# Access to GPIO on Raspberry Pi? **URL:** \ **Category:** Discussion\ **Tags:** raspberry-pi\ **Created:** [September 13, 2021, 12:13pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438 "2021-09-13T12:13:33Z")\ **Posts on this page:** 13\ **Page:** 1
**Author:** ![MrProLogic](https://avatars.discourse-cdn.com/v4/letter/m/848f3c/32.png) [@MrProLogic](https://swi-prolog.discourse.group/u/MrProLogic)\ **Post date:** [September 13, 2021, 12:13pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/1 "2021-09-13T12:13:33Z")
I am doing a little reading around before beginning a project and looking to use Prolog for a control system. Does SWI-Prolog give you access to the GPIO pins on a Raspberry Pi? Is it possible to initiate Prolog searches by a GPIO input and produce a signal on a pin as the result of a rule? ---
**Author:** ![grossdan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/grossdan/32/23_2.png) [@grossdan](https://swi-prolog.discourse.group/u/grossdan)\ **Post date:** [September 13, 2021, 12:36pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/2 "2021-09-13T12:36:28Z")
I don’t think you can access the pin directly from Prolog but if there exists a c based library to access the pins, then the FFI package could possibly be used. [https://www.swi-prolog.org/pack/list?p=ffi](https://www.swi-prolog.org/pack/list?p=ffi) Edit – Perhaps its possible to use a setup that is combined with Python. Python accessess the pins and forwards the values to Prolog --with Prolog embedded or called from Python. ---
**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\ **Post date:** [September 13, 2021, 1:15pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/3 "2021-09-13T13:15:04Z")
A quick Google search leads me to [How to Control GPIO Hardware from C or C++ | ICS](https://www.ics.com/blog/how-control-gpio-hardware-c-or-c), which tells me that the GPIO pins are simply controlled through the filesystem. Translating the code there gives something like below. Not tested (I have a Pi, but nothing to connect to the GPIO pins directly ready ☹ ). ```prolog blink :- setup_call_cleanup( open('/sys/class/gpio/export', write, Out1), write(Out1, 24), close(Out1)), setup_call_cleanup( open('/sys/class/gpio24/direction', write, Out2), write(Out2, out2), close(Out2)), setup_call_cleanup( open('/sys/class/gpio/gpio24/value', write, Out3), blink(100, Out3), close(Out3)). blink(N, Out) :- N > 0, !, write(Out, 1), flush_output(Out), sleep(0.05), write(Out, 0), flush_output(Out), sleep(0.05), N2 is N - 1, blink(N2, Out). blink(_,_). ``` ---
**Author:** ![grossdan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/grossdan/32/23_2.png) [@grossdan](https://swi-prolog.discourse.group/u/grossdan)\ **Post date:** [September 13, 2021, 1:25pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/4 "2021-09-13T13:25:07Z")
thats great. I tried to see if reading from the file system would then be reading input – and i guess the approach is polling driven … to identify a change in input. ---
**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\ **Post date:** [September 13, 2021, 1:42pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/5 "2021-09-13T13:42:09Z")
A more complete overview of the fs API is at [GPIO Programming: Using the sysfs Interface | ICS](https://www.ics.com/blog/gpio-programming-using-sysfs-interface). Seems you can read `'/sys/class/gpio/gpioNN/value'` and get a 0 or a 1. That seems to be a polling interface ☹ I don’t see what you need to poll again, i.e., do you have to go through the open/close or maybe you can use seek/4 to go back to the start and try again. One needs better docs or a Pi to tell 🙂 Easy enough to build a nice library and submit as a pack. Someone? ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [September 19, 2021, 7:29pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/6 "2021-09-19T19:29:36Z")
The `sysfs` gpio interface has been [deprecated](https://www.kernel.org/doc/html/latest/admin-guide/gpio/sysfs.html). The new kernel inteface is through a character device that can be easily accessible using [libgpiod](https://git.kernel.org/pub/scm/libs/libgpiod/libgpiod.git/). It is easy enough to use by calling the executables included with `libgpiod` or it can be called directly from C with ffi. One of the nice executables provided by libgpiod (besides `gpioget` and `gpioset`) is called `gpiomon`. It allows monitoring rising edges, falling edges and even setup pull down or pull up biases, printing to stdout as events happen: ```prolog NAME lt-gpiomon - manual page for lt-gpiomon v1.6.3 SYNOPSIS lt-gpiomon [OPTIONS] ... DESCRIPTION Wait for events on GPIO lines and print them to standard output OPTIONS -h, --help: display this message and exit -v, --version: display the version and exit -l, --active-low: set the line active state to low -B, --bias=[as-is|disable|pull-down|pull-up] (defaults to 'as-is'): set the line bias -n, --num-events=NUM: exit after processing NUM events -s, --silent: don't print event info -r, --rising-edge: only process rising edge events -f, --falling-edge: only process falling edge events -b, --line-buffered: set standard output as line buffered -F, --format=FMT specify custom output format Biases: as-is: leave bias unchanged disable: disable bias pull-up: enable pull-up pull-down: enable pull-down [...] ``` Here is an [article](https://blog.adafruit.com/2018/11/26/sysfs-is-dead-long-live-libgpiod-libgpiod-for-linux-circuitpython/) showing why libgpiod is much better than sysfs. ---
**Author:** ![grossdan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/grossdan/32/23_2.png) [@grossdan](https://swi-prolog.discourse.group/u/grossdan)\ **Post date:** [September 20, 2021, 4:57am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/7 "2021-09-20T04:57:15Z")
A short tutorial on the use of FFI to directly access c would be great. Dan ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [September 20, 2021, 5:11am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/8 "2021-09-20T05:11:18Z")
Check out the [ffi add-on page](https://www.swi-prolog.org/pack/list?p=ffi). It has a simple example to call the [`statfs`](https://man7.org/linux/man-pages/man2/statfs.2.html) libc function: ```prolog :- use_module(library(ffi)). :- use_module(library(cerror)). :- c_import("#include ", % include file where C functions are found [libc], % dynamic library(ies) to load [statfs(string, -struct(statfs), [int]) % function(s) to import, and description % of parameters and return value ]). statfs(File, FsStat) :- % regular prolog predicate statfs(File, FsStat, Status), % now we call the C function posix_status(Status, statfs, file, File). % covert posix errors to prolog exception ``` Query: ```prolog ?- statfs(".", FsStat), c_load(FsStat[f_bavail], A). FsStat = (0x55a2c9d5bae0), A = 66197957. ``` More detailed documentation for the ffi add-on is found in this [pdf](https://github.com/JanWielemaker/ffi/wiki/files/ffi.pdf). **EDIT** : By the way, strictly speaking you don’t have to use posix\_status/4 above, it is just used to convert posix error codes into prolog exceptions. ---
**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\ **Post date:** [September 20, 2021, 8:54am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/9 "2021-09-20T08:54:07Z")
> [@swi](#): > > The new kernel inteface is through a character device that can be easily accessible using [libgpiod](https://git.kernel.org/pub/scm/libs/libgpiod/libgpiod.git/). Possibly simply opening, reading and writing to this character device is easier than using libgpio? Sometimes it is … The advantage of using a library is than mostly that if the kernel interface changes again, the library is probably changed to deal with that. Usually, kernel interfaces are pretty stable though. ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [September 20, 2021, 8:39pm UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/10 "2021-09-20T20:39:10Z")
The advantage of using libgpiod, instead of the character device directly, is that libgpiod provides for managing for simultaneous access to the gpio from several processes/threads. As it says in the [linux docs](https://www.kernel.org/doc/html/latest/driver-api/gpio/using-gpio.html): > For structured and managed applications, we recommend that you make use of the [libgpiod](https://git.kernel.org/pub/scm/libs/libgpiod/libgpiod.git/) library. This provides helper abstractions, command line utlities and arbitration for multiple simultaneous consumers on the same GPIO chip. ---
**Author:** ![nicos](https://avatars.discourse-cdn.com/v4/letter/n/7feea3/32.png) [@nicos](https://swi-prolog.discourse.group/u/nicos)\ **Post date:** [September 27, 2021, 10:30am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/11 "2021-09-27T10:30:41Z")
Is this example slightly incomplete ? As far as I can see posix\_status/4 is defined in ffi/prolog/cerror.pl and should be loaded with something like ```prolog :- use_module(library(cerror)). ``` ? Regards, ## Nicos Angelopoulos > **[Nicos Angelopoulos](https://stoics.org.uk/~nicos/)** > > logic programming ---
**Author:** ![swi](https://avatars.discourse-cdn.com/v4/letter/s/51bf81/32.png) [@swi](https://swi-prolog.discourse.group/u/swi)\ **Post date:** [September 28, 2021, 5:51am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/12 "2021-09-28T05:51:55Z")
Nicos, you are correct, fixed the example above. Should probably be fixed on the ffi add-ons README also. Thanks for pointing it out. ---
**Author:** ![jan](https://yyz2.discourse-cdn.com/free1/user_avatar/swi-prolog.discourse.group/jan/32/4_2.png) [@jan](https://swi-prolog.discourse.group/u/jan)\ **Post date:** [September 28, 2021, 7:07am UTC](https://swi-prolog.discourse.group/t/access-to-gpio-on-raspberry-pi/4438/13 "2021-09-28T07:07:49Z")
> [@swi](#): > > Should probably be fixed on the ffi add-ons README also. Fixed. Also updated version and pushed a new version with the issue solved for Nicos using this package on the Raspberry pi using Ubuntu 21.04 (variation in `ldconfig` output and deal with systems where `char` is by default _unsigned_). The library now provides a primitive type `schar` next to `uchar` and `char`, where `schar` is explicitly signed and `char` does what the compiler does by default.