Introduce a function for comparing two vsock definitions.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Allow hotplugging the vsock device.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Alter qemuBuildVsockDevStr to allow passing a prefix for
the vhostfd file descriptor name. Domain startup uses
the numeric value of fd without a prefix, but hotplug
will need to use a prefix because passed file descriptor
names cannot start with a number.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Split out the device string building to allow reuse for hotplug.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Commit 656151bf fixed formatting of the <cmdline> element. Perhaps it
would have been noticed and fixed earlier if we had a test. With this
change, all possible cases of formatting <cmdline> from xmconfig are
covered

1. no 'extra=' or 'root=' in xm.cfg
2. 'extra=' but no 'root=' in xm.cfg
3. 'root=' but no 'extra=' in xm.cfg
4. both 'root=' and 'extra=' in xm.cfg

Case 1 is covered by all existing paravirt tests since they have no
'extra=' or 'root='. Case 2 is covered by adding 'extra=' to a few
of the existing paravirt tests. Cases 3 and 4 are covered by new
tests that only test conversion of xm.cfg to xml.
Signed-off-by: NJim Fehlig <jfehlig@suse.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Per-domain log files were introduced in commit a30b08b7. The FILE
objects associated with these log files are stored in a hash table
using domid as a key. When a domain is shutdown, destroyed, or
otherwise powered-off, the FILE object is removed from the hash table,
where the free function will close the FILE.

Unfortunately the call to remove the FILE from the hash table occurs
after setting domid=-1 in the libxlDomainCleanup() function. The
object is never removed from the hash table, the free function is
never called, and the underlying fd is leaked. Fix by removing the
FILE object from the hash table before setting domid=-1.
Signed-off-by: NJim Fehlig <jfehlig@suse.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

* docs/news.xml: updated for release
Signed-off-by: NDaniel Veillard <veillard@redhat.com>

My commit b8b42ca0 added support for formatting the vsock
command line without actually checking if it's supported.

Add it to the per-device validation function.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Reported-by: NJohn Ferlan <jferlan@redhat.com>
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>

It's "a hard_limit", not "an hard_limit".  Probably that was just a typo.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

This has been broken since commit v4.0.0-165-g93412bb8 which added
jobInfo->statsType enum to distinguish various statistics types. During
migration the type will always be QEMU_DOMAIN_JOB_STATS_TYPE_MIGRATION,
however the destination code consuming the statistics data from
migration cookie failed to properly set the type. So even though
everything was filled in, the type remained *_NONE and any attempt to
fetch the statistics data of a completed migration on the destination
host failed.

https://bugzilla.redhat.com/show_bug.cgi?id=1584071Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

At least since Fedora 26 (maybe earlier, but we don't support older
Fedora releases), the "tc" tool is provided by a separate iproute-tc
package.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

Document the addition of vsock.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Acked-by: NPeter Krempa <pkrempa@redhat.com>

To avoid the <source> vs. <target> confusion,
change <source auto='no' cid='3'/> to:
<cid auto='no' address='3'/>
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Suggested-by: NDaniel P. Berrangé <berrange@redhat.com>
Acked-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

The first feature is SCSI persistent reservation, the other is
support for multihead screenshots.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

In commit 8bebb2b7 I've refactored how the detach of disk with a
managed persistent reservations object is handled. After the commit if
any disk with a managed PR object would be removed libvirt would also
attempt to remove the shared 'pr-manager-helper' object potentially used
by other disks.

Thankfully this should not have practical impact as qemu should reject
deletion of the object if it was still used and the rest of the code is
correct.

Fix this by removing the disk from the definition earlier and checking
if the shared/managed pr-manager-helper object is still needed.

This basically splits the detach code for the managed PR object from the
unmanaged ones. The same separation will follow for the attachment code
as well as it greatly simplifies -blockdev support for this.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

When preparing qemuCaps for test cases the following is
happening:

qemuTestParseCapabilitiesArch() is called, which calls
virQEMUCapsLoadCache() which in turn calls
virQEMUCapsInitHostCPUModel() which sets qemuCaps->kvmCPU and
qemuCaps->tcgCPU.

But then the code tries to update the capabilities:

testCompareXMLToArgv() calls testUpdateQEMUCaps() which calls
virQEMUCapsInitHostCPUModel() again overwriting previously
allocated memory. The solution is to free host cpuData in
testUpdateQEMUCaps().
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>

This patch fixes the USB input bus check, the bug was introduced by commit 317badb2Signed-off-by: NXiao Feng Ren <renxiaof@linux.vnet.ibm.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>
Signed-off-by: NJán Tomko <jtomko@redhat.com>

There was a missing enum for mdev causing a strange 'unknown device type'
warning when hot-plugging mdev.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1583927Signed-off-by: NErik Skultety <eskultet@redhat.com>

There's no point in calling testInitQEMUCaps() (which sets
info.qemuCaps) only to overwrite (and leak) it on the very next
line.

==12962== 296 (208 direct, 88 indirect) bytes in 1 blocks are definitely lost in loss record 265 of 331
==12962==    at 0x4C2CF26: calloc (vg_replace_malloc.c:711)
==12962==    by 0x5D28D9F: virAllocVar (viralloc.c:560)
==12962==    by 0x5D96AB4: virObjectNew (virobject.c:239)
==12962==    by 0x56DB7C7: virQEMUCapsNew (qemu_capabilities.c:1480)
==12962==    by 0x112A5B: testInitQEMUCaps (qemuxml2argvtest.c:361)
==12962==    by 0x1371C8: mymain (qemuxml2argvtest.c:2871)
==12962==    by 0x13AD0B: virTestMain (testutils.c:1120)
==12962==    by 0x1372FD: main (qemuxml2argvtest.c:2883)

Removing the function call renders @gic argument unused therefore
it's removed from the macro (and all its callers).
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

We need to free return value of virXPathString().

==12962== 37 bytes in 1 blocks are definitely lost in loss record 156 of 331
==12962==    at 0x4C2AF0F: malloc (vg_replace_malloc.c:299)
==12962==    by 0x91E8439: strdup (in /lib64/libc-2.25.so)
==12962==    by 0x5DBD551: virStrdup (virstring.c:977)
==12962==    by 0x5DD3E5E: virXPathString (virxml.c:84)
==12962==    by 0x5E178AB: virDomainDefParseXML (domain_conf.c:19110)
==12962==    by 0x5E1E985: virDomainDefParseNode (domain_conf.c:20885)
==12962==    by 0x5E1E7CB: virDomainDefParse (domain_conf.c:20827)
==12962==    by 0x5E1E871: virDomainDefParseFile (domain_conf.c:20853)
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Introduced by:
commit d4abb7b4
    conf: introduce <vsock> element
commit b8b42ca0
    qemu: add support for vhost-vsock-pci
Signed-off-by: NJán Tomko <jtomko@redhat.com>

We allocate a temporary bitmap but never free it.

Introduced by <commit 1f8a1a95>:
    qemu: Do not use qemuMonitorSetMigrationCapability
Signed-off-by: NJán Tomko <jtomko@redhat.com>

Found by cppcheck:
[tests/metadatatest.c:284]: (error) Uninitialized variable: test
[tests/objecteventtest.c:855]: (error) Uninitialized variable: test
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NJiri Denemark <jdenemar@redhat.com>

The nwfilter XMLs in /etc are defined as %ghost in the spec file, which
means rpm will not install them, but it will record its existence and
permissions in the database. During installation the files are copied in
a %post scriptlet from /usr/share/libvirt/nwfilter, but once libvirtd is
restarted, it will rewrite the files to add generated UUIDs.

While RPM recorded 644 mode for the XMLs, libvirt saves them with 600
and thus any future attempt to verify the libvirt-daemon-config-nwfilter
package would fail. We need to tell RPM the ghost files are supposed to
have 600 permissions.

https://bugzilla.redhat.com/show_bug.cgi?id=1559284Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1560946

Similar to the the Logical backend, use qemu-img on the created
disk partition device to set up for LUKS encryption. Secret mgmt
for the device can be complicated by a reboot possibly changing
the path to the device if the infrastructure changes.
Signed-off-by: NJohn Ferlan <jferlan@redhat.com>
ACKed-by: NPeter Krempa <pkrempa@redhat.com>

Create a new vsock endpoint by opening /dev/vhost-vsock,
set the requested CID via ioctl (or assign a free one if auto='yes'),
pass the file descriptor to QEMU and build the command line.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

A file for vsock-related helper functions.
virVsockSetGuestCid to set an already-known CID,
virVsockAcquireGuestCid that will use the first available CID

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

Add a new capability flag for vhost-vsock-device

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

Introduce a structure and a class that will be used to store
the private data.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

Add a new 'vsock' element for the vsock device.
The 'model' attribute is optional.
A <source cid> subelement should be used to specify the guest cid,
or <source auto='yes'/> should be used.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

An object for storing driver-specific data in the vsock definition.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

A type to represent the new vsock device.
Also implement an allocation function to allow future addition
of private data.

https://bugzilla.redhat.com/show_bug.cgi?id=1291851Signed-off-by: NJán Tomko <jtomko@redhat.com>

Remove the locks since they are unnecessary and would cause
a hang for a driver reload/restart when a transient pool was
previously active as a result of the call:

virStoragePoolUpdateInactive:
...
    if (!virStoragePoolObjGetConfigFile(obj)) {
        virStoragePoolObjRemove(driver->pools, obj);
...

stack trace:

Thread 17 (Thread 0x7fffcc574700 (LWP 12465)):
...pthread_rwlock_wrlock
...virRWLockWrite
...virObjectRWLockWrite
...virStoragePoolObjRemove
...virStoragePoolUpdateInactive
...storagePoolUpdateStateCallback
...virStoragePoolObjListForEachCb
...virHashForEach
...virStoragePoolObjListForEach
...storagePoolUpdateAllState
...storageStateInitialize
...virStateInitialize
...daemonRunStateInit
...virThreadHelper
...start_thread
...clone

Introduced by commit id 4b2e0ed6.
Signed-off-by: NJohn Ferlan <jferlan@redhat.com>
ACKed-by: NMichal Privoznik <mprivozn@redhat.com>

When status XML was parsed the post-parse callbacks could not access
qemu caps and potentially upgrade the definition according to the
present caps. Implement the callback to pass it in.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

The status XML parser function virDomainObjParseXML could not pass in
parseOpaque into the post parse callbacks. Add a callback which will
allow hypervisor drivers to fill it from the 'virDomainObj' data.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

When parsing status XML the post-parse callbacks can't access any
private data present in the status XML as the private bits were parsed
after invoking post-parse callbacks.

Move the invocation so that everything is parsed first.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

Its only use is now to check for duplicate boot order values,
which is now also done in virDomainDefPostParseCommon.

Remove it completely.
Signed-off-by: NJán Tomko <jtomko@redhat.com>

As the function signature of virDomainDefPostParseInternal does not
differ from virDomainDefPostParse now, the wrapper can be dropped.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Signed-off-by: NJán Tomko <jtomko@redhat.com>

Move the check for boot elements into a separate function
and remove its dependency on the parser-supplied bootHash table.

Reconstructing the hash table from the domain definition
effectively duplicates the check for duplicate boot order
values, also present in virDomainDeviceBootParseXML.

Now it will also be run on domains created by other means than XML
parsing, since it will be run even for code paths that did not supply
the bootHash table before.
Signed-off-by: NJán Tomko <jtomko@redhat.com>