'auto-read-only' qemu feature doesn't apply od the 'data-file' blockdevs, thus libvirt needs to handle them explicitly - lookup of the node name associated with the <dataStore> volume in the chains (for block threshold event handling) Peter Krempa (8): qemuBlockReopenAccess: Add debug log entry about state of the image qemuBlockReopenAccess: Fix update of 'readonly' state qemuBlockReopenAccess: Don't require backing chain terminator for non-chained images qemu: block: Ensure that <dataStore> is in appropriate state qemu: snapshot: Change 'data-file' to read-only after snapshot qemuDomainVirStorageSourceFindByNodeName: Extract nodename matching qemuDomainVirStorageSourceFindByNodeName: Match also '<dataStore>' sources qemuDomainGetStorageSourceByDevstr: Lookup also '<dataStore>' src/qemu/qemu_block.c | 60 +++++++++++++++++++++++++++++++--------- src/qemu/qemu_blockjob.c | 16 ++++++++++- src/qemu/qemu_domain.c | 38 +++++++++++++++++++++---- src/qemu/qemu_snapshot.c | 12 ++++++-- 4 files changed, 105 insertions(+), 21 deletions(-) -- 2.47.0

2 14

[PATCH for 10.10.0 0/5] Fix broken XML schemas and 'virschematest'
by Peter Krempa 28 Nov '24

28 Nov '24

'virschematest' didn't test the majority of files for some time so some schema problems snuck in. Peter Krempa (5): schemas: domain: Make <identity> subelement of NFS disk source optional schemas: domaincaps: Add schema for 'panic' device schemas: domaincaps: Add schema for 'canonical' cpu model name schemas: domaincaps: Add schema for CPU 'blockers' virschematest: Don't skip all "directory" tests src/conf/schemas/domaincaps.rng | 29 +++++++++++++++++++++++++++++ src/conf/schemas/domaincommon.rng | 26 ++++++++++++++------------ tests/virschematest.c | 9 +++++---- 3 files changed, 48 insertions(+), 16 deletions(-) -- 2.47.0

2 7