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.
7 ---------------------------
9 What: x86 floppy disable_hlt
11 Why: ancient workaround of dubious utility clutters the
12 code used by everybody else.
13 Who: Len Brown <len.brown@intel.com>
15 ---------------------------
17 What: CONFIG_APM_CPU_IDLE, and its ability to call APM BIOS in idle
19 Why: This optional sub-feature of APM is of dubious reliability,
20 and ancient APM laptops are likely better served by calling HLT.
21 Deleting CONFIG_APM_CPU_IDLE allows x86 to stop exporting
22 the pm_idle function pointer to modules.
23 Who: Len Brown <len.brown@intel.com>
25 ----------------------------
27 What: x86_32 "no-hlt" cmdline param
29 Why: remove a branch from idle path, simplify code used by everybody.
30 This option disabled the use of HLT in idle and machine_halt()
31 for hardware that was flakey 15-years ago. Today we have
32 "idle=poll" that removed HLT from idle, and so if such a machine
33 is still running the upstream kernel, "idle=poll" is likely sufficient.
34 Who: Len Brown <len.brown@intel.com>
36 ----------------------------
41 Why: prism54 FullMAC PCI / Cardbus devices used to be supported only by the
42 prism54 wireless driver. After Intersil stopped selling these
43 devices in preference for the newer more flexible SoftMAC devices
44 a SoftMAC device driver was required and prism54 did not support
45 them. The p54pci driver now exists and has been present in the kernel for
46 a while. This driver supports both SoftMAC devices and FullMAC devices.
47 The main difference between these devices was the amount of memory which
48 could be used for the firmware. The SoftMAC devices support a smaller
49 amount of memory. Because of this the SoftMAC firmware fits into FullMAC
50 devices's memory. p54pci supports not only PCI / Cardbus but also USB
51 and SPI. Since p54pci supports all devices prism54 supports
52 you will have a conflict. I'm not quite sure how distributions are
53 handling this conflict right now. prism54 was kept around due to
54 claims users may experience issues when using the SoftMAC driver.
55 Time has passed users have not reported issues. If you use prism54
56 and for whatever reason you cannot use p54pci please let us know!
57 E-mail us at: linux-wireless@vger.kernel.org
59 For more information see the p54 wiki page:
61 http://wireless.kernel.org/en/users/Drivers/p54
63 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
65 ---------------------------
67 What: IRQF_SAMPLE_RANDOM
68 Check: IRQF_SAMPLE_RANDOM
71 Why: Many of IRQF_SAMPLE_RANDOM users are technically bogus as entropy
72 sources in the kernel's current entropy model. To resolve this, every
73 input point to the kernel's entropy pool needs to better document the
74 type of entropy source it actually is. This will be replaced with
75 additional add_*_randomness functions in drivers/char/random.c
77 Who: Robin Getz <rgetz@blackfin.uclinux.org> & Matt Mackall <mpm@selenic.com>
79 ---------------------------
81 What: Deprecated snapshot ioctls
84 Why: The ioctls in kernel/power/user.c were marked as deprecated long time
85 ago. Now they notify users about that so that they need to replace
86 their userspace. After some more time, remove them completely.
88 Who: Jiri Slaby <jirislaby@gmail.com>
90 ---------------------------
92 What: The ieee80211_regdom module parameter
93 When: March 2010 / desktop catchup
95 Why: This was inherited by the CONFIG_WIRELESS_OLD_REGULATORY code,
96 and currently serves as an option for users to define an
97 ISO / IEC 3166 alpha2 code for the country they are currently
98 present in. Although there are userspace API replacements for this
99 through nl80211 distributions haven't yet caught up with implementing
100 decent alternatives through standard GUIs. Although available as an
101 option through iw or wpa_supplicant its just a matter of time before
102 distributions pick up good GUI options for this. The ideal solution
103 would actually consist of intelligent designs which would do this for
104 the user automatically even when travelling through different countries.
105 Until then we leave this module parameter as a compromise.
107 When userspace improves with reasonable widely-available alternatives for
108 this we will no longer need this module parameter. This entry hopes that
109 by the super-futuristically looking date of "March 2010" we will have
110 such replacements widely available.
112 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
114 ---------------------------
116 What: dev->power.power_state
118 Why: Broken design for runtime control over driver power states, confusing
119 driver-internal runtime power management with: mechanisms to support
120 system-wide sleep state transitions; event codes that distinguish
121 different phases of swsusp "sleep" transitions; and userspace policy
122 inputs. This framework was never widely used, and most attempts to
123 use it were broken. Drivers should instead be exposing domain-specific
124 interfaces either to kernel or to userspace.
125 Who: Pavel Machek <pavel@ucw.cz>
127 ---------------------------
129 What: Video4Linux obsolete drivers using V4L1 API
131 Files: drivers/staging/se401/* drivers/staging/usbvideo/*
132 Check: drivers/staging/se401/se401.c drivers/staging/usbvideo/usbvideo.c
133 Why: There are some drivers still using V4L1 API, despite all efforts we've done
134 to migrate. Those drivers are for obsolete hardware that the old maintainer
135 didn't care (or not have the hardware anymore), and that no other developer
136 could find any hardware to buy. They probably have no practical usage today,
137 and people with such old hardware could probably keep using an older version
138 of the kernel. Those drivers will be moved to staging on 2.6.38 and, if nobody
139 cares enough to port and test them with V4L2 API, they'll be removed on 2.6.39.
140 Who: Mauro Carvalho Chehab <mchehab@infradead.org>
142 ---------------------------
144 What: Video4Linux: Remove obsolete ioctl's
146 Files: include/media/videodev2.h
147 Why: Some ioctl's were defined wrong on 2.6.2 and 2.6.6, using the wrong
148 type of R/W arguments. They were fixed, but the old ioctl names are
149 still there, maintained to avoid breaking binary compatibility:
150 #define VIDIOC_OVERLAY_OLD _IOWR('V', 14, int)
151 #define VIDIOC_S_PARM_OLD _IOW('V', 22, struct v4l2_streamparm)
152 #define VIDIOC_S_CTRL_OLD _IOW('V', 28, struct v4l2_control)
153 #define VIDIOC_G_AUDIO_OLD _IOWR('V', 33, struct v4l2_audio)
154 #define VIDIOC_G_AUDOUT_OLD _IOWR('V', 49, struct v4l2_audioout)
155 #define VIDIOC_CROPCAP_OLD _IOR('V', 58, struct v4l2_cropcap)
156 There's no sense on preserving those forever, as it is very doubtful
157 that someone would try to use a such old binary with a modern kernel.
158 Removing them will allow us to remove some magic done at the V4L ioctl
161 Who: Mauro Carvalho Chehab <mchehab@infradead.org>
163 ---------------------------
167 Option: CONFIG_SYSCTL_SYSCALL
168 Why: The same information is available in a more convenient from
169 /proc/sys, and none of the sysctl variables appear to be
170 important performance wise.
172 Binary sysctls are a long standing source of subtle kernel
173 bugs and security issues.
175 When I looked several months ago all I could find after
176 searching several distributions were 5 user space programs and
177 glibc (which falls back to /proc/sys) using this syscall.
179 The man page for sysctl(2) documents it as unusable for user
182 sysctl(2) is not generally ABI compatible to a 32bit user
183 space application on a 64bit and a 32bit kernel.
185 For the last several months the policy has been no new binary
186 sysctls and no one has put forward an argument to use them.
188 Binary sysctls issues seem to keep happening appearing so
189 properly deprecating them (with a warning to user space) and a
190 2 year grace warning period will mean eventually we can kill
191 them and end the pain.
193 In the mean time individual binary sysctls can be dealt with
194 in a piecewise fashion.
196 Who: Eric Biederman <ebiederm@xmission.com>
198 ---------------------------
200 What: /proc/<pid>/oom_adj
202 Why: /proc/<pid>/oom_adj allows userspace to influence the oom killer's
203 badness heuristic used to determine which task to kill when the kernel
206 The badness heuristic has since been rewritten since the introduction of
207 this tunable such that its meaning is deprecated. The value was
208 implemented as a bitshift on a score generated by the badness()
209 function that did not have any precise units of measure. With the
210 rewrite, the score is given as a proportion of available memory to the
211 task allocating pages, so using a bitshift which grows the score
212 exponentially is, thus, impossible to tune with fine granularity.
214 A much more powerful interface, /proc/<pid>/oom_score_adj, was
215 introduced with the oom killer rewrite that allows users to increase or
216 decrease the badness() score linearly. This interface will replace
219 A warning will be emitted to the kernel log if an application uses this
220 deprecated interface. After it is printed once, future warnings will be
221 suppressed until the kernel is rebooted.
223 ---------------------------
225 What: CS5535/CS5536 obsolete GPIO driver
227 Files: drivers/staging/cs5535_gpio/*
228 Check: drivers/staging/cs5535_gpio/cs5535_gpio.c
229 Why: A newer driver replaces this; it is drivers/gpio/cs5535-gpio.c, and
230 integrates with the Linux GPIO subsystem. The old driver has been
231 moved to staging, and will be removed altogether around 2.6.40.
232 Please test the new driver, and ensure that the functionality you
233 need and any bugfixes from the old driver are available in the new
235 Who: Andres Salomon <dilinger@queued.net>
237 --------------------------
239 What: remove EXPORT_SYMBOL(kernel_thread)
241 Files: arch/*/kernel/*_ksyms.c
243 Why: kernel_thread is a low-level implementation detail. Drivers should
244 use the <linux/kthread.h> API instead which shields them from
245 implementation details and provides a higherlevel interface that
246 prevents bugs and code duplication
247 Who: Christoph Hellwig <hch@lst.de>
249 ---------------------------
251 What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
252 (temporary transition config option provided until then)
253 The transition config option will also be removed at the same time.
255 Why: Unused symbols are both increasing the size of the kernel binary
256 and are often a sign of "wrong API"
257 Who: Arjan van de Ven <arjan@linux.intel.com>
259 ---------------------------
261 What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
263 Why: The stacking of class devices makes these values misleading and
265 Class devices should not carry any of these properties, and bus
266 devices have SUBSYTEM and DRIVER as a replacement.
267 Who: Kay Sievers <kay.sievers@suse.de>
269 ---------------------------
271 What: ACPI procfs interface
273 Why: ACPI sysfs conversion should be finished by January 2008.
274 ACPI procfs interface will be removed in July 2008 so that
275 there is enough time for the user space to catch up.
276 Who: Zhang Rui <rui.zhang@intel.com>
278 ---------------------------
280 What: CONFIG_ACPI_PROCFS_POWER
282 Why: sysfs I/F for ACPI power devices, including AC and Battery,
283 has been working in upstream kenrel since 2.6.24, Sep 2007.
284 In 2.6.37, we make the sysfs I/F always built in and this option
286 Remove this option and the ACPI power procfs interface in 2.6.39.
287 Who: Zhang Rui <rui.zhang@intel.com>
289 ---------------------------
291 What: /proc/acpi/button
293 Why: /proc/acpi/button has been replaced by events to the input layer
295 Who: Len Brown <len.brown@intel.com>
297 ---------------------------
299 What: /proc/acpi/event
301 Why: /proc/acpi/event has been replaced by events via the input layer
302 and netlink since 2.6.23.
303 Who: Len Brown <len.brown@intel.com>
305 ---------------------------
307 What: i386/x86_64 bzImage symlinks
310 Why: The i386/x86_64 merge provides a symlink to the old bzImage
311 location so not yet updated user space tools, e.g. package
312 scripts, do not break.
313 Who: Thomas Gleixner <tglx@linutronix.de>
315 ---------------------------
317 What: GPIO autorequest on gpio_direction_{input,output}() in gpiolib
319 Why: All callers should use explicit gpio_request()/gpio_free().
320 The autorequest mechanism in gpiolib was provided mostly as a
321 migration aid for legacy GPIO interfaces (for SOC based GPIOs).
322 Those users have now largely migrated. Platforms implementing
323 the GPIO interfaces without using gpiolib will see no changes.
324 Who: David Brownell <dbrownell@users.sourceforge.net>
325 ---------------------------
327 What: b43 support for firmware revision < 410
328 When: The schedule was July 2008, but it was decided that we are going to keep the
329 code as long as there are no major maintanance headaches.
330 So it _could_ be removed _any_ time now, if it conflicts with something new.
331 Why: The support code for the old firmware hurts code readability/maintainability
332 and slightly hurts runtime performance. Bugfixes for the old firmware
333 are not provided by Broadcom anymore.
334 Who: Michael Buesch <mb@bu3sch.de>
336 ---------------------------
338 What: /sys/o2cb symlink
340 Why: /sys/fs/o2cb is the proper location for this information - /sys/o2cb
341 exists as a symlink for backwards compatibility for old versions of
342 ocfs2-tools. 2 years should be sufficient time to phase in new versions
343 which know to look in /sys/fs/o2cb.
344 Who: ocfs2-devel@oss.oracle.com
346 ---------------------------
348 What: Ability for non root users to shm_get hugetlb pages based on mlock
351 Why: Non root users need to be part of /proc/sys/vm/hugetlb_shm_group or
352 have CAP_IPC_LOCK to be able to allocate shm segments backed by
353 huge pages. The mlock based rlimit check to allow shm hugetlb is
354 inconsistent with mmap based allocations. Hence it is being
356 Who: Ravikiran Thirumalai <kiran@scalex86.org>
358 ---------------------------
360 What: CONFIG_THERMAL_HWMON
362 Why: This option was introduced just to allow older lm-sensors userspace
363 to keep working over the upgrade to 2.6.26. At the scheduled time of
364 removal fixed lm-sensors (2.x or 3.x) should be readily available.
365 Who: Rene Herman <rene.herman@gmail.com>
367 ---------------------------
369 What: Code that is now under CONFIG_WIRELESS_EXT_SYSFS
370 (in net/core/net-sysfs.c)
371 When: After the only user (hal) has seen a release with the patches
372 for enough time, probably some time in 2010.
373 Why: Over 1K .text/.data size reduction, data is available in other
375 Who: Johannes Berg <johannes@sipsolutions.net>
377 ---------------------------
379 What: sysfs ui for changing p4-clockmod parameters
381 Why: See commits 129f8ae9b1b5be94517da76009ea956e89104ce8 and
382 e088e4c9cdb618675874becb91b2fd581ee707e6.
383 Removal is subject to fixing any remaining bugs in ACPI which may
384 cause the thermal throttling not to happen at the right time.
385 Who: Dave Jones <davej@redhat.com>, Matthew Garrett <mjg@redhat.com>
387 -----------------------------
389 What: fakephp and associated sysfs files in /sys/bus/pci/slots/
391 Why: In 2.6.27, the semantics of /sys/bus/pci/slots was redefined to
392 represent a machine's physical PCI slots. The change in semantics
393 had userspace implications, as the hotplug core no longer allowed
394 drivers to create multiple sysfs files per physical slot (required
395 for multi-function devices, e.g.). fakephp was seen as a developer's
396 tool only, and its interface changed. Too late, we learned that
397 there were some users of the fakephp interface.
399 In 2.6.30, the original fakephp interface was restored. At the same
400 time, the PCI core gained the ability that fakephp provided, namely
401 function-level hot-remove and hot-add.
403 Since the PCI core now provides the same functionality, exposed in:
406 /sys/bus/pci/devices/.../remove
407 /sys/bus/pci/devices/.../rescan
409 there is no functional reason to maintain fakephp as well.
411 We will keep the existing module so that 'modprobe fakephp' will
412 present the old /sys/bus/pci/slots/... interface for compatibility,
413 but users are urged to migrate their applications to the API above.
415 After a reasonable transition period, we will remove the legacy
417 Who: Alex Chiang <achiang@hp.com>
419 ---------------------------
421 What: CONFIG_RFKILL_INPUT
423 Why: Should be implemented in userspace, policy daemon.
424 Who: Johannes Berg <johannes@sipsolutions.net>
426 ----------------------------
428 What: sound-slot/service-* module aliases and related clutters in
431 Why: OSS sound_core grabs all legacy minors (0-255) of SOUND_MAJOR
432 (14) and requests modules using custom sound-slot/service-*
433 module aliases. The only benefit of doing this is allowing
434 use of custom module aliases which might as well be considered
435 a bug at this point. This preemptive claiming prevents
436 alternative OSS implementations.
438 Till the feature is removed, the kernel will be requesting
439 both sound-slot/service-* and the standard char-major-* module
440 aliases and allow turning off the pre-claiming selectively via
441 CONFIG_SOUND_OSS_CORE_PRECLAIM and soundcore.preclaim_oss
444 After the transition phase is complete, both the custom module
445 aliases and switches to disable it will go away. This removal
446 will also allow making ALSA OSS emulation independent of
447 sound_core. The dependency will be broken then too.
448 Who: Tejun Heo <tj@kernel.org>
450 ----------------------------
452 What: Support for lcd_switch and display_get in asus-laptop driver
454 Why: These two features use non-standard interfaces. There are the
455 only features that really need multiple path to guess what's
456 the right method name on a specific laptop.
458 Removing them will allow to remove a lot of code an significantly
461 This will affect the backlight code which won't be able to know
462 if the backlight is on or off. The platform display file will also be
463 write only (like the one in eeepc-laptop).
465 This should'nt affect a lot of user because they usually know
466 when their display is on or off.
468 Who: Corentin Chary <corentin.chary@gmail.com>
470 ----------------------------
472 What: sysfs-class-rfkill state file
474 Files: net/rfkill/core.c
475 Why: Documented as obsolete since Feb 2010. This file is limited to 3
476 states while the rfkill drivers can have 4 states.
477 Who: anybody or Florian Mickler <florian@mickler.org>
479 ----------------------------
481 What: sysfs-class-rfkill claim file
483 Files: net/rfkill/core.c
484 Why: It is not possible to claim an rfkill driver since 2007. This is
485 Documented as obsolete since Feb 2010.
486 Who: anybody or Florian Mickler <florian@mickler.org>
488 ----------------------------
492 Files: drivers/isdn/capi/capifs.*
493 Why: udev fully replaces this special file system that only contains CAPI
494 NCCI TTY device nodes. User space (pppdcapiplugin) works without
495 noticing the difference.
496 Who: Jan Kiszka <jan.kiszka@web.de>
498 ----------------------------
500 What: KVM paravirt mmu host support
502 Why: The paravirt mmu host support is slower than non-paravirt mmu, both
503 on newer and older hardware. It is already not exposed to the guest,
504 and kept only for live migration purposes.
505 Who: Avi Kivity <avi@redhat.com>
507 ----------------------------
509 What: iwlwifi 50XX module parameters
511 Why: The "..50" modules parameters were used to configure 5000 series and
512 up devices; different set of module parameters also available for 4965
513 with same functionalities. Consolidate both set into single place
514 in drivers/net/wireless/iwlwifi/iwl-agn.c
516 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
518 ----------------------------
520 What: iwl4965 alias support
522 Why: Internal alias support has been present in module-init-tools for some
523 time, the MODULE_ALIAS("iwl4965") boilerplate aliases can be removed
526 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
528 ---------------------------
531 Files: net/netfilter/xt_NOTRACK.c
533 Why: Superseded by xt_CT
534 Who: Netfilter developer team <netfilter-devel@vger.kernel.org>
536 ----------------------------
540 Why: The flag is a NOOP as we run interrupt handlers with interrupts disabled
541 Who: Thomas Gleixner <tglx@linutronix.de>
543 ----------------------------
545 What: The acpi_sleep=s4_nonvs command line option
547 Files: arch/x86/kernel/acpi/sleep.c
548 Why: superseded by acpi_sleep=nonvs
549 Who: Rafael J. Wysocki <rjw@sisk.pl>
551 ----------------------------
553 What: PCI DMA unmap state API
555 Why: PCI DMA unmap state API (include/linux/pci-dma.h) was replaced
556 with DMA unmap state API (DMA unmap state API can be used for
558 Who: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
560 ----------------------------
562 What: DMA_xxBIT_MASK macros
564 Why: DMA_xxBIT_MASK macros were replaced with DMA_BIT_MASK() macros.
565 Who: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
567 ----------------------------
569 What: namespace cgroup (ns_cgroup)
571 Why: The ns_cgroup leads to some problems:
572 * cgroup creation is out-of-control
573 * cgroup name can conflict when pids are looping
574 * it is not possible to have a single process handling
575 a lot of namespaces without falling in a exponential creation time
576 * we may want to create a namespace without creating a cgroup
578 The ns_cgroup is replaced by a compatibility flag 'clone_children',
579 where a newly created cgroup will copy the parent cgroup values.
580 The userspace has to manually create a cgroup and add a task to
582 Who: Daniel Lezcano <daniel.lezcano@free.fr>
584 ----------------------------
586 What: iwlwifi disable_hw_scan module parameters
588 Why: Hareware scan is the prefer method for iwlwifi devices for
589 scanning operation. Remove software scan support for all the
592 Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
594 ----------------------------
596 What: access to nfsd auth cache through sys_nfsservctl or '.' files
597 in the 'nfsd' filesystem.
599 Why: This is a legacy interface which have been replaced by a more
600 dynamic cache. Continuing to maintain this interface is an
602 Who: NeilBrown <neilb@suse.de>
604 ----------------------------
608 Why: This field is deprecated. I2C device drivers shouldn't change their
609 behavior based on the underlying I2C adapter. Instead, the I2C
610 adapter driver should instantiate the I2C devices and provide the
611 needed platform-specific information.
612 Who: Jean Delvare <khali@linux-fr.org>
614 ----------------------------
616 What: cancel_rearming_delayed_work[queue]()
619 Why: The functions have been superceded by cancel_delayed_work_sync()
620 quite some time ago. The conversion is trivial and there is no
622 Who: Tejun Heo <tj@kernel.org>
624 ----------------------------
626 What: Legacy, non-standard chassis intrusion detection interface.
628 Why: The adm9240, w83792d and w83793 hardware monitoring drivers have
629 legacy interfaces for chassis intrusion detection. A standard
630 interface has been added to each driver, so the legacy interface
632 Who: Jean Delvare <khali@linux-fr.org>
634 ----------------------------
636 What: noswapaccount kernel command line parameter
638 Why: The original implementation of memsw feature enabled by
639 CONFIG_CGROUP_MEM_RES_CTLR_SWAP could be disabled by the noswapaccount
640 kernel parameter (introduced in 2.6.29-rc1). Later on, this decision
641 turned out to be not ideal because we cannot have the feature compiled
642 in and disabled by default and let only interested to enable it
643 (e.g. general distribution kernels might need it). Therefore we have
644 added swapaccount[=0|1] parameter (introduced in 2.6.37) which provides
645 the both possibilities. If we remove noswapaccount we will have
646 less command line parameters with the same functionality and we
647 can also cleanup the parameter handling a bit ().
648 Who: Michal Hocko <mhocko@suse.cz>
650 ----------------------------