Currently, memory device (def->mems) part of cmd line is
generated before any controller. In majority of cases it doesn't
matter because neither of memory devices live on a bus that's
created by an exposed controller (e.g. there's no DIMM
controller, at least not exposed). Except for virtio-mem and
virtio-pmem, which do have a PCI address. And if it so happens
that the device goes onto non-default bus (pci.0) starting such
guest fails, because the controller that creates the desired bus
wasn't processed yet. QEMU processes arguments in order.
For instance, if virtio-mem has address with bus='0x01' QEMU
refuses to start with the following message:
Bus 'pci.1' not found
Similarly for virtio-pmem. I've successfully tested migration and
changing the order does not affect migration stream.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2047271
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Now that qemu fixed device unplug when JSON syntax is used with -device
we can re-enable the feature.
Since the old capability string representation is condemned by
suggesting filtering it as a workaround we must introduce a new string.
To achieve this the original capability position is renamed to
X_QEMU_CAPS_DEVICE_JSON_BROKEN_HOTPLUG and a new position with the
original name QEMU_CAPS_DEVICE_JSON is introduced to prevent us having
to change the rest of the code.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When -device is configured via JSON a bug [1] is triggered in qemu were
the DEVICE_DELETED event for the removal of the device frontend is no
longer delivered to libvirt. Without the DEVICE_DELETED event we don't
remove the corresponding entries in the VM XML.
Until qemu will be fixed we must stop using the JSON syntax for -device.
This patch removes the detection of the capability. The capability is
used only during startup of a fresh VM so we don't need to consider any
compaitibility steps for existing VMs.
For users who wish to use 'libvirt-7.9' and 'libvirt-7.10' with
'qemu-6.2' there are two possible workarounds:
- filter out the 'device.json' qemu capability '/etc/libvirt/qemu.conf':
capability_filters = [ "device.json" ]
- filter out the 'device.json' qemu capability via qemu namespace XML:
<domain type='kvm' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'>
[...]
<qemu:capabilities>
<qemu:del capability='device.json'/>
</qemu:capabilities>
</domain>
We must never again use the same capability name as we are now
instructing users to filter it as a workaround so once qemu is fixed
we'll need to pick a new capability value for it.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=2036669
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2035237
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Upcoming patches will modify how we populate the capability cache in
tests to be more saner. This also means that the emulator binary and
architecture used in the test files using real capabilities must match
what the real capabilities have.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We currently use -machine accel=XXX which is just a syntax sugar
for -accel XXX. The former doesn't allow specifying arguments for
accelerator, because all arguments passed to -machine are
treated as arguments of machine itself.
The -accel argument was introduced in QEMU commit
v2.9.0-rc0~70^2~19 and since our minimum required version is
newer (2.11.0) we can safely assume its existence and use it
without any capability.
Resolves: https://gitlab.com/libvirt/libvirt/-/issues/233
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Tested-by: Kashyap Chamarthy <kchamart@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
'-audiodev' as a modern implementation based on QAPI already takes JSON
as the argument. Convert our code to use it directly.
The declaration of the QAPI types can be found in
'qemu.git/qapi/audio.json'.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Starting with QEMU-6.2 started accepting a JSON object as argument for
'-device' which will also become the only syntax considered stable by
qemu in the future.
Since libvirt was recently converted to generate the properties via JSON
to begin wit we can start using it on the commandline as well, by simply
enabling the QEMU_CAPS_DEVICE_JSON capability, which we do by probing
for the 'json-cli' feature flag of 'device_add'.
Normally a change which changes a commandline output should be happening
only after the impacted real-caps test files are forked in the version
preceding the change, but in this case it's not necessary as the logic
for generating the device properties stays identical and we just change
the output format (avoid conversion). Additionally we still have a lot
of tests validating the conversion to the old commandline options.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
virCommandToString has the possibility to return an already wrapped
string with better format than what we get from the test wrapper script.
The main advantage is that arguments for an option are always on the
same line which makes it more easy to see what changed in a diff and
prevents re-wrapping of the line if a wrapping point moves over the
threshold.
Additionally the used output is the same we have in the VM log file when
a VM is starting.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Base the detection on the presence of the 'secret' qom-type entry, which
isn't conditionally compiled in qemu.
All caps-based test now switch to using JSON for -object.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The -audiodev argument is replacing the QEMU_AUDIO_DRV env variable (and
its relations).
Sadly we still have to use the SDL_AUDIODRIVER env variable because that
wasn't mapped into QAPI schema.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
QEMU has long accepted many different values for boolean properties, but
set accepted has been different depending on which QEMU parser you hit.
The on|off values were supported by all QEMU parsers. The yes|no, y|n,
true|false values were only partially supported:
https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg01012.html
Thus we should standardize on on|off everywhere since that is most
widely supported in QEMU.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The preferred syntax for boolean options is to set the value "on" or
"off". QEMU 7.1.0 will deprecate the short format we currently use.
The long format has been supported with -chardev since at least 1.5.3,
so we don't need to check for it.
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Currently, we configure QEMU to prealloc memory almost by
default. Well, by default for NVDIMMs, hugepages and if user
asked us to (via memoryBacking <allocation mode="immediate"/>).
However, when guest's NVDIMM is backed by real life NVDIMM this
approach is not the best. In this case users should put <pmem/>
into the <memory/> device <source/>, like this:
<memory model='nvdimm' access='shared'>
<source>
<path>/dev/pmem0</path>
<pmem/>
</source>
</memory>
Instructing QEMU to do prealloc in this case means that each
page of the NVDIMM is "touched" (the first byte is read and
written back - see QEMU commit v2.9.0-rc1~26^2) which cripples
device wear.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1894053
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
In 88957116c9 I've switched to -machine memory-backend=ID and
-object memory-backend-* because QEMU is obsoleting -mem-path
and -mem-prealloc. However, what I did not foresee was that using
-machine memory-backend in combination with -numa is not allowed
in QEMU. This was reported upstream and fortunately not released
yet.
The problem is that if domain has NUMA nodes then we will
generate memory-backend-* objects for NUMA nodes (because if QEMU
is new enough to expose default RAM ID it also supports -numa
memdev=) and adding non-NUMA memory backend is wrong.
Reported-by: Masayoshi Mizuma <msys.mizuma@gmail.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
So far, Libvirt configures memory-backend-* for memory hotplug,
possibly NUMA nodes and in a few other cases. This patch
switches to constructing the memory-backend-* command line for
all cases. To keep ability to migrate guests a little hack is
used: the ID of the object is set to the one that QEMU uses
internally anyways. These IDs are stable (first started to appear
somewhere around v0.13.0-rc0~96) and can't change.
In fact, this patch does exactly what QEMU does internally. The
reason for moving the logic into Libvirt is that QEMU wants to
deprecate the old style of specifying memory.
So far, only x84_64 test cases are changed, because tests for
other architectures use older capabilities, which still lack the
QEMU_CAPS_MACHINE_MEMORY_BACKEND capability and they don't report
the RAM ID.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1836043
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
All three memory backends (-file, -ram and -memfd) have .prealloc
attribute. Since we are setting it only for -file, the
corresponding code lives only under if() that handles that
specific backend. But in near future we will want to set the
attribute for other backends too. Therefore, move the
corresponding code outside of the if().
This causes some .argv files to be changed, but the only change
happening there is move of the attribute (best viewed with:
'git show --color-words=.').
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Note that 'numa-mem-supported' turned off for certain machine types
which in turn forced us to generate a newer command line in certain
tests.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Commit 82576d8f35 used a string "on" to enable the 'pmem' property.
This is okay for the command line visitor, but the property is declared
as boolean in qemu and thus it will not work when using QMP.
Modify the type to boolean. This changes the command line, but
fortunately the command line visitor in qemu parses both 'yes' and 'on'
as true for the property.
https://bugzilla.redhat.com/show_bug.cgi?id=1854684
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The previous approac of just purging the alias combined with the fact
that we filled in fake machine types in the test data meant that if a
test case used an alias machine type such as 'pc' or 'q35' it would not
properly resolve to the actual data returned by qemu.
This started to be a problem since the CPU driver now looks at the
default CPU reported with the machine type.
This patch replaces the original approach of just removing the alias by
replacing it with a copy of the machine type data which the type would
alias to. This means that we are using the real data while we don't
modify the test output after every qemu upgrade.
Additionally this change will allow us to drop adding the fake machine
types later.
The test fallout is from actually excercising the CPU driver with
actual data.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
QEMU since 4.1.0 supports the "dies" parameter for -smp
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The 32-bit x86 binary is called qemu-system-i386, not
qemu-system-i686. This mistake across many test XML files was
not noticed because the mistake was also made in testutilsqemu.c
when mocking the capabilities.
Reviewed-by: Erik Skultety <eskultet@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 latter is deprecated and will be removed soon. The advised
replacement is '-overcommit mem-lock=on|off'.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
By default, qemu user's home dir points to '/' which shouldn't be used
at all. We therefore pass the HOME variable from the current variable
iff not running as SUID, which means that for systemd we never set it.
This patch makes sure, that for system QEMU this is always set to
libDir/<driver>, session mode is left untouched.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
For session mode, only XDG_CACHE_HOME is set, because we want to remain
integrating with services in user session, but for system mode, this
would have become reading/writing to '/' which carries the obvious issue
with permissions (also, '/' is the wrong location in 99.9% cases anyway).
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
According to the result parsing from xml, add pmem property
into QEMU command line:
-object memory-backend-file,...[,pmem=on]
Signed-off-by: Luyao Zhong <luyao.zhong@intel.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
