blob: 54630095aa3c8f841c097e723eef518f660a30a7 [file] [log] [blame]
David Brownell4c203862007-02-12 00:53:11 -08001GPIO Interfaces
2
3This provides an overview of GPIO access conventions on Linux.
4
David Brownell7560fa62008-03-04 14:28:27 -08005These calls use the gpio_* naming prefix. No other calls should use that
6prefix, or the related __gpio_* prefix.
7
David Brownell4c203862007-02-12 00:53:11 -08008
9What is a GPIO?
10===============
11A "General Purpose Input/Output" (GPIO) is a flexible software-controlled
12digital signal. They are provided from many kinds of chip, and are familiar
13to Linux developers working with embedded and custom hardware. Each GPIO
14represents a bit connected to a particular pin, or "ball" on Ball Grid Array
15(BGA) packages. Board schematics show which external hardware connects to
16which GPIOs. Drivers can be written generically, so that board setup code
17passes such pin configuration data to drivers.
18
19System-on-Chip (SOC) processors heavily rely on GPIOs. In some cases, every
20non-dedicated pin can be configured as a GPIO; and most chips have at least
21several dozen of them. Programmable logic devices (like FPGAs) can easily
22provide GPIOs; multifunction chips like power managers, and audio codecs
23often have a few such pins to help with pin scarcity on SOCs; and there are
24also "GPIO Expander" chips that connect using the I2C or SPI serial busses.
25Most PC southbridges have a few dozen GPIO-capable pins (with only the BIOS
26firmware knowing how they're used).
27
28The exact capabilities of GPIOs vary between systems. Common options:
29
30 - Output values are writable (high=1, low=0). Some chips also have
31 options about how that value is driven, so that for example only one
32 value might be driven ... supporting "wire-OR" and similar schemes
David Brownell1668be72007-04-11 23:28:42 -070033 for the other value (notably, "open drain" signaling).
David Brownell4c203862007-02-12 00:53:11 -080034
35 - Input values are likewise readable (1, 0). Some chips support readback
36 of pins configured as "output", which is very useful in such "wire-OR"
37 cases (to support bidirectional signaling). GPIO controllers may have
David Brownell7c2db752008-02-04 22:28:21 -080038 input de-glitch/debounce logic, sometimes with software controls.
David Brownell4c203862007-02-12 00:53:11 -080039
40 - Inputs can often be used as IRQ signals, often edge triggered but
41 sometimes level triggered. Such IRQs may be configurable as system
42 wakeup events, to wake the system from a low power state.
43
44 - Usually a GPIO will be configurable as either input or output, as needed
45 by different product boards; single direction ones exist too.
46
47 - Most GPIOs can be accessed while holding spinlocks, but those accessed
48 through a serial bus normally can't. Some systems support both types.
49
50On a given board each GPIO is used for one specific purpose like monitoring
51MMC/SD card insertion/removal, detecting card writeprotect status, driving
52a LED, configuring a transceiver, bitbanging a serial bus, poking a hardware
53watchdog, sensing a switch, and so on.
54
55
56GPIO conventions
57================
58Note that this is called a "convention" because you don't need to do it this
59way, and it's no crime if you don't. There **are** cases where portability
60is not the main issue; GPIOs are often used for the kind of board-specific
61glue logic that may even change between board revisions, and can't ever be
62used on a board that's wired differently. Only least-common-denominator
63functionality can be very portable. Other features are platform-specific,
64and that can be critical for glue logic.
65
David Brownell7c2db752008-02-04 22:28:21 -080066Plus, this doesn't require any implementation framework, just an interface.
David Brownell4c203862007-02-12 00:53:11 -080067One platform might implement it as simple inline functions accessing chip
68registers; another might implement it by delegating through abstractions
David Brownell7c2db752008-02-04 22:28:21 -080069used for several very different kinds of GPIO controller. (There is some
70optional code supporting such an implementation strategy, described later
71in this document, but drivers acting as clients to the GPIO interface must
72not care how it's implemented.)
David Brownell4c203862007-02-12 00:53:11 -080073
74That said, if the convention is supported on their platform, drivers should
David Brownell7560fa62008-03-04 14:28:27 -080075use it when possible. Platforms must declare GENERIC_GPIO support in their
76Kconfig (boolean true), and provide an <asm/gpio.h> file. Drivers that can't
77work without standard GPIO calls should have Kconfig entries which depend
78on GENERIC_GPIO. The GPIO calls are available, either as "real code" or as
79optimized-away stubs, when drivers use the include file:
David Brownell4c203862007-02-12 00:53:11 -080080
David Brownell7560fa62008-03-04 14:28:27 -080081 #include <linux/gpio.h>
David Brownell4c203862007-02-12 00:53:11 -080082
83If you stick to this convention then it'll be easier for other developers to
84see what your code is doing, and help maintain it.
85
David Brownella0a99832007-07-19 01:47:52 -070086Note that these operations include I/O barriers on platforms which need to
87use them; drivers don't need to add them explicitly.
88
David Brownell4c203862007-02-12 00:53:11 -080089
90Identifying GPIOs
91-----------------
92GPIOs are identified by unsigned integers in the range 0..MAX_INT. That
93reserves "negative" numbers for other purposes like marking signals as
David Brownellf5de6112007-02-16 01:27:14 -080094"not available on this board", or indicating faults. Code that doesn't
95touch the underlying hardware treats these integers as opaque cookies.
David Brownell4c203862007-02-12 00:53:11 -080096
97Platforms define how they use those integers, and usually #define symbols
98for the GPIO lines so that board-specific setup code directly corresponds
99to the relevant schematics. In contrast, drivers should only use GPIO
100numbers passed to them from that setup code, using platform_data to hold
101board-specific pin configuration data (along with other board specific
102data they need). That avoids portability problems.
103
104So for example one platform uses numbers 32-159 for GPIOs; while another
105uses numbers 0..63 with one set of GPIO controllers, 64-79 with another
106type of GPIO controller, and on one particular board 80-95 with an FPGA.
107The numbers need not be contiguous; either of those platforms could also
108use numbers 2000-2063 to identify GPIOs in a bank of I2C GPIO expanders.
109
110Whether a platform supports multiple GPIO controllers is currently a
111platform-specific implementation issue.
112
113
114Using GPIOs
115-----------
116One of the first things to do with a GPIO, often in board setup code when
117setting up a platform_device using the GPIO, is mark its direction:
118
119 /* set as input or output, returning 0 or negative errno */
120 int gpio_direction_input(unsigned gpio);
David Brownell28735a72007-03-16 13:38:14 -0800121 int gpio_direction_output(unsigned gpio, int value);
David Brownell4c203862007-02-12 00:53:11 -0800122
123The return value is zero for success, else a negative errno. It should
124be checked, since the get/set calls don't have error returns and since
David Brownell83c65902007-05-16 22:11:13 -0700125misconfiguration is possible. You should normally issue these calls from
126a task context. However, for spinlock-safe GPIOs it's OK to use them
127before tasking is enabled, as part of early board setup.
David Brownell4c203862007-02-12 00:53:11 -0800128
David Brownell28735a72007-03-16 13:38:14 -0800129For output GPIOs, the value provided becomes the initial output value.
130This helps avoid signal glitching during system startup.
131
David Brownell7c2db752008-02-04 22:28:21 -0800132For compatibility with legacy interfaces to GPIOs, setting the direction
133of a GPIO implicitly requests that GPIO (see below) if it has not been
134requested already. That compatibility may be removed in the future;
135explicitly requesting GPIOs is strongly preferred.
136
David Brownell4c203862007-02-12 00:53:11 -0800137Setting the direction can fail if the GPIO number is invalid, or when
138that particular GPIO can't be used in that mode. It's generally a bad
139idea to rely on boot firmware to have set the direction correctly, since
140it probably wasn't validated to do more than boot Linux. (Similarly,
141that board setup code probably needs to multiplex that pin as a GPIO,
142and configure pullups/pulldowns appropriately.)
143
144
145Spinlock-Safe GPIO access
146-------------------------
147Most GPIO controllers can be accessed with memory read/write instructions.
148That doesn't need to sleep, and can safely be done from inside IRQ handlers.
David Brownell7c2db752008-02-04 22:28:21 -0800149(That includes hardirq contexts on RT kernels.)
David Brownell4c203862007-02-12 00:53:11 -0800150
151Use these calls to access such GPIOs:
152
153 /* GPIO INPUT: return zero or nonzero */
154 int gpio_get_value(unsigned gpio);
155
156 /* GPIO OUTPUT */
157 void gpio_set_value(unsigned gpio, int value);
158
159The values are boolean, zero for low, nonzero for high. When reading the
160value of an output pin, the value returned should be what's seen on the
161pin ... that won't always match the specified output value, because of
David Brownell7c2db752008-02-04 22:28:21 -0800162issues including open-drain signaling and output latencies.
David Brownell4c203862007-02-12 00:53:11 -0800163
164The get/set calls have no error returns because "invalid GPIO" should have
David Brownellbe1ff382007-07-23 18:43:57 -0700165been reported earlier from gpio_direction_*(). However, note that not all
David Brownell4c203862007-02-12 00:53:11 -0800166platforms can read the value of output pins; those that can't should always
David Brownellf5de6112007-02-16 01:27:14 -0800167return zero. Also, using these calls for GPIOs that can't safely be accessed
168without sleeping (see below) is an error.
David Brownell4c203862007-02-12 00:53:11 -0800169
David Brownellf5de6112007-02-16 01:27:14 -0800170Platform-specific implementations are encouraged to optimize the two
David Brownell4c203862007-02-12 00:53:11 -0800171calls to access the GPIO value in cases where the GPIO number (and for
172output, value) are constant. It's normal for them to need only a couple
173of instructions in such cases (reading or writing a hardware register),
174and not to need spinlocks. Such optimized calls can make bitbanging
175applications a lot more efficient (in both space and time) than spending
176dozens of instructions on subroutine calls.
177
178
179GPIO access that may sleep
180--------------------------
181Some GPIO controllers must be accessed using message based busses like I2C
182or SPI. Commands to read or write those GPIO values require waiting to
183get to the head of a queue to transmit a command and get its response.
184This requires sleeping, which can't be done from inside IRQ handlers.
185
186Platforms that support this type of GPIO distinguish them from other GPIOs
David Brownell7c2db752008-02-04 22:28:21 -0800187by returning nonzero from this call (which requires a valid GPIO number,
188either explicitly or implicitly requested):
David Brownell4c203862007-02-12 00:53:11 -0800189
190 int gpio_cansleep(unsigned gpio);
191
192To access such GPIOs, a different set of accessors is defined:
193
194 /* GPIO INPUT: return zero or nonzero, might sleep */
195 int gpio_get_value_cansleep(unsigned gpio);
196
197 /* GPIO OUTPUT, might sleep */
198 void gpio_set_value_cansleep(unsigned gpio, int value);
199
200Other than the fact that these calls might sleep, and will not be ignored
201for GPIOs that can't be accessed from IRQ handlers, these calls act the
202same as the spinlock-safe calls.
203
204
205Claiming and Releasing GPIOs (OPTIONAL)
206---------------------------------------
207To help catch system configuration errors, two calls are defined.
208However, many platforms don't currently support this mechanism.
209
210 /* request GPIO, returning 0 or negative errno.
211 * non-null labels may be useful for diagnostics.
212 */
213 int gpio_request(unsigned gpio, const char *label);
214
215 /* release previously-claimed GPIO */
216 void gpio_free(unsigned gpio);
217
218Passing invalid GPIO numbers to gpio_request() will fail, as will requesting
219GPIOs that have already been claimed with that call. The return value of
David Brownell83c65902007-05-16 22:11:13 -0700220gpio_request() must be checked. You should normally issue these calls from
221a task context. However, for spinlock-safe GPIOs it's OK to request GPIOs
222before tasking is enabled, as part of early board setup.
David Brownell4c203862007-02-12 00:53:11 -0800223
224These calls serve two basic purposes. One is marking the signals which
225are actually in use as GPIOs, for better diagnostics; systems may have
226several hundred potential GPIOs, but often only a dozen are used on any
David Brownell7c2db752008-02-04 22:28:21 -0800227given board. Another is to catch conflicts, identifying errors when
228(a) two or more drivers wrongly think they have exclusive use of that
229signal, or (b) something wrongly believes it's safe to remove drivers
230needed to manage a signal that's in active use. That is, requesting a
231GPIO can serve as a kind of lock.
David Brownell4c203862007-02-12 00:53:11 -0800232
233These two calls are optional because not not all current Linux platforms
234offer such functionality in their GPIO support; a valid implementation
235could return success for all gpio_request() calls. Unlike the other calls,
236the state they represent doesn't normally match anything from a hardware
237register; it's just a software bitmap which clearly is not necessary for
238correct operation of hardware or (bug free) drivers.
239
240Note that requesting a GPIO does NOT cause it to be configured in any
241way; it just marks that GPIO as in use. Separate code must handle any
242pin setup (e.g. controlling which pin the GPIO uses, pullup/pulldown).
243
David Brownell7c2db752008-02-04 22:28:21 -0800244Also note that it's your responsibility to have stopped using a GPIO
245before you free it.
246
David Brownell4c203862007-02-12 00:53:11 -0800247
248GPIOs mapped to IRQs
249--------------------
250GPIO numbers are unsigned integers; so are IRQ numbers. These make up
251two logically distinct namespaces (GPIO 0 need not use IRQ 0). You can
252map between them using calls like:
253
254 /* map GPIO numbers to IRQ numbers */
255 int gpio_to_irq(unsigned gpio);
256
257 /* map IRQ numbers to GPIO numbers */
258 int irq_to_gpio(unsigned irq);
259
260Those return either the corresponding number in the other namespace, or
261else a negative errno code if the mapping can't be done. (For example,
David Brownell7c2db752008-02-04 22:28:21 -0800262some GPIOs can't be used as IRQs.) It is an unchecked error to use a GPIO
David Brownellbe1ff382007-07-23 18:43:57 -0700263number that wasn't set up as an input using gpio_direction_input(), or
David Brownell4c203862007-02-12 00:53:11 -0800264to use an IRQ number that didn't originally come from gpio_to_irq().
265
266These two mapping calls are expected to cost on the order of a single
267addition or subtraction. They're not allowed to sleep.
268
269Non-error values returned from gpio_to_irq() can be passed to request_irq()
270or free_irq(). They will often be stored into IRQ resources for platform
271devices, by the board-specific initialization code. Note that IRQ trigger
272options are part of the IRQ interface, e.g. IRQF_TRIGGER_FALLING, as are
273system wakeup capabilities.
274
275Non-error values returned from irq_to_gpio() would most commonly be used
David Brownellf5de6112007-02-16 01:27:14 -0800276with gpio_get_value(), for example to initialize or update driver state
277when the IRQ is edge-triggered.
David Brownell4c203862007-02-12 00:53:11 -0800278
279
David Brownell1668be72007-04-11 23:28:42 -0700280Emulating Open Drain Signals
281----------------------------
282Sometimes shared signals need to use "open drain" signaling, where only the
283low signal level is actually driven. (That term applies to CMOS transistors;
284"open collector" is used for TTL.) A pullup resistor causes the high signal
285level. This is sometimes called a "wire-AND"; or more practically, from the
286negative logic (low=true) perspective this is a "wire-OR".
287
288One common example of an open drain signal is a shared active-low IRQ line.
289Also, bidirectional data bus signals sometimes use open drain signals.
290
291Some GPIO controllers directly support open drain outputs; many don't. When
292you need open drain signaling but your hardware doesn't directly support it,
293there's a common idiom you can use to emulate it with any GPIO pin that can
294be used as either an input or an output:
295
296 LOW: gpio_direction_output(gpio, 0) ... this drives the signal
297 and overrides the pullup.
298
299 HIGH: gpio_direction_input(gpio) ... this turns off the output,
300 so the pullup (or some other device) controls the signal.
301
302If you are "driving" the signal high but gpio_get_value(gpio) reports a low
303value (after the appropriate rise time passes), you know some other component
304is driving the shared signal low. That's not necessarily an error. As one
305common example, that's how I2C clocks are stretched: a slave that needs a
306slower clock delays the rising edge of SCK, and the I2C master adjusts its
307signaling rate accordingly.
308
David Brownell4c203862007-02-12 00:53:11 -0800309
310What do these conventions omit?
311===============================
312One of the biggest things these conventions omit is pin multiplexing, since
313this is highly chip-specific and nonportable. One platform might not need
314explicit multiplexing; another might have just two options for use of any
315given pin; another might have eight options per pin; another might be able
316to route a given GPIO to any one of several pins. (Yes, those examples all
317come from systems that run Linux today.)
318
319Related to multiplexing is configuration and enabling of the pullups or
320pulldowns integrated on some platforms. Not all platforms support them,
321or support them in the same way; and any given board might use external
322pullups (or pulldowns) so that the on-chip ones should not be used.
David Brownell7c2db752008-02-04 22:28:21 -0800323(When a circuit needs 5 kOhm, on-chip 100 kOhm resistors won't do.)
David Brownell7560fa62008-03-04 14:28:27 -0800324Likewise drive strength (2 mA vs 20 mA) and voltage (1.8V vs 3.3V) is a
325platform-specific issue, as are models like (not) having a one-to-one
326correspondence between configurable pins and GPIOs.
David Brownell4c203862007-02-12 00:53:11 -0800327
328There are other system-specific mechanisms that are not specified here,
329like the aforementioned options for input de-glitching and wire-OR output.
330Hardware may support reading or writing GPIOs in gangs, but that's usually
David Brownellf5de6112007-02-16 01:27:14 -0800331configuration dependent: for GPIOs sharing the same bank. (GPIOs are
David Brownell4c203862007-02-12 00:53:11 -0800332commonly grouped in banks of 16 or 32, with a given SOC having several such
David Brownell7c2db752008-02-04 22:28:21 -0800333banks.) Some systems can trigger IRQs from output GPIOs, or read values
334from pins not managed as GPIOs. Code relying on such mechanisms will
335necessarily be nonportable.
David Brownell4c203862007-02-12 00:53:11 -0800336
David Brownell7c2db752008-02-04 22:28:21 -0800337Dynamic definition of GPIOs is not currently standard; for example, as
David Brownell4c203862007-02-12 00:53:11 -0800338a side effect of configuring an add-on board with some GPIO expanders.
339
340These calls are purely for kernel space, but a userspace API could be built
David Brownell7c2db752008-02-04 22:28:21 -0800341on top of them.
342
343
344GPIO implementor's framework (OPTIONAL)
345=======================================
346As noted earlier, there is an optional implementation framework making it
347easier for platforms to support different kinds of GPIO controller using
348the same programming interface.
349
350As a debugging aid, if debugfs is available a /sys/kernel/debug/gpio file
351will be found there. That will list all the controllers registered through
352this framework, and the state of the GPIOs currently in use.
353
354
355Controller Drivers: gpio_chip
356-----------------------------
357In this framework each GPIO controller is packaged as a "struct gpio_chip"
358with information common to each controller of that type:
359
360 - methods to establish GPIO direction
361 - methods used to access GPIO values
362 - flag saying whether calls to its methods may sleep
363 - optional debugfs dump method (showing extra state like pullup config)
364 - label for diagnostics
365
366There is also per-instance data, which may come from device.platform_data:
367the number of its first GPIO, and how many GPIOs it exposes.
368
369The code implementing a gpio_chip should support multiple instances of the
370controller, possibly using the driver model. That code will configure each
371gpio_chip and issue gpiochip_add(). Removing a GPIO controller should be
372rare; use gpiochip_remove() when it is unavoidable.
373
374Most often a gpio_chip is part of an instance-specific structure with state
375not exposed by the GPIO interfaces, such as addressing, power management,
376and more. Chips such as codecs will have complex non-GPIO state,
377
378Any debugfs dump method should normally ignore signals which haven't been
379requested as GPIOs. They can use gpiochip_is_requested(), which returns
380either NULL or the label associated with that GPIO when it was requested.
381
382
383Platform Support
384----------------
385To support this framework, a platform's Kconfig will "select HAVE_GPIO_LIB"
386and arrange that its <asm/gpio.h> includes <asm-generic/gpio.h> and defines
387three functions: gpio_get_value(), gpio_set_value(), and gpio_cansleep().
388They may also want to provide a custom value for ARCH_NR_GPIOS.
389
390Trivial implementations of those functions can directly use framework
391code, which always dispatches through the gpio_chip:
392
393 #define gpio_get_value __gpio_get_value
394 #define gpio_set_value __gpio_set_value
395 #define gpio_cansleep __gpio_cansleep
396
397Fancier implementations could instead define those as inline functions with
398logic optimizing access to specific SOC-based GPIOs. For example, if the
399referenced GPIO is the constant "12", getting or setting its value could
400cost as little as two or three instructions, never sleeping. When such an
401optimization is not possible those calls must delegate to the framework
402code, costing at least a few dozen instructions. For bitbanged I/O, such
403instruction savings can be significant.
404
405For SOCs, platform-specific code defines and registers gpio_chip instances
406for each bank of on-chip GPIOs. Those GPIOs should be numbered/labeled to
407match chip vendor documentation, and directly match board schematics. They
408may well start at zero and go up to a platform-specific limit. Such GPIOs
409are normally integrated into platform initialization to make them always be
410available, from arch_initcall() or earlier; they can often serve as IRQs.
411
412
413Board Support
414-------------
415For external GPIO controllers -- such as I2C or SPI expanders, ASICs, multi
416function devices, FPGAs or CPLDs -- most often board-specific code handles
417registering controller devices and ensures that their drivers know what GPIO
418numbers to use with gpiochip_add(). Their numbers often start right after
419platform-specific GPIOs.
420
421For example, board setup code could create structures identifying the range
422of GPIOs that chip will expose, and passes them to each GPIO expander chip
423using platform_data. Then the chip driver's probe() routine could pass that
424data to gpiochip_add().
425
426Initialization order can be important. For example, when a device relies on
427an I2C-based GPIO, its probe() routine should only be called after that GPIO
428becomes available. That may mean the device should not be registered until
429calls for that GPIO can work. One way to address such dependencies is for
430such gpio_chip controllers to provide setup() and teardown() callbacks to
431board specific code; those board specific callbacks would register devices
432once all the necessary resources are available.