In the aspeed_rtc device we store a difference between two time_t
values in an 'int'. This is not really correct when time_t could
be 64 bits. Enlarge the field to 'int64_t'.

This is a migration compatibility break for the aspeed boards.
While we are changing the vmstate, remove the accidental
duplicate of the offset field.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>

The SRC device is normally used to start the secondary CPU.

When running Linux directly, QEMU is emulating a PSCI interface that UBOOT
is installing at boot time and therefore the fact that the SRC device is
unimplemented is hidden as Qemu respond directly to PSCI requets without
using the SRC device.

But if you try to run a more bare metal application (maybe uboot itself),
then it is not possible to start the secondary CPU as the SRC is an
unimplemented device.

This patch adds the ability to start the secondary CPU through the SRC
device so that you can use this feature in bare metal applications.

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: ce9a0162defd2acee5dc7f8a674743de0cded569.1692964892.git.jcd@tribudubois.net
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

* Add TZASC as unimplemented device.
  - Allow bare metal application to access this (unimplemented) device
* Add CSU as unimplemented device.
  - Allow bare metal application to access this (unimplemented) device
* Add various memory segments
  - OCRAM
  - OCRAM EPDC
  - OCRAM PXP
  - OCRAM S
  - ROM
  - CAAM

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: f887a3483996ba06d40bd62ffdfb0ecf68621987.1692964892.git.jcd@tribudubois.net
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

* Add Addr and size definition for all i.MX7 devices in i.MX7 header file.
* Use those newly defined named constants whenever possible.
* Standardize the way we init a familly of unimplemented devices
  - SAI
  - PWM
  - CAN
* Add/rework few comments

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Message-id: 59e195d33e4d486a8d131392acd46633c8c10ed7.1692964892.git.jcd@tribudubois.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

* Add TZASC as unimplemented device.
  - Allow bare metal application to access this (unimplemented) device
* Add CSU as unimplemented device.
  - Allow bare metal application to access this (unimplemented) device
* Add 4 missing PWM devices

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: 59e4dc56e14eccfefd379275ec19048dff9c10b3.1692964892.git.jcd@tribudubois.net
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

* Add Addr and size definition for most i.MX6UL devices in i.MX6UL header file.
* Use those newly defined named constants whenever possible.
* Standardize the way we init a familly of unimplemented devices
  - SAI
  - PWM
  - CAN
* Add/rework few comments

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Message-id: d579043fbd4e4b490370783fda43fc02c8e9be75.1692964892.git.jcd@tribudubois.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

i.MX7 IOMUX GPR device is not equivalent to i.MX6UL IOMUXC GPR device.
In particular, register 22 is not present on i.MX6UL and this is actualy
The only register that is really emulated in the i.MX7 IOMUX GPR device.

Note: The i.MX6UL code is actually also implementing the IOMUX GPR device
as an unimplemented device at the same bus adress and the 2 instantiations
were actualy colliding. So we go back to the unimplemented device for now.

Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Message-id: 48681bf51ee97646479bb261bee19abebbc8074e.1692964892.git.jcd@tribudubois.net
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

When target image is using subclusters, and we align the request during
copy-on-read, it makes sense to align to subcluster_size rather than
cluster_size.  Otherwise we end up with unnecessary allocations.

This commit renames bdrv_round_to_clusters() to bdrv_round_to_subclusters()
and utilizes subcluster_size field of BlockDriverInfo to make necessary
alignments.  It affects copy-on-read as well as mirror job (which is
using bdrv_round_to_clusters()).

This change also fixes the following bug with failing assert (covered by
the test in the subsequent commit):

qemu-img create -f qcow2 base.qcow2 64K
qemu-img create -f qcow2 -o extended_l2=on,backing_file=base.qcow2,backing_fmt=qcow2 img.qcow2 64K
qemu-io -c "write -P 0xaa 0 2K" img.qcow2
qemu-io -C -c "read -P 0x00 2K 62K" img.qcow2

qemu-io: ../block/io.c:1236: bdrv_co_do_copy_on_readv: Assertion `skip_bytes < pnum' failed.

Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230711172553.234055-3-andrey.drobyshev@virtuozzo.com>

This is going to be used in the subsequent commit as requests alignment
(in particular, during copy-on-read).  This value only makes sense for
the formats which support subclusters (currently QCOW2 only).  If this
field isn't set by driver's own bdrv_get_info() implementation, we
simply set it equal to the cluster size thus treating each cluster as
having a single subcluster.

Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230711172553.234055-2-andrey.drobyshev@virtuozzo.com>

This reverts commit fc15bfb6.

This patch caused a regression for tricore-softmmu,
./tests/tcg/tricore-softmmu/test_boot_to_main.c.tst.

Reported-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

After system startup, run the update to memory_dispatch
and the tlb_flush on the cpu.  This eliminates a race,
wherein a running cpu sees the memory_dispatch change
but has not yet seen the tlb_flush.

Since the update now happens on the cpu, we need not use
qatomic_rcu_read to protect the read of memory_dispatch.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1826
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1834
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1846


Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Implement PRAGMA_DISABLE_PACKED_WARNING and
PRAGMA_REENABLE_PACKED_WARNING macros in include/qemu/compiler.h.

Signed-off-by: Kyle Evans <kevans@FreeBSD.org>
Signed-off-by: Karim Taha <kariem.taha2.7@gmail.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>

move _WANT_FREEBSD macros from bsd-user/freebsd/os-syscall.c to
include/qemu/osdep.h in order to pull some struct defintions needed
later in the build.

Signed-off-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Karim Taha <kariem.taha2.7@gmail.com>
Acked-by: Richard Henderson <richard.henderson@linaro.org>

Introduce a new opcode for negative setcond.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Replace the separate defines with TCG_TARGET_HAS_extr_i64_i32,
so that the two parts of backend-specific type changing cannot
be out of sync.

Reported-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: <20230822175127.1173698-1-richard.henderson@linaro.org>

tlb_addr is changed from target_ulong to uint64_t to match the type of
a CPUTLBEntry value, and the addressed is changed to vaddr.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-8-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

In system mode, abi_ptr is primarily used for representing addresses
when accessing guest memory with cpu_[st|ld]*(). Widening it from
target_ulong to vaddr reduces the target dependence of these functions
and is step towards building accel/ once for system mode.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-7-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Changes the address type of the guest memory read/write functions from
target_ulong to abi_ptr. (abi_ptr is currently typedef'd to target_ulong
but that will change in a following commit.) This will reduce the
coupling between accel/ and target/.

Note: Function pointers that point to cpu_[st|ld]*() in target/riscv and
target/rx are also updated in this commit.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-6-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Changes the signature of the target-defined functions for
inserting/removing hvf hw breakpoints. The address and length arguments
are now of vaddr type, which both matches the type used internally in
accel/hvf/hvf-all.c and makes the api target-agnostic.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-5-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Changes the signature of the target-defined functions for
inserting/removing kvm hw breakpoints. The address and length arguments
are now of vaddr type, which both matches the type used internally in
accel/kvm/kvm-all.c and makes the api target-agnostic.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-4-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Widens the pc and saved_insn fields of hvf_sw_breakpoint from
target_ulong to vaddr. Other hvf_* functions accessing hvf_sw_breakpoint
are also widened to match.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-3-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Widens the pc and saved_insn fields of kvm_sw_breakpoint from
target_ulong to vaddr. The pc argument of kvm_find_sw_breakpoint is also
widened to match.

Signed-off-by: Anton Johansson <anjo@rev.ng>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230807155706.9580-2-anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Using "-device virtio-gpu,blob=true" currently does not work on big
endian hosts (like s390x). The guest kernel prints an error message
like:

 [drm:virtio_gpu_dequeue_ctrl_func [virtio_gpu]] *ERROR* response 0x1200 (command 0x10c)

and the display stays black. When running QEMU with "-d guest_errors",
it shows an error message like this:

 virtio_gpu_create_mapping_iov: nr_entries is too big (83886080 > 16384)

which indicates that this value has not been properly byte-swapped.
And indeed, the virtio_gpu_create_blob_bswap() function (that should
swap the fields in the related structure) fails to swap some of the
entries. After correctly swapping all missing values here, too, the
virtio-gpu device is now also working with blob=true on s390x hosts.

Fixes: e0933d91 ("virtio-gpu: Add virtio_gpu_resource_create_blob")
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2230469


Message-Id: <20230815122007.928049-1-thuth@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

Add 8.2 machine types for arm/i440fx/m68k/q35/s390x/spapr.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20230718142235.135319-1-cohuck@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Acked-by: Laurent Vivier <laurent@vivier.eu>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

kvm_arch_get_default_type() returns the default KVM type. This hook is
particularly useful to derive a KVM type that is valid for "none"
machine model, which is used by libvirt to probe the availability of
KVM.

For MIPS, the existing mips_kvm_type() is reused. This function ensures
the availability of VZ which is mandatory to use KVM on the current
QEMU.

Cc: qemu-stable@nongnu.org
Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-id: 20230727073134.134102-2-akihiko.odaki@daynix.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: added doc comment for new function]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

Implement nRF51 DETECT signal in the GPIO peripheral.

The reference manual makes mention of a per-pin DETECT signal, but these
are not exposed to the user. See https://devzone.nordicsemi.com/f/nordic-q-a/39858/gpio-per-pin-detect-signal-available


for more information. Currently, I don't see a reason to model these.

Signed-off-by: Chris Laplante <chris@laplante.io>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20230728160324.1159090-2-chris@laplante.io
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Rather than using a zero tuple to end the table, use a macro
to apply ARRAY_SIZE and pass that on to the convert functions.

This fixes two bugs in which the conversion functions required
that both the target and host masks be non-zero in order to
continue, rather than require both target and host masks be
zero in order to terminate.

This affected mmap_flags_tbl when the host does not support
all of the flags we wish to convert (e.g. MAP_UNINITIALIZED).
Mapping these flags to zero is good enough, and matches how
the kernel ignores bits that are unknown.

Fixes: 4b840f96 ("linux-user: Populate more bits in mmap_flags_tbl")
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

In order for our emulation of MAP_FIXED_NOREPLACE to succeed within
linux-user target_mmap, we require a non-zero value.  This does not
require host kernel support, merely the bit being defined.

MAP_FIXED_NOREPLACE was added with glibc 2.28.  From repology.org:

                  Fedora 36: 2.35
          CentOS 8 (RHEL-8): 2.28
                  Debian 11: 2.31
         OpenSUSE Leap 15.4: 2.31
           Ubuntu LTS 20.04: 2.31

Reported-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230808164418.69989-1-richard.henderson@linaro.org>

We will want to be able to search the set of mappings.
For this patch, the two users iterate the tree in order.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Calling OpenGL from different threads can have bad consequences if not
carefully reviewed. It's not generally supported. In my case, I was
debugging a crash in glDeleteTextures from OPENGL32.DLL, where I asked
qemu for gl=es, and thus ANGLE implementation was expected. libepoxy did
resolution of the global pointer for glGenTexture to the GLES version
from the main thread. But it resolved glDeleteTextures to the GL
version, because it was done from a different thread without correct
context. Oops.

Let's stick to the main thread for GL calls by using a BH.

Note: I didn't use atomics for reset_finished check, assuming the BQL
will provide enough of sync, but I might be wrong.

Acked-by: Dongwon Kim <dongwon.kim@intel.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20230726173929.690601-3-marcandre.lureau@redhat.com>

Acked-by: Dongwon Kim <dongwon.kim@intel.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20230726173929.690601-2-marcandre.lureau@redhat.com>

OpenRISC (or1k) has long long alignment to 4 bytes, but currently not
defined in abitypes.h. This lead to incorrect packing of /epoll_event/
structure and eventually infinite loop while waiting for file
descriptor[s] event[s].

Fixed also CRIS alignments (1 byte for all types).

Signed-off-by: Luca Bonissi <qemu@bonslack.org>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1770


Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

The first bitfield here is supposed to be used as a 64-bit equivalent
to the "uint64_t msi_addr" in the union. To make this work correctly
on big endian hosts, too, the __addr_hi field has to be part of the
bitfield, and the the bitfield members must be declared with "uint64_t"
instead of "uint32_t" - otherwise the values are placed in the wrong
bytes on big endian hosts.

Same applies to the 32-bit "msi_data" field: __resved1 must be part
of the bitfield, and the members must be declared with "uint32_t"
instead of "uint16_t".

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230802135723.178083-7-thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>

The code already tries to do some endianness handling here, but
currently fails badly:
- While it already swaps the data when logging errors / tracing, it fails
  to byteswap the value before e.g. accessing entry->irte.present
- entry->irte.source_id is swapped with le32_to_cpu(), though this is
  a 16-bit value
- The whole union is apparently supposed to be swapped via the 64-bit
  data[2] array, but the struct is a mixture between 32 bit values
  (the first 8 bytes) and 64 bit values (the second 8 bytes), so this
  cannot work as expected.

Fix it by converting the struct to two proper 64-bit bitfields, and
by swapping the values only once for everybody right after reading
the data from memory.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230802135723.178083-3-thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>

The QEMU CI fails in virtio-scmi test occasionally.  As reported by
Thomas Huth, this happens most likely when the system is loaded and it
fails with the following error:

  qemu-system-aarch64: ../../devel/qemu/hw/pci/msix.c:659:
  msix_unset_vector_notifiers: Assertion `dev->msix_vector_use_notifier && dev->msix_vector_release_notifier' failed.
  ../../devel/qemu/tests/qtest/libqtest.c:200: kill_qemu() detected QEMU death from signal 6 (Aborted) (core dumped)

As discovered by Fabiano Rosas, the cause is a duplicate invocation of
msix_unset_vector_notifiers via duplicate vu_scmi_stop calls:

  msix_unset_vector_notifiers
  virtio_pci_set_guest_notifiers
  vu_scmi_stop
  vu_scmi_disconnect
  ...
  qemu_chr_write_buffer

  msix_unset_vector_notifiers
  virtio_pci_set_guest_notifiers
  vu_scmi_stop
  vu_scmi_set_status
  ...
  qemu_cleanup

While vu_scmi_stop calls are protected by vhost_dev_is_started()
check, it's apparently not enough.  vhost-user-blk and vhost-user-gpio
use an extra protection, see f5b22d06 (vhost: recheck dev state in
the vhost_migration_log routine) for the motivation.  Let's use the
same in vhost-user-scmi, which fixes the failure above.

Fixes: a5dab090 ("hw/virtio: Add boilerplate for vhost-user-scmi device")
Signed-off-by: Milan Zamazal <mzamazal@redhat.com>
Message-Id: <20230720101037.2161450-1-mzamazal@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230730180329.851576-1-sw@weilnetz.de>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>

The TLS handshake make take some time to complete, during which time an
I/O watch might be registered with the main loop. If the owner of the
I/O channel invokes qio_channel_close() while the handshake is waiting
to continue the I/O watch must be removed. Failing to remove it will
later trigger the completion callback which the owner is not expecting
to receive. In the case of the VNC server, this results in a SEGV as
vnc_disconnect_start() tries to shutdown a client connection that is
already gone / NULL.

CVE-2023-3354
Reported-by: jiangyegen <jiangyegen@huawei.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

Coverity points out (CID 1513106, 1513107) that MemoryListener is a
192 byte struct which we are passing around by value.  Switch to
passing a const pointer into xen_register_ioreq() and then to
xen_do_ioreq_register().  We can also make the file-scope
MemoryListener variables const, since nothing changes them.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230718101057.1110979-1-peter.maydell@linaro.org>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>

Extend query-migrate to provide throttle time and estimated
ring full time with dirty-limit capability enabled, through which
we can observe if dirty limit take effect during live migration.

Signed-off-by: Hyman Huang(黄勇) <yong.huang@smartx.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-ID: <168733225273.5845.15871826788879741674-8@git.sr.ht>
Signed-off-by: Juan Quintela <quintela@redhat.com>

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>