xref: /openbmc/qemu/docs/about/deprecated.rst (revision 33a24910)
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
14Prior to the 2.10.0 release there was no official policy on how
15long features would be deprecated prior to their removal, nor
16any documented list of which features were deprecated. Thus
17any features deprecated prior to 2.10.0 will be treated as if
18they were first deprecated in the 2.10.0 release.
19
20What follows is a list of all features currently marked as
21deprecated.
22
23System emulator command line arguments
24--------------------------------------
25
26Short-form boolean options (since 6.0)
27''''''''''''''''''''''''''''''''''''''
28
29Boolean options such as ``share=on``/``share=off`` could be written
30in short form as ``share`` and ``noshare``.  This is now deprecated
31and will cause a warning.
32
33``delay`` option for socket character devices (since 6.0)
34'''''''''''''''''''''''''''''''''''''''''''''''''''''''''
35
36The replacement for the ``nodelay`` short-form boolean option is ``nodelay=on``
37rather than ``delay=off``.
38
39``-smp`` ("parameter=0" SMP configurations) (since 6.2)
40'''''''''''''''''''''''''''''''''''''''''''''''''''''''
41
42Specified CPU topology parameters must be greater than zero.
43
44In the SMP configuration, users should either provide a CPU topology
45parameter with a reasonable value (greater than zero) or just omit it
46and QEMU will compute the missing value.
47
48However, historically it was implicitly allowed for users to provide
49a parameter with zero value, which is meaningless and could also possibly
50cause unexpected results in the -smp parsing. So support for this kind of
51configurations (e.g. -smp 8,sockets=0) is deprecated since 6.2 and will
52be removed in the near future, users have to ensure that all the topology
53members described with -smp are greater than zero.
54
55Plugin argument passing through ``arg=<string>`` (since 6.1)
56''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
57
58Passing TCG plugins arguments through ``arg=`` is redundant is makes the
59command-line less readable, especially when the argument itself consist of a
60name and a value, e.g. ``-plugin plugin_name,arg="arg_name=arg_value"``.
61Therefore, the usage of ``arg`` is redundant. Single-word arguments are treated
62as short-form boolean values, and passed to plugins as ``arg_name=on``.
63However, short-form booleans are deprecated and full explicit ``arg_name=on``
64form is preferred.
65
66QEMU Machine Protocol (QMP) commands
67------------------------------------
68
69``blockdev-open-tray``, ``blockdev-close-tray`` argument ``device`` (since 2.8)
70'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
71
72Use argument ``id`` instead.
73
74``eject`` argument ``device`` (since 2.8)
75'''''''''''''''''''''''''''''''''''''''''
76
77Use argument ``id`` instead.
78
79``blockdev-change-medium`` argument ``device`` (since 2.8)
80''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
81
82Use argument ``id`` instead.
83
84``block_set_io_throttle`` argument ``device`` (since 2.8)
85'''''''''''''''''''''''''''''''''''''''''''''''''''''''''
86
87Use argument ``id`` instead.
88
89``blockdev-add`` empty string argument ``backing`` (since 2.10)
90'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
91
92Use argument value ``null`` instead.
93
94``block-commit`` arguments ``base`` and ``top`` (since 3.1)
95'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
96
97Use arguments ``base-node`` and ``top-node`` instead.
98
99``nbd-server-add`` and ``nbd-server-remove`` (since 5.2)
100''''''''''''''''''''''''''''''''''''''''''''''''''''''''
101
102Use the more generic commands ``block-export-add`` and ``block-export-del``
103instead.  As part of this deprecation, where ``nbd-server-add`` used a
104single ``bitmap``, the new ``block-export-add`` uses a list of ``bitmaps``.
105
106``query-qmp-schema`` return value member ``values`` (since 6.2)
107'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
108
109Member ``values`` in return value elements with meta-type ``enum`` is
110deprecated.  Use ``members`` instead.
111
112``drive-backup`` (since 6.2)
113''''''''''''''''''''''''''''
114
115Use ``blockdev-backup`` in combination with ``blockdev-add`` instead.
116This change primarily separates the creation/opening process of the backup
117target with explicit, separate steps. ``blockdev-backup`` uses mostly the
118same arguments as ``drive-backup``, except the ``format`` and ``mode``
119options are removed in favor of using explicit ``blockdev-create`` and
120``blockdev-add`` calls. See :doc:`/interop/live-block-operations` for
121details.
122
123Incorrectly typed ``device_add`` arguments (since 6.2)
124''''''''''''''''''''''''''''''''''''''''''''''''''''''
125
126Due to shortcomings in the internal implementation of ``device_add``, QEMU
127incorrectly accepts certain invalid arguments: Any object or list arguments are
128silently ignored. Other argument types are not checked, but an implicit
129conversion happens, so that e.g. string values can be assigned to integer
130device properties or vice versa.
131
132This is a bug in QEMU that will be fixed in the future so that previously
133accepted incorrect commands will return an error. Users should make sure that
134all arguments passed to ``device_add`` are consistent with the documented
135property types.
136
137QEMU Machine Protocol (QMP) events
138----------------------------------
139
140``MEM_UNPLUG_ERROR`` (since 6.2)
141''''''''''''''''''''''''''''''''''''''''''''''''''''''''
142
143Use the more generic event ``DEVICE_UNPLUG_GUEST_ERROR`` instead.
144
145``vcpu`` trace events (since 8.1)
146'''''''''''''''''''''''''''''''''
147
148The ability to instrument QEMU helper functions with vCPU-aware trace
149points was removed in 7.0. However QMP still exposed the vcpu
150parameter. This argument has now been deprecated and the remaining
151remaining trace points that used it are selected just by name.
152
153Host Architectures
154------------------
155
156BE MIPS (since 7.2)
157'''''''''''''''''''
158
159As Debian 10 ("Buster") moved into LTS the big endian 32 bit version of
160MIPS moved out of support making it hard to maintain our
161cross-compilation CI tests of the architecture. As we no longer have
162CI coverage support may bitrot away before the deprecation process
163completes. The little endian variants of MIPS (both 32 and 64 bit) are
164still a supported host architecture.
165
166System emulation on 32-bit x86 hosts (since 8.0)
167''''''''''''''''''''''''''''''''''''''''''''''''
168
169Support for 32-bit x86 host deployments is increasingly uncommon in mainstream
170OS distributions given the widespread availability of 64-bit x86 hardware.
171The QEMU project no longer considers 32-bit x86 support for system emulation to
172be an effective use of its limited resources, and thus intends to discontinue
173it. Since all recent x86 hardware from the past >10 years is capable of the
17464-bit x86 extensions, a corresponding 64-bit OS should be used instead.
175
176
177System emulator CPUs
178--------------------
179
180Nios II CPU (since 8.2)
181'''''''''''''''''''''''
182
183The Nios II architecture is orphan. The ``nios2`` guest CPU support is
184deprecated and will be removed in a future version of QEMU.
185
186``power5+`` and ``power7+`` CPU names (since 9.0)
187'''''''''''''''''''''''''''''''''''''''''''''''''
188
189The character "+" in device (and thus also CPU) names is not allowed
190in the QEMU object model anymore. ``power5+``, ``power5+_v2.1``,
191``power7+`` and ``power7+_v2.1`` are currently still supported via
192an alias, but for consistency these will get removed in a future
193release, too. Use ``power5p_v2.1`` and ``power7p_v2.1`` instead.
194
195CRIS CPU architecture (since 9.0)
196'''''''''''''''''''''''''''''''''
197
198The CRIS architecture was pulled from Linux in 4.17 and the compiler
199is no longer packaged in any distro making it harder to run the
200``check-tcg`` tests. Unless we can improve the testing situation there
201is a chance the code will bitrot without anyone noticing.
202
203System emulator machines
204------------------------
205
206Arm ``virt`` machine ``dtb-kaslr-seed`` property (since 7.1)
207''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
208
209The ``dtb-kaslr-seed`` property on the ``virt`` board has been
210deprecated; use the new name ``dtb-randomness`` instead. The new name
211better reflects the way this property affects all random data within
212the device tree blob, not just the ``kaslr-seed`` node.
213
214``pc-i440fx-2.0`` up to ``pc-i440fx-2.3`` (since 8.2)
215'''''''''''''''''''''''''''''''''''''''''''''''''''''
216
217These old machine types are quite neglected nowadays and thus might have
218various pitfalls with regards to live migration. Use a newer machine type
219instead.
220
221Nios II ``10m50-ghrd`` and ``nios2-generic-nommu`` machines (since 8.2)
222'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
223
224The Nios II architecture is orphan.
225
226``shix`` (since 9.0)
227''''''''''''''''''''
228
229The machine is no longer in existence and has been long unmaintained
230in QEMU. This also holds for the TC51828 16MiB flash that it uses.
231
232``pseries-2.1`` up to ``pseries-2.11`` (since 9.0)
233''''''''''''''''''''''''''''''''''''''''''''''''''
234
235Older pseries machines before version 2.12 have undergone many changes
236to correct issues, mostly regarding migration compatibility. These are
237no longer maintained and removing them will make the code easier to
238read and maintain. Use versions 2.12 and above as a replacement.
239
240Backend options
241---------------
242
243Using non-persistent backing file with pmem=on (since 6.1)
244''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
245
246This option is used when ``memory-backend-file`` is consumed by emulated NVDIMM
247device. However enabling ``memory-backend-file.pmem`` option, when backing file
248is (a) not DAX capable or (b) not on a filesystem that support direct mapping
249of persistent memory, is not safe and may lead to data loss or corruption in case
250of host crash.
251Options are:
252
253    - modify VM configuration to set ``pmem=off`` to continue using fake NVDIMM
254      (without persistence guaranties) with backing file on non DAX storage
255    - move backing file to NVDIMM storage and keep ``pmem=on``
256      (to have NVDIMM with persistence guaranties).
257
258Device options
259--------------
260
261Emulated device options
262'''''''''''''''''''''''
263
264``-device virtio-blk,scsi=on|off`` (since 5.0)
265^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
266
267The virtio-blk SCSI passthrough feature is a legacy VIRTIO feature.  VIRTIO 1.0
268and later do not support it because the virtio-scsi device was introduced for
269full SCSI support.  Use virtio-scsi instead when SCSI passthrough is required.
270
271Note this also applies to ``-device virtio-blk-pci,scsi=on|off``, which is an
272alias.
273
274``-device nvme-ns,eui64-default=on|off`` (since 7.1)
275^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
276
277In QEMU versions 6.1, 6.2 and 7.0, the ``nvme-ns`` generates an EUI-64
278identifier that is not globally unique. If an EUI-64 identifier is required, the
279user must set it explicitly using the ``nvme-ns`` device parameter ``eui64``.
280
281``-device nvme,use-intel-id=on|off`` (since 7.1)
282^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
283
284The ``nvme`` device originally used a PCI Vendor/Device Identifier combination
285from Intel that was not properly allocated. Since version 5.2, the controller
286has used a properly allocated identifier. Deprecate the ``use-intel-id``
287machine compatibility parameter.
288
289``-device cxl-type3,memdev=xxxx`` (since 8.0)
290^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
291
292The ``cxl-type3`` device initially only used a single memory backend.  With
293the addition of volatile memory support, it is now necessary to distinguish
294between persistent and volatile memory backends.  As such, memdev is deprecated
295in favor of persistent-memdev.
296
297``-fsdev proxy`` and ``-virtfs proxy`` (since 8.1)
298^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
299
300The 9p ``proxy`` filesystem backend driver has been deprecated and will be
301removed (along with its proxy helper daemon) in a future version of QEMU. Please
302use ``-fsdev local`` or ``-virtfs local`` for using the 9p ``local`` filesystem
303backend, or alternatively consider deploying virtiofsd instead.
304
305The 9p ``proxy`` backend was originally developed as an alternative to the 9p
306``local`` backend. The idea was to enhance security by dispatching actual low
307level filesystem operations from 9p server (QEMU process) over to a separate
308process (the virtfs-proxy-helper binary). However this alternative never gained
309momentum. The proxy backend is much slower than the local backend, hasn't seen
310any development in years, and showed to be less secure, especially due to the
311fact that its helper daemon must be run as root, whereas with the local backend
312QEMU is typically run as unprivileged user and allows to tighten behaviour by
313mapping permissions et al by using its 'mapped' security model option.
314
315Nowadays it would make sense to reimplement the ``proxy`` backend by using
316QEMU's ``vhost`` feature, which would eliminate the high latency costs under
317which the 9p ``proxy`` backend currently suffers. However as of to date nobody
318has indicated plans for such kind of reimplementation unfortunately.
319
320RISC-V 'any' CPU type ``-cpu any`` (since 8.2)
321^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
322
323The 'any' CPU type was introduced back in 2018 and has been around since the
324initial RISC-V QEMU port. Its usage has always been unclear: users don't know
325what to expect from a CPU called 'any', and in fact the CPU does not do anything
326special that isn't already done by the default CPUs rv32/rv64.
327
328After the introduction of the 'max' CPU type, RISC-V now has a good coverage
329of generic CPUs: rv32 and rv64 as default CPUs and 'max' as a feature complete
330CPU for both 32 and 64 bit builds. Users are then discouraged to use the 'any'
331CPU type starting in 8.2.
332
333RISC-V CPU properties which start with capital 'Z' (since 8.2)
334^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
335
336All RISC-V CPU properties which start with capital 'Z' are being deprecated
337starting in 8.2. The reason is that they were wrongly added with capital 'Z'
338in the past. CPU properties were later added with lower-case names, which
339is the format we want to use from now on.
340
341Users which try to use these deprecated properties will receive a warning
342recommending to switch to their stable counterparts:
343
344- "Zifencei" should be replaced with "zifencei"
345- "Zicsr" should be replaced with "zicsr"
346- "Zihintntl" should be replaced with "zihintntl"
347- "Zihintpause" should be replaced with "zihintpause"
348- "Zawrs" should be replaced with "zawrs"
349- "Zfa" should be replaced with "zfa"
350- "Zfh" should be replaced with "zfh"
351- "Zfhmin" should be replaced with "zfhmin"
352- "Zve32f" should be replaced with "zve32f"
353- "Zve64f" should be replaced with "zve64f"
354- "Zve64d" should be replaced with "zve64d"
355
356``-device pvrdma`` and the rdma subsystem (since 8.2)
357^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
358
359The pvrdma device and the whole rdma subsystem are in a bad shape and
360without active maintenance. The QEMU project intends to remove this
361device and subsystem from the code base in a future release without
362replacement unless somebody steps up and improves the situation.
363
364
365Block device options
366''''''''''''''''''''
367
368``"backing": ""`` (since 2.12)
369^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
370
371In order to prevent QEMU from automatically opening an image's backing
372chain, use ``"backing": null`` instead.
373
374``rbd`` keyvalue pair encoded filenames: ``""`` (since 3.1)
375^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
376
377Options for ``rbd`` should be specified according to its runtime options,
378like other block drivers.  Legacy parsing of keyvalue pair encoded
379filenames is useful to open images with the old format for backing files;
380These image files should be updated to use the current format.
381
382Example of legacy encoding::
383
384  json:{"file.driver":"rbd", "file.filename":"rbd:rbd/name"}
385
386The above, converted to the current supported format::
387
388  json:{"file.driver":"rbd", "file.pool":"rbd", "file.image":"name"}
389
390``iscsi,password=xxx`` (since 8.0)
391^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
392
393Specifying the iSCSI password in plain text on the command line using the
394``password`` option is insecure. The ``password-secret`` option should be
395used instead, to refer to a ``--object secret...`` instance that provides
396a password via a file, or encrypted.
397
398Character device options
399''''''''''''''''''''''''
400
401Backend ``memory`` (since 9.0)
402^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
403
404``memory`` is a deprecated synonym for ``ringbuf``.
405
406CPU device properties
407'''''''''''''''''''''
408
409``pmu-num=n`` on RISC-V CPUs (since 8.2)
410^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
411
412In order to support more flexible counter configurations this has been replaced
413by a ``pmu-mask`` property. If set of counters is continuous then the mask can
414be calculated with ``((2 ^ n) - 1) << 3``. The least significant three bits
415must be left clear.
416
417
418Backwards compatibility
419-----------------------
420
421Runnability guarantee of CPU models (since 4.1)
422'''''''''''''''''''''''''''''''''''''''''''''''
423
424Previous versions of QEMU never changed existing CPU models in
425ways that introduced additional host software or hardware
426requirements to the VM.  This allowed management software to
427safely change the machine type of an existing VM without
428introducing new requirements ("runnability guarantee").  This
429prevented CPU models from being updated to include CPU
430vulnerability mitigations, leaving guests vulnerable in the
431default configuration.
432
433The CPU model runnability guarantee won't apply anymore to
434existing CPU models.  Management software that needs runnability
435guarantees must resolve the CPU model aliases using the
436``alias-of`` field returned by the ``query-cpu-definitions`` QMP
437command.
438
439While those guarantees are kept, the return value of
440``query-cpu-definitions`` will have existing CPU model aliases
441point to a version that doesn't break runnability guarantees
442(specifically, version 1 of those CPU models).  In future QEMU
443versions, aliases will point to newer CPU model versions
444depending on the machine type, so management software must
445resolve CPU model aliases before starting a virtual machine.
446
447QEMU guest agent
448----------------
449
450``--blacklist`` command line option (since 7.2)
451'''''''''''''''''''''''''''''''''''''''''''''''
452
453``--blacklist`` has been replaced by ``--block-rpcs`` (which is a better
454wording for what this option does). The short form ``-b`` still stays
455the same and thus is the preferred way for scripts that should run with
456both, older and future versions of QEMU.
457
458``blacklist`` config file option (since 7.2)
459''''''''''''''''''''''''''''''''''''''''''''
460
461The ``blacklist`` config file option has been renamed to ``block-rpcs``
462(to be in sync with the renaming of the corresponding command line
463option).
464
465Migration
466---------
467
468``skipped`` MigrationStats field (since 8.1)
469''''''''''''''''''''''''''''''''''''''''''''
470
471``skipped`` field in Migration stats has been deprecated.  It hasn't
472been used for more than 10 years.
473
474``inc`` migrate command option (since 8.2)
475''''''''''''''''''''''''''''''''''''''''''
476
477Use blockdev-mirror with NBD instead.
478
479As an intermediate step the ``inc`` functionality can be achieved by
480setting the ``block-incremental`` migration parameter to ``true``.
481But this parameter is also deprecated.
482
483``blk`` migrate command option (since 8.2)
484''''''''''''''''''''''''''''''''''''''''''
485
486Use blockdev-mirror with NBD instead.
487
488As an intermediate step the ``blk`` functionality can be achieved by
489setting the ``block`` migration capability to ``true``.  But this
490capability is also deprecated.
491
492block migration (since 8.2)
493'''''''''''''''''''''''''''
494
495Block migration is too inflexible.  It needs to migrate all block
496devices or none.
497
498Please see "QMP invocation for live storage migration with
499``blockdev-mirror`` + NBD" in docs/interop/live-block-operations.rst
500for a detailed explanation.
501
502old compression method (since 8.2)
503''''''''''''''''''''''''''''''''''
504
505Compression method fails too much.  Too many races.  We are going to
506remove it if nobody fixes it.  For starters, migration-test
507compression tests are disabled because they fail randomly.  If you need
508compression, use multifd compression methods.
509