Peter Krempa [Wed, 15 Sep 2021 13:09:00 +0000 (15:09 +0200)]
virshDomainGetVcpuBitmap: Return bitmap when taking the fallback path
In case the specific VCPU states are not present in the XML we were
taking a fallback code path just noting that all cpus of the VM are
enabled.
This was broken by a mistake in a recent refactor where a 'goto cleanup'
was mistakenly replaced by a 'return NULL'. This broke reporting of cpus
and also caused a memory leak.
Return the fallback cpu map.
Fixes: bd1f40fe7d4
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2004429 Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Jiri Denemark <jdenemar@redhat.com> Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Peter Krempa [Wed, 15 Sep 2021 11:11:40 +0000 (13:11 +0200)]
qemu: validate: Allow 'preserve' action for on_crash lifecycle action
In fact keeping the VM around for debugging is a desirable configuration
and actually the implementation has no code as we keep the VM around.
Remove the validation and add a note that it's actually used.
Fixes: b1b85a475fb251b9068b75f629479f5c452f1b43 Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Peter Krempa <pkrempa@redhat.com> Acked-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: Boris Fiuczynski <fiuczy@linux.ibm.com>
Nick Chevsky [Mon, 13 Sep 2021 06:16:18 +0000 (01:16 -0500)]
qemu_tpm: Start swtpm(8) daemon with --terminate switch
Launch swtpm(8) with the --terminate switch, which guarantees that
the daemon will shut itself down when QEMU dies (current behavior).
We had so far been getting this "for free" (i.e. without --terminate)
due to a defect in upstream's connection handling logic [1], on which
libvirt should not rely since it will eventually be fixed. Adding
--terminate preserves and guarantees the current behavior.
Michal Privoznik [Sun, 12 Sep 2021 06:28:48 +0000 (08:28 +0200)]
node_device_conf: Don't prealloc @vfs in virNodeDeviceGetPCISRIOVCaps()
The array of virtual functions @vfs in
virNodeDeviceGetPCISRIOVCaps() is allocated twice: the first time
during its declaration and the second time inside
virPCIGetVirtualFunctions() which leads to a memleak:
==16691== 1,128 bytes in 47 blocks are definitely lost in loss record 1,771 of 1,803
==16691== at 0x4844CC1: calloc (vg_replace_malloc.c:1117)
==16691== by 0x4E50070: g_malloc0 (in /usr/lib64/libglib-2.0.so.0.6800.3)
==16691== by 0x4A7B034: virNodeDeviceGetPCISRIOVCaps (node_device_conf.c:2649)
==16691== by 0x4A7B5E2: virNodeDeviceGetPCIDynamicCaps (node_device_conf.c:2762)
==16691== by 0xA7F6E18: udevProcessPCI (node_device_udev.c:418)
Fixes: c97518d9b833a607f29b9bb02e3fbe74c011c088 Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Fri, 10 Sep 2021 12:55:02 +0000 (14:55 +0200)]
docs: Format @variable properly
When documenting our public API in some places we use '@' to
refer to the variable. For instance:
* This API tries to set guest time to the given value. The time
* to set (@seconds and @nseconds) should be in seconds relative
* to the Epoch of 1970-01-01 00:00:00 in UTC.
However, when generating HTML documentation these tokens are
copied verbatim. What we can do is drop the '@' character and
wrap the variable in <code/> so that it is formatted properly.
Due to the way we 'parse' docs a token might actually be slightly
more than just '@variable'. For instance in the example above we
will have the following tokens: '(@seconds' and '@nseconds)'.
Thus we need to handle possible substring before and after
variable.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
These two public APIs are implemented for almost all other objects that
have a concept of persistent definition and activatability. Now that we
have node devices (mdevs) that can be defined and inactive, it will be
useful to query the persistent/active state of node devices as well.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Jim Fehlig [Tue, 14 Sep 2021 17:11:11 +0000 (11:11 -0600)]
libxl: Don't autostart domains on driver reload
When libxlAutostartDomain was introduced with commit fb92307f0d, one hunk
mistakenly added a call site in libxlStateReload. Domains should not be
autostarted when reloading the driver, so remove the offending hunk.
Signed-off-by: Jim Fehlig <jfehlig@suse.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Jim Fehlig [Mon, 13 Sep 2021 22:04:41 +0000 (16:04 -0600)]
libxl: Fix driver reload
On reload, the libxl driver calls virDomainObjListLoadAllConfigs to load
all configs from /etc/libvirt/libxl/ but incorrectly passes 'true' for
the liveStatus parameter, resulting in error messages such as
libvirtd[21053]: XML error: unexpected root element <domain>, expecting <domstatus>
libvirtd[21053]: Failed to load config for domain 'sles15sp3'
Fix by not requesting live status when re-reading the persistent VM config
files.
Signed-off-by: Jim Fehlig <jfehlig@suse.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Jim Fehlig [Fri, 10 Sep 2021 20:34:09 +0000 (14:34 -0600)]
libxl: Improve reporting of die_id in capabilities
On Xen, libvirt runs in a VM (typically dom0) and does not have an accurate
picture of numa and cpu topology of the underlying physical machine using
the "usual" mechanisms. numa info and cpu toplogy are retrieved from libxl
and used to populate the libvirt conterparts. Commit 7b79ee2f78b introduced
support for reporting die_id in capabilities, but did not account for
special handling of numa and cpu topology in libxl.
Currently, Xen does not report die_id in the libxl_cputopology structure.
In the meantime, set die_id to 0, which was suggested by the Xen developers
and is slightly better than random garbage such as
Andrea Bolognani [Fri, 10 Sep 2021 14:48:25 +0000 (16:48 +0200)]
ci: Add Debian 11 builds
The armv7l and ppc64le cross-builds as well as the Clang build
are adopted from Debian 10, while the mips64el build is adopted
from Debian sid. As always, the way jobs are distributed across
Debian versions is fairly arbitrary.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Andrea Bolognani [Fri, 10 Sep 2021 13:35:49 +0000 (15:35 +0200)]
ci: Avoid use of magic constants
The value 3 is the length of the "ci-" prefix, which is present
in the items returned by get_registry_images() but not in those
returned by get_dockerfiles().
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Beraldo Leal <bleal@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Andrea Bolognani [Fri, 10 Sep 2021 13:27:44 +0000 (15:27 +0200)]
ci: Restore information about Coverity integration
These were removed along with the outdated information on how
to regenerate the Dockerfiles contained in the repository, but
this part is still relevant.
Reverts: 30856d2865122f82ec25c258c7cd2cceb52aa7a5 (partially) Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Sun, 12 Sep 2021 07:30:29 +0000 (09:30 +0200)]
docs: Fix dimm example
In the example for <memory model='dimm'/> we show how to
configure hugepages as backend. In the example we show 4MiB
hugepages which are non-standard and thus at the first glance may
mislead users thinking that a regular sized pages (4K) will be
used. Use 2MiB as the value instead.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
src: network_conf: propagate only bool to virNetworkDefParseString()
We don't need to propagate all public flags, only the information
about the presence of the validation one, which can differ from
function to function. This patch makes it easier and more
readable in case of a future additions of validation flags.
This change was suggested by Daniel.
Signed-off-by: Kristina Hanicova <khanicov@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Fri, 10 Sep 2021 13:33:17 +0000 (15:33 +0200)]
qemuDomainChangeDiskLive: Modify 'startupPolicy' before changing source
We don't support all startup policies with all source types so to
correctly allow switching from a 'file' based cdrom with 'optional'
startup policy to a 'block' based one which doesn't support optional we
must update the startup policy field first. Obviously we need to have
fallback if the update fails.
Reported-by: Vojtech Juranek <vjuranek@redhat.com> Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Fri, 10 Sep 2021 12:16:50 +0000 (14:16 +0200)]
qemuProcessQMPInit: Use long options for '-qmp' when probing qemu.
'-qmp' in this case behaves the same as '-chardev' so it should have
been converted the same way as others were in 43c9c0859f2d53321 since
short options are deprecated.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
ci: remove obsolete logic for refreshing containers/cirrus
We now use lcitool's manifest feature to generate files. The logic
for checking for stale containers in the registry, however, is still
relevant so that is propagated to a standalone command.
Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
gitlab: fold sanitizer jobs into regular ubuntu jobs
The sanitizer jobs run in ubuntu 20.04 containers and thus overlap with
testing already done for the regular ubuntu 20.04 build job. Fold the
sanitizer run for GCC into the regular build job and add a second
ubuntu 20.04 build job for CLang sanitizers.
Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Michal Privoznik [Fri, 10 Sep 2021 08:58:55 +0000 (10:58 +0200)]
libvirt-domain.c: Fix typo in virDomainAttachDevice doc
In the documentation to virDomainAttachDevice() we refer to a
non-existent virDomainUpdateDeviceFlag() function. The correct
name is virDomainUpdateDeviceFlags().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Stefan Hajnoczi [Thu, 9 Sep 2021 14:58:29 +0000 (15:58 +0100)]
docs: virtiofs: use the preferred virtiofs spelling
The virtiofs project started off using "virtio-fs" but later switched to
the "virtiofs" spelling because it matches the spelling of the mount -t
virtiofs command-line. Update the kbase article with the new spelling so
it matches the virtiofs website.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Stefan Hajnoczi [Thu, 9 Sep 2021 14:58:28 +0000 (15:58 +0100)]
docs: virtiofs: move legacy docs to the bottom
A number of legacy issues make the virtiofs kbase article hard to
understand. Most users don't need to configure NUMA or a memory backend
other than memfd. Move that information to the bottom of the article so
the recommended syntax is most prominent.
Suggested-by: Richard W.M. Jones <rjones@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Richard W.M. Jones <rjones@redhat.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
qemu: switch to use -M graphics=off instead of -device sga
SeaBIOS >= 1.11 has built-in support for outputting to the serial
console when QEMU sets -M graphics=off. Our minimum QEMU version
is 2.11.0, which bundled SeaBIOS 1.11. Thus we have no need to
use '-device sga' anymore.
This change results in a slight layout difference for option ROMs
in memory, however, it does not affect the migration data stream
format on the wire and once migration is complete the target QEMU
memory layout for ROMs matches the source QEMU once again.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
qemu: tweak error message to be more general purpose
The BIOS serial console output is currently implemented using the QEMU
'sga' device, but this is going to change in future patches, so the
error message ought to be more generically phrased.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
qemu: prevent use of <bios useserial='yes'> on non-x86 arches
The <bios useserial='yes'> config results in use of the '-device sga'
QEMU options. This in turn causes QEMU go load the sgabios.bin option
ROM, which contains x86 machine code. This cannot work on non-x86
arches, thus we should block the bad config.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
William Douglas [Wed, 8 Sep 2021 18:01:23 +0000 (11:01 -0700)]
ch_driver: Handle validation failure correctly
When validation like deviceValidateCallback fails, the vm will not be
set and so the call to virDomainObjListRemove will be passed a NULL
pointer causing a segfault. To prevent this add a check that the vm is
defined before calling out to virDomainObjListRemove.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: William Douglas <william.douglas@intel.com>
William Douglas [Wed, 8 Sep 2021 18:01:20 +0000 (11:01 -0700)]
ch_process: Handle enabled console devices
Add functionality to allow libvirt console to connect to the
cloud-hypervisor created PTY associated with a VM by updating the
domain with console path information. This has to be run after the VM
is created by cloud-hypervisor.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: William Douglas <william.douglas@intel.com>
William Douglas [Wed, 8 Sep 2021 18:01:18 +0000 (11:01 -0700)]
ch_monitor: Use virCHMonitorGet to access cloud-hypervisor API
Now that virCHMonitorGet is capable of handling data returned by the
cloud-hypervisor API, make use of this via virCHMonitorGetInfo to call
into the vm.info endpoint.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: William Douglas <william.douglas@intel.com>
William Douglas [Wed, 8 Sep 2021 18:01:17 +0000 (11:01 -0700)]
ch_monitor: Update virCHMonitorGet to handle accept a response
The virCHMonitorGet function needed to be able to return data from the
hypervisor. This functionality is needed in order for the driver to
support PTY enablement and getting details about the VM state.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: William Douglas <william.douglas@intel.com>
Peter Krempa [Wed, 8 Sep 2021 15:57:30 +0000 (17:57 +0200)]
virstoragetest: Reinstate testing of images without 'backing_fmt'
There are important security implications when we'd misprobe those
images. This commit reinstates the tests removed by commit 979d1ba3ae13
since 'qemu-img' refused to format them.
With the new testing approach with stored images we won't run into that
problem.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 9 Sep 2021 07:53:30 +0000 (09:53 +0200)]
virstoragetest: Remove pointless goto from mymain
Improve the error message and abort the test. Continuing here is not
desired as without chdiring into the appropriate directory the test
would fail anyways and worse could attempt stat-ing random files on the
host.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 8 Sep 2021 15:11:00 +0000 (17:11 +0200)]
virstoragetest: Assume that 'qemu-img' supports '-o compat='
All supported qemu versions have the parameter, so we don't need to
check. This allows us to simplify the code used for formating real
images for virstoragetest.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 8 Sep 2021 14:49:03 +0000 (16:49 +0200)]
virstoragetest: Stop rewriting images in 'mymain'
For testing of real images formatted by 'qemu-img' it's now sufficient
to format them once without the need to rewrtie them since we use the
real images only for testing of one scenario.
This allows us to also remove most of the global variables holding the
path to the images which was necessary when they were being rewritten.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 8 Sep 2021 14:44:13 +0000 (16:44 +0200)]
virstoragetest: Unify testing of QCOW2 images with absolute backing
We have 3 test cases for this currently:
1) "qcow2->raw"
1.1) VIR_STORAGE_FILE_QCOW2 as top level format
1.2) VIR_STORAGE_FILE_AUTO as top level format
2) "wrap->qcow2->raw" whith just VIR_STORAGE_FILE_QCOW2
This patch adds also testing of VIR_STORAGE_FILE_AUTO for case 2) and
removes both 1) subcases as they are being actually tested as part of
2).
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Wed, 8 Sep 2021 07:06:08 +0000 (09:06 +0200)]
virstoragetest: Use a pre-formatted QED file for testing backing store extraction
The QED format isn't really being developed any more. Use a
pre-formatted image to test the existing code. In this instance we
switch to using a relative backing path for simplicity.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Tue, 7 Sep 2021 16:58:32 +0000 (18:58 +0200)]
virstoragetest: Rework TEST_LOOKUP* cases to work on fake backing chain
Rather than using 'qemu-img' and rewriting the chain we can use fake
data and few empty files to ensure the same level of coverage. This is
possible since we've already tested that the metadata parsing from files
works properly and the only thing we are testing here is that the
symlink resolution works properly.
Additionally after the refactor of 'virstoragetest' is complete
additional tests on real data will be added.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Mon, 6 Sep 2021 12:13:29 +0000 (14:13 +0200)]
virstoragetest: Store output of TEST_CHAIN in output files
The TEST_CHAIN cases were storing the expected output (or rather data
to generate the expected output) in code. This made the code really hard
to follow and even harder to modify to add new cases.
This patch modifies the code to store the expected output in text files
(using the same generator as we've used to) and uses
'virTestCompareToFile' to check the outputs.
The result is that the code is way simpler and doesn't require fiddling
with 'testFileData' structs when adding new cases. Additionally this
removes mixing of code and declaration so we can stop disabling the
warning for this file.
Another advantage is that the tests are now named so it's easier to
figure out if one of them breaks.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Tue, 7 Sep 2021 13:31:42 +0000 (15:31 +0200)]
testutils: Introduce helper for stripping bulilddir/srcdir from test outputs
In certain cases we want to be able to compare test output containing
real paths against a static output file and thus we need a helper which
strips srcdir/builddir from given path.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Mon, 6 Sep 2021 11:35:09 +0000 (13:35 +0200)]
virstoragetest: Drop testing of NBD backends via parsing real images
We now have specific tests for the backing store parser and previous
tests cover the extraction of the backing store string so there's no
need for these particular tests.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
projects / thirdparty / libvirt.git / log
