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: IRQF_SAMPLE_RANDOM
10 Check: IRQF_SAMPLE_RANDOM
13 Why: Many of IRQF_SAMPLE_RANDOM users are technically bogus as entropy
14 sources in the kernel's current entropy model. To resolve this, every
15 input point to the kernel's entropy pool needs to better document the
16 type of entropy source it actually is. This will be replaced with
17 additional add_*_randomness functions in drivers/char/random.c
19 Who: Robin Getz <rgetz@blackfin.uclinux.org> & Matt Mackall <mpm@selenic.com>
21 ---------------------------
23 What: The ieee80211_regdom module parameter
24 When: March 2010 / desktop catchup
26 Why: This was inherited by the CONFIG_WIRELESS_OLD_REGULATORY code,
27 and currently serves as an option for users to define an
28 ISO / IEC 3166 alpha2 code for the country they are currently
29 present in. Although there are userspace API replacements for this
30 through nl80211 distributions haven't yet caught up with implementing
31 decent alternatives through standard GUIs. Although available as an
32 option through iw or wpa_supplicant its just a matter of time before
33 distributions pick up good GUI options for this. The ideal solution
34 would actually consist of intelligent designs which would do this for
35 the user automatically even when travelling through different countries.
36 Until then we leave this module parameter as a compromise.
38 When userspace improves with reasonable widely-available alternatives for
39 this we will no longer need this module parameter. This entry hopes that
40 by the super-futuristically looking date of "March 2010" we will have
41 such replacements widely available.
43 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
45 ---------------------------
47 What: CONFIG_WIRELESS_OLD_REGULATORY - old static regulatory information
48 When: March 2010 / desktop catchup
50 Why: The old regulatory infrastructure has been replaced with a new one
51 which does not require statically defined regulatory domains. We do
52 not want to keep static regulatory domains in the kernel due to the
53 the dynamic nature of regulatory law and localization. We kept around
54 the old static definitions for the regulatory domains of:
60 and used by default the US when CONFIG_WIRELESS_OLD_REGULATORY was
61 set. We will remove this option once the standard Linux desktop catches
62 up with the new userspace APIs we have implemented.
64 Who: Luis R. Rodriguez <lrodriguez@atheros.com>
66 ---------------------------
68 What: dev->power.power_state
70 Why: Broken design for runtime control over driver power states, confusing
71 driver-internal runtime power management with: mechanisms to support
72 system-wide sleep state transitions; event codes that distinguish
73 different phases of swsusp "sleep" transitions; and userspace policy
74 inputs. This framework was never widely used, and most attempts to
75 use it were broken. Drivers should instead be exposing domain-specific
76 interfaces either to kernel or to userspace.
77 Who: Pavel Machek <pavel@suse.cz>
79 ---------------------------
81 What: Video4Linux API 1 ioctls and from Video devices.
83 Files: include/linux/videodev.h
84 Check: include/linux/videodev.h
85 Why: V4L1 AP1 was replaced by V4L2 API during migration from 2.4 to 2.6
86 series. The old API have lots of drawbacks and don't provide enough
87 means to work with all video and audio standards. The newer API is
88 already available on the main drivers and should be used instead.
89 Newer drivers should use v4l_compat_translate_ioctl function to handle
90 old calls, replacing to newer ones.
91 Decoder iocts are using internally to allow video drivers to
92 communicate with video decoders. This should also be improved to allow
93 V4L2 calls being translated into compatible internal ioctls.
94 Compatibility ioctls will be provided, for a while, via
96 Who: Mauro Carvalho Chehab <mchehab@infradead.org>
98 ---------------------------
100 What: PCMCIA control ioctl (needed for pcmcia-cs [cardmgr, cardctl])
102 Files: drivers/pcmcia/: pcmcia_ioctl.c
103 Why: With the 16-bit PCMCIA subsystem now behaving (almost) like a
104 normal hotpluggable bus, and with it using the default kernel
105 infrastructure (hotplug, driver core, sysfs) keeping the PCMCIA
106 control ioctl needed by cardmgr and cardctl from pcmcia-cs is
107 unnecessary, and makes further cleanups and integration of the
108 PCMCIA subsystem into the Linux kernel device driver model more
109 difficult. The features provided by cardmgr and cardctl are either
110 handled by the kernel itself now or are available in the new
111 pcmciautils package available at
112 http://kernel.org/pub/linux/utils/kernel/pcmcia/
113 Who: Dominik Brodowski <linux@brodo.de>
115 ---------------------------
119 Option: CONFIG_SYSCTL_SYSCALL
120 Why: The same information is available in a more convenient from
121 /proc/sys, and none of the sysctl variables appear to be
122 important performance wise.
124 Binary sysctls are a long standing source of subtle kernel
125 bugs and security issues.
127 When I looked several months ago all I could find after
128 searching several distributions were 5 user space programs and
129 glibc (which falls back to /proc/sys) using this syscall.
131 The man page for sysctl(2) documents it as unusable for user
134 sysctl(2) is not generally ABI compatible to a 32bit user
135 space application on a 64bit and a 32bit kernel.
137 For the last several months the policy has been no new binary
138 sysctls and no one has put forward an argument to use them.
140 Binary sysctls issues seem to keep happening appearing so
141 properly deprecating them (with a warning to user space) and a
142 2 year grace warning period will mean eventually we can kill
143 them and end the pain.
145 In the mean time individual binary sysctls can be dealt with
146 in a piecewise fashion.
148 Who: Eric Biederman <ebiederm@xmission.com>
150 ---------------------------
152 What: remove EXPORT_SYMBOL(kernel_thread)
154 Files: arch/*/kernel/*_ksyms.c
156 Why: kernel_thread is a low-level implementation detail. Drivers should
157 use the <linux/kthread.h> API instead which shields them from
158 implementation details and provides a higherlevel interface that
159 prevents bugs and code duplication
160 Who: Christoph Hellwig <hch@lst.de>
162 ---------------------------
164 What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
165 (temporary transition config option provided until then)
166 The transition config option will also be removed at the same time.
168 Why: Unused symbols are both increasing the size of the kernel binary
169 and are often a sign of "wrong API"
170 Who: Arjan van de Ven <arjan@linux.intel.com>
172 ---------------------------
174 What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
176 Why: The stacking of class devices makes these values misleading and
178 Class devices should not carry any of these properties, and bus
179 devices have SUBSYTEM and DRIVER as a replacement.
180 Who: Kay Sievers <kay.sievers@suse.de>
182 ---------------------------
184 What: ACPI procfs interface
186 Why: ACPI sysfs conversion should be finished by January 2008.
187 ACPI procfs interface will be removed in July 2008 so that
188 there is enough time for the user space to catch up.
189 Who: Zhang Rui <rui.zhang@intel.com>
191 ---------------------------
193 What: /proc/acpi/button
195 Why: /proc/acpi/button has been replaced by events to the input layer
197 Who: Len Brown <len.brown@intel.com>
199 ---------------------------
201 What: /proc/acpi/event
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>
207 ---------------------------
209 What: libata spindown skipping and warning
211 Why: Some halt(8) implementations synchronize caches for and spin
212 down libata disks because libata didn't use to spin down disk on
213 system halt (only synchronized caches).
214 Spin down on system halt is now implemented. sysfs node
215 /sys/class/scsi_disk/h:c:i:l/manage_start_stop is present if
216 spin down support is available.
217 Because issuing spin down command to an already spun down disk
218 makes some disks spin up just to spin down again, libata tracks
219 device spindown status to skip the extra spindown command and
221 This is to give userspace tools the time to get updated and will
222 be removed after userspace is reasonably updated.
223 Who: Tejun Heo <htejun@gmail.com>
225 ---------------------------
227 What: i386/x86_64 bzImage symlinks
230 Why: The i386/x86_64 merge provides a symlink to the old bzImage
231 location so not yet updated user space tools, e.g. package
232 scripts, do not break.
233 Who: Thomas Gleixner <tglx@linutronix.de>
235 ---------------------------
238 - "forwarding" header files like ipt_mac.h in
239 include/linux/netfilter_ipv4/ and include/linux/netfilter_ipv6/
241 - xt_recent: the old ipt_recent proc dir
242 (superseded by /proc/net/xt_recent)
244 When: January 2009 or Linux 2.7.0, whichever comes first
245 Why: Superseded by newer revisions or modules
246 Who: Jan Engelhardt <jengelh@computergmbh.de>
248 ---------------------------
250 What: GPIO autorequest on gpio_direction_{input,output}() in gpiolib
252 Why: All callers should use explicit gpio_request()/gpio_free().
253 The autorequest mechanism in gpiolib was provided mostly as a
254 migration aid for legacy GPIO interfaces (for SOC based GPIOs).
255 Those users have now largely migrated. Platforms implementing
256 the GPIO interfaces without using gpiolib will see no changes.
257 Who: David Brownell <dbrownell@users.sourceforge.net>
258 ---------------------------
260 What: b43 support for firmware revision < 410
261 When: The schedule was July 2008, but it was decided that we are going to keep the
262 code as long as there are no major maintanance headaches.
263 So it _could_ be removed _any_ time now, if it conflicts with something new.
264 Why: The support code for the old firmware hurts code readability/maintainability
265 and slightly hurts runtime performance. Bugfixes for the old firmware
266 are not provided by Broadcom anymore.
267 Who: Michael Buesch <mb@bu3sch.de>
269 ---------------------------
271 What: usedac i386 kernel parameter
273 Why: replaced by allowdac and no dac combination
274 Who: Glauber Costa <gcosta@redhat.com>
276 ---------------------------
278 What: print_fn_descriptor_symbol()
280 Why: The %pF vsprintf format provides the same functionality in a
281 simpler way. print_fn_descriptor_symbol() is deprecated but
282 still present to give out-of-tree modules time to change.
283 Who: Bjorn Helgaas <bjorn.helgaas@hp.com>
285 ---------------------------
287 What: /sys/o2cb symlink
289 Why: /sys/fs/o2cb is the proper location for this information - /sys/o2cb
290 exists as a symlink for backwards compatibility for old versions of
291 ocfs2-tools. 2 years should be sufficient time to phase in new versions
292 which know to look in /sys/fs/o2cb.
293 Who: ocfs2-devel@oss.oracle.com
295 ---------------------------
297 What: SCTP_GET_PEER_ADDRS_NUM_OLD, SCTP_GET_PEER_ADDRS_OLD,
298 SCTP_GET_LOCAL_ADDRS_NUM_OLD, SCTP_GET_LOCAL_ADDRS_OLD
300 Why: A newer version of the options have been introduced in 2005 that
301 removes the limitions of the old API. The sctp library has been
302 converted to use these new options at the same time. Any user
303 space app that directly uses the old options should convert to using
305 Who: Vlad Yasevich <vladislav.yasevich@hp.com>
307 ---------------------------
309 What: Ability for non root users to shm_get hugetlb pages based on mlock
312 Why: Non root users need to be part of /proc/sys/vm/hugetlb_shm_group or
313 have CAP_IPC_LOCK to be able to allocate shm segments backed by
314 huge pages. The mlock based rlimit check to allow shm hugetlb is
315 inconsistent with mmap based allocations. Hence it is being
317 Who: Ravikiran Thirumalai <kiran@scalex86.org>
319 ---------------------------
321 What: CONFIG_THERMAL_HWMON
323 Why: This option was introduced just to allow older lm-sensors userspace
324 to keep working over the upgrade to 2.6.26. At the scheduled time of
325 removal fixed lm-sensors (2.x or 3.x) should be readily available.
326 Who: Rene Herman <rene.herman@gmail.com>
328 ---------------------------
330 What: Code that is now under CONFIG_WIRELESS_EXT_SYSFS
331 (in net/core/net-sysfs.c)
332 When: After the only user (hal) has seen a release with the patches
333 for enough time, probably some time in 2010.
334 Why: Over 1K .text/.data size reduction, data is available in other
336 Who: Johannes Berg <johannes@sipsolutions.net>
338 ---------------------------
340 What: CONFIG_NF_CT_ACCT
342 Why: Accounting can now be enabled/disabled without kernel recompilation.
343 Currently used only to set a default value for a feature that is also
344 controlled by a kernel/module/sysfs/sysctl parameter.
345 Who: Krzysztof Piotr Oledzki <ole@ans.pl>
347 ---------------------------
349 What: fscher and fscpos drivers
351 Why: Deprecated by the new fschmd driver.
352 Who: Hans de Goede <hdegoede@redhat.com>
353 Jean Delvare <khali@linux-fr.org>
355 ---------------------------
357 What: sysfs ui for changing p4-clockmod parameters
359 Why: See commits 129f8ae9b1b5be94517da76009ea956e89104ce8 and
360 e088e4c9cdb618675874becb91b2fd581ee707e6.
361 Removal is subject to fixing any remaining bugs in ACPI which may
362 cause the thermal throttling not to happen at the right time.
363 Who: Dave Jones <davej@redhat.com>, Matthew Garrett <mjg@redhat.com>
365 -----------------------------
367 What: __do_IRQ all in one fits nothing interrupt handler
369 Why: __do_IRQ was kept for easy migration to the type flow handlers.
370 More than two years of migration time is enough.
371 Who: Thomas Gleixner <tglx@linutronix.de>
373 -----------------------------
375 What: obsolete generic irq defines and typedefs
377 Why: The defines and typedefs (hw_interrupt_type, no_irq_type, irq_desc_t)
378 have been kept around for migration reasons. After more than two years
379 it's time to remove them finally
380 Who: Thomas Gleixner <tglx@linutronix.de>
382 ---------------------------
384 What: fakephp and associated sysfs files in /sys/bus/pci/slots/
386 Why: In 2.6.27, the semantics of /sys/bus/pci/slots was redefined to
387 represent a machine's physical PCI slots. The change in semantics
388 had userspace implications, as the hotplug core no longer allowed
389 drivers to create multiple sysfs files per physical slot (required
390 for multi-function devices, e.g.). fakephp was seen as a developer's
391 tool only, and its interface changed. Too late, we learned that
392 there were some users of the fakephp interface.
394 In 2.6.30, the original fakephp interface was restored. At the same
395 time, the PCI core gained the ability that fakephp provided, namely
396 function-level hot-remove and hot-add.
398 Since the PCI core now provides the same functionality, exposed in:
401 /sys/bus/pci/devices/.../remove
402 /sys/bus/pci/devices/.../rescan
404 there is no functional reason to maintain fakephp as well.
406 We will keep the existing module so that 'modprobe fakephp' will
407 present the old /sys/bus/pci/slots/... interface for compatibility,
408 but users are urged to migrate their applications to the API above.
410 After a reasonable transition period, we will remove the legacy
412 Who: Alex Chiang <achiang@hp.com>
414 ---------------------------
416 What: i2c-voodoo3 driver
418 Why: Superseded by tdfxfb. I2C/DDC support used to live in a separate
419 driver but this caused driver conflicts.
420 Who: Jean Delvare <khali@linux-fr.org>
421 Krzysztof Helt <krzysztof.h1@wp.pl>
423 ---------------------------
425 What: CONFIG_RFKILL_INPUT
427 Why: Should be implemented in userspace, policy daemon.
428 Who: Johannes Berg <johannes@sipsolutions.net>
430 ----------------------------
432 What: CONFIG_X86_OLD_MCE
434 Why: Remove the old legacy 32bit machine check code. This has been
435 superseded by the newer machine check code from the 64bit port,
436 but the old version has been kept around for easier testing. Note this
437 doesn't impact the old P5 and WinChip machine check handlers.
438 Who: Andi Kleen <andi@firstfloor.org>