- Devel - libvirt List Archives

[libvirt] [PATCH 1/2] qemuDomainDetachChrDevice: Don't leak @charAlias
by Michal Privoznik 18 Jul '13

18 Jul '13

Moreover, since virAsprintf now does report OOM error, there's no need to call virReportOOMError in error path. --- src/qemu/qemu_hotplug.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 788ad47..a98de9d 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -3155,10 +3155,8 @@ int qemuDomainDetachChrDevice(virQEMUDriverPtr driver, if (qemuBuildChrDeviceStr(&devstr, vm->def, chr, priv->qemuCaps) < 0) return ret; - if (virAsprintf(&charAlias, "char%s", tmpChr->info.alias) < 0) { - virReportOOMError(); - return ret; - } + if (virAsprintf(&charAlias, "char%s", tmpChr->info.alias) < 0) + goto cleanup; qemuDomainObjEnterMonitor(driver, vm); if (devstr && qemuMonitorDelDevice(priv->mon, tmpChr->info.alias) < 0) { -- 1.8.1.5

2 3

[libvirt] [PATCH 0/6] Enable fd passing / socket activation with LXC guest
by Daniel P. Berrange 18 Jul '13

18 Jul '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> Systemd has a concept of socket activation whereby systemd will listen on a TCP socket in the host. When a client arrives on the socket, systemd will run a service, passing it the pre-opened TCP server socket. The service can then accept the client connection. This patch series adds the ability to pass pre-opened file descriptors into LXC guests. The file descriptors will be made available to the 'init' process in the container, starting from STDERR_FILENO + 1. For example, assuming you have pre-opened a file descriptors in your shell # exec 10>/tmp/foo # exec 20>/tmp/bar # exec 30>/tmp/wizz You can then start a container with: # virsh -c lxc:/// start --pass-fds 10,20,30 demo Inside that container the FDs will appear as 3, 4, 5: # virsh -c lxc:/// console demo Connected to domain demo Escape character is ^] sh-4.2# lsof -p 1 | grep /tmp sh 1 root 3w REG 0,32 0 90226444 /tmp/foo sh 1 root 4w REG 0,32 0 90238163 /tmp/bar sh 1 root 5w REG 0,32 0 90238164 /tmp/wizz Finally, if you run systemd inside the container, it can then use these pre-opened file descriptors, passing them along when launching services inside the container. So you have end-to-end socket activation between the host & guest systemd instances. Daniel P. Berrange (6): Introduce new domain create APIs to pass pre-opened FDs to LXC Introduce remote protocol support for virDomainCreate{XML}WithFiles Fix impl of virDomainCreateWithFlags remote client helper LXC: Wire up the virDomainCreate{XML}WithFiles methods Enable FD passing when starting guests with virsh Merge virCommandPreserveFD / virCommandTransferFD daemon/remote.c | 104 ++++++++++++++++++++++ include/libvirt/libvirt.h.in | 10 +++ python/generator.py | 3 + python/libvirt-override-virConnect.py | 30 +++++++ python/libvirt-override-virDomain.py | 38 ++++++++ python/libvirt-override.c | 89 +++++++++++++++++++ src/driver.h | 13 +++ src/fdstream.c | 3 +- src/libvirt.c | 154 ++++++++++++++++++++++++++++++++ src/libvirt_private.syms | 3 +- src/libvirt_public.syms | 6 ++ src/lxc/lxc_container.c | 136 ++++++++++++++++++++++------- src/lxc/lxc_container.h | 6 +- src/lxc/lxc_controller.c | 36 +++++++- src/lxc/lxc_driver.c | 45 ++++++++-- src/lxc/lxc_process.c | 20 ++++- src/lxc/lxc_process.h | 1 + src/qemu/qemu_command.c | 16 ++-- src/remote/remote_driver.c | 91 +++++++++++++++---- src/remote/remote_protocol.x | 32 ++++++- src/remote_protocol-structs | 16 ++++ src/uml/uml_conf.c | 3 +- src/util/vircommand.c | 159 ++++++++++++++++------------------ src/util/vircommand.h | 13 +-- tests/commandtest.c | 5 +- tools/virsh-domain.c | 82 +++++++++++++++++- tools/virsh.pod | 13 ++- 27 files changed, 960 insertions(+), 167 deletions(-) -- 1.8.1.4

3 13

[libvirt] [PATCH 0/9] Fine grained locking in LXC driver
by Michal Privoznik 18 Jul '13

18 Jul '13

The first round of my patches that breaks LXC driver into smaller pieces. With my testing on 10 domains which are started up and destroyed in a loop with 10 iterations (each loop is run in a separate thread in client): Before: real 0m45.973s user 0m0.080s sys 0m0.020s After: real 0m14.951s user 0m0.080s sys 0m0.020s Michal Privoznik (9): qemu: Move close callbacks handling into util/virclosecallbacks.c Introduce a virLXCDriverConfigPtr object lxc: Use atomic ops for driver->nactive Introduce annotations for virLXCDriverPtr fields lxc: switch to virCloseCallbacks API Stop accessing driver->caps directly in LXC driver lxc: Make activeUsbHostdevs use locks Remove lxcDriverLock from almost everywhere Introduce lxcDomObjFromDomain po/POTFILES.in | 1 + src/Makefile.am | 4 +- src/libvirt_private.syms | 8 + src/lxc/lxc_conf.c | 120 ++++++-- src/lxc/lxc_conf.h | 64 ++-- src/lxc/lxc_controller.c | 2 +- src/lxc/lxc_driver.c | 705 +++++++++++++++---------------------------- src/lxc/lxc_hostdev.c | 8 + src/lxc/lxc_process.c | 181 ++++------- src/lxc/lxc_process.h | 1 - src/qemu/qemu_conf.c | 295 +----------------- src/qemu/qemu_conf.h | 25 +- src/qemu/qemu_driver.c | 4 +- src/qemu/qemu_migration.c | 18 +- src/qemu/qemu_migration.h | 2 +- src/qemu/qemu_process.c | 14 +- src/util/virclosecallbacks.c | 356 ++++++++++++++++++++++ src/util/virclosecallbacks.h | 56 ++++ 18 files changed, 900 insertions(+), 964 deletions(-) create mode 100644 src/util/virclosecallbacks.c create mode 100644 src/util/virclosecallbacks.h -- 1.8.1.5

2 20

[libvirt] [libvirt-sandbox][PATCH] Docs: update EXAMPLES section of virt-sandbox man page
by Alex Jia 18 Jul '13

18 Jul '13

Signed-off-by: Alex Jia <ajia(a)redhat.com> --- bin/virt-sandbox.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/bin/virt-sandbox.c b/bin/virt-sandbox.c index 883e6c0..b51465d 100644 --- a/bin/virt-sandbox.c +++ b/bin/virt-sandbox.c @@ -438,7 +438,9 @@ Run an interactive shell under LXC, replace $HOME with the contents of $HOME/scratch # mkdir $HOME/scratch - # virt-sandbox -c lxc:/// --host-bind $HOME=$HOME/scratch /bin/sh + # echo "hello" > $HOME/scratch/foo + # echo "sandbox" > $HOME/scratch/bar + # virt-sandbox -c lxc:/// -m host-bind:$HOME=$HOME/scratch -i $HOME/scratch/foo -i $HOME/scratch/bar /bin/sh Convert an OGG file to WAV inside QEMU -- 1.7.1

2 2

[libvirt] [sandbox][PATCH] Docs: remove duplicate -u item in create man page
by Wayne Sun 18 Jul '13

18 Jul '13

This related to bug: https://bugzilla.redhat.com/show_bug.cgi?id=916651#c11 Signed-off-by: Wayne Sun <gsun(a)redhat.com> --- bin/virt-sandbox-service-create.pod | 4 ---- 1 files changed, 0 insertions(+), 4 deletions(-) diff --git a/bin/virt-sandbox-service-create.pod b/bin/virt-sandbox-service-create.pod index ad081ea..4dbfcb3 100644 --- a/bin/virt-sandbox-service-create.pod +++ b/bin/virt-sandbox-service-create.pod @@ -187,10 +187,6 @@ static,label=system_u:system_r:svirt_t:s0:c412,c355 Create file system image file of this size to store container content. -=item B<-u UNITFILE>, B<--unitfile UNITFILE> - -systemd Unit file to run within the container - =item B<-P PACKAGE>, B<--package PACKAGE> Package(s) to be used within the container. -- 1.7.1

2 2

[libvirt] Schedule for next release 1.1.1
by Daniel Veillard 18 Jul '13

18 Jul '13

Time to plan the release at the end of the month, for a change I would prefer to release a bit earlier and shoot for Tue 30th July. So this would mean that we enter the freeze on Wednesday 24 next week, hopefully that is fine for everybody, if people have pending patche(s) that still need reviews or ACKs to be pushed in time for the release please raise the issue ! thanks, 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/

1 0

[libvirt] [PATCH v3 0/3] libxl: implement some chuncks of the NUMA interface
by Dario Faggioli 18 Jul '13

18 Jul '13

Hi everyone again, 3rd round for the initial NUMA support in the libxl driver. I think I addressed all the comments that have been raised during v2 review (more details in the signle changelogs). Let me know if there's more. Thanks and Regards, Dario --- Dario Faggioli (3): libxl: implement NUMA capabilities reporting libxl: advertise the support for VIR_TYPED_PARAM_STRING libxl: implement virDomainGetNumaParameters src/libxl/libxl_conf.c | 148 ++++++++++++++++++++++++++++++++++++++++++ src/libxl/libxl_driver.c | 161 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 307 insertions(+), 2 deletions(-) -- <<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

3 11

[libvirt] [PATCH] maint: Make ctags work out of the box
by Jiri Denemark 18 Jul '13

18 Jul '13

The .ctags file specifies default options for ctags so that it does not ignore libvirt.h.in and ignores uninteresting files. As a result, you can just run "ctags" and navigating to a public API won't get you to a useless entry in api.html. --- .ctags | 5 +++++ 1 file changed, 5 insertions(+) create mode 100644 .ctags diff --git a/.ctags b/.ctags new file mode 100644 index 0000000..75e9753 --- /dev/null +++ b/.ctags @@ -0,0 +1,5 @@ +--recurse +--exclude=*.orig +--exclude=*.html +--exclude=*.html.in +--langmap=c:+.h.in -- 1.8.3.2

2 2

[libvirt] [PATCH] esx: Support for disk-only and quiescing snapshots.
by Geoff Hickey 18 Jul '13

18 Jul '13

Add support for creating disk-only (no memory) snapshots in esx, and for quiescing the VM before taking the snapshot. The VMware API supports these operations directly, so adding support to libvirt is just a matter of setting the flags correctly when calling VMware. VIR_DOMAIN_SNAPSHOT_CREATE_DISK_ONLY and VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE are now valid flags for esx. --- src/esx/esx_driver.c | 14 ++++++++++---- 1 files changed, 10 insertions(+), 4 deletions(-) diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index fbe43c2..d69576d 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -4209,9 +4209,14 @@ esxDomainSnapshotCreateXML(virDomainPtr domain, const char *xmlDesc, esxVI_TaskInfoState taskInfoState; char *taskInfoErrorMessage = NULL; virDomainSnapshotPtr snapshot = NULL; + bool diskOnly = (flags & VIR_DOMAIN_SNAPSHOT_CREATE_DISK_ONLY) != 0; + bool quiesce = (flags & VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE) != 0; - /* ESX has no snapshot metadata, so this flag is trivial. */ - virCheckFlags(VIR_DOMAIN_SNAPSHOT_CREATE_NO_METADATA, NULL); + /* ESX supports disk-only and quiesced snapshots, but has no snapshot * + * metadata. */ + virCheckFlags(VIR_DOMAIN_SNAPSHOT_CREATE_DISK_ONLY | + VIR_DOMAIN_SNAPSHOT_CREATE_QUIESCE | + VIR_DOMAIN_SNAPSHOT_CREATE_NO_METADATA, NULL); if (esxVI_EnsureSession(priv->primary) < 0) { return NULL; @@ -4249,8 +4254,9 @@ esxDomainSnapshotCreateXML(virDomainPtr domain, const char *xmlDesc, if (esxVI_CreateSnapshot_Task(priv->primary, virtualMachine->obj, def->name, def->description, - esxVI_Boolean_True, - esxVI_Boolean_False, &task) < 0 || + diskOnly ? esxVI_Boolean_False : esxVI_Boolean_True, + quiesce ? esxVI_Boolean_True : esxVI_Boolean_False, + &task) < 0 || esxVI_WaitForTaskCompletion(priv->primary, task, domain->uuid, esxVI_Occurrence_RequiredItem, priv->parsedUri->autoAnswer, &taskInfoState, -- 1.7.1

2 2

Re: [libvirt] [PATCH] add console support in libxl
by Jim Fehlig 18 Jul '13

18 Jul '13

Bamvor Jian Zhang wrote: > Hi, Jim > > thanks your reply. one comment below. > > >>>已写入 "Jim Fehlig <jfehlig(a)suse.com>"> On 07/04/2013 05:58 AM, Bamvor Jian Zhang wrote: > >> BTW, do all of these types work with Xen? I've only tested with type 'pty', >> >> which works fine with both pv and hvm guests. >> > i only test the pty type too. lots of type is introduced in 2b84e445(Add libxenlight driver), i add the missing type compare with qemu driver. maybe it will be used in future. for now, only pty is needed for my console patch. > I found some time today to test type={unix,tcp,file,udp} and they seem to work fine with Xen 4.3, but HVM guests only. I suppose that is expected since qemu is not being used for serial ports for PV guests. Can you fix all my previous comments, rebase the patch, and send a V2? Oh, and one more thing... > @@ -4739,6 +4857,7 @@ static virDriver libxlDriver = { > .domainManagedSave = libxlDomainManagedSave, /* 0.9.2 */ > .domainHasManagedSaveImage = libxlDomainHasManagedSaveImage, /* 0.9.2 */ > .domainManagedSaveRemove = libxlDomainManagedSaveRemove, /* 0.9.2 */ > + .domainOpenConsole = libxlDomainOpenConsole, /* 1.0.8 */ > .domainIsActive = libxlDomainIsActive, /* 0.9.0 */ > .domainIsPersistent = libxlDomainIsPersistent, /* 0.9.0 */ > .domainIsUpdated = libxlDomainIsUpdated, /* 0.9.0 */ AFAIK, the next release will be 1.1.1. Regards, Jim

2 1