- Devel - libvirt List Archives

[libvirt] virNodeDevice APIs and VFIO
by Laine Stump 24 Apr '13

24 Apr '13

Yesterday for the first time I consciously noticed the virNodeDeviceDettach and virNodeDeviceReAttach APIs, and found that they are hardcoded to bind to/unbind from the pci-stub driver for qemu, and the "pciback" driver for Xen. If we want these APIs to be useful for VFIO, they will need to bind to the vfio driver instead, but there is currently no method in those APIs to specify which driver to bind to. I guess we could do this with new virNodeDeviceDetachFlags() and virNodeDeviceReAttachFlags() APIs which have a flag to indicate vfio, but before going to that trouble I'd like to know if these APIs are actually used or if they are deprecated (they don't seem to be of any use if the hostdev devices you're assigning have "managed='yes'" - as far as I can see, setting managed='yes' just makes the bind/unbind from the stub driver an automatic part of assigning/un-assigning the device to a guest). (Beyond that, this would be exposing a lower level of detail in the APIs than the original functions, and I'm not sure if it's appropriate - the choice of stub driver was previously hidden in each hypervisor driver).

3 6

[libvirt] [PATCH] util: Error out if the numa nodeset is out of range
by Osier Yang 24 Apr '13

24 Apr '13

Instead of a silent warning, it's better to error out if the numa nodeset is out of range. Just like for numa node larger than NUMA_NUM_NODES. --- src/util/virnuma.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) diff --git a/src/util/virnuma.c b/src/util/virnuma.c index bace06f..902ed43 100644 --- a/src/util/virnuma.c +++ b/src/util/virnuma.c @@ -89,7 +89,6 @@ virNumaSetupMemoryPolicy(virNumaTuneDef numatune, int ret = -1; int i = 0; int maxnode = 0; - bool warned = false; virBitmapPtr tmp_nodemask = NULL; if (numatune.memory.placement_mode == @@ -113,20 +112,17 @@ virNumaSetupMemoryPolicy(virNumaTuneDef numatune, } maxnode = numa_max_node() + 1; + /* Convert nodemask to NUMA bitmask. */ nodemask_zero(&mask); i = -1; while ((i = virBitmapNextSetBit(tmp_nodemask, i)) >= 0) { - if (i > NUMA_NUM_NODES) { + if (i > maxnode || i > NUMA_NUM_NODES) { virReportError(VIR_ERR_INTERNAL_ERROR, - _("Host cannot support NUMA node %d"), i); + _("Nodeset is out of range, host cannot support " + "NUMA node bigger than %d"), i); return -1; } - if (i > maxnode && !warned) { - VIR_WARN("nodeset is out of range, there is only %d NUMA " - "nodes on host", maxnode); - warned = true; - } nodemask_set(&mask, i); } -- 1.8.1.4

2 3

[libvirt] [PATCHv2] selinux: Don't mask errors of virSecuritySELinuxGenNewContext
by Peter Krempa 24 Apr '13

24 Apr '13

Since cbe67ff9b0a5a94911afd4d12388bf182bcca86c virSecuritySELinuxGenNewContext reports good error messages. Fix callers that mask the errors by generic error message. --- src/security/security_selinux.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) diff --git a/src/security/security_selinux.c b/src/security/security_selinux.c index c620a2e..61ff1de 100644 --- a/src/security/security_selinux.c +++ b/src/security/security_selinux.c @@ -685,13 +685,10 @@ virSecuritySELinuxGenSecurityLabel(virSecurityManagerPtr mgr, } } - seclabel->label = - virSecuritySELinuxGenNewContext(baselabel, mcs, false); - if (!seclabel->label) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot generate selinux context for %s"), mcs); + seclabel->label = virSecuritySELinuxGenNewContext(baselabel, mcs, false); + if (!seclabel->label) goto cleanup; - } + break; case VIR_DOMAIN_SECLABEL_NONE: @@ -709,11 +706,8 @@ virSecuritySELinuxGenSecurityLabel(virSecurityManagerPtr mgr, seclabel->imagelabel = virSecuritySELinuxGenNewContext(data->file_context, mcs, true); - if (!seclabel->imagelabel) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot generate selinux context for %s"), mcs); + if (!seclabel->imagelabel) goto cleanup; - } } if (!seclabel->model && -- 1.8.2.1

2 2

[libvirt] Fuse file system mounted at /proc/meminfo needs correct label.
by dwalsh＠redhat.com 24 Apr '13

24 Apr '13

Second Pass... [PATCH 1/1] Fuse file system mounted at /proc/meminfo needs correct

1 1

[libvirt] [PATCH] selinux: Don't mask errors of virSecuritySELinuxGenNewContext
by Peter Krempa 24 Apr '13

24 Apr '13

virSecuritySELinuxGenNewContext() reports sensible errors but at virSecuritySELinuxGenSecurityLabel() the error is masked with a non-specific one. --- src/security/security_selinux.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/src/security/security_selinux.c b/src/security/security_selinux.c index c620a2e..89d63ab 100644 --- a/src/security/security_selinux.c +++ b/src/security/security_selinux.c @@ -685,13 +685,10 @@ virSecuritySELinuxGenSecurityLabel(virSecurityManagerPtr mgr, } } - seclabel->label = - virSecuritySELinuxGenNewContext(baselabel, mcs, false); - if (!seclabel->label) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot generate selinux context for %s"), mcs); + seclabel->label = virSecuritySELinuxGenNewContext(baselabel, mcs, false); + if (!seclabel->label) goto cleanup; - } + break; case VIR_DOMAIN_SECLABEL_NONE: -- 1.8.2.1

2 1

[libvirt] [PATCH] qemuBuildCommandLine: Don't overwrite errors with NWFilter's one
by Michal Privoznik 24 Apr '13

24 Apr '13

Currently, if there has been an error in building command line process after virtual interfaces has been created, the flow jumps to 'error' label, where virDomainConfNWFilterTeardown() is called. This may report an error as well, but should not overwrite the original cause why we jumped to 'error' label. --- src/qemu/qemu_command.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 741fa82..92b15d9 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -5872,6 +5872,7 @@ qemuBuildCommandLine(virConnectPtr conn, virDomainSnapshotObjPtr snapshot, enum virNetDevVPortProfileOp vmop) { + virErrorPtr originalError = NULL; int i, j; const char *emulator; char uuid[VIR_UUID_STRING_BUFLEN]; @@ -7802,13 +7803,16 @@ qemuBuildCommandLine(virConnectPtr conn, virObjectUnref(cfg); return cmd; - no_memory: +no_memory: virReportOOMError(); - error: +error: virObjectUnref(cfg); - /* free up any resources in the network driver */ + /* free up any resources in the network driver + * but don't overwrite the original error */ + originalError = virSaveLastError(); for (i = 0; i <= last_good_net; i++) virDomainConfNWFilterTeardown(def->nets[i]); + virSetError(originalError); virCommandFree(cmd); return NULL; } -- 1.8.1.5

2 1

[libvirt] [PATCH 0/6] Cleanup of internal API naming conventions
by Daniel P. Berrange 24 Apr '13

24 Apr '13

The naming conventions used for APIs in both src/driver.h and src/remote/remote_protocol.x should follow the public API naming, but there are a great many violations of this. This series fixes them all, and adds some tests which will help prevent regressions.

2 18

[libvirt] [PATCH] qemu: Generate agent socket path if missing
by Michal Privoznik 24 Apr '13

24 Apr '13

It's not desired to force users imagine path for a socket they are not even supposed to connect to. On the other hand, we already have a release where the qemu agent socket path is exposed to XML, so we cannot silently drop it from there. The new path is generated in form: $LOCALSTATEDIR/lib/libvirt/qemu/channel/target/$domain.$name --- libvirt.spec.in | 1 + src/Makefile.am | 1 + src/conf/domain_conf.c | 34 ++++++++++++++++------------------ src/qemu/qemu_domain.c | 16 ++++++++++++++++ 4 files changed, 34 insertions(+), 18 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 328b009..11fc12d 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1719,6 +1719,7 @@ fi %if %{with_qemu} %ghost %dir %attr(0700, root, root) %{_localstatedir}/run/libvirt/qemu/ %dir %attr(0750, %{qemu_user}, %{qemu_group}) %{_localstatedir}/lib/libvirt/qemu/ +%dir %attr{0750, %{qemu_user}, %{qemu_group}) %{_localstatedir}/lib/libvirt/qemu/channel/target/ %dir %attr(0750, %{qemu_user}, %{qemu_group}) %{_localstatedir}/cache/libvirt/qemu/ %endif %if %{with_lxc} diff --git a/src/Makefile.am b/src/Makefile.am index fc6b846..9576da8 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -2097,6 +2097,7 @@ if WITH_SANLOCK endif if WITH_QEMU $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/qemu" + $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/qemu/channel/target" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/run/libvirt/qemu" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/cache/libvirt/qemu" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/log/libvirt/qemu" diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 03e5740..3ff548a 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -6464,7 +6464,9 @@ virDomainChrSourceDefParseXML(virDomainChrSourceDefPtr def, break; case VIR_DOMAIN_CHR_TYPE_UNIX: - if (path == NULL) { + /* path can be auto generated */ + if (!path && + chr_def->targetType != VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Missing source path attribute for char device")); goto error; @@ -6559,7 +6561,6 @@ virDomainChrDefParseXML(xmlXPathContextPtr ctxt, char *type = NULL; const char *nodeName; virDomainChrDefPtr def; - int remaining; bool seenTarget = false; if (!(def = virDomainChrDefNew())) @@ -6583,29 +6584,26 @@ virDomainChrDefParseXML(xmlXPathContextPtr ctxt, } cur = node->children; - remaining = virDomainChrSourceDefParseXML(&def->source, cur, flags, - def, ctxt, - vmSeclabels, nvmSeclabels); - if (remaining < 0) - goto error; - if (remaining) { - while (cur != NULL) { - if (cur->type == XML_ELEMENT_NODE) { - if (xmlStrEqual(cur->name, BAD_CAST "target")) { - seenTarget = true; - if (virDomainChrDefParseTargetXML(def, cur) < 0) { - goto error; - } + while (cur != NULL) { + if (cur->type == XML_ELEMENT_NODE) { + if (xmlStrEqual(cur->name, BAD_CAST "target")) { + seenTarget = true; + if (virDomainChrDefParseTargetXML(def, cur) < 0) { + goto error; } } - cur = cur->next; } + cur = cur->next; } if (!seenTarget && ((def->targetType = virDomainChrDefaultTargetType(def->deviceType)) < 0)) goto cleanup; + if (virDomainChrSourceDefParseXML(&def->source, node->children, flags, def, + ctxt, vmSeclabels, nvmSeclabels) < 0) + goto error; + if (def->source.type == VIR_DOMAIN_CHR_TYPE_SPICEVMC) { if (def->targetType != VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", @@ -13897,8 +13895,8 @@ virDomainChrSourceDefFormat(virBufferPtr buf, case VIR_DOMAIN_CHR_TYPE_UNIX: virBufferAsprintf(buf, " <source mode='%s'", def->data.nix.listen ? "bind" : "connect"); - virBufferEscapeString(buf, " path='%s'/>\n", - def->data.nix.path); + virBufferEscapeString(buf, " path='%s'", def->data.nix.path); + virBufferAddLit(buf, "/>\n"); break; } diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 3d6eef4..967890f 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -738,6 +738,22 @@ qemuDomainDeviceDefPostParse(virDomainDeviceDefPtr dev, (def->os.arch == VIR_ARCH_S390 || def->os.arch == VIR_ARCH_S390X)) dev->data.chr->targetType = VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_VIRTIO; + /* auto generate unix socket path */ + if (dev->type == VIR_DOMAIN_DEVICE_CHR && + dev->data.chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CHANNEL && + dev->data.chr->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO && + dev->data.chr->source.type == VIR_DOMAIN_CHR_TYPE_UNIX && + !dev->data.chr->source.data.nix.path && + (cfg || (driver && (cfg = virQEMUDriverGetConfig(driver))))) { + + if (virAsprintf(&dev->data.chr->source.data.nix.path, + "%s/channel/target/%s.%s", + cfg->libDir, def->name, + dev->data.chr->target.name) < 0) + goto no_memory; + dev->data.chr->source.data.nix.listen = true; + } + ret = 0; cleanup: -- 1.8.1.5

3 3

[libvirt] Using leases with virtlockd
by Jim Fehlig 24 Apr '13

24 Apr '13

Nearly a question for the users list, but I might be encountering a bug too. I'm using virtlockd with 'auto_disk_leases = 0' and struggle with the correct configuration for a lease in the domXML. Using <lease> <lockspace>test-ls</lockspace> <key>test-lock</key> <target path='/var/lib/libvirt/lockd'/> </lease> results in the following error when starting the domain internal error Lockspace for path /var/lib/libvirt/lockd/test-ls It appears the path attribute + lockspace are used as the Lockspace for this lease. But I'm not sure what "Lockspace" means in this context. Is it a directory? A file? I assume the latter since it is just flock()ed when the domain is started, right? I'll be happy to provide patches to fix any bugs that might exist here (docs included), but need understand the correct configuration of leases with virtlockd first :). BTW, nice piece of work Daniel! Regards, Jim

2 1

[libvirt] [PATCH] Use camelCase for XML attribute numQueues
by Martin Kletzander 24 Apr '13

24 Apr '13

In commit d4bf0a9, we used num_queues for an attribute in the XML, but the consensus is that we use camelCase for that. Since there was no release yet (the above commit describes as v1.0.4-65-gd4bf0a9), we still have time to change it. Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- This patch applies on top of Laine's RNG tightening patch [1] and it was proposed in that thread as well. [1] http://www.redhat.com/archives/libvir-list/2013-April/msg01320.html --- docs/formatdomain.html.in | 2 +- docs/schemas/domaincommon.rng | 2 +- src/conf/domain_conf.c | 6 +++--- src/qemu/qemu_command.c | 2 +- tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml | 2 +- 5 files changed, 7 insertions(+), 7 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 0cc56d9..a5be162 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -2135,7 +2135,7 @@ controller. A "scsi" controller has an optional attribute <code>model</code>, which is one of "auto", "buslogic", "ibmvscsi", "lsilogic", "lsisas1068", "lsisas1078", "virtio-scsi" or - "vmpvscsi". The attribute <code>num_queues</code> + "vmpvscsi". The attribute <code>numQueues</code> (<span class="since">1.0.5 (QEMU and KVM only)</span>) specifies the number of queues for the controller. For best performance, it's recommended to specify a value matching the number of vCPUs. A "usb" diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 3976b82..bcb1453 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -1444,7 +1444,7 @@ </attribute> </optional> <optional> - <attribute name="num_queues"> + <attribute name="numQueues"> <ref name="unsignedInt"/> </attribute> </optional> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 1643f30..0487053 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5195,10 +5195,10 @@ virDomainControllerDefParseXML(xmlNodePtr node, def->model = -1; } - if ((num_queues = virXMLPropString(node, "num_queues"))) { + if ((num_queues = virXMLPropString(node, "numQueues"))) { if (virStrToLong_ui(num_queues, NULL, 10, &def->num_queues) < 0) { virReportError(VIR_ERR_XML_ERROR, - _("Malformed 'num_queues' value '%s'"), num_queues); + _("Malformed 'numQueues' value '%s'"), num_queues); goto error; } } @@ -13531,7 +13531,7 @@ virDomainControllerDefFormat(virBufferPtr buf, } if (def->num_queues) - virBufferAsprintf(buf, " num_queues='%u'", def->num_queues); + virBufferAsprintf(buf, " numQueues='%u'", def->num_queues); switch (def->type) { case VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL: diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 009d42d..65675b7 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3574,7 +3574,7 @@ qemuBuildControllerDevStr(virDomainDefPtr domainDef, !(def->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI && def->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", - _("'num_queues' is only supported by virtio-scsi controller")); + _("'numQueues' is only supported by virtio-scsi controller")); return NULL; } diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml index dfa9cf1..063ccf7 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml @@ -20,7 +20,7 @@ <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> <controller type='usb' index='0'/> - <controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/> + <controller type='scsi' index='0' model='virtio-scsi' numQueues='8'/> <memballoon model='virtio'/> </devices> </domain> -- 1.8.1.5

5 12

[libvirt] internal error Cannot find suitable CPU model for given data
by yue 24 Apr '13

24 Apr '13

hi,all: if xml configuration is not right? <cpu match="exact"> <model>Nehalem</model> <topology cores="1" sockets="1" threads="1"/> </cpu> thanks

2 1

[libvirt] Patch sets proper SELinux context on /proc/meminfo (proc_t)
by dwalsh＠redhat.com 24 Apr '13

24 Apr '13

[PATCH] Fuse file system mounted at /proc/meminfo needs correct

2 2

[libvirt] [PATCH 0/2] Prefer -machine accel={tcg, kvm} over -{dis, en}able-kvm
by Jiri Denemark 24 Apr '13

24 Apr '13

Jiri Denemark (2): qemu: Move -enable-kvm and friends earlier in the command line qemu: Use -machine accel=tcg|kvm when available src/qemu/qemu_command.c | 141 ++++++++++++--------- .../qemuxml2argv-cpu-host-kvmclock.args | 2 +- .../qemuxml2argv-cpu-kvmclock.args | 2 +- tests/qemuxml2argvdata/qemuxml2argv-kvm.args | 4 + tests/qemuxml2argvdata/qemuxml2argv-kvm.xml | 22 ++++ .../qemuxml2argv-machine-core-off.args | 2 +- .../qemuxml2argv-machine-core-on.args | 2 +- .../qemuxml2argv-machine-usb-opt.args | 2 +- tests/qemuxml2argvtest.c | 1 + 9 files changed, 113 insertions(+), 65 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-kvm.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-kvm.xml -- 1.8.2.1

2 5

Re: [libvirt] [PATCHv2] Configure native vlan modes on Open vSwitch ports
by james robson 24 Apr '13

24 Apr '13

> Date: Thu, 18 Apr 2013 14:14:32 -0400 > From: Laine Stump <laine(a)laine.org> > To: libvir-list(a)redhat.com > Subject: Re: [libvirt] [PATCHv2] Configure native vlan modes on Open > vSwitch ports > Message-ID: <51703808.2000504(a)laine.org> > Content-Type: text/plain; charset=ISO-8859-1 > > On 04/18/2013 01:44 PM, james robson wrote: > > Hello, > > Has any one been able to review this yet? I realise that the 'Since > > 1.0.3' in the doc page is now out of date, but is the code itself > > acceptable? > > I was hoping that someone with more knowledge of Open vSwitch and/or > vlan tagging/trunking/native mode would repond to the message (Kyle?) > but there was silence instead... > > > > > > > > diff --git a/tests/networkxml2xmlin/openvswitch-net.xml > > b/tests/networkxml2xmlin/openvswitch-net.xml > > index a3d82b1..93c49d5 100644 > > --- a/tests/networkxml2xmlin/openvswitch-net.xml > > +++ b/tests/networkxml2xmlin/openvswitch-net.xml > > @@ -21,4 +21,13 @@ > > <parameters profileid='alice-profile'/> > > </virtualport> > > </portgroup> > > + <portgroup name='tagged'> > > + <vlan native_mode='tagged' native_tag='123'> > > + <tag id='555'/> > > + <tag id='444'/> > > + </vlan> > > + <virtualport> > > + <parameters profileid='tagged-profile'/> > > + </virtualport> > > + </portgroup> > > As brought up again in a separate conversation today, we prefer to use > camelCase rather than underscored in attribute and element names. So, if > we were to use the layout you're proposing, the attributes should be > called "nativeMode" and "nativeTag". > > However, I'm wondering if there might be a better way to structure it. > What about this? > > > <vlan trunk='yes'> > <tag id='123' native='tagged|untagged'/> (or whatever values are > appropriate) > <tag id='555'/> > <tag id='444'/> > </vlan> > > Do I understand correctly that native mode is telling what to do with > packets that come in untagged, and that (using your nomenclature > "native_mode='yes' native_tag='123'" means "when an untagged packet come > in from this interface, it should be tagged as 123 before forwarding"? That is correct, setting the native vlan changes how an untagged packet is handled when it enters the port. The difference between the 'tagged' and 'untagged' modes is in how packets on the native vlan are processed before exiting the port. > And what happens when native_mode='yes' but there is no native_tag? In that case you configuration is invalid, and will get an error. > (that's what I was trying to describe with <tag id='123' > native='untagged'/>, but I don't even know if that makes sense, because > I don't know exactly what is the native vlan tag and what is done with > it :-) That arrangement would make sense, I chose the arrangement I did for two main reasons. There can only be one native vlan on a port, making it an attribute of the 'vlan' tag enforces this. Also, I wanted to keep the validation and processing separate rather than add 'if native' branches to the loops that operate on the vlan id list. I can see the advantage of having a single setting to configure the native vlan, rather than the two attributes I proposed. If the new suggestion is preferred I can rework my patch to use that format. > Also, is it valid to have a native_mode/native_tag if trunk='no'? (right > now trunk is automatically set to 'yes' if there is more than one vlan tag) It isn't valid to have trunk='no' and the native settings. Therefore "<vlan trunk='no' native_mode='tagged' native_tag='123'>" will get an error if you try to enter it. If no "trunk" attribute is set explicitly then it will be set to 'yes'. This means "<vlan native_mode='tagged' native_tag='123'>" is equivalent to "<vlan trunk='yes' native_mode='tagged' native_tag='123'>". > > > </network> > > diff --git a/tests/networkxml2xmlout/openvswitch-net.xml > > b/tests/networkxml2xmlout/openvswitch-net.xml > > index a3d82b1..ab3d797 100644 > > --- a/tests/networkxml2xmlout/openvswitch-net.xml > > +++ b/tests/networkxml2xmlout/openvswitch-net.xml > > @@ -21,4 +21,13 @@ > > <parameters profileid='alice-profile'/> > > </virtualport> > > </portgroup> > > + <portgroup name='tagged'> > > + <vlan trunk='yes' native_mode='tagged' native_tag='123'> > > + <tag id='555'/> > > + <tag id='444'/> > > + </vlan> > > + <virtualport> > > + <parameters profileid='tagged-profile'/> > > + </virtualport> > > + </portgroup> > > </network> > > > Protected by Websense Hosted Email Security -- www.websense.com

2 4

[libvirt] [PATCHv2 0/6] Add support for guest-agent based cpu hot(un)plug
by Peter Krempa 24 Apr '13

24 Apr '13

This version contains fixes requested by Eric on patches 1/6, 3/6 and 4/6. Patch 2/6 is new in the series. I tweaked some details in 5/6 and 6/6 but they weren't reviewed in version 1. Changes are described in the patches themselves. Peter Krempa (6): virsh-domain: Refactor cmdVcpucount and fix output on inactive domains conf: Reword error message to be more universal qemu_agent: Introduce helpers for agent based CPU hot(un)plug API: Introduce VIR_DOMAIN_VCPU_AGENT, for agent based CPU hot(un)plug qemu: Implement request of vCPU state using the guest agent qemu: Implement support for VIR_DOMAIN_VCPU in qemuDomainSetVcpusFlags include/libvirt/libvirt.h.in | 1 + src/conf/domain_conf.c | 2 +- src/libvirt.c | 24 +++- src/qemu/qemu_agent.c | 148 ++++++++++++++++++++++ src/qemu/qemu_agent.h | 12 ++ src/qemu/qemu_driver.c | 189 +++++++++++++++++++++++++--- tools/virsh-domain.c | 292 +++++++++++++++++++++---------------------- tools/virsh.pod | 13 +- 8 files changed, 510 insertions(+), 171 deletions(-) -- 1.8.1.5

3 15

[libvirt] What about next release ?
by Daniel Veillard 23 Apr '13

23 Apr '13

We are getting quite close to the end of the month, if we want to stick to the 'end of month release' I guess we need to enter the freeze for 1.0.5 at the end of the week say Friday, and then push the release, assuming there is no big problem on Thursday next week (May 2, I could probably do it on May 1 but most civilized people will enjoy the May 1 day off so that should not change that much ;-) Daniel -- Daniel Veillard | Open Source and Standards, Red Hat veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

3 3

[libvirt] domblkstat not reporting disk errors
by SHREE DUTH AWASTHI 23 Apr '13

23 Apr '13

Hi All, We are using libvirt 0.10.2 with hypervisor QEMU 0.14.1 We see that our hardware doesnt report any disk errors ( virsh domblkstat) CENTOS(Libvirt version 0.9.4) virsh # domblkstat fp-cla0 hda hda rd_req 59837 hda rd_bytes 477478400 hda wr_req 462795 hda wr_bytes 5631972352 *hda errs 0 * Our hardware (libvirt version 0.10.2) virsh # domblkstat CLA-0 hda hda rd_req 11 hda rd_bytes 217088 hda wr_req 0 hda wr_bytes 0 Could you please let us know if it is a known issue or does libvirt have any dependency for this ? Please throw more light on this. Thanks and Regards, Shree Duth Awasthi.

1 0

[libvirt] [PATCH] Omitting braces with a single-line body
by harryxiyou＠gmail.com 23 Apr '13

23 Apr '13

From: Harry Wei <harryxiyou(a)gmail.com> After i read libvirt/HACKING file, i find we should Omit braces with a single-line body. So this patch fix this coding style problem for Sheepdog storage backend driver. Signed-off-by: Harry Wei <harryxiyou(a)gmail.com> --- src/storage/storage_backend_sheepdog.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) diff --git a/src/storage/storage_backend_sheepdog.c b/src/storage/storage_backend_sheepdog.c index 35a3a04..15fa29c 100644 --- a/src/storage/storage_backend_sheepdog.c +++ b/src/storage/storage_backend_sheepdog.c @@ -97,12 +97,10 @@ virStorageBackendSheepdogAddHostArg(virCommandPtr cmd, const char *address = "localhost"; int port = 7000; if (pool->def->source.nhost > 0) { - if (pool->def->source.hosts[0].name != NULL) { + if (pool->def->source.hosts[0].name != NULL) address = pool->def->source.hosts[0].name; - } - if (pool->def->source.hosts[0].port) { + if (pool->def->source.hosts[0].port) port = pool->def->source.hosts[0].port; - } } virCommandAddArg(cmd, "-a"); virCommandAddArgFormat(cmd, "%s", address); @@ -210,11 +208,10 @@ virStorageBackendSheepdogParseVdiList(virStorageVolDefPtr vol, continue; /* skip space */ - if (p + 2 < next) { + if (p + 2 < next) p += 2; - } else { + else return -1; - } /* skip name */ while (*p != '\0' && *p != ' ') { -- 1.7.9.5

4 4

[libvirt] [libvirt-designer] Add missing files to EXTRA_DIST
by Christophe Fergeau 23 Apr '13

23 Apr '13

This fixes make distcheck --- I've pushed this under the trivial rule. Christophe Makefile.am | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Makefile.am b/Makefile.am index 9ef319a..726d475 100644 --- a/Makefile.am +++ b/Makefile.am @@ -10,6 +10,8 @@ EXTRA_DIST = \ $(PACKAGE).spec \ $(pkgconfig_DATA:%.pc=%.pc.in) \ AUTHORS.in \ + GNUmakefile \ + maint.mk \ README.in \ $(NULL) -- 1.8.1.4

1 0

[libvirt] [PATCH 0/3] Ignore libvirt logs when reading QEMU error output
by Jiri Denemark 23 Apr '13

23 Apr '13

Jiri Denemark (3): qemu: Move QEMU log reading into a separate function qemu: Ignore libvirt logs when reading QEMU error output logging: Make log regexp more compact (and readable) src/qemu/qemu_process.c | 81 +++++++++++++++++++++++++------------------------ src/util/virlog.c | 4 +-- 2 files changed, 44 insertions(+), 41 deletions(-) -- 1.8.2.1

2 7

[libvirt] [PATCH] update input ip processing
by Gene Czarcinski 23 Apr '13

23 Apr '13

1. Handle invalid ULong prefix specified. When parsing for @prefix as a ULong, a -2 can be returned if the specification is not a valid ULong. 2. Error out if address= is not specified. 3. Merge netmask process/tests under family tests. 4. Max sure that prefix does not exceed maximum. . Signed-off-by: Gene Czarcinski <gene(a)czarc.net> --- src/conf/network_conf.c | 118 +++++++++++++++++++++++++++--------------------- 1 file changed, 66 insertions(+), 52 deletions(-) diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 2b56ca7..1503ece 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -1136,7 +1136,8 @@ virNetworkIPDefParseXML(const char *networkName, xmlNodePtr cur, save; char *address = NULL, *netmask = NULL; - unsigned long prefix; + unsigned long prefix = 0; + int prefixRc; int result = -1; save = ctxt->node; @@ -1144,14 +1145,8 @@ virNetworkIPDefParseXML(const char *networkName, /* grab raw data from XML */ def->family = virXPathString("string(./@family)", ctxt); - address = virXPathString("string(./@address)", ctxt); - if (virXPathULong("string(./@prefix)", ctxt, &prefix) < 0) - def->prefix = 0; - else - def->prefix = prefix; - - netmask = virXPathString("string(./@netmask)", ctxt); + address = virXPathString("string(./@address)", ctxt); if (address) { if (virSocketAddrParse(&def->address, address, AF_UNSPEC) < 0) { virReportError(VIR_ERR_XML_ERROR, @@ -1160,23 +1155,66 @@ virNetworkIPDefParseXML(const char *networkName, goto cleanup; } + } else { + virReportError(VIR_ERR_XML_ERROR, + _("Network address must be specified in definition of network '%s'"), + networkName); + goto cleanup; + } + + netmask = virXPathString("string(./@netmask)", ctxt); + if (netmask) { + if (virSocketAddrParse(&def->netmask, netmask, AF_UNSPEC) < 0) { + virReportError(VIR_ERR_XML_ERROR, + _("Bad netmask '%s' in definition of network '%s'"), + netmask, networkName); + goto cleanup; + } + + } + + prefixRc = virXPathULong("string(./@prefix)", ctxt, &prefix); + if (prefixRc == -2) { + virReportError(VIR_ERR_XML_ERROR, + _("Invalid ULong value specified for prefix in definition of network '%s'"), + networkName); + goto cleanup; } + if (prefixRc < 0) + def->prefix = 0; + else + def->prefix = prefix; - /* validate family vs. address */ - if (def->family == NULL) { + /* validate address, etc. for each family */ + if ((def->family == NULL) || (STREQ(def->family, "ipv4"))) { if (!(VIR_SOCKET_ADDR_IS_FAMILY(&def->address, AF_INET) || VIR_SOCKET_ADDR_IS_FAMILY(&def->address, AF_UNSPEC))) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("no family specified for non-IPv4 address '%s' in network '%s'"), - address, networkName); + _("%s family specified for non-IPv4 address '%s' in network '%s'"), + def->family == NULL? "no" : "ipv4", address, networkName); goto cleanup; } - } else if (STREQ(def->family, "ipv4")) { - if (!VIR_SOCKET_ADDR_IS_FAMILY(&def->address, AF_INET)) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("family 'ipv4' specified for non-IPv4 address '%s' in network '%s'"), - address, networkName); - goto cleanup; + if (netmask) { + if (!VIR_SOCKET_ADDR_IS_FAMILY(&def->netmask, AF_INET)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("network '%s' has invalid netmask '%s' for address '%s' (both must be IPv4)"), + networkName, netmask, address); + goto cleanup; + } + if (def->prefix > 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("network '%s' cannot have both a prefix and a netmask"), + networkName); + goto cleanup; + } + } + else { + if (def->prefix > 32 ) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("Invalid IPv4 prefix '%lu'specified in network'%s'"), + prefix, networkName); + goto cleanup; + } } } else if (STREQ(def->family, "ipv6")) { if (!VIR_SOCKET_ADDR_IS_FAMILY(&def->address, AF_INET6)) { @@ -1185,47 +1223,23 @@ virNetworkIPDefParseXML(const char *networkName, address, networkName); goto cleanup; } - } else { - virReportError(VIR_ERR_XML_ERROR, - _("Unrecognized family '%s' in definition of network '%s'"), - def->family, networkName); - goto cleanup; - } - - /* parse/validate netmask */ - if (netmask) { - if (address == NULL) { - /* netmask is meaningless without an address */ - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("netmask specified without address in network '%s'"), - networkName); - goto cleanup; - } - - if (!VIR_SOCKET_ADDR_IS_FAMILY(&def->address, AF_INET)) { + if (netmask) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("netmask not supported for address '%s' in network '%s' (IPv4 only)"), + _("specifying netmask invalid for IPv6 address '%s' in network '%s'"), address, networkName); goto cleanup; } - - if (def->prefix > 0) { - /* can't have both netmask and prefix at the same time */ + if (def->prefix > 128 ) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("network '%s' cannot have both prefix='%u' and a netmask"), - networkName, def->prefix); - goto cleanup; - } - - if (virSocketAddrParse(&def->netmask, netmask, AF_UNSPEC) < 0) - goto cleanup; - - if (!VIR_SOCKET_ADDR_IS_FAMILY(&def->netmask, AF_INET)) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("network '%s' has invalid netmask '%s' for address '%s' (both must be IPv4)"), - networkName, netmask, address); + _("Invalid IPv6 prefix '%lu'specified in network'%s'"), + prefix, networkName); goto cleanup; } + } else { + virReportError(VIR_ERR_XML_ERROR, + _("Unrecognized family '%s' in definition of network '%s'"), + def->family, networkName); + goto cleanup; } cur = node->children; -- 1.8.1.4

2 1

[libvirt] [libvirt-designer PATCHv2 0/3] Improve disk bus type generation
by Christophe Fergeau 23 Apr '13

23 Apr '13

Here is a new version of this series rebased against master. Christophe

3 9

[libvirt] [libvirt-designer 1/3] Fix 'dependancy' typo in README file
by Christophe Fergeau 23 Apr '13

23 Apr '13

--- README | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README b/README index ed39b9d..b7ca356 100644 --- a/README +++ b/README @@ -24,7 +24,7 @@ Or to install into a private user specific location make make install -The following mandatory dependancies are required in order to +The following mandatory dependencies are required in order to build libvirt-designer libvirt-gconfig >= 0.1.3 (part of libvirt-glib project releases) -- 1.8.1.4

2 5

[libvirt] [PATCH] qemu: Fix setting of memory tunables
by Peter Krempa 23 Apr '13

23 Apr '13

Refactoring done in 19c6ad9ac7e7eb2fd3c8262bff5f087b508ad07f didn't correctly take into account the order cgroup limit modification needs to be done in. This resulted into errors when decreasing the limits. The operations need to take place in this order: decrease hard limit change swap hard limit or change swap hard limit increase hard limit This patch also fixes the check if the hard_limit is less than swap_hard_limit to print better error messages. For this purpose I introduced a helper function virCompareLimitUlong to compare limit values where value of 0 is equal to unlimited. Additionally the check is now applied also when the user does not provide all of the tunables through the API and in that case the currently set values are used. This patch resolves: https://bugzilla.redhat.com/show_bug.cgi?id=950478 --- src/libvirt_private.syms | 1 + src/qemu/qemu_driver.c | 95 ++++++++++++++++++++++++------------------------ src/util/virutil.c | 12 ++++++ src/util/virutil.h | 2 + 4 files changed, 63 insertions(+), 47 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 30fdcd7..ec644ee 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1835,6 +1835,7 @@ safezero; virArgvToString; virAsprintf; virBuildPathInternal; +virCompareLimitUlong; virDirCreate; virDoubleToStr; virEnumFromString; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index cee5557..a174fb0 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -7103,11 +7103,11 @@ qemuDomainSetMemoryParameters(virDomainPtr dom, virDomainDefPtr persistentDef = NULL; virDomainObjPtr vm = NULL; unsigned long long swap_hard_limit; - unsigned long long memory_hard_limit; - unsigned long long memory_soft_limit; + unsigned long long hard_limit = 0; + unsigned long long soft_limit = 0; bool set_swap_hard_limit = false; - bool set_memory_hard_limit = false; - bool set_memory_soft_limit = false; + bool set_hard_limit = false; + bool set_soft_limit = false; virQEMUDriverConfigPtr cfg = NULL; int ret = -1; int rc; @@ -7157,62 +7157,63 @@ qemuDomainSetMemoryParameters(virDomainPtr dom, set_ ## VALUE = true; VIR_GET_LIMIT_PARAMETER(VIR_DOMAIN_MEMORY_SWAP_HARD_LIMIT, swap_hard_limit) - VIR_GET_LIMIT_PARAMETER(VIR_DOMAIN_MEMORY_HARD_LIMIT, memory_hard_limit) - VIR_GET_LIMIT_PARAMETER(VIR_DOMAIN_MEMORY_SOFT_LIMIT, memory_soft_limit) + VIR_GET_LIMIT_PARAMETER(VIR_DOMAIN_MEMORY_HARD_LIMIT, hard_limit) + VIR_GET_LIMIT_PARAMETER(VIR_DOMAIN_MEMORY_SOFT_LIMIT, soft_limit) #undef VIR_GET_LIMIT_PARAMETER + /* Swap hard limit must be greater than hard limit. + * Note that limit of 0 denotes unlimited */ + if (set_swap_hard_limit || set_hard_limit) { + unsigned long long mem_limit = vm->def->mem.hard_limit; + unsigned long long swap_limit = vm->def->mem.swap_hard_limit; - /* It will fail if hard limit greater than swap hard limit anyway */ - if (set_swap_hard_limit && set_memory_hard_limit && - memory_hard_limit > swap_hard_limit) { - virReportError(VIR_ERR_INVALID_ARG, "%s", - _("memory hard_limit tunable value must be lower than " - "swap_hard_limit")); - goto cleanup; - } + if (set_swap_hard_limit) + swap_limit = swap_hard_limit; - if (set_swap_hard_limit) { - if (flags & VIR_DOMAIN_AFFECT_LIVE) { - if ((rc = virCgroupSetMemSwapHardLimit(priv->cgroup, swap_hard_limit)) < 0) { - virReportSystemError(-rc, "%s", - _("unable to set memory swap_hard_limit tunable")); - goto cleanup; - } - vm->def->mem.swap_hard_limit = swap_hard_limit; + if (set_hard_limit) + mem_limit = hard_limit; + + if (virCompareLimitUlong(mem_limit, swap_limit) > 0) { + virReportError(VIR_ERR_INVALID_ARG, "%s", + _("memory hard_limit tunable value must be lower " + "than swap_hard_limit")); + goto cleanup; } + } - if (flags & VIR_DOMAIN_AFFECT_CONFIG) - persistentDef->mem.swap_hard_limit = swap_hard_limit; +#define QEMU_SET_MEM_PARAMETER(FUNC, VALUE) \ + if (set_ ## VALUE) { \ + if (flags & VIR_DOMAIN_AFFECT_LIVE) { \ + if ((rc = FUNC(priv->cgroup, VALUE)) < 0) { \ + virReportSystemError(-rc, _("unable to set memory %s tunable"), \ + #VALUE); \ + \ + goto cleanup; \ + } \ + vm->def->mem.VALUE = VALUE; \ + } \ + \ + if (flags & VIR_DOMAIN_AFFECT_CONFIG) \ + persistentDef->mem.VALUE = VALUE; \ } - if (set_memory_hard_limit) { - if (flags & VIR_DOMAIN_AFFECT_LIVE) { - if ((rc = virCgroupSetMemoryHardLimit(priv->cgroup, memory_hard_limit)) < 0) { - virReportSystemError(-rc, "%s", - _("unable to set memory hard_limit tunable")); - goto cleanup; - } - vm->def->mem.hard_limit = memory_hard_limit; - } + /* Soft limit doesn't clash with the others */ + QEMU_SET_MEM_PARAMETER(virCgroupSetMemorySoftLimit, soft_limit); - if (flags & VIR_DOMAIN_AFFECT_CONFIG) - persistentDef->mem.hard_limit = memory_hard_limit; + /* set hard limit before swap hard limit if decreasing it */ + if (virCompareLimitUlong(vm->def->mem.hard_limit, hard_limit) > 0) { + QEMU_SET_MEM_PARAMETER(virCgroupSetMemoryHardLimit, hard_limit); + /* inhibit changing the limit a second time */ + set_hard_limit = false; } - if (set_memory_soft_limit) { - if (flags & VIR_DOMAIN_AFFECT_LIVE) { - if ((rc = virCgroupSetMemorySoftLimit(priv->cgroup, memory_soft_limit)) < 0) { - virReportSystemError(-rc, "%s", - _("unable to set memory soft_limit tunable")); - goto cleanup; - } - vm->def->mem.soft_limit = memory_soft_limit; - } + QEMU_SET_MEM_PARAMETER(virCgroupSetMemSwapHardLimit, swap_hard_limit); - if (flags & VIR_DOMAIN_AFFECT_CONFIG) - persistentDef->mem.soft_limit = memory_soft_limit; - } + /* otherwise increase it after swap hard limit */ + QEMU_SET_MEM_PARAMETER(virCgroupSetMemoryHardLimit, hard_limit); + +#undef QEMU_SET_MEM_PARAMETER if (flags & VIR_DOMAIN_AFFECT_CONFIG && virDomainSaveConfig(cfg->configDir, persistentDef) < 0) diff --git a/src/util/virutil.c b/src/util/virutil.c index 9cc3672..491a004 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -3832,3 +3832,15 @@ virFindFCHostCapableVport(const char *sysfs_prefix ATTRIBUTE_UNUSED) } #endif /* __linux__ */ + +int +virCompareLimitUlong(unsigned long long a, unsigned long b) +{ + if (a == b) + return 0; + + if (a == 0 || a > b) + return 1; + + return -1; +} diff --git a/src/util/virutil.h b/src/util/virutil.h index 7b37d65..39033db 100644 --- a/src/util/virutil.h +++ b/src/util/virutil.h @@ -324,4 +324,6 @@ char *virGetFCHostNameByWWN(const char *sysfs_prefix, char *virFindFCHostCapableVport(const char *sysfs_prefix); +int virCompareLimitUlong(unsigned long long a, unsigned long b); + #endif /* __VIR_UTIL_H__ */ -- 1.8.1.5

2 2

[libvirt] Add UID/GID support to virt-sandbox-service
by dwalsh＠redhat.com 23 Apr '13

23 Apr '13

This patch set is adding support for UID/GID/USERNAME/USERDIR for use with openshift containers Also fixes virt-sandbox-service to not complain if the destdir has been precreated. Finally we also do not want excess processes running withing containers (/bin/sh). [sandbox PATCH 1/3] Add UID/GID support for use with interactive [sandbox PATCH 2/3] Only create the destination path if it does not [sandbox PATCH 3/3] Do not run a shell within a lxc container by

2 6