December 2015 - Devel - libvirt List Archives

[libvirt] [PATCH 0/6] Add multiqueue support for macvtaps
by Michal Privoznik 08 Dec '15

08 Dec '15

Some patches are easier to review when showed with --ignore-space-change. Michal Privoznik (6): virNetDevMacVLanCreateWithVPortProfile: Turn vnet_hdr into flag virNetDevMacVLanTapOpen: Slightly rework virNetDevMacVLanTapOpen: Rework to support multiple FDs virNetDevMacVLanTapSetup: Rework to support multiple FDs virNetDevMacVLanCreateWithVPortProfile: Rework to support multiple FDs qemu: Enable multiqueue for macvtaps src/lxc/lxc_process.c | 3 +- src/qemu/qemu_command.c | 61 +++++++++------- src/qemu/qemu_command.h | 2 + src/qemu/qemu_hotplug.c | 16 +++-- src/util/virnetdevmacvlan.c | 171 ++++++++++++++++++++++++-------------------- src/util/virnetdevmacvlan.h | 13 ++-- 6 files changed, 150 insertions(+), 116 deletions(-) -- 2.4.10

6 21

[libvirt] [PATCH] rpm: explicitly enable & start virlockd/virtlogd sockets on install
by Daniel P. Berrange 08 Dec '15

08 Dec '15

When installing the libvirt-daemon RPM, we have a %post rule to enable the libvirtd.service, virtlockd.socket and virtlogd.socket files. This is only done, however, when the RPM is first installed, not when upgrading RPMs. This adds a separate rule to explicitly enable the virtlockd.socket and virtlogd.socket units upon RPM upgrade, if libvirtd.service is marked as enabled. It also adds a rule to start the socket units if libvirtd.service is running. The downside with taking this approach is that it is impossible for a user to have libvirtd.service enabled and disable socket activation, because we'll re-enable it on every upgrade. Similarly we'll be starting the socket listeners on every upgrade if libvirtd is running. This is a pretty nasty thing to have to keep in libvirt for the long term, given this upgrade problem is a one-time only pain point. For this reason I am recommending *NOT* applying this patch to GIT. Instead, I think we just have to release note the fact that after upgrading you should enable the virtlockd.socket unit. --- libvirt.spec.in | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 1f76f24..48e5632 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1703,6 +1703,18 @@ if [ $1 -eq 1 ] ; then libvirtd.service >/dev/null 2>&1 || : fi %endif + +# In upgrade scenario we must explicitly enable virtlockd/virtlogd +# sockets, if libvirtd is already enabled and start them if +# libvirtd is running, otherwise you'll get failures to start +# guests +if [ $1 -ge 1 ] ; then + /bin/systemctl is-enabled libvirtd.service 1>/dev/null 2>&1 && + /bin/systemctl enable virtlockd.socket virtlogd.socket + /bin/systemctl is-active libvirtd.service 1>/dev/null 2>&1 && + /bin/systemctl start virtlockd.socket virtlogd.socket +fi + %else %if %{with_cgconfig} # Starting with Fedora 16/RHEL-7, systemd automounts all cgroups, -- 2.5.0

2 5

[libvirt] Sparse image volDownload
by Michal Privoznik 08 Dec '15

08 Dec '15

Dear list, I'd like to hear your opinion on the following bug: https://bugzilla.redhat.com/show_bug.cgi?id=1282859 Long story short, Imagine the following scenario: 1. Create 4GB file full of zeroes 2. virsh vol-download it What happens is that all those 4GB are transferred byte after byte through our RPC system. Not only this puts needles pressure on our event loop, it's suboptimal for network and other resources too. I'd like to explore our options here keeping in mind that the original volume might have been sparse and we ought to keep it sparse on the destination too. In the bug the reporter (Matthew Booth) suggests introducing new type of RPC message that will let us keep our APIs unchanged. The source will scan the file for windows of zeroes bigger than some value. When found the new type of message is passed to client without need to copy those zeroes. Yes, this is very similar to RLE. If we are going that way, should we enable users to put a compression program in between read()/write() and our RPC? Well, should we let users to choose what compression program we will put there? Because there are better compression algorithms than RLE. Michal

3 6

[libvirt] Entering freeze for libvirt-1.3.0
by Daniel Veillard 07 Dec '15

07 Dec '15

As planned I have tagged RC1 in git and pushed signed tarball and rpms to the usual place: ftp://libvirt.org/libvirt/ there is significant changes in this release so people are really encouraged to give it a good try. I will try to push rc2 possibly in the week-end and shooting for a release on Tuesday next week i.e. Dec 8th, thanks for testing it ! Daniel -- Daniel Veillard | Open Source and Standards, Red Hat veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

4 8

[libvirt] [PATCH] rpm: explicitly enable & start virtlogd on install
by Daniel P. Berrange 07 Dec '15

07 Dec '15

When installing the libvirt-daemon RPM, we have a %post rule to enable the libvirtd.service, virtlockd.socket and virtlogd.socket files. This is only done, however, when the RPM is first installed, not when upgrading RPMs. So virtlogd will not get activated on upgrading, which is a problem as libvirt qemu driver will expect it to be available by default. This adds a trigger that is run when uninstalling libvirt-daemon older than 1.3.0 that will enable & start virtlogd.socket if libvirtd is enabled and/or started. Using the trigger rather than %post ensures that it only runs once, allowing admins to disable it explicitly thereafter without future upgrades re-enabling it. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- libvirt.spec.in | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 1f76f24..ff25881 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1779,6 +1779,25 @@ if [ "$1" -ge "1" ]; then fi %endif +# In upgrade scenario we must explicitly enable virtlockd/virtlogd +# sockets, if libvirtd is already enabled and start them if +# libvirtd is running, otherwise you'll get failures to start +# guests +%triggerpostun daemon -- libvirt-daemon < 1.3.0 +if [ $1 -ge 1 ] ; then + %if %{with_systemd} + /bin/systemctl is-enabled libvirtd.service 1>/dev/null 2>&1 && + /bin/systemctl enable virtlogd.socket || : + /bin/systemctl is-active libvirtd.service 1>/dev/null 2>&1 && + /bin/systemctl start virtlogd.socket || : + %else + /sbin/chkconfig libvirtd 1>/dev/null 2>&1 && + /bin/chkconfig virtlogd on || : + /sbin/service libvirtd status 1>/dev/null 2>&1 && + /bin/service virtlogd start || : + %endif +fi + %if %{with_network} %post daemon-config-network if test $1 -eq 1 && test ! -f %{_sysconfdir}/libvirt/qemu/networks/default.xml ; then -- 2.5.0

1 0

[libvirt] libvirt-php 0.5.1 uneeded files
by Remi Collet 07 Dec '15

07 Dec '15

Installing the new 0.5.1 version: libvirt-php.la libvirt-php.so -> libvirt-php.so.0.0.0 libvirt-php.so.0 -> libvirt-php.so.0.0.0 libvirt-php.so.0.0.0 This is not a versioned library but a PHP extension so having those have no sense. Only libvirt-php.so is needed. Notice: I think it will be simpler to switch to the standard PHP extension build system (phpize; configure; make) Especially, this will allow to use the --with-php-config option (an thus, build for not standard PHP, e.g. TS build or SCL build) Remi.

3 8

[libvirt] [PATCH] libvirtd: enable virtlockd/virtlogd socket activation on install
by Daniel P. Berrange 07 Dec '15

07 Dec '15

When someone does 'systemctl enable libvirtd.service' we should also enable virtlockd.socket/virtlogd.socket, so that they can be auto-activated if libvirtd tries to access the sockets. Without this, people have to manually enable the units themselves via 'systemctl enable virtdlogd.socket'. This also ensures that if distros uses 'systemctl preset' for enabling 'libvirtd.service', then the virtdlogd.socket gets enabled without having to wait for the distro to update their presets file. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- daemon/libvirtd.service.in | 2 ++ 1 file changed, 2 insertions(+) diff --git a/daemon/libvirtd.service.in b/daemon/libvirtd.service.in index 9e67e43..608221c 100644 --- a/daemon/libvirtd.service.in +++ b/daemon/libvirtd.service.in @@ -22,3 +22,5 @@ Restart=on-failure [Install] WantedBy=multi-user.target +Also=virtlockd.socket +Also=virtlogd.socket -- 2.5.0

2 1

[libvirt] [libvirt-glib][PATCH] Drop unused finalize function
by Michal Privoznik 07 Dec '15

07 Dec '15

After 1225a160ac3b4 where I'm dropping some unused variables, the functions I was touching became sort of redundant too. Specifically, gvir_domain_disk_finalize() and gvir_domain_interface_finalize(). Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- libvirt-gobject/libvirt-gobject-domain-disk.c | 8 -------- libvirt-gobject/libvirt-gobject-domain-interface.c | 8 -------- 2 files changed, 16 deletions(-) diff --git a/libvirt-gobject/libvirt-gobject-domain-disk.c b/libvirt-gobject/libvirt-gobject-domain-disk.c index 7f34ebd..bb2a682 100644 --- a/libvirt-gobject/libvirt-gobject-domain-disk.c +++ b/libvirt-gobject/libvirt-gobject-domain-disk.c @@ -49,16 +49,8 @@ gvir_domain_disk_error_quark(void) return g_quark_from_static_string("gvir-domain-disk"); } -static void gvir_domain_disk_finalize(GObject *object) -{ - G_OBJECT_CLASS(gvir_domain_disk_parent_class)->finalize(object); -} - static void gvir_domain_disk_class_init(GVirDomainDiskClass *klass) { - GObjectClass *object_class = G_OBJECT_CLASS (klass); - - object_class->finalize = gvir_domain_disk_finalize; g_type_class_add_private(klass, sizeof(GVirDomainDiskPrivate)); } diff --git a/libvirt-gobject/libvirt-gobject-domain-interface.c b/libvirt-gobject/libvirt-gobject-domain-interface.c index 443a2b6..beae11e 100644 --- a/libvirt-gobject/libvirt-gobject-domain-interface.c +++ b/libvirt-gobject/libvirt-gobject-domain-interface.c @@ -49,16 +49,8 @@ gvir_domain_interface_error_quark(void) return g_quark_from_static_string("gvir-domain-interface"); } -static void gvir_domain_interface_finalize(GObject *object) -{ - G_OBJECT_CLASS(gvir_domain_interface_parent_class)->finalize(object); -} - static void gvir_domain_interface_class_init(GVirDomainInterfaceClass *klass) { - GObjectClass *object_class = G_OBJECT_CLASS (klass); - - object_class->finalize = gvir_domain_interface_finalize; g_type_class_add_private(klass, sizeof(GVirDomainInterfacePrivate)); } -- 2.4.10

2 1

[libvirt] [libvirt-glib][PATCH] gobject: Drop some unused variables
by Michal Privoznik 07 Dec '15

07 Dec '15

After abe983ddf8a569 there are some variables, that are unused: CC libvirt_gobject_1_0_la-libvirt-gobject-domain-disk.lo CC libvirt_gobject_1_0_la-libvirt-gobject-domain-interface.lo CC libvirt_gobject_1_0_la-libvirt-gobject-domain.lo CC libvirt_gobject_1_0_la-libvirt-gobject-interface.lo ../../libvirt-gobject/libvirt-gobject-domain-interface.c: In function 'gvir_domain_interface_finalize': ../../libvirt-gobject/libvirt-gobject-domain-interface.c:54:26: error: unused variable 'self' [-Werror=unused-variable] GVirDomainInterface *self = GVIR_DOMAIN_INTERFACE(object); ^ cc1: all warnings being treated as errors Makefile:724: recipe for target 'libvirt_gobject_1_0_la-libvirt-gobject-domain-interface.lo' failed make[3]: *** [libvirt_gobject_1_0_la-libvirt-gobject-domain-interface.lo] Error 1 make[3]: *** Waiting for unfinished jobs.... ../../libvirt-gobject/libvirt-gobject-domain-disk.c: In function 'gvir_domain_disk_finalize': ../../libvirt-gobject/libvirt-gobject-domain-disk.c:54:21: error: unused variable 'self' [-Werror=unused-variable] GVirDomainDisk *self = GVIR_DOMAIN_DISK(object); ^ cc1: all warnings being treated as errors Makefile:717: recipe for target 'libvirt_gobject_1_0_la-libvirt-gobject-domain-disk.lo' failed Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Pushed under trivial rule. libvirt-gobject/libvirt-gobject-domain-disk.c | 2 -- libvirt-gobject/libvirt-gobject-domain-interface.c | 2 -- 2 files changed, 4 deletions(-) diff --git a/libvirt-gobject/libvirt-gobject-domain-disk.c b/libvirt-gobject/libvirt-gobject-domain-disk.c index 770acff..7f34ebd 100644 --- a/libvirt-gobject/libvirt-gobject-domain-disk.c +++ b/libvirt-gobject/libvirt-gobject-domain-disk.c @@ -51,8 +51,6 @@ gvir_domain_disk_error_quark(void) static void gvir_domain_disk_finalize(GObject *object) { - GVirDomainDisk *self = GVIR_DOMAIN_DISK(object); - G_OBJECT_CLASS(gvir_domain_disk_parent_class)->finalize(object); } diff --git a/libvirt-gobject/libvirt-gobject-domain-interface.c b/libvirt-gobject/libvirt-gobject-domain-interface.c index 57896fd..443a2b6 100644 --- a/libvirt-gobject/libvirt-gobject-domain-interface.c +++ b/libvirt-gobject/libvirt-gobject-domain-interface.c @@ -51,8 +51,6 @@ gvir_domain_interface_error_quark(void) static void gvir_domain_interface_finalize(GObject *object) { - GVirDomainInterface *self = GVIR_DOMAIN_INTERFACE(object); - G_OBJECT_CLASS(gvir_domain_interface_parent_class)->finalize(object); } -- 2.4.10

2 1

[libvirt] [PATCH v2] libxl: free ifname on libxlDomainMigrationPrepareDef
by Joao Martins 06 Dec '15

06 Dec '15

Commit d2e5538b1 changes virDomainDef to include ifnames that autogenerated by libxl, and that are also cleared on domain cleanup. One place that's missing is on migration, when domain xml is sent to dst libvirtd and would contain ifnames from the source libvirtd. This would lead to erronous behaviour (as seen in osstest CI) such as failing to migrate when a vif with the same name existed (belonging to another domain) on destination. This patch adds a helper libxlDomainFreeIfaceNames for clearing ifnames on both migration prepare phase and domain cleanup. It is done on prepare phase so that we don't affect source domain (specially in cases when the migration fails), while still allowing interface stats to be consulted during the migration. Signed-off-by: Joao Martins <joao.m.martins(a)oracle.com> --- Changes since v1: - Commit message is changed - Clear ifnames on prepare phase as opposed to begin phase. --- src/libxl/libxl_domain.c | 32 ++++++++++++++++++++++---------- src/libxl/libxl_domain.h | 3 +++ src/libxl/libxl_migration.c | 2 ++ 3 files changed, 27 insertions(+), 10 deletions(-) diff --git a/src/libxl/libxl_domain.c b/src/libxl/libxl_domain.c index ef92974..487589d 100644 --- a/src/libxl/libxl_domain.c +++ b/src/libxl/libxl_domain.c @@ -728,16 +728,7 @@ libxlDomainCleanup(libxlDriverPrivatePtr driver, } } - if ((vm->def->nnets)) { - size_t i; - - for (i = 0; i < vm->def->nnets; i++) { - virDomainNetDefPtr net = vm->def->nets[i]; - - if (STRPREFIX(net->ifname, "vif")) - VIR_FREE(net->ifname); - } - } + libxlDomainFreeIfaceNames(vm->def); if (virAsprintf(&file, "%s/%s.xml", cfg->stateDir, vm->def->name) > 0) { if (unlink(file) < 0 && errno != ENOENT && errno != ENOTDIR) @@ -923,6 +914,27 @@ libxlDomainCreateIfaceNames(virDomainDefPtr def, libxl_domain_config *d_config) } } +/* + * Removes autogenerated interface names for the network devices in + * parameter def. User-provided interface names are skipped. + */ +void +libxlDomainFreeIfaceNames(virDomainDefPtr def) +{ + size_t i; + + for (i = 0; i < def->nnets; i++) { + virDomainNetDefPtr net = def->nets[i]; + + if (!net->ifname) + continue; + + if (STRPREFIX(net->ifname, "vif")) + VIR_FREE(net->ifname); + } +} + + /* * Start a domain through libxenlight. diff --git a/src/libxl/libxl_domain.h b/src/libxl/libxl_domain.h index 44b3e0b..70c139c 100644 --- a/src/libxl/libxl_domain.h +++ b/src/libxl/libxl_domain.h @@ -135,6 +135,9 @@ int libxlDomainSetVcpuAffinities(libxlDriverPrivatePtr driver, virDomainObjPtr vm); +void +libxlDomainFreeIfaceNames(virDomainDefPtr def); + int libxlDomainStart(libxlDriverPrivatePtr driver, virDomainObjPtr vm, diff --git a/src/libxl/libxl_migration.c b/src/libxl/libxl_migration.c index 0d23e5f..1ca31ab 100644 --- a/src/libxl/libxl_migration.c +++ b/src/libxl/libxl_migration.c @@ -288,6 +288,8 @@ libxlDomainMigrationPrepareDef(libxlDriverPrivatePtr driver, VIR_DOMAIN_DEF_PARSE_INACTIVE))) goto cleanup; + libxlDomainFreeIfaceNames(def); + if (dname) { name = def->name; if (VIR_STRDUP(def->name, dname) < 0) { -- 2.1.4

2 4