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: dev->power.power_state
11 Why: Broken design for runtime control over driver power states, confusing
12 driver-internal runtime power management with: mechanisms to support
13 system-wide sleep state transitions; event codes that distinguish
14 different phases of swsusp "sleep" transitions; and userspace policy
15 inputs. This framework was never widely used, and most attempts to
16 use it were broken. Drivers should instead be exposing domain-specific
17 interfaces either to kernel or to userspace.
18 Who: Pavel Machek <pavel@suse.cz>
20 ---------------------------
22 What: old NCR53C9x driver
24 Why: Replaced by the much better esp_scsi driver. Actual low-level
25 driver can be ported over almost trivially.
26 Who: David Miller <davem@davemloft.net>
27 Christoph Hellwig <hch@lst.de>
29 ---------------------------
31 What: Video4Linux API 1 ioctls and video_decoder.h from Video devices.
33 Files: include/linux/video_decoder.h include/linux/videodev.h
34 Check: include/linux/video_decoder.h include/linux/videodev.h
35 Why: V4L1 AP1 was replaced by V4L2 API during migration from 2.4 to 2.6
36 series. The old API have lots of drawbacks and don't provide enough
37 means to work with all video and audio standards. The newer API is
38 already available on the main drivers and should be used instead.
39 Newer drivers should use v4l_compat_translate_ioctl function to handle
40 old calls, replacing to newer ones.
41 Decoder iocts are using internally to allow video drivers to
42 communicate with video decoders. This should also be improved to allow
43 V4L2 calls being translated into compatible internal ioctls.
44 Compatibility ioctls will be provided, for a while, via
46 Who: Mauro Carvalho Chehab <mchehab@infradead.org>
48 ---------------------------
50 What: PCMCIA control ioctl (needed for pcmcia-cs [cardmgr, cardctl])
52 Files: drivers/pcmcia/: pcmcia_ioctl.c
53 Why: With the 16-bit PCMCIA subsystem now behaving (almost) like a
54 normal hotpluggable bus, and with it using the default kernel
55 infrastructure (hotplug, driver core, sysfs) keeping the PCMCIA
56 control ioctl needed by cardmgr and cardctl from pcmcia-cs is
57 unnecessary, and makes further cleanups and integration of the
58 PCMCIA subsystem into the Linux kernel device driver model more
59 difficult. The features provided by cardmgr and cardctl are either
60 handled by the kernel itself now or are available in the new
61 pcmciautils package available at
62 http://kernel.org/pub/linux/utils/kernel/pcmcia/
63 Who: Dominik Brodowski <linux@brodo.de>
65 ---------------------------
69 Option: CONFIG_SYSCTL_SYSCALL
70 Why: The same information is available in a more convenient from
71 /proc/sys, and none of the sysctl variables appear to be
72 important performance wise.
74 Binary sysctls are a long standing source of subtle kernel
75 bugs and security issues.
77 When I looked several months ago all I could find after
78 searching several distributions were 5 user space programs and
79 glibc (which falls back to /proc/sys) using this syscall.
81 The man page for sysctl(2) documents it as unusable for user
84 sysctl(2) is not generally ABI compatible to a 32bit user
85 space application on a 64bit and a 32bit kernel.
87 For the last several months the policy has been no new binary
88 sysctls and no one has put forward an argument to use them.
90 Binary sysctls issues seem to keep happening appearing so
91 properly deprecating them (with a warning to user space) and a
92 2 year grace warning period will mean eventually we can kill
93 them and end the pain.
95 In the mean time individual binary sysctls can be dealt with
96 in a piecewise fashion.
98 Who: Eric Biederman <ebiederm@xmission.com>
100 ---------------------------
102 What: remove EXPORT_SYMBOL(kernel_thread)
104 Files: arch/*/kernel/*_ksyms.c
106 Why: kernel_thread is a low-level implementation detail. Drivers should
107 use the <linux/kthread.h> API instead which shields them from
108 implementation details and provides a higherlevel interface that
109 prevents bugs and code duplication
110 Who: Christoph Hellwig <hch@lst.de>
112 ---------------------------
114 What: eepro100 network driver
116 Why: replaced by the e100 driver
117 Who: Adrian Bunk <bunk@stusta.de>
119 ---------------------------
121 What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
122 (temporary transition config option provided until then)
123 The transition config option will also be removed at the same time.
125 Why: Unused symbols are both increasing the size of the kernel binary
126 and are often a sign of "wrong API"
127 Who: Arjan van de Ven <arjan@linux.intel.com>
129 ---------------------------
132 When: Soon, provided in-kernel callers have been converted
133 Why: This interface is replaced by vm_ops.fault, but it has been around
134 forever, is used by a lot of drivers, and doesn't cost much to
136 Who: Nick Piggin <npiggin@suse.de>
138 ---------------------------
140 What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
142 Why: The stacking of class devices makes these values misleading and
144 Class devices should not carry any of these properties, and bus
145 devices have SUBSYTEM and DRIVER as a replacement.
146 Who: Kay Sievers <kay.sievers@suse.de>
148 ---------------------------
150 What: ACPI procfs interface
152 Why: ACPI sysfs conversion should be finished by January 2008.
153 ACPI procfs interface will be removed in July 2008 so that
154 there is enough time for the user space to catch up.
155 Who: Zhang Rui <rui.zhang@intel.com>
157 ---------------------------
159 What: /proc/acpi/button
161 Why: /proc/acpi/button has been replaced by events to the input layer
163 Who: Len Brown <len.brown@intel.com>
165 ---------------------------
167 What: /proc/acpi/event
169 Why: /proc/acpi/event has been replaced by events via the input layer
170 and netlink since 2.6.23.
171 Who: Len Brown <len.brown@intel.com>
173 ---------------------------
175 What: libata spindown skipping and warning
177 Why: Some halt(8) implementations synchronize caches for and spin
178 down libata disks because libata didn't use to spin down disk on
179 system halt (only synchronized caches).
180 Spin down on system halt is now implemented. sysfs node
181 /sys/class/scsi_disk/h:c:i:l/manage_start_stop is present if
182 spin down support is available.
183 Because issuing spin down command to an already spun down disk
184 makes some disks spin up just to spin down again, libata tracks
185 device spindown status to skip the extra spindown command and
187 This is to give userspace tools the time to get updated and will
188 be removed after userspace is reasonably updated.
189 Who: Tejun Heo <htejun@gmail.com>
191 ---------------------------
193 What: The arch/ppc and include/asm-ppc directories
195 Why: The arch/powerpc tree is the merged architecture for ppc32 and ppc64
196 platforms. Currently there are efforts underway to port the remaining
197 arch/ppc platforms to the merged tree. New submissions to the arch/ppc
198 tree have been frozen with the 2.6.22 kernel release and that tree will
199 remain in bug-fix only mode until its scheduled removal. Platforms
200 that are not ported by June 2008 will be removed due to the lack of an
201 interested maintainer.
202 Who: linuxppc-dev@ozlabs.org
204 ---------------------------
206 What: i386/x86_64 bzImage symlinks
209 Why: The i386/x86_64 merge provides a symlink to the old bzImage
210 location so not yet updated user space tools, e.g. package
211 scripts, do not break.
212 Who: Thomas Gleixner <tglx@linutronix.de>
214 ---------------------------
216 What: i2c-i810, i2c-prosavage and i2c-savage4
218 Why: These drivers are superseded by i810fb, intelfb and savagefb.
219 Who: Jean Delvare <khali@linux-fr.org>
221 ---------------------------
224 - include/linux/netfilter_ipv4/ipt_TOS.h ipt_tos.h header files
225 (superseded by xt_TOS/xt_tos target & match)
227 - "forwarding" header files like ipt_mac.h in
228 include/linux/netfilter_ipv4/ and include/linux/netfilter_ipv6/
230 - xt_CONNMARK match revision 0
231 (superseded by xt_CONNMARK match revision 1)
233 - xt_MARK target revisions 0 and 1
234 (superseded by xt_MARK match revision 2)
236 - xt_connmark match revision 0
237 (superseded by xt_connmark match revision 1)
239 - xt_conntrack match revision 0
240 (superseded by xt_conntrack match revision 1)
242 - xt_iprange match revision 0,
243 include/linux/netfilter_ipv4/ipt_iprange.h
244 (superseded by xt_iprange match revision 1)
246 - xt_mark match revision 0
247 (superseded by xt_mark match revision 1)
249 When: January 2009 or Linux 2.7.0, whichever comes first
250 Why: Superseded by newer revisions or modules
251 Who: Jan Engelhardt <jengelh@computergmbh.de>
253 ---------------------------
255 What: b43 support for firmware revision < 410
257 Why: The support code for the old firmware hurts code readability/maintainability
258 and slightly hurts runtime performance. Bugfixes for the old firmware
259 are not provided by Broadcom anymore.
260 Who: Michael Buesch <mb@bu3sch.de>
262 ---------------------------
264 What: Solaris/SunOS syscall and binary support on Sparc
266 Why: Largely unmaintained and almost entirely unused. File system
267 layering used to divert library and dynamic linker searches to
268 /usr/gnemul is extremely buggy and unfixable. Making it work
269 is largely pointless as without a lot of work only the most
270 trivial of Solaris binaries can work with the emulation code.
271 Who: David S. Miller <davem@davemloft.net>
273 ---------------------------
277 Why: Not used in-tree. The current out-of-tree users used it to
278 work around problems in the CPA code which should be resolved
279 by now. One usecase was described to provide verification code
280 of the CPA operation. That's a good idea in general, but such
281 code / infrastructure should be in the kernel and not in some
283 Who: Thomas Gleixner <tglx@linutronix.de>
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