- Devel - libvirt List Archives

[libvirt] [PATCH] tests: monitor: json: Fix error message when returning json in json
by Peter Krempa 03 Jun '14

03 Jun '14

The qemu JSON monitor test allows to test also expected command arguments. As the error from the monitor simulator is returned as a simulated qemu error (in JSON) all other JSON contained in the error message needs to be escaped. This will happen if the monitor command under test receives a JSON array as an argument. This will improve the error message from: libvirt: error : internal error: cannot parse json { "error": { "desc": "Invalid value of argument 'keys' of command 'send-key': expected 'ble' got '[{"type":"number","data":43},{"type":"number","data":26}, {"type":"number","data":46},{"type":"number","data":32}]'", "class": "UnexpectedCommand" } }: lexical error: invalid string in json text. To: libvirt: QEMU Driver error : internal error: unable to execute QEMU command 'send-key': Invalid value of argument 'keys' of command 'send-key': expected 'ble' got '[{"type":"number","data":43}, {"type":"number","data":26},{"type":"number","data":46}, {"type":"number","data":32}]' This improvement will not have any effect on tests executing as expected, but it will help test development. --- tests/qemumonitortestutils.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/tests/qemumonitortestutils.c b/tests/qemumonitortestutils.c index 30c2153..8155a69 100644 --- a/tests/qemumonitortestutils.c +++ b/tests/qemumonitortestutils.c @@ -145,6 +145,12 @@ qemuMonitorReportError(qemuMonitorTestPtr test, const char *errmsg, ...) goto cleanup; if (test->agent || test->json) { + char *tmp = msg; + msg = qemuMonitorEscapeArg(tmp); + VIR_FREE(tmp); + if (!msg) + goto cleanup; + if (virAsprintf(&jsonmsg, "{ \"error\": " " { \"desc\": \"%s\", " " \"class\": \"UnexpectedCommand\" } }", -- 1.9.3

2 2

[libvirt] parallels: add 2 functions needed by OpenStack
by Alexander Burluka 03 Jun '14

03 Jun '14

Hello, I would like to apologize my for my inattention, that led to unformatted subjects. I'll try to prevent this in the future. -- Regards, Alexander Burluka

2 5

Re: [libvirt] [PATCH 1/2] Parallels: add domainGetVcpus().
by Daniel P. Berrange 03 Jun '14

03 Jun '14

Re-adding the mailing list - please don't drop the list CC when replying. On Mon, Jun 02, 2014 at 09:22:35PM +0400, aburluka wrote: > >Hmm, what error did you get from openstack ? The two uses of the > >'dom.vcpus' function are both wrapped in try/except so that it is > >considered non-fatal if libvirt doesn't provide this. > I have caught this error on Havana release. I would try to reproduce > this error, but it will take some short time to test it on Icehouse. Ok, then I think you should find the lack of this API is non-fatal in Icehouse / Juno now. > >Are you talking about container based virt here or full machine > >based virt ? IIUC Parallels can do both ? With container based > >virt, does parallels have the concept of 'vcpus' at all ? We > >don't have that in LXC at least. > Parallels Cloud Server has that concept in both types of virt. With the full machine virt does Parallels have a separate process or thread for each vCPU, and if so can we identify the PIDs for the vCPUs to let us do proper pCPU<->vCPU mapping & reporting. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

2 2

[libvirt] [PATCH 0/3] Expose distance between host NUMA nodes
by Michal Privoznik 03 Jun '14

03 Jun '14

*** BLURB HERE *** Michal Privoznik (3): virnuma.c: Fix some comments virnuma: Introduce virNumaGetDistances virCaps: Expose distance between host NUMA nodes docs/schemas/capability.rng | 11 ++++++++ src/conf/capabilities.c | 13 +++++++++- src/conf/capabilities.h | 13 +++++++++- src/libvirt_private.syms | 1 + src/libxl/libxl_conf.c | 4 +-- src/nodeinfo.c | 61 ++++++++++++++++++++++++++++++++++++++++++--- src/test/test_driver.c | 2 +- src/util/virnuma.c | 54 +++++++++++++++++++++++++++++++++++++-- src/util/virnuma.h | 3 +++ src/xen/xend_internal.c | 3 ++- tests/vircapstest.c | 4 +-- 11 files changed, 156 insertions(+), 13 deletions(-) -- 2.0.0

2 8

[libvirt] [PATCH] network: bridge: Avoid memory leak from networkBuildDhcpDaemonCommandLine
by Peter Krempa 03 Jun '14

03 Jun '14

If the leasehelper_path couldn't be found the code would leak the freshly constructed command structure. Re-arrange code to avoid the problem. Found by coverity, broken by baafe668fa56767c031468ccd5df3e62eaa11370. --- src/network/bridge_driver.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index 51f01fb..4fc4c9a 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -1273,15 +1273,14 @@ networkBuildDhcpDaemonCommandLine(virNetworkObjPtr network, goto cleanup; } - cmd = virCommandNew(dnsmasqCapsGetBinaryPath(caps)); - virCommandAddArgFormat(cmd, "--conf-file=%s", configfile); - /* This helper is used to create custom leases file for libvirt */ if (!(leaseshelper_path = virFileFindResource("libvirt_leaseshelper", "src", LIBEXECDIR))) goto cleanup; + cmd = virCommandNew(dnsmasqCapsGetBinaryPath(caps)); + virCommandAddArgFormat(cmd, "--conf-file=%s", configfile); virCommandAddArgFormat(cmd, "--dhcp-script=%s", leaseshelper_path); *cmdout = cmd; -- 1.9.3

2 2

[libvirt] [PATCH] qemu: monitor: Fix type of holdtime argument in qemuMonitorJSONSendKey
by Peter Krempa 03 Jun '14

03 Jun '14

qemuMonitorJSONSendKey declares the "holdtime" argument as unsigned int while the command was constructed in qemuMonitorJSONMakeCommand using the "P" modifier which took a unsigned long from the variable arguments which then made it possible to access uninitialized memory. This broke the qemumonitorjsontest on 32bit fedora 20: 64) qemuMonitorJSONSendKey ... libvirt: QEMU Driver error : internal error: unsupported data type 'W' for arg 'WVSì D$0èwÿÿÃAå' FAILED Uncovered by upstream commit f744b831c66d9e82453f7a96cab5eddf7570c253. Additionally add test for the hold-time option. --- src/qemu/qemu_monitor_json.c | 2 +- tests/qemumonitorjsontest.c | 33 +++++++++++++++++++++++++++++++++ 2 files changed, 34 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 0c44b11..bedd959 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -3610,7 +3610,7 @@ int qemuMonitorJSONSendKey(qemuMonitorPtr mon, cmd = qemuMonitorJSONMakeCommand("send-key", "a:keys", keys, - "P:hold-time", holdtime, + "p:hold-time", holdtime, NULL); if (!cmd) goto cleanup; diff --git a/tests/qemumonitorjsontest.c b/tests/qemumonitorjsontest.c index 47d7481..2099dc8 100644 --- a/tests/qemumonitorjsontest.c +++ b/tests/qemumonitorjsontest.c @@ -1960,6 +1960,38 @@ testQemuMonitorJSONqemuMonitorJSONSendKey(const void *data) } static int +testQemuMonitorJSONqemuMonitorJSONSendKeyHoldtime(const void *data) +{ + virDomainXMLOptionPtr xmlopt = (virDomainXMLOptionPtr)data; + qemuMonitorTestPtr test = qemuMonitorTestNewSimple(true, xmlopt); + int ret = -1; + unsigned int keycodes[] = {43, 26, 46, 32}; + + if (!test) + return -1; + + if (qemuMonitorTestAddItemParams(test, "send-key", + "{\"return\":{}}", + "hold-time", "31337", + "keys", "[{\"type\":\"number\",\"data\":43}," + "{\"type\":\"number\",\"data\":26}," + "{\"type\":\"number\",\"data\":46}," + "{\"type\":\"number\",\"data\":32}]", + NULL, NULL) < 0) + goto cleanup; + + if (qemuMonitorJSONSendKey(qemuMonitorTestGetMonitor(test), + 31337, keycodes, + ARRAY_CARDINALITY(keycodes)) < 0) + goto cleanup; + + ret = 0; + cleanup: + qemuMonitorTestFree(test); + return ret; +} + +static int testQemuMonitorJSONqemuMonitorJSONGetDumpGuestMemoryCapability(const void *data) { virDomainXMLOptionPtr xmlopt = (virDomainXMLOptionPtr)data; @@ -2230,6 +2262,7 @@ mymain(void) DO_TEST(qemuMonitorJSONGetVirtType); DO_TEST(qemuMonitorJSONSendKey); DO_TEST(qemuMonitorJSONGetDumpGuestMemoryCapability); + DO_TEST(qemuMonitorJSONSendKeyHoldtime); DO_TEST_CPU_DATA("host"); DO_TEST_CPU_DATA("full"); -- 1.9.3

2 2

[libvirt] [PATCH] tests: Build virstoragetest only when storage driver is compiled too
by Peter Krempa 03 Jun '14

03 Jun '14

virstoragetest now requires parts of the storage driver to be built. Without this change the test can't be compiled on platforms that don't build the storage driver (mingw). make[2]: *** No rule to make target `../src/libvirt_driver_storage_impl.la', needed by `virstoragetest.exe'. Stop. Broken by commit 713cc3b0a7ff8ad42b4c13429b624d1b2b5a99f2 --- tests/Makefile.am | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/Makefile.am b/tests/Makefile.am index f9f2b84..1fdfd3b 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -162,7 +162,6 @@ test_programs = virshtest sockettest \ virstringtest \ virportallocatortest \ sysinfotest \ - virstoragetest \ virnetdevbandwidthtest \ virkmodtest \ vircapstest \ @@ -279,7 +278,8 @@ test_programs += nwfilterxml2firewalltest endif WITH_NWFILTER if WITH_STORAGE -test_programs += storagevolxml2argvtest +test_programs += storagevolxml2argvtest \ + virstoragetest endif WITH_STORAGE if WITH_LINUX -- 1.9.3

2 2

[libvirt] Idea for representing VMX virtualHW.version in domain XML
by Daniel P. Berrange 03 Jun '14

03 Jun '14

At the Xen Hackathon we were discussing dealing hypervisors which have a "virtual hardware version" concept, like VMWares' virtualHW.version. Currently our code mostly ignores the virtualHW.version when reading VMX files, and when writing them, just uses the hypervisor's default virtual hardware version. This is sub-optimal since that means the guest's hardware version could change if the hypervisor is upgraded, which could in turn cause Windows re-activation. It was suggested to me that the virtualHW.version is really akin to QEMU's machine type concept, and so we should map it that way in the XML. eg if the VMX file contains virtualHW.version = "4" Then we might represent that in the XML as <domain type='vmware'> ... <os> <type machine="vmware4">hvm</type> </os> ... </domain> ie we just invent a machine type prefix and append the version number to it. I'm not planning on implementing this myself, just sending this mail so the idea is in the historical record and some other motivated person can wire it up.... Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|

1 0

[libvirt] [PATCH 0/5] Testing libvirt XML -> libxl_domain_config conversion
by Daniel P. Berrange 03 Jun '14

03 Jun '14

At the Xen Hackathon I learnt that libxl provides an API which can serialize a libxl_domain_config instance to a JSON document. This is exactly what we need for testing the XML -> libxl_domain_config conversion process, so I spent the afternoon trying to get such a test working. The result is that we can now just add pairs of XML, JSON files to libvirt to test handling of new config features. I hit a couple of small issues with libxl, which I worked around, when writing this test which I why I'm copying xen-devel - libxl_ctx_alloc() will call xs_daemon_open and xc_interface_open, and stat /var/run/xenstored.pid to see if Xen is actually running. This fails when run on non-Xen hosts (and also possibly if run unprivileged). I used an evil LD_PRELOAD hack to stub out xs_daemon_open and xc_interface_open to return (void*)0x1, and also turn xc_interface_close and xs_daemon_close to no-ops, and make stat() always return success for xenstored.pid. This works (evidenced by the fact that if something was needing these xs/xc handles they would have crashed referencing 0x1), but at the same time it might be an idea to have an officially supported "non live" mode for libxl_ctx_alloc() turned on by a flag of some sort. - The libxl_json.h header file is relying on conditionals that are only set by Xen's build process eg HAVE_YAJL_YAJL_VERSION_H I hacked around this, but it is a little dirty too. libvirt already links to libyajl for the QEMU driver, but we don't really need the raw YAJL objects. It'd be nice to have a char * libxl_domain_config_as_json(libxl_domain_config *p) as a higher level wrapper around libxl_domain_config_gen_json avoiding the pain of dealing with YAJL's different APIs. Ian J mentioned to me that he thought there was already such a method, but AFAICT, the only such code is in the 'xl' command line tool itself (xl_cmdimpl.c - printf_info_one_json) A few further ideas that could be done as a followup - Make virConnectDomainXMLToNative accept 'xl-json" as a data format, so you can feed in a libvirt XML and get back out a JSON document. This could be a useful debugging tool for Xen developers trying to identify bugs in libvirt. - Write out the JSON document to /var/log/libvirt/libxl/$GUEST.log whenever starting a guest. Again this would be a useful debugging aid to Xen developers / support people trying to identify why a guest might be mis-behaving Regards, Daniel Daniel P. Berrange (5): Don't pass virDomainObjPtr to libxlBuildDomainConfig Don't pass libxlDriverPrivatePtr into libxlBuildDomainConfig libxl: Move virDomainXMLOptionNew into libxlCreateXMLConf Add more test suite mock helpers Add a test suite for libxl option generator src/libxl/libxl_conf.c | 38 +++--- src/libxl/libxl_conf.h | 10 +- src/libxl/libxl_domain.c | 7 +- src/libxl/libxl_driver.c | 4 +- tests/Makefile.am | 25 +++- tests/libxlxml2jsondata/minimal.json | 172 +++++++++++++++++++++++++++ tests/libxlxml2jsondata/minimal.xml | 36 ++++++ tests/libxlxml2jsontest.c | 219 +++++++++++++++++++++++++++++++++++ tests/virfirewalltest.c | 4 +- tests/virmock.h | 54 ++++++--- tests/virmocklibxl.c | 87 ++++++++++++++ tests/virsystemdtest.c | 4 +- 12 files changed, 617 insertions(+), 43 deletions(-) create mode 100644 tests/libxlxml2jsondata/minimal.json create mode 100644 tests/libxlxml2jsondata/minimal.xml create mode 100644 tests/libxlxml2jsontest.c create mode 100644 tests/virmocklibxl.c -- 1.9.3

3 17

[libvirt] [libvirt-glib] [PATCH v3] GVirDomainSnapshot: Add gvir_domain_snapshot_delete
by Timm Bäder 03 Jun '14

03 Jun '14

--- libvirt-gobject/libvirt-gobject-domain-snapshot.c | 29 +++++++++++++++++++++++ libvirt-gobject/libvirt-gobject-domain-snapshot.h | 16 +++++++++++++ libvirt-gobject/libvirt-gobject.sym | 6 +++++ 3 files changed, 51 insertions(+) diff --git a/libvirt-gobject/libvirt-gobject-domain-snapshot.c b/libvirt-gobject/libvirt-gobject-domain-snapshot.c index fcf70ed..f835b58 100644 --- a/libvirt-gobject/libvirt-gobject-domain-snapshot.c +++ b/libvirt-gobject/libvirt-gobject-domain-snapshot.c @@ -205,3 +205,32 @@ GVirConfigDomainSnapshot *gvir_domain_snapshot_get_config free(xml); return conf; } + +/** + * gvir_domain_snapshot_delete: + * @snapshot: The domain snapshot + * @flags: Bitwise or of #GVirDomainSnapshotDeleteFlags + * @error: (allow-none): Place-holder for error or NULL + * + * Returns: TRUE on success, FALSE otherwise + */ +gboolean gvir_domain_snapshot_delete (GVirDomainSnapshot *snapshot, + guint flags, + GError **error) +{ + GVirDomainSnapshotPrivate *priv; + int status; + + g_return_if_fail(GVIR_IS_DOMAIN_SNAPSHOT (snapshot)); + g_return_if_fail(error == NULL || *error == NULL); + + priv = snapshot->priv; + status = virDomainSnapshotDelete(priv->handle, flags); + if (status < 0) { + gvir_set_error(error, GVIR_DOMAIN_SNAPSHOT_ERROR, 0, + "Unable to delete snapshot `%s'", + gvir_domain_snapshot_get_name(snapshot)); + return FALSE; + } + return TRUE; +} diff --git a/libvirt-gobject/libvirt-gobject-domain-snapshot.h b/libvirt-gobject/libvirt-gobject-domain-snapshot.h index 5bd827c..b3ebe7f 100644 --- a/libvirt-gobject/libvirt-gobject-domain-snapshot.h +++ b/libvirt-gobject/libvirt-gobject-domain-snapshot.h @@ -58,6 +58,18 @@ struct _GVirDomainSnapshotClass gpointer padding[20]; }; +/** + * GVirDomainSnapshotDeleteFlags: + * @GVIR_DOMAIN_SNAPSHOT_DELETE_CHILDREN: Also delete children + * @GVIR_DOMAIN_SNAPSHOT_DELETE_METADATA_ONLY: Delete just metadata + * @GVIR_DOMAIN_SNAPSHOT_DELETE_CHILDREN_ONLY: Delete just children + */ +typedef enum { + GVIR_DOMAIN_SNAPSHOT_DELETE_CHILDREN = 1, + GVIR_DOMAIN_SNAPSHOT_DELETE_METADATA_ONLY = 2, + GVIR_DOMAIN_SNAPSHOT_DELETE_CHILDREN_ONLY = 4 +} GVirDomainSnapshotDeleteFlags; + GType gvir_domain_snapshot_get_type(void); GType gvir_domain_snapshot_handle_get_type(void); @@ -69,6 +81,10 @@ GVirConfigDomainSnapshot *gvir_domain_snapshot_get_config guint flags, GError **err); +gboolean gvir_domain_snapshot_delete (GVirDomainSnapshot *snapshot, + guint flags, + GError **error); + G_END_DECLS #endif /* __LIBVIRT_GOBJECT_DOMAIN_SNAPSHOT_H__ */ diff --git a/libvirt-gobject/libvirt-gobject.sym b/libvirt-gobject/libvirt-gobject.sym index f2419ac..232e63b 100644 --- a/libvirt-gobject/libvirt-gobject.sym +++ b/libvirt-gobject/libvirt-gobject.sym @@ -234,4 +234,10 @@ LIBVIRT_GOBJECT_0.1.5 { gvir_connection_open_read_only_finish; } LIBVIRT_GOBJECT_0.1.4; +LIBVIRT_GOBJECT_0.1.9 { + global: + gvir_domain_snapshot_delete_flags_get_type; + gvir_domain_snapshot_delete; +} LIBVIRT_GOBJECT_0.1.5; + # .... define new API here using predicted next version number .... -- 2.0.0

2 2