1 This is a place for planning the ongoing long-term work in the GPIO
7 Starting with commit 79a9becda894 the GPIO subsystem embarked on a journey
8 to move away from the global GPIO numberspace and toward a descriptor-based
9 approach. This means that GPIO consumers, drivers and machine descriptions
10 ideally have no use or idea of the global GPIO numberspace that has/was
11 used in the inception of the GPIO subsystem.
13 The numberspace issue is the same as to why irq is moving away from irq
14 numbers to IRQ descriptors.
16 The underlying motivation for this is that the GPIO numberspace has become
17 unmanageable: machine board files tend to become full of macros trying to
18 establish the numberspace at compile-time, making it hard to add any numbers
19 in the middle (such as if you missed a pin on a chip) without the numberspace
22 Machine descriptions such as device tree or ACPI does not have a concept of the
23 Linux GPIO number as those descriptions are external to the Linux kernel
24 and treat GPIO lines as abstract entities.
26 The runtime-assigned GPIO numberspace (what you get if you assign the GPIO
27 base as -1 in struct gpio_chip) has also became unpredictable due to factors
28 such as probe ordering and the introduction of -EPROBE_DEFER making probe
29 ordering of independent GPIO chips essentially unpredictable, as their base
30 number will be assigned on a first come first serve basis.
32 The best way to get out of the problem is to make the global GPIO numbers
33 unimportant by simply not using them. GPIO descriptors deal with this.
37 - Convert all GPIO device drivers to only #include <linux/gpio/driver.h>
39 - Convert all consumer drivers to only #include <linux/gpio/consumer.h>
41 - Convert all machine descriptors in "boardfiles" to only
42 #include <linux/gpio/machine.h>, the other option being to convert it
43 to a machine description such as device tree, ACPI or fwnode that
44 implicitly does not use global GPIO numbers.
46 - When this work is complete (will require some of the items in the
47 following ongoing work as well) we can delete the old global
48 numberspace accessors from <linux/gpio.h> and eventually delete
49 <linux/gpio.h> altogether.
52 Get rid of <linux/of_gpio.h>
54 This header and helpers appeared at one point when there was no proper
55 driver infrastructure for doing simpler MMIO GPIO devices and there was
56 no core support for parsing device tree GPIOs from the core library with
57 the [devm_]gpiod_get() calls we have today that will implicitly go into
58 the device tree back-end. It is legacy and should not be used in new code.
62 - Get rid of struct of_mm_gpio_chip altogether: use the generic MMIO
63 GPIO for all current users (see below). Delete struct of_mm_gpio_chip,
64 to_of_mm_gpio_chip(), of_mm_gpiochip_add_data(), of_mm_gpiochip_add()
65 of_mm_gpiochip_remove() from the kernel.
67 - Change all consumer drivers that #include <linux/of_gpio.h> to
68 #include <linux/gpio/consumer.h> and stop doing custom parsing of the
69 GPIO lines from the device tree. This can be tricky and often ivolves
70 changing boardfiles, etc.
72 - Pull semantics for legacy device tree (OF) GPIO lookups into
73 gpiolib-of.c: in some cases subsystems are doing custom flags and
74 lookups for polarity inversion, open drain and what not. As we now
75 handle this with generic OF bindings, pull all legacy handling into
76 gpiolib so the library API becomes narrow and deep and handle all
77 legacy bindings internally. (See e.g. commits 6953c57ab172,
80 - Delete <linux/of_gpio.h> when all the above is complete and everything
81 uses <linux/gpio/consumer.h> or <linux/gpio/driver.h> instead.
84 Get rid of <linux/gpio.h>
86 This legacy header is a one stop shop for anything GPIO is closely tied
87 to the global GPIO numberspace. The endgame of the above refactorings will
88 be the removal of <linux/gpio.h> and from that point only the specialized
89 headers under <linux/gpio/*.h> will be used. This requires all the above to
90 be completed and is expected to take a long time.
95 Collect GPIO drivers from arch/* and other places that should be placed
96 in drivers/gpio/gpio-*. Augment platforms to create platform devices or
97 similar and probe a proper driver in the gpiolib subsystem.
99 In some cases it makes sense to create a GPIO chip from the local driver
100 for a few GPIOs. Those should stay where they are.
102 At the same time it makes sense to get rid of code duplication in existing or
103 new coming drivers. For example, gpio-ml-ioh should be incorporated into
109 The GPIO drivers can utilize the generic MMIO helper library in many
110 cases, and the helper library should be as helpful as possible for MMIO
111 drivers. (drivers/gpio/gpio-mmio.c)
115 - Look over and identify any remaining easily converted drivers and
116 dry-code conversions to MMIO GPIO for maintainers to test
118 - Expand the MMIO GPIO or write a new library for regmap-based I/O
119 helpers for GPIO drivers on regmap that simply use offsets
120 0..n in some register to drive GPIO lines
122 - Expand the MMIO GPIO or write a new library for port-mapped I/O
123 helpers (x86 inb()/outb()) and convert port-mapped I/O drivers to use
124 this with dry-coding and sending to maintainers to test
129 The GPIOLIB irqchip is a helper irqchip for "simple cases" that should
130 try to cover any generic kind of irqchip cascaded from a GPIO.
132 - Look over and identify any remaining easily converted drivers and
133 dry-code conversions to gpiolib irqchip for maintainers to test
136 Increase integration with pin control
138 There are already ways to use pin control as back-end for GPIO and
139 it may make sense to bring these subsystems closer. One reason for
140 creating pin control as its own subsystem was that we could avoid any
141 use of the global GPIO numbers. Once the above is complete, it may
142 make sense to simply join the subsystems into one and make pin
143 multiplexing, pin configuration, GPIO, etc selectable options in one
144 and the same pin control and GPIO subsystem.
147 Debugfs in place of sysfs
149 The old sysfs code that enables simple uses of GPIOs from the
150 command line is still popular despite the existance of the proper
151 character device. The reason is that it is simple to use on
152 root filesystems where you only have a minimal set of tools such
153 as "cat", "echo" etc.
155 The old sysfs still need to be strongly deprecated and removed
156 as it relies on the global GPIO numberspace that assume a strict
157 order of global GPIO numbers that do not change between boots
158 and is independent of probe order.
160 To solve this and provide an ABI that people can use for hacks
161 and development, implement a debugfs interface to manipulate
162 GPIO lines that can do everything that sysfs can do today: one
163 directory per gpiochip and one file entry per line:
165 /sys/kernel/debug/gpiochip/gpiochip0
166 /sys/kernel/debug/gpiochip/gpiochip0/gpio0
167 /sys/kernel/debug/gpiochip/gpiochip0/gpio1
168 /sys/kernel/debug/gpiochip/gpiochip0/gpio2
169 /sys/kernel/debug/gpiochip/gpiochip0/gpio3
171 /sys/kernel/debug/gpiochip/gpiochip1
172 /sys/kernel/debug/gpiochip/gpiochip1/gpio0
173 /sys/kernel/debug/gpiochip/gpiochip1/gpio1
176 The exact files and design of the debugfs interface can be
177 discussed but the idea is to provide a low-level access point
178 for debugging and hacking and to expose all lines without the
179 need of any exporting. Also provide ample ammunition to shoot
180 oneself in the foot, because this is debugfs after all.
183 Moving over to immutable irq_chip structures
185 Most of the gpio chips implementing interrupt support rely on gpiolib
186 intercepting some of the irq_chip callbacks, preventing the structures
187 from being made read-only and forcing duplication of structures that
188 should otherwise be unique.
190 The solution is to call into the gpiolib code when needed (resource
191 management, enable/disable or unmask/mask callbacks), and to let the
192 core code know about that by exposing a flag (IRQCHIP_IMMUTABLE) in
193 the irq_chip structure. The irq_chip structure can then be made unique
196 A small number of drivers have been converted (pl061, tegra186, msm,
197 amd, apple), and can be used as examples of how to proceed with this
198 conversion. Note that drivers using the generic irqchip framework
199 cannot be converted yet, but watch this space!