Change the signature of the function-like macro SHPC_VMSTATE(), so that we
can produce and expect this field conditionally in the migration stream,
starting with an upcoming patch.

Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Marcel Apfelbaum <marcel@redhat.com>
Signed-off-by: NLaszlo Ersek <lersek@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

There is no _TEST() variant of VMSTATE_BUFFER_UNSAFE_INFO() yet, but we'll
soon need it. Introduce it and rebase the original
VMSTATE_BUFFER_UNSAFE_INFO() on top.

The parameter order of the new function-like macro follows that of
VMSTATE_SINGLE_TEST(): "_test" is introduced between "_state" and
"_version".

Cc: Juan Quintela <quintela@redhat.com>
Cc: Amit Shah <amit.shah@redhat.com>
Cc: Marcel Apfelbaum <marcel@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: NLaszlo Ersek <lersek@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Simplifies multiseat configuration, see
docs/multiseat.txt update for details.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

remove some code duplication in acpi-build.c and drop 5
ASL and binary blobs files with TPM ACPI device description,
replacing them with 1 small hunk written in AML API.
Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Igor agreed to help review ACPI patches, add an entry to MAINTAINERS
with all ACPI stuff I could think of.
Note: I listed ARM ACPI files here just to make sure we are Cc'd, no
plan to maintain ACPI for ARM through my tree :)
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

We override the error value r in fail_vq, this will cause the caller
can't detect the failure which may cause the caller may disable the
notifiers twice if vhost is failed to start. Fix this by using another
variable to keep track the return value of set_host_notifier().

Fixes b0b3db79 ("vhost-net: cleanup
host notifiers at last step")

Cc: qemu-stable@nongnu.org
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: NJason Wang <jasowang@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Excessive virtio_balloon inflation can cause invocation of OOM-killer,
when Linux is under severe memory pressure. Various mechanisms are
responsible for correct virtio_balloon memory management. Nevertheless it
is often the case that these control tools does not have enough time to
react on fast changing memory load. As a result OS runs out of memory and
invokes OOM-killer. The balancing of memory by use of the virtio balloon
should not cause the termination of processes while there are pages in the
balloon. Now there is no way for virtio balloon driver to free memory at
the last moment before some process get killed by OOM-killer.

This does not provide a security breach as balloon itself is running
inside Guest OS and is working in the cooperation with the host. Thus
some improvements from Guest side should be considered as normal.

To solve the problem, introduce a virtio_balloon callback which is
expected to be called from the oom notifier call chain in out_of_memory()
function. If virtio balloon could release some memory, it will make the
system return and retry the allocation that forced the out of memory
killer to run.

This behavior should be enabled if and only if appropriate feature bit
is set on the device. It is off by default.

This functionality was recently merged into vanilla Linux.

  commit 5a10b7dbf904bfe01bb9fcc6298f7df09eed77d5
  Author: Raushaniya Maksudova <rmaksudova@parallels.com>
  Date:   Mon Nov 10 09:36:29 2014 +1030

This patch adds respective control bits into QEMU. It introduces
deflate-on-oom option for balloon device which does the trick.
Signed-off-by: NDenis V. Lunev <den@openvz.org>
CC: Raushaniya Maksudova <rmaksudova@parallels.com>
CC: Anthony Liguori <aliguori@amazon.com>
CC: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Acked-by: NJames Bottomley <JBottomley@Odin.com>
Reviewed-by: NChristian Borntraeger <borntraeger@de.ibm.com>

Potentially overflowing expression "1 << prop->bitnr" with
type "int" (32 bits, signed) is evaluated using 32-bit arithmetic,
and then used in a context that expects an expression of type
"uint64_t" (64 bits, unsigned).

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Overrunning array "proxy->guest_features" of 2 4-byte
elements at element index 2 (byte offset 8) using index
"proxy->gfselect" (which evaluates to 2). Normally, the
Linux kernel driver just read/write '0' or '1' as the
"proxy->gfselect" values, so using '<' instead of '=<' to
make coverity happy and avoid potential harm.

Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

The action to potentially switch sp register is not occurring at the correct
point in the interrupt entry or exception exit sequences.

For the interrupt entry case the sp on entry is used to create the stack
exception frame - but this may well be the user stack pointer, since we
haven't done the switch yet. Re-order the flow to switch the sp regs then
use the current sp to create the exception frame.

For the return from exception case the code is unwinding the sp after
switching sp registers. But it should always unwind the supervisor sp
first, then carry out any required sp switch.

Note that these problems don't effect operation unless the user sp bit is
set in the CACR register. Only a single sp is used in the default power up
state. Previously Linux only used this single sp mode. But modern versions
of Linux use the user sp mode now, so we need correct behavior for Linux
to work.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: NLaurent Vivier <laurent@vivier.eu>
Tested-by: NLaurent Vivier <laurent@vivier.eu>
Message-id: 1434721406-25288-4-git-send-email-gerg@uclinux.org
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Fill out the code support for the move to/from usp instructions. They are
being decoded, but there is no code to support there actions. So add it.

Current versions of Linux running on the ColdFire 5208 use these instructions.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Reviewed-by: NRichard Henderson <rth@twiddle.net>
Reviewed-by: NLaurent Vivier <laurent@vivier.eu>
Tested-by: NLaurent Vivier <laurent@vivier.eu>
Message-id: 1434721406-25288-3-git-send-email-gerg@uclinux.org
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Implement the SIMR and CIMR registers of the 5208 interrupt controller.
These are used by modern versions of Linux running on ColdFire (not sure
of the exact version they were introduced, but they have been in for quite
a while now).

Without this change when attempting to run a linux-3.5 kernel you will
see:

  qemu: hardware error: mcf_intc_write: Bad write offset 28

and execution will stop and dump out.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Tested-by: NLaurent Vivier <laurent@vivier.eu>
Message-id: 1434721406-25288-2-git-send-email-gerg@uclinux.org
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

QOM infrastructure fixes and device conversions

* Changes to name string ownership for alias properties
* Improvements around enum properties
* Cleanups around -object handling
* New helper functions
* Cleanups of qdev init helper functions
* Add path argument to qom-tree script
* QTest cleanup to use new qtest_add_data_func() consistently

# gpg: Signature made Fri Jun 19 18:14:38 2015 BST using RSA key ID 3E7E013F
# gpg: Good signature from "Andreas Färber <afaerber@suse.de>"
# gpg:                 aka "Andreas Färber <afaerber@suse.com>"

* remotes/afaerber/tags/qom-devices-for-peter:
  qdev: Un-deprecate qdev_init_nofail()
  qdev: Deprecated qdev_init() is finally unused, drop
  qom: Don't pass string table to object_get_enum() function
  qom: Add an object_property_add_enum() helper function
  qom: Make enum string tables const-correct
  qom: Add object_new_with_props() / object_new_withpropv() helpers
  qom: Add helper function for getting user objects root
  vl: Create (most) objects before creating chardev backends
  doc: Document user creatable object types in help text
  backends: Fix typename of 'policy' enum property in hostmem obj
  scripts: Add support for path as argument of qom-tree
  tests: Use qtest_add_data_func() consistently
  qdev: Free property names after registering gpio aliases
  qom: strdup() target property name on object_property_add_alias()
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

* i8254 security fix
* Avoid long 100% CPU wait after restarting guests that use the periodic timer
* Fixes for access clamping (WinXP, MIPS)
* wixl/.msi support for qemu-ga on Windows

# gpg: Signature made Fri Jun 19 11:30:53 2015 BST using RSA key ID 78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream:
  exec: clamp accesses against the MemoryRegionSection
  exec: do not clamp accesses to MMIO regions
  mc146818rtc: Reset the periodic timer on load
  qemu-timer: Call clock reset notifiers on forward jumps
  tests: virtio-scsi: Add test for unaligned WRITE SAME
  tests: virtio-scsi: Move start/stop to individual test functions
  libqos: Complete virtio device ID definition list
  libqos: Allow calling guest_free on NULL pointer
  tests: Link libqos virtio object to virtio-scsi-test
  i8254: fix out-of-bounds memory access in pit_ioport_read()
  qemu-ga: Building Windows MSI installation with configure/Makefile
  qemu-ga: Introduce Windows MSI script
  qemu-ga: debug printouts to help troubleshoot installation
  qemu-ga: adding vss-[un]install options
  qemu-log: Open file for logging when specified
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

This code is already being run in the mb_cpu_realizefn()
function. As PVR registers are preserved on reset this
code is not required.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Remove the hardcoded values from the machine specific reset
function, as the same values are already set in the standard
MicroBlaze reset.

This also allows the entire reset function to be deleted, as
PVR registers are now preserved on reset.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Move the hard coded register values to the init function.
This also allows the entire reset function to be deleted, as
PVR registers are now preserved on reset.

The hardcoded PVR0 values can be removed as they are setting
the endianness and stack protection, which is already done
or invalid.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally the pvr-full PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally the version_mask PVR bits were manually set for each
machine. This is a hassle and difficult to read, instead set them
based on the CPU properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally the endi PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally  the dcache-writeback PVR bits were manually set for each machine.
This is a hassle and difficult to read, instead set them based on the CPU
properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally the use-mmu PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Rename the usefpu variable to use_fpu.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Stack protection is not available when the MMU is enabled.
As the MMU is enabled by default, disable stack protection
by default.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Originally the use-fpu PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Rename the "xlnx.base-vectors" string to "base-vectors" and
move the base_vectors variable into the cfg struct.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Microblaze stack protection is configurable and isn't always enabled.
This patch allows the stack protection to be disabled from the
CPU properties.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Move the Microblaze PVR registers to the end of the CPUMBState
and preserve them during reset. This is similar to what the
QEMU ARM model does with some of it's registers.

This allows the Microblaze PVR registers to only be set once
at realise instead of constantly at reset.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Fix up the incorrect indentation level in the helper_stackprot() function.
Signed-off-by: NAlistair Francis <alistair.francis@xilinx.com>
Reviewed-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

Instantiate and realise the CPU directly, rather than using
cpu_mb_init. Microblazes cpu_model argument is a dummy so remove the
default cpu_model set logic.
Reviewed-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: NPeter Crosthwaite <crosthwaite.peter@gmail.com>
Signed-off-by: NEdgar E. Iglesias <edgar.iglesias@xilinx.com>

It's a perfectly sensible helper function.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

qdev_init() is a wrapper around setting property "realized" to true,
plus error handling that passes errors to qerror_report_err().
qerror_report_err() is a transitional interface to help with
converting existing monitor commands to QMP.  It should not be used
elsewhere.

All code has been modernized to avoid qdev_init() and its
inappropriate error handling.  We can finally drop it.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

Now that properties can be explicitly registered as an enum
type, there is no need to pass the string table to the
object_get_enum() function. The object property registration
already has a pointer to the string table.

In changing this method signature, the hostmem backend object
has to be converted to use the new enum property registration
code, which simplifies it somewhat.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

A QOM property can be parsed as enum using the visit_type_enum()
helper function, but this forces callers to use the more complex
generic object_property_add() method when registering it. It
also requires that users of that object have access to the
string map when they want to read the property value.

This patch introduces a specialized object_property_add_enum()
method which simplifies the use of enum properties, so the
setters/getters directly get passed the int value.

  typedef enum {
     MYDEV_TYPE_FROG,
     MYDEV_TYPE_ALLIGATOR,
     MYDEV_TYPE_PLATYPUS,

     MYDEV_TYPE_LAST
  } MyDevType;

Then provide a table of enum <-> string mappings

  static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = {
     [MYDEV_TYPE_FROG] = "frog",
     [MYDEV_TYPE_ALLIGATOR] = "alligator",
     [MYDEV_TYPE_PLATYPUS] = "platypus",
     [MYDEV_TYPE_LAST] = NULL,
  };

Assuming an object struct of

   typedef struct {
      Object parent_obj;
      MyDevType devtype;
      ...other fields...
   } MyDev;

The property can then be registered as follows:

   static int mydev_prop_get_devtype(Object *obj,
                                     Error **errp G_GNUC_UNUSED)
   {
       MyDev *dev = MYDEV(obj);

       return dev->devtype;
   }

   static void mydev_prop_set_devtype(Object *obj,
                                      int value,
                                      Error **errp G_GNUC_UNUSED)
   {
       MyDev *dev = MYDEV(obj);

       dev->devtype = value;
   }

   object_property_add_enum(obj, "devtype",
                            mydevtypemap, "MyDevType",
                            mydev_prop_get_devtype,
                            mydev_prop_set_devtype,
                            NULL);

Note there is no need to check the range of 'value' in
the setter, because the string->enum conversion code will
have already done that and reported an error as required.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

The enum string table parameters in various QOM/QAPI methods
are declared 'const char *strings[]'. This results in const
warnings if passed a variable that was declared as

   static const char * const strings[] = { .... };

Add the extra const annotation to the parameters, since
neither the string elements, nor the array itself should
ever be modified.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

It is reasonably common to want to create an object, set a
number of properties, register it in the hierarchy and then
mark it as complete (if a user creatable type). This requires
quite a lot of error prone, verbose, boilerplate code to achieve.

First a pair of functions object_set_props() / object_set_propv()
are added which allow for a list of objects to be set in
one single API call.

Then object_new_with_props() / object_new_with_propv() constructors
are added which simplify the sequence of calls to create an
object, populate properties, register in the object composition
tree and mark the object complete, into a single method call.

Usage would be:

   Error *err = NULL;
   Object *obj;
   obj = object_new_with_propv(TYPE_MEMORY_BACKEND_FILE,
                               object_get_objects_root(),
                               "hostmem0",
                               &err,
                               "share", "yes",
                               "mem-path", "/dev/shm/somefile",
                               "prealloc", "yes",
                               "size", "1048576",
                               NULL);

Note all property values are passed in string form and will
be parsed into their required data types, using normal QOM
semantics for parsing from string format.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

Add object_get_objects_root() function which is a convenience for
obtaining the Object * located at /objects in the object
composition tree. Convert existing code over to use the new
API where appropriate.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

Some types of object must be created before chardevs, other types of
object must be created after chardevs. As such there is no option but
to create objects in two phases.

This takes the decision to create as many object types as possible
right away before anyother backends are created, and only delay
creation of those few which have an explicit dependency on the
chardevs. Hopefully the set which need delaying will remain small
over time.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

The QEMU help for -object is essentially useless, just giving users
the generic syntax. Move it down into its own section and introduce
a nested table where each user creatable object can be documented.
The existing memory-backend-file, rng-random and rng-egd object
types are documented.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>

The 'policy' property was being registered with a typename of
'str', but it is in fact an enum of the 'HostMemPolicy' type.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NAndreas Färber <afaerber@suse.de>