- Devel - libvirt List Archives

[PATCH 0/6] Be tolerant to CH_CMD missing
by Michal Privoznik 04 Jun '21

04 Jun '21

See 4/6 for more info. The rest is just cleanup. Michal Prívozník (6): ch_conf: Move error reporting into chExtractVersionInfo() chExtractVersionInfo: Don't check for retversion != NULL ch_conf: Dissolve chExtractVersionInfo() in chExtractVersion() ch_driver: Don't error out if CH_CMD was not found chExtractVersion: use g_auto* chExtractVersion: Drop @ret src/ch/ch_conf.c | 53 +++++++++++++++++----------------------------- src/ch/ch_driver.c | 15 +++++++++---- 2 files changed, 30 insertions(+), 38 deletions(-) -- 2.31.1

2 7

[PATCH 0/3] Replace some libvirt handling functions with GLib APIs
by Luke Yue 04 Jun '21

04 Jun '21

Note that the g_find_program_in_path() will search file in current dir when PATH env is not set, while virFindFileInPath() won't. The virFileAbsPath() could be replaced by g_canonicalize_file() when not take the return value into account, and it just simply return 0 now, so maybe remove the function is a better choice? Related issue: https://gitlab.com/libvirt/libvirt/-/issues/12 Luke Yue (3): virfile: Use g_build_filename() when building paths virfile: Simplify virFindFileInPath() with g_find_program_in_path() virfile: Use g_canonicalize_file() to simplify virFileAbsPath() src/util/virfile.c | 71 +++++++++++++--------------------------------- 1 file changed, 19 insertions(+), 52 deletions(-) -- 2.31.1

2 10

[libvirt PATCH 0/2] docs: small fixes to audio docs
by Daniel P. Berrangé 04 Jun '21

04 Jun '21

Daniel P. Berrangé (2): docs: tweak heading name for <audio> element docs: add missing docs for some sound devices docs/formatdomain.rst | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) -- 2.31.1

2 4

[PATCH v2] Add basic driver for the Cloud-Hypervisor
by William Douglas 04 Jun '21

04 Jun '21

Cloud-Hypervisor is a KVM virtualization using hypervisor. It functions similarly to qemu and the libvirt Cloud-Hypervisor driver uses a very similar structure to the libvirt driver. The biggest difference from the libvirt perspective is that the "monitor" socket is seperated into two sockets one that commands are issued to and one that events are notified from. The current implementation only uses the command socket (running over a REST API with json encoded data) with future changes to add support for the event socket (to better handle shutdowns from inside the VM). This patch adds support for the following initial VM actions using the Cloud-Hypervsior API: * vm.create * vm.delete * vm.boot * vm.shutdown * vm.reboot * vm.pause * vm.resume To use the Cloud-Hypervisor driver, the v15.0 release of Cloud-Hypervisor is required to be installed. Some additional notes: * The curl handle is persistent but not useful to detect ch process shutdown/crash (a future patch will address this shortcoming) * On a 64-bit host Cloud-Hypervisor needs to support PVH and so can emulate 32-bit mode but it isn't fully tested (a 64-bit kernel and 32-bit userspace is fine, a 32-bit kernel isn't validated) Signed-off-by: William Douglas <william.douglas(a)intel.com> --- The original RFC is https://listman.redhat.com/archives/libvir-list/2020-August/msg01040.html The v1 patch is https://listman.redhat.com/archives/libvir-list/2021-April/msg01271.html Changes since v1: * Removed specfile changes as cloud-hypervisor isn't included in distros yet * Clarified booting a 32-bit kernel is untested * Removed unused ch_cmd variable from a previous refactor * Updated version detection based on latest cloud-hypervisor release * Update the libvirt version cloud-hypervisor driver features were added in * Updated the supported URI schemas to remove "Cloud-Hypervisor:///" * Updated build system to only build the cloud-hypervisor driver on x86_64 and aarch64 --- docs/drivers.html.in | 1 + docs/drvch.rst | 55 ++ docs/meson.build | 1 + include/libvirt/virterror.h | 1 + meson.build | 43 ++ meson_options.txt | 3 + po/POTFILES.in | 5 + src/ch/ch_conf.c | 251 ++++++++ src/ch/ch_conf.h | 85 +++ src/ch/ch_domain.c | 203 ++++++ src/ch/ch_domain.h | 65 ++ src/ch/ch_driver.c | 930 ++++++++++++++++++++++++++++ src/ch/ch_driver.h | 24 + src/ch/ch_monitor.c | 837 +++++++++++++++++++++++++ src/ch/ch_monitor.h | 60 ++ src/ch/ch_process.c | 126 ++++ src/ch/ch_process.h | 31 + src/ch/meson.build | 74 +++ src/ch/virtchd.service.in | 47 ++ src/ch/virtchd.sysconf | 3 + src/meson.build | 1 + src/remote/remote_daemon.c | 4 + src/remote/remote_daemon_dispatch.c | 3 +- src/util/virerror.c | 1 + tools/virsh.c | 3 + 25 files changed, 2856 insertions(+), 1 deletion(-) create mode 100644 docs/drvch.rst create mode 100644 src/ch/ch_conf.c create mode 100644 src/ch/ch_conf.h create mode 100644 src/ch/ch_domain.c create mode 100644 src/ch/ch_domain.h create mode 100644 src/ch/ch_driver.c create mode 100644 src/ch/ch_driver.h create mode 100644 src/ch/ch_monitor.c create mode 100644 src/ch/ch_monitor.h create mode 100644 src/ch/ch_process.c create mode 100644 src/ch/ch_process.h create mode 100644 src/ch/meson.build create mode 100644 src/ch/virtchd.service.in create mode 100644 src/ch/virtchd.sysconf diff --git a/docs/drivers.html.in b/docs/drivers.html.in index 34f98f60b6..824604998e 100644 --- a/docs/drivers.html.in +++ b/docs/drivers.html.in @@ -37,6 +37,7 @@ <li><a href="drvhyperv.html">Microsoft Hyper-V</a></li> <li><a href="drvvirtuozzo.html">Virtuozzo</a></li> <li><a href="drvbhyve.html">Bhyve</a> - The BSD Hypervisor</li> + <li><a href="drvch.html">Cloud Hypervisor</a></li> </ul> </body> diff --git a/docs/drvch.rst b/docs/drvch.rst new file mode 100644 index 0000000000..bb13599e6f --- /dev/null +++ b/docs/drvch.rst @@ -0,0 +1,55 @@ +======================= +Cloud Hypervisor driver +======================= + +.. contents:: + +Cloud Hypervisor is an open source Virtual Machine Monitor (VMM) that +runs on top of KVM. The project focuses on exclusively running modern, +cloud workloads, on top of a limited set of hardware architectures and +platforms. Cloud workloads refers to those that are usually run by +customers inside a cloud provider. For our purposes this means modern +operating systems with most I/O handled by paravirtualised devices +(i.e. virtio), no requirement for legacy devices, and 64-bit CPUs. + +The libvirt Cloud Hypervisor driver is intended to be run as a session +driver without privileges. The cloud-hypervisor binary itself should be +``setcap cap_net_admin+ep`` (in order to create tap interfaces). + +Expected connection URI would be + +``ch:///session`` + + +Example guest domain XML configurations +======================================= + +The Cloud Hypervisor driver in libvirt is in its early stage under active +development only supporting a limited number of Cloud Hypervisor features. + +Firmware is from +`hypervisor-fw <https://github.com/cloud-hypervisor/rust-hypervisor-firmware/releases>`__ + +**Note: Only virtio devices are supported** + +:: + + <domain type='kvm'> + <name>cloudhypervisor</name> + <uuid>4dea22b3-1d52-d8f3-2516-782e98ab3fa0</uuid> + <os> + <type>hvm</type> + <kernel>hypervisor-fw</kernel> + </os> + <memory unit='G'>2</memory> + <devices> + <disk type='file'> + <source file='disk.raw'/> + <target dev='vda' bus='virtio'/> + </disk> + <interface type='ethernet'> + <model type='virtio'/> + </interface> + </devices> + <vcpu>2</vcpu> + </domain> diff --git a/docs/meson.build b/docs/meson.build index f550629d8e..bee0d80d53 100644 --- a/docs/meson.build +++ b/docs/meson.build @@ -111,6 +111,7 @@ docs_rst_files = [ 'daemons', 'developer-tooling', 'drvqemu', + 'drvch', 'formatbackup', 'formatcheckpoint', 'formatdomain', diff --git a/include/libvirt/virterror.h b/include/libvirt/virterror.h index 524a7bf9e8..57986931fd 100644 --- a/include/libvirt/virterror.h +++ b/include/libvirt/virterror.h @@ -136,6 +136,7 @@ typedef enum { VIR_FROM_TPM = 70, /* Error from TPM */ VIR_FROM_BPF = 71, /* Error from BPF code */ + VIR_FROM_CH = 72, /* Error from Cloud-Hypervisor driver */ # ifdef VIR_ENUM_SENTINELS VIR_ERR_DOMAIN_LAST diff --git a/meson.build b/meson.build index 1f97842319..933055ef33 100644 --- a/meson.build +++ b/meson.build @@ -1525,6 +1525,48 @@ elif get_option('driver_lxc').enabled() error('linux and remote_driver are required for LXC') endif +if not get_option('driver_ch').disabled() and host_machine.system() == 'linux' and (host_machine.cpu_family() == 'x86_64' or host_machine.cpu_family() == 'aarch64') + use_ch = true + + if not conf.has('WITH_LIBVIRTD') + use_ch = false + if get_option('driver_ch').enabled() + error('libvirtd is required to build Cloud-Hypervisor driver') + endif + endif + + if not yajl_dep.found() + use_ch = false + if get_option('driver_ch').enabled() + error('YAJL 2 is required to build Cloud-Hypervisor driver') + endif + endif + + if not curl_dep.found() + use_ch = false + if get_option('driver_ch').enabled() + error('curl is required to build Cloud-Hypervisor driver') + endif + endif + + if use_ch + conf.set('WITH_CH', 1) + + default_ch_user = 'root' + default_ch_group = 'root' + ch_user = get_option('ch_user') + if ch_user == '' + ch_user = default_ch_user + endif + ch_group = get_option('ch_group') + if ch_group == '' + ch_group = default_ch_group + endif + conf.set_quoted('CH_USER', ch_user) + conf.set_quoted('CH_GROUP', ch_group) + endif +endif + # there's no use compiling the network driver without the libvirt # daemon, nor compiling it for macOS, where it breaks the compile if not get_option('driver_network').disabled() and conf.has('WITH_LIBVIRTD') and host_machine.system() != 'darwin' @@ -2178,6 +2220,7 @@ driver_summary = { 'VBox': conf.has('WITH_VBOX'), 'libxl': conf.has('WITH_LIBXL'), 'LXC': conf.has('WITH_LXC'), + 'Cloud-Hypervisor': conf.has('WITH_CH'), 'ESX': conf.has('WITH_ESX'), 'Hyper-V': conf.has('WITH_HYPERV'), 'vz': conf.has('WITH_VZ'), diff --git a/meson_options.txt b/meson_options.txt index 2606648b64..cd0b4b33be 100644 --- a/meson_options.txt +++ b/meson_options.txt @@ -53,6 +53,9 @@ option('driver_interface', type: 'feature', value: 'auto', description: 'host in option('driver_libvirtd', type: 'feature', value: 'auto', description: 'libvirtd driver') option('driver_libxl', type: 'feature', value: 'auto', description: 'libxenlight driver') option('driver_lxc', type: 'feature', value: 'auto', description: 'Linux Container driver') +option('driver_ch', type: 'feature', value: 'auto', description: 'Cloud-Hypervisor driver') +option('ch_user', type: 'string', value: '', description: 'username to run Cloud-Hypervisor system instance as') +option('ch_group', type: 'string', value: '', description: 'groupname to run Cloud-Hypervisor system instance as') option('driver_network', type: 'feature', value: 'auto', description: 'virtual network driver') option('driver_openvz', type: 'feature', value: 'auto', description: 'OpenVZ driver') option('driver_qemu', type: 'feature', value: 'auto', description: 'QEMU/KVM driver') diff --git a/po/POTFILES.in b/po/POTFILES.in index 413783ee35..c200d7452a 100644 --- a/po/POTFILES.in +++ b/po/POTFILES.in @@ -18,6 +18,11 @@ @SRCDIR@src/bhyve/bhyve_monitor.c @SRCDIR@src/bhyve/bhyve_parse_command.c @SRCDIR@src/bhyve/bhyve_process.c +@SRCDIR@src/ch/ch_conf.c +@SRCDIR@src/ch/ch_domain.c +@SRCDIR@src/ch/ch_driver.c +@SRCDIR@src/ch/ch_monitor.c +@SRCDIR@src/ch/ch_process.c @SRCDIR@src/conf/backup_conf.c @SRCDIR@src/conf/capabilities.c @SRCDIR@src/conf/checkpoint_conf.c diff --git a/src/ch/ch_conf.c b/src/ch/ch_conf.c new file mode 100644 index 0000000000..2dd104b8a8 --- /dev/null +++ b/src/ch/ch_conf.c @@ -0,0 +1,251 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_conf.c: functions for Cloud-Hypervisor configuration + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#include <config.h> + +#include "configmake.h" +#include "viralloc.h" +#include "vircommand.h" +#include "virlog.h" +#include "virobject.h" +#include "virstring.h" +#include "virutil.h" + +#include "ch_conf.h" +#include "ch_domain.h" + +#define VIR_FROM_THIS VIR_FROM_CH + +VIR_LOG_INIT("ch.ch_conf"); + +static virClass *virCHDriverConfigClass; +static void virCHDriverConfigDispose(void *obj); + +static int virCHConfigOnceInit(void) +{ + if (!VIR_CLASS_NEW(virCHDriverConfig, virClassForObject())) + return -1; + + return 0; +} + +VIR_ONCE_GLOBAL_INIT(virCHConfig); + + +/* Functions */ +virCaps *virCHDriverCapsInit(void) +{ + virCaps *caps; + virCapsGuest *guest; + + if ((caps = virCapabilitiesNew(virArchFromHost(), + false, false)) == NULL) + goto cleanup; + + if (!(caps->host.numa = virCapabilitiesHostNUMANewHost())) + goto cleanup; + + if (virCapabilitiesInitCaches(caps) < 0) + goto cleanup; + + if ((guest = virCapabilitiesAddGuest(caps, + VIR_DOMAIN_OSTYPE_HVM, + caps->host.arch, + NULL, + NULL, + 0, + NULL)) == NULL) + goto cleanup; + + if (virCapabilitiesAddGuestDomain(guest, + VIR_DOMAIN_VIRT_KVM, + NULL, + NULL, + 0, + NULL) == NULL) + goto cleanup; + + return caps; + + cleanup: + virObjectUnref(caps); + return NULL; +} + +/** + * virCHDriverGetCapabilities: + * + * Get a reference to the virCaps instance for the + * driver. If @refresh is true, the capabilities will be + * rebuilt first + * + * The caller must release the reference with virObjetUnref + * + * Returns: a reference to a virCaps instance or NULL + */ +virCaps *virCHDriverGetCapabilities(virCHDriver *driver, + bool refresh) +{ + virCaps *ret; + if (refresh) { + virCaps *caps = NULL; + if ((caps = virCHDriverCapsInit()) == NULL) + return NULL; + + chDriverLock(driver); + virObjectUnref(driver->caps); + driver->caps = caps; + } else { + chDriverLock(driver); + } + + ret = virObjectRef(driver->caps); + chDriverUnlock(driver); + return ret; +} + +virDomainXMLOption * +chDomainXMLConfInit(virCHDriver *driver) +{ + virCHDriverDomainDefParserConfig.priv = driver; + return virDomainXMLOptionNew(&virCHDriverDomainDefParserConfig, + &virCHDriverPrivateDataCallbacks, + NULL, NULL, NULL); +} + +virCHDriverConfig * +virCHDriverConfigNew(bool privileged) +{ + virCHDriverConfig *cfg; + + if (virCHConfigInitialize() < 0) + return NULL; + + if (!(cfg = virObjectNew(virCHDriverConfigClass))) + return NULL; + + cfg->uri = g_strdup(privileged ? "ch:///system" : "ch:///session"); + if (privileged) { + if (virGetUserID(CH_USER, &cfg->user) < 0) + return NULL; + if (virGetGroupID(CH_GROUP, &cfg->group) < 0) + return NULL; + } else { + cfg->user = (uid_t)-1; + cfg->group = (gid_t)-1; + } + + if (privileged) { + cfg->logDir = g_strdup_printf("%s/log/libvirt/ch", LOCALSTATEDIR); + cfg->stateDir = g_strdup_printf("%s/libvirt/ch", RUNSTATEDIR); + + } else { + g_autofree char *rundir = NULL; + g_autofree char *cachedir = NULL; + + cachedir = virGetUserCacheDirectory(); + + cfg->logDir = g_strdup_printf("%s/ch/log", cachedir); + + rundir = virGetUserRuntimeDirectory(); + cfg->stateDir = g_strdup_printf("%s/ch/run", rundir); + } + + return cfg; +} + +virCHDriverConfig *virCHDriverGetConfig(virCHDriver *driver) +{ + virCHDriverConfig *cfg; + chDriverLock(driver); + cfg = virObjectRef(driver->config); + chDriverUnlock(driver); + return cfg; +} + +static void +virCHDriverConfigDispose(void *obj) +{ + virCHDriverConfig *cfg = obj; + + g_free(cfg->stateDir); + g_free(cfg->logDir); +} + +#define MIN_VERSION ((15 * 1000000) + (0 * 1000) + (0)) + +static int +chExtractVersionInfo(int *retversion) +{ + int ret = -1; + unsigned long version; + char *help = NULL; + char *tmp = NULL; + g_autofree char *ch_cmd = g_find_program_in_path(CH_CMD); + virCommand *cmd = virCommandNewArgList(ch_cmd, "--version", NULL); + + if (retversion) + *retversion = 0; + + virCommandAddEnvString(cmd, "LC_ALL=C"); + virCommandSetOutputBuffer(cmd, &help); + + if (virCommandRun(cmd, NULL) < 0) + goto cleanup; + + tmp = help; + + /* expected format: cloud-hypervisor v<major>.<minor>.<micro> */ + if ((tmp = STRSKIP(tmp, "cloud-hypervisor v")) == NULL) + goto cleanup; + + if (virParseVersionString(tmp, &version, true) < 0) + goto cleanup; + + if (version < MIN_VERSION) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Cloud-Hypervisor version is too old (v15.0 is the minimum supported version)")); + goto cleanup; + } + + if (retversion) + *retversion = version; + + ret = 0; + + cleanup: + virCommandFree(cmd); + + return ret; +} + +int chExtractVersion(virCHDriver *driver) +{ + if (driver->version > 0) + return 0; + + if (chExtractVersionInfo(&driver->version) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Could not extract Cloud-Hypervisor version")); + return -1; + } + + return 0; +} diff --git a/src/ch/ch_conf.h b/src/ch/ch_conf.h new file mode 100644 index 0000000000..d856825377 --- /dev/null +++ b/src/ch/ch_conf.h @@ -0,0 +1,85 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_conf.h: header file for Cloud-Hypervisor configuration + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#pragma once + +#include "virdomainobjlist.h" +#include "virthread.h" + +#define CH_DRIVER_NAME "CH" +#define CH_CMD "cloud-hypervisor" + +typedef struct _virCHDriver virCHDriver; + +typedef struct _virCHDriverConfig virCHDriverConfig; + +struct _virCHDriverConfig { + GObject parent; + + char *stateDir; + char *logDir; + char *uri; + + uid_t user; + gid_t group; +}; + +struct _virCHDriver +{ + virMutex lock; + + /* Require lock to get a reference on the object, + * lockless access thereafter */ + virCaps *caps; + + /* Immutable pointer, Immutable object */ + virDomainXMLOption *xmlopt; + + /* Immutable pointer, self-locking APIs */ + virDomainObjList *domains; + + /* Cloud-Hypervisor version */ + int version; + + /* Require lock to get reference on 'config', + * then lockless thereafter */ + virCHDriverConfig *config; + + /* pid file FD, ensures two copies of the driver can't use the same root */ + int lockFD; +}; + +virCaps *virCHDriverCapsInit(void); +virCaps *virCHDriverGetCapabilities(virCHDriver *driver, + bool refresh); +virDomainXMLOption *chDomainXMLConfInit(virCHDriver *driver); +virCHDriverConfig *virCHDriverConfigNew(bool privileged); +virCHDriverConfig *virCHDriverGetConfig(virCHDriver *driver); +int chExtractVersion(virCHDriver *driver); + +static inline void chDriverLock(virCHDriver *driver) +{ + virMutexLock(&driver->lock); +} + +static inline void chDriverUnlock(virCHDriver *driver) +{ + virMutexUnlock(&driver->lock); +} diff --git a/src/ch/ch_domain.c b/src/ch/ch_domain.c new file mode 100644 index 0000000000..9d0b091699 --- /dev/null +++ b/src/ch/ch_domain.c @@ -0,0 +1,203 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_domain.c: Domain manager functions for Cloud-Hypervisor driver + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#include <config.h> + +#include "ch_domain.h" +#include "viralloc.h" +#include "virlog.h" +#include "virtime.h" + +#define VIR_FROM_THIS VIR_FROM_CH + +VIR_ENUM_IMPL(virCHDomainJob, + CH_JOB_LAST, + "none", + "query", + "destroy", + "modify", +); + +VIR_LOG_INIT("ch.ch_domain"); + +static int +virCHDomainObjInitJob(virCHDomainObjPrivate *priv) +{ + memset(&priv->job, 0, sizeof(priv->job)); + + if (virCondInit(&priv->job.cond) < 0) + return -1; + + return 0; +} + +static void +virCHDomainObjResetJob(virCHDomainObjPrivate *priv) +{ + struct virCHDomainJobObj *job = &priv->job; + + job->active = CH_JOB_NONE; + job->owner = 0; +} + +static void +virCHDomainObjFreeJob(virCHDomainObjPrivate *priv) +{ + ignore_value(virCondDestroy(&priv->job.cond)); +} + +/* + * obj must be locked before calling, virCHDriver must NOT be locked + * + * This must be called by anything that will change the VM state + * in any way + * + * Upon successful return, the object will have its ref count increased. + * Successful calls must be followed by EndJob eventually. + */ +int +virCHDomainObjBeginJob(virDomainObj *obj, enum virCHDomainJob job) +{ + virCHDomainObjPrivate *priv = obj->privateData; + unsigned long long now; + unsigned long long then; + + if (virTimeMillisNow(&now) < 0) + return -1; + then = now + CH_JOB_WAIT_TIME; + + while (priv->job.active) { + VIR_DEBUG("Wait normal job condition for starting job: %s", + virCHDomainJobTypeToString(job)); + if (virCondWaitUntil(&priv->job.cond, &obj->parent.lock, then) < 0) + goto error; + } + + virCHDomainObjResetJob(priv); + + VIR_DEBUG("Starting job: %s", virCHDomainJobTypeToString(job)); + priv->job.active = job; + priv->job.owner = virThreadSelfID(); + + return 0; + + error: + VIR_WARN("Cannot start job (%s) for domain %s;" + " current job is (%s) owned by (%d)", + virCHDomainJobTypeToString(job), + obj->def->name, + virCHDomainJobTypeToString(priv->job.active), + priv->job.owner); + + if (errno == ETIMEDOUT) + virReportError(VIR_ERR_OPERATION_TIMEOUT, + "%s", _("cannot acquire state change lock")); + else + virReportSystemError(errno, + "%s", _("cannot acquire job mutex")); + return -1; +} + +/* + * obj must be locked and have a reference before calling + * + * To be called after completing the work associated with the + * earlier virCHDomainBeginJob() call + */ +void +virCHDomainObjEndJob(virDomainObj *obj) +{ + virCHDomainObjPrivate *priv = obj->privateData; + enum virCHDomainJob job = priv->job.active; + + VIR_DEBUG("Stopping job: %s", + virCHDomainJobTypeToString(job)); + + virCHDomainObjResetJob(priv); + virCondSignal(&priv->job.cond); +} + +static void * +virCHDomainObjPrivateAlloc(void *opaque G_GNUC_UNUSED) +{ + virCHDomainObjPrivate *priv; + + priv = g_new0(virCHDomainObjPrivate, 1); + + if (virCHDomainObjInitJob(priv) < 0) { + g_free(priv); + return NULL; + } + + return priv; +} + +static void +virCHDomainObjPrivateFree(void *data) +{ + virCHDomainObjPrivate *priv = data; + + virCHDomainObjFreeJob(priv); + g_free(priv); +} + +virDomainXMLPrivateDataCallbacks virCHDriverPrivateDataCallbacks = { + .alloc = virCHDomainObjPrivateAlloc, + .free = virCHDomainObjPrivateFree, +}; + +static int +virCHDomainDefPostParseBasic(virDomainDef *def, + void *opaque G_GNUC_UNUSED) +{ + /* check for emulator and create a default one if needed */ + if (!def->emulator) { + if (!(def->emulator = g_find_program_in_path(CH_CMD))) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("No emulator found for cloud-hypervisor")); + return 1; + } + } + + return 0; +} + +static int +virCHDomainDefPostParse(virDomainDef *def, + unsigned int parseFlags G_GNUC_UNUSED, + void *opaque, + void *parseOpaque G_GNUC_UNUSED) +{ + virCHDriver *driver = opaque; + g_autoptr(virCaps) caps = virCHDriverGetCapabilities(driver, false); + if (!caps) + return -1; + if (!virCapabilitiesDomainSupported(caps, def->os.type, + def->os.arch, + def->virtType)) + return -1; + + return 0; +} + +virDomainDefParserConfig virCHDriverDomainDefParserConfig = { + .domainPostParseBasicCallback = virCHDomainDefPostParseBasic, + .domainPostParseCallback = virCHDomainDefPostParse, +}; diff --git a/src/ch/ch_domain.h b/src/ch/ch_domain.h new file mode 100644 index 0000000000..b4e0d4c212 --- /dev/null +++ b/src/ch/ch_domain.h @@ -0,0 +1,65 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_domain.h: header file for domain manager's Cloud-Hypervisor driver functions + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#pragma once + +#include "ch_conf.h" +#include "ch_monitor.h" + +/* Give up waiting for mutex after 30 seconds */ +#define CH_JOB_WAIT_TIME (1000ull * 30) + +/* Only 1 job is allowed at any time + * A job includes *all* ch.so api, even those just querying + * information, not merely actions */ + +enum virCHDomainJob { + CH_JOB_NONE = 0, /* Always set to 0 for easy if (jobActive) conditions */ + CH_JOB_QUERY, /* Doesn't change any state */ + CH_JOB_DESTROY, /* Destroys the domain (cannot be masked out) */ + CH_JOB_MODIFY, /* May change state */ + CH_JOB_LAST +}; +VIR_ENUM_DECL(virCHDomainJob); + + +struct virCHDomainJobObj { + virCond cond; /* Use to coordinate jobs */ + enum virCHDomainJob active; /* Currently running job */ + int owner; /* Thread which set current job */ +}; + + +typedef struct _virCHDomainObjPrivate virCHDomainObjPrivate; +struct _virCHDomainObjPrivate { + struct virCHDomainJobObj job; + + virCHMonitor *monitor; +}; + +extern virDomainXMLPrivateDataCallbacks virCHDriverPrivateDataCallbacks; +extern virDomainDefParserConfig virCHDriverDomainDefParserConfig; + +int +virCHDomainObjBeginJob(virDomainObj *obj, enum virCHDomainJob job) + G_GNUC_WARN_UNUSED_RESULT; + +void +virCHDomainObjEndJob(virDomainObj *obj); diff --git a/src/ch/ch_driver.c b/src/ch/ch_driver.c new file mode 100644 index 0000000000..a4f03d644d --- /dev/null +++ b/src/ch/ch_driver.c @@ -0,0 +1,930 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_driver.c: Core Cloud-Hypervisor driver functions + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#include <config.h> + +#include "ch_conf.h" +#include "ch_domain.h" +#include "ch_driver.h" +#include "ch_monitor.h" +#include "ch_process.h" +#include "datatypes.h" +#include "driver.h" +#include "viraccessapicheck.h" +#include "viralloc.h" +#include "virbuffer.h" +#include "vircommand.h" +#include "virerror.h" +#include "virfile.h" +#include "virlog.h" +#include "virnetdevtap.h" +#include "virobject.h" +#include "virstring.h" +#include "virtypedparam.h" +#include "viruri.h" +#include "virutil.h" +#include "viruuid.h" + +#define VIR_FROM_THIS VIR_FROM_CH + +VIR_LOG_INIT("ch.ch_driver"); + +static int chStateInitialize(bool privileged, + const char *root, + virStateInhibitCallback callback, + void *opaque); +static int chStateCleanup(void); +virCHDriver *ch_driver = NULL; + +static virDomainObj * +chDomObjFromDomain(virDomain *domain) +{ + virDomainObj *vm; + virCHDriver *driver = domain->conn->privateData; + char uuidstr[VIR_UUID_STRING_BUFLEN]; + + vm = virDomainObjListFindByUUID(driver->domains, domain->uuid); + if (!vm) { + virUUIDFormat(domain->uuid, uuidstr); + virReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching uuid '%s' (%s)"), + uuidstr, domain->name); + return NULL; + } + + return vm; +} + +/* Functions */ +static int +chConnectURIProbe(char **uri) +{ + if (ch_driver == NULL) + return 0; + + *uri = g_strdup("ch:///system"); + return 1; +} + +static virDrvOpenStatus chConnectOpen(virConnectPtr conn, + virConnectAuthPtr auth G_GNUC_UNUSED, + virConf *conf G_GNUC_UNUSED, + unsigned int flags) +{ + virCheckFlags(VIR_CONNECT_RO, VIR_DRV_OPEN_ERROR); + + /* URI was good, but driver isn't active */ + if (ch_driver == NULL) { + virReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("Cloud-Hypervisor state driver is not active")); + return VIR_DRV_OPEN_ERROR; + } + + if (virConnectOpenEnsureACL(conn) < 0) + return VIR_DRV_OPEN_ERROR; + + conn->privateData = ch_driver; + + return VIR_DRV_OPEN_SUCCESS; +} + +static int chConnectClose(virConnectPtr conn) +{ + conn->privateData = NULL; + return 0; +} + +static const char *chConnectGetType(virConnectPtr conn) +{ + if (virConnectGetTypeEnsureACL(conn) < 0) + return NULL; + + return "CH"; +} + +static int chConnectGetVersion(virConnectPtr conn, + unsigned long *version) +{ + virCHDriver *driver = conn->privateData; + + if (virConnectGetVersionEnsureACL(conn) < 0) + return -1; + + chDriverLock(driver); + *version = driver->version; + chDriverUnlock(driver); + return 0; +} + +static char *chConnectGetHostname(virConnectPtr conn) +{ + if (virConnectGetHostnameEnsureACL(conn) < 0) + return NULL; + + return virGetHostname(); +} + +static int chConnectNumOfDomains(virConnectPtr conn) +{ + virCHDriver *driver = conn->privateData; + + if (virConnectNumOfDomainsEnsureACL(conn) < 0) + return -1; + + return virDomainObjListNumOfDomains(driver->domains, true, + virConnectNumOfDomainsCheckACL, conn); +} + +static int chConnectListDomains(virConnectPtr conn, int *ids, int nids) +{ + virCHDriver *driver = conn->privateData; + + if (virConnectListDomainsEnsureACL(conn) < 0) + return -1; + + return virDomainObjListGetActiveIDs(driver->domains, ids, nids, + virConnectListDomainsCheckACL, conn); +} + +static int +chConnectListAllDomains(virConnectPtr conn, + virDomainPtr **domains, + unsigned int flags) +{ + virCHDriver *driver = conn->privateData; + + virCheckFlags(VIR_CONNECT_LIST_DOMAINS_FILTERS_ALL, -1); + + if (virConnectListAllDomainsEnsureACL(conn) < 0) + return -1; + + return virDomainObjListExport(driver->domains, conn, domains, + virConnectListAllDomainsCheckACL, flags); +} + +static int chNodeGetInfo(virConnectPtr conn, + virNodeInfoPtr nodeinfo) +{ + if (virNodeGetInfoEnsureACL(conn) < 0) + return -1; + + return virCapabilitiesGetNodeInfo(nodeinfo); +} + +static char *chConnectGetCapabilities(virConnectPtr conn) +{ + virCHDriver *driver = conn->privateData; + virCaps *caps; + char *xml; + + if (virConnectGetCapabilitiesEnsureACL(conn) < 0) + return NULL; + + if (!(caps = virCHDriverGetCapabilities(driver, true))) + return NULL; + + xml = virCapabilitiesFormatXML(caps); + + virObjectUnref(caps); + return xml; +} + +/** + * chDomainCreateXML: + * @conn: pointer to connection + * @xml: XML definition of domain + * @flags: bitwise-OR of supported virDomainCreateFlags + * + * Creates a domain based on xml and starts it + * + * Returns a new domain object or NULL in case of failure. + */ +static virDomainPtr +chDomainCreateXML(virConnectPtr conn, + const char *xml, + unsigned int flags) +{ + virCHDriver *driver = conn->privateData; + virDomainDef *vmdef = NULL; + virDomainObj *vm = NULL; + virDomainPtr dom = NULL; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE; + + virCheckFlags(VIR_DOMAIN_START_VALIDATE, NULL); + + if (flags & VIR_DOMAIN_START_VALIDATE) + parse_flags |= VIR_DOMAIN_DEF_PARSE_VALIDATE_SCHEMA; + + + if ((vmdef = virDomainDefParseString(xml, driver->xmlopt, + NULL, parse_flags)) == NULL) + goto cleanup; + + if (virDomainCreateXMLEnsureACL(conn, vmdef) < 0) + goto cleanup; + + if (!(vm = virDomainObjListAdd(driver->domains, + vmdef, + driver->xmlopt, + VIR_DOMAIN_OBJ_LIST_ADD_LIVE | + VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE, + NULL))) + goto cleanup; + + vmdef = NULL; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + if (virCHProcessStart(driver, vm, VIR_DOMAIN_RUNNING_BOOTED) < 0) + goto cleanup; + + dom = virGetDomain(conn, vm->def->name, vm->def->uuid, vm->def->id); + + virCHDomainObjEndJob(vm); + + cleanup: + if (!dom) { + virDomainObjListRemove(driver->domains, vm); + } + virDomainDefFree(vmdef); + virDomainObjEndAPI(&vm); + chDriverUnlock(driver); + return dom; +} + +static int +chDomainCreateWithFlags(virDomainPtr dom, unsigned int flags) +{ + virCHDriver *driver = dom->conn->privateData; + virDomainObj *vm; + int ret = -1; + + virCheckFlags(0, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainCreateWithFlagsEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + ret = virCHProcessStart(driver, vm, VIR_DOMAIN_RUNNING_BOOTED); + + virCHDomainObjEndJob(vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainCreate(virDomainPtr dom) +{ + return chDomainCreateWithFlags(dom, 0); +} + +static virDomainPtr +chDomainDefineXMLFlags(virConnectPtr conn, const char *xml, unsigned int flags) +{ + virCHDriver *driver = conn->privateData; + virDomainDef *vmdef = NULL; + virDomainObj *vm = NULL; + virDomainPtr dom = NULL; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE; + + virCheckFlags(VIR_DOMAIN_DEFINE_VALIDATE, NULL); + + if (flags & VIR_DOMAIN_START_VALIDATE) + parse_flags |= VIR_DOMAIN_DEF_PARSE_VALIDATE_SCHEMA; + + if ((vmdef = virDomainDefParseString(xml, driver->xmlopt, + NULL, parse_flags)) == NULL) + goto cleanup; + + if (virXMLCheckIllegalChars("name", vmdef->name, "\n") < 0) + goto cleanup; + + if (virDomainDefineXMLFlagsEnsureACL(conn, vmdef) < 0) + goto cleanup; + + if (!(vm = virDomainObjListAdd(driver->domains, vmdef, + driver->xmlopt, + 0, NULL))) + goto cleanup; + + vmdef = NULL; + vm->persistent = 1; + + dom = virGetDomain(conn, vm->def->name, vm->def->uuid, vm->def->id); + + cleanup: + virDomainDefFree(vmdef); + virDomainObjEndAPI(&vm); + return dom; +} + +static virDomainPtr +chDomainDefineXML(virConnectPtr conn, const char *xml) +{ + return chDomainDefineXMLFlags(conn, xml, 0); +} + +static int +chDomainUndefineFlags(virDomainPtr dom, + unsigned int flags) +{ + virCHDriver *driver = dom->conn->privateData; + virDomainObj *vm; + int ret = -1; + + virCheckFlags(0, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainUndefineFlagsEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (!vm->persistent) { + virReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("Cannot undefine transient domain")); + goto cleanup; + } + + if (virDomainObjIsActive(vm)) { + vm->persistent = 0; + } else { + virDomainObjListRemove(driver->domains, vm); + } + + ret = 0; + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainUndefine(virDomainPtr dom) +{ + return chDomainUndefineFlags(dom, 0); +} + +static int chDomainIsActive(virDomainPtr dom) +{ + virCHDriver *driver = dom->conn->privateData; + virDomainObj *vm; + int ret = -1; + + chDriverLock(driver); + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainIsActiveEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + ret = virDomainObjIsActive(vm); + + cleanup: + virDomainObjEndAPI(&vm); + chDriverUnlock(driver); + return ret; +} + +static int +chDomainShutdownFlags(virDomainPtr dom, + unsigned int flags) +{ + virCHDomainObjPrivate *priv; + virDomainObj *vm; + virDomainState state; + int ret = -1; + + virCheckFlags(VIR_DOMAIN_SHUTDOWN_ACPI_POWER_BTN, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + priv = vm->privateData; + + if (virDomainShutdownFlagsEnsureACL(dom->conn, vm->def, flags) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto endjob; + + state = virDomainObjGetState(vm, NULL); + if (state != VIR_DOMAIN_RUNNING && state != VIR_DOMAIN_PAUSED) { + virReportError(VIR_ERR_OPERATION_INVALID, "%s", + _("only can shutdown running/paused domain")); + goto endjob; + } else { + if (virCHMonitorShutdownVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to shutdown guest VM")); + goto endjob; + } + } + + virDomainObjSetState(vm, VIR_DOMAIN_SHUTDOWN, VIR_DOMAIN_SHUTDOWN_USER); + + ret = 0; + + endjob: + virCHDomainObjEndJob(vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainShutdown(virDomainPtr dom) +{ + return chDomainShutdownFlags(dom, 0); +} + + +static int +chDomainReboot(virDomainPtr dom, unsigned int flags) +{ + virCHDomainObjPrivate *priv; + virDomainObj *vm; + virDomainState state; + int ret = -1; + + virCheckFlags(VIR_DOMAIN_REBOOT_ACPI_POWER_BTN, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + priv = vm->privateData; + + if (virDomainRebootEnsureACL(dom->conn, vm->def, flags) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto endjob; + + state = virDomainObjGetState(vm, NULL); + if (state != VIR_DOMAIN_RUNNING && state != VIR_DOMAIN_PAUSED) { + virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", + _("only can reboot running/paused domain")); + goto endjob; + } else { + if (virCHMonitorRebootVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to reboot domain")); + goto endjob; + } + } + + if (state == VIR_DOMAIN_RUNNING) + virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, VIR_DOMAIN_RUNNING_BOOTED); + else + virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, VIR_DOMAIN_RUNNING_UNPAUSED); + + ret = 0; + + endjob: + virCHDomainObjEndJob(vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainSuspend(virDomainPtr dom) +{ + virCHDomainObjPrivate *priv; + virDomainObj *vm; + int ret = -1; + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + priv = vm->privateData; + + if (virDomainSuspendEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto endjob; + + if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_RUNNING) { + virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", + _("only can suspend running domain")); + goto endjob; + } else { + if (virCHMonitorSuspendVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to suspend domain")); + goto endjob; + } + } + + virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_USER); + + ret = 0; + + endjob: + virCHDomainObjEndJob(vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainResume(virDomainPtr dom) +{ + virCHDomainObjPrivate *priv; + virDomainObj *vm; + int ret = -1; + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + priv = vm->privateData; + + if (virDomainResumeEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_MODIFY) < 0) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto endjob; + + if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_PAUSED) { + virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", + _("only can resume paused domain")); + goto endjob; + } else { + if (virCHMonitorResumeVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to resume domain")); + goto endjob; + } + } + + virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, VIR_DOMAIN_RUNNING_UNPAUSED); + + ret = 0; + + endjob: + virCHDomainObjEndJob(vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +/** + * chDomainDestroyFlags: + * @dom: pointer to domain to destroy + * @flags: extra flags; not used yet. + * + * Sends SIGKILL to Cloud-Hypervisor process to terminate it + * + * Returns 0 on success or -1 in case of error + */ +static int +chDomainDestroyFlags(virDomainPtr dom, unsigned int flags) +{ + virCHDriver *driver = dom->conn->privateData; + virDomainObj *vm; + int ret = -1; + + virCheckFlags(0, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainDestroyFlagsEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (virCHDomainObjBeginJob(vm, CH_JOB_DESTROY) < 0) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto endjob; + + ret = virCHProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_DESTROYED); + + endjob: + virCHDomainObjEndJob(vm); + if (!vm->persistent) + virDomainObjListRemove(driver->domains, vm); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int +chDomainDestroy(virDomainPtr dom) +{ + return chDomainDestroyFlags(dom, 0); +} + +static virDomainPtr chDomainLookupByID(virConnectPtr conn, + int id) +{ + virCHDriver *driver = conn->privateData; + virDomainObj *vm; + virDomainPtr dom = NULL; + + chDriverLock(driver); + vm = virDomainObjListFindByID(driver->domains, id); + chDriverUnlock(driver); + + if (!vm) { + virReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching id '%d'"), id); + goto cleanup; + } + + if (virDomainLookupByIDEnsureACL(conn, vm->def) < 0) + goto cleanup; + + dom = virGetDomain(conn, vm->def->name, vm->def->uuid, vm->def->id); + + cleanup: + virDomainObjEndAPI(&vm); + return dom; +} + +static virDomainPtr chDomainLookupByName(virConnectPtr conn, + const char *name) +{ + virCHDriver *driver = conn->privateData; + virDomainObj *vm; + virDomainPtr dom = NULL; + + chDriverLock(driver); + vm = virDomainObjListFindByName(driver->domains, name); + chDriverUnlock(driver); + + if (!vm) { + virReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching name '%s'"), name); + goto cleanup; + } + + if (virDomainLookupByNameEnsureACL(conn, vm->def) < 0) + goto cleanup; + + dom = virGetDomain(conn, vm->def->name, vm->def->uuid, vm->def->id); + + cleanup: + virDomainObjEndAPI(&vm); + return dom; +} + +static virDomainPtr chDomainLookupByUUID(virConnectPtr conn, + const unsigned char *uuid) +{ + virCHDriver *driver = conn->privateData; + virDomainObj *vm; + virDomainPtr dom = NULL; + + chDriverLock(driver); + vm = virDomainObjListFindByUUID(driver->domains, uuid); + chDriverUnlock(driver); + + if (!vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(uuid, uuidstr); + virReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching uuid '%s'"), uuidstr); + goto cleanup; + } + + if (virDomainLookupByUUIDEnsureACL(conn, vm->def) < 0) + goto cleanup; + + dom = virGetDomain(conn, vm->def->name, vm->def->uuid, vm->def->id); + + cleanup: + virDomainObjEndAPI(&vm); + return dom; +} + +static int +chDomainGetState(virDomainPtr dom, + int *state, + int *reason, + unsigned int flags) +{ + virDomainObj *vm; + int ret = -1; + + virCheckFlags(0, -1); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainGetStateEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + *state = virDomainObjGetState(vm, reason); + ret = 0; + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static char *chDomainGetXMLDesc(virDomainPtr dom, + unsigned int flags) +{ + virCHDriver *driver = dom->conn->privateData; + virDomainObj *vm; + char *ret = NULL; + + virCheckFlags(VIR_DOMAIN_XML_COMMON_FLAGS, NULL); + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainGetXMLDescEnsureACL(dom->conn, vm->def, flags) < 0) + goto cleanup; + + ret = virDomainDefFormat(vm->def, driver->xmlopt, + virDomainDefFormatConvertXMLFlags(flags)); + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int chDomainGetInfo(virDomainPtr dom, + virDomainInfoPtr info) +{ + virDomainObj *vm; + int ret = -1; + + if (!(vm = chDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainGetInfoEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + info->state = virDomainObjGetState(vm, NULL); + + info->cpuTime = 0; + + info->maxMem = virDomainDefGetMemoryTotal(vm->def); + info->memory = vm->def->mem.cur_balloon; + info->nrVirtCpu = virDomainDefGetVcpus(vm->def); + + ret = 0; + + cleanup: + virDomainObjEndAPI(&vm); + return ret; +} + +static int chStateCleanup(void) +{ + if (ch_driver == NULL) + return -1; + + virObjectUnref(ch_driver->domains); + virObjectUnref(ch_driver->xmlopt); + virObjectUnref(ch_driver->caps); + virObjectUnref(ch_driver->config); + virMutexDestroy(&ch_driver->lock); + g_free(ch_driver); + + return 0; +} + +static int chStateInitialize(bool privileged, + const char *root, + virStateInhibitCallback callback G_GNUC_UNUSED, + void *opaque G_GNUC_UNUSED) +{ + if (root != NULL) { + virReportError(VIR_ERR_INVALID_ARG, "%s", + _("Driver does not support embedded mode")); + return -1; + } + + ch_driver = g_new0(virCHDriver, 1); + + if (virMutexInit(&ch_driver->lock) < 0) { + g_free(ch_driver); + return VIR_DRV_STATE_INIT_ERROR; + } + + if (!(ch_driver->domains = virDomainObjListNew())) + goto cleanup; + + if (!(ch_driver->caps = virCHDriverCapsInit())) + goto cleanup; + + if (!(ch_driver->xmlopt = chDomainXMLConfInit(ch_driver))) + goto cleanup; + + if (!(ch_driver->config = virCHDriverConfigNew(privileged))) + goto cleanup; + + if (chExtractVersion(ch_driver) < 0) + goto cleanup; + + return VIR_DRV_STATE_INIT_COMPLETE; + + cleanup: + chStateCleanup(); + return VIR_DRV_STATE_INIT_ERROR; +} + +/* Function Tables */ +static virHypervisorDriver chHypervisorDriver = { + .name = "CH", + .connectURIProbe = chConnectURIProbe, + .connectOpen = chConnectOpen, /* 7.3.0 */ + .connectClose = chConnectClose, /* 7.3.0 */ + .connectGetType = chConnectGetType, /* 7.3.0 */ + .connectGetVersion = chConnectGetVersion, /* 7.3.0 */ + .connectGetHostname = chConnectGetHostname, /* 7.3.0 */ + .connectNumOfDomains = chConnectNumOfDomains, /* 7.3.0 */ + .connectListAllDomains = chConnectListAllDomains, /* 7.3.0 */ + .connectListDomains = chConnectListDomains, /* 7.3.0 */ + .connectGetCapabilities = chConnectGetCapabilities, /* 7.3.0 */ + .domainCreateXML = chDomainCreateXML, /* 7.3.0 */ + .domainCreate = chDomainCreate, /* 7.3.0 */ + .domainCreateWithFlags = chDomainCreateWithFlags, /* 7.3.0 */ + .domainShutdown = chDomainShutdown, /* 7.3.0 */ + .domainShutdownFlags = chDomainShutdownFlags, /* 7.3.0 */ + .domainReboot = chDomainReboot, /* 7.3.0 */ + .domainSuspend = chDomainSuspend, /* 7.3.0 */ + .domainResume = chDomainResume, /* 7.3.0 */ + .domainDestroy = chDomainDestroy, /* 7.3.0 */ + .domainDestroyFlags = chDomainDestroyFlags, /* 7.3.0 */ + .domainDefineXML = chDomainDefineXML, /* 7.3.0 */ + .domainDefineXMLFlags = chDomainDefineXMLFlags, /* 7.3.0 */ + .domainUndefine = chDomainUndefine, /* 7.3.0 */ + .domainUndefineFlags = chDomainUndefineFlags, /* 7.3.0 */ + .domainLookupByID = chDomainLookupByID, /* 7.3.0 */ + .domainLookupByUUID = chDomainLookupByUUID, /* 7.3.0 */ + .domainLookupByName = chDomainLookupByName, /* 7.3.0 */ + .domainGetState = chDomainGetState, /* 7.3.0 */ + .domainGetXMLDesc = chDomainGetXMLDesc, /* 7.3.0 */ + .domainGetInfo = chDomainGetInfo, /* 7.3.0 */ + .domainIsActive = chDomainIsActive, /* 7.3.0 */ + .nodeGetInfo = chNodeGetInfo, /* 7.3.0 */ +}; + +static virConnectDriver chConnectDriver = { + .localOnly = true, + .uriSchemes = (const char *[]){"ch", NULL}, + .hypervisorDriver = &chHypervisorDriver, +}; + +static virStateDriver chStateDriver = { + .name = "cloud-hypervisor", + .stateInitialize = chStateInitialize, + .stateCleanup = chStateCleanup, +}; + +int chRegister(void) +{ + if (virRegisterConnectDriver(&chConnectDriver, false) < 0) + return -1; + if (virRegisterStateDriver(&chStateDriver) < 0) + return -1; + return 0; +} diff --git a/src/ch/ch_driver.h b/src/ch/ch_driver.h new file mode 100644 index 0000000000..933be3953b --- /dev/null +++ b/src/ch/ch_driver.h @@ -0,0 +1,24 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_driver.h: header file for Cloud-Hypervisor driver functions + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#pragma once + +/* Function declarations */ +int chRegister(void); diff --git a/src/ch/ch_monitor.c b/src/ch/ch_monitor.c new file mode 100644 index 0000000000..2968c2aae7 --- /dev/null +++ b/src/ch/ch_monitor.c @@ -0,0 +1,837 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_monitor.c: Manage Cloud-Hypervisor interactions + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#include <config.h> + +#include <stdio.h> +#include <unistd.h> +#include <curl/curl.h> + +#include "ch_conf.h" +#include "ch_monitor.h" +#include "viralloc.h" +#include "vircommand.h" +#include "virerror.h" +#include "virfile.h" +#include "virjson.h" +#include "virlog.h" +#include "virstring.h" +#include "virtime.h" + +#define VIR_FROM_THIS VIR_FROM_CH + +VIR_LOG_INIT("ch.ch_monitor"); + +static virClass *virCHMonitorClass; +static void virCHMonitorDispose(void *obj); + +static int virCHMonitorOnceInit(void) +{ + if (!VIR_CLASS_NEW(virCHMonitor, virClassForObjectLockable())) + return -1; + + return 0; +} + +VIR_ONCE_GLOBAL_INIT(virCHMonitor); + +int virCHMonitorShutdownVMM(virCHMonitor *mon); +int virCHMonitorPutNoContent(virCHMonitor *mon, const char *endpoint); +int virCHMonitorGet(virCHMonitor *mon, const char *endpoint); + +static int +virCHMonitorBuildCPUJson(virJSONValue *content, virDomainDef *vmdef) +{ + virJSONValue *cpus; + unsigned int maxvcpus = 0; + unsigned int nvcpus = 0; + virDomainVcpuDef *vcpu; + size_t i; + + /* count maximum allowed number vcpus and enabled vcpus when boot.*/ + maxvcpus = virDomainDefGetVcpusMax(vmdef); + for (i = 0; i < maxvcpus; i++) { + vcpu = virDomainDefGetVcpu(vmdef, i); + if (vcpu->online) + nvcpus++; + } + + if (maxvcpus != 0 || nvcpus != 0) { + cpus = virJSONValueNewObject(); + if (virJSONValueObjectAppendNumberInt(cpus, "boot_vcpus", nvcpus) < 0) + goto cleanup; + if (virJSONValueObjectAppendNumberInt(cpus, "max_vcpus", vmdef->maxvcpus) < 0) + goto cleanup; + if (virJSONValueObjectAppend(content, "cpus", &cpus) < 0) + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(cpus); + return -1; +} + +static int +virCHMonitorBuildKernelRelatedJson(virJSONValue *content, virDomainDef *vmdef) +{ + virJSONValue *kernel = virJSONValueNewObject(); + virJSONValue *cmdline = virJSONValueNewObject(); + virJSONValue *initramfs = virJSONValueNewObject(); + + if (vmdef->os.kernel == NULL) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Kernel image path in this domain is not defined")); + goto cleanup; + } else { + kernel = virJSONValueNewObject(); + if (virJSONValueObjectAppendString(kernel, "path", vmdef->os.kernel) < 0) + goto cleanup; + if (virJSONValueObjectAppend(content, "kernel", &kernel) < 0) + goto cleanup; + } + + if (vmdef->os.cmdline) { + if (virJSONValueObjectAppendString(cmdline, "args", vmdef->os.cmdline) < 0) + goto cleanup; + if (virJSONValueObjectAppend(content, "cmdline", &cmdline) < 0) + goto cleanup; + } + + if (vmdef->os.initrd != NULL) { + initramfs = virJSONValueNewObject(); + if (virJSONValueObjectAppendString(initramfs, "path", vmdef->os.initrd) < 0) + goto cleanup; + if (virJSONValueObjectAppend(content, "initramfs", &initramfs) < 0) + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(kernel); + virJSONValueFree(cmdline); + virJSONValueFree(initramfs); + + return -1; +} + +static int +virCHMonitorBuildMemoryJson(virJSONValue *content, virDomainDef *vmdef) +{ + virJSONValue *memory; + unsigned long long total_memory = virDomainDefGetMemoryInitial(vmdef) * 1024; + + if (total_memory != 0) { + memory = virJSONValueNewObject(); + if (virJSONValueObjectAppendNumberUlong(memory, "size", total_memory) < 0) + goto cleanup; + if (virJSONValueObjectAppend(content, "memory", &memory) < 0) + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(memory); + return -1; +} + +static int +virCHMonitorBuildDiskJson(virJSONValue *disks, virDomainDiskDef *diskdef) +{ + virJSONValue *disk = virJSONValueNewObject(); + + if (!diskdef->src) + goto cleanup; + + switch (diskdef->src->type) { + case VIR_STORAGE_TYPE_FILE: + if (!diskdef->src->path) { + virReportError(VIR_ERR_INVALID_ARG, "%s", + _("Missing disk file path in domain")); + goto cleanup; + } + if (diskdef->bus != VIR_DOMAIN_DISK_BUS_VIRTIO) { + virReportError(VIR_ERR_INVALID_ARG, + _("Only virtio bus types are supported for '%s'"), diskdef->src->path); + goto cleanup; + } + if (virJSONValueObjectAppendString(disk, "path", diskdef->src->path) < 0) + goto cleanup; + if (diskdef->src->readonly) { + if (virJSONValueObjectAppendBoolean(disk, "readonly", true) < 0) + goto cleanup; + } + if (virJSONValueArrayAppend(disks, &disk) < 0) + goto cleanup; + + break; + case VIR_STORAGE_TYPE_NONE: + case VIR_STORAGE_TYPE_BLOCK: + case VIR_STORAGE_TYPE_DIR: + case VIR_STORAGE_TYPE_NETWORK: + case VIR_STORAGE_TYPE_VOLUME: + case VIR_STORAGE_TYPE_NVME: + case VIR_STORAGE_TYPE_VHOST_USER: + default: + virReportEnumRangeError(virStorageType, diskdef->src->type); + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(disk); + return -1; +} + +static int +virCHMonitorBuildDisksJson(virJSONValue *content, virDomainDef *vmdef) +{ + virJSONValue *disks; + size_t i; + + if (vmdef->ndisks > 0) { + disks = virJSONValueNewArray(); + + for (i = 0; i < vmdef->ndisks; i++) { + if (virCHMonitorBuildDiskJson(disks, vmdef->disks[i]) < 0) + goto cleanup; + } + if (virJSONValueObjectAppend(content, "disks", &disks) < 0) + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(disks); + return -1; +} + +static int +virCHMonitorBuildNetJson(virJSONValue *nets, virDomainNetDef *netdef) +{ + virDomainNetType netType = virDomainNetGetActualType(netdef); + char macaddr[VIR_MAC_STRING_BUFLEN]; + virJSONValue *net; + + // check net type at first + net = virJSONValueNewObject(); + + switch (netType) { + case VIR_DOMAIN_NET_TYPE_ETHERNET: + if (netdef->guestIP.nips == 1) { + const virNetDevIPAddr *ip = netdef->guestIP.ips[0]; + g_autofree char *addr = NULL; + virSocketAddr netmask; + g_autofree char *netmaskStr = NULL; + if (!(addr = virSocketAddrFormat(&ip->address))) + goto cleanup; + if (virJSONValueObjectAppendString(net, "ip", addr) < 0) + goto cleanup; + + if (virSocketAddrPrefixToNetmask(ip->prefix, &netmask, AF_INET) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Failed to translate net prefix %d to netmask"), + ip->prefix); + goto cleanup; + } + if (!(netmaskStr = virSocketAddrFormat(&netmask))) + goto cleanup; + if (virJSONValueObjectAppendString(net, "mask", netmaskStr) < 0) + goto cleanup; + } else if (netdef->guestIP.nips > 1) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("ethernet type supports a single guest ip")); + } + break; + case VIR_DOMAIN_NET_TYPE_VHOSTUSER: + if ((virDomainChrType)netdef->data.vhostuser->type != VIR_DOMAIN_CHR_TYPE_UNIX) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("vhost_user type support UNIX socket in this CH")); + goto cleanup; + } else { + if (virJSONValueObjectAppendString(net, "vhost_socket", netdef->data.vhostuser->data.nix.path) < 0) + goto cleanup; + if (virJSONValueObjectAppendBoolean(net, "vhost_user", true) < 0) + goto cleanup; + } + break; + case VIR_DOMAIN_NET_TYPE_BRIDGE: + case VIR_DOMAIN_NET_TYPE_NETWORK: + case VIR_DOMAIN_NET_TYPE_DIRECT: + case VIR_DOMAIN_NET_TYPE_USER: + case VIR_DOMAIN_NET_TYPE_SERVER: + case VIR_DOMAIN_NET_TYPE_CLIENT: + case VIR_DOMAIN_NET_TYPE_MCAST: + case VIR_DOMAIN_NET_TYPE_INTERNAL: + case VIR_DOMAIN_NET_TYPE_HOSTDEV: + case VIR_DOMAIN_NET_TYPE_UDP: + case VIR_DOMAIN_NET_TYPE_VDPA: + case VIR_DOMAIN_NET_TYPE_LAST: + default: + virReportEnumRangeError(virDomainNetType, netType); + goto cleanup; + } + + if (netdef->ifname != NULL) { + if (virJSONValueObjectAppendString(net, "tap", netdef->ifname) < 0) + goto cleanup; + } + if (virJSONValueObjectAppendString(net, "mac", virMacAddrFormat(&netdef->mac, macaddr)) < 0) + goto cleanup; + + + if (netdef->virtio != NULL) { + if (netdef->virtio->iommu == VIR_TRISTATE_SWITCH_ON) { + if (virJSONValueObjectAppendBoolean(net, "iommu", true) < 0) + goto cleanup; + } + } + if (netdef->driver.virtio.queues) { + if (virJSONValueObjectAppendNumberInt(net, "num_queues", netdef->driver.virtio.queues) < 0) + goto cleanup; + } + + if (netdef->driver.virtio.rx_queue_size || netdef->driver.virtio.tx_queue_size) { + if (netdef->driver.virtio.rx_queue_size != netdef->driver.virtio.tx_queue_size) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("virtio rx_queue_size option %d is not same with tx_queue_size %d"), + netdef->driver.virtio.rx_queue_size, + netdef->driver.virtio.tx_queue_size); + goto cleanup; + } + if (virJSONValueObjectAppendNumberInt(net, "queue_size", netdef->driver.virtio.rx_queue_size) < 0) + goto cleanup; + } + + if (virJSONValueArrayAppend(nets, &net) < 0) + goto cleanup; + + return 0; + + cleanup: + virJSONValueFree(net); + return -1; +} + +static int +virCHMonitorBuildNetsJson(virJSONValue *content, virDomainDef *vmdef) +{ + virJSONValue *nets; + size_t i; + + if (vmdef->nnets > 0) { + nets = virJSONValueNewArray(); + + for (i = 0; i < vmdef->nnets; i++) { + if (virCHMonitorBuildNetJson(nets, vmdef->nets[i]) < 0) + goto cleanup; + } + if (virJSONValueObjectAppend(content, "net", &nets) < 0) + goto cleanup; + } + + return 0; + + cleanup: + virJSONValueFree(nets); + return -1; +} + +static int +virCHMonitorDetectUnsupportedDevices(virDomainDef *vmdef) +{ + int ret = 0; + + if (vmdef->ngraphics > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support graphics")); + ret = 1; + } + if (vmdef->ncontrollers > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support controllers")); + ret = 1; + } + if (vmdef->nfss > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support fss")); + ret = 1; + } + if (vmdef->ninputs > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support inputs")); + ret = 1; + } + if (vmdef->nsounds > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support sounds")); + ret = 1; + } + if (vmdef->naudios > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support audios")); + ret = 1; + } + if (vmdef->nvideos > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support videos")); + ret = 1; + } + if (vmdef->nhostdevs > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support hostdevs")); + ret = 1; + } + if (vmdef->nredirdevs > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support redirdevs")); + ret = 1; + } + if (vmdef->nsmartcards > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support smartcards")); + ret = 1; + } + if (vmdef->nserials > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support serials")); + ret = 1; + } + if (vmdef->nparallels > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support parallels")); + ret = 1; + } + if (vmdef->nchannels > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support channels")); + ret = 1; + } + if (vmdef->nconsoles > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support consoles")); + ret = 1; + } + if (vmdef->nleases > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support leases")); + ret = 1; + } + if (vmdef->nhubs > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support hubs")); + ret = 1; + } + if (vmdef->nseclabels > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support seclabels")); + ret = 1; + } + if (vmdef->nrngs > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support rngs")); + ret = 1; + } + if (vmdef->nshmems > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support shmems")); + ret = 1; + } + if (vmdef->nmems > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support mems")); + ret = 1; + } + if (vmdef->npanics > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support panics")); + ret = 1; + } + if (vmdef->nsysinfo > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Cloud-Hypervisor doesn't support sysinfo")); + ret = 1; + } + + return ret; +} + +static int +virCHMonitorBuildVMJson(virDomainDef *vmdef, char **jsonstr) +{ + virJSONValue *content = virJSONValueNewObject(); + int ret = -1; + + if (vmdef == NULL) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("VM is not defined")); + goto cleanup; + } + + if (virCHMonitorDetectUnsupportedDevices(vmdef)) + goto cleanup; + + if (virCHMonitorBuildCPUJson(content, vmdef) < 0) + goto cleanup; + + if (virCHMonitorBuildMemoryJson(content, vmdef) < 0) + goto cleanup; + + if (virCHMonitorBuildKernelRelatedJson(content, vmdef) < 0) + goto cleanup; + + if (virCHMonitorBuildDisksJson(content, vmdef) < 0) + goto cleanup; + + if (virCHMonitorBuildNetsJson(content, vmdef) < 0) + goto cleanup; + + if (!(*jsonstr = virJSONValueToString(content, false))) + goto cleanup; + + ret = 0; + + cleanup: + virJSONValueFree(content); + return ret; +} + +static int +chMonitorCreateSocket(const char *socket_path) +{ + struct sockaddr_un addr; + socklen_t addrlen = sizeof(addr); + int fd; + + if ((fd = socket(AF_UNIX, SOCK_STREAM, 0)) < 0) { + virReportSystemError(errno, "%s", + _("Unable to create UNIX socket")); + goto error; + } + + memset(&addr, 0, sizeof(addr)); + addr.sun_family = AF_UNIX; + if (virStrcpyStatic(addr.sun_path, socket_path) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("UNIX socket path '%s' too long"), + socket_path); + goto error; + } + + if (unlink(socket_path) < 0 && errno != ENOENT) { + virReportSystemError(errno, + _("Unable to unlink %s"), + socket_path); + goto error; + } + + if (bind(fd, (struct sockaddr *)&addr, addrlen) < 0) { + virReportSystemError(errno, + _("Unable to bind to UNIX socket path '%s'"), + socket_path); + goto error; + } + + if (listen(fd, 1) < 0) { + virReportSystemError(errno, + _("Unable to listen to UNIX socket path '%s'"), + socket_path); + goto error; + } + + /* We run cloud-hypervisor with umask 0002. Compensate for the umask + * libvirtd might be running under to get the same permission + * cloud-hypervisor would have. */ + if (virFileUpdatePerm(socket_path, 0002, 0664) < 0) + goto error; + + return fd; + + error: + VIR_FORCE_CLOSE(fd); + return -1; +} + +virCHMonitor * +virCHMonitorNew(virDomainObj *vm, const char *socketdir) +{ + virCHMonitor *ret = NULL; + virCHMonitor *mon = NULL; + virCommand *cmd = NULL; + int socket_fd = 0; + + if (virCHMonitorInitialize() < 0) + return NULL; + + if (!(mon = virObjectLockableNew(virCHMonitorClass))) + return NULL; + + if (!vm->def) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("VM is not defined")); + return NULL; + } + + /* prepare to launch Cloud-Hypervisor socket */ + mon->socketpath = g_strdup_printf("%s/%s-socket", socketdir, vm->def->name); + if (g_mkdir_with_parents(socketdir, 0777) < 0) { + virReportSystemError(errno, + _("Cannot create socket directory '%s'"), + socketdir); + goto cleanup; + } + + cmd = virCommandNew(vm->def->emulator); + virCommandSetUmask(cmd, 0x002); + socket_fd = chMonitorCreateSocket(mon->socketpath); + if (socket_fd < 0) { + virReportSystemError(errno, + _("Cannot create socket '%s'"), + mon->socketpath); + goto cleanup; + } + + virCommandAddArg(cmd, "--api-socket"); + virCommandAddArgFormat(cmd, "fd=%d", socket_fd); + virCommandPassFD(cmd, socket_fd, VIR_COMMAND_PASS_FD_CLOSE_PARENT); + + /* launch Cloud-Hypervisor socket */ + if (virCommandRunAsync(cmd, &mon->pid) < 0) + goto cleanup; + + /* get a curl handle */ + mon->handle = curl_easy_init(); + + /* now has its own reference */ + virObjectRef(mon); + mon->vm = virObjectRef(vm); + + ret = mon; + + cleanup: + virCommandFree(cmd); + return ret; +} + +static void virCHMonitorDispose(void *opaque) +{ + virCHMonitor *mon = opaque; + + VIR_DEBUG("mon=%p", mon); + virObjectUnref(mon->vm); +} + +void virCHMonitorClose(virCHMonitor *mon) +{ + if (!mon) + return; + + if (mon->pid > 0) { + /* try cleaning up the Cloud-Hypervisor process */ + virProcessAbort(mon->pid); + mon->pid = 0; + } + + if (mon->handle) + curl_easy_cleanup(mon->handle); + + if (mon->socketpath) { + if (virFileRemove(mon->socketpath, -1, -1) < 0) { + VIR_WARN("Unable to remove CH socket file '%s'", + mon->socketpath); + } + g_free(mon->socketpath); + } + + virObjectUnref(mon); + if (mon->vm) + virObjectUnref(mon->vm); +} + +static int +virCHMonitorCurlPerform(CURL *handle) +{ + CURLcode errorCode; + long responseCode = 0; + + errorCode = curl_easy_perform(handle); + + if (errorCode != CURLE_OK) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("curl_easy_perform() returned an error: %s (%d)"), + curl_easy_strerror(errorCode), errorCode); + return -1; + } + + errorCode = curl_easy_getinfo(handle, CURLINFO_RESPONSE_CODE, + &responseCode); + + if (errorCode != CURLE_OK) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("curl_easy_getinfo(CURLINFO_RESPONSE_CODE) returned an " + "error: %s (%d)"), curl_easy_strerror(errorCode), + errorCode); + return -1; + } + + if (responseCode < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("curl_easy_getinfo(CURLINFO_RESPONSE_CODE) returned a " + "negative response code")); + return -1; + } + + return responseCode; +} + +int +virCHMonitorPutNoContent(virCHMonitor *mon, const char *endpoint) +{ + g_autofree char *url = NULL; + int responseCode = 0; + int ret = -1; + + url = g_strdup_printf("%s/%s", URL_ROOT, endpoint); + + virObjectLock(mon); + + /* reset all options of a libcurl session handle at first */ + curl_easy_reset(mon->handle); + + curl_easy_setopt(mon->handle, CURLOPT_UNIX_SOCKET_PATH, mon->socketpath); + curl_easy_setopt(mon->handle, CURLOPT_URL, url); + curl_easy_setopt(mon->handle, CURLOPT_PUT, true); + curl_easy_setopt(mon->handle, CURLOPT_HTTPHEADER, NULL); + + responseCode = virCHMonitorCurlPerform(mon->handle); + + virObjectUnlock(mon); + + if (responseCode == 200 || responseCode == 204) + ret = 0; + + return ret; +} + +int +virCHMonitorGet(virCHMonitor *mon, const char *endpoint) +{ + g_autofree char *url = NULL; + int responseCode = 0; + int ret = -1; + + url = g_strdup_printf("%s/%s", URL_ROOT, endpoint); + + virObjectLock(mon); + + /* reset all options of a libcurl session handle at first */ + curl_easy_reset(mon->handle); + + curl_easy_setopt(mon->handle, CURLOPT_UNIX_SOCKET_PATH, mon->socketpath); + curl_easy_setopt(mon->handle, CURLOPT_URL, url); + + responseCode = virCHMonitorCurlPerform(mon->handle); + + virObjectUnlock(mon); + + if (responseCode == 200 || responseCode == 204) + ret = 0; + + return ret; +} + +int +virCHMonitorShutdownVMM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VMM_SHUTDOWN); +} + +int +virCHMonitorCreateVM(virCHMonitor *mon) +{ + g_autofree char *url = NULL; + int responseCode = 0; + int ret = -1; + g_autofree char *payload = NULL; + struct curl_slist *headers = NULL; + + url = g_strdup_printf("%s/%s", URL_ROOT, URL_VM_CREATE); + headers = curl_slist_append(headers, "Accept: application/json"); + headers = curl_slist_append(headers, "Content-Type: application/json"); + + if (virCHMonitorBuildVMJson(mon->vm->def, &payload) != 0) + return -1; + + virObjectLock(mon); + + /* reset all options of a libcurl session handle at first */ + curl_easy_reset(mon->handle); + + curl_easy_setopt(mon->handle, CURLOPT_UNIX_SOCKET_PATH, mon->socketpath); + curl_easy_setopt(mon->handle, CURLOPT_URL, url); + curl_easy_setopt(mon->handle, CURLOPT_CUSTOMREQUEST, "PUT"); + curl_easy_setopt(mon->handle, CURLOPT_HTTPHEADER, headers); + curl_easy_setopt(mon->handle, CURLOPT_POSTFIELDS, payload); + + responseCode = virCHMonitorCurlPerform(mon->handle); + + virObjectUnlock(mon); + + if (responseCode == 200 || responseCode == 204) + ret = 0; + + curl_slist_free_all(headers); + return ret; +} + +int +virCHMonitorBootVM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VM_BOOT); +} + +int +virCHMonitorShutdownVM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VM_SHUTDOWN); +} + +int +virCHMonitorRebootVM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VM_REBOOT); +} + +int +virCHMonitorSuspendVM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VM_Suspend); +} + +int +virCHMonitorResumeVM(virCHMonitor *mon) +{ + return virCHMonitorPutNoContent(mon, URL_VM_RESUME); +} diff --git a/src/ch/ch_monitor.h b/src/ch/ch_monitor.h new file mode 100644 index 0000000000..e717e11cbc --- /dev/null +++ b/src/ch/ch_monitor.h @@ -0,0 +1,60 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_monitor.h: header file for managing Cloud-Hypervisor interactions + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#pragma once + +#include <curl/curl.h> + +#include "virobject.h" +#include "domain_conf.h" + +#define URL_ROOT "http://localhost/api/v1" +#define URL_VMM_SHUTDOWN "vmm.shutdown" +#define URL_VM_CREATE "vm.create" +#define URL_VM_DELETE "vm.delete" +#define URL_VM_BOOT "vm.boot" +#define URL_VM_SHUTDOWN "vm.shutdown" +#define URL_VM_REBOOT "vm.reboot" +#define URL_VM_Suspend "vm.pause" +#define URL_VM_RESUME "vm.resume" + +typedef struct _virCHMonitor virCHMonitor; + +struct _virCHMonitor { + virObjectLockable parent; + + CURL *handle; + + char *socketpath; + + pid_t pid; + + virDomainObj *vm; +}; + +virCHMonitor *virCHMonitorNew(virDomainObj *vm, const char *socketdir); +void virCHMonitorClose(virCHMonitor *mon); + +int virCHMonitorCreateVM(virCHMonitor *mon); +int virCHMonitorBootVM(virCHMonitor *mon); +int virCHMonitorShutdownVM(virCHMonitor *mon); +int virCHMonitorRebootVM(virCHMonitor *mon); +int virCHMonitorSuspendVM(virCHMonitor *mon); +int virCHMonitorResumeVM(virCHMonitor *mon); diff --git a/src/ch/ch_process.c b/src/ch/ch_process.c new file mode 100644 index 0000000000..93b1f7f97e --- /dev/null +++ b/src/ch/ch_process.c @@ -0,0 +1,126 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_process.c: Process controller for Cloud-Hypervisor driver + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#include <config.h> + +#include <unistd.h> +#include <fcntl.h> + +#include "ch_domain.h" +#include "ch_monitor.h" +#include "ch_process.h" +#include "viralloc.h" +#include "virerror.h" +#include "virlog.h" + +#define VIR_FROM_THIS VIR_FROM_CH + +VIR_LOG_INIT("ch.ch_process"); + +#define START_SOCKET_POSTFIX ": starting up socket\n" +#define START_VM_POSTFIX ": starting up vm\n" + + + +static virCHMonitor * +virCHProcessConnectMonitor(virCHDriver *driver, + virDomainObj *vm) +{ + virCHMonitor *monitor = NULL; + virCHDriverConfig *cfg = virCHDriverGetConfig(driver); + + monitor = virCHMonitorNew(vm, cfg->stateDir); + + virObjectUnref(cfg); + return monitor; +} + +/** + * virCHProcessStart: + * @driver: pointer to driver structure + * @vm: pointer to virtual machine structure + * @reason: reason for switching vm to running state + * + * Starts Cloud-Hypervisor listen on a local socket + * + * Returns 0 on success or -1 in case of error + */ +int virCHProcessStart(virCHDriver *driver, + virDomainObj *vm, + virDomainRunningReason reason) +{ + int ret = -1; + virCHDomainObjPrivate *priv = vm->privateData; + + if (!priv->monitor) { + /* And we can get the first monitor connection now too */ + if (!(priv->monitor = virCHProcessConnectMonitor(driver, vm))) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to create connection to CH socket")); + goto cleanup; + } + + if (virCHMonitorCreateVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to create guest VM")); + goto cleanup; + } + } + + if (virCHMonitorBootVM(priv->monitor) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to boot guest VM")); + goto cleanup; + } + + vm->pid = priv->monitor->pid; + vm->def->id = vm->pid; + virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, reason); + + return 0; + + cleanup: + if (ret) + virCHProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED); + + return ret; +} + +int virCHProcessStop(virCHDriver *driver G_GNUC_UNUSED, + virDomainObj *vm, + virDomainShutoffReason reason) +{ + virCHDomainObjPrivate *priv = vm->privateData; + + VIR_DEBUG("Stopping VM name=%s pid=%d reason=%d", + vm->def->name, (int)vm->pid, (int)reason); + + if (priv->monitor) { + virCHMonitorClose(priv->monitor); + priv->monitor = NULL; + } + + vm->pid = -1; + vm->def->id = -1; + + virDomainObjSetState(vm, VIR_DOMAIN_SHUTOFF, reason); + + return 0; +} diff --git a/src/ch/ch_process.h b/src/ch/ch_process.h new file mode 100644 index 0000000000..abc4915979 --- /dev/null +++ b/src/ch/ch_process.h @@ -0,0 +1,31 @@ +/* + * Copyright Intel Corp. 2020-2021 + * + * ch_process.h: header file for Cloud-Hypervisor's process controller + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library. If not, see + * <http://www.gnu.org/licenses/>. + */ + +#pragma once + +#include "ch_conf.h" +#include "internal.h" + +int virCHProcessStart(virCHDriver *driver, + virDomainObj *vm, + virDomainRunningReason reason); +int virCHProcessStop(virCHDriver *driver, + virDomainObj *vm, + virDomainShutoffReason reason); diff --git a/src/ch/meson.build b/src/ch/meson.build new file mode 100644 index 0000000000..e34974d56c --- /dev/null +++ b/src/ch/meson.build @@ -0,0 +1,74 @@ +ch_driver_sources = [ + 'ch_conf.c', + 'ch_conf.h', + 'ch_domain.c', + 'ch_domain.h', + 'ch_driver.c', + 'ch_driver.h', + 'ch_monitor.c', + 'ch_monitor.h', + 'ch_process.c', + 'ch_process.h', +] + +driver_source_files += files(ch_driver_sources) + +stateful_driver_source_files += files(ch_driver_sources) + +if conf.has('WITH_CH') + ch_driver_impl = static_library( + 'virt_driver_ch_impl', + [ + ch_driver_sources, + ], + dependencies: [ + access_dep, + curl_dep, + log_dep, + src_dep, + ], + include_directories: [ + conf_inc_dir, + ], + ) + + virt_modules += { + 'name': 'virt_driver_ch', + 'link_whole': [ + ch_driver_impl, + ], + 'link_args': [ + libvirt_no_undefined, + ], + } + + virt_daemons += { + 'name': 'virtchd', + 'c_args': [ + '-DDAEMON_NAME="virtchd"', + '-DMODULE_NAME="ch"', + ], + } + + virt_daemon_confs += { + 'name': 'virtchd', + } + + virt_daemon_units += { + 'service': 'virtchd', + 'service_in': files('virtchd.service.in'), + 'name': 'Libvirt ch', + 'sockprefix': 'virtchd', + 'sockets': [ 'main', 'ro', 'admin' ], + } + + sysconf_files += { + 'name': 'virtchd', + 'file': files('virtchd.sysconf'), + } + + virt_install_dirs += [ + localstatedir / 'lib' / 'libvirt' / 'ch', + runstatedir / 'libvirt' / 'ch', + ] +endif diff --git a/src/ch/virtchd.service.in b/src/ch/virtchd.service.in new file mode 100644 index 0000000000..cc1e85d1df --- /dev/null +++ b/src/ch/virtchd.service.in @@ -0,0 +1,47 @@ +[Unit] +Description=Virtualization Cloud-Hypervisor daemon +Conflicts=libvirtd.service +Requires=virtchd.socket +Requires=virtchd-ro.socket +Requires=virtchd-admin.socket +Wants=systemd-machined.service +Before=libvirt-guests.service +After=network.target +After=dbus.service +After=apparmor.service +After=local-fs.target +After=remote-fs.target +After=systemd-logind.service +After=systemd-machined.service +Documentation=man:libvirtd(8) +Documentation=https://libvirt.org + +[Service] +Type=notify +EnvironmentFile=-@sysconfdir@/sysconfig/virtchd +ExecStart=@sbindir@/virtchd $VIRTCHD_ARGS +ExecReload=/bin/kill -HUP $MAINPID +KillMode=process +Restart=on-failure +# At least 2 FD per guest (eg ch monitor + ch socket). +# eg if we want to support 4096 guests, we'll typically need 8192 FDs +# If changing this, also consider virtlogd.service & virtlockd.service +# limits which are also related to number of guests +LimitNOFILE=8192 +# The cgroups pids controller can limit the number of tasks started by +# the daemon, which can limit the number of domains for some hypervisors. +# A conservative default of 8 tasks per guest results in a TasksMax of +# 32k to support 4096 guests. +TasksMax=32768 +# With cgroups v2 there is no devices controller anymore, we have to use +# eBPF to control access to devices. In order to do that we create a eBPF +# hash MAP which locks memory. The default map size for 64 devices together +# with program takes 12k per guest. After rounding up we will get 64M to +# support 4096 guests. +LimitMEMLOCK=64M + +[Install] +WantedBy=multi-user.target +Also=virtchd.socket +Also=virtchd-ro.socket +Also=virtchd-admin.socket diff --git a/src/ch/virtchd.sysconf b/src/ch/virtchd.sysconf new file mode 100644 index 0000000000..5ee44be5cf --- /dev/null +++ b/src/ch/virtchd.sysconf @@ -0,0 +1,3 @@ +# Customizations for the virtchd.service systemd unit + +VIRTCHD_ARGS="--timeout 120" diff --git a/src/meson.build b/src/meson.build index c7ff9e978c..2bd88e6699 100644 --- a/src/meson.build +++ b/src/meson.build @@ -271,6 +271,7 @@ subdir('esx') subdir('hyperv') subdir('libxl') subdir('lxc') +subdir('ch') subdir('openvz') subdir('qemu') subdir('test') diff --git a/src/remote/remote_daemon.c b/src/remote/remote_daemon.c index ec2661d0a8..7076fe3294 100644 --- a/src/remote/remote_daemon.c +++ b/src/remote/remote_daemon.c @@ -169,6 +169,10 @@ static int daemonInitialize(void) if (virDriverLoadModule("qemu", "qemuRegister", false) < 0) return -1; # endif +# ifdef WITH_CH + if (virDriverLoadModule("ch", "chRegister", false) < 0) + return -1; +# endif # ifdef WITH_LXC if (virDriverLoadModule("lxc", "lxcRegister", false) < 0) return -1; diff --git a/src/remote/remote_daemon_dispatch.c b/src/remote/remote_daemon_dispatch.c index 1b4f5256c3..e500b41564 100644 --- a/src/remote/remote_daemon_dispatch.c +++ b/src/remote/remote_daemon_dispatch.c @@ -2139,7 +2139,8 @@ remoteDispatchConnectOpen(virNetServer *server G_GNUC_UNUSED, STREQ(type, "VBOX") || STREQ(type, "bhyve") || STREQ(type, "vz") || - STREQ(type, "Parallels")) { + STREQ(type, "Parallels") || + STREQ(type, "CH")) { VIR_DEBUG("Hypervisor driver found, setting URIs for secondary drivers"); if (getuid() == 0) { priv->interfaceURI = "interface:///system"; diff --git a/src/util/virerror.c b/src/util/virerror.c index 1746487f7d..d9e2c65dc8 100644 --- a/src/util/virerror.c +++ b/src/util/virerror.c @@ -145,6 +145,7 @@ VIR_ENUM_IMPL(virErrorDomain, "TPM", /* 70 */ "BPF", + "Cloud-Hypervisor Driver", ); diff --git a/tools/virsh.c b/tools/virsh.c index 7d7109cfdf..70355a606b 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -506,6 +506,9 @@ virshShowVersion(vshControl *ctl G_GNUC_UNUSED) #ifdef WITH_OPENVZ vshPrint(ctl, " OpenVZ"); #endif +#ifdef WITH_CH + vshPrint(ctl, " Cloud-Hypervisor"); +#endif #ifdef WITH_VZ vshPrint(ctl, " Virtuozzo"); #endif -- 2.31.1

2 1

Arbitrary truncating of SCSI disk serial number
by Peter Krempa 03 Jun '21

03 Jun '21

Hi, recently I've got a report that an upgrade of libvirt (and qemu) caused a guest-visible change in the SCSI disk identification when a very long serial number is used. I've traced it back to the point where libvirt started to use the 'device_id=' property of the SCSI disk to pass in the alias of the disk when the serial is not configured and the serial if it is. https://gitlab.com/libvirt/libvirt/-/commit/a1dce96236f6d35167924fa7e6a70f5… The change is caused by the fact that when serial is configured via the 'serial=' property it's being silently truncated. Now there are two distinct VPD pages which report the serial number: 0x83 - device identification This one used to report only the device alias in the beginning but starting from qemu commit: commit fd9307912d0a2ffa0310f9e20935d96d5af0a1ca Author: Paolo Bonzini <pbonzini(a)redhat.com> Date: Fri Mar 16 19:12:43 2012 +0100 scsi: copy serial number into VPD page 0x83 Currently QEMU passes the qdev device id to the guest in an ASCII-string designator in page 0x83. While this is fine, it does not match what real hardware does; usually the ASCII-string designator there hosts another copy of the serial number (there can be other designators, for example with a world-wide name). Do the same for QEMU SCSI disks. ATAPI does not support VPD pages, so it does not matter there. Signed-off-by: Paolo Bonzini <pbonzini(a)redhat.com> it reports the serial number instead of the device alias when the serial is configured. Now this historically copied the IDE(?) limit of 20 characters. Now with the change to use 'device_id' which overrides the behavior the length of the reported value is limited to the technical limit of 255-8 which creates the problem. Libvirt uses 'device_id' because when -blockdev is used the disk alias which was configured via the -drive is no longer configured and thus would be missing. Libvirt also (unfortunately in this case I'd say) started to pass the serial number via this property. 0x80 - device serial (optional, only when serial is configured) This one started similarly to the 0x83 page to report the serial truncated to 20, but later in commit: commit 48b6206305b8d56524ac2ee347b68e6e0a528559 Author: Rony Weng <ronyweng(a)synology.com> Date: Mon Aug 29 15:52:18 2016 +0800 scsi-disk: change disk serial length from 20 to 36 Openstack Cinder assigns volume a 36 characters uuid as serial. QEMU will shrinks the uuid to 20 characters, which does not match the original uuid. Note that there is no limit to the length of the serial number in the SCSI spec. 20 was copy-pasted from virtio-blk which in turn was copy-pasted from ATA; 36 is even more arbitrary. However, bumping it up too much might cause issues (e.g. 252 seems to make sense because then the maximum amount of returned data is 256; but who knows there's no off-by-one somewhere for such a nicely rounded number). Signed-off-by: Rony Weng <ronyweng(a)synology.com> Message-Id: <1472457138-23386-1-git-send-email-ronyweng(a)synology.com> Cc: qemu-stable(a)nongnu.org Signed-off-by: Paolo Bonzini <pbonzini(a)redhat.com> qemu actually changed the silent truncation to another arbitrary value of 36, which is the length of the UUID. Thus qemu isn't inocent either in these regards. Now based on the fact that the above mentioned libvirt commit is contained in libvirt-5.1 (and qemu-4.0 adds support for 'device_id') reverting to truncation to 20 characters would IMO also be considerable as regression, based on the fact that there are users who changed qemu to lessen the truncation. As of such I don't think libvirt should revert to using the trucated serial despite an ABI change. On the other hand QEMU should IMO: 1) unify the truncation to a single length; preferrably the technical limit 2) add possibility to report error when the serial is too long (libvirt can accept a new property for example) I'm open to other suggestions though. Thanks. Peter

2 1

[libvirt PATCH 0/2] Add missing audio timer period setting
by Daniel P. Berrangé 03 Jun '21

03 Jun '21

Somehow I missed this one property during initial impl of audio Daniel P. Berrangé (2): conf: add support for timer period audio setting qemu: wire up support for timer period audio setting docs/formatdomain.rst | 6 +++++- docs/schemas/domaincommon.rng | 5 +++++ src/conf/domain_conf.c | 7 +++++++ src/conf/domain_conf.h | 2 ++ src/qemu/qemu_command.c | 8 ++++++++ tests/qemuxml2argvdata/audio-alsa-best.args | 1 + tests/qemuxml2argvdata/audio-alsa-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-alsa-best.xml | 2 +- tests/qemuxml2argvdata/audio-coreaudio-best.args | 1 + .../audio-coreaudio-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-coreaudio-best.xml | 2 +- tests/qemuxml2argvdata/audio-file-best.args | 1 + tests/qemuxml2argvdata/audio-file-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-file-best.xml | 2 +- tests/qemuxml2argvdata/audio-none-best.args | 1 + tests/qemuxml2argvdata/audio-none-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-none-best.xml | 2 +- tests/qemuxml2argvdata/audio-oss-best.args | 1 + tests/qemuxml2argvdata/audio-oss-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-oss-best.xml | 2 +- tests/qemuxml2argvdata/audio-pulseaudio-best.args | 1 + .../audio-pulseaudio-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-pulseaudio-best.xml | 2 +- tests/qemuxml2argvdata/audio-sdl-best.args | 1 + tests/qemuxml2argvdata/audio-sdl-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-sdl-best.xml | 2 +- tests/qemuxml2argvdata/audio-spice-best.args | 1 + .../qemuxml2argvdata/audio-spice-best.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/audio-spice-best.xml | 2 +- 29 files changed, 51 insertions(+), 17 deletions(-) -- 2.31.1

2 3

Storing PCI(e) VPD Board Serial Numbers
by Dmitrii Shcherbakov 03 Jun '21

03 Jun '21

Hello Libvirt Developers, I am looking for some feedback on a planned enhancement to Libvirt: the aim is to store a portion of PCI(e) Vital Product Data (VPD) for each device along with other PCI/PCIe device information already collected. Specifically, the SN (Serial Number) read-only field of a VPD data structure of a device is of interest which is described in PCI/PCIe specs (PCI local bus 2.1+ and PCIe 4.0+). The context for this is the cross-project work in OpenStack (Nova, Neutron), OVS and OVN to support for off-path SmartNIC DPUs ([1], [2], [3], [4]). The Nova specification [1] provides an overview of the relevant hardware and the use-case for board serial numbers, however, VPD is the standard capability in the PCI/PCIe specifications not tied to the use-case in particular so the suggestion from the Nova core team was to aim at introducing means of collecting this information via Libvirt. It can then be retrieved by the respective virt driver in Nova via Libvirt without having to introduce this code into Nova itself. Quoting the PCI(e) specs: * "Vital Product Data (VPD) is the information that uniquely defines items such as the hardware, software, and microcode elements of a system."; * "Vital Product Data is made up of Small and Large Resource Data Types."; * "Large resource type VPD-R Tag: This tag contains the read only VPD keywords for an add-in card." * SN read-only field: "The characters are alphanumeric and represent the unique add-in card Serial Number." The VPD capability is optional per the specification so it may or may not appear for PCI(e) endpoints. The devices of interest (SmartNIC DPUs), however, generally have it exposed. The PCI/PCIe specs define a binary format for VPD and a sysfs entry exposing a binary blob in that format has been available since kernel v2.6.26 [5]. The relevant sections of specs are: * "6.4. Vital Product Data" in the PCI Local Bus specification; * "6.28 Vital Product Data (VPD)" in the PCIe 4.0 Base Specification. Note that the serial number stored in VPD is not identical to the information stored in the Device Serial Number (DSN) capability also present in the specs as it may identify a component on a board which presents a multi-function device but the board itself may have multiple components ([9] also makes a distinction between a board serial and a device serial). As a reference, there is some code to parse and print the VPD in lspci [6] and there is a prototype along those lines in Python [7], a polished version of which I plan to use in Nova until the relevant functionality appears in Libvirt. Likewise, the devlink kernel infrastructure, which is already used in Libvirt to query additional device capabilities [8] (e.g. the presence of an eswitch and its switchdev mode) has a devlink-info API [9] that exposes a way to query a board serial number if a device driver exposes it (in turn, by querying controller firmware or via PCIe VPD). This allows doing that in a bus-independent manner (e.g. it would work for PCIe, platform devices or other I/O interconnects) but in the context of devices that implement devlink API only (which are not necessarily network devices [10] but most of them currently are). I would like to suggest the following to be done in Libvirt: 1) adding the code for extracting a serial number from VPD for PCI/PCIe devices in general and storing it for exposure via the Libvirt API; More specifically, I propose adding a nested capability called "vpd" under VIR_NODE_DEV_CAP_PCI_DEV: <capability type='pci'> <capability type='vpd'> <serial>UNIQUESERIAL</serial>  </capability>  </capability> 2) (optional) implementing functionality to obtain a board serial number via devlink-info for PCIe devices if they do not expose a VPD capability but the device driver can retrieve it via firmware. The board serial number can be stored in the same element as suggested above. Not all devices expose the devlink API and even fewer do expose board serial via devlink-info: * devlink was added in 4.10 [11]; * devlink-info was introduced in 5.1 [12]; * querying for board.serial_number was added in kernel 5.9 [13] and iproute2 5.9.0 [14]; * Besides the generic devlink infrastructure support above, device drivers also need to support exposing this field. Therefore, implementing two approaches (sysfs VPD, devlink) is preferable for better compatibility. I would appreciate any feedback on whether this potential addition makes sense. If so, I can look into implementing this. [1] https://review.opendev.org/c/openstack/nova-specs/+/787458 [2] https://review.opendev.org/c/openstack/neutron-specs/+/788821 [3] https://patchwork.ozlabs.org/project/openvswitch/patch/20210323145032.45312… [4] https://patchwork.ozlabs.org/project/ovn/patch/20210509140305.1910796-1-fro… [5] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… [6] https://github.com/pciutils/pciutils/blob/v3.7.0/ls-vpd.c#L95-L216 [7] https://gist.github.com/dshcherb/40e982989599a757e5b1e25999501019 [8] https://github.com/libvirt/libvirt/blob/v7.3.0/src/util/virnetdev.c#L3167-L… [9] https://www.kernel.org/doc/html/latest/networking/devlink/devlink-info.html [10] https://www.kernel.org/doc/html/latest/networking/devlink/index.html [11] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… [12] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… [13] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… [14] https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/commit/?id=733… Best Regards, Dmitrii Shcherbakov LP: ~dmitriis

2 2

[PATCH] tests: Update expected output of virsh-optparse
by Michal Privoznik 02 Jun '21

02 Jun '21

In the previous commit I've changed what API is called from 'virsh setmem' command. However, since virsh-optparse test is ran only when expensive tests are enabled I've completely missed that the expected output for virsh-optparse test must be updated too as it contains the API. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Pushed as a trivial build breaker fix. tests/virsh-optparse | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/virsh-optparse b/tests/virsh-optparse index fed71a8509..3fe38b2931 100755 --- a/tests/virsh-optparse +++ b/tests/virsh-optparse @@ -215,7 +215,7 @@ compare exp-err err || fail=1 # it has been parsed correctly cat <<\EOF > exp-err || framework_failure error: Unable to change MaxMemorySize -error: memory in virDomainSetMaxMemory must not be zero +error: memory in virDomainSetMemoryFlags must not be zero EOF $VIRSH -q -c $test_url setmaxmem test 0 >out 2>err && fail=1 test -s out && fail=1 -- 2.31.1

1 0

[PATCH 0/5] Support for launchSecurity type papr-pef
by Daniel Henrique Barboza 02 Jun '21

02 Jun '21

Hi, This is the implementation of ppc64 secure guest execution, known as PAPR-PEF. This implementation is based on the new confidential-guest-support machine option that is available in QEMU 6.0.0. The patches were done on top of Boris s390-pv patches that were sent by review [1] given that Boris provided a common base for new security launches there. The series is also available in gitlab: https://gitlab.com/danielhb/libvirt/-/tree/papr_pef_v1 Note: If someone wishes to give this series a try, considering that you have a host with proper PowerPC PEF hardware, be aware that there is a QEMU bug in which launching a PAPR-PEF guest will trigger an early assert [2]. The fix is not merged upstream at the moment of this posting, so you'll need to apply the patch yourself or build QEMU using David Gibson's ppc-for-6.1 tree [3]. [1] https://listman.redhat.com/archives/libvir-list/2021-May/msg00570.html [2] https://lists.gnu.org/archive/html/qemu-devel/2021-05/msg08491.html [3] https://gitlab.com/dgibson/qemu/-/tree/ppc-for-6.1 Daniel Henrique Barboza (5): tests: Add QEMU and domain ppc64 capapbilities for qemu 6.1 qemu: add 'pef-guest' capability conf, qemu: add 'papr-pef' launch security type tests: add tests for 'papr-pef' support docs/formatdomain.rst: add 'papr-pef' launchSecurity docs docs/formatdomain.rst | 4 + docs/schemas/domaincommon.rng | 1 + src/conf/domain_conf.c | 3 + src/conf/domain_conf.h | 1 + src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 26 + src/qemu/qemu_namespace.c | 1 + src/qemu/qemu_process.c | 1 + src/qemu/qemu_validate.c | 8 + tests/domaincapsdata/qemu_6.1.0.ppc64.xml | 133 + .../launch-security-papr-pef.xml | 22 + tests/genericxml2xmltest.c | 1 + .../caps_6.1.0.ppc64.replies | 30368 ++++++++++++++++ .../qemucapabilitiesdata/caps_6.1.0.ppc64.xml | 1175 + ...default-video-type-ppc64.ppc64-latest.args | 5 +- ...y-papr-pef-ignore-policy.ppc64-latest.args | 30 + ...launch-security-papr-pef-ignore-policy.xml | 19 + ...launch-security-papr-pef.ppc64-latest.args | 30 + .../launch-security-papr-pef.xml | 16 + ...ault-cpu-kvm-pseries-2.7.ppc64-latest.args | 5 +- ...ault-cpu-kvm-pseries-3.1.ppc64-latest.args | 5 +- ...ault-cpu-kvm-pseries-4.2.ppc64-latest.args | 5 +- ...ault-cpu-tcg-pseries-2.7.ppc64-latest.args | 5 +- ...ault-cpu-tcg-pseries-3.1.ppc64-latest.args | 5 +- ...ault-cpu-tcg-pseries-4.2.ppc64-latest.args | 5 +- .../ppc64-pseries-graphics.ppc64-latest.args | 7 +- .../ppc64-pseries-headless.ppc64-latest.args | 7 +- .../ppc64-tpmproxy-single.ppc64-latest.args | 5 +- .../ppc64-tpmproxy-with-tpm.ppc64-latest.args | 5 +- .../tpm-emulator-spapr.ppc64-latest.args | 5 +- tests/qemuxml2argvtest.c | 3 + 32 files changed, 31871 insertions(+), 38 deletions(-) create mode 100644 tests/domaincapsdata/qemu_6.1.0.ppc64.xml create mode 100644 tests/genericxml2xmlindata/launch-security-papr-pef.xml create mode 100644 tests/qemucapabilitiesdata/caps_6.1.0.ppc64.replies create mode 100644 tests/qemucapabilitiesdata/caps_6.1.0.ppc64.xml create mode 100644 tests/qemuxml2argvdata/launch-security-papr-pef-ignore-policy.ppc64-latest.args create mode 100644 tests/qemuxml2argvdata/launch-security-papr-pef-ignore-policy.xml create mode 100644 tests/qemuxml2argvdata/launch-security-papr-pef.ppc64-latest.args create mode 100644 tests/qemuxml2argvdata/launch-security-papr-pef.xml -- 2.31.1

1 5

[PATCH v2 0/3] virsh: Fix logic wrt to --current flag in cmdSetmem
by Michal Privoznik 02 Jun '21

02 Jun '21

v2 of: https://listman.redhat.com/archives/libvir-list/2021-May/msg00484.html diff to v1: - Work in Jano's and Peter's review suggestions in 1/3. - Patches 2/3 and 3/3 are new. I've done 2/3 before 3/3 because it's still worth merging even if 3/3 is disliked. Michal Prívozník (3): virsh: Fix logic wrt to --current flag in cmdSetmem virsh-domain: Fix @ret handling in cmdSetmem and cmdSetmaxmem virsh-domain: Drop support for old APIs in cmdSetmem and cmdSetmaxmem tools/virsh-domain.c | 48 +++++++++++++++++--------------------------- 1 file changed, 18 insertions(+), 30 deletions(-) -- 2.26.3

3 7

Recommended volume permissions (being created for vagrant-libvirt via fog-libvirt)
by Darragh Bailey 02 Jun '21

02 Jun '21

Hi, A request has come up recently in vagrant-libvirt about changing the permissions used for the VM volume image file. Currently there is a backing image file uploaded that gets 744 as the file permissions, and then the VM domain is created using this as the backing file for any changes. The file containing the changes for the VM gets 600, so accessing what is contained is limited to libvirt and thus to those that can connect to libvirt. The request is to change this to be 744, it appears to have been triggered due to a desire to try and use virt-v2v to create a portable XML and export the disks. However I'm a little hesitant as in general I would default to more secure rather than less secure to avoid creating security concerns down the line. Even though vagrant-libvirt is typically used for development, it wouldn't surprise me to see it being used on CI build infrastructure and given the shared nature of that, making things less secure may cause issues for some users. Of course working out who would be impacted is virtually impossible without making the change and seeing who is concerned. And that might be several months down the line before it's raised. Rather than just merging this, wondering if there are any security guidelines on the file permissions for VM image files? That or something that can outline the risks, or even clarify that it's unnecessary to worry about? -- Darragh Bailey "Nothing is foolproof to a sufficiently talented fool"

2 3

[PATCH 0/4] node_device_udev: Also process ID_TYPE=cd/dvd in udevProcessStorage()
by Michal Privoznik 02 Jun '21

02 Jun '21

See the last one which fixes a real bug. The rest is just cleanups. Michal Prívozník (4): node_device_udev: Make udevGenerateDeviceName() return void node_device_udev: Make udevGetStringProperty() return void node_device_udev: Don't overwrite @ret in udevProcessStorage() node_device_udev: Also process ID_TYPE=cd/dvd in udevProcessStorage() src/node_device/node_device_udev.c | 114 +++++++++++------------------ 1 file changed, 43 insertions(+), 71 deletions(-) -- 2.31.1

2 7

[libvirt PATCH v2] meson: Don't include warning flags in summary
by Andrea Bolognani 01 Jun '21

01 Jun '21

The list takes up half a screen on my machine and it doesn't seem useful enough for that to be warranted, especially when it results in other, arguably more relevant parts of the summary being pushed off-screen. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> Reviewed-by: Pavel Hrdina <phrdina(a)redhat.com> --- Changes from [v1]: * move out of the series, which has mostly been pushed now. [v1] https://listman.redhat.com/archives/libvir-list/2021-June/msg00002.html meson.build | 1 - 1 file changed, 1 deletion(-) diff --git a/meson.build b/meson.build index 2fa379f47a..b09d766b11 100644 --- a/meson.build +++ b/meson.build @@ -2313,7 +2313,6 @@ else loader_res = '' endif misc_summary = { - 'Warning Flags': supported_cc_flags, 'docs': gen_docs, 'tests': build_tests, 'DTrace': conf.has('WITH_DTRACE_PROBES'), -- 2.31.1

1 0

[libvirt PATCH 0/2] fix disk XML formatting for qemuDomainBlockCopy API
by Pavel Hrdina 01 Jun '21

01 Jun '21

Pavel Hrdina (2): domain_conf: extract disk driver source bits to its own function virDomainDiskDefParseSource: parse source bits from driver element src/conf/domain_conf.c | 52 ++++++++++++++++++++++++++++++------------ 1 file changed, 37 insertions(+), 15 deletions(-) -- 2.31.1

2 3

[libvirt PATCH 00/26] meson: Various fixes and improvements
by Andrea Bolognani 01 Jun '21

01 Jun '21

Test pipeline: https://gitlab.com/abologna/libvirt/-/pipelines/312508056 Andrea Bolognani (26): spec: Be explicit about more features meson: Whitespace tweaks meson: Don't use 'required: true' meson: Drop curl workaround meson: Drop netinet workaround meson: Use get_pkgconfig_variable('cflags') meson: Use built-in pcap detection meson: Make libm a required dependency meson: Drop numactl_version meson: Fix vstorage detection meson: Fix sanlock detection meson: Rewrite polkit check meson: Tweak XDR check meson: Fix disabling netcf meson: Rewrite firewalld check meson: Fix firewalld check meson: Rewrite libacl check meson: Use dependency() when possible meson: Rewrite apparmor_profiles check meson: Turn apparmor_profiles into a feature meson: Switch to autodetection for apparmor_profiles meson: Switch to autodetection for driver_remote meson: Switch to autodetection for driver_test meson: Style tweaks meson: Reorganize summary meson: Don't include warning flags in summary libvirt.spec.in | 3 + meson.build | 301 +++++++++++++++++++------------------ meson_options.txt | 6 +- mingw-libvirt.spec.in | 2 + src/locking/meson.build | 4 +- src/qemu/qemu_namespace.c | 2 +- src/util/virfile.c | 8 +- src/util/virnetdevbridge.c | 18 --- 8 files changed, 166 insertions(+), 178 deletions(-) -- 2.31.1

2 32

Release of libvirt-7.4.0
by Jiri Denemark 01 Jun '21

01 Jun '21

The 7.4.0 release of both libvirt and libvirt-python is tagged and signed tarballs and source RPMs are available at https://libvirt.org/sources/ https://libvirt.org/sources/python/ Thanks everybody who helped with this release by sending patches, reviewing, testing, or providing any other feedback. Your work is greatly appreciated. * Removed features * qemu: Remove support for QEMU < 2.11 In accordance with our platform support policy, the oldest supported QEMU version is now bumped from 1.5 to 2.11. * New features * qemu: Add support for hotplugging ``<transient/>`` disks The disk hotplug code in the qemu driver now can handle hotplug of disks with automatically added overlay. * qemu: Add support for sharing base image of ``<transient/>`` disks Users can use ``<transient shareBacking='yes'/>`` to tell the qemu driver to never open the base image in write mode thus multiple VMs can share the same image. Note that the disk will be hotplugged during startup. * Improvements * Add win-dmp crashdump format New ``win-dmp`` format for ``virDomainCoreDumpWithFormat`` API and/or virsh ``dump --format`` was introduced. * Bug fixes * Allow 0 offset in XML schema for ``<slice type='storage' offset='0' size='321'/>`` Having a 0 offset so that the size of the image can be limited is a valid configuration so it was allowed in the XML schema. Enjoy. Jirka

1 0

[libvirt PATCH] docs: introduce stubs for new libvirt Go packages
by Daniel P. Berrangé 01 Jun '21

01 Jun '21

Currently we expose libvirt Go packages at libvirt.org/libvirt-go libvirt.org/libvirt-go-xml These packages have not supported Go modules historically and when we tried to introduce modules, we hit the problem that we're not using semver for versioning. The only way around this is to introduce new packages under a different namespace, that will have the exact same code, but be tagged with a different version numbering scheme. This change proposes: libvirt.org/go/libvirt libvirt.org/go/libvirtxml Note the hyphen is removed so that the import basename matches the Go package name. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- The corresponding repos exist https://gitlab.com/libvirt/libvirt-go-module https://gitlab.com/libvirt/libvirt-go-xml-module but I'm leaving them empty until we've completed this release, then I'll populate with content matching 7.4.0, minus the existing tags. This change should also wait until after the release. docs/go/libvirt.rst | 13 +++++++++++ docs/go/libvirtxml.rst | 10 ++++++++ docs/go/meson.build | 53 ++++++++++++++++++++++++++++++++++++++++++ docs/meson.build | 1 + 4 files changed, 77 insertions(+) create mode 100644 docs/go/libvirt.rst create mode 100644 docs/go/libvirtxml.rst create mode 100644 docs/go/meson.build diff --git a/docs/go/libvirt.rst b/docs/go/libvirt.rst new file mode 100644 index 0000000000..46250691c1 --- /dev/null +++ b/docs/go/libvirt.rst @@ -0,0 +1,13 @@ +======================= +Libvirt Go Language API +======================= + +The `Go <https://golang.org/>`__ package ``libvirt.org/go/libvirt`` provides +`CGo <https://golang.org/cmd/cgo/>`__ binding from the OS native Libvirt API. + +In general the Go representation is a direct 1-1 mapping from native API +concepts to Go, so the native API documentation should serve as a reference +for most behaviour. + +For details of Go specific behaviour consult the +`Go package documentation <https://pkg.go.dev/libvirt.org/go/libvirt>`__. diff --git a/docs/go/libvirtxml.rst b/docs/go/libvirtxml.rst new file mode 100644 index 0000000000..7583433f56 --- /dev/null +++ b/docs/go/libvirtxml.rst @@ -0,0 +1,10 @@ +========================== +Libvirt Go XML parsing API +========================== + +The `Go <https://golang.org/>`__ package ``libvirt.org/go/libvirtxml`` provides +annotated Go struct definitions for parsing (and formatting) XML documents used +with libvirt APIs. + +For details of Go specific behaviour consult the +`Go package documentation <https://pkg.go.dev/libvirt.org/go/libvirtxml>`__. diff --git a/docs/go/meson.build b/docs/go/meson.build new file mode 100644 index 0000000000..99fa1b074a --- /dev/null +++ b/docs/go/meson.build @@ -0,0 +1,53 @@ +docs_go_files = [ + 'libvirt', + 'libvirtxml', +] + +html_xslt_gen_xslt = subsite_xsl +html_xslt_gen_install_dir = docs_html_dir / 'go' +html_xslt_gen = [] + +foreach name : docs_go_files + rst_file = '@0@.rst'.format(name) + + html_xslt_gen += { + 'name': name, + 'file': docs_rst2html_gen.process(rst_file), + 'source': 'docs/go' / rst_file, + } +endforeach + +# keep the XSLT processing code block in sync with docs/meson.build + +# --- begin of XSLT processing --- + +foreach data : html_xslt_gen + html_filename = data['name'] + '.html' + + html_file = custom_target( + html_filename, + input: data.get('file', data['name'] + '.html.in'), + output: html_filename, + command: [ + xsltproc_prog, + '--stringparam', 'pagesrc', data.get('source', ''), + '--stringparam', 'builddir', meson.build_root(), + '--stringparam', 'timestamp', docs_timestamp, + '--nonet', + html_xslt_gen_xslt, + '@INPUT@', + ], + depends: data.get('depends', []), + depend_files: [ page_xsl ], + capture: true, + install: true, + install_dir: html_xslt_gen_install_dir, + ) + + install_web_deps += html_file + install_web_files += html_file.full_path() + ':' + html_xslt_gen_install_dir +endforeach + +html_xslt_gen = [] + +# --- end of XSLT processing --- diff --git a/docs/meson.build b/docs/meson.build index f550629d8e..d7afdbd323 100644 --- a/docs/meson.build +++ b/docs/meson.build @@ -296,6 +296,7 @@ html_xslt_gen = [] # --- end of XSLT processing --- subdir('fonts') +subdir('go') subdir('html') subdir('internals') subdir('js') -- 2.31.1

2 1

[PATCH] NEWS: disks: Mention <transient> improvements and <slice> XML fix
by Peter Krempa 01 Jun '21

01 Jun '21

Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- NEWS.rst | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/NEWS.rst b/NEWS.rst index 8358cbd369..66ce95d7cb 100644 --- a/NEWS.rst +++ b/NEWS.rst @@ -20,6 +20,17 @@ v7.4.0 (unreleased) * **New features** + * qemu: Add support for hotplugging ``<transient/>`` disks + + The disk hotplug code in the qemu driver now can handle hotplug of disks + with automatically added overlay. + + * qemu: Add support for sharing base image of ``<transient/>`` disks + + Users can use ``<transient shareBacking='yes'/>`` to tell the qemu driver to + never open the base image in write mode thus multiple VMs can share the + same image. Note that the disk will be hotplugged during startup. + * **Improvements** * Add win-dmp crashdump format @@ -29,6 +40,10 @@ v7.4.0 (unreleased) * **Bug fixes** + * Allow 0 offset in XML schema for ``<slice type='storage' offset='0' size='321'/>`` + + Having a 0 offset so that the size of the image can be limited is a + valid configuration so it was allowd in the XML schema. v7.3.0 (2021-05-03) =================== -- 2.31.1

2 1

[PATCH] qemuProcessSetupDisksTransientSnapshot: Skip enabling transientOverlayCreated flag
by Masayoshi Mizuma 01 Jun '21

01 Jun '21

From: Masayoshi Mizuma <m.mizuma(a)jp.fujitsu.com> QEMU_DOMAIN_DISK_PRIVATE(disk)->transientOverlayCreated flag gets true unexpectedly on qemuProcessSetupDisksTransientSnapshot() when the disk has <transient shareBacking='yes'> option. The flag should be enabled on qemuDomainAttachDiskGeneric() after the overlay setup is completed. Skip enabling transientOverlayCreated for the disk here. Fixes: 75871da0ecb8b552f9e304d0f83e216839bbf82d Signed-off-by: Masayoshi Mizuma <m.mizuma(a)jp.fujitsu.com> --- src/qemu/qemu_process.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 9b069fe7ce..c37687f249 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -7039,7 +7039,8 @@ qemuProcessSetupDisksTransientSnapshot(virDomainObj *vm, for (i = 0; i < vm->def->ndisks; i++) { virDomainDiskDef *domdisk = vm->def->disks[i]; - if (!domdisk->transient) + if (!domdisk->transient || + domdisk->transientShareBacking == VIR_TRISTATE_BOOL_YES) continue; QEMU_DOMAIN_DISK_PRIVATE(domdisk)->transientOverlayCreated = true; -- 2.27.0

2 1

[libvirt PATCH 00/10] remote: various refactoring & bugs for remote driver
by Daniel P. Berrangé 27 May '21

27 May '21

This is in prep for switching to use the modular daemons by default. In testing that setup I hit a couple of minor bugs. Daniel P. Berrangé (10): remote: use absolute path to check for daemons remote: stop checking for errors from socket allocations remote: move open flags enum to sockets header remote: consistently use flags for passing ro/user/autostart props remote: change socket helper to return full daemon path remote: don't populate daemon path if autostart is not required rpc: remove "spawnDaemon" parameter remote: move proxy/mode defaults after URI parsing remote: fix regression connecting to remote session daemon remote: use virtproxyd if no URI is given src/admin/admin_remote.c | 2 +- src/locking/lock_driver_lockd.c | 1 - src/logging/log_manager.c | 1 - src/lxc/lxc_monitor.c | 2 +- src/qemu/qemu_migration.c | 3 +- src/remote/remote_daemon_dispatch.c | 4 +- src/remote/remote_driver.c | 135 ++++++++++++---------------- src/remote/remote_sockets.c | 95 +++++++++++++------- src/remote/remote_sockets.h | 15 ++-- src/remote/remote_ssh_helper.c | 16 ++-- src/rpc/virnetclient.c | 5 +- src/rpc/virnetclient.h | 3 +- src/rpc/virnetsocket.c | 26 ++---- src/rpc/virnetsocket.h | 3 +- tests/virnetsockettest.c | 4 +- 15 files changed, 157 insertions(+), 158 deletions(-) -- 2.31.1

3 15

Entering freeze for libvirt-7.4.0
by Jiri Denemark 27 May '21

27 May '21

I have just tagged v7.4.0-rc1 in the repository and pushed signed tarballs and source RPMs to https://libvirt.org/sources/ 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

2 3

[libvirt PATCH 0/2] Revert mdev autostart support
by Jonathon Jongsma 27 May '21

27 May '21

As Daniel Berrange mentioned, the method of supporting autostart for node devices was not consistent with other libvirt objects. This patch series reverts this feature before it gets into a release. I will submit a follow-up patch soon which adds the feature back with a consistent API. Jonathon Jongsma (2): Partial Revert of "tests: nodedevxml2xmltest: test more mdev files" nodedev: Revert auto-start property for mdevs docs/formatnode.html.in | 10 --------- docs/schemas/nodedev.rng | 11 ---------- src/conf/node_device_conf.c | 21 +------------------ src/conf/node_device_conf.h | 12 ----------- src/libvirt_private.syms | 2 -- src/node_device/node_device_driver.c | 7 +------ .../mdevctl-list-multiple.out.xml | 4 ---- tests/nodedevschemadata/mdev_autostart.xml | 10 --------- tests/nodedevschemadata/mdev_manual.xml | 10 --------- ...v_3627463d_b7f0_4fea_b468_f1da537d301b.xml | 10 +-------- tests/nodedevxml2xmlout/mdev_autostart.xml | 11 ---------- ...v_d069d019_36ea_4111_8f0a_8c9a70e21366.xml | 1 - ...v_d2441d39_495e_4243_ad9f_beb3f14c23d9.xml | 1 - ...v_ee0b88c4_f554_4dc1_809d_b2a01e8e48ad.xml | 11 +--------- ...v_fedc4916_1ca8_49ac_b176_871d16c13076.xml | 1 - tests/nodedevxml2xmlout/mdev_manual.xml | 11 ---------- tests/nodedevxml2xmltest.c | 2 -- 17 files changed, 4 insertions(+), 131 deletions(-) delete mode 100644 tests/nodedevschemadata/mdev_autostart.xml delete mode 100644 tests/nodedevschemadata/mdev_manual.xml mode change 100644 => 120000 tests/nodedevxml2xmlout/mdev_3627463d_b7f0_4fea_b468_f1da537d301b.xml delete mode 100644 tests/nodedevxml2xmlout/mdev_autostart.xml mode change 100644 => 120000 tests/nodedevxml2xmlout/mdev_ee0b88c4_f554_4dc1_809d_b2a01e8e48ad.xml delete mode 100644 tests/nodedevxml2xmlout/mdev_manual.xml -- 2.31.1

3 8

[PATCH for 7.4.0] virCapabilitiesHostNUMAInitReal: Don't jump over cleanup
by Michal Privoznik 27 May '21

27 May '21

In one of my recent commits I've done some renaming. But whilst doing so I also mistakenly replaced 'goto cleanup' with 'return -1' in virCapabilitiesHostNUMAInitReal() which was incorrect. Fixes: fe25224fdaa53bbeceed3ddeef1b3a150665e656 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/conf/capabilities.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/conf/capabilities.c b/src/conf/capabilities.c index 1290c9c15d..2f9a1e7d1f 100644 --- a/src/conf/capabilities.c +++ b/src/conf/capabilities.c @@ -1623,7 +1623,7 @@ virCapabilitiesHostNUMAInitReal(virCapsHostNUMA *caps) } if (virCapabilitiesGetNUMADistances(n, &distances, &ndistances) < 0) - return -1; + goto cleanup; if (virCapabilitiesGetNUMAPagesInfo(n, &pageinfo, &npageinfo) < 0) goto cleanup; -- 2.31.1

2 1

[PATCH] qemuDomainAttachDiskGenericTransient: Add NULL check in case the overlay disk already exists
by Masayoshi Mizuma 27 May '21

27 May '21

From: Masayoshi Mizuma <m.mizuma(a)jp.fujitsu.com> When <transient shareBacking='yes'> is set to a disk and the overlay disk already exists because of something abnormal, libvirt is terminated by Segmentation fault. # virsh start Test0 error: Disconnected from qemu:///system due to end of file error: Failed to start domain 'Test0' error: End of file while reading data: Input/output error Add NULL check for snapdiskdef so that the rollback can work correctly. Signed-off-by: Masayoshi Mizuma <m.mizuma(a)jp.fujitsu.com> --- src/qemu/qemu_hotplug.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 9ff6175dd3..0640cdd9f7 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -695,7 +695,9 @@ qemuDomainAttachDiskGenericTransient(virDomainObj *vm, g_autoptr(qemuSnapshotDiskContext) snapctxt = NULL; g_autoptr(virDomainSnapshotDiskDef) snapdiskdef = NULL; - snapdiskdef = qemuSnapshotGetTransientDiskDef(disk, vm->def->name); + if (!(snapdiskdef = qemuSnapshotGetTransientDiskDef(disk, vm->def->name))) + return NULL; + snapctxt = qemuSnapshotDiskContextNew(1, vm, asyncJob); if (qemuSnapshotDiskPrepareOne(snapctxt, disk, snapdiskdef, -- 2.27.0

2 1

[PATCH V5] libxl: adjust handling of libxl_device_nic objects
by Jim Fehlig 26 May '21

26 May '21

libxl objects are supposed to be initialized and disposed. Adjust libxlMakeNic to use an already initialized object owned by the caller. Adjust libxlMakeNicList to initialize the list of objects, before they are filled by libxlMakeNic. The libxl_domain_config object passed to libxlMakeNicList is owned by the caller and will be disposed with libxl_domain_config_dispose, which also disposes embedded objects such as libxl_device_nic. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/libxl/libxl_conf.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 91cebf1498..fc4268db01 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -1313,8 +1313,6 @@ libxlMakeNic(virDomainDef *def, return -1; } - libxl_device_nic_init(x_nic); - virMacAddrGetRaw(&l_nic->mac, x_nic->mac); /* @@ -1502,6 +1500,7 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) size_t nnics = def->nnets; libxl_device_nic *x_nics; size_t i, nvnics = 0; + int ret = -1; x_nics = g_new0(libxl_device_nic, nnics); @@ -1509,8 +1508,9 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) if (virDomainNetGetActualType(l_nics[i]) == VIR_DOMAIN_NET_TYPE_HOSTDEV) continue; + libxl_device_nic_init(&x_nics[nvnics]); if (libxlMakeNic(def, l_nics[i], &x_nics[nvnics], false)) - goto error; + goto out; /* * The devid (at least right now) will not get initialized by * libxl in the setup case but is required for starting the @@ -1521,18 +1521,14 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) nvnics++; } + ret = 0; + out: VIR_SHRINK_N(x_nics, nnics, nnics - nvnics); d_config->nics = x_nics; d_config->num_nics = nvnics; - return 0; - - error: - for (i = 0; i < nnics; i++) - libxl_device_nic_dispose(&x_nics[i]); - VIR_FREE(x_nics); - return -1; + return ret; } int -- 2.31.1

2 2

[PATCH v4] libxl: adjust handling of libxl_device_nic objects
by Olaf Hering 26 May '21

26 May '21

libxl objects are supposed to be initialized and disposed. Adjust libxlMakeNic to use an already initialized object, it is owned by the caller. Adjust libxlMakeNicList to initialize the list of objects, before they are filled by libxlMakeNic. In case of error the objects are disposed by libxl_domain_config_dispose. Signed-off-by: Olaf Hering <olaf(a)aepfle.de> --- src/libxl/libxl_conf.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 2ecbcf6911..c672bafbe9 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -1335,8 +1335,6 @@ libxlMakeNic(virDomainDef *def, return -1; } - libxl_device_nic_init(x_nic); - virMacAddrGetRaw(&l_nic->mac, x_nic->mac); /* @@ -1531,8 +1529,9 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) if (virDomainNetGetActualType(l_nics[i]) == VIR_DOMAIN_NET_TYPE_HOSTDEV) continue; + libxl_device_nic_init(&x_nics[nvnics]); if (libxlMakeNic(def, l_nics[i], &x_nics[nvnics], false)) - goto error; + return -1; /* * The devid (at least right now) will not get initialized by * libxl in the setup case but is required for starting the @@ -1549,12 +1548,6 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) d_config->num_nics = nvnics; return 0; - - error: - for (i = 0; i < nnics; i++) - libxl_device_nic_dispose(&x_nics[i]); - VIR_FREE(x_nics); - return -1; } int

2 2

[PATCH RESEND v3 0/8] cleanup meson checks for runtime binaries
by Pavel Hrdina 26 May '21

26 May '21

Recent attempt to add a lot of meson options to specify different runtime paths motivated me enough to cleanup this from meson. Changes in v3: - some patches were already pushed - removed patch that moved virFindFileInPath from testutilsqemu.c Changes in v2: - split and rework patch 16/17 to address review comments - added a new patch to cleanup libvirt.spec.in file Pavel Hrdina (8): virfile: introduce virFindFileInPathFull() qemu_conf: use virFindFileInPathFull for runtime binaries meson: drop check for runtime binary dependencies meson: move iscsiadm check into storage_iscsi condition meson: stop setting runtime binaries defines during compilation meson: use runtime binaries to only resolve features with "auto" value meson: optional_programs should be used only for building libvirt libvirt.spec: drop no longer required build dependencies libvirt.spec.in | 31 ---- meson.build | 214 +++++++++---------------- src/bhyve/bhyve_command.c | 4 + src/libvirt_private.syms | 2 +- src/locking/lock_driver_lockd.c | 12 +- src/network/bridge_driver.c | 2 + src/node_device/node_device_driver.c | 2 + src/qemu/qemu_conf.c | 23 ++- src/storage/storage_backend_logical.c | 13 ++ src/storage/storage_backend_sheepdog.c | 2 + src/storage/storage_backend_zfs.c | 3 + src/storage/storage_util.c | 2 + src/storage/storage_util.h | 6 + src/util/virdnsmasq.c | 1 + src/util/virfile.c | 16 +- src/util/virfile.h | 5 +- src/util/virfirewall.h | 4 + src/util/viriscsi.h | 2 + src/util/virkmod.h | 3 + src/util/virnetdevbandwidth.h | 2 + src/util/virnetdevip.c | 2 + src/util/virnetdevmidonet.c | 2 + src/util/virnetdevopenvswitch.c | 2 + src/util/virnuma.c | 1 + src/util/virsysinfo.c | 1 + src/util/virutil.c | 2 + tests/testutilsqemu.c | 3 +- tests/virfirewallmock.c | 3 +- 28 files changed, 173 insertions(+), 192 deletions(-) -- 2.31.1

4 21

[PATCH] meson.build: Fix the -Wvla-larger-than flag
by Thomas Huth 25 May '21

25 May '21

It's "...-than=..." and not "...-then=...". Fixes: 8dd259d0c6 ("meson: add manywarnings") Signed-off-by: Thomas Huth <thuth(a)redhat.com> --- meson.build | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 4f23f9104e..8ee0416700 100644 --- a/meson.build +++ b/meson.build @@ -390,7 +390,7 @@ cc_flags += [ '-Wvariadic-macros', '-Wvector-operation-performance', '-Wvla', - '-Wvla-larger-then=4031', + '-Wvla-larger-than=4031', '-Wvolatile-register-var', '-Wwrite-strings', ] -- 2.27.0

2 1

availability of cluster_size support (re. [libvirt PATCH v2 0/2] storage_driver: add support for cluster_size QCOW2 option)
by Geoff Macartney1 25 May '21

25 May '21

1 0

[PATCH v3] libxl: adjust handling of libxl_device_nic objects
by Olaf Hering 25 May '21

25 May '21

libxl objects are supposed to be initialized and disposed. Adjust libxlMakeNic to use an already initialized object, it is owned by the caller. Adjust libxlMakeNicList to initialize the list of objects, before they are filled by libxlMakeNic. In case of error the objects are disposed by libxl_domain_config_dispose. Signed-off-by: Olaf Hering <olaf(a)aepfle.de> --- src/libxl/libxl_conf.c | 34 +++++++++++++--------------------- 1 file changed, 13 insertions(+), 21 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index f29a28a841..8dc7e26cea 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -1339,8 +1339,6 @@ libxlMakeNic(virDomainDef *def, return -1; } - libxl_device_nic_init(x_nic); - virMacAddrGetRaw(&l_nic->mac, x_nic->mac); /* @@ -1526,39 +1524,33 @@ libxlMakeNicList(virDomainDef *def, libxl_domain_config *d_config) { virDomainNetDef **l_nics = def->nets; size_t nnics = def->nnets; - libxl_device_nic *x_nics; size_t i, nvnics = 0; - x_nics = g_new0(libxl_device_nic, nnics); - for (i = 0; i < nnics; i++) { if (virDomainNetGetActualType(l_nics[i]) == VIR_DOMAIN_NET_TYPE_HOSTDEV) continue; + nvnics++; + } + if (!nvnics) + return 0; - if (libxlMakeNic(def, l_nics[i], &x_nics[nvnics], false)) - goto error; + d_config->nics = g_new0(libxl_device_nic, nvnics); + d_config->num_nics = nvnics; + + for (i = 0; i < nvnics; i++) { + libxl_device_nic_init(&d_config->nics[i]); + if (libxlMakeNic(def, l_nics[i], &d_config->nics[i], false)) + return -1; /* * The devid (at least right now) will not get initialized by * libxl in the setup case but is required for starting the * device-model. */ - if (x_nics[nvnics].devid < 0) - x_nics[nvnics].devid = nvnics; - - nvnics++; + if (d_config->nics[i].devid < 0) + d_config->nics[i].devid = i; } - VIR_SHRINK_N(x_nics, nnics, nnics - nvnics); - d_config->nics = x_nics; - d_config->num_nics = nvnics; - return 0; - - error: - for (i = 0; i < nnics; i++) - libxl_device_nic_dispose(&x_nics[i]); - VIR_FREE(x_nics); - return -1; } int

2 2

[PATCH] spec: Fix %endif indentation
by Michal Privoznik 25 May '21

25 May '21

In recent commit f772c1fd2a a misaligned %endif sneaked in which upsets syntax-check. Align it properly. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Pushed as trivial and build breaker fix. libvirt.spec.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index cdca7d8db6..7920599498 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -73,7 +73,7 @@ %else %ifnarch %{arches_qemu_kvm} %define with_storage_gluster 0 - %endif + %endif %endif %endif -- 2.26.3

1 0

[PATCH for 7.4.0 v2 0/3] Avoid double indentation of <metadata/> element
by Michal Privoznik 25 May '21

25 May '21

v2 of: https://listman.redhat.com/archives/libvir-list/2021-May/msg00767.html diff to v1: - Deduplicated the code per Pavel's suggestion Michal Prívozník (3): virxml: Introduce and use virXMLFormatMetadata() virxml: Report error if virXMLFormatMetadata() fails virxml: Avoid double indentation of <metadata/> element src/conf/domain_conf.c | 23 ++---------------- src/conf/network_conf.c | 23 ++---------------- src/libvirt_private.syms | 1 + src/util/virxml.c | 51 ++++++++++++++++++++++++++++++++++++++++ src/util/virxml.h | 3 +++ 5 files changed, 59 insertions(+), 42 deletions(-) -- 2.26.3

3 6

[PATCH for 7.4.0] conf: Avoid double indentation of <metadata/> element
by Michal Privoznik 25 May '21

25 May '21

There was a recent change in libxml2 that caused a trouble for us. To us, <metadata/> in domain or network XMLs are just opaque value where management application can store whatever data it finds fit. At XML parser/formatter level, we just make a copy of the element during parsing and then format it back. For formatting we use xmlNodeDump() which allows caller to specify level of indentation. Previously, the indentation was not applied onto the very first line, but as of v2.9.12-2-g85b1792e libxml2 is applying indentation also on the first line. This does not work well with out virBuffer because as soon as we call virBufferAsprintf() to append <metadata/> element, virBufferAsprintf() will apply another level of indentation. Instead of version checking, let's skip any indentation added by libxml2 before virBufferAsprintf() is called. Note, the problem is only when telling xmlNodeDump() to use indentation, i.e. level argument is not zero. Therefore, virXMLNodeToString() which also calls xmlNodeDump() is safe as it passes zero. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- I've raised this issue with libxml2 team: https://gitlab.gnome.org/GNOME/libxml2/-/commit/85b1792e37b131e7a51af98a37f… but I'm not sure it'll be fixed. Unfortunately, the patch that's causing us trouble is being backported all over the place, because it's supposedly fixing a regression. src/conf/domain_conf.c | 9 ++++++++- src/conf/network_conf.c | 9 ++++++++- 2 files changed, 16 insertions(+), 2 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 4f78b7b43d..84a8c269be 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -27808,6 +27808,7 @@ virDomainDefFormatInternalSetRootName(virDomainDef *def, if (def->metadata) { g_autoptr(xmlBuffer) xmlbuf = NULL; + const char *xmlbufContent = NULL; int oldIndentTreeOutput = xmlIndentTreeOutput; /* Indentation on output requires that we previously set @@ -27824,7 +27825,13 @@ virDomainDefFormatInternalSetRootName(virDomainDef *def, xmlIndentTreeOutput = oldIndentTreeOutput; return -1; } - virBufferAsprintf(buf, "%s\n", (char *) xmlBufferContent(xmlbuf)); + + /* After libxml2-v2.9.12-2-g85b1792e even the first line is indented. + * But virBufferAsprintf() also adds indentation. Skip one of them. */ + xmlbufContent = (const char *) xmlBufferContent(xmlbuf); + virSkipSpaces(&xmlbufContent); + + virBufferAsprintf(buf, "%s\n", xmlbufContent); xmlIndentTreeOutput = oldIndentTreeOutput; } diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index a9eadff29c..20c6dc091a 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -2488,6 +2488,7 @@ virNetworkDefFormatBuf(virBuffer *buf, if (def->metadata) { g_autoptr(xmlBuffer) xmlbuf = NULL; + const char *xmlbufContent = NULL; int oldIndentTreeOutput = xmlIndentTreeOutput; /* Indentation on output requires that we previously set @@ -2504,7 +2505,13 @@ virNetworkDefFormatBuf(virBuffer *buf, xmlIndentTreeOutput = oldIndentTreeOutput; return -1; } - virBufferAsprintf(buf, "%s\n", (char *) xmlBufferContent(xmlbuf)); + + /* After libxml2-v2.9.12-2-g85b1792e even the first line is indented. + * But virBufferAsprintf() also adds indentation. Skip one of them. */ + xmlbufContent = (const char *) xmlBufferContent(xmlbuf); + virSkipSpaces(&xmlbufContent); + + virBufferAsprintf(buf, "%s\n", xmlbufContent); xmlIndentTreeOutput = oldIndentTreeOutput; } -- 2.26.3

5 6

RFC: Qemu backup interface plans
by Vladimir Sementsov-Ogievskiy 25 May '21

25 May '21

Hi all! I'd like to share and discuss some plans on Qemu backup interface I have. (actually, most of this I've presented on KVM Forum long ago.. But now I'm a lot closer to realization:) I'd start with a reminder about image fleecing: We have image fleecing scheme to export point-in-time state of active disk (iotest 222): backup(sync=none) ┌───────────────────────────────────────┐ ▼ │ ┌────────────┐ ┌────────────────┐ backing ┌─────────────┐ │ NBD export │ ─── │ temp qcow2 img │ ───────────────────▶ │ active disk │ └────────────┘ └────────────────┘ └─────────────┘ ▲ ┌────────────┐ │ │ guest blk │ ──────────────────────────────────────────────┘ └────────────┘ Actually, backup job inserts a backup-top filter, so in detail it looks like: backup(sync=none) ┌───────────────────────────────────────┐ ▼ │ ┌────────────┐ ┌────────────────┐ backing ┌─────────────┐ │ NBD export │ ─── │ temp qcow2 img │ ───────────────────▶ │ active disk │ └────────────┘ └────────────────┘ └─────────────┘ ▲ ▲ │ target │ │ │ ┌────────────┐ ┌────────────────┐ backing │ │ guest blk │ ──▶ │ backup-top │ ───────────────────────┘ └────────────┘ └────────────────┘ This scheme is also called external backup or pull backup. It allows some external tool to write data to actual backup, and Qemu only provides this data. We support also incremental external backup: Qemu can manage dirty bitmaps in any way user wants, and we can export bitmaps through NBD protocol. So, client of NBD export can get the bitmap, and read only "dirty" regions of exported image. What we lack in this scheme: 1. handling dirty bitmap in backup-top filter: backup-top does copy-before-write operation on any guest write, when actually we are interested only in "dirty" regions for incremental backup Probable solution would allowing specifying bitmap for sync=none mode of backup, but I think what I propose below is better. 2. [actually it's a tricky part of 1]: possibility to not do copy-before-write operations for regions that was already copied to final backup. With normal Qemu backup job, this is achieved by the fact that block-copy state with its internal bitmap is shared between backup job and copy-before-write filter. 3. Not a real problem but fact: backup block-job does nothing in the scheme, the whole job is done by filter. So, it would be interesting to have a possibility to simply insert/remove the filter, and avoid block-job creation and managing at all for external backup. (and I'd like to send another RFC on how to insert/remove filters, let's not discuss it here). Next. Think about internal backup. It has one drawback too: 4. If target is remote with slow connection, copy-before-write operations will slow down guest writes appreciably. It may be solved with help of image fleecing: we create temporary qcow2 image, setup fleecing scheme, and instead of exporting temp image through NBD we start a second backup with source = temporary image and target would be real backup target (NBD for example). Still, with such solution there are same [1,2] problems, 3 becomes worse: 5. We'll have two jobs and two automatically inserted filters, when actually one filter and one job are enough (as first job is needed only to insert a filter, second job doesn't need a filter at all). Note also, that this (starting two backup jobs to make push backup with fleecing) doesn't work now, op-blockers will be against. It's simple to fix (and in Virtuozzo we live with downstream-only patch, which allows push backup with fleecing, based on starting two backup jobs).. But I never send a patch, as I have better plan, which will solve all listed problems. So, what I propose: 1. We make backup-top filter public, so that it could be inserted/removed where user wants through QMP (how to properly insert/remove filter I'll post another RFC, as backup-top is not the only filter that can be usefully inserted somewhere). For this first step I've sent a series today: subject: [PATCH 00/21] block: publish backup-top filter id: <20210517064428.16223-1-vsementsov(a)virtuozzo.com> patchew: https://patchew.org/QEMU/20210517064428.16223-1-vsementsov@virtuozzo.com/ (note, that one of things in this series is rename s/backup-top/copy-before-write/, still, I call it backup-top in this letter) This solves [3]. [4, 5] are solved partly: we still have one extra filter, created by backup block jobs, and also I didn't test does this work, probably some op-blockers or permissions should be tuned. So, let it be step 2: 2. Test, that we can start backup job with source = (target of backup-top filter), so that we have "push backup with fleecing". Make an option for backup to start without a filter, when we don't need copy-before-write operations, to not create extra superfluous filter. 3. Support bitmap in backup-top filter, to solve [1] 3.1 and make it possible to modify the bitmap externally, so that consumer of fleecing can say to backup-top filter: I've already copied these blocks, don't bother with copying them to temp image". This is to solve [2]. Still, how consumer of fleecing will reset shared bitmap after copying blocks? I have the following idea: we make a "discard-bitmap-filter" filter driver, that own some bitmap and on discard request unset corresponding bits. Also, on read, if read from the region with unset bits the EINVAL returned immediately. This way both consumers (backup job and NBD client) are able to use this interface: Backup job can simply call discard on source, we can add an option for this. External backup tool will send TRIM request after reading some region. This way disk space will be freed and no extra copy-before-write operations will be done. I also have a side idea that we can implement READ_ONCE flag, so that READ and TRIM can be done in one NBD command. But this works only for clients that don't want to implement any kind of retrying. So, finally, how will it look (here I call backup-top with a new name, and "file" child is used instead of "backing", as this change I propose in "[PATCH 00/21] block: publish backup-top filter"): Pull backup: ┌────────────────────────────────────┐ │ NBD export │ └────────────────────────────────────┘ │ │ ┌────────────────────────────────────┐ file ┌───────────────────────────────────────┐ backing ┌─────────────┐ │ discard-bitmap filter (bitmap=bm0) │ ──────▶ │ temp qcow2 img │ ─────────▶ │ active disk │ └────────────────────────────────────┘ └───────────────────────────────────────┘ └─────────────┘ ▲ ▲ │ target │ │ │ ┌────────────────────────────────────┐ ┌───────────────────────────────────────┐ file │ │ guest blk │ ──────▶ │ copy-before-write filter (bitmap=bm0) │ ─────────────┘ └────────────────────────────────────┘ └───────────────────────────────────────┘ Operations: - Create temp qcow2 image - blockdev-add to add the new image, setup its backing, and add two filters - some command to actually set backup-top filter as child of guest blk. That's a "point-in-time" of backup. Should be done during fs-freeze. - start NBD export on top of discard filter (and we can export bitmap bm0 as well, for the client) Now NBD client (our external backup tool) can: - import the bitmap - READ the data - send DISCARD requests on already handled areas to save disk space and avoid extra copy-before-write operations on host node Push backup with fleecing: ┌─────────────────────┐ │ final backup target │ └─────────────────────┘ ▲ │ backup job (bitmap=bm0, insert-filter=False, discard-source=True) │ ┌────────────────────────────────────┐ file ┌───────────────────────────────────────┐ backing ┌─────────────┐ │ discard-bitmap filter (bitmap=bm0) │ ──────▶ │ temp qcow2 img │ ─────────▶ │ active disk │ └────────────────────────────────────┘ └───────────────────────────────────────┘ └─────────────┘ ▲ ▲ │ target │ │ │ ┌────────────────────────────────────┐ ┌───────────────────────────────────────┐ file │ │ guest blk │ ──────▶ │ copy-before-write filter (bitmap=bm0) │ ─────────────┘ └────────────────────────────────────┘ └───────────────────────────────────────┘ Note, that the whole fleecing part is the same, we only need to run backup job instead of NBD export. Additional future idea. Why we need push backup with fleecing? To handle cases with slow connection to backup target. In any case when writing to remote target is slower than writing to local file, push-backup-with-fleecing will less disturb the running guest than simple backup job. But this is not free: 1. We need additional disk space on source. No way to fix that (that's a core idea of the scheme, store data locally), still discard-source option for backup job will help 2. If connection is not too slow than probably part of CBW (copy before write) operations could go to final target immediately. But with the scheme above all CBW operations go to qcow2 temporary image. This can be solved with help of ram-cache format driver (to be implemented, of course): ┌─────────────────────┐ │ final backup target │ └─────────────────────┘ ▲ │ backup job (bitmap=bm0, insert-filter=False, discard-source=True) │ ┌────────────────────────────────────┐ ┌───────────────────────────────────────┐ backing ┌─────────────┐ │ discard-bitmap filter (bitmap=bm0) │ │ temp qcow2 img │ ─────────▶ │ active disk │ └────────────────────────────────────┘ └───────────────────────────────────────┘ └─────────────┘ │ ▲ ▲ │ │ file │ │ │ │ │ file ┌───────────┐ │ └───────────▶ │ ram-cache │ │ └───────────┘ │ ▲ │ │ target │ │ │ ┌────────────────────────────────────┐ ┌───────────────────────────────────────┐ file │ │ guest blk │ ──────▶ │ copy-before-write filter (bitmap=bm0) │ ─────────────┘ └────────────────────────────────────┘ └───────────────────────────────────────┘ This way data from copy-before-write filter goes first to ram-cache, and backup job could read it from ram. ram-cache will automatically flush data to temp qcow2 image, when ram-usage limit is reached. We'll also need a way to say backup-job that it should first read clusters that are cached in ram, and only then other clusters. So, we'll have a priory for clusters to be copied by block-copy: 1. clusters in ram-cache 2. clusters not in temp img (to avoid copy-before-write operations in future) 3. clusters in temp img. This will be a kind of block_status() thing, that allows a block driver to give recommendations on sequence of reading to be effective. Not also, that there is another benefit of such thing: we'll implement this callback in qcow2 driver, so that backup will read clusters not in guest cluster order, but in host cluster order, to read more sequentially, which should bring better performance on rotating disks. -- Best regards, Vladimir

3 8

[libvirt PATCH 0/4] adjust max memlock limit when hotplugging a vDPA device
by Laine Stump 25 May '21

25 May '21

see patch 4 for the why. Found during testing of https://bugzilla.redhat.com/1939776 Laine Stump (4): qemu_hotplug.c: don't skip cleanup on failures of qemuDomainAttachNetDevice conf: new function virDomainNetRemoveByObj() qemu_hotplug.c: add net devices to the domain list earlier qemu: adjust the maxmemlock limit when hotplugging a vDPA device src/conf/domain_conf.c | 16 ++++++++++++++++ src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 1 + src/qemu/qemu_hotplug.c | 41 +++++++++++++++++++++++++++++++--------- 4 files changed, 50 insertions(+), 9 deletions(-) -- 2.31.1

2 5

[PATCH 0/3] capabilities: Turn @cpus in _virCapsHostNUMACell into GArray
by Michal Privoznik 24 May '21

24 May '21

Honestly, I don't really like GArray-s. I find plain C arrays better (smaller memory footprint, direct access to items, no obscure g_array_*() bugs, fewer lines of code, etc.). But I'm posting these to stir discussion. However, 1/3 is still proper patch worth merging. Michal Prívozník (3): libxl: Break down an if() in libxlCapsInitNuma() capabilities: Turn @cpus in _virCapsHostNUMACell into GArray virCapabilitiesHostNUMAInitReal: Drop cleanup label src/conf/capabilities.c | 145 +++++++++++++++++---------------- src/conf/capabilities.h | 11 +-- src/libvirt_private.syms | 2 +- src/libxl/libxl_capabilities.c | 34 ++++---- src/test/test_driver.c | 12 +-- tests/testutils.c | 23 +++--- 6 files changed, 121 insertions(+), 106 deletions(-) -- 2.26.3

2 4

Ceph RBD Connection
by Mr. Gecko 24 May '21

24 May '21

Hello, I setup a Ceph Cluster on my system in hopes of using it with libvirtd, however I'm finding myself unable to have libvirtd make the connection. [root@server ~]# virsh pool-start "${CEPH_POOL}" error: Failed to start pool libvirt-pool error: failed to connect to the RADOS monitor on: localhost:6789,: No such file or directory I know that rbd is working as I am able to use it with the rbd utility. [root@server ~]# rbd list libvirt-pool Kolab [root@server ~]# rbd device map libvirt-pool/Kolab /dev/rbd0 [root@server ~]# lsblk /dev/rbd0 NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT rbd0 253:0 0 20G 0 disk ├─rbd0p1 253:1 0 1G 0 part └─rbd0p2 253:2 0 19G 0 part The libvirtd logs are not providing any useful information on the issue, I tried changing log-level to 1 and enabling the log_filters/adding "1:rbd" to the log filters. I have tried both the libvirt version 7.3.0 in my repository and the bleeding edge with a compile from git. Both returns the same error. Any help in troubleshooting this issue? I'm somewhat new when it comes to all of this. Thank you, James Coleman

2 3

[PATCH] hw/arm/raspi: Remove deprecated raspi2/raspi3 aliases
by Philippe Mathieu-Daudé 24 May '21

24 May '21

Remove the raspi2/raspi3 machine aliases, deprecated since commit 155e1c82ed0. Signed-off-by: Philippe Mathieu-Daudé <f4bug(a)amsat.org> --- docs/system/deprecated.rst | 7 ------- docs/system/removed-features.rst | 7 +++++++ hw/arm/raspi.c | 2 -- 3 files changed, 7 insertions(+), 9 deletions(-) diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index 80cae862528..7895bd4d849 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -238,13 +238,6 @@ this CPU is also deprecated. System emulator machines ------------------------ -Raspberry Pi ``raspi2`` and ``raspi3`` machines (since 5.2) -''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' - -The Raspberry Pi machines come in various models (A, A+, B, B+). To be able -to distinguish which model QEMU is implementing, the ``raspi2`` and ``raspi3`` -machines have been renamed ``raspi2b`` and ``raspi3b``. - Device options -------------- diff --git a/docs/system/removed-features.rst b/docs/system/removed-features.rst index 29e90601a51..8a8b8ca627b 100644 --- a/docs/system/removed-features.rst +++ b/docs/system/removed-features.rst @@ -312,6 +312,13 @@ This machine has been renamed ``fuloong2e``. These machine types were very old and likely could not be used for live migration from old QEMU versions anymore. Use a newer machine type instead. +Raspberry Pi ``raspi2`` and ``raspi3`` machines (removed in 6.1) +'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' + +The Raspberry Pi machines come in various models (A, A+, B, B+). To be able +to distinguish which model QEMU is implementing, the ``raspi2`` and ``raspi3`` +machines have been renamed ``raspi2b`` and ``raspi3b``. + linux-user mode CPUs -------------------- diff --git a/hw/arm/raspi.c b/hw/arm/raspi.c index 990509d3852..20bba0316f1 100644 --- a/hw/arm/raspi.c +++ b/hw/arm/raspi.c @@ -342,7 +342,6 @@ static void raspi2b_machine_class_init(ObjectClass *oc, void *data) MachineClass *mc = MACHINE_CLASS(oc); RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc); - mc->alias = "raspi2"; rmc->board_rev = 0xa21041; raspi_machine_class_common_init(mc, rmc->board_rev); }; @@ -362,7 +361,6 @@ static void raspi3b_machine_class_init(ObjectClass *oc, void *data) MachineClass *mc = MACHINE_CLASS(oc); RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc); - mc->alias = "raspi3"; rmc->board_rev = 0xa02082; raspi_machine_class_common_init(mc, rmc->board_rev); }; -- 2.26.3

3 4

[PATCH 0/4] conf: Deduplicate NUMA distance code
by Michal Privoznik 24 May '21

24 May '21

While the number of saved lines is not that great, it prepares the code for what I'm working on - exposing HMAT in capabilities XML. My idea is to reuse XML formatters there too so that mgmt app can just take whatever we produce in capabilities XML and paste it into domain XML. Michal Prívozník (4): capabilities: Rename siblings to distances numa_conf: Rename virDomainNumaDistance to virNumaDistance numa_conf: Expose virNumaDistance formatter conf: Deduplicate NUMA distance code src/conf/capabilities.c | 60 ++++++++++++++------------------ src/conf/capabilities.h | 13 +++---- src/conf/numa_conf.c | 63 ++++++++++++++++++---------------- src/conf/numa_conf.h | 10 ++++++ src/conf/virconftypes.h | 2 -- src/libvirt_private.syms | 1 + src/libxl/libxl_capabilities.c | 20 +++++------ 7 files changed, 83 insertions(+), 86 deletions(-) -- 2.26.3

3 12

[PATCH 00/35] qemu: Implement support for transient disk hotplug and backing store sharing
by Peter Krempa 24 May '21

24 May '21

This series refactors the qemu disk hotplug code so that it can be easily extended for hotplug of transient disks and then also used at startup to plug in disks where the users want to share the backing image. Masayoshi Mizuma (1): qemu_snapshot: Add the guest name to the transient disk path Peter Krempa (34): qemu: snapshot: Extract setup of snapshot disk definition for transient disks qemu: process: Setup transient disks only when starting a fresh VM qemuSnapshotDiskPrepareOne: Pass in qemuSnapshotDiskContext qemuSnapshotDiskContext: Store virQEMUDriverConfig in the struct qemuSnapshotDiskPrepareOne: Use data from qemuSnapshotDiskContext qemuSnapshotDiskCreate: Use 'cfg' from the qemuSnapshotDiskContext qemu: snapshot: move transient snapshot code to qemu_process qemu: Move 'bootindex' handling for disks out of command line formatter qemu: Move bootindex usage logic into qemuBuildDiskDeviceStr qemuxml2argvtest: Remove pointless tests for keywrapping on s390 qemu: Move iothread and s390 address validation for disk devices into the validator Replace virDomainDiskInsertPreAlloced by virDomainDiskInsert conf: remove virDomainDiskInsertPreAlloced qemuBlockStorageSourceChainData: Add handling of 'copy-on-read' filter layer qemuDomainAttachDiskGeneric: Move 'copy-on-read' handling to qemuBlockStorageSourceChainData qemuDomainRemoveDiskDevice: Move 'copy-on-read' handling to qemuBlockStorageSourceChainData qemuDomainAttachDeviceDiskLiveInternal: Absorb qemuDomainAttachUSBMassStorageDevice qemuDomainAttachDeviceDiskLiveInternal: Absorb qemuDomainAttachVirtioDiskDevice qemuDomainAttachDeviceDiskLiveInternal: Absorb qemuDomainAttachSCSIDisk qemuDomainAttachDeviceDiskLiveInternal: Simplify call to qemuDomainAttachDiskGeneric qemuDomainAttachDiskGeneric: Move setup of disk into qemuDomainAttachDeviceDiskLiveInternal qemu: hotplug: Move post-insertion steps of disk hotplug to qemuDomainAttachDeviceDiskLiveInternal qemuDomainAttachDiskGeneric: Fix whitespace qemuDomainAttachDiskGeneric: Refactor cleanup qemuDomainAttachDiskGeneric: Move PR helper attach into qemuDomainAttachDeviceDiskLiveInternal qemuDomainAttachDiskGeneric: Refactor rollback handling qemuDomainAttachDiskGeneric: Split up frontend and backend attachment qemu: Track creation of <transient> disk overlay individually qemuDomainAttachDiskGeneric: Implement hotplug of <transient> disk qemuDomainAttachDiskGeneric: Pass the qemu async job type qemuDomainAttachDiskGeneric: Export conf: Introduce 'shareBacking' for <transient> disks qemu: Allow <transient> disks with images shared accross VMs tests: Add qemuxml2argv and qemuxml2xml test for <transient shareBacking='yes'> docs/formatdomain.rst | 6 + docs/schemas/domaincommon.rng | 5 + src/conf/domain_conf.c | 33 +- src/conf/domain_conf.h | 3 +- src/libvirt_private.syms | 1 - src/libxl/libxl_driver.c | 4 +- src/lxc/lxc_driver.c | 6 +- src/qemu/qemu_block.c | 12 + src/qemu/qemu_block.h | 4 + src/qemu/qemu_command.c | 134 ++----- src/qemu/qemu_command.h | 1 - src/qemu/qemu_domain.c | 2 - src/qemu/qemu_domain.h | 11 +- src/qemu/qemu_hotplug.c | 335 +++++++++--------- src/qemu/qemu_hotplug.h | 6 + src/qemu/qemu_process.c | 203 ++++++++++- src/qemu/qemu_snapshot.c | 160 +++------ src/qemu/qemu_snapshot.h | 27 +- src/qemu/qemu_validate.c | 73 ++++ .../disk-transient.x86_64-4.1.0.err | 2 +- .../disk-transient.x86_64-latest.args | 4 +- tests/qemuxml2argvdata/disk-transient.xml | 6 + tests/qemuxml2argvtest.c | 8 - .../disk-transient.x86_64-latest.xml | 48 +++ tests/qemuxml2xmltest.c | 3 +- 25 files changed, 644 insertions(+), 453 deletions(-) create mode 100644 tests/qemuxml2xmloutdata/disk-transient.x86_64-latest.xml -- 2.31.1

3 72

[PATCH] kbase: Fix broken link of migration doc
by Han Han 24 May '21

24 May '21

Signed-off-by: Han Han <hhan(a)redhat.com> --- docs/kbase/index.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/kbase/index.rst b/docs/kbase/index.rst index d483ca94de..91083ee49d 100644 --- a/docs/kbase/index.rst +++ b/docs/kbase/index.rst @@ -66,4 +66,4 @@ Internals / Debugging `VM migration internals <migrationinternals.html>`__ VM migration implementation details, complementing the info in - `migration <migration.html>`__ + `migration <../migration.html>`__ -- 2.31.1

2 1

[libvirt PATCH v2 0/2] storage_driver: add support for cluster_size QCOW2 option
by Pavel Hrdina 22 May '21

22 May '21

Pavel Hrdina (2): storage: add support for QCOW2 cluster_size option storage_file: add support to probe cluster_size from QCOW2 images docs/formatstorage.html.in | 6 ++ docs/schemas/storagecommon.rng | 7 ++ docs/schemas/storagevol.rng | 3 + src/conf/storage_conf.c | 12 ++++ src/storage/storage_util.c | 8 +++ src/storage_file/storage_file_probe.c | 70 ++++++++++++++----- .../qcow2-clusterSize.argv | 6 ++ tests/storagevolxml2argvtest.c | 4 ++ .../vol-qcow2-clusterSize.xml | 17 +++++ .../vol-qcow2-clusterSize.xml | 17 +++++ tests/storagevolxml2xmltest.c | 1 + 11 files changed, 134 insertions(+), 17 deletions(-) create mode 100644 tests/storagevolxml2argvdata/qcow2-clusterSize.argv create mode 100644 tests/storagevolxml2xmlin/vol-qcow2-clusterSize.xml create mode 100644 tests/storagevolxml2xmlout/vol-qcow2-clusterSize.xml -- 2.31.1

3 6

Qemu block filter insertion/removal API
by Vladimir Sementsov-Ogievskiy 21 May '21

21 May '21

Hi all! I'd like to be sure that we know where we are going to. In blockdev-era where qemu user is aware about block nodes, all nodes have good names and controlled by user we can efficiently use block filters. We already have some useful filters: copy-on-read, throttling, compress. In my parallel series I make backup-top filter public and useful without backup block jobs. But now filters could be inserted only together with opening their child. We can specify filters in qemu cmdline, or filter can take place in the block node chain created by blockdev-add. Still, it would be good to insert/remove filters on demand. Currently we are going to use x-blockdev-reopen for this. Still it can't be used to insert a filter above root node (as x-blockdev-reopen can change only block node options and their children). In my series "[PATCH 00/21] block: publish backup-top filter" I propose (as Kevin suggested) to modify qom-set, so that it can set drive option of running device. That's not difficult, but it means that we have different scenario of inserting/removing filters: 1. filter above root node X: inserting: - do blockdev-add to add a filter (and specify X as its child) - do qom-set to set new filter as a rood node instead of X removing - do qom-set to make X a root node again - do blockdev-del to drop a filter 2. filter between two block nodes P and X. (For example, X is a backing child of P) inserting - do blockdev-add to add a filter (and specify X as its child) - do blockdev-reopen to set P.backing = filter remvoing - do blockdev-reopen to set P.backing = X - do blockdev-del to drop a filter And, probably we'll want transaction support for all these things. Is it OK? Or do we need some kind of additional blockdev-replace command, that can replace one node by another, so in both cases we will do inserting: - blockdev-add filter - blockdev-replace (make all parents of X to point to the new filter instead (except for the filter itself of course) removing - blockdev-replace (make all parante of filter to be parents of X instead) - blockdev-del filter It's simple to implement, and it seems for me that it is simpler to use. Any thoughts? -- Best regards, Vladimir

3 8

[PATCH v1] libxl: remove libxl_domain_build_info_init from libxlMakeDomBuildInfo
by Olaf Hering 21 May '21

21 May '21

The used libxl_domain_build_info, which is contained in libxl_domain_config, is owned and already initialized by the caller. Signed-off-by: Olaf Hering <olaf(a)aepfle.de> --- src/libxl/libxl_conf.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 1fd0322b6a..3598ca395d 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -346,8 +346,6 @@ libxlMakeDomBuildInfo(virDomainDef *def, size_t i; size_t nusbdevice = 0; - libxl_domain_build_info_init(b_info); - if (hvm) { libxl_domain_build_info_init_type(b_info, LIBXL_DOMAIN_TYPE_HVM); } else if (pvh) {

2 1

[PATCH v1] libxl: remove libxl_domain_create_info_init from libxlMakeDomCreateInfo
by Olaf Hering 21 May '21

21 May '21

The passed libxl_domain_create_info is owned, and already initialized, by the caller. Signed-off-by: Olaf Hering <olaf(a)aepfle.de> --- src/libxl/libxl_conf.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 8dc7e26cea..1fd0322b6a 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -129,8 +129,6 @@ libxlMakeDomCreateInfo(libxl_ctx *ctx, { char uuidstr[VIR_UUID_STRING_BUFLEN]; - libxl_domain_create_info_init(c_info); - if (def->os.type == VIR_DOMAIN_OSTYPE_HVM || def->os.type == VIR_DOMAIN_OSTYPE_XENPVH) { #ifdef WITH_XEN_PVH

2 1

[PATCH v2] libxl: adjust handling of libxl_device_disk objects
by Olaf Hering 21 May '21

21 May '21

libxl objects are supposed to be initialized and disposed. Correct the usage of libxl_device_disk objects which are allocated on the stack. Initialize each one prior usage, and dispose them once done. Adjust libxlMakeDisk to use an already initialized object, it is owned by the caller. Adjust libxlMakeDiskList to initialize the list of objects, before they are filled by libxlMakeDisk. In case of error, the objects are disposed by libxl_domain_config_dispose. Signed-off-by: Olaf Hering <olaf(a)aepfle.de> --- src/libxl/libxl_conf.c | 20 +++++--------------- src/libxl/libxl_driver.c | 6 ++++++ 2 files changed, 11 insertions(+), 15 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 3149ee3b4a..f29a28a841 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -1114,8 +1114,6 @@ libxlMakeDisk(virDomainDiskDef *l_disk, libxl_device_disk *x_disk) int format = virDomainDiskGetFormat(l_disk); int actual_type = virStorageSourceGetActualType(l_disk->src); - libxl_device_disk_init(x_disk); - if (actual_type == VIR_STORAGE_TYPE_NETWORK) { if (STRNEQ_NULLABLE(driver, "qemu")) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", @@ -1265,26 +1263,18 @@ libxlMakeDiskList(virDomainDef *def, libxl_domain_config *d_config) { virDomainDiskDef **l_disks = def->disks; int ndisks = def->ndisks; - libxl_device_disk *x_disks; size_t i; - x_disks = g_new0(libxl_device_disk, ndisks); + d_config->disks = g_new0(libxl_device_disk, ndisks); + d_config->num_disks = ndisks; for (i = 0; i < ndisks; i++) { - if (libxlMakeDisk(l_disks[i], &x_disks[i]) < 0) - goto error; + libxl_device_disk_init(&d_config->disks[i]); + if (libxlMakeDisk(l_disks[i], &d_config->disks[i]) < 0) + return -1; } - d_config->disks = x_disks; - d_config->num_disks = ndisks; - return 0; - - error: - for (i = 0; i < ndisks; i++) - libxl_device_disk_dispose(&x_disks[i]); - VIR_FREE(x_disks); - return -1; } /* diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index d54cd41785..2b844bb3b5 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -2978,6 +2978,7 @@ libxlDomainChangeEjectableMedia(virDomainObj *vm, virDomainDiskDef *disk) size_t i; int ret = -1; + libxl_device_disk_init(&x_disk); for (i = 0; i < vm->def->ndisks; i++) { if (vm->def->disks[i]->bus == disk->bus && STREQ(vm->def->disks[i]->dst, disk->dst)) { @@ -3018,6 +3019,7 @@ libxlDomainChangeEjectableMedia(virDomainObj *vm, virDomainDiskDef *disk) ret = 0; cleanup: + libxl_device_disk_dispose(&x_disk); virObjectUnref(cfg); return ret; } @@ -3030,6 +3032,7 @@ libxlDomainAttachDeviceDiskLive(virDomainObj *vm, virDomainDeviceDef *dev) libxl_device_disk x_disk; int ret = -1; + libxl_device_disk_init(&x_disk); switch (l_disk->device) { case VIR_DOMAIN_DISK_DEVICE_CDROM: ret = libxlDomainChangeEjectableMedia(vm, l_disk); @@ -3091,6 +3094,7 @@ libxlDomainAttachDeviceDiskLive(virDomainObj *vm, virDomainDeviceDef *dev) } cleanup: + libxl_device_disk_dispose(&x_disk); virObjectUnref(cfg); return ret; } @@ -3329,6 +3333,7 @@ libxlDomainDetachDeviceDiskLive(virDomainObj *vm, virDomainDeviceDef *dev) int idx; int ret = -1; + libxl_device_disk_init(&x_disk); switch (dev->data.disk->device) { case VIR_DOMAIN_DISK_DEVICE_DISK: if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_XEN) { @@ -3380,6 +3385,7 @@ libxlDomainDetachDeviceDiskLive(virDomainObj *vm, virDomainDeviceDef *dev) } cleanup: + libxl_device_disk_dispose(&x_disk); virObjectUnref(cfg); return ret; }

2 1

[PATCH v2] conf: Report alias name of the detached device in error
by Kristina Hanicova 21 May '21

21 May '21

This is v2 from: https://listman.redhat.com/archives/libvir-list/2021-May/msg00481.html I have reworked the code a bit to have only one error report instead of multiple ones with different combinations of possible matching items. Suggested by Laine. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1942367 Signed-off-by: Kristina Hanicova <khanicov(a)redhat.com> --- src/conf/domain_conf.c | 67 ++++++++++++++++-------------------------- 1 file changed, 25 insertions(+), 42 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b3ed3a9c5a..9631e40815 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -15645,6 +15645,12 @@ virDomainNetFindIdx(virDomainDef *def, virDomainNetDef *net) VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI); bool CCWAddrSpecified = virDomainDeviceAddressIsValid(&net->info, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW); + g_autofree char *addr = NULL; + const char *macAddr = _("(<null>)"); + const char *alias = _("(<null>)"); + + if (MACAddrSpecified) + macAddr = virMacAddrFormat(&net->mac, mac); for (i = 0; i < def->nnets; i++) { if (MACAddrSpecified && @@ -15676,7 +15682,7 @@ virDomainNetFindIdx(virDomainDef *def, virDomainNetDef *net) if (MACAddrSpecified) { virReportError(VIR_ERR_OPERATION_FAILED, _("multiple devices matching MAC address %s found"), - virMacAddrFormat(&net->mac, mac)); + macAddr); } else { virReportError(VIR_ERR_OPERATION_FAILED, "%s", _("multiple matching devices found")); @@ -15688,47 +15694,24 @@ virDomainNetFindIdx(virDomainDef *def, virDomainNetDef *net) matchidx = i; } - if (matchidx < 0) { - if (MACAddrSpecified && PCIAddrSpecified) { - virReportError(VIR_ERR_DEVICE_MISSING, - _("no device matching MAC address %s found on " - VIR_PCI_DEVICE_ADDRESS_FMT), - virMacAddrFormat(&net->mac, mac), - net->info.addr.pci.domain, - net->info.addr.pci.bus, - net->info.addr.pci.slot, - net->info.addr.pci.function); - } else if (MACAddrSpecified && CCWAddrSpecified) { - virReportError(VIR_ERR_DEVICE_MISSING, - _("no device matching MAC address %s found on " - VIR_CCW_DEVICE_ADDRESS_FMT), - virMacAddrFormat(&net->mac, mac), - net->info.addr.ccw.cssid, - net->info.addr.ccw.ssid, - net->info.addr.ccw.devno); - } else if (PCIAddrSpecified) { - virReportError(VIR_ERR_DEVICE_MISSING, - _("no device found on " VIR_PCI_DEVICE_ADDRESS_FMT), - net->info.addr.pci.domain, - net->info.addr.pci.bus, - net->info.addr.pci.slot, - net->info.addr.pci.function); - } else if (CCWAddrSpecified) { - virReportError(VIR_ERR_DEVICE_MISSING, - _("no device found on " VIR_CCW_DEVICE_ADDRESS_FMT), - net->info.addr.ccw.cssid, - net->info.addr.ccw.ssid, - net->info.addr.ccw.devno); - } else if (MACAddrSpecified) { - virReportError(VIR_ERR_DEVICE_MISSING, - _("no device matching MAC address %s found"), - virMacAddrFormat(&net->mac, mac)); - } else { - virReportError(VIR_ERR_DEVICE_MISSING, "%s", - _("no matching device found")); - } - } - return matchidx; + if (matchidx >= 0) + return matchidx; + + if (net->info.alias) + alias = net->info.alias; + + if (CCWAddrSpecified) + addr = virDomainCCWAddressAsString(&net->info.addr.ccw); + else if (PCIAddrSpecified) + addr = virPCIDeviceAddressAsString(&net->info.addr.pci); + else + addr = g_strdup(_("(<null>)")); + + virReportError(VIR_ERR_DEVICE_MISSING, + _("no device found at address '%s' matching MAC address" + " '%s' and alias '%s'"), + addr, macAddr, alias); + return -1; } bool -- 2.31.1

3 2

[libvirt PATCH] Revert "qemuDomainSetNumaParamsLive: set nodeset for root cgroup"
by Pavel Hrdina 21 May '21

21 May '21

This reverts commit <1b22dd6dd44202094e0f78f887cbe790c00e9ebc>. First of all, the reverted commit is incomplete. It only sets cpuset.mems in the VM root cgroup when the API is used but there is no code that would do the same when the VM is started. Libvirt never places any process into the VM root cgroup directly. All the supporting processes like slirp-helper or dbus-daemon are placed into the emulator sub-cgroup and all the QEMU threads are distributed between emulator, vcpu* and iothread* sub-cgroups. The scenario described in the reverted commit can happen only if someone manually adds any process there which we should not care about. If we would like to set the limit in the VM root cgroup we need to introduce better logic: - set both (old and new) numa group in the VM root cgroup - change the numa group in all sub-cgroups to new value - finally set only the new value in the VM root cgroup The simplest fix now is to revert the commit. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/qemu/qemu_driver.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index e10e699a1a..a972662c3f 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -8797,10 +8797,6 @@ qemuDomainSetNumaParamsLive(virDomainObj *vm, return -1; } - /* set nodeset for root cgroup */ - if (virCgroupSetCpusetMems(priv->cgroup, nodeset_str) < 0) - return -1; - return 0; } -- 2.31.1

2 1

RFC: qemu log permissions
by Kristina Hanicova 21 May '21

21 May '21

Hello everyone, I came across this issue: https://gitlab.com/libvirt/libvirt/-/issues/71 and I would like your opinion on implementing support to define mode (maybe even user/group/other?) in config, which the qemu log files would be created with. In this case, it is desired so that this tool 'Elasticsearch' can read logs. Is it worth implementing? What are your thoughts? Kristina

1 0

[PATCH] schema: Allow '0' offset for a <slice> of <disk>
by Peter Krempa 21 May '21

21 May '21

Using slice to cut off the end of the image is a perfectly vaid configuration. Use 'unsignedInt' instead of 'positiveInteger' for the 'offset' attribute in the XML schema and modify one test case to cover this use case. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1960993 Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- docs/schemas/domaincommon.rng | 2 +- tests/qemuxml2argvdata/disk-slices.x86_64-latest.args | 2 +- tests/qemuxml2argvdata/disk-slices.xml | 2 +- tests/qemuxml2xmloutdata/disk-slices.x86_64-latest.xml | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index a2e5c50c1d..475f7f1a42 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -1638,7 +1638,7 @@ <define name="diskSourceSlice"> <attribute name="offset"> - <ref name="positiveInteger"/> + <ref name="unsignedInt"/> </attribute> <attribute name="size"> <ref name="positiveInteger"/> diff --git a/tests/qemuxml2argvdata/disk-slices.x86_64-latest.args b/tests/qemuxml2argvdata/disk-slices.x86_64-latest.args index d337ea91a7..f41bbeea09 100644 --- a/tests/qemuxml2argvdata/disk-slices.x86_64-latest.args +++ b/tests/qemuxml2argvdata/disk-slices.x86_64-latest.args @@ -28,7 +28,7 @@ XDG_CONFIG_HOME=/tmp/lib/domain--1-QEMUGuest1/.config \ -boot strict=on \ -device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 \ -blockdev '{"driver":"file","filename":"/var/lib/libvirt/images/raw.img","node-name":"libvirt-6-storage","auto-read-only":true,"discard":"unmap"}' \ --blockdev '{"node-name":"libvirt-6-format","read-only":false,"driver":"raw","offset":1234,"size":321,"file":"libvirt-6-storage"}' \ +-blockdev '{"node-name":"libvirt-6-format","read-only":false,"driver":"raw","offset":0,"size":321,"file":"libvirt-6-storage"}' \ -device virtio-blk-pci,bus=pci.0,addr=0x2,drive=libvirt-6-format,id=virtio-disk0,bootindex=1 \ -blockdev '{"driver":"file","filename":"/var/lib/libvirt/images/raw.img","node-name":"libvirt-5-storage","auto-read-only":true,"discard":"unmap"}' \ -blockdev '{"driver":"raw","node-name":"libvirt-5-slice-sto","offset":9876,"size":123456789,"file":"libvirt-5-storage","auto-read-only":true,"discard":"unmap"}' \ diff --git a/tests/qemuxml2argvdata/disk-slices.xml b/tests/qemuxml2argvdata/disk-slices.xml index 1675766a46..016aa1b905 100644 --- a/tests/qemuxml2argvdata/disk-slices.xml +++ b/tests/qemuxml2argvdata/disk-slices.xml @@ -18,7 +18,7 @@ <driver name='qemu' type='raw'/> <source file='/var/lib/libvirt/images/raw.img'> <slices> - <slice type='storage' offset='1234' size='321'/> + <slice type='storage' offset='0' size='321'/> </slices> </source> <backingStore/> diff --git a/tests/qemuxml2xmloutdata/disk-slices.x86_64-latest.xml b/tests/qemuxml2xmloutdata/disk-slices.x86_64-latest.xml index 5c200ad31c..be5cd25084 100644 --- a/tests/qemuxml2xmloutdata/disk-slices.x86_64-latest.xml +++ b/tests/qemuxml2xmloutdata/disk-slices.x86_64-latest.xml @@ -21,7 +21,7 @@ <driver name='qemu' type='raw'/> <source file='/var/lib/libvirt/images/raw.img'> <slices> - <slice type='storage' offset='1234' size='321'/> + <slice type='storage' offset='0' size='321'/> </slices> </source> <backingStore/> -- 2.31.1

2 1