xref: /openbmc/qemu/docs/about/deprecated.rst (revision 94ae227e)
1.. _Deprecated features:
2
3Deprecated features
4===================
5
6In general features are intended to be supported indefinitely once
7introduced into QEMU. In the event that a feature needs to be removed,
8it will be listed in this section. The feature will remain functional for the
9release in which it was deprecated and one further release. After these two
10releases, the feature is liable to be removed. Deprecated features may also
11generate warnings on the console when QEMU starts up, or if activated via a
12monitor command, however, this is not a mandatory requirement.
13
14As a special exception to this general timeframe, rather than have an
15indefinite lifetime, versioned machine types are only intended to be
16supported for a period of 6 years, equivalent to 18 QEMU releases. All
17versioned machine types will be automatically marked deprecated after an
18initial 3 years (9 QEMU releases) has passed, and will then be deleted after
19a further 3 year period has passed. It is recommended that a deprecated
20machine type is only used for incoming migrations and restore of saved state,
21for pre-existing VM deployments. They should be scheduled for updating to a
22newer machine type during an appropriate service window. Newly deployed VMs
23should exclusively use a non-deprecated machine type, with use of the most
24recent version highly recommended. Non-versioned machine types follow the
25general feature deprecation policy.
26
27Prior to the 2.10.0 release there was no official policy on how
28long features would be deprecated prior to their removal, nor
29any documented list of which features were deprecated. Thus
30any features deprecated prior to 2.10.0 will be treated as if
31they were first deprecated in the 2.10.0 release.
32
33What follows is a list of all features currently marked as
34deprecated.
35
36System emulator command line arguments
37--------------------------------------
38
39Short-form boolean options (since 6.0)
40''''''''''''''''''''''''''''''''''''''
41
42Boolean options such as ``share=on``/``share=off`` could be written
43in short form as ``share`` and ``noshare``.  This is now deprecated
44and will cause a warning.
45
46``delay`` option for socket character devices (since 6.0)
47'''''''''''''''''''''''''''''''''''''''''''''''''''''''''
48
49The replacement for the ``nodelay`` short-form boolean option is ``nodelay=on``
50rather than ``delay=off``.
51
52Plugin argument passing through ``arg=<string>`` (since 6.1)
53''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
54
55Passing TCG plugins arguments through ``arg=`` is redundant is makes the
56command-line less readable, especially when the argument itself consist of a
57name and a value, e.g. ``-plugin plugin_name,arg="arg_name=arg_value"``.
58Therefore, the usage of ``arg`` is redundant. Single-word arguments are treated
59as short-form boolean values, and passed to plugins as ``arg_name=on``.
60However, short-form booleans are deprecated and full explicit ``arg_name=on``
61form is preferred.
62
63``-smp`` (Unsupported "parameter=1" SMP configurations) (since 9.0)
64'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
65
66Specified CPU topology parameters must be supported by the machine.
67
68In the SMP configuration, users should provide the CPU topology parameters that
69are supported by the target machine.
70
71However, historically it was allowed for users to specify the unsupported
72topology parameter as "1", which is meaningless. So support for this kind of
73configurations (e.g. -smp drawers=1,books=1,clusters=1 for x86 PC machine) is
74marked deprecated since 9.0, users have to ensure that all the topology members
75described with -smp are supported by the target machine.
76
77``-runas`` (since 9.1)
78----------------------
79
80Use ``-run-with user=..`` instead.
81
82
83User-mode emulator command line arguments
84-----------------------------------------
85
86``-p`` (since 9.0)
87''''''''''''''''''
88
89The ``-p`` option pretends to control the host page size.  However,
90it is not possible to change the host page size, and using the
91option only causes failures.
92
93QEMU Machine Protocol (QMP) commands
94------------------------------------
95
96``blockdev-open-tray``, ``blockdev-close-tray`` argument ``device`` (since 2.8)
97'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
98
99Use argument ``id`` instead.
100
101``eject`` argument ``device`` (since 2.8)
102'''''''''''''''''''''''''''''''''''''''''
103
104Use argument ``id`` instead.
105
106``blockdev-change-medium`` argument ``device`` (since 2.8)
107''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
108
109Use argument ``id`` instead.
110
111``block_set_io_throttle`` argument ``device`` (since 2.8)
112'''''''''''''''''''''''''''''''''''''''''''''''''''''''''
113
114Use argument ``id`` instead.
115
116``blockdev-add`` empty string argument ``backing`` (since 2.10)
117'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
118
119Use argument value ``null`` instead.
120
121``block-commit`` arguments ``base`` and ``top`` (since 3.1)
122'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
123
124Use arguments ``base-node`` and ``top-node`` instead.
125
126``nbd-server-add`` and ``nbd-server-remove`` (since 5.2)
127''''''''''''''''''''''''''''''''''''''''''''''''''''''''
128
129Use the more generic commands ``block-export-add`` and ``block-export-del``
130instead.  As part of this deprecation, where ``nbd-server-add`` used a
131single ``bitmap``, the new ``block-export-add`` uses a list of ``bitmaps``.
132
133``query-qmp-schema`` return value member ``values`` (since 6.2)
134'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
135
136Member ``values`` in return value elements with meta-type ``enum`` is
137deprecated.  Use ``members`` instead.
138
139``drive-backup`` (since 6.2)
140''''''''''''''''''''''''''''
141
142Use ``blockdev-backup`` in combination with ``blockdev-add`` instead.
143This change primarily separates the creation/opening process of the backup
144target with explicit, separate steps. ``blockdev-backup`` uses mostly the
145same arguments as ``drive-backup``, except the ``format`` and ``mode``
146options are removed in favor of using explicit ``blockdev-create`` and
147``blockdev-add`` calls. See :doc:`/interop/live-block-operations` for
148details.
149
150Incorrectly typed ``device_add`` arguments (since 6.2)
151''''''''''''''''''''''''''''''''''''''''''''''''''''''
152
153Due to shortcomings in the internal implementation of ``device_add``, QEMU
154incorrectly accepts certain invalid arguments: Any object or list arguments are
155silently ignored. Other argument types are not checked, but an implicit
156conversion happens, so that e.g. string values can be assigned to integer
157device properties or vice versa.
158
159This is a bug in QEMU that will be fixed in the future so that previously
160accepted incorrect commands will return an error. Users should make sure that
161all arguments passed to ``device_add`` are consistent with the documented
162property types.
163
164Host Architectures
165------------------
166
167BE MIPS (since 7.2)
168'''''''''''''''''''
169
170As Debian 10 ("Buster") moved into LTS the big endian 32 bit version of
171MIPS moved out of support making it hard to maintain our
172cross-compilation CI tests of the architecture. As we no longer have
173CI coverage support may bitrot away before the deprecation process
174completes. The little endian variants of MIPS (both 32 and 64 bit) are
175still a supported host architecture.
176
177System emulation on 32-bit x86 hosts (since 8.0)
178''''''''''''''''''''''''''''''''''''''''''''''''
179
180Support for 32-bit x86 host deployments is increasingly uncommon in mainstream
181OS distributions given the widespread availability of 64-bit x86 hardware.
182The QEMU project no longer considers 32-bit x86 support for system emulation to
183be an effective use of its limited resources, and thus intends to discontinue
184it. Since all recent x86 hardware from the past >10 years is capable of the
18564-bit x86 extensions, a corresponding 64-bit OS should be used instead.
186
187
188System emulator CPUs
189--------------------
190
191``power5+`` and ``power7+`` CPU names (since 9.0)
192'''''''''''''''''''''''''''''''''''''''''''''''''
193
194The character "+" in device (and thus also CPU) names is not allowed
195in the QEMU object model anymore. ``power5+``, ``power5+_v2.1``,
196``power7+`` and ``power7+_v2.1`` are currently still supported via
197an alias, but for consistency these will get removed in a future
198release, too. Use ``power5p_v2.1`` and ``power7p_v2.1`` instead.
199
200``Sun-UltraSparc-IIIi+`` and ``Sun-UltraSparc-IV+`` CPU names (since 9.1)
201'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
202
203The character "+" in device (and thus also CPU) names is not allowed
204in the QEMU object model anymore. ``Sun-UltraSparc-IIIi+`` and
205``Sun-UltraSparc-IV+`` are currently still supported via a workaround,
206but for consistency these will get removed in a future release, too.
207Use ``Sun-UltraSparc-IIIi-plus`` and ``Sun-UltraSparc-IV-plus`` instead.
208
209CRIS CPU architecture (since 9.0)
210'''''''''''''''''''''''''''''''''
211
212The CRIS architecture was pulled from Linux in 4.17 and the compiler
213is no longer packaged in any distro making it harder to run the
214``check-tcg`` tests. Unless we can improve the testing situation there
215is a chance the code will bitrot without anyone noticing.
216
217System emulator machines
218------------------------
219
220Arm ``virt`` machine ``dtb-kaslr-seed`` property (since 7.1)
221''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
222
223The ``dtb-kaslr-seed`` property on the ``virt`` board has been
224deprecated; use the new name ``dtb-randomness`` instead. The new name
225better reflects the way this property affects all random data within
226the device tree blob, not just the ``kaslr-seed`` node.
227
228``pc-i440fx-2.4`` up to ``pc-i440fx-2.12`` (since 9.1)
229''''''''''''''''''''''''''''''''''''''''''''''''''''''
230
231These old machine types are quite neglected nowadays and thus might have
232various pitfalls with regards to live migration. Use a newer machine type
233instead.
234
235``shix`` (since 9.0)
236''''''''''''''''''''
237
238The machine is no longer in existence and has been long unmaintained
239in QEMU. This also holds for the TC51828 16MiB flash that it uses.
240
241``pseries-2.1`` up to ``pseries-2.12`` (since 9.0)
242''''''''''''''''''''''''''''''''''''''''''''''''''
243
244Older pseries machines before version 3.0 have undergone many changes
245to correct issues, mostly regarding migration compatibility. These are
246no longer maintained and removing them will make the code easier to
247read and maintain. Use versions 3.0 and above as a replacement.
248
249Arm machines ``akita``, ``borzoi``, ``cheetah``, ``connex``, ``mainstone``, ``n800``, ``n810``, ``spitz``, ``terrier``, ``tosa``, ``verdex``, ``z2`` (since 9.0)
250''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
251
252QEMU includes models of some machine types where the QEMU code that
253emulates their SoCs is very old and unmaintained. This code is now
254blocking our ability to move forward with various changes across
255the codebase, and over many years nobody has been interested in
256trying to modernise it. We don't expect any of these machines to have
257a large number of users, because they're all modelling hardware that
258has now passed away into history. We are therefore dropping support
259for all machine types using the PXA2xx and OMAP2 SoCs. We are also
260dropping the ``cheetah`` OMAP1 board, because we don't have any
261test images for it and don't know of anybody who does; the ``sx1``
262and ``sx1-v1`` OMAP1 machines remain supported for now.
263
264PPC 405 ``ref405ep`` machine (since 9.1)
265''''''''''''''''''''''''''''''''''''''''
266
267The ``ref405ep`` machine and PPC 405 CPU have no known users, firmware
268images are not available, OpenWRT dropped support in 2019, U-Boot in
2692017, Linux also is dropping support in 2024. It is time to let go of
270this ancient hardware and focus on newer CPUs and platforms.
271
272Arm ``tacoma-bmc`` machine (since 9.1)
273''''''''''''''''''''''''''''''''''''''''
274
275The ``tacoma-bmc`` machine was a board including an AST2600 SoC based
276BMC and a witherspoon like OpenPOWER system. It was used for bring up
277of the AST2600 SoC in labs.  It can be easily replaced by the
278``rainier-bmc`` machine which is a real product.
279
280Backend options
281---------------
282
283Using non-persistent backing file with pmem=on (since 6.1)
284''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
285
286This option is used when ``memory-backend-file`` is consumed by emulated NVDIMM
287device. However enabling ``memory-backend-file.pmem`` option, when backing file
288is (a) not DAX capable or (b) not on a filesystem that support direct mapping
289of persistent memory, is not safe and may lead to data loss or corruption in case
290of host crash.
291Options are:
292
293    - modify VM configuration to set ``pmem=off`` to continue using fake NVDIMM
294      (without persistence guaranties) with backing file on non DAX storage
295    - move backing file to NVDIMM storage and keep ``pmem=on``
296      (to have NVDIMM with persistence guaranties).
297
298Device options
299--------------
300
301Emulated device options
302'''''''''''''''''''''''
303
304``-device nvme-ns,eui64-default=on|off`` (since 7.1)
305^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
306
307In QEMU versions 6.1, 6.2 and 7.0, the ``nvme-ns`` generates an EUI-64
308identifier that is not globally unique. If an EUI-64 identifier is required, the
309user must set it explicitly using the ``nvme-ns`` device parameter ``eui64``.
310
311``-device nvme,use-intel-id=on|off`` (since 7.1)
312^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
313
314The ``nvme`` device originally used a PCI Vendor/Device Identifier combination
315from Intel that was not properly allocated. Since version 5.2, the controller
316has used a properly allocated identifier. Deprecate the ``use-intel-id``
317machine compatibility parameter.
318
319``-device cxl-type3,memdev=xxxx`` (since 8.0)
320^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
321
322The ``cxl-type3`` device initially only used a single memory backend.  With
323the addition of volatile memory support, it is now necessary to distinguish
324between persistent and volatile memory backends.  As such, memdev is deprecated
325in favor of persistent-memdev.
326
327``-fsdev proxy`` and ``-virtfs proxy`` (since 8.1)
328^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
329
330The 9p ``proxy`` filesystem backend driver has been deprecated and will be
331removed (along with its proxy helper daemon) in a future version of QEMU. Please
332use ``-fsdev local`` or ``-virtfs local`` for using the 9p ``local`` filesystem
333backend, or alternatively consider deploying virtiofsd instead.
334
335The 9p ``proxy`` backend was originally developed as an alternative to the 9p
336``local`` backend. The idea was to enhance security by dispatching actual low
337level filesystem operations from 9p server (QEMU process) over to a separate
338process (the virtfs-proxy-helper binary). However this alternative never gained
339momentum. The proxy backend is much slower than the local backend, hasn't seen
340any development in years, and showed to be less secure, especially due to the
341fact that its helper daemon must be run as root, whereas with the local backend
342QEMU is typically run as unprivileged user and allows to tighten behaviour by
343mapping permissions et al by using its 'mapped' security model option.
344
345Nowadays it would make sense to reimplement the ``proxy`` backend by using
346QEMU's ``vhost`` feature, which would eliminate the high latency costs under
347which the 9p ``proxy`` backend currently suffers. However as of to date nobody
348has indicated plans for such kind of reimplementation unfortunately.
349
350RISC-V 'any' CPU type ``-cpu any`` (since 8.2)
351^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
352
353The 'any' CPU type was introduced back in 2018 and has been around since the
354initial RISC-V QEMU port. Its usage has always been unclear: users don't know
355what to expect from a CPU called 'any', and in fact the CPU does not do anything
356special that isn't already done by the default CPUs rv32/rv64.
357
358After the introduction of the 'max' CPU type, RISC-V now has a good coverage
359of generic CPUs: rv32 and rv64 as default CPUs and 'max' as a feature complete
360CPU for both 32 and 64 bit builds. Users are then discouraged to use the 'any'
361CPU type starting in 8.2.
362
363RISC-V CPU properties which start with capital 'Z' (since 8.2)
364^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
365
366All RISC-V CPU properties which start with capital 'Z' are being deprecated
367starting in 8.2. The reason is that they were wrongly added with capital 'Z'
368in the past. CPU properties were later added with lower-case names, which
369is the format we want to use from now on.
370
371Users which try to use these deprecated properties will receive a warning
372recommending to switch to their stable counterparts:
373
374- "Zifencei" should be replaced with "zifencei"
375- "Zicsr" should be replaced with "zicsr"
376- "Zihintntl" should be replaced with "zihintntl"
377- "Zihintpause" should be replaced with "zihintpause"
378- "Zawrs" should be replaced with "zawrs"
379- "Zfa" should be replaced with "zfa"
380- "Zfh" should be replaced with "zfh"
381- "Zfhmin" should be replaced with "zfhmin"
382- "Zve32f" should be replaced with "zve32f"
383- "Zve64f" should be replaced with "zve64f"
384- "Zve64d" should be replaced with "zve64d"
385
386``-device sd-card,spec_version=1`` (since 9.1)
387^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
388
389SD physical layer specification v2.00 supersedes the v1.10 one.
390v2.00 is the default since QEMU 3.0.0.
391
392Block device options
393''''''''''''''''''''
394
395``"backing": ""`` (since 2.12)
396^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
397
398In order to prevent QEMU from automatically opening an image's backing
399chain, use ``"backing": null`` instead.
400
401``rbd`` keyvalue pair encoded filenames: ``""`` (since 3.1)
402^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
403
404Options for ``rbd`` should be specified according to its runtime options,
405like other block drivers.  Legacy parsing of keyvalue pair encoded
406filenames is useful to open images with the old format for backing files;
407These image files should be updated to use the current format.
408
409Example of legacy encoding::
410
411  json:{"file.driver":"rbd", "file.filename":"rbd:rbd/name"}
412
413The above, converted to the current supported format::
414
415  json:{"file.driver":"rbd", "file.pool":"rbd", "file.image":"name"}
416
417``iscsi,password=xxx`` (since 8.0)
418^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
419
420Specifying the iSCSI password in plain text on the command line using the
421``password`` option is insecure. The ``password-secret`` option should be
422used instead, to refer to a ``--object secret...`` instance that provides
423a password via a file, or encrypted.
424
425Character device options
426''''''''''''''''''''''''
427
428Backend ``memory`` (since 9.0)
429^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
430
431``memory`` is a deprecated synonym for ``ringbuf``.
432
433CPU device properties
434'''''''''''''''''''''
435
436``pcommit`` on x86 (since 9.1)
437^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
438
439The PCOMMIT instruction was never included in any physical processor.
440It was implemented as a no-op instruction in TCG up to QEMU 9.0, but
441only with ``-cpu max`` (which does not guarantee migration compatibility
442across versions).
443
444``pmu-num=n`` on RISC-V CPUs (since 8.2)
445^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
446
447In order to support more flexible counter configurations this has been replaced
448by a ``pmu-mask`` property. If set of counters is continuous then the mask can
449be calculated with ``((2 ^ n) - 1) << 3``. The least significant three bits
450must be left clear.
451
452
453Backwards compatibility
454-----------------------
455
456Runnability guarantee of CPU models (since 4.1)
457'''''''''''''''''''''''''''''''''''''''''''''''
458
459Previous versions of QEMU never changed existing CPU models in
460ways that introduced additional host software or hardware
461requirements to the VM.  This allowed management software to
462safely change the machine type of an existing VM without
463introducing new requirements ("runnability guarantee").  This
464prevented CPU models from being updated to include CPU
465vulnerability mitigations, leaving guests vulnerable in the
466default configuration.
467
468The CPU model runnability guarantee won't apply anymore to
469existing CPU models.  Management software that needs runnability
470guarantees must resolve the CPU model aliases using the
471``alias-of`` field returned by the ``query-cpu-definitions`` QMP
472command.
473
474While those guarantees are kept, the return value of
475``query-cpu-definitions`` will have existing CPU model aliases
476point to a version that doesn't break runnability guarantees
477(specifically, version 1 of those CPU models).  In future QEMU
478versions, aliases will point to newer CPU model versions
479depending on the machine type, so management software must
480resolve CPU model aliases before starting a virtual machine.
481
482RISC-V "virt" board "riscv,delegate" DT property (since 9.1)
483''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
484
485The "riscv,delegate" DT property was added in QEMU 7.0 as part of
486the AIA APLIC support.  The property changed name during the review
487process in Linux and the correct name ended up being
488"riscv,delegation".  Changing the DT property name will break all
489available firmwares that are using the current (wrong) name.  The
490property is kept as is in 9.1, together with "riscv,delegation", to
491give more time for firmware developers to change their code.
492
493Migration
494---------
495
496``fd:`` URI when used for file migration (since 9.1)
497''''''''''''''''''''''''''''''''''''''''''''''''''''
498
499The ``fd:`` URI can currently provide a file descriptor that
500references either a socket or a plain file. These are two different
501types of migration. In order to reduce ambiguity, the ``fd:`` URI
502usage of providing a file descriptor to a plain file has been
503deprecated in favor of explicitly using the ``file:`` URI with the
504file descriptor being passed as an ``fdset``. Refer to the ``add-fd``
505command documentation for details on the ``fdset`` usage.
506