<interface> devices (virDomainNetDef) are a bit different from other
types of devices in that their actual type may come from a network (in
the form of a port connection), and that doesn't happen until the
domain is started. This means that any validation of an <interface> at
parse time needs to be a bit liberal in what it accepts - when
type='network', you could think that something is/isn't allowed, but
once the domain is started and a port is created by the configured
network, the opposite might be true.
To solve this problem hypervisor drivers need to do an extra
validation step when the domain is being started. I recently (commit
3cff23f7, libvirt 5.7.0) added a function to peform such validation
for all interfaces to the QEMU driver -
qemuDomainValidateActualNetDef() - but while that function is a good
single point to call for the multiple places that need to "start" an
interface (domain startup, device hotplug, device update), it can't be
called by the other hypervisor drivers, since 1) it's in the QEMU
driver, and 2) it contains some checks specific to QEMU. For
validation that applies to network devices on *all* hypervisors, we
need yet another interface validation function that can be called by
any hypervisor driver (not just QEMU) right after its network port has
been created during domain startup or hotplug. This patch adds that
function - virDomainActualNetDefValidate(), in the conf directory,
and calls it in appropriate places in the QEMU, lxc, and libxl
drivers.
This new function is the place to put all network device validation
that 1) is hypervisor agnostic, and 2) can't be done until we know the
"actual type" of an interface.
There is no framework for validation at domain startup as there is for
post-parse validation, but I don't want to create a whole elaborate
system that will only be used by one type of device. For that reason,
I just made a single function that should be called directly from the
hypervisors, when they are initializing interfaces to start a domain,
right after conditionally allocating the network port (and regardless
of whether or not that was actually needed). In the case of the QEMU
driver, qemuDomainValidateActualNetDef() is already called in all the
appropriate places, so we can just call the new function from
there. In the case of the other hypervisors, we search for
virDomainNetAllocateActualDevice() (which is the hypervisor-agnostic
function that calls virNetworkPortCreateXML()), and add the call to our
new function right after that.
The new function itself could be plunked down into many places in the
code, but we already have 3 validation functions for network devices
in 2 different places (not counting any basic validation done in
virDomainNetDefParseXML() itself):
1) post-parse hypervisor-agnostic
(virDomainNetDefValidate() - domain_conf.c:6145)
2) post-parse hypervisor-specific
(qemuDomainDeviceDefValidateNetwork() - qemu_domain.c:5498)
3) domain-start hypervisor-specific
(qemuDomainValidateActualNetDef() - qemu_domain.c:5390)
I placed (3) right next to (2) when I added it, specifically to avoid
spreading validation all over the code. For the same reason, I decided
to put this new function right next to (1) - this way if someone needs
to add validation specific to qemu, they go to one location, and if
they need to add validation applying to everyone, they go to the
other. It looks a bit strange to have a public function in between a
bunch of statics, but I think it's better than the alternative of
further fragmentation. (I'm open to other ideas though, of course.)
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
This also isn't required (due to the vportprofile being stored in the
NetDef as a pointer rather than being directly contained), but it
seemed dishonest to not mark it as const (and thus permit users to
modify its contents)
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
In this case, the virNetDevBandwidthPtr that is returned is not to a
region within the virDomainNetDef arg, but points elsewhere (the
NetDef has the pointer, not the entire object), so technically it's
not necessary to make the return value a const, but it's a bit
disingenuous to *not* do it.
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
This makes it easier to understand which interface's config caused the
error.
Signed-off-by: Laine Stump <laine@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
The cpuModels member of _virQEMUCapsAccel struct is not a
virObject but regular struct with a free function defined:
qemuMonitorCPUDefsFree(). Use that when clearing parent structure
instead of virObjectUnref() to avoid a memleak:
==212322== 57,275 (48 direct, 57,227 indirect) bytes in 3 blocks are definitely lost in loss record 623 of 627
==212322== at 0x4838B86: calloc (vg_replace_malloc.c:762)
==212322== by 0x554A158: g_malloc0 (in /usr/lib64/libglib-2.0.so.0.6000.6)
==212322== by 0x17B14BF5: qemuMonitorCPUDefsNew (qemu_monitor.c:3587)
==212322== by 0x17B27BA7: qemuMonitorJSONGetCPUDefinitions (qemu_monitor_json.c:5616)
==212322== by 0x17B14B0B: qemuMonitorGetCPUDefinitions (qemu_monitor.c:3559)
==212322== by 0x17A6AFBB: virQEMUCapsFetchCPUDefinitions (qemu_capabilities.c:2571)
==212322== by 0x17A6B2CC: virQEMUCapsProbeQMPCPUDefinitions (qemu_capabilities.c:2629)
==212322== by 0x17A70C00: virQEMUCapsInitQMPMonitorTCG (qemu_capabilities.c:4769)
==212322== by 0x17A70DDF: virQEMUCapsInitQMPSingle (qemu_capabilities.c:4820)
==212322== by 0x17A70E99: virQEMUCapsInitQMP (qemu_capabilities.c:4848)
==212322== by 0x17A71044: virQEMUCapsNewForBinaryInternal (qemu_capabilities.c:4891)
==212322== by 0x17A7119C: virQEMUCapsNewData (qemu_capabilities.c:4923)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
On s390 machines host-passthrough and host-model CPUs result in the same
guest ABI (with QEMU new enough to be able to tell us what "host" CPU is
expanded to, which was implemented around 2.9.0). So instead of using
host-passthrough CPU when there's no CPU specified in a domain XML we
can safely use host-model and benefit from CPU compatibility checks
during migration, snapshot restore and similar operations.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
If a graphics device was added to XML that had no video device, libvirt
automatically added a video device which was always of type 'cirrus' on
x86_64, even if the underlying qemu didn't support cirrus.
This patch refines a bit the decision about the type of the video device.
Based on QEMU capabilities, cirrus is still preferred but only added if
QEMU supports it, otherwise VGA is used if supported by QEMU. There is now
no fallback as libvirt only aspires to generate a basic working config and
leaves anything more specific up to higher-level management tools.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Pavel Mores <pmores@redhat.com>
The default video device type selection algorithm we're about to deploy will
increase the amount of code dedicated to the task by amount enough to warrant
factoring the whole thing into its own function so as not to pollute the
caller qemuDomainDeviceVideoDefPostParse(). Do it now so that the actual
algorithm change later on is in a clean commit by itself and easy to review.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Pavel Mores <pmores@redhat.com>
This previous commit introduced a simpler free callback for
hash data with only 1 arg, the value to free:
commit 49288fac96
Author: Peter Krempa <pkrempa@redhat.com>
Date: Wed Oct 9 15:26:37 2019 +0200
util: hash: Add possibility to use simpler data free function in virHash
It missed two functions in the hash table code which need
to call the alternate data free function, virHashRemoveEntry
and virHashRemoveSet.
After the previous patch though, there is no code that
makes functional use of the 2nd key arg in the data
free function. There is merely one log message that can
be dropped.
We can thus purge the current virHashDataFree callback
entirely, and rename virHashDataFreeSimple to replace
it.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Now that all pieces are in place (hopefully) let's enable -blockdev.
We base the capability on presence of the fix for 'auto-read-only' on
files so that blockdev works properly, mandate that qemu supports
explicit SCSI id strings to avoid ABI regression and that the fix for
'savevm' is present so that internal snapshots work.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The 'savevm' HMP command didn't work properly with blockdev as it tried
to do snapshot of everything including the protocol nodes accessing
files which are not snapshottable. Qemu fixed this bug so now we need to
detect it to allow enabling blockdev.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The top level commands now can have 'feature' flags for fixes so add
support for querying those as well.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Initial implementation of 'auto-read-only' didn't reopen the backing
files when needed. For '-blockdev' to work we need to be able to tel
qemu to open a file read-only and change it during blockjobs as we label
backing chains with a sVirt label which does not allow writing. The
dynamic auto-read-only supports this as it reopens files when writing
is demanded.
Add a capability to detect that the posix file based backends support
the dynamic part.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The qemu driver will obey <backingStore> when we support blockdev.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
If user starts a blockcommit or a blockcopy then we modify access
for qemu on both images and leave it like that until the job
terminates. So far so good. Problem is, if user instead of
terminating the job (where we would modify the access again so
that the state before the job is restored) calls destroy on the
domain or if qemu dies whilst executing the block job. In this
case we don't ever clear the access we granted at the beginning.
To fix this, maybe a bit harsh approach is used, but it works:
after all labels were restored (that is after
qemuSecurityRestoreAllLabel() was called), we iterate over each
disk in the domain and remove XATTRs from the whole backing chain
and also from any file the disk is being mirrored to.
This would have been done at the time of pivot, but it isn't
because user decided to kill the domain instead. If we don't do
this and leave some XATTRs behind the domain might be unable to
start.
Also, secdriver can't do this because it doesn't know if there is
any job running. It's outside of its scope - the hypervisor
driver is responsible for calling secdriver's APIs.
Moreover, this is safe to call because we don't remember labels
for any member of a backing chain except of the top layer. But
that one was restored in qemuSecurityRestoreAllLabel() call done
earlier. Therefore, not only we don't remember labels (and thus
this is basically a NOP for other images in the backing chain) it
is also safe to call this when no blockjob was started in the
first place, or if some parts of the backing chain are shared
with some other domains - this is NOP, unless a block job is
active at the time of domain destroy.
https://bugzilla.redhat.com/show_bug.cgi?id=1741456#c19
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
There are four places where we remove image XATTRs and in all of
them we have the same for() loop with the same body. Move it into
a separate function because I'm about to introduce fifth place
where the same needs to be done.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Install the convertor function which enables the internals that will use
-blockdev to make qemu open the firmware image and stop using -drive.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The old way to instantiate a pflash device via -drive was a hack since
it's a platform device.
The modern approach calls for configuring it via -machine and takes the
node name as an argument.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
As a first step we will build the blockdevs which will be supposed to
back the pflash drives when moving away from -drive.
This code is similar to the way we build the blockdevs for the disk, but
skips the copy-on-read layer and doesn't implement any legacy approach.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Add a helper which will covert the PFLASH code file and variable file
into the virStorageSource objects stored in private data so that we can
use them with -blockdev while keeping the infrastructure to determine
the path to the loaders intact.
This is a temporary solution until we will want to do snapshots of the
pflash where we will be forced do track the full backing chain in the
XML.
In the meanwhile just convert it partially so that we can stop using
-drive.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
To allow converting the pflash drives to blockdev we will need a
virStorageSource to allow using our helpers. Temporarily prior to
coverting loader data to a virStorageSoruce add private data which will
house this.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Extract the old way to instantiate pflash devices to hold the firmware
via -drive to a separate function so that it can later be conditionally
disabled when -blockdev will be used.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Commit 5751a0b6b1 added a helper function
called virDomainCapsFeaturesInitUnsupported which initialized all domain
capability features as unsupported.
When adding a new feature this would initialize it as unsupported also
for hypervisor drivers which the original author possibly didn't intend
to modify. To prevent accidental wrong value being reported in such case
revert back to initializing individual features in the hypervisor
drivers themselves.
This is not a straight revert as additonal patches modified how we store
the capabilities.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain without a CPU model specified in the domain XML,
QEMU will choose a default one. Which is fine unless the domain gets
migrated to another host because libvirt doesn't perform any CPU ABI
checks and the virtual CPU provided by QEMU on the destination host can
differ from the one on the source host.
With QEMU 4.2.0 we can probe for the default CPU model used by QEMU for
a particular machine type and store it in the domain XML. This way the
chosen CPU model is more visible to users and libvirt will make sure
the guest will see the exact same CPU after migration.
Architecture specific notes
- aarch64: We only set the default CPU for TCG domains as KVM requires
explicit "-cpu host" to work.
- ppc64: The default CPU for KVM is "host" thanks to some hacks in QEMU,
we will translate the default model to the model corresponding to the
host CPU ("POWER8" on a Power8 host, "POWER9" on Power9 host, etc.).
This is not a problem as the corresponding CPU model is in fact an
alias for "host". This is probably not ideal, but it's not wrong and
the default virtual CPU configured by libvirt is the same QEMU would
use. TCG uses various CPU models depending on machine type and its
version.
- s390x: The default CPU for KVM is "host" while TCG defaults to "qemu".
- x86_64: The default CPU model (qemu64) is not runnable on any host
with KVM, but QEMU just disables unavailable features and starts
happily.
https://bugzilla.redhat.com/show_bug.cgi?id=1598151https://bugzilla.redhat.com/show_bug.cgi?id=1598162
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
QEMU 4.2.0 will report default CPU types used by each machine type and
we will want to start using it.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Almost all TCG query-machines replies match KVM. The only exceptions are
4.2.0 replies on s390x which differ in the reported default CPU type.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Some specifics of machine types may depend on the accelerator and thus
the data should be moved to virQEMUCapsAccel. The TCG machine types are
just copied from the ones probed for KVM to simplify the changes to
qemucapabilitiestest data files.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The function copies machine type data from one QEMU caps structure to
another.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In preparation for making machine types dependent on the accelerator,
the <machine> elements are formatted between <cpu type='kvm'> and
<cpu type='tcg'>.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
All the code for formatting machine type data was moved to a standalone
virQEMUCapsFormatMachines function.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
All the code for loading machine type data was moved to a standalone
virQEMUCapsLoadMachines function.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
To avoid duplicating code which selects the right virQEMUCapsAccel data
to be filled during probing.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
It is a tiny wrapper around virQEMUCapsProbeQMPCPUDefinitions which will
soon get private parameters and thus it cannot be exposed outside
qemu_capabilities.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
And make it use virQEMUCapsGetAccel once rather than repeating the same
code in all functions called from virQEMUCapsFormatAccel.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
And make it use virQEMUCapsGetAccel once rather than repeating the same
code in all functions called from virQEMUCapsLoadAccel.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The function can be used to get the pointer to all data which depend on
the accelerator.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This is container for capabilities data that depend on the accelerator.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The new functions are designed to load and format capabilities which
depend on the accelerator (host CPU expansion and CPU models).
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We need to create a mapping between CPU model names and their
corresponding QOM types.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We will need to keep some QEMU-specific data for each CPU model
supported by a QEMU binary. Instead of complicating the generic
virDomainCapsCPUModelsPtr, we can just directly store
qemuMonitorCPUDefsPtr returned by the capabilities probing code.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
