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.
   6
   7 ---------------------------
   8
   9 What:   V4L2 VIDIOC_G_MPEGCOMP and VIDIOC_S_MPEGCOMP
  10 When:   October 2007
  11 Why:    Broken attempt to set MPEG compression parameters. These ioctls are
  12         not able to implement the wide variety of parameters that can be set
  13         by hardware MPEG encoders. A new MPEG control mechanism was created
  14         in kernel 2.6.18 that replaces these ioctls. See the V4L2 specification
  15         (section 1.9: Extended controls) for more information on this topic.
  16 Who:    Hans Verkuil <hverkuil@xs4all.nl> and
  17         Mauro Carvalho Chehab <mchehab@infradead.org>
  18
  19 ---------------------------
  20
  21 What:   /sys/devices/.../power/state
  22         dev->power.power_state
  23         dpm_runtime_{suspend,resume)()
  24 When:   July 2007
  25 Why:    Broken design for runtime control over driver power states, confusing
  26         driver-internal runtime power management with:  mechanisms to support
  27         system-wide sleep state transitions; event codes that distinguish
  28         different phases of swsusp "sleep" transitions; and userspace policy
  29         inputs.  This framework was never widely used, and most attempts to
  30         use it were broken.  Drivers should instead be exposing domain-specific
  31         interfaces either to kernel or to userspace.
  32 Who:    Pavel Machek <pavel@suse.cz>
  33
  34 ---------------------------
  35
  36 What:   RAW driver (CONFIG_RAW_DRIVER)
  37 When:   December 2005
  38 Why:    declared obsolete since kernel 2.6.3
  39         O_DIRECT can be used instead
  40 Who:    Adrian Bunk <bunk@stusta.de>
  41
  42 ---------------------------
  43
  44 What:   raw1394: requests of type RAW1394_REQ_ISO_SEND, RAW1394_REQ_ISO_LISTEN
  45 When:   June 2007
  46 Why:    Deprecated in favour of the more efficient and robust rawiso interface.
  47         Affected are applications which use the deprecated part of libraw1394
  48         (raw1394_iso_write, raw1394_start_iso_write, raw1394_start_iso_rcv,
  49         raw1394_stop_iso_rcv) or bypass libraw1394.
  50 Who:    Dan Dennedy <dan@dennedy.org>, Stefan Richter <stefanr@s5r6.in-berlin.de>
  51
  52 ---------------------------
  53
  54 What:   Video4Linux API 1 ioctls and video_decoder.h from Video devices.
  55 When:   December 2006
  56 Why:    V4L1 AP1 was replaced by V4L2 API. during migration from 2.4 to 2.6
  57         series. The old API have lots of drawbacks and don't provide enough
  58         means to work with all video and audio standards. The newer API is
  59         already available on the main drivers and should be used instead.
  60         Newer drivers should use v4l_compat_translate_ioctl function to handle
  61         old calls, replacing to newer ones.
  62         Decoder iocts are using internally to allow video drivers to
  63         communicate with video decoders. This should also be improved to allow
  64         V4L2 calls being translated into compatible internal ioctls.
  65 Who:    Mauro Carvalho Chehab <mchehab@brturbo.com.br>
  66
  67 ---------------------------
  68
  69 What:   PCMCIA control ioctl (needed for pcmcia-cs [cardmgr, cardctl])
  70 When:   November 2005
  71 Files:  drivers/pcmcia/: pcmcia_ioctl.c
  72 Why:    With the 16-bit PCMCIA subsystem now behaving (almost) like a
  73         normal hotpluggable bus, and with it using the default kernel
  74         infrastructure (hotplug, driver core, sysfs) keeping the PCMCIA
  75         control ioctl needed by cardmgr and cardctl from pcmcia-cs is
  76         unnecessary, and makes further cleanups and integration of the
  77         PCMCIA subsystem into the Linux kernel device driver model more
  78         difficult. The features provided by cardmgr and cardctl are either
  79         handled by the kernel itself now or are available in the new
  80         pcmciautils package available at
  81         http://kernel.org/pub/linux/utils/kernel/pcmcia/
  82 Who:    Dominik Brodowski <linux@brodo.de>
  83
  84 ---------------------------
  85
  86 What:   remove EXPORT_SYMBOL(kernel_thread)
  87 When:   August 2006
  88 Files:  arch/*/kernel/*_ksyms.c
  89 Why:    kernel_thread is a low-level implementation detail.  Drivers should
  90         use the <linux/kthread.h> API instead which shields them from
  91         implementation details and provides a higherlevel interface that
  92         prevents bugs and code duplication
  93 Who:    Christoph Hellwig <hch@lst.de>
  94
  95 ---------------------------
  96
  97 What:   CONFIG_FORCED_INLINING
  98 When:   June 2006
  99 Why:    Config option is there to see if gcc is good enough. (in january
 100         2006). If it is, the behavior should just be the default. If it's not,
 101         the option should just go away entirely.
 102 Who:    Arjan van de Ven
 103
 104 ---------------------------
 105
 106 What:   eepro100 network driver
 107 When:   January 2007
 108 Why:    replaced by the e100 driver
 109 Who:    Adrian Bunk <bunk@stusta.de>
 110
 111 ---------------------------
 112
 113 What:  drivers depending on OSS_OBSOLETE_DRIVER
 114 When:  options in 2.6.20, code in 2.6.22
 115 Why:   OSS drivers with ALSA replacements
 116 Who:   Adrian Bunk <bunk@stusta.de>
 117
 118 ---------------------------
 119
 120 What:   pci_module_init(driver)
 121 When:   January 2007
 122 Why:    Is replaced by pci_register_driver(pci_driver).
 123 Who:    Richard Knutsson <ricknu-0@student.ltu.se> and Greg Kroah-Hartman <gregkh@suse.de>
 124
 125 ---------------------------
 126
 127 What:   Usage of invalid timevals in setitimer
 128 When:   March 2007
 129 Why:    POSIX requires to validate timevals in the setitimer call. This
 130         was never done by Linux. The invalid (e.g. negative timevals) were
 131         silently converted to more or less random timeouts and intervals.
 132         Until the removal a per boot limited number of warnings is printed
 133         and the timevals are sanitized.
 134
 135 Who:    Thomas Gleixner <tglx@linutronix.de>
 136
 137 ---------------------------
 138
 139 What:   Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
 140         (temporary transition config option provided until then)
 141         The transition config option will also be removed at the same time.
 142 When:   before 2.6.19
 143 Why:    Unused symbols are both increasing the size of the kernel binary
 144         and are often a sign of "wrong API"
 145 Who:    Arjan van de Ven <arjan@linux.intel.com>
 146
 147 ---------------------------
 148
 149 What:   mount/umount uevents
 150 When:   February 2007
 151 Why:    These events are not correct, and do not properly let userspace know
 152         when a file system has been mounted or unmounted.  Userspace should
 153         poll the /proc/mounts file instead to detect this properly.
 154 Who:    Greg Kroah-Hartman <gregkh@suse.de>
 155
 156 ---------------------------
 157
 158 What:   USB driver API moves to EXPORT_SYMBOL_GPL
 159 When:   February 2008
 160 Files:  include/linux/usb.h, drivers/usb/core/driver.c
 161 Why:    The USB subsystem has changed a lot over time, and it has been
 162         possible to create userspace USB drivers using usbfs/libusb/gadgetfs
 163         that operate as fast as the USB bus allows.  Because of this, the USB
 164         subsystem will not be allowing closed source kernel drivers to
 165         register with it, after this grace period is over.  If anyone needs
 166         any help in converting their closed source drivers over to use the
 167         userspace filesystems, please contact the
 168         linux-usb-devel@lists.sourceforge.net mailing list, and the developers
 169         there will be glad to help you out.
 170 Who:    Greg Kroah-Hartman <gregkh@suse.de>
 171
 172 ---------------------------
 173
 174 What:   Interrupt only SA_* flags
 175 When:   Januar 2007
 176 Why:    The interrupt related SA_* flags are replaced by IRQF_* to move them
 177         out of the signal namespace.
 178
 179 Who:    Thomas Gleixner <tglx@linutronix.de>
 180
 181 ---------------------------
 182
 183 What:   PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
 184 When:   October 2008
 185 Why:    The stacking of class devices makes these values misleading and
 186         inconsistent.
 187         Class devices should not carry any of these properties, and bus
 188         devices have SUBSYTEM and DRIVER as a replacement.
 189 Who:    Kay Sievers <kay.sievers@suse.de>
 190
 191 ---------------------------
 192
 193 What:   i2c-isa
 194 When:   December 2006
 195 Why:    i2c-isa is a non-sense and doesn't fit in the device driver
 196         model. Drivers relying on it are better implemented as platform
 197         drivers.
 198 Who:    Jean Delvare <khali@linux-fr.org>
 199
 200 ---------------------------
 201
 202 What:   i2c_adapter.dev
 203         i2c_adapter.list
 204 When:   July 2007
 205 Why:    Superfluous, given i2c_adapter.class_dev:
 206           * The "dev" was a stand-in for the physical device node that legacy
 207             drivers would not have; but now it's almost always present.  Any
 208             remaining legacy drivers must upgrade (they now trigger warnings).
 209           * The "list" duplicates class device children.
 210         The delay in removing this is so upgraded lm_sensors and libsensors
 211         can get deployed.  (Removal causes minor changes in the sysfs layout,
 212         notably the location of the adapter type name and parenting the i2c
 213         client hardware directly from their controller.)
 214 Who:    Jean Delvare <khali@linux-fr.org>,
 215         David Brownell <dbrownell@users.sourceforge.net>
 216
 217 ---------------------------
 218
 219 What:  drivers depending on OBSOLETE_OSS
 220 When:  options in 2.6.22, code in 2.6.24
 221 Why:   OSS drivers with ALSA replacements
 222 Who:   Adrian Bunk <bunk@stusta.de>
 223
 224 ---------------------------
 225
 226 What:   ACPI hooks (X86_SPEEDSTEP_CENTRINO_ACPI) in speedstep-centrino driver
 227 When:   December 2006
 228 Why:    Speedstep-centrino driver with ACPI hooks and acpi-cpufreq driver are
 229         functionally very much similar. They talk to ACPI in same way. Only
 230         difference between them is the way they do frequency transitions.
 231         One uses MSRs and the other one uses IO ports. Functionaliy of
 232         speedstep_centrino with ACPI hooks is now merged into acpi-cpufreq.
 233         That means one common driver will support all Intel Enhanced Speedstep
 234         capable CPUs. That means less confusion over name of
 235         speedstep-centrino driver (with that driver supposed to be used on
 236         non-centrino platforms). That means less duplication of code and
 237         less maintenance effort and no possibility of these two drivers
 238         going out of sync.
 239         Current users of speedstep_centrino with ACPI hooks are requested to
 240         switch over to acpi-cpufreq driver. speedstep-centrino will continue
 241         to work using older non-ACPI static table based scheme even after this
 242         date.
 243
 244 Who:    Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
 245
 246 ---------------------------
 247
 248 What:   /sys/firmware/acpi/namespace
 249 When:   2.6.21
 250 Why:    The ACPI namespace is effectively the symbol list for
 251         the BIOS.  The device names are completely arbitrary
 252         and have no place being exposed to user-space.
 253
 254         For those interested in the BIOS ACPI namespace,
 255         the BIOS can be extracted and disassembled with acpidump
 256         and iasl as documented in the pmtools package here:
 257         http://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/utils
 258 Who:    Len Brown <len.brown@intel.com>
 259
 260 ---------------------------
 261
 262 What:   ACPI procfs interface
 263 When:   July 2007
 264 Why:    After ACPI sysfs conversion, ACPI attributes will be duplicated
 265         in sysfs and the ACPI procfs interface should be removed.
 266 Who:    Zhang Rui <rui.zhang@intel.com>
 267
 268 ---------------------------
 269
 270 What:   /proc/acpi/button
 271 When:   August 2007
 272 Why:    /proc/acpi/button has been replaced by events to the input layer
 273         since 2.6.20.
 274 Who:    Len Brown <len.brown@intel.com>
 275
 276 ---------------------------
 277
 278 What:   sk98lin network driver
 279 When:   July 2007
 280 Why:    In kernel tree version of driver is unmaintained. Sk98lin driver
 281         replaced by the skge driver.
 282 Who:    Stephen Hemminger <shemminger@osdl.org>
 283
 284 ---------------------------
 285
 286 What:   Compaq touchscreen device emulation
 287 When:   Oct 2007
 288 Files:  drivers/input/tsdev.c
 289 Why:    The code says it was obsolete when it was written in 2001.
 290         tslib is a userspace library which does anything tsdev can do and
 291         much more besides in userspace where this code belongs. There is no
 292         longer any need for tsdev and applications should have converted to
 293         use tslib by now.
 294         The name "tsdev" is also extremely confusing and lots of people have
 295         it loaded when they don't need/use it.
 296 Who:    Richard Purdie <rpurdie@rpsys.net>
 297
 298 ---------------------------
 299
 300 What:   i8xx_tco watchdog driver
 301 When:   in 2.6.22
 302 Why:    the i8xx_tco watchdog driver has been replaced by the iTCO_wdt
 303         watchdog driver.
 304 Who:    Wim Van Sebroeck <wim@iguana.be>
 305
 306 ---------------------------
 307
 308 What:   Multipath cached routing support in ipv4
 309 When:   in 2.6.23
 310 Why:    Code was merged, then submitter immediately disappeared leaving
 311         us with no maintainer and lots of bugs.  The code should not have
 312         been merged in the first place, and many aspects of it's
 313         implementation are blocking more critical core networking
 314         development.  It's marked EXPERIMENTAL and no distribution
 315         enables it because it cause obscure crashes due to unfixable bugs
 316         (interfaces don't return errors so memory allocation can't be
 317         handled, calling contexts of these interfaces make handling
 318         errors impossible too because they get called after we've
 319         totally commited to creating a route object, for example).
 320         This problem has existed for years and no forward progress
 321         has ever been made, and nobody steps up to try and salvage
 322         this code, so we're going to finally just get rid of it.
 323 Who:    David S. Miller <davem@davemloft.net>
 324
 325 ---------------------------