June 2014 - Devel - libvirt List Archives

[libvirt] [PATCH 0/4] catch bogus 'blockcommit --base=a --top=a'
by Eric Blake 16 Jun '14

16 Jun '14

This patch series was first hinted at here: https://www.redhat.com/archives/libvir-list/2014-June/msg00589.html It fixes a minor bug that has existed since blockcommit was first added, where we had an off-by-one that caused us to generate different error messages in some cases where a user was passing an impossible request to blockcommit (basically, we tried to filter bogus cases up front to give a nice message, but missed one case that resulted in reporting a different message from qemu instead). I apologize in advance for the conflicts this creates with Peter's pending series, but think that the testsuite is both more legible and more thorough with the changes added in this series. It's split into several patches to make review easier. And can I just add that it is annoying dealing with gcc's: virstoragetest.c:1076:1: error: the frame size of 6608 bytes is larger than 4096 bytes [-Werror=frame-larger-than=] (although I agree with the need for keeping that warning) Eric Blake (4): storage: add alias for less typing storage: renumber lookup tests storage: better tests of lookup blockcommit: require base below top src/qemu/qemu_driver.c | 3 +- src/util/virstoragefile.c | 33 +++++---- tests/virstoragetest.c | 184 ++++++++++++++++++++++++++++------------------ 3 files changed, 135 insertions(+), 85 deletions(-) -- 1.9.3

2 9

[libvirt] [PATCH] qemu: snapshot: Don't mark all block disks for metadata reuse
by Peter Krempa 16 Jun '14

16 Jun '14

For block devices used as snapshot source the new snapshot code would set the reuse flag. This inhibits to take snapshot without specially preparing the block image before taking the snapshot. Fortunately this is not a regression as only the new way of specifying snapshot source is affected. For the followin snapshot XML: <domainsnapshot> <disks> <disk name='vda' type='block'> <driver type='qcow2'/> <source dev="/dev/andariel/testsnap" /> </disk> </disks> </domainsnapshot> You'd get: error: internal error: unable to execute QEMU command 'transaction': Image is not in qcow2 format After this patch the snapshot is created successfully. --- src/qemu/qemu_driver.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 7bf2020..7ac7aa2 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -12885,13 +12885,11 @@ qemuDomainSnapshotCreateSingleDiskActive(virQEMUDriverPtr driver, switch ((virStorageType)snap->src->type) { case VIR_STORAGE_TYPE_BLOCK: - reuse = true; - /* fallthrough */ case VIR_STORAGE_TYPE_FILE: /* create the stub file and set selinux labels; manipulate disk in * place, in a way that can be reverted on failure. */ - if (!reuse) { + if (!reuse && snap->src->type != VIR_STORAGE_TYPE_BLOCK) { fd = qemuOpenFile(driver, vm, source, O_WRONLY | O_TRUNC | O_CREAT, &need_unlink, NULL); if (fd < 0) -- 1.9.3

2 2

[libvirt] [PATCH v2 0/2] Expose PCI Express capabilities
by Michal Privoznik 16 Jun '14

16 Jun '14

diff to v1: -Matin's review suggestions worked in -I've found that (testing this on my new HW) that some PCI-Express devices doesn't have to necessarily export link info. That's why I'm inventing this virPCIDeviceHasPCIExpressLink() which tells us if that's the case. So for a PCI Express device without link info exposed, the XML will simply contain <pci-express/>. Michal Privoznik (2): virpci: Introduce virPCIDeviceIsPCIExpress and friends nodedev: Introduce <pci-express/> to PCI devices docs/formatnode.html.in | 19 +++ docs/schemas/nodedev.rng | 26 ++++ src/conf/node_device_conf.c | 138 ++++++++++++++++++++- src/conf/node_device_conf.h | 33 ++++- src/libvirt_private.syms | 3 + src/node_device/node_device_udev.c | 34 +++++ src/util/virpci.c | 96 +++++++++++++- src/util/virpci.h | 8 ++ .../pci_8086_0c0c_snd_hda_intel.xml | 16 +++ .../pci_8086_4238_pcie_wireless.xml | 26 ++++ tests/nodedevxml2xmltest.c | 2 + 11 files changed, 397 insertions(+), 4 deletions(-) create mode 100644 tests/nodedevschemadata/pci_8086_0c0c_snd_hda_intel.xml create mode 100644 tests/nodedevschemadata/pci_8086_4238_pcie_wireless.xml -- 1.8.5.5

2 4

[libvirt] [PATCH] virsh: Parse "--mode readonly, shareable" in virsh attach-disk.
by Dongsheng Yang 16 Jun '14

16 Jun '14

We can not attach a disk with mode of both readonly and shareable. Before: # virsh attach-disk --domain jeos --source attach.img --target vdb --persistent --mode readonly --mode shareable error: option --mode already seen # virsh attach-disk --domain jeos --source attach.img --target vdb --persistent --mode readonly --shareable error: option --mode already seen # virsh attach-disk --domain jeos --source attach.img --target vdb --persistent --mode readonly,shareable error: No support for readonly,shareable in command 'attach-disk' After: # virsh attach-disk --domain jeos --source attach.img --target vdb --persistent --mode readonly,shareable Disk attached successfully Signed-off-by: Dongsheng Yang <yangds.fnst(a)cn.fujitsu.com> --- tools/virsh-domain.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 6b3dd70..62a9824 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -515,6 +515,7 @@ cmdAttachDisk(vshControl *ctl, const vshCmd *cmd) bool config = vshCommandOptBool(cmd, "config"); bool live = vshCommandOptBool(cmd, "live"); bool persistent = vshCommandOptBool(cmd, "persistent"); + char **modes = NULL, **tmp; VSH_EXCLUSIVE_OPTIONS_VAR(persistent, current); @@ -553,12 +554,9 @@ cmdAttachDisk(vshControl *ctl, const vshCmd *cmd) goto cleanup; } - if (mode) { - if (STRNEQ(mode, "readonly") && STRNEQ(mode, "shareable")) { - vshError(ctl, _("No support for %s in command 'attach-disk'"), - mode); - goto cleanup; - } + if (mode && !(modes = virStringSplit(mode, ",", 0))) { + vshError(ctl, "%s", _("Cannot parse mode string")); + goto cleanup; } if (wwn && !virValidateWWN(wwn)) @@ -591,8 +589,19 @@ cmdAttachDisk(vshControl *ctl, const vshCmd *cmd) virBufferAsprintf(&buf, "<source %s='%s'/>\n", isFile ? "file" : "dev", source); virBufferAsprintf(&buf, "<target dev='%s'/>\n", target); - if (mode) - virBufferAsprintf(&buf, "<%s/>\n", mode); + + tmp = modes; + while (tmp && *tmp) { + mode = *tmp; + if (STREQ(mode, "readonly") || STREQ(mode, "shareable")) { + virBufferAsprintf(&buf, "<%s/>\n", mode); + } else { + vshError(ctl, _("Unknown mode %s value, expecting " + "'readonly', 'shareable'"), mode); + goto cleanup; + } + tmp++; + } if (serial) virBufferAsprintf(&buf, "<serial>%s</serial>\n", serial); -- 1.8.3.1

2 1

[libvirt] [PATCH] leaseshelper: fix another creash
by Pavel Hrdina 16 Jun '14

16 Jun '14

We create a 'lease_new' when we are adding new lease entry, then later in the code we add the 'lease_new' into a 'leases_array_new' which leades into the creash because we double free the 'lease_new'. To prevent the double free we set the 'lease_new' to NULL after successful append into the 'leases_array_new'. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- src/network/leaseshelper.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/network/leaseshelper.c b/src/network/leaseshelper.c index 69081c3..bf1842b 100644 --- a/src/network/leaseshelper.c +++ b/src/network/leaseshelper.c @@ -331,6 +331,7 @@ main(int argc, char **argv) _("failed to create json")); goto cleanup; } + lease_new = NULL; } if (!(leases_str = virJSONValueToString(leases_array_new, true))) { -- 1.8.5.5

2 3

[libvirt] [libvirt-sandbox PATCH v2] Only set SELinux seclabel if supported by the host.
by Cédric Bosdonnat 16 Jun '14

16 Jun '14

This code depends on new API in libvirt-gconfig to extract the secmodels handled by the host. --- Diff to v1: * Adapted the naming change from libvirt-gconfig libvirt-sandbox/libvirt-sandbox-builder.c | 68 ++++++++++++++++++++----------- 1 file changed, 45 insertions(+), 23 deletions(-) diff --git a/libvirt-sandbox/libvirt-sandbox-builder.c b/libvirt-sandbox/libvirt-sandbox-builder.c index 48b3acc..327f144 100644 --- a/libvirt-sandbox/libvirt-sandbox-builder.c +++ b/libvirt-sandbox/libvirt-sandbox-builder.c @@ -323,38 +323,60 @@ static gboolean gvir_sandbox_builder_construct_devices(GVirSandboxBuilder *build } -static gboolean gvir_sandbox_builder_construct_security(GVirSandboxBuilder *builder G_GNUC_UNUSED, +static gboolean gvir_sandbox_builder_construct_security(GVirSandboxBuilder *builder, GVirSandboxConfig *config G_GNUC_UNUSED, const gchar *statedir G_GNUC_UNUSED, GVirConfigDomain *domain, - GError **error G_GNUC_UNUSED) + GError **error) { GVirConfigDomainSeclabel *sec = gvir_config_domain_seclabel_new(); const char *label = gvir_sandbox_config_get_security_label(config); + GVirConnection *connection = gvir_sandbox_builder_get_connection(builder); + GVirConfigCapabilities *configCapabilities; + GVirConfigCapabilitiesHost *hostCapabilities; + GList *secmodels, *iter; + gboolean supportsSelinux = FALSE; + + /* What security models are available on the host? */ + if (!(configCapabilities = gvir_connection_get_capabilities(connection, error))) { + g_object_unref(sec); + return FALSE; + } + + hostCapabilities = gvir_config_capabilities_get_host(configCapabilities); - gvir_config_domain_seclabel_set_model(sec, "selinux"); - if (gvir_sandbox_config_get_security_dynamic(config)) { - gvir_config_domain_seclabel_set_type(sec, - GVIR_CONFIG_DOMAIN_SECLABEL_DYNAMIC); - if (label) - gvir_config_domain_seclabel_set_baselabel(sec, label); - else if (gvir_config_domain_get_virt_type(domain) == - GVIR_CONFIG_DOMAIN_VIRT_LXC) - gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_lxc_net_t:s0"); - else if (gvir_config_domain_get_virt_type(domain) == - GVIR_CONFIG_DOMAIN_VIRT_QEMU) - gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_tcg_t:s0"); - else if (gvir_config_domain_get_virt_type(domain) == - GVIR_CONFIG_DOMAIN_VIRT_KVM) - gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_t:s0"); - } else { - gvir_config_domain_seclabel_set_type(sec, - GVIR_CONFIG_DOMAIN_SECLABEL_STATIC); - if (label) - gvir_config_domain_seclabel_set_label(sec, label); + secmodels = gvir_config_capabilities_host_get_secmodels(hostCapabilities); + for (iter = secmodels; iter != NULL; iter = iter->next) { + supportsSelinux = g_str_equal(gvir_config_capabilities_host_secmodel_get_model( + GVIR_CONFIG_CAPABILITIES_HOST_SECMODEL(iter->data)), "selinux"); + g_object_unref(iter->data); } - gvir_config_domain_set_seclabel(domain, sec); + if (supportsSelinux) { + gvir_config_domain_seclabel_set_model(sec, "selinux"); + if (gvir_sandbox_config_get_security_dynamic(config)) { + gvir_config_domain_seclabel_set_type(sec, + GVIR_CONFIG_DOMAIN_SECLABEL_DYNAMIC); + if (label) + gvir_config_domain_seclabel_set_baselabel(sec, label); + else if (gvir_config_domain_get_virt_type(domain) == + GVIR_CONFIG_DOMAIN_VIRT_LXC) + gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_lxc_net_t:s0"); + else if (gvir_config_domain_get_virt_type(domain) == + GVIR_CONFIG_DOMAIN_VIRT_QEMU) + gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_tcg_t:s0"); + else if (gvir_config_domain_get_virt_type(domain) == + GVIR_CONFIG_DOMAIN_VIRT_KVM) + gvir_config_domain_seclabel_set_baselabel(sec, "system_u:system_r:svirt_t:s0"); + } else { + gvir_config_domain_seclabel_set_type(sec, + GVIR_CONFIG_DOMAIN_SECLABEL_STATIC); + if (label) + gvir_config_domain_seclabel_set_label(sec, label); + } + + gvir_config_domain_set_seclabel(domain, sec); + } g_object_unref(sec); return TRUE; -- 1.8.4.5

3 2

[libvirt] [PATCH 0/5] Expose host's huge pages capability
by Michal Privoznik 16 Jun '14

16 Jun '14

*** BLURB HERE *** Michal Privoznik (5): virnuma: Introduce huge pages helpers virCaps: expose huge page info Introduce virNodeGetFreePages virsh: Expose virNodeGetFreePages qemu: Implement virNodeGetFreePages daemon/remote.c | 52 ++++++++++ docs/schemas/capability.rng | 21 ++++ include/libvirt/libvirt.h.in | 7 ++ src/conf/capabilities.c | 25 ++++- src/conf/capabilities.h | 15 ++- src/driver.h | 10 ++ src/libvirt.c | 95 +++++++++++++++++ src/libvirt_private.syms | 2 + src/libvirt_public.syms | 4 + src/libxl/libxl_conf.c | 1 + src/nodeinfo.c | 41 +++++++- src/qemu/qemu_capabilities.c | 29 +++++- src/qemu/qemu_driver.c | 43 ++++++++ src/remote/remote_driver.c | 50 +++++++++ src/remote/remote_protocol.x | 20 +++- src/remote_protocol-structs | 16 +++ src/test/test_driver.c | 2 +- src/util/virnuma.c | 242 +++++++++++++++++++++++++++++++++++++++++++ src/util/virnuma.h | 10 ++ src/xen/xend_internal.c | 1 + tests/vircaps2xmltest.c | 3 +- tests/vircapstest.c | 1 + tools/virsh-host.c | 167 +++++++++++++++++++++++++++++ tools/virsh.pod | 8 ++ 24 files changed, 856 insertions(+), 9 deletions(-) -- 1.8.5.5

3 15

[libvirt] [PATCH v3] QEMU: parse '-device vfio-pci' and '-device pci-assign'
by Olivia Yin 16 Jun '14

16 Jun '14

Signed-off-by: Olivia Yin <Hong-Hua.Yin(a)freescale.com> Signed-off-by: Laine Stump <laine(a)laine.org> Modify the existing function qemuParseCommandLinePCI(), which works with -pcidevice, to support for "-device pci-assign" and "-device pci-assign". Change test cases 'hostdev-vfio' and 'hostdev-pci-address-device' to validate the new function. The case related to QEMU_CAPS_HOST_PCI_MULTIDOMAIN which uses 'host=domain:bus:slot.func' is not supported yet. --- src/qemu/qemu_command.c | 50 +++++++++++++++++++--- tests/qemuargv2xmltest.c | 3 +- .../qemuxml2argv-hostdev-pci-address-device.xml | 6 +++ .../qemuxml2argvdata/qemuxml2argv-hostdev-vfio.xml | 1 + 4 files changed, 53 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index e1d7e1b..3a4bc61 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -10193,7 +10193,8 @@ qemuParseCommandLineNet(virDomainXMLOptionPtr xmlopt, * Tries to parse a QEMU PCI device */ static virDomainHostdevDefPtr -qemuParseCommandLinePCI(const char *val) +qemuParseCommandLinePCI(const char *val, + virDomainHostdevSubsysPCIBackendType backend) { int bus = 0, slot = 0, func = 0; const char *start; @@ -10222,10 +10223,20 @@ qemuParseCommandLinePCI(const char *val) goto error; } start = end + 1; - if (virStrToLong_i(start, NULL, 16, &func) < 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot extract PCI device function '%s'"), val); - goto error; + + if (backend) { + if (virStrToLong_i(start, &end, 16, &func) < 0 || *end != ',') { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("cannot extract PCI device function '%s'"), val); + goto error; + } else + def->source.subsys.u.pci.backend = backend; + } else { + if (virStrToLong_i(start, NULL, 16, &func) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("cannot extract PCI device function '%s'"), val); + goto error; + } } def->mode = VIR_DOMAIN_HOSTDEV_MODE_SUBSYS; @@ -11347,7 +11358,34 @@ qemuParseCommandLine(virCapsPtr qemuCaps, } else if (STREQ(arg, "-pcidevice")) { virDomainHostdevDefPtr hostdev; WANT_VALUE(); - if (!(hostdev = qemuParseCommandLinePCI(val))) + if (!(hostdev = qemuParseCommandLinePCI(val, + VIR_DOMAIN_HOSTDEV_PCI_BACKEND_DEFAULT))) + goto error; + if (VIR_APPEND_ELEMENT(def->hostdevs, def->nhostdevs, hostdev) < 0) { + virDomainHostdevDefFree(hostdev); + goto error; + } + } else if (STREQ(arg, "-device") && progargv[i+1] && + STRPREFIX(progargv[i+1], "vfio-pci")) { + const char *start; + virDomainHostdevDefPtr hostdev; + WANT_VALUE(); + start = val + strlen("vfio-pci,"); + if (!(hostdev = qemuParseCommandLinePCI(start, + VIR_DOMAIN_HOSTDEV_PCI_BACKEND_VFIO))) + goto error; + if (VIR_APPEND_ELEMENT(def->hostdevs, def->nhostdevs, hostdev) < 0) { + virDomainHostdevDefFree(hostdev); + goto error; + } + } else if (STREQ(arg, "-device") && progargv[i+1] && + STRPREFIX(progargv[i+1], "pci-assign")) { + const char *start; + virDomainHostdevDefPtr hostdev; + WANT_VALUE(); + start = val + strlen("pci-assign,"); + if (!(hostdev = qemuParseCommandLinePCI(start, + VIR_DOMAIN_HOSTDEV_PCI_BACKEND_KVM))) goto error; if (VIR_APPEND_ELEMENT(def->hostdevs, def->nhostdevs, hostdev) < 0) { virDomainHostdevDefFree(hostdev); diff --git a/tests/qemuargv2xmltest.c b/tests/qemuargv2xmltest.c index 2cbbe3d..80188be 100644 --- a/tests/qemuargv2xmltest.c +++ b/tests/qemuargv2xmltest.c @@ -274,8 +274,9 @@ mymain(void) DO_TEST("watchdog"); DO_TEST("hostdev-usb-address"); - DO_TEST("hostdev-pci-address"); + DO_TEST("hostdev-pci-address-device"); + DO_TEST("hostdev-vfio"); DO_TEST("smp"); diff --git a/tests/qemuxml2argvdata/qemuxml2argv-hostdev-pci-address-device.xml b/tests/qemuxml2argvdata/qemuxml2argv-hostdev-pci-address-device.xml index b29ef58..b9a221a 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-hostdev-pci-address-device.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-hostdev-pci-address-device.xml @@ -15,10 +15,16 @@ <devices> <emulator>/usr/bin/qemu</emulator> <disk type='block' device='disk'> + <driver name='qemu' type='raw'/> <source dev='/dev/HostVG/QEMUGuest2'/> <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> + <controller type='usb' index='0'/> + <controller type='ide' index='0'/> + <controller type='pci' index='0' model='pci-root'/> <hostdev mode='subsystem' type='pci' managed='yes'> + <driver name='kvm'/> <source> <address domain='0x0000' bus='0x06' slot='0x12' function='0x5'/> </source> diff --git a/tests/qemuxml2argvdata/qemuxml2argv-hostdev-vfio.xml b/tests/qemuxml2argvdata/qemuxml2argv-hostdev-vfio.xml index 8daa53a..b99f798 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-hostdev-vfio.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-hostdev-vfio.xml @@ -15,6 +15,7 @@ <devices> <emulator>/usr/bin/qemu</emulator> <disk type='block' device='disk'> + <driver name='qemu' type='raw'/> <source dev='/dev/HostVG/QEMUGuest2'/> <target dev='hda' bus='ide'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> -- 1.8.5

2 1

[libvirt] [PATCH] bhyve: silent destroy command errors on cleanup
by Roman Bogorodskiy 13 Jun '14

13 Jun '14

When virBhyveProcessStart() fails, it tries to unload a guest that could have been already loaded using bhyveload(8) to make sure not to leave it hanging in memory. However, we could fail before loading a VM into memory, so 'bhyvectl --destroy' command will fail and print an error message that looks confusing to users. So ignore errors when running this in cleanup. --- src/bhyve/bhyve_process.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/bhyve/bhyve_process.c b/src/bhyve/bhyve_process.c index b8db076..36c2c21 100644 --- a/src/bhyve/bhyve_process.c +++ b/src/bhyve/bhyve_process.c @@ -189,12 +189,13 @@ virBhyveProcessStart(virConnectPtr conn, cleanup: if (ret < 0) { + int exitstatus = -1; virCommandPtr destroy_cmd; if ((destroy_cmd = virBhyveProcessBuildDestroyCmd(driver, vm->def)) != NULL) { virCommandSetOutputFD(load_cmd, &logfd); virCommandSetErrorFD(load_cmd, &logfd); - ignore_value(virCommandRun(destroy_cmd, NULL)); + ignore_value(virCommandRun(destroy_cmd, &exitstatus)); virCommandFree(destroy_cmd); } -- 1.9.0

2 2

[libvirt] [PATCH 0/2] bhyve: fix crashes when running without networking
by Roman Bogorodskiy 13 Jun '14

13 Jun '14

Craig Rodrigues reported a problem on freebsd-virtualization list when libvirtd with bhyve driver crashes when the bridge does not exist [1]. 1: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=1431191+0+current/freebsd-virt… These two patches addresses the problem. Roman Bogorodskiy (2): bhyve: fix crash in bhyveBuildNetArgStr bhyve: do not cleanup unallocated networks on fail src/bhyve/bhyve_command.c | 3 +-- src/bhyve/bhyve_process.c | 10 ++++++---- 2 files changed, 7 insertions(+), 6 deletions(-) -- 1.9.0

2 6