linux/Documentation/ABI
Linus Torvalds 69afef4af4 gpio updates for v6.9
Serialization rework:
 - use SRCU to serialize access to the global GPIO device list, to GPIO device
   structs themselves and to GPIO descriptors
 - make the GPIO subsystem resilient to the GPIO providers being unbound while
   the API calls are in progress
 - don't dereference the SRCU-protected chip pointer if the information we need
   can be obtained from the GPIO device structure
 - move some of the information contained in struct gpio_chip to struct
   gpio_device to further reduce the need to dereference the former
 - pass the GPIO device struct instead of the GPIO chip to sysfs callback to,
   again, reduce the need for accessing the latter
 - get GPIO descriptors from the GPIO device, not from the chip for the same
   reason
 - allow for mostly lockless operation of the GPIO driver API: assure
   consistency with SRCU and atomic operations
 - remove the global GPIO spinlock
 - remove the character device RW semaphore
 
 Core GPIOLIB:
 - constify pointers in GPIO API where applicable
 - unify the GPIO counting APIs for ACPI and OF
 - provide a macro for iterating over all GPIOs, not only the ones that are
   requested
 - remove leftover typedefs
 - pass the consumer device to GPIO core in devm_fwnode_gpiod_get_index() for
   improved logging
 - constify the GPIO bus type
 - don't warn about removing GPIO chips with descriptors still held by users as
   we can now handle this situation gracefully
 - remove unused logging helpers
 - unexport functions that are only used internally in the GPIO subsystem
 - set the device type (assign the relevant struct device_type) for GPIO devices
 
 New drivers:
 - add the ChromeOS EC GPIO driver
 
 Driver improvements:
 - allow building gpio-vf610 with COMPILE_TEST as well as disabling it in
   menuconfig (before it was always built for i.MX cofigs)
 - count the number of EICs using the device properties instead of hard-coding
   it in gpio-eic-sprd
 - improve the device naming, extend the debugfs output and add lockdep asserts
   to gpio-sim
 
 DT bindings:
 - document the 'label' property for gpio-pca9570
 - convert aspeed,ast2400-gpio bindings to DT schema
 - disallow unevaluated properties for gpio-mvebu
 - document a new model in renesas,rcar-gpio
 
 Documentation:
 - improve the character device kerneldocs in user-space headers
 - add proper documentation for the character device uAPI (both v1 and v2)
 - move the sysfs and gpio-mockup docs into the "obsolete" section
 - improve naming consistency for GPIO terms
 - clarify the line values description for sysfs
 - minor docs improvements
 - improve the driver API contract for setting GPIO direction
 - mark unsafe APIs as deprecated in kerneldocs and suggest replacements
 
 Other:
 - remove an obsolete test from selftests
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEEFp3rbAvDxGAT0sefEacuoBRx13IFAmXu1ecACgkQEacuoBRx
 13JR7w//R3TswZ1uC9qkRjat9eA2KZUaI2QChlS7V/yXVcDHynuTlO/ZQmnnMdYL
 ch7T2cjPcW0OCt0UhcjamUmYtWaxe1e5GU3E42EosWUsojEzgGs0iNKe0R4SHYzv
 whlkFqO8+8IctYhiMpAU1PzP9N4YBqypwgCrTqHIrYuhz3MbPQxtCMkr7g0LTo8u
 Z3K0D3Y0LuwISWNYYhA20Bwemn1fEHXJ9f3pTeNaGh2dGZek9k9xd0zWcCxwhaYD
 CBTBiZXf57TUTJ2u+JG+au1ghEmmvBPlMpza+fazypbcvsiQxdGvv5QH1bTwyt4B
 woGq+biLLvlwfJ8BT7+09uni7gUyNL3wWkixlx/8Slkyti4xWqgZQ3WnhwN8yS4Y
 DbkTtzH/PIsjr1dZw6rnGoXi80lBEaok7LeI0QhybopTXQI+CnIbE/RBhzly8Mf8
 1cAVFjrF2gPuaTuheakRBw4LOhegf4a485fadJVEUeEpeF7/p9rDQWAbgohYUnCE
 gHPwkTOJuOZp+BlsTOyspnqxWnDVMtCnmi+q1o7JvEgXqAtzU7+1gz/wDpfsHgHQ
 oze6V2JvD2R3JkHmdqcIzq5yNwk1rOguOY3saNiwSk95JY+A8vhAe/gVykklKDXX
 oX/DPwlVd/0OR+0HCQ3r0pXK8BSRQ9qm/nUZNnLB+Rts9K1peIU=
 =LX+L
 -----END PGP SIGNATURE-----

Merge tag 'gpio-updates-for-v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux

Pull gpio updates from Bartosz Golaszewski:
 "The biggest feature is the locking overhaul. Up until now the
  synchronization in the GPIO subsystem was broken. There was a single
  spinlock "protecting" multiple data structures but doing it wrong (as
  evidenced by several places where it would be released when a sleeping
  function was called and then reacquired without checking the protected
  state).

  We tried to use an RW semaphore before but the main issue with GPIO is
  that we have drivers implementing the interfaces in both sleeping and
  non-sleeping ways as well as user-facing interfaces that can be called
  both from process as well as atomic contexts. Both ends converge in
  the same code paths that can use neither spinlocks nor mutexes. The
  only reasonable way out is to use SRCU and go mostly lockless. To that
  end: we add several SRCU structs in relevant places and use them to
  assure consistency between API calls together with atomic reads and
  writes of GPIO descriptor flags where it makes sense.

  This code has spent several weeks in next and has received several
  fixes in the first week or two after which it stabilized nicely. The
  GPIO subsystem is now resilient to providers being suddenly unbound.
  We managed to also remove the existing character device RW semaphore
  and the obsolete global spinlock.

  Other than the locking rework we have one new driver (for Chromebook
  EC), much appreciated documentation improvements from Kent and the
  regular driver improvements, DT-bindings updates and GPIOLIB core
  tweaks.

  Serialization rework:
   - use SRCU to serialize access to the global GPIO device list, to
     GPIO device structs themselves and to GPIO descriptors
   - make the GPIO subsystem resilient to the GPIO providers being
     unbound while the API calls are in progress
   - don't dereference the SRCU-protected chip pointer if the
     information we need can be obtained from the GPIO device structure
   - move some of the information contained in struct gpio_chip to
     struct gpio_device to further reduce the need to dereference the
     former
   - pass the GPIO device struct instead of the GPIO chip to sysfs
     callback to, again, reduce the need for accessing the latter
   - get GPIO descriptors from the GPIO device, not from the chip for
     the same reason
   - allow for mostly lockless operation of the GPIO driver API: assure
     consistency with SRCU and atomic operations
   - remove the global GPIO spinlock
   - remove the character device RW semaphore

  Core GPIOLIB:
   - constify pointers in GPIO API where applicable
   - unify the GPIO counting APIs for ACPI and OF
   - provide a macro for iterating over all GPIOs, not only the ones
     that are requested
   - remove leftover typedefs
   - pass the consumer device to GPIO core in
     devm_fwnode_gpiod_get_index() for improved logging
   - constify the GPIO bus type
   - don't warn about removing GPIO chips with descriptors still held by
     users as we can now handle this situation gracefully
   - remove unused logging helpers
   - unexport functions that are only used internally in the GPIO
     subsystem
   - set the device type (assign the relevant struct device_type) for
     GPIO devices

  New drivers:
   - add the ChromeOS EC GPIO driver

  Driver improvements:
   - allow building gpio-vf610 with COMPILE_TEST as well as disabling it
     in menuconfig (before it was always built for i.MX cofigs)
   - count the number of EICs using the device properties instead of
     hard-coding it in gpio-eic-sprd
   - improve the device naming, extend the debugfs output and add
     lockdep asserts to gpio-sim

  DT bindings:
   - document the 'label' property for gpio-pca9570
   - convert aspeed,ast2400-gpio bindings to DT schema
   - disallow unevaluated properties for gpio-mvebu
   - document a new model in renesas,rcar-gpio

  Documentation:
   - improve the character device kerneldocs in user-space headers
   - add proper documentation for the character device uAPI (both v1 and v2)
   - move the sysfs and gpio-mockup docs into the "obsolete" section
   - improve naming consistency for GPIO terms
   - clarify the line values description for sysfs
   - minor docs improvements
   - improve the driver API contract for setting GPIO direction
   - mark unsafe APIs as deprecated in kerneldocs and suggest
     replacements

  Other:
   - remove an obsolete test from selftests"

* tag 'gpio-updates-for-v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (79 commits)
  gpio: sysfs: repair export returning -EPERM on 1st attempt
  selftest: gpio: remove obsolete gpio-mockup test
  gpiolib: Deduplicate cleanup for-loop in gpiochip_add_data_with_key()
  dt-bindings: gpio: aspeed,ast2400-gpio: Convert to DT schema
  gpio: acpi: Make acpi_gpio_count() take firmware node as a parameter
  gpio: of: Make of_gpio_get_count() take firmware node as a parameter
  gpiolib: Pass consumer device through to core in devm_fwnode_gpiod_get_index()
  gpio: sim: use for_each_hwgpio()
  gpio: provide for_each_hwgpio()
  gpio: don't warn about removing GPIO chips with active users anymore
  gpio: sim: delimit the fwnode name with a ":" when generating labels
  gpio: sim: add lockdep asserts
  gpio: Add ChromeOS EC GPIO driver
  gpio: constify of_phandle_args in of_find_gpio_device_by_xlate()
  gpio: fix memory leak in gpiod_request_commit()
  gpio: constify opaque pointer "data" in gpio_device_find()
  gpio: cdev: fix a NULL-pointer dereference with DEBUG enabled
  gpio: uapi: clarify default_values being logical
  gpio: sysfs: fix inverted pointer logic
  gpio: don't let lockdep complain about inherently dangerous RCU usage
  ...
2024-03-13 11:14:55 -07:00
..
obsolete Documentation: ABI: update sysfs-gpio to reference gpio-cdev 2024-01-22 10:49:03 +01:00
removed docs: update ocfs2-devel mailing list address 2023-07-08 09:29:29 -07:00
stable dmaengine: idxd: add wq driver name support for accel-config user tool 2023-10-04 13:00:34 +05:30
testing gpio updates for v6.9 2024-03-13 11:14:55 -07:00
README

This directory attempts to document the ABI between the Linux kernel and
userspace, and the relative stability of these interfaces.  Due to the
everchanging nature of Linux, and the differing maturity levels, these
interfaces should be used by userspace programs in different ways.

We have four different levels of ABI stability, as shown by the four
different subdirectories in this location.  Interfaces may change levels
of stability according to the rules described below.

The different levels of stability are:

  stable/
	This directory documents the interfaces that the developer has
	defined to be stable.  Userspace programs are free to use these
	interfaces with no restrictions, and backward compatibility for
	them will be guaranteed for at least 2 years.  Most interfaces
	(like syscalls) are expected to never change and always be
	available.

  testing/
	This directory documents interfaces that are felt to be stable,
	as the main development of this interface has been completed.
	The interface can be changed to add new features, but the
	current interface will not break by doing this, unless grave
	errors or security problems are found in them.  Userspace
	programs can start to rely on these interfaces, but they must be
	aware of changes that can occur before these interfaces move to
	be marked stable.  Programs that use these interfaces are
	strongly encouraged to add their name to the description of
	these interfaces, so that the kernel developers can easily
	notify them if any changes occur (see the description of the
	layout of the files below for details on how to do this.)

  obsolete/
	This directory documents interfaces that are still remaining in
	the kernel, but are marked to be removed at some later point in
	time.  The description of the interface will document the reason
	why it is obsolete and when it can be expected to be removed.

  removed/
	This directory contains a list of the old interfaces that have
	been removed from the kernel.

Every file in these directories will contain the following information:

What:		Short description of the interface
Date:		Date created
KernelVersion:	Kernel version this feature first showed up in.
Contact:	Primary contact for this interface (may be a mailing list)
Description:	Long description of the interface and how to use it.
Users:		All users of this interface who wish to be notified when
		it changes.  This is very important for interfaces in
		the "testing" stage, so that kernel developers can work
		with userspace developers to ensure that things do not
		break in ways that are unacceptable.  It is also
		important to get feedback for these interfaces to make
		sure they are working in a proper way and do not need to
		be changed further.


Note:
   The fields should be use a simple notation, compatible with ReST markup.
   Also, the file **should not** have a top-level index, like::

	===
	foo
	===

How things move between levels:

Interfaces in stable may move to obsolete, as long as the proper
notification is given.

Interfaces may be removed from obsolete and the kernel as long as the
documented amount of time has gone by.

Interfaces in the testing state can move to the stable state when the
developers feel they are finished.  They cannot be removed from the
kernel tree without going through the obsolete state first.

It's up to the developer to place their interfaces in the category they
wish for it to start out in.


Notable bits of non-ABI, which should not under any circumstances be considered
stable:

- Kconfig.  Userspace should not rely on the presence or absence of any
  particular Kconfig symbol, in /proc/config.gz, in the copy of .config
  commonly installed to /boot, or in any invocation of the kernel build
  process.

- Kernel-internal symbols.  Do not rely on the presence, absence, location, or
  type of any kernel symbol, either in System.map files or the kernel binary
  itself.  See Documentation/process/stable-api-nonsense.rst.