Although I have not experienced this in real life, there is a
possible race condition when creating new device, getting its XML
or parent or listing its capabilities. If the nodedev driver is
still enumerating devices (in a separate thread) and one of
virNodeDeviceGetXMLDesc(), virNodeDeviceGetParent(),
virNodeDeviceNumOfCaps(), virNodeDeviceListCaps() or
virNodeDeviceCreate() is called then it can lead to spurious
results because the device enumeration thread is removing devices
from or adding them to the internal list of devices (among with
their states).
Therefore, wait for things to settle down before proceeding with
any of the APIs.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This is more academic dispute than a real bug, but this is taken
from pthread_cond_broadcast(3p) man:
The pthread_cond_broadcast() or pthread_cond_signal() functions
may be called by a thread whether or not it currently owns the
mutex that threads calling pthread_cond_wait() or
pthread_cond_timedwait() have associated with the condition
variable during their waits; however, if predictable scheduling
behavior is required, then that mutex shall be locked by the
thread calling pthread_cond_broadcast() or
pthread_cond_signal().
Therefore, broadcast the initCond while the nodedev driver is
still locked.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
The consensus is to put the verb last. Therefore, the new name is
nodeDeviceInitWait(). This allows us to introduce new function
(done later in a separate commit) that will "complete" the device
initialization.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Historically, we declared pointer type to our types:
typedef struct _virXXX virXXX;
typedef virXXX *virXXXPtr;
But usefulness of such declaration is questionable, at best.
Unfortunately, we can't drop every such declaration - we have to
carry some over, because they are part of public API (e.g.
virDomainPtr). But for internal types - we can do drop them and
use what every other C project uses 'virXXX *'.
This change was generated by a very ugly shell script that
generated sed script which was then called over each file in the
repository. For the shell script refer to the cover letter:
https://listman.redhat.com/archives/libvir-list/2021-March/msg00537.html
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The comment for that option states that the default value is 'none' but
it was not set by the code. By default the value is NULL which results
into the following warning:
warning : qemuBuildCompatDeprecatedCommandLine:10393 : Unsupported deprecation behavior '(null)' for VM 'test'
Fixes: 7004504493
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The g_path_is_absolute() considers more situations
than just a simply "path[0] == '/'".
Related issue: https://gitlab.com/libvirt/libvirt/-/issues/12
Signed-off-by: Luke Yue <lukedyue@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When doing a blockpull with NULL base the full contents of the disk are
pulled into the topmost image which then becomes fully self-contained.
qemuBlockJobProcessEventCompletedPull doesn't install the backing chain
terminators though, although it's guaranteed that there will be no
backing chain behind disk->src.
Add the terminators for completness and for disabling backing chain
detection on further boots.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
When doing a full block pull job (base == NULL) and the config XML
contains a compatible disk, the completer function would leave a
dangling pointer in 'cfgdisk->src->backingStore' as cfgdisk->src would
be set to the value of 'cfgbase' which was always set to
'cfgdisk->src->backingStore'.
This is wrong though since for the live definition XML we set the
respective counterpart to 'job->data.pull.base' which is NULL in the
above scenario.
This leads to a invalid pointer read when saving the config XML and may
end up in a crash.
Resolve it by setting 'cfgbase' only when 'job->data.pull.base' is
non-NULL.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1946918
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
During its initialization, the nodedev driver tries to set up
monitors for /etc/mdevctl.d directory, so that it can register
mdevs as they come and go. However, if the file doesn't exist
there is nothing to monitor and therefore we can exit early. In
fact, we have to otherwise monitorFileRecursively() fails and
whole driver initialization fails with it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
During the nodedev driver initialization two threads are created:
one for listening on udev events (like device plug/unplug) and
the other for enumerating devices (so that the main thread doing
the driver init is not blocked). If something goes wrong at any
point then nodeStateCleanup() is called which joins those two
threads (possibly) created before. But it tries to join them even
they weren't created which is undefined behaviour (and it just so
happens that it crashes on my system).
If those two virThread variables are turned into pointers then we
can use comparison against NULL to detect whether threads were
created.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
The nodedev driver private data object @priv is created by
calling udevEventDataNew(). After that, driver->privateData
pointer is set to the freshly allocated object and only a few
lines after all of this the object is locked. Technically it is
safe because there should not be any other thread at this point,
but defensive style of programming says it's better if the object
is locked before driver's privateData is set.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
If initialization of priv->mdevctlMonitors fails, then the
control jumps over to cleanup label where nodeStateCleanup() is
called which tries to lock @priv. But since @priv was already
locked before taking the jump a deadlock occurs. The solution is
to jump onto @unlock label, just like the code around is doing.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
In commit ad80bba90a I mistakenly didn't delete '**' from the
variable declaration when converting it to 'GStrv' and deleted the
'separator' variable since it was declared on the same line as a
different variable.
Fixes: ad80bba90a
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Move calls to virStorageBackendFileSystemMountAddOptions earlier so that
the options are formatted before the positional arguments.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Put multiple values for an option if followed by another option as used
in certain iptables arguments.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
virCommandToStringFull used internally when virCommandSetDryRun is
requested allows to strip command path and wrap lines nicely. Expose
these via virCommandSetDryRun so that tests can use those features
instead of local hacks.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
While virCommandSetDryRun is used in tests only, there were some cases
when error paths would not call the function with NULL arguments to
reset the dry run infrastructure.
Introduce virCommandDryRunToken type which must be allocated via
virCommandDryRunTokenNew and passed to virCommandSetDryRun.
This way we can use automatic variable cleaning to trigger the cleanup
of virCommandSetDryRun parameters and also the use of the token variable
ensures that all callers of virCommandSetDryRun clean up after
themselves and also that the token isn't left unused in the code.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
In tests we don't want to use the full path to commands as it's
unpleasant to keep that working on all systems.
Add an integrated way to strip the prefix which will be used to replace
virTestClearCommandPath() as a more systemic solution.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Callers which need the count of elements now count it in place.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While the code invokes the string list length calculation twice, it
happens only on error path, which by itself should never happen.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use g_strsplit instead of virStringSplitCount and automatically free the
temporary string list.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In 3 of 4 instances the code didn't even need the count of the elements.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Refactor the handling of variables so that the cleanup section can be
sanitized.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Remove 'cleanup' and 'error' labels by switching 'ret' to automatic
pointer and stealing it in the return statement.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Move variables into the loop which uses them and use automatic freeing
for temporarily allocated variables.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Move variables into the loop which uses them and use automatic freeing
for temporarily allocated variables.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Both instances just check the length once. Replicate that faithfully.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use str(r)chr to find the correct bit rather than fully splitting the
string.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Unfortunately here we do need the count of elements. Use g_strv_length
to calculate it so that virStringSplitCount can be removed later.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Rewrite the code to remove the need to calculate the string list count.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Rewrite the code to remove the need to calculate the string list count.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The presence of the second element can be checked by looking at it
directly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The presence of the second element can be checked by looking at it
directly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Prevent unbounded chains by limiting the recursion depth of
virStorageSourceGetMetadataRecurse to the maximum number of image layers
we limit anyways.
This removes the last use of virStorageSourceGetUniqueIdentifier which
will allow us to delete some crusty old infrastructure which isn't
really needed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The function is used only inside of the file. We can open-code it and
remove it as it's not very useful.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Add a simpler algorithm converting the JSON array to bitmap so that
virJSONValueGetArrayAsBitmap can be removed in next step.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Follow best practices and add a unsigned int flags parameter to these
new APIs that have not been in a release yet.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The comments mistakenly say 7.2.0, when they were actually merged during
the 7.3 development cycle.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
When calling virNodeDeviceDefineXML() to define a new mediated device,
we call virMdevctlDefine() and then wait for the new device to appear in
the driver's device list before returning. This caused long delays due
to the behavior of nodeDeviceFindNewMediatedDevice(). This function
checks to see if the device is in the list and then waits for 5s before
checking again.
Because mdevctl is relatively slow to query the list of defined
devices[0], the newly-defined device was generally not in the device
list when we first checked. This results in libvirt almost always taking
at least 5s to complete this API call for mediated devices, which is
unacceptable.
In order to avoid this long delay, we resort to a workaround. If the
call to virMdevctlDefine() was successful, we can assume that this new
device will exist the next time we query mdevctl for new devices. So we
simply add this provisional device definition directly to the nodedev
driver's device list and return from the function. At some point in the
future, the mdevctl handler will run and the "official" device will be
processed, which will update the provisional device if any new details
need to be added.
The reason that this is not necessary for virNodeDeviceCreateXML() is
because detecting newly-created (not defined) mdevs happens through
udev instead of mdevctl. And nodeDeviceFindNewMediatedDevice() always
calls 'udevadm settle' before checking to see whether the device is in
the list. This allows us to wait just long enough for all udev events to
be processed, so the device is almost always in the list the first time
we check and so we almost never end up hitting the 5s sleep.
[0] on my machine, 'mdevctl list --defined' took around 0.8s to
complete for only 3 defined mdevs.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
To accomodate re-use of this functionality in a following patch, split
out the processing of an individual mdev definition into a separate
function.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Calling `mdevctl stop` for a mediated device that is in use by an active
domain will block until that vm exits (or the vm closes the device).
Since the nodedev driver cannot query the hypervisor driver to see
whether any active domains are using the device, we resort to a
workaround that relies on the fact that a vfio group can only be opened
by one user at a time. If we get an EBUSY error when attempting to open
the group file, we assume the device is in use and refuse to try to
destroy that device.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Use the new <uuid> element in the mdev caps to define and start devices
with a specific UUID.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
It will be useful to be able to specify a particular UUID for a mediated
device when defining the node device. To accomodate that, allow this to
be specified in the xml schema. This patch also parses and formats that
value to the xml, but does not yet use it.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This new API function provides a way to start a persistently-defined
mediate device that was defined by virNodeDeviceDefineXML() (or one that
was defined externally via mdevctl)
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This interface allows you to undefine a persistently defined (but
inactive) mediated devices. It is implemented via 'mdevctl'
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
With mediated devices, we can now define persistent node devices that
can be started and stopped. In order to take advantage of this, we need
an API to define new node devices.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Abstract out the function used to generate the commandline for 'mdevctl
start' since they take the same arguments. Add tests to ensure that
we're generating the command properly.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
We need to query mdevctl for changes to device definitions since an
administrator can define new devices by executing mdevctl outside of
libvirt.
In the future, mdevctl may add a way to signal device add/remove via
events, but for now we resort to a bit of a workaround: monitoring the
mdevctl config directory for changes to files. When a change is
detected, we query mdevctl and update our device list. The mdevctl
querying is handled in a throwaway thread, and these threads are
synchronized with a mutex.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
mdevctl does not currently provide any events when the list of defined
devices changes, so we will need to poll mdevctl for the list of defined
devices periodically. When a mediated device no longer exists from one
iteration to the next, we need to treat it as an "undefine" event.
When we get such an event, we remove the device from the list if it's
not active. Otherwise, we simply mark it as non-persistent.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
When a mediated device is stopped or undefined by an application outside
of libvirt, we need to remove it from our list of node devices within
libvirt. This patch introduces virNodeDeviceObjListRemoveLocked() and
virNodeDeviceObjListForEachRemove() (which are analogous to other types
of object lists in libvirt) to facilitate that. They will be used in
coming commits.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
At startup, query devices that are defined by 'mdevctl' and add them to
the node device list.
This adds a complication: we now have two potential sources of
information for a node device:
- udev for all devices and for activated mediated devices
- mdevctl for persistent mediated devices
Unfortunately, neither backend returns full information for a mediated
device. For example, if a persistent mediated device in the list (with
information provided from mdevctl) is 'started', that same device will
now be detected by udev. If we simply overwrite the existing device
definition with the new one provided by the udev backend, we will lose
extra information that was provided by mdevctl (e.g. attributes, etc).
To avoid this, make sure to copy the extra information into the new
device definition.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Consistent with other objects (e.g. virDomainObj), add a field to
indicate whether the node device is persistent or transient.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This adds an internal API to query for persistent mediated devices
that are defined by mdevctl. Upcoming commits will make use of this
information.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This function will parse the list of mediated devices that are returned
by mdevctl and convert it into our internal node device representation.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Expose a helper function that can be used by udev and mdevctl to
generate device names for node devices.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
It doesn't make sense to list all of the flag values in the function
documentation. This is unnecessary duplication, we already refer to the
enum type. Also, remove reference to exclusive groups of flags, since
that does not apply to this API.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Add two flag values for virConnectListAllNodeDevices() so that we can
list only node devices that are active or inactive.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
we will be able to define mediated devices that can be started or
stopped, so we need to be able to indicate whether the device is active
or not, similar to other resources (storage pools, domains, etc.)
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
When an mdevctl command fails, there is not much information available
to the user about why it failed. This is partly because we were not
making use of the error message that mdevctl itself prints upon failure.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This makes it possible to enable stable NIC device names in most modern
Linux distros.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This property is exposed by QEMU on any PCI device, but we have to pick
some specific device(s) to probe it against. We expect that at least one
of the virtio devices will be present, so probe against them.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The compiler can more easily optimize a switch, and more importantly can
also warn when new address types are added which are not handled.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
PCI devices can be associated with a unique integer index that is
exposed via ACPI. In Linux OS with systemd, this value is used for
provide a NIC device naming scheme that is stable across changes
in PCI slot configuration.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Enable '-compat' if requested in qemu.conf and supported by qemu to
instruct qemu to crash when a deprecated command is used and stop
returning deprecated fields.
This setting is meant for libvirt developers and such.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Similar to the qemu.conf knob 'deprecation_behavior' add a per-VM knob
in the QEMU namespace:
<qemu:deprecation behavior='...'/>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
New QEMU supports a harsh, but hard to ignore way to notify that the
QMP user used a deprecated command. This is useful e.g. for developers
to see that something needs to be fixed.
This patch introduces a qemu.conf option to enable the setting in cases
when qemu supports it so that developers and continiuous integration
efforts are notified about use of deprecated fields before it's too
late.
The option is deliberately stored as string and not validated to prevent
failures when downgrading qemu or libvirt versions. While we don't
support this, the knob isn't meant for public consumption anyways.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
The capability is asserted if qemu supports the -compat
deprecated-input= and deprecated-output= settings to control what should
happen if deprecated fields are used in QMP.
This will be used for a developer/tester-oriented setting which will
aid us in catching use of deprecated settings sooner.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_set_memory_target, which changed the storage size of
parameter "target_memkb" in Xen 4.8.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_send_trigger, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_set_vcpuonline, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_get_free_memory, which changed storage size of parameter
"memkb" in Xen 4.8.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_need_memory, which changed the storage size of
"need_memkb" in Xen 4.8. With Xen 4.12 the libxl_domain_config
parameter was changed
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_unpause, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_pause, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_reboot, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_shutdown, which got a new parameter
"ao_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_retrieve_domain_configuration, which got a new parameter
"libxl_asyncop_how" in Xen 4.12. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Upcoming changes will use different LIBXL_API_VERSION variants.
Prepare libxl_domain_create_restore, which got a new parameter
"send_back_fd" in Xen 4.7. libvirt does not use this parameter.
No functional change intended.
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
This is available in QEMU with "ide-hd" and "scsi-hd" device
types. It was originally mistakenly added to the "scsi-block"
device type too, but later removed. This doesn't affect libvirt
since we restrict usage to device=disk.
When this property is not set then QEMU's default behaviour
is to not report any rotation rate information, which
causes most guest OS to assume rotational storage.
https://bugzilla.redhat.com/show_bug.cgi?id=1498955
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This lets the app expose the virtual SCSI or IDE disks as solid state
devices by setting a rate of '1', or rotational media by setting a
rate between 1025 and 65534.
https://bugzilla.redhat.com/show_bug.cgi?id=1498955
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Previously, we accepted empty bridge name, because some old versions of
VMWare Workstation did not put it into the config. But this doesn't make
much sense - to have an interface type bridge with no name. We
circumvented this problem by generating an empty name but that is
equally wrong.
Therefore, fill in missing bridge names (according to the documentation
[1] the default bridge name is VMnet0) and error out if bridge name is
missing.
This partially reverts f246cdb5ac
1: https://docs.vmware.com/en/VMware-Workstation-Player-for-Linux/16.0/com.vmware.player.linux.using.doc/GUID-BAFA66C3-81F0-4FCA-84C4-D9F7D258A60A.html
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
The
<os firmware='efi'>
<firmware type='efi'>
<feature enabled='no' name='enrolled-keys'/>
</firmware>
</os>
repeats the firmware attribute twice. This has no functional benefit, as
evidenced by fact that we use a single struct field to store both
attributes, while needlessly introducing an error scenario. The XML can
just be simplified to:
<os firmware='efi'>
<firmware>
<feature enabled='no' name='enrolled-keys'/>
</firmware>
</os>
which also means that we don't need to emit the empty element
<firmware type='efi'/> for all existing configs too.
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This API talks to QEMU and changes its internal state. Therefore,
it should acquire QEMU_JOB_MODIFY instead of QEMU_JOB_QUERY.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
This API interacts with the hypervisor and makes changes to its
behaviour, so must be protected by the write permission.
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Since the functions provided by libpciaccess are not thread-safe,
when the udev-event and nodedev-init threads of libvirt call the
pci_get_strings function provided by libpaciaccess at the same
time the following can happen:
nodedev-init thread:
nodeStateInitializeEnumerate ->
udevEnumerateDevices->
udevProcessDeviceListEntry ->
udevAddOneDevice ->
udevGetDeviceDetails->
udevProcessPCI ->
udevTranslatePCIIds ->
pci_get_strings -> (libpciaccess)
find_device_name ->
populate_vendor ->
d = realloc( vend->devices, (vend->num_devices + 1), * sizeof( struct pci_device_leaf ) );
vend->num_devices++;
udev-event thread:
udevEventHandleThread ->
udevHandleOneDevice ->
udevAddOneDevice->
udevGetDeviceDetails->
udevProcessPCI ->
udevTranslatePCIIds ->
pci_get_strings -> (libpciaccess)
find_device_name ->
populate_vendor ->
d = realloc( vend->devices, (vend->num_devices + 1), * sizeof( struct pci_device_leaf ) );
vend->num_devices++;
Signed-off-by: WangJian <wangjian161@huawei.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Recently, a few commits back I've switched bunch of code to
g_steal_pointer() using coccinelle. Problem was that the semantic
patch used was slightly off:
@@
expression a, b;
@@
+ b = g_steal_pointer(&a);
- b = a;
... when != a
- a = NULL;
Problem is that, "... when != a" is supposed to jump over those
lines, which don't contain expression a. My idea was to replace
the following pattern too:
ptrX = ptrY;
if (something(ptrZ) < 0) goto error;
ptrY = NULL;
But what I missed is that the following pattern is also matched
and replaced:
ptrX = ptrY;
if (something(ptrX) < 0) goto error;
ptrY = NULL;
This is not necessarily correct - as demonstrated by our hotplug
code. The real problem is ambiguous memory ownership transfer
(functions which add device to domain def take ownership only on
success), but to not tackle the real issue let's revert those
parts.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Turns out, the way that glib implements g_steal_pointer() is not
compatible with function callbacks. And that's what my recent
patch did in virNetSocketEventFree(). Revert that part.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In one of my recent patches I've introduced new connection
feature VIR_DRV_FEATURE_NETWORK_UPDATE_HAS_CORRECT_ORDER.
However, I forgot to add corresponding case into a switch in
vzConnectSupportsFeature().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The order in which virNetworkUpdate() accepts @section and
@command arguments is not the same as in which it passes them
onto networkUpdate() callback. Until recently, it did not really
matter, because calling the API on client side meant arguments
were encoded in reversed order (compared to the public API), but
then on the server it was fixed again - because the server
decoded RPC (still swapped), called public API (still swapped)
and in turn called the network driver callback (with reversing
the order - so magically fixing the order).
Long story short, if the public API is called even number of
times those swaps cancel each other out. The problem is when the
API is called an odd numbed of times - which happens with split
daemons and the right URI. There's one call in the client (e.g.
virsh net-update), the other in a hypervisor daemon (say
virtqemud) which ends up calling the API in the virnetworkd.
The fix is obvious - fix the order in which arguments are passed
to the callback.
But, to maintain compatibility with older, yet unfixed, daemons
new connection feature is introduced. The feature is detected
just before calling the callback and allows client to pass
arguments in correct order (talking to fixed daemon) or in
reversed order (talking to older daemon).
Unfortunately, older client talking to newer daemon can't be
fixed. Let's hope that it's less frequent scenario.
Fixes: 574b9bc66b
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1870552
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
So far, it was not needed, but shortly a client will want to know
whether virNetworkUpdate() API is fixed or not. See next commits
for more info.
Side note, this driver's implementation is called only when using
sub-driver's connection, i.e. "network:///system". For any other
URI the corresponding hypervisor's driver callback is called.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Just like VFIO devices, vDPA devices may need to have all guest memory
pages locked/pinned in order to operate properly. In the case of VFIO
devices (including mdev and NVME, which also use VFIO) libvirt
automatically increases the locked memory limit when one of those
devices is present. This patch modifies that code to also increase the
limit if there are any vDPA devices present.
Resolves: https://bugzilla.redhat.com/1939776
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
This function is a specialized version of
qemuDomainGetMemLockLimitBytes() for PPC64. Simplifying it in the same
manner as the previous patch has the nice side effect of accounting
for the possibility of an mdev device
(I don't know if mdev devices are supported on PPC, but even if not
then a) the additional check for mdev devices gained by using
qemuDomainNeedsVFIO() in place of open coding will be an effective
NOP, and b) if mdev devices are supported on PPC64 in the future, this
function will be prepared for it).
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
This function goes through a loop checking if each hostdev is a VFIO
or mdev device, and then later it calls virDomainDefHasNVMEDisk(). The
function qemuDomainNeedsVFIO() does exactly the same thing, so let's
just call that instead.
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
This function returns true if the domain has any interfaces that are
type='vdpa'.
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Generated by the following spatch:
@@
expression a, b;
@@
+ b = g_steal_pointer(&a);
- b = a;
... when != a
- a = NULL;
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
The ESX implementation of virConnectListAllDomains() follows
pretty much implementations in other drivers: it has local array
of virDomainPtr-s which (if requested by caller) is filled by
actual domains or not (if the caller is interested only in the
count of domains).
Anyway, in case of the former, the passed @domains argument is
set to the local array, which is then set to NULL to prevent it
from freeing under cleanup label. Pretty standard pattern.
Except, the local array is set to NULL always. Even if the local
array is not stolen. Fortunately, this doesn't lead to a memory
leak, because if caller is not interested in the array, none is
allocated. But it doesn't set good example and also breaks my
spatch rules.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Some SRIOV PFs don't have a netdev associated with them (the spec
apparently doesn't require it). In most cases when libvirt is dealing
with an SRIOV VF, that VF must have a PF, and the PF *must* have an
associated netdev (the only way to set the MAC address of a VF is by
sending a netlink message to the netdev of that VF's PF). But there
are times when we don't need for the PF to have a netdev; in
particular, when we're just getting the Switchdev Features for a VF,
we don't need the PF netdev - the netdev of the VF (apparently) works
just as well.
Commit 6452e2f5 (libvirt 5.1.0) *kind of* made libvirt work around PFs
with no netdevs in this case - if virNetDevGetPhysicalFunction
returned an error when setting up to retrieve Switchdev feature info,
it would ignore the error, and then check if the PF netdev name was
NULL and, if so it would reset the error object and continue on rather
than returning early with a failure. The problem is that by the time
this special handling occured, the error message about missing netdev
had already been logged, which was harmless to proper operation, but
confused the user.
Fortunately there are only 2 users of virNetDevGetPhysicalFunction, so
it is easy to redefine it's API to state that a missing netdev name is
*not* an error - in that case it will still return success, but the
caller must be prepared for the PF netdev name to be NULL. After
making this change, we can modify the two callers to behave properly
with the new semantics (for one of the callers it *is* still an error,
so the error message is moved there, but for the other it is okay to
continue), and our spurious error messages are a thing of the past.
Resolves: https://bugzilla.redhat.com/1924616
Fixes: 6452e2f5e1
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
With this, incomplete XML without <frames/> for <rx/> in coalesce
won't raise error as before. It will leave the coalesce parameter
empty, thanks to passing it as a parameter and return an integer
to indicate error state - previously it returned pointer (or NULL
for both error and incomplete XML).
I also added a test case to test this functionality in the
qemuxml2xmltest.
The code went through some refactoring:
* change of a condition
* addition of a parameter
* change of order, that allowed removal of VIR_FREE
* removal of redundant labels and variables
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1535930
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Commit ac87d3520a consolidated common cgroup code between the QEMU and
lxc drivers in domain_cgroup.c. In this process, in
virDomainCgroupSetupDomainBlkioParameters(), a call to
virCgroupGetBlkioWeight() went missing.
The result is that 'virsh blkiotune' is setting the blkio.weight for the
guest in the host cgroup, but not on the domain XML, because
virCgroupGetBlkioWeight() is also used to write the blkio.weight value
in the domain object.
Fix it by adding the virCgroupGetBlkioWeight() call in the
virDomainCgroupSetupDomainBlkioParameters() helper.
Fixes: ac87d3520a
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1941407
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Previously we'd assign the default checkpoint bitmap names in
virDomainCheckpointAlignDisks. In cases when the checkpoint is redefined
without a domain XML virDomainCheckpointAlignDisks is not called.
Add an explicit call to virDomainCheckpointDefAssignBitmapNames to
restore functionality.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When a checkpoint is redefined without providing the domain XML, we
might end up with a definition where the per-disk bitmap name is not
set. Trying to delete such checkpoint would lead to a crash.
Refuse such deletion.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1941600
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Base the detection on the presence of the 'secret' qom-type entry, which
isn't conditionally compiled in qemu.
All caps-based test now switch to using JSON for -object.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Skip the lossy conversion to legacy commandline arguments by using the
JSON props directly when -object is QAPIfied. This avoids issues with
conversion of bitmaps and also allows validation of the generated JSON
against the QMP schema in the tests.
Since the new approach is triggered by a qemu capability the code
from 'virQEMUBuildObjectCommandlineFromJSON' in util/virqemu.c was moved
to 'qemuBuildObjectCommandlineFromJSON' in qemu/qemu_command.c which has
the virQEMUCaps type.
Some functions needed to be modified to propagate qemuCaps.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Set 'objectAddNoWrap' when the capability is present.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
There's just one caller left. Since qemuBuildMemoryBackendProps is too
complex to be modified for now, just move the adding of 'id' and 'qom'
type directly into the function.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Construct the JSON object which is used for object-add without the
'props' wrapper and add the wrapper only in the monitor code.
This simplifies the JSON->commandline generator in the first place and
also prepares for upcoming qemu where 'props' will be removed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Starting from qemu-6.0 the parameters of -object/object-add are formally
described by the QAPI schema. Additionally this changes the nesting of
the properties as the 'props' nested object will be flattened to the
parent.
We'll need to detect whether qemu switched to this new approach to
generate the objects with proper nesting and also allow testing.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The firmwareFeatures member of virDomainOSDef struct is allocated
in virDomainDefParseBootFirmwareOptions() but never freed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
The virDomainDefFree() function frees individual members of
virDomainDef struct. The function is already long enough, move
code that handles def->os member into a separate function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Our reallocation APIs already abort on OOM and thus can only return 0.
There's no need to force callers to check the result.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
virQEMUCapsGet checks for qemuCaps itself, no need to do it explicitly.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
In case an async job spans multiple APIs (e.g., incoming migration) the
API that started the job is recorded as the asyncOwnerAPI even though it
is no longer running and the owner thread is updated properly to the one
currently handling the job. Let's also update asyncOwnerAPI to make it
more obvious which is the current (or the most recent) API involved in
the job.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Attempting to set the memlock limit might fail if we're running
in a containerized environment where CAP_SYS_RESOURCE is not
available, and if the limit is already high enough there's no
point in trying to raise it anyway.
https://bugzilla.redhat.com/show_bug.cgi?id=1916346
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Store the current memory locking limit and the desired one
separately, which will help with later changes.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Now that we've implemented a fallback for the function that
obtains the information from /proc, there is no reason we would
get a failure unless there's something seriously wrong with the
environment we're running in, in which case we're better off
reporting the issue to the user rather than pretending
everything is fine.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Calling prlimit() requires elevated privileges, specifically
CAP_SYS_RESOURCE, and getrlimit() only works for the current
process which is too limiting for our needs; /proc/$pid/limits,
on the other hand, can be read by any process, so implement
parsing that file as a fallback for when prlimit() fails.
This is useful in containerized environments.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Implement "<os firmware='efi'>" support for bhyve driver.
As there are not really lot of options, try to find
"BHYVE_UEFI.fd" firmware which is installed by the
sysutils/uefi-edk2-bhyve FreeBSD port.
If not found, just use the first found firmware
in the firmwares directory (which is configurable via
config file).
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When the backup job is terminated normally the security label is
restored by the blockjob finishing handler.
If the VM dies or is destroyed that wouldn't happen as the blockjob
handler wouldn't be called.
Restore the security label on disk store where we remember that the job
was running at the point when 'qemuBackupJobTerminate' was called.
Not resetting the security label means that we also leak the xattr
attributes remembering the label which prevents any further use of the
file, which is a problem for block devices.
This also requires that the call to 'qemuBackupJobTerminate' from
'qemuProcessStop' happens only after 'vm->pid' was reset as otherwise
the security subdrivers attempt to enter the process namespace which
fails if the process isn't running any more.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1939082
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
qemuBackupBegin can take a full backup of the disks (excluding any
operations with bitmaps) without the need to wait for the
blockdev-reopen support in qemu.
Add a check that no checkpoint creation is required and the disk backup
mode isn't VIR_DOMAIN_BACKUP_DISK_BACKUP_MODE_INCREMENTAL.
Call to virDomainBackupAlignDisks is moved earlier as it initializes the
disk backup mode if not present in user config.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Upcoming commit will enable full backup support (incremental part
requires blockdev-reopen, which won't happen in qemu for at least
another release).
Add a capability that the 'blockdev-backup' job is supported by qemu
capped, but limited to when qemu supports QEMU_CAPS_BLOCKDEV.
We can also use it in the expression to enable
QEMU_CAPS_INCREMENTAL_BACKUP since it's a pre-requisite too.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When the firmware auto-selection was introduced it always picked first
usable firmware based on the JSON descriptions on the host. It is
possible to add/remove/change the JSON files but it will always be for
the whole host.
This patch introduces support for configuring the auto-selection per VM
by adding users an option to limit what features they would like to have
available in the firmware.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The original code used a lot of conditions and was not that obvious
when each XML bits are parsed.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
With this, XML fails if non-virtio video devices have virtio
options. Previously it didn't raise error.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1922093
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Move this function in order to use it in the next patch before
its previous declaration.
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Currently, virDomainDeviceInfoParseXML() uses node->children
evaluation which is too verbose. Use XPath evaluation which is
nicer.
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Function virDomainDeviceInfoParseXML() will need it soon, because it
will be doing XPath evaluation.
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
If libvirtd is terminated before the node driver finishes
initialization, it can crash with a backtrace similar to the following:
Stack trace of thread 1922933:
#0 0x00007f8515178774 g_hash_table_find (libglib-2.0.so.0)
#1 0x00007f851593ea98 virHashSearch (libvirt.so.0)
#2 0x00007f8515a1dd83 virNodeDeviceObjListSearch (libvirt.so.0)
#3 0x00007f84cceb40a1 udevAddOneDevice (libvirt_driver_nodedev.so)
#4 0x00007f84cceb5fae nodeStateInitializeEnumerate (libvirt_driver_nodedev.so)
#5 0x00007f85159840cb virThreadHelper (libvirt.so.0)
#6 0x00007f8511c7d14a start_thread (libpthread.so.0)
#7 0x00007f851442bdb3 __clone (libc.so.6)
Stack trace of thread 1922863:
#0 0x00007f851442651d syscall (libc.so.6)
#1 0x00007f85159842d4 virThreadSelfID (libvirt.so.0)
#2 0x00007f851594e240 virLogFormatString (libvirt.so.0)
#3 0x00007f851596635d vir_object_finalize (libvirt.so.0)
#4 0x00007f8514efe8e9 g_object_unref (libgobject-2.0.so.0)
#5 0x00007f85159667f8 virObjectUnref (libvirt.so.0)
#6 0x00007f851517755f g_hash_table_remove_all_nodes.part.0 (libglib-2.0.so.0)
#7 0x00007f8515177e62 g_hash_table_unref (libglib-2.0.so.0)
#8 0x00007f851596637e vir_object_finalize (libvirt.so.0)
#9 0x00007f8514efe8e9 g_object_unref (libgobject-2.0.so.0)
#10 0x00007f85159667f8 virObjectUnref (libvirt.so.0)
#11 0x00007f84cceb2b42 nodeStateCleanup (libvirt_driver_nodedev.so)
#12 0x00007f8515b37950 virStateCleanup (libvirt.so.0)
#13 0x00005648085348e8 main (libvirtd)
#14 0x00007f8514352493 __libc_start_main (libc.so.6)
#15 0x00005648085350fe _start (libvirtd)
This is because the initial population of the device list is done in a
separate initialization thread. If we attempt to exit libvirtd before
this init thread has completed, we'll try to free the device list while
accessing it from the other thread. In order to guarantee that this
init thread is not accessing the device list when we're cleaning up the
nodedev driver, make it joinable and wait for it to finish before
proceding with the cleanup. This is similar to how we handle the udev
event handler thread.
The separate initialization thread was added in commit
9f0ae0b1.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1836865
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Previously, if xml node passed to the virXMLNodeContentString()
was not of type XML_ELEMENT_NODE, @ret could have caused a memory
leak because xmlNodeGetContent() works for other types of nodes
as well.
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Extend dirtyrate statistics for domGetStats to display the information
of a domain's memory dirty rate produced by domainStartDirtyRateCalc.
Signed-off-by: Hao Wang <wanghao232@huawei.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Introduce virDomainStartDirtyRateCalc API for start calculation of
a domain's memory dirty rate with a specified time.
Signed-off-by: Hao Wang <wanghao232@huawei.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We don't need to go to the trouble of telling users about existance of
insecure SASL mechanisms only to then say that they shouldn't be used.
We should only tell people about the GSSAPI mechanism for TCP sockets.
For the SCRAM mechanism we should be telling people about the SHA256
variant only, and also warning that the password database stores the
passwords in clear text.
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
If running libvirtd via systemd, it gets a 64 MB memlock limit, but if
running from the shell it will only get 64 KB on a Fedora 33 system.
The latter low limit causes any attempt to use BPF to fail and it is
not obvious why.
This improves the error message thus:
# virsh -c lxc:/// start sh
error: Failed to start domain 'sh'
error: internal error: guest failed to start: Failure in libvirt_lxc startup: failed to initialize device BPF map; locked memory limit for libvirtd probably needs to be raised: Operation not permitted
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The g_idle_add function adds a callback to the primary GMainContext.
To workaround the GSource unref bugs, we need to add our callbacks
to the GMainContext that is associated with the GSource being
unref'd. Thus code using the per-VM virEventThread must use its
private GMainContext.
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Somehow, command argument was not printed into debug logs. It is
imperative that all arguments are logged.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When connecting to the monitor, a timeout is calculated that is
bigger the more memory guest has (because QEMU has to allocate
and possibly zero out the memory and what not, empirically
deducted). However, when computing the timeout the @total_memory
mmember is accessed directly even though
virDomainDefGetMemoryTotal() should have been used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When switching to g_autoptr this was incorrectly changed from
'continue;' into 'return -1;' resulting into an error when user tries
to set vcpu_quota of running VM:
error: An error occurred, but the cause is unknown
Fixes: e4a8bbfaf2
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
In short, virXXXPtr type is going away. With big bang. And to
help us rewrite the code with a sed script, it's better if each
variable is declared on its own line.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The name is supposed to be virCapsGuestArchPtr not ..ptr.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The use of virXXXPtr is going away soon, therefore use 'virXXX *'
instead.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
What we are using really is heap allocated structure rather than
stack allocated. And for that it's better to use g_autoptr() +
G_DEFINE_AUTOPTR_CLEANUP_FUNC() combo, as Glib documentation for
g_auto() reads:
This is meant to be used with stack-allocated structures and
non-pointer types. For the (more commonly used) pointer
version, see g_autoptr().
This will be even more visible, when virSysinfoDefPtr type is
gone. Stay tuned.
Fixes: cee3a900a0
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The qemu shim spawns a separate thread in which the event loop is
ran. The virEventRunDefaultImpl() call is wrapped in a while()
loop, just like it should. There are few lines of code around
which try to ensure that domain is destroyed (when quitting) and
that the last round of event loop is ran after the
virDomainDestroy() call. Only after that the loop is quit from
and the thread quits.
However, if domain creation fails, there is no @dom to call
destroy over, the @quit flag is never set and while() never
exits. Set the flag regardless of @dom pointer.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1920337
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
The commandline generator for 'iothread' objects has a private
implementation of the properties. Convert it to JSON so that it can be
later validated.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While the 'sev0' sev-guest object will never be hotplugged, but we want
to generate it through JSON so that we'll be able to validate all
parameters of '-object' against the QAPI schema once 'object-add' is
qapified in qemu.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While the 'masterKey0' secret object will never be hotplugged we want to
generate it through JSON so that we'll be able to validate all
parameters of '-object' against the QAPI schema once 'object-add' is
qapified in qemu.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Commit 94e45d1042 broke exec-restart of virtlogd and virtlockd as the
code waiting for the daemon shutdown closed the daemons before
exec-restarting.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1912243
Fixes: 94e45d1042
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Recent changes which meant to fix daemon shutdown broke the exec-restart
capability of virtlogd and virtlockd, since the code actually closed all
the sockets and shut down all the internals.
Add virNetDaemonQuitExecRestart, which requests a shutdown of the
process, but keeps all the services open and registered since they are
preserved across the restart.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
This problem is reproducible only with secret driver. When
starting a domain via virt-qemu-run and both secret and
(nonexistent) root directory specified this is what happens:
1) virt-qemu-run opens "secret:///embed?root=$rootdir"
connection, which results in the secret driver initialization
(done in secretStateInitialize()). During this process, the
driver creates its own configDir (derived from $rootdir)
including those parents which don't exists yet. This is all
done with the mode S_IRWXU and thus results in the $rootdir
being created with very restrictive mode (specifically, +x is
missing for group and others).
2) now, virt-qemu-run opens "qemu:///embed?root=$rootdir" and
calls virDomainCreateXML(). This results in the master-key.aes
being written somewhere under the $rootdir and telling qemu
where to find it.
But because the secret driver created $rootdir with too
restrictive mode, qemu can't access the file (even though it
knows the full path) and fails to start.
It looks like the best solution is to pre-create the root
directory before opening any connection (letting any driver
initialize itself) and set its mode to something less
restrictive.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1859873
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
In theory, users might want to use a relative path as a root
directory for embed drivers. But in practice, nothing in driver
initialization (specifically QEMU driver since it's the only one
that supports embedding now), is prepared for that. Document and
enforce absolute paths.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1883725
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
'res->owners' is allocated to 'res->nOwners' elements, but unfortunately
'res->nOwners' doesn't contain the proper value until after the
allocation so 0 elements are allocated. The following loop which assumes
that the array has the right number of elements then accesses the
pointer out of bounds. The bug was also faithfully converted from
VIR_ALLOC_N to g_new0.
Fixes: 4a3d6ed5ee
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Recent refactor marked 'object' which is returned from the function as
autofree but forgot to use g_steal_pointer in the return statement to
prevent freeing it.
Fixes: 9a1651f64d
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Commit cb29e4e801 didn't take into account that the VM can be inactive
when it's destroyed. This means that the job would remain active also
when the VM became inactive.
To fix this properly:
1) Remove the bogus VM liveness check and early return
(reverts the aforementioned commit)
2) Conditionalize the stats assignment only when the stats object is
present
(properly fix the crash when VM dies when reconnecting)
3) end the asyncjob only when it was already set
(prevent corruption of priv->jobs_queued)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1937598
Fixes: cb29e4e801
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
'qemuBackupJobTerminate' needs the API flags to see whether
VIR_DOMAIN_BACKUP_BEGIN_REUSE_EXTERNAL. Unfortunately when called via
qemuProcessReconnect()->qemuProcessStop() early (e.g. if the qemu
process died while we were reconnecting) the job is cleared temporarily
so that other APIs can be called. This would mean that we couldn't clean
up the files in some cases.
Save the 'apiFlags' inside the backup object and set it from the
'qemuDomainJobObj' 'apiFlags' member when reconnecting to a VM.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
g_variant_new_parsed uses '%t' for a uint64_t rather than printf-like
%llu. Additionally ensure that the passed value is a uint64_t since the
argument used is a 'unsigned int'.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1937287
Fixes: bf5f2ed09c
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
The function is now unused and motivated users to write crazy parsers
which were hard to understand, had pointless error paths just to avoid
few memory allocations.
Remove the function as we're fine with g_strndup and virStrcpy.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use g_strsplit to split the string and avoid use of stack'd strings.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The problem is that g_get_host_name() caches the hostname in a
thread local variable. Therefore, it doesn't reflect any
subsequent hostname changes. While this might be acceptable for
logs where the hostname is printed exactly once when the libvirtd
starts up, it is not optimal for virGetHostnameImpl() which is
what our public virConnectGetHostname() API calls. If the
hostname at the moment of the first API invocation happens to
start with "localhost" or contains a dot, then no further
hostname changes will ever be reflected.
This reverts 26d9748ff1, partially.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We already assume that 'retr_passphrase.result' is a string, thus we can
use virStrcpy instead.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use g_strndup with a freed buffer instead of the more complex approach
using virStrncpy.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Make the temporary string an autofree-ing pointer and copy the contents.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Rewrite so that the parser doesn't use virStrncpy by employing
g_strsplit.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Copy the input string so that we don't have to use a static buffer and
virStrncpy.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
With this, XML fails if config video type 'ramfb' contains
address, since address is not supported for 'ramfb' video
devices. Previously it didn't raise error.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1891416
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Switch to using the 'g_auto*' helpers.
Signed-off-by: Yi Li <yili@winhong.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
This introduces support for the QEMU audio settings that are common to
all audio backends. These are expressed in the QAPI schema as settings
common to all backends, but in reality some backends ignore some of
them. For example, some backends are output only. The parser isn't
attempting to apply restrictions that QEMU itself doesn't apply.
<audio id='1' type='pulseaudio'>
<input mixingEngine='yes' fixedSettings='yes' voices='1' bufferLength='100'>
<settings frequency='44100' channels='2' format='s16'/>
</input>
<output mixingEngine='yes' fixedSettings='yes' voices='2' bufferLength='100'>
<settings frequency='22050' channels='4' format='f32'/>
</output>
</audio>
The <settings> child is only valid if fixedSettings='yes'
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The -audiodev argument is replacing the QEMU_AUDIO_DRV env variable (and
its relations).
Sadly we still have to use the SDL_AUDIODRIVER env variable because that
wasn't mapped into QAPI schema.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The -audiodev arg is a new way to configure audio devices in QEMU to
replace the QEMU_AUDIO_DRV env variable. This arg is not visible in
the "query-command-line-options" output since it is entirely QAPI
driven, not QemuOpts. It also isn't in "query-qmp-schema" though
since there's no QMP command that uses the Audiodev type yet.
So probe for the existance of this feature by looking for the
-vnc "audiodev" property. This won't let us determine which
precise audio backends QEMU has been built with, but for now
that's no worse than with env variables today.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Currently the QEMU driver secretly sets the QEMU_AUDIO_DRV env variable
- VNC - set to "none", unless passthrough of host env variable is set
- SPICE - always set to "spice"
- SDL - always passthrough host env
- No graphics - set to "none", unless passthrough of host env variable is set
The setting of the QEMU_AUDIO_DRV env variable is done in the code which
configures graphics.
If no <audio> element is present, we now auto-populate <audio> elements
to reflect this historical default config. This avoids need to set audio
env when processing graphics.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Currently the QEMU driver secretly sets the QEMU_AUDIO_DRV env variable
depending on how <graphics> are configured.
This introduces support for configuring audio backends from the <audio>
elements in the XML config.
The existing default behaviour is now only used if no <audio> element is
present.
All except the 'jack' audio driver are supported via QEMU's old env
variable config.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The virDomainDefFindAudioForSound only takes a virDomainSoundDefPtr as
its arg, but we want to use the same functionality for VNC graphics.
In addition if audio ID is zero, then we want to return the first
available audio backend.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Validate that if a non-zero audio ID is given for <sound> or <graphics>
elements, it must map to an <audio> backend that exists.
Validate that audio IDs given in <audio> are unique.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
When there are multiple <audio> backends specified, it is possible to
assign a specific one to the VNC server using
<graphics type='vnc'...>
<audio id='1'/>
</graphics>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The current <audio> element only allows an "OSS" audio backend, as this
is all that BHyve needed. This is now extended to cover most QEMU audio
backends. These backends all have a variety of attributes they support,
but this initial impl does the bare minimum, relying on built-in
defaults for everything. The only QEMU backend omitted is "dsound" since
the libvirt QEMU driver is not built on Windows platforms.
The SDL audio driver names are based on the SDL 2.0 drivers. It is not
intended to support SDL 1.2 drivers.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
To prepare for the introduction for more backend specific audio options,
move the OSS options into a dedicated struct and introduce separate
helper methods for parse/format/free.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The check for ICH6 || ICH9 is repeated in many places in the code. The
new virDomainSoundModelSupportsCodecs() method provides a helper to
standardize this check.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The attributes on the elements are optional, so we should not force the
elements themselves to be present, especially since we omit them when
formating the XML thus breaking round-tripping.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Check for varuous mandatory elements and improve error message
clarity
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This patch adds delay time (steal time inside guest) to libvirt
domain per-vcpu stats. Delay time is an important performance metric.
It is a consequence of the overloaded CPU. Knowledge of the delay
time of a virtual machine helps to understand if it is affected and
estimate the impact.
As a result, it is possible to react exactly when needed and
rebalance the load between hosts. This is used by cloud providers
to provide quality of service, especially when the CPU is
oversubscribed.
It's more convenient to work with this metric in a context of a
libvirt domain. Any monitoring software may use this information.
Signed-off-by: Aleksei Zakharov <zaharov@selectel.ru>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The preprocessor macro we use to check whether we're on Linux
has not been spelled properly, and so we will always report the
error message intended for other platforms.
Fixes: 879bcee08c
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Failure of 'qemuMigrationSetDBusVMState' would jump to 'exit_monitor'
but the function isn't called inside of the monitor context.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
When generating TC rules for domain's outbound traffic, Libvirt
will use the 'average' as the default for 'burst' - it's been
this way since the feature introduction in v0.9.4-rc1~22. The
reason is that 'average' considers 'burst' for policing. However,
when parsing its command line TC uses an unsigned int (with
overflow detection) to store the 'burst' size. This means, that
the upper limit for the value is UINT_MAX, well UINT_MAX / 1024
because we are putting the value in KiB onto the command line.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1912210
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
Up until now we've implicitly relied on the fact that failures
reported from this function were simply ignored, but that's
about to change and so we need a proper mock.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
That's more consistent with our usual naming convention.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
This behavior reflects the needs of the QEMU driver and has no
place in a generic module such as virProcess.
Thanks to the changes made with the previous commit, it is now
safe to remove these checks and make all virProcessSetMax*()
functions finally behave the same way.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The current code is written under the assumption that, for all
limits except the core size, asking for the limit to be set to
zero is a no-op, and so the operation is performed
unconditionally.
While this is the behavior we want for the QEMU driver, the
virCommand and virProcess facilities are generic, and should not
implement this kind of policy: asking for a limit to be set to
zero should result in that limit being set to zero every single
time.
Add some checks in the QEMU driver, effectively moving the
policy where it belongs.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Currently this only happens for the core size, but we want the
behavior to be consistent for other limits as well.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
qemuProcessLaunch() is the correct place to set process limits,
and in fact is where we were dealing with almost all of them,
but the memory locking limit was handled in
qemuBuildCommandLine() instead for some reason.
The code is rewritten so that the desired limit is calculated
and applied in separated steps, which will help with further
changes, but this doesn't alter the behavior.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Doing this now will make the next changes nicer.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
These functions abstract part of the existing logic, which is
the same in all virProcessSetMax*() functions, and changes it
so that which underlying syscall is used depends on their
availability rather than on the context in which they are
called: since prlimit() and {g,s}etrlimit() have slightly
different requirements, using the same one every single time
should make for a more consistent experience.
As part of the change, we also remove the special case for
passing zero to virProcessSetMax*() functions: we have removed
all callers that depended on that functionality in the previous
commit, so this is now safe to do and makes the semantics
simpler.
This commit is better viewed with 'git show -w'.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Currently, the functions accept either an explicit pid or zero,
in which case the current process should be modified: the latter
might sound like a convenient little feature, but in reality
obtaining the pid of the current process is a single additional
function call away, so it hardly makes a difference.
Removing the few cases in which we're passing zero will allow us
to simplify and improve the functions later.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
This allows the VNC client user to perform a shutdown, reboot and reset
of the VM from the host side.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The <graphics type="vnc" .... powerControl="yes"/> option instructs the
VNC server to enable an extension that lets the client perform a
graceful shutdown, reboot and hard reset.
This is enabled by default since it cannot be assumed that the VNC
client user has administrator rights over the guest OS. In the case
where the VNC user is a guest administrator though, it is reasonable
to allow direct power control host side too.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Calling a stub should always result in ENOSYS being raised,
regardless of what arguments are passed to it.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We're going to change their behavior, so it's good to have the
current one documented to serve as baseline.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
For reasons unknown, when rewriting this code and dropping
libdevmapper I've mistakenly used incorrect length of dm.name. In
linux/dm-ioctl.h the dm_ioctl struct is defined as follows:
#define DM_NAME_LEN 128
struct dm_ioctl {
...
char name[DM_NAME_LEN]; /* device name */
...
};
However, when copying string into this member, DM_TABLE_DEPS was
used, which is defined as follows:
#define DM_TABLE_DEPS _IOWR(DM_IOCTL, DM_TABLE_DEPS_CMD, struct dm_ioctl)
After decryption, this results in the following size: 3241737483.
Fixes: 2249455654
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Tools depend on keycode generated sources, so declare that as an
explicit dependency, otherwise it might fail with:
../tools/virsh-completer-domain.c:35:10: fatal error: 'virkeynametable_linux.h' file not found
^~~~~~~~~~~~~~~~~~~~~~~~~
Fixes: b0f4cf25a6
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Commit bbc25f0d03 juggled around some
error reporting. Unfortunately virFirewallApply tries to report the
errno stored in the firewall object and we'd try to do that when the
firewall object is NULL too. Report EINVAL if 'firewall' is NULL.
Found by Coverity.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
There's an optimization in virBufferAdd which returns early when the
length of the added string is 0 (given that auto-indent is disabled).
The optimization causes inconsistent behaviour between these two cases:
virBufferAdd(buf, "", 0); // this doesn't initialize the buffer
and
virBufferAdd(buf, "", -1); //this initializes the buffer
Since using an empty string is used to prime the buffer to an empty
string it can be confusing. Remove the optimization.
This fixes such a wrong initialization done in x86FeatureNames.
Note that our code in many places expects that if no virBuffer APIs are
used on a buffer object, then NULL should be retured, so we can't always
prime the buffer to an empty string.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Simplify the logic picking which element form to format by using
virBuffers for the partial properties and virXMLFormatElement for
combining them.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use an allocated buffer for 'cpu_header' so that g_strdup(_printf) can
be used to fill it.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Use a dynamic string helper so that we don't have to calculate the
string lengths and then iterate from the rear.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We can remove the check that 'idx' is negative by forcing callers to
pass unsigned numbers, which they do already or have a check that 'idx'
is positive.
This in turn allows us to remove most return value NULL checks.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The function is used to automatically feed a buffer into a pipe which
can be used by the command to read contents of the buffer.
Rather than passing in a pipe, let's create the pipe inside
virCommandSetSendBuffer and directly associate the reader end with the
command. This way the ownership of both ends of the pipe will end up
with the virCommand right away reducing the need of cleanup in callers.
The returned value then can be used just to format the appropriate
arguments without worrying about cleanup or failure.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The function can't fail nowadays. Remove the return value and adjust the
only caller which ensures that @cmd is non-NULL and @fd is positive.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Extract the check and reporting of error from the individual virCommand
APIs into a separate helper. This will aid future refactors.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
If WITH_PIPE2 is not defined we attempt to set the pipe to nonblocking
operation after they are created. We errorneously rewrote the existing
error message on failure to do so or even reported an error if quiet
mode was requested.
Fixes: ab36f72947
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The function is constructing an error message from a prefix and the
contents of the qemu log file. Marking just two string modifiers as
translatable is pointless and will certainly confuse translators.
Remove the marking and add a comment which bypasses the
sc_libvirt_unmarked_diagnostics check.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Now that error message formatting doesn't use fixed size buffers we can
drop the math for calculating the maximum chunk of log to report in the
error message and use a round number. This also makes it obvious that
the chosen number is arbitrary.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use of VIR_ERROR_MAX_LENGTH is actually misleading to the readers
because it implies that the strings in virError are 1024 bytes at most.
That isn't true at least for the 'message' field as it's constructed
from concatenating the detail string which (was) max 1024 bytes with
the string variant of the error code without limiting to 1024.
Use a local copy for declaring the struct for error transport with a
comment so that's obvious that it's a local decision to use 1k buffers.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Some error message reporting functions already have allocated buffers
which were used to format the error message, so copying the strings is
redundant.
Extract the internals from 'virRaiseErrorFull' to
'virRaiseErrorInternal' which takes allocated strings as arguments and
steals them, so that callers can reuse the buffers.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This was (probably) a relict from times when we cared about OOM
conditions and the possibility to report the error. Nowadays it doesn't
make sense as virRaiseErrorFull will do an allocated copy of the strings
and also concatenate the error message prefix with the detail which
doesn't guarantee that the result will be less than 1024 chars.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use glib functions to do the relative name lookup instead of manual
assembly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Passing 'strlen(src)' for length makes it equivalent to virStrcpy.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
virStrncpy was called with -1 for length of the copied source which is
equivalent to virStrcpy.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We want a (possibly truncated) copy of the full source string so
virStrcpy is a better fit.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
15 out of 72 invocations of virStrcpy(Static) ignore the return value as
it's either impossible to fail or in certain cases a truncated copy is
still good enough. Unfortunately virStrcpy doesn't copy anything in
such case as the checks are done first.
Fix this by using g_strlcpy for the implementation and removing
G_GNUC_WARN_UNUSED_RESULT from the function so that callers can decide
when it's okay.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Similarly to the crash workaround:
commit 0db4743645
Author: Daniel P. Berrangé <berrange@redhat.com>
Date: Tue Jul 28 16:52:47 2020 +0100
util: avoid crash due to race in glib event loop code
we need to do this in the other event loop as crash in that one was also
reported:
https://bugzilla.redhat.com/show_bug.cgi?id=1931331
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This way it can be used from other places as well.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Generated using the following spatch:
@@
expression path;
@@
- virFileMakePath(path)
+ g_mkdir_with_parents(path, 0777)
However, 14 occurrences were not replaced, e.g. in
virHostdevManagerNew(). I don't really understand why.
Fixed by hand afterwards.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
These functions are identical. Made using this spatch:
@@
expression path, mode;
@@
- virFileMakePathWithMode(path, mode)
+ g_mkdir_with_parents(path, mode)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Refractoring includes:
* removal of VIR_FREE
* inversion of the condition
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
By:
* declaration of an autofreed variable in for loop
* use of a new variable
* removal of VIR_FREE
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
VIR_DOMAIN_HYPERV_STIMER happens to have the same numerical value as
VIR_DOMAIN_FEATURE_HYPERV, resulting in the if-block to always being
executed when a "<hyperv>" tag is found, whether or not it actually
contained a "<stimer>" tag. This had no ill effects, as virXPathNodeSet()
would simply return 0 if that tag does not exist.
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Variables using `g_autofree` should not be manually VIR_FREE'd and reused.
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Previous commit removed all usage of this function so we can remove it.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Now that we enforce the cpu.shares range kernel will no longer silently
change the value that libvirt configures so there is no need to read
the value back to get the actual configuration.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Before the conversion to using systemd DBus API to set the cpu.shares
there was some magic conversion done by kernel which was documented in
virsh manpage as well. Now systemd errors out if the value is out of
range.
Since we enforce the range for other cpu cgroup attributes 'quota' and
'period' it makes sense to do the same for 'shares' as well.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Commit a208176ca1 introduced this feature
with an incorrect "svme-addr-check" spelling.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Tim Wiederhake <twiederh@redhat.com>
Commit 887dd0d331 caused a small regression in NodeDeviceDetach in the libxl
driver when the 'driver' parameter is not specified. E.g.
# virsh nodedev-detach pci_0000_0a_10_0
error: Failed to detach device pci_0000_0a_10_0
error: An error occurred, but the cause is unknown
If the driver name is not specified, NULL is passed to
virDomainDriverNodeDeviceDetachFlags, in which case virPCIDeviceSetStubDriver
is never called to set the stub to pciback. Fix it by setting the driver to
"xen" if it is not specified when invoking NodeDeviceDetach.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Introduced in QEMU 6.0.0 by 623972ceae091b31331ae4a1dc94fe5cbb891937
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Trying to report an OOM error is pointless since our infrastructure to
report error needs to allocate memory to report the error.
In addition our code mistakenly reported OOM errors even in cases where
a function could fail for another reason, which would make issues harder
to debug.
Remove the virReportOOMError and backend so that programmers are forced
to think about what can happen. In case when there's another failure
possible a specific error should be reported and otherwise a direct
abort() is better since the logger would abort on g_new anyways.
This patch also removes the syntas-check which forces use of
virReportOOMError instead of using VIR_ERR_NO_MEMORY with other
functions. This allows possible future use when we'd end up in a
situation where trying to recover from an OOM would make sense, such as
when attempting to allocate a massive buffer.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
The function has also non-OOM failure case when the passed string has 0
length, so reporting OOM error is not correct.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
OOM isn't the only failure glfs_new can encounter. Report an error which
might give more insight. libgfapi seems to be setting errno but
reporting a system error migt be misleading.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
The yajl library returns a wide range of error codes so reporting OOM on
any failure is wrong. In case the error was really based by memory issue
the error reporting will probably cause an abort anyways. Change the
error message so that we know that it happened in JSON at least.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
