January 2023 - Devel - libvirt List Archives

[PATCH] NEWS: document addition of passt support
by Laine Stump 11 Jan '23

11 Jan '23

Signed-off-by: Laine Stump <laine(a)redhat.com> --- NEWS.rst | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/NEWS.rst b/NEWS.rst index fecca64344..768208ed03 100644 --- a/NEWS.rst +++ b/NEWS.rst @@ -23,6 +23,16 @@ v9.0.0 (unreleased) ``virDomainSnapshotDelete()``. Flags that allow deleting children or children only are not supported. + * QEMU: support passt (https://passt.top) + + passt can be used to connect an emulated network device to the + host's network without requiring libvirt to have any sort of + elevated privileges. This is configured with:: + + <interface type='user'> + <backend type='passt'> + ... + * **Improvements** * qemu: Prefer PNG for domain screenshots -- 2.38.1

2 3

[PATCH] spec: Restart libvirtd on upgrade without socket activation
by Martin Kletzander 10 Jan '23

10 Jan '23

The %posttrans phase needs has a special case for upgrading libvirt daemon with --listen, but it forgot to also restart the daemon in order to run the new installed version. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1820437 Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- libvirt.spec.in | 1 + 1 file changed, 1 insertion(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index d9529fc76c4a..77b274b6f91d 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1449,6 +1449,7 @@ then libvirtd-admin.socket \ libvirtd-tls.socket \ libvirtd-tcp.socket >/dev/null 2>&1 || : + /bin/systemctl try-restart libvirtd.service >/dev/null 2>&1 || : else # Old libvirtd owns the sockets and will delete them on # shutdown. Can't use a try-restart as libvirtd will simply -- 2.39.0

3 2

[libvirt PATCH 0/9] src: misc cleanups related to remote protocol/client
by Daniel P. Berrangé 10 Jan '23

10 Jan '23

Just before xmas I did some work on an rpcgen replacement so we can use a non-sucky API instead of xdr_*. This short series has some general cleanups I found useful. The rpcgen code will come later. While this is all non-functional changes, I'm NOT proposing it for 9.0, it can safely wait as it isn't fixing any real bugs (that I've realized). Daniel P. Berrangé (9): remote: remove redundant initialization of args variable lxc: fix XDR protocol compliance logging: remove redundant XDR typedef rpc: use VIR_LOCK_GUARD in remote client code remote: use VIR_LOCK_GUARD in client code admin: use VIR_LOCK_GUARD in client code rpc: use struct zero initializer instead of memset remote: use struct zero initializer instead of memset admin: use struct zero initializer instead of memset src/admin/admin_remote.c | 92 +-- src/logging/log_protocol.x | 2 - src/lxc/lxc_monitor_protocol.x | 6 +- src/remote/remote_driver.c | 1280 +++++++++----------------------- src/rpc/gendispatch.pl | 18 +- 5 files changed, 397 insertions(+), 1001 deletions(-) -- 2.38.1

2 10

[PATCH] qemu: Fix handling of passed FDs in remoteDispatchDomainFdAssociate
by Peter Krempa 10 Jan '23

10 Jan '23

To ensure same behaviour when remote driver is or is not used we must not steal the FDs and array holding them passed to qemuDomainFDAssociate but rather duplicate them. At the same time the remote driver must close and free them to prevent leak. Pointed out by Coverity as FD leak on error path: *** CID 404348: Resource leaks (RESOURCE_LEAK) /src/remote/remote_daemon_dispatch.c: 7484 in remoteDispatchDomainFdAssociate() 7478 rv = 0; 7479 7480 cleanup: 7481 if (rv < 0) 7482 virNetMessageSaveError(rerr); 7483 virObjectUnref(dom); >>> CID 404348: Resource leaks (RESOURCE_LEAK) >>> Variable "fds" going out of scope leaks the storage it points to. 7484 return rv; Fixes: abd9025c2fd Fixes: f762f87534e Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- src/qemu/qemu_driver.c | 15 ++++++++++++--- src/remote/remote_daemon_dispatch.c | 3 +++ 2 files changed, 15 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index a88c9ebe64..d6879175fe 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -20442,7 +20442,8 @@ qemuDomainFDAssociate(virDomainPtr domain, { virDomainObj *vm = NULL; qemuDomainObjPrivate *priv; - virStorageSourceFDTuple *new; + g_autoptr(virStorageSourceFDTuple) new = NULL; + size_t i; int ret = -1; virCheckFlags(VIR_DOMAIN_FD_ASSOCIATE_SECLABEL_RESTORE | @@ -20460,8 +20461,16 @@ qemuDomainFDAssociate(virDomainPtr domain, priv = vm->privateData; new = virStorageSourceFDTupleNew(); - new->fds = fds; new->nfds = nfds; + new->fds = g_new0(int, new->nfds); + for (i = 0; i < new->nfds; i++) { + if ((new->fds[i] = dup(fds[i])) < 0) { + virReportSystemError(errno, + _("failed to duplicate passed fd with index '%zu'"), + i); + goto cleanup; + } + } new->conn = domain->conn; new->writable = flags & VIR_DOMAIN_FD_ASSOCIATE_SECLABEL_WRITABLE; @@ -20469,7 +20478,7 @@ qemuDomainFDAssociate(virDomainPtr domain, virCloseCallbacksDomainAdd(vm, domain->conn, qemuDomainFDHashCloseConnect); - g_hash_table_insert(priv->fds, g_strdup(name), new); + g_hash_table_insert(priv->fds, g_strdup(name), g_steal_pointer(&new)); ret = 0; diff --git a/src/remote/remote_daemon_dispatch.c b/src/remote/remote_daemon_dispatch.c index 40c734ce6b..6c56e9ec3e 100644 --- a/src/remote/remote_daemon_dispatch.c +++ b/src/remote/remote_daemon_dispatch.c @@ -7478,6 +7478,9 @@ remoteDispatchDomainFdAssociate(virNetServer *server G_GNUC_UNUSED, rv = 0; cleanup: + for (i = 0; i < nfds; i++) + VIR_FORCE_CLOSE(fds[i]); + g_free(fds); if (rv < 0) virNetMessageSaveError(rerr); virObjectUnref(dom); -- 2.38.1

2 1

[PATCH 00/12] migration/migrationpin:support migration pin
by Jiang Jiacheng 10 Jan '23

10 Jan '23

Support set the CPU affinity of the live migration thread to improve the migration performance in specific cases. By default, the migration thread shares CPU resources with the VM process. With those API, support pin migration thread to expected CPU list to avoid preempting CPU resources of VM process. New API 'pinMigrationThread' and virsh command 'virsh migrationpin' is used to pin migration thread to expected CPU list before or during migration. New migration param 'migration.pin' is used to support migration pin via interface 'virDomainMigrateToURI3'. Jiang Jiacheng (2): migration/migration-pin: support migration thread pin by virsh command migration/migration-pin/multifd-pin: add migration pin status handle zhengchuan (10): migration/migration-pin: get migration pid for migration pin migration/migration-pin: pin migration pid by given cpumap migration/migration-pin: add qemu monitor callback functions migration/migration-pin: add migrationpin for migration parameters migration/migration-pin: get cpumap from migration.pin migration/migration-pin: add domainMigrationPid for qemuMonitorCallbacks migration/multifd-pin: get multifd pid for migration pin migration/multifd-pin: pin multifd pid by given cpumap migration/multifd-pin: add qemu monitor callback functions migration/multifd-pin: support migration multifd thread pin include/libvirt/libvirt-domain.h | 19 ++ src/conf/domain_conf.c | 9 + src/conf/domain_conf.h | 11 ++ src/conf/virconftypes.h | 2 + src/driver-hypervisor.h | 16 ++ src/libvirt-domain.c | 144 +++++++++++++++ src/libvirt_private.syms | 1 + src/libvirt_public.syms | 7 + src/qemu/qemu_domain.c | 5 + src/qemu/qemu_domain.h | 5 + src/qemu/qemu_driver.c | 175 ++++++++++++++++++ src/qemu/qemu_migration.c | 5 + src/qemu/qemu_migration.h | 1 + src/qemu/qemu_migration_params.c | 21 +++ src/qemu/qemu_migration_params.h | 4 + src/qemu/qemu_monitor.c | 20 ++ src/qemu/qemu_monitor.h | 13 ++ src/qemu/qemu_monitor_json.c | 32 ++++ src/qemu/qemu_process.c | 303 +++++++++++++++++++++++++++++++ src/qemu/qemu_process.h | 15 ++ src/remote/remote_driver.c | 3 + src/remote/remote_protocol.x | 43 ++++- src/remote_protocol-structs | 28 +++ src/util/vircgroup.c | 3 + src/util/vircgroup.h | 1 + tools/virsh-domain.c | 69 +++++++ 26 files changed, 954 insertions(+), 1 deletion(-) -- 2.33.0

2 17

[PATCH] conf: check vhost-user queues with vcpus
by Jiang Jiacheng 10 Jan '23

10 Jan '23

With kernel without the ref patch, if queues > vcpus, interrupts will be centralized on one vcpu affecting guest performance. After the ref patch merged, the queues whose number is greater than the number of vcpus will not be used. Considering the above, it's better to check the counts of vhost-user queues and vcpus. ref: https://patchwork.kernel.org/project/linux-scsi/cover/1553682995-5682-1-git… Signed-off-by: Jiang Jiacheng <jiangjiacheng(a)huawei.com> --- src/conf/domain_validate.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/src/conf/domain_validate.c b/src/conf/domain_validate.c index 95b8d9b419..6106e79999 100644 --- a/src/conf/domain_validate.c +++ b/src/conf/domain_validate.c @@ -308,7 +308,7 @@ virSecurityDeviceLabelDefValidate(virSecurityDeviceLabelDef **seclabels, static int -virDomainDiskVhostUserValidate(const virDomainDiskDef *disk) +virDomainDiskVhostUserValidate(const virDomainDef *def, const virDomainDiskDef *disk) { if (disk->bus != VIR_DOMAIN_DISK_BUS_VIRTIO) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", @@ -465,6 +465,12 @@ virDomainDiskVhostUserValidate(const virDomainDiskDef *disk) return -1; } + if (disk->queues > virDomainDefGetVcpus(def)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("vhost-user disk queues must <= vcpus")); + return -1; + } + return 0; } @@ -807,7 +813,7 @@ virDomainDiskDefValidate(const virDomainDef *def, } if (disk->src->type == VIR_STORAGE_TYPE_VHOST_USER && - virDomainDiskVhostUserValidate(disk) < 0) { + virDomainDiskVhostUserValidate(def, disk) < 0) { return -1; } -- 2.33.0

2 2

[PATCH 0/3] define g_autoptr for virNWFilterDef and virNWFilterRuleDef
by Jiang Jiacheng 10 Jan '23

10 Jan '23

Define and use g_autoptr() for virNWFilterDef and virNWFilterRuleDef, and remove unnecessary label/variable. Those patches are followup with: https://gitlab.com/libvirt/libvirt/-/commit/a9027d447be789cd11c0aa18ceb83ed… Jiang Jiacheng (3): conf: define g_autoptr for virNWFilterDef and virNWFilterRuleDef src/tests: use g_autoptr for virNWFilterDef and virNWFilterRuleDef virNWFilterSaveConfig: remove the unnecessary variable 'ret' src/conf/nwfilter_conf.c | 52 ++++++++++++++-------------------- src/conf/nwfilter_conf.h | 2 ++ src/conf/virnwfilterobj.c | 19 ++++++------- src/nwfilter/nwfilter_driver.c | 7 ++--- tests/nwfilterxml2xmltest.c | 22 ++++++-------- 5 files changed, 42 insertions(+), 60 deletions(-) -- 2.33.0

2 6

[PATCH] remote: fix double free of migration params on error
by Daniel P. Berrangé 10 Jan '23

10 Jan '23

The remote_*_args methods will generally borrow pointers passed in the caller, so should not be freed. On failure of the virTypedParamsSerialize method, however, xdr_free was being called. This is presumably because it was thought that the params may have been partially serialized and need cleaning up. This is incorrect, as virTypedParamsSerialize takes care to cleanup partially serialized data. This xdr_free call would lead to free'ing the borrowed cookie pointers, which would be a double free. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- src/remote/remote_driver.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index b0dba9057b..bb44d0004f 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -6919,8 +6919,6 @@ remoteDomainMigrateBegin3Params(virDomainPtr domain, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_begin3_params_args, - (char *) &args); goto cleanup; } @@ -6981,8 +6979,6 @@ remoteDomainMigratePrepare3Params(virConnectPtr dconn, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_prepare3_params_args, - (char *) &args); goto cleanup; } @@ -7063,8 +7059,6 @@ remoteDomainMigratePrepareTunnel3Params(virConnectPtr dconn, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_prepare_tunnel3_params_args, - (char *) &args); goto cleanup; } @@ -7149,8 +7143,6 @@ remoteDomainMigratePerform3Params(virDomainPtr dom, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_perform3_params_args, - (char *) &args); goto cleanup; } @@ -7216,8 +7208,6 @@ remoteDomainMigrateFinish3Params(virConnectPtr dconn, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_finish3_params_args, - (char *) &args); goto cleanup; } @@ -7284,8 +7274,6 @@ remoteDomainMigrateConfirm3Params(virDomainPtr domain, (struct _virTypedParameterRemote **) &args.params.params_val, &args.params.params_len, VIR_TYPED_PARAM_STRING_OKAY) < 0) { - xdr_free((xdrproc_t) xdr_remote_domain_migrate_confirm3_params_args, - (char *) &args); goto cleanup; } -- 2.38.1

2 1

[PATCH] admin: fix leak of typed parameters on error
by Daniel P. Berrangé 10 Jan '23

10 Jan '23

A few admin client methods had the xdr_free call the wrong side of the cleanup label, so typed parameters would not be freed on error. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- src/admin/admin_remote.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/admin/admin_remote.c b/src/admin/admin_remote.c index 012488bff3..f99b93eef6 100644 --- a/src/admin/admin_remote.c +++ b/src/admin/admin_remote.c @@ -269,9 +269,9 @@ remoteAdminServerGetThreadPoolParameters(virAdmServerPtr srv, goto cleanup; rv = 0; - xdr_free((xdrproc_t)xdr_admin_server_get_threadpool_parameters_ret, (char *) &ret); cleanup: + xdr_free((xdrproc_t)xdr_admin_server_get_threadpool_parameters_ret, (char *) &ret); virObjectUnlock(priv); return rv; } @@ -342,9 +342,9 @@ remoteAdminClientGetInfo(virAdmClientPtr client, goto cleanup; rv = 0; - xdr_free((xdrproc_t)xdr_admin_client_get_info_ret, (char *) &ret); cleanup: + xdr_free((xdrproc_t)xdr_admin_client_get_info_ret, (char *) &ret); virObjectUnlock(priv); return rv; } @@ -380,10 +380,10 @@ remoteAdminServerGetClientLimits(virAdmServerPtr srv, goto cleanup; rv = 0; - xdr_free((xdrproc_t) xdr_admin_server_get_client_limits_ret, - (char *) &ret); cleanup: + xdr_free((xdrproc_t) xdr_admin_server_get_client_limits_ret, + (char *) &ret); virObjectUnlock(priv); return rv; } -- 2.38.1

2 1

[libvirt PATCH 0/2] docs: submitting-patches: shorten slightly
by Ján Tomko 10 Jan '23

10 Jan '23

Ján Tomko (2): docs: submimtting-patches: Remove emphasis on not cc'ing developers docs: submitting-patches: remove note about --patience docs/submitting-patches.rst | 9 +-------- 1 file changed, 1 insertion(+), 8 deletions(-) -- 2.39.0

3 4