During his review of the iothreads pin setting code, Pavel noted that
there was a potential memory leak with respect to how the newVcpuPin
is handled and the goto endjob's in failure paths which would not free
the memory. For reference, See:
http://www.redhat.com/archives/libvir-list/2015-March/msg00415.html
The memory sizes in qemu are aligned up to 1 MiB boundaries. There are
two places where this was done once for the total size and then for
individual NUMA cell sizes.
Add a function that will align the sizes in one place so that it's clear
where the sizes are aligned.
As there are two possible approaches to define a domain's memory size -
one used with legacy, non-NUMA VMs configured in the <memory> element
and per-node based approach on NUMA machines - the user needs to make
sure that both are specified correctly in the NUMA case.
To avoid this burden on the user I'd like to replace the NUMA case with
automatic totaling of the memory size. To achieve this I need to replace
direct access to the virDomainMemtune's 'max_balloon' field with
two separate getters depending on the desired size.
The two sizes are needed as:
1) Startup memory size doesn't include memory modules in some
hypervisors.
2) After startup these count as the usable memory size.
Note that the comments for the functions are future aware and document
state that will be present after a few later patches.
Surprisingly we did not grab a VM job when a block job finished and we'd
happily rewrite the backing chain data. This made it possible to crash
libvirt when queueing two backing chains tightly and other badness.
To fix it, add yet another handler to the helper thread that handles
monitor events that require a job.
We interpret port values as signed int (convert them from char *),
so if a negative value is provided in network disk's configuration,
we accept it as valid, however there's an 'unknown cause' error raised later.
This error is only accidental because we return the port value in the return code.
This patch adds just a minor tweak to the already existing check so we
reject negative values the same way as we reject non-numerical strings.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1163553
A helper that never returns an error and treats bits out of bitmap range
as false.
Use it everywhere we use ignore_value on virBitmapGetBit, or loop over
the bitmap size.
Now that qemuDomainBlocksStatsGather provides functions of both
qemuMonitorGetBlockStatsParamsNumber and qemuMonitorGetBlockStatsInfo we
can reuse it and kill a lot of code.
Additionally as a bonus qemuDomainBlockStatsFlags will now support
summary statistics so add a statement to the virsh man page about that.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1142636
In the LXC driver, if the disk path is not provided the API returns
total statistics for all disks of the domain. With the new text monitor
implementation this can be now done in the qemu driver too.
Add code that wil total the stats for all disks if the path is not
provided.
Extract the code to look up the disk alias and return the block stats
struct so that it can be reused later in qemuDomainBlockStatsFlags.
The function uses qemuMonitorGetAllBlockStatsInfo instead of
qemuMonitorGetBlockStatsInfo.
Our virDomainBlockStatsFlags API uses the old approach where, when it's
called without the typed parameter array, returns the count of parameters
supported by qemu.
The supported parameter count is obtained via separate monitor calls
which is a waste since we can calculate it when gathering the data.
This patch adds code to the qemuMonitorGetAllBlockStatsInfo workers that
allows to track the count of supported fields reported by qemu and will
allow to remove the old duplicate code.
The function that is extracting block stats data from the QMP monitor
reply contains a lot of repeated code. Since I'd be changing each of the
copies in the next patch, lets convert it to a macro right away.
Add a different version of parser for "info blockstats" that basically
parses the same information as the existing copy of the function.
This will allow us to remove the single device version
qemuMonitorGetBlockStatsInfo in the future.
The new implementation uses few new helpers so it should be more
understandable and provides a test case to verify that it works.
Allocate the hash table in the monitor wrapper function instead of the
worker itself so that the text monitor impl that will be added in the
next patch doesn't have to duplicate it.
Error messages are already set in all code paths returning -1 from
networkGetNetworkAddress, so we don't want to overwrite them.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: John Ferlan <jferlan@redhat.com>
When creating qemu capabilities, a dummy virDomainObj is created just
because our monitor code expects that. However, the object is created
locked already. Then, under cleanup label, we simply unref the object
which results in whole domain object to be disposed. The object lock
is destroyed subsequently, but hey - it's still locked:
==24845== Thread #14's call to pthread_mutex_destroy failed
==24845== with error code 16 (EBUSY: Device or resource busy)
==24845== at 0x4C3024E: pthread_mutex_destroy (in /usr/lib64/valgrind/vgpreload_helgrind-amd64-linux.so)
==24845== by 0x531F72E: virMutexDestroy (virthread.c:83)
==24845== by 0x5302977: virObjectLockableDispose (virobject.c:237)
==24845== by 0x5302A89: virObjectUnref (virobject.c:265)
==24845== by 0x1DD37866: virQEMUCapsInitQMP (qemu_capabilities.c:3397)
==24845== by 0x1DD37CC6: virQEMUCapsNewForBinary (qemu_capabilities.c:3481)
==24845== by 0x1DD381E2: virQEMUCapsCacheLookup (qemu_capabilities.c:3609)
==24845== by 0x1DD30F8A: virQEMUCapsInitGuest (qemu_capabilities.c:744)
==24845== by 0x1DD31889: virQEMUCapsInit (qemu_capabilities.c:1020)
==24845== by 0x1DD7DD36: virQEMUDriverCreateCapabilities (qemu_conf.c:888)
==24845== by 0x1DDC57C0: qemuStateInitialize (qemu_driver.c:803)
==24845== by 0x53DC743: virStateInitialize (libvirt.c:777)
==24845==
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Commit 4bbe1029f fixed a problem in commit f7afeddc by moving the call
to virNetDevGetIndex() to a location common to all interface types (so
that the nicindex array would be filled in for macvtap as well as tap
interfaces), but the location was *too* common, as the original call
to virNetDevGetIndex() had been in a section qualified by "if
(cfg->privileged)". The result was that the "fixed" libvirtd would try
to call virNetDevGetIndex() even for session mode libvirtd, and end up
failing with the log message:
Unable to open control socket: Operation not permitted
To remedy that, this patch qualifies the call to virNetDevGetIndex()
in its new location with cfg->privileged.
This resolves https://bugzilla.redhat.com/show_bug.cgi?id=1198244
By adding a call and check of return of virBitmapToData to the
IOThreads code, my Coverity checker lets me know qemuDomainHelperGetVcpus
also needs to check the status...
Depending on the flags passed, either attempt to return the active/live
IOThread data for the domain or the config data.
The active/live path will call into the Monitor in order to get the
IOThread data and then correlate the thread_id's returned from the
monitor to the currently running system/threads in order to ascertain
the affinity for each iothread_id.
The config path will map each of the configured IOThreads and return
any configured iothreadspin data
Signed-off-by: John Ferlan <jferlan@redhat.com>
There was a mess in the way how we store unlimited value for memory
limits and how we handled values provided by user. Internally there
were two possible ways how to store unlimited value: as 0 value or as
VIR_DOMAIN_MEMORY_PARAM_UNLIMITED. Because we chose to store memory
limits as unsigned long long, we cannot use -1 to represent unlimited.
It's much easier for us to say that everything greater than
VIR_DOMAIN_MEMORY_PARAM_UNLIMITED means unlimited and leave 0 as valid
value despite that it makes no sense to set limit to 0.
Remove unnecessary function virCompareLimitUlong. The update of test
is to prevent the 0 to be miss-used as unlimited in future.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1146539
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Pass the TPM file descriptor to QEMU via command line.
Instead of passing /dev/tpm0 we now pass /dev/fdset/10 and the additional
parameters -add-fd set=10,fd=20.
This addresses the use case when QEMU is started with non-root privileges
and QEMU cannot open /dev/tpm0 for example.
Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
When the domain's source disk type is network, if source protocol is rbd
or sheepdog, the 'if().. break' will end the current case, which lead to
miss check the driver type is raw or qcow2. Libvirt will allow to create
internal snapshot for a running domain with raw format disk which based
on rbd storage.
While both protocols support internal snapshots of the disk qemu is not
able to use it as it requires some place to store the memory image. The
check if the disk is backed by a qcow2 image needs to be executed
always.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1179533
Signed-off-by: Shanzhi Yu <shyu@redhat.com>
Previously when a domain would get stuck in a domain job due to a
programming mistake we'd report the following control state:
$ virsh domcontrol domain
occupied (1424343406.150s)
The timestamp is invalid as the monitor was not entered for that domain.
We can use that to detect that the domain has an active job and report a
better error instead:
$ virsh domcontrol domain
error: internal (locking) error
https://bugzilla.redhat.com/show_bug.cgi?id=1197600
So, libvirt uses pid file to track pid of started qemus. Whenever
a domain is started, its pid is put into corresponding pid file.
The pid file path is generated based on domain name and stored
into domain object internals. However, it's not stored in the
status XML and therefore lost on daemon restarts. Hence, later,
when domain is being shut down, the daemon does not know which
pid file to unlink, and the correct pid file is left behind. To
avoid this, lets generate the pid file path again in
qemuProcessReconnect().
Reported-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Instead of checking defaultMode for every channel that has no mode
configured, test it only once outside of channel loop. This fixes a bug
that in case all possible channels are fore example set to insecure, but
defaultMode is set to secure, we wouldn't auto-generate TLS port. This
results in failure while starting a guest.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1143832
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
We have two different places that needs to be updated while touching
code for allocation spice ports. Add a bool option to
'qemuProcessSPICEAllocatePorts' function to switch between true and fake
allocation so we can use this function also in qemu_driver to generate
native domain definition.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Since adding the support for scheduler policy settings in commit
8680ea97, there are two enums with the same information. That was
caused by rewriting the patch since first draft.
Find out thanks to clang, but there was no impact whatsoever.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
The problem here was that when opening a channel, we were checking
whether the channel given is alias (can't be NULL for running domain) or
it's name, which can be NULL (for example with spicevmc). In case of
such domain qemuDomainOpenChannel() made the daemon crash.
STREQ_NULLABLE() is safe to use since the code in question is wrapped in
"if (name)" and is more readable, so use that instead of checking for
non-NULL "vm->def->channels[i]->target.name".
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1142631
This patch resolves a situation where the same "<target dev='$name'...>"
can be used for multiple disks in the domain.
While the $name is "mostly" advisory regarding the expected order that
the disk is added to the domain and not guaranteed to map to the device
name in the guest OS, it still should be unique enough such that other
domblk* type operations can be performed.
Without the patch, the domblklist will list the same Target twice:
$ virsh domblklist $dom
Target Source
------------------------------------------------
sda /var/lib/libvirt/images/file.qcow2
sda /var/lib/libvirt/images/file.img
Additionally, getting domblkstat, domblkerror, domblkinfo, and other block*
type calls will not be able to reference the second target.
Fortunately, hotplug disallows adding a "third" sda value:
$ qemu-img create -f raw /var/lib/libvirt/images/file2.img 10M
$ virsh attach-disk $dom /var/lib/libvirt/images/file2.img sda
error: Failed to attach disk
error: operation failed: target sda already exists
$
BUT, it since 'sdb' doesn't exist one would get the following on the same
hotplug attempt, but changing to use 'sdb' instead of 'sda'
$ virsh attach-disk $dom /var/lib/libvirt/images/file2.img sdb
error: Failed to attach disk
error: internal error: unable to execute QEMU command 'device_add': Duplicate ID 'scsi0-0-1' for device
$
Since we cannot fix this issue at parsing time, the best that can be done so
as to not "lose" a domain is to make the check prior to starting the guest
with the results as follows:
$ virsh start $dom
error: Failed to start domain $dom
error: XML error: target 'sda' duplicated for disk sources '/var/lib/libvirt/images/file.qcow2' and '/var/lib/libvirt/images/file.img'
$
Running 'make check' found a few more instances in the tests where this
duplicated target dev value was being used. These also exhibited some
duplicated 'id=' values (negating the uniqueness argument of aliases) in
the corresponding .args file and of course the *xmlout version of a few
input XML files.
NUMA enabled guest configuration explicitly specifies memory sizes for
individual nodes. Allowing the virDomainSetMemoryFlags API (and friends)
to change the total doesn't make sense as the individual node configs
are not updated in that case.
Forbid use of the API in case NUMA is specified.
If we combine the boot order on the command line with other
boot options, we prepend order= in front of it.
Instead of checking if the number of added arguments is between
0 and 2, separate the strings for boot order and options
and prepend boot order only if both strings are not empty.
Commit f7afeddc added code to report to systemd an array of interface
indexes for all tap devices used by a guest. Unfortunately it not only
didn't add code to report the ifindexes for macvtap interfaces
(interface type='direct') or the tap devices used by type='ethernet',
it ended up sending "-1" as the ifindex for each macvtap or hostdev
interface. This resulted in a failure to start any domain that had a
macvtap or hostdev interface (or actually any type other than
"network" or "bridge").
This patch does the following with the nicindexes array:
1) Modify qemuBuildInterfaceCommandLine() to only fill in the
nicindexes array if given a non-NULL pointer to an array (and modifies
the test jig calls to the function to send NULL). This is because
there are tests in the test suite that have type='ethernet' and still
have an ifname specified, but that device of course doesn't actually
exist on the test system, so attempts to call virNetDevGetIndex() will
fail.
2) Even then, only add an entry to the nicindexes array for
appropriate types, and to do so for all appropriate types ("network",
"bridge", and "direct"), but only if the ifname is known (since that
is required to call virNetDevGetIndex().
libvirt was unconditionally calling virNetDevBandwidthClear() for
every interface (and network bridge) of a type that supported
bandwidth, whether it actually had anything set or not. This doesn't
hurt anything (unless ifname == NULL!), but is wasteful.
This patch makes sure that all calls to virNetDevBandwidthClear() are
qualified by checking that the interface really had some bandwidth
setup done, and checks for a null ifname inside
virNetDevBandwidthClear(), silently returning success if it is null
(as well as removing the ATTRIBUTE_NONNULL from that function's
prototype, since we can't guarantee that it is never null,
e.g. sometimes a type='ethernet' interface has no ifname as it is
provided on the fly by qemu).
If the qemu binary on x86 does not support lsi SCSI controller,
but it supports virtio-scsi, we reject the virtio-specific attributes
for no reason.
Move the default controller assignment before the check.
https://bugzilla.redhat.com/show_bug.cgi?id=1168849
https://bugzilla.redhat.com/show_bug.cgi?id=1183869
Soo. you've successfully started yourself a domain. And since you want
to use it on your host exclusively you are confident enough to
passthrough the host CPU model, like this:
<cpu mode='host-passthrough'/>
Then, after a while, you want to save the domain into a file (e.g.
virsh save dom dom.save). And here comes the trouble. The file consist
of two parts: Libvirt header (containing domain XML among other
things), and qemu migration data. Now, the domain XML in the header is
formatted using special flags (VIR_DOMAIN_XML_SECURE |
VIR_DOMAIN_XML_UPDATE_CPU | VIR_DOMAIN_XML_INACTIVE |
VIR_DOMAIN_XML_MIGRATABLE).
Then, on your way back from the bar, you think of changing something
in the XML in the saved file (we have a command for it after all), say
listen address for graphics console. So you successfully type in the
command:
virsh save-image-edit dom.save
Change all the bits, and exit the editor. But instead of success
you're left with sad error message:
error: unsupported configuration: Target CPU model <null> does not
match source Pentium Pro
Sigh. Digging into the code you see lines, where we check for ABI
stability. The new XML you've produced is compared with the old one
from the saved file to see if qemu ABI will break or not. Wait, what?
We are using different flags to parse the XML you've provided so we
were just lucky it worked in some cases? Yep, that's right.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In commit cc41c648 I've re-factored qemuMonitorFindBalloonObjectPath, but
missed that there is a memory leak. The "nextpath" variable is
overwritten while looping in for cycle and we have to free it before next
cycle.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Making use of the ARCH_IS_S390 macro introduced with
e808357528
Signed-off-by: Stefan Zimmermann <stzi@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Since s390 does not support usb the default creation of a usb controller
for a domain should not occur.
Also adjust s390 test cases by removing usb device instances since
usb devices are no longer created by default for s390 the s390
test cases need to be adjusted.
Signed-off-by: Stefan Zimmermann <stzi@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
For historical reasons data regarding NUMA configuration were split
between the CPU definition and numatune. We cannot do anything about the
XML still being split, but we certainly can at least store the relevant
data in one place.
This patch moves the NUMA stuff to the right place.
As virDomainNumatuneSet now doesn't allocate the virDomainNuma object
any longer it's not necessary to pass the pointer to a pointer to store
the object as it will not change any longer.
While touching the parameter definitions I've also changed the name of
the parameter to "numa".
Name it virNumaMemAccess and add it to conf/numa_conf.[ch]
Note that to avoid a circular dependency the type of the NUMA cell
memAccess variable was changed to int. It will be turned back later
after the circular dependency will not exist.
Not all machine types support all devices, device properties, backends,
etc. So until we create a matrix of [machineType, qemuCaps], lets just
filter out some capabilities before we return them to the consumer
(which is going to make decisions based on them straight away).
Currently, as qemu is unable to tell which capabilities are (not)
enabled for given machine types, it's us who has to hardcode the matrix.
One day maybe the hardcoding will go away and we can create the matrix
dynamically on the fly based on a few monitor calls.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1179678
When migrating with storage, libvirt iterates over domain disks and
instruct qemu to migrate the ones we are interested in (shared, RO and
source-less disks are skipped). The disks are migrated in series. No
new disk is transferred until the previous one hasn't been quiesced.
This is checked on the qemu monitor via 'query-jobs' command. If the
disk has been quiesced, it practically went from copying its content
to mirroring state, where all disk writes are mirrored to the other
side of migration too. Having said that, there's one inherent error in
the design. The monitor command we use reports only active jobs. So if
the job fails for whatever reason, we will not see it anymore in the
command output. And this can happen fairly simply: just try to migrate
a domain with storage. If the storage migration fails (e.g. due to
ENOSPC on the destination) we resume the host on the destination and
let it run on partly copied disk.
The proper fix is what even the comment in the code says: listen for
qemu events instead of polling. If storage migration changes state an
event is emitted and we can act accordingly: either consider disk
copied and continue the process, or consider disk mangled and abort
the migration.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Upon BLOCK_JOB_COMPLETED event delivery, we check if the job has
completed (in qemuMonitorJSONHandleBlockJobImpl()). For better image,
the event looks something like this:
"timestamp": {"seconds": 1423582694, "microseconds": 372666}, "event":
"BLOCK_JOB_COMPLETED", "data": {"device": "drive-virtio-disk0", "len":
8412790784, "offset": 409993216, "speed": 8796093022207, "type":
"mirror", "error": "No space left on device"}}
If "len" does not equal "offset" it's considered an error, and we can
clearly see "error" field filled in. However, later in the event
processing this case was handled no differently to case of job being
aborted via separate API. It's time that we start differentiate these
two because of the future work.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Currently, upon BLOCK_JOB_* event, disk->mirrorState is not updated
each time. The callback code handling the events checks if a blockjob
was started via our public APIs prior to setting the mirrorState.
However, some block jobs may be started internally (e.g. during
storage migration), in which case we don't bother with setting
disk->mirror (there's nothing we can set it to anyway), or other
fields. But it will come handy if we update the mirrorState in these
cases too. The event wasn't delivered just for fun - we've started the
job after all.
So, in this commit, the mirrorState is set to whatever job status
we've obtained. Of course, there are some actions on some statuses
that we want to perform. But instead of if {} else if {} else {} ...
enumeration, let's move to switch().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If 'virNumaGetHostNodeset()' fails then the error path will try to free
uninitialized pointer mem_mask. Introduced by commit af2a1f058.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
PowerPC : Forbid NULL CPU model with 'host-model' mode in qemu command line.
This ensures that an XML such as following:
...
<cpu mode='host-model'>
<model fallback='allow'/>
</cpu>
...
will not generate a '-cpu host,compat=(null)' command line with qemu-system-ppc64.
Signed-off-by: Prerna Saxena <prerna@linux.vnet.ibm.com>
PowerPC : Explicitly associate 'qemu-system-ppc64' as the
default emulator for all 64-bit PowerPC guests ( both Big & Little Endian )
Signed-off-by: Prerna Saxena <prerna@linux.vnet.ibm.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1126762
Commit 43b67f introduced a deadlock issue when we use numatune
to change numa settings to a vm in session mode.
Jump to endjob instead of jump to cleanup.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
So, when building the '-numa' command line, the
qemuBuildMemoryBackendStr() function does quite a lot of checks to
chose the best backend, or to check if one is in fact needed. However,
it returned that backend is needed even for this little fella:
<numatune>
<memory mode="strict" nodeset="0,2"/>
</numatune>
This can be guaranteed via CGroups entirely, there's no need to use
memory-backend-ram to let qemu know where to get memory from. Well, as
long as there's no <memnode/> element, which explicitly requires the
backend. Long story short, we wouldn't have to care, as qemu works
either way. However, the problem is migration (as always). Previously,
libvirt would have started qemu with:
-numa node,memory=X
in this case and restricted memory placement in CGroups. Today, libvirt
creates more complicated command line:
-object memory-backend-ram,id=ram-node0,size=X
-numa node,memdev=ram-node0
Again, one wouldn't find anything wrong with these two approaches.
Both work just fine. Unless you try to migrated from the older libvirt
into the newer one. These two approaches are, unfortunately, not
compatible. My suggestion is, in order to allow users to migrate, lets
use the older approach for as long as the newer one is not needed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We do have a check for valid per-domain security model, however we still
do permit an invalid security model for a domain's device (those which
are specified with <source> element).
This patch introduces a new function virSecurityManagerCheckAllLabel
which compares user specified security model against currently
registered security drivers. That being said, it also permits 'none'
being specified as a device security model.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1165485
Signed-off-by: Ján Tomko <jtomko@redhat.com>
The qemuDomainHelperGetVcpus attempted to report an error when the
vcpupids info was NULL. Unfortunately earlier code would clamp the
value of 'maxinfo' to 0 when nvcpupids was 0, so the error reporting
would end up being skipped.
This lead to 'virsh vcpuinfo <dom>' just returning an empty list
instead of giving the user a clear error.
If a previous commit I fixed the incorrect handling of vcpu pids
for TCG mode QEMU:
commit b07f3d821d
Author: Daniel P. Berrange <berrange@redhat.com>
Date: Thu Dec 18 16:34:39 2014 +0000
Don't setup fake CPU pids for old QEMU
The code assumes that def->vcpus == nvcpupids, so when we setup
fake CPU pids for old QEMU with nvcpupids == 1, we cause the
later code to read off the end of the array. This has fun results
like sche_setaffinity(0, ...) which changes libvirtd's own CPU
affinity, or even better sched_setaffinity($RANDOM, ...) which
changes the affinity of a random OS process.
The intent was that this would merely disable the ability to set
per-vCPU affinity. It should still have been possible to set VM
level host CPU affinity.
Unfortunately, when you set <vcpu cpuset='0-1'>4</vcpu>, the XML
parser will internally take this & initialize an entry in the
def->cputune.vcpupin array for every VCPU. IOW this is implicitly
being treated as
<cputune>
<vcpupin cpuset='0-1' vcpu='0'/>
<vcpupin cpuset='0-1' vcpu='1'/>
<vcpupin cpuset='0-1' vcpu='2'/>
<vcpupin cpuset='0-1' vcpu='3'/>
</cputune>
Even more fun, the faked cputune elements are hidden from view when
querying the live XML, because their cpuset mask is the same as the
VM default cpumask.
The upshot was that it was impossible to set VM level CPU affinity.
To fix this we must update qemuProcessSetVcpuAffinities so that it
only reports a fatal error if the per-VCPU cpu mask is different
from the VM level cpu mask.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
In the event we're falling into the code that tries to create the file
in a forked environment (VIR_FILE_OPEN_FORK) we pass different mode bits,
but those are never set because the virFileOpenForceOwnerMode has a check
if the OPEN_FORCE_MODE bit is set before attempting to change the mode.
Since this is a special case it seems reasonable to set u+rw,g+rw,o
https://bugzilla.redhat.com/show_bug.cgi?id=1191355
When we attempt to migrate a vm with a migrateuri that has no scheme:
# virsh migrate test4 --live qemu+ssh://lhuang/system --migrateuri 127.0.0.1
target libvirtd will crash because uri->scheme is NULL in
qemuMigrationPrepareDirect on this line:
if (STRNEQ(uri->scheme, "tcp") &&
Add a value check before this line. Also fix a bug like this in
doNativeMigrate, that could only happen when destination libvirtd
returned an incorrect URI.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Export the required helpers and add backend code to hotplug RNG devices.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
As the RNG device is using an -object as backend refactor the code to
use the JSON to commandline generator so that we can reuse the code
later in hotplug.
Move the alias name right after the object type for rng-egd backend so
that we can later use the JSON to commandline generator to create the
command line.
Libvirt didn't prefix the random number generator backend object alias
with any string thus the device alias and object alias were identical.
To avoid possible problems, rename the alias for the backend object and
tweak tests to comply with the change.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Rename qemuBuildRNGDeviceArgs to qemuBuildRNGDevStr and change the
return type so that it can be reused in the device hotplug code later.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
This function is used to assign an alias for a RNG device. It will be
later reused when hotplugging RNGs.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
It is only usable for NETWORK and BRIDGE type interfaces.
Error out when trying to start a domain where the custom
tap device path is specified for interfaces of other types,
or when the daemon is not privileged.
Note that this cannot be checked at definition time, because
the comparison is against actual type.
https://bugzilla.redhat.com/show_bug.cgi?id=1147195
It is often helpful to know which version of libvirt and QEMU
was present when a guest was first launched. Ensure this info
is written into the QEMU log file for each guest.
Add the missing jump to the error label when the uuid in the
migration cookie XML does not match the uuid of the migrated
domain.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Some code paths have special logic depending on the page size
reported by sysconf, which in turn affects the test results.
We must mock this so tests always have a consistent page size.
Change done by commit f309db1f4d wrongly
assumes that qemu can start with a combination of NUMA nodes specified
with the "memdev" option and the appropriate backends, and the legacy
way by specifying only "mem" as a size argument. QEMU rejects such
commandline though:
$ /usr/bin/qemu-system-x86_64 -S -M pc -m 1024 -smp 2 \
-numa node,nodeid=0,cpus=0,mem=256 \
-object memory-backend-ram,id=ram-node1,size=12345 \
-numa node,nodeid=1,cpus=1,memdev=ram-node1
qemu-system-x86_64: -numa node,nodeid=1,cpus=1,memdev=ram-node1: qemu: memdev option must be specified for either all or no nodes
To fix this issue we need to check if any of the nodes requires the new
definition with the backend and if so, then all other nodes have to use
it too.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1182467
With the new JSON to argv formatter we are now able to represent the
memory backend definitions in the JSON object format that is reusable
for monitor use (hotplug) and then convert it into the shell string.
This will avoid having two separate instances of the same code that
would create the different formats.
Previous refactors now allow to make this step without changes to the
test suite.
QEMU's command line visitor as well as the JSON interface take bytes by
default for memory object sizes. Convert mebibytes to bytes so that we
can later refactor the existing code for hotplug purposes.
QEMU's qapi visitor code allows yes/on/y for true and no/off/n for false
value of boolean properities. Unify the used style so that we can
generate it later and fix test cases.
Extract the memory backend device code into a separate function so that
it can be later easily refactored and reused.
Few small changes for future reusability, namely:
- new (currently unused) parameter for user specified page size
- size of the memory is specified in kibibytes, divided up in the
function
- new (currently unused) parameter for user specifed source nodeset
- option to enforce capability check
Unlike -device, qemu uses a JSON object to add backend "objects" via the
monitor rather than the string that would be passed on the commandline.
To be able to reuse code parts that configure backends for various
devices, this patch adds a helper that will allow generating the command
line representations from the JSON property object.
This patch enables synchronization of the host macvtap
device options with the guest device's in response to the
NIC_RX_FILTER_CHANGED event.
The following device options will be synchronized:
* PROMISC
* MULTICAST
* ALLMULTI
Signed-off-by: Tony Krowiak <akrowiak@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1158034
If we're expecting to create a file somewhere and that fails for some
reason during qemuOpenFileAs, then we unlink the path we're attempting
to create leaving no way to determine what the "existing" privileges,
protections, or labels are that caused the failure (open, change owner
and group, change mode, etc.).
Furthermore, if we fall into the path where we'll be opening / creating
the file using VIR_FILE_OPEN_FORK, we need to first unlink/delete the file
we created in the first path; otherwise, the attempt by the child process
to open as some specific user:group may fail because the file was already
created using nfsnobody:nfsnobody. Again, if we didn't create the file we
don't want to blindly delete what already exists. Thus, a second reason for
the original check to set need_unlink to false when we find the file with
CREAT set, but already existing.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Commit id '540c339a' to fix issues with reference counting and transient
domains moved the qemuDomainObjEndAsyncJob call prior to the attempt to
restart the guest CPU's resulting in an error:
error: Failed to save domain rhel70 to /tmp/pl/rhel70.save
error: internal error: unexpected async job 3
when (ret != 0) - eg, the error path from qemuDomainSaveMemory.
This patch will adjust the logic to call the EndAsyncJob only after
we've tried to restart the guest CPUs. It also needs to adjust the
test for qemuDomainRemoveInactive to add the ret == 0 condition.
Additionally, if we get to endjob: because of some error earlier, then
we need to save that error in the event the CPU restart logic fails.
We don't want to return the error from CPU restart failure, rather we
want to return the error from the failed save that caused us to fall
into the retry to start the CPU logic.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Well, even though users can pass the list of UEFI:NVRAM pairs at the
configure time, we may maintain the list of widely available UEFI
ourselves too. And as arm64 begin to rises, OVMF was ported there too.
With a slight name change - it's called AAVMF, with AAVMF_CODE.fd
being the UEFI firmware and AAVMF_VARS.fd being the NVRAM store file.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Up until now there are just two ways how to specify UEFI paths to
libvirt. The first one is editing qemu.conf, the other is editing
qemu_conf.c and recompile which is not that fancy. So, new
configure option is introduced: --with-loader-nvram which takes a
list of pairs of UEFI firmware and NVRAM store. This way, the
compiled in defaults can be passed during compile time without
need to change the code itself.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1183890
When we try to update a xml to a image file, we will clear the
graphics passwd settings, because we do not pass VIR_DOMAIN_XML_SECURE
to qemuDomainDefCopy, qemuDomainDefFormatBuf won't format the passwd.
Add VIR_DOMAIN_XML_SECURE flag when we call qemuDomainDefCopy
in qemuDomainSaveImageUpdateDef.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1161024
This way the device is in vmdef only if ret = 0 and the caller
(qemuDomainAttachDeviceFlags) does not free it.
Otherwise it might get double freed by qemuProcessStop
and qemuDomainAttachDeviceFlags if the domain crashed
in monitor after we've added it to vm->def.
Do the allocation first, then add the actual device.
The second part should never fail. This is good
for live hotplug where we don't want to remove the device
on OOM after the monitor command succeeded.
The only change in behavior is that on failure, the
vmdef->consoles array is freed, not just the first console.
For stateless, client side drivers, it is never correct to
probe for secondary drivers. It is only ever appropriate to
use the secondary driver that is associated with the
hypervisor in question. As a result the ESX & HyperV drivers
have both been forced to do hacks where they register no-op
drivers for the ones they don't implement.
For stateful, server side drivers, we always just want to
use the same built-in shared driver. The exception is
virtualbox which is really a stateless driver and so wants
to use its own server side secondary drivers. To deal with
this virtualbox has to be built as 3 separate loadable
modules to allow registration to work in the right order.
This can all be simplified by introducing a new struct
recording the precise set of secondary drivers each
hypervisor driver wants
struct _virConnectDriver {
virHypervisorDriverPtr hypervisorDriver;
virInterfaceDriverPtr interfaceDriver;
virNetworkDriverPtr networkDriver;
virNodeDeviceDriverPtr nodeDeviceDriver;
virNWFilterDriverPtr nwfilterDriver;
virSecretDriverPtr secretDriver;
virStorageDriverPtr storageDriver;
};
Instead of registering the hypervisor driver, we now
just register a virConnectDriver instead. This allows
us to remove all probing of secondary drivers. Once we
have chosen the primary driver, we immediately know the
correct secondary drivers to use.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The code modifies the domain configuration but doesn't take a MODIFY
type job to do so.
This patch also fixes a few very long lines of code around the touched
parts.
For distros that want to add versioned machine types, they will add
(downstream) machine types like "virt-foo-1.2.3". Detect these as
MMIO too.
Signed-off-by: Richard W.M. Jones <rjones@redhat.com>
Previous patch of this series fixed the issue with adding a new PCI bridge
when all the slots were reserved by devices with user specified addresses.
In case there are still some PCI devices waiting to get a slot reserved
by qemuAssignDevicePCISlots, this means a new bus needs to be
created along with a corresponding bridge controller. By adding an
additional check, this scenario now results in a reasonable error
instead of generating wrong qemu command line.
Commit 93c8ca tried to fix the issue with auto-adding of a PCI bridge
controller, but didn't work properly in all scenarios.
This patch provides a better fix of the issue when all slots on a PCI bus
are reserved by devices with user specified addresses and no additional
bridges need to be created.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1132900
In order to be able to test for fully reserved PCI buses, assignment of
PCI slots for integrated devices needs to be moved to a separate function.
This also might be a good preparation if we decide to add support for
other chipsets as well.
Move qemuDomainAssignPCIAddresses after the definition
of the static function qemuDomainValidateDevicePCISlotsQ35.
This lets us define a new static function using
qemuDomainValidateDevicePCISlots* and use it in
qemuDomainAssignPCIAddresses without a forward declaration.
Signed-off-by: Ján Tomko <jtomko@redhat.com>
As it turned out, fix of dead code 419a22 changed the affected condition
from "never true" to "always true", so better fix would be to change the
return code of virDomainMaybeAddController from 0 to 1 if
a new bridge has been added, thus distinguishing case when we didn't need to
add any controller and case we successfully added one.
The return code is changed in the next commit
The ACL check didn't check the VIR_DOMAIN_XML_SECURE flag and the
appropriate permission for it. Found via code inspection while fixing
permissions for save images.
https://bugzilla.redhat.com/show_bug.cgi?id=1164627
When using 'virsh attach-device' to hotplug an unsupported console type
into a qemu guest the attachment would succeed as the command line
formatter didn't report error in such case.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1130390
The listen address is not mandatory for <interface type='server'>
but when it's not specified, we've been formatting it as:
-netdev socket,listen=(null):5558,id=hostnet0
which failed with:
Device 'socket' could not be initialized
Omit the address completely and only format the port in the listen
attribute.
Also fix the schema to allow specifying a model.
Using the same driver multiple times is pointless and
it can result in confusing errors:
$ virsh start test
error: Failed to start domain test
error: internal error: security label already defined for VM
https://bugzilla.redhat.com/show_bug.cgi?id=1153891
Depending on the context, either error out if the domain
has disappeared in the meantime, or just ignore the value
to allow marking the function as ATTRIBUTE_RETURN_CHECK.
https://bugzilla.redhat.com/show_bug.cgi?id=1161024
If the domain crashed while we were in monitor,
we cannot rely on the REALLOC done on live definition,
since vm->def now points to the persistent definition.
Skip adding the attached devices to domain definition
if the domain crashed.
In AttachChrDevice, the chardev was already added to the
live definition and freed by qemuProcessStop in the case
of a crash. Skip the device removal in that case.
Also skip audit if the domain crashed in the meantime.
https://bugzilla.redhat.com/show_bug.cgi?id=1161024
In the device type-specific functions, exit early
if the domain has disappeared, because the cleanup
should have been done by qemuProcessStop.
Check the return value in processDeviceDeletedEvent
and qemuProcessUpdateDevices.
Skip audit and removing the device from live def because
it has already been cleaned up.
Ploop is a pseudo device which makeit possible to access
to an image in a file as a block device. Like loop devices,
but with additional features, like snapshots, write tracker
and without double-caching.
It used in PCS for containers and in OpenVZ. You can manage
ploop devices and images with ploop utility
(http://git.openvz.org/?p=ploop).
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
We tested for positive return value from virDomainMaybeAddController,
but it returns 0 or -1 only resulting in a dead code.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In case we find out, there are more PCI devices to be connected
than there are available slots on the default PCI bus, we automatically add a
new bus and a related PCI bridge controller as well. As there are no free slots
left on the default PCI bus, PCI bridge controller gets a free slot on a
newly created PCI bus which causes qemu to refuse to start the guest.
This fix introduces a new function qemuDomainPCIBusFullyReserved which
is checked right before we possibly try to reserve a slot for PCI bridge
controller.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1132900
The virDomainDefineXMLFlags and virDomainCreateXML APIs both
gain new flags allowing them to be told to validate XML.
This updates all the drivers to turn on validation in the
XML parser when the flags are set
systemd-machined introduced a new method CreateMachineWithNetwork
that obsoletes CreateMachine. It expects to be given a list of
VETH/TAP device indexes for the host side device(s) associated
with a container/machine.
This falls back to the old CreateMachine method when the new
one is not supported.
https://bugzilla.redhat.com/show_bug.cgi?id=1181182
When we meet error in qemuMigrationPrepareAny and goto
cleanup with rc < 0, we forget clear the priv->origname and this
will make this vm migrate fail next time because leave a wrong
origname in priv, and will Generate a wrong cookie when do
migrate next time.
This patch will make priv->origname is NULL when migrate fail
in target host.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Make local copy of the disk alias in qemuProcessInitPasswords,
instead of referencing the one in domain definition, which
might get freed if the domain crashes while we're in monitor.
Also copy the memballoon period value.
Make a local copy of the disk alias instead of pointing
to the domain definition, which might get freed if
the domain dies while we're in monitor.
Also exit early if that happens.
Exit the monitor right after we've done with it to get
the virDomainObjPtr lock back, otherwise we might be accessing
vm->def while it's being cleaned up by qemuProcessStop.
If the domain crashed while we were in the monitor, exit
early instead of changing vm->def which is now the persistent
definition.
The domain might disappear during the time in monitor when
the virDomainObjPtr is unlocked, so the caller needs to check
if it's still alive.
Since most of the callers are going to need it, put the
check inside qemuDomainObjExitMonitor and return -1 if
the domain died in the meantime.
QEMU internally updates the size of video memory if the domain XML had
provided too low memory size or there are some dependencies for a QXL
devices 'vgamem' and 'ram' size. We need to know about the changes and
store them into the status XML to not break migration or managedsave
through different libvirt versions.
The values would be loaded only if the "vgamem_mb" property exists for
the device. The presence of the "vgamem_mb" also tells that the
"ram_size" and "vram_size" exists for QXL devices.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The search is done recursively only through QOM object that has a type
prefixed with "child<" as this indicate that the QOM is a parent for
other QOM objects.
The usage is that you give known device name with starting path where to
search.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Commit e3435caf fixed hot-plugging of vcpus with strict memory pinning
on NUMA hosts, but unfortunately it also broke updating number of vcpus
for offline guests using our API.
The issue is that we try to create a cpu cgroup for non-running guest
which fails as there are no cgroups for that domain. We should create
cgroups and update cpuset.mems only if we are hot-plugging.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1165993
So, there are still plenty of vNIC types that we don't know how to set
bandwidth on. Let's warn explicitly in case user has requested it
instead of pretending everything was set.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When create inactive external snapshot, after update disk definitions,
virDomainSaveConfig is needed, if not after restart libvirtd the new snapshot
file definitions in xml will be lost.
Reproduce steps:
1. prepare a shut off guest
$ virsh domstate rhel7 && virsh domblklist rhel7
shut off
Target Source
------------------------------------------------
vda /var/lib/libvirt/images/rhel7.img
2. create external disk snapshot
$ virsh snapshot-create rhel7 --disk-only && virsh domblklist rhel7
Domain snapshot 1417882967 created
Target Source
------------------------------------------------
vda /var/lib/libvirt/images/rhel7.1417882967
3. restart libvirtd then check guest source file
$ service libvirtd restart && virsh domblklist rhel7
Redirecting to /bin/systemctl restart libvirtd.service
Target Source
------------------------------------------------
vda /var/lib/libvirt/images/rhel7.img
This was first reported by Eric Blake
http://www.redhat.com/archives/libvir-list/2014-December/msg00369.html
Signed-off-by: Shanzhi Yu <shyu@redhat.com>
The virDomainDefParse* and virDomainDefFormat* methods both
accept the VIR_DOMAIN_XML_* flags defined in the public API,
along with a set of other VIR_DOMAIN_XML_INTERNAL_* flags
defined in domain_conf.c.
This is seriously confusing & error prone for a number of
reasons:
- VIR_DOMAIN_XML_SECURE, VIR_DOMAIN_XML_MIGRATABLE and
VIR_DOMAIN_XML_UPDATE_CPU are only relevant for the
formatting operation
- Some of the VIR_DOMAIN_XML_INTERNAL_* flags only apply
to parse or to format, but not both.
This patch cleanly separates out the flags. There are two
distint VIR_DOMAIN_DEF_PARSE_* and VIR_DOMAIN_DEF_FORMAT_*
flags that are used by the corresponding methods. The
VIR_DOMAIN_XML_* flags received via public API calls must
be converted to the VIR_DOMAIN_DEF_FORMAT_* flags where
needed.
The various calls to virDomainDefParse which hardcoded the
use of the VIR_DOMAIN_XML_INACTIVE flag change to use the
VIR_DOMAIN_DEF_PARSE_INACTIVE flag.
https://bugzilla.redhat.com/show_bug.cgi?id=1135339 documents some
confusing behavior when a user tries to start an inactive block
commit in a second connection while there is already an on-going
active commit from a first connection. Eventually, qemu will
support multiple simultaneous block jobs, but as of now, it does
not; furthermore, libvirt also needs an overhaul before we can
support simultaneous jobs. So, the best way to avoid confusing
ourselves is to quit relying on qemu to tell us about the situation
(where we risk getting in weird states) and instead forbid a
duplicate block commit ourselves.
Note that we are still relying on qemu to diagnose attempts to
interrupt an inactive commit (since we only track XML of an active
commit), but as inactive commit is less confusing for libvirt to
manage, there is less that can go wrong by leaving that detection
up to qemu.
* src/qemu/qemu_driver.c (qemuDomainBlockCommit): Hoist check for
active commit to occur earlier outside of conditions.
Signed-off-by: Eric Blake <eblake@redhat.com>
QEMU supports feature specification with -cpu host and we just skip
using that. Since QEMU developers themselves would like to use this
feature, this patch modifies the code to work.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1178850
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
The default value should be 16 MiB instead of 8 MiB. Only really old
version of upstream QEMU used the 8 MiB as default for vga framebuffer.
Without this change if you update your libvirt where we introduced the
"vgamem" attribute for QXL video device the value will be set to 8 MiB,
but previously your guest had 16 MiB because we didn't pass any value to
QEMU command line which means QEMU used its own 16 MiB as default.
This will affect all users with guest's display resolution higher than
1920x1080.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
In one of my previous commits (311b4a67) I've tried to allow to
pass regular system pages to <hugepages>. However, there was a
little bug that wasn't caught. If domain has guest NUMA topology
defined, qemuBuildNumaArgStr() function takes care of generating
corresponding command line. The hugepages backing for guest NUMA
nodes is handled there too. And here comes the bug: the hugepages
setting from XML is stored in KiB internally, however, the system
pages size was queried and stored in Bytes. So the check whether
these two are equal was failing even if it shouldn't.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In commit 540c339a25 the whole domain
reference counting was refactored in the qemu driver. Domain jobs now
don't need to reference the domain object as they now expect the
reference from the calling function.
However, the patch forgot to remove the unref call in case we exit the
monitor when we were acquiring a nested job. This caused the daemon to
crash on a subsequent access to the domain object once we've done an
operation requiring a nested job for a monitor access.
An easy reproducer case:
1) Start a vm with qcow disks
2) virsh snapshot-create-as DOMNAME
3) virsh dumpxml DOMNAME
4) daemon crashes in a semi-random spot while accessing a now-removed VM
object.
Fortunately, the commit wasn't released yet, so there are no security
implications.
Reported-by: Shanzi Yu <shyu@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1177723
When setting new bandwidth limits via
virDomainSetInterfaceParameters, the old ones are cleared first.
However, if setting the new ones fails, the old are already gone
and interface is left in inconsistent state. Therefore, right
before failing we ought to try to restore the old bandwidth.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1178652
We will get a warning when we have a guest in paused
status (caused by kernel panic) and restart libvirtd,
warning message like this:
Qemu reported unknown VM status: 'guest-panicked'
and this seems because we set a wrong status name in
qemu_monitor.c, and from qemu qapi-schema.json file
we know this status should named 'guest-panicked'.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Add the possibility to have more than one IP address configured for a
domain network interface. IP addresses can also have a prefix to define
the corresponding netmask.
There is one problem that causes various errors in the daemon. When
domain is waiting for a job, it is unlocked while waiting on the
condition. However, if that domain is for example transient and being
removed in another API (e.g. cancelling incoming migration), it get's
unref'd. If the first call, that was waiting, fails to get the job, it
unref's the domain object, and because it was the last reference, it
causes clearing of the whole domain object. However, when finishing the
call, the domain must be unlocked, but there is no way for the API to
know whether it was cleaned or not (unless there is some ugly temporary
variable, but let's scratch that).
The root cause is that our APIs don't ref the objects they are using and
all use the implicit reference that the object has when it is in the
domain list. That reference can be removed when the API is waiting for
a job. And because each domain doesn't do its ref'ing, it results in
the ugly checking of the return value of virObjectUnref() that we have
everywhere.
This patch changes qemuDomObjFromDomain() to ref the domain (using
virDomainObjListFindByUUIDRef()) and adds qemuDomObjEndAPI() which
should be the only function in which the return value of
virObjectUnref() is checked. This makes all reference counting
deterministic and makes the code a bit clearer.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Although QMP returns info about vCPU threads in TCG mode, the
data it returns is mostly lies. Only the first vCPU has a valid
thread_id returned. The thread_id given for the other vCPUs is
in fact the main emulator thread. All vCPUs actually run under
the same thread in TCG mode.
Our vCPU pinning code is not at all able to cope with this
so if you try to set CPU affinity per-vCPU you end up with
wierd errors
error: Failed to start domain instance-00000007
error: cannot set CPU affinity on process 24365: Invalid argument
Since few people will care about the performance of TCG with
strict CPU pinning, lets just disable that for now, so we get
a clear error message
error: Failed to start domain instance-00000007
error: Requested operation is not valid: cpu affinity is not supported
The code assumes that def->vcpus == nvcpupids, so when we setup
fake CPU pids for old QEMU with nvcpupids == 1, we cause the
later code to read off the end of the array. This has fun results
like sche_setaffinity(0, ...) which changes libvirtd's own CPU
affinity, or even better sched_setaffinity($RANDOM, ...) which
changes the affinity of a random OS process.
Libvirt BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1175397
QEMU BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1170093
In qemu there are two interesting arguments:
1) -numa to create a guest NUMA node
2) -object memory-backend-{ram,file} to tell qemu which memory
region on which host's NUMA node it should allocate the guest
memory from.
Combining these two together we can instruct qemu to create a
guest NUMA node that is tied to a host NUMA node. And it works
just fine. However, depending on machine type used, there might
be some issued during migration when OVMF is enabled (see QEMU
BZ). While this truly is a QEMU bug, we can help avoiding it. The
problem lies within the memory backend objects somewhere. Having
said that, fix on our side consists on putting those objects on
the command line if and only if needed. For instance, while
previously we would construct this (in all ways correct) command
line:
-object memory-backend-ram,size=256M,id=ram-node0 \
-numa node,nodeid=0,cpus=0,memdev=ram-node0
now we create just:
-numa node,nodeid=0,cpus=0,mem=256
because the backend object is obviously not tied to any specific
host NUMA node.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Coverity flagged commit 0282ca45 as introducing a memory leak;
in all my refactoring to make capacity probing conditional on
whether the image is non-raw, I missed deleting the unconditional
probe.
* src/qemu/qemu_driver.c (qemuStorageLimitsRefresh): Drop
redundant assignment.
Signed-off-by: Eric Blake <eblake@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1174569
There's nothing we need to do for shared iSCSI devices in
qemuAddSharedHostdev and qemuRemoveSharedHostdev. The iSCSI layer
takes care about that for us.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Wire up backing chain recursion. For the first time, it is now
possible to get libvirt to expose that qemu tracks read statistics
on backing files, as well as report maximum extent written on a
backing file during a block-commit operation.
For a running domain, where one of the two images has a backing
file, I see the traditional output:
$ virsh domstats --block testvm2
Domain: 'testvm2'
block.count=2
block.0.name=vda
block.0.path=/tmp/wrapper.qcow2
block.0.rd.reqs=1
block.0.rd.bytes=512
block.0.rd.times=28858
block.0.wr.reqs=0
block.0.wr.bytes=0
block.0.wr.times=0
block.0.fl.reqs=0
block.0.fl.times=0
block.0.allocation=0
block.0.capacity=1310720000
block.0.physical=200704
block.1.name=vdb
block.1.path=/dev/sda7
block.1.rd.reqs=0
block.1.rd.bytes=0
block.1.rd.times=0
block.1.wr.reqs=0
block.1.wr.bytes=0
block.1.wr.times=0
block.1.fl.reqs=0
block.1.fl.times=0
block.1.allocation=0
block.1.capacity=1310720000
vs. the new output:
$ virsh domstats --block --backing testvm2
Domain: 'testvm2'
block.count=3
block.0.name=vda
block.0.path=/tmp/wrapper.qcow2
block.0.rd.reqs=1
block.0.rd.bytes=512
block.0.rd.times=28858
block.0.wr.reqs=0
block.0.wr.bytes=0
block.0.wr.times=0
block.0.fl.reqs=0
block.0.fl.times=0
block.0.allocation=0
block.0.capacity=1310720000
block.0.physical=200704
block.1.name=vda
block.1.path=/dev/sda6
block.1.backingIndex=1
block.1.rd.reqs=0
block.1.rd.bytes=0
block.1.rd.times=0
block.1.wr.reqs=0
block.1.wr.bytes=0
block.1.wr.times=0
block.1.fl.reqs=0
block.1.fl.times=0
block.1.allocation=327680
block.1.capacity=786432000
block.2.name=vdb
block.2.path=/dev/sda7
block.2.rd.reqs=0
block.2.rd.bytes=0
block.2.rd.times=0
block.2.wr.reqs=0
block.2.wr.bytes=0
block.2.wr.times=0
block.2.fl.reqs=0
block.2.fl.times=0
block.2.allocation=0
block.2.capacity=1310720000
I may later do a patch that trims the output to avoid 0 stats,
particularly for backing files (which are more likely to have
0 stats, at least for write statistics when no block-commit
is performed). Also, I still plan to expose physical size
information (qemu doesn't expose it yet, so it requires a stat,
and for block devices, a further open/seek operation). But
this patch is good enough without worrying about that yet.
* src/qemu/qemu_driver.c (QEMU_DOMAIN_STATS_BACKING): New internal
enum bit.
(qemuConnectGetAllDomainStats): Recognize new user flag, and pass
details to...
(qemuDomainGetStatsBlock): ...here, where we can do longer recursion.
(qemuDomainGetStatsOneBlock): Output new field.
Signed-off-by: Eric Blake <eblake@redhat.com>
In order to report stats on backing chains, we need to separate
the output of stats for one block from how we traverse blocks.
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Split...
(qemuDomainGetStatsOneBlock): ...into new helper.
Signed-off-by: Eric Blake <eblake@redhat.com>
A coming patch will make it optionally possible to list backing
chain block stats; in this mode of operation, block.counts is no
longer the number of <disks> in the domain, but the number of
blocks in the array being reported. We still want block.count
listed first, but rather than iterate the tree twice (once to
count, and once to list stats), it's easier to just touch things
up after the fact.
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Compute count
after the fact.
Signed-off-by: Eric Blake <eblake@redhat.com>
The prior refactoring can now be put to use. With the same domain
as the earlier commit 7b49926 (one qcow2 disk and an empty
cdrom drive):
$ virsh domstats --block foo
Domain: 'foo'
block.count=2
block.0.name=hda
block.0.path=/var/lib/libvirt/images/foo.qcow2
block.0.allocation=1309614080
block.0.capacity=42949672960
block.0.physical=1309671424
block.1.name=hdc
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Use
qemuStorageLimitsRefresh to report offline statistics.
Signed-off-by: Eric Blake <eblake@redhat.com>
Create a helper function that can be reused for gathering block
info from virDomainListGetStats.
* src/qemu/qemu_driver.c (qemuDomainGetBlockInfo): Split guts...
(qemuStorageLimitsRefresh): ...into new helper function.
Signed-off-by: Eric Blake <eblake@redhat.com>
The documentation for virDomainBlockInfo was confusing: it stated
that 'physical' was the size of the container, then gave an example
of it being the amount of storage used by a sparse file (that is,
for a sparse raw image on a regular file, the wording implied
capacity==physical, while allocation was smaller; but the example
instead claimed physical==allocation). Since we use 'physical' for
the last offset of a block device, we should do likewise for
regular files.
Furthermore, the example claimed that for a qcow2 regular file,
allocation==physical. At the time the code was first written,
this was true (qcow2 files were allocated sequentially, and were
never sparse, so the last sector written happened to also match
the disk space occupied); but modern qemu does much better and
can punch holes for a qcow2 with allocation < physical.
Basically, after this patch, the three fields are now reliably
mapped as:
'capacity' - how much storage the guest can see (equal to
physical for raw images, determined by image metadata otherwise)
'allocation' - how much storage the image occupies (similar to
what 'du' would report)
'physical' - the last offset of the image (similar to what 'ls'
would report)
'capacity' can be larger than 'physical' (such as for a qcow2
image that does not vary much from a backing file) or smaller
(such as for a qcow2 file with lots of internal snapshots).
Likewise, 'allocation' can be (slightly) larger than 'physical'
(such as counting the tail of cluster allocations required to
round a file size up to filesystem granularity) or smaller
(for a sparse file). A block-resize operation changes capacity
(which, for raw images, also changes physical); many non-raw
images automatically grow physical and allocation as necessary
when starting with an allocation smaller than capacity; and even
when capacity and physical stay unchanged, allocation can change
when converting sectors from holes to data or back.
Note that this does not change semantics for qcow2 images stored
on block devices; there, we still rely on qemu to report the
highest written extent for allocation. So using this API to
track when to extend a block device because a qcow2 image is
about to exceed a threshold will not see any changes.
Also, note that virStorageVolInfo is unfortunately limited to
just 'capacity' and 'allocation' (we can't expand it to add
'physical', although we can expand the XML to add it there);
historically, that struct's 'allocation' value has reported
file size for qcow2 files (what this patch terms 'physical'
for a domain block device), but disk usage for raw files (what
this patch terms 'allocation'). So follow-up patches will be
needed to make storage volumes report the same allocation
values and get at physical values, where those differ.
* include/libvirt/libvirt-domain.h (_virDomainBlockInfo): Tweak
documentation to match saner definition.
* src/qemu/qemu_driver.c (qemuDomainGetBlockInfo): For regular
files, physical size is capacity, not allocation.
Signed-off-by: Eric Blake <eblake@redhat.com>
Ultimately, we want to avoid read()ing a file while qemu is running.
We still have to open() block devices to determine their physical
size, but that is safer. This patch rearranges code to group
together all code that reads the image, to make it easier for later
patches to skip the metadata collection when possible.
* src/qemu/qemu_driver.c (qemuDomainGetBlockInfo): Check for empty
disk up front. Place metadata reading next to use.
Signed-off-by: Eric Blake <eblake@redhat.com>
When requested in a later patch, the QMP command results are now
examined recursively. As qemu_driver will eventually have to
read items out of the hash table as stored by this patch, the
computation of backing alias string is done in a shared location.
* src/qemu/qemu_domain.h (qemuDomainStorageAlias): New prototype.
* src/qemu/qemu_domain.c (qemuDomainStorageAlias): Implement it.
* src/qemu/qemu_monitor_json.c
(qemuMonitorJSONGetOneBlockStatsInfo)
(qemuMonitorJSONBlockStatsUpdateCapacityOne): Perform recursion.
(qemuMonitorJSONGetAllBlockStatsInfo)
(qemuMonitorJSONBlockStatsUpdateCapacity): Update callers.
Signed-off-by: Eric Blake <eblake@redhat.com>
A future patch will allow recursion into backing chains when
collecting block stats. This patch should not change behavior,
but merely moves out the common code that will be reused once
recursion is enabled, and adds the parameter that will turn on
recursion.
* src/qemu/qemu_monitor.h (qemuMonitorGetAllBlockStatsInfo)
(qemuMonitorBlockStatsUpdateCapacity): Add recursion parameter,
although it is ignored for now.
* src/qemu/qemu_monitor.h (qemuMonitorGetAllBlockStatsInfo)
(qemuMonitorBlockStatsUpdateCapacity): Likewise.
* src/qemu/qemu_monitor_json.h
(qemuMonitorJSONGetAllBlockStatsInfo)
(qemuMonitorJSONBlockStatsUpdateCapacity): Likewise.
* src/qemu/qemu_monitor_json.c
(qemuMonitorJSONGetAllBlockStatsInfo)
(qemuMonitorJSONBlockStatsUpdateCapacity): Add parameter, and
split...
(qemuMonitorJSONGetOneBlockStatsInfo)
(qemuMonitorJSONBlockStatsUpdateCapacityOne): ...into helpers.
(qemuMonitorJSONGetBlockStatsInfo): Update caller.
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Update caller.
* src/qemu/qemu_migration.c (qemuMigrationCookieAddNBD): Likewise.
Signed-off-by: Eric Blake <eblake@redhat.com>
Right now, grabbing blockinfo always calls stat on the disk, then
opens the image to determine the capacity, using a throw-away
virStorageSourcePtr. This has a couple of drawbacks:
1. We are calling stat and opening a file on every invocation of
the API. However, there are cases where the stats should NOT be
changing between successive calls (if a domain is running, no
one should be changing the physical size of a block device or raw
image behind our backs; capacity of read-only files should not
be changing; and we are the gateway to the block-resize command
to know when the capacity of read-write files should be changing).
True, we still have to use stat in some cases (a sparse raw file
changes allocation if it is read-write and the amount of holes is
changing, and a read-write qcow2 image stored in a file changes
physical size if it was not fully pre-allocated). But for
read-only images, even this should be something we can remember
from the previous time, rather than repeating every call.
2. We want to enhance the power of virDomainListGetStats, by
sharing code. But we already have a virStorageSourcePtr for
each disk, and it would be easier to reuse the common structure
than to have to worry about the one-off virDomainBlockInfoPtr.
While this patch does not optimize reuse of information in point
1, it does get us closer to being able to do so; by updating a
structure that survives between consecutive calls.
* src/util/virstoragefile.h (_virStorageSource): Add physical, to
mirror virDomainBlockInfo; rearrange fields to match public struct.
(virStorageSourceCopy): Copy the new field.
* src/qemu/qemu_driver.c (qemuDomainGetBlockInfo): Store into
storage source, then copy to block info.
Signed-off-by: Eric Blake <eblake@redhat.com>
In order for a future patch to virDomainListGetStats to reuse
some code for determining disk usage of offline domains, we
need to make it easier to pull out part of the guts of grabbing
blockinfo. The current implementation grabs a job fairly late
in the game, while getstats will already own a job; reordering
things so that the job is always grabbed up front in both
functions will make it easier to pull out the common code.
This patch results in grabbing a job in cases where one was not
previously needed, but as it is a query job, it should not be
noticeably slower.
This patch touches the same code as the fix for CVE-2014-6458
(commit b799259); in that patch, we avoided hotplug changing
a disk reference during the time of obtaining a monitor lock
by copying all data we needed and no longer referencing disk;
this patch goes the other way and ensures that by holding the
job, the disk cannot be changed so we no longer need to worry
about the disk being invalidated across the monitor lock.
* src/qemu/qemu_driver.c (qemuDomainGetBlockInfo): Rearrange job
control to be outside of disk information.
Signed-off-by: Eric Blake <eblake@redhat.com>
Commit e3435caf added cleanup code to qemuDomainSetVcpusFlags() that was
not supposed to reset the error. Usual procedure was done, saving the
error to temporary variable, but it was never free'd, but rather leaked.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Commit af2a1f05 tried clearly separating each condition in
qemuRestoreCgroupState() for the sake of readability, however somehow
one condition body was missing. That means that the body of the next
condition got executed only if both of there were true, which is
impossible, thus resulting in a dead code and a logic error.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
When hot-plugging a VCPU into the guest, kvm needs to allocate some data
from the DMA zone, which might be in a memory node that's not allowed in
cpuset.mems. Basically the same problem as there was with starting the
domain and due to which commit 7e72ac7878
exists. This patch just extends it to hotplugging as well.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1161540
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Instead of setting the value of cpuset.mems once when the domain starts
and then re-calculating the value every time we need to change the child
cgroup values, leave the cgroup alone and rather set the child data
every time there is new cgroup created. We don't leave any task in the
parent group anyway. This will ease both current and future code.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1174154
When we use attach-device add a hostdev or chr device which have a
iscsi address or others (just like guest agent, subsys iscsi disk...),
we will find there is no basic controller for our new attached device.
Somtimes this will make guest cannot start after we add them (although
they can start at the second time).
Signed-off-by: Luyao Huang <lhuang@redhat.com>
When libvirt is managing a bridge's forwarding database (FDB)
(macTableManager='libvirt'), if we add FDB entries for a new guest
interface even before the qemu process is created, then in the case of
a migration any other guest attached to the "destination" bridge will
have its traffic immediately sent to the destination of the migration
even while the source domain is still running (and the destination, of
course, isn't). To make sure that traffic from other guests on the new
host continues flowing to the old guest until the new one is ready, we
have to wait until the new guest CPUs are started to add the FDB
entries.
Conversely, we need to remove the FDB entries from the bridge any time
the guest CPUs are stopped; among other things, this will assure
proper operation during a post-copy migration (which is just the
opposite of the problem described in the previous paragraph).
We can change vnc password by using virDomainUpdateDeviceFlags API with
live flag. But it can't be changed with config flag. Error is reported as
below.
error: Operation not supported: persistent update of device 'graphics' is not supported
This patch supports the graphics arguments changed with config flag.
Signed-off-by: Wang Rui <moon.wangrui@huawei.com>
It's not supported to change some graphics arguments with '--live'.
Replace some error code VIR_ERR_INTERNAL_ERROR and VIR_ERR_INVALID_ARG
with VIR_ERR_OPERATION_UNSUPPORTED.
Signed-off-by: Wang Rui <moon.wangrui@huawei.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1173507
It occurred to me that OpenStack uses the following XML when not using
regular huge pages:
<memoryBacking>
<hugepages>
<page size='4' unit='KiB'/>
</hugepages>
</memoryBacking>
However, since we are expecting to see huge pages only, we fail to
startup the domain with following error:
libvirtError: internal error: Unable to find any usable hugetlbfs
mount for 4 KiB
While regular system pages are not huge pages technically, our code is
prepared for that and if it helps OpenStack (or other management
applications) we should cope with that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1160995
In our config files users are expected to pass several integer values
for different configuration knobs. However, majority of them expect a
nonnegative number and only a few of them accept a negative number too
(notably keepalive_interval in libvirtd.conf).
Therefore, a new type to config value is introduced: VIR_CONF_ULONG
that is set whenever an integer is positive or zero. With this
approach knobs accepting VIR_CONF_LONG should accept VIR_CONF_ULONG
too.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We now have a qemuInterfaceStartDevices() which does the final
activation needed for the host-side tap/macvtap devices that are used
for qemu network connections. It will soon make sense to have the
converse qemuInterfaceStopDevices() which will undo whatever was done
during qemuInterfaceStartDevices().
A function to "stop" a single device has also been added, and is
called from the appropriate place in qemuDomainDetachNetDevice(),
although this is currently unnecessary - the device is going to
immediately be deleted anyway, so any extra "deactivation" will be for
naught. The call is included for completeness, though, in anticipation
that in the future there may be some required action that *isn't*
nullified by deleting the device.
This patch is a part of a more complete fix for:
https://bugzilla.redhat.com/show_bug.cgi?id=1081461
The patch that added qemuInterfaceStartDevices() (upstream commit
82977058f5) had an extra conditional to
prevent calling it if the reason for starting the CPUs was
VIR_DOMAIN_RUNNING_UNPAUSED or VIR_DOMAIN_RUNNING_SAVE_CANCELED. This
was put in by the author as the result of a reviewer asking if it was
necessary to ifup the interfaces in *all* occasions (because these
were the two cases where the CPU would have already been started (and
stopped) once, so the interface would already be ifup'ed).
It turns out that, as long as there is no corresponding
qemuInterfaceStopDevices() to ifdown the interfaces anytime the CPUs
are stopped, neglecting to ifup when reason is RUNNING_UNPAUSED or
RUNNING_SAVE_CANCELED doesn't cause any problems (because it just
happens that the interface will have already been ifup'ed by a prior
call when the CPU was previously started for some other reason).
However, it also doesn't *help*, and there will soon be a
qemuInterfaceStopDevices() function which *will* ifdown these
interfaces when the guest CPUs are stopped, and once that is done, the
interfaces will be left down in some cases when they should be up (for
example, if a domain is paused and then unpaused).
So, this patch is removing the condition in favor of always calling
qemuInterfaeStartDevices() when the guest CPUs are started.
This patch (and the aforementioned patch) resolve:
https://bugzilla.redhat.com/show_bug.cgi?id=1081461
A logic bug in qemuConnectGetAllDomainStats makes the code mark the
monitor as available when qemuDomainObjBeginJob fails, instead of when
it succeeds, as the correct flow requires.
This patch fixes the check and updates the code documentation
accordingly.
Broken by commit 57023c0a3a.
Signed-off-by: Francesco Romani <fromani@redhat.com>
Currently, MAC registration occurs during device creation, which is
early enough that, during live migration, you end up with duplicate
MAC addresses on still-running source and target devices, even though
the target device isn't actually being used yet.
This patch proposes to defer MAC registration until right before
the guest can actually use the device -- In other words, right
before starting guest CPUs.
Signed-off-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com>
Signed-off-by: Laine Stump <laine@laine.org>
When user doesn't have read access on one of the domains he requested,
the for loop could exit abruptly or continue and override pointer which
pointed to locked object.
This patch fixed two issues at once. One is that domflags might have
had QEMU_DOMAIN_STATS_HAVE_JOB even when there was no job started (this
is fixed by doing domflags |= QEMU_DOMAIN_STATS_HAVE_JOB only when the
job was acquired and cleaning domflags on every start of the loop.
Second one is that the domain is kept locked when
virConnectGetAllDomainStatsCheckACL() fails and continues the loop when
it didn't end. Adding a simple virObjectUnlock() and clearing the
pointer ought to do.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Avoid leaving the domain locked on a failed ACL check in
qemuDomainMigratePerform() and qemuDomainMigrateFinish2().
Introduced in commit abf75aea24 (Add ACL checks into the QEMU driver).
qemuNetworkIfaceConnect() used to have a special case for
actualType='network' (a network with forward mode of route, nat, or
isolated) to call the libvirt public API to retrieve the bridge being
used by a network. That is no longer necessary - since all network
types that use a bridge and tap device now get the bridge name stored
in the ActualNetDef, we can just always use
virDomainNetGetActualBridgeName() instead.
(an audit of the two callers to qemuNetworkIfaceConnect() confirms
that it is never called for any other type of network, so the dead
code in the else statement (logging an internal error if it is called
for any other type of network) is eliminated in the process.)
When libvirt is managing the MAC table of a Linux host bridge, it must
turn off learning and unicast_flood for each tap device attached to
that bridge, then add a Forwarding Database (fdb) entry for the tap
device using the MAC address from the domain interface config.
Once we have disabled learning and flooding, any packet that has a
destination MAC address not present in the fdb will be dropped by the
bridge. This, along with the opportunistic disabling of promiscuous
mode[*], can result in enhanced network performance. and a potential
slight security improvement.
[*] If there is only one device on the bridge with learning/unicast_flood
enabled, then that device will automatically have promiscuous mode
disabled. If there are *no* devices with learning/unicast_flood
enabled (e.g. for a libvirt "route", "nat", or isolated network that
has no physical device attached), then all non-tap devices will have
promiscuous mode disabled (tap devices always have promiscuous mode
enabled, which may be a bug in the kernel, but in practice has 0
effect).
None of this has any effect for kernels prior to 3.15 (upstream kernel
commit 2796d0c648c940b4796f84384fbcfb0a2399db84 "bridge: Automatically
manage port promiscuous mode"). Even after that, until kernel 3.17
(upstream commit 5be5a2df40f005ea7fb7e280e87bbbcfcf1c2fc0 "bridge: Add
filtering support for default_pvid") traffic will not be properly
forwarded without manually adding vlan table entries. Unfortunately,
although the presence of the first patch is signalled by existence of
the "learning" and "unicast_flood" options in sysfs, there is no
reliable way to query whether or not the system's kernel has the
second of those patches installed, the only thing that can be done is
to try the setting and see if traffic continues to pass.
I'm about to make block stats optionally more complex to cover
backing chains, where block.count will no longer equal the number
of <disks> for a domain. For these reasons, it is nicer if the
statistics output includes the source path (for local files).
This patch doesn't add anything for network disks, although we
may decide to add that later.
With this patch, I now see the following for the same domain as
in the previous patch (one qcow2 file, and an empty cdrom drive):
$ virsh domstats --block foo
Domain: 'foo'
block.count=2
block.0.name=hda
block.0.path=/var/lib/libvirt/images/foo.qcow2
block.1.name=hdc
* src/libvirt-domain.c (virConnectGetAllDomainStats): Document
new field.
* tools/virsh.pod (domstats): Document new field.
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Return the new
stat for local files/block devices.
(QEMU_ADD_NAME_PARAM): Add parameter.
(qemuDomainGetStatsInterface): Update caller.
Signed-off-by: Eric Blake <eblake@redhat.com>
I noticed that for an offline domain, 'virsh domstats --block $dom'
was producing just the domain name, with no stats. But the older
'virsh domblkinfo' works just fine on offline domains. This patch
starts to get us closer, by at least reporting the disk names for
an offline domain.
With this patch, I now see the following for an offline domain
with one qcow2 disk and an empty cdrom drive:
$ virsh domstats --block foo
Domain: 'foo'
block.count=2
block.0.name=hda
block.1.name=hdc
* src/qemu/qemu_driver.c (qemuDomainGetStatsBlock): Don't short-circuit
output of block name.
Signed-off-by: Eric Blake <eblake@redhat.com>
qemuDomainGetStatsBlock() could leak a stats hash table if it
encountered OOM while populating the virTypedParameters.
Oddly, the fix doesn't even touch qemuDomainGetStatsBlock :)
* src/qemu/qemu_driver.c (QEMU_ADD_COUNT_PARAM)
(QEMU_ADD_NAME_PARAM): Don't return early.
(qemuDomainGetStatsInterface): Adjust caller.
Signed-off-by: Eric Blake <eblake@redhat.com>
If probing capabilities via QMP fails, we now have a check
that prevents us falling back to -help parsing. Unfortunately
the error message
"Failed to probe capabilities for /usr/bin/qemu-kvm:
unsupported configuration: QEMU 2.1.2 is too new for help parsing"
is proving rather unhelpful to the user. We need to be telling
them why QMP failed (the root cause), rather than they can't
use -help (the side effect).
To do this we should capture stderr during QMP probing, and
if -help parsing then sees a new QEMU version, we know that
QMP should have worked, and so we can show the messages from
stderr. The message thus becomes
"Failed to probe capabilities for /usr/bin/qemu-kvm:
internal error: QEMU / QMP failed: Could not access
KVM kernel module: No such file or directory
failed to initialize KVM: No such file or directory"
When attempting to create internal system checkpoint with a passthrough
device qemu will report the following error:
error: operation failed: Error -22 while writing VM
This patch calls the function to check if migration is possible with
given VM and thus improves the error to:
error: Requested operation is not valid: domain has assigned non-USB host devices
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=874418#c19
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Move entering the job into the thread to simplify the program flow. Also
as the code holds a separate reference to the domain object some
conditions can be simplified.
After this patch qemuDomainObjTransferJob is no longer needed so this
patch removes it.
If someone removes blockcopy storage file when still in mirroring phase
and then requesting blockjob abort using pivot, virsh cmd freezes. This
is not an issue with older qemu versions which did not support
asynchronous jobs (which we prefer by default).
As we have reached the mirroring phase successfully, polling monitor for
blockjob info always returns 1 and the loop never ends.
This fix introduces a check for qemuDomainBlockPivot return code, possibly
skipping the asynchronous waiting completely, if an error occurred and
asynchronous waiting was the preferred method.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1139567
Reconnect to the VM is a possibly long-running job spawned in a separate
thread. We should reload the snapshot defs and managedsave state prior
to spawning the thread to avoid blocking of the daemon startup which
would serialize on the VM lock.
Also the reloading code would violate the domain job held while
reconnecting as the loader functions don't create jobs.
Based on previous commit, we can now precreate missing volumes. While
digging out the functionality from storage driver would be nicer, if
you've seen the code it's nearly impossible. So I'm going from the
other end:
1) For given disk target, disk path is looked up.
2) For the disk path, storage pool is looked up, a volume XML is
constructed and then passed to virStorageVolCreateXML() which has all
the knowledge how to create raw images, (encrypted) qcow(2) images,
etc.
One of the advantages of this approach is, we don't have to care about
image conversion - qemu does that for us. So for instance, users can
transform qcow2 into raw on migration (if the correct XML is passed to
the migration API).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Up 'til now, users need to precreate non-shared storage on migration
themselves. This is not very friendly requirement and we should do
something about it. In this patch, the migration cookie is extended,
so that <nbd/> section does not only contain NBD port, but info on
disks being migrated. This patch sends a list of pairs of:
<disk target; disk size>
to the destination. The actual storage allocation is left for next
commit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The function queries the block devices visible to qemu
('query-block') and parses the qemu's output. The info is
returned in a hash table which is expected to be pre-filled by
qemuMonitorJSONGetAllBlockStatsInfo(). However, in the next patch
we are not going to call the latter function at all, so we should
make the former function add devices into the hash table if not
found there.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Since virDomainSnapshotFree will call virObjectUnref anyway, let's just use
that directly so as to avoid the possibility that we inadvertently clear out
a pending error message when using the public API.
Since virNetworkFree will call virObjectUnref anyway, let's just use that
directly so as to avoid the possibility that we inadvertently clear out
a pending error message when using the public API.
Since virDomainFree will call virObjectUnref anyway, let's just use that
directly so as to avoid the possibility that we inadvertently clear out
a pending error message when using the public API.
There is a race condition between the fopen and fscanf calls
in qemuGetProcessInfo. If fopen succeeds, there is a small
possibility that the file no longer exists before reading from it.
Now, if either fopen or fscanf calls fail, the function will behave
just as only fopen had failed.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1169055
Signed-off-by: Eric Blake <eblake@redhat.com>
Coverity complained that because the cfg->macFilter call checked
net->ifname != NULL before calling ebtablesRemoveForwardAllowIn, then
the virNetDevOpenvswitchRemovePort call should have the same check.
However, if I move the ebtables call prior to the check for TYPE_DIRECT
(where there is a VIR_FREE(net->ifname)), then it seems Coverity is
happy. Since firewall info is tacked on last during setup, removing
it in the opposite order of initialization seems to be natural anyway
There are some small issue in qemuProcessAttach:
1.Fix virSecurityManagerGetProcessLabel always get pid = 0,
move 'vm->pid = pid' before call virSecurityManagerGetProcessLabel.
2.Use virSecurityManagerGenLabel to get image label.
3.Fix always set selinux label for other security driver label.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
When a block{commit,copy} job was aborted on a domain, block job handler
did not process it correctly, leaving a phantom job in the background.
Any further calls to any blockjob causes "block <jobtype> still active"
error. This patch fixes the blockjob handler so that it checks not only
for VIR_DOMAIN_BLOCK_JOB_FAILED status, but VIR_DOMAIN_BLOCK_JOB_CANCELED
status as well, followed by our existing cleanup routine.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1135169
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
If job is failed in qemuMigrationRun, we expect the jobinfo type as
FAILED. But jobinfo type won't be updated until entering
qemuMigrationWaitForCompletion. We should make it updated in all
conditions. Moreover, we can't use qemuMigrationUpdateJobStatus
here because job may fail in libvirt, so we can't query job status
from QEMU.
Signed-off-by: Wang Rui <moon.wangrui@huawei.com>
The migration job status is traced in qemuMigrationUpdateJobStatus
which is called in qemuMigrationRun. But if migration is cancelled
before the trace such as in qemuMigrationDriveMirror, the jobinfo
type won't be updated to CANCELLED. After this patch, we can get
jobinfo type CANCELLED if migration is cancelled during drive
mirror. Moreover, we can't use qemuMigrationUpdateJobStatus
because from qemu's point of view it's just the drive mirror being
cancelled and the migration hasn't even started yet.
Signed-off-by: Wang Rui <moon.wangrui@huawei.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1160084
As of b6d4dad11b (1.2.5) we are trying to keep the status of FSFreeze
in the guest. Even though I've tried to fixed couple of corner cases
(6ea54769ba), it occurred to me just recently, that the approach is
broken by design. Firstly, there are many other ways to talk to
qemu-ga (even through libvirt) that filesystems can be thawed (e.g.
qemu-agent-command) without libvirt noticing. Moreover, there are
plenty of ways to thaw filesystems without even qemu-ga noticing (yes,
qemu-ga keeps internal track of FSFreeze status). So, instead of
keeping the track ourselves, or asking qemu-ga for stale state, it's
the best to let qemu-ga deal with that (and possibly let guest kernel
propagate an error).
Moreover, there's one bug with the following approach, if fsfreeze
command failed, we've executed fsthaw subsequently. So issuing
domfsfreeze in virsh gave the following result:
virsh # domfsfreeze gentoo
Froze 1 filesystem(s)
virsh # domfsfreeze gentoo
error: Unable to freeze filesystems
error: internal error: unable to execute QEMU agent command 'guest-fsfreeze-freeze': The command guest-fsfreeze-freeze has been disabled for this instance
virsh # domfsfreeze gentoo
Froze 1 filesystem(s)
virsh # domfsfreeze gentoo
error: Unable to freeze filesystems
error: internal error: unable to execute QEMU agent command 'guest-fsfreeze-freeze': The command guest-fsfreeze-freeze has been disabled for this instance
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
virReportSystemError is reserved for reporting system errors, calling it
with VIR_ERR_* error codes produces error messages that do not make any
sense, such as
internal error: guest failed to start: Kernel doesn't support user
namespace: Link has been severed
We should prohibit wrong usage with a syntax-check rule.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Commit 6fcddfcd refactored job statistics but missed the jobinfo type updated
in qemuDomainGetJobInfo. After this patch, we can use virDomainGetJobInfo to
get jobinfo type again.
Signed-off-by: Wang Rui <moon.wangrui@huawei.com>
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Add attribute to set vgamem_mb parameter of QXL device for QEMU. This
value sets the size of VGA framebuffer for QXL device. Default value in
QEMU is 8MB so reuse it also in libvirt to not break things.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
So far we didn't have any option to set video memory size for qemu video
devices. There was only the vram (ram for QXL) attribute but it was valid
only for the QXL video device.
To provide this feature to users QEMU has a dedicated device attribute
called 'vgamem_mb' to set the video memory size. We will use the 'vram'
attribute for setting video memory size for other QEMU video devices.
For the cirrus device we will ignore the vram value because it has
hardcoded video size in QEMU.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
QEMU has two different type of QXL display device. The first "qxl-vga"
is for primary video device and second "qxl" is for secondary video
device.
There are also two different ways how to specify those devices on qemu
command line, the first one and obsolete is using "-vga" option and the
current new one is using "-device" option. The "-vga" could be used only
to setup primary video device, so the "-vga qxl" equal to
"-device qxl-vga". Unfortunately the "-vga qxl" doesn't support setting
additional parameters for the device and "-global" option must be used
for this purpose. It's mandatory to use "-global qxl-vga...." to set the
parameters of primary video device previously defined with "-vga qxl".
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The vram attribute was introduced to set the video memory but it is
usable only for few hypervisors excluding QEMU/KVM and the old XEN
driver. Only in case of QEMU the vram was used for QXL.
This patch updates the documentation to reflect current code in libvirt
and also changes the cases when we will set the default vram attribute.
It also fixes existing strange default value for VGA devices 9MB to 16MB
because the video ram should be rounded to power of two.
The change of default value could affect migrations but I found out that
QEMU always round the video ram to power of two internally so it's safe
to change the default value to the next closest power of two and also
silently correct every domain XML definition. And it's also safe because
we don't pass the value to QEMU.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Get mounted filesystems list, which contains hardware info of disks and its
controllers, from QEMU guest agent 2.2+. Then, convert the hardware info
to corresponding device aliases for the disks.
Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama@hds.com>
Improve the monitor function to also retrieve the guest state of
character device (if provided) so that we can refresh the state of
virtio-serial channels and perhaps react to changes in the state in
future patches.
This patch changes the returned data from qemuMonitorGetChardevInfo to
return a structure containing the pty path and the state for all the
character devices.
The change to the testsuite makes sure that the data is parsed
correctly.
