- Devel - libvirt List Archives

[libvirt] [PATCH] cpu: Add Intel Haswell cpu model
by Peter Krempa 26 Nov '12

26 Nov '12

The new model supports following features in addition to those supported by SandyBridge: fma, movbe, fsgsbase, bmi1, hle, avx2, smep, bmi2, erms, invpcid, rtm --- Based on: https://lists.gnu.org/archive/html/qemu-devel/2012-11/msg01328.html --- src/cpu/cpu_map.xml | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/src/cpu/cpu_map.xml b/src/cpu/cpu_map.xml index 7ff91be..eb69a34 100644 --- a/src/cpu/cpu_map.xml +++ b/src/cpu/cpu_map.xml @@ -495,6 +495,22 @@ <feature name='rdtscp'/> </model> + <model name='Haswell'> + <model name='SandyBridge'/> + <feature name='fma'/> + <feature name='pcid'/> + <feature name='movbe'/> + <feature name='fsgsbase'/> + <feature name='bmi1'/> + <feature name='hle'/> + <feature name='avx2'/> + <feature name='smep'/> + <feature name='bmi2'/> + <feature name='erms'/> + <feature name='invpcid'/> + <feature name='rtm'/> + </model> +  <model name='athlon'> <model name='pentiumpro'/> -- 1.8.0

2 4

[libvirt] [PATCH] storage: fix logical volume cloning
by Ján Tomko 26 Nov '12

26 Nov '12

Commit 258e06c removed setting of the volume type to VIR_STORAGE_VOL_BLOCK, which leads to failures in storageVolumeCreateXMLFrom. The type (and target.format) of the volume was set to zero. In virStorageBackendGetBuildVolFromFunction, this gets interpreted as VIR_STORAGE_FILE_NONE and the qemu-img tool is called with unknown "none" format. Bug: https://bugzilla.redhat.com/show_bug.cgi?id=879780 --- src/storage/storage_backend_logical.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/storage/storage_backend_logical.c b/src/storage/storage_backend_logical.c index de43c3a..fd5cbd1 100644 --- a/src/storage/storage_backend_logical.c +++ b/src/storage/storage_backend_logical.c @@ -708,6 +708,8 @@ virStorageBackendLogicalCreateVol(virConnectPtr conn, return -1; } + vol->type = VIR_STORAGE_VOL_BLOCK; + if (vol->target.path != NULL) { /* A target path passed to CreateVol has no meaning */ VIR_FREE(vol->target.path); -- 1.7.8.6

3 2

[libvirt] [PATCH] build: fix build --without-network
by Ján Tomko 26 Nov '12

26 Nov '12

bridge_driver.h: silence gcc warnings: statement with no effect [-Wunused-value] unused variable 'net' [-Wunused-variable] virdrivermoduletest.c: don't require network driver module if it hasn't been built. --- src/network/bridge_driver.h | 4 ++-- tests/virdrivermoduletest.c | 8 ++++++++ 2 files changed, 10 insertions(+), 2 deletions(-) diff --git a/src/network/bridge_driver.h b/src/network/bridge_driver.h index 0fae275..1c42d81 100644 --- a/src/network/bridge_driver.h +++ b/src/network/bridge_driver.h @@ -53,8 +53,8 @@ int networkBuildDhcpDaemonCommandLine(virNetworkObjPtr network, # else /* Define no-op replacements that don't drag in any link dependencies. */ # define networkAllocateActualDevice(iface) 0 -# define networkNotifyActualDevice(iface) 0 -# define networkReleaseActualDevice(iface) 0 +# define networkNotifyActualDevice(iface) (iface=iface,0) +# define networkReleaseActualDevice(iface) (iface=iface,0) # define networkGetNetworkAddress(netname, netaddr) (-2) # define networkBuildDhcpDaemonCommandLine(network, cmdout, pidfile, dctx) 0 # endif diff --git a/tests/virdrivermoduletest.c b/tests/virdrivermoduletest.c index 9c14e79..0123c36 100644 --- a/tests/virdrivermoduletest.c +++ b/tests/virdrivermoduletest.c @@ -83,10 +83,18 @@ mymain(void) TEST("interface", NULL); #endif #ifdef WITH_QEMU +# ifdef WITH_NETWORK TEST("qemu", "network"); +# else + TEST("qemu", NULL); +# endif #endif #ifdef WITH_LXC +# ifdef WITH_NETWORK TEST("lxc", "network"); +# else + TEST("lxc", NULL); +# endif #endif #ifdef WITH_UML TEST("uml", NULL); -- 1.7.8.6

4 3

[libvirt] libvirt [PATCHv2 0/2] Refactor ESX storage driver and append support for iSCSI devices.
by Ata E Husain Bohra 26 Nov '12

26 Nov '12

Major changes: Split changes to two parts: a. Refactor ESX storage driver to implement facade design. b. Add backend driver to support iSCSI devices. Ata E Husain Bohra (2): Refactor ESX storage driver to implement facade pattern Add iSCSI backend storage driver for ESX. daemon/remote.c | 6 +- src/Makefile.am | 2 + src/conf/storage_conf.c | 3 +- src/datatypes.c | 25 +- src/datatypes.h | 24 +- src/esx/esx_driver.c | 6 +- src/esx/esx_storage_backend_iscsi.c | 807 +++++++++++++++++++ src/esx/esx_storage_backend_iscsi.h | 29 + src/esx/esx_storage_backend_vmfs.c | 1491 +++++++++++++++++++++++++++++++++++ src/esx/esx_storage_backend_vmfs.h | 30 + src/esx/esx_storage_driver.c | 1325 +++++-------------------------- src/esx/esx_vi.c | 337 +++++++- src/esx/esx_vi.h | 21 +- src/esx/esx_vi_generator.input | 302 +++++++ src/esx/esx_vi_generator.py | 19 + src/parallels/parallels_storage.c | 24 +- src/remote/remote_driver.c | 6 +- src/storage/storage_driver.c | 28 +- src/test/test_driver.c | 30 +- src/vbox/vbox_tmpl.c | 14 +- 20 files changed, 3357 insertions(+), 1172 deletions(-) create mode 100644 src/esx/esx_storage_backend_iscsi.c create mode 100644 src/esx/esx_storage_backend_iscsi.h create mode 100644 src/esx/esx_storage_backend_vmfs.c create mode 100644 src/esx/esx_storage_backend_vmfs.h -- 1.7.9.5

3 2

[libvirt] [PATCH 0/6 v2] Unprivileged SG_IO support
by Osier Yang 26 Nov '12

26 Nov '12

Hi, As a result of RFC [1], this implements the unprivleged SG_IO support. Though it is a bit rushed, and I still have no chance to test it yet, but I'd like see an earlier reviewing. @Paolo, please let me known once you have a kernel build with the new sysfs knob supported. Thanks! Osier Yang (6): unpriv_sgio: Add docs and rng schema for new XML unpriv_sgio unpriv_sgio: Parse and format the new XML unpriv_sgio: Prepare helpers for unpriv_sgio setting unpriv_sgio: Manage unpriv_sgio in domain's lifecyle unpriv_sgio: Do not restore unpriv_sgio if the disk is being used unpriv_sgio: Error out if the unpriv_sgio setting conflicts with others docs/formatdomain.html.in | 10 +- docs/schemas/domaincommon.rng | 52 ++++-- src/conf/domain_conf.c | 168 ++++++++++++++++++-- src/conf/domain_conf.h | 18 ++ src/libvirt_private.syms | 6 + src/qemu/qemu_process.c | 54 +++++++ src/util/util.c | 154 ++++++++++++++++++ src/util/util.h | 7 + ...l2argv-disk-scsi-lun-passthrough-upriv-sgio.xml | 32 ++++ tests/qemuxml2xmltest.c | 1 + 10 files changed, 472 insertions(+), 30 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-scsi-lun-passthrough-upriv-sgio.xml [1] https://www.redhat.com/archives/libvir-list/2012-November/msg00988.html Regards, Osier

1 6

[libvirt] [PATCH] util: Use virReportSystemError for system error in pci.c
by Osier Yang 26 Nov '12

26 Nov '12

--- src/util/pci.c | 14 ++++++-------- 1 files changed, 6 insertions(+), 8 deletions(-) diff --git a/src/util/pci.c b/src/util/pci.c index d1ad121..191f99d 100644 --- a/src/util/pci.c +++ b/src/util/pci.c @@ -1860,10 +1860,9 @@ pciGetPciConfigAddressFromSysfsDeviceLink(const char *device_link, device_path = canonicalize_file_name(device_link); if (device_path == NULL) { memset(errbuf, '\0', sizeof(errbuf)); - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Failed to resolve device link '%s': '%s'"), - device_link, virStrerror(errno, errbuf, - sizeof(errbuf))); + virReportSystemError(errno, + _("Failed to resolve device link '%s'"), + device_link); return ret; } @@ -1941,10 +1940,9 @@ pciGetVirtualFunctions(const char *sysfs_path, dir = opendir(sysfs_path); if (dir == NULL) { memset(errbuf, '\0', sizeof(errbuf)); - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Failed to open dir '%s': '%s'"), - sysfs_path, virStrerror(errno, errbuf, - sizeof(errbuf))); + virReportSystemError(errno, + _("Failed to open dir '%s'"), + sysfs_path); return ret; } -- 1.7.7.6

2 2

[libvirt] [PATCH] util: Fix the indention
by Osier Yang 25 Nov '12

25 Nov '12

--- Pushed under trivial rule. --- src/util/cgroup.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/src/util/cgroup.c b/src/util/cgroup.c index 3f7b5f7..41a6e4a 100644 --- a/src/util/cgroup.c +++ b/src/util/cgroup.c @@ -1189,10 +1189,10 @@ int virCgroupSetBlkioDeviceWeight(virCgroupPtr group, return ret; } #else - int - virCgroupSetBlkioDeviceWeight(virCgroupPtr group ATTRIBUTE_UNUSED, - const char *path ATTRIBUTE_UNUSED, - unsigned int weight ATTRIBUTE_UNUSED) +int +virCgroupSetBlkioDeviceWeight(virCgroupPtr group ATTRIBUTE_UNUSED, + const char *path ATTRIBUTE_UNUSED, + unsigned int weight ATTRIBUTE_UNUSED) { return -ENOSYS; } -- 1.7.7.6

1 0

[libvirt] image transfer incompletely in live migration
by yli＠cloudiya.com 25 Nov '12

25 Nov '12

Hi, I am using libvirt version 0.10.1 and qemu-kvm version 1.2.0 with RHEL 6.3 When I use libvirt API to make block migrate with a VM, it failed frequently. I check the migrated VM and found that the filesystem damaged. But sometimes it works very fine, I don't know why, it seems to be random. All failed tests have the same condition, image transfer incompletely in live migration. The image I used is 10GB size, after a failed migration it has only about 400MB to 3GB in remote host. Here is the code: import libvirt from xml.dom import minidom name = "vm-01" host = libvirt.open("qemu:///system") vm = host.lookupByNmae(name) vm_xml_file = "/etc/libvirt/qemu/%s.xml" % name vm_xml = minidom.parse(vm_xml_file).documentElement.toxml() remote_host = "qemu+ssh://host02/system" remote = libvirt.open(remote_host) try: vm.migrate2(remote, vm_xml, 89, name, None, 0) except Exception,e: print str(e) Before run the code I have already create a image for migration in remote host 'host02'. Path, size, and privilege all OK. Here is the related log: in source host 'host01': 2012-11-24 05:14:40.001+0000: 24888: warning : qemuDomainObjBeginJobInternal:838 : Cannot start job (query, none) for domain wrksp-776; current job is (async nested, migration out) owned by (24890, 24890) 2012-11-24 05:16:16.398+0000: 24888: error : qemuDomainObjBeginJobInternal:842 : Timed out during operation: cannot acquire state change lock 2012-11-24 05:16:21.317+0000: 24887: error : virNetSocketReadWire:1184 : End of file while reading data: Input/output error in remote host 'host02': 2012-11-24 05:14:00.776+0000: 11065: warning : qemuDomainObjEnterMonitorInternal:993 : This thread seems to be the async job owner; entering monitor without asking for a nested job is dangerous 2012-11-24 05:14:01.196+0000: 11065: warning : qemuDomainObjEnterMonitorInternal:993 : This thread seems to be the async job owner; entering monitor without asking for a nested job is dangerous 2012-11-24 05:14:01.198+0000: 11065: warning : qemuDomainObjEnterMonitorInternal:993 : This thread seems to be the async job owner; entering monitor without asking for a nested job is dangerous 2012-11-24 05:14:01.208+0000: 11065: warning : qemuDomainObjEnterMonitorInternal:993 : This thread seems to be the async job owner; entering monitor without asking for a nested job is dangerous 2012-11-24 05:14:01.211+0000: 11065: warning : qemuDomainObjEnterMonitorInternal:993 : This thread seems to be the async job owner; entering monitor without asking for a nested job is dangerous I try to use command "virsh migrate vm-01 --live --copy-storage-all --persistent --undefinesource qemu+ssh://host02/system --verbose", the test results have no change, common but not regular. Even after my once successful test, it will certainly fail when I try to make a second migrate with the same VM. How can I solve this problem？ Anything is helpful. Thanks.

1 0

[libvirt] [PATCH 0/6] Misc unrelated fixes
by Daniel P. Berrange 25 Nov '12

25 Nov '12

The following are miscellaneous patches fixes various issues I've come across today, while hacking on LXC

4 12

[libvirt] [BUG] storage.xml: owner|group=-1 → 2^32-1 on 32 Bit
by Philipp Hahn 23 Nov '12

23 Nov '12

Hello, when you install and use libvirt in a 32 bit environment, the handling of owner and group in /etc/libvirt/storage/*.xml is broken: The initial -1 is printed as an "unsigned int" and is thus converted to +(2^32-1): > # cat /etc/libvirt/storage/default.xml >  > > <pool type='dir'> > <name>default</name> > <uuid>4e2670b8-9dc3-528e-6fc2-58268e99f44f</uuid> > <capacity unit='bytes'>0</capacity> > <allocation unit='bytes'>0</allocation> > <available unit='bytes'>0</available> > <source> > </source> > <target> > <path>/var/lib/libvirt/images</path> > <permissions> > <mode>0700</mode> > <owner>4294967295</owner> > <group>4294967295</group> > </permissions> > </target> > </pool> When you next restart libvirtd virStorageDefParsePerms() tries to parse that as an "signed int", which fails in virXPathLongBase() because of > 182 if (*value != obj->floatval) { > (gdb) print *value > $11 = -2147483648 > (gdb) print obj->floatval > $12 = 4294967295 > (gdb) print /x *value > $13 = 0x80000000 > (gdb) print /x obj->floatval > $14 = 0xffffffff This then prints the following error messages: > 2012-11-23 14:03:18.877+0000: 25536: error : virStorageDefParsePerms:613 : > XML error: malformed owner element > 2012-11-23 14:03:19.581+0000: > 25529: error : storagePoolSetAutostart:1056 : Failed to create symlink > '/etc/libvirt/storage/autostart/default.xml' to '/etc/libvirt/storage/def The change to print owner and group as %d instead of %u was introdeces in commit 37a10129: > Update xml schemas according to libvirt source ... > storage_conf: Print uid_t and gid_t as signed to storage pool XML. Care to revert that change? Patch is attached. Sincerely Philipp -- Philipp Hahn Open Source Software Engineer hahn(a)univention.de Univention GmbH be open. fon: +49 421 22 232- 0 Mary-Somerville-Str.1 D-28359 Bremen fax: +49 421 22 232-99 http://www.univention.de/ diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 7944555..b07a7aa 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -1036,10 +1036,10 @@ virStoragePoolDefFormat(virStoragePoolDefPtr def) { virBufferAddLit(&buf," <permissions>\n"); virBufferAsprintf(&buf," <mode>0%o</mode>\n", def->target.perms.mode); - virBufferAsprintf(&buf," <owner>%d</owner>\n", - (int) def->target.perms.uid); - virBufferAsprintf(&buf," <group>%d</group>\n", - (int) def->target.perms.gid); + virBufferAsprintf(&buf," <owner>%u</owner>\n", + (unsigned int) def->target.perms.uid); + virBufferAsprintf(&buf," <group>%u</group>\n", + (unsigned int) def->target.perms.gid);

2 1