Michal Privoznik [Tue, 25 Jan 2022 10:05:21 +0000 (11:05 +0100)]
qemuDomainDetachDeviceLive: Handle hostevs with unassigned type of address
A <hostdev/> can have <address type='unassigned'/> which means
libvirt manages the device detach from/reattach to the host but
the device is never exposed to the guest. This means that we have
to take a shortcut during hotunplug (e.g. never ask QEMU on the
monitor to detach the device, or never wait for DEVICE_DELETED
event).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 25 Jan 2022 07:51:26 +0000 (08:51 +0100)]
qemuDomainAttachHostPCIDevice: Handle hostevs with unassigned type of address
A <hostdev/> can have <address type='unassigned'/> which means
libvirt manages the device detach from/reattach to the host but
the device is never exposed to the guest. This means that we have
to take a shortcut during hotplug, similar to the one we are
taking when constructing the command line (see
qemuBuildHostdevCommandLine()).
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2040548 Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
We document that <address type='unassigned'/> can be used only
for <hostdev/>-s. However, corresponding validation rule is
missing. Let's put the rule into hypervisor agnostic part of
validation process so that all drivers can benefit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Andrea Bolognani [Wed, 26 Jan 2022 14:14:13 +0000 (15:14 +0100)]
meson: Require that both qemu_user and qemu_group are provided
Or that neither is. The current implementation, where if only
one of the two is provided the other one will be based on
OS-specific defaults is more likely to cause confusion than it
is to be helpful.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Wed, 26 Jan 2022 14:11:17 +0000 (15:11 +0100)]
meson: Don't check whether qemu_user and qemu_group exist
QEMU might not be installed on the build system, in which case
the user and group will not be present. We should avoid falling
back to root:root in that case, and assume the user and group
are going to be present in the target system instead.
Suggested-by: Olaf Hering <olaf@aepfle.de> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Wed, 26 Jan 2022 14:01:40 +0000 (15:01 +0100)]
meson: Don't require /etc/os-release to be present
It might be part of some non-mandatory package on certain
distros, and our logic deals just fine with its contents not
being available.
Fixes: 4c69d64efa3731d074d198f871fd42e74c4a39f6 Reported-by: Olaf Hering <olaf@aepfle.de> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Michal Privoznik [Wed, 26 Jan 2022 12:47:33 +0000 (13:47 +0100)]
node_device: Treat NVMe disks as regular disks
Unfortunately, udev doesn't set ID_TYPE attribute for NVMe disks,
therefore we have to add another case into udevKludgeStorageType()
to treat /dev/nvme* devlinks as any other disk.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2045953 Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Michal Privoznik [Wed, 26 Jan 2022 13:00:13 +0000 (14:00 +0100)]
node_device: Rework udevKludgeStorageType()
The udevKludgeStorageType() function looks at devlink name
(/dev/XXX) and guesses the type of the (storage) device using a
series of STRPREFIX() calls. Well those can be turn into an array
and a for() loop, especially if we are about to add a new case
(in the next commit).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Michal Privoznik [Tue, 30 Nov 2021 17:16:00 +0000 (18:16 +0100)]
qemu: Allow prealloc for virtio-mem-pci
There are a some scenarios in which we want to prealloc guest
memory (e.g. when requested in domain XML, when using hugepages,
etc.). With 'regular' <memory/> models (like 'dimm', 'nvdimm' or
'virtio-pmem') or regular guest memory it is corresponding
memory-backend-* object that ends up with .prealloc attribute
set. And that's desired because neither of those devices can
change its size on the fly. However, with virtio-mem model things
are a bit different. While one can set .prealloc attribute on
corresponding memory-backend-* object it doesn't make much sense,
because virtio-mem can inflate/deflate on the fly, i.e. change
how big of a portion of the memory-backend-* object is exposed to
the guest. For instance, from a say 4GiB module only a half can
be exposed to the guest. Therefore, it doesn't make much sense to
preallocate whole 4GiB and keep them allocated. But we still want
the part exposed to the guest preallocated (when conditions
described at the beginning are met).
Having said that, with new enough QEMU the virtio-mem-pci device
gained new attribute ".prealloc" which instructs the device to
talk to the memory backend object and allocate only the requested
portion of memory.
Now, that our algorithm for setting .prealloc was isolated in a
single function, the function can be called when constructing cmd
line for virtio-mem-pci device.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
qemu: Move @prealloc into qemuBuildMemoryGetPagesize()
The qemuBuildMemoryGetPagesize() function has everything is needs
to decide whether preallocation is needed or not. Move the logic
from qemuBuildMemoryBackendProps() into
qemuBuildMemoryGetPagesize().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Mon, 24 Jan 2022 06:15:04 +0000 (07:15 +0100)]
conf: Separate out virDomainClockDef formatting
Currently, virDomainClockDef is formatted inside
virDomainDefFormatInternalSetRootName() which is already long
enough. Move the code into a new function
(virDomainClockDefFormat()) and make the code use
virXMLFormatElement() while at it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 21:08:03 +0000 (22:08 +0100)]
conf: Fix @mode member of _virDomainTimerDef struct
The @mode member of the _virDomainTimerDef struct stores
values of the virDomainTimerModeType enum, or -1 for the
default value (when user provided no value in XML).
This is needlessly complicated. Introduce new value to the enum
which reflects the default state.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 20:56:54 +0000 (21:56 +0100)]
conf: Fix @track member of _virDomainTimerDef struct
The @track member of the _virDomainTimerDef struct stores
values of the virDomainTimerTrackType enum, or -1 for the
default value (when user provided no value in XML).
This is needlessly complicated. Introduce new value to the enum
which reflects the default state.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 20:34:29 +0000 (21:34 +0100)]
conf: Fix @tickpolicy member of _virDomainTimerDef struct
The @tickpolicy member of the _virDomainTimerDef struct stores
values of the virDomainTimerTickpolicyType enum, or -1 for the
default value (when user provided no value in XML).
This is needlessly complicated. Introduce new value to the enum
which reflects the default state.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 20:15:10 +0000 (21:15 +0100)]
conf: Fix type of @present in _virDomainTimerDef struct
In the _virDomainTimerDef structure we have @present member which
is like virTristateBool, except it's an integer and has values
shifted by one. This is harder to read. Retype the member to
virTristateBool which we are familiar with.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Mon, 24 Jan 2022 07:58:00 +0000 (08:58 +0100)]
virDomainChrSourceDefCopy: return void
This function never returns an error, make it void then. And
while at it, make the @src argument const to make it obvious it's
never changed inside the function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Mon, 24 Jan 2022 07:55:52 +0000 (08:55 +0100)]
virDomainChrSourceDefCopy: Don't check arguments against NULL
The only caller of this function
(qemuProcessFindCharDevicePTYsMonitor()) doesn't pass NULL.
Remove corresponding check from virDomainChrSourceDefCopy().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 13:58:36 +0000 (14:58 +0100)]
virDomainChrSourceDefCopy: Copy more struct members
The aim of virDomainChrSourceDefCopy() is to make a deep copy of
given virDomainChrSourceDef. However, some types were not copied
at all (VIR_DOMAIN_CHR_TYPE_SPICEVMC and
VIR_DOMAIN_CHR_TYPE_SPICEPORT) and some members weren't copied
either (@logfile, @logappend).
After this, there are still some members that are not copied
(seclabels and private data), but the sole caller
qemuProcessFindCharDevicePTYsMonitor() doesn't seem to care.
Therefore, just document this behavior so that future user is
aware.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Sun, 23 Jan 2022 11:54:27 +0000 (12:54 +0100)]
virDomainInputDefParseXML: Move validation into validator
There is some code that validates whether parsed @bus <input/>
makes sense (e.g. some hypervisors have their own type of bus).
But this code should not live in the parser, but validator
rather. That way, we can also validate that the value we compute
(if user didn't provide any) is valid.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Tue, 25 Jan 2022 16:49:00 +0000 (17:49 +0100)]
qemuDomainSetupDisk: Initialize 'targetPaths'
Compiler isn't able to see that 'virDevMapperGetTargets' in cases e.g.
when the devmapper isn't available may not initialize the value in the
pointer passed as the second argument.
The usage 'qemuDomainSetupDisk' lead to an accidental infinite loop as
previous calls apparently doctored the stack to a point where
'g_slist_concat' would end up in an infinite loop trying to find the end
of the list.
Fixes: 6c49c2ee9fcb88de02cdc333f666a8e95d60a3b0 Closes: https://gitlab.com/libvirt/libvirt/-/issues/268 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Roman Bolshakov [Mon, 19 Nov 2018 11:13:53 +0000 (14:13 +0300)]
docs: Note hvf support for domain elements
Many domain elements have "QEMU and KVM only" or "QEMU/KVM since x.y.z"
remarks. Most of the elements work for HVF domain, so it makes sense to
add respective notices for HVF domain.
All the elements have been manually tested.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The new DO_TEST_MACOS() macro makes it possible to create test
cases that verify the behavior of libvirt on a macOS machine
with HVF support available.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This exposes a couple of macOS-specific variants of existing
APIs, which can be used when implementing test programs and
result in HVF support being advertised.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This new enumeration provides a way to specify the host OS
that a specific test case expects. The default is Linux, which
has been the implicit host OS until now; when Linux is selected
as the host OS, KVM support is advertised in capabilies data
exposed to test cases.
This commit doesn't result in any functional change, and simply
sets the stage for introducing macOS host OS support later.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Roman Bolshakov [Fri, 19 Oct 2018 12:07:53 +0000 (15:07 +0300)]
qemu: Query hvf capability on macOS
There's no QMP command for querying if hvf is supported, therefore we
use sysctl interface that tells if Hypervisor.framework works/available
on the host.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Roman Bolshakov [Thu, 18 Oct 2018 22:14:01 +0000 (01:14 +0300)]
conf: Add hvf domain type
QEMU supports Hypervisor.framework since 2.12 as hvf accel.
Hypervisor.framework provides a lightweight interface to run a virtual
cpu on macOS without the need to install third-party kernel
extensions (KEXTs).
It's supported since macOS 10.10 on machines with Intel VT-x feature
set that includes Extended Page Tables (EPT) and Unrestricted Mode.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Roman Bolshakov [Wed, 21 Nov 2018 18:45:02 +0000 (21:45 +0300)]
qemu: Add KVM CPUs into cache only if KVM is present
virQEMUCapsFormatCache/virQEMUCapsLoadCache adds/reads KVM CPUs to/from
capabilities cache regardless of QEMU_CAPS_KVM. That can cause undesired
side-effects when KVM CPUs are present in the cache on a platform that
doesn't support it, e.g. macOS or Linux without KVM support.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We already know it's not going to be available on other
platforms.
Suggested-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Andrea Bolognani <abologna@redhat.com> Tested-by: Brad Laue <brad@brad-x.com> Tested-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Ani Sinha <ani@anisinha.ca> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Jim Fehlig [Mon, 24 Jan 2022 19:03:20 +0000 (12:03 -0700)]
libxl: Add lock process indicator to saved VM state
Commit fa58f571ee added a lock processes indicator to the
libxlDomainObjPrivate struct to note that a lock process was
successfully started for the VM. However, the commit neglected to
add the indicator to the VM's saved state file. As a result, the
indicator is lost on libvirtd restart, along with the knowledge of
whether a lock process was started for the VM.
This change adds support for the indicator in the domainObjPrivate
data parse and format callbacks, ensuring its value survives libvirtd
restarts.
Signed-off-by: Jim Fehlig <jfehlig@suse.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
conf: network: Allow hostnames to start with a number
RFC952 mandated that hostnames would start with an alpha character.
This requirement was later relaxed by RFC1123 which allowed hostnames to
start with a number as well.
Michal Privoznik [Tue, 25 Jan 2022 09:48:36 +0000 (10:48 +0100)]
qemuDomainAttachHostPCIDevice: Fix coding style
Our coding style requires that a body of an if() longer than two
lines is wrapped in a curly braces. There's one offender in
qemuDomainAttachHostPCIDevice(). Fortunately, there was no
functional problem because one of the lines is a comment.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
shenjiatong [Sat, 22 Jan 2022 04:10:35 +0000 (12:10 +0800)]
qemu: Remove scsi parameter for vhost-user-blk for qemu < 4.2
When trying to attach vhost-user-blk device to virtual machine using
qemu < 4.2 libvirt would mistakenly add a scsi=off parameter, which is
not supported by qemu.
Fixes: https://gitlab.com/libvirt/libvirt/-/issues/265 Signed-off-by: shenjiatong <yshxxsjt715@gmail.com> Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Peter Krempa [Thu, 13 Jan 2022 11:47:17 +0000 (12:47 +0100)]
qemuxml2argvtest: disk-vhostuser: Add invocation for qemu-4.2
With qemu versions prior to qemu-5.0 we'll format 'scsi=off' for
virtio-blk disks, but also for vhost-user-blk. This is a bug as it's not
supported.
Add a test case to show that wrong configuration is generated by adding
running 'disk-vhostuser' test case on capabilities from qemu-4.2.
For this to be possible it's required to enable shared memory via NUMA
configuration as old QEMU's don't allow configuration of the default
memory backend. This is achieved by adding a copy of the
'disk-vhostuser' XML with NUMA enabled.
An update to meson 0.61.1 meant that it started showing warnings due to the fact
that the default for run_command's 'check' parameter is going to change. It
unveiled the fact that we were even missing that parameter in some calls where
we expected different outcome. To make sure the behaviour does not change
specify the parameter explicitly. In places where we check for the return code
the parameter should be 'false' so that meson does not fail. In all other cases
the parameter should be set to 'true' to make sure possible failure also stops
meson.
The warning in meson was added in https://github.com/mesonbuild/meson/pull/9304
Signed-off-by: Martin Kletzander <mkletzan@redhat.com> Reviewed-by: Tim Wiederhake <twiederh@redhat.com>
The lookups in esx_vi work a bit differently that we are used to. The filters
(travelsalSpec and selectSet) choose how to look up the objects, but given a
root object the lookup lists all the objects of a requested type inside it as
well as the root object itself. We then go through the results and find the one
which has the same name as was requested. However in a case with nested folders
of a same name this could break when the first returned object in the list is
the parent folder as we'd select it only based on the name. To avoid this also
add a check that the candidate we are trying to pick is not exactly the same
object (reference) as the root object.
Michal Privoznik [Thu, 20 Jan 2022 12:58:59 +0000 (13:58 +0100)]
conf: Convert virDomainHostdevDefParseXMLSubsys() to virXMLProp*()
After previous cleanups, the virDomainHostdevDefParseXMLSubsys()
function uses a mixture of virXMLProp*() and the old
virXMLPropString() + virXXXTypeFromString() patterns. Rework it
so that virXMLProp*() is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Fri, 21 Jan 2022 09:19:34 +0000 (10:19 +0100)]
conf: Convert virNetworkPortDefParseXML() to virXMLProp*()
After previous cleanups, the virNetworkPortDefParseXML() function
uses a mixture of virXMLProp*() and the old virXMLPropString() +
virXXXTypeFromString() patterns. Rework it so that virXMLProp*()
is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 12:15:03 +0000 (13:15 +0100)]
conf: Convert virDomainNetDefParseXML() to virXMLProp*()
After previous cleanups, the virDomainNetDefParseXML() function
uses a mixture of virXMLProp*() and the old virXMLPropString() +
virXXXTypeFromString() patterns. Rework it so that virXMLProp*()
is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 12:14:53 +0000 (13:14 +0100)]
conf: Convert virDomainFSDefParseXML() to virXMLProp*()
After previous cleanups, the virDomainFSDefParseXML() function
uses a mixture of virXMLProp*() and the old virXMLPropString() +
virXXXTypeFromString() patterns. Rework it so that virXMLProp*()
is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 20:57:49 +0000 (21:57 +0100)]
conf: Convert virDomainDefParseBootXML() to virXMLProp*()
After previous cleanups, the virDomainDefParseBootXML() function
uses a mixture of virXMLProp*() and the old virXMLPropString() +
virXXXTypeFromString() patterns. Rework it so that virXMLProp*()
is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 20:57:35 +0000 (21:57 +0100)]
conf: Convert virCPUDefParseXML() to virXMLProp*()
After previous cleanups, the virCPUDefParseXML() function uses a
mixture of virXMLProp*() and the old virXMLPropString() +
virXXXTypeFromString() patterns. Rework it so that virXMLProp*()
is used.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 12:58:49 +0000 (13:58 +0100)]
lib: Almost eliminate use of virTristateBoolTypeFromString()
There are couple of places where virTristateBoolTypeFromString()
is called. Well, the same result can be achieved by
virXMLPropTristateBool() and on fewer lines.
Note there are couple of places left untouched because those
don't care about error reporting and thus are shorter they way
they are now.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 10:24:55 +0000 (11:24 +0100)]
lib: Eliminate use of virTristateSwitchTypeFromString()
There are couple of places (all of them in XML parsing) where
virTristateSwitchTypeFromString() is called. Well, the same
result can be achieved by virXMLPropTristateSwitch() and on fewer
lines.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Thu, 20 Jan 2022 13:14:31 +0000 (14:14 +0100)]
qemu: Use virTristateBool instead of virTristateSwitch in a few places
Both @accel2d and @accel3d are parsed as virTristateBool, but in
a few places (qemuDeviceVideoGetModel() and
qemuValidateDomainDeviceDefVideo()) they are compared to
virTristateSwitch enum either directly or via a variable of that
type. Clear this confusion by using the correct enum.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Fri, 21 Jan 2022 11:22:07 +0000 (12:22 +0100)]
qemuValidateDomainDeviceDefFS: Use correct enum for fs->multidevs comparison
During validation of a virDomainFSDef QEMU capabilities are check
for multidevs support if the FS definition has it enabled.
However, the fs->multidevs is really type of virDomainFSMultidevs
but is compared against virDomainFSModel enum. Fortunately, both
values are the same so no user visible harm done here.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 20 Jan 2022 15:20:22 +0000 (16:20 +0100)]
man: virt-admin: Mention monolithic daemon URIs
Hint users that they can use 'virt-admin' also for the new monolithic
daemons.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2038045 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com>
Peter Krempa [Mon, 17 Jan 2022 15:19:59 +0000 (16:19 +0100)]
kbase: debuglogs: Fix for modular daemons
Modular daemons are now the default in many new installations, thus we
need to include steps how to determine that modular daemons are used and
modular-daemon specific locations for the config files and admin URIs.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com>
Peter Krempa [Mon, 17 Jan 2022 09:24:14 +0000 (10:24 +0100)]
docs: daemons: Add section on figuring out whether modular or monolithic daemon is in use
Since we are at a transition period where some users may be running
monolithic libvirtd and others already the modular topology we need a
section that allows users to figure out which is in use.
This will be particularly important in the document about enabling
logging, as the active log file depends on which daemon is in use.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com>
Peter Krempa [Thu, 20 Jan 2022 17:17:47 +0000 (18:17 +0100)]
man: virt-admin: Remove misleading paragraph about unix socket permissions
The socket permissions are controlled by 'unix_sock_admin_perms', but
regardless the code requires that 'geteuid() != clientuid' to allow
clients thus it doesn't make sense to make users aware of it.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Turns out, the commit did more harm than good. It changed
semantics on some public APIs. For instance, while
qemuDomainGetInfo() previously did not returned an error it does
now. While the calls to virProcessGetStatInfo() is guarded with
virDomainObjIsActive() it doesn't necessarily mean that QEMU's
PID is still alive. QEMU might be gone but we just haven't
realized it (e.g. because the eof handler thread is waiting for a
job).
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2041610 Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Ján Tomko [Thu, 20 Jan 2022 13:53:33 +0000 (14:53 +0100)]
qemu: fix inactive snapshot revert
The commit splitting out the qemuSnapshotRevertInactive function
dropped the 'defined = true' line by accident and instead
returned -1, leaving the user with a cryptic error:
error: An error occurred, but the cause is unknown
Peter Krempa [Thu, 20 Jan 2022 10:46:51 +0000 (11:46 +0100)]
storage: Implement 'checkPool' method for 'disk' type pools
If 'checkPool' is not implemented, the pool will be made inactive when
restarting libvirtd and subsequently re-loading the state from the pool
state XML.
Base the 'checkPool' implementation on logic similar to 'startPool'.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1910856 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 20 Jan 2022 10:43:22 +0000 (11:43 +0100)]
virStoragePoolObjSourceFindDuplicateCb: Fix handling of VIR_STORAGE_POOL_ISCSI_DIRECT
The direct SCSI pool doesn't expose the volumes in the host attempting
to match it using 'virStoragePoolObjSourceMatchTypeDEVICE' which in turn
uses 'virStoragePoolSourceFindDuplicateDevices' doesn't make sense.
Remove it from the source matching completely as we can open multiple
connections to the target.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 20 Jan 2022 10:36:29 +0000 (11:36 +0100)]
storageDriverAutostartCallback: Refactor control flow
Use early returns to decrease the indentation level and make it more
obvious that the 'cleanup' path is a noop in those cases.
'virStoragePoolObjSetStarting' was called only when the code wanted to
start the pool, so if that was skipped, cleanup is noop as it's
conditional on the return value of 'virStoragePoolObjIsStarting'.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 19 Jan 2022 08:49:31 +0000 (09:49 +0100)]
docs: man: virsh: Document more carefully that 'guestinfo' can return nothing
When invoking 'virsh guestinfo $VM' without explicitly specifying a
group of information to return, virsh always reports success even when
the guest agent doesn't report any information in the current state.
This is desired in situations when you are okay with stats being missing
and avoids spurious errors being reported.
Clarify that this is really desired in the man page.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2041665 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Peter Krempa [Mon, 17 Jan 2022 11:15:17 +0000 (12:15 +0100)]
kbase: debuglogs: Section off the settings for legacy libvirt
Move out the settings required to pre libvirt-4.4.0 deployments into a
separate section so that the main point is not cluttered by now mostly
irrelevant settings.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com>
Peter Krempa [Wed, 19 Jan 2022 12:41:28 +0000 (13:41 +0100)]
spec: Run pre/post-install stuff on 'daemon-driver-storage-core'
'daemon-driver-storage' is just a meta-package, the actual daemon is in
the 'daemon-driver-storage-core' package, so without installing the meta
package the storage daemon is not being enabled.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2025644 Fixes: 50eae3f8859, b8ef6255444 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Peter Krempa [Tue, 18 Jan 2022 14:00:21 +0000 (15:00 +0100)]
remote: systemd: Remove unix sockets from filesystem when disabling a '.socket' unit
The existence of the unix socket path is used by the remote driver to
determine whether modular daemons are in use, so if the socket file
stays behind and the user decided to switch from modular to monolithic
daemon which was socket activated, the remote driver will insist on
picking '/var/run/libvirt/virtqemud-sock', even when it's no longer in
use:
# systemctl start libvirtd.service
# virsh list
Id Name State
--------------------
# systemctl stop libvirtd.service
Warning: Stopping libvirtd.service, but it can still be activated by:
libvirtd.socket
libvirtd-ro.socket
libvirtd-admin.socket
# systemctl start virtqemud.socket
# virsh list
Id Name State
--------------------
# systemctl stop virtqemud.socket
# systemctl start libvirtd.service
# virsh list
error: failed to connect to the hypervisor
error: Failed to connect socket to '/var/run/libvirt/virtqemud-sock': Connection refused
# virsh -c 'qemu:///system?socket=/var/run/libvirt/libvirt-sock' list
Id Name State
--------------------
Fix this by instructing systemd to delete the socket file when
deactivating the unit file for the socket.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ani Sinha <ani@anisinha.ca>
Peter Krempa [Tue, 18 Jan 2022 15:59:35 +0000 (16:59 +0100)]
qemuValidateDomainDeviceDefVideo: Reject non-default video head count for devices not supporting it
Only QXL and virtio-vga actually propagate the 'heads' attribute as
'max_outputs' to the commandline of qemu. Reject the setting when
non-default value is used for any other video type.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2036300 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
projects / libvirt.git / log