Root/Documentation/feature-removal-schedule.txt

Source at commit 2604e7f9a98c27be50a0c3ff7503b6a5ea8f6cfe created 12 years 7 months ago.
By Maarten ter Huurne, cpufreq_stats: Support runtime changes to frequency table
1The following is a list of files and features that are going to be
2removed in the kernel source tree. Every entry should contain what
3exactly is going away, why it is happening, and who is going to be doing
4the work. When the feature is removed from the kernel, it should also
5be removed from this file. The suggested deprecation period is 3 releases.
6
7---------------------------
8
9What: ddebug_query="query" boot cmdline param
10When: v3.8
11Why: obsoleted by dyndbg="query" and module.dyndbg="query"
12Who: Jim Cromie <jim.cromie@gmail.com>, Jason Baron <jbaron@redhat.com>
13
14---------------------------
15
16What: /proc/sys/vm/nr_pdflush_threads
17When: 2012
18Why: Since pdflush is deprecated, the interface exported in /proc/sys/vm/
19     should be removed.
20Who: Wanpeng Li <liwp@linux.vnet.ibm.com>
21
22---------------------------
23
24What: CONFIG_APM_CPU_IDLE, and its ability to call APM BIOS in idle
25When: 2012
26Why: This optional sub-feature of APM is of dubious reliability,
27    and ancient APM laptops are likely better served by calling HLT.
28    Deleting CONFIG_APM_CPU_IDLE allows x86 to stop exporting
29    the pm_idle function pointer to modules.
30Who: Len Brown <len.brown@intel.com>
31
32----------------------------
33
34What: x86_32 "no-hlt" cmdline param
35When: 2012
36Why: remove a branch from idle path, simplify code used by everybody.
37    This option disabled the use of HLT in idle and machine_halt()
38    for hardware that was flakey 15-years ago. Today we have
39    "idle=poll" that removed HLT from idle, and so if such a machine
40    is still running the upstream kernel, "idle=poll" is likely sufficient.
41Who: Len Brown <len.brown@intel.com>
42
43----------------------------
44
45What: x86 "idle=mwait" cmdline param
46When: 2012
47Why: simplify x86 idle code
48Who: Len Brown <len.brown@intel.com>
49
50----------------------------
51
52What: PRISM54
53When: 2.6.34
54
55Why: prism54 FullMAC PCI / Cardbus devices used to be supported only by the
56    prism54 wireless driver. After Intersil stopped selling these
57    devices in preference for the newer more flexible SoftMAC devices
58    a SoftMAC device driver was required and prism54 did not support
59    them. The p54pci driver now exists and has been present in the kernel for
60    a while. This driver supports both SoftMAC devices and FullMAC devices.
61    The main difference between these devices was the amount of memory which
62    could be used for the firmware. The SoftMAC devices support a smaller
63    amount of memory. Because of this the SoftMAC firmware fits into FullMAC
64    devices's memory. p54pci supports not only PCI / Cardbus but also USB
65    and SPI. Since p54pci supports all devices prism54 supports
66    you will have a conflict. I'm not quite sure how distributions are
67    handling this conflict right now. prism54 was kept around due to
68    claims users may experience issues when using the SoftMAC driver.
69    Time has passed users have not reported issues. If you use prism54
70    and for whatever reason you cannot use p54pci please let us know!
71    E-mail us at: linux-wireless@vger.kernel.org
72
73    For more information see the p54 wiki page:
74
75    http://wireless.kernel.org/en/users/Drivers/p54
76
77Who: Luis R. Rodriguez <lrodriguez@atheros.com>
78
79---------------------------
80
81What: The ieee80211_regdom module parameter
82When: March 2010 / desktop catchup
83
84Why: This was inherited by the CONFIG_WIRELESS_OLD_REGULATORY code,
85    and currently serves as an option for users to define an
86    ISO / IEC 3166 alpha2 code for the country they are currently
87    present in. Although there are userspace API replacements for this
88    through nl80211 distributions haven't yet caught up with implementing
89    decent alternatives through standard GUIs. Although available as an
90    option through iw or wpa_supplicant its just a matter of time before
91    distributions pick up good GUI options for this. The ideal solution
92    would actually consist of intelligent designs which would do this for
93    the user automatically even when travelling through different countries.
94    Until then we leave this module parameter as a compromise.
95
96    When userspace improves with reasonable widely-available alternatives for
97    this we will no longer need this module parameter. This entry hopes that
98    by the super-futuristically looking date of "March 2010" we will have
99    such replacements widely available.
100
101Who: Luis R. Rodriguez <lrodriguez@atheros.com>
102
103---------------------------
104
105What: dev->power.power_state
106When: July 2007
107Why: Broken design for runtime control over driver power states, confusing
108    driver-internal runtime power management with: mechanisms to support
109    system-wide sleep state transitions; event codes that distinguish
110    different phases of swsusp "sleep" transitions; and userspace policy
111    inputs. This framework was never widely used, and most attempts to
112    use it were broken. Drivers should instead be exposing domain-specific
113    interfaces either to kernel or to userspace.
114Who: Pavel Machek <pavel@ucw.cz>
115
116---------------------------
117
118What: /proc/<pid>/oom_adj
119When: August 2012
120Why: /proc/<pid>/oom_adj allows userspace to influence the oom killer's
121    badness heuristic used to determine which task to kill when the kernel
122    is out of memory.
123
124    The badness heuristic has since been rewritten since the introduction of
125    this tunable such that its meaning is deprecated. The value was
126    implemented as a bitshift on a score generated by the badness()
127    function that did not have any precise units of measure. With the
128    rewrite, the score is given as a proportion of available memory to the
129    task allocating pages, so using a bitshift which grows the score
130    exponentially is, thus, impossible to tune with fine granularity.
131
132    A much more powerful interface, /proc/<pid>/oom_score_adj, was
133    introduced with the oom killer rewrite that allows users to increase or
134    decrease the badness score linearly. This interface will replace
135    /proc/<pid>/oom_adj.
136
137    A warning will be emitted to the kernel log if an application uses this
138    deprecated interface. After it is printed once, future warnings will be
139    suppressed until the kernel is rebooted.
140
141---------------------------
142
143What: remove EXPORT_SYMBOL(kernel_thread)
144When: August 2006
145Files: arch/*/kernel/*_ksyms.c
146Check: kernel_thread
147Why: kernel_thread is a low-level implementation detail. Drivers should
148        use the <linux/kthread.h> API instead which shields them from
149    implementation details and provides a higherlevel interface that
150    prevents bugs and code duplication
151Who: Christoph Hellwig <hch@lst.de>
152
153---------------------------
154
155What: Unused EXPORT_SYMBOL/EXPORT_SYMBOL_GPL exports
156    (temporary transition config option provided until then)
157    The transition config option will also be removed at the same time.
158When: before 2.6.19
159Why: Unused symbols are both increasing the size of the kernel binary
160    and are often a sign of "wrong API"
161Who: Arjan van de Ven <arjan@linux.intel.com>
162
163---------------------------
164
165What: PHYSDEVPATH, PHYSDEVBUS, PHYSDEVDRIVER in the uevent environment
166When: October 2008
167Why: The stacking of class devices makes these values misleading and
168    inconsistent.
169    Class devices should not carry any of these properties, and bus
170    devices have SUBSYTEM and DRIVER as a replacement.
171Who: Kay Sievers <kay.sievers@suse.de>
172
173---------------------------
174
175What: ACPI procfs interface
176When: July 2008
177Why: ACPI sysfs conversion should be finished by January 2008.
178    ACPI procfs interface will be removed in July 2008 so that
179    there is enough time for the user space to catch up.
180Who: Zhang Rui <rui.zhang@intel.com>
181
182---------------------------
183
184What: CONFIG_ACPI_PROCFS_POWER
185When: 2.6.39
186Why: sysfs I/F for ACPI power devices, including AC and Battery,
187        has been working in upstream kernel since 2.6.24, Sep 2007.
188    In 2.6.37, we make the sysfs I/F always built in and this option
189    disabled by default.
190    Remove this option and the ACPI power procfs interface in 2.6.39.
191Who: Zhang Rui <rui.zhang@intel.com>
192
193---------------------------
194
195What: /proc/acpi/event
196When: February 2008
197Why: /proc/acpi/event has been replaced by events via the input layer
198    and netlink since 2.6.23.
199Who: Len Brown <len.brown@intel.com>
200
201---------------------------
202
203What: i386/x86_64 bzImage symlinks
204When: April 2010
205
206Why: The i386/x86_64 merge provides a symlink to the old bzImage
207    location so not yet updated user space tools, e.g. package
208    scripts, do not break.
209Who: Thomas Gleixner <tglx@linutronix.de>
210
211---------------------------
212
213What: GPIO autorequest on gpio_direction_{input,output}() in gpiolib
214When: February 2010
215Why: All callers should use explicit gpio_request()/gpio_free().
216    The autorequest mechanism in gpiolib was provided mostly as a
217    migration aid for legacy GPIO interfaces (for SOC based GPIOs).
218    Those users have now largely migrated. Platforms implementing
219    the GPIO interfaces without using gpiolib will see no changes.
220Who: David Brownell <dbrownell@users.sourceforge.net>
221---------------------------
222
223What: b43 support for firmware revision < 410
224When: The schedule was July 2008, but it was decided that we are going to keep the
225        code as long as there are no major maintanance headaches.
226    So it _could_ be removed _any_ time now, if it conflicts with something new.
227Why: The support code for the old firmware hurts code readability/maintainability
228    and slightly hurts runtime performance. Bugfixes for the old firmware
229    are not provided by Broadcom anymore.
230Who: Michael Buesch <m@bues.ch>
231
232---------------------------
233
234What: Ability for non root users to shm_get hugetlb pages based on mlock
235    resource limits
236When: 2.6.31
237Why: Non root users need to be part of /proc/sys/vm/hugetlb_shm_group or
238    have CAP_IPC_LOCK to be able to allocate shm segments backed by
239    huge pages. The mlock based rlimit check to allow shm hugetlb is
240    inconsistent with mmap based allocations. Hence it is being
241    deprecated.
242Who: Ravikiran Thirumalai <kiran@scalex86.org>
243
244---------------------------
245
246What: sysfs ui for changing p4-clockmod parameters
247When: September 2009
248Why: See commits 129f8ae9b1b5be94517da76009ea956e89104ce8 and
249    e088e4c9cdb618675874becb91b2fd581ee707e6.
250    Removal is subject to fixing any remaining bugs in ACPI which may
251    cause the thermal throttling not to happen at the right time.
252Who: Dave Jones <davej@redhat.com>, Matthew Garrett <mjg@redhat.com>
253
254-----------------------------
255
256What: fakephp and associated sysfs files in /sys/bus/pci/slots/
257When: 2011
258Why: In 2.6.27, the semantics of /sys/bus/pci/slots was redefined to
259    represent a machine's physical PCI slots. The change in semantics
260    had userspace implications, as the hotplug core no longer allowed
261    drivers to create multiple sysfs files per physical slot (required
262    for multi-function devices, e.g.). fakephp was seen as a developer's
263    tool only, and its interface changed. Too late, we learned that
264    there were some users of the fakephp interface.
265
266    In 2.6.30, the original fakephp interface was restored. At the same
267    time, the PCI core gained the ability that fakephp provided, namely
268    function-level hot-remove and hot-add.
269
270    Since the PCI core now provides the same functionality, exposed in:
271
272        /sys/bus/pci/rescan
273        /sys/bus/pci/devices/.../remove
274        /sys/bus/pci/devices/.../rescan
275
276    there is no functional reason to maintain fakephp as well.
277
278    We will keep the existing module so that 'modprobe fakephp' will
279    present the old /sys/bus/pci/slots/... interface for compatibility,
280    but users are urged to migrate their applications to the API above.
281
282    After a reasonable transition period, we will remove the legacy
283    fakephp interface.
284Who: Alex Chiang <achiang@hp.com>
285
286---------------------------
287
288What: CONFIG_RFKILL_INPUT
289When: 2.6.33
290Why: Should be implemented in userspace, policy daemon.
291Who: Johannes Berg <johannes@sipsolutions.net>
292
293----------------------------
294
295What: sound-slot/service-* module aliases and related clutters in
296    sound/sound_core.c
297When: August 2010
298Why: OSS sound_core grabs all legacy minors (0-255) of SOUND_MAJOR
299    (14) and requests modules using custom sound-slot/service-*
300    module aliases. The only benefit of doing this is allowing
301    use of custom module aliases which might as well be considered
302    a bug at this point. This preemptive claiming prevents
303    alternative OSS implementations.
304
305    Till the feature is removed, the kernel will be requesting
306    both sound-slot/service-* and the standard char-major-* module
307    aliases and allow turning off the pre-claiming selectively via
308    CONFIG_SOUND_OSS_CORE_PRECLAIM and soundcore.preclaim_oss
309    kernel parameter.
310
311    After the transition phase is complete, both the custom module
312    aliases and switches to disable it will go away. This removal
313    will also allow making ALSA OSS emulation independent of
314    sound_core. The dependency will be broken then too.
315Who: Tejun Heo <tj@kernel.org>
316
317----------------------------
318
319What: sysfs-class-rfkill state file
320When: Feb 2014
321Files: net/rfkill/core.c
322Why: Documented as obsolete since Feb 2010. This file is limited to 3
323    states while the rfkill drivers can have 4 states.
324Who: anybody or Florian Mickler <florian@mickler.org>
325
326----------------------------
327
328What: sysfs-class-rfkill claim file
329When: Feb 2012
330Files: net/rfkill/core.c
331Why: It is not possible to claim an rfkill driver since 2007. This is
332    Documented as obsolete since Feb 2010.
333Who: anybody or Florian Mickler <florian@mickler.org>
334
335----------------------------
336
337What: iwlwifi 50XX module parameters
338When: 3.0
339Why: The "..50" modules parameters were used to configure 5000 series and
340    up devices; different set of module parameters also available for 4965
341    with same functionalities. Consolidate both set into single place
342    in drivers/net/wireless/iwlwifi/iwl-agn.c
343
344Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
345
346----------------------------
347
348What: iwl4965 alias support
349When: 3.0
350Why: Internal alias support has been present in module-init-tools for some
351    time, the MODULE_ALIAS("iwl4965") boilerplate aliases can be removed
352    with no impact.
353
354Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
355
356---------------------------
357
358What: xt_NOTRACK
359Files: net/netfilter/xt_NOTRACK.c
360When: April 2011
361Why: Superseded by xt_CT
362Who: Netfilter developer team <netfilter-devel@vger.kernel.org>
363
364----------------------------
365
366What: IRQF_DISABLED
367When: 2.6.36
368Why: The flag is a NOOP as we run interrupt handlers with interrupts disabled
369Who: Thomas Gleixner <tglx@linutronix.de>
370
371----------------------------
372
373What: PCI DMA unmap state API
374When: August 2012
375Why: PCI DMA unmap state API (include/linux/pci-dma.h) was replaced
376    with DMA unmap state API (DMA unmap state API can be used for
377    any bus).
378Who: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
379
380----------------------------
381
382What: iwlwifi disable_hw_scan module parameters
383When: 3.0
384Why: Hareware scan is the prefer method for iwlwifi devices for
385    scanning operation. Remove software scan support for all the
386    iwlwifi devices.
387
388Who: Wey-Yi Guy <wey-yi.w.guy@intel.com>
389
390----------------------------
391
392What: Legacy, non-standard chassis intrusion detection interface.
393When: June 2011
394Why: The adm9240, w83792d and w83793 hardware monitoring drivers have
395    legacy interfaces for chassis intrusion detection. A standard
396    interface has been added to each driver, so the legacy interface
397    can be removed.
398Who: Jean Delvare <khali@linux-fr.org>
399
400----------------------------
401
402What: i2c_driver.attach_adapter
403    i2c_driver.detach_adapter
404When: September 2011
405Why: These legacy callbacks should no longer be used as i2c-core offers
406    a variety of preferable alternative ways to instantiate I2C devices.
407Who: Jean Delvare <khali@linux-fr.org>
408
409----------------------------
410
411What: Opening a radio device node will no longer automatically switch the
412    tuner mode from tv to radio.
413When: 3.3
414Why: Just opening a V4L device should not change the state of the hardware
415    like that. It's very unexpected and against the V4L spec. Instead, you
416    switch to radio mode by calling VIDIOC_S_FREQUENCY. This is the second
417    and last step of the move to consistent handling of tv and radio tuners.
418Who: Hans Verkuil <hans.verkuil@cisco.com>
419
420----------------------------
421
422What: CONFIG_CFG80211_WEXT
423When: as soon as distributions ship new wireless tools, ie. wpa_supplicant 1.0
424    and NetworkManager/connman/etc. that are able to use nl80211
425Why: Wireless extensions are deprecated, and userland tools are moving to
426    using nl80211. New drivers are no longer using wireless extensions,
427    and while there might still be old drivers, both new drivers and new
428    userland no longer needs them and they can't be used for an feature
429    developed in the past couple of years. As such, compatibility with
430    wireless extensions in new drivers will be removed.
431Who: Johannes Berg <johannes@sipsolutions.net>
432
433----------------------------
434
435What: g_file_storage driver
436When: 3.8
437Why: This driver has been superseded by g_mass_storage.
438Who: Alan Stern <stern@rowland.harvard.edu>
439
440----------------------------
441
442What: threeg and interface sysfs files in /sys/devices/platform/acer-wmi
443When: 2012
444Why: In 3.0, we can now autodetect internal 3G device and already have
445    the threeg rfkill device. So, we plan to remove threeg sysfs support
446    for it's no longer necessary.
447
448    We also plan to remove interface sysfs file that exposed which ACPI-WMI
449    interface that was used by acer-wmi driver. It will replaced by
450    information log when acer-wmi initial.
451Who: Lee, Chun-Yi <jlee@novell.com>
452
453---------------------------
454
455What: /sys/devices/platform/_UDC_/udc/_UDC_/is_dualspeed file and
456    is_dualspeed line in /sys/devices/platform/ci13xxx_*/udc/device file.
457When: 3.8
458Why: The is_dualspeed file is superseded by maximum_speed in the same
459    directory and is_dualspeed line in device file is superseded by
460    max_speed line in the same file.
461
462    The maximum_speed/max_speed specifies maximum speed supported by UDC.
463    To check if dualspeeed is supported, check if the value is >= 3.
464    Various possible speeds are defined in <linux/usb/ch9.h>.
465Who: Michal Nazarewicz <mina86@mina86.com>
466
467----------------------------
468
469What: The XFS nodelaylog mount option
470When: 3.3
471Why: The delaylog mode that has been the default since 2.6.39 has proven
472    stable, and the old code is in the way of additional improvements in
473    the log code.
474Who: Christoph Hellwig <hch@lst.de>
475
476----------------------------
477
478What: iwlagn alias support
479When: 3.5
480Why: The iwlagn module has been renamed iwlwifi. The alias will be around
481    for backward compatibility for several cycles and then dropped.
482Who: Don Fry <donald.h.fry@intel.com>
483
484----------------------------
485
486What: pci_scan_bus_parented()
487When: 3.5
488Why: The pci_scan_bus_parented() interface creates a new root bus. The
489    bus is created with default resources (ioport_resource and
490    iomem_resource) that are always wrong, so we rely on arch code to
491    correct them later. Callers of pci_scan_bus_parented() should
492    convert to using pci_scan_root_bus() so they can supply a list of
493    bus resources when the bus is created.
494Who: Bjorn Helgaas <bhelgaas@google.com>
495
496----------------------------
497
498What: Low Performance USB Block driver ("CONFIG_BLK_DEV_UB")
499When: 3.6
500Why: This driver provides support for USB storage devices like "USB
501    sticks". As of now, it is deactivated in Debian, Fedora and
502        Ubuntu. All current users can switch over to usb-storage
503        (CONFIG_USB_STORAGE) which only drawback is the additional SCSI
504        stack.
505Who: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
506
507----------------------------
508
509What: get_robust_list syscall
510When: 2013
511Why: There appear to be no production users of the get_robust_list syscall,
512    and it runs the risk of leaking address locations, allowing the bypass
513    of ASLR. It was only ever intended for debugging, so it should be
514    removed.
515Who: Kees Cook <keescook@chromium.org>
516
517----------------------------
518
519What: Removing the pn544 raw driver.
520When: 3.6
521Why: With the introduction of the NFC HCI and SHDL kernel layers, pn544.c
522    is being replaced by pn544_hci.c which is accessible through the netlink
523    and socket NFC APIs. Moreover, pn544.c is outdated and does not seem to
524    work properly with the latest Android stacks.
525    Having 2 drivers for the same hardware is confusing and as such we
526    should only keep the one following the kernel NFC APIs.
527Who: Samuel Ortiz <sameo@linux.intel.com>
528
529----------------------------
530
531What: setitimer accepts user NULL pointer (value)
532When: 3.6
533Why: setitimer is not returning -EFAULT if user pointer is NULL. This
534    violates the spec.
535Who: Sasikantha Babu <sasikanth.v19@gmail.com>
536
537----------------------------
538
539What: remove bogus DV presets V4L2_DV_1080I29_97, V4L2_DV_1080I30 and
540    V4L2_DV_1080I25
541When: 3.6
542Why: These HDTV formats do not exist and were added by a confused mind
543    (that was me, to be precise...)
544Who: Hans Verkuil <hans.verkuil@cisco.com>
545
546----------------------------
547
548What: V4L2_CID_HCENTER, V4L2_CID_VCENTER V4L2 controls
549When: 3.7
550Why: The V4L2_CID_VCENTER, V4L2_CID_HCENTER controls have been deprecated
551    for about 4 years and they are not used by any mainline driver.
552    There are newer controls (V4L2_CID_PAN*, V4L2_CID_TILT*) that provide
553    similar functionality.
554Who: Sylwester Nawrocki <sylvester.nawrocki@gmail.com>
555
556----------------------------
557
558What: cgroup option updates via remount
559When: March 2013
560Why: Remount currently allows changing bound subsystems and
561    release_agent. Rebinding is hardly useful as it only works
562    when the hierarchy is empty and release_agent itself should be
563    replaced with conventional fsnotify.
564
565----------------------------
566
567What: xt_recent rev 0
568When: 2013
569Who: Pablo Neira Ayuso <pablo@netfilter.org>
570Files: net/netfilter/xt_recent.c
571
572----------------------------
573
574What: KVM debugfs statistics
575When: 2013
576Why: KVM tracepoints provide mostly equivalent information in a much more
577        flexible fashion.
578
579----------------------------
580
581What: at91-mci driver ("CONFIG_MMC_AT91")
582When: 3.8
583Why: There are two mci drivers: at91-mci and atmel-mci. The PDC support
584    was added to atmel-mci as a first step to support more chips.
585    Then at91-mci was kept only for old IP versions (on at91rm9200 and
586    at91sam9261). The support of these IP versions has just been added
587    to atmel-mci, so atmel-mci can be used for all chips.
588Who: Ludovic Desroches <ludovic.desroches@atmel.com>
589
590----------------------------
591
592What: net/wanrouter/
593When: June 2013
594Why: Unsupported/unmaintained/unused since 2.6
595
596----------------------------
597
598What: V4L2 selections API target rectangle and flags unification, the
599    following definitions will be removed: V4L2_SEL_TGT_CROP_ACTIVE,
600    V4L2_SEL_TGT_COMPOSE_ACTIVE, V4L2_SUBDEV_SEL_*, V4L2_SUBDEV_SEL_FLAG_*
601    in favor of common V4L2_SEL_TGT_* and V4L2_SEL_FLAG_* definitions.
602    For more details see include/linux/v4l2-common.h.
603When: 3.8
604Why: The regular V4L2 selections and the subdev selection API originally
605    defined distinct names for the target rectangles and flags - V4L2_SEL_*
606    and V4L2_SUBDEV_SEL_*. Although, it turned out that the meaning of these
607    target rectangles is virtually identical and the APIs were consolidated
608    to use single set of names - V4L2_SEL_*. This didn't involve any ABI
609    changes. Alias definitions were created for the original ones to avoid
610    any instabilities in the user space interface. After few cycles these
611    backward compatibility definitions will be removed.
612Who: Sylwester Nawrocki <sylvester.nawrocki@gmail.com>
613
614----------------------------
615
616What: Using V4L2_CAP_VIDEO_CAPTURE and V4L2_CAP_VIDEO_OUTPUT flags
617    to indicate a V4L2 memory-to-memory device capability
618When: 3.8
619Why: New drivers should use new V4L2_CAP_VIDEO_M2M capability flag
620    to indicate a V4L2 video memory-to-memory (M2M) device and
621    applications can now identify a M2M video device by checking
622    for V4L2_CAP_VIDEO_M2M, with VIDIOC_QUERYCAP ioctl. Using ORed
623    V4L2_CAP_VIDEO_CAPTURE and V4L2_CAP_VIDEO_OUTPUT flags for M2M
624    devices is ambiguous and may lead, for example, to identifying
625    a M2M device as a video capture or output device.
626Who: Sylwester Nawrocki <s.nawrocki@samsung.com>
627
628----------------------------
629
630What: OMAP private DMA implementation
631When: 2013
632Why: We have a DMA engine implementation; all users should be updated
633    to use this rather than persisting with the old APIs. The old APIs
634    block merging the old DMA engine implementation into the DMA
635    engine driver.
636Who: Russell King <linux@arm.linux.org.uk>,
637    Santosh Shilimkar <santosh.shilimkar@ti.com>
638
639----------------------------
640

Archive Download this file



interactive