1 GPIO Sysfs Interface for Userspace
2 ==================================
4 Platforms which use the "gpiolib" implementors framework may choose to
5 configure a sysfs user interface to GPIOs. This is different from the
6 debugfs interface, since it provides control over GPIO direction and
7 value instead of just showing a gpio state summary. Plus, it could be
8 present on production systems without debugging support.
10 Given appropriate hardware documentation for the system, userspace could
11 know for example that GPIO #23 controls the write protect line used to
12 protect boot loader segments in flash memory. System upgrade procedures
13 may need to temporarily remove that protection, first importing a GPIO,
14 then changing its output state, then updating the code before re-enabling
15 the write protection. In normal use, GPIO #23 would never be touched,
16 and the kernel would have no need to know about it.
18 Again depending on appropriate hardware documentation, on some systems
19 userspace GPIO can be used to determine system configuration data that
20 standard kernels won't know about. And for some tasks, simple userspace
21 GPIO drivers could be all that the system really needs.
23 Note that standard kernel drivers exist for common "LEDs and Buttons"
24 GPIO tasks: "leds-gpio" and "gpio_keys", respectively. Use those
25 instead of talking directly to the GPIOs; they integrate with kernel
26 frameworks better than your userspace code could.
31 There are three kinds of entry in /sys/class/gpio:
33 - Control interfaces used to get userspace control over GPIOs;
35 - GPIOs themselves; and
37 - GPIO controllers ("gpio_chip" instances).
39 That's in addition to standard files including the "device" symlink.
41 The control interfaces are write-only:
45 "export" ... Userspace may ask the kernel to export control of
46 a GPIO to userspace by writing its number to this file.
48 Example: "echo 19 > export" will create a "gpio19" node
49 for GPIO #19, if that's not requested by kernel code.
51 "unexport" ... Reverses the effect of exporting to userspace.
53 Example: "echo 19 > unexport" will remove a "gpio19"
54 node exported using the "export" file.
56 GPIO signals have paths like /sys/class/gpio/gpio42/ (for GPIO #42)
57 and have the following read/write attributes:
59 /sys/class/gpio/gpioN/
61 "direction" ... reads as either "in" or "out". This value may
62 normally be written. Writing as "out" defaults to
63 initializing the value as low. To ensure glitch free
64 operation, values "low" and "high" may be written to
65 configure the GPIO as an output with that initial value.
67 Note that this attribute *will not exist* if the kernel
68 doesn't support changing the direction of a GPIO, or
69 it was exported by kernel code that didn't explicitly
70 allow userspace to reconfigure this GPIO's direction.
72 "value" ... reads as either 0 (low) or 1 (high). If the GPIO
73 is configured as an output, this value may be written;
74 any nonzero value is treated as high.
76 If the pin can be configured as interrupt-generating interrupt
77 and if it has been configured to generate interrupts (see the
78 description of "edge"), you can poll(2) on that file and
79 poll(2) will return whenever the interrupt was triggered. If
80 you use poll(2), set the events POLLPRI and POLLERR. If you
81 use select(2), set the file descriptor in exceptfds. After
82 poll(2) returns, either lseek(2) to the beginning of the sysfs
83 file and read the new value or close the file and re-open it
86 "edge" ... reads as either "none", "rising", "falling", or
87 "both". Write these strings to select the signal edge(s)
88 that will make poll(2) on the "value" file return.
90 This file exists only if the pin can be configured as an
91 interrupt generating input pin.
93 "active_low" ... reads as either 0 (false) or 1 (true). Write
94 any nonzero value to invert the value attribute both
95 for reading and writing. Existing and subsequent
96 poll(2) support configuration via the edge attribute
97 for "rising" and "falling" edges will follow this
100 GPIO controllers have paths like /sys/class/gpio/gpiochip42/ (for the
101 controller implementing GPIOs starting at #42) and have the following
102 read-only attributes:
104 /sys/class/gpio/gpiochipN/
106 "base" ... same as N, the first GPIO managed by this chip
108 "label" ... provided for diagnostics (not always unique)
110 "ngpio" ... how many GPIOs this manges (N to N + ngpio - 1)
112 Board documentation should in most cases cover what GPIOs are used for
113 what purposes. However, those numbers are not always stable; GPIOs on
114 a daughtercard might be different depending on the base board being used,
115 or other cards in the stack. In such cases, you may need to use the
116 gpiochip nodes (possibly in conjunction with schematics) to determine
117 the correct GPIO number to use for a given signal.
120 Exporting from Kernel code
121 --------------------------
122 Kernel code can explicitly manage exports of GPIOs which have already been
123 requested using gpio_request():
125 /* export the GPIO to userspace */
126 int gpiod_export(struct gpio_desc *desc, bool direction_may_change);
128 /* reverse gpio_export() */
129 void gpiod_unexport(struct gpio_desc *desc);
131 /* create a sysfs link to an exported GPIO node */
132 int gpiod_export_link(struct device *dev, const char *name,
133 struct gpio_desc *desc);
135 After a kernel driver requests a GPIO, it may only be made available in
136 the sysfs interface by gpiod_export(). The driver can control whether the
137 signal direction may change. This helps drivers prevent userspace code
138 from accidentally clobbering important system state.
140 This explicit exporting can help with debugging (by making some kinds
141 of experiments easier), or can provide an always-there interface that's
142 suitable for documenting as part of a board support package.
144 After the GPIO has been exported, gpiod_export_link() allows creating
145 symlinks from elsewhere in sysfs to the GPIO sysfs node. Drivers can
146 use this to provide the interface under their own device in sysfs with