Prepare the full backing chain by instantiating authentication and TLS
transport secrets and other necessary objects so that we can add the
full backing chain explicitly to qemu. This also includes allocation of
nodenames for the individual backing chain members.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The copy-on-read feature is expressed by adding a new node layer in
qemu when using -blockdev. Since we will keep these per-disk (as opposed
to per storage source) we need to store the appropriate node names in
the disk definition.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

To allow referring to the drives via the QOM id we need to setup the
floppy drives with a proper ID. This means that -device should be used
for them.

There are the following quirks:
- FDC needs to be instantiated prior to any floppy device
- floppy drive specified via -device does not support 'bootindex'
    (hacked around by passing bootindexA=1 to the FDC)
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

When we stop using -drive qemu stops reporting it in some of the monitor
commands. To allow referring the disk frontends and the corresponding
block backends we need to know these names. Unfortunately different
buses require different names.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

When using -blockdev you need to use the qom path to refer to the disk
fronends. Add means for storing the path and getting it after restart.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Use the index stored in virStorageSource struct rather than
recalculating it. Currently we'd report proper numbers but that will
change with blockdev.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Node names for block objects in qemu need to be unique for an instance
of the qemu process. Add a counter to generate objects sequentially and
store it in the status XML so that we can restore it.

The helpers added allow to create new node names and reset the counter
after the VM process terminates.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

We'll specify them ourselves so it's pointless to attempt to redetect
them.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

We need to load the backing chain from the XML when using -blockdev.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

SD cards are currently passed by using -drive only which would not be
compatible with using -blockdev fully.

Clear QEMU_CAPS_BLOCKDEV if the VM has such devices.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The capability currently is not enabled so that we can add individual
bits first.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

It should be impossible to lack an alias in the domain definition. Other
disk types don't generate it so remove it here as well.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Allow referring to individual node name to resize.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The 'device' field reported by 'query-block' is empty when -blockdev is
used. Add an argument which will allow matching disk by using the qdev
id so we can use this code with -blockdev.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The 'device' argument matches only the legacy drive alias. For blockdev
we need to set the throttling for a QOM id and thus we'll need to use
the 'id' field.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

The wrapper executes the command and does error detection so there's no
need to open-code all of those things.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Move the preparation steps from qemuDomainAttachDiskGeneric up into
qemuDomainAttachDeviceDiskLive so that also media changing can use the
prepared file.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Use qemuDomainAttachDeviceDiskLive to change the media in
qemuDomainChangeDiskLive as the former function already does all the
necessary steps to prepare the new medium.

This also allows us to turn qemuDomainChangeEjectableMedia static.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Turns out that 'query-nodes' is not what we want and the
'query-blockstats' command was in fact buggy. Revert the new field since
it's not needed.

This reverts commit 50edca13.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

We don't use it for anything useful so it does not make much sense to
extract it.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

QEMU supports 'block_resize' since 0.14 so we don't need to do explicit
checking. Additionally the caller did not use the different value at
all.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Remove the pointless "empty path" check and use a better error message
if the disk was not found.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Currently we'd report the alias of the drive which is backing the cdrom
rather than the device itself:

 $ virsh event ds tray-change --loop
 event 'tray-change' for domain ds disk drive-ide0-0-1: opened
 event 'tray-change' for domain ds disk drive-ide0-0-1: closed

Report the disk device alias as we document in the API docs:

https://libvirt.org/html/libvirt-libvirt-domain.html#virConnectDomainEventTrayChangeCallbackSigned-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Since we're not saving the platform-specific data into a cache, we're
not going to populate the structure, which in turn will cause a crash
upon calling virNodeGetSEVInfo because of a NULL pointer dereference.
Ultimately, we should start caching this data along with host-specific
capabilities like NUMA and SELinux stuff into a separate cache, but for
the time being, this is a semi-proper fix for a potential crash.

Backtrace (requires libvirtd restart to load qemu caps from cache):
    #0 qemuGetSEVInfoToParams
    #1 qemuNodeGetSEVInfo
    #2 virNodeGetSEVInfo
    #3 remoteDispatchNodeGetSevInfo
    #4 remoteDispatchNodeGetSevInfoHelper
    #5 virNetServerProgramDispatchCall
    #6 virNetServerProgramDispatch
    #7 virNetServerProcessMsg
    #8 virNetServerHandleJob
    #9 virThreadPoolWorker
    #10 virThreadHelper

https: //bugzilla.redhat.com/show_bug.cgi?id=1612009
Signed-off-by: NErik Skultety <eskultet@redhat.com>
Acked-by: NPeter Krempa <pkrempa@redhat.com>
Tested-by: NBrijesh Singh <brijesh.singh@amd.com>

So the procedure to detect SEV support works like this:
1) we detect that sev-guest is among the QOM types and set the cap flag
2) we probe the monitor for SEV support
    - this is tricky, because QEMU with compiled SEV support will always
    report -object sev-guest and query-sev-capabilities command, that
    however doesn't mean SEV is supported
3) depending on what the monitor returned, we either keep or clear the
capability flag for SEV

Commit a349c6c2 added an explicit check for "GenericError" in the
monitor reply to prevent libvirtd to spam logs about missing
'query-sev-capabilities' command. At the same time though, it returned
success in this case which means that we didn't clear the capability
flag afterwards and happily formatted SEV into qemuCaps. Therefore,
adjust all the relevant callers to handle -1 on errors, 0 on SEV being
unsupported and 1 on SEV being supported.
Signed-off-by: NErik Skultety <eskultet@redhat.com>
Acked-by: NPeter Krempa <pkrempa@redhat.com>

Keep with the recent effort of replacing as many explicit *Free
functions with their automatic equivalents.
Signed-off-by: NErik Skultety <eskultet@redhat.com>
Acked-by: NPeter Krempa <pkrempa@redhat.com>

commit 5c81c342 forgot to skip the detaching of the shmem backend
when async unplug is requested which meant that we've tried to unplug
the backend prior to delivery of the DEVICE_DELETED event.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1618622Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

Qemu-3.0 supports Hyper-V-style PV TLB flush, Windows guests can benefit
from this feature as KVM knows which vCPUs are not currently scheduled (and
thus don't require any immediate action).
Signed-off-by: NVitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Qemu-3.0 supports so-called 'Reenlightenment' notifications and this (in
conjunction with 'hv-frequencies') can be used make Hyper-V on KVM pass
stable TSC page clocksource to L2 guests.
Signed-off-by: NVitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Qemu-2.12 gained 'hv-frequencies' cpu flag to enable Hyper-V frequency
MSRs. These MSRs are required (but not sufficient) to make Hyper-V on
KVM pass stable TSC page clocksource to L2 guests.
Signed-off-by: NVitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

There are some path where the buffer is not passed to
virCommandAddArgBuffer and therefore the buffer might leak.

==191201== 1,010 bytes in 1 blocks are definitely lost in loss record 826 of 836
==191201==    at 0x4C2CE3F: malloc (vg_replace_malloc.c:298)
==191201==    by 0x4C2F1BF: realloc (vg_replace_malloc.c:785)
==191201==    by 0x5D39E82: virReallocN (viralloc.c:245)
==191201==    by 0x5D3E8F2: virBufferGrow (virbuffer.c:150)
==191201==    by 0x5D3E9C8: virBufferAdd (virbuffer.c:185)
==191201==    by 0x56EAC98: qemuBuildFloppyCommandLineControllerOptions (qemu_command.c:2162)
==191201==    by 0x56EB3E1: qemuBuildDisksCommandLine (qemu_command.c:2370)
==191201==    by 0x570055E: qemuBuildCommandLine (qemu_command.c:10315)
==191201==    by 0x575EA7F: qemuProcessCreatePretendCmd (qemu_process.c:6777)
==191201==    by 0x113DAB: testCompareXMLToArgv (qemuxml2argvtest.c:598)
==191201==    by 0x13A75B: virTestRun (testutils.c:180)
==191201==    by 0x138BE8: mymain (qemuxml2argvtest.c:2975)
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

Resctrl not only supports cache tuning, but also memory bandwidth
tuning. Renaming cachetune to resctrl to reflect that. With resctrl,
all allocation for different resources (cache, memory bandwidth) are
aggregated and represented by a virResctrlAllocPtr inside
virDomainResctrlDef.
Signed-off-by: NBing Niu <bing.niu@intel.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

If a domain has hugepages configured and we're currently building
memory-backend-file for a nvdimm device that domain has we will
put hugepages path onto the command line. It should have been
nvdimm path configured in the XML.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

This reverts commit 9cf38263.

Jansson cannot parse QEMU's quirky JSON.
Revert back to yajl.

https://bugzilla.redhat.com/show_bug.cgi?id=1614569Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

When a domain is killed on the source host while it is being migrated
and libvirtd is waiting for the migration to finish (waiting for the
domain condition in qemuMigrationSrcWaitForCompletion), the run-time
state including priv->job.current may already be freed once
virDomainObjWait returns with -1. Thus the priv->job.current pointer
cached in jobInfo is no longer valid and setting jobInfo->status may
crash the daemon.

https://bugzilla.redhat.com/show_bug.cgi?id=1593137Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

We can safely validate the hugepage nodeset attribute at a define time.
This validation is not done for already existing domains when the daemon
is restarted.

All the changes to the tests are necessary because we move the error
from domain start into XML parse.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

After commit caccbba6 this variable is always NULL.
Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>

Starting from pc-q35-2.4 the floppy controller is not enabled by
default. Fix the version check so that it does not match 2.11 as being
2.1.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Fix regression introduced in <42fd5a58>. With q35 machine type which
requires the explicitly specified FDC we'd format twoisa-fdc
controllers to the command line as the code was moved to a place where
it's called per-disk.

Move the call back after formatting all disks and reiterate the disks to
find the floppy controllers.

This also moves the '-global' directive which sets up the default
ISA-FDC to the end after all the disks but since we are modifying the
properties it is safe to do so.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Now that the argument is unused we can remove it transitively from all
the call graphs.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>