A couple of places in LXC setup for filesystems did not do
a "goto cleanup" after reporting errors. While fixing this,
also add in many more debug statements to aid troubleshooting
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
With the majority of fields in the virLXCDriverPtr struct
now immutable or self-locking, there is no need for practically
any methods to be using the LXC driver lock. Only a handful
of helper APIs now need it.
The activeUsbHostdevs item in LXCDriver are lockable, but the lock has
to be called explicitly. Call the virObject(Un)Lock() in order to
achieve mutual exclusion once lxcDriverLock is removed.
The 'driver->caps' pointer can be changed on the fly. Accessing
it currently requires the global driver lock. Isolate this
access in a single helper, so a future patch can relax the
locking constraints.
Currently the virLXCDriverPtr struct contains an wide variety
of data with varying access needs. Move all the static config
data into a dedicated virLXCDriverConfigPtr object. The only
locking requirement is to hold the driver lock, while obtaining
an instance of virLXCDriverConfigPtr. Once a reference is held
on the config object, it can be used completely lockless since
it is immutable.
NB, not all APIs correctly hold the driver lock while getting
a reference to the config object in this patch. This is safe
for now since the config is never updated on the fly. Later
patches will address this fully.
Merge the virCommandPreserveFD / virCommandTransferFD methods
into a single virCommandPasFD method, and use a new
VIR_COMMAND_PASS_FD_CLOSE_PARENT to indicate their difference
in behaviour
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the new virDomainCreate{XML}WithFiles methods in the
LXC driver, so that FDs get passed down to the init process.
The lxc_container code needs to do a little dance in order
to renumber the file descriptors it receives into linear
order, starting from STDERR_FILENO + 1.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Commit 75c1256 states that virGetGroupList must not be called
between fork and exec, then commit ee777e99 promptly violated
that for lxc.
Patch originally posted by Eric Blake <eblake@redhat.com>.
Currently, if the primary security driver is 'none', we skip
initializing caps->host.secModels. This means, later, when LXC domain
XML is parsed and <seclabel type='none'/> is found (see
virSecurityLabelDefsParseXML), the model name is not copied to the
seclabel. This leads to subsequent crash in virSecurityManagerGenLabel
where we call STREQ() over the model (note, that we are expecting model
to be !NULL).
Otherwise the container will fail to start if we
enable user namespace, since there is no rights to
do mknod in uninit user namespace.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
lxc driver will use this function to change the owner
of hot added devices.
Move virLXCControllerChown to lxc_container.c and Rename
it to lxcContainerChown.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
When failing to start a container due to inaccessible root
filesystem path, we did not log any meaningful error. Add a
few debug statements to assist diagnosis
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=964358
POSIX states that multi-threaded apps should not use functions
that are not async-signal-safe between fork and exec, yet we
were using getpwuid_r and initgroups. Although rare, it is
possible to hit deadlock in the child, when it tries to grab
a mutex that was already held by another thread in the parent.
I actually hit this deadlock when testing multiple domains
being started in parallel with a command hook, with the following
backtrace in the child:
Thread 1 (Thread 0x7fd56bbf2700 (LWP 3212)):
#0 __lll_lock_wait ()
at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:136
#1 0x00007fd5761e7388 in _L_lock_854 () from /lib64/libpthread.so.0
#2 0x00007fd5761e7257 in __pthread_mutex_lock (mutex=0x7fd56be00360)
at pthread_mutex_lock.c:61
#3 0x00007fd56bbf9fc5 in _nss_files_getpwuid_r (uid=0, result=0x7fd56bbf0c70,
buffer=0x7fd55c2a65f0 "", buflen=1024, errnop=0x7fd56bbf25b8)
at nss_files/files-pwd.c:40
#4 0x00007fd575aeff1d in __getpwuid_r (uid=0, resbuf=0x7fd56bbf0c70,
buffer=0x7fd55c2a65f0 "", buflen=1024, result=0x7fd56bbf0cb0)
at ../nss/getXXbyYY_r.c:253
#5 0x00007fd578aebafc in virSetUIDGID (uid=0, gid=0) at util/virutil.c:1031
#6 0x00007fd578aebf43 in virSetUIDGIDWithCaps (uid=0, gid=0, capBits=0,
clearExistingCaps=true) at util/virutil.c:1388
#7 0x00007fd578a9a20b in virExec (cmd=0x7fd55c231f10) at util/vircommand.c:654
#8 0x00007fd578a9dfa2 in virCommandRunAsync (cmd=0x7fd55c231f10, pid=0x0)
at util/vircommand.c:2247
#9 0x00007fd578a9d74e in virCommandRun (cmd=0x7fd55c231f10, exitstatus=0x0)
at util/vircommand.c:2100
#10 0x00007fd56326fde5 in qemuProcessStart (conn=0x7fd53c000df0,
driver=0x7fd55c0dc4f0, vm=0x7fd54800b100, migrateFrom=0x0, stdin_fd=-1,
stdin_path=0x0, snapshot=0x0, vmop=VIR_NETDEV_VPORT_PROFILE_OP_CREATE,
flags=1) at qemu/qemu_process.c:3694
...
The solution is to split the work of getpwuid_r/initgroups into the
unsafe portions (getgrouplist, called pre-fork) and safe portions
(setgroups, called post-fork).
* src/util/virutil.h (virSetUIDGID, virSetUIDGIDWithCaps): Adjust
signature.
* src/util/virutil.c (virSetUIDGID): Add parameters.
(virSetUIDGIDWithCaps): Adjust clients.
* src/util/vircommand.c (virExec): Likewise.
* src/util/virfile.c (virFileAccessibleAs, virFileOpenForked)
(virDirCreate): Likewise.
* src/security/security_dac.c (virSecurityDACSetProcessLabel):
Likewise.
* src/lxc/lxc_container.c (lxcContainerSetID): Likewise.
* configure.ac (AC_CHECK_FUNCS_ONCE): Check for setgroups, not
initgroups.
Signed-off-by: Eric Blake <eblake@redhat.com>
Recent changes uncovered a NEGATIVE_RETURNS in the return from sysconf()
when processing a for loop in virtTestCaptureProgramExecChild() in
testutils.c
Code review uncovered 3 other code paths with the same condition that
weren't found by Covirity, so fixed those as well.
Convert the type of loop iterators named 'i', 'j', k',
'ii', 'jj', 'kk', to be 'size_t' instead of 'int' or
'unsigned int', also santizing 'ii', 'jj', 'kk' to use
the normal 'i', 'j', 'k' naming
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Create parent directroy for hostdev automatically when we
start a lxc domain or attach a hostdev to a lxc domain.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
This helper function is used to create parent directory for
the hostdev which will be added to the container. If the
parent directory of this hostdev doesn't exist, the mknod of
the hostdev will fail. eg with /dev/net/tun
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Many applications use /dev/tty to read from stdin.
e.g. zypper on openSUSE.
Let's create this device node to unbreak those applications.
As /dev/tty is a synonym for the current controlling terminal
it cannot harm the host or any other containers.
Signed-off-by: Richard Weinberger <richard@nod.at>
libivrt lxc can only set generic weight for container,
This patch allows user to setup per device blkio
weigh for container.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
User namespaces will deny the ability to mount the SELinux
filesystem. This is harmless for libvirt's LXC needs, so the
error can be ignored.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Since these devices are created for the container.
the owner should be the root user of the container.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
container will create /dev/pts directory in /dev.
the owner of /dev should be the root user of container.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Since these tty devices will be used by container,
the owner of them should be the root user of container.
This patch also adds a new function virLXCControllerChown,
we can use this general function to change the owner of
files.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
user namespace doesn't allow to create devices in
uninit userns. We should create devices on host side.
We first mount tmpfs on dev directroy under state dir
of container. then create devices under this dev dir.
Finally in container, mount the dev directroy created
on host to the /dev/ directroy of container.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
This patch introduces new helper function
virLXCControllerSetupUserns, in this function,
we set the files uid_map and gid_map of the init
task of container.
lxcContainerSetID is used for creating cred for
tasks running in container. Since after setuid/setgid,
we may be a new user. This patch calls lxcContainerSetUserns
at first to make sure the new created files belong to
right user.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
User namespace will be enabled only when the idmap exist
in configuration.
If you want disable user namespace,just remove these
elements from XML.
If kernel doesn't support user namespace and idmap exist
in configuration file, libvirt lxc will start failed and
return "Kernel doesn't support user namespace" message.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
As a consequence of the cgroup layout changes from commit 'cfed9ad4', the
lxcDomainGetSchedulerParameters[Flags]()' and lxcGetSchedulerType() APIs
failed to return data for a non running domain. This can be seen through
a 'virsh schedinfo <domain>' command which returns:
Scheduler : Unknown
error: Requested operation is not valid: cgroup CPU controller is not mounted
Prior to that change a non running domain would return:
Scheduler : posix
cpu_shares : 0
vcpu_period : 0
vcpu_quota : 0
emulator_period: 0
emulator_quota : 0
This patch will restore the capability to return configuration only data
for a non running domain regardless of whether cgroups are available.
By default files in a FUSE mount can only be accessed by the
user which created them, even if the file permissions would
otherwise allow it. To allow other users to access the FUSE
mount the 'allow_other' mount option must be used. This bug
prevented non-root users in an LXC container from reading
the /proc/meminfo file.
https://bugzilla.redhat.com/show_bug.cgi?id=967977
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Earlier commit f7e8653f dropped support for using LXC with
kernels having single-instance devpts filesystem from the
LXC controller. It forgot to remove the same code from the
LXC container setup.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
After commit c131525bec
"Auto-add a root <filesystem> element to LXC containers on startup"
for libvirt lxc, root must be existent.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Re-add the selinux header to lxc_container.c since other
functions now use it, beyond the patch that was just
reverted.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Before trying to mount the selinux filesystem in a container
use is_selinux_enabled() to check if the machine actually
has selinux support (eg not booted with selinux=0)
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
During startup, the LXC driver uses paths such as
/.oldroot/var/run/libvirt/lxc/...
to access directories from the previous root filesystem
after doing a pivot_root(). Unfortunately if /var/run
is an absolute symlink to /run, instead of a relative
symlink to ../run, these paths break.
At least one Linux distro is known to use an absolute
symlink for /var/run, so workaround this, by resolving
all symlinks before doing the pivot_root().
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
We do not want to allow contained applications to be able to read fusefs_t.
So we want /proc/meminfo label to match the system default proc_t.
Fix checking of error codes
The lxcContainerMountAllFS method had a 'bool skipRoot'
flag to control whether it mounts the / filesystem. Since
removal of the non-pivot root container setup codepaths,
this flag is obsolete as the only caller always passes
'true'.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Many methods accept a string parameter specifying the
old root directory prefix. Since removal of the non-pivot
root container setup codepaths, this parameter is obsolete
in many methods where the callers always pass "/.oldroot".
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The lxcContainerMountBasicFS method had a 'bool pivotRoot'
flag to control whether it mounted a private /dev. Since
removal of the non-pivot root container setup codepaths,
this flag is obsolete as the only caller always passes
'true'.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The LXC driver can already configure <disk> or <filesystem>
devices to use the loop device. This extends it to also allow
for use of the NBD device, to support non-raw formats.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The current code for setting up loop devices to LXC disks first
does a switch() based on the disk format, then looks at the
disk driver name. Reverse this so it first looks at the driver
name, and then the disk format. This is more useful since the
list of supported disk formats depends on what driver is used.
The code for setting loop devices for LXC fs entries also needs
to have the same logic added, now the XML schema supports this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The LXC, QEMU, and LibXL drivers have all merged their handling of
the attach/update/modify device APIs into one large
'xxxxDomainModifyDeviceFlags'
which then does a 'switch()' based on the actual API being invoked.
While this saves some lines of code, it is not really all that
significant in the context of the driver API impls as a whole.
This merger of the handling of different APIs creates pain when
wanting to automated analysis of the code and do things which
are specific to individual APIs. The slight duplication of code
from unmerged the API impls, is preferrable to allow for easier
automated analysis.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The individual hypervisor drivers were directly referencing
APIs in virnodesuspend.c in their virDriverPtr struct. Separate
these methods, so there is always a wrapper in the hypervisor
driver. This allows the unused virConnectPtr args to be removed
from the virnodesuspend.c file. Again this will ensure that
ACL checks will only be performed on invocations that are
directly associated with public API usage.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The individual hypervisor drivers were directly referencing
APIs in src/nodeinfo.c in their virDriverPtr struct. Separate
these methods, so there is always a wrapper in the hypervisor
driver. This allows the unused virConnectPtr args to be
removed from the nodeinfo.c file. Again this will ensure that
ACL checks will only be performed on invocations that are
directly associated with public API usage.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the virGetHostname() API has a bogus virConnectPtr
parameter. This is because virtualization drivers directly
reference this API in their virDriverPtr tables, tieing its
API design to the public virConnectGetHostname API design.
This also causes problems for access control checks since
these must only be done for invocations from the public
API, not internal invocation.
Remove the bogus virConnectPtr parameter, and make each
hypervisor driver provide a dedicated function for the
driver API impl. This will allow access control checks
to be easily inserted later.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The source code base needs to be adapted as well. Some files
include virutil.h just for the string related functions (here,
the include is substituted to match the new file), some include
virutil.h without any need (here, the include is removed), and
some require both.
Ensure that all drivers implementing public APIs use a
naming convention for their implementation that matches
the public API name.
eg for the public API virDomainCreate make sure QEMU
uses qemuDomainCreate and not qemuDomainStart
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Ensure that the driver struct field names match the public
API names. For an API virXXXX we must have a driver struct
field xXXXX. ie strip the leading 'vir' and lowercase any
leading uppercase letters.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The change in commit aed4986322
was incomplete, missing a couple of cases of /system. This
caused failure to start VMs.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
After discussions with systemd developers it was decided that
a better default policy for resource partitions is to have
3 default partitions at the top level
/system - system services
/machine - virtual machines / containers
/user - user login session
This ensures that the default policy isolates guest from
user login sessions & system services, so a mis-behaving
guest can't consume 100% of CPU usage if other things are
contending for it.
Thus we change the default partition from /system to
/machine
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
http://www.uhv.edu/ac/newsletters/writing/grammartip2009.07.01.htm
(and several other sites) give hints that 'onto' is best used if
you can also add 'up' just before it and still make sense. In many
cases in the code base, we really want the two-word form, or even
a simplification to just 'on' or 'to'.
* docs/hacking.html.in: Use correct 'on to'.
* python/libvirt-override.c: Likewise.
* src/lxc/lxc_controller.c: Likewise.
* src/util/virpci.c: Likewise.
* daemon/THREADS.txt: Use simpler 'on'.
* docs/formatdomain.html.in: Better usage.
* docs/internals/rpc.html.in: Likewise.
* src/conf/domain_event.c: Likewise.
* src/rpc/virnetclient.c: Likewise.
* tests/qemumonitortestutils.c: Likewise.
* HACKING: Regenerate.
Signed-off-by: Eric Blake <eblake@redhat.com>
The LXC driver currently has code to detect cgroups mounts
and then re-mount them inside the new root filesystem. Replace
this fragile code with a call to virCgroupIsolateMount.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The virCgroupNewDriver method had a 'bool privileged' param.
If a false value was ever passed in, it would simply not
work, since non-root users don't have any privileges to create
new cgroups. Just delete this broken code entirely and make
the QEMU driver skip cgroup setup in non-privileged mode
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
A resource partition is an absolute cgroup path, ignoring the
current process placement. Expose a virCgroupNewPartition API
for constructing such cgroups
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Rename all the virCgroupForXXX methods to use the form
virCgroupNewXXX since they are all constructors. Also
make sure the output parameter is the last one in the
list, and annotate all pointers as non-null. Fix up
all callers, and make sure they use true/false not 0/1
for the boolean parameters
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Instead of calling virCgroupForDomain every time we need
the virCgrouPtr instance, just do it once at Vm startup
and cache a reference to the object in virLXCDomainObjPrivatePtr
until shutdown of the VM. Removing the virCgroupPtr from
the LXC driver state also means we don't have stale mount
info, if someone mounts the cgroups filesystem after libvirtd
has been started
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
If the user requests a mount for /run, this may hide any existing
mounts that are lower down in /run. The result is that the
container still sees the mounts in /proc/mounts, but cannot
access them
sh-4.2# df
df: '/run/user/501/gvfs': No such file or directory
df: '/run/media/berrange/LIVE': No such file or directory
df: '/run/media/berrange/SecureDiskA1': No such file or directory
df: '/run/libvirt/lxc/sandbox': No such file or directory
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/mapper/vg_t500wlan-lv_root 151476396 135390200 8384900 95% /
tmpfs 1970888 3204 1967684 1% /run
/dev/sda1 194241 155940 28061 85% /boot
devfs 64 0 64 0% /dev
tmpfs 64 0 64 0% /sys/fs/cgroup
tmpfs 1970888 1200 1969688 1% /etc/libvirt-sandbox/scratch
Before mounting any filesystem at a particular location, we
must recursively unmount anything at or below the target mount
point
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Ensure lxcContainerUnmountSubtree is at the top of the
lxc_container.c file so it is easily referenced from
any other method. No functional change
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This allows a container-type domain to have exclusive access to one of
the host's NICs.
Wire <hostdev caps=net> with the lxc_controller - when moving the newly
created veth devices into a new namespace, also look for any hostdev
devices that should be moved. Note: once the container domain has been
destroyed, there is no code that moves the interfaces back to the
original namespace. This does happen, though, probably due to default
cleanup on namespace destruction.
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@freescale.com>
The virCgroupMounted method is badly named, since a controller can be
mounted, but disabled in the current object. Rename the method to be
virCgroupHasController. Also make it tolerant to a NULL virCgroupPtr
and out-of-range controller index, to avoid duplication of these
checks in all callers
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently when getting an instance of virCgroupPtr we will
create the path in all cgroup controllers. Only at the virt
driver layer are we attempting to filter controllers. This
is bad because the mere act of creating the dirs in the
controllers can have a functional impact on the kernel,
particularly for performance.
Update the virCgroupForDriver() method to accept a bitmask
of controllers to use. Only create dirs in the controllers
that are requested. When creating cgroups for domains,
respect the active controller list from the parent cgroup
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The virCgroupGetAppRoot is not clear in its meaning. Change
to virCgroupForSelf to highlight that this returns the
cgroup config for the caller's process
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This patch refactors various places to allow removing of the
defaultConsoleTargetType callback from the virCaps structure.
A new console character device target type is introduced -
VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_NONE - to mark that no type was
specified in the XML. This type is at the end converted to the standard
VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL. Other types that are
different from this default have to be processed separately in the
device post parse callback.
Use the virDomainXMLConf structure to hold this data and tweak the code
to avoid semantic change.
Without configuration the KVM mac prefix is used by default. I chose it
as it's in the privately administered segment so it should be usable for
any purposes.
This patch removes the emulatorRequired field and associated
infrastructure from the virCaps object. Instead the driver specific
callbacks are used as this field isn't enforced by all drivers.
This patch implements the appropriate callbacks in the qemu and lxc
driver and moves to check to that location.
This patch adds instrumentation that will allow hypervisor drivers to
fill and validate domain and device definitions after parsed by the XML
parser.
With this patch, after the XML is parsed, a callback to the driver is
issued requesting to fill and validate driver specific details of the
configuration. This allows to use sensible defaults and checks on a per
driver basis at the time the XML is parsed.
Two callback pointers are stored in the new virDomainXMLConf object:
* virDomainDeviceDefPostParseCallback (devicesPostParseCallback)
- called for a single device parsed and for every single device in a
domain config. A virDomainDeviceDefPtr is passed along with the
domain definition and virCaps.
* virDomainDefPostParseCallback, (domainPostParseCallback)
- A callback that is meant to process the domain config after it's
parsed. A virDomainDefPtr is passed along with virCaps.
Both types of callbacks support arbitrary opaque data passed for the
callback functions.
Errors may be reported in those callbacks resulting in a XML parsing
failure.
This patch is the result of running:
for i in $(git ls-files | grep -v html | grep -v \.po$ ); do
sed -i -e "s/virDomainXMLConf/virDomainXMLOption/g" -e "s/xmlconf/xmlopt/g" $i
done
and a few manual tweaks.
This reverts commit c9c87376f2.
Now that we force all containers to have a root filesystem,
there is no way the host's /dev is ever exposed
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the LXC container code has two codepaths, depending on
whether there is a <filesystem> element with a target path of '/'.
If we automatically add a <filesystem> device with src=/ and dst=/,
for any container which has not specified a root filesystem, then
we only need one codepath for setting up the filesystem.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Early on kernel support for private devpts was not widespread,
so we had compatibiltiy codepaths. Such old kernels are not
seriously used for LXC these days, so the compat code can go
away
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
For a root filesystem with type=file or type=block, the LXC
container was forgetting to actually mount it, before doing
the pivot root step.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the lxc controller sets up the devpts instance on
$rootfsdef->src, but this only works if $rootfsdef is using
type=mount. To support type=block or type=file for the root
filesystem, we must use /var/lib/libvirt/lxc/$NAME.devpts
for the temporary devpts mount in the controller
Instead of using /var/lib/libvirt/lxc/$NAME for the FUSE
filesystem, use /var/lib/libvirt/lxc/$NAME.fuse. This allows
room for other temporary mounts in the same directory
Some of the LXC callbacks did not lock the virDomainObjPtr
instance. This caused transient errors like
error: Failed to start domain busy-mount
error: cannot rename file '/var/run/libvirt/lxc/busy-mount.xml.new' as '/var/run/libvirt/lxc/busy-mount.xml': No such file or directory
as 2 threads tried to update the status file concurrently
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The 'nodeset' variable was never initialized, causing a later
VIR_FREE(nodeset) to free uninitialized memory.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Intend to reduce the redundant code,use virNumaSetupMemoryPolicy
to replace virLXCControllerSetupNUMAPolicy and
qemuProcessInitNumaMemoryPolicy.
This patch also moves the numa related codes to the
file virnuma.c and virnuma.h
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Allow lxc using the advisory nodeset from querying numad,
this means if user doesn't specify the numa nodes that
the lxc domain should assign to, libvirt will automatically
bind the lxc domain to the advisory nodeset which queried from
numad.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
The LXC controller is closing loop devices as soon as the
container has started. This is fine if the loop device
was setup as a mounted filesystem, but if we're just passing
through the loop device as a disk, nothing else is keeping
it open. Thus we must keep the loop device FDs open for as
long the libvirt_lxc process is running.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the LXC controller creates the cgroup, configures the
resources and adds the task all in one go. This is not sufficiently
flexible for the forthcoming NBD integration. We need to make sure
the NBD process gets into the right cgroup immediately, but we can
not have limits (in particular the device ACL) applied at the point
where we start qemu-nbd. So create a virLXCCgroupCreate method
which creates the cgroup and adds the current task to be called
early, and leave virLXCCgroupSetup to only do resource config.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The naming used in the RPC protocols for the LXC monitor and
lock daemon confused the script used to generate systemtap
helper functions. Rename the LXC monitor protocol symbols to
reduce confusion. Adapt the gensystemtap.pl script to cope
with the LXC monitor / lock daemon naming conversions.
This has no functional impact on RPC wire protocol, since
names are only used in the C layer
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
If an LXC domain failed to start because of a bogus SELinux
label, virLXCProcessStart would call VIR_CLOSE(0) by mistake.
This is because the code which initializes the member of the
ttyFDs array to -1 got moved too far away from the place where
the array is first allocated.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
In some startup failure modes, the fuse thread may get itself
wedged. This will cause the entire libvirt_lxc process to
hang trying to the join the thread. There is no compelling
reason to wait for the thread to exit if the whole process
is exiting, so just daemonize the fuse thread instead.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The virDomainGetSecurityLabel method is currently (mistakenly)
showing the label of the libvirt_lxc process:
...snip...
Security model: selinux
Security DOI: 0
Security label: system_u:system_r:virtd_t:s0-s0:c0.c1023 (permissive)
when it should be showing the init process label
...snip...
Security model: selinux
Security DOI: 0
Security label: system_u:system_r:svirt_t:s0:c724,c995 (permissive)
The virCaps structure gathered a ton of irrelevant data over time that.
The original reason is that it was propagated to the XML parser
functions.
This patch aims to create a new data structure virDomainXMLConf that
will contain immutable data that are used by the XML parser. This will
allow two things we need:
1) Get rid of the stuff from virCaps
2) Allow us to add callbacks to check and add driver specific stuff
after domain XML is parsed.
This first attempt removes pointers to private data allocation functions
to this new structure and update all callers and function that require
them.
When setting up disks with loop devices for LXC, one of the
switch cases was missing a 'break' causing it to fallthrough
to an error condition.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
otherwise we crash with
#0 virUSBDeviceListFind (list=0x0, dev=dev@entry=0x8193d70) at util/virusb.c:526
#1 0xb1a4995b in virLXCPrepareHostdevUSBDevices (driver=driver@entry=0x815d9a0, name=0x815dbf8 "debian-700267", list=list@entry=0x81d8f08) at lxc/lxc_hostdev.c:88
#2 0xb1a49fce in virLXCPrepareHostUSBDevices (def=0x8193af8, driver=0x815d9a0) at lxc/lxc_hostdev.c:261
#3 virLXCPrepareHostDevices (driver=driver@entry=0x815d9a0, def=0x8193af8) at lxc/lxc_hostdev.c:328
#4 0xb1a4c5b1 in virLXCProcessStart (conn=0x817d3f8, driver=driver@entry=0x815d9a0, vm=vm@entry=0x8190908, autoDestroy=autoDestroy@entry=false, reason=reason@entry=VIR_DOMAIN_RUNNING_BOOTED)
at lxc/lxc_process.c:1068
#5 0xb1a57e00 in lxcDomainStartWithFlags (dom=dom@entry=0x815e460, flags=flags@entry=0) at lxc/lxc_driver.c:1014
#6 0xb1a57fc3 in lxcDomainStart (dom=0x815e460) at lxc/lxc_driver.c:1046
#7 0xb79c8375 in virDomainCreate (domain=domain@entry=0x815e460) at libvirt.c:8450
#8 0x08078959 in remoteDispatchDomainCreate (args=0x81920a0, rerr=0xb65c21d0, client=0xb0d00490, server=<optimized out>, msg=<optimized out>) at remote_dispatch.h:1066
#9 remoteDispatchDomainCreateHelper (server=0x80c4928, client=0xb0d00490, msg=0xb0d005b0, rerr=0xb65c21d0, args=0x81920a0, ret=0x815d208) at remote_dispatch.h:1044
#10 0xb7a36901 in virNetServerProgramDispatchCall (msg=0xb0d005b0, client=0xb0d00490, server=0x80c4928, prog=0x80c6438) at rpc/virnetserverprogram.c:432
#11 virNetServerProgramDispatch (prog=0x80c6438, server=server@entry=0x80c4928, client=0xb0d00490, msg=0xb0d005b0) at rpc/virnetserverprogram.c:305
#12 0xb7a300a7 in virNetServerProcessMsg (msg=<optimized out>, prog=<optimized out>, client=<optimized out>, srv=0x80c4928) at rpc/virnetserver.c:162
#13 virNetServerHandleJob (jobOpaque=0xb0d00510, opaque=0x80c4928) at rpc/virnetserver.c:183
#14 0xb7924f98 in virThreadPoolWorker (opaque=opaque@entry=0x80a94b0) at util/virthreadpool.c:144
#15 0xb7924515 in virThreadHelper (data=0x80a9440) at util/virthreadpthread.c:161
#16 0xb7887c39 in start_thread (arg=0xb65c2b70) at pthread_create.c:304
#17 0xb77eb78e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:130
when adding a domain with a usb device. This is Debian bug
http://bugs.debian.org/700267
This fixes the build on Debian Wheezy which otherwise fails with:
CC libvirt_driver_lxc_impl_la-lxc_process.lo
lxc/lxc_process.c: In function 'virLXCProcessGetNsInode':
lxc/lxc_process.c:648:5: error: implicit declaration of function 'stat' [-Werror=implicit-function-declaration]
lxc/lxc_process.c:648:5: error: nested extern declaration of 'stat' [-Werror=nested-externs]
cc1: all warnings being treated as errors
By using a loopback device, disks backed by plain files can
be made available to LXC containers. We make no attempt to
auto-detect format if <driver type="raw"/> is not set,
instead we unconditionally treat that as meaning raw. This
is to avoid the security issues inherent with format
auto-detection
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently we rely on a VIR_ERROR message being logged by the
virRaiseError function to report LXC startup errors. This gives
the right message, but is rather ugly and can be truncated
if lots of log messages are written. Change the LXC controller
to explicitly print any virErrorPtr message to stderr. Then
change the driver to skip over anything that looks like a log
message.
The result is that this
error: Failed to start domain busy
error: internal error guest failed to start: 2013-03-04 19:46:42.846+0000: 1734: info : libvirt version: 1.0.2
2013-03-04 19:46:42.846+0000: 1734: error : virFileLoopDeviceAssociate:600 : Unable to open /root/disk.raw: No such file or directory
changes to
error: Failed to start domain busy
error: internal error guest failed to start: Unable to open /root/disk.raw: No such file or directory
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
In the LXC container startup code when switching stdio
streams, we call VIR_FORCE_CLOSE on all FDs. This triggers
a huge number of warnings, but we don't see them because
stdio is closed at this point. strace() however shows them
which can confuse people debugging the code. Switch to
VIR_MASS_CLOSE to avoid this
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
To enable locking to be introduced to the security manager
objects later, turn virSecurityManager into a virObjectLockable
class
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
To enable virCapabilities instances to be reference counted,
turn it into a virObject. All cases of virCapabilitiesFree
turn into virObjectUnref
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
To allow modifications to the lists to be synchronized, convert
virPCIDeviceList and virUSBDeviceList into virObjectLockable
classes. The locking, however, will not be self-contained. The
users of these classes will have to call virObjectLock/Unlock
in the critical regions.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
When iterating over USB host devices to setup cgroups, the
usbDevice object was leaked in both LXC and QEMU driers
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The duplicate VM checking should be done atomically with
virDomainObjListAdd, so shoud not be a separate function.
Instead just use flags to indicate what kind of checks are
required.
This pair, used in virDomainCreateXML:
if (virDomainObjListIsDuplicate(privconn->domains, def, 1) < 0)
goto cleanup;
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def, false)))
goto cleanup;
Changes to
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def,
VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE,
NULL)))
goto cleanup;
This pair, used in virDomainRestoreFlags:
if (virDomainObjListIsDuplicate(privconn->domains, def, 1) < 0)
goto cleanup;
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def, true)))
goto cleanup;
Changes to
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def,
VIR_DOMAIN_OBJ_LIST_ADD_LIVE |
VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE,
NULL)))
goto cleanup;
This pair, used in virDomainDefineXML:
if (virDomainObjListIsDuplicate(privconn->domains, def, 0) < 0)
goto cleanup;
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def, false)))
goto cleanup;
Changes to
if (!(dom = virDomainObjListAdd(privconn->domains,
privconn->caps,
def,
0, NULL)))
goto cleanup;
As a step towards making virDomainObjList thread-safe turn it
into an opaque virObject, preventing any direct access to its
internals.
As part of this a new method virDomainObjListForEach is
introduced to replace all existing usage of virHashForEach
A followon to commit id: 68dceb635 - if iface->iname is NULL, then
neither virNetDevOpenvswitchRemovePort() nor virNetDevVethDelete()
should be called. Found by Coverity.
The use of switch statements inside a bounded for loop resulted in some
false positives regarding the "default:" label which cannot be reached
since each of the other case statements use the possible for loop values.
The virDomainObj, qemuAgent, qemuMonitor, lxcMonitor classes
all require a mutex, so can be switched to use virObjectLockable
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently all classes must directly inherit from virObject.
This allows for arbitrarily deep hierarchy. There's not much
to this aside from chaining up the 'dispose' handlers from
each class & providing APIs to check types.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
libvirt lxc will fail to start when selinux is disabled.
error: Failed to start domain noroot
error: internal error guest failed to start: PATH=/bin:/sbin TERM=linux container=lxc-libvirt container_uuid=b9873916-3516-c199-8112-1592ff694a9e LIBVIRT_LXC_UUID=b9873916-3516-c199-8112-1592ff694a9e LIBVIRT_LXC_NAME=noroot /bin/sh
2013-01-09 11:04:05.384+0000: 1: info : libvirt version: 1.0.1
2013-01-09 11:04:05.384+0000: 1: error : lxcContainerMountBasicFS:546 : Failed to mkdir /sys/fs/selinux: No such file or directory
2013-01-09 11:04:05.384+0000: 7536: info : libvirt version: 1.0.1
2013-01-09 11:04:05.384+0000: 7536: error : virLXCControllerRun:1466 : error receiving signal from container: Input/output error
2013-01-09 11:04:05.404+0000: 7536: error : virCommandWait:2287 : internal error Child process (ip link del veth1) unexpected exit status 1: Cannot find device "veth1"
fix this problem by checking if selinuxfs is mounted
in host before we try to create dir /sys/fs/selinux.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
The virDomainLxcOpenNamespace method needs to open every file
in /proc/$INITPID/ns and return the open file descriptor to the
client application.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This patch introduces support for LXC specific public APIs. In
common with what was done for QEMU, this creates a libvirt_lxc.so
library and libvirt/libvirt-lxc.h header file.
The actual APIs are
int virDomainLxcOpenNamespace(virDomainPtr domain,
int **fdlist,
unsigned int flags);
int virDomainLxcEnterNamespace(virDomainPtr domain,
unsigned int nfdlist,
int *fdlist,
unsigned int *noldfdlist,
int **oldfdlist,
unsigned int flags);
which provide a way to use the setns() system call to move the
calling process into the container's namespace. It is not
practical to write in a generically applicable manner. The
nearest that we could get to such an API would be an API which
allows to pass a command + argv to be executed inside a
container. Even if we had such a generic API, this LXC specific
API is still useful, because it allows the caller to maintain
the current process context, in particular any I/O streams they
have open.
NB the virDomainLxcEnterNamespace() API is special in that it
runs client side, so does not involve the internal driver API.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
when we has no host's src mapped to container.
there is no .oldroot dir,so libvirt lxc will fail
to start when mouting meminfo.
in this case,the parameter srcprefix of function
lxcContainerMountProcFuse should be NULL.and make
this method handle NULL correctly.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
To avoid confusion between the LXC driver <-> controller
monitor RPC protocol and the libvirt-lxc.so <-> libvirtd public
RPC protocol, rename the former to lxc_monitor_protocol.x
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The code for setting up a private /dev/pts for the containers
is also responsible for making the LXC controller have a
private mount namespace. Unfortunately the /dev/pts code is
not run if launching a container without a custom root. This
causes the LXC FUSE mount to leak into the host FS.
To bring in line with new naming practice, rename the=
src/util/cgroup.{h,c} files to vircgroup.{h,c}
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Convert the host capabilities and domain config structs to
use the virArch datatype. Update the parsers and all drivers
to take account of datatype change
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach device drivers in LXC to support the
hotplug/unplug of host misc devices.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach device drivers in LXC to support the
hotplug/unplug of host storage devices.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach device drivers in LXC to support the
hotplug/unplug of USB host devices.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach/update device APIs to support changing
of hostdevs in the persistent config file
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach/update device APIs to support changing
of disks in the persistent config file
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Wire up the attach/detach/update device APIs to support changing
of network interfaces in the persistent config file
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This wires up the LXC driver to support the domain device attach/
detach/update APIs, following the same code design as used in
the QEMU driver. No actual changes are possible with this commit,
it is only providing the framework
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This extends support for host device passthrough with LXC to
cover misc devices. In this case all we need todo is a
mknod in the container's /dev and whitelist the device in
cgroups
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This extends support for host device passthrough with LXC to
cover storage devices. In this case all we need todo is a
mknod in the container's /dev and whitelist the device in
cgroups
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This adds support for host device passthrough with the
LXC driver. Since there is only a single kernel image,
it doesn't make sense to pass through PCI devices, but
USB devices are fine. For the latter we merely need to
make the /dev/bus/usb/NNN/MMM character device exist
in the container's /dev
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently LXC guests can be given arbitrary pre-mounted
filesystems, however, for some usecases it is more appropriate
to provide block devices which the container can mount itself.
This first impl only allows for <disk type='block'>, in other
words exposing a host disk device to a container. Since LXC
does not have device namespace virtualization, we are cheating
a little bit. If the XML specifies /dev/sdc4 to be given to
the container as /dev/sda1, when we do the mknod /dev/sda1
in the container's /dev, we actually use the major:minor
number of /dev/sdc4, not /dev/sda1.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
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>
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.
Currently to deal with auto-shutdown libvirtd must periodically
poll all stateful drivers. Thus sucks because it requires
acquiring both the driver lock and locks on every single virtual
machine. Instead pass in a "inhibit" callback to virStateInitialize
which drivers can invoke whenever they want to inhibit shutdown
due to existance of active VMs.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add support for doing controlled shutdown / reboot in the LXC
driver. The default behaviour is to try talking to /dev/initctl
inside the container's virtual root (/proc/$INITPID/root). This
works with sysvinit or systemd. If that file does not exist
then send SIGTERM (for shutdown) or SIGHUP (for reboot). These
signals are not any kind of particular standard for shutdown
or reboot, just something apps can choose to handle. The new
virDomainSendProcessSignal allows for sending custom signals.
We might allow the choice of SIGTERM/HUP to be configured for
LXC containers via the XML in the future.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The virStateInitialize method and several cgroups methods were
using an 'int privileged' parameter or similar for dual-state
values. These are better represented with the bool type.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Implement the new API for sending signals to processes in a guest
for the LXC driver. Only support sending signals to the init
process for now, because
- The kernel does not appear to expose the mapping between
container PID numbers and host PID numbers anywhere in the
host OS namespace
- There is no race-free way to validate whether a host PID
corresponds to a process in a container.
* src/lxc/lxc_driver.c: Allow sending processes signals
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
we already have virtualize meminfo for container through fuse filesystem,
add function lxcContainerMountProcFuse to mount this meminfo file to
the container's /proc/meminfo.
So we can isolate container's /proc/meminfo from host now.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
with this patch,container's meminfo will be shown based on
containers' mem cgroup.
Right now,it's impossible to virtualize all values in meminfo,
I collect some values such as MemTotal,MemFree,Cached,Active,
Inactive,Active(anon),Inactive(anon),Active(file),Inactive(anon),
Active(file),Inactive(file),Unevictable,SwapTotal,SwapFree.
if I miss something, please let me know.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
this patch addes fuse support for libvirt lxc.
we can use fuse filesystem to generate sysinfo dynamically,
So we can isolate /proc/meminfo,cpuinfo and so on through
fuse filesystem.
we mount fuse filesystem for every container.
the mount name is libvirt,mount point is
localstatedir/run/libvirt/lxc/containername.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
When starting an LXC guest with a virNetwork based NIC device,
if the network was not active, the virNetworkPtr device would
be leaked
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
When starting a container, newDef is initialized to a
copy of 'def', but when startup fails newDef is never
removed. This cause later attempts to use 'virDomainDefine'
to lose the new data being defined.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
A mistaken initialization of 'ret' caused failure to create
macvtap devices to be ignored. The libvirt_lxc process
would later fail to start due to missing devices
Also make sure code checks '< 0' and not '!= 0' since only
-1 is considered an error condition
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
If the <interface> device did not contain any <target>
element, LXC would crash on a NULL pointer if starting
the container failed
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The LXC driver relies on use of cgroups to kill off LXC processes
in shutdown. If cgroups aren't available, we're unable to kill
off processes, so we must treat lack of cgroups as a fatal startup
error.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The code setting up LXC cgroups used an 'rc' variable both
for capturing the return value of methods it calls, and
its own return status. The result was that several failures
in setting up cgroups would actually result in success being
returned.
Use a separate 'ret' for tracking return value as per normal
code design in other parts of libvirt
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The initpid will be required long term to enable LXC to
implement various hotplug operations. Thus it needs to be
persisted in the domain status XML. LXC has not used the
domain status XML before, so this introduces use of the
helpers.
Currently the lxcContainerSetupMounts method uses the
virSecurityManagerPtr instance to obtain the mount options
string and then only passes the string down into methods
it calls. As functionality in LXC grows though, those
methods need to have direct access to the virSecurityManagerPtr
instance. So push the code down a level.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The impls of virSecurityManagerGetMountOptions had no way to
return errors, since the code was treating 'NULL' as a success
value. This is somewhat pointless, since the calling code did
not want NULL in the first place and has to translate it into
the empty string "". So change the code so that the impls can
return "" directly, allowing use of NULL for error reporting
once again
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
When no security driver is specified libvirt_lxc segfaults as a debug
message tries to access security labels for the container that are not
present.
This problem was introduced in commit 6c3cf57d6c.
Early jumps to the cleanup label caused a crash of the libvirt_lxc
container helper as the cleanup section called
virLXCControllerDeleteInterfaces(ctrl) without checking the ctrl argument
for NULL. The argument was de-referenced soon after.
$ /usr/libexec/libvirt_lxc
/usr/libexec/libvirt_lxc: missing --name argument for configuration
Segmentation fault
The virLXCControllerClientCloseHook method was mistakenly
assuming that the private data associated with the network
client was the virLXCControllerPtr. In fact it was just a
dummy int, so we were derefencing a bogus struct. The
frequent result of this was that we would never quit, because
we tried to arm a non-existant timer.
Fix the code by removing the dummy private data and just
using the virLXCControllerPtr instance as private data
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the LXC driver logs audit messages when a container
is started or stopped. These audit messages, however, contain
the PID of the libvirt_lxc supervisor process. To enable
sysadmins to correlate with audit messages generated by
processes /inside/ the container, we need to include the
container init process PID.
We can't do this in the main 'start' audit message, since
the init PID is not available at that point. Instead we output
a completely new audit record, that lists both PIDs.
type=VIRT_CONTROL msg=audit(1353433750.071:363): pid=20180 uid=0 auid=501 ses=3 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='virt=lxc op=init vm="busy" uuid=dda7b947-0846-1759-2873-0f375df7d7eb vm-pid=20371 init-pid=20372 exe="/home/berrange/src/virt/libvirt/daemon/.libs/lt-libvirtd" hostname=? addr=? terminal=pts/6 res=success'
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The LXC controller code currently directly invokes the
libvirt main loop code. The problem is that this misses
the cleanup of virNetServerClient connections that
virNetServerRun takes care of.
The result is that when libvirtd is stopped, the
libvirt_lxc controller process gets stuck in a I/O loop.
When libvirtd is then started again, it fails to connect
to the controller and thus kills off the entire domain.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
For S390, the default console target type cannot be of type 'serial'.
It is necessary to at least interpret the 'arch' attribute
value of the os/type element to produce the correct default type.
Therefore we need to extend the signature of defaultConsoleTargetType
to account for architecture. As a consequence all the drivers
supporting this capability function must be updated.
Despite the amount of changed files, the only change in behavior is
that for S390 the default console target type will be 'virtio'.
N.B.: A more future-proof approach could be to to use hypervisor
specific capabilities to determine the best possible console type.
For instance one could add an opaque private data pointer to the
virCaps structure (in case of QEMU to hold capsCache) which could
then be passed to the defaultConsoleTargetType callback to determine
the console target type.
Seems to be however a bit overengineered for the use case...
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
The libvirt coding standard is to use 'function(...args...)'
instead of 'function (...args...)'. A non-trivial number of
places did not follow this rule and are fixed in this patch.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This needs to be done before the container starts. Turning
off the mknod capability is noticed by systemd, which will
no longer attempt to create device nodes.
This eliminates SELinux AVC messages and ugly failure messages in the journal.
Add two new APIs virNetServerClientNewPostExecRestart and
virNetServerClientPreExecRestart which allow a virNetServerClientPtr
object to be created from a JSON object and saved to a
JSON object, for the purpose of re-exec'ing a process.
This includes serialization of the connected socket associated
with the client
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
There was an inverted return value in lxcCgroupControllerActive().
The function assumes cgroups are active and do couple of checks
to prove that. If any of them fails, false is returned. Therefore,
at the end, after all checks are done we must return true, not false.
Many parts of virDomainDefPtr were using 'int' variables as
array length counts. Replace all these with size_t and update
various format strings & API signatures to adapt
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Depending on the scenario in which LXC containers exit, it is
possible for the EOF callback of the LXC monitor to deadlock
the driver.
#0 0x00000038a0a0de4d in __lll_lock_wait () from /lib64/libpthread.so.0
#1 0x00000038a0a09ca6 in _L_lock_840 () from /lib64/libpthread.so.0
#2 0x00000038a0a09ba8 in pthread_mutex_lock () from /lib64/libpthread.so.0
#3 0x00007f4bd9579d55 in virMutexLock (m=<optimized out>) at util/threads-pthread.c:85
#4 0x00007f4bcacc7597 in lxcDriverLock (driver=0x7f4bc40c8290) at lxc/lxc_conf.h:81
#5 virLXCProcessMonitorEOFNotify (mon=<optimized out>, vm=0x7f4bb4000b00) at lxc/lxc_process.c:581
#6 0x00007f4bd9645c91 in virNetClientCloseLocked (client=client@entry=0x7f4bb4009e60)
at rpc/virnetclient.c:554
#7 0x00007f4bd96460f8 in virNetClientIOEventLoopPassTheBuck (thiscall=0x0, client=0x7f4bb4009e60)
at rpc/virnetclient.c:1306
#8 virNetClientIOEventLoopPassTheBuck (client=0x7f4bb4009e60, thiscall=0x0)
at rpc/virnetclient.c:1287
#9 0x00007f4bd96467a2 in virNetClientCloseInternal (reason=3, client=0x7f4bb4009e60)
at rpc/virnetclient.c:589
#10 virNetClientCloseInternal (client=0x7f4bb4009e60, reason=3) at rpc/virnetclient.c:561
#11 0x00007f4bcacc4a82 in virLXCMonitorClose (mon=0x7f4bb4000a00) at lxc/lxc_monitor.c:201
#12 0x00007f4bcacc55ac in virLXCProcessCleanup (reason=<optimized out>, vm=0x7f4bb4000b00,
driver=0x7f4bc40c8290) at lxc/lxc_process.c:240
#13 virLXCProcessStop (driver=0x7f4bc40c8290, vm=vm@entry=0x7f4bb4000b00,
reason=reason@entry=VIR_DOMAIN_SHUTOFF_DESTROYED) at lxc/lxc_process.c:735
#14 0x00007f4bcacc5bd2 in virLXCProcessAutoDestroyDom (payload=<optimized out>,
name=0x7f4bb4003c80, opaque=0x7fff41af2df0) at lxc/lxc_process.c:94
#15 0x00007f4bd9586649 in virHashForEach (table=0x7f4bc409b270,
iter=iter@entry=0x7f4bcacc5ab0 <virLXCProcessAutoDestroyDom>, data=data@entry=0x7fff41af2df0)
at util/virhash.c:514
#16 0x00007f4bcacc52d7 in virLXCProcessAutoDestroyRun (driver=driver@entry=0x7f4bc40c8290,
conn=conn@entry=0x7f4bb8000ab0) at lxc/lxc_process.c:120
#17 0x00007f4bcacca628 in lxcClose (conn=0x7f4bb8000ab0) at lxc/lxc_driver.c:128
#18 0x00007f4bd95e67ab in virReleaseConnect (conn=conn@entry=0x7f4bb8000ab0) at datatypes.c:114
When the driver calls virLXCMonitorClose, there is really no
need for the EOF callback to be invoked in this case, since
the caller can easily handle events itself. In changing this,
the monitor needs to take a deep copy of the callback list,
not merely a reference.
Also adds debug statements in various places to aid
troubleshooting
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Asynchronously setting priv->mon to NULL was pointless,
just remove the destroy callback entirely.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Continue consolidation of process functions by moving some
helpers out of command.{c,h} into virprocess.{c,h}
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Some kernel versions (at least RHEL-6 2.6.32) do not let you over-mount
an existing selinuxfs instance with a new one. Thus we must unmount the
existing instance inside our namespace.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
https://www.gnu.org/licenses/gpl-howto.html recommends that
the 'If not, see <url>.' phrase be a separate sentence.
* tests/securityselinuxhelper.c: Remove doubled line.
* tests/securityselinuxtest.c: Likewise.
* globally: s/; If/. If/
The introduction of /sys/fs/cgroup came in fairly recent kernels.
Prior to that time distros would pick a custom directory like
/cgroup or /dev/cgroup. We need to auto-detect where this is,
rather than hardcoding it
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This bug was revealed by the crash described in
https://bugzilla.redhat.com/show_bug.cgi?id=852383
The vlan info pointer sent to virNetDevOpenvswitchAddPort should never
be non-NULL unless there is at least one tag. The factthat such a vlan
info pointer was receveid pointed out that a caller was passing the
wrong pointer. Instead of sending &net->vlan, the result of
virDomainNetGetActualVlan(net) should be sent - that function will
look for vlan info in net->data.network.actual->vlan, and in cany case
return NULL instead of a pointer if the vlan info it finds has no
tags.
Aside from causing the crash, sending a hardcoded &net->vlan has the
effect of ignoring vlan info from a <network> or <portgroup> config.
This patch updates the structures that store information about each
domain and each hypervisor to support multiple security labels and
drivers. It also updates all the remaining code to use the new fields.
Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>
Add the ability to support VLAN tags for Open vSwitch virtual port
types. To accomplish this, modify virNetDevOpenvswitchAddPort and
virNetDevTapCreateInBridgePort to take a virNetDevVlanPtr
argument. When adding the port to the OVS bridge, setup either a
single VLAN or a trunk port based on the configuration from the
virNetDevVlanPtr.
Signed-off-by: Kyle Mestery <kmestery@cisco.com>
Currently there is a hook function that is invoked when a
new client connection comes in, which allows an app to
setup private data. This setup will make it difficult to
serialize client state during process re-exec(). Change to
a model where the app registers a callback when creating
the virNetServerPtr instance, which is used to allocate
the client private data immediately during virNetClientPtr
construction.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the virNetClientPtr constructor will always register
the async IO event handler and the keepalive objects. In the
case of the lock manager, there will be no event loop available
nor keepalive support required. Split this setup out of the
constructor and into separate methods.
The remote driver will enable async IO and keepalives, while
the LXC driver will only enable async IO
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
As the consensus in:
https://www.redhat.com/archives/libvir-list/2012-July/msg01692.html,
this patch is to destroy conf/virdomainlist.[ch], folding the
helpers into conf/domain_conf.[ch].
* src/Makefile.am:
- Various indention fixes incidentally
- Add macro DATATYPES_SOURCES (datatypes.[ch])
- Link datatypes.[ch] for libvirt_lxc
* src/conf/domain_conf.c:
- Move all the stuffs from virdomainlist.c into it
- Use virUnrefDomain and virUnrefDomainSnapshot instead of
virDomainFree and virDomainSnapshotFree, which are defined
in libvirt.c, and we don't want to link to it.
- Remove "if" before "free" the object, as virObjectUnref
is in the list "useless_free_options".
* src/conf/domain_conf.h:
- Move all the stuffs from virdomainlist.h into it
- s/LIST_FILTER/LIST_DOMAINS_FILTER/
* src/libxl/libxl_driver.c:
- s/LIST_FILTER/LIST_DOMAINS_FILTER/
- no (include "virdomainlist.h")
* src/libxl/libxl_driver.c: Likewise
* src/lxc/lxc_driver.c: Likewise
* src/openvz/openvz_driver.c: Likewise
* src/parallels/parallels_driver.c: Likewise
* src/qemu/qemu_driver.c: Likewise
* src/test/test_driver.c: Likewise
* src/uml/uml_driver.c: Likewise
* src/vbox/vbox_tmpl.c: Likewise
* src/vmware/vmware_driver.c: Likewise
* tools/virsh-domain-monitor.c: Likewise
* tools/virsh.c: Likewise
The meat of this patch is just moving the calls to
virNWFilterRegisterCallbackDriver from each hypervisor's "register"
function into its "initialize" function. The rest is just code
movement to allow that, and a new virNWFilterUnRegisterCallbackDriver
function to undo what the register function does.
The long explanation:
There is an array in nwfilter called callbackDrvArray that has
pointers to a table of functions for each hypervisor driver that are
called by nwfilter. One of those function pointers is to a function
that will lock the hypervisor driver. Entries are added to the table
by calling each driver's "register" function, which happens quite
early in libvirtd's startup.
Sometime later, each driver's "initialize" function is called. This
function allocates a driver object and stores a pointer to it in a
static variable that was previously initialized to NULL. (and here's
the important part...) If the "initialize" function fails, the driver
object is freed, and that pointer set back to NULL (but the entry in
nwfilter's callbackDrvArray is still there).
When the "lock the driver" function mentioned above is called, it
assumes that the driver was successfully loaded, so it blindly tries
to call virMutexLock on "driver->lock".
BUT, if the initialize never happened, or if it failed, "driver" is
NULL. And it just happens that "lock" is always the first field in
driver so it is also NULL.
Boom.
To fix this, the call to virNWFilterRegisterCallbackDriver for each
driver shouldn't be called until the end of its (*already guaranteed
successful*) "initialize" function, not during its "register" function
(which is currently the case). This implies that there should also be
a virNWFilterUnregisterCallbackDriver() function that is called in a
driver's "shutdown" function (although in practice, that function is
currently never called).
Switch virDomainObjPtr to use the virObject APIs for reference
counting. The main change is that virObjectUnref does not return
the reference count, merely a bool indicating whether the object
still has any refs left. Checking the return value is also not
mandatory.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
All callers used the same initialization seed (well, the new
viratomictest forgot to look at getpid()); so we might as well
make this value automatic. And while it may feel like we are
giving up functionality, I documented how to get it back in the
unlikely case that you actually need to debug with a fixed
pseudo-random sequence. I left that crippled by default, so
that a stray environment variable doesn't cause a lack of
randomness to become a security issue.
* src/util/virrandom.c (virRandomInitialize): Rename...
(virRandomOnceInit): ...and make static, with one-shot call.
Document how to do fixed-seed debugging.
* src/util/virrandom.h (virRandomInitialize): Drop prototype.
* src/libvirt_private.syms (virrandom.h): Don't export it.
* src/libvirt.c (virInitialize): Adjust caller.
* src/lxc/lxc_controller.c (main): Likewise.
* src/security/virt-aa-helper.c (main): Likewise.
* src/util/iohelper.c (main): Likewise.
* tests/seclabeltest.c (main): Likewise.
* tests/testutils.c (virtTestMain): Likewise.
* tests/viratomictest.c (mymain): Likewise.
Otherwise, a build may fail with:
lxc/lxc_conatiner.c: In function 'lxcContainerDropCapabilities':
lxc/lxc_container.c:1662:46: error: unused parameter 'keepReboot' [-Werror=unused-parameter]
* src/lxc/lxc_container.c (lxcContainerDropCapabilities): Mark
parameter unused.
If an LXC container is using a virtual network and that network
is not active, currently the user gets a rather unhelpful
error message about tap device setup failure. Add an explicit
check for whether the network is active, in exactly the same
way as the QEMU driver
The reboot() syscall is allowed by new kernels for LXC containers.
The LXC controller can detect whether a reboot was requested
(instead of a normal shutdown) by looking at the "init" process
exit status. If a reboot was triggered, the exit status will
record SIGHUP as the kill reason.
The LXC controller has cleared all its capabilities, and the
veth network devices will no longer exist at this time. Thus
it cannot restart the container init process itself. Instead
it emits an event which is picked up by the LXC driver in
libvirtd. This will then re-create the container, using the
same configuration as it was previously running with (ie it
will not activate 'newDef').
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Check whether the reboot() system call is virtualized, and if
it is, then allow the container to keep CAP_SYS_REBOOT.
Based on an original patch by Serge Hallyn
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This defines a new RPC protocol to be used between the LXC
controller and the libvirtd LXC driver. There is only a
single RPC message defined thus far, an asynchronous "EXIT"
event that is emitted just before the LXC controller process
exits. This provides the LXC driver with details about how
the container shutdown - normally, or abnormally (crashed),
thus allowing the driver to emit better libvirt events.
Emitting the event in the LXC controller requires a few
little tricks with the RPC service. Simply calling the
virNetServiceClientSendMessage does not work, since this
merely queues the message for asynchronous processing.
In addition the main event loop is no longer running at
the point the event is emitted, so no I/O is processed.
Thus after invoking virNetServiceClientSendMessage it is
necessary to mark the client as being in "delayed close"
mode. Then the event loop is run again, until the client
completes its close - this happens only after the queued
message has been fully transmitted. The final complexity
is that it is not safe to run virNetServerQuit() from the
client close callback, since that is invoked from a
context where the server is locked. Thus a zero-second
timer is used to trigger shutdown of the event loop,
causing the controller to finally exit.
* src/Makefile.am: Add rules for generating RPC protocol
files and dispatch methods
* src/lxc/lxc_controller.c: Emit an RPC event immediately
before exiting
* src/lxc/lxc_domain.h: Record the shutdown reason
given by the controller
* src/lxc/lxc_monitor.c, src/lxc/lxc_monitor.h: Register
RPC program and event handler. Add callback to let
driver receive EXIT event.
* src/lxc/lxc_process.c: Use monitor exit event to decide
what kind of domain event to emit
* src/lxc/lxc_protocol.x: Define wire protocol for LXC
controller monitor.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Move the code that handles the LXC monitor out of the
lxc_process.c file and into lxc_monitor.{c,h}
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Update the LXC driver to use the virNetClient APIs for
connecting to the libvirt_lxc monitor, instead of the
low-level socket APIs. This is a step towards running
a full RPC protocol with libvirt_lxc
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
