Commit Graph

42240 Commits

Author SHA1 Message Date
Michal Privoznik
ccb8acc966 numa_conf: Rename virDomainNumaDistance to virNumaDistance
There's nothing domain specific about NUMA distances. Rename the
virDomainNumaDistance structure to just virNumaDistance.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-24 14:49:28 +02:00
Michal Privoznik
fe25224fda capabilities: Rename siblings to distances
The virCapsHostNUMACellSiblingInfo structure really represents
distance to other NUMA node. Rename the structure and variables
of that type to make it more obvious.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-24 14:49:28 +02:00
Pavel Hrdina
8b118c909a Revert "qemuDomainSetNumaParamsLive: set nodeset for root cgroup"
This reverts commit <1b22dd6dd44202094e0f78f887cbe790c00e9ebc>.

First of all, the reverted commit is incomplete. It only sets
cpuset.mems in the VM root cgroup when the API is used but there is no
code that would do the same when the VM is started.

Libvirt never places any process into the VM root cgroup directly. All
the supporting processes like slirp-helper or dbus-daemon are placed
into the emulator sub-cgroup and all the QEMU threads are distributed
between emulator, vcpu* and iothread* sub-cgroups. The scenario
described in the reverted commit can happen only if someone manually
adds any process there which we should not care about.

If we would like to set the limit in the VM root cgroup we need to
introduce better logic:

    - set both (old and new) numa group in the VM root cgroup
    - change the numa group in all sub-cgroups to new value
    - finally set only the new value in the VM root cgroup

The simplest fix now is to revert the commit.

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-24 14:31:42 +02:00
Han Han
6ff937747c kbase: Fix broken link of migration doc
Signed-off-by: Han Han <hhan@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
2021-05-24 12:47:15 +02:00
Olaf Hering
d8c468d58c libxl: remove libxl_domain_build_info_init from libxlMakeDomBuildInfo
The used libxl_domain_build_info, which is contained in
libxl_domain_config, is owned and already initialized by the caller.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
2021-05-21 10:59:47 -06:00
Olaf Hering
2b51050b10 libxl: remove libxl_domain_create_info_init from libxlMakeDomCreateInfo
The passed libxl_domain_create_info is owned, and already initialized,
by the caller.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
2021-05-21 10:58:53 -06:00
Olaf Hering
2d0decccba libxl: adjust handling of libxl_device_disk objects
libxl objects are supposed to be initialized and disposed.
Correct the usage of libxl_device_disk objects which are allocated on
the stack. Initialize each one prior usage, and dispose them once done.

Adjust libxlMakeDisk to use an already initialized object, it is owned
by the caller.

Adjust libxlMakeDiskList to initialize the list of objects, before they
are filled by libxlMakeDisk. In case of error, the objects are disposed
by libxl_domain_config_dispose.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
2021-05-21 10:57:36 -06:00
Kristina Hanicova
7ab1ba112f conf: Report alias name of the detached device in error
This is v2 from:
https://listman.redhat.com/archives/libvir-list/2021-May/msg00481.html

I have reworked the code a bit to have only one error report
instead of multiple ones with different combinations of possible
matching items. Suggested by Laine.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1942367

Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
2021-05-21 12:46:31 -04:00
Kristina Hanicova
b6b6725c95 qemu: Return -EINVAL to keep qemuDomainOpenFile() consistent
The description of the function says that the return value is a
file descriptor on success and negative errno on failure which is
not true. If the 'if' case with check on security labels fails,
the return value is -1 not -errno. The solution is to return
'-EINVAL' instead.

Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-21 14:38:09 +02:00
Kristina Hanicova
bcdaa91a27 qemu: Use qemuDomainOpenFile() in qemuPrepareNVRAM()
Previously, nvram file was created with user/group owner as
'root', rather than specifications defined in libvirtd.conf. The
solution is to call qemuDomainOpenFile(), which creates file with
defined permissions and qemuSecurityDomainSetPathLabel() to set
security label for created nvram file.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1783255

Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-21 14:36:57 +02:00
Kristina Hanicova
483e943884 qemu: Build command line for object input-linux
Resolves: https://gitlab.com/libvirt/libvirt/-/issues/142

Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
2021-05-21 14:29:09 +02:00
Kristina Hanicova
cff0444e51 conf: Parse/format XML input type 'evdev'
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-21 14:29:09 +02:00
Kristina Hanicova
3bfbc3c0ef qemu_capabilities: Add QEMU_CAPS_INPUT_LINUX
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-21 14:29:09 +02:00
Pavel Hrdina
93344aed27 storage_file: add support to probe cluster_size from QCOW2 images
From QEMU docs/interop/qcow2.txt :

   Byte  20 - 23:   cluster_bits
                    Number of bits that are used for addressing an offset
                    within a cluster (1 << cluster_bits is the cluster size).

With this patch libvirt will be able to report the current cluster_size
for all existing storage volumes managed by storage driver.

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-21 14:00:55 +02:00
Pavel Hrdina
3e1d2c93a3 storage: add support for QCOW2 cluster_size option
The default value hard-coded in QEMU (64KiB) is not always the ideal.
Having a possibility to set the cluster_size by user may in specific
use-cases improve performance for QCOW2 images.

QEMU internally has some limits, the value has to be between 512B and
2048KiB and must by power of two, except when the image has Extended L2
Entries the minimal value has to be 16KiB.

Since qemu-img ensures the value is correct and the limit is not always
the same libvirt will not duplicate any of these checks as the error
message from qemu-img is good enough:

    Cluster size must be a power of two between 512 and 2048k

Resolves: https://gitlab.com/libvirt/libvirt/-/issues/154

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-21 14:00:43 +02:00
Kristina Hanicova
19967f64f4 qemu: Add check for needed paths for memory devices
When building a commandline for a DIMM memory device with
non-default access mode, the qemuBuildMemoryBackendProps() will
tell QEMU to allocate memory from per-domain memory backing dir.
But later, when preparing the host, the
qemuProcessNeedMemoryBackingPath() does not check for memory
devices at all resulting in per-domain memory backing dir not
being created which upsets QEMU.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1961114

Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-21 08:51:11 +02:00
Ján Tomko
015fe0439f qemu: fs: do not try to fill binary path if we have a socket
We do not need to look for a suitable binary in the vhost-user
description files, if we aren't the ones starting it.
Otherwise startup will fail with:

error: Failed to start domain 'vm1'
error: operation failed: Unable to find a satisfying virtiofsd

https://bugzilla.redhat.com/show_bug.cgi?id=1855789

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 16:27:21 +02:00
Michal Privoznik
7166b1ec7e NEWS: Document new memory dump format
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 16:24:11 +02:00
Michal Privoznik
3de70da32d virsh: Expose new win-dmp dump format
In previous commit the virDomainCoreDumpWithFormat() API gained
new format. Expose it.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 16:24:06 +02:00
Michal Privoznik
c2707c7e3c lib: Add win-dmp crashdump format
QEMU gained support for 'win-dmp' format in it's release of 3.0,
but libvirt doesn't implement it yet. Fortunately, there not much
needed: new value to virDomainCoreDumpFormat public enum, which
unfortunately means that QEMU driver has to be updated in the
same commit, because of VIR_ENUM_IMPL().

Luckily, we don't need any extra QEMU capability - the code
already checks supported formats via
'query-dump-guest-memory-capability' just before issuing
'dump-guest-memory'.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 16:23:32 +02:00
Michal Privoznik
8e75ecefd7 include: Fix copy-paste error in comment to virDomainCoreDumpFormat enum
The comment to virDomainCoreDumpFormat enum says that new values
can be introduced in the future "as new events are added". Well,
it should have been "formats" instead of "events", obviously.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 16:21:03 +02:00
Michal Privoznik
9d4cd1d1cd virNodeDevCapMdevParseXML: Use virXMLPropEnum() for ./start/@type
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 16:08:02 +02:00
Peter Krempa
7eb281f91c conf: node_device: Fix build with clang
Clang complains:

  ../libvirt/src/conf/node_device_conf.c:1945:74: error: result of comparison of unsigned enum expression < 0 is always false [-Werror,-Wtautological-unsigned-enum-zero-compare]
        if ((mdev->start = virNodeDevMdevStartTypeFromString(starttype)) < 0) {

Fixes: 42a5585499
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
2021-05-20 12:17:09 +02:00
Tim Wiederhake
3d5591aede virDomainDeviceSpaprVioAddressParseXML: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`ULLONG_MAX + value + 1`) for attribute `reg`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute, as it
refers to a 32 bit address space.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:46 +02:00
Tim Wiederhake
99037fb2b9 virStorageAdapterParseXML: Use virXMLProp*
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:44 +02:00
Tim Wiederhake
058d4ed47e virStorageAdapterFCHost: Change type of "type" to virStorageAdapterType
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:41 +02:00
Tim Wiederhake
8e2d030e26 virStorageAdapterParseXML: Use g_autofree
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:38 +02:00
Tim Wiederhake
b127e50290 virDomainIOMMUDefParseXML: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `aw_bits`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:35 +02:00
Tim Wiederhake
5ad6db604a virDomainAudioDefParseXML: Don't ignore return value of virDomainAudio*Parse()
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:32 +02:00
Tim Wiederhake
e663717cb1 virDomainAudioDefParseXML: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `id`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:29 +02:00
Tim Wiederhake
e1d589509a virDomainAudioDef: Change type of "sdl.driver" to virDomainAudioSDLDriver
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:27 +02:00
Tim Wiederhake
8f1b810333 virDomainAudioDef: Change type of "type" to virDomainAudioType
Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:23 +02:00
Tim Wiederhake
15670ebb17 virDomainAudioPulseAudioParse: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `latency`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:52:21 +02:00
Peng Liang
667dea5a1e virnetdevopenvswitch: Remove redundant declaration
virNetDevOpenvswitchInterfaceGetMaster is declared twice in
src/util/virnetdevopenvswitch.h.  Remove the last one.

Signed-off-by: Peng Liang <liangpeng10@huawei.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:30:01 +02:00
Jonathon Jongsma
bb8c3b6120 tests: nodedevxml2xmltest: test more mdev files
Add the rest of the mdev xml files to the xml2xml test, and include 2
new test cases: one that explicitly specifies 'manual' start, and one
that explicitly specifies 'auto' start.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:11:44 +02:00
Boris Fiuczynski
42a5585499 nodedev: support auto-start property for mdevs
This adds a new element to the mdev capabilities xml schema that
represents the start policy for a defined mediated device. The actual
auto-start functionality is handled behind the scenes by mdevctl, but it
wasn't yet hooked up in libvirt.

Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com>
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 11:10:08 +02:00
Jonathon Jongsma
86ecc06c06 test: move nodedev xml2xml output to a separate dir
Currently, we're loading and parsing the xml from the input file, and
then formatting it and then comparing it directly back to the input
file. This works for now, but is severely limiting as it relies on the
input file being fully-specified and in the exact order as the output
xml format.

If optional elements are ommitted in the input XML, the output xml
may include default values for the ommitted elements and thus the output
will not match the input.

In order to allow more flexibility in testing, save the expected output
to a seprate 'out' directory similar to what most of the other xml2xml
tests are already doing.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-20 10:59:45 +02:00
Olaf Hering
0ad0204ce7 libxl: remove libxl_domain_config_init from libxlBuildDomainConfig
The passed libxl_domain_config is owned, and already initialized, by the
caller.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
2021-05-19 15:57:47 -06:00
Michal Privoznik
01a2339e1f virsh-domain: Fix @ret handling in cmdSetmem and cmdSetmaxmem
These functions initialize @ret to true and only after something
fails either they call cleanup code (which consists only from
virshDomainFree()) and return false, or they set ret = false and
carry on (when the failure occurred close to cleanup code).

Switch them to the usual pattern in which ret is initialized to
failure, goto cleanup is used and ret is set to true only after
everything succeeded.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-19 15:31:36 +02:00
Michal Privoznik
168190e19c virsh: Fix logic wrt to --current flag in cmdSetmem
In my commit of v7.1.0-rc1~376 I've simplified the logic of
handling @flags. My assumption back then was that calling
virDomainSetMemory() is equivalent to
virDomainSetMemoryFlags(flags = 0). But that is not the case,
because it is equivalent to virDomainSetMemoryFlags(flags =
VIR_DOMAIN_AFFECT_LIVE). Fix the condition that calls the old
API.

Fixes: b5e267e8c5
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1961118
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2021-05-19 15:31:17 +02:00
Jonathon Jongsma
ae02689568 docs: nodedev: document mdev uuid property
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-05-19 10:42:18 +02:00
Shalini Chellathurai Saroja
8a7fb16ada tests: Add capabilities for QEMU 6.0.0 on s390x
Introduce replies and xml files for QEMU 6.0.0 on s390x.

Signed-off-by: Shalini Chellathurai Saroja <shalini@linux.ibm.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2021-05-19 10:40:50 +02:00
Peter Krempa
d53afa1e66 tests: qemucapabilities: Add test-data for the qemu-6.1 cycle
Add test data based on qemu commit v6.0.0-540-g6005ee07c3.

Notable changes are the removal of 'sheepdog' disk storage protocol.

Additionally the cpu model reported when probing seems to have changed
from:

"model-id": "AMD Ryzen 9 3900X 12-Core Processor            "

to:

"model-id": "QEMU TCG CPU version 2.5+"

despite building on the same machine. This probably also results in the
2 test changes in the CPU definition which popped up in this update.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2021-05-19 10:21:32 +02:00
Peter Krempa
985fd74bd3 qemuxml2argvtest: Limit 'disk-network-sheepdog' testcase to qemu-6.0.0
QEMU is dropping sheepdog support in 6.1 so we need to limit the test
case to the latest version supporting sheepdog as it won't be described
by the QMP schema any more.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2021-05-19 10:19:47 +02:00
Peter Krempa
aca69e4420 qemublocktest: Drop 'network-sheepdog-qcow2' image creation test case
QEMU dropped sheepdog support for the 6.1 release. Since we use schema
validation in the image creation it would create test failures.

In this instance we just drop the test altogether as adding versioned
capabilities would be a bit too overkill for this scenario.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2021-05-19 10:19:47 +02:00
Peter Krempa
0dc45670f1 testQemuInfoSetArgs: Strip default machine alias only for 'latest' test cases
For the real-capabilities test cases testing 'latest' capabilities we
strip off the alias from 'pc' to the appropriate versioned machine type
to prevent update to all tests when bumping qemu capabilities.

Recenly we also started caching the capabilities to prevent re-parsing
the XML all the time. The commit adding the caching kept the alias
stripping prior to cache insertion, thus the cache contains the stripped
alias.

This leads to problem when a test case is added where the 'latest'
equals to the selected version.

Move the machine alias stripping after we create a local copy thus
stripping it only for 'latest' tests.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2021-05-19 10:19:47 +02:00
Olaf Hering
5663be9f3a libxl: fix refcounting in libxlDomainChangeEjectableMedia
The initial variant of libxlDomainChangeEjectableMedia could just leave
the function earlier. With refcounting this does not work anymore.

Fixes commit a5bf06ba34

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
2021-05-18 15:35:19 -06:00
Tim Wiederhake
c6b0b7ad1a virNodeDevCapPCIDevIommuGroupParseXML: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `number`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
2021-05-18 12:27:14 -04:00
Tim Wiederhake
cd4c756fd5 virDomainAudioOSSParse: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `bufferCount`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
2021-05-18 12:27:14 -04:00
Tim Wiederhake
23e763eb46 virDomainAudioCoreAudioParse: Use virXMLProp*
This strictens the parser to disallow negative values (interpreted as
`UINT_MAX + value + 1`) for attribute `bufferCount`. Allowing negative
numbers to be interpreted this way makes no sense for this attribute.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
2021-05-18 12:27:13 -04:00