]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/blob - Documentation/feature-removal-schedule.txt
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
[mirror_ubuntu-bionic-kernel.git] / Documentation / feature-removal-schedule.txt
1 The following is a list of files and features that are going to be
2 removed in the kernel source tree. Every entry should contain what
3 exactly is going away, why it is happening, and who is going to be doing
4 the work. When the feature is removed from the kernel, it should also
5 be removed from this file. The suggested deprecation period is 3 releases.
6
7 ---------------------------
8
9 What: ddebug_query="query" boot cmdline param
10 When: v3.8
11 Why: obsoleted by dyndbg="query" and module.dyndbg="query"
12 Who: Jim Cromie <jim.cromie@gmail.com>, Jason Baron <jbaron@redhat.com>
13
14 ---------------------------
15
16 What: CONFIG_APM_CPU_IDLE, and its ability to call APM BIOS in idle
17 When: 2012
18 Why: This optional sub-feature of APM is of dubious reliability,
19 and ancient APM laptops are likely better served by calling HLT.
20 Deleting CONFIG_APM_CPU_IDLE allows x86 to stop exporting
21 the pm_idle function pointer to modules.
22 Who: Len Brown <len.brown@intel.com>
23
24 ----------------------------
25
26 What: x86_32 "no-hlt" cmdline param
27 When: 2012
28 Why: remove a branch from idle path, simplify code used by everybody.
29 This option disabled the use of HLT in idle and machine_halt()
30 for hardware that was flakey 15-years ago. Today we have
31 "idle=poll" that removed HLT from idle, and so if such a machine
32 is still running the upstream kernel, "idle=poll" is likely sufficient.
33 Who: Len Brown <len.brown@intel.com>
34
35 ----------------------------
36
37 What: x86 "idle=mwait" cmdline param
38 When: 2012
39 Why: simplify x86 idle code
40 Who: Len Brown <len.brown@intel.com>
41
42 ----------------------------
43
44 What: PRISM54
45 When: 2.6.34
46
47 Why: prism54 FullMAC PCI / Cardbus devices used to be supported only by the
48 prism54 wireless driver. After Intersil stopped selling these
49 devices in preference for the newer more flexible SoftMAC devices
50 a SoftMAC device driver was required and prism54 did not support
51 them. The p54pci driver now exists and has been present in the kernel for
52 a while. This driver supports both SoftMAC devices and FullMAC devices.
53 The main difference between these devices was the amount of memory which
54 could be used for the firmware. The SoftMAC devices support a smaller
55 amount of memory. Because of this the SoftMAC firmware fits into FullMAC
56 devices's memory. p54pci supports not only PCI / Cardbus but also USB
57 and SPI. Since p54pci supports all devices prism54 supports
58 you will have a conflict. I'm not quite sure how distributions are
59 handling this conflict right now. prism54 was kept around due to
60 claims users may experience issues when using the SoftMAC driver.
61 Time has passed users have not reported issues. If you use prism54
62 and for whatever reason you cannot use p54pci please let us know!
63 E-mail us at: linux-wireless@vger.kernel.org
64
65 For more information see the p54 wiki page:
66
67 http://wireless.kernel.org/en/users/Drivers/p54
68
69 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
70
71 ---------------------------
72
73 What: IRQF_SAMPLE_RANDOM
74 Check: IRQF_SAMPLE_RANDOM
75 When: July 2009
76
77 Why: Many of IRQF_SAMPLE_RANDOM users are technically bogus as entropy
78 sources in the kernel's current entropy model. To resolve this, every
79 input point to the kernel's entropy pool needs to better document the
80 type of entropy source it actually is. This will be replaced with
81 additional add_*_randomness functions in drivers/char/random.c
82
83 Who: Robin Getz <rgetz@blackfin.uclinux.org> & Matt Mackall <mpm@selenic.com>
84
85 ---------------------------
86
87 What: The ieee80211_regdom module parameter
88 When: March 2010 / desktop catchup
89
90 Why: This was inherited by the CONFIG_WIRELESS_OLD_REGULATORY code,
91 and currently serves as an option for users to define an
92 ISO / IEC 3166 alpha2 code for the country they are currently
93 present in. Although there are userspace API replacements for this
94 through nl80211 distributions haven't yet caught up with implementing
95 decent alternatives through standard GUIs. Although available as an
96 option through iw or wpa_supplicant its just a matter of time before
97 distributions pick up good GUI options for this. The ideal solution
98 would actually consist of intelligent designs which would do this for
99 the user automatically even when travelling through different countries.
100 Until then we leave this module parameter as a compromise.
101
102 When userspace improves with reasonable widely-available alternatives for
103 this we will no longer need this module parameter. This entry hopes that
104 by the super-futuristically looking date of "March 2010" we will have
105 such replacements widely available.
106
107 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
108
109 ---------------------------
110
111 What: dev->power.power_state
112 When: July 2007
113 Why: Broken design for runtime control over driver power states, confusing
114 driver-internal runtime power management with: mechanisms to support
115 system-wide sleep state transitions; event codes that distinguish
116 different phases of swsusp "sleep" transitions; and userspace policy
117 inputs. This framework was never widely used, and most attempts to
118 use it were broken. Drivers should instead be exposing domain-specific
119 interfaces either to kernel or to userspace.
120 Who: Pavel Machek <pavel@ucw.cz>
121
122 ---------------------------
123
124 What: /proc/<pid>/oom_adj
125 When: August 2012
126 Why: /proc/<pid>/oom_adj allows userspace to influence the oom killer's
127 badness heuristic used to determine which task to kill when the kernel
128 is out of memory.
129
130 The badness heuristic has since been rewritten since the introduction of
131 this tunable such that its meaning is deprecated. The value was
132 implemented as a bitshift on a score generated by the badness()
133 function that did not have any precise units of measure. With the
134 rewrite, the score is given as a proportion of available memory to the
135 task allocating pages, so using a bitshift which grows the score
136 exponentially is, thus, impossible to tune with fine granularity.
137
138 A much more powerful interface, /proc/<pid>/oom_score_adj, was
139 introduced with the oom killer rewrite that allows users to increase or
140 decrease the badness score linearly. This interface will replace
141 /proc/<pid>/oom_adj.
142
143 A warning will be emitted to the kernel log if an application uses this
144 deprecated interface. After it is printed once, future warnings will be
145 suppressed until the kernel is rebooted.
146
147 ---------------------------
148
149 What: remove EXPORT_SYMBOL(kernel_thread)
150 When: August 2006
151 Files: arch/*/kernel/*_ksyms.c
152 Check: kernel_thread
153 Why: kernel_thread is a low-level implementation detail. Drivers should
154 use the <linux/kthread.h> API instead which shields them from
155 implementation details and provides a higherlevel interface that
156 prevents bugs and code duplication
157 Who: Christoph Hellwig <hch@lst.de>
158
159 ---------------------------
160
161 What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
162 (temporary transition config option provided until then)
163 The transition config option will also be removed at the same time.
164 When: before 2.6.19
165 Why: Unused symbols are both increasing the size of the kernel binary
166 and are often a sign of "wrong API"
167 Who: Arjan van de Ven <arjan@linux.intel.com>
168
169 ---------------------------
170
171 What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
172 When: October 2008
173 Why: The stacking of class devices makes these values misleading and
174 inconsistent.
175 Class devices should not carry any of these properties, and bus
176 devices have SUBSYTEM and DRIVER as a replacement.
177 Who: Kay Sievers <kay.sievers@suse.de>
178
179 ---------------------------
180
181 What: ACPI procfs interface
182 When: July 2008
183 Why: ACPI sysfs conversion should be finished by January 2008.
184 ACPI procfs interface will be removed in July 2008 so that
185 there is enough time for the user space to catch up.
186 Who: Zhang Rui <rui.zhang@intel.com>
187
188 ---------------------------
189
190 What: CONFIG_ACPI_PROCFS_POWER
191 When: 2.6.39
192 Why: sysfs I/F for ACPI power devices, including AC and Battery,
193 has been working in upstream kernel since 2.6.24, Sep 2007.
194 In 2.6.37, we make the sysfs I/F always built in and this option
195 disabled by default.
196 Remove this option and the ACPI power procfs interface in 2.6.39.
197 Who: Zhang Rui <rui.zhang@intel.com>
198
199 ---------------------------
200
201 What: /proc/acpi/event
202 When: February 2008
203 Why: /proc/acpi/event has been replaced by events via the input layer
204 and netlink since 2.6.23.
205 Who: Len Brown <len.brown@intel.com>
206
207 ---------------------------
208
209 What: i386/x86_64 bzImage symlinks
210 When: April 2010
211
212 Why: The i386/x86_64 merge provides a symlink to the old bzImage
213 location so not yet updated user space tools, e.g. package
214 scripts, do not break.
215 Who: Thomas Gleixner <tglx@linutronix.de>
216
217 ---------------------------
218
219 What: GPIO autorequest on gpio_direction_{input,output}() in gpiolib
220 When: February 2010
221 Why: All callers should use explicit gpio_request()/gpio_free().
222 The autorequest mechanism in gpiolib was provided mostly as a
223 migration aid for legacy GPIO interfaces (for SOC based GPIOs).
224 Those users have now largely migrated. Platforms implementing
225 the GPIO interfaces without using gpiolib will see no changes.
226 Who: David Brownell <dbrownell@users.sourceforge.net>
227 ---------------------------
228
229 What: b43 support for firmware revision < 410
230 When: The schedule was July 2008, but it was decided that we are going to keep the
231 code as long as there are no major maintanance headaches.
232 So it _could_ be removed _any_ time now, if it conflicts with something new.
233 Why: The support code for the old firmware hurts code readability/maintainability
234 and slightly hurts runtime performance. Bugfixes for the old firmware
235 are not provided by Broadcom anymore.
236 Who: Michael Buesch <m@bues.ch>
237
238 ---------------------------
239
240 What: Ability for non root users to shm_get hugetlb pages based on mlock
241 resource limits
242 When: 2.6.31
243 Why: Non root users need to be part of /proc/sys/vm/hugetlb_shm_group or
244 have CAP_IPC_LOCK to be able to allocate shm segments backed by
245 huge pages. The mlock based rlimit check to allow shm hugetlb is
246 inconsistent with mmap based allocations. Hence it is being
247 deprecated.
248 Who: Ravikiran Thirumalai <kiran@scalex86.org>
249
250 ---------------------------
251
252 What: sysfs ui for changing p4-clockmod parameters
253 When: September 2009
254 Why: See commits 129f8ae9b1b5be94517da76009ea956e89104ce8 and
255 e088e4c9cdb618675874becb91b2fd581ee707e6.
256 Removal is subject to fixing any remaining bugs in ACPI which may
257 cause the thermal throttling not to happen at the right time.
258 Who: Dave Jones <davej@redhat.com>, Matthew Garrett <mjg@redhat.com>
259
260 -----------------------------
261
262 What: fakephp and associated sysfs files in /sys/bus/pci/slots/
263 When: 2011
264 Why: In 2.6.27, the semantics of /sys/bus/pci/slots was redefined to
265 represent a machine's physical PCI slots. The change in semantics
266 had userspace implications, as the hotplug core no longer allowed
267 drivers to create multiple sysfs files per physical slot (required
268 for multi-function devices, e.g.). fakephp was seen as a developer's
269 tool only, and its interface changed. Too late, we learned that
270 there were some users of the fakephp interface.
271
272 In 2.6.30, the original fakephp interface was restored. At the same
273 time, the PCI core gained the ability that fakephp provided, namely
274 function-level hot-remove and hot-add.
275
276 Since the PCI core now provides the same functionality, exposed in:
277
278 /sys/bus/pci/rescan
279 /sys/bus/pci/devices/.../remove
280 /sys/bus/pci/devices/.../rescan
281
282 there is no functional reason to maintain fakephp as well.
283
284 We will keep the existing module so that 'modprobe fakephp' will
285 present the old /sys/bus/pci/slots/... interface for compatibility,
286 but users are urged to migrate their applications to the API above.
287
288 After a reasonable transition period, we will remove the legacy
289 fakephp interface.
290 Who: Alex Chiang <achiang@hp.com>
291
292 ---------------------------
293
294 What: CONFIG_RFKILL_INPUT
295 When: 2.6.33
296 Why: Should be implemented in userspace, policy daemon.
297 Who: Johannes Berg <johannes@sipsolutions.net>
298
299 ----------------------------
300
301 What: sound-slot/service-* module aliases and related clutters in
302 sound/sound_core.c
303 When: August 2010
304 Why: OSS sound_core grabs all legacy minors (0-255) of SOUND_MAJOR
305 (14) and requests modules using custom sound-slot/service-*
306 module aliases. The only benefit of doing this is allowing
307 use of custom module aliases which might as well be considered
308 a bug at this point. This preemptive claiming prevents
309 alternative OSS implementations.
310
311 Till the feature is removed, the kernel will be requesting
312 both sound-slot/service-* and the standard char-major-* module
313 aliases and allow turning off the pre-claiming selectively via
314 CONFIG_SOUND_OSS_CORE_PRECLAIM and soundcore.preclaim_oss
315 kernel parameter.
316
317 After the transition phase is complete, both the custom module
318 aliases and switches to disable it will go away. This removal
319 will also allow making ALSA OSS emulation independent of
320 sound_core. The dependency will be broken then too.
321 Who: Tejun Heo <tj@kernel.org>
322
323 ----------------------------
324
325 What: sysfs-class-rfkill state file
326 When: Feb 2014
327 Files: net/rfkill/core.c
328 Why: Documented as obsolete since Feb 2010. This file is limited to 3
329 states while the rfkill drivers can have 4 states.
330 Who: anybody or Florian Mickler <florian@mickler.org>
331
332 ----------------------------
333
334 What: sysfs-class-rfkill claim file
335 When: Feb 2012
336 Files: net/rfkill/core.c
337 Why: It is not possible to claim an rfkill driver since 2007. This is
338 Documented as obsolete since Feb 2010.
339 Who: anybody or Florian Mickler <florian@mickler.org>
340
341 ----------------------------
342
343 What: iwlwifi 50XX module parameters
344 When: 3.0
345 Why: The "..50" modules parameters were used to configure 5000 series and
346 up devices; different set of module parameters also available for 4965
347 with same functionalities. Consolidate both set into single place
348 in drivers/net/wireless/iwlwifi/iwl-agn.c
349
350 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
351
352 ----------------------------
353
354 What: iwl4965 alias support
355 When: 3.0
356 Why: Internal alias support has been present in module-init-tools for some
357 time, the MODULE_ALIAS("iwl4965") boilerplate aliases can be removed
358 with no impact.
359
360 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
361
362 ---------------------------
363
364 What: xt_NOTRACK
365 Files: net/netfilter/xt_NOTRACK.c
366 When: April 2011
367 Why: Superseded by xt_CT
368 Who: Netfilter developer team <netfilter-devel@vger.kernel.org>
369
370 ----------------------------
371
372 What: IRQF_DISABLED
373 When: 2.6.36
374 Why: The flag is a NOOP as we run interrupt handlers with interrupts disabled
375 Who: Thomas Gleixner <tglx@linutronix.de>
376
377 ----------------------------
378
379 What: PCI DMA unmap state API
380 When: August 2012
381 Why: PCI DMA unmap state API (include/linux/pci-dma.h) was replaced
382 with DMA unmap state API (DMA unmap state API can be used for
383 any bus).
384 Who: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
385
386 ----------------------------
387
388 What: iwlwifi disable_hw_scan module parameters
389 When: 3.0
390 Why: Hareware scan is the prefer method for iwlwifi devices for
391 scanning operation. Remove software scan support for all the
392 iwlwifi devices.
393
394 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
395
396 ----------------------------
397
398 What: Legacy, non-standard chassis intrusion detection interface.
399 When: June 2011
400 Why: The adm9240, w83792d and w83793 hardware monitoring drivers have
401 legacy interfaces for chassis intrusion detection. A standard
402 interface has been added to each driver, so the legacy interface
403 can be removed.
404 Who: Jean Delvare <khali@linux-fr.org>
405
406 ----------------------------
407
408 What: i2c_driver.attach_adapter
409 i2c_driver.detach_adapter
410 When: September 2011
411 Why: These legacy callbacks should no longer be used as i2c-core offers
412 a variety of preferable alternative ways to instantiate I2C devices.
413 Who: Jean Delvare <khali@linux-fr.org>
414
415 ----------------------------
416
417 What: Opening a radio device node will no longer automatically switch the
418 tuner mode from tv to radio.
419 When: 3.3
420 Why: Just opening a V4L device should not change the state of the hardware
421 like that. It's very unexpected and against the V4L spec. Instead, you
422 switch to radio mode by calling VIDIOC_S_FREQUENCY. This is the second
423 and last step of the move to consistent handling of tv and radio tuners.
424 Who: Hans Verkuil <hans.verkuil@cisco.com>
425
426 ----------------------------
427
428 What: CONFIG_CFG80211_WEXT
429 When: as soon as distributions ship new wireless tools, ie. wpa_supplicant 1.0
430 and NetworkManager/connman/etc. that are able to use nl80211
431 Why: Wireless extensions are deprecated, and userland tools are moving to
432 using nl80211. New drivers are no longer using wireless extensions,
433 and while there might still be old drivers, both new drivers and new
434 userland no longer needs them and they can't be used for an feature
435 developed in the past couple of years. As such, compatibility with
436 wireless extensions in new drivers will be removed.
437 Who: Johannes Berg <johannes@sipsolutions.net>
438
439 ----------------------------
440
441 What: g_file_storage driver
442 When: 3.8
443 Why: This driver has been superseded by g_mass_storage.
444 Who: Alan Stern <stern@rowland.harvard.edu>
445
446 ----------------------------
447
448 What: threeg and interface sysfs files in /sys/devices/platform/acer-wmi
449 When: 2012
450 Why: In 3.0, we can now autodetect internal 3G device and already have
451 the threeg rfkill device. So, we plan to remove threeg sysfs support
452 for it's no longer necessary.
453
454 We also plan to remove interface sysfs file that exposed which ACPI-WMI
455 interface that was used by acer-wmi driver. It will replaced by
456 information log when acer-wmi initial.
457 Who: Lee, Chun-Yi <jlee@novell.com>
458
459 ---------------------------
460
461 What: /sys/devices/platform/_UDC_/udc/_UDC_/is_dualspeed file and
462 is_dualspeed line in /sys/devices/platform/ci13xxx_*/udc/device file.
463 When: 3.8
464 Why: The is_dualspeed file is superseded by maximum_speed in the same
465 directory and is_dualspeed line in device file is superseded by
466 max_speed line in the same file.
467
468 The maximum_speed/max_speed specifies maximum speed supported by UDC.
469 To check if dualspeeed is supported, check if the value is >= 3.
470 Various possible speeds are defined in <linux/usb/ch9.h>.
471 Who: Michal Nazarewicz <mina86@mina86.com>
472
473 ----------------------------
474
475 What: The XFS nodelaylog mount option
476 When: 3.3
477 Why: The delaylog mode that has been the default since 2.6.39 has proven
478 stable, and the old code is in the way of additional improvements in
479 the log code.
480 Who: Christoph Hellwig <hch@lst.de>
481
482 ----------------------------
483
484 What: iwlagn alias support
485 When: 3.5
486 Why: The iwlagn module has been renamed iwlwifi. The alias will be around
487 for backward compatibility for several cycles and then dropped.
488 Who: Don Fry <donald.h.fry@intel.com>
489
490 ----------------------------
491
492 What: pci_scan_bus_parented()
493 When: 3.5
494 Why: The pci_scan_bus_parented() interface creates a new root bus. The
495 bus is created with default resources (ioport_resource and
496 iomem_resource) that are always wrong, so we rely on arch code to
497 correct them later. Callers of pci_scan_bus_parented() should
498 convert to using pci_scan_root_bus() so they can supply a list of
499 bus resources when the bus is created.
500 Who: Bjorn Helgaas <bhelgaas@google.com>
501
502 ----------------------------
503
504 What: Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
505 When: 3.6
506 Why: This driver provides support for USB storage devices like "USB
507 sticks". As of now, it is deactivated in Debian, Fedora and
508 Ubuntu. All current users can switch over to usb-storage
509 (CONFIG_USB_STORAGE) which only drawback is the additional SCSI
510 stack.
511 Who: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
512
513 ----------------------------
514
515 What: kmap_atomic(page, km_type)
516 When: 3.5
517 Why: The old kmap_atomic() with two arguments is deprecated, we only
518 keep it for backward compatibility for few cycles and then drop it.
519 Who: Cong Wang <amwang@redhat.com>
520
521 ----------------------------
522
523 What: get_robust_list syscall
524 When: 2013
525 Why: There appear to be no production users of the get_robust_list syscall,
526 and it runs the risk of leaking address locations, allowing the bypass
527 of ASLR. It was only ever intended for debugging, so it should be
528 removed.
529 Who: Kees Cook <keescook@chromium.org>
530
531 ----------------------------
532
533 What: Removing the pn544 raw driver.
534 When: 3.6
535 Why: With the introduction of the NFC HCI and SHDL kernel layers, pn544.c
536 is being replaced by pn544_hci.c which is accessible through the netlink
537 and socket NFC APIs. Moreover, pn544.c is outdated and does not seem to
538 work properly with the latest Android stacks.
539 Having 2 drivers for the same hardware is confusing and as such we
540 should only keep the one following the kernel NFC APIs.
541 Who: Samuel Ortiz <sameo@linux.intel.com>
542
543 ----------------------------
544
545 What: setitimer accepts user NULL pointer (value)
546 When: 3.6
547 Why: setitimer is not returning -EFAULT if user pointer is NULL. This
548 violates the spec.
549 Who: Sasikantha Babu <sasikanth.v19@gmail.com>
550
551 ----------------------------
552
553 What: remove bogus DV presets V4L2_DV_1080I29_97, V4L2_DV_1080I30 and
554 V4L2_DV_1080I25
555 When: 3.6
556 Why: These HDTV formats do not exist and were added by a confused mind
557 (that was me, to be precise...)
558 Who: Hans Verkuil <hans.verkuil@cisco.com>
559
560 ----------------------------
561
562 What: V4L2_CID_HCENTER, V4L2_CID_VCENTER V4L2 controls
563 When: 3.7
564 Why: The V4L2_CID_VCENTER, V4L2_CID_HCENTER controls have been deprecated
565 for about 4 years and they are not used by any mainline driver.
566 There are newer controls (V4L2_CID_PAN*, V4L2_CID_TILT*) that provide
567 similar functionality.
568 Who: Sylwester Nawrocki <sylvester.nawrocki@gmail.com>
569
570 ----------------------------
571
572 What: cgroup option updates via remount
573 When: March 2013
574 Why: Remount currently allows changing bound subsystems and
575 release_agent. Rebinding is hardly useful as it only works
576 when the hierarchy is empty and release_agent itself should be
577 replaced with conventional fsnotify.
578
579 ----------------------------
580
581 What: xt_recent rev 0
582 When: 2013
583 Who: Pablo Neira Ayuso <pablo@netfilter.org>
584 Files: net/netfilter/xt_recent.c
585
586 ----------------------------
587
588 What: KVM debugfs statistics
589 When: 2013
590 Why: KVM tracepoints provide mostly equivalent information in a much more
591 flexible fashion.
592
593 ----------------------------
594
595 What: at91-mci driver ("CONFIG_MMC_AT91")
596 When: 3.7
597 Why: There are two mci drivers: at91-mci and atmel-mci. The PDC support
598 was added to atmel-mci as a first step to support more chips.
599 Then at91-mci was kept only for old IP versions (on at91rm9200 and
600 at91sam9261). The support of these IP versions has just been added
601 to atmel-mci, so atmel-mci can be used for all chips.
602 Who: Ludovic Desroches <ludovic.desroches@atmel.com>
603
604 ----------------------------
605
606 What: net/wanrouter/
607 When: June 2013
608 Why: Unsupported/unmaintained/unused since 2.6
609
610 ----------------------------