The 'auto-read-only' blockdev option is available in all supported qemu
versions so we can remove the migration hack which disabled it.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We want at least one file to always be present, so that it can
serve as a pointer for users. Ensure that this is the case by
unconditionally using the value of the respective keys.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
It's somewhat confusing that some of the services have a
corresponding foo.service.extra.in and foo.socket.extra.in, some
have just one of the two, and some have neither.
In order to make things more approachable, make sure that both
files exists for each service.
In most cases the extra units are currently unused, so they will
just contain a comment briefly explaining their purpose and
pointing users to meson.build, where they can find more
information. The same comment is also added to the top of
extra units that already have some contents in them for
consistency.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Now that the underlying script is able to merge an arbitrary
number of units into the base template, expose this possibility
in the build system.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
It uses custom templates which already hardcode the correct
value.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Starting v18, cloud-hypervisor supports serial and console devices in
parallel. Drop related check based on ch version.
Signed-off-by: Praveen K Paladugu <prapal@linux.microsoft.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Starting with v28.0 cloud-hypervisor requires the use of "payload" api to pass
kernel, initramfs and cmdline options. Extend ch driver to use the new
api based on ch version.
Signed-off-by: Praveen K Paladugu <prapal@linux.microsoft.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The function virHostCPUGetPhysAddrSize was introduced with commit be1b7d5b18
fails on architectures other than x86 and SuperH. The commit 8417c1394c
fixed the issue only for s390 but the problem is still seen on other
architectures like ppc which does not report Physical address size in their
cpuinfo output.
command:
systemctl restart libvirtd.service
Output :
<snip>
dnsmasq[2377]: read /var/lib/libvirt/dnsmasq/default.addnhosts - 0
addresses
dnsmasq-dhcp[2377]: read /var/lib/libvirt/dnsmasq/default.hostsfile
libvirtd[3163]: libvirt version: 9.8.0
libvirtd[3163]: hostname: xxxxxxxxxx
libvirtd[3163]: internal error: Missing or invalid CPU address size in
/proc/cpuinfo
libvirtd.service: Deactivated successfully.
</snip>
This patch fixes this issue by returning the size=0 for architectures
other than x86 and SuperH.
Signed-off-by: Narayana Murty N <nnmlinux@linux.ibm.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
vmDef->fss[i]->src->path may be NULL,
so check is needed before passing it to VIR_DEBUG.
Also removed checking vmDef->fss[i]->src for NULL, since it may not be NULL.
Fixes: 57487085dc ("lxc: don't try to reference NULL when mounting filesystems")
Signed-off-by: Dmitry Frolov <frolov@swemel.ru>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Currently, libvirt doesn't send events when devices are attached,
detached or updated. Thus, any services that listen to events are
unaware of the change to persistent config.
Signed-off-by: Fima Shevrin <efim.shevrin@virtuozzo.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Considering that at the virSecuritySELinuxSetFilecon() function can only
return 0 or -1 and so does the virSecuritySELinuxFSetFilecon(), the check
for '1' at the end of virSecuritySELinuxSetImageLabelInternal() is
effectively a dead code. Drop it.
Co-developed-by: sdl.qemu <sdl.qemu@linuxtesting.org>
Signed-off-by: Sergey Mironov <mironov@fintech.ru>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
While the name itself doesn't matter, this rename is done to prove that
all places using 'nodeformat' were converted to the appropriate
accessors.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The persistent bitmaps are stored in the format layer, using 'effective'
bitmap name is the most reasonable approach in this case.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The blockjob, NBD export and setup of the cookie data all care about the
effective nodename.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The frontend device needs to access the blocks directly so it cares
about the effective nodename.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In most cases the bitmap operations are relevant only on qcow2 images
thus the 'format' layer will be present. Although in certain specific
cases temporary bitmaps can be created on top of other images as well,
thus we use the 'effective' bitmap name in all cases for bitmap
operations.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
I case of statistics we're interested in the statistics of the effective
bitmap whatever it happens to be.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The disk backend setup code is concerned only about the effective
nodename. Doing this conversion will also simplify further changes
needed to drop the 'raw' layer in cases when it's not really needed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The code setting the nodenames needs to use the 'true' nodename of the
format layer.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Convert the main -blockdev JSON object setup code to use the new
accessors. In these we use mainly the real 'format' layer node name.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use the effective nodename for naming the job as we use that one now.
It doesn't matter too much which one we pick, because it's used just for
the name of the job, which we preserve in the status XML.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Both modified cases in this patch require the effective nodename as they
deal with the data being backed up.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Introduce a set of accessors, which return node names based on
semantics. This will allow to us to modify how we setup the backing
chain in cases when e.g. the format driver can be omitted, without
breaking all the code.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While the name itself doesn't matter, this rename is done to prove that
all places using 'nodestorage' were converted to the appropriate
accessors.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We need to keep setting the block threshold on the real storage layer
per semantics of the API. Use the appropriate accessor.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In all cases we want to probe stats from the 'storage' layer as we're
interested in the 'threshold' value, which we set there.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use the new nodename accessors for any storage layer helper object.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Refactor the code settin up data structures used to attach/detach disks
and SCSI hostdevs.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Refactor the code which assigns the 'storage' layer nodenames for disks.
scsi hostdevs and pflash backend.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We need to use the 'effective' storage nodename (one which includes the
optional storage slice 'raw' intermediate layer) in the code which
formats the 'format' layer props.
All other cases need the real storage driver nodename as they either
generate the 'storage' layer props, or the storage slice, which refers
to the proper storage backend.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use the new accessors in the private XML formatters and parsers and the
recovery code.
Specifically in all instances we use the proper (not effective) storage
nodename. In the virStorageSource private data it is what we need to
store. In blockjobs status XML it simply serves us to find the
appropriate 'virStorageSource' struct so using the storage layer node
name is simpler.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use 'qemuBlockStorageSourceGetEffectiveStorageNodename' in all the JSON
props formatters for setting up a 'blockdev-create' job of a format
layer.
In case of the blockjob name designator we're okay to use just the
storage layer nodename as that serves only to find the appropriate
entry.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The lookup by nodename requires the proper storage nodename which we use
also in status XML.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Introduce a set of accessors, which return node names based on
semantics. This will allow to us to modify how we setup the backing
chain in cases when e.g. the format driver can be omitted, without
breaking all the code.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use qemuBlockStorageSourceGetFormatProps as it formats the properties of
the 'format' driver in qemu. Adjust the comment which was hinting
otherwise.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Restructure the conditions so that we can use virJSONValueObjectAdd with
a clearer logic for backing store control.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use the node name of the storage access driver to identify the block job
volumes. This will prepare the blockjob code for the possibility that the
format layer may be missing. Our lookup code can find either of them,
thus we can safely switch.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The 'virt-aa-helper' process gets a XML of the VM it needs to create a
profile for. For a disk type='volume' this XML contained only the
pool and volume name.
The 'virt-aa-helper' needs a local path though for anything it needs to
label. This means that we'd either need to invoke connection to the
storage driver and re-resolve the volume. Alternative which makes more
sense is to pass the proper data in the XML already passed to it via the
new XML formatter and parser flags.
This was indirectly reported upstream in
https://gitlab.com/libvirt/libvirt/-/issues/546
The configuration in the issue above was created by Cockpit on Debian.
Since Cockpit is getting more popular it's more likely that users will
be impacted by this problem.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Re-translating the disk source pools when reconnecting to a VM makes no
sense as the volume might have changed or pool became inactive. The VM
still uses the original volume though. Failing to re-translate the pool
also causes the VM to be killed.
Fix this by storing the original translation in the status XML.
Resolves: https://issues.redhat.com/browse/RHEL-7345
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Re-translating a disk type='volume' definition from a storage pool is
not a good idea in cases when the volume might have changed or we might
not have access to the storage driver.
Specific cases are if a storage pool is not activated on daemon restart,
then re-connecting to a VM fails, or if the virt-aa-helper program tries
to setup labelling for apparmor.
Add a new flag which will preserve the translated data in the
definition.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
If a disk definition was already translated re-doing it makes no sense.
Skip the translation if the 'actualtype' is already populated.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Register autoptr cleanup function for virStorageSourcePoolDef and
refactor the parser to simplify the logic.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use proper enum type and refactor the formatter accordingly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When loading a secret value fails, the control jumps over to the
'cleanup' label where explicit call to virSecretDefFree()
happens. This is unnecessary as the corresponding variable can be
declared with g_autoptr() after which all error paths can just
return NULL instantly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When loading virSecret configs, the @path variable holds path to
individual config files. In each iteration it is freed explicitly
using VIR_FREE(). Switch it to g_autofree and remove those
explicit calls.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
After previous cleanup the virBuildPathInternal() function is no
longer used. Drop it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Our virBuildPath() constructs a path from given arguments.
Exactly like g_build_filename(), except the latter is more
generic as it uses backslashes on Windows. Therefore, replace the
former with the latter.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In my recent commit of v9.8.0-rc1~7 I've introduced validation
wrt other memory devices. And mostly works, except when doing
memory device update ('virsh update-memory-device') because then
@mem is just parsed <memory/> device XML and thus its pointer is
not in the vm->def->mem, yet. Thus my algorithm which skips over
the same entry fails. Fortunately, we require full device XML on
device update and thus we can use device address and aliases to
detect duplicity.
Fixes: 3fd64fb0e2
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While qemu is still reporting the 'device' field in the tray even the
code was not ready for the possibility of it missing. Fix the condition
for clearing 'devAlias' if qemu doesn't report the 'device' field.
Signed-off-by: Sergey Mironov <mironov@fintech.ru>
Comparison "if (ret == -1)" is always false.
This statement was forgotten during switching to g_new0()
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 0275b06a55 ("util: command: use g_new0")
Signed-off-by: Dmitry Frolov <frolov@swemel.ru>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Now that deleting and reverting external snapshots is implemented we can
report that in capabilities so management applications can use that
information and start using external snapshots.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Original code assumed that the memory state file is only migration
stream but it has additional metadata stored by libvirt. To correctly
load the memory state file we need to reuse code that is used when
restoring domain from saved image.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
We need to skip all disks that have snapshot type other than 'external'.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When used with internal snapshots there is no memory state file so we
have no data to load and decompression is not needed.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When called from snapshot code we will need to pass snapshot object in
order to make internal snapshots work correctly.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When called by snapshot code we will need to use different reason.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The function will no longer be used only when restoring VM as it will
be used when reverting snapshot as well so move it to qemu_process
and rename it accordingly.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
These new helpers separates the code from the logic used to start new
QEMU process with memory state and will make it easier to move
qemuSaveImageStartProcess() into qemu_process.c file.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Part of qemuSaveImageStartVM() function will be used when reverting
external snapshots. To avoid duplicating code and logic extract the
shared bits into separate function.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Currently, nbdkit support will automatically be enabled as long as
the pidfd_open(2) syscall is available. Optionally, libnbd is used
to generate more user-friendly error messages.
In theory this is all good, since use of nbdkit is supposed to be
transparent to the user. In practice, however, there is a problem:
if support for it is enabled at build time and the necessary
runtime components are installed, nbdkit will always be preferred,
with no way for the user to opt out.
This will arguably be fine in the long run, but right now none of
the platforms that we target ships with a SELinux policy that
allows libvirt to launch nbdkit, and the AppArmor policy that we
maintain ourselves hasn't been updated either.
So, in practice, as of today having nbdkit installed on the host
makes network disks completely unusable unless you're willing to
compromise the overall security of the system by disabling
SELinux/AppArmor.
In order to make the transition smoother, provide a convenient
way for users and distro packagers to disable nbdkit support at
compile time until SELinux and AppArmor are ready.
In the process, detection is completely overhauled. libnbd is
made mandatory when nbdkit support is enabled, since availability
across operating systems is comparable and offering users the
option to make error messages worse doesn't make a lot of sense;
we also make sure that an explicit request from the user to
enable/disable nbdkit support is either complied with, or results
in a build failure when that's not possible. Last but not least,
we avoid linking against libnbd when nbdkit support is disabled.
At the RPM level, we disable the feature when building against
anything older than Fedora 40, which still doesn't have the
necessary SELinux bits but will hopefully gain them by the time
it's released. We also allow nbdkit support to be disabled at
build time the same way as other optional features, that is, by
passing "--define '_without_nbdkit 1'" to rpmbuild. Finally, if
nbdkit support has been disabled, installing libvirt will no
longer drag it in as a (weak) dependency.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Wrap the macro body in a new block and move the declaration of 'tmp'
into it, to avoid the need to mix g_autofree with manual freeing.
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
The function is supposed to return the number of items filled into the
array and not zero. Also change the initialization of the "randomness"
to be based on the startCell so that the values are different for each
cell even for separate calls.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
(cherry picked from commit 208569b07b6479e0acd05c5a7e1978b0b641e188)
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
virProcessKillPainfullyDelay() currently almost always returns 1 or -1,
even though the documentation indicates that it should return 0 if the
process was terminated gracefully. But the computation of the return
code is faulty and the only case where it currently returns 0 is when it
is called with the pid of a process that does not exist.
Since no callers ever even distinguish between the 0 and 1 response
codes, simply get rid of the distinction and return 0 for both cases.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Like the Description, these are intended to be displayed to the
user, so it makes sense to have them towards the top of the file
before all the information that systemd will parse to calculate
dependencies.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Hypervisors are referred to by their user-facing name rather
than the name of their libvirt driver, the monolithic daemon is
explicitly referred to as legacy, and a consistent format is
used throughout.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Currently we only set this for the main sockets, which means
that
$ systemctl stop virtqemud.socket
will make the socket disappear from the filesystem while
$ systemctl stop virtqemud-ro.socket
won't. Get rid of this inconsistency.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This results in all sockets for a service being enabled when a
single one of them is.
The -tcp and -tls sockets are intentionally excluded, because
enabling them should require explicit action on the
administrator's part; moreover, disabling them should not result
in the local sockets being disabled too.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
systemd will automatically infer this dependency based on the
socket's Service=foo.service setting.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We have already declared the mirror relationship, so this one
is now redundant.
Moreover, this version was incomplete: it only ever worked for
the monolithic daemon, but the modular daemons for QEMU and Xen
also want the sockets to be active.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Requires/Wants only tells systemd that the corresponding unit
should be started when the current one is, but that could very
well happen in parallel. For virtlogd/virtlockd, we want the
socket to be already active when the hypervisor driver is
started.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Only the main socket is actually necessary for the service to be
usable.
In the past, we've had security issues that could be exploited via
access to the read-only socket, so a security-minded administrator
might consider disabling all optional sockets. This change makes
such a setup possible.
Note that the services will still try to activate all their
sockets on startup, even if they have been disabled. To make sure
that the optional sockets are never started, they will have to be
masked.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This is the strongest relationship that can be declared between
two units, and causes the service to be terminated immediately
if its main socket disappears. This is the behavior we want.
Note that we don't do the same for the read-only/admin sockets,
because those are not as critical for the core functionality of
services as the main socket it.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We want to make sure that, at any given time, we have either the
modular daemons or the monolithic one running, never both. In
order to achieve that, make every single modular unit conflict
with the corresponding libvirtd unit.
We set both Conflicts=libvirtd.unit and After=libvirtd.unit: this
tells systemd that, whenever virtfood.unit and libvirtd.unit are
part of the same transaction, the former should win out.
Thanks to this, if both the modular daemons and the monolithic
one have been enabled because of outdated automation or a simple
mistake of the administrator, the request to start libvirtd at
boot will be ignored and the result will be a regular modular
deployment.
If the request to start libvirtd is made when the modular daemons
are already running, we have no way to prevent systemd from
complying with that request; however, thanks to the way the
conflict relationship has been declared, they will be shut down
cleanly before libvirtd is started. From the user's point of
view, the transition from modular to monolithic will be
completely transparent: it's basically the same scenario as a
regular package upgrade, just with an extra twist.
Note that, while switching from modular to monolithic at runtime
happens automatically, going back requires manual intervention,
i.e. starting all the necessary sockets one by one. That's okay:
the goal here is to prevent misconfiguration and force of habit
to accidentally disrupt a working setup, not to encourage the
scenario. In a correctly configured and managed host, it should
never occur.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Up until now the files have been used as template for most
services, but now that those have been converted to common
templates we can drop parametrization and make it clear that
these files are for libvirtd only.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Now that the migration to common templates has been completed,
we no longer need these.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
All services are still listing their input files explicitly, so
no changes to the output files will occur yet.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We already use templating to generate sockets, which are all
based off libvirtd's. Push the idea further, and extend it to
cover services as well.
This is more challenging, as the various modular daemons each have
their own needs in terms of what system services needs to be
available before they can be started, which other components of
libvirt they depend on, and so on.
In order to make this sort of per-service tweaks possible, we
introduce a Python script that can merge two systemd units
together. The script is aware of the semantics of systemd's unit
definition format, so it can intelligently merge sections
together.
This generic systemd unit merging mechanism will also supersede
the extremely ad-hoc @deps@ variable, which is currently used in
a single scenario.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We're about to change the defaults and start migrating to common
templates: in order to be able to switch units over one at a
time, make the input files that are currently used explicit
rather than implicit.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
These will be useful during the upcoming migration to common
templates for systemd units and will be dropped as soon as all
services have been converted.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
It is currently considered required, but we're soon going to
provide a default that will be suitable for most services.
Since all services currently provide a value explicitly, we
can implement a default without breaking anything.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
They're similar to the existing socket_in/socket_out variables
and will make future changes nicer.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The idea behind these is to prevent running both modular daemons
and monolithic daemon at the same time. We will implement a more
effective solution for that shortly.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
virBitmapFormat returns the string that should be freed.
All strings in three ADD_BITMAP calls in qemuDomainGetGuestVcpusParams
are contained in tmp. So memory leak is possible here without VIR_FREE.
Fixes: 0108deb944
Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Now that providing the value is optional, we can remove almost
all uses.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
For most services, the socket paths can be derived trivially from
the name of the daemon: for virtqemud, for example, they will be
/run/libvirt/virtqemud-sock
/run/libvirt/virtqemud-sock-ro
/run/libvirt/virtqemud-admin-sock
libvirtd and virtproxyd are the exceptions, since their socket
paths will be
/run/libvirt/libvirt-sock
/run/libvirt/libvirt-sock-ro
/run/libvirt/libvirt-admin-sock
So we still need to be able to provide a custom @sockprefix@ in
those cases, but in the most common scenario we can do away with
the requirement by introducing a sensible default.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
For most services, the value provided explicitly matches the
documented default.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The decision is based only on whether Polkit support is enabled,
so there's no need to go through it again for every single
service.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The meaning of the _def suffix might not be immediately obvious,
especially since it's also used to refer to the output of the
meson-gen-def.py script elsewhere in the same file.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The information is not used anywhere right now, but the
documentation for virt_daemon_units claims it's mandatory.
We also intend to actually start using it later on.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This tells systemd that the services in question support the
native socket activation protocol.
virtlogd and virtlockd, just like all the other daemons, implement
the necessary handshake.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
While systemd will automatically match foo.socket with foo.service
based on their names, it's nicer to connect the two explicitly.
This is what we do for all services, with virtlogd and virtlockd
being the only exceptions.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This annotation being missing resulted in virtlogd and virtlockd
being marked as "indirect" services, i.e. services that cannot
be started directly but have to be socket activated instead.
While this is our preferred configuration, we shouldn't prevent
the admin to start them at boot if they want to.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
When libvirtd, virtlog and virtlockd are enabled, we want their
admin sockets to be enabled for socket activation as well.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Commit 93af79fb removed a cleanup label in favor of returning error
values directly in certain cases. But the final return value was changed
from -1 to 0. If we get to the end of the function, that means that
we've waited for the process to exit but it still exists. So we should
return -1. The error message was still being set correctly, but we were
returning a success status (0).
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In a modular daemon configuration, virtxend does not support the
virNetwork* APIs. It should open a connection to virtnetworkd when
using those APIs, but currently always opens a connection to
"xen:///system". Switch to using virGetConnectNetwork to obtain a
valid connection instead of using the hardcoded URI.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewing the code I found that return value of function
udev_device_get_sysattr_value() is dereferenced without a check.
udev_device_get_sysattr_value() may return NULL by number of reasons.
v2: VIR_DEBUG added, replaced STREQ(NULLSTR()) with STREQ_NULLABLE()
v3: More checks added, to skip earlier. More verbose VIR_DEBUG.
Signed-off-by: Dmitry Frolov <frolov@swemel.ru>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
As of v9.4.0-rc2~5 it is possible to specify guest address where
a virtio-mem/virtio-pmem memory device is mapped to. What that
commit forgot to introduce was a check for overlaps.
And yes, this is technically an O(n^2) algorithm, as
virDomainMemoryDefValidate() is called over each memory device
and after this, virDomainMemoryDefValidate() also iterates over
each memory device. But given there's usually only a handful of
such devices, and this runs only when parsing domain XML I guess
code readability wins over some less obvious solution.
Resolves: https://issues.redhat.com/browse/RHEL-4452
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Inside of virDomainMemoryDefValidate() there's a check that
address where a virtio-mem memory device is mapped to is a
multiple of its block size. But this check is off by a couple of
bits, because the memory address is in bytes while the block size
is in kibibytes. Therefore, when checking whether address is a
multiple of the block size, the latter has to be multiplied by a
factor of 1024.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
QEMU mandates the VIRTIO_PMEM address is aligned to a pagesize.
This is a very reasonable requirement. So much so, that it
deserves to be in hypervisor agnostic validation code
(virDomainMemoryDefValidate()). Not that any other hypervisor
would support VIRTIO_PMEM yet. But even if they did, this would
surely be still valid.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The current message can be misleading, because it seems to suggest
that no firmware of the requested type is available on the system.
What actually happens most of the time, however, is that despite
having multiple firmwares of the right type to choose from, none
of them is suitable because of lacking some specific feature or
being incompatible with some setting that the user has explicitly
enabled.
Providing an error message that describes exactly the problem is
not feasible, since we would have to list each candidate along
with the reason why we rejected it, which would get out of hand
quickly.
As a small but hopefully helpful improvement over the current
situation, reword the error message to make it clearer that the
culprit is not necessarily the firmware type, but rather the
overall domain configuration.
Suggested-by: Michael Kjörling <7d1340278307@ewoof.net>
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Our test driver lacks implementation for
virConnectGetDomainCapabilities(). Provide one, though a trivial
one. Mostly so that something else than VIR_ERR_NO_SUPPORT error
is returned.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
As can be seen from previous commits, it's fairly easy to pass a
different type to virReportEnumRangeError() than the actual
variable is of. So far, we have a sizeof() hack to check if some
nonsensical types are not passed, e.g. it catches cases where a
function name is passed instead of an enum. Extend the hack to
check whether proper enum was passed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The @backend member of _virDomainVideoDef struct is of type
virDomainVideoBackendType. Pass the proper type to
virReportEnumRangeError().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The @virtPortOp variable inside of virNetDevVPortProfileOp8021Qbh
is of type virNetDevVPortProfileLinkOp. Pass the proper type to
virReportEnumRangeError().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This allows us to declare variables without using 'enum
virNetDev....' and will become more useful in the near future
(when virReportEnumRangeError() is fixed).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Function virGetConnectSecret() can return NULL so we need to check it
since in virSecretGetSecretString() it gets dereferenced.
Reported-by: coverity
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
It's not possible to use password-protected ssh keys directly with
libvirt because libvirt doesn't have any way to prompt a user for the
password. To accomodate password-protected key files, an administrator
can add these keys to an ssh agent and then configure the domain with
the path to the ssh-agent socket.
Note that this requires an administrator or management app to
configure the ssh-agent with an appropriate socket path and add the
necessary keys to it. In addition, it does not currently work with
selinux enabled. The ssh-agent socket would need a label that libvirt
would be allowed to access rather than unconfined_t.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Add the ability to specify a path to a ssh-agent socket in order to use
the ssh-agent to authenticate to remote ssh disks. Example
configuration:
<disk type='network'>
</source protocol='ssh' ...>
<identity username='myusername' agentsock='/path/to/socket'/>
...
</source>
...
</disk>
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
For ssh disks that are served by nbdkit, we can support logging in with
an ssh key file. Pass the path to the configured key file and the
username to the nbdkit process.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Authenticating via key file to an ssh server is often preferable to
logging in via password. In order to support this functionality add a
new <identity> xml element for ssh disks that allows the user to specify
a keyfile and username. Example configuration:
<disk type='network'>
<source protocol='ssh' ...>
<identity keyfile='/path/to/id_rsa' username='myusername'/>
...
</source>
...
</disk>
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
For ssh disks that are served by nbdkit, use the configured value for
knownHosts and pass it to the nbdkit process.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In order to make ssh disks usable, we need to be able to validate a
remote host. To do this, add a <knownHosts> xml element for ssh disks to
allow the user to specify a location for a file that contains known host
keys. Implementation to follow.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
For ssh disks that are served by nbdkit, lookup the password from the
configured secret and securely pass it to the nbdkit process using fd
passing.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Right now, ssh network disks are not usable. There is some basic support
in libvirt that is meant to support disk chains that have backing disks
located at ssh urls, but there is no real way for a user to configure a
ssh-based disk. This commit allows users to configure an ssh disk with
password authentication. Implementation will follow.
<disk type='network'>
<source protocol='ssh' ...>
<auth username='myusername'>
<secret type='iscsi' usage='secretname'/>
</auth>
</disk>
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When using nbdkit to serve a network disk source, the nbdkit process
will start and wait for an nbd connection before actually attempting to
connect to the (remote) disk location. Because of this, nbdkit will not
report an error until after qemu is launched and tries to read from the
disk. This results in a fairly user-unfriendly error saying that qemu
was unable to start because "Requested export not available".
Ideally we'd like to be able to tell the user *why* the export is not
available, but this sort of information is only available to nbdkit, not
qemu. It could be because the url was incorrect, or because of an
authentication failure, or one of many other possibilities.
To make this friendlier for users and easier to detect
misconfigurations, try to connect to nbdkit immediately after starting
nbdkit and before we try to start qemu. This requires adding a
dependency on libnbd. If an error occurs when connecting to nbdkit, read
back from the nbdkit error log and provide that information in the error
report from qemuNbdkitProcessStart().
User-visible change demonstrated below:
Previous error:
$ virsh start nbdkit-test
2023-01-18 19:47:45.778+0000: 30895: error : virNetClientProgramDispatchError:172 : internal
error: process exited while connecting to monitor: 2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}: Requested export not
available
error: Failed to start domain 'nbdkit-test'
error: internal error: process exited while connecting to monitor: 2023-01-18T19:47:45.704658Z
qemu-system-x86_64: -blockdev {"driver":"nbd","server":{"type":"unix",
"path":"/var/lib/libvirt/qemu/domain-1-nbdkit-test/nbdkit-libvirt-1-storage.socket"},
"node-name":"libvirt-1-storage","auto-read-only":true,"discard":"unmap"}: Requested export not
available
After this change:
$ virsh start nbdkit-test
2023-01-18 19:44:36.242+0000: 30895: error : virNetClientProgramDispatchError:172 : internal
error: Failed to connect to nbdkit for 'http://localhost:8888/nonexistent.iso': nbdkit: curl[1]:
error: problem doing HEAD request to fetch size of URL [http://localhost:8888/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
error: Failed to start domain 'nbdkit-test'
error: internal error: Failed to connect to nbdkit for 'http://localhost:8888/nonexistent.iso]:
error: problem doing HEAD request to fetch size of URL [http://localhost:8888/nonexistent.iso]:
HTTP response code said error: The requested URL returned error: 404
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Adds the ability to monitor the nbdkit process so that we can take
action in case the child exits unexpectedly.
When the nbdkit process exits, we pause the vm, restart nbdkit, and then
resume the vm. This allows the vm to continue working in the event of a
nbdkit failure.
Eventually we may want to generalize this functionality since we may
need something similar for e.g. qemu-storage-daemon, etc.
The process is monitored with the pidfd_open() syscall if it exists
(since linux 5.3). Otherwise it resorts to checking whether the process
is alive once a second. The one-second time period was chosen somewhat
arbitrarily.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Since the restart handler will trigger at an arbitrary time (when the
nbdkit process crashes, for instance), it's difficult to provide
feedback to the user if the restart is unsuccessful. Rather than just
relying on a warning in the log, taint the domain so that there will be
a slightly more user-visible notification.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
We were testing the arguments that were being passed to qemu when a disk
was being served by nbdkit, but the arguments used to start nbdkit
itself were not testable. This adds a test to ensure that we're invoking
nbdkit correctly for various disk source definitions.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Add a private function to peek at the list of send buffers in virCommand
so that it is testable
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
For virStorageSource objects that contain an nbdkitProcess, start that
nbdkit process to serve that network drive and then pass the nbdkit
socket to qemu rather than sending the network url to qemu directly.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Rather than passing passwords and cookies (which could contain
passwords) to nbdkit via commandline arguments, use the alternate format
that nbdkit supports where we can specify a file descriptor which nbdkit
will read to get the password or cookies.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
All users of virCommandSetSendBuffer() are using it to send sensitive
data to a child process. So, since these buffers contain sensitive
information, clear it with virSecureErase().
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Add xml to the private data for a disk source to represent the nbdkit
process so that the state can be re-created if the libvirt daemon is
restarted. Format:
<nbdkit>
<pidfile>/path/to/nbdkit.pid</pidfile>
<socketfile>/path/to/nbdkit.socket</socketfile>
</nbdkit>
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
This prepares encryption secrets and authentication secrets. When we add
nbdkit-backed network storage sources, we will not need to send
authentication secrets to qemu, since they will be sent to nbdkit
instead. So split this into two different functions.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Add new DO_TEST_CAPS_LATEST_NBDKIT macro to test xml2argv for various
nbdkit capability scenarios.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
