Introduces two new -machine option parameters to the QEMU command to
enable/disable the CPACF protected key management operations for a guest:

    aes-key-wrap='on|off'
    dea-key-wrap='on|off'

The QEMU code maps the corresponding domain configuration elements to the
QEMU -machine option parameters to create the QEMU command:

    <cipher name='aes' state='on'>   --> aes-key-wrap=on
    <cipher name='aes' state='off'>  --> aes-key-wrap=off
    <cipher name='dea' state='on'>   --> dea-key-wrap=on
    <cipher name='dea' state='off'>  --> dea-key-wrap=off
Signed-off-by: NTony Krowiak <akrowiak@linux.vnet.ibm.com>
Signed-off-by: NDaniel Hansel <daniel.hansel@linux.vnet.ibm.com>
Signed-off-by: NBoris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: NBoris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Two new domain configuration XML elements are added to enable/disable
the protected key management operations for a guest:

    <domain>
      ...
      <keywrap>
        <cipher name='aes|dea' state='on|off'/>
      </keywrap>
      ...
    </domain>
Signed-off-by: NTony Krowiak <akrowiak@linux.vnet.ibm.com>
Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
Signed-off-by: NDaniel Hansel <daniel.hansel@linux.vnet.ibm.com>
Reviewed-by: NBoris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Currently, the libxl driver does not support any security drivers.
When the qemu driver has no security driver configued,
nodeGetSecurityModel succeeds but returns an empty virSecurityModel
object.  Do the same in the libxl driver instead of reporting

this function is not supported by the connection driver:
virNodeGetSecurityModel

We have previously effectively ignored all <controller type='ide'>
elements in a domain definition.

On the i440fx-based machinetypes there is an IDE controller that is
included in the chipset and can't be removed (which is the ide
controller with index='0'>), so it makes sense to ignore that one
controller. However, if an i440fx domain definition has a 2nd
controller, nothing catches this error (unless you also have a disk
attached to it, in which case qemu will complain that you're trying to
use the ide controller named "ide1", which doesn't exist), and if any
other type of domain has even a single controller defined, it will be
incorrectly ignored.

Ignoring a bogus controller definition isn't such a big problem, as
long as an error is logged when any disk is attached to that
non-existent controller. But in the case of q35-based machinetypes,
the hardcoded id ("alias" in libvirt terms) of its builtin SATA
controller is "ide", which happens to be the same id as the builtin
IDE controller on i440fx machinetypes. So libvirt creates a
commandline believing that it is connecting the disk to the builtin
(but actually nonexistent) IDE controller, qemu thinks that libvirt
wanted that disk connected to the builtin SATA controller, and
everybody is happy.

Until you try to connect a 2nd disk to the IDE controller. Then qemu
will complain that you're trying to set unit=1 on a controller that
requires unit=0 (SATA controllers are organized differently than IDE
controllers).

After this patch, if a domain has an IDE controller defined for a
machinetype that has no IDE controllers, libvirt will log an error
about the controller itself as it is building the qemu commandline
(rather than a (possible) error from qemu about disks attached to that
controller). This is done by adding IDE to the list of controller
types that are handled in the loop that creates controller command
strings in qemuBuildCommandline() (previously it would *always* skip
IDE controllers). Then qemuBuildControllerDevStr() is modified to log
an appropriate error in the case of IDE controllers.

In the future, if we add support for extra IDE controllers (piix3-ide
and/or piix4-ide) we can just add it into the IDE case in
qemuBuildControllerDevStr(). For now, nobody seems anxious to add
extra support for an aging and very slow controller, when there are so
many better options available.

Resolves:

https://bugzilla.redhat.com/show_bug.cgi?id=1176071 (Fedora)

Reorganize the loop that builds controller args to remove unnecessary
duplicated code and superfluous else clauses. No functional change.

Back in 2013, commit 877bc089 added in some tests that made sure no
error was generated on a domain definition that had an automatically
added usb controller if that domain didn't have a PCI bus to attach
the usb controller to. This was done because, at that time, libvirt
was automatically adding a usb controller to *any* domain definition
that didn't have one.  Along with permitting the controller, two
s390-specific tests were added to ensure this behavior was maintained
- one with <controller type='usb' model='none'/> and another (called
"s390-piix-controllers") that had both usb and ide controllers, but
nothing attached to them.

Then in February of this year, commit 09ab9dcc eliminated the annoying
auto-adding of a usb device for s390 and s390x machines, stating:

 "Since s390 does not support usb the default creation of a usb
  controller for a domain should not occur."

Although, as verified here, the s390 doesn't support usb, and usb
controllers aren't currently added to s390 domain definitions
automatically, there are likely still some domain definitions in the
wild that have a usb controller (which was added *by libvirt*, not by
the user), so we will keep the tests verifying that behavior for
now. But this patch changes the names of the tests to reflect that
they don't actually contain a valid s390 config; this way future
developers won't propagate the incorrect idea that an s390 virtual
machine can have a USB (or IDE) bus.

In the case of the IDE controller, though, libvirt has never
automatically added an IDE controller unless a user added an IDE disk
(which itself would have caused an error), and we specifically *do*
want to begin generating an error when someone tries to add an IDE
controller to a domain that can't support one. For that reason, while
renaming the sz390-piix-controllers patch, this patch removes the
<controller type='ide'...> from it (otherwise the upcoming patch would
break make check)

This makes sure that that the commandlines generated for devices and
controller devices are all using the alias that has been set in the
controller's object as the id of the controller, rather than
hardcoding a printf (or worse, encoding exceptions to the standard
${controller}${index} into the logic)

Since this "fixes" the controller name used for the sata controller,
the commandline arg for the sata controller in the sata test case had
to be adjusted to be "sata0" instead of "ahci0". All other tests
remain unchanged, verifying that the patch causes no other functional
change.

Because the function that finds a controller alias based on a device
def requires a pointer to the full domainDef in order to get the list
of controllers, the arglist of a few functions had to have this added.

There are a few extra exceptions that weren't being accounted for when
creating the alias for a controller. This resulted in 1) incorrect
status XML, and 2) exceptions/printfs of what *should* have been
directly available in the controller alias when constructing device
commandline arguments:

1) The primary (and only) IDE controller on a 440FX machinetype is
hardcoded to be "ide" in qemu.

2) The primary SATA controller on a 440FX machinetype is also
hardcoded to be "ide" in qemu.

3) On machinetypes that don't support multiple PCI buses, the PCI bus
is hardcoded in qemu to have the name "pci".

4) The first usb master controller is "usb", all others are the normal
"usb%d". (note that usb controllers that are not a "master" will have
the same index, and thus alias, as the master).

We needed to pass in the full domainDef and qemuCaps in order to
properly make the decisions about these exceptions.

Because there are multiple potential reasons for an error, this
function logs any errors before returning NULL (since the caller won't
have the information needed to determine which was the reason for
failure).

When cancelling drive mirror, always try to do that for all disks even
if it fails for some of them. Report the first error we saw.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

Instead of redoing the same filtering over and over everytime we need to
walk through all disks which are being migrated.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

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

And move it to qemu_domain.[ch] because this API is QEMU-only.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

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

The APIs take the memory value in KiB and we store it in KiB
internally, but we cannot parse the whole ULONG_MAX range
on 64-bit systems, because virDomainParseScaledValue
needs to fit the value in bytes in an unsigned long long.

https://bugzilla.redhat.com/show_bug.cgi?id=1176739

Again, a clean-up for which we don't have proper syntax-check.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

We don't allow it in normal code, why would it need to be in the
generated one.  IT also splits the line in perl code so it's readable.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Since we don't have syntax-check for this, it has to be checked
manually.  Let's hope this is the only place it happened.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

This only affected the servers that re-exec themselves, which is only
virtlockd and it didn't do any mess, so this is mostly a clenaup.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1218577

Treat pinning an IOThread via API as if someone added an IOThread to
ensure the iothreadid doesn't cause the guest to disappear

Since 'autofill'd iothreadid entries are not written during XML format
processing, it is possible that if an iothreadid in the middle of an
autofilled list would then change it's id on a subsequent restart.

Thus during the iothreadid deletion, if we determine the delete is not
the "last" thread, then clear the autofill bit for all iothreadid's
following the one being deleted (either the first or one in the middle).
This way, iothreadid's will be printed/saved.

When parsing a cpulist, the virBitmapParse is used. On an invalid
bitmap an error is reported, but the error gets cleared
immediately by subsequent public APIs call, e.g. virDomainFree().
Moreover, we don't check whether bitmap fits into maximal CPU ID
on the host. Therefore the following examples failed without any
error:

 # virsh vcpupin test3 1 aaa

 # virsh vcpupin test3 1 1000
Signed-off-by: NLuyao Huang <lhuang@redhat.com>

We have a lot of passing arguments code just to pass connection
object cause it holds jobTimeout. Taking into account that
right now this value is defined at compile time let's just
get rid of it and make arguments list more clear in many
places.

In case we later need some runtime configurable timeout
value we can provide this value through arguments
function already operate such as a parallels domain
object etc as this timeouts are operation( and thus
object) specific in practice.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@parallels.com>

As of eeb008db the variable is not used anymore. Drop it.
Signed-off-by: NWang Yufei <james.wangyufei@huawei.com>
Signed-off-by: NZhang Bo <oscar.zhangbo@huawei.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

This removes some noise when you're working on the repository
and also have a bunch of source files open in Vim in another
terminal.

In order not to bring in any link dependencies, bridge driver doesn't
use the usual stubs as other conditionally-built code does.  However,
having the function as a macro imposes a problem with possibly unused
variables if just defined as "0".  This was worked around by using
(dom=dom, iface=iface, 0) which should act like a 0 if used in a
condition.  However, gcc still bugs about that, so I came up with
another way how to fix that.

Using static inline functions in the header won't collide with anything,
it fixes the bug and does one thing that the macro didn't do.  It checks
whenther passed variables are pointers of compatible type.  It has only
one downside, and that is that we need to either a) define it with
ATTRIBUTE_UNUSED, which needs an exception in cfg.mk or b) do something
like ignore_value(variable); in the function body.  I went with the
first variant.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

Otherwise we might allow coldplugging a device
that uses an address that is already occupied, creating
an unstartable domain.

https://bugzilla.redhat.com/show_bug.cgi?id=1220195

In the XML we have the vnc port number, but QEMU takes on command line
a vnc screen number, it's port-5900.  We should fail with error message
that only ports in range [5900,65535] are valid.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1164966Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

VIR_STRDUP plays nicely with NULLs. Theres no need to guard its
call with check for non-NULL.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

While implementing support for SPICE, I noticed VNC passwd was
never copied to libxl_device_vfb's vnc.passwd field.
Signed-off-by: NJim Fehlig <jfehlig@suse.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1171984
https://bugzilla.redhat.com/show_bug.cgi?id=1188463

Remove the check for the source host name for iSCSI source XML processing
declaring duplicate sources when the source device path and if present the
initiator of a proposed storage pool matches an existing storage pool.

The backend iSCSI storage driver uses 'iscsiadm --mode session' to query
available iscsid target sessions. The output displayed is the IP address
and the IQN (target path) of known targets. The displayed IP address
is a resolved address based on the session --login. Additionally, iscsid
keeps track of the various ways to define the host name (IPv4 Address,
IPv6 Address, /etc/hosts, etc.) for that IQN (see output of an 'iscsiadm
--mode node'). If an incoming IQN matches and the host name provided by
libvirt is resolved to the existing IQN, then iscsid will "reuse" the
session. Although libvirt could do the same name resolution, if there
is a difference, iscsid could still declare two seemingly different sources
to be the same and not create a new session which means libvirt now has
two storage pools looking at the same source. Thus to avoid any strange
host name resolution issues, just rely on iscsid for that and do not
allow multiple pools on the same host to use the same device path (IQN).

Only perform the port number check if the incoming definition actually
provides it. Since the port number is optional we could erroneously pass
a duplicate source host check since some storage pool backends which fill
in the default port number (e.g., iSCSI and sheepdog) for the started pool.

https://bugzilla.redhat.com/show_bug.cgi?id=1220809

When cold-plugging an RNG device but something fails in
qemuDomainAssignAddresses, we will double free the RNG device.
Once a device is plugged into the domain, we should set the
device pointer to NULL to fix this issue.

...
5  0x00007fb7d180ac8a in virFree at util/viralloc.c:582
6  0x00007fb7d1895cdd in virDomainRNGDefFree at conf/domain_conf.c:19786
7  0x00007fb7d1895d99 in virDomainDeviceDefFree at conf/domain_conf.c:2022
8  0x00007fb7b92b8baf in qemuDomainAttachDeviceFlags at qemu/qemu_driver.c:8785
9  0x00007fb7d190c5d7 in virDomainAttachDeviceFlags at libvirt-domain.c:8488
10 0x00007fb7d23af9d2 in remoteDispatchDomainAttachDeviceFlags at remote_dispatch.h:2842
...
Signed-off-by: NLuyao Huang <lhuang@redhat.com>

Similarly to other error codes that notify the user that the object does
not exist lower the priority of VIR_ERR_NO_DOMAIN_METADATA to
VIR_LOG_DEBUG when writing the log entry.

Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

There is a lot of places, were it's pretty easy for user to enter some
characters that we need to escape to create a valid XML description.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1197580Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

On the s/390x architecture, libvirt may already return 0 in the
node_info->mhz field (see src/nodeinfo.c:linuxNodeInfoCPUPopulate).

We may also want to return this on aarch64 in future, because
calculating the proper value requires SMBIOS, which is not available
on non-server-class systems (specifically on systems which don't
adhere to the SBSA standard).

Therefore this change documents the existing behaviour and provides a
valid path for aarch64.
Signed-off-by: NRichard W.M. Jones <rjones@redhat.com>
Bug-URL: https://bugzilla.redhat.com/1206353

https://bugzilla.redhat.com/show_bug.cgi?id=1191227

Since 0fa15b19 we have this variable SYNC_TIME which allows users to
synchronize time on domain resume. However, despite what documentation
says, it's by default on because it's never initialized. Fix this by
setting it to zero at the beginning of the libvirt-guests script.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

The code to add device type to the commandline was identical for lsi
and other models of SCSI controllers, but was duplicated (with the
exception of a minor ordering difference of the if-else clauses) for
the two cases. This patch replaces those two with a single instance of
the code just before the if().

This patch makes qemuValideDevicePCISlotsChipsets() more consistent in
appearance by replacing several clauses of an if with the equivalent
call to qemuDomainMachineIsI440FX. The if was checking exactly the
same items, just in a slightly different order.