March 2011 - Devel - libvirt List Archives

[libvirt] [PATCH v2 0/4] configure inactive domains' maximum memory size
by Taku Izumi 07 Apr '11

07 Apr '11

Hi all, This patchset enables us to configure inactive domain's maximum memory size. I redesigned according to Daniel-san's commnet. *[PATCH 1/4] [RESEND] setmaxmem: remove the code to invoke virDomainSetMemory in cmdSetmaxmem => http://www.redhat.com/archives/libvir-list/2011-March/msg00747.html *[PATCH 2/4] maxmem: introduces VIR_DOMAIN_MEM_MAXIMUM flag *[PATCH 3/4] maxmem: implement virDomainSetMaxMemory API of the qemu driver *[PATCH 4/4] setmaxmem: add the new options to "virsh setmaxmem" command The following patchset is the prerequisite of these. => http://www.redhat.com/archives/libvir-list/2011-March/msg01057.html Best regards, Taku Izumi

2 7

[libvirt] [PATCH] setmaxmem: remove the code to invoke virDomainSetMemory in cmdSetmaxmem
by Taku Izumi 07 Apr '11

07 Apr '11

When the new maximum memory size becomes less than the current memory size, I think it is not the libvirt client but the each driver that decides the behavior (reject the operation or shrink the current memory size). Signed-off-by: Taku Izumi <izumi.taku(a)jp.fujitsu.com> --- tools/virsh.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) Index: libvirt/tools/virsh.c =================================================================== --- libvirt.orig/tools/virsh.c +++ libvirt/tools/virsh.c @@ -3030,15 +3030,7 @@ cmdSetmaxmem(vshControl *ctl, const vshC if (virDomainSetMaxMemory(dom, kilobytes) != 0) { vshError(ctl, "%s", _("Unable to change MaxMemorySize")); - virDomainFree(dom); - return FALSE; - } - - if (kilobytes < info.memory) { - if (virDomainSetMemory(dom, kilobytes) != 0) { - vshError(ctl, "%s", _("Unable to shrink current MemorySize")); - ret = FALSE; - } + ret = FALSE; } virDomainFree(dom);

3 2

[libvirt] [BUG] Managed save qemu state gets deleted after failed resume
by Philipp Hahn 07 Apr '11

07 Apr '11

Hello, I haven't had time to provide a fix, but still want you to inform you about a bug: If resuming a saved VM fails with Qemu-0.14, the managed save state file /var/lib/libvirt/qemu/save/$VM.save is still deleted. I think it would be better to only delete the state after an successful resume. /var/log/libvirt/qemu/ucs2.3-0_basis_amd64.log: 2011-03-30 09:37:56.960: starting up LC_ALL=C PATH=/sbin:/bin:/usr/sbin:/usr/bin QEMU_AUDIO_DRV=none /usr/bin/kvm -S -M pc-0.14 -enable-kvm -m 2048 -smp 2,sockets=2,cores=1,threads=1 -name ucs2.3-0_basis_amd64 -uuid 656957aa-13a0-6922-5d08-3a39561f9775 -nodefconfig -nodefaults -chardev socket,id=monitor,path=/var/lib/libvirt/qemu/ucs2.3-0_basis_amd64.monitor,server,nowait -mon chardev=monitor,mode=readline -rtc base=utc -boot cnd -drive file=/var/lib/libvirt/images/ucs_2.3-0-091215-dvd-amd64.iso,if=none,media=cdrom,id=drive-ide0-0-0,readonly=on,format=raw -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -drive file=/var/lib/libvirt/images/ucs230basis.qcow2,if=none,id=drive-ide0-0-1,format=qcow2 -device ide-drive,bus=ide.0,unit=1,drive=drive-ide0-0-1,id=ide0-0-1 -netdev tap,fd=24,id=hostnet0 -device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:12:7c:03,bus=pci.0,addr=0x3 -usb -device usb-tablet,id=input0 -vnc 0.0.0.0:5 -k de -vga cirrus -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x4 -option-rom /usr/share/kvm/pxe-rtl8139.bin Failed to allocate 2147483648 B: Cannot allocate memory /var/log/univention/virtual-machine-manager-daemon-errors.log: libvir: QEMU error : operation failed: migration to 'exec:cat | { dd bs=4096 seek=1 if=/dev/null && dd bs=1048576; } 1<>'/var/lib/libvirt/qemu/save/ucs2.3-0_basis_amd64.save'' failed: migration failed libvir: QEMU error : operation failed: failed to retrieve chardev info in qemu with 'info chardev' This issue is tracked in our (German) bug-tracker at <https://forge.univention.org/bugzilla/show_bug.cgi?id=22021> Sincerely Philipp Hahn -- Philipp Hahn Open Source Software Engineer hahn(a)univention.de Univention GmbH Linux for Your Business fon: +49 421 22 232- 0 Mary-Somerville-Str.1 D-28359 Bremen fax: +49 421 22 232-99 http://www.univention.de/

2 2

[libvirt] [PATCH 0/6] v3: Allow data upload/download to/from storage volumes
by Daniel P. Berrange 07 Apr '11

07 Apr '11

An update of http://www.redhat.com/archives/libvir-list/2011-March/msg00880.html Addressing all the feedback from v2 and this time with the correct set of patches!

4 15

[libvirt] [PATCH 0/3] reattach pci devices when qemuPrepareHostdevPCIDevices() failed
by Wen Congyang 07 Apr '11

07 Apr '11

When I try to attach a pci device to guest OS, libvirtd resets the pci failed. But I can not use this pci device after attaching pci device failed. I must rollback the operation by hand. I think libvirtd should auto rollback the operation when it failed. Wen Congyang (3): remove devices from driver->activePciHostdevs when qemuPrepareHostdevPCIDevices() failed reattach pci device when pciBindDeviceToStub() failed reattach pci devices when qemuPrepareHostdevPCIDevices() failed src/qemu/qemu_hostdev.c | 33 ++++++++++++++--- src/util/pci.c | 90 +++++++++++++++++++++++++++++++++++++++++----- 2 files changed, 107 insertions(+), 16 deletions(-)

2 12

[libvirt] [libvirt-snmp][PATCH v4 0/4] Add SNMP trap/notification support
by Michal Privoznik 06 Apr '11

06 Apr '11

The fourth version. In case we build against libvirt older than 0.9.0 (which have old event API), we need to include some sources from libvirt git. So we are able to run even on older RHELs, like 5.6. Those files were copied to our git and modified very slighty. Finally, we have a proof-of-concept implementation of SNMP trap. Two new files were first generated using mib2c then edited, so be careful when re-generating them. Destination of traps is defined in snmpd.conf (trap2sink), and to receive them snmptrapd must be configured and up. The idea behind is - create a thread in which we poll for domain events. Once we receive event notification, we just fire up trap sending. Michal Privoznik (4): Add notification-type object to libvirt MIB Add libvirt error handling function Create functions to fill in and send notification packets. Add SNMP trap/notification support. INSTALL.1st | 9 +- configure.ac | 17 +- libvirt-snmp.spec.in | 7 +- src/LIBVIRT-MIB.txt | 9 + src/Makefile.am | 27 ++- src/README.txt | 9 +- src/event.c | 193 ++++++++++++ src/event.h | 101 ++++++ src/event_poll.c | 724 ++++++++++++++++++++++++++++++++++++++++++++ src/event_poll.h | 132 ++++++++ src/ignore-value.h | 64 ++++ src/internal.h | 265 ++++++++++++++++ src/libvirtNotifications.c | 122 ++++++++ src/libvirtNotifications.h | 33 ++ src/libvirtSnmp.c | 136 ++++++++- src/libvirtSnmpError.c | 34 ++ src/libvirtSnmpError.h | 31 ++ src/memory.c | 313 +++++++++++++++++++ src/memory.h | 212 +++++++++++++ src/threads.c | 251 +++++++++++++++ src/threads.h | 101 ++++++ src/util.c | 105 +++++++ src/util.h | 38 +++ 23 files changed, 2919 insertions(+), 14 deletions(-) create mode 100644 src/event.c create mode 100644 src/event.h create mode 100644 src/event_poll.c create mode 100644 src/event_poll.h create mode 100644 src/ignore-value.h create mode 100644 src/internal.h create mode 100644 src/libvirtNotifications.c create mode 100644 src/libvirtNotifications.h create mode 100644 src/libvirtSnmpError.c create mode 100644 src/libvirtSnmpError.h create mode 100644 src/memory.c create mode 100644 src/memory.h create mode 100644 src/threads.c create mode 100644 src/threads.h create mode 100644 src/util.c create mode 100644 src/util.h -- 1.7.4

3 6

[libvirt] [PATCH] qemu: unlock qemu driver before return
by Hu Tao 06 Apr '11

06 Apr '11

--- src/qemu/qemu_driver.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 104e92d..6e3edde 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2068,12 +2068,14 @@ static int qemudDomainSave(virDomainPtr dom, const char *path) qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("Invalid save image format specified " "in configuration file")); + qemuDriverUnlock(driver); return -1; } if (!qemudCompressProgramAvailable(compressed)) { qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("Compression program for image format " "in configuration file isn't available")); + qemuDriverUnlock(driver); return -1; } } -- 1.7.3.1

3 11

[libvirt] [PATCH 0/2] Log enhancement
by Naoya Horiguchi 06 Apr '11

06 Apr '11

Hi, Following patches improve logging feature of libvirt. Please see individual patches for more details. [PATCH 1/2] extend logging to record configuration-related events on host machine [PATCH 2/2] qemu: extend logging to record guest configuration events Thanks, Naoya Horiguchi

3 5

[libvirt] [PATCH] Allow relative path for qemu backing file
by Jesse Cook 05 Apr '11

05 Apr '11

This patch enables the relative backing file path support provided by qemu-img create. If a relative path is specified for the backing file, it is converted to an absolute path using the storage pool path. The absolute path is used to verify that the backing file exists. If the backing file exists, the relative path is allowed and will be provided to qemu-img create. This patch takes the place of a previous patch: http://www.redhat.com/archives/libvir-list/2011-March/msg00179.html --- src/storage/storage_backend.c | 18 +++++++++++++++++- 1 files changed, 17 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 2eede74..abdc2dd 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -666,6 +666,9 @@ virStorageBackendCreateQemuImg(virConnectPtr conn, } if (vol->backingStore.path) { + int accessRetCode = -1; + + char *absolutePath = NULL; /* XXX: Not strictly required: qemu-img has an option a different * backing store, not really sure what use it serves though, and it @@ -686,7 +689,20 @@ virStorageBackendCreateQemuImg(virConnectPtr conn, vol->backingStore.format); return -1; } - if (access(vol->backingStore.path, R_OK) != 0) { + + /* Convert relative backing store paths to absolute paths for access + * validation. + */ + if ('/' != *(vol->backingStore.path)) { + virAsprintf(&absolutePath, "%s/%s", pool->def->target.path, + vol->backingStore.path); + + } else { + virAsprintf(&absolutePath, "%s", vol->backingStore.path); + } + accessRetCode = access(absolutePath, R_OK); + VIR_FREE(absolutePath); + if (accessRetCode != 0) { virReportSystemError(errno, _("inaccessible backing store volume %s"), vol->backingStore.path); -- 1.7.4.1

3 4

[libvirt] [PATCH] virsh: Increase device-detach intelligence
by Michal Privoznik 05 Apr '11

05 Apr '11

Up to now users have to give a full XML description on input when device-detaching. If they omited something it lead to unclear error messages (like generated MAC wasn't found, etc.). With this patch users can specify only those information which specify one device sufficiently precise. Remaining information is completed from domain. --- tools/virsh.c | 260 +++++++++++++++++++++++++++++++++++++++++++++++++++++---- 1 files changed, 245 insertions(+), 15 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 50ca50f..95d27f7 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -8702,6 +8702,224 @@ cmdAttachDevice(vshControl *ctl, const vshCmd *cmd) return TRUE; } +/** + * Check if n1 is superset of n2, meaning n1 contains all elements and + * attributes as n2 at lest. Including children. + * @n1 first node + * @n2 second node + * return 1 in case n1 covers n2, 0 otherwise. + */ +static int +vshNodeIsSuperset(xmlNodePtr n1, xmlNodePtr n2) { + xmlNodePtr child1, child2; + xmlAttrPtr attr1, attr2; + int found; + + if (!n1 && !n2) + return 1; + + if (!n1 || !n2) + return 0; + + if (!xmlStrEqual(n1->name, n2->name)) + return 0; + + attr2 = n2->properties; + while (attr2) { + if (attr2->type == XML_ATTRIBUTE_NODE) { + attr1 = n1->properties; + found = 0; + while (attr1) { + if (xmlStrEqual(attr1->name, attr2->name)) { + found = 1; + break; + } + attr1 = attr1->next; + } + if (!found) + return 0; + if (!xmlStrEqual(BAD_CAST virXMLPropString(n1, (const char *) attr1->name), + BAD_CAST virXMLPropString(n2, (const char *) attr2->name))) + return 0; + } + attr2 = attr2->next; + } + + child2 = n2->children; + while (child2) { + if (child2->type == XML_ELEMENT_NODE) { + child1 = n1->children; + found = 0; + while (child1) { + if (child1->type == XML_ELEMENT_NODE && + xmlStrEqual(child1->name, child2->name)) { + found = 1; + break; + } + child1 = child1->next; + } + if (!found) + return 0; + if (!vshNodeIsSuperset(child1, child2)) + return 0; + } + child2 = child2->next; + } + + return 1; +} + +/** + * To given domain and (probably incomplete) device XML specification try to + * find such device in domain and complete missing parts. This is however + * possible when given device XML is sufficiently precise so it addresses only + * one device. + * @ctl vshControl for error messages printing + * @dom domain + * @oldXML device XML before + * @newXML and after completion + * Returns -2 when no such device exists in domain, -3 when given XML selects many + * (is too ambiguous), 0 in case of success. Otherwise returns -1. @newXML + * is touched only in case of success. + */ +static int +vshCompleteXMLFromDomain(vshControl *ctl, virDomainPtr dom, char *oldXML, + char **newXML) { + int funcRet = -1; + char *domXML = NULL; + xmlDocPtr domDoc = NULL, devDoc = NULL; + xmlNodePtr node = NULL; + xmlXPathContextPtr domCtxt = NULL, devCtxt = NULL; + xmlNodePtr *devices = NULL; + xmlSaveCtxtPtr sctxt = NULL; + int devices_size; + char *xpath; + xmlBufferPtr buf = NULL; + + if (!(domXML = virDomainGetXMLDesc(dom, 0))) { + vshError(ctl, _("couldn't get XML description of domain %s"), + virDomainGetName(dom)); + goto error; + } + + if (!(domDoc = xmlReadDoc(BAD_CAST domXML, "domain.xml", NULL, + XML_PARSE_NOENT | XML_PARSE_NONET | + XML_PARSE_NOERROR | XML_PARSE_NOWARNING))) { + vshError(ctl, "%s", _("could not parse domain XML")); + goto error; + } + + if (!(devDoc = xmlReadDoc(BAD_CAST oldXML, "device.xml", NULL, + XML_PARSE_NOENT | XML_PARSE_NONET | + XML_PARSE_NOERROR | XML_PARSE_NOWARNING))) { + vshError(ctl, "%s", _("could not parse device XML")); + goto error; + } + + node = xmlDocGetRootElement(domDoc); + if (!node) { + vshError(ctl, "%s", _("failed to get domain root element")); + goto error; + } + + domCtxt = xmlXPathNewContext(domDoc); + if (!domCtxt) { + vshError(ctl, "%s", _("failed to create context")); + goto error; + } + domCtxt->node = node; + + node = xmlDocGetRootElement(devDoc); + if (!node) { + vshError(ctl, "%s", _("failed to get device root element")); + goto error; + } + + devCtxt = xmlXPathNewContext(devDoc); + if (!devCtxt) { + vshError(ctl, "%s", _("failed to create context")); + goto error; + } + devCtxt->node = node; + + buf = xmlBufferCreate(); + if (!buf) { + vshError(ctl, "%s", _("out of memory")); + goto error; + } + + xmlBufferCat(buf, BAD_CAST "/domain/devices/"); + xmlBufferCat(buf, node->name); + xpath = (char *) xmlBufferContent(buf); + /* we get all possible devices */ + devices_size = virXPathNodeSet(xpath, domCtxt, &devices); + xmlBufferEmpty(buf); + + if (devices_size < 0) { + /* error */ + vshError(ctl, "%s", _("error when selecting nodes")); + goto error; + } else if (devices_size < 1) { + /* no such device */ + funcRet = -2; + goto error; + } + + /* and refine */ + int i = 0; + while (i < devices_size) { + if (!vshNodeIsSuperset(devices[i], node)) { + if (devices_size == 1) { + VIR_FREE(devices); + devices_size = 0; + } else { + memmove(devices + i, devices + i + 1, + sizeof(*devices) * (devices_size-i-1)); + devices_size--; + if (VIR_REALLOC_N(devices, devices_size) < 0) { + /* ignore, harmless */ + } + } + } else { + i++; + } + } + + if (!devices_size) { + /* no such device */ + funcRet = -2; + goto error; + } else if (devices_size > 1) { + /* ambiguous */ + funcRet = -3; + goto error; + } + + if (newXML) { + sctxt = xmlSaveToBuffer(buf, NULL, 0); + if (!sctxt) { + vshError(ctl, "%s", _("failed to create document saving context")); + goto error; + } + + xmlSaveTree(sctxt, devices[0]); + xmlSaveClose(sctxt); + *newXML = (char *) xmlBufferContent(buf); + buf->content = NULL; + } + + funcRet = 0; + +error: + xmlBufferFree(buf); + VIR_FREE(devices); + xmlXPathFreeContext(devCtxt); + xmlXPathFreeContext(domCtxt); + xmlFreeDoc(devDoc); + xmlFreeDoc(domDoc); + VIR_FREE(domXML); + return funcRet; +} /* * "detach-device" command @@ -8724,7 +8942,7 @@ cmdDetachDevice(vshControl *ctl, const vshCmd *cmd) { virDomainPtr dom; const char *from = NULL; - char *buffer; + char *buffer = NULL, *new_buffer = NULL; int ret = TRUE; unsigned int flags; @@ -8734,37 +8952,49 @@ cmdDetachDevice(vshControl *ctl, const vshCmd *cmd) if (!(dom = vshCommandOptDomain(ctl, cmd, NULL))) return FALSE; - if (vshCommandOptString(cmd, "file", &from) <= 0) { - virDomainFree(dom); - return FALSE; - } + if (vshCommandOptString(cmd, "file", &from) <= 0) + goto error; if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) { virshReportError(ctl); - virDomainFree(dom); - return FALSE; + goto error; + } + + ret = vshCompleteXMLFromDomain(ctl, dom, buffer, &new_buffer); + if (ret < 0) { + if (ret == -2) { + vshError(ctl, _("no such device in %s"), virDomainGetName(dom)); + } else if (ret == -3) { + vshError(ctl, "%s", _("given XML selects too many devices. " + "Please, be more specific")); + } + ret = FALSE; + goto error; } if (vshCommandOptBool(cmd, "persistent")) { flags = VIR_DOMAIN_DEVICE_MODIFY_CONFIG; if (virDomainIsActive(dom) == 1) flags |= VIR_DOMAIN_DEVICE_MODIFY_LIVE; - ret = virDomainDetachDeviceFlags(dom, buffer, flags); + ret = virDomainDetachDeviceFlags(dom, new_buffer, flags); } else { - ret = virDomainDetachDevice(dom, buffer); + ret = virDomainDetachDevice(dom, new_buffer); } - VIR_FREE(buffer); if (ret < 0) { vshError(ctl, _("Failed to detach device from %s"), from); - virDomainFree(dom); - return FALSE; - } else { - vshPrint(ctl, "%s", _("Device detached successfully\n")); + ret = FALSE; + goto error; } + vshPrint(ctl, "%s", _("Device detached successfully\n")); + ret = TRUE; + +error: + VIR_FREE(new_buffer); + VIR_FREE(buffer); virDomainFree(dom); - return TRUE; + return ret; } -- 1.7.4

3 2