Peter Krempa [Tue, 22 Sep 2020 09:04:17 +0000 (11:04 +0200)]
virStorageSourceNew: Abort on failure
Add an abort() on the class/object allocation failures so that
virStorageSourceNew() always returns a virStorageSource and remove
checks from all callers.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
qemu: substitute missing model name for host-passthrough
Before:
$ uname -m
s390x
$ cat passthrough-cpu.xml
<cpu check="none" mode="host-passthrough" />
$ virsh hypervisor-cpu-compare passthrough-cpu.xml
error: Failed to compare hypervisor CPU with passthrough-cpu.xml
error: internal error: unable to execute QEMU command 'query-cpu-model-comp
arison': Invalid parameter type for 'modelb.name', expected: string
After:
$ virsh hypervisor-cpu-compare passthrough-cpu.xml
CPU described in passthrough-cpu.xml is identical to the CPU provided by hy
pervisor on the host
domain_conf.c: auto-align pSeries NVDIMM in virDomainMemoryDefPostParse()
The alignment for the pSeries NVDIMM does not depend on runtime
constraints. This means that it can be done in device parse
time, instead of runtime, allowing the domain XML to reflect
what the auto-alignment would do when the domain starts.
This brings consistency between the NVDIMM size reported by the
domain XML and what the guest sees, without impacting existing
guests that are using an unaligned size - they'll work as usual,
but the domain XML will be updated with the actual size of the
NVDIMM.
Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
conf, qemu: move qemuDomainNVDimmAlignSizePseries to domain_conf.c
We'll use the auto-alignment function during parse time, in
domain_conf.c. Let's move the function to that file, renaming
it to virDomainNVDimmAlignSizePseries(). This will also make it
clearer that, although QEMU is the only driver that currently
supports it, pSeries NVDIMM restrictions aren't tied to QEMU.
Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Ján Tomko [Wed, 23 Sep 2020 14:30:22 +0000 (16:30 +0200)]
util: do not unref event thread after joining it
g_thread_join() eats a reference.
==295055== Invalid read of size 4
==295055== at 0x4DA4AE4: g_thread_unref (in /usr/lib64/libglib-2.0.so.0.6400.5)
==295055== by 0x491D5FA: vir_event_thread_finalize (vireventthread.c:47)
==295055== by 0x4E6BCFF: g_object_unref (in /usr/lib64/libgobject-2.0.so.0.6400.5)
==295055== by 0x22F35CF4: qemuProcessQMPFree (qemu_process.c:8525)
==295055== by 0x22E71B58: glib_autoptr_clear_qemuProcessQMP (qemu_process.h:237)
...
==295055== by 0x22E98A29: qemuDomainPostParseDataAlloc (qemu_domain.c:5476)
==295055== by 0x49ABF83: virDomainDefPostParse (domain_conf.c:6023)
==295055== Address 0x2acb1c68 is 24 bytes inside a block of size 88 free'd
==295055== at 0x483B9F5: free (vg_replace_malloc.c:538)
==295055== by 0x4D80A4C: g_free (in /usr/lib64/libglib-2.0.so.0.6400.5)
...
==295055== by 0x491D5F1: vir_event_thread_finalize (vireventthread.c:46)
==295055== by 0x4E6BCFF: g_object_unref (in /usr/lib64/libgobject-2.0.so.0.6400.5)
==295055== by 0x22F35CF4: qemuProcessQMPFree (qemu_process.c:8525)
==295055== by 0x22E71B58: glib_autoptr_clear_qemuProcessQMP (qemu_process.h:237)
...
==295055== Block was alloc'd at
==295055== at 0x483A809: malloc (vg_replace_malloc.c:307)
==295055== by 0x4D80958: g_malloc (in /usr/lib64/libglib-2.0.so.0.6400.5)
...
==295055== by 0x4DA4C32: g_thread_try_new (in /usr/lib64/libglib-2.0.so.0.6400.5)
==295055== by 0x491D3BC: virEventThreadStart (vireventthread.c:159)
==295055== by 0x491D3BC: virEventThreadNew (vireventthread.c:185)
...
Signed-off-by: Ján Tomko <jtomko@redhat.com> Fixes: f4fc3db9204407874181117085756c9ced78adad Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Pavel Hrdina [Wed, 23 Sep 2020 14:31:21 +0000 (16:31 +0200)]
virnetdaemon: fix memory leak in virNetDaemonCallInhibit
g_variant_new() returns a weak reference which can be consumed by passing
to other g_variant* functions or to g_dbus_connection_call* functions.
This make it possible to call g_variant_new() directly as argument to
the functions above. Because this might be confusing I explicitly call
g_variant_ref_sink() to make it normal reference in both
virGDBusCallMethod() and virGDBusCallMethodWithFD() so the caller is
always responsible for the data.
Reported-by: Peter Krempa <pkrempa@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Peter Krempa [Wed, 23 Sep 2020 13:13:57 +0000 (15:13 +0200)]
docs: manpages: Strip table of contents from manpages
After meson conversion the man pages started to contain the table of
contents.
In autoconf we prevented this by a 'grep -v ::contents' in the command
building the manpages.
A more cultured solution is to strip out the 'contents' docutils element
directly.
Reported-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Ján Tomko [Wed, 23 Sep 2020 08:29:56 +0000 (10:29 +0200)]
tests: esxutilstest: depend on esx_gen_headers
Sometimes parallel compilation randomly fails on platforms
that do not have many drivers enabled, like macOS:
In file included from ../tests/esxutilstest.c:13:
../src/esx/esx_vi_types.h:62:10: fatal error: 'esx_vi_types.generated.typedef' file not found
#include "esx_vi_types.generated.typedef"
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
List esx_gen_headers as a source to stop meson from building
it before the headers are generated.
Pavel Hrdina [Mon, 21 Sep 2020 12:39:49 +0000 (14:39 +0200)]
virfirewalld: fix g_variant_get call
We need to pass pointer to `array`.
Reported-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Tested-by: Ján Tomko <jtomko@redhat.com>
Andrea Bolognani [Tue, 22 Sep 2020 11:41:49 +0000 (13:41 +0200)]
tests: Don't advertise VIR_TEST_EXPENSIVE to users
Right now, the logic that takes care of deciding whether expensive
tests should be run or not is not working correctly: more
specifically, it's not possible to use something like
$ VIR_TEST_EXPENSIVE=1 ninja test
to override the default choice, because in meson.build we always
pass an explicit value that overrides whatever is present in the
environment.
We could implement logic to make this work properly, but that
would require some refactoring of our test infrastructure and is
arguably of little value given that running
$ meson build -Dexpensive_tests=enabled
is very fast, so let's just stop telling users about the variable
instead and call it a day.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Matt Coleman [Tue, 22 Sep 2020 02:01:46 +0000 (22:01 -0400)]
libvirt: ensure defresult is used in virConnectAuthCallbackDefault
A previous change to this function's password handling broke the use of
default values for credential types other than VIR_CRED_PASSPHRASE and
VIR_CRED_NOECHOPROMPT.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Neal Gompa <ngompa13@gmail.com> Signed-off-by: Matt Coleman <matt@datto.com>
Fabian Freyer [Wed, 6 May 2020 13:35:55 +0000 (13:35 +0000)]
bhyve: add VNC password support
Support setting a password for the VNC framebuffer using the passwd
attribute on the <graphics/> element, if the driver has the
BHYVE_CAP_VNC_PASSWORD capability.
Note that virsh domxml-from-native does not output the password in the
generated XML, as VIR_DOMAIN_DEF_FORMAT_SECURE is not set when
formatting the domain definition.
Signed-off-by: Fabian Freyer <fabian.freyer@physik.tu-berlin.de> Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Fabian Freyer [Wed, 6 May 2020 13:35:54 +0000 (13:35 +0000)]
bhyve: probe for VNC password capability
Introduces the BHYVE_CAP_VNC_PASSWORD capability, which is probed by
parsing the error message from the bhyve command. When it is not
supported, bhyve -s 0,fbuf,password= will return an error message.
Signed-off-by: Fabian Freyer <fabian.freyer@physik.tu-berlin.de> Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Fabian Freyer [Wed, 6 May 2020 13:35:53 +0000 (13:35 +0000)]
bhyve: add support for setting fbuf resolution
The resolution of the VNC framebuffer can now be set via the resolution
definition introduced in 5.9.0.
Also, add "gop" to the list of model types the <resolution/>
sub-element is valid for.
Signed-off-by: Fabian Freyer <fabian.freyer@physik.tu-berlin.de> Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Fabian Freyer [Wed, 6 May 2020 13:35:52 +0000 (13:35 +0000)]
bhyve: support parsing fbuf PCI device
Add a new helper function, bhyveParsePCIFbuf, to parse the bhyve-argv
parameters for a frame-buffer device to <graphics/> and <video/>
definitions.
For now, only the listen address, port, and vga mode are detected.
Unsupported parameters are silently skipped.
This involves upgrading the private API to expose the
virDomainGraphicsDefNew helper function, which is used by
bhyveParsePCIFbuf.
Signed-off-by: Fabian Freyer <fabian.freyer@physik.tu-berlin.de> Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
In [1], changes were made to remove the existing auto-alignment
for pSeries NVDIMM devices. That design promotes strange situations
where the NVDIMM size reported in the domain XML is different
from what QEMU is actually using. We removed the auto-alignment
and relied on standard size validation.
However, this goes against Libvirt design philosophy of not
tampering with existing guest behavior, as pointed out by Daniel
in [2]. Since we can't know for sure whether there are guests that
are relying on the auto-alignment feature to work, the changes
made in [1] are a direct violation of this rule.
This patch reverts [1] entirely, re-enabling auto-alignment for
pSeries NVDIMM as it was before. Changes will be made to ease
the limitations of this design without hurting existing
guests.
tests: avoid close of bad file handle in commandtest
Closed file handles need to be initialized to -1, not 0. This caused a
inappropriate double close of stdin, which is not desirable, although
it had no ill effects.
Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
tests: don't mix FILE* and UNIX FD I/O on same stream
There is currently a hang in test27 that exhibits itself on FreeBSD 11.4
only. The behaviour is that virCommandProcessIO gets POLLIN on the
FD for stdout, but read() blocks. Meanwhile commandtest also blocks
in write for stderr because the pipe buffers are full.
This fix in commandhelper likely does not really address the root cause
just hides it due to the buffering done by FILE *. Mixing UNIX FD I/O
and FILE * I/O is bad practice regardless.
Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Currently, slot 1 is only allowed to be used by the LPC device.
Relax this requirement and allow to use slot 1 if it was explicitly
specified by the user for any other device type. In this case the LPC
device will have the next available address.
If slot 1 was not used by the user, it'll be reserved for the LPC
device, even if it is not configured to make address assignment
consistent in case the LPC device becomes necessary (e.g. the user
adds a console or a video device which require LPC).
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Support modeling of the 'isa' controller for bhyve. User can manually
define any PCI slot for the 'isa' controller, including PCI slot 1,
but other devices are not allowed to use this address.
When domain configuration requires the 'isa' controller to be present,
automatically add it on domain post-parse stage.
Now, as this controller is always available when needed, it's not
necessary to implicitly add it to the bhyve command line, so remove
bhyveBuildLPCArgStr().
Also, make bhyveDomainDefNeedsISAController() static as it's no longer
used outside of bhyve_domain.c.
As more than one ISA controller is not supported by bhyve,
and multiple controllers with the same index are forbidden,
so forbid ISA controllers with non-zero index for bhyve.
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Currently, this is needed for the bhyve driver to allow choosing a
specific PCI address for that. In bhyve, this controller is used to
attach serial ports and a boot ROM.
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
remote: slightly improve debugging of socket selection
The current debug message reports the "mode" after selection has
completed, however, the "mode" value can be changed by the selection
logic. It is thus beneficial to report most values upfront, and only
report newly changed values at the end.
Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The storage driver was wired up to support creating raw volumes in LUKS
format, but was never adapted to support LUKS-in-qcow2. This is trivial
as it merely requires the encryption properties to be prefixed with
the "encrypt." prefix, and "encrypt.format=luks" when creating the
volume.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The two removed files have exactly the same config as other LUKS volume
data files, simply with different file names. Consolidate down to just
two LUKS volume data files as that's all that we need for the test
coverage.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
b_info->u.hvm.{acpi,apic} are deprecated. But also, on recent libxl
version (4.14) the old one seems to be broken. While libxl part should
be fixed too, update the usage here and at some point drop support for
the old version.
b_info->acpi was added in Xen 4.8
b_info->apic was added in Xen 4.10
Xen 4.10 is the oldest version that still has security support (until
December 2020).
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Pavel Hrdina [Fri, 11 Sep 2020 12:12:07 +0000 (14:12 +0200)]
tests/virmock: extend number of arguments
Rewrite to use GLib DBus instead of libdbus will introduce function with
large number of arguments that we will have to mock for our tests so we
need to extend the number of arguments for our macros.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Pavel Hrdina [Tue, 8 Sep 2020 11:55:24 +0000 (13:55 +0200)]
util: introduce helpers for GLib DBus implementation
With libdbus our wrappers had a special syntax to create the DBus
messages by defining the DBus message signature followed by list
of arguments providing data based on the signature.
There will be no similar helper with GLib implementation as they
provide same functionality via GVariant APIs. The syntax is slightly
different mostly for how arrays, variadic types and dictionaries are
created/parsed.
Additional difference is that with GLib DBus everything is wrapped in
extra tuple (struct). For more details refer to the documentation [1].
Pavel Hrdina [Tue, 15 Sep 2020 11:55:53 +0000 (13:55 +0200)]
tests: mock libdbus in networkxml2firewalltest
This test calls into src/util/virfirewalld.c where it uses DBus to
figure out if firewalld is registered. Without the mock it luckily
fails and the test works correctly.
To isolate the tests from host environment we should mock the DBus
calls.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Pavel Hrdina [Thu, 3 Sep 2020 12:20:43 +0000 (14:20 +0200)]
remove HAL node device driver
There was one attempt a year ago done by me to drop HAL [1] but it was
never resolved. There was another time when Dan suggested to drop HAL
driver [2] but it was decided to keep it around in case device
assignment will be implemented for FreeBSD and the fact that
virt-manager uses node device driver [3].
I checked git history and code and it doesn't look like bhyve supports
device assignment so from that POV it should not block removing HAL.
The argument about virt-manager is not strong as well because libvirt
installed from FreeBSD packages doesn't have HAL support so it will not
affect these users as well [4].
The only users affected by this change would be the ones compiling
libvirt from GIT on FreeBSD.
I looked into alternatives and there is libudev-devd package on FreeBSD
but unfortunately it doesn't work as it doesn't list any devices when
used with libvirt. It provides libudev APIs using devd.
I also looked into devd directly and it provides some APIs but there are
no APIs for device monitoring and events so that would have to be
somehow done by libvirt.
Main motivation for dropping HAL support is to replace libdbus with GLib
dbus implementation and it cannot be done with HAL driver present in
libvirt because HAL APIs heavily depends on symbols provided by libdbus.
access: ensure ACL files are rebuilt when protocol changes
Meson is not told that the .x protocol files are an input for the
generator, so it doesn't know to setup a rebuild dependancy.
Reviewed-by: Erik Skultety <eskultet@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
qemu: fix concurrency crash bug in force snapshot revert
This patch is just revert of [1]. Actually we should NOT pass
QEMU_ASYNC_JOB_NONE as that patch suggests while we are in async job in order
to acquire nested jobs correctly. The patch tries to fix issues introduced by
another patch [2] where jobs are mistakenly cleared out in qemuProcessStop.
Later patch [3] fixed the issue introduced by patch [2]. Now we need to revert
[1] as well as we now still have same concurrency crash issues as [3] described
but for the force revert.
[1] 0c4408c83: qemu: Don't use asyncJob after stop during snapshot revert
[2] 888aa4b6b: qemuDomainObjPrivateDataClear: Don't leak @migParams
[3] d75f865fb: qemu: fix concurrency crash bug in snapshot revert
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Tue, 15 Sep 2020 15:58:04 +0000 (17:58 +0200)]
qemuBuildHostdevSCSIAttachPrepare: Propagate 'readonly' flag also for iSCSI
The 'readonly' hostdev property is stored separately from the
virStorageSource as some hostdevs are not described by a virStorage
source. We need to propagate the flag to the virStorage source also for
iSCSI backends as it's used to generate the backend properties.
Peter Krempa [Thu, 10 Sep 2020 13:43:54 +0000 (15:43 +0200)]
qemuDomainPrepareHostdev: Don't base backend nodename on device alias
QEMU's blockdev nodenames which are used to back SCSI/iSCSI hostdevs are
limited to 32 characters. If a user passes a very long user alias as
name of the host device it's easy to end up with a too-long nodename.
To prevent this from happening don't base the nodename on the possibly
user-specified alias but on the normal sequential node name generator.
We then store the name in the status XML for further use.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 10 Sep 2020 10:32:04 +0000 (12:32 +0200)]
qemu: domain: Extract preparation of hostdev specific data to a separate function
Historically we've prepared secrets for all objects in one place. This
doesn't make much sense and it's semantically more appealing to prepare
everything for a single device type in one place.
Move the setup of the (iSCSI|SCSI) hostdev secrets into a new function
which will be used to setup other things as well in the future.
This is a similar approach we do for disks.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
This was a hack when we were locally regenerating the nodename so that
it's not leaked. Now that we use proper virStorageSource with
persistence it's no longer required.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 10 Sep 2020 06:25:37 +0000 (08:25 +0200)]
qemuBuildHostdevSCSI(A|De)tachPrepare: Use virStorageSource in def for SCSI hostdevs
Modify the attach/detach data generators to actually use the
virStorageSourceStructure embedded in the SCSI config data rather than
creating an ad-hoc internal one.
The modification will allow us to properly store the nodename used for
the backend in the status XML rather than re-generating it all the
time.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 9 Sep 2020 15:58:34 +0000 (17:58 +0200)]
qemu: domain: Fill in (i)SCSI backend nodename if it is not present in status XML
For upgrade reasons so that we can modify the used nodename we must
generate the old version for all status XMLs which don't have it stored
explicitly.
The change will be required as using the user-provided alias may result
in too-long nodenames which will be rejected by qemu.
Add code which fills in the appropriate old value and add test cases to
validate that it's added and also that existing nodenames are not
overwritten.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 9 Sep 2020 13:27:45 +0000 (15:27 +0200)]
conf: Add virStorageSource member for SCSI host device config data
The backend for the SCSI host device is a storage source. While the
definition doesn't look like that it's converted to a storage source
when the VM is running.
Add the storage source to the definition object and also parse/format
its private data which will be used for internal state storage while
the VM is running.
Note that the virStorageSourcePtr may not be allocated all the time so
the private data parser allocates it if there is any private data
present.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
projects / thirdparty / libvirt.git / log
