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 ---------------------------
12 Why: USER_SCHED was implemented as a proof of concept for group scheduling.
13 The effect of USER_SCHED can already be achieved from userspace with
14 the help of libcgroup. The removal of USER_SCHED will also simplify
15 the scheduler code with the removal of one major ifdef. There are also
16 issues USER_SCHED has with USER_NS. A decision was taken not to fix
17 those and instead remove USER_SCHED. Also new group scheduling
18 features will not be implemented for USER_SCHED.
20 Who: Dhaval Giani <dhaval@linux.vnet.ibm.com>
22 ---------------------------
27 Why: prism54 FullMAC PCI / Cardbus devices used to be supported only by the
28 prism54 wireless driver. After Intersil stopped selling these
29 devices in preference for the newer more flexible SoftMAC devices
30 a SoftMAC device driver was required and prism54 did not support
31 them. The p54pci driver now exists and has been present in the kernel for
32 a while. This driver supports both SoftMAC devices and FullMAC devices.
33 The main difference between these devices was the amount of memory which
34 could be used for the firmware. The SoftMAC devices support a smaller
35 amount of memory. Because of this the SoftMAC firmware fits into FullMAC
36 devices's memory. p54pci supports not only PCI / Cardbus but also USB
37 and SPI. Since p54pci supports all devices prism54 supports
38 you will have a conflict. I'm not quite sure how distributions are
39 handling this conflict right now. prism54 was kept around due to
40 claims users may experience issues when using the SoftMAC driver.
41 Time has passed users have not reported issues. If you use prism54
42 and for whatever reason you cannot use p54pci please let us know!
43 E-mail us at: linux-wireless@vger.kernel.org
45 For more information see the p54 wiki page:
47 http://wireless.kernel.org/en/users/Drivers/p54
49 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
51 ---------------------------
53 What: IRQF_SAMPLE_RANDOM
54 Check: IRQF_SAMPLE_RANDOM
57 Why: Many of IRQF_SAMPLE_RANDOM users are technically bogus as entropy
58 sources in the kernel's current entropy model. To resolve this, every
59 input point to the kernel's entropy pool needs to better document the
60 type of entropy source it actually is. This will be replaced with
61 additional add_*_randomness functions in drivers/char/random.c
63 Who: Robin Getz <rgetz@blackfin.uclinux.org> & Matt Mackall <mpm@selenic.com>
65 ---------------------------
67 What: The ieee80211_regdom module parameter
68 When: March 2010 / desktop catchup
70 Why: This was inherited by the CONFIG_WIRELESS_OLD_REGULATORY code,
71 and currently serves as an option for users to define an
72 ISO / IEC 3166 alpha2 code for the country they are currently
73 present in. Although there are userspace API replacements for this
74 through nl80211 distributions haven't yet caught up with implementing
75 decent alternatives through standard GUIs. Although available as an
76 option through iw or wpa_supplicant its just a matter of time before
77 distributions pick up good GUI options for this. The ideal solution
78 would actually consist of intelligent designs which would do this for
79 the user automatically even when travelling through different countries.
80 Until then we leave this module parameter as a compromise.
82 When userspace improves with reasonable widely-available alternatives for
83 this we will no longer need this module parameter. This entry hopes that
84 by the super-futuristically looking date of "March 2010" we will have
85 such replacements widely available.
87 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
89 ---------------------------
91 What: CONFIG_WIRELESS_OLD_REGULATORY - old static regulatory information
92 When: March 2010 / desktop catchup
94 Why: The old regulatory infrastructure has been replaced with a new one
95 which does not require statically defined regulatory domains. We do
96 not want to keep static regulatory domains in the kernel due to the
97 the dynamic nature of regulatory law and localization. We kept around
98 the old static definitions for the regulatory domains of:
104 and used by default the US when CONFIG_WIRELESS_OLD_REGULATORY was
105 set. We will remove this option once the standard Linux desktop catches
106 up with the new userspace APIs we have implemented.
108 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
110 ---------------------------
112 What: dev->power.power_state
114 Why: Broken design for runtime control over driver power states, confusing
115 driver-internal runtime power management with: mechanisms to support
116 system-wide sleep state transitions; event codes that distinguish
117 different phases of swsusp "sleep" transitions; and userspace policy
118 inputs. This framework was never widely used, and most attempts to
119 use it were broken. Drivers should instead be exposing domain-specific
120 interfaces either to kernel or to userspace.
121 Who: Pavel Machek <pavel@suse.cz>
123 ---------------------------
125 What: Video4Linux API 1 ioctls and from Video devices.
127 Files: include/linux/videodev.h
128 Check: include/linux/videodev.h
129 Why: V4L1 AP1 was replaced by V4L2 API during migration from 2.4 to 2.6
130 series. The old API have lots of drawbacks and don't provide enough
131 means to work with all video and audio standards. The newer API is
132 already available on the main drivers and should be used instead.
133 Newer drivers should use v4l_compat_translate_ioctl function to handle
134 old calls, replacing to newer ones.
135 Decoder iocts are using internally to allow video drivers to
136 communicate with video decoders. This should also be improved to allow
137 V4L2 calls being translated into compatible internal ioctls.
138 Compatibility ioctls will be provided, for a while, via
140 Who: Mauro Carvalho Chehab <mchehab@infradead.org>
142 ---------------------------
144 What: PCMCIA control ioctl (needed for pcmcia-cs [cardmgr, cardctl])
146 Files: drivers/pcmcia/: pcmcia_ioctl.c
147 Why: With the 16-bit PCMCIA subsystem now behaving (almost) like a
148 normal hotpluggable bus, and with it using the default kernel
149 infrastructure (hotplug, driver core, sysfs) keeping the PCMCIA
150 control ioctl needed by cardmgr and cardctl from pcmcia-cs is
151 unnecessary, and makes further cleanups and integration of the
152 PCMCIA subsystem into the Linux kernel device driver model more
153 difficult. The features provided by cardmgr and cardctl are either
154 handled by the kernel itself now or are available in the new
155 pcmciautils package available at
156 http://kernel.org/pub/linux/utils/kernel/pcmcia/
157 Who: Dominik Brodowski <linux@brodo.de>
159 ---------------------------
163 Option: CONFIG_SYSCTL_SYSCALL
164 Why: The same information is available in a more convenient from
165 /proc/sys, and none of the sysctl variables appear to be
166 important performance wise.
168 Binary sysctls are a long standing source of subtle kernel
169 bugs and security issues.
171 When I looked several months ago all I could find after
172 searching several distributions were 5 user space programs and
173 glibc (which falls back to /proc/sys) using this syscall.
175 The man page for sysctl(2) documents it as unusable for user
178 sysctl(2) is not generally ABI compatible to a 32bit user
179 space application on a 64bit and a 32bit kernel.
181 For the last several months the policy has been no new binary
182 sysctls and no one has put forward an argument to use them.
184 Binary sysctls issues seem to keep happening appearing so
185 properly deprecating them (with a warning to user space) and a
186 2 year grace warning period will mean eventually we can kill
187 them and end the pain.
189 In the mean time individual binary sysctls can be dealt with
190 in a piecewise fashion.
192 Who: Eric Biederman <ebiederm@xmission.com>
194 ---------------------------
196 What: remove EXPORT_SYMBOL(kernel_thread)
198 Files: arch/*/kernel/*_ksyms.c
200 Why: kernel_thread is a low-level implementation detail. Drivers should
201 use the <linux/kthread.h> API instead which shields them from
202 implementation details and provides a higherlevel interface that
203 prevents bugs and code duplication
204 Who: Christoph Hellwig <hch@lst.de>
206 ---------------------------
208 What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
209 (temporary transition config option provided until then)
210 The transition config option will also be removed at the same time.
212 Why: Unused symbols are both increasing the size of the kernel binary
213 and are often a sign of "wrong API"
214 Who: Arjan van de Ven <arjan@linux.intel.com>
216 ---------------------------
218 What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
220 Why: The stacking of class devices makes these values misleading and
222 Class devices should not carry any of these properties, and bus
223 devices have SUBSYTEM and DRIVER as a replacement.
224 Who: Kay Sievers <kay.sievers@suse.de>
226 ---------------------------
228 What: ACPI procfs interface
230 Why: ACPI sysfs conversion should be finished by January 2008.
231 ACPI procfs interface will be removed in July 2008 so that
232 there is enough time for the user space to catch up.
233 Who: Zhang Rui <rui.zhang@intel.com>
235 ---------------------------
237 What: /proc/acpi/button
239 Why: /proc/acpi/button has been replaced by events to the input layer
241 Who: Len Brown <len.brown@intel.com>
243 ---------------------------
245 What: /proc/acpi/event
247 Why: /proc/acpi/event has been replaced by events via the input layer
248 and netlink since 2.6.23.
249 Who: Len Brown <len.brown@intel.com>
251 ---------------------------
253 What: i386/x86_64 bzImage symlinks
256 Why: The i386/x86_64 merge provides a symlink to the old bzImage
257 location so not yet updated user space tools, e.g. package
258 scripts, do not break.
259 Who: Thomas Gleixner <tglx@linutronix.de>
261 ---------------------------
264 - xt_recent: the old ipt_recent proc dir
265 (superseded by /proc/net/xt_recent)
267 When: January 2009 or Linux 2.7.0, whichever comes first
268 Why: Superseded by newer revisions or modules
269 Who: Jan Engelhardt <jengelh@computergmbh.de>
271 ---------------------------
273 What: GPIO autorequest on gpio_direction_{input,output}() in gpiolib
275 Why: All callers should use explicit gpio_request()/gpio_free().
276 The autorequest mechanism in gpiolib was provided mostly as a
277 migration aid for legacy GPIO interfaces (for SOC based GPIOs).
278 Those users have now largely migrated. Platforms implementing
279 the GPIO interfaces without using gpiolib will see no changes.
280 Who: David Brownell <dbrownell@users.sourceforge.net>
281 ---------------------------
283 What: b43 support for firmware revision < 410
284 When: The schedule was July 2008, but it was decided that we are going to keep the
285 code as long as there are no major maintanance headaches.
286 So it _could_ be removed _any_ time now, if it conflicts with something new.
287 Why: The support code for the old firmware hurts code readability/maintainability
288 and slightly hurts runtime performance. Bugfixes for the old firmware
289 are not provided by Broadcom anymore.
290 Who: Michael Buesch <mb@bu3sch.de>
292 ---------------------------
294 What: usedac i386 kernel parameter
296 Why: replaced by allowdac and no dac combination
297 Who: Glauber Costa <gcosta@redhat.com>
299 ---------------------------
301 What: print_fn_descriptor_symbol()
303 Why: The %pF vsprintf format provides the same functionality in a
304 simpler way. print_fn_descriptor_symbol() is deprecated but
305 still present to give out-of-tree modules time to change.
306 Who: Bjorn Helgaas <bjorn.helgaas@hp.com>
308 ---------------------------
310 What: /sys/o2cb symlink
312 Why: /sys/fs/o2cb is the proper location for this information - /sys/o2cb
313 exists as a symlink for backwards compatibility for old versions of
314 ocfs2-tools. 2 years should be sufficient time to phase in new versions
315 which know to look in /sys/fs/o2cb.
316 Who: ocfs2-devel@oss.oracle.com
318 ---------------------------
320 What: SCTP_GET_PEER_ADDRS_NUM_OLD, SCTP_GET_PEER_ADDRS_OLD,
321 SCTP_GET_LOCAL_ADDRS_NUM_OLD, SCTP_GET_LOCAL_ADDRS_OLD
323 Why: A newer version of the options have been introduced in 2005 that
324 removes the limitions of the old API. The sctp library has been
325 converted to use these new options at the same time. Any user
326 space app that directly uses the old options should convert to using
328 Who: Vlad Yasevich <vladislav.yasevich@hp.com>
330 ---------------------------
332 What: Ability for non root users to shm_get hugetlb pages based on mlock
335 Why: Non root users need to be part of /proc/sys/vm/hugetlb_shm_group or
336 have CAP_IPC_LOCK to be able to allocate shm segments backed by
337 huge pages. The mlock based rlimit check to allow shm hugetlb is
338 inconsistent with mmap based allocations. Hence it is being
340 Who: Ravikiran Thirumalai <kiran@scalex86.org>
342 ---------------------------
344 What: CONFIG_THERMAL_HWMON
346 Why: This option was introduced just to allow older lm-sensors userspace
347 to keep working over the upgrade to 2.6.26. At the scheduled time of
348 removal fixed lm-sensors (2.x or 3.x) should be readily available.
349 Who: Rene Herman <rene.herman@gmail.com>
351 ---------------------------
353 What: Code that is now under CONFIG_WIRELESS_EXT_SYSFS
354 (in net/core/net-sysfs.c)
355 When: After the only user (hal) has seen a release with the patches
356 for enough time, probably some time in 2010.
357 Why: Over 1K .text/.data size reduction, data is available in other
359 Who: Johannes Berg <johannes@sipsolutions.net>
361 ---------------------------
363 What: CONFIG_NF_CT_ACCT
365 Why: Accounting can now be enabled/disabled without kernel recompilation.
366 Currently used only to set a default value for a feature that is also
367 controlled by a kernel/module/sysfs/sysctl parameter.
368 Who: Krzysztof Piotr Oledzki <ole@ans.pl>
370 ---------------------------
372 What: sysfs ui for changing p4-clockmod parameters
374 Why: See commits 129f8ae9b1b5be94517da76009ea956e89104ce8 and
375 e088e4c9cdb618675874becb91b2fd581ee707e6.
376 Removal is subject to fixing any remaining bugs in ACPI which may
377 cause the thermal throttling not to happen at the right time.
378 Who: Dave Jones <davej@redhat.com>, Matthew Garrett <mjg@redhat.com>
380 -----------------------------
382 What: __do_IRQ all in one fits nothing interrupt handler
384 Why: __do_IRQ was kept for easy migration to the type flow handlers.
385 More than two years of migration time is enough.
386 Who: Thomas Gleixner <tglx@linutronix.de>
388 -----------------------------
390 What: fakephp and associated sysfs files in /sys/bus/pci/slots/
392 Why: In 2.6.27, the semantics of /sys/bus/pci/slots was redefined to
393 represent a machine's physical PCI slots. The change in semantics
394 had userspace implications, as the hotplug core no longer allowed
395 drivers to create multiple sysfs files per physical slot (required
396 for multi-function devices, e.g.). fakephp was seen as a developer's
397 tool only, and its interface changed. Too late, we learned that
398 there were some users of the fakephp interface.
400 In 2.6.30, the original fakephp interface was restored. At the same
401 time, the PCI core gained the ability that fakephp provided, namely
402 function-level hot-remove and hot-add.
404 Since the PCI core now provides the same functionality, exposed in:
407 /sys/bus/pci/devices/.../remove
408 /sys/bus/pci/devices/.../rescan
410 there is no functional reason to maintain fakephp as well.
412 We will keep the existing module so that 'modprobe fakephp' will
413 present the old /sys/bus/pci/slots/... interface for compatibility,
414 but users are urged to migrate their applications to the API above.
416 After a reasonable transition period, we will remove the legacy
418 Who: Alex Chiang <achiang@hp.com>
420 ---------------------------
422 What: i2c-voodoo3 driver
424 Why: Superseded by tdfxfb. I2C/DDC support used to live in a separate
425 driver but this caused driver conflicts.
426 Who: Jean Delvare <khali@linux-fr.org>
427 Krzysztof Helt <krzysztof.h1@wp.pl>
429 ---------------------------
431 What: CONFIG_RFKILL_INPUT
433 Why: Should be implemented in userspace, policy daemon.
434 Who: Johannes Berg <johannes@sipsolutions.net>
436 ---------------------------
440 Why: last user (audit) will be converted to the newer more generic
441 and more easily maintained fsnotify subsystem
442 Who: Eric Paris <eparis@redhat.com>
444 ----------------------------
446 What: lock_policy_rwsem_* and unlock_policy_rwsem_* will not be
447 exported interface anymore.
449 Why: cpu_policy_rwsem has a new cleaner definition making it local to
450 cpufreq core and contained inside cpufreq.c. Other dependent
451 drivers should not use it in order to safely avoid lockdep issues.
452 Who: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
454 ----------------------------
456 What: sound-slot/service-* module aliases and related clutters in
459 Why: OSS sound_core grabs all legacy minors (0-255) of SOUND_MAJOR
460 (14) and requests modules using custom sound-slot/service-*
461 module aliases. The only benefit of doing this is allowing
462 use of custom module aliases which might as well be considered
463 a bug at this point. This preemptive claiming prevents
464 alternative OSS implementations.
466 Till the feature is removed, the kernel will be requesting
467 both sound-slot/service-* and the standard char-major-* module
468 aliases and allow turning off the pre-claiming selectively via
469 CONFIG_SOUND_OSS_CORE_PRECLAIM and soundcore.preclaim_oss
472 After the transition phase is complete, both the custom module
473 aliases and switches to disable it will go away. This removal
474 will also allow making ALSA OSS emulation independent of
475 sound_core. The dependency will be broken then too.
476 Who: Tejun Heo <tj@kernel.org>
478 ----------------------------
480 What: Support for VMware's guest paravirtuliazation technique [VMI] will be
482 When: 2.6.37 or earlier.
483 Why: With the recent innovations in CPU hardware acceleration technologies
484 from Intel and AMD, VMware ran a few experiments to compare these
485 techniques to guest paravirtualization technique on VMware's platform.
486 These hardware assisted virtualization techniques have outperformed the
487 performance benefits provided by VMI in most of the workloads. VMware
488 expects that these hardware features will be ubiquitous in a couple of
489 years, as a result, VMware has started a phased retirement of this
490 feature from the hypervisor. We will be removing this feature from the
491 Kernel too. Right now we are targeting 2.6.37 but can retire earlier if
492 technical reasons (read opportunity to remove major chunk of pvops)
495 Please note that VMI has always been an optimization and non-VMI kernels
496 still work fine on VMware's platform.
497 Latest versions of VMware's product which support VMI are,
498 Workstation 7.0 and VSphere 4.0 on ESX side, future maintainence
499 releases for these products will continue supporting VMI.
501 For more details about VMI retirement take a look at this,
502 http://blogs.vmware.com/guestosguide/2009/09/vmi-retirement.html
504 Who: Alok N Kataria <akataria@vmware.com>
506 ----------------------------