May 2012 - Devel - libvirt List Archives

[libvirt] [Patch] Libvirt - Fix locking for readonly devices
by David Weber 08 May '12

08 May '12

Hi, I'm currently working on getting sanlock into Debian/Ubuntu. While testing, I noticed that I wasn't able to add a readonly or shared device: "internal error unsupported configuration: Readonly leases are not supported". After looking into the source, it seems to be the following situation: - Libvirt passes every device to the sanlock plugin, even if it is readonly - The sanlock plugin rejects to add a lease for the readonly device, returning an error so the machine starts to fail The attached patch rejects passing readonly and shared devices to the lock-plugin so they shouldn't be a problem anymore. Would this be an appropriated way to fix this problem? (I'm neither experienced in libvirt nor C) What seems odd to me is that Fedora either doesn't hit this or nobody has ever reported it. I'm not member of this list so please CC me. Thanks! Cheers, David ----------------------------------------- diff --git a/src/locking/domain_lock.c b/src/locking/domain_lock.c index de1937c..45617e6 100644 --- a/src/locking/domain_lock.c +++ b/src/locking/domain_lock.c @@ -79,9 +79,9 @@ static int virDomainLockManagerAddDisk(virLockManagerPtr lock, return 0; if (disk->readonly) - diskFlags |= VIR_LOCK_MANAGER_RESOURCE_READONLY; + return 0; if (disk->shared) - diskFlags |= VIR_LOCK_MANAGER_RESOURCE_SHARED; + return 0; VIR_DEBUG("Add disk %s", disk->src); if (virLockManagerAddResource(lock,

1 0

[libvirt] [PATCH] numad: Set memory policy according to the advisory nodeset from numad
by Osier Yang 08 May '12

08 May '12

Though numad will manage the memory allocation of task dynamically, but it wants management application (libvirt) to pre-set the memory policy according to the advisory nodeset returned from querying numad, (just like pre-bind CPU nodeset for domain process), and thus the performance could benifit much more from it. This patch introduces new XML tag 'placement', value 'auto' indicates whether to set the memory policy with the advisory nodeset from numad, and its value defaults to 'static'. e.g. <numatune> <memory placement='auto' mode='interleave'/> </numatune> Just like what current "numatune" does, the 'auto' numa memory policy setting uses libnuma's API too. So, to full drive numad, one needs to specify placement='auto' for both "<vcpu>" and "<numatune><memory .../></numatune>". It's a bit inconvenient, but makes sense from semantics' point of view. --- An alternative way is to not introduce the new XML tag, and pre-set the memory policy implicitly with "<vcpu placement='auto'>4</vcpu>", but IMHO it implies too much, and I'd not like go this way unless the new XML tag is not accepted. --- docs/formatdomain.html.in | 11 ++- docs/schemas/domaincommon.rng | 39 +++++++--- libvirt.spec.in | 1 + src/conf/domain_conf.c | 96 ++++++++++++++++-------- src/conf/domain_conf.h | 9 ++ src/libvirt_private.syms | 2 + src/qemu/qemu_process.c | 79 +++++++++++-------- tests/qemuxml2argvdata/qemuxml2argv-numad.args | 4 + tests/qemuxml2argvdata/qemuxml2argv-numad.xml | 31 ++++++++ tests/qemuxml2argvtest.c | 1 + 10 files changed, 194 insertions(+), 79 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-numad.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-numad.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index e1fe0c4..01b3124 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -580,9 +580,14 @@ The optional <code>memory</code> element specifies how to allocate memory for the domain process on a NUMA host. It contains two attributes, attribute <code>mode</code> is either 'interleave', 'strict', - or 'preferred', - attribute <code>nodeset</code> specifies the NUMA nodes, it leads same - syntax with attribute <code>cpuset</code> of element <code>vcpu</code>. + or 'preferred', attribute <code>nodeset</code> specifies the NUMA nodes, + it leads same syntax with attribute <code>cpuset</code> of element + <code>vcpu</code>, the optional attribute <code>placement</code> can be + used to indicate the memory placement mode for domain process, its value + can be either "static" or "auto", defaults to "static". "auto" indicates + the domain process will only allocate memory from the advisory nodeset + returned from querying numad, and the value of attribute <code>nodeset</code> + will be ignored if it's specified. Since 0.9.3 </dd> </dl> diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 8419ccc..9b509f1 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -562,16 +562,35 @@ <element name="numatune"> <optional> <element name="memory"> - <attribute name="mode"> - <choice> - <value>strict</value> - <value>preferred</value> - <value>interleave</value> - </choice> - </attribute> - <attribute name="nodeset"> - <ref name="cpuset"/> - </attribute> + <choice> + <group> + <attribute name="mode"> + <choice> + <value>strict</value> + <value>preferred</value> + <value>interleave</value> + </choice> + </attribute> + <attribute name='placement'> + <choice> + <value>static</value> + <value>auto</value> + </choice> + </attribute> + </group> + <group> + <attribute name="mode"> + <choice> + <value>strict</value> + <value>preferred</value> + <value>interleave</value> + </choice> + </attribute> + <attribute name="nodeset"> + <ref name="cpuset"/> + </attribute> + </group> + </choice> </element> </optional> </element> diff --git a/libvirt.spec.in b/libvirt.spec.in index e7e0a55..0b119c2 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -454,6 +454,7 @@ BuildRequires: scrub %if %{with_numad} BuildRequires: numad +BuildRequires: numactl-devel %endif %description diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 3fce7e5..b728cb6 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -640,6 +640,11 @@ VIR_ENUM_IMPL(virDomainDiskTray, VIR_DOMAIN_DISK_TRAY_LAST, "closed", "open"); +VIR_ENUM_IMPL(virDomainNumatuneMemPlacementMode, + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_LAST, + "static", + "auto"); + #define virDomainReportError(code, ...) \ virReportErrorHelper(VIR_FROM_DOMAIN, code, __FILE__, \ __FUNCTION__, __LINE__, __VA_ARGS__) @@ -8023,30 +8028,22 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, while (cur != NULL) { if (cur->type == XML_ELEMENT_NODE) { if (xmlStrEqual(cur->name, BAD_CAST "memory")) { - tmp = virXMLPropString(cur, "nodeset"); - + tmp = virXMLPropString(cur, "placement"); + int placement_mode; if (tmp) { - char *set = tmp; - int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; - - if (VIR_ALLOC_N(def->numatune.memory.nodemask, - nodemasklen) < 0) { - virReportOOMError(); + if ((placement_mode = + virDomainNumatuneMemPlacementModeTypeFromString(tmp)) < 0) { + virDomainReportError(VIR_ERR_XML_ERROR, + _("Unsupported memory placement " + "mode '%s'"), tmp); + VIR_FREE(tmp); goto error; } - - /* "nodeset" leads same syntax with "cpuset". */ - if (virDomainCpuSetParse(set, 0, - def->numatune.memory.nodemask, - nodemasklen) < 0) - goto error; VIR_FREE(tmp); } else { - virDomainReportError(VIR_ERR_XML_ERROR, "%s", - _("nodeset for NUMA memory " - "tuning must be set")); - goto error; + placement_mode = VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC; } + def->numatune.memory.placement_mode = placement_mode; tmp = virXMLPropString(cur, "mode"); if (tmp) { @@ -8055,13 +8052,40 @@ static virDomainDefPtr virDomainDefParseXML(virCapsPtr caps, virDomainReportError(VIR_ERR_XML_ERROR, _("Unsupported NUMA memory " "tuning mode '%s'"), - tmp); + tmp); goto error; } VIR_FREE(tmp); } else { def->numatune.memory.mode = VIR_DOMAIN_NUMATUNE_MEM_STRICT; } + + if (placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC) { + tmp = virXMLPropString(cur, "nodeset"); + if (tmp) { + char *set = tmp; + int nodemasklen = VIR_DOMAIN_CPUMASK_LEN; + + if (VIR_ALLOC_N(def->numatune.memory.nodemask, + nodemasklen) < 0) { + virReportOOMError(); + goto error; + } + + /* "nodeset" leads same syntax with "cpuset". */ + if (virDomainCpuSetParse(set, 0, + def->numatune.memory.nodemask, + nodemasklen) < 0) + goto error; + VIR_FREE(tmp); + } else { + virDomainReportError(VIR_ERR_XML_ERROR, "%s", + _("nodeset for NUMA memory " + "tuning must be set")); + goto error; + } + } } else { virDomainReportError(VIR_ERR_XML_ERROR, _("unsupported XML element %s"), @@ -12491,25 +12515,33 @@ virDomainDefFormatInternal(virDomainDefPtr def, def->cputune.period || def->cputune.quota) virBufferAddLit(buf, " </cputune>\n"); - if (def->numatune.memory.nodemask) { + if (def->numatune.memory.nodemask || + (def->numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_AUTO)) { virBufferAddLit(buf, " <numatune>\n"); const char *mode; char *nodemask = NULL; - - nodemask = virDomainCpuSetFormat(def->numatune.memory.nodemask, - VIR_DOMAIN_CPUMASK_LEN); - if (nodemask == NULL) { - virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", - _("failed to format nodeset for " - "NUMA memory tuning")); - goto cleanup; - } + const char *placement; mode = virDomainNumatuneMemModeTypeToString(def->numatune.memory.mode); - virBufferAsprintf(buf, " <memory mode='%s' nodeset='%s'/>\n", - mode, nodemask); - VIR_FREE(nodemask); + virBufferAsprintf(buf, " <memory mode='%s' ", mode); + if (def->numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC) { + nodemask = virDomainCpuSetFormat(def->numatune.memory.nodemask, + VIR_DOMAIN_CPUMASK_LEN); + if (nodemask == NULL) { + virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("failed to format nodeset for " + "NUMA memory tuning")); + goto cleanup; + } + virBufferAsprintf(buf, "nodeset='%s'/>\n", nodemask); + VIR_FREE(nodemask); + } else { + placement = virDomainNumatuneMemPlacementModeTypeToString(def->numatune.memory.placement_mode); + virBufferAsprintf(buf, "placement='%s'/>\n", placement); + } virBufferAddLit(buf, " </numatune>\n"); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 5aa8fc1..9aade99 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1416,6 +1416,13 @@ enum virDomainCpuPlacementMode { VIR_DOMAIN_CPU_PLACEMENT_MODE_LAST, }; +enum virDomainNumatuneMemPlacementMode { + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC = 0, + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_AUTO, + + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_LAST, +}; + typedef struct _virDomainTimerCatchupDef virDomainTimerCatchupDef; typedef virDomainTimerCatchupDef *virDomainTimerCatchupDefPtr; struct _virDomainTimerCatchupDef { @@ -1504,6 +1511,7 @@ struct _virDomainNumatuneDef { struct { char *nodemask; int mode; + int placement_mode; } memory; /* Future NUMA tuning related stuff should go here. */ @@ -2176,6 +2184,7 @@ VIR_ENUM_DECL(virDomainGraphicsSpiceStreamingMode) VIR_ENUM_DECL(virDomainGraphicsSpiceClipboardCopypaste) VIR_ENUM_DECL(virDomainGraphicsSpiceMouseMode) VIR_ENUM_DECL(virDomainNumatuneMemMode) +VIR_ENUM_DECL(virDomainNumatuneMemPlacementMode) VIR_ENUM_DECL(virDomainSnapshotState) /* from libvirt.h */ VIR_ENUM_DECL(virDomainState) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 88f8a21..c9c1486 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -400,6 +400,8 @@ virDomainNostateReasonTypeFromString; virDomainNostateReasonTypeToString; virDomainNumatuneMemModeTypeFromString; virDomainNumatuneMemModeTypeToString; +virDomainNumatuneMemPlacementModeTypeFromString; +virDomainNumatuneMemPlacementModeTypeToString; virDomainObjAssignDef; virDomainObjCopyPersistentDef; virDomainObjGetPersistentDef; diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index f1401e1..72beb83 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1657,7 +1657,8 @@ qemuProcessDetectVcpuPIDs(struct qemud_driver *driver, */ #if HAVE_NUMACTL static int -qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) +qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm, + const char *nodemask) { nodemask_t mask; int mode = -1; @@ -1666,9 +1667,18 @@ qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) int i = 0; int maxnode = 0; bool warned = false; + virDomainNumatuneDef numatune = vm->def->numatune; + const char *tmp_nodemask = NULL; - if (!vm->def->numatune.memory.nodemask) - return 0; + if (numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC) { + if (!numatune.memory.nodemask) + return 0; + tmp_nodemask = numatune.memory.nodemask; + } else if (numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_AUTO) { + tmp_nodemask = nodemask; + } VIR_DEBUG("Setting NUMA memory policy"); @@ -1679,11 +1689,10 @@ qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) } maxnode = numa_max_node() + 1; - /* Convert nodemask to NUMA bitmask. */ nodemask_zero(&mask); for (i = 0; i < VIR_DOMAIN_CPUMASK_LEN; i++) { - if (vm->def->numatune.memory.nodemask[i]) { + if (tmp_nodemask[i]) { if (i > NUMA_NUM_NODES) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("Host cannot support NUMA node %d"), i); @@ -1693,12 +1702,12 @@ qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm) VIR_WARN("nodeset is out of range, there is only %d NUMA " "nodes on host", maxnode); warned = true; - } + } nodemask_set(&mask, i); } } - mode = vm->def->numatune.memory.mode; + mode = numatune.memory.mode; if (mode == VIR_DOMAIN_NUMATUNE_MEM_STRICT) { numa_set_bind_policy(1); @@ -1789,7 +1798,8 @@ qemuGetNumadAdvice(virDomainDefPtr def ATTRIBUTE_UNUSED) */ static int qemuProcessInitCpuAffinity(struct qemud_driver *driver, - virDomainObjPtr vm) + virDomainObjPtr vm, + const char *nodemask) { int ret = -1; int i, hostcpus, maxcpu = QEMUD_CPUMASK_LEN; @@ -1815,27 +1825,6 @@ qemuProcessInitCpuAffinity(struct qemud_driver *driver, } if (vm->def->placement_mode == VIR_DOMAIN_CPU_PLACEMENT_MODE_AUTO) { - char *nodeset = NULL; - char *nodemask = NULL; - - nodeset = qemuGetNumadAdvice(vm->def); - if (!nodeset) - goto cleanup; - - if (VIR_ALLOC_N(nodemask, VIR_DOMAIN_CPUMASK_LEN) < 0) { - virReportOOMError(); - VIR_FREE(nodeset); - goto cleanup; - } - - if (virDomainCpuSetParse(nodeset, 0, nodemask, - VIR_DOMAIN_CPUMASK_LEN) < 0) { - VIR_FREE(nodemask); - VIR_FREE(nodeset); - goto cleanup; - } - VIR_FREE(nodeset); - /* numad returns the NUMA node list, convert it to cpumap */ int prev_total_ncpus = 0; for (i = 0; i < driver->caps->host.nnumaCell; i++) { @@ -1852,8 +1841,6 @@ qemuProcessInitCpuAffinity(struct qemud_driver *driver, } prev_total_ncpus += cur_ncpus; } - - VIR_FREE(nodemask); } else { if (vm->def->cpumask) { /* XXX why don't we keep 'cpumask' in the libvirt cpumap @@ -2564,6 +2551,8 @@ static int qemuProcessHook(void *data) struct qemuProcessHookData *h = data; int ret = -1; int fd; + char *nodeset = NULL; + char *nodemask = NULL; /* Some later calls want pid present */ h->vm->pid = getpid(); @@ -2597,14 +2586,34 @@ static int qemuProcessHook(void *data) if (qemuAddToCgroup(h->driver, h->vm->def) < 0) goto cleanup; + if ((h->vm->def->placement_mode == + VIR_DOMAIN_CPU_PLACEMENT_MODE_AUTO) || + (h->vm->def->numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_AUTO)) { + nodeset = qemuGetNumadAdvice(h->vm->def); + if (!nodeset) + goto cleanup; + + VIR_DEBUG("Nodeset returned from numad: %s", nodeset); + + if (VIR_ALLOC_N(nodemask, VIR_DOMAIN_CPUMASK_LEN) < 0) { + virReportOOMError(); + goto cleanup; + } + + if (virDomainCpuSetParse(nodeset, 0, nodemask, + VIR_DOMAIN_CPUMASK_LEN) < 0) + goto cleanup; + } + /* This must be done after cgroup placement to avoid resetting CPU * affinity */ VIR_DEBUG("Setup CPU affinity"); - if (qemuProcessInitCpuAffinity(h->driver, h->vm) < 0) + if (qemuProcessInitCpuAffinity(h->driver, h->vm, nodemask) < 0) goto cleanup; - if (qemuProcessInitNumaMemoryPolicy(h->vm) < 0) - return -1; + if (qemuProcessInitNumaMemoryPolicy(h->vm, nodemask) < 0) + goto cleanup; VIR_DEBUG("Setting up security labelling"); if (virSecurityManagerSetProcessLabel(h->driver->securityManager, h->vm->def) < 0) @@ -2614,6 +2623,8 @@ static int qemuProcessHook(void *data) cleanup: VIR_DEBUG("Hook complete ret=%d", ret); + VIR_FREE(nodeset); + VIR_FREE(nodemask); return ret; } diff --git a/tests/qemuxml2argvdata/qemuxml2argv-numad.args b/tests/qemuxml2argvdata/qemuxml2argv-numad.args new file mode 100644 index 0000000..23bcb70 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-numad.args @@ -0,0 +1,4 @@ +LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test /usr/bin/qemu -S -M \ +pc -m 214 -smp 2 -nographic -monitor \ +unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -hda \ +/dev/HostVG/QEMUGuest1 -net none -serial none -parallel none -usb diff --git a/tests/qemuxml2argvdata/qemuxml2argv-numad.xml b/tests/qemuxml2argvdata/qemuxml2argv-numad.xml new file mode 100644 index 0000000..c87ec49 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-numad.xml @@ -0,0 +1,31 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory unit='KiB'>219136</memory> + <currentMemory unit='KiB'>219136</currentMemory> + <vcpu placement='auto'>2</vcpu> + <numatune> + <memory mode="interleave" placement='auto'/> + </numatune> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <cpu> + <topology sockets='2' cores='1' threads='1'/> + </cpu> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='block' device='disk'> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' target='0' unit='0'/> + </disk> + <controller type='ide' index='0'/> + <memballoon model='virtio'/> + </devices> +</domain> diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index e47a385..3e529e2 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -732,6 +732,7 @@ mymain(void) DO_TEST("blkiotune-device", false, QEMU_CAPS_NAME); DO_TEST("cputune", false, QEMU_CAPS_NAME); DO_TEST("numatune-memory", false, NONE); + DO_TEST("numad", false, NONE); DO_TEST("blkdeviotune", false, QEMU_CAPS_NAME, QEMU_CAPS_DEVICE, QEMU_CAPS_DRIVE, QEMU_CAPS_DRIVE_IOTUNE); -- 1.7.7.3

2 2

[libvirt] internal error hostname of destination resolved to localhost, but migration requires an FQDN
by William Herry 08 May '12

08 May '12

HI I use virsh to migrate kvm virtual machine, it show me this error I already use ip address in my command: virsh migrate --live kvm-test qemu+ssh://root@my_ip:port/system after a few tries I make it work by change the hostname of dest to something else rather than localhost, this looks like a bug cause a already this virsh the ip address, or any thing I misunderstand? regards -- =========================== William Herry WilliamHerryChina(a)Gmail.com

2 2

[libvirt] [PATCHv6 0/8] live block migration
by Eric Blake 07 May '12

07 May '12

v5: https://www.redhat.com/archives/libvir-list/2012-April/msg00753.html Differences in v6: - rebased on top of accepted patches v5:1-4/23 and latest tree - corresponds to patches v5:7-14/23 - patch 5/8 is new - patch v5:12/23 is dropped; qemu is giving us something better, but I still need to finish writing that patch - patch v5:11/23 comments were incorporated, with better cleanup on error - tweak series to deal with potential for qemu 1.1 to support copy but not pivot - limit series to just the bare minimum for now; patches from v5:15/23 are still on my tree, but not worth submitting until we know more about what qemu will provide I'm posting this more for reference for my efforts to backport this to RHEL 6 where I have tested against a candidate qemu build, and don't this is ready for upstream until we have confirmation that actual patches have gone into at least the qemu block queue for 1.1. Eric Blake (8): blockjob: react to active block copy blockjob: add qemu capabilities related to block jobs blockjob: return appropriate event and info blockjob: support pivot operation on cancel blockjob: make drive-reopen safer blockjob: implement block copy for qemu blockjob: allow for existing files blockjob: allow mirroring under SELinux src/conf/domain_conf.c | 12 ++ src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 1 + src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 2 + src/qemu/qemu_driver.c | 396 +++++++++++++++++++++++++++++++++++++++++- src/qemu/qemu_hotplug.c | 7 + src/qemu/qemu_monitor.c | 37 ++++ src/qemu/qemu_monitor.h | 11 ++ src/qemu/qemu_monitor_json.c | 67 +++++++ src/qemu/qemu_monitor_json.h | 18 ++- 11 files changed, 549 insertions(+), 6 deletions(-) -- 1.7.7.6

3 15

[libvirt] [PATCH] virsh: avoid heap corruption leading to virsh abort
by Jim Meyering 07 May '12

07 May '12

Investigating a build problem reported by Laine, I was surprised to see "make check" fail on F17 due to a glibc invalid free abort. Ok to push to master? >From 61a559e0b2f4bded3059c5be7c958e2276f7fd16 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 7 May 2012 21:22:09 +0200 Subject: [PATCH] virsh: avoid heap corruption leading to virsh abort * tools/virsh.c (vshParseSnapshotDiskspec): Fix off-by-3 memmove that would corrupt heap when parsing escaped --diskspec comma. Bug introduced via commit v0.9.4-260-g35d52b5. --- tools/virsh.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/virsh.c b/tools/virsh.c index 1207ac9..dd9292a 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -16107,7 +16107,7 @@ vshParseSnapshotDiskspec(vshControl *ctl, virBufferPtr buf, const char *str) while ((tmp = strchr(tmp, ','))) { if (tmp[1] == ',') { /* Recognize ,, as an escape for a literal comma */ - memmove(&tmp[1], &tmp[2], len - (tmp - spec) + 2); + memmove(&tmp[1], &tmp[2], len - (tmp - spec) - 2 + 1); len--; tmp++; continue; -- 1.7.10.1.457.g8275905

2 2

[libvirt] [PATCHv3 0/4] util: fix libvirtd startup failure due to netlink error
by Laine Stump 07 May '12

07 May '12

The initial patch in this series is the same as the single patch I provided before, but in addition there are 3 new patches that take care of the fallout from 1/4 by: 2/4) removing hard-coding of the src nl_pid when virNetLinkCommand is sending a message 3/4) adding a helper function to retrieve the local nl_pid from a libnl socket and 4/4) using that value when appropriate (I think). As before, I'm unable to fully test this myself, so I won't push unless/until I get verification it works.

3 10

[libvirt] [PATCH] openvz: simplify openvzDomainDefineCmd by using virCommandPtr
by Guido Günther 07 May '12

07 May '12

--- While working on setting filesystem quotas I came accross this and I figured I'll send this in advance because of the diffstat. Cheers, -- Guido src/openvz/openvz_driver.c | 79 +++++++++++--------------------------------- 1 file changed, 19 insertions(+), 60 deletions(-) diff --git a/src/openvz/openvz_driver.c b/src/openvz/openvz_driver.c index 91f5d49..8cceed6 100644 --- a/src/openvz/openvz_driver.c +++ b/src/openvz/openvz_driver.c @@ -100,68 +100,31 @@ static void cmdExecFree(const char *cmdExec[]) /* generate arguments to create OpenVZ container return -1 - error 0 - OK + Caller has to free the cmd */ -static int -openvzDomainDefineCmd(const char *args[], - int maxarg, virDomainDefPtr vmdef) +static virCommandPtr +openvzDomainDefineCmd(virDomainDefPtr vmdef) { - int narg; - - for (narg = 0; narg < maxarg; narg++) - args[narg] = NULL; + virCommandPtr cmd = virCommandNewArgList(VZCTL, + "--quiet", + "create", + NULL); if (vmdef == NULL) { openvzError(VIR_ERR_INTERNAL_ERROR, "%s", _("Container is not defined")); - return -1; + virCommandFree(cmd); + return NULL; } -#define ADD_ARG(thisarg) \ - do { \ - if (narg >= maxarg) \ - goto no_memory; \ - args[narg++] = thisarg; \ - } while (0) - -#define ADD_ARG_LIT(thisarg) \ - do { \ - if (narg >= maxarg) \ - goto no_memory; \ - if ((args[narg++] = strdup(thisarg)) == NULL) \ - goto no_memory; \ - } while (0) - - narg = 0; - ADD_ARG_LIT(VZCTL); - ADD_ARG_LIT("--quiet"); - ADD_ARG_LIT("create"); - - ADD_ARG_LIT(vmdef->name); - ADD_ARG_LIT("--name"); - ADD_ARG_LIT(vmdef->name); + virCommandAddArgList(cmd, vmdef->name, "--name", vmdef->name, NULL); if (vmdef->nfss == 1 && vmdef->fss[0]->type == VIR_DOMAIN_FS_TYPE_TEMPLATE) { - ADD_ARG_LIT("--ostemplate"); - ADD_ARG_LIT(vmdef->fss[0]->src); - } -#if 0 - if ((vmdef->profile && *(vmdef->profile))) { - ADD_ARG_LIT("--config"); - ADD_ARG_LIT(vmdef->profile); + virCommandAddArgList(cmd, "--ostemplate", vmdef->fss[0]->src, NULL); } -#endif - ADD_ARG(NULL); - return 0; - -no_memory: - openvzError(VIR_ERR_INTERNAL_ERROR, - _("Could not put argument to %s"), VZCTL); - return -1; - -#undef ADD_ARG -#undef ADD_ARG_LIT + return cmd; } @@ -170,8 +133,7 @@ static int openvzSetInitialConfig(virDomainDefPtr vmdef) int ret = -1; int vpsid; char * confdir = NULL; - const char *prog[OPENVZ_MAX_ARG]; - prog[0] = NULL; + virCommandPtr cmd = NULL; if (vmdef->nfss > 1) { openvzError(VIR_ERR_INTERNAL_ERROR, "%s", @@ -210,24 +172,21 @@ static int openvzSetInitialConfig(virDomainDefPtr vmdef) _("Could not set the source dir for the filesystem")); goto cleanup; } - } - else - { - if (openvzDomainDefineCmd(prog, OPENVZ_MAX_ARG, vmdef) < 0) { - VIR_ERROR(_("Error creating command for container")); + } else { + cmd = openvzDomainDefineCmd(vmdef); + if (cmd == NULL) goto cleanup; - } - if (virRun(prog, NULL) < 0) { + if (virCommandRun(cmd, NULL) < 0) goto cleanup; - } } ret = 0; cleanup: VIR_FREE(confdir); - cmdExecFree(prog); + virCommandFree(cmd); + return ret; } -- 1.7.10

2 2

[libvirt] [PATCH] docs: Add 'maintenance releases' link in 'News' sidebar
by Cole Robinson 07 May '12

07 May '12

Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- docs/sitemap.html.in | 8 +++++++- 1 files changed, 7 insertions(+), 1 deletions(-) diff --git a/docs/sitemap.html.in b/docs/sitemap.html.in index 8f58d46..206040c 100644 --- a/docs/sitemap.html.in +++ b/docs/sitemap.html.in @@ -13,8 +13,14 @@ Details of new features and bugs fixed in each release <ul> <li> + <a href="http://wiki.libvirt.org/page/Maintenance_Releases">Maintenance Releases</a> + Details about libvirt maintenance releases + </li> + </ul> + <ul> + <li> <a href="http://libvirt.org/git/?p=libvirt.git;a=log">Git log</a> - Latest commit messages from the source repository + Latest commit messages from the source repository </li> </ul> </li> -- 1.7.7.6

2 2

[libvirt] [PATCH] rpm: Handle different source URLs for maint releases
by Cole Robinson 07 May '12

07 May '12

Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- libvirt.spec.in | 6 +++++- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index e7e0a55..95d8af4 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -276,10 +276,14 @@ Version: @VERSION@ Release: 1%{?dist}%{?extra_release} License: LGPLv2+ Group: Development/Libraries -Source: http://libvirt.org/sources/libvirt-%{version}.tar.gz BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root URL: http://libvirt.org/ +%if %(echo %{version} | grep -o \\. | wc -l) == 3 +%define mainturl stable_updates/ +%endif +Source: http://libvirt.org/sources/%{?mainturl}libvirt-%{version}.tar.gz + %if %{with_libvirtd} Requires: libvirt-daemon = %{version}-%{release} %if %{with_network} -- 1.7.7.6

2 2

[libvirt] [PATCH 0/2] Fix migration to older versions of libvirt
by Jiri Denemark 07 May '12

07 May '12

This series deals with migrating to older versions of libvirt broken by unconditionally added USB controller into domain XML. Jiri Denemark (2): qemu: Don't use virDomainDefFormat* directly qemu: Emit compatible XML when migrating a domain src/qemu/qemu_domain.c | 88 ++++++++++++++++++++++++++++++++++++++++----- src/qemu/qemu_domain.h | 15 ++++++-- src/qemu/qemu_driver.c | 26 ++++++++------ src/qemu/qemu_migration.c | 30 +++++++++------ src/qemu/qemu_process.c | 8 ++-- 5 files changed, 128 insertions(+), 39 deletions(-) -- 1.7.8.6

2 5