Commit Graph

42826 Commits

Author SHA1 Message Date
Peter Krempa
795e9e05c3 virPCIGetVirtualFunctions: Fetch also network interface name if needed
'virNetDevGetVirtualFunctions' calls 'virPCIGetVirtualFunctions' and
then re-iterates the returned list to fetch the interface names for the
returned virtual functions.

If we move the fetching of the interface name into
virPCIGetVirtualFunctions we can simplify the code and remove a bunch of
impossible error states.

To accomplish this the function is renamed to
'virPCIGetVirtualFunctionsFull' while keeping a wrapper with original
name and if the physical port ID is passed the interface name is fetched
too without the need to re-convert the address into a sysfs link.

For now 'virNetDevGetVirtualFunctions' still converts the returned data
into two lists.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-09 10:09:00 +02:00
Peter Krempa
7ef618f30a virPCIGetNetName: Make 'physPortID' argument const
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-09 10:09:00 +02:00
Peter Krempa
9766d0bac7 virPCIGetVirtualFunctionIndex: Refactor cleanup
The 'ret' variable and 'out' label can be removed.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-09 10:09:00 +02:00
Peter Krempa
c97518d9b8 virPCIGetVirtualFunctions: Simplify cleanup of returned data
Introduce a struct for holding the list of VFs returned by
virPCIGetVirtualFunctions so that we can employ automatic memory
clearing and also allow querying more information at once.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-09 10:09:00 +02:00
Peter Krempa
00f804c1cf virNetDevGetVirtualFunctions: Remove 'max_vfs' argument
The only caller doesn't use it.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-09 10:09:00 +02:00
Jonathon Jongsma
fcae7a44c3 nodedev: look up mdevs by UUID and parent
Since UUID is not guaranteed to be unique by mdevctl, we may have more
than one nodedev with the same UUID. Therefore, we need to disambiguate
when looking up mdevs by specifying the UUID and parent address, which
mdevctl guarantees to be a unique combination.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:41 -05:00
Jonathon Jongsma
3bd8181bc5 nodedev: Handle inactive mdevs with the same UUID
Unfortunately, mdevctl supports defining more than one mdev with the
same UUID as long as they have different parent devices. (Only one of
these devices can be active at any given time).

This means that we can't use the UUID alone as a way to uniquely
identify mdev node devices. Append the parent address to ensure
uniqueness. For example:

    Before: mdev_88a6b868_46bd_4015_8e5b_26107f82da38
    After:  mdev_88a6b868_46bd_4015_8e5b_26107f82da38_0000_00_02_0

Related: https://bugzilla.redhat.com/show_bug.cgi?id=1979440

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:38 -05:00
Jonathon Jongsma
08d29eb3b1 nodedev: add PostParse callback for nodedev parsing
This can be used similarly to other postparse callbacks in libvirt --
filling in additional information that can be determined by using the
information provided in the XML. In this case, we determine the address
of the parent device and cache it in the mdev caps so that we can use it
for generating a unique name and interacting with mdevctl.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:36 -05:00
Jonathon Jongsma
d5ae634ba2 nodedev: Add parser validation for node devices
At the moment, this is only for mediated devices. When a new mediated
device is created or defined, the xml is expected specify the nodedev
name of an existing device as its parent. We were not previously
validating this and were simply accepting any string here.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:32 -05:00
Jonathon Jongsma
e44408dd9c nodedev: cache parent address in mdev caps
mdevctl can report multiple defined devices with the same UUID
but different parents, including parents that don't actually exist on
the host machine. Libvirt sets the parent to the 'computer' device for
all of the mdevs that have nonexistent parents. Because of this, it's
possible that there are multiple devices with the same UUID and the same
'computer' device as their parent, so the combination of uuid and parent
nodedev name is not guaranteed to be a unique name.

We need to ensure that each nodedev has a unique name. If we can't use
the UUID as a unique nodedev name, and we can't use the combination of
UUID and nodedev parent name, we need to find another solution. By
caching and using the parent name reported by mdevctl in combination
with the UUID, we can achieve a unique name. mdevctl guarantees that its
uuid/parent combination is unique.

This value will be used to set the mdev nodedev name in a following commit.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:29 -05:00
Jonathon Jongsma
a9c1febcf4 nodedev: fix xml output for mdev parents in test suite
Commit 51fbbfdce8 attempted to get the proper nodedev name for the
parent of an defined mdev by traversing the filesystem and looking for a
device that had the appropriate sysfs path.  This works, but it would be
cleaner to to avoid mucking around in the filesystem and instead just
just examine the list of devices we have in memory.

We already had a function nodeDeviceFindAddressByName() which constructs
an address for parent device in a format that can be used with mdevctl.
So if we refactor this function into a a function that simply formats an
address for an arbitrary virNodeDeviceObj*, then we can use this
function as a predicate for our new virNodeDeviceObjListFind() function
from the previous commit. This will search our list of devices for one
whose address matches the address we get from mdevctl.

One nice benefit of this approach is that our test cases will now
display xml output with the proper parent name for mdevs (assuming that
we've added the appropriate mock parent devices to the test driver).
Previously they just displayed 'computer' for the parent because the
alternative would have required specially constructing a mock filesystem
environment with a sysfs that mapped to the appropriate parent.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:02:23 -05:00
Jonathon Jongsma
fdfe4b2837 nodedev: add internal virNodeDeviceObjListFind()
This is a generic function that you can provide your own predicate
function to search for a particular device. It will be used in an
upcoming commit.

Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 15:01:52 -05:00
Daniel P. Berrangé
1280a631ef src: stop checking virIdentityNew return value
This method will always succeed.

Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-06 12:11:58 +01:00
Daniel P. Berrangé
821414f0cd rpm: add conditionals around post scripts
The hypervisor drivers can be disabled in certain build scenarios, so
their corresponding post scripts need to match.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-06 12:00:53 +01:00
Daniel P. Berrangé
5df2c49263 util: directly query KVM for TSC scaling support
We currently query the host MSRs to determine if TSC scaling is
supported. This works OK when running privileged and can open
the /dev/cpu/0/msr. When unprivileged we fallback to querying
MSRs from /dev/kvm. This is incorrect because /dev/kvm only
reports accurate info for MSRs that are valid to use from inside
a guest.  The TSC scaling support MSR is not, thus we always end
up reporting lack of TSC scaling when unprivileged.

The solution to this is easy, because KVM can directly report
whether TSC scaling is available, which matches what QEMU will
do at startup.

Closes: https://gitlab.com/libvirt/libvirt/-/issues/188
Reported-by: Roman Mohr <rmohr@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-06 12:00:53 +01:00
Daniel P. Berrangé
a29af66ffe Revert "remote: remove probing logic from virtproxyd dispatcher"
This reverts commit 05bd8db60b.

It is true that the remote driver client now contains logic for probing
the driver to connect to when using modular daemons. This logic, however,
only runs when the remote driver is NOT running inside a daemon since we
don't want it activated inside libvirtd. Since the same remote driver
build is used in all daemons, we can't rely on it in virtproxyd either.
Thus we need to keep the virtproxyd probing logic

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-06 12:00:53 +01:00
Peter Krempa
51733511d1 virStorageBackendLogicalParseVolExtents: Remove 'cleanup' and 'ret'
The function was inconsistently using 'return -1' and 'goto cleanup;'
unify it by removing the cleanup label and 'ret' variable.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
e03e54c9a2 virStorageBackendLogicalParseVolExtents: Move 'extents' inside the loop
It's used only inside the loop filling the extents, move it there and
restructure the code so that 'extent.path' doesn't have to be freed.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
bbd89d7894 virStorageBackendLogicalParseVolExtents: Declare one variable per line
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
bd83527521 lxcNetworkParseDataIPs: Automatically free string list
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
30b6be3f8c virNetServerGetClients: Remove pointless cleanup
'list' will always be NULL when reaching 'virObjectListFreeCount' thus
we can remove the call as well as the 'ret' variable which was only ever
equal to 'nclients' at the point when we returned the value.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
a257668ede qemuProcessSetupHotpluggableVcpus: Use automatic memory freeing
'bootHotplug' can be auto-freed when terminating the function and moving
the declaration of 'vcpuprops' to the loop which uses it along with
automatic freeing allows us to simplify cleanup in certain cases.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
0b6888451f virNWFilterIncludeDefToRuleInst: Refactor cleanup
Use automatic memory freeing for 'tmpvars' and move the allocation of
tmpvars earlier so that we are guaranteed that 'obj' will always be
appended to 'inst->filters' and thus don't need cleanup for it.

By moving the reset of 'inst' to the block when virNWFilterDefToInst
fails we can get rid of the rest of the cleanup section and remove the
'ret' variable.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
abd045030e virNWFilterRuleDefToRuleInst: Restructure code to avoid cleanup
Construct the 'ruleinst->vars' hash table separately in a temporary
variable so that 'ruleinst' can be allocated on success. This allows us
to get rid of the cleanup section.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
8d4468d007 virObjectEventCallbackListAddID: Remove pointless cleanup of 'cb'
'cb' is always NULL when 'virObjectEventCallbackListAddID' is called.
Remove the call.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
0b450b0ad1 virNWFilterRuleDefToRuleInst: Remove pointless assignment
'ruleinst' is NULLed by VIR_APPEND_ELEMENT

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
aa76e5d0a4 qemuDomainAttachDeviceConfig: Remove pointless assignment
'dev->data.rng' is NULLed by VIR_APPEND_ELEMENT

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
4e7576a741 virSecuritySELinuxContextListAppend: Remove unreachable cleanup
'item' is always NULLed-out by VIR_APPEND_ELEMENT and 'ret' variable is
always 0 when used so both can be removed.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
966ba852b7 qemuDomainUSBAddressAddHubs: Refactor cleanup
'hub' doesn't need to be freed any more because it's always consumed and
NULLed-out by VIR_APPEND element. This also makes the 'ret' variable
obsolete.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
18e65eef04 xenParseXMDisk: Use automatic memory clearing and remove 'ret' variable
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
a31864e60a prlsdkAddDomainVideoInfoCt: Remove pointless cleanup section
'video' will only ever be NULL after the 'cleanup' label thus there's no
need to use 'virDomainVideoDefFree'. In fact we can fully remove the
cleanup section and 'ret' variable by returning directly from failure
points.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:26 +02:00
Peter Krempa
f90de01763 util: alloc: Completely replace VIR_APPEND_ELEMENT_QUIET by VIR_APPEND_ELEMENT
VIR_APPEND_ELEMENT doesn't report any errors now so we can remove
VIR_APPEND_ELEMENT_QUIET and replace all uses by VIR_APPEND_ELEMENT

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
98f6f2081d util: alloc: Reimplement VIR_APPEND_ELEMENT using virAppendElement
Use virAppendElement instead of virInsertElementsN to implement
VIR_APPEND_ELEMENT which allows us to remove error handling as the
only relevant errors were removed when switching to aborting memory
allocation functions.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
8c22b28dbe util: alloc: Reimplement VIR_APPEND_ELEMENT_QUIET using virAppendElement
For now it was an alias to VIR_APPEND_ELEMENT. Use virAppendElement
directly until VIR_APPEND_ELEMENT is refactored too and we'll be able to
get rid of VIR_APPEND_ELEMENT_QUIET completely.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
592517636f util: alloc: Reimplement VIR_APPEND_ELEMENT_COPY using virAppendElement
Use virAppendElement instead of virInsertElementsN to implement
VIR_APPEND_ELEMENT_COPY which allows us to remove error handling as the
only relevant errors were removed when switching to aborting memory
allocation functions.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
0215139bc6 util: alloc: Reimplement VIR_APPEND_ELEMENT_(COPY_)INPLACE using virAppendElement
VIR_APPEND_ELEMENT_INPLACE and VIR_APPEND_ELEMENT_COPY_INPLACE already
ignore the return value from 'virInsertElementsN' which allows a trivial
conversion to virAppendElement without the need for 'ignore_value'.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
51c8245847 util: alloc: Introduce virAppendElement helper
The new wrapper calls virInsertElementInternal with the appropriate
arguments without any checks which are unnecessary for appension. This
allows to have no return value.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
853fb577d8 virInsertElementsN: Split out actual insertion code
Split out the code doing the movement of the elements and insertion from
the range checks. This will help in adding an error-free version for
appension.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Peter Krempa
05b69f3ec4 virInsertElementsN: Rename 'add' argument
The idea of @add was that the insersion/appension macros would allow
adding more than one element but this feature was never implemented.

'add' is nowadays used as a dummy variable consuming the result of the
VIR_TYPEMATCH compile time check.

Make it obvious that we don't use 'add' by renaming it to
'typematchDummy', marking it as unused and replacing all occurences
where the value was used by literal '1'.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-06 08:53:25 +02:00
Michal Privoznik
73890139bf tests: Don't leak cpu defs
There are cpu definitions that are allocated in
qemuTestDriverInit() but are missing corresponding
virCPUDefFree() call in qemuTestDriverFree(). It's safe to call
the free function because the definitions contain a refcounter
and thus even if they were still in use the refcounter would be
just decreased.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2021-08-05 14:53:19 +02:00
Ján Tomko
ca4f9e403f util: fix typo
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-05 11:18:13 +02:00
Ján Tomko
ca43a5198a qemu: virtiofs: kill the whole process group
Send TERM/KILL to virtiofsd and its child processes too
and do not exit until they are all dead.

https://bugzilla.redhat.com/show_bug.cgi?id=1940276

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Tested-by: Masayoshi Mizuma <m.mizuma@jp.fujitsu.com>
2021-08-05 11:18:13 +02:00
Ján Tomko
ff7b8043b6 util: virPidFileForceCleanupPath: add group argument
Add a version of virPidFileForceCleanupPath that takes
a 'group' bool argument and propagate it all the way
down to virProcessKillPainfullyDelay.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-05 11:18:09 +02:00
Ján Tomko
4b39c2aa2e util: introduce virProcessGroupGet
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-05 11:18:05 +02:00
Ján Tomko
fc180c741f util: Introduce virProcessGroupKill
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2021-08-05 11:18:00 +02:00
Daniel P. Berrangé
5998497d5f rpm: use direct remote connection for Fedora >= 35 / RHEL >= 9
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-04 16:06:11 +01:00
Daniel P. Berrangé
50eae3f885 rpm: handle enabling/disabling modular daemons post/postun-install
We need to enable or disable the modular daemons with systemd after the
RPM install/uninstall.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-04 15:59:15 +01:00
Daniel P. Berrangé
a97c2b14b2 rpm: macroize logic for enabling/disabling daemons post/postun-install
The patterns for enabling/disabling daemons post/postun-install has a
bit of duplication across the different part of the spec, due to the
number of socket units involved. This is going to get much worse with
the need to enable/disalbe modular daemons, so benefits from macroization.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-04 15:59:11 +01:00
Daniel P. Berrangé
b8ef625544 rpm: restart modular daemons on upgrade
The daemons all need restarting to ensure they pick up the newly
installed code.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-04 15:58:28 +01:00
Daniel P. Berrangé
c2eafaf7a3 rpm: restart virtnwfilter/virnetworkd if configs change
Currently we restart libvirtd if the nwfilter/network configs have
changed. We need to take account of possibility that the modular
daemons are in use instead though.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2021-08-04 15:57:45 +01:00