- Devel - libvirt List Archives

[libvirt] [PATCH] storage: document existing pools
by Eric Blake 16 Oct '13

16 Oct '13

We forgot to document several pool types. Furthermore, the RNG grammar did not allow arbitrary interleaving, which makes it harder than necessary to create a new pool from XML. * docs/formatstorage.html.in: Add docs for scsi, mpath, rbd, and sheepdog. * docs/schemas/storagepool.rng: Allow interleaving. * tests/storagepoolxml2xmlin/pool-sheepdog.xml: Test interleave. * tests/storagepoolxml2xmlin/pool-iscsi-auth.xml: Likewise. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- docs/formatstorage.html.in | 44 +-- docs/schemas/storagepool.rng | 356 ++++++++++++++----------- tests/storagepoolxml2xmlin/pool-iscsi-auth.xml | 12 +- tests/storagepoolxml2xmlin/pool-sheepdog.xml | 6 +- 4 files changed, 238 insertions(+), 180 deletions(-) diff --git a/docs/formatstorage.html.in b/docs/formatstorage.html.in index 3300d67..6f79b0c 100644 --- a/docs/formatstorage.html.in +++ b/docs/formatstorage.html.in @@ -18,9 +18,12 @@ The top level tag for a storage pool document is 'pool'. It has a single attribute <code>type</code>, which is one of <code>dir</code>, <code>fs</code>, <code>netfs</code>, <code>disk</code>, - <code>iscsi</code>, <code>logical</code>. This corresponds to the + <code>iscsi</code>, <code>logical</code>, <code>scsi</code> + (all <span class="since">since 0.4.1</span>), <code>mpath</code> + (<span class="since">since 0.7.1</span>), <code>rbd</code> + (<span class="since">since 0.9.13</span>), or <code>sheepdog</code> + (<span class="since">since 0.10.0</span>). This corresponds to the storage backend drivers listed further along in this document. - The storage pool XML format is available <span class="since">since 0.4.1</span> </p> <h3><a name="StoragePoolFirst">General metadata</a></h3> @@ -64,7 +67,8 @@ <p> A single <code>source</code> element is contained within the top level <code>pool</code> element. This tag is used to describe the source of - the storage pool. It can contain the following child elements: + the storage pool. The set of child elements that it will contain + depend on the pool type, but come from the following child elements: </p> <pre> @@ -89,17 +93,18 @@ <dl> <dt><code>device</code></dt> - <dd>Provides the source for pools backed by physical devices. + <dd>Provides the source for pools backed by physical devices + (fs, logical, disk, iscsi). May be repeated multiple times depending on backend driver. Contains a single attribute <code>path</code> which is the fully qualified path to the block device node. <span class="since">Since 0.4.1</span></dd> - <dt><code>directory</code></dt> - <dd>Provides the source for pools backed by directories. May + <dt><code>dir</code></dt> + <dd>Provides the source for pools backed by directories (dir). May only occur once. Contains a single attribute <code>path</code> which is the fully qualified path to the backing directory. <span class="since">Since 0.4.1</span></dd> <dt><code>adapter</code></dt> - <dd>Provides the source for pools backed by SCSI adapters. May + <dd>Provides the source for pools backed by SCSI adapters (scsi). May only occur once. Attribute <code>name</code> is the SCSI adapter name (ex. "scsi_host1". NB, although a name such as "host1" is still supported for backwards compatibility, it is not recommended). @@ -120,7 +125,8 @@ <span class="since">Since 0.6.2</span></dd> <dt><code>host</code></dt> <dd>Provides the source for pools backed by storage from a - remote server. Will be used in combination with a <code>directory</code> + remote server (netfs, iscsi, rbd, sheepdog). Will be + used in combination with a <code>directory</code> or <code>device</code> element. Contains an attribute <code>name</code> which is the hostname or IP address of the server. May optionally contain a <code>port</code> attribute for the protocol specific @@ -128,18 +134,19 @@ <dt><code>auth</code></dt> <dd>If present, the <code>auth</code> element provides the authentication credentials needed to access the source by the - setting of the <code>type</code> attribute. The <code>type</code> - must be either "chap" or "ceph". Additionally a mandatory attribute + setting of the <code>type</code> attribute (iscsi, + rbd). The <code>type</code> + must be either "chap" or "ceph". Use "ceph" for + Ceph RBD (Rados Block Device) network sources and use "iscsi" for CHAP + (Challenge-Handshake Authentication Protocol) iSCSI + targets. Additionally a mandatory attribute <code>username</code> identifies the username to use during authentication as well as a sub-element <code>secret</code> with a mandatory attribute <code>type</code>, to tie back to a <a href="formatsecret.html">libvirt secret object</a> that holds the actual password or other credentials. The domain XML intentionally does not expose the password, only the reference - to the object that manages the password. The secret element - <code>type</code> must be either "ceph" or "iscsi". Use "ceph" for - Ceph RBD (Rados Block Device) network sources and use "iscsi" for CHAP - (Challenge-Handshake Authentication Protocol) iSCSI targets. + to the object that manages the password. The <code>secret</code> element requires either a <code>uuid</code> attribute with the UUID of the secret object or a <code>usage</code> attribute matching the key that was specified in the @@ -148,11 +155,12 @@ </dd> <dt><code>name</code></dt> <dd>Provides the source for pools backed by storage from a - named element (e.g., a logical volume group name). + named element (logical, rbd, sheepdog). Contains a string identifier. <span class="since">Since 0.4.5</span></dd> <dt><code>format</code></dt> - <dd>Provides information about the format of the pool. This + <dd>Provides information about the format of the pool (fs, + netfs, disk, logical). This contains a single attribute <code>type</code> whose value is backend specific. This is typically used to indicate filesystem type, or network filesystem type, or partition table type, or @@ -174,7 +182,9 @@ <p> A single <code>target</code> element is contained within the top level - <code>pool</code> element. This tag is used to describe the mapping of + <code>pool</code> element for some types of pools (dir, fs, + netfs, logical, disk, iscsi, scsi, mpath). This tag is used to + describe the mapping of the storage pool into the host filesystem. It can contain the following child elements: </p> diff --git a/docs/schemas/storagepool.rng b/docs/schemas/storagepool.rng index a4ef5af..66d3c22 100644 --- a/docs/schemas/storagepool.rng +++ b/docs/schemas/storagepool.rng @@ -29,192 +29,224 @@ <attribute name='type'> <value>dir</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcedir'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcedir'/> + <ref name='target'/> + </interleave> </define> <define name='poolfs'> <attribute name='type'> <value>fs</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcefs'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcefs'/> + <ref name='target'/> + </interleave> </define> <define name='poolnetfs'> <attribute name='type'> <value>netfs</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcenetfs'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcenetfs'/> + <ref name='target'/> + </interleave> </define> <define name='poollogical'> <attribute name='type'> <value>logical</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcelogical'/> - <ref name='targetlogical'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcelogical'/> + <ref name='targetlogical'/> + </interleave> </define> <define name='pooldisk'> <attribute name='type'> <value>disk</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcedisk'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcedisk'/> + <ref name='target'/> + </interleave> </define> <define name='pooliscsi'> <attribute name='type'> <value>iscsi</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourceiscsi'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourceiscsi'/> + <ref name='target'/> + </interleave> </define> <define name='poolscsi'> <attribute name='type'> <value>scsi</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcescsi'/> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcescsi'/> + <ref name='target'/> + </interleave> </define> <define name='poolmpath'> <attribute name='type'> <value>mpath</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <optional> - <ref name='sourcempath'/> - </optional> - <ref name='target'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <optional> + <ref name='sourcempath'/> + </optional> + <ref name='target'/> + </interleave> </define> <define name='poolrbd'> <attribute name='type'> <value>rbd</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcerbd'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcerbd'/> + </interleave> </define> <define name='poolsheepdog'> <attribute name='type'> <value>sheepdog</value> </attribute> - <ref name='commonmetadata'/> - <ref name='sizing'/> - <ref name='sourcesheepdog'/> + <interleave> + <ref name='commonmetadata'/> + <ref name='sizing'/> + <ref name='sourcesheepdog'/> + </interleave> </define> <define name='sourceinfovendor'> - <optional> - <element name='vendor'> - <attribute name='name'> - <text/> - </attribute> - </element> - </optional> - <optional> - <element name='product'> - <attribute name='name'> - <text/> - </attribute> - </element> - </optional> + <interleave> + <optional> + <element name='vendor'> + <attribute name='name'> + <text/> + </attribute> + </element> + </optional> + <optional> + <element name='product'> + <attribute name='name'> + <text/> + </attribute> + </element> + </optional> + </interleave> </define> <define name='commonmetadata'> - <element name='name'> - <ref name='genericName'/> - </element> - <optional> - <element name='uuid'> - <ref name='UUID'/> + <interleave> + <element name='name'> + <ref name='genericName'/> </element> - </optional> + <optional> + <element name='uuid'> + <ref name='UUID'/> + </element> + </optional> + </interleave> </define> <define name='sizing'> - <optional> - <element name='capacity'> - <ref name='scaledInteger'/> - </element> - </optional> - <optional> - <element name='allocation'> - <ref name='scaledInteger'/> - </element> - </optional> - <optional> - <element name='available'> - <ref name='scaledInteger'/> - </element> - </optional> + <interleave> + <optional> + <element name='capacity'> + <ref name='scaledInteger'/> + </element> + </optional> + <optional> + <element name='allocation'> + <ref name='scaledInteger'/> + </element> + </optional> + <optional> + <element name='available'> + <ref name='scaledInteger'/> + </element> + </optional> + </interleave> </define> <define name='permissions'> <optional> <element name='permissions'> - <element name='mode'> - <ref name='octalMode'/> - </element> - <element name='owner'> - <choice> - <ref name='unsignedInt'/> - <value>-1</value> - </choice> - </element> - <element name='group'> - <choice> - <ref name='unsignedInt'/> - <value>-1</value> - </choice> - </element> - <optional> - <element name='label'> - <text/> - </element> - </optional> + <interleave> + <element name='mode'> + <ref name='octalMode'/> + </element> + <element name='owner'> + <choice> + <ref name='unsignedInt'/> + <value>-1</value> + </choice> + </element> + <element name='group'> + <choice> + <ref name='unsignedInt'/> + <value>-1</value> + </choice> + </element> + <optional> + <element name='label'> + <text/> + </element> + </optional> + </interleave> </element> </optional> </define> <define name='target'> <element name='target'> - <element name='path'> - <ref name='absFilePath'/> - </element> - <ref name='permissions'/> + <interleave> + <element name='path'> + <ref name='absFilePath'/> + </element> + <ref name='permissions'/> + </interleave> </element> </define> <define name='targetlogical'> <element name='target'> - <optional> - <element name='path'> - <ref name='absFilePath'/> - </element> - </optional> - <ref name='permissions'/> + <interleave> + <optional> + <element name='path'> + <ref name='absFilePath'/> + </element> + </optional> + <ref name='permissions'/> + </interleave> </element> </define> @@ -413,74 +445,86 @@ <define name='sourcefs'> <element name='source'> - <ref name='sourceinfodev'/> - <ref name='sourcefmtfs'/> - <optional> - <ref name='sourceinfovendor'/> - </optional> + <interleave> + <ref name='sourceinfodev'/> + <ref name='sourcefmtfs'/> + <optional> + <ref name='sourceinfovendor'/> + </optional> + </interleave> </element> </define> <define name='sourcenetfs'> <element name='source'> - <ref name='sourceinfohost'/> - <ref name='sourceinfodir'/> - <ref name='sourcefmtnetfs'/> - <optional> - <ref name='sourceinfovendor'/> - </optional> + <interleave> + <ref name='sourceinfohost'/> + <ref name='sourceinfodir'/> + <ref name='sourcefmtnetfs'/> + <optional> + <ref name='sourceinfovendor'/> + </optional> + </interleave> </element> </define> <define name='sourcelogical'> <element name='source'> - <oneOrMore> + <interleave> + <oneOrMore> + <optional> + <ref name='sourceinfoname'/> + </optional> + <optional> + <ref name='sourceinfodev'/> + </optional> + </oneOrMore> + <ref name='sourcefmtlogical'/> <optional> - <ref name='sourceinfoname'/> - </optional> - <optional> - <ref name='sourceinfodev'/> + <ref name='sourceinfovendor'/> </optional> - </oneOrMore> - <ref name='sourcefmtlogical'/> - <optional> - <ref name='sourceinfovendor'/> - </optional> + </interleave> </element> </define> <define name='sourcedisk'> <element name='source'> - <ref name='sourceinfodev'/> - <ref name='sourcefmtdisk'/> - <optional> - <ref name='sourceinfovendor'/> - </optional> + <interleave> + <ref name='sourceinfodev'/> + <ref name='sourcefmtdisk'/> + <optional> + <ref name='sourceinfovendor'/> + </optional> + </interleave> </element> </define> <define name='sourceiscsi'> <element name='source'> - <ref name='sourceinfohost'/> - <ref name='sourceinfodev'/> - <optional> - <ref name='initiatorinfo'/> - </optional> - <optional> - <ref name='sourceinfoauth'/> - </optional> - <optional> - <ref name='sourceinfovendor'/> - </optional> + <interleave> + <ref name='sourceinfohost'/> + <ref name='sourceinfodev'/> + <optional> + <ref name='initiatorinfo'/> + </optional> + <optional> + <ref name='sourceinfoauth'/> + </optional> + <optional> + <ref name='sourceinfovendor'/> + </optional> + </interleave> </element> </define> <define name='sourcescsi'> <element name='source'> - <ref name='sourceinfoadapter'/> - <optional> - <ref name='sourceinfovendor'/> - </optional> + <interleave> + <ref name='sourceinfoadapter'/> + <optional> + <ref name='sourceinfovendor'/> + </optional> + </interleave> </element> </define> @@ -492,18 +536,22 @@ <define name='sourcerbd'> <element name='source'> - <ref name='sourceinfoname'/> - <ref name='sourceinfohost'/> - <optional> - <ref name='sourceinfoauth'/> - </optional> + <interleave> + <ref name='sourceinfoname'/> + <ref name='sourceinfohost'/> + <optional> + <ref name='sourceinfoauth'/> + </optional> + </interleave> </element> </define> <define name='sourcesheepdog'> <element name='source'> - <ref name='sourceinfohost'/> - <ref name='sourceinfoname'/> + <interleave> + <ref name='sourceinfohost'/> + <ref name='sourceinfoname'/> + </interleave> </element> </define> diff --git a/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml b/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml index c81eb60..5e5d336 100644 --- a/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml +++ b/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml @@ -1,19 +1,19 @@ <pool type='iscsi'> - <name>virtimages</name> - <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> <source> - <host name="iscsi.example.com"/> - <device path="demo-target"/> <auth type='chap' username='admin'> <secret uuid='2ec115d7-3a88-3ceb-bc12-0ac909a6fd87'/> </auth> + <device path="demo-target"/> + <host name="iscsi.example.com"/> </source> + <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> <target> - <path>/dev/disk/by-path</path> <permissions> - <mode>0700</mode> <owner>0</owner> + <mode>0700</mode> <group>0</group> </permissions> + <path>/dev/disk/by-path</path> </target> + <name>virtimages</name> </pool> diff --git a/tests/storagepoolxml2xmlin/pool-sheepdog.xml b/tests/storagepoolxml2xmlin/pool-sheepdog.xml index 1287047..49b6baf 100644 --- a/tests/storagepoolxml2xmlin/pool-sheepdog.xml +++ b/tests/storagepoolxml2xmlin/pool-sheepdog.xml @@ -1,8 +1,8 @@ <pool type='sheepdog'> - <name>sheepdog</name> - <uuid>65fcba04-5b13-bd93-cff3-52ce48e11ad7</uuid> <source> - <host name='localhost' port='7000'/> <name>sheepdog</name> + <host name='localhost' port='7000'/> </source> + <uuid>65fcba04-5b13-bd93-cff3-52ce48e11ad7</uuid> + <name>sheepdog</name> </pool> -- 1.8.3.1

2 2

[libvirt] [PATCH] remote-driver: Fix 'leav' typo in comment
by Christophe Fergeau 16 Oct '13

16 Oct '13

--- Pushed under the trivial rule. src/remote/remote_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index 96ccb99..87ef5a9 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -3721,7 +3721,7 @@ static int remoteAuthCredSASL2Vir(int vircred) * * Build up the SASL callback structure. We register one callback for * each credential type that the libvirt client indicated they support. - * We explicitly leav the callback function pointer at NULL though, + * We explicitly leave the callback function pointer at NULL though, * because we don't actually want to get SASL callbacks triggered. * Instead, we want the start/step functions to return SASL_INTERACT. * This lets us give the libvirt client a list of all required -- 1.8.3.1

1 0

[libvirt] [PATCH v2] rpc: Correct the wrong payload size checking
by Osier Yang 16 Oct '13

16 Oct '13

<...> /* Size of message length field. Not counted in VIR_NET_MESSAGE_MAX * and VIR_NET_MESSAGE_INITIAL. */ const VIR_NET_MESSAGE_LEN_MAX = 4; </...> However, msg->bufferLength includes the length word. The wrong checking was introduced by commit e914dcfd. * src/rpc/virnetmessage.c: - Correct the checking in virNetMessageEncodePayloadRaw - Use a new variable to track the new payload length in virNetMessageEncodePayloadRaw --- src/rpc/virnetmessage.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) diff --git a/src/rpc/virnetmessage.c b/src/rpc/virnetmessage.c index 8f4e4bc..d60366b 100644 --- a/src/rpc/virnetmessage.c +++ b/src/rpc/virnetmessage.c @@ -346,15 +346,16 @@ int virNetMessageEncodePayload(virNetMessagePtr msg, /* Try to encode the payload. If the buffer is too small increase it. */ while (!(*filter)(&xdr, data)) { - if ((msg->bufferLength - VIR_NET_MESSAGE_LEN_MAX) * 4 > VIR_NET_MESSAGE_MAX) { + unsigned int newlen = (msg->bufferLength - VIR_NET_MESSAGE_LEN_MAX) * 4; + + if (newlen > VIR_NET_MESSAGE_MAX) { virReportError(VIR_ERR_RPC, "%s", _("Unable to encode message payload")); goto error; } xdr_destroy(&xdr); - msg->bufferLength = (msg->bufferLength - VIR_NET_MESSAGE_LEN_MAX) * 4 + - VIR_NET_MESSAGE_LEN_MAX; + msg->bufferLength = newlen + VIR_NET_MESSAGE_LEN_MAX; if (VIR_REALLOC_N(msg->buffer, msg->bufferLength) < 0) goto error; @@ -426,10 +427,15 @@ int virNetMessageEncodePayloadRaw(virNetMessagePtr msg, /* If the message buffer is too small for the payload increase it accordingly. */ if ((msg->bufferLength - msg->bufferOffset) < len) { - if ((msg->bufferOffset + len) > VIR_NET_MESSAGE_MAX) { + if ((msg->bufferOffset + len) > + (VIR_NET_MESSAGE_MAX + VIR_NET_MESSAGE_LEN_MAX)) { virReportError(VIR_ERR_RPC, - _("Stream data too long to send (%zu bytes needed, %zu bytes available)"), - len, (VIR_NET_MESSAGE_MAX - msg->bufferOffset)); + _("Stream data too long to send " + "(%zu bytes needed, %zu bytes available)"), + len, + VIR_NET_MESSAGE_MAX + + VIR_NET_MESSAGE_LEN_MAX - + msg->bufferOffset); return -1; } -- 1.8.1.4

2 2

[libvirt] [PATCH] qemu: Fix the wrong logic of building cmd line for volume type disk
by Osier Yang 16 Oct '13

16 Oct '13

It builds a "libiscsi" uri if the volume type is "file" and "disk->auth.username" is specified, but I even see no reason to build the "libiscsi" uri for a "file" type volume, and the logic to check whether "disk->auth.username" is specified is also beyond understanding. It produces unexpected result like below: Disk config (pool is of 'fs' type): <disk type='volume' device='disk'> <driver name='qemu' type='raw'/> <auth username='myuser'> </auth> <source pool='default' volume='f18.img'/> <target dev='sda' bus='scsi'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> % virsh start f18 error: Failed to start domain f18 error: internal error: iscsi accepts only one host --- src/qemu/qemu_command.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index abb62e9..9b6a27d 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3806,13 +3806,7 @@ qemuBuildVolumeString(virConnectPtr conn, } break; case VIR_STORAGE_VOL_FILE: - if (disk->auth.username) { - if (qemuBuildISCSIString(conn, disk, opt) < 0) - goto cleanup; - virBufferAddChar(opt, ','); - } else { - virBufferEscape(opt, ',', ",", "file=%s,", disk->src); - } + virBufferEscape(opt, ',', ",", "file=%s,", disk->src); break; case VIR_STORAGE_VOL_NETWORK: /* Keep the compiler quite, qemuTranslateDiskSourcePool already -- 1.8.1.4

2 1

[libvirt] [PATCH v2] Introduce new OOM testing support
by Daniel P. Berrange 16 Oct '13

16 Oct '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> The previous OOM testing support would re-run the entire "main" method each iteration, failing a different malloc each time. When a test suite has 'n' allocations, the number of repeats requires is (n * (n + 1) ) / 2. This gets very large, very quickly. This new OOM testing support instead integrates at the virtTestRun level, so each individual test case gets repeated, instead of the entire test suite. This means the values of 'n' are orders of magnitude smaller. The simple usage is $ VIR_TEST_OOM=1 ./qemuxml2argvtest ... 29) QEMU XML-2-ARGV clock-utc ... OK Test OOM for nalloc=36 .................................... OK 30) QEMU XML-2-ARGV clock-localtime ... OK Test OOM for nalloc=36 .................................... OK 31) QEMU XML-2-ARGV clock-france ... OK Test OOM for nalloc=38 ...................................... OK ... the second lines reports how many mallocs have to be failed, and thus how many repeats of the test will be run. If it crashes, then running under valgrind will often show the problem $ VIR_TEST_OOM=1 ../run valgrind ./qemuxml2argvtest When debugging problems it is also helpful to select an individual test case $ VIR_TEST_RANGE=30 VIR_TEST_OOM=1 ../run valgrind ./qemuxml2argvtest When things get really tricky, it is possible to request that just specific allocs are failed. eg to fail allocs 5 -> 12, use $ VIR_TEST_RANGE=30 VIR_TEST_OOM=1:5-12 ../run valgrind ./qemuxml2argvtest In the worse case, you might want to know the stack trace of the alloc which was failed then VIR_TEST_OOM_TRACE can be set. If it is set to 1 then it will only print if it thinks a mistake happened. This is often not reliable, so setting it to 2 will make it print the stack trace for every alloc that is failed. $ VIR_TEST_OOM_TRACE=2 VIR_TEST_RANGE=30 VIR_TEST_OOM=1:5-5 ../run valgrind ./qemuxml2argvtest 30) QEMU XML-2-ARGV clock-localtime ... OK Test OOM for nalloc=36 !virAllocN /home/berrange/src/virt/libvirt/src/util/viralloc.c:180 virHashCreateFull /home/berrange/src/virt/libvirt/src/util/virhash.c:144 virDomainDefParseXML /home/berrange/src/virt/libvirt/src/conf/domain_conf.c:11745 virDomainDefParseNode /home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12646 virDomainDefParse /home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12590 testCompareXMLToArgvFiles /home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:106 virtTestRun /home/berrange/src/virt/libvirt/tests/testutils.c:250 mymain /home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:418 (discriminator 2) virtTestMain /home/berrange/src/virt/libvirt/tests/testutils.c:750 ?? ??:0 _start ??:? FAILED Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- docs/internals/oomtesting.html.in | 213 ++++++++++++++++++++++++++++++++++++++ docs/sitemap.html.in | 4 + tests/qemuargv2xmltest.c | 14 +-- tests/qemuxml2argvtest.c | 12 ++- tests/qemuxmlnstest.c | 18 ++-- tests/testutils.c | 195 +++++++++++++++++++++++++++++++++- tests/testutils.h | 2 + 7 files changed, 437 insertions(+), 21 deletions(-) create mode 100644 docs/internals/oomtesting.html.in diff --git a/docs/internals/oomtesting.html.in b/docs/internals/oomtesting.html.in new file mode 100644 index 0000000..652aff6 --- /dev/null +++ b/docs/internals/oomtesting.html.in @@ -0,0 +1,213 @@ +<?xml version="1.0" encoding="UTF-8"?> +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> +<html xmlns="http://www.w3.org/1999/xhtml"> + <body> + <h1>Out of memory testing</h1> + + <ul id="toc"></ul> + + + <p> + This page describes how to use the test suite todo out of memory + testing. + </p> + + <h2>Building with OOM testing</h2> + + <p> + Since OOM testing requires hooking into the malloc APIs, it is + not enabled by default. The flag <code>--enable-test-oom</code> + must be given to <code>configure</code>. When this is done the + libvirt allocation APIs will have some hooks enabled. + </p> + + <pre> +$ ./configure --enable-test-oom +</pre> + + + <h2><a name="basicoom">Basic OOM testing support</a></h2> + + <p> + The first step in validating OOM usage is to run a test suite + with full OOM testing enabled. This is done by setting the + <code>VIR_TEST_OOM=1</code> environment variable. The way this + works is that it runs the test once normally to "prime" any + static memory allocations. Then it runs it once more counting + the total number of memory allocations. Then it runs it in a + loop failing a different memory allocation each time. For every + memory allocation failure triggered, it expects the test case + to return an error. OOM testing is quite slow requiring each + test case to be executed O(n) times, where 'n' is the total + number of memory allocations. This results in a total number + of memory allocations of '(n * (n + 1) ) / 2' + </p> + + <pre> +$ VIR_TEST_OOM=1 ./qemuxml2argvtest + 1) QEMU XML-2-ARGV minimal ... OK + Test OOM for nalloc=42 .......................................... OK + 2) QEMU XML-2-ARGV minimal-s390 ... OK + Test OOM for nalloc=28 ............................ OK + 3) QEMU XML-2-ARGV machine-aliases1 ... OK + Test OOM for nalloc=38 ...................................... OK + 4) QEMU XML-2-ARGV machine-aliases2 ... OK + Test OOM for nalloc=38 ...................................... OK + 5) QEMU XML-2-ARGV machine-core-on ... OK + Test OOM for nalloc=37 ..................................... OK +...snip... +</pre> + + <p> + In this output, the first line shows the normal execution and + the test number, and the second line shows the total number + of memory allocations from that test case. + </p> + + <h3><a name="valgrind">Tracking failures with valgrind</a></h3> + + <p> + The test suite should obviously *not* crash during OOM testing. + If it does crash, then to assist in tracking down the problem + it is worth using valgrind and only running a single test case. + For example, supposing test case 5 crashed. Then re-run the + test with + </p> + + <pre> +$ VIR_TEST_OOM=1 VIR_TEST_RANGE=5 ../run valgrind ./qemuxml2argvtest +...snip... + 5) QEMU XML-2-ARGV machine-core-on ... OK + Test OOM for nalloc=37 ..................................... OK +...snip... + </pre> + + <p> + Valgrind should report the cause of the crash - for example a + double free or use of uninitialized memory or NULL pointer + access. + </p> + + <h3><a name="stacktraces">Tracking failures with stack traces</a></h3> + + <p> + With some really difficult bugs valgrind is not sufficient to + identify the cause. In this case, it is useful to identify the + precise allocation which was failed, to allow the code path + to the error to be traced. The <code>VIR_TEST_OOM</code> + env variable can be given a range of memory allocations to + test. So if a test case has 150 allocations, it can be told + to only test allocation numbers 7-10. The <code>VIR_TEST_OOM_TRACE</code> + variable can be used to print out stack traces. + </p> + + <pre> +$ VIR_TEST_OOM_TRACE=2 VIR_TEST_OOM=1:7-10 VIR_TEST_RANGE=5 \ + ../run valgrind ./qemuxml2argvtest + 5) QEMU XML-2-ARGV machine-core-on ... OK + Test OOM for nalloc=37 !virAllocN +/home/berrange/src/virt/libvirt/src/util/viralloc.c:180 +virDomainDefParseXML +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:11786 (discriminator 1) +virDomainDefParseNode +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12677 +virDomainDefParse +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12621 +testCompareXMLToArgvFiles +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:107 +virtTestRun +/home/berrange/src/virt/libvirt/tests/testutils.c:266 +mymain +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:388 (discriminator 2) +virtTestMain +/home/berrange/src/virt/libvirt/tests/testutils.c:791 +__libc_start_main +??:? +_start +??:? +!virAlloc +/home/berrange/src/virt/libvirt/src/util/viralloc.c:133 +virDomainDiskDefParseXML +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:4790 +virDomainDefParseXML +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:11797 +virDomainDefParseNode +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12677 +virDomainDefParse +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12621 +testCompareXMLToArgvFiles +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:107 +virtTestRun +/home/berrange/src/virt/libvirt/tests/testutils.c:266 +mymain +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:388 (discriminator 2) +virtTestMain +/home/berrange/src/virt/libvirt/tests/testutils.c:791 +__libc_start_main +??:? +_start +??:? +!virAllocN +/home/berrange/src/virt/libvirt/src/util/viralloc.c:180 +virXPathNodeSet +/home/berrange/src/virt/libvirt/src/util/virxml.c:609 +virDomainDefParseXML +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:11805 +virDomainDefParseNode +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12677 +virDomainDefParse +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12621 +testCompareXMLToArgvFiles +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:107 +virtTestRun +/home/berrange/src/virt/libvirt/tests/testutils.c:266 +mymain +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:388 (discriminator 2) +virtTestMain +/home/berrange/src/virt/libvirt/tests/testutils.c:791 +__libc_start_main +??:? +_start +??:? +!virAllocN +/home/berrange/src/virt/libvirt/src/util/viralloc.c:180 +virDomainDefParseXML +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:11808 (discriminator 1) +virDomainDefParseNode +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12677 +virDomainDefParse +/home/berrange/src/virt/libvirt/src/conf/domain_conf.c:12621 +testCompareXMLToArgvFiles +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:107 +virtTestRun +/home/berrange/src/virt/libvirt/tests/testutils.c:266 +mymain +/home/berrange/src/virt/libvirt/tests/qemuxml2argvtest.c:388 (discriminator 2) +virtTestMain +/home/berrange/src/virt/libvirt/tests/testutils.c:791 +__libc_start_main +??:? +_start +??:? + </pre> + + <h3><a name="noncrash">Non-crash related problems</a></h3> + + <p> + Not all memory allocation bugs result in code crashing. Sometimes + the code will be silently ignoring the allocation failure, resulting + in incorrect data being produced. For example the XML parser may + mistakenly treat an allocation failure as indicating that an XML + attribute was not set in the input document. It is hard to identify + these problems from the test suite automatically. For this, the + test suites should be run with <code>VIR_TEST_DEBUG=1</code> set + and then stderr analysed for any unexpected data. For example, + the XML conversion may show an embedded "(null)" literal, or the + test suite might complain about missing elements / attributes + in the actual vs expected data. These are all signs of bugs in + OOM handling. In the future the OOM tests will be enhanced to + validate that an error VIR_ERR_NO_MEMORY is returned for each + allocation failed, rather than some other error. + </p> + </body> +</html> diff --git a/docs/sitemap.html.in b/docs/sitemap.html.in index d821a9e..d885f02 100644 --- a/docs/sitemap.html.in +++ b/docs/sitemap.html.in @@ -338,6 +338,10 @@ <a href="internals/locking.html">Lock managers</a> <span>Use lock managers to protect disk content</span> </li> + <li> + <a href="internals/oomtesting.html">Out of memory testing</a> + <span>Simulating OOM conditions in the test suite</span> + </li> </ul> </li> <li> diff --git a/tests/qemuargv2xmltest.c b/tests/qemuargv2xmltest.c index 6dd8bb0..92fc89a 100644 --- a/tests/qemuargv2xmltest.c +++ b/tests/qemuargv2xmltest.c @@ -42,7 +42,6 @@ static int testCompareXMLToArgvFiles(const char *xml, char *cmd = NULL; int ret = -1; virDomainDefPtr vmdef = NULL; - char *log; if (virtTestLoadFile(cmdfile, &cmd) < 0) goto fail; @@ -53,13 +52,16 @@ static int testCompareXMLToArgvFiles(const char *xml, cmd, NULL, NULL, NULL))) goto fail; - if ((log = virtTestLogContentAndReset()) == NULL) - goto fail; - if ((*log != '\0') != expect_warning) { + if (!virtTestOOMActive()) { + char *log; + if ((log = virtTestLogContentAndReset()) == NULL) + goto fail; + if ((*log != '\0') != expect_warning) { + VIR_FREE(log); + goto fail; + } VIR_FREE(log); - goto fail; } - VIR_FREE(log); if (!(actualxml = virDomainDefFormat(vmdef, 0))) goto fail; diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 060acf2..83ddbad 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -107,7 +107,8 @@ static int testCompareXMLToArgvFiles(const char *xml, if (!(vmdef = virDomainDefParseFile(xml, driver.caps, driver.xmlopt, QEMU_EXPECTED_VIRT_TYPES, VIR_DOMAIN_XML_INACTIVE))) { - if (flags & FLAG_EXPECT_PARSE_ERROR) + if (!virtTestOOMActive() && + (flags & FLAG_EXPECT_PARSE_ERROR)) goto ok; goto out; } @@ -170,7 +171,8 @@ static int testCompareXMLToArgvFiles(const char *xml, migrateFrom, migrateFd, NULL, VIR_NETDEV_VPORT_PROFILE_OP_NO_OP, &testCallbacks))) { - if (flags & FLAG_EXPECT_FAILURE) { + if (!virtTestOOMActive() && + (flags & FLAG_EXPECT_FAILURE)) { ret = 0; if (virTestGetDebug() > 1) fprintf(stderr, "Got expected error: %s\n", @@ -184,7 +186,8 @@ static int testCompareXMLToArgvFiles(const char *xml, goto out; } - if (!!virGetLastError() != !!(flags & FLAG_EXPECT_ERROR)) { + if (!virtTestOOMActive() && + (!!virGetLastError() != !!(flags & FLAG_EXPECT_ERROR))) { if (virTestGetDebug() && (log = virtTestLogContentAndReset())) fprintf(stderr, "\n%s", log); goto out; @@ -205,7 +208,8 @@ static int testCompareXMLToArgvFiles(const char *xml, } ok: - if (flags & FLAG_EXPECT_ERROR) { + if (!virtTestOOMActive() && + (flags & FLAG_EXPECT_ERROR)) { /* need to suppress the errors */ virResetLastError(); } diff --git a/tests/qemuxmlnstest.c b/tests/qemuxmlnstest.c index 2cc15d1..38c98bd 100644 --- a/tests/qemuxmlnstest.c +++ b/tests/qemuxmlnstest.c @@ -117,15 +117,17 @@ static int testCompareXMLToArgvFiles(const char *xml, &testCallbacks))) goto fail; - if (!!virGetLastError() != expectError) { - if (virTestGetDebug() && (log = virtTestLogContentAndReset())) - fprintf(stderr, "\n%s", log); - goto fail; - } + if (!virtTestOOMActive()) { + if (!!virGetLastError() != expectError) { + if (virTestGetDebug() && (log = virtTestLogContentAndReset())) + fprintf(stderr, "\n%s", log); + goto fail; + } - if (expectError) { - /* need to suppress the errors */ - virResetLastError(); + if (expectError) { + /* need to suppress the errors */ + virResetLastError(); + } } if (!(actualargv = virCommandToString(cmd))) diff --git a/tests/testutils.c b/tests/testutils.c index 5d634b4..bc1e34e 100644 --- a/tests/testutils.c +++ b/tests/testutils.c @@ -47,6 +47,13 @@ #include "virprocess.h" #include "virstring.h" +#ifdef TEST_OOM +# ifdef TEST_OOM_TRACE +#include <dlfcn.h> +#include <execinfo.h> +# endif +#endif + #ifdef HAVE_PATHS_H # include <paths.h> #endif @@ -64,6 +71,18 @@ static unsigned int testDebug = -1; static unsigned int testVerbose = -1; static unsigned int testExpensive = -1; +#ifdef TEST_OOM +static unsigned int testOOM = 0; +static unsigned int testOOMStart = -1; +static unsigned int testOOMEnd = -1; +static unsigned int testOOMTrace = 0; +# ifdef TEST_OOM_TRACE +void *testAllocStack[30]; +int ntestAllocStack; +# endif +#endif +static bool testOOMActive = false; + static size_t testCounter = 0; static size_t testStart = 0; static size_t testEnd = 0; @@ -71,6 +90,19 @@ static size_t testEnd = 0; char *progname; char *abs_srcdir; +bool virtTestOOMActive(void) +{ + return testOOMActive; +} + +#ifdef TEST_OOM_TRACE +static void virTestAllocHook(int nalloc ATTRIBUTE_UNUSED, + void *opaque ATTRIBUTE_UNUSED) +{ + ntestAllocStack = backtrace(testAllocStack, ARRAY_CARDINALITY(testAllocStack)); +} +#endif + void virtTestResult(const char *name, int ret, const char *msg, ...) { va_list vargs; @@ -109,6 +141,35 @@ void virtTestResult(const char *name, int ret, const char *msg, ...) va_end(vargs); } +#ifdef TEST_OOM_TRACE +static void +virTestShowTrace(void) +{ + size_t j; + for (j = 2; j < ntestAllocStack; j++) { + Dl_info info; + char *cmd; + + dladdr(testAllocStack[j], &info); + if (info.dli_fname && + strstr(info.dli_fname, ".so")) { + if (virAsprintf(&cmd, "addr2line -f -e %s %p", + info.dli_fname, + ((void*)((unsigned long long)testAllocStack[j] + - (unsigned long long)info.dli_fbase))) < 0) + continue; + } else { + if (virAsprintf(&cmd, "addr2line -f -e %s %p", + (char*)(info.dli_fname ? info.dli_fname : "<unknown>"), + testAllocStack[j]) < 0) + continue; + } + ignore_value(system(cmd)); + VIR_FREE(cmd); + } +} +#endif + /* * Runs test * @@ -155,7 +216,7 @@ virtTestRun(const char *title, !((testCounter-1) % 40)) { fprintf(stderr, " %-3zu\n", (testCounter-1)); fprintf(stderr, " "); - } + } if (ret == 0) fprintf(stderr, "."); else if (ret == EXIT_AM_SKIP) @@ -164,6 +225,77 @@ virtTestRun(const char *title, fprintf(stderr, "!"); } +#ifdef TEST_OOM + if (testOOM && ret != EXIT_AM_SKIP) { + int nalloc; + int oomret; + int start, end; + size_t i; + virResetLastError(); + virAllocTestInit(); +# ifdef TEST_OOM_TRACE + virAllocTestHook(virTestAllocHook, NULL); +# endif + oomret = body(data); + nalloc = virAllocTestCount(); + fprintf(stderr, " Test OOM for nalloc=%d ", nalloc); + if (testOOMStart == -1 || + testOOMEnd == -1) { + start = 0; + end = nalloc; + } else { + start = testOOMStart; + end = testOOMEnd + 1; + } + testOOMActive = true; + for (i = start; i < end; i++) { + bool missingFail = false; +# ifdef TEST_OOM_TRACE + memset(testAllocStack, 0, ARRAY_CARDINALITY(testAllocStack)); + ntestAllocStack = 0; +# endif + virAllocTestOOM(i + 1, 1); + oomret = body(data); + + /* fprintf() disabled because XML parsing APIs don't allow + * distinguish between element / attribute not present + * in the XML (which is non-fatal), vs OOM / malformed + * which should be fatal. Thus error reporting for + * optionally present XML is mostly broken. + */ + if (oomret == 0) { + missingFail = true; +# if 0 + fprintf(stderr, " alloc %zu failed but no err status\n", i + 1); +# endif + } else { + virErrorPtr lerr = virGetLastError(); + if (!lerr) { +# if 0 + fprintf(stderr, " alloc %zu failed but no error report\n", i + 1); +# endif + missingFail = true; + } + } + if ((missingFail && testOOMTrace) || (testOOMTrace > 1)) { + fprintf(stderr, "%s", "!"); +# ifdef TEST_OOM_TRACE + virTestShowTrace(); +# endif + ret = -1; + } else { + fprintf(stderr, "%s", "."); + } + } + testOOMActive = false; + if (ret == 0) + fprintf(stderr, " OK\n"); + else + fprintf(stderr, " FAILED\n"); + virAllocTestInit(); + } +#endif /* TEST_OOM */ + return ret; } @@ -192,7 +324,6 @@ virtTestLoadFile(const char *file, char **buf) tmplen = buflen = st.st_size + 1; if (VIR_ALLOC_N(*buf, buflen) < 0) { - fprintf(stderr, "%s: larger than available memory (> %d)\n", file, buflen); VIR_FORCE_FCLOSE(fp); return -1; } @@ -468,7 +599,8 @@ virtTestLogOutput(virLogSource source ATTRIBUTE_UNUSED, { struct virtTestLogData *log = data; virCheckFlags(VIR_LOG_STACK_TRACE,); - virBufferAsprintf(&log->buf, "%s: %s", timestamp, str); + if (!testOOMActive) + virBufferAsprintf(&log->buf, "%s: %s", timestamp, str); } static void @@ -537,6 +669,9 @@ int virtTestMain(int argc, int ret; bool abs_srcdir_cleanup = false; char *testRange = NULL; +#ifdef TEST_OOM + char *oomstr; +#endif abs_srcdir = getenv("abs_srcdir"); if (!abs_srcdir) { @@ -597,6 +732,60 @@ int virtTestMain(int argc, } } +#ifdef TEST_OOM + if ((oomstr = getenv("VIR_TEST_OOM")) != NULL) { + char *next; + if (testDebug == -1) + testDebug = 1; + testOOM = 1; + if (oomstr[0] != '\0' && + oomstr[1] == ':') { + if (virStrToLong_ui(oomstr + 2, &next, 10, &testOOMStart) < 0) { + fprintf(stderr, "Cannot parse range %s\n", oomstr); + return EXIT_FAILURE; + } + if (*next == '\0') { + testOOMEnd = testOOMStart; + } else { + if (*next != '-') { + fprintf(stderr, "Cannot parse range %s\n", oomstr); + return EXIT_FAILURE; + } + if (virStrToLong_ui(next+1, NULL, 10, &testOOMEnd) < 0) { + fprintf(stderr, "Cannot parse range %s\n", oomstr); + return EXIT_FAILURE; + } + } + } else { + testOOMStart = -1; + testOOMEnd = -1; + } + } + +# ifdef TEST_OOM_TRACE + if ((oomstr = getenv("VIR_TEST_OOM_TRACE")) != NULL) { + if (virStrToLong_ui(oomstr, NULL, 10, &testOOMTrace) < 0) { + fprintf(stderr, "Cannot parse oom trace %s\n", oomstr); + return EXIT_FAILURE; + } + } +# else + if (getenv("VIR_TEST_OOM_TRACE")) { + fprintf(stderr, "%s", "OOM test tracing not enabled in this build\n"); + return EXIT_FAILURE; + } +# endif +#else /* TEST_OOM */ + if (getenv("VIR_TEST_OOM")) { + fprintf(stderr, "%s", "OOM testing not enabled in this build\n"); + return EXIT_FAILURE; + } + if (getenv("VIR_TEST_OOM_TRACE")) { + fprintf(stderr, "%s", "OOM test tracing not enabled in this build\n"); + return EXIT_FAILURE; + } +#endif /* TEST_OOM */ + ret = (func)(); if (abs_srcdir_cleanup) diff --git a/tests/testutils.h b/tests/testutils.h index 478b53c..e6f494d 100644 --- a/tests/testutils.h +++ b/tests/testutils.h @@ -40,6 +40,8 @@ extern char *progname; extern char *abs_srcdir; +bool virtTestOOMActive(void); + void virtTestResult(const char *name, int ret, const char *msg, ...) ATTRIBUTE_FMT_PRINTF(3,4); int virtTestRun(const char *title, -- 1.8.3.1

1 0

[libvirt] [PATCH 0/6] Try to fix selinux/securityfs mount handling in LXC
by Daniel P. Berrange 16 Oct '13

16 Oct '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> This series tries to address the problem of LXC containers failing to start because 'securityfs' is not compiled in. In doing this some general cleanup was performed on the LXC mount code. Daniel P. Berrange (6): Add virFileIsMountPoint function Remove unused 'opts' field from LXC basic mounts struct Remove pointless 'srcpath' variable in lxcContainerMountBasicFS Remove duplicate entries in lxcBasicMounts array Add flag to lxcBasicMounts to control use in user namespaces Skip any files which are not mounted on the host src/libvirt_private.syms | 1 + src/lxc/lxc_container.c | 86 +++++++++++++++++++++++++++++------------------- src/util/virfile.c | 58 ++++++++++++++++++++++++++++++++ src/util/virfile.h | 2 ++ 4 files changed, 113 insertions(+), 34 deletions(-) -- 1.8.3.1

4 16

[libvirt] [PATCHv2 00/23] Better approach to add the "pvticketlocks" feature
by Peter Krempa 16 Oct '13

16 Oct '13

This version resolves conflicts with Eric's const correctnes series. The paravirtual spinlock feature was introduced recently into the qemu and linux code base. To test it you need the most recent -rc version of the linux kernel and the git version of qemu. As the requirements to activate this feature in the guest are really volatile (kernel support, qemu support .. .) I decided to improve the original version so that it detects if the feature was actually enabled in the guest. This avoids possible problems with breaking guest ABI when migrating between two hosts with different kernels. Additionally we may start enforcin other features too. This patchset is partialy based on Jiri's work on getting actual cpu definition from a running qemu. Jiri Denemark (8): cpu: Add support for loading and storing CPU data cpu: x86: Rename struct cpuX86cpuid as virCPUx86CPUID cpu: x86: Rename struct cpuX86Data as virCPUx86Data cpu: x86: Rename x86DataFree() as virCPUx86DataFree() cpu: x86: Rename x86MakeCPUData as virCPUx86MakeData cpu: x86: Rename x86DataAddCpuid as virCPUx86DataAddCPUID cpu: Export few x86-specific APIs qemu: Add monitor APIs to fetch CPUID data from QEMU Peter Krempa (15): schema: Rename option to make it reusable conf: Clean up few error messages qemu: command: Fix macro indentation cpu: x86: Rename data_iterator and DATA_ITERATOR_INIT cpu: x86: Fix return types of x86cpuidMatch and x86cpuidMatchMasked cpu: x86: Fix function header formatting and whitespace cpu: x86: Use whitespace to clarify context and use consistent labels cpu: x86: Clean up error messages in x86VendorLoad() qemu: Clean up check of maximum cpu count supported by a machine type cpu_x86: Refactor storage of CPUID data to add support for KVM features cpu: x86: Parse the CPU feature map only once cpu: x86: Add internal CPUID features support and KVM feature bits conf: Refactor storing and usage of feature flags qemu: Add support for paravirtual spinlocks in the guest qemu: process: Validate specific CPUID flags of a guest docs/formatdomain.html.in | 7 + docs/schemas/domaincommon.rng | 18 +- src/conf/domain_conf.c | 200 ++++-- src/conf/domain_conf.h | 3 +- src/cpu/cpu.c | 41 ++ src/cpu/cpu.h | 16 +- src/cpu/cpu_x86.c | 699 ++++++++++++--------- src/cpu/cpu_x86.h | 9 + src/cpu/cpu_x86_data.h | 26 +- src/libvirt_private.syms | 8 + src/libxl/libxl_conf.c | 9 +- src/lxc/lxc_container.c | 6 +- src/qemu/qemu_command.c | 31 +- src/qemu/qemu_monitor.c | 21 + src/qemu/qemu_monitor.h | 3 + src/qemu/qemu_monitor_json.c | 107 ++++ src/qemu/qemu_monitor_json.h | 3 + src/qemu/qemu_process.c | 50 +- src/vbox/vbox_tmpl.c | 45 +- src/xenapi/xenapi_driver.c | 10 +- src/xenapi/xenapi_utils.c | 22 +- src/xenxs/xen_sxpr.c | 20 +- src/xenxs/xen_xm.c | 30 +- tests/Makefile.am | 1 + .../qemumonitorjson-getcpu-empty.data | 2 + .../qemumonitorjson-getcpu-empty.json | 46 ++ .../qemumonitorjson-getcpu-filtered.data | 4 + .../qemumonitorjson-getcpu-filtered.json | 46 ++ .../qemumonitorjson-getcpu-full.data | 5 + .../qemumonitorjson-getcpu-full.json | 46 ++ .../qemumonitorjson-getcpu-host.data | 6 + .../qemumonitorjson-getcpu-host.json | 45 ++ tests/qemumonitorjsontest.c | 76 +++ .../qemuxml2argv-pv-spinlock-disabled.args | 5 + .../qemuxml2argv-pv-spinlock-disabled.xml | 26 + .../qemuxml2argv-pv-spinlock-enabled.args | 5 + .../qemuxml2argv-pv-spinlock-enabled.xml | 26 + tests/qemuxml2argvtest.c | 2 + tests/qemuxml2xmltest.c | 2 + 39 files changed, 1279 insertions(+), 448 deletions(-) create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-empty.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-empty.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-filtered.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-filtered.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-full.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-full.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-host.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-host.json create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-pv-spinlock-disabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-pv-spinlock-disabled.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-pv-spinlock-enabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-pv-spinlock-enabled.xml -- 1.8.3.2

3 50

[libvirt] [PATCH] virprocess: make virProcessGetStartTime work on Mac OS X
by Ryota Ozaki 16 Oct '13

16 Oct '13

This patch addresses the following warning output by libvirtd: warning : virProcessGetStartTime:843 : Process start time of pid XXXXX not available on this platform For Mac OS X, we can use similar instructions to FreeBSD to get the start time of a process. The difference between them is struct kinfo_proc; kinfo_proc.ki_start is the start time of a process for FreeBSD while kinfo_proc.kp_proc.p_starttime for Mac OS X. Note that this patch works for Lion and Mountain Lion, however, doesn't work for Snow Leopard for some reason; sysctlnametomib("kern.proc.pid", ...) fails on Snow Leopard with the following error: error : virProcessGetStartTime:822 : Unable to get MIB of kern.proc.pid: No such file or directory This is unexpected. man 3 sysctl of Snow Leopard says it should work... Anyway libvirtd is able to launch on Snow Leopard regardless of the error. Signed-off-by: Ryota Ozaki <ozaki.ryota(a)gmail.com> --- src/util/virprocess.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/src/util/virprocess.c b/src/util/virprocess.c index 9fc3207..5a0ed0d 100644 --- a/src/util/virprocess.c +++ b/src/util/virprocess.c @@ -32,7 +32,7 @@ #endif #include <sched.h> -#ifdef __FreeBSD__ +#if defined(__FreeBSD__) || defined(__APPLE__) # include <sys/param.h> # include <sys/sysctl.h> # include <sys/user.h> @@ -809,7 +809,7 @@ cleanup: VIR_FREE(buf); return ret; } -#elif defined(__FreeBSD__) +#elif defined(__FreeBSD__) || defined(__APPLE__) int virProcessGetStartTime(pid_t pid, unsigned long long *timestamp) { @@ -817,7 +817,12 @@ int virProcessGetStartTime(pid_t pid, int mib[4]; size_t len = 4; - sysctlnametomib("kern.proc.pid", mib, &len); + /* FIXME: It doesn't work on Snow Leopard for some reason */ + if (sysctlnametomib("kern.proc.pid", mib, &len) < 0) { + virReportSystemError(errno, "%s", + _("Unable to get MIB of kern.proc.pid")); + return -1; + } len = sizeof(struct kinfo_proc); mib[3] = pid; @@ -828,7 +833,11 @@ int virProcessGetStartTime(pid_t pid, return -1; } +# if defined(__FreeBSD__) *timestamp = (unsigned long long)p.ki_start.tv_sec; +# else + *timestamp = (unsigned long long)p.kp_proc.p_starttime.tv_sec; +# endif return 0; -- 1.8.4

2 2

[libvirt] [PATCH 0/3] leave cgroup value checking to kernel
by Chen Hanxiao 16 Oct '13

16 Oct '13

From: Chen Hanxiao <chenhanxiao(a)cn.fujitsu.com> Since 2.6.39, kernel changed the minimum weight of device blkio while libvirt hard-coded the value range checking. So we should leave the job of checking to kernel. Chen Hanxiao (3): [libvirt]docs: change the minimum weight description for blkio [libvirt]cgroup: show error when EINVAL caught [libvirt]cgroup: leave value checking to kernel docs/formatdomain.html.in | 6 ++++-- src/util/vircgroup.c | 26 +++++++++++--------------- tools/virsh.pod | 11 ++++++----- 3 files changed, 21 insertions(+), 22 deletions(-) -- 1.8.2.1

2 7

[libvirt] [PATCH v3]LXC: Helper function for checking permission of dir when userns enabled
by Chen Hanxiao 16 Oct '13

16 Oct '13

From: Chen Hanxiao <chenhanxiao(a)cn.fujitsu.com> If we enable userns, the process with uid/gid in idmap should have enough permission to access dir we provided for containers. Currently, the debug log is very implicit or misleading sometimes. This patch will help clarify this for us when using debug log or virsh. v2: syntax-check clean v3: reliable method for checking permission of dir Signed-off-by: Chen Hanxiao <chenhanxiao(a)cn.fujitsu.com> --- src/lxc/lxc_container.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 88 insertions(+) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 8abaea0..9a05e30 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -110,6 +110,13 @@ struct __lxc_child_argv { int handshakefd; }; +typedef struct __lxc_userns_DirPermCheck_argv lxc_userns_DirPermCheck_argv_t; +struct __lxc_userns_DirPermCheck_argv { + uid_t uid; + gid_t gid; + virDomainDefPtr vmDef; +}; + static int lxcContainerMountFSBlock(virDomainFSDefPtr fs, const char *srcprefix); @@ -1829,6 +1836,84 @@ lxcNeedNetworkNamespace(virDomainDefPtr def) return false; } +static +int lxcContainerCheckDirPermissionChild(void *argv) +{ + size_t i; + lxc_userns_DirPermCheck_argv_t *args = argv; + uid_t uid = args->uid; + uid_t gid = args->gid; + virDomainDefPtr vmDef = args->vmDef; + char *path; + + if (virSetUIDGID(uid, gid, NULL, 0) < 0) { + virReportSystemError(errno, "%s", + _("setuid or setgid failed")); + _exit(-1); + } + + for (i = 0; i < vmDef->nfss; i++) { + path = vmDef->fss[i]->src; + if (access(path, R_OK) || access(path, W_OK) || virFileIsExecutable(path)) { + VIR_DEBUG("Src dir '%s' does not belong to uid/gid: %d/%d", + vmDef->fss[i]->src, uid, gid); + _exit(-1); + } + } + + _exit(0); +} + +/* + * Helper function for helping check + * whether we have enough privilege + * to operate the source dir when userns enabled + * @vmDef: pointer to vm definition structure + * Returns 0 on success or -1 in case of error + */ +static int +lxcContainerCheckDirPermission(virDomainDefPtr vmDef) +{ + uid_t uid; + gid_t gid; + int cpid = 0; + int status; + char *childStack; + char *stack; + int flags = SIGCHLD; + + uid = vmDef->idmap.uidmap[0].target; + gid = vmDef->idmap.gidmap[0].target; + + lxc_userns_DirPermCheck_argv_t args = { + .uid = uid, + .gid = gid, + .vmDef = vmDef + }; + + if (VIR_ALLOC_N(stack, getpagesize() * 4) < 0) + return -1; + + childStack = stack + (getpagesize() * 4); + cpid = clone(lxcContainerCheckDirPermissionChild, childStack, flags, &args); + VIR_FREE(stack); + if (cpid < 0) { + virReportSystemError(errno, "%s", + _("Unable to clone to check permission of directory")); + return -1; + } else if (virProcessWait(cpid, &status) < 0) { + return -1; + } + + if (WEXITSTATUS(status) != 0) { + virReportSystemError(errno, "%s", + _("Check the permission of source dir provided for container")); + return -1; + } + + return 0; +} + /** * lxcContainerStart: * @def: pointer to virtual machine structure @@ -1880,6 +1965,9 @@ int lxcContainerStart(virDomainDefPtr def, if (userns_supported()) { VIR_DEBUG("Enable user namespace"); cflags |= CLONE_NEWUSER; + if (lxcContainerCheckDirPermission(def) < 0) { + return -1; + } } else { virReportSystemError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("Kernel doesn't support user namespace")); -- 1.8.2.1

2 4