Skip bulk relabelling of resources in SELinux driver when used with LXC
The virSecurityManager{Set,Restore}AllLabel methods are invoked
at domain startup/shutdown to relabel resources associated with
a domain. This works fine with QEMU, but with LXC they are in
fact both currently no-ops since LXC does not support disks,
hostdevs, or kernel/initrd files. Worse, when LXC gains support
for disks/hostdevs, they will do the wrong thing, since they
run in host context, not container context. Thus this patch
turns then into a formal no-op when used with LXC. The LXC
controller will call out to specific security manager labelling
APIs as required during startup.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Refactor LXC NIC creation to allow reuse by hotplug code
The code for creating veth/macvlan devices is part of the
LXC process startup code. Refactor this a little and export
the methods to the rest of the LXC driver. This allows them
to be reused for NIC hotplug code
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The <hostdev> device type has long had a redundant "mode"
attribute, which has always been "subsys". This finally
introduces a new mode "capabilities", which will be used
by the LXC driver for device assignment. Since container
based virtualization uses a single kernel, the idea of
assigning physical PCI devices doesn't make sense. It is
still reasonable to assign USB devices, but for assigning
arbitrary nodes in /dev, the new 'capabilities' mode is
to be used.
The first capability support is 'storage', which is for
assignment of block devices. Functionally this is really
pretty similar to the <disk> support. The only difference
is the device node name is identical in both host and
container namespaces.
The second capability support is 'misc', which is for
assignment of character devices. There is no existing
parallel to this. Again the device node is the same
inside & outside the container.
The reason for keeping the char & storage devices
separate in the domain XML, is to mirror the split
in the node device XML. NB the node device XML does
not yet report character devices, but that's another
new patch to come
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Test cases for virSysinfoRead. Initially, there are tests for
x86 (DMI based) and s390 (/proc/... based).
In lack of PPC data, I have stubbed out the test for it, but it
can be added with a minimal effort.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
There was a double free issue caused by virSysinfoRead on s390,
as the same manufacturer string instance was assigned to more
than one processor record.
Cleaned up other potential memory issues and restructured the sysinfo
parsing code by moving repeating patterns into a helper function.
The restructuring made it necessary to conditionally disable
-Wlogical-op for some older GCC versions, using pragma GCC diagnostic.
This is a GCC specific pragma, which is acceptable, since we're
using it to work around a GCC specific bug.
Finally, added a function virSysinfoSetup to configure the sysinfo
data source files/script during run time, to facilitate writing test
programs. This function is not published in sysinfo.h and only
there for testing.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
Peter Krempa [Mon, 17 Dec 2012 10:49:33 +0000 (11:49 +0100)]
conf: cpu: Refactor parsing of vendor_id and fallback attributes
This patch simplifies the code that parses the fallback and vendor_id
attributes from the domain xml cpu definition.
Changes done:
- free temp variables in the cleanup section instead of local use
- remove checking for presence of the attribute to directly getting the
value (saving call to virXPathBoolean)
- replace loop used to check for ',' in the vendor_id string with strchr
Support custom 'svirt_tcg_t' context for TCG based guests
The current SELinux policy only works for KVM guests, since
TCG requires the 'execmem' privilege. There is a 'virt_use_execmem'
boolean to turn this on globally, but that is unpleasant for users.
This changes libvirt to automatically use a new 'svirt_tcg_t'
context for TCG based guests. This obsoletes the previous
boolean tunable and makes things 'just work(tm)'
Since we can't assume we run with new enough policy, I also
make us log a warning message (once only) if we find the policy
lacks support. In this case we fallback to the normal label and
expect users to set the boolean tunable
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Guannan Ren [Fri, 14 Dec 2012 07:08:01 +0000 (15:08 +0800)]
qemu: use newer -device video device in qemu commandline
'-device VGA' maps to '-vga std'
'-device cirrus-vga' maps to '-vga cirrus'
'-device qxl-vga' maps to '-vga qxl'
(there is also '-device qxl' for secondary devices)
'-device vmware-svga' maps to '-vga vmware'
For qemu(>=1.2), we can use -device to replace -vga for video
device. For the primary video device, the patch tries to use 0x2
slot for matching old qemu. If the 0x2 slot is allocated already,
the addr property could help for using any available slot.
For qemu(< 1.2), we keep using -vga for primary device.
Guannan Ren [Mon, 17 Dec 2012 06:01:20 +0000 (14:01 +0800)]
conf: add optional attribte primary to video <model> element
If there are multiple video devices
primary = 'yes' marks this video device as the primary one.
The rest are secondary video devices. No more than one could be
mark as primary. If none of them has primary attribute, the first
one will be the primary by default like what it was.
The reason of this changing is that for qemu, only one primary video
device is permitted which can be of any type. For secondary video
devices, only qxl is allowd. Primary attribute removes the restriction
that the first have to be the primary one.
We always put the primary video device into the first position of
video device structure array after parsing.
This adds an implementation of virNetSocketGetUNIXIdentity()
using LOCAL_PEERCRED socket option and xucred struct, defined
in <sys/ucred.h> on systems that have it.
Laine Stump [Fri, 14 Dec 2012 17:44:03 +0000 (12:44 -0500)]
network: fix (non)update of dnsmasq config during virDomainUpdateDeviceFlags
A forgotten "!" in recently-modified code at the top of
networkRefreshDaemon() meant an improper early return, which led to 1)
dnsmasq config files not being updated from the newly modified config,
and 2) dnsmasq not being sent a SIGHUP so that it could learn about
the changes to the config.
virNetworkDefGetIpByIndex() returns NULL if there are no ip objects of
the requested type, and if there are no IP elements, then dnsmasq
shouldn't be running, so we can return early. Otherwise we should
rewrite the config files and send a SIGHUP.
Michal Privoznik [Fri, 14 Dec 2012 11:17:55 +0000 (12:17 +0100)]
sanlock: Re-add lockspace unconditionally
Currently, if sanlock is already registering a lockspace other
libvirtd instances (from other hosts) obtain -EINPROGRESS. On
sufficiently new sanlock, sanlock_inq_lockspace() is called,
which suspend execution until lockspace state is changed. With
current libvirt implementation, we fail to retry adding the
lockspace again but continue in error path. Therefore we produce
meaningless error message:
virLockManagerSanlockSetupLockspace:363 : Unable to add lockspace
/var/lib/libvirt/sanlock/__LIBVIRT__DISKS__: Success
qemudLoadDriverConfig:558 : Failed to load lock manager sanlock
We should try to re-add the lockspace after its state change to
be sure it was added successfully. In fact, with sufficiently new
sanlock we can just avoid dummy usleep() which is used if there's
no inquire API.
Eric Blake [Thu, 13 Dec 2012 23:28:27 +0000 (16:28 -0700)]
install: fix virtlockd installation
The virtlockd daemon scripts were lousy, when compared to their
counterparts in daemon/Makefile.am. In particular, when init
scripts were selected, this resulted in 'make distcheck' failing
due to failure to clean up src/virtlockd.init.
* src/Makefile.am (install-systemd): Fix dependencies. Use MKDIR_P.
(uninstall-systemd): Remove empty directory. Use fewer processes.
(install-init, install-sysconfig): Use MKDIR_P.
(uninstall-init): Remove correct file, and also empty directory.
(uninstall-sysconfig): Remove empty directory.
(DISTCLEANFILES): Clean up trivially built sources.
Laine Stump [Fri, 14 Dec 2012 00:09:51 +0000 (19:09 -0500)]
qemu: don't fail update netdev on bridge detach failure
When a network device's bridge connection is changed by
virDomainUpdateDevice, libvirt first removes the netdev's tap from its
old bridge, then adds it to the new bridge. Sometimes, due to a
network being destroyed while a guest device is still attached, the
tap may already be "removed" from the old bridge (or the old bridge
may not even exist any more); the existing code was needlessly failing
the update when this happened, making it impossible to recover from
the situation without completely detaching (i.e. removing) the netdev
from the guest and re-attaching.
Instead of failing the entire operation when removal of the tap from
the old bridge fails, this patch changes qemuDomainChangeNetBridge to
just log a warning and continue, allowing a reasonable recover from
the situation.
(you'll appreciate this change if you ever accidentally destroy a
network while your guests are still using it).
which was originally committed upstream in commit 753ff83a50263d6975f88d6605d4b5ddfcc97560. That commit improperly
removed the "--except-interface lo" from dnsmasq commandlines when
--bind-dynamic was used (based on comments in the latter bug).
It turns out that the problem reported in the CVE could be eliminated
without removing "--except-interface lo", and removing it actually
caused each instance of dnsmasq to listen on localhost on port 53,
which created a new problem:
If another instance of dnsmasq using "bind-interfaces" (instead of
"bind-dynamic") had already been started (or if another instance
started later used "bind-dynamic"), this wouldn't have any immediately
visible ill effects, but if you tried to start another dnsmasq
instance using "bind-interfaces" *after* starting any libvirt
networks, the new dnsmasq would fail to start, because there was
already another process listening on port 53.
(Subsequent to the CVE fix, another patch changed the network driver
to put dnsmasq options in a conf file rather than directly on the
dnsmasq commandline, but preserved the same options.)
This patch changes the network driver to *always* add
"except-interface=lo" to dnsmasq conf files, regardless of whether we use
bind-dynamic or bind-interfaces. This way no libvirt dnsmasq instances
are listening on localhost (and the CVE is still fixed).
The actual code change is miniscule, but must be propogated through all
of the test files as well.
Ján Tomko [Thu, 13 Dec 2012 13:08:01 +0000 (14:08 +0100)]
virsh: use vshReconnect for non-default connections too
For non-default connections (specified by the environment variable or
the command line option) we call virConnectOpenAuth without registering
the vshCatchDisconnect callback.
This calls vshReconnect instead which takes care of it.
Ján Tomko [Thu, 13 Dec 2012 13:08:00 +0000 (14:08 +0100)]
virsh: don't lie about reconnection in vshReconnect
Since we (ab)use vshReconnect for the default URI connection, if it
fails it might print 'Failed to reconnect to the hypervisor' even if we
were never connected before.
This changes it to only mention reconnection on the first try after
getting disconnected.
The default lockd driver behavour is to acquire leases
directly on the disk files. This introduces an alternative
mode, where leases are acquire indirectly on a file that
is based on a SHA256 hash of the disk filename.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This adds a 'lockd' lock driver which is just a client which
talks to the lockd daemon to perform all locking. This will
be the default lock driver for any hypervisor which needs one.
Add support for re-exec() of virtlockd upon SIGUSR1
The virtlockd daemon maintains file locks on behalf of libvirtd
and any VMs it is running. These file locks must be held for as
long as any VM is running. If virtlockd itself ever quits, then
it is expected that a node would be fenced/rebooted. Thus to
allow for software upgrads on live systemd, virtlockd needs the
ability to re-exec() itself.
Upon receipt of SIGUSR1, virtlockd will save its current live
state out to a file /var/run/virtlockd-restart-exec.json
It then re-exec()'s itself with exactly the same argv as it
originally had, and loads the state file, reconstructing any
objects as appropriate.
The state file contains information about all locks held and
all network services and clients currently active. An example
state document is
This enhancement virtlockd so that it can receive a pre-opened
UNIX domain socket from systemd at launch time, and adds the
systemd service/socket unit files
Define a wire protocol for talking to the virtlockd daemon
The virtlockd daemon will be responsible for managing locks
on virtual machines. Communication will be via the standard
RPC infrastructure. This provides the XDR protocol definition
* src/locking/lock_protocol.x: Wire protocol for virtlockd
* src/Makefile.am: Include lock_protocol.[ch] in virtlockd
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Introduce basic infrastructure for virtlockd daemon
The virtlockd daemon will maintain locks on behalf of libvirtd.
There are two reasons for it to be separate
- Avoid risk of other libvirtd threads accidentally
releasing fcntl() locks by opening + closing a file
that is locked
- Ensure locks can be preserved across libvirtd restarts.
virtlockd will need to be able to re-exec itself while
maintaining locks. This is simpler to achieve if its
sole job is maintaining locks
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Refactor virLockManagerPluginNew() so that the caller does
not need to pass in the config file path itself - just the
config directory and driver name.
Fix QEMU to actually pass in a config file when creating the
default lock manager plugin, rather than NULL.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The current virStorageFileGet{LVM,SCSI}Key methods return
the key as the return value. Unfortunately it is desirable
for "NULL" to be a valid return value, as well as an error
indicator. Thus the returned key must instead be provided
as an out-parameter.
When we invoke lvs or scsi_id to extract ID for block devices,
we don't want virCommandWait logging errors messages. Thus we
must explicitly check 'status != 0', rather than letting
virCommandWait do it.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Jim Fehlig [Wed, 12 Dec 2012 21:28:24 +0000 (14:28 -0700)]
Support network boot for HVM guests in libxl
The libxl driver ignored boot devices in the domain config,
preventing PXE booting HVM domains. This patch accounts for
user-specified boot devices when building the libxl domain
configuration.
The QED file format is non-versioned, so although the magic
value matched, libvirt rejected it due to lack of a version
number to compare against. We need to distinguish this case
by allowing a value of '-2' to indicate a non-versioned file
where only the magic is required to match
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Change virCgroupGetAppRoot stub on non-Linux to avoid unused param warning
Fully stub out the virCgroupGetAppRoot method as done with other
methods in the file, rather than just the body. This lets us
annotate the unused parameter to avoid a warning
* Autotools changes:
- Don't assume Qemu is Linux-only
- Check Linux headers only on Linux
- Disable firewalld on FreeBSD
* Initctl:
Initctl seem to present only on Linux, so stub it on other platforms
* Raw I/O: Linux-only as well
* Headers cleanup
Peter Krempa [Tue, 11 Dec 2012 18:20:53 +0000 (19:20 +0100)]
qemu: Add a new domain lookup helper and improve the docs
This patch adds a new domain lookup helper qemuDomObjFromDomainDriver
that lookups the domain and leaves the driver locked. The driver is
returned as the second argument of that function. If the lookup fails
the driver is unlocked to help avoid cleanup codepaths.
Peter Krempa [Tue, 11 Dec 2012 20:26:52 +0000 (21:26 +0100)]
util: rework error reporting in virGet(User|Group)IDByName
This patch gets rid of the undeterministic error reporting code done on
return values of get(pw|gr)nam_r. With this patch, if the group record
is not returned by the corresponding function this error is not
considered fatal even if errno != 0. The error is logged in such case.
Move the code for matching hostdev instances out of virDomainHostdevFind
and into virDomainHostdevMatch method, which in turn calls out to other
helper methods depending on the type of hostdev.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Rename virDomainHostdevPartsParse to virDomainHostdevDefParseSubsys
to reflect the fact that it only deals with hostdevs uing the
traditional mode=subsystem, and not mode=capabilities
Rename virDomainHostSourceFormat to virDomainHostdevDefFormatSubsys
for the same reason.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Cole Robinson [Sun, 21 Oct 2012 02:29:51 +0000 (22:29 -0400)]
tools: Only install guests init script if --with-init=script=redhat
Most of this deals with moving the libvirt-guests.sh script which
does all the work to /usr/libexec, so it can be shared by both
systemd and traditional init. Previously systemd depended on
the script being in /etc/init.d
Required to fix https://bugzilla.redhat.com/show_bug.cgi?id=789747
Serge Hallyn [Tue, 11 Dec 2012 20:20:29 +0000 (20:20 +0000)]
add security hook for permitting hugetlbfs access
When a qemu domain is backed by huge pages, apparmor needs to grant the domain
rw access to files under the hugetlbfs mount point. Add a hook, called in
qemu_process.c, which ends up adding the read-write access through
virt-aa-helper. Qemu will be creating a randomly named file under the
mountpoint and unlinking it as soon as it has mmap()d it, therefore we
cannot predict the full pathname, but for the same reason it is generally
safe to provide access to $path/**.
Eric Blake [Thu, 6 Dec 2012 23:33:07 +0000 (16:33 -0700)]
maint: mention when HACKING is rebuilt during make
I noticed that on Fedora 18, xlstproc decides to regenerate
HACKING with additional whitespace. I haven't figured out why
that is happening (although fixing it would probably be a task
for xlstproc), but in the process of investigating, I noticed
that 'make HACKING' was completely silent, for no good reason.
* Makefile.am (gen-ChangeLog, gen-AUTHORS, NEWS)
($(top_srcdir)/HACKING): Mention which files we are generating.
Peter Krempa [Fri, 7 Dec 2012 10:59:24 +0000 (11:59 +0100)]
qemu: Make qemuMigrationIsAllowed more reusable
This patch exports qemuMigrationIsAllowed and adds a new parameter to it
to denote if it's a remote migration or a local migration. Local
migrations are used in snapshots and saving of the machine state and
have fewer restrictions. This patch also adjusts callers of the function
and tweaks some error messages to be more universal.
Ján Tomko [Mon, 10 Dec 2012 15:28:55 +0000 (16:28 +0100)]
qemu: assume seccomp sandbox is supported since qemu 1.2
Currently there is no way to detect it via QMP and requesting "-sandbox
off" works correctly even if it was compiled out, so this will work
unless someone both requests the sandbox in qemu.conf and builds QEMU
without the support for it.
Interfaces keeps a class_id, which is an ID from which bridge
part of QoS settings is derived. We need to store class_id
in domain status file, so we can later pass it to
virNetDevBandwidthUnplug.
Michal Privoznik [Sun, 18 Nov 2012 15:59:21 +0000 (16:59 +0100)]
network: Create real network status files
Currently, we are only keeping a inactive XML configuration
in status dir. This is no longer enough as we need to keep
this class_id attribute so we don't overwrite old entries
when the daemon restarts. However, since there has already
been release which has just <network/> as root element,
and we want to keep things compatible, detect that loaded
status file is older one, and don't scream about it.
Network should be notified if we plug in or unplug an
interface, so it can perform some action, e.g. set/unset
network part of QoS. However, we are doing this in very
early stage, so iface->ifname isn't filled in yet. So
whenever we want to report an error, we must use a different
identifier, e.g. the MAC address.
Michal Privoznik [Fri, 16 Nov 2012 11:52:44 +0000 (12:52 +0100)]
bandwidth: Create rate update function
This will be used whenever a NIC with guaranteed throughput is to
be plugged into a bridge. It will adjust the average throughput of
non guaranteed NICs (classid 1:2) to meet new requirements.
Michal Privoznik [Fri, 16 Nov 2012 11:37:51 +0000 (12:37 +0100)]
bandwidth: Create (un)plug functions
These set bridge part of QoS when bringing domain's interface up.
Long story short, if there's a 'floor' set, a new QoS class is created.
ClassID MUST be unique within the bridge and should be kept for
unplug phase.
Michal Privoznik [Fri, 16 Nov 2012 10:36:02 +0000 (11:36 +0100)]
bandwidth: Create hierarchical shaping classes
These classes can borrow unused bandwidth. Basically,
only egress qdsics can have classes, therefore we can
do this kind of traffic shaping only on host's outgoing,
that is domain's incoming traffic.
Michal Privoznik [Fri, 16 Nov 2012 09:37:15 +0000 (10:37 +0100)]
bandwidth: add new 'floor' attribute
This is however supported only on domain interfaces with
type='network'. Moreover, target network needs to have at least
inbound QoS set. This is required by hierarchical traffic shaping.
From now on, the required attribute for <inbound/> is either 'average'
(old) or 'floor' (new). This new attribute can be used just for
interfaces type of network (<interface type='network'/>) currently.
Michal Privoznik [Mon, 19 Nov 2012 11:16:07 +0000 (12:16 +0100)]
bandwidth: Attach sfq to leaf node
Stochastic Fairness Queuing (SFQ) is queuing discipline
(qdisc) which doesn't really shape any traffic but 'just'
re-arrange packets in sending buffer so no stream starve.
The goal is to ensure fairness. There is basically only one
configuration parameter (perturb) which is set to advised
value of 10.
Dmitry Guryanov [Tue, 11 Dec 2012 10:59:48 +0000 (14:59 +0400)]
parallels: add routed pseudo network
Historically if traffic from the adapter is routed to LAN without
NAT, it isn't connected to any virtual networks, but has a 'type'
instead. Sinse libvirt has special virtual network type for such case,
let's add pseudo network 'routed' to fit libvirt's API well.
Dmitry Guryanov [Tue, 11 Dec 2012 10:59:45 +0000 (14:59 +0400)]
parallels: add network driver
Parallels Cloud Server uses virtual networks model for network
configuration. It uses own tools for virtual network management.
So add network driver, which will be responsible for listing
virtual networks and performing different operations on them
(in consequent patched).
This patch only allows listing virtual network names, without
any parameters like DHCP server settings.
Dmitry Guryanov [Tue, 11 Dec 2012 10:59:43 +0000 (14:59 +0400)]
parallels: add support of network interfaces to parallelsDomainDefineXML
Allow changing network interfaces in domain configuration.
ifname is used as iterface identifier: if there is interface
with some ifname in old config and there are no interfaces with
such name in the new config - issue prlctl command to delete
the network interface. And vice versa - if interface with
some ifname exists only in new config - issue prlctl command
to create it.
Dmitry Guryanov [Tue, 11 Dec 2012 10:59:42 +0000 (14:59 +0400)]
parallels: parse information about network interfaces
Parse network interfaces info from prlctl output.
Parallels Cloud Server uses virtual networks model for
network configuration: You can add network adapter to
VM and connect it to some predefined virtual network.
Fill type, mac, network name and linkstate fields of
virDomainNetDef structure.
Peter Krempa [Tue, 11 Dec 2012 11:02:16 +0000 (12:02 +0100)]
qemu: snapshot: Remove memory image if external checkpoint fails
When the disk snapshot part of an external system checkpoint fails the
memory image is retained. This patch adds code to remove the image in
such case.
Peter Krempa [Tue, 11 Dec 2012 10:59:58 +0000 (11:59 +0100)]
qemu: snapshot: Don't leak XML definition if restarting of CPUs fails
In case the snapshot code isn't able to restart CPUs after an external
checkpoint we would leak a copy of the domains XML definition. This
patch fixes the cleanup path.
Ján Tomko [Tue, 11 Dec 2012 11:50:53 +0000 (12:50 +0100)]
qemu: fix uninitialized variable warning in doPeer2PeerMigrate
False positive, but it breaks the build with gcc-4.6.3.
qemu/qemu_migration.c:2931:37: error: 'offline' may be used
uninitialized in this function [-Werror=uninitialized]
qemu/qemu_migration.c:2887:10: note: 'offline' was declared here