- Devel - libvirt List Archives

[libvirt] [PATCH v2 0/2] Add support for qcow2 cache
by Liu Qing 04 Sep '17

04 Sep '17

Qcow2 small IO random write performance will drop dramatically if the l2 cache table could not cover the whole disk. This will be a lot of l2 cache table RW operations if cache miss happens frequently. This patch exports the qcow2 driver parameter l2-cache-size/refcount-cache-size, first added in Qemu 2.2, and cache-clean-interval, first added in Qemu 2.5, in libvirt. *** BLURB HERE *** Liu Qing (2): conf, docs: Add qcow2 cache configuration support qemu: add capability checking for qcow2 cache configuration docs/formatdomain.html.in | 21 ++++++++++ docs/schemas/domaincommon.rng | 24 +++++++++++ src/conf/domain_conf.c | 30 ++++++++++++++ src/qemu/qemu_capabilities.c | 15 +++++++ src/qemu/qemu_capabilities.h | 3 ++ src/qemu/qemu_command.c | 47 ++++++++++++++++++++++ src/util/virstoragefile.c | 3 ++ src/util/virstoragefile.h | 4 ++ tests/qemucapabilitiesdata/caps_2.4.0.x86_64.xml | 2 + tests/qemucapabilitiesdata/caps_2.5.0.x86_64.xml | 3 ++ .../caps_2.6.0-gicv2.aarch64.xml | 3 ++ .../caps_2.6.0-gicv3.aarch64.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.6.0.ppc64le.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.6.0.x86_64.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.7.0.s390x.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.7.0.x86_64.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.8.0.s390x.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.8.0.x86_64.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.9.0.ppc64le.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.9.0.s390x.xml | 3 ++ tests/qemucapabilitiesdata/caps_2.9.0.x86_64.xml | 3 ++ ...argv-disk-drive-qcow2-cache-clean-interval.args | 27 +++++++++++++ ...2argv-disk-drive-qcow2-cache-clean-interval.xml | 37 +++++++++++++++++ ...emuxml2argv-disk-drive-qcow2-l2-cache-size.args | 27 +++++++++++++ ...qemuxml2argv-disk-drive-qcow2-l2-cache-size.xml | 37 +++++++++++++++++ ...2argv-disk-drive-qcow2-refcount-cache-size.args | 27 +++++++++++++ ...l2argv-disk-drive-qcow2-refcount-cache-size.xml | 37 +++++++++++++++++ tests/qemuxml2argvtest.c | 6 +++ ...mlout-disk-drive-qcow2-cache-clean-interval.xml | 41 +++++++++++++++++++ ...muxml2xmlout-disk-drive-qcow2-l2-cache-size.xml | 41 +++++++++++++++++++ ...xmlout-disk-drive-qcow2-refcount-cache-size.xml | 41 +++++++++++++++++++ tests/qemuxml2xmltest.c | 3 ++ 32 files changed, 509 insertions(+) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-cache-clean-interval.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-cache-clean-interval.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-l2-cache-size.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-l2-cache-size.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-refcount-cache-size.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-qcow2-refcount-cache-size.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-drive-qcow2-cache-clean-interval.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-drive-qcow2-l2-cache-size.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-drive-qcow2-refcount-cache-size.xml -- 1.8.3.1

1 0

[libvirt] [PATCH v2] news: Prepare for 3.7.0 release
by Andrea Bolognani 02 Sep '17

02 Sep '17

Documents some changes that have slipped through the cracks during the development cycle. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- Changes from [v1]: * rebase on top of master * remove the part about guests no longer disappearing if the QEMU binary is missing, since Peter already documented that [v1] https://www.redhat.com/archives/libvir-list/2017-September/msg00030.html docs/news.xml | 76 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 76 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index dd8967e25..141b1a42a 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -69,6 +69,18 @@ you can configure reconnect timeout if the connection is closed. </description> </change> + <change> + <summary> + qemu: Implement editing guest configuration for managed save files + </summary> + <description> + New <code>managedsave-define</code>, <code>managedsave-edit</code> + and <code>managedsave-dumpxml</code> commands have been added to + <code>virsh</code> to allow editing the guest configuration for + managed save files just like it was already possible for unmanaged + save files. + </description> + </change> </section> <section title="Improvements"> <change> @@ -81,6 +93,43 @@ indication for users (and devs). </description> </change> + <change> + <summary> + apparmor: Update for QEMU 2.10 compatibility + </summary> + <description> + Starting with version 2.10, QEMU locks disk images and NVRAM files + to prevent them from being corrupted; however, file locking needs to + be explicitly allowed through <code>virt-aa-helper</code> or AppArmor + will reject the requests and the guest will not be able to run. + </description> + </change> + <change> + <summary> + virsh: List Unix sockets in 'domdisplay' output + </summary> + <description> + VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets + as connection endpoints, but such a configuration was not handled + correctly by <code>virsh domdisplay</code>, causing the respective + endpoints to be missing from the output. + </description> + </change> + <change> + <summary> + qemu: Don't check whether offline migration is safe + </summary> + <description> + Since offline migration only copies the guest definition to the + destination host, data corruption is not a concern and the operation + can always be performed safely. + </description> + </change> + <change> + <summary> + virt-host-validate: Fix IOMMU detection on ppc64 + </summary> + </change> </section> <section title="Bug fixes"> <change> @@ -115,6 +164,33 @@ are missing during startup. </description> </change> + <change> + <summary> + qemu: Prevent pSeries guests from disappearing in some situations + </summary> + <description> + pSeries guest would disappear if any of the host devices they were + configured to use was not available during libvirtd startup, which + could easily happen for SR-IOV Virtual Functions. This scenario is + now handled correctly. + </description> + </change> + <change> + <summary> + qemu: Honor <code><on_reboot/></code> setting + </summary> + <description> + The setting was accepted by the parser, but not actually implemented. + </description> + </change> + <change> + <summary> + daemon: Fix <code>--verbose</code> option + </summary> + <description> + The option had not been working as intended from v3.0.0 onwards. + </description> + </change> </section> </release> <release version="v3.6.0" date="2017-08-02"> -- 2.13.5

2 2

[libvirt] [PATCH v2 0/2] Fix docs/news.xml template structure and add new features
by Kothapally Madhu Pavan 02 Sep '17

02 Sep '17

This patchset will fix docs/news.xml template structure which is broken by commit b7e779c1a51793 and add support for managedsave-define, managedsave-edit and managedsave-dumpxml commands. Kothapally Madhu Pavan (2): docs: Fix docs/news.xml template structure docs: Document managedsave-edit commands support docs/news.xml | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) -- 1.8.3.1

2 3

[libvirt] [PATCH] news: Prepare for 3.7.0 release
by Andrea Bolognani 01 Sep '17

01 Sep '17

Documents some changes that have slipped through the cracks during the development cycle. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- I'll push this by the end of tomorrow under the "nobody reads the release notes" rule, unless I get either an ACK or a NACK earlier. docs/news.xml | 77 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 77 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index 4b48f0fb3..40dec513f 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -69,8 +69,57 @@ you can configure reconnect timeout if the connection is closed. </description> </change> + <change> + <summary> + qemu: Implement editing guest configuration for managed save files + </summary> + <description> + New <code>managedsave-define</code>, <code>managedsave-edit</code> + and <code>managedsave-dumpxml</code> commands have been added to + <code>virsh</code> to allow editing the guest configuration for + managed save files just like it was already possible for unmanaged + save files. + </description> + </change> </section> <section title="Improvements"> + <change> + <summary> + apparmor: Update for QEMU 2.10 compatibility + </summary> + <description> + Starting with version 2.10, QEMU locks disk images and NVRAM files + to prevent them from being corrupted; however, file locking needs to + be explicitly allowed through <code>virt-aa-helper</code> or AppArmor + will reject the requests and the guest will not be able to run. + </description> + </change> + <change> + <summary> + virsh: List Unix sockets in 'domdisplay' output + </summary> + <description> + VNC and SPICE graphics can use Unix sockets instead of TCP/IP sockets + as connection endpoints, but such a configuration was not handled + correctly by <code>virsh domdisplay</code>, causing the respective + endpoints to be missing from the output. + </description> + </change> + <change> + <summary> + qemu: Don't check whether offline migration is safe + </summary> + <description> + Since offline migration only copies the guest definition to the + destination host, data corruption is not a concern and the operation + can always be performed safely. + </description> + </change> + <change> + <summary> + virt-host-validate: Fix IOMMU detection on ppc64 + </summary> + </change> </section> <section title="Bug fixes"> <change> @@ -93,6 +142,34 @@ The path is now generated with shortened name instead. </description> </change> + <change> + <summary> + qemu: Prevent guests from disappearing in some situations + </summary> + <description> + pSeries guest would disappear if any of the host devices they were + configured to use was not available during libvirtd startup, which + could easily happen for SR-IOV Virtual Functions; additionally, + guests would disappear if the QEMU binary was removed or renamed. + Both scenarios are now handled correctly. + </description> + </change> + <change> + <summary> + qemu: Honor <code><on_reboot/></code> setting + </summary> + <description> + The setting was accepted by the parser, but not actually implemented. + </description> + </change> + <change> + <summary> + daemon: Fix <code>--verbose</code> option + </summary> + <description> + The option had not been working as intended from v3.0.0 onwards. + </description> + </change> </section> </release> <release version="v3.6.0" date="2017-08-02"> -- 2.13.5

1 0

[libvirt] [PATCH 2/3] libvirtaio: add allow for moving callbacks to other event loop
by Wojtek Porczyk 01 Sep '17

01 Sep '17

The virEvent implementation is tied to a particular loop. When spinning another loop, the callbacks have to be moved to another implementation, so they will have a chance to be invoked, should they be scheduled. If not, file descriptors will be leaking. Signed-off-by: Wojtek Porczyk <woju(a)invisiblethingslab.com> --- libvirtaio.py | 64 +++++++++++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 51 insertions(+), 13 deletions(-) diff --git a/libvirtaio.py b/libvirtaio.py index fc868bd..d161cd1 100644 --- a/libvirtaio.py +++ b/libvirtaio.py @@ -195,9 +195,10 @@ class FDCallback(Callback): return '<{} iden={} fd={} event={}>'.format( self.__class__.__name__, self.iden, self.descriptor.fd, self.event) - def update(self, event): + def update(self, event=None): '''Update the callback and fix descriptor's watchers''' - self.event = event + if event is not None: + self.event = event self.descriptor.update() # @@ -238,20 +239,21 @@ class TimeoutCallback(Callback): self.cb(self.iden, self.opaque) self.impl.log.debug('timer %r callback ended', self.iden) - def update(self, timeout): + def update(self, timeout=None): '''Start or the timer, possibly updating timeout''' - self.timeout = timeout - - if self.timeout >= 0 and self._task is None: - self.impl.log.debug('timer %r start', self.iden) - self._task = ensure_future(self._timer(), - loop=self.impl.loop) + if timeout is not None: + self.timeout = timeout - elif self.timeout < 0 and self._task is not None: + if self._task is not None: self.impl.log.debug('timer %r stop', self.iden) self._task.cancel() # pylint: disable=no-member self._task = None + if self.timeout >= 0: + self.impl.log.debug('timer %r start', self.iden) + self._task = ensure_future(self._timer(), + loop=self.impl.loop) + def close(self): '''Stop the timer and call ff callback''' super(TimeoutCallback, self).close() @@ -274,6 +276,7 @@ class virEventAsyncIOImpl(object): self.callbacks = {} self.descriptors = DescriptorDict(self) self.log = logging.getLogger(self.__class__.__name__) + self.pending_tasks = set() def register(self): '''Register this instance as event loop implementation''' @@ -284,9 +287,30 @@ class virEventAsyncIOImpl(object): self._add_timeout, self._update_timeout, self._remove_timeout) return self + def takeover(self, other): + '''Take over other implementation, probably registered on another loop + + :param virEventAsyncIOImpl other: other implementation to be taken over + ''' + self.log.warning('%r taking over %r', self, other) + + while other.callbacks: + iden, callback = other.callbacks.popitem() + self.log.debug(' takeover %d %r', iden, callback) + assert callback.iden == iden + callback.impl = self + self.callbacks[iden] = callback + + if isinstance(callback, FDCallback): + fd = callback.descriptor.fd + assert callback is other.descriptors[fd].remove_handle(iden) + self.descriptors[fd].add_handle(callback) + def schedule_ff_callback(self, iden, opaque): '''Schedule a ff callback from one of the handles or timers''' - ensure_future(self._ff_callback(iden, opaque), loop=self.loop) + fut = ensure_future(self._ff_callback(iden, opaque), loop=self.loop) + self.pending_tasks.add(fut) + fut.add_done_callback(self.pending_tasks.remove) @asyncio.coroutine def _ff_callback(self, iden, opaque): @@ -297,13 +321,19 @@ class virEventAsyncIOImpl(object): self.log.debug('ff_callback(iden=%d, opaque=...)', iden) return libvirt.virEventInvokeFreeCallback(opaque) + @asyncio.coroutine + def drain(self): + self.log.debug('drain()') + if self.pending_tasks: + yield from asyncio.wait(self.pending_tasks, loop=self.loop) + def is_idle(self): '''Returns False if there are leftovers from a connection Those may happen if there are sematical problems while closing a connection. For example, not deregistered events before .close(). ''' - return not self.callbacks + return not self.callbacks and not self.pending_tasks def _add_handle(self, fd, event, cb, opaque): '''Register a callback for monitoring file handle events @@ -403,10 +433,18 @@ class virEventAsyncIOImpl(object): callback = self.callbacks.pop(timer) callback.close() + +_current_impl = None def virEventRegisterAsyncIOImpl(loop=None): '''Arrange for libvirt's callbacks to be dispatched via asyncio event loop The implementation object is returned, but in normal usage it can safely be discarded. ''' - return virEventAsyncIOImpl(loop=loop).register() + global _current_impl + impl = virEventAsyncIOImpl(loop=loop) + impl.register() + if _current_impl is not None: + impl.takeover(_current_impl) + _current_impl = impl + return impl -- 2.9.4

2 4

[libvirt] [PATCH] news: Add stuff for 3.7.0
by Peter Krempa 01 Sep '17

01 Sep '17

--- docs/news.xml | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index 4b48f0fb3..dd8967e25 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -71,6 +71,16 @@ </change> </section> <section title="Improvements"> + <change> + <summary> + qemu: Report a clear error when dropping a VM during startup + </summary> + <description> + "Failed to load config for domain 'DOMNAME'" is now reported if a VM + config can't be parsed for some reason, and thus provides a clear + indication for users (and devs). + </description> + </change> </section> <section title="Bug fixes"> <change> @@ -93,6 +103,18 @@ The path is now generated with shortened name instead. </description> </change> + <change> + <summary> + qemu: Tolerate missing emulator binary during libvirtd restart + </summary> + <description> + For some time libvirt required qemu capabilities being present when + parsing VM configs during startup. As a side effect VM configs would + fail to parse and thus vanish, if the emulator binary would be + uninstalled or broken. Libvirt now tolerates when capabilities + are missing during startup. + </description> + </change> </section> </release> <release version="v3.6.0" date="2017-08-02"> -- 2.14.1

2 1

[libvirt] [PATCH] virnetdaemon: Don't deadlock when talking to D-Bus
by Michal Privoznik 01 Sep '17

01 Sep '17

https://bugzilla.redhat.com/show_bug.cgi?id=1487322 In ace45e67abbd I've tried to fix a problem that we get the reply to a D-Bus call while we were sleeping. In that case the callback was never set. So I've changed the code that the callback is called directly in this case. However, I hadn't realized that since callback is called out of order it lock the virNetDaemon. Exactly the very same virNetDaemon that we are dealing with right now and have it locked already (in virNetDaemonAddShutdownInhibition()) Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/rpc/virnetdaemon.c | 21 +++++++++++++++------ 1 file changed, 15 insertions(+), 6 deletions(-) diff --git a/src/rpc/virnetdaemon.c b/src/rpc/virnetdaemon.c index 00247cfc3..e3b9390af 100644 --- a/src/rpc/virnetdaemon.c +++ b/src/rpc/virnetdaemon.c @@ -439,14 +439,12 @@ virNetDaemonAutoShutdown(virNetDaemonPtr dmn, #if defined(WITH_DBUS) && defined(DBUS_TYPE_UNIX_FD) static void -virNetDaemonGotInhibitReply(DBusPendingCall *pending, - void *opaque) +virNetDaemonGotInhibitReplyLocked(DBusPendingCall *pending, + virNetDaemonPtr dmn) { - virNetDaemonPtr dmn = opaque; DBusMessage *reply; int fd; - virObjectLock(dmn); dmn->autoShutdownCallingInhibit = false; VIR_DEBUG("dmn=%p", dmn); @@ -470,11 +468,22 @@ virNetDaemonGotInhibitReply(DBusPendingCall *pending, virDBusMessageUnref(reply); cleanup: - virObjectUnlock(dmn); dbus_pending_call_unref(pending); } +static void +virNetDaemonGotInhibitReply(DBusPendingCall *pending, + void *opaque) +{ + virNetDaemonPtr dmn = opaque; + + virObjectLock(dmn); + virNetDaemonGotInhibitReplyLocked(pending, dmn); + virObjectUnlock(dmn); +} + + /* As per: http://www.freedesktop.org/wiki/Software/systemd/inhibit */ static void virNetDaemonCallInhibit(virNetDaemonPtr dmn, @@ -516,7 +525,7 @@ virNetDaemonCallInhibit(virNetDaemonPtr dmn, 25 * 1000) && pendingReply) { if (dbus_pending_call_get_completed(pendingReply)) { - virNetDaemonGotInhibitReply(pendingReply, dmn); + virNetDaemonGotInhibitReplyLocked(pendingReply, dmn); } else { dbus_pending_call_set_notify(pendingReply, virNetDaemonGotInhibitReply, -- 2.13.5

3 2

[libvirt] [PATCH] vmx: Expose VMware Managed Object Reference (moref) in XML.
by Richard W.M. Jones 01 Sep '17

01 Sep '17

If you use the VDDK library to access virtual machines remotely, you really need to know the Managed Object Reference ("moref") of the VM. This must be passed each time you connect to the API. For example nbdkit's VDDK plugin requires a moref to be passed to mount up a VM's disk remotely: nbdkit vddk user=root password=+/tmp/rootpw \ server=esxi.example.com thumbprint=xx:xx:xx:... \ vm=moref=2 \ file="[datastore1] Fedora/Fedora.vmdk" Getting the moref is a huge pain. To get some idea of what it is, why it is needed, and how much trouble it is to get it, see: https://blogs.vmware.com/vsphere/2012/02/uniquely-identifying-virtual-machi… https://blogs.vmware.com/vsphere/2012/02/uniquely-identifying-virtual-machi… However the moref is available conveniently in the internals of the libvirt VMX driver. This patch exposes it as a custom XML element using the same "vmware:" namespace which was previously used for the datacenterpath (see libvirt commit 636a99058758a044). It appears in the XML like this: <domain type='vmware' xmlns:vmware='http://libvirt.org/schemas/domain/vmware/1.0'> <name>Fedora</name> ... <vmware:datacenterpath>ha-datacenter</vmware:datacenterpath> <vmware:moref>2</vmware:moref> </domain> Note that the moref can appear as either a simple ID (for esx:// connections) or as a "vm-<ID>" (for vpx:// connections). It should be treated by users as an opaque string. --- src/esx/esx_driver.c | 7 +++++++ src/esx/esx_vi.c | 15 +++++++++++++++ src/esx/esx_vi.h | 4 ++++ src/vmx/vmx.c | 51 +++++++++++++++++++++++++++++++++++++++++---------- src/vmx/vmx.h | 2 ++ 5 files changed, 69 insertions(+), 10 deletions(-) diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index 1f4f2c7a7..0cce0a41a 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -2645,6 +2645,7 @@ esxDomainGetXMLDesc(virDomainPtr domain, unsigned int flags) esxVI_ObjectContent *virtualMachine = NULL; esxVI_VirtualMachinePowerState powerState; int id; + char *moref = NULL; char *vmPathName = NULL; char *datastoreName = NULL; char *directoryName = NULL; @@ -2670,6 +2671,7 @@ esxDomainGetXMLDesc(virDomainPtr domain, unsigned int flags) esxVI_LookupVirtualMachineByUuid(priv->primary, domain->uuid, propertyNameList, &virtualMachine, esxVI_Occurrence_RequiredItem) < 0 || + esxVI_GetVirtualMachineMORef(virtualMachine, &moref) < 0 || esxVI_GetVirtualMachinePowerState(virtualMachine, &powerState) < 0 || esxVI_GetVirtualMachineIdentity(virtualMachine, &id, NULL, NULL) < 0 || esxVI_GetStringValue(virtualMachine, "config.files.vmPathName", @@ -2715,6 +2717,7 @@ esxDomainGetXMLDesc(virDomainPtr domain, unsigned int flags) ctx.formatFileName = NULL; ctx.autodetectSCSIControllerModel = NULL; ctx.datacenterPath = priv->primary->datacenterPath; + ctx.moref = moref; def = virVMXParseConfig(&ctx, priv->xmlopt, priv->caps, vmx); @@ -2732,6 +2735,7 @@ esxDomainGetXMLDesc(virDomainPtr domain, unsigned int flags) esxVI_String_Free(&propertyNameList); esxVI_ObjectContent_Free(&virtualMachine); + VIR_FREE(moref); VIR_FREE(datastoreName); VIR_FREE(directoryName); VIR_FREE(directoryAndFileName); @@ -2774,6 +2778,7 @@ esxConnectDomainXMLFromNative(virConnectPtr conn, const char *nativeFormat, ctx.formatFileName = NULL; ctx.autodetectSCSIControllerModel = NULL; ctx.datacenterPath = NULL; + ctx.moref = NULL; def = virVMXParseConfig(&ctx, priv->xmlopt, priv->caps, nativeConfig); @@ -2830,6 +2835,7 @@ esxConnectDomainXMLToNative(virConnectPtr conn, const char *nativeFormat, ctx.formatFileName = esxFormatVMXFileName; ctx.autodetectSCSIControllerModel = esxAutodetectSCSIControllerModel; ctx.datacenterPath = NULL; + ctx.moref = NULL; vmx = virVMXFormatConfig(&ctx, priv->xmlopt, def, virtualHW_version); @@ -3077,6 +3083,7 @@ esxDomainDefineXMLFlags(virConnectPtr conn, const char *xml, unsigned int flags) ctx.formatFileName = esxFormatVMXFileName; ctx.autodetectSCSIControllerModel = esxAutodetectSCSIControllerModel; ctx.datacenterPath = NULL; + ctx.moref = NULL; vmx = virVMXFormatConfig(&ctx, priv->xmlopt, def, virtualHW_version); diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index 8586e3ff0..77bcfd9b6 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -2390,6 +2390,21 @@ esxVI_GetVirtualMachineQuestionInfo } +int +esxVI_GetVirtualMachineMORef(esxVI_ObjectContent *virtualMachine, + char **moref) +{ + for (; virtualMachine != NULL; virtualMachine = virtualMachine->_next) { + if (virtualMachine->obj && + STREQ(virtualMachine->obj->type, "VirtualMachine") && + virtualMachine->obj->value) { + if (VIR_STRDUP(*moref, virtualMachine->obj->value) < 0) + return -1; + return 0; + } + } + return -1; +} int esxVI_GetBoolean(esxVI_ObjectContent *objectContent, const char *propertyName, diff --git a/src/esx/esx_vi.h b/src/esx/esx_vi.h index 7c53f3781..47d518dd1 100644 --- a/src/esx/esx_vi.h +++ b/src/esx/esx_vi.h @@ -334,6 +334,10 @@ int esxVI_GetVirtualMachineQuestionInfo (esxVI_ObjectContent *virtualMachine, esxVI_VirtualMachineQuestionInfo **questionInfo); +int esxVI_GetVirtualMachineMORef + (esxVI_ObjectContent *virtualMachine, + char **moref); + int esxVI_GetBoolean(esxVI_ObjectContent *objectContent, const char *propertyName, esxVI_Boolean *value, esxVI_Occurrence occurrence); diff --git a/src/vmx/vmx.c b/src/vmx/vmx.c index d1d8184c5..6f96f4cbf 100644 --- a/src/vmx/vmx.c +++ b/src/vmx/vmx.c @@ -553,23 +553,41 @@ static virDomainDefParserConfig virVMXDomainDefParserConfig = { VIR_DOMAIN_DEF_FEATURE_NAME_SLASH), }; +struct virVMXDomainDefNamespaceData { + char *datacenterPath; + char *moref; +}; + static void virVMXDomainDefNamespaceFree(void *nsdata) { - VIR_FREE(nsdata); + struct virVMXDomainDefNamespaceData *data = nsdata; + + if (data) { + VIR_FREE(data->datacenterPath); + VIR_FREE(data->moref); + } + VIR_FREE(data); } static int virVMXDomainDefNamespaceFormatXML(virBufferPtr buf, void *nsdata) { - const char *datacenterPath = nsdata; + struct virVMXDomainDefNamespaceData *data = nsdata; - if (!datacenterPath) + if (!data) return 0; - virBufferAddLit(buf, "<vmware:datacenterpath>"); - virBufferEscapeString(buf, "%s", datacenterPath); - virBufferAddLit(buf, "</vmware:datacenterpath>\n"); + if (data->datacenterPath) { + virBufferAddLit(buf, "<vmware:datacenterpath>"); + virBufferEscapeString(buf, "%s", data->datacenterPath); + virBufferAddLit(buf, "</vmware:datacenterpath>\n"); + } + if (data->moref) { + virBufferAddLit(buf, "<vmware:moref>"); + virBufferEscapeString(buf, "%s", data->moref); + virBufferAddLit(buf, "</vmware:moref>\n"); + } return 0; } @@ -1304,7 +1322,6 @@ virVMXParseConfig(virVMXContext *ctx, bool hgfs_disabled = true; long long sharedFolder_maxNum = 0; int cpumasklen; - char *namespaceData; if (ctx->parseFileName == NULL) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", @@ -1802,12 +1819,26 @@ virVMXParseConfig(virVMXContext *ctx, } /* ctx:datacenterPath -> def:namespaceData */ - if (ctx->datacenterPath) { - if (VIR_STRDUP(namespaceData, ctx->datacenterPath) < 0) + if (ctx->datacenterPath || ctx->moref) { + struct virVMXDomainDefNamespaceData *nsdata; + + if (VIR_ALLOC(nsdata) < 0) goto cleanup; + nsdata->datacenterPath = NULL; + nsdata->moref = NULL; + + if (ctx->datacenterPath) { + if (VIR_STRDUP(nsdata->datacenterPath, ctx->datacenterPath) < 0) + goto cleanup; + } + + if (ctx->moref) { + if (VIR_STRDUP(nsdata->moref, ctx->moref) < 0) + goto cleanup; + } def->ns = *virDomainXMLOptionGetNamespace(xmlopt); - def->namespaceData = namespaceData; + def->namespaceData = nsdata; } if (virDomainDefPostParse(def, caps, VIR_DOMAIN_DEF_PARSE_ABI_UPDATE, diff --git a/src/vmx/vmx.h b/src/vmx/vmx.h index 08b627765..6b3d880d6 100644 --- a/src/vmx/vmx.h +++ b/src/vmx/vmx.h @@ -45,6 +45,7 @@ typedef int (*virVMXAutodetectSCSIControllerModel)(virDomainDiskDefPtr def, * formatFileName is only used by virVMXFormatConfig. * autodetectSCSIControllerModel is optionally used by virVMXFormatConfig. * datacenterPath is only used by virVMXFormatConfig. + * moref is only used by virVMXFormatConfig. */ struct _virVMXContext { void *opaque; @@ -52,6 +53,7 @@ struct _virVMXContext { virVMXFormatFileName formatFileName; virVMXAutodetectSCSIControllerModel autodetectSCSIControllerModel; const char *datacenterPath; /* including folders */ + const char *moref; }; -- 2.13.2

2 1

[libvirt] [PATCH] qemu: handle -1 for pid in qemuDomainGetMachineName
by Nikolay Shirokovskiy 01 Sep '17

01 Sep '17

We call qemuDomainGetMachineName on domain start. On first start (after daemon start) pid is 0 and virSystemdGetMachineNameByPID don't get called. But after domain shutting down pid became -1 so on next start virSystemdGetMachineNameByPID is called and returned an error. Error is ignored so it is not critical. But at least on my system (systemd-219 with extra patches) systemd-machined is crashed on this request. This behaviour is triggered by eaf2c9f89. --- src/qemu/qemu_domain.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index e580a5b..fe44d36 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -9696,7 +9696,7 @@ qemuDomainGetMachineName(virDomainObjPtr vm) virQEMUDriverPtr driver = priv->driver; char *ret = NULL; - if (vm->pid) { + if (vm->pid > 0) { ret = virSystemdGetMachineNameByPID(vm->pid); if (!ret) virResetLastError(); -- 1.8.3.1

3 2

[libvirt] [PATCH 3/3] libvirtaio: cache the list of callbacks when calling
by Wojtek Porczyk 31 Aug '17

31 Aug '17

When the callback causes something that results in changes wrt registered handles, python aborts iteration. Relevant error message: Exception in callback None() handle: <Handle cancelled> Traceback (most recent call last): File "/usr/lib64/python3.5/asyncio/events.py", line 126, in _run self._callback(*self._args) File "/usr/lib64/python3.5/site-packages/libvirtaio.py", line 99, in _handle for callback in self.callbacks.values(): RuntimeError: dictionary changed size during iteration QubesOS/qubes-issues#2805 Signed-off-by: Wojtek Porczyk <woju(a)invisiblethingslab.com> --- libvirtaio.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libvirtaio.py b/libvirtaio.py index d161cd1..46923d8 100644 --- a/libvirtaio.py +++ b/libvirtaio.py @@ -96,7 +96,7 @@ class Descriptor(object): :param int event: The event (from libvirt's constants) being dispatched ''' - for callback in self.callbacks.values(): + for callback in list(self.callbacks.values()): if callback.event is not None and callback.event & event: callback.cb(callback.iden, self.fd, event, callback.opaque) -- 2.9.4

1 0