The QEMU driver has a list of devices nodes that are whitelisted
for all guests. The kernel has recently started returning an
error if you try to whitelist a device which does not exist.
This causes a warning in libvirt logs and an audit error for
any missing devices. eg
2013-02-27 16:08:26.515+0000: 29625: warning : virDomainAuditCgroup:451 : success=no virt=kvm resrc=cgroup reason=allow vm="vm031714" uuid=9d8f1de0-44f4-a0b1-7d50-e41ee6cd897b cgroup="/sys/fs/cgroup/devices/libvirt/qemu/vm031714/" class=path path=/dev/kqemu rdev=? acl=rw
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The code for putting the emulator threads in a separate cgroup
would spam the logs with warnings
2013-02-27 16:08:26.731+0000: 29624: warning : virCgroupMoveTask:887 : no vm cgroup in controller 3
2013-02-27 16:08:26.731+0000: 29624: warning : virCgroupMoveTask:887 : no vm cgroup in controller 4
2013-02-27 16:08:26.732+0000: 29624: warning : virCgroupMoveTask:887 : no vm cgroup in controller 6
This is because it has only created child cgroups for 3 of the
controllers, but was trying to move the processes from all the
controllers. The fix is to only try to move threads in the
controllers we actually created. Also remove the warning and
make it return a hard error to avoid such lazy callers in the
future.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=896685 points out
a regression caused by commit 38c4a9c - libvirt only labels
the backing chain if the backing chain cache is populated, but
the code to populate the cache was only conditionally performed
if cgroup labeling was necessary.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Hoist cache setup...
* src/qemu/qemu_process.c (qemuProcessStart): ...earlier into
caller, where it is now unconditional.
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>
Currently the virQEMUDriverPtr struct contains an wide variety
of data with varying access needs. Move all the static config
data into a dedicated virQEMUDriverConfigPtr object. The only
locking requirement is to hold the driver lock, while obtaining
an instance of virQEMUDriverConfigPtr. 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.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
While OOM can have knock-on effects that trash a system, generally
the first symptom is one of memory thrashing.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Reword slightly.
Currently, if there's no hard memory limit defined for a domain,
libvirt tries to calculate one, based on domain definition and magic
equation and set it upon the domain startup. The rationale behind was,
if there's a memory leak or exploit in qemu, we should prevent the
host system trashing. However, the equation was too tightening, as it
didn't reflect what the kernel counts into the memory used by a
process. Since many hosts do have a swap, nobody hasn't noticed
anything, because if hard memory limit is reached, process can
continue allocating memory on a swap. However, if there is no swap on
the host, the process gets killed by OOM killer. In our case, the qemu
process it is.
To prevent this, we need to relax the hard RSS limit. Moreover, we
should reflect more precisely the kernel way of accounting the memory
for process. That is, even the kernel caches are counted within the
memory used by a process (within cgroups at least). Hence the magic
equation has to be changed:
limit = 1.5 * (domain memory + total video memory) + (32MB for cache
per each disk) + 200MB
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>
When LXC labels USB devices during hotplug, it is running in
host context, so it needs to pass in a vroot path to the
container root.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Remove the obsolete 'qemud' naming prefix and underscore
based type name. Introduce virQEMUDriverPtr as the replacement,
in common with LXC driver naming style
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>
When the cpu placement model is "auto", it sets the affinity for
domain process with the advisory nodeset from numad, however,
creating cgroup for the domain process (called emulator thread
in some contexts) later overrides that with pinning it to all
available pCPUs.
How to reproduce:
* Configure the domain with "auto" placement for <vcpu>, e.g.
<vcpu placement='auto'>4</vcpu>
* % virsh start dom
* % cat /proc/$dompid/status
Though the emulator cgroup cause conflicts, but we can't simply
prohibit creating it, as other tunables are still useful, such
as "emulator_period", which is used by API
virDomainSetSchedulerParameter. So this patch doesn't prohibit
creating the emulator cgroup, but inherit the nodeset from numad,
and reset the affinity for domain process.
* src/qemu/qemu_cgroup.h: Modify definition of qemuSetupCgroupForEmulator
to accept the passed nodenet
* src/qemu/qemu_cgroup.c: Set the affinity with the passed nodeset
We used to walk the backing file chain at least twice per disk,
once to set up cgroup device whitelisting, and once to set up
security labeling. Rather than walk the chain every iteration,
which possibly includes calls to fork() in order to open root-squashed
NFS files, we can exploit the cache of the previous patch.
* src/conf/domain_conf.h (virDomainDiskDefForeachPath): Alter
signature.
* src/conf/domain_conf.c (virDomainDiskDefForeachPath): Require caller
to supply backing chain via disk, if recursion is desired.
* src/security/security_dac.c
(virSecurityDACSetSecurityImageLabel): Adjust caller.
* src/security/security_selinux.c
(virSecuritySELinuxSetSecurityImageLabel): Likewise.
* src/security/virt-aa-helper.c (get_files): Likewise.
* src/qemu/qemu_cgroup.c (qemuSetupDiskCgroup)
(qemuTeardownDiskCgroup): Likewise.
(qemuSetupCgroup): Pre-populate chain.
According to our recent changes (clarifications), we should be pinning
qemu's emulator processes using the <vcpu> 'cpuset' attribute in case
there is no <emulatorpin> specified. This however doesn't work
entirely as expected and this patch should resolve all the remaining
issues.
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/
This is another fix for the emulator-pin series. When going through
the cputune pinning settings, the current code is trying to pin all
the CPUs, even when not all of them are specified. This causes error
in the subsequent function which, of course, cannot find the cpu to
pin. Since it's enough to pass the correct VCPU ID to the function,
the fix is trivial.
When domain XML contains any of the elements for setting up CPU
scheduling parameters (period, quota, emulator_period, or
emulator_quota) we need cpu cgroup to enforce the configuration.
However, the existing code would just ignore silently such settings if
either cgroups were not available at all cpu cgroup was not available.
Moreover, APIs for manipulating CPU scheduler parameters were already
failing if cpu cgroup was not available. This patch makes cpu cgroup
mandatory for all domains that use CPU scheduling elements in their XML.
If cgroups are enabled in general but cpu cgroup is disabled in
qemu.conf or not mounted at all, libvirt would refuse to start any
domain even though scheduler parameters are not set in domain XML.
This patch makes cpu cgroup mandatory only for domains that actually
want to use it.
Commit 4b03d59167 changed the pinning
behavior in a way that makes some machines non-startable.
The comment mentioning that we cannot control each vcpu when there is
not VCPU<-> PID mapping available is true, however, this isn't
necessarily an error, because this can be caused by old QEMU without
support for "query-cpus" command as well as a software emulated
machines that don't create more than one process.
This patch introduces support of setting emulator's period and
quota to limit cpu bandwidth when the vm starts. Also updates
XML Schema for new entries and docs.
This patch changes the behaviour of xml element cputune.period
and cputune.quota to limit cpu bandwidth only for vcpus, and no
longer limit cpu bandwidth for the whole guest.
The reasons to do this are:
- This matches docs of cputune.period and cputune.quota.
- The other parts excepting vcpus are treated as "emulator",
and there are separate period/quota settings for emulator
in the subsequent patches
Introduce qemuSetupCgroupEmulatorPin() function to add emulator
threads pin info to cpuset cgroup, the same as vcpupin.
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
vcpu threads pin are implemented using sched_setaffinity(), but
not controlled by cgroup. This patch does the following things:
1) enable cpuset cgroup
2) reflect all the vcpu threads pin info to cgroup
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Create a new cgroup and move all emulator threads to the new cgroup.
And then we can do the other things:
1. limit only vcpu usage rather than the whole qemu
2. limit for emulator threads(include vhost-net threads)
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
If there's a memory leak in qemu or qemu is exploited the host's
system will sooner or later start trashing instead of killing
the bad process. This however has impact on performance and other
guests as well. Therefore we should set a reasonable RSS limit
even when user hasn't set any. It's better to be secure by default.
Any time we have a string with no % passed through gettext, a
translator can inject a % to cause a stack overread. When there
is nothing to format, it's easier to ask for a string that cannot
be used as a formatter, by using a trivial "%s" format instead.
In the past, we have used --disable-nls to catch some of the
offenders, but that doesn't get run very often, and many more
uses have crept in. Syntax check to the rescue!
The syntax check can catch uses such as
virReportError(code,
_("split "
"string"));
by using a sed script to fold context lines into one pattern
space before checking for a string without %.
This patch is just mechanical insertion of %s; there are probably
several messages touched by this patch where we would be better
off giving the user more information than a fixed string.
* cfg.mk (sc_prohibit_diagnostic_without_format): New rule.
* src/datatypes.c (virUnrefConnect, virGetDomain)
(virUnrefDomain, virGetNetwork, virUnrefNetwork, virGetInterface)
(virUnrefInterface, virGetStoragePool, virUnrefStoragePool)
(virGetStorageVol, virUnrefStorageVol, virGetNodeDevice)
(virGetSecret, virUnrefSecret, virGetNWFilter, virUnrefNWFilter)
(virGetDomainSnapshot, virUnrefDomainSnapshot): Add %s wrapper.
* src/lxc/lxc_driver.c (lxcDomainSetBlkioParameters)
(lxcDomainGetBlkioParameters): Likewise.
* src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML)
(virDomainDiskDefParseXML, virDomainGraphicsDefParseXML):
Likewise.
* src/conf/network_conf.c (virNetworkDNSHostsDefParseXML)
(virNetworkDefParseXML): Likewise.
* src/conf/nwfilter_conf.c (virNWFilterIsValidChainName):
Likewise.
* src/conf/nwfilter_params.c (virNWFilterVarValueCreateSimple)
(virNWFilterVarAccessParse): Likewise.
* src/libvirt.c (virDomainSave, virDomainSaveFlags)
(virDomainRestore, virDomainRestoreFlags)
(virDomainSaveImageGetXMLDesc, virDomainSaveImageDefineXML)
(virDomainCoreDump, virDomainGetXMLDesc)
(virDomainMigrateVersion1, virDomainMigrateVersion2)
(virDomainMigrateVersion3, virDomainMigrate, virDomainMigrate2)
(virStreamSendAll, virStreamRecvAll)
(virDomainSnapshotGetXMLDesc): Likewise.
* src/nwfilter/nwfilter_dhcpsnoop.c (virNWFilterSnoopReqLeaseDel)
(virNWFilterDHCPSnoopReq): Likewise.
* src/openvz/openvz_driver.c (openvzUpdateDevice): Likewise.
* src/openvz/openvz_util.c (openvzKBPerPages): Likewise.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Likewise.
* src/qemu/qemu_command.c (qemuBuildHubDevStr, qemuBuildChrChardevStr)
(qemuBuildCommandLine): Likewise.
* src/qemu/qemu_driver.c (qemuDomainGetPercpuStats): Likewise.
* src/qemu/qemu_hotplug.c (qemuDomainAttachNetDevice): Likewise.
* src/rpc/virnetsaslcontext.c (virNetSASLSessionGetIdentity):
Likewise.
* src/rpc/virnetsocket.c (virNetSocketNewConnectUNIX)
(virNetSocketSendFD, virNetSocketRecvFD): Likewise.
* src/storage/storage_backend_disk.c
(virStorageBackendDiskBuildPool): Likewise.
* src/storage/storage_backend_fs.c
(virStorageBackendFileSystemProbe)
(virStorageBackendFileSystemBuild): Likewise.
* src/storage/storage_backend_rbd.c
(virStorageBackendRBDOpenRADOSConn): Likewise.
* src/storage/storage_driver.c (storageVolumeResize): Likewise.
* src/test/test_driver.c (testInterfaceChangeBegin)
(testInterfaceChangeCommit, testInterfaceChangeRollback):
Likewise.
* src/vbox/vbox_tmpl.c (vboxListAllDomains): Likewise.
* src/xenxs/xen_sxpr.c (xenFormatSxprDisk, xenFormatSxpr):
Likewise.
* src/xenxs/xen_xm.c (xenXMConfigGetUUID, xenFormatXMDisk)
(xenFormatXM): Likewise.
Per the FSF address could be changed from time to time, and GNU
recommends the following now: (http://www.gnu.org/licenses/gpl-howto.html)
You should have received a copy of the GNU General Public License
along with Foobar. If not, see <http://www.gnu.org/licenses/>.
This patch removes the explicit FSF address, and uses above instead
(of course, with inserting 'Lesser' before 'General').
Except a bunch of files for security driver, all others are changed
automatically, the copyright for securify files are not complete,
that's why to do it manually:
src/security/security_selinux.h
src/security/security_driver.h
src/security/security_selinux.c
src/security/security_apparmor.h
src/security/security_apparmor.c
src/security/security_driver.c
The only useful translation of "%s" as a format string is "%s" (I
suppose you could claim "%1$s" is also valid, but why bother). So
it is not worth translating; fixing this exposes some instances
where we were failing to translate real error messages. This makes
the fix of commit 097da1ab more generic, as well as ensuring no
future regressions.
* cfg.mk (sc_prohibit_useless_translation): New rule.
* src/lxc/lxc_driver.c (lxcSetVcpuBWLive): Fix offender.
* src/openvz/openvz_conf.c (openvzReadFSConf): Likewise.
* src/qemu/qemu_cgroup.c (qemuSetupCgroupForVcpu): Likewise.
* src/qemu/qemu_driver.c (qemuSetVcpusBWLive): Likewise.
* src/xenapi/xenapi_utils.c (xenapiSessionErrorHandle): Likewise.
Like for 'static' placement, when the memory policy mode is
'strict', set the memory policy by writing the advisory nodeset
returned from numad to cgroup file cpuset.mems,
This eliminates the warning message reported in:
https://bugzilla.redhat.com/show_bug.cgi?id=624447
It was caused by a failure to open an image file that is not
accessible by root (the uid libvirtd is running as) because it's on a
root-squash NFS share, owned by a different user, with permissions of
660 (or maybe 600).
The solution is to use virFileOpenAs() rather than open(). The
codepath that generates the error is during qemuSetupDiskCGroup(), but
the actual open() is in a lower-level generic function called from
many places (virDomainDiskDefForeachPath), so some other pieces of the
code were touched just to add dummy (or possibly useful) uid and gid
arguments.
Eliminating this warning message has the nice side effect that the
requested operation may even succeed (which in this case isn't
necessary, but shouldn't hurt anything either).
filter 0-device-weight when:
- getting blkio parameters with --config
- starting up a domain
When testing with blkio, I found these issues:
(dom is down)
virsh blkiotune dom --device-weights /dev/sda,300,/dev/sdb,500
virsh blkiotune dom --device-weights /dev/sda,300,/dev/sdb,0
virsh blkiotune dom
weight : 800
device_weight : /dev/sda,200,/dev/sdb,0
# issue 1: shows 0 device weight of /dev/sdb that may confuse user
(continued)
virsh start dom
# issue 2: If /dev/sdb doesn't exist, libvirt refuses to bring the
# dom up because it wants to set the device weight to 0 of a
# non-existing device. Since 0 means no weight-limit, we really don't
# have to set it.
The cpu bandwidth is applied at the vcpu group level. We should apply it
at the vm group level too, because the vm may do heavy I/O, and it will affect
the other vm.
We apply cpu bandwidth at the vcpu and the vm group level, so we must ensure
that max(child_quota) <= parent_quota when we modify cpu bandwidth.
The LXC and UML drivers can both make use of auditing. Move
the qemu_audit.{c,h} files to src/conf/domain_audit.{c,h}
* src/conf/domain_audit.c: Rename from src/qemu/qemu_audit.c
* src/conf/domain_audit.h: Rename from src/qemu/qemu_audit.h
* src/Makefile.am: Remove qemu_audit.{c,h}, add domain_audit.{c,h}
* src/qemu/qemu_audit.h, src/qemu/qemu_cgroup.c,
src/qemu/qemu_command.c, src/qemu/qemu_driver.c,
src/qemu/qemu_hotplug.c, src/qemu/qemu_migration.c,
src/qemu/qemu_process.c: Update for changed audit API names
Detected by Coverity. All existing callers happen to be in
range, so this isn't too serious.
* src/qemu/qemu_cgroup.c (qemuCgroupControllerActive): Check
bounds before dereference.
These VIR_XXXX0 APIs make us confused, use the non-0-suffix APIs instead.
How do these coversions works? The magic is using the gcc extension of ##.
When __VA_ARGS__ is empty, "##" will swallow the "," in "fmt," to
avoid compile error.
example: origin after CPP
high_level_api("%d", a_int) low_level_api("%d", a_int)
high_level_api("a string") low_level_api("a string")
About 400 conversions.
8 special conversions:
VIR_XXXX0("") -> VIR_XXXX("msg") (avoid empty format) 2 conversions
VIR_XXXX0(string_literal_with_%) -> VIR_XXXX(%->%%) 0 conversions
VIR_XXXX0(non_string_literal) -> VIR_XXXX("%s", non_string_literal)
(for security) 6 conversions
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Clang warned about a dead assignment. In the process, I noticed
that we are only using the function for a bool value. I audited
all other callers in qemu_{migration,cgroup,driver,hotplug), and
all were making the call in a bool context.
Also, do bounds checking on the argument.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Delete dead
assignment.
(qemuCgroupControllerActive): Change return type to bool.
* src/qemu/qemu_cgroup.h (qemuCgroupControllerActive): Likewise.
It throws errors as long as the cgroup controller is not available,
regardless of whether we really want to use it to do setup or not,
which is not what we want, fixing it with throwing error when need
to use the controller.
And change "VIR_WARN" to "qemuReportError" for memory controller
incidentally.
When domain startup, setting cpu affinity and cpu shares according
to the cputune xml specified in domain xml.
Modify "qemudDomainPinVcpu" to update domain config for vcpupin,
and modify "qemuSetSchedulerParameters" to update domain config
for cpu shares.
v1 - v2:
* Use "VIR_ALLOC_N" instead of "VIR_ALLOC_VAR"
* But keep raising error when it fails on adding vcpupin xml
entry, as I still don't have a better idea yet.
* Correct the documentation for cgroup: the swap_hard_limit indicates
mem+swap_hard_limit.
* Change cgroup private apis to: virCgroupGet/SetMemSwapHardLimit
Signed-off-by: Nikunj A. Dadhania <nikunj@linux.vnet.ibm.com>
Since libvirt always passes /dev/net/tun to qemu via fd, we should
never trigger the cases where qemu tries to directly open the
device. Therefore, it is safer to deny the cgroup device ACL.
* src/qemu/qemu_cgroup.c (defaultDeviceACL): Remove /dev/net/tun.
* src/qemu/qemu.conf (cgroup_device_acl): Reflect this change.
Adding audit points showed that we were granting too much privilege
to qemu; it should not need any mknod rights to recreate any
devices. On the other hand, lxc should have all device privileges.
The solution is adding a flag parameter.
This also lets us restrict write access to read-only disks.
* src/util/cgroup.h (virCgroup*Device*): Adjust prototypes.
* src/util/cgroup.c (virCgroupAllowDevice)
(virCgroupAllowDeviceMajor, virCgroupAllowDevicePath)
(virCgroupDenyDevice, virCgroupDenyDeviceMajor)
(virCgroupDenyDevicePath): Add parameter.
* src/qemu/qemu_driver.c (qemudDomainSaveFlag): Update clients.
* src/lxc/lxc_controller.c (lxcSetContainerResources): Likewise.
* src/qemu/qemu_cgroup.c: Likewise.
(qemuSetupDiskPathAllow): Also, honor read-only disks.
Device names can be manipulated, so it is better to also log
the major/minor device number corresponding to the cgroup ACL
changes that libvirt made. This required some refactoring
of the relatively new qemu cgroup audit code.
Also, qemuSetupChardevCgroup was only auditing on failure, not success.
* src/qemu/qemu_audit.h (qemuDomainCgroupAudit): Delete.
(qemuAuditCgroup, qemuAuditCgroupMajor, qemuAuditCgroupPath): New
prototypes.
* src/qemu/qemu_audit.c (qemuDomainCgroupAudit): Rename...
(qemuAuditCgroup): ...and drop a parameter.
(qemuAuditCgroupMajor, qemuAuditCgroupPath): New functions, to
allow listing device major/minor in audit.
(qemuAuditGetRdev): New helper function.
* src/qemu/qemu_driver.c (qemudDomainSaveFlag): Adjust callers.
* src/qemu/qemu_cgroup.c (qemuSetupDiskPathAllow)
(qemuSetupHostUsbDeviceCgroup, qemuSetupCgroup)
(qemuTeardownDiskPathDeny): Likewise.
(qemuSetupChardevCgroup): Likewise, fixing missing audit.
When a SPICE or VNC graphics controller is present, and sound is
piggybacked over a channel to the graphics device rather than
directly accessing host hardware, then there is no need to grant
host hardware access to that qemu process.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Prevent sound with
spice, and with vnc when vnc_allow_host_audio is 0.
Reported by Daniel Berrange.
Regression in commit caa805ea let a lot of bad messages slip in.
* cfg.mk (msg_gen_function): Fix function name.
* src/qemu/qemu_cgroup.c (qemuRemoveCgroup): Fix fallout from
'make syntax-check'.
* src/qemu/qemu_driver.c (qemudDomainGetInfo)
(qemuDomainWaitForMigrationComplete, qemudStartVMDaemon)
(qemudDomainSaveFlag, qemudDomainAttachDevice)
(qemuDomainUpdateDeviceFlags): Likewise.
* src/qemu/qemu_hotplug.c (qemuDomainAttachHostUsbDevice)
(qemuDomainDetachPciDiskDevice, qemuDomainDetachSCSIDiskDevice):
Likewise.
The QEMU driver file is far too large. Move all the cgroup
helper code out into a separate file. No functional change.
* src/qemu/qemu_cgroup.c, src/qemu/qemu_cgroup.h,
src/Makefile.am: Add cgroup helper file
* src/qemu/qemu_driver.c: Delete cgroup code
