May 2012 - Devel - libvirt List Archives

[libvirt] [PATCH v4 0/3] usb devices with same vendorID, productID hotplug support
by Guannan Ren 16 May '12

16 May '12

v4 updates: make each patch not break build make usbDeviceSearch flags filtering code simpler https://bugzilla.redhat.com/show_bug.cgi?id=815755 The set of patch tries to fix the issue when multiple usb devices with same idVendor, idProduct are availible on host, the usb device with lowest bus:device will be attached to guest if usb xml file is given like this: <hostdev mode='subsystem' type='usb' managed='yes'> <source> <vendor id='0x15e1'/> <product id='0x2007'/> </source> </hostdev> The reason is that the usb hotplug function searchs usb device in system files to match vendor and product id, the file with lowest number is always found first. After fix, in this case, libvirt will report an error like: # virsh attach-device rhel6u1 /tmp/usb.xml error: Failed to attach device from /tmp/usb.xml error: operation failed: multiple USB devices for 15e1:2007, use <address> to specify one At the same time, the usb part of domain initilization is also updated in patch 2/3 These patches also fix the problem when using the following xml, the usb device could be hotplugged in two domains without raising errors <hostdev mode='subsystem' type='usb' managed='yes'> <source> <address bus='6' device='7'/> </source> </hostdev> # virsh attach-device rhel6u12nd /tmp/usb_by_bus.xml error: Failed to attach device from /tmp/usb_by_bus.xml error: Requested operation is not valid: USB device 006:007 is in use by domain rhel6u1

3 11

[libvirt] [PATCH v2] Remount cgroups controllers after setting up new /sys in LXC
by Daniel P. Berrange 16 May '12

16 May '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> Normal practice is for cgroups controllers to be mounted at /sys/fs/cgroup. When setting up a container, /sys is mounted with a new sysfs instance, thus we must re-mount all the cgroups controllers. The complexity is that we must mount them in the same layout as the host OS. ie if 'cpu' and 'cpuacct' were mounted at the same location in the host we must preserve this in the container. Also if any controllers are co-located we must setup symlinks from the individual controller name to the co-located mount-point Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/lxc/lxc_container.c | 243 +++++++++++++++++++++++++++++++++++++++++++++-- src/util/cgroup.h | 2 + 2 files changed, 236 insertions(+), 9 deletions(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index a3ca76c..e606108 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -35,6 +35,7 @@ #include <sys/stat.h> #include <unistd.h> #include <mntent.h> +#include <dirent.h> /* Yes, we want linux private one, for _syscall2() macro */ #include <linux/unistd.h> @@ -1122,6 +1123,192 @@ cleanup: } +struct lxcContainerCGroup { + const char *dir; + const char *linkDest; +}; + + +static void lxcContainerCGroupFree(struct lxcContainerCGroup *mounts, + size_t nmounts) +{ + size_t i; + + if (!mounts) + return; + + for (i = 0 ; i < nmounts ; i++) { + VIR_FREE(mounts[i].dir); + VIR_FREE(mounts[i].linkDest); + } + VIR_FREE(mounts); +} + + +static int lxcContainerIdentifyCGroups(struct lxcContainerCGroup **mountsret, + size_t *nmountsret) +{ + FILE *procmnt = NULL; + struct mntent mntent; + struct dirent *dent; + char mntbuf[1024]; + int ret = -1; + struct lxcContainerCGroup *mounts = NULL; + size_t nmounts = 0; + DIR *dh = NULL; + char *path = NULL; + + *mountsret = NULL; + *nmountsret = 0; + + VIR_DEBUG("Finding cgroups mount points under %s", VIR_CGROUP_SYSFS_MOUNT); + + if (!(procmnt = setmntent("/proc/mounts", "r"))) { + virReportSystemError(errno, "%s", + _("Failed to read /proc/mounts")); + return -1; + } + + while (getmntent_r(procmnt, &mntent, mntbuf, sizeof(mntbuf)) != NULL) { + VIR_DEBUG("Got %s", mntent.mnt_dir); + if (STRNEQ(mntent.mnt_type, "cgroup") || + !STRPREFIX(mntent.mnt_dir, VIR_CGROUP_SYSFS_MOUNT)) + continue; + + /* Skip named mounts with no controller since they're + * for application use only ie systemd */ + if (strstr(mntent.mnt_opts, "name=")) + continue; + + if (VIR_EXPAND_N(mounts, nmounts, 1) < 0) { + virReportOOMError(); + goto cleanup; + } + if (!(mounts[nmounts-1].dir = strdup(mntent.mnt_dir))) { + virReportOOMError(); + goto cleanup; + } + VIR_DEBUG("Grabbed %s", mntent.mnt_dir); + } + + VIR_DEBUG("Checking for symlinks in %s", VIR_CGROUP_SYSFS_MOUNT); + if (!(dh = opendir(VIR_CGROUP_SYSFS_MOUNT))) { + virReportSystemError(errno, + _("Unable to read directory %s"), + VIR_CGROUP_SYSFS_MOUNT); + goto cleanup; + } + + while ((dent = readdir(dh)) != NULL) { + ssize_t rv; + /* The cgroups links are just relative to the local + * dir so we don't need a large buf */ + char linkbuf[100]; + + if (dent->d_name[0] == '.') + continue; + + VIR_DEBUG("Checking entry %s", dent->d_name); + if (virAsprintf(&path, "%s/%s", VIR_CGROUP_SYSFS_MOUNT, dent->d_name) < 0) { + virReportOOMError(); + goto cleanup; + } + + if ((rv = readlink(path, linkbuf, sizeof(linkbuf)-1)) < 0) { + if (errno != EINVAL) { + virReportSystemError(errno, + _("Unable to resolve link %s"), + path); + VIR_FREE(path); + goto cleanup; + } + /* Ok not a link */ + VIR_FREE(path); + } else { + linkbuf[rv] = '\0'; + VIR_DEBUG("Got a link %s to %s", path, linkbuf); + if (VIR_EXPAND_N(mounts, nmounts, 1) < 0) { + virReportOOMError(); + goto cleanup; + } + if (!(mounts[nmounts-1].linkDest = strdup(linkbuf))) { + virReportOOMError(); + goto cleanup; + } + mounts[nmounts-1].dir = path; + path = NULL; + } + } + + *mountsret = mounts; + *nmountsret = nmounts; + ret = 0; + +cleanup: + closedir(dh); + endmntent(procmnt); + VIR_FREE(path); + + if (ret < 0) + lxcContainerCGroupFree(mounts, nmounts); + return ret; +} + + +static int lxcContainerMountCGroups(struct lxcContainerCGroup *mounts, + size_t nmounts) +{ + size_t i; + + VIR_DEBUG("Mounting cgroups at '%s'", VIR_CGROUP_SYSFS_MOUNT); + + if (virFileMakePath(VIR_CGROUP_SYSFS_MOUNT) < 0) { + virReportSystemError(errno, + _("Unable to create directory %s"), + VIR_CGROUP_SYSFS_MOUNT); + return -1; + } + + if (mount("tmpfs", VIR_CGROUP_SYSFS_MOUNT, "tmpfs", MS_NOSUID|MS_NODEV|MS_NOEXEC, "mode=755") < 0) { + virReportSystemError(errno, + _("Failed to mount %s on %s type %s"), + "tmpfs", VIR_CGROUP_SYSFS_MOUNT, "tmpfs"); + return -1; + } + + for (i = 0 ; i < nmounts ; i++) { + if (mounts[i].linkDest) { + VIR_DEBUG("Link mount point '%s' to '%s'", + mounts[i].dir, mounts[i].linkDest); + if (symlink(mounts[i].linkDest, mounts[i].dir) < 0) { + virReportSystemError(errno, + _("Unable to symlink directory %s to %s"), + mounts[i].dir, mounts[i].linkDest); + return -1; + } + } else { + VIR_DEBUG("Create mount point '%s'", mounts[i].dir); + if (virFileMakePath(mounts[i].dir) < 0) { + virReportSystemError(errno, + _("Unable to create directory %s"), + mounts[i].dir); + return -1; + } + + if (mount("cgroup", mounts[i].dir, "cgroup", + 0, mounts[i].dir + strlen(VIR_CGROUP_SYSFS_MOUNT) + 1) < 0) { + virReportSystemError(errno, + _("Failed to mount %s on %s"), + "cgroup", mounts[i].dir); + return -1; + } + } + } + + return 0; +} + + /* Got a FS mapped to /, we're going the pivot_root * approach to do a better-chroot-than-chroot * this is based on this thread http://lkml.org/lkml/2008/3/5/29 @@ -1132,31 +1319,49 @@ static int lxcContainerSetupPivotRoot(virDomainDefPtr vmDef, size_t nttyPaths, virSecurityManagerPtr securityDriver) { + struct lxcContainerCGroup *mounts = NULL; + size_t nmounts = 0; + int ret = -1; + + /* Before pivoting we need to identify any + * cgroups controllers that are mounted */ + if (lxcContainerIdentifyCGroups(&mounts, &nmounts) < 0) + goto cleanup; + /* Gives us a private root, leaving all parent OS mounts on /.oldroot */ if (lxcContainerPivotRoot(root) < 0) - return -1; + goto cleanup; /* Mounts the core /proc, /sys, etc filesystems */ if (lxcContainerMountBasicFS(vmDef, true, securityDriver) < 0) - return -1; + goto cleanup; + + /* Now we can re-mount the cgroups controllers in the + * same configuration as before */ + if (lxcContainerMountCGroups(mounts, nmounts) < 0) + goto cleanup; /* Mounts /dev/pts */ if (lxcContainerMountFSDevPTS(root) < 0) - return -1; + goto cleanup; /* Populates device nodes in /dev/ */ if (lxcContainerPopulateDevices(ttyPaths, nttyPaths) < 0) - return -1; + goto cleanup; /* Sets up any non-root mounts from guest config */ if (lxcContainerMountAllFS(vmDef, "/.oldroot", true) < 0) - return -1; + goto cleanup; /* Gets rid of all remaining mounts from host OS, including /.oldroot itself */ if (lxcContainerUnmountSubtree("/.oldroot", true) < 0) - return -1; + goto cleanup; - return 0; + ret = 0; + +cleanup: + lxcContainerCGroupFree(mounts, nmounts); + return ret; } @@ -1166,6 +1371,10 @@ static int lxcContainerSetupExtraMounts(virDomainDefPtr vmDef, virDomainFSDefPtr root, virSecurityManagerPtr securityDriver) { + int ret = -1; + struct lxcContainerCGroup *mounts = NULL; + size_t nmounts = 0; + VIR_DEBUG("def=%p", vmDef); /* * This makes sure that any new filesystems in the @@ -1190,19 +1399,35 @@ static int lxcContainerSetupExtraMounts(virDomainDefPtr vmDef, if (lxcContainerMountAllFS(vmDef, "", false) < 0) return -1; + /* Before replacing /sys we need to identify any + * cgroups controllers that are mounted */ + if (lxcContainerIdentifyCGroups(&mounts, &nmounts) < 0) + goto cleanup; + /* Gets rid of any existing stuff under /proc, since we need new * namespace aware versions of those. We must do /proc second * otherwise we won't find /proc/mounts :-) */ if (lxcContainerUnmountSubtree("/sys", false) < 0 || lxcContainerUnmountSubtree("/proc", false) < 0) - return -1; + goto cleanup; /* Mounts the core /proc, /sys, etc filesystems */ if (lxcContainerMountBasicFS(vmDef, false, securityDriver) < 0) - return -1; + goto cleanup; + + /* Now we can re-mount the cgroups controllers in the + * same configuration as before */ + if (lxcContainerMountCGroups(mounts, nmounts) < 0) + goto cleanup; VIR_DEBUG("Mounting completed"); return 0; + + ret = 0; + +cleanup: + lxcContainerCGroupFree(mounts, nmounts); + return ret; } diff --git a/src/util/cgroup.h b/src/util/cgroup.h index 8486c42..857945d 100644 --- a/src/util/cgroup.h +++ b/src/util/cgroup.h @@ -16,6 +16,8 @@ struct virCgroup; typedef struct virCgroup *virCgroupPtr; +#define VIR_CGROUP_SYSFS_MOUNT "/sys/fs/cgroup" + enum { VIR_CGROUP_CONTROLLER_CPU, VIR_CGROUP_CONTROLLER_CPUACCT, -- 1.7.10.1

2 1

[libvirt] [PATCH 0/4] Improve mount setup in LXC containers
by Daniel P. Berrange 16 May '12

16 May '12

These 4 patches improve the way we setup special mounts like /sys and /proc in LXC containers. This should help ensure systemd is happier with the cgroups setup These apply ontop of the earlier series to improve SELinux driver integration

3 10

[libvirt] [libvirt-glib] Add getter for GVirConfigDomain.virt_type
by Zeeshan Ali (Khattak) 16 May '12

16 May '12

From: "Zeeshan Ali (Khattak)" <zeeshanak(a)gnome.org> --- libvirt-gconfig/libvirt-gconfig-domain.c | 12 ++++++++++++ libvirt-gconfig/libvirt-gconfig-domain.h | 1 + libvirt-gconfig/libvirt-gconfig.sym | 2 ++ 3 files changed, 15 insertions(+), 0 deletions(-) diff --git a/libvirt-gconfig/libvirt-gconfig-domain.c b/libvirt-gconfig/libvirt-gconfig-domain.c index b6b43f0..7477400 100644 --- a/libvirt-gconfig/libvirt-gconfig-domain.c +++ b/libvirt-gconfig/libvirt-gconfig-domain.c @@ -187,6 +187,18 @@ GVirConfigDomain *gvir_config_domain_new(void) return GVIR_CONFIG_DOMAIN(object); } +GVirConfigDomainVirtType gvir_config_domain_get_virt_type(GVirConfigDomain *domain) +{ + g_return_val_if_fail(GVIR_CONFIG_IS_DOMAIN(domain), + GVIR_CONFIG_DOMAIN_VIRT_QEMU); + + return gvir_config_object_get_attribute_genum + (GVIR_CONFIG_OBJECT(domain), + NULL, + "type", + GVIR_CONFIG_TYPE_DOMAIN_VIRT_TYPE, + GVIR_CONFIG_DOMAIN_VIRT_QEMU); +} void gvir_config_domain_set_virt_type(GVirConfigDomain *domain, GVirConfigDomainVirtType type) { diff --git a/libvirt-gconfig/libvirt-gconfig-domain.h b/libvirt-gconfig/libvirt-gconfig-domain.h index bdb842b..5deb58e 100644 --- a/libvirt-gconfig/libvirt-gconfig-domain.h +++ b/libvirt-gconfig/libvirt-gconfig-domain.h @@ -99,6 +99,7 @@ GType gvir_config_domain_get_type(void); GVirConfigDomain *gvir_config_domain_new_from_xml(const gchar *xml, GError **error); GVirConfigDomain *gvir_config_domain_new(void); +GVirConfigDomainVirtType gvir_config_domain_get_virt_type(GVirConfigDomain *domain); void gvir_config_domain_set_virt_type(GVirConfigDomain *domain, GVirConfigDomainVirtType type); const char *gvir_config_domain_get_name(GVirConfigDomain *domain); void gvir_config_domain_set_name(GVirConfigDomain *domain, const char *name); diff --git a/libvirt-gconfig/libvirt-gconfig.sym b/libvirt-gconfig/libvirt-gconfig.sym index 21fff46..f1af2ea 100644 --- a/libvirt-gconfig/libvirt-gconfig.sym +++ b/libvirt-gconfig/libvirt-gconfig.sym @@ -376,6 +376,8 @@ LIBVIRT_GCONFIG_0.0.8 { LIBVIRT_GCONFIG_0.0.9 { global: + gvir_config_domain_get_virt_type; + gvir_config_domain_get_os; gvir_config_domain_os_get_boot_devices; -- 1.7.7.6

3 2

[libvirt] [PATCH] Set a sensible default master start port for ehci companion controllers
by Daniel P. Berrange 16 May '12

16 May '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> The uhci1, uhci2, uhci3 companion controllers for ehci1 must have a master start port set. Since this value is predictable we should set it automatically if the app does not supply it --- src/conf/domain_conf.c | 22 ++++++++++++++++++ .../qemuxml2argv-usb-ich9-ehci-addr.xml | 24 +++++--------------- 2 files changed, 28 insertions(+), 18 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 54ac1db..f4775be 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -4066,6 +4066,28 @@ virDomainControllerDefParseXML(xmlNodePtr node, VIR_FREE(vectors); break; } + case VIR_DOMAIN_CONTROLLER_TYPE_USB: { + /* If the XML has a uhci1, uhci2, uhci3 controller and no + * master port was given, we should set a sensible one */ + int masterPort = -1; + switch (def->model) { + case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1: + masterPort = 0; + break; + case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI2: + masterPort = 2; + break; + case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI3: + masterPort = 4; + break; + } + if (masterPort != -1 && + def->info.mastertype == VIR_DOMAIN_CONTROLLER_MASTER_NONE) { + def->info.mastertype = VIR_DOMAIN_CONTROLLER_MASTER_USB; + def->info.master.usb.startport = masterPort; + } + break; + } default: break; diff --git a/tests/qemuxml2argvdata/qemuxml2argv-usb-ich9-ehci-addr.xml b/tests/qemuxml2argvdata/qemuxml2argv-usb-ich9-ehci-addr.xml index 8eff1d7..ad85d63 100644 --- a/tests/qemuxml2argvdata/qemuxml2argv-usb-ich9-ehci-addr.xml +++ b/tests/qemuxml2argvdata/qemuxml2argv-usb-ich9-ehci-addr.xml @@ -19,19 +19,13 @@ <controller type='usb' index='1' model='ich9-ehci1'> </controller> - <controller type='usb' index='0' model='ich9-uhci1'> - <master startport='0'/> - </controller> - <controller type='usb' index='1' model='ich9-uhci1'> - <master startport='0'/> - </controller> + <controller type='usb' index='0' model='ich9-uhci1'/> + <controller type='usb' index='1' model='ich9-uhci1'/> <controller type='usb' index='2' model='ich9-uhci1'> <master startport='0'/> </controller> - <controller type='usb' index='0' model='ich9-uhci3'> - <master startport='4'/> - </controller> + <controller type='usb' index='0' model='ich9-uhci3'/> <controller type='usb' index='1' model='ich9-uhci3'> <master startport='4'/> </controller> @@ -39,15 +33,9 @@ <master startport='4'/> </controller> - <controller type='usb' index='2' model='ich9-uhci2'> - <master startport='2'/> - </controller> - <controller type='usb' index='1' model='ich9-uhci2'> - <master startport='2'/> - </controller> - <controller type='usb' index='0' model='ich9-uhci2'> - <master startport='2'/> - </controller> + <controller type='usb' index='2' model='ich9-uhci2'/> + <controller type='usb' index='1' model='ich9-uhci2'/> + <controller type='usb' index='0' model='ich9-uhci2'/> <memballoon model='virtio'/> </devices> </domain> -- 1.7.10.1

3 10

[libvirt] [PATCH libvirt 0/9] domain/qemu enhancements
by Marc-André Lureau 15 May '12

15 May '12

Hi, I have sent those patches before 0.12, and updated them according to comments (hopefully you don't mind I send them all revised in the same thread) They bring: - domain audio <codec> selection, for ICH6 devices - domain port <forward>, for user-mode networking - storage <preallocation> mode for QCOW2 images Marc-André Lureau (9): qemu: test CAPS_HDA_MICRO domain: add <codec> sound sub-element qemu: honour sound <codec> sub-elements tests: add qemu-1.1-rc0 data tests: add ich6 codec type test to qemuxml2argv-sound-device storage: add preallocation element storage: learn to create qcow2 with preallocation domain: add <forward> element for user mode networking qemu: implement user mode port <forward> docs/formatdomain.html.in | 40 +++ docs/formatstorage.html.in | 6 + docs/schemas/domaincommon.rng | 58 +++- docs/schemas/storagevol.rng | 18 ++ src/conf/domain_conf.c | 286 +++++++++++++++++++- src/conf/domain_conf.h | 45 +++ src/conf/storage_conf.c | 26 ++ src/conf/storage_conf.h | 1 + src/libvirt_private.syms | 3 + src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 213 +++++++++++++-- src/storage/storage_backend.c | 70 ++++- src/util/storage_file.c | 4 + src/util/storage_file.h | 10 + tests/qemuhelpdata/qemu-1.1 | 266 ++++++++++++++++++ tests/qemuhelpdata/qemu-1.1-device | 95 +++++++ tests/qemuhelptest.c | 70 +++++ tests/qemuxml2argvdata/qemuxml2argv-net-user.args | 5 +- tests/qemuxml2argvdata/qemuxml2argv-net-user.xml | 2 + .../qemuxml2argv-sound-device.args | 7 +- .../qemuxml2argvdata/qemuxml2argv-sound-device.xml | 7 + tests/qemuxml2argvtest.c | 3 +- tests/qemuxml2xmltest.c | 1 + tests/storagevolxml2xmlin/vol-qcow2.xml | 1 + tests/storagevolxml2xmlout/vol-qcow2.xml | 1 + 26 files changed, 1201 insertions(+), 41 deletions(-) create mode 100644 tests/qemuhelpdata/qemu-1.1 create mode 100644 tests/qemuhelpdata/qemu-1.1-device -- 1.7.10.1

2 11

[libvirt] [PATCH 0/2] qemu: Add support for -no-user-config
by Jiri Denemark 15 May '12

15 May '12

Eduardo submitted patches[1] for qemu implementing -no-user-config as a better alternative to all-or-nothing -nodefconfig. With this new option, we are finally able to use modern CPU models defined in qemu's configuration file without allowing user-supplied qemu configuration to mess up with qemu processes started by libvirt. The qemu patches are not committed upstream yet (and this patchset won't be committed either until qemu support is finished) but they are very close and are expected to make it into qemu-1.1. [1] https://www.redhat.com/archives/libvir-list/2012-April/msg01293.html Jiri Denemark (2): qemu: Use common helper when probing qemu capabilities qemu: Add support for -no-user-config src/qemu/qemu_capabilities.c | 64 +++++++++++++++++++++++++++-------------- src/qemu/qemu_capabilities.h | 6 ++++ src/qemu/qemu_command.c | 11 ++++--- src/qemu/qemu_driver.c | 9 +++++- 4 files changed, 62 insertions(+), 28 deletions(-) -- 1.7.8.5

2 6

[libvirt] [PATCH libvirt 0/5] Add ich6 <codec> sub-element
by Marc-André Lureau 15 May '12

15 May '12

Hi, Upcoming qemu 1.1 will have hda-micro codec device type, which helps to solve recording with some applications under Windows 7. The following series allows to specify the codecs attached to a ICH6 audio device, using the syntax described in commit comment 6cabc0b0d0fbc35a68bf1c0646898ac148d2fa3f: <sound model='ich6'> <codec type='micro'/> </sound> The codec type 'micro' maps to qemu 'hda-micro' device kind. See also https://bugzilla.redhat.com/show_bug.cgi?id=816503 Marc-André Lureau (5): qemu: test CAPS_HDA_MICRO domain: add <codec> sound sub-element qemu: honour sound <codec> sub-elements tests: add qemu-1.1-rc0 data tests: add ich6 codec type test to qemuxml2argv-sound-device docs/formatdomain.html.in | 16 ++ docs/schemas/domaincommon.rng | 29 ++- src/conf/domain_conf.c | 119 ++++++++- src/conf/domain_conf.h | 20 ++ src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 74 ++++-- tests/qemuhelpdata/qemu-1.1 | 266 ++++++++++++++++++++ tests/qemuhelpdata/qemu-1.1-device | 95 +++++++ tests/qemuhelptest.c | 70 ++++++ .../qemuxml2argv-sound-device.args | 7 +- .../qemuxml2argvdata/qemuxml2argv-sound-device.xml | 7 + tests/qemuxml2argvtest.c | 3 +- tests/qemuxml2xmltest.c | 1 + 14 files changed, 682 insertions(+), 29 deletions(-) create mode 100644 tests/qemuhelpdata/qemu-1.1 create mode 100644 tests/qemuhelpdata/qemu-1.1-device -- 1.7.10

3 9

[libvirt] [PATCH] Remove bogus udev.target dep from libvirtd unit
by Daniel P. Berrange 15 May '12

15 May '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> There is no 'udev.target' unit in systemd (only 'udev.service') yet libvirtd's unit file had a dep on one. There's no compelling reason for a dep on udev, so remove it altogether. Reported-by: Avi Kivity <avi(a)redhat.com> --- daemon/libvirtd.service.in | 1 - 1 file changed, 1 deletion(-) diff --git a/daemon/libvirtd.service.in b/daemon/libvirtd.service.in index d6e46ab..64dbd27 100644 --- a/daemon/libvirtd.service.in +++ b/daemon/libvirtd.service.in @@ -6,7 +6,6 @@ [Unit] Description=Virtualization daemon After=syslog.target -After=udev.target After=avahi.target After=dbus.target Before=libvirt-guests.service -- 1.7.10.1

2 1

[libvirt] The config of IP address for 'macvtap' network
by Lei Li 15 May '12

15 May '12

Hi guys, I know the macvtap network is supported by libvirt as forward mode 'passthrough', I wonder is there anyway to configure the IP address for its interface? For example, If I create a network as below: <network> <name>vdsm-testnet</name> <uuid>31f6b3b3-e959-0dd1-ad3a-bf95db660415</uuid> <forward dev='eth0.8' mode='passthrough'> <interface dev='eth0.8'/> </forward> </network> For now, I have to set the IP address by 'ifconfig eth0.8 XXX.XXX.XXX.XXX' after defining the network. How can I set IP address for this VLAN device 'eth0.8' by libvirt, I mean does libvirt support to assign IP address in this mode now? If does, what is the xml format for it? It'd appreciate a lot if anybody could read my post and give me some suggestions! -- Lei

3 5