- Devel - libvirt List Archives

[libvirt] [PATCH] storage: handle NULL return from virGetStorageVol
by Michael Chapman 12 Feb '14

12 Feb '14

virGetStorageVol can return NULL on out-of-memory. If it does, cleanly abort the volume clone operation. Signed-off-by: Michael Chapman <mike(a)very.puzzling.org> --- src/storage/storage_driver.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index c83aa8a..2f7b2e5 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -1828,6 +1828,10 @@ storageVolCreateXMLFrom(virStoragePoolPtr obj, pool->volumes.objs[pool->volumes.count++] = newvol; volobj = virGetStorageVol(obj->conn, pool->def->name, newvol->name, newvol->key, NULL, NULL); + if (!volobj) { + pool->volumes.count--; + goto cleanup; + } /* Drop the pool lock during volume allocation */ pool->asyncjobs++; -- 1.8.3.1

2 1

[libvirt] [PATCH 0/2] Partially revert separation of volume creating and building
by Ján Tomko 12 Feb '14

12 Feb '14

The storage driver expects 'createVol' to fill out the volume key. For disk and logical backends, we generate the key only after the volume has been built. Revert the separation commits for these backends. Ján Tomko (2): Revert "storage: lvm: Separate creating of the volume from building" Revert "storage: disk: Separate creating of the volume from building" src/storage/storage_backend_disk.c | 44 ++++++++----------------- src/storage/storage_backend_logical.c | 60 ++++++++++++++--------------------- 2 files changed, 37 insertions(+), 67 deletions(-) -- 1.8.3.2

2 4

[libvirt] [PATCH] Fix leaks in vircapstest
by Ján Tomko 12 Feb '14

12 Feb '14

Coverity complains about cell_cpus being leaked on error and valgrind shows 'caps' is leaked on success. Introduced in eb64e87. --- tests/vircapstest.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/tests/vircapstest.c b/tests/vircapstest.c index a40771d..4264e9e 100644 --- a/tests/vircapstest.c +++ b/tests/vircapstest.c @@ -40,7 +40,7 @@ static virCapsPtr buildNUMATopology(int seq) { virCapsPtr caps; - virCapsHostNUMACellCPUPtr cell_cpus; + virCapsHostNUMACellCPUPtr cell_cpus = NULL; int core_id, cell_id; int id; @@ -75,6 +75,8 @@ buildNUMATopology(int seq) return caps; error: + virCapabilitiesClearHostNUMACellCPUTopology(cell_cpus, MAX_CPUS_IN_CELL); + VIR_FREE(cell_cpus); virObjectUnref(caps); return NULL; @@ -87,7 +89,7 @@ test_virCapabilitiesGetCpusForNodemask(const void *data ATTRIBUTE_UNUSED) const char *nodestr = "3,4,5,6"; virBitmapPtr nodemask = NULL; virBitmapPtr cpumap = NULL; - virCapsPtr caps; + virCapsPtr caps = NULL; int mask_size = 8; int ret = -1; @@ -107,6 +109,7 @@ test_virCapabilitiesGetCpusForNodemask(const void *data ATTRIBUTE_UNUSED) ret = 0; error: + virObjectUnref(caps); virBitmapFree(nodemask); virBitmapFree(cpumap); return ret; -- 1.8.3.2

2 2

[libvirt] [PATCH] AppArmor: Fix the place where the template should be installed
by Cédric Bosdonnat 12 Feb '14

12 Feb '14

The security driver expects /etc/apparmor.d/libvirt/TEMPLATE but we installed it to /etc/apparmor.d/libvirtd/TEMPLATE. Move the template to the expected place since that code was here long before. --- examples/apparmor/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/apparmor/Makefile.am b/examples/apparmor/Makefile.am index 6e69440..2630fef 100644 --- a/examples/apparmor/Makefile.am +++ b/examples/apparmor/Makefile.am @@ -32,7 +32,7 @@ abstractions_DATA = \ libvirt-qemu \ $(NULL) -templatesdir = $(apparmordir)/libvirtd +templatesdir = $(apparmordir)/libvirt templates_DATA = \ TEMPLATE \ $(NULL) -- 1.8.5.2

2 1

[libvirt] [PATCH v2 0/2] Handle non-sequential NUMA node numbers
by Pradipta Kr. Banerjee 12 Feb '14

12 Feb '14

v2 *Add test case as suggested by Daniel *Minor change in comments Pradipta Kr. Banerjee (2): Handle non-sequential NUMA node numbers vircapstest: Introduce virCapabilitiesGetCpusForNodemask test src/conf/capabilities.c | 12 ++++- src/qemu/qemu_driver.c | 5 +- src/qemu/qemu_process.c | 2 +- tests/Makefile.am | 5 ++ tests/vircapstest.c | 129 ++++++++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 148 insertions(+), 5 deletions(-) create mode 100644 tests/vircapstest.c -- 1.8.5.3

4 6

[libvirt] [PATCH 0/8] storage: separate storage volume creation from building
by Peter Krempa 12 Feb '14

12 Feb '14

This series splits the internal storage volume metadata creation from actual volume building and adds support for deletion of guster volumes. Peter Krempa (8): storage: fs: Fix comment for virStorageBackendFileSystemDelete storage: Support deletion of volumes on gluster pools storage: lvm: Avoid forward decl of virStorageBackendLogicalDeleteVol storage: lvm: Separate creating of the volume from building storage: disk: Separate creating of the volume from building storage: RBD: Separate creating of the volume from building storage: Sheepdog: Separate creating of the volume from building storage: Improve error message when a storage backend is missing src/storage/storage_backend.c | 3 +- src/storage/storage_backend_disk.c | 44 ++++++++---- src/storage/storage_backend_fs.c | 6 +- src/storage/storage_backend_gluster.c | 64 +++++++++++++++++ src/storage/storage_backend_logical.c | 127 ++++++++++++++++++--------------- src/storage/storage_backend_rbd.c | 41 +++++++++-- src/storage/storage_backend_sheepdog.c | 29 +++++++- 7 files changed, 226 insertions(+), 88 deletions(-) -- 1.8.5.2

4 23

[libvirt] [PATCH 00/11] libxl: add job support
by Jim Fehlig 12 Feb '14

12 Feb '14

This patch series adds job support to the libxl driver, using techiques from the qemu driver. One benefit is no longer blocking get operations during long running modify operations. E.g. with these patches 'vish dominfo dom' will work while 'virsh save dom ...' is in progress. The first patch adds the job support machinery, followed by several patches that make use of it. I initially had all but the first in a single "use-job-support" patch, but hope breaking it out eases review. Jim Fehlig (11): libxl: Add job support to libxl driver libxl: use job functions in libxlVmStart libxl: use job functions in libxlDomainSetMemoryFlags libxl: use job functions in libxlDomain{Suspend,Resume} libxl: use job functions in libxlDomainDestroyFlags libxl: use job functions in domain save operations libxl: use job functions in libxlDomainCoreDump libxl: use job functions in vcpu set and pin functions libxl: use job functions in device attach and detach functions libxl: use job functions in libxlDomainSetAutostart libxl: use job functions in libxlDomainSetSchedulerParametersFlags src/libxl/libxl_domain.c | 128 ++++++++++++++++++ src/libxl/libxl_domain.h | 37 +++++ src/libxl/libxl_driver.c | 346 +++++++++++++++++++++++++++++++---------------- 3 files changed, 395 insertions(+), 116 deletions(-) -- 1.8.1.4

3 37

[libvirt] [python PATCH] examples: demonstrate network events
by Eric Blake 12 Feb '14

12 Feb '14

Commit 6ea5be0 added network event callback support, so we might as well demonstrate that it works by updating our example. * examples/event-test.py: Add network event, fix typos. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- examples/event-test.py | 70 ++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 51 insertions(+), 19 deletions(-) diff --git a/examples/event-test.py b/examples/event-test.py index 1402c04..101dbc0 100644 --- a/examples/event-test.py +++ b/examples/event-test.py @@ -2,6 +2,6 @@ # # # -################################################################################# -# Start off by implementing a general purpose event loop for anyones use -################################################################################# +############################################################################## +# Start off by implementing a general purpose event loop for anyone's use +############################################################################## import sys import getopt @@ -158,7 +158,7 @@ class virEventLoopPure: # This is the heart of the event loop, performing one single # iteration. It asks when the next timeout is due, and then - # calcuates the maximum amount of time it is able to sleep + # calculates the maximum amount of time it is able to sleep # for in poll() pending file handle events. # # It then goes into the poll() sleep. @@ -167,9 +167,9 @@ class virEventLoopPure: # events which need to be dispatched to registered callbacks # It may also be time to fire some periodic timers. # - # Due to the coarse granularity of schedular timeslices, if + # Due to the coarse granularity of scheduler timeslices, if # we ask for a sleep of 500ms in order to satisfy a timer, we - # may return up to 1 schedular timeslice early. So even though + # may return up to 1 scheduler timeslice early. So even though # our sleep timeout was reached, the registered timer may not # technically be at its expiry point. This leads to us going # back around the loop with a crazy 5ms sleep. So when checking @@ -227,7 +227,7 @@ class virEventLoopPure: self.runningPoll = False - # Actually the event loop forever + # Actually run the event loop forever def run_loop(self): self.quit = False while not self.quit: @@ -429,8 +429,8 @@ def virEventLoopNativeStart(): ########################################################################## # Everything that now follows is a simple demo of domain lifecycle events ########################################################################## -def eventToString(event): - eventStrings = ( "Defined", +def domEventToString(event): + domEventStrings = ( "Defined", "Undefined", "Started", "Suspended", @@ -438,11 +438,12 @@ def eventToString(event): "Stopped", "Shutdown", "PMSuspended", - "Crashed" ) - return eventStrings[event] + "Crashed", + ) + return domEventStrings[event] -def detailToString(event, detail): - eventStrings = ( +def domDetailToString(event, detail): + domEventStrings = ( ( "Added", "Updated" ), ( "Removed", ), ( "Booted", "Migrated", "Restored", "Snapshot", "Wakeup" ), @@ -451,19 +452,19 @@ def detailToString(event, detail): ( "Shutdown", "Destroyed", "Crashed", "Migrated", "Saved", "Failed", "Snapshot"), ( "Finished", ), ( "Memory", "Disk" ), - ( "Panicked", ) + ( "Panicked", ), ) - return eventStrings[event][detail] + return domEventStrings[event][detail] def myDomainEventCallback1 (conn, dom, event, detail, opaque): print("myDomainEventCallback1 EVENT: Domain %s(%s) %s %s" % (dom.name(), dom.ID(), - eventToString(event), - detailToString(event, detail))) + domEventToString(event), + domDetailToString(event, detail))) def myDomainEventCallback2 (conn, dom, event, detail, opaque): print("myDomainEventCallback2 EVENT: Domain %s(%s) %s %s" % (dom.name(), dom.ID(), - eventToString(event), - detailToString(event, detail))) + domEventToString(event), + domDetailToString(event, detail))) def myDomainEventRebootCallback(conn, dom, opaque): print("myDomainEventRebootCallback: Domain %s(%s)" % (dom.name(), dom.ID())) @@ -501,6 +502,35 @@ def myDomainEventDeviceRemovedCallback(conn, dom, dev, opaque): print("myDomainEventDeviceRemovedCallback: Domain %s(%s) device removed: %s" % ( dom.name(), dom.ID(), dev)) +########################################################################## +# Network events +########################################################################## +def netEventToString(event): + netEventStrings = ( "Defined", + "Undefined", + "Started", + "Stopped", + ) + return netEventStrings[event] + +def netDetailToString(event, detail): + netEventStrings = ( + ( "Added", ), + ( "Removed", ), + ( "Started", ), + ( "Stopped", ), + ) + return netEventStrings[event][detail] + +def myNetworkEventLifecycleCallback(conn, net, event, detail, opaque): + print("myNetworkEventLifecycleCallback: Network %s %s %s" % (net.name(), + netEventToString(event), + netDetailToString(event, detail))) + +########################################################################## +# Set up and run the program +########################################################################## + run = True def myConnectionCloseCallback(conn, reason, opaque): @@ -577,6 +607,8 @@ def main(): vc.domainEventRegisterAny(None, libvirt.VIR_DOMAIN_EVENT_ID_PMSUSPEND_DISK, myDomainEventPMSuspendDiskCallback, None) vc.domainEventRegisterAny(None, libvirt.VIR_DOMAIN_EVENT_ID_DEVICE_REMOVED, myDomainEventDeviceRemovedCallback, None) + vc.networkEventRegisterAny(None, libvirt.VIR_NETWORK_EVENT_ID_LIFECYCLE, myNetworkEventLifecycleCallback, None) + vc.setKeepAlive(5, 3) # The rest of your app would go here normally, but for sake -- 1.8.5.3

2 2

[libvirt] Make systemd work with LXC user namespaces
by Richard Weinberger 11 Feb '14

11 Feb '14

These two patches fix the issue that control groups are unusable if user namespaces are enabled. We have to chown() the control group to the correct user. As the container mounts the control group and only the controller is allowed to chown() the mount point we need a new barrier to synchronize them after the container has setup the control groups. Thanks, //richard [PATCH 1/2] lxc: Add another barrier [PATCH 2/2] lxc: Add virCgroupSetOwner()

2 4

[libvirt] [PATCH] qemu: Implement a stub cpuArchDriver.baseline() handler for aarch64
by Oleg Strikov 11 Feb '14

11 Feb '14

Openstack Nova calls virConnectBaselineCPU() during initialization of the instance to get a full list of CPU features. This patch adds a stub to aarch64-specific code to handle this request (no actual work is done). That's enough to have this stub with limited functionality because qemu/kvm backend supports only 'host-passthrough' cpu mode on aarch64. Signed-off-by: Oleg Strikov <oleg.strikov(a)canonical.com> --- src/cpu/cpu_aarch64.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) diff --git a/src/cpu/cpu_aarch64.c b/src/cpu/cpu_aarch64.c index f674bff..3c3e749 100644 --- a/src/cpu/cpu_aarch64.c +++ b/src/cpu/cpu_aarch64.c @@ -85,6 +85,29 @@ AArch64GuestData(virCPUDefPtr host ATTRIBUTE_UNUSED, return VIR_CPU_COMPARE_IDENTICAL; } +static virCPUDefPtr +AArch64Baseline(virCPUDefPtr *cpus, + unsigned int ncpus ATTRIBUTE_UNUSED, + const char **models ATTRIBUTE_UNUSED, + unsigned int nmodels ATTRIBUTE_UNUSED, + unsigned int flags) +{ + virCPUDefPtr cpu = NULL; + + virCheckFlags(VIR_CONNECT_BASELINE_CPU_EXPAND_FEATURES, NULL); + + if (VIR_ALLOC(cpu) < 0 || + VIR_STRDUP(cpu->model, cpus[0]->model) < 0) { + virCPUDefFree(cpu); + return NULL; + } + + cpu->type = VIR_CPU_TYPE_GUEST; + cpu->match = VIR_CPU_MATCH_EXACT; + + return cpu; +} + struct cpuArchDriver cpuDriverAARCH64 = { .name = "aarch64", .arch = archs, @@ -95,7 +118,7 @@ struct cpuArchDriver cpuDriverAARCH64 = { .free = AArch64DataFree, .nodeData = AArch64NodeData, .guestData = AArch64GuestData, - .baseline = NULL, + .baseline = AArch64Baseline, .update = AArch64Update, .hasFeature = NULL, }; -- 1.7.9.5

2 1