1GPIO Interfaces 2 3This provides an overview of GPIO access conventions on Linux. 4 5 6What is a GPIO? 7=============== 8A "General Purpose Input/Output" (GPIO) is a flexible software-controlled 9digital signal. They are provided from many kinds of chip, and are familiar 10to Linux developers working with embedded and custom hardware. Each GPIO 11represents a bit connected to a particular pin, or "ball" on Ball Grid Array 12(BGA) packages. Board schematics show which external hardware connects to 13which GPIOs. Drivers can be written generically, so that board setup code 14passes such pin configuration data to drivers. 15 16System-on-Chip (SOC) processors heavily rely on GPIOs. In some cases, every 17non-dedicated pin can be configured as a GPIO; and most chips have at least 18several dozen of them. Programmable logic devices (like FPGAs) can easily 19provide GPIOs; multifunction chips like power managers, and audio codecs 20often have a few such pins to help with pin scarcity on SOCs; and there are 21also "GPIO Expander" chips that connect using the I2C or SPI serial busses. 22Most PC southbridges have a few dozen GPIO-capable pins (with only the BIOS 23firmware knowing how they're used). 24 25The exact capabilities of GPIOs vary between systems. Common options: 26 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 30 for the other value (notably, "open drain" signaling). 31 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. 36 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. 40 41 - Usually a GPIO will be configurable as either input or output, as needed 42 by different product boards; single direction ones exist too. 43 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. 46 47On a given board each GPIO is used for one specific purpose like monitoring 48MMC/SD card insertion/removal, detecting card writeprotect status, driving 49a LED, configuring a transceiver, bitbanging a serial bus, poking a hardware 50watchdog, sensing a switch, and so on. 51 52 53GPIO conventions 54================ 55Note that this is called a "convention" because you don't need to do it this 56way, and it's no crime if you don't. There **are** cases where portability 57is not the main issue; GPIOs are often used for the kind of board-specific 58glue logic that may even change between board revisions, and can't ever be 59used on a board that's wired differently. Only least-common-denominator 60functionality can be very portable. Other features are platform-specific, 61and that can be critical for glue logic. 62 63Plus, this doesn't define an implementation framework, just an interface. 64One platform might implement it as simple inline functions accessing chip 65registers; another might implement it by delegating through abstractions 66used for several very different kinds of GPIO controller. 67 68That said, if the convention is supported on their platform, drivers should 69use it when possible. Platforms should declare GENERIC_GPIO support in 70Kconfig (boolean true), which multi-platform drivers can depend on when 71using the include file: 72 73 #include <asm/gpio.h> 74 75If you stick to this convention then it'll be easier for other developers to 76see what your code is doing, and help maintain it. 77 78 79Identifying GPIOs 80----------------- 81GPIOs are identified by unsigned integers in the range 0..MAX_INT. That 82reserves "negative" numbers for other purposes like marking signals as 83"not available on this board", or indicating faults. Code that doesn't 84touch the underlying hardware treats these integers as opaque cookies. 85 86Platforms define how they use those integers, and usually #define symbols 87for the GPIO lines so that board-specific setup code directly corresponds 88to the relevant schematics. In contrast, drivers should only use GPIO 89numbers passed to them from that setup code, using platform_data to hold 90board-specific pin configuration data (along with other board specific 91data they need). That avoids portability problems. 92 93So for example one platform uses numbers 32-159 for GPIOs; while another 94uses numbers 0..63 with one set of GPIO controllers, 64-79 with another 95type of GPIO controller, and on one particular board 80-95 with an FPGA. 96The numbers need not be contiguous; either of those platforms could also 97use numbers 2000-2063 to identify GPIOs in a bank of I2C GPIO expanders. 98 99Whether a platform supports multiple GPIO controllers is currently a 100platform-specific implementation issue. 101 102 103Using GPIOs 104----------- 105One of the first things to do with a GPIO, often in board setup code when 106setting up a platform_device using the GPIO, is mark its direction: 107 108 /* set as input or output, returning 0 or negative errno */ 109 int gpio_direction_input(unsigned gpio); 110 int gpio_direction_output(unsigned gpio, int value); 111 112The return value is zero for success, else a negative errno. It should 113be checked, since the get/set calls don't have error returns and since 114misconfiguration is possible. You should normally issue these calls from 115a task context. However, for spinlock-safe GPIOs it's OK to use them 116before tasking is enabled, as part of early board setup. 117 118For output GPIOs, the value provided becomes the initial output value. 119This helps avoid signal glitching during system startup. 120 121Setting the direction can fail if the GPIO number is invalid, or when 122that particular GPIO can't be used in that mode. It's generally a bad 123idea to rely on boot firmware to have set the direction correctly, since 124it probably wasn't validated to do more than boot Linux. (Similarly, 125that board setup code probably needs to multiplex that pin as a GPIO, 126and configure pullups/pulldowns appropriately.) 127 128 129Spinlock-Safe GPIO access 130------------------------- 131Most GPIO controllers can be accessed with memory read/write instructions. 132That doesn't need to sleep, and can safely be done from inside IRQ handlers. 133 134Use these calls to access such GPIOs: 135 136 /* GPIO INPUT: return zero or nonzero */ 137 int gpio_get_value(unsigned gpio); 138 139 /* GPIO OUTPUT */ 140 void gpio_set_value(unsigned gpio, int value); 141 142The values are boolean, zero for low, nonzero for high. When reading the 143value of an output pin, the value returned should be what's seen on the 144pin ... that won't always match the specified output value, because of 145issues including wire-OR and output latencies. 146 147The get/set calls have no error returns because "invalid GPIO" should have 148been reported earlier in gpio_set_direction(). However, note that not all 149platforms can read the value of output pins; those that can't should always 150return zero. Also, using these calls for GPIOs that can't safely be accessed 151without sleeping (see below) is an error. 152 153Platform-specific implementations are encouraged to optimize the two 154calls to access the GPIO value in cases where the GPIO number (and for 155output, value) are constant. It's normal for them to need only a couple 156of instructions in such cases (reading or writing a hardware register), 157and not to need spinlocks. Such optimized calls can make bitbanging 158applications a lot more efficient (in both space and time) than spending 159dozens of instructions on subroutine calls. 160 161 162GPIO access that may sleep 163-------------------------- 164Some GPIO controllers must be accessed using message based busses like I2C 165or SPI. Commands to read or write those GPIO values require waiting to 166get to the head of a queue to transmit a command and get its response. 167This requires sleeping, which can't be done from inside IRQ handlers. 168 169Platforms that support this type of GPIO distinguish them from other GPIOs 170by returning nonzero from this call: 171 172 int gpio_cansleep(unsigned gpio); 173 174To access such GPIOs, a different set of accessors is defined: 175 176 /* GPIO INPUT: return zero or nonzero, might sleep */ 177 int gpio_get_value_cansleep(unsigned gpio); 178 179 /* GPIO OUTPUT, might sleep */ 180 void gpio_set_value_cansleep(unsigned gpio, int value); 181 182Other than the fact that these calls might sleep, and will not be ignored 183for GPIOs that can't be accessed from IRQ handlers, these calls act the 184same as the spinlock-safe calls. 185 186 187Claiming and Releasing GPIOs (OPTIONAL) 188--------------------------------------- 189To help catch system configuration errors, two calls are defined. 190However, many platforms don't currently support this mechanism. 191 192 /* request GPIO, returning 0 or negative errno. 193 * non-null labels may be useful for diagnostics. 194 */ 195 int gpio_request(unsigned gpio, const char *label); 196 197 /* release previously-claimed GPIO */ 198 void gpio_free(unsigned gpio); 199 200Passing invalid GPIO numbers to gpio_request() will fail, as will requesting 201GPIOs that have already been claimed with that call. The return value of 202gpio_request() must be checked. You should normally issue these calls from 203a task context. However, for spinlock-safe GPIOs it's OK to request GPIOs 204before tasking is enabled, as part of early board setup. 205 206These calls serve two basic purposes. One is marking the signals which 207are actually in use as GPIOs, for better diagnostics; systems may have 208several hundred potential GPIOs, but often only a dozen are used on any 209given board. Another is to catch conflicts between drivers, reporting 210errors when drivers wrongly think they have exclusive use of that signal. 211 212These two calls are optional because not not all current Linux platforms 213offer such functionality in their GPIO support; a valid implementation 214could return success for all gpio_request() calls. Unlike the other calls, 215the state they represent doesn't normally match anything from a hardware 216register; it's just a software bitmap which clearly is not necessary for 217correct operation of hardware or (bug free) drivers. 218 219Note that requesting a GPIO does NOT cause it to be configured in any 220way; it just marks that GPIO as in use. Separate code must handle any 221pin setup (e.g. controlling which pin the GPIO uses, pullup/pulldown). 222 223 224GPIOs mapped to IRQs 225-------------------- 226GPIO numbers are unsigned integers; so are IRQ numbers. These make up 227two logically distinct namespaces (GPIO 0 need not use IRQ 0). You can 228map between them using calls like: 229 230 /* map GPIO numbers to IRQ numbers */ 231 int gpio_to_irq(unsigned gpio); 232 233 /* map IRQ numbers to GPIO numbers */ 234 int irq_to_gpio(unsigned irq); 235 236Those return either the corresponding number in the other namespace, or 237else a negative errno code if the mapping can't be done. (For example, 238some GPIOs can't used as IRQs.) It is an unchecked error to use a GPIO 239number that hasn't been marked as an input using gpio_set_direction(), or 240to use an IRQ number that didn't originally come from gpio_to_irq(). 241 242These two mapping calls are expected to cost on the order of a single 243addition or subtraction. They're not allowed to sleep. 244 245Non-error values returned from gpio_to_irq() can be passed to request_irq() 246or free_irq(). They will often be stored into IRQ resources for platform 247devices, by the board-specific initialization code. Note that IRQ trigger 248options are part of the IRQ interface, e.g. IRQF_TRIGGER_FALLING, as are 249system wakeup capabilities. 250 251Non-error values returned from irq_to_gpio() would most commonly be used 252with gpio_get_value(), for example to initialize or update driver state 253when the IRQ is edge-triggered. 254 255 256Emulating Open Drain Signals 257---------------------------- 258Sometimes shared signals need to use "open drain" signaling, where only the 259low signal level is actually driven. (That term applies to CMOS transistors; 260"open collector" is used for TTL.) A pullup resistor causes the high signal 261level. This is sometimes called a "wire-AND"; or more practically, from the 262negative logic (low=true) perspective this is a "wire-OR". 263 264One common example of an open drain signal is a shared active-low IRQ line. 265Also, bidirectional data bus signals sometimes use open drain signals. 266 267Some GPIO controllers directly support open drain outputs; many don't. When 268you need open drain signaling but your hardware doesn't directly support it, 269there's a common idiom you can use to emulate it with any GPIO pin that can 270be used as either an input or an output: 271 272 LOW: gpio_direction_output(gpio, 0) ... this drives the signal 273 and overrides the pullup. 274 275 HIGH: gpio_direction_input(gpio) ... this turns off the output, 276 so the pullup (or some other device) controls the signal. 277 278If you are "driving" the signal high but gpio_get_value(gpio) reports a low 279value (after the appropriate rise time passes), you know some other component 280is driving the shared signal low. That's not necessarily an error. As one 281common example, that's how I2C clocks are stretched: a slave that needs a 282slower clock delays the rising edge of SCK, and the I2C master adjusts its 283signaling rate accordingly. 284 285 286What do these conventions omit? 287=============================== 288One of the biggest things these conventions omit is pin multiplexing, since 289this is highly chip-specific and nonportable. One platform might not need 290explicit multiplexing; another might have just two options for use of any 291given pin; another might have eight options per pin; another might be able 292to route a given GPIO to any one of several pins. (Yes, those examples all 293come from systems that run Linux today.) 294 295Related to multiplexing is configuration and enabling of the pullups or 296pulldowns integrated on some platforms. Not all platforms support them, 297or support them in the same way; and any given board might use external 298pullups (or pulldowns) so that the on-chip ones should not be used. 299 300There are other system-specific mechanisms that are not specified here, 301like the aforementioned options for input de-glitching and wire-OR output. 302Hardware may support reading or writing GPIOs in gangs, but that's usually 303configuration dependent: for GPIOs sharing the same bank. (GPIOs are 304commonly grouped in banks of 16 or 32, with a given SOC having several such 305banks.) Some systems can trigger IRQs from output GPIOs. Code relying on 306such mechanisms will necessarily be nonportable. 307 308Dynamic definition of GPIOs is not currently supported; for example, as 309a side effect of configuring an add-on board with some GPIO expanders. 310 311These calls are purely for kernel space, but a userspace API could be built 312on top of it. 313