3 This provides an overview of GPIO access conventions on Linux.
8 A "General Purpose Input/Output" (GPIO) is a flexible software-controlled
9 digital signal. They are provided from many kinds of chip, and are familiar
10 to Linux developers working with embedded and custom hardware. Each GPIO
11 represents a bit connected to a particular pin, or "ball" on Ball Grid Array
12 (BGA) packages. Board schematics show which external hardware connects to
13 which GPIOs. Drivers can be written generically, so that board setup code
14 passes such pin configuration data to drivers.
16 System-on-Chip (SOC) processors heavily rely on GPIOs. In some cases, every
17 non-dedicated pin can be configured as a GPIO; and most chips have at least
18 several dozen of them. Programmable logic devices (like FPGAs) can easily
19 provide GPIOs; multifunction chips like power managers, and audio codecs
20 often have a few such pins to help with pin scarcity on SOCs; and there are
21 also "GPIO Expander" chips that connect using the I2C or SPI serial busses.
22 Most PC southbridges have a few dozen GPIO-capable pins (with only the BIOS
23 firmware knowing how they're used).
25 The exact capabilities of GPIOs vary between systems. Common options:
27 - Output values are writable (high=1, low=0). Some chips also have
28 options about how that value is driven, so that for example only one
29 value might be driven ... supporting "wire-OR" and similar schemes
32 - Input values are likewise readable (1, 0). Some chips support readback
33 of pins configured as "output", which is very useful in such "wire-OR"
34 cases (to support bidirectional signaling). GPIO controllers may have
35 input de-glitch logic, sometimes with software controls.
37 - Inputs can often be used as IRQ signals, often edge triggered but
38 sometimes level triggered. Such IRQs may be configurable as system
39 wakeup events, to wake the system from a low power state.
41 - Usually a GPIO will be configurable as either input or output, as needed
42 by different product boards; single direction ones exist too.
44 - Most GPIOs can be accessed while holding spinlocks, but those accessed
45 through a serial bus normally can't. Some systems support both types.
47 On a given board each GPIO is used for one specific purpose like monitoring
48 MMC/SD card insertion/removal, detecting card writeprotect status, driving
49 a LED, configuring a transceiver, bitbanging a serial bus, poking a hardware
50 watchdog, sensing a switch, and so on.
55 Note that this is called a "convention" because you don't need to do it this
56 way, and it's no crime if you don't. There **are** cases where portability
57 is not the main issue; GPIOs are often used for the kind of board-specific
58 glue logic that may even change between board revisions, and can't ever be
59 used on a board that's wired differently. Only least-common-denominator
60 functionality can be very portable. Other features are platform-specific,
61 and that can be critical for glue logic.
63 Plus, this doesn't define an implementation framework, just an interface.
64 One platform might implement it as simple inline functions accessing chip
65 registers; another might implement it by delegating through abstractions
66 used for several very different kinds of GPIO controller.
68 That said, if the convention is supported on their platform, drivers should
73 If you stick to this convention then it'll be easier for other developers to
74 see what your code is doing, and help maintain it.
79 GPIOs are identified by unsigned integers in the range 0..MAX_INT. That
80 reserves "negative" numbers for other purposes like marking signals as
81 "not available on this board", or indicating faults. Code that doesn't
82 touch the underlying hardware treats these integers as opaque cookies.
84 Platforms define how they use those integers, and usually #define symbols
85 for the GPIO lines so that board-specific setup code directly corresponds
86 to the relevant schematics. In contrast, drivers should only use GPIO
87 numbers passed to them from that setup code, using platform_data to hold
88 board-specific pin configuration data (along with other board specific
89 data they need). That avoids portability problems.
91 So for example one platform uses numbers 32-159 for GPIOs; while another
92 uses numbers 0..63 with one set of GPIO controllers, 64-79 with another
93 type of GPIO controller, and on one particular board 80-95 with an FPGA.
94 The numbers need not be contiguous; either of those platforms could also
95 use numbers 2000-2063 to identify GPIOs in a bank of I2C GPIO expanders.
97 Whether a platform supports multiple GPIO controllers is currently a
98 platform-specific implementation issue.
103 One of the first things to do with a GPIO, often in board setup code when
104 setting up a platform_device using the GPIO, is mark its direction:
106 /* set as input or output, returning 0 or negative errno */
107 int gpio_direction_input(unsigned gpio);
108 int gpio_direction_output(unsigned gpio, int value);
110 The return value is zero for success, else a negative errno. It should
111 be checked, since the get/set calls don't have error returns and since
112 misconfiguration is possible. (These calls could sleep.)
114 For output GPIOs, the value provided becomes the initial output value.
115 This helps avoid signal glitching during system startup.
117 Setting the direction can fail if the GPIO number is invalid, or when
118 that particular GPIO can't be used in that mode. It's generally a bad
119 idea to rely on boot firmware to have set the direction correctly, since
120 it probably wasn't validated to do more than boot Linux. (Similarly,
121 that board setup code probably needs to multiplex that pin as a GPIO,
122 and configure pullups/pulldowns appropriately.)
125 Spinlock-Safe GPIO access
126 -------------------------
127 Most GPIO controllers can be accessed with memory read/write instructions.
128 That doesn't need to sleep, and can safely be done from inside IRQ handlers.
130 Use these calls to access such GPIOs:
132 /* GPIO INPUT: return zero or nonzero */
133 int gpio_get_value(unsigned gpio);
136 void gpio_set_value(unsigned gpio, int value);
138 The values are boolean, zero for low, nonzero for high. When reading the
139 value of an output pin, the value returned should be what's seen on the
140 pin ... that won't always match the specified output value, because of
141 issues including wire-OR and output latencies.
143 The get/set calls have no error returns because "invalid GPIO" should have
144 been reported earlier in gpio_set_direction(). However, note that not all
145 platforms can read the value of output pins; those that can't should always
146 return zero. Also, using these calls for GPIOs that can't safely be accessed
147 without sleeping (see below) is an error.
149 Platform-specific implementations are encouraged to optimize the two
150 calls to access the GPIO value in cases where the GPIO number (and for
151 output, value) are constant. It's normal for them to need only a couple
152 of instructions in such cases (reading or writing a hardware register),
153 and not to need spinlocks. Such optimized calls can make bitbanging
154 applications a lot more efficient (in both space and time) than spending
155 dozens of instructions on subroutine calls.
158 GPIO access that may sleep
159 --------------------------
160 Some GPIO controllers must be accessed using message based busses like I2C
161 or SPI. Commands to read or write those GPIO values require waiting to
162 get to the head of a queue to transmit a command and get its response.
163 This requires sleeping, which can't be done from inside IRQ handlers.
165 Platforms that support this type of GPIO distinguish them from other GPIOs
166 by returning nonzero from this call:
168 int gpio_cansleep(unsigned gpio);
170 To access such GPIOs, a different set of accessors is defined:
172 /* GPIO INPUT: return zero or nonzero, might sleep */
173 int gpio_get_value_cansleep(unsigned gpio);
175 /* GPIO OUTPUT, might sleep */
176 void gpio_set_value_cansleep(unsigned gpio, int value);
178 Other than the fact that these calls might sleep, and will not be ignored
179 for GPIOs that can't be accessed from IRQ handlers, these calls act the
180 same as the spinlock-safe calls.
183 Claiming and Releasing GPIOs (OPTIONAL)
184 ---------------------------------------
185 To help catch system configuration errors, two calls are defined.
186 However, many platforms don't currently support this mechanism.
188 /* request GPIO, returning 0 or negative errno.
189 * non-null labels may be useful for diagnostics.
191 int gpio_request(unsigned gpio, const char *label);
193 /* release previously-claimed GPIO */
194 void gpio_free(unsigned gpio);
196 Passing invalid GPIO numbers to gpio_request() will fail, as will requesting
197 GPIOs that have already been claimed with that call. The return value of
198 gpio_request() must be checked. (These calls could sleep.)
200 These calls serve two basic purposes. One is marking the signals which
201 are actually in use as GPIOs, for better diagnostics; systems may have
202 several hundred potential GPIOs, but often only a dozen are used on any
203 given board. Another is to catch conflicts between drivers, reporting
204 errors when drivers wrongly think they have exclusive use of that signal.
206 These two calls are optional because not not all current Linux platforms
207 offer such functionality in their GPIO support; a valid implementation
208 could return success for all gpio_request() calls. Unlike the other calls,
209 the state they represent doesn't normally match anything from a hardware
210 register; it's just a software bitmap which clearly is not necessary for
211 correct operation of hardware or (bug free) drivers.
213 Note that requesting a GPIO does NOT cause it to be configured in any
214 way; it just marks that GPIO as in use. Separate code must handle any
215 pin setup (e.g. controlling which pin the GPIO uses, pullup/pulldown).
220 GPIO numbers are unsigned integers; so are IRQ numbers. These make up
221 two logically distinct namespaces (GPIO 0 need not use IRQ 0). You can
222 map between them using calls like:
224 /* map GPIO numbers to IRQ numbers */
225 int gpio_to_irq(unsigned gpio);
227 /* map IRQ numbers to GPIO numbers */
228 int irq_to_gpio(unsigned irq);
230 Those return either the corresponding number in the other namespace, or
231 else a negative errno code if the mapping can't be done. (For example,
232 some GPIOs can't used as IRQs.) It is an unchecked error to use a GPIO
233 number that hasn't been marked as an input using gpio_set_direction(), or
234 to use an IRQ number that didn't originally come from gpio_to_irq().
236 These two mapping calls are expected to cost on the order of a single
237 addition or subtraction. They're not allowed to sleep.
239 Non-error values returned from gpio_to_irq() can be passed to request_irq()
240 or free_irq(). They will often be stored into IRQ resources for platform
241 devices, by the board-specific initialization code. Note that IRQ trigger
242 options are part of the IRQ interface, e.g. IRQF_TRIGGER_FALLING, as are
243 system wakeup capabilities.
245 Non-error values returned from irq_to_gpio() would most commonly be used
246 with gpio_get_value(), for example to initialize or update driver state
247 when the IRQ is edge-triggered.
251 What do these conventions omit?
252 ===============================
253 One of the biggest things these conventions omit is pin multiplexing, since
254 this is highly chip-specific and nonportable. One platform might not need
255 explicit multiplexing; another might have just two options for use of any
256 given pin; another might have eight options per pin; another might be able
257 to route a given GPIO to any one of several pins. (Yes, those examples all
258 come from systems that run Linux today.)
260 Related to multiplexing is configuration and enabling of the pullups or
261 pulldowns integrated on some platforms. Not all platforms support them,
262 or support them in the same way; and any given board might use external
263 pullups (or pulldowns) so that the on-chip ones should not be used.
265 There are other system-specific mechanisms that are not specified here,
266 like the aforementioned options for input de-glitching and wire-OR output.
267 Hardware may support reading or writing GPIOs in gangs, but that's usually
268 configuration dependent: for GPIOs sharing the same bank. (GPIOs are
269 commonly grouped in banks of 16 or 32, with a given SOC having several such
270 banks.) Some systems can trigger IRQs from output GPIOs. Code relying on
271 such mechanisms will necessarily be nonportable.
273 Dynamic definition of GPIOs is not currently supported; for example, as
274 a side effect of configuring an add-on board with some GPIO expanders.
276 These calls are purely for kernel space, but a userspace API could be built