April 2016 - Devel - libvirt List Archives

[libvirt] [PATCH] storage: Fix regression cloning volume into a logical pool
by John Ferlan 10 May '16

10 May '16

https://bugzilla.redhat.com/show_bug.cgi?id=1318993 Commit id 'dd519a294' caused a regression cloning a volume into a logical pool by removing just the 'allocation' adjustment during storageVolCreateXMLFrom. Combined with the change to not require the new volume input XML to have a capacity listed (commit id 'e3f1d2a8') left the possibility that a copy from a larger volume into a smaller volume would create a thin/sparse logical volume. If a thin lv becomes fully populated, then LVM will set the partition 'inactive' and the subsequent fdatasync() would fail. In order to fix this in a backend agnostic manner, only adjust the new capacity if not provided. Likewise, if the new allocation is not provided, then use the capacity value as we document that if omitted, the volume will be fully allocated at time of creation. For a logical backend, that creation time is 'createVol', while for a file backend, creation doesn't set the size, but the 'createRaw' called during buildVolFrom will decide whether the file is sparse or not based on the provided capacity and allocation value. For volume clones that provide different allocation and capacity values to allow for sparse files, there is no change. Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- NOTE: Related bz's to previous changes: https://bugzilla.redhat.com/show_bug.cgi?id=1130739 https://bugzilla.redhat.com/show_bug.cgi?id=1298065 I have tested using 'virsh vol-clone' and 'virt-clone' with and without this patch applied. Both tests have the same results for the 'qemu-img info' on the cloned file. src/storage/storage_driver.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index fcc0991..4415a4c 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -2034,11 +2034,22 @@ storageVolCreateXMLFrom(virStoragePoolPtr obj, goto cleanup; } - /* Use the original volume's capacity in case the new capacity - * is less than that, or it was omitted */ - if (newvol->target.capacity < origvol->target.capacity) + /* Use the original volume's capacity if the new capacity was omitted. + * If the provided newvol capacity is less than original, we cannot just + * use the original since the subsequent createVol has "competing needs" + * for backends. A logical volume backend could then create a thin lv + * that has different characteristics when copying from the original + * volume than perhaps a raw disk file. */ + if (newvol->target.capacity == 0) newvol->target.capacity = origvol->target.capacity; + /* If the new allocation is 0, then use capacity as it's specifically + * documented "If omitted when creating a volume, the volume will be + * fully allocated at time of creation.". This is especially important + * for logical volume creation. */ + if (newvol->target.allocation == 0) + newvol->target.allocation = newvol->target.capacity; + if (!backend->buildVolFrom) { virReportError(VIR_ERR_NO_SUPPORT, "%s", _("storage pool does not support" -- 2.5.5

3 5

[libvirt] [PATCH 00/27] Introduce sparse streams
by Michal Privoznik 09 May '16

09 May '16

So, after couple of sleepless nights and headaches I'm proud to announce that finally got this working. What? Our regular streams that are can be used to transfer disk images for domains are unaware of any sparseness. Therefore they have two limitations: a) transferring big but sparse image can take ages as all the holes (interpreted by kernel as '\0') have to go through our event loop. b) resulting volume is not sparse even if the source was. How? I went by verified approach that linux kernel has. One way to look at our streams is just like read() and write() with a different names: virStreamRecv() and virStreamSend(). They even have the same prototype (if 'int fd' is substituted with 'virStreamPtr'). Now, holes in files are created and detected via third API: lseek(). Therefore I'm introducing new virStreamSkip() API that mimics the missing primitive. Now, because our streams do not necessarily have to work over files (they are for generic data transfer), I had to let users register a callback that is called whenever the other side calls virStreamSkip(). So now that we have all three primitives, we can focus on making life easier for our users. Nobody is actually using bare virStreamSend() and virStreamRecv() rather than our wrappers: virStreamSendAll() and virStreamRecvAll(). With my approach described above just virStreamSendAll() needs to be adjusted so that it's 'sparse file' aware. The virStreamRecvAll() will only get the data to write (just like it is now) with skip callback called automatically whenever needed. In order for virStreamSendAll() to skip holes I'm introducing yet another callback: virStreamInDataFunc(). This callback will help us to create a map of a file: before each virStreamSend() it checks whether we are in a data section or a hole and calls virStreamSend() or virStreamSkip() respectively. Do not worry - it will all become clear once you see the code. Now question is - how will users enable this feature? I mean, we have take into account that we might be talking to an older daemon that does not know how to skip a hole. Or vice versa - older client. The solution I came up with is to introduce flags to APIs where sparse streams make sense. I guess it makes sense for volume upload and download, but almost certainly makes no sense for virDomainOpenConsole(). Code? >From users POV they just need to pass correct argument to 'vol-upload' or 'vol-download' virsh commands. One layer down, on programming level they need to merely: st = virStreamNew(conn, 0); virStreamRegisterSkip(st, skipFunc, &fd); virStorageVolDownload(st, ...); virStreamRecvAll(st, sinkFunc, &fd); where: int skipFunc(virStreamPtr st, unsigned long long offset, void *opaque) { int *fd = opaque; return lseek(*fd, offset, SEEK_CUR) == (off_t) -1 ? -1 : 0; } And for uploading it's slightly more verbose - see patch 24. Limitations? While testing this on my machine with XFS, I've noticed that the resulting map of a transferred file is not exactly the same as the source's. Checksums are the same though. After digging deeper I found this commit in the kernel: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=0… Thing is, as we transfer the file, we are practically just seeking at EOF and thus creating holes. But if the hole size is small enough, XFS will use some speculative file allocation algorithm and eventually fully allocate the blocks even if we intended to create a hole. This does not occur when punching a hole into a file though. Well, I guess XFS devels have some reasons to do that. This behaviour has not been observed on EXT4. Notes? Oh, patches 01-03 have been ACKed already, but are not pushed yet because of the freeze. But since this feature build on the top of them, I'm sending them too. Also the whole patch set is accessible at my github: https://github.com/zippy2/libvirt/tree/sparse_streams4 Michal Privoznik (27): Revert "rpc: Fix slow volume download (virsh vol-download)" virnetclientstream: Process stream messages later virStream{Recv,Send}All: Increase client buffer Introduce virStreamSkip Introduce virStreamRegisterSkip and virStreamSkipCallback Introduce virStreamInData and virStreamRegisterInData virNetClientStreamNew: Track origin stream Track if stream is seekable RPC: Introduce virNetStreamSkip Introduce VIR_NET_STREAM_SKIP message type Teach wireshark plugin about VIR_NET_STREAM_SKIP daemon: Implement VIR_NET_STREAM_SKIP handling daemon: Introduce virNetServerProgramSendStreamSkip virnetclientstream: Introduce virNetClientStreamSendSkip virnetclientstream: Introduce virNetClientStreamHandleSkip remote_driver: Implement virStreamSkip daemonStreamHandleRead: Wire up seekable stream virNetClientStream: Wire up VIR_NET_STREAM_SKIP virStreamSendAll: Wire up sparse streams fdstream: Implement seek gendispatch: Introduce @sparseflag for our calls Introduce virStorageVol{Download,Upload}Flags virsh: Implement sparse stream to vol-download virsh: Implement sparse stream to vol-upload fdstream: Suppress use of IO helper for sparse streams daemon: Don't call virStreamInData so often storage: Enable sparse streams for virStorageVol{Download,Upload} daemon/remote.c | 2 +- daemon/stream.c | 134 +++++++++++++-- daemon/stream.h | 3 +- include/libvirt/libvirt-storage.h | 9 + include/libvirt/libvirt-stream.h | 47 ++++++ src/datatypes.h | 8 + src/driver-stream.h | 5 + src/fdstream.c | 156 +++++++++++++++--- src/fdstream.h | 3 +- src/libvirt-storage.c | 4 +- src/libvirt-stream.c | 238 ++++++++++++++++++++++++++- src/libvirt_internal.h | 7 + src/libvirt_private.syms | 2 + src/libvirt_public.syms | 7 + src/libvirt_remote.syms | 2 + src/remote/remote_driver.c | 41 ++++- src/remote/remote_protocol.x | 2 + src/rpc/gendispatch.pl | 21 ++- src/rpc/virnetclient.c | 1 + src/rpc/virnetclientstream.c | 308 ++++++++++++++++++++++++----------- src/rpc/virnetclientstream.h | 10 +- src/rpc/virnetprotocol.x | 16 +- src/rpc/virnetserverprogram.c | 33 ++++ src/rpc/virnetserverprogram.h | 7 + src/storage/storage_backend.c | 12 +- src/storage/storage_driver.c | 4 +- src/virnetprotocol-structs | 4 + tools/virsh-volume.c | 40 ++++- tools/virsh.c | 79 +++++++++ tools/virsh.h | 12 ++ tools/virsh.pod | 6 +- tools/wireshark/src/packet-libvirt.c | 40 +++++ tools/wireshark/src/packet-libvirt.h | 2 + 33 files changed, 1104 insertions(+), 161 deletions(-) -- 2.8.1

4 45

[libvirt] [PATCH v2 0/8] Add support for fetching statistics of completed jobs
by Jiri Denemark 09 May '16

09 May '16

Using virDomainGetJobStats, we can monitor running jobs but sometimes it may be useful to get statistics about a job that already finished, for example, to get the final amount of data transferred during migration or to get an idea about total downtime. This is what the following patches are about. Version 2: - changed according to John's review (see individual patches for details) Jiri Denemark (8): Refactor job statistics qemu: Avoid incrementing jobs_queued if virTimeMillisNow fails Add support for fetching statistics of completed jobs qemu: Silence coverity on optional migration stats virsh: Add support for completed job stats qemu: Transfer migration statistics to destination qemu: Recompute downtime and total time when migration completes qemu: Transfer recomputed stats back to source include/libvirt/libvirt.h.in | 11 ++ src/libvirt.c | 11 +- src/qemu/qemu_domain.c | 189 ++++++++++++++++++++++++++- src/qemu/qemu_domain.h | 32 ++++- src/qemu/qemu_driver.c | 130 ++++-------------- src/qemu/qemu_migration.c | 304 ++++++++++++++++++++++++++++++++++++------- src/qemu/qemu_monitor_json.c | 10 +- src/qemu/qemu_process.c | 9 +- tools/virsh-domain.c | 27 +++- tools/virsh.pod | 10 +- 10 files changed, 557 insertions(+), 176 deletions(-) -- 2.1.0

3 12

[libvirt] [PATCH] conf: Skip generating random MAC in DetachDevice xml
by Kothapally Madhu Pavan 06 May '16

06 May '16

When we try to detach a network device without specifying the mac address, random mac address is generated. As the generated mac address will not be available in the running vm, detaching device will fail erroring out "error: operation failed: no device matching mac address xx:xx:xx:xx:xx:xx found". Signed-off-by: Kothapally Madhu Pavan <kmp(a)linux.vnet.ibm.com> --- src/conf/domain_conf.c | 4 ++++ src/conf/domain_conf.h | 3 +++ src/libxl/libxl_driver.c | 12 ++++++------ src/lxc/lxc_driver.c | 7 ++++--- src/qemu/qemu_driver.c | 2 ++ src/uml/uml_driver.c | 6 ++++-- src/vbox/vbox_common.c | 6 ++++-- src/vz/vz_driver.c | 4 +++- src/xen/xend_internal.c | 6 ++++-- src/xen/xm_internal.c | 8 ++++---- 10 files changed, 38 insertions(+), 20 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 28248c8..512d877 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -8784,6 +8784,10 @@ virDomainNetDefParseXML(virDomainXMLOptionPtr xmlopt, (const char *)macaddr); goto error; } + } else if (flags & VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR) { + virReportError(VIR_ERR_XML_ERROR, "%s", + _("mac address not specified in the device xml")); + goto error; } else { virDomainNetGenerateMAC(xmlopt, &def->mac); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 1986f53..74692f1 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2679,6 +2679,9 @@ typedef enum { VIR_DOMAIN_DEF_PARSE_SKIP_OSTYPE_CHECKS = 1 << 8, /* allow updates in post parse callback that would break ABI otherwise */ VIR_DOMAIN_DEF_PARSE_ABI_UPDATE = 1 << 9, + /* don't generate random mac address when a network device without mac address + * is detached */ + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR = 1 << 10, } virDomainDefParseFlags; typedef enum { diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index bf97c9c..507edcf 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -3726,6 +3726,8 @@ libxlDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, virDomainDefPtr vmdef = NULL; virDomainDeviceDefPtr dev = NULL; int ret = -1; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; virCheckFlags(VIR_DOMAIN_DEVICE_MODIFY_LIVE | VIR_DOMAIN_DEVICE_MODIFY_CONFIG, -1); @@ -3743,9 +3745,8 @@ libxlDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, goto endjob; if (flags & VIR_DOMAIN_DEVICE_MODIFY_CONFIG) { - if (!(dev = virDomainDeviceDefParse(xml, vm->def, - cfg->caps, driver->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE))) + if (!(dev = virDomainDeviceDefParse(xml, vm->def, cfg->caps, + driver->xmlopt, parse_flags))) goto endjob; /* Make a copy for updated domain. */ @@ -3760,9 +3761,8 @@ libxlDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, if (flags & VIR_DOMAIN_DEVICE_MODIFY_LIVE) { /* If dev exists it was created to modify the domain config. Free it. */ virDomainDeviceDefFree(dev); - if (!(dev = virDomainDeviceDefParse(xml, vm->def, - cfg->caps, driver->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE))) + if (!(dev = virDomainDeviceDefParse(xml, vm->def, cfg->caps, + driver->xmlopt, parse_flags))) goto endjob; if (libxlDomainDetachDeviceLive(driver, vm, dev) < 0) diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c index ef48812..23f0d80 100644 --- a/src/lxc/lxc_driver.c +++ b/src/lxc/lxc_driver.c @@ -5196,6 +5196,8 @@ static int lxcDomainDetachDeviceFlags(virDomainPtr dom, virDomainDefPtr vmdef = NULL; virDomainDeviceDefPtr dev = NULL, dev_copy = NULL; int ret = -1; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; virLXCDriverConfigPtr cfg = virLXCDriverGetConfig(driver); virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | @@ -5213,9 +5215,8 @@ static int lxcDomainDetachDeviceFlags(virDomainPtr dom, if (!(caps = virLXCDriverGetCapabilities(driver, false))) goto cleanup; - dev = dev_copy = virDomainDeviceDefParse(xml, vm->def, - caps, driver->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE); + dev = dev_copy = virDomainDeviceDefParse(xml, vm->def, caps, + driver->xmlopt, parse_flags); if (dev == NULL) goto cleanup; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 862c44c..df85fd5 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -8486,6 +8486,8 @@ static int qemuDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, !(flags & VIR_DOMAIN_AFFECT_LIVE)) parse_flags |= VIR_DOMAIN_DEF_PARSE_INACTIVE; + parse_flags |= VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; + dev = dev_copy = virDomainDeviceDefParse(xml, vm->def, caps, driver->xmlopt, parse_flags); diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index 84e1df8..c232435 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -2346,6 +2346,8 @@ static int umlDomainDetachDevice(virDomainPtr dom, const char *xml) virDomainObjPtr vm; virDomainDeviceDefPtr dev = NULL; int ret = -1; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; umlDriverLock(driver); vm = virDomainObjListFindByUUID(driver->domains, dom->uuid); @@ -2366,8 +2368,8 @@ static int umlDomainDetachDevice(virDomainPtr dom, const char *xml) goto cleanup; } - dev = virDomainDeviceDefParse(xml, vm->def, driver->caps, driver->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE); + dev = virDomainDeviceDefParse(xml, vm->def, driver->caps, + driver->xmlopt, parse_flags); if (dev == NULL) goto cleanup; diff --git a/src/vbox/vbox_common.c b/src/vbox/vbox_common.c index 0cead10..4c539ef 100644 --- a/src/vbox/vbox_common.c +++ b/src/vbox/vbox_common.c @@ -4228,6 +4228,8 @@ static int vboxDomainDetachDevice(virDomainPtr dom, const char *xml) virDomainDeviceDefPtr dev = NULL; nsresult rc; int ret = -1; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; if (!data->vboxObj) return ret; @@ -4238,8 +4240,8 @@ static int vboxDomainDetachDevice(virDomainPtr dom, const char *xml) def->os.type = VIR_DOMAIN_OSTYPE_HVM; - dev = virDomainDeviceDefParse(xml, def, data->caps, data->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE); + dev = virDomainDeviceDefParse(xml, def, data->caps, + data->xmlopt, parse_flags); if (dev == NULL) goto cleanup; diff --git a/src/vz/vz_driver.c b/src/vz/vz_driver.c index ffa6f45..f5f5395 100644 --- a/src/vz/vz_driver.c +++ b/src/vz/vz_driver.c @@ -1172,6 +1172,8 @@ static int vzDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, vzConnPtr privconn = dom->conn->privateData; virDomainDeviceDefPtr dev = NULL; virDomainObjPtr privdom = NULL; + unsigned int parse_flags = VIR_DOMAIN_XML_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | VIR_DOMAIN_AFFECT_CONFIG, -1); @@ -1184,7 +1186,7 @@ static int vzDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, goto cleanup; dev = virDomainDeviceDefParse(xml, privdom->def, privconn->driver->caps, - privconn->driver->xmlopt, VIR_DOMAIN_XML_INACTIVE); + privconn->driver->xmlopt, parse_flags); if (dev == NULL) goto cleanup; diff --git a/src/xen/xend_internal.c b/src/xen/xend_internal.c index cf7cdd0..216c4f4 100644 --- a/src/xen/xend_internal.c +++ b/src/xen/xend_internal.c @@ -2327,6 +2327,8 @@ xenDaemonDetachDeviceFlags(virConnectPtr conn, int ret = -1; char *xendev = NULL; virBuffer buf = VIR_BUFFER_INITIALIZER; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | VIR_DOMAIN_AFFECT_CONFIG, -1); @@ -2354,8 +2356,8 @@ xenDaemonDetachDeviceFlags(virConnectPtr conn, NULL))) goto cleanup; - if (!(dev = virDomainDeviceDefParse(xml, def, priv->caps, priv->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE))) + if (!(dev = virDomainDeviceDefParse(xml, def, priv->caps, + priv->xmlopt, parse_flags); goto cleanup; if (virDomainXMLDevID(conn, minidef, dev, class, ref, sizeof(ref))) diff --git a/src/xen/xm_internal.c b/src/xen/xm_internal.c index ef1a460..ae142a3 100644 --- a/src/xen/xm_internal.c +++ b/src/xen/xm_internal.c @@ -1322,6 +1322,8 @@ xenXMDomainDetachDeviceFlags(virConnectPtr conn, int ret = -1; size_t i; xenUnifiedPrivatePtr priv = conn->privateData; + unsigned int parse_flags = VIR_DOMAIN_DEF_PARSE_INACTIVE | + VIR_DOMAIN_DEF_PARSE_SKIP_GENERATE_MACADDR; virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | VIR_DOMAIN_AFFECT_CONFIG, -1); @@ -1340,10 +1342,8 @@ xenXMDomainDetachDeviceFlags(virConnectPtr conn, goto cleanup; def = entry->def; - if (!(dev = virDomainDeviceDefParse(xml, entry->def, - priv->caps, - priv->xmlopt, - VIR_DOMAIN_DEF_PARSE_INACTIVE))) + if (!(dev = virDomainDeviceDefParse(xml, entry->def, priv->caps, + priv->xmlopt, parse_flags))) goto cleanup; switch (dev->type) {

2 1

[libvirt] [PATCH 0/4] rpc: Fix fd leak and potential crash
by Cole Robinson 05 May '16

05 May '16

Patch 1 was already sent, I'd like to get it in for the release, it should be quite safe. Remaining patches centralize some cleanup code and fix a potential crasher, but since it should be rare I'd like to wait till after the release. Previous thread: http://www.redhat.com/archives/libvir-list/2016-April/msg01618.html Ben Gray (1): rpc: Don't leak fd via CreateXMLWithFiles Cole Robinson (3): rpc: Add virNetMessageClearPayload rpc: Clear more in virNetMessageClearPayload rpc: use virNetMessageClearPayload in client src/libvirt_remote.syms | 1 + src/rpc/virnetclient.c | 5 +---- src/rpc/virnetmessage.c | 28 +++++++++++++++++++--------- src/rpc/virnetmessage.h | 2 ++ 4 files changed, 23 insertions(+), 13 deletions(-) -- 2.7.4

2 10

[libvirt] [PATCH v2] bhyve: implement domainShutdown
by Roman Bogorodskiy 04 May '16

04 May '16

Bhyve supports ACPI shutdown by issuing SIGTERM signal to the bhyve process. Add the bhyveDomainShutdown() function and virBhyveProcessShutdown() helper function that just sends SIGTERM to VM's bhyve process. If a guest supports ACPI shutdown then process will be terminated and this event will be noticed by the bhyve monitor code that will handle setting proper status and clean up VM's resources. Also, remove usage of virProcessKillPainfully() from domainDestroy. First, it sends SIGTERM to the process that actually triggers ACPI reset and that's not we want to do. Second, we're doing bhyvectl --destroy later and it kills bhyve process, so there's no need to manually kill it. --- Changes from v1: - Sending SIGKILL dropped completely for destroy due to reasons described in the commit message. This is mainly based on the results of this (ongoing) discussion: https://lists.freebsd.org/pipermail/freebsd-virtualization/2016-April/00434… src/bhyve/bhyve_driver.c | 27 +++++++++++++++++++++++++++ src/bhyve/bhyve_process.c | 39 +++++++++++++++++++++++++++------------ src/bhyve/bhyve_process.h | 2 ++ 3 files changed, 56 insertions(+), 12 deletions(-) diff --git a/src/bhyve/bhyve_driver.c b/src/bhyve/bhyve_driver.c index 5526bb0..4fc504e 100644 --- a/src/bhyve/bhyve_driver.c +++ b/src/bhyve/bhyve_driver.c @@ -1018,6 +1018,32 @@ bhyveDomainDestroy(virDomainPtr dom) } static int +bhyveDomainShutdown(virDomainPtr dom) +{ + virDomainObjPtr vm; + int ret = -1; + + if (!(vm = bhyveDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainShutdownEnsureACL(dom->conn, vm->def) < 0) + goto cleanup; + + if (!virDomainObjIsActive(vm)) { + virReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("Domain is not running")); + goto cleanup; + } + + ret = virBhyveProcessShutdown(vm); + + cleanup: + if (vm) + virObjectUnlock(vm); + return ret; +} + +static int bhyveDomainOpenConsole(virDomainPtr dom, const char *dev_name ATTRIBUTE_UNUSED, virStreamPtr st, @@ -1502,6 +1528,7 @@ static virHypervisorDriver bhyveHypervisorDriver = { .domainCreateWithFlags = bhyveDomainCreateWithFlags, /* 1.2.3 */ .domainCreateXML = bhyveDomainCreateXML, /* 1.2.4 */ .domainDestroy = bhyveDomainDestroy, /* 1.2.2 */ + .domainShutdown = bhyveDomainShutdown, /* 1.3.3 */ .domainLookupByUUID = bhyveDomainLookupByUUID, /* 1.2.2 */ .domainLookupByName = bhyveDomainLookupByName, /* 1.2.2 */ .domainLookupByID = bhyveDomainLookupByID, /* 1.2.3 */ diff --git a/src/bhyve/bhyve_process.c b/src/bhyve/bhyve_process.c index 14588a9..e42ed74 100644 --- a/src/bhyve/bhyve_process.c +++ b/src/bhyve/bhyve_process.c @@ -281,23 +281,15 @@ virBhyveProcessStop(bhyveConnPtr driver, if ((priv != NULL) && (priv->mon != NULL)) bhyveMonitorClose(priv->mon); - /* First, try to kill 'bhyve' process */ - if (virProcessKillPainfully(vm->pid, true) != 0) - VIR_WARN("Failed to gracefully stop bhyve VM '%s' (pid: %d)", - vm->def->name, - (int)vm->pid); - - /* Cleanup network interfaces */ - bhyveNetCleanup(vm); - - /* No matter if shutdown was successful or not, we - * need to unload the VM */ if (!(cmd = virBhyveProcessBuildDestroyCmd(driver, vm->def))) - goto cleanup; + return -1; if (virCommandRun(cmd, NULL) < 0) goto cleanup; + /* Cleanup network interfaces */ + bhyveNetCleanup(vm); + ret = 0; virCloseCallbacksUnset(driver->closeCallbacks, vm, @@ -317,6 +309,29 @@ virBhyveProcessStop(bhyveConnPtr driver, } int +virBhyveProcessShutdown(virDomainObjPtr vm) +{ + if (vm->pid <= 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Invalid PID %d for VM"), + (int)vm->pid); + return -1; + } + + /* Bhyve tries to perform ACPI shutdown when it receives + * SIGTERM signal. So we just issue SIGTERM here and rely + * on the bhyve monitor to clean things up if process disappears. + */ + if (virProcessKill(vm->pid, SIGTERM) != 0) { + VIR_WARN("Failed to terminate bhyve process for VM '%s': %s", + vm->def->name, virGetLastErrorMessage()); + return -1; + } + + return 0; +} + +int virBhyveGetDomainTotalCpuStats(virDomainObjPtr vm, unsigned long long *cpustats) { diff --git a/src/bhyve/bhyve_process.h b/src/bhyve/bhyve_process.h index cfa80af..ebabe17 100644 --- a/src/bhyve/bhyve_process.h +++ b/src/bhyve/bhyve_process.h @@ -34,6 +34,8 @@ int virBhyveProcessStop(bhyveConnPtr driver, virDomainObjPtr vm, virDomainShutoffReason reason); +int virBhyveProcessShutdown(virDomainObjPtr vm); + int virBhyveGetDomainTotalCpuStats(virDomainObjPtr vm, unsigned long long *cpustats); -- 2.7.4

2 3

[libvirt] [PATCH 0/3] Admin API: Add support for client disconnect
by Erik Skultety 04 May '16

04 May '16

This series adds a simple support for a client force disconnect. As with the series about client identity retrieval, this series is rebased on listing clients due to some gendispatch stuff needed in all of them. To test it, you checkout my remote branch https://github.com/eskultety/libvirt/tree/list-clients-info-disconnect Erik Skultety (3): admin: Introduce virAdmServerLookupClient admin: Introduce virAdmClientClose API virt-admin: Introduce client-disconnect command daemon/admin.c | 6 ++++ daemon/admin_server.c | 19 +++++++++++ daemon/admin_server.h | 7 ++++ include/libvirt/libvirt-admin.h | 7 ++++ include/libvirt/virterror.h | 1 + src/admin/admin_protocol.x | 27 +++++++++++++++- src/admin/admin_remote.c | 10 ++++++ src/admin_protocol-structs | 14 ++++++++ src/libvirt-admin.c | 66 +++++++++++++++++++++++++++++++++++++ src/libvirt_admin_private.syms | 3 ++ src/libvirt_admin_public.syms | 2 ++ src/rpc/virnetserver.c | 23 +++++++++++++ src/rpc/virnetserver.h | 3 ++ src/util/virerror.c | 6 ++++ tools/virt-admin.c | 72 +++++++++++++++++++++++++++++++++++++++++ tools/virt-admin.pod | 16 +++++++++ 16 files changed, 281 insertions(+), 1 deletion(-) -- 2.4.11

2 4

[libvirt] RFC filesystem pool proposal description.
by Olga Krishtal 04 May '16

04 May '16

In this proposal description we would like to introduce a separate pool type:fspool. Fspool provides and manages filesystems. Below are listed possible examples of fspool, depending on source type: volume, directory or network. Volume fspool uses storage pool volume to reside upon. <fspool type = 'volume'> <name>fspool</name> <uuid>cb07b799-8e03-485c-aaa0-7c88f60b37b0</uuid> <allocation>0</allocation> <capacity>0<capacity> <source> <pool ='pool name'/> <volume='volume name'/> <format type = 'ext4'/> (ext4, zfs - depending on type of volume that is used as a backend). <source/> <target> <path>/path/to/fspool</path> <permissions> <mode>0777</mode> <owner>0</owner> <group>0</group> </permissions> </target> </fspool> Directory fspool represents a host directory: <fspool type = 'dir'> <name>fspool</name> <uuid>cb07b799-8e03-485c-aaa0-7c88f60b37b0</uuid> <target> <path>/path/to/fspool/root</path> <permissions> <mode>0777</mode> <owner>0</owner> <group>0</group> </permissions> </target> </fspool> Network fspool is similar to previous one, except it is remote: <fspool type='network'> <name>fspool</name> <uuid>b27a72b5-a92e-4e58-a172-a80ba73b30fe</uuid> <source> <host name='X.X.X.X'/> - if the pool is not on the host <dir path='some/directory/exposed/to/our/host'/> </source> <target> <path>/fspool</path> </target> </fspool> basic API for fspool management: fspool-define fspool-build fspool-start fspool-stop fspool-destroy fspool-undefine Fspools are divided into fs items. Single item is a directory inside the fspool root directory. Fs item example is listed below: <fs> <name>item</name> <allocation>0</allocation> <capacity>0<capacity> <source>fspool='name'</source> <target> <path>/fspool</path> </target> <permissions> <mode>0777</mode> <owner>0</owner> <group>0</group> <cgroups> Limits</cgroups> </permissions> </fs> - allocation/capacity - is omitted when fspools type is directory. - target path is a path to fs item from the fspool root. If it is not present then the content of item is placed directly in fspool root. - cgroups can be used in order to manipulate space usage politics. Fs items will be manipulated via fs API: fs-define fs-delete fs-list In addition to containers use-cases, fspools are also useful for libvirt itself - they will help store and easily manage some internal metadata, such as snapshots or internal connection driver information (logs, domain configurations, etc). Please, take a look and share your opinion. Olga

4 7

[libvirt] [PATCH 0/7] Admin API: Add support for client identity info retrieval
by Erik Skultety 04 May '16

04 May '16

This series adds support for client identity retrieval, i.e. information like remote IP (if connected remotely), uid,gid,pid, as well as username if connected locally and also information regarding authentication (if used). The series is rebased on the listing clients series, because it relies on the gendispatch stuff, so for testing purposes checkout my remote branch https://github.com/eskultety/libvirt/tree/list-clients-info-disconnect which also covers the next series about client disconnect. Erik Skultety (7): admin: Introduce virAdmServerLookupClient admin: include: Introduce some client's identity related typed params macros virnetsocket: Provide socket address format in a more standard form virneserverclient: Introduce virNetServerClientHasSASLSession virnetserverclient: Add an internal method to retrieve client's identity admin: Introduce virAdmClientGetInfo API virt-admin: Introduce command client-info daemon/admin.c | 59 ++++++++++++++++++ daemon/admin_server.c | 102 +++++++++++++++++++++++++++++++ daemon/admin_server.h | 9 +++ daemon/remote.c | 13 +++- include/libvirt/libvirt-admin.h | 130 ++++++++++++++++++++++++++++++++++++++++ include/libvirt/virterror.h | 1 + src/admin/admin_protocol.x | 34 ++++++++++- src/admin/admin_remote.c | 47 +++++++++++++++ src/admin_protocol-structs | 20 +++++++ src/libvirt-admin.c | 75 +++++++++++++++++++++++ src/libvirt_admin_private.syms | 4 ++ src/libvirt_admin_public.syms | 2 + src/remote/remote_driver.c | 7 +++ src/rpc/virnetclient.c | 10 ++++ src/rpc/virnetclient.h | 2 + src/rpc/virnetserver.c | 23 +++++++ src/rpc/virnetserver.h | 3 + src/rpc/virnetserverclient.c | 48 +++++++++++++++ src/rpc/virnetserverclient.h | 6 ++ src/rpc/virnetsocket.c | 17 +++++- src/rpc/virnetsocket.h | 2 + src/util/virerror.c | 6 ++ src/util/virsocketaddr.c | 24 ++++++-- tests/virnetsockettest.c | 10 ++-- tools/virt-admin.c | 91 ++++++++++++++++++++++++++++ 25 files changed, 731 insertions(+), 14 deletions(-) -- 2.4.11

3 15

[libvirt] [PATCH v2] send default USB controller in xml to destination during migration
by Shivaprasad G Bhat 04 May '16

04 May '16

The default USB controller is not sent to destination as the older versions of libvirt(0.9.4 or earlier as I see in commit log of 409b5f54) didn't support them. For some archs where the support started much later can safely send the USB controllers without this worry. So, send the controller to destination for all archs except x86. Moreover this is not very applicable to x86 as the USB controller has model ich9_ehci1 on q35 and for pc-i440fx, there cant be any slots before USB as it is fixed on slot 1. The patch fixes a bug that, if the USB controller happens to occupy a slot after disks/interfaces and one of them is hot-unplugged, then the default USB controller added on destination takes the smallest slot number and that would lead to savestate mismatch and migration failure. Seen and verified on PPC64. Signed-off-by: Shivaprasad G Bhat <sbhat(a)linux.vnet.ibm.com> --- src/qemu/qemu_domain.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 6262bfe..963ff35 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -2599,7 +2599,11 @@ qemuDomainDefFormatBuf(virQEMUDriverPtr driver, usb = def->controllers[i]; } } - if (usb && usb->idx == 0 && usb->model == -1) { + /* The original purpose of the check was the migration compatibility + * with libvirt <= 0.9.4. Limitation doesn't apply to other archs + * and can cause problems on PPC64. + */ + if (ARCH_IS_X86(def->os.arch) && usb && usb->idx == 0 && usb->model == -1) { VIR_DEBUG("Removing default USB controller from domain '%s'" " for migration compatibility", def->name); toremove++;

5 5