Our code relies on mount events propagating into the namespace we
create for a domain. However, there's one caveat. In v8.8.0-rc1~8
I've tried to make us detect differences in mount tables between
the namespace in which libvirtd runs and the domain namespace.
This is crucial for any mounts that happen after the domain was
started (for instance new hugetlbfs can be mounted on say
/dev/hugepages1G).
Therefore, we take a look into /proc/$(pgrep qemu)/mounts to see
what filesystems are mounted under /dev. Now, since we don't
umount the original /dev, just mount a tmpfs over it, we get all
the events (e.g. aforementioned hugetlbfs mount on
/dev/hugepages1G), but we are not really able to access it
because of the tmpfs that's placed on top. This then confuses our
algorithm for detecting which filesystems are mounted (the
algorithm is implemented in qemuDomainGetPreservedMounts()).
To break the link between host's and guest's /dev we just need to
umount() the original /dev in the namespace. Just before our
artificially created tmpfs is moved into its place.
Fixes: 46b03819ae
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2151869#c6
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Inside of qemuCaps (for the corresponding accelerator) we have
full host CPU expansion stored, among with supported Hyper-V
Enlightenments. To report them in the domain capabilities, we
just have to pick those starting with "hv-" and see if we know
them.
You may notice that neither of our domaincapsdata test shows any
enlightenment. This is because the test works by parsing
corresponding qemucapabilitiesdata/caps_*.xml file and none of
these store the full host CPU expansion (hostCPU.fullQEMU)
because that is runtime piece of information and not formatted
into virQEMUCaps XML.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1717611
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Now that we have qemuMonitorGetCPUModelExpansion() aware of
Hyper-V Enlightenments, we can start querying it. Two conditions
need to be met:
1) KVM is in use,
2) Arch is either x86 or arm.
It may look like modifying the first call to
qemuMonitorGetCPUModelExpansion() inside of
virQEMUCapsProbeQMPHostCPU() would be sufficient but it is not.
We really need to ask QEMU for full expansion and the first call
does not guarantee that.
For the test data, I've just copied whatever
'query-cpu-model-expansion' returned earlier, therefore there are
no hv-* props. But that's okay - the full expansion is not stored
in cache (and thus not formatted in
tests/qemucapabilitiesdata/caps_*.replies files either). This is
purely runtime thing.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
This continues and finishes propagation of the @hv_passthrough
argument started in the previous commit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Apart from setting @migratable prop to the
query-cpu-model-expansion command, we will need @hv-passthrough
so that we can query for expansion of Hyper-V Enlightenments
supported on the current host. The idea is to run:
{
"execute": "query-cpu-model-expansion",
"arguments": {
"type": "full",
"model": {
"name": "host",
"props": {
"hv-passthrough": true
}
}
}
}
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The virDomainCapsEnumFormat() function does not return anything
but zero and none of its callers is interested in the failure
anyways. Switch its return type from integer to void.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
We are formatting <enum/> element and its children using
virBufferAddLit(), virBufferAsprintf(), virBufferAdjustIndent(),
etc. Well, we can avoid that when switching to
virXMLFormatElement().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In a recent commit, when ditching virXPathULong() the parsing of
<selfvers/> was changed. But it was changed to virXMLPropUInt()
which is not correct because the value we're interested in is not
in an attribute but element itself.
Fixes: a3c7426839
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
Since we really only need to handle key skipping in the top level object
the caller doesn't at this point even pass it to the array formatting
helper function. Remove the unused argument.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Skipping of a specific key is needed only for the top level object to
specially handle the object type. We must not pass it to any recursed
printing of nested objects as skipping keys there might be surprising
and also is unhandlable later when formatting the commandline.
Until now this did not pose a problem but was discovered when adding a
new netdev backend which has a nested config object which also has the
'type' key which was being skipped.
Modern usage will prefer JSON directly but fix the commandline generator
to prevent surprises.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The @hash variable inside of virQEMUCapsProbeQMPHostCPU() is used
only within a block, but declared at the beginning of the
function. Bring the variable declaration into the said block.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
There's nothing qemu specific about
qemuDomainCapsFeatureFormatSimple() and in fact, the function
lives in hypervisor agnostic location and thus mustn't have qemu
prefix.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
There are some network FSs (ceph, CIFS) that propagate XATTRs
properly and thus SELinux labels too. In such case using dynamic
seclabels would get in the way of migration as new seclabel is
assigned to the domain on the destination and thus two processes
with different labels (the source and the destination QEMU/helper
process) would try to access the same file. One of them is
necessarily going to be denied access.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
After previous cleanup this function is no longer used and thus
can be dropped.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When starting swtpm binary, the qemuSecurityStartTPMEmulator() is
called which sets seclabel on the TPM state and then uses
qemuSecurityCommandRun() to execute the swtpm binary with proper
seclabel. Well, the aim is to ditch
qemuSecurityStartTPMEmulator() because it entangles two distinct
operations. Just call functions for them separately.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
If swtpm binary fails to start after successful exec() (e.g. it
fails to initialize itself), the seclabels set in
qemuSecurityStartTPMEmulator() are not restored. This is due to
lacking qemuSecurityRestoreTPMLabels() call in the error path.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Now that we have qemuSecurityRestoreTPMLabels() we might as well
have qemuSecuritySetTPMLabels(). The aim here is to remove
qemuSecurityStartTPMEmulator() which couples two separate things
into a single function call.
Therefore, introduce qemuSecuritySetTPMLabels() which does only
set seclabels on the TPM state.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The qemuSecurityCleanupTPMEmulator() function calls
virSecurityManagerRestoreTPMLabels() and thus the proper name is
qemuSecurityRestoreTPMLabels(). Rename it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Currently, qemuSecurityCleanupTPMEmulator() returns nothing which
means a caller (well, there's only one - qemuExtTPMStop()) can't
produce a warning when restoring seclabels on TPM state failed.
True, qemuSecurityCleanupTPMEmulator() does report a warning
itself, but only in one specific error path.
Make the function return an integer, just like the rest of
qemuSecurity*Restore() functions.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
qemu is about to deprecate the '-no-hpet' option in favor of configuring
the timer via '-machine'.
Use the QEMU_CAPS_MACHINE_HPET capability to switch to the new syntax
and mask out the old QEMU_CAPS_NO_HPET capability at the same time to
prevent using the old syntax.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
The capability represents that qemu accepts the configuration of the
HPET timer via -machine hpet=on/off rather than the
soon-to-be-deprecated '-no-hpet' option.
The capability is detected from 'query-command-line-options' which
recently added the 'hpet' option.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Add test data based on qemu commit v7.2.0-333-g222059a0fc
- query-command-line-options now reports more accurate data
- machine types for the 8.0 cycle were added
- vhost-vdpa device support was added
- default value of 'noreboot' changed from 'true' to 'false'
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Currently translated at 22.8% (2369 of 10368 strings)
Translation: libvirt/libvirt
Translate-URL: https://translate.fedoraproject.org/projects/libvirt/libvirt/fi/
Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com>
Signed-off-by: Jan Kuparinen <copper_fin@hotmail.com>
That way it actually fits with what the condition checks for.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This way we actually check for the proper error, not any error like invalid JSON
format.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
It just so happens that our JSON snippets in
qemucapabilitiesdata/*.replies files are separated by an empty
line. These empty lines are then overwritten to make a single
line JSON. Nevertheless, the line counter @line is not
incremented which then leads to a misleading numbers in errors.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Our secret driver divides secrets into two groups: ephemeral
(stored only in memory) and persistent (stored on disk). Now, the
aim of ephemeral secrets is to define them shortly before being
used and then undefine them. But 'shortly before being used' is a
very vague time frame. And since we default to socket activation
and thus pass '--timeout 120' to every daemon it may happen that
just defined ephemeral secret is gone among with the virtsecretd.
This is no problem for persistent secrets as their definition
(and value) is restored when the virtsecretd starts again, but
ephemeral secrets can't be restored.
Therefore, we could view ephemeral secrets as active objects that
the daemon manages and thus inhibit automatic shutdown (just like
hypervisor daemons do when a guest is running).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Xen 4.17 has strict parsing of 'soundhw' option that allows only
specific values (instead of passing through any value directly to
qemu's -soundhw option, it uses -device now). For 'intel-hda' audio
device, it requires "hda" string. "hda" works with older libxl too.
Other supported models are the same as in libvirt XML.
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Xen supports only subset of libvirt's sound devices, and starting with
Xen 4.17 it is enforced by libxl. Verify it early.
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Automatically free 'sec' and remove the 'cleanup' section and 'ret'
variables.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
'virStorageBackendRBDRADOSConfSet' logs its arguments but it's also
used to set the RBD secret/key.
All the security theatre with securely erasing the string we do to fetch
the secret would be quite pointless if we log it thus introduce
virStorageBackendRBDRADOSConfSetQuiet and use it to avoid logging the
password.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The initialization vector is not optional thus we also don't need to
check whether the caller passed it in. Additionally we can use c99
initializers for the gnutls_datum_t structs.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
'gnutls_datum_t' simply holds pointers to the encryption key and its
length. There's absolutely no point in securely erasing that.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Introduce a new backend type 'external' for connecting to a swtpm daemon
not managed by libvirtd.
Mostly in one commit, thanks to -Wswitch and the way we generate
capabilities.
https://bugzilla.redhat.com/show_bug.cgi?id=2063723
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
QEMU 7.2 was released, update the capabilities data to the final state.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
List the various options so that the most likely ones come
first.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Users are likely more interested in the main deployment
scenarios than in the detailed list of every existing RPM
package. Reorder sections accordingly.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Since the takeover of the bird site, the bulk of tech people who want
a more friendly and inclusive media site have jumped over to Mastodon.
With its decentralized nature, there's no one replacement that captures
everything, but the fosstodon.org site is a topic relevant choice.
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The storage-backend/ and storage-file/ directories are currently
considered unowned by RPM. Have the libvirt-daemon package take
ownership of them, just as it already owns the connection-driver/
and lock-driver/ directories.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
In the past, the preferred policy
(VIR_DOMAIN_NUMATUNE_MEM_PREFERRED) required exactly one (host)
NUMA node. This made sense because:
1) the libnuma API - numa_set_preferred() allowed exactly one
node, because
2) corresponding kernel syscall (__NR_set_mempolicy) accepted
exactly one node (for MPOL_PREFERRED mode).
But things have changed since then. Firstly, kernel introduced
new MPOL_PREFERRED_MANY mode (v5.15-rc1~107^2~21) which was then
exposed in libnuma as numa_set_preferred_many() (v2.0.15~24).
Fortunately, libnuma also exposes numa_has_preferred_many() which
returns whether the kernel has support for the new mode (1) or
not (0).
Putting this all together, we can lift our check for sufficiently
new kernel and libnuma.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2151064
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>