- Devel - libvirt List Archives

[libvirt] New repositories and comments for Code Search
by Richard W.M. Jones 14 Oct '11

14 Oct '11

Enter your code repository type (SVN, CVS, Hg, Git, Archive) and URL here: GIT git://libvirt.org/libvirt.git Libvirt is the virtualization API. Comment: You already have some random forks of libvirt.git in the archive, but as far as I can see, not the real upstream git repo. GIT git://git.annexia.org/git/libguestfs.git libguestfs is a set of tools for manipulating virtual machine disk images. Comment: You are already taking the Fedora packaging git archive (eg: git://pkgs.fedoraproject.org/libguestfs.git) but those aren't the sources, they contain the spec files and tarballs. Now, if you were to just unpack all those tarballs, you'd have the source for 10,000s of verified open source projects!

2 1

[libvirt] [PATCH] macvtap: plug memory leak for 802.1Qbh
by Eric Blake 14 Oct '11

14 Oct '11

Detected by Coverity. Leak present since commit ca3b22b. * src/util/macvtap.c (doPortProfileOp8021Qbh): Release device name. --- getPhysfnDev allocates physfndev, but nothing freed it. Pushing under the trivial rule. src/util/macvtap.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/util/macvtap.c b/src/util/macvtap.c index 9bf7fa6..a020c90 100644 --- a/src/util/macvtap.c +++ b/src/util/macvtap.c @@ -983,27 +983,27 @@ doPortProfileOp8021Qbh(const char *ifname, # ifndef IFLA_VF_PORT_MAX (void)ifname; (void)macaddr; (void)virtPort; (void)vm_uuid; (void)virtPortOp; macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", _("Kernel VF Port support was missing at compile time.")); rc = 1; # else /* IFLA_VF_PORT_MAX */ - char *physfndev; + char *physfndev = NULL; unsigned char hostuuid[VIR_UUID_BUFLEN]; int32_t vf; bool nltarget_kernel = true; int ifindex; int vlanid = -1; rc = getPhysfnDev(ifname, &vf, &physfndev); if (rc) goto err_exit; if (ifaceGetIndex(true, physfndev, &ifindex) < 0) { rc = 1; goto err_exit; @@ -1049,26 +1049,27 @@ doPortProfileOp8021Qbh(const char *ifname, NULL, NULL, vf, PORT_REQUEST_DISASSOCIATE); break; default: macvtapError(VIR_ERR_INTERNAL_ERROR, _("operation type %d not supported"), virtPortOp); rc = 1; } err_exit: + VIR_FREE(physfndev); # endif /* IFLA_VF_PORT_MAX */ return rc; } /** * vpAssociatePortProfile * * @macvtap_ifname: The name of the macvtap device * @virtPort: pointer to the object holding port profile parameters * @vmuuid : the UUID of the virtual machine * @vmOp : The VM operation (i.e., create, no-op) -- 1.7.4.4

2 5

Re: [libvirt] [PATCH] Bug Fix: Fix getPhysfnDev to always allocate memory for physfndev on success
by Eric Blake 14 Oct '11

14 Oct '11

On 10/14/2011 04:12 PM, Roopa Prabhu wrote: > From: Roopa Prabhu<roprabhu(a)cisco.com> > > getPhysfnDev was conditionally allocating memory for physfndev on success. > Recently added unconditional free in the caller of getPhysfnDev > can result in a crash. So fix getPhysfnDev to unconditionally allocate > memory for physfndev on success > > Signed-off-by: Roopa Prabhu<roprabhu(a)cisco.com> > Acked-by: Eric Blake<eblake(a)redhat.com> > --- > src/util/macvtap.c | 6 +++++- > 1 files changed, 5 insertions(+), 1 deletions(-) Already pushed. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] [PATCH] Clarify semantics of virDomainMigrate2
by Jiri Denemark 14 Oct '11

14 Oct '11

Explicitly disallow conflicts between domain name from dxml and dname. --- src/libvirt.c | 3 +++ src/qemu/qemu_driver.c | 4 ++-- src/qemu/qemu_migration.c | 17 ++++++++++++++--- src/qemu/qemu_migration.h | 1 + 4 files changed, 20 insertions(+), 5 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index f1e6a6b..69cffc5 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -4854,6 +4854,9 @@ error: * in accessing the underlying storage. The migration will fail * if @dxml would cause any guest-visible changes. Pass NULL * if no changes are needed to the XML between source and destination. + * @dxml cannot be used to rename the domain during migration (use + * @dname for that purpose). Domain name in @dxml must either match the + * original domain name or @dname if it was specified. * * Returns the new domain object if the migration was successful, * or NULL in case of error. Note that the new domain object diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 98f4d7f..f833655 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -8043,7 +8043,7 @@ qemuDomainMigrateBegin3(virDomainPtr domain, char **cookieout, int *cookieoutlen, unsigned long flags, - const char *dname ATTRIBUTE_UNUSED, + const char *dname, unsigned long resource ATTRIBUTE_UNUSED) { struct qemud_driver *driver = domain->conn->privateData; @@ -8083,7 +8083,7 @@ qemuDomainMigrateBegin3(virDomainPtr domain, if (qemuDomainCheckEjectableMedia(driver, vm) < 0) goto endjob; - if (!(xml = qemuMigrationBegin(driver, vm, xmlin, + if (!(xml = qemuMigrationBegin(driver, vm, xmlin, dname, cookieout, cookieoutlen))) goto endjob; diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index f0a0e49..3770bb6 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -993,6 +993,7 @@ qemuDomainMigrateGraphicsRelocate(struct qemud_driver *driver, char *qemuMigrationBegin(struct qemud_driver *driver, virDomainObjPtr vm, const char *xmlin, + const char *dname, char **cookieout, int *cookieoutlen) { @@ -1001,8 +1002,10 @@ char *qemuMigrationBegin(struct qemud_driver *driver, virDomainDefPtr def = NULL; qemuDomainObjPrivatePtr priv = vm->privateData; - VIR_DEBUG("driver=%p, vm=%p, xmlin=%s, cookieout=%p, cookieoutlen=%p", - driver, vm, NULLSTR(xmlin), cookieout, cookieoutlen); + VIR_DEBUG("driver=%p, vm=%p, xmlin=%s, dname=%s," + " cookieout=%p, cookieoutlen=%p", + driver, vm, NULLSTR(xmlin), NULLSTR(dname), + cookieout, cookieoutlen); /* Only set the phase if we are inside QEMU_ASYNC_JOB_MIGRATION_OUT. * Otherwise we will start the async job later in the perform phase losing @@ -1028,6 +1031,14 @@ char *qemuMigrationBegin(struct qemud_driver *driver, VIR_DOMAIN_XML_INACTIVE))) goto cleanup; + if (STRNEQ(def->name, vm->def->name) && + (!dname || STRNEQ(def->name, dname))) { + qemuReportError(VIR_ERR_INVALID_ARG, "%s", + "target domain name doesn't match source name" + " nor destination name"); + goto cleanup; + } + if (!virDomainDefCheckABIStability(vm->def, def)) goto cleanup; @@ -2012,7 +2023,7 @@ static int doPeer2PeerMigrate3(struct qemud_driver *driver, * bit here, because we are already running inside the context of * a single job. */ - dom_xml = qemuMigrationBegin(driver, vm, xmlin, + dom_xml = qemuMigrationBegin(driver, vm, xmlin, dname, &cookieout, &cookieoutlen); if (!dom_xml) goto cleanup; diff --git a/src/qemu/qemu_migration.h b/src/qemu/qemu_migration.h index ec70422..f806ca1 100644 --- a/src/qemu/qemu_migration.h +++ b/src/qemu/qemu_migration.h @@ -79,6 +79,7 @@ int qemuMigrationSetOffline(struct qemud_driver *driver, char *qemuMigrationBegin(struct qemud_driver *driver, virDomainObjPtr vm, const char *xmlin, + const char *dname, char **cookieout, int *cookieoutlen); -- 1.7.7

2 2

[libvirt] [PATCH] docs: fix network XML documentation
by Laine Stump 14 Oct '11

14 Oct '11

A few people have attempted to use the new forwarding modes with older versions of libvirt. The docs where the modes are described have always stated the minimum required libvirt version, but the examples at the end didn't, which I believe is what has caused the confusion. Similarly, the section on portgroups now has a version tag added at the beginning. I also noticed that there was no example of defining a <dns> hostname, so I added one, as well as making the domain name example more recognizable (by adding ".com" to the domain). --- docs/formatnetwork.html.in | 15 ++++++++++++--- 1 files changed, 12 insertions(+), 3 deletions(-) diff --git a/docs/formatnetwork.html.in b/docs/formatnetwork.html.in index 99031d0..e06392b 100644 --- a/docs/formatnetwork.html.in +++ b/docs/formatnetwork.html.in @@ -58,7 +58,7 @@ <pre> ... <bridge name="virbr0" stp="on" delay="5"/> - <domain name="example"/> + <domain name="example.com"/> <forward mode="nat" dev="eth0"/> ...</pre> @@ -214,7 +214,7 @@ As mentioned above, a <code><forward></code> element can have multiple <code><interface></code> subelements, each one giving the name of a physical interface that can be used - for this network<span class="since">Since 0.9.4</span>: + for this network <span class="since">Since 0.9.4</span>: <pre> ... <forward mode='passthrough'> @@ -301,9 +301,11 @@ ...</pre> <p> + <span class="since">Since 0.9.4</span> A portgroup provides a method of easily putting guest connections to the network into different classes, with each - class potentially having a different level/type of service. Each + class potentially having a different level/type of service. + <span class="since">Since 0.9.4</span> Each network can have multiple portgroup elements (and one of those can optionally be designated as the 'default' portgroup for the network), and each portgroup has a name, as well as various @@ -340,8 +342,12 @@ <pre> ... <mac address='00:16:3E:5D:C7:9E'/> + <domain name="example.com"/> <dns> <txt name="example" value="example value" /> + <host ip='192.168.122.2'> + <hostname>myhost</hostname> + <hostname>myhostalias</hostname> </dns> <ip address="192.168.122.1" netmask="255.255.255.0"> <dhcp> @@ -542,6 +548,7 @@ <h3><a name="examplesBridge">Using an existing host bridge</a></h3> <p> + <span class="since">Since 0.9.4</span> This shows how to use a pre-existing host bridge "br0". The guests will effectively be directly connected to the physical network (i.e. their IP addresses will all be on the subnet of @@ -559,6 +566,8 @@ <h3><a name="examplesDirect">Using a macvtap "direct" connection</a></h3> <p> + <span class="since">Since 0.9.4, QEMU and KVM only, requires + Linux kernel 2.6.34 or newer</span> This shows how to use macvtap to connect to the physical network directly through one of a group of physical devices (without using a host bridge device). As with the host bridge network, -- 1.7.6.4

2 1

[libvirt] [PATCH] util: Make getaddrinfo failure nonfatal in virGetHostname
by Jiri Denemark 14 Oct '11

14 Oct '11

Setting a hostname that cannot be resolved is not the best configuration but since virGetHostname only calls getaddrinfo to get host's canonical name and we do not fail if the returned canonical name is NULL or "localhost", there is no reason why we should fail if getaddrinfo itself fails. --- src/util/util.c | 16 ++++++++-------- 1 files changed, 8 insertions(+), 8 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index 1ff287d..fd4d7fa 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -1858,10 +1858,10 @@ char *virIndexToDiskName(int idx, const char *prefix) * try to resolve this to a fully-qualified name. Therefore we pass it * to getaddrinfo(). There are two possible responses: * a) getaddrinfo() resolves to a FQDN - return the FQDN - * b) getaddrinfo() resolves to localhost - in this case, the data we got - * from gethostname() is actually more useful than what we got from - * getaddrinfo(). Return the value from gethostname() and hope for - * the best. + * b) getaddrinfo() files or resolves to localhost - in this case, the + * data we got from gethostname() is actually more useful than what + * we got from getaddrinfo(). Return the value from gethostname() + * and hope for the best. */ char *virGetHostname(virConnectPtr conn ATTRIBUTE_UNUSED) { @@ -1897,10 +1897,10 @@ char *virGetHostname(virConnectPtr conn ATTRIBUTE_UNUSED) hints.ai_family = AF_UNSPEC; r = getaddrinfo(hostname, NULL, &hints, &info); if (r != 0) { - virUtilError(VIR_ERR_INTERNAL_ERROR, - _("getaddrinfo failed for '%s': %s"), - hostname, gai_strerror(r)); - return NULL; + VIR_WARN("getaddrinfo failed for '%s': %s", + hostname, gai_strerror(r)); + result = strdup(hostname); + goto check_and_return; } /* Tell static analyzers about getaddrinfo semantics. */ -- 1.7.7

2 2

[libvirt] [PATCH] qemu: Make sure BeginJob is always followed by EndJob
by Jiri Denemark 14 Oct '11

14 Oct '11

Otherwise we can end up with a dangling job that can only be cleared by restarting libvirtd. --- src/qemu/qemu_driver.c | 24 ++++++++++++------------ 1 files changed, 12 insertions(+), 12 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 0e307e1..98f4d7f 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2010,42 +2010,42 @@ static int qemuDomainSendKey(virDomainPtr domain, if (!vm) { char uuidstr[VIR_UUID_STRING_BUFLEN]; virUUIDFormat(domain->uuid, uuidstr); qemuReportError(VIR_ERR_NO_DOMAIN, _("no domain with matching uuid '%s'"), uuidstr); goto cleanup; } priv = vm->privateData; if (qemuDomainObjBeginJobWithDriver(driver, vm, QEMU_JOB_MODIFY) < 0) goto cleanup; if (!virDomainObjIsActive(vm)) { qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); - goto cleanup; + goto endjob; } qemuDomainObjEnterMonitorWithDriver(driver, vm); ret = qemuMonitorSendKey(priv->mon, holdtime, keycodes, nkeycodes); qemuDomainObjExitMonitorWithDriver(driver, vm); - if (qemuDomainObjEndJob(driver, vm) == 0) { + +endjob: + if (qemuDomainObjEndJob(driver, vm) == 0) vm = NULL; - goto cleanup; - } cleanup: if (vm) virDomainObjUnlock(vm); qemuDriverUnlock(driver); return ret; } static int qemudDomainGetInfo(virDomainPtr dom, virDomainInfoPtr info) { struct qemud_driver *driver = dom->conn->privateData; virDomainObjPtr vm; int ret = -1; int err; unsigned long balloon; @@ -7208,103 +7208,103 @@ qemudDomainBlockStatsFlags (virDomainPtr dom, &errs); qemuDomainObjExitMonitor(driver, vm); if (ret < 0) goto endjob; /* Field 'errs' is meaningless for QEMU, won't set it. */ for (i = 0; i < *nparams; i++) { virTypedParameterPtr param = &params[i]; switch (i) { case 0: /* fill write_bytes here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_WRITE_BYTES) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field write bytes too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = wr_bytes; break; case 1: /* fill wr_operations here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_WRITE_REQ) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field write requests too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = wr_req; break; case 2: /* fill read_bytes here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_READ_BYTES) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field read bytes too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = rd_bytes; break; case 3: /* fill rd_operations here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_READ_REQ) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field read requests too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = rd_req; break; case 4: /* fill flush_operations here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_FLUSH_REQ) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field flush requests too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = flush_req; break; case 5: /* fill wr_total_times_ns here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_WRITE_TOTAL_TIMES) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field write total times too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = wr_total_times; break; case 6: /* fill rd_total_times_ns here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_READ_TOTAL_TIMES) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field read total times too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = rd_total_times; break; case 7: /* fill flush_total_times_ns here */ if (virStrcpyStatic(param->field, VIR_DOMAIN_BLOCK_STATS_READ_TOTAL_TIMES) == NULL) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Field flush total times too long for destination")); - goto cleanup; + goto endjob; } param->type = VIR_TYPED_PARAM_LLONG; param->value.l = flush_total_times; break; default: break; /* should not hit here */ } } endjob: if (qemuDomainObjEndJob(driver, vm) == 0) vm = NULL; cleanup: -- 1.7.7

2 2

[libvirt] [PATCH] qemu: Log debug messages when changing job
by Jiri Denemark 14 Oct '11

14 Oct '11

Log debug messages anytime we call *BeginJob* or *EndJob* so that it's easier to spot incorrect usage of domain job APIs. --- src/qemu/qemu_domain.c | 12 ++++++++++++ 1 files changed, 12 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 85bebd6..5abc900 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -761,8 +761,13 @@ retry: qemuDomainObjResetJob(priv); if (job != QEMU_JOB_ASYNC) { + VIR_DEBUG("Starting job: %s (async=%s)", + qemuDomainJobTypeToString(job), + qemuDomainAsyncJobTypeToString(priv->job.asyncJob)); priv->job.active = job; } else { + VIR_DEBUG("Starting async job: %s", + qemuDomainAsyncJobTypeToString(asyncJob)); qemuDomainObjResetAsyncJob(priv); priv->job.asyncJob = asyncJob; priv->job.start = now; @@ -873,6 +878,10 @@ int qemuDomainObjEndJob(struct qemud_driver *driver, virDomainObjPtr obj) priv->jobs_queued--; + VIR_DEBUG("Stopping job: %s (async=%s)", + qemuDomainJobTypeToString(priv->job.active), + qemuDomainAsyncJobTypeToString(priv->job.asyncJob)); + qemuDomainObjResetJob(priv); qemuDomainObjSaveJob(driver, obj); virCondSignal(&priv->job.cond); @@ -887,6 +896,9 @@ qemuDomainObjEndAsyncJob(struct qemud_driver *driver, virDomainObjPtr obj) priv->jobs_queued--; + VIR_DEBUG("Stopping async job: %s", + qemuDomainAsyncJobTypeToString(priv->job.asyncJob)); + qemuDomainObjResetAsyncJob(priv); qemuDomainObjSaveJob(driver, obj); virCondBroadcast(&priv->job.asyncCond); -- 1.7.7

2 2

[libvirt] [PATCH v2 0/2] Xen: skip xenHypervisor version init in tests
by Philipp Hahn 14 Oct '11

14 Oct '11

Several tests fail when run as root on a Xen-dom0-system, since virInitialize() then succeeds to open /proc/xen/privcmd and returns the actual supported features instead of the faked one when calling xenHypervisorMakeCapabilitiesInternal(). Since Xen-3.3 supports additional features like "hap" and "viridian", the xencapstest fails. v2: Skip initialization of static version variables for Xen Hypervisor and use provided values for unit test cases. Philipp Hahn (2): Xen: move versions to struct Xen: Fake versions in xencapstest src/xen/xen_driver.c | 2 +- src/xen/xen_hypervisor.c | 307 ++++++++++++++++++++++++---------------------- src/xen/xen_hypervisor.h | 10 ++- tests/xencapstest.c | 10 ++ 4 files changed, 180 insertions(+), 149 deletions(-)

2 4

[libvirt] [PATCH] esx: drop dead code to silence Coverity
by Eric Blake 14 Oct '11

14 Oct '11

Coverity detected that the only way to get to the cleanup label is if objectSpec had been successfully allocated, so the null check was dead code. * src/esx/esx_vi.c (esxVI_LookupObjectContentByType): Drop redundant null check. --- I'll wait for a review on this one. src/esx/esx_vi.c | 7 ++----- 1 files changed, 2 insertions(+), 5 deletions(-) diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index 4a8c709..8bcd76c 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -1763,11 +1763,8 @@ esxVI_LookupObjectContentByType(esxVI_Context *ctx, * Remove values given by the caller from the data structures to prevent * them from being freed by the call to esxVI_PropertyFilterSpec_Free(). */ - if (objectSpec != NULL) { - objectSpec->obj = NULL; - objectSpec->selectSet = NULL; - } - + objectSpec->obj = NULL; + objectSpec->selectSet = NULL; if (propertySpec != NULL) { propertySpec->type = NULL; propertySpec->pathSet = NULL; -- 1.7.4.4

3 3