Jetson TX1, TX2, AGX Xavier, and Nano development boards contain a 40 pin GPIO header, similar to the 40 pin header in the Raspberry Pi. These GPIOs can be controlled for digital input and output using the Python library provided in the Jetson GPIO Library package. The library has the same API as the RPi.GPIO library for Raspberry Pi in order to provide an easy way to move applications running on the Raspberry Pi to the Jetson board.

This document walks through what is contained in The Jetson GPIO library package, how to configure the system and run the provided sample applications, and the library API.

In addition to this document, the Jetson GPIO library package contains the following:

1. The lib/python/ subdirectory contains the Python modules that implement all library functionality. The gpio.py module is the main component that will be imported into an application and provides the needed APIs. The gpio_event.py and gpio_pin_data.py modules are used by the gpio.py module and must not be imported directly in to an application.

2. The samples/ subdirectory contains sample applications to help in getting familiar with the library API and getting started on an application. The simple_input.py and simple_output.py applications show how to perform read and write to a GPIO pin respectively, while the button_led.py, button_event.py and button_interrupt.py show how a button press may be used to blink an LED using busy-waiting, blocking wait and interrupt callbacks respectively.

The easiest way to install this library is using pip:

sudo pip install Jetson.GPIO

You may clone this git repository, or download a copy of it as an archive file and decompress it. You may place the library files anywhere you like on your system. You may use the library directly from this directory by manually setting PYTHONPATH, or install it using setup.py:

sudo python3 setup.py install

In order to use the Jetson GPIO Library, the correct user permissions/groups must be set first.

Create a new gpio user group. Then add your user to the newly created group.

sudo groupadd -f -r gpio
sudo usermod -a -G gpio your_user_name

Install custom udev rules by copying the 99-gpio.rules file into the rules.d directory.

If you have downloaded the source to Jetson.GPIO:

sudo cp lib/python/Jetson/GPIO/99-gpio.rules /etc/udev/rules.d/

If you installed Jetson.GPIO from a package, e.g. using pip into a virtual environment:

sudo cp venv/lib/pythonNN/site-packages/Jetson/GPIO/99-gpio.rules /etc/udev/rules.d/

For the new rule to take place, you either need to reboot or reload the udev rules by running:

sudo udevadm control --reload-rules && sudo udevadm trigger

With the permissions set as needed, the sample applications provided in the samples/ directory can be used. The following describes the operation of each application:

1. simple_input.py: This application uses the BCM pin numbering mode and reads the value at pin 12 of the 40 pin header and prints the value to the screen.

2. simple_out.py: This application uses the BCM pin numbering mode from Raspberry Pi and outputs alternating high and low values at BCM pin 18 (or board pin 12 on the header) every 2 seconds.

3. button_led.py: This application uses the BOARD pin numbering. It requires a button connected to pin 18 and GND, a pull-up resistor connecting pin 18 to 3V3 and an LED and current limiting resistor connected to pin 12. The application reads the button state and keeps the LED on for 1 second every time the button is pressed.

4. button_event.py: This application uses the BOARD pin numbering. It requires a button connected to pin 18 and GND, a pull-up resistor connecting the button to 3V3 and an LED and current limiting resistor connected to pin 12. The application performs the same function as the button_led.py but performs a blocking wait for the button press event instead of continuously checking the value of the pin in order to reduce CPU usage.

5. button_interrupt.py: This application uses the BOARD pin numbering. It requires a button connected to pin 18 and GND, a pull-up resistor connecting the button to 3V3, an LED and current limiting resistor connected to pin 12 and a second LED and current limiting resistor connected to pin 13. The application slowly blinks the first LED continuously and rapidly blinks the second LED five times only when the button is pressed.

To run these sample applications if Jetson.GPIO is added to the PYTHONPATH:

python3 <name_of_application_to_run>

Alternatively, if Jetson.GPIO is not added to the PYTHONPATH, the run_sample.sh script can be used to run these sample applications. This can be done with the following command when in the samples/ directory:

./run_sample.sh <name_of_application_to_run>

The usage of the script can also be viewed by using:

./run_sample.sh -h
./run_sample.sh --help

The Jetson GPIO library provides all public APIs provided by the RPi.GPIO library. The following discusses the use of each API:

To import the Jetson.GPIO module use:

import Jetson.GPIO as GPIO

This way, you can refer to the module as GPIO throughout the rest of the application. The module can also be imported using the name RPi.GPIO instead of Jetson.GPIO for existing code using the RPi library.

The Jetson GPIO library provides four ways of numbering the I/O pins. The first two correspond to the modes provided by the RPi.GPIO library, i.e BOARD and BCM which refer to the pin number of the 40 pin GPIO header and the Broadcom SoC GPIO numbers respectively. The remaining two modes, CVM and TEGRA_SOC use strings instead of numbers which correspond to signal names on the CVM/CVB connector and the Tegra SoC respectively.

To specify which mode you are using (mandatory), use the following function call:

GPIO.setmode(GPIO.BOARD)
# or
GPIO.setmode(GPIO.BCM)
# or
GPIO.setmode(GPIO.CVM)
# or
GPIO.setmode(GPIO.TEGRA_SOC)

To check which mode has be set, you can call:

mode = GPIO.getmode()

The mode must be one of GPIO.BOARD, GPIO.BCM, GPIO.CVM, GPIO.TEGRA_SOC or None.

It is possible that the GPIO you are trying to use is already being used external to the current application. In such a condition, the Jetson GPIO library will warn you if the GPIO being used is configured to anything but the default direction (input). It will also warn you if you try cleaning up before setting up the mode and channels. To disable warnings, call:

GPIO.setwarnings(False)

The GPIO channel must be set up before use as input or output. To configure the channel as input, call:

# (where channel is based on the pin numbering mode discussed above)
GPIO.setup(channel, GPIO.IN)

To set up a channel as output, call:

GPIO.setup(channel, GPIO.OUT)

It is also possible to specify an initial value for the output channel:

GPIO.setup(channel, GPIO.OUT, initial=GPIO.HIGH)

When setting up a channel as output, it is also possible to set up more than one channel at once:

# add as many as channels as needed. You can also use tuples: (18,12,13)
channels = [18, 12, 13]
GPIO.setup(channels, GPIO.OUT)

To read the value of a channel, use:

GPIO.input(channel)

This will return either GPIO.LOW or GPIO.HIGH.

To set the value of a pin configured as output, use:

GPIO.output(channel, state)

where state can be GPIO.LOW or GPIO.HIGH.

You can also output to a list or tuple of channels:

channels = [18, 12, 13] # or use tuples
GPIO.output(channels, GPIO.HIGH) # or GPIO.LOW
# set first channel to LOW and rest to HIGH
GPIO.output(channel, (GPIO.LOW, GPIO.HIGH, GPIO.HIGH))

At the end of the program, it is good to clean up the channels so that all pins are set in their default state. To clean up all channels used, call:

GPIO.cleanup()

If you don't want to clean all channels, it is also possible to clean up individual channels or a list or tuple of channels:

GPIO.cleanup(chan1) # cleanup only chan1
GPIO.cleanup([chan1, chan2]) # cleanup only chan1 and chan2
GPIO.cleanup((chan1, chan2))  # does the same operation as previous statement

To get information about the Jetson module, use/read:

GPIO.JETSON_INFO

This provides a Python dictionary with the following keys: P1_REVISION, RAM, REVISION, TYPE, MANUFACTURER and PROCESSOR. All values in the dictionary are strings with the exception of P1_REVISION which is an integer.

To get information about the library version, use/read:

GPIO.VERSION

This provides a string with the X.Y.Z version format.

Aside from busy-polling, the library provides three additional ways of monitoring an input event:

This function blocks the calling thread until the provided edge(s) is detected. The function can be called as follows:

GPIO.wait_for_edge(channel, GPIO.RISING)

The second parameter specifies the edge to be detected and can be GPIO.RISING, GPIO.FALLING or GPIO.BOTH. If you only want to limit the wait to a specified amount of time, a timeout can be optionally set:

# timeout is in milliseconds
GPIO.wait_for_edge(channel, GPIO.RISING, timeout=500)

The function returns the channel for which the edge was detected or None if a timeout occurred.

This function can be used to periodically check if an event occurred since the last call. The function can be set up and called as follows:

# set rising edge detection on the channel
GPIO.add_event_detect(channel, GPIO.RISING)
run_other_code()
if GPIO.event_detected(channel):
    do_something()

As before, you can detect events for GPIO.RISING, GPIO.FALLING or GPIO.BOTH.

This feature can be used to run a second thread for callback functions. Hence, the callback function can be run concurrent to your main program in response to an edge. This feature can be used as follows:

# define callback function
def callback_fn(channel):
    print("Callback called from channel %s" % channel)

# add rising edge detection
GPIO.add_event_detect(channel, GPIO.RISING, callback=callback_fn)

More than one callback can also be added if required as follows:

def callback_one(channel):
    print("First Callback")

def callback_two(channel):
    print("Second Callback")

GPIO.add_event_detect(channel, GPIO.RISING)
GPIO.add_event_callback(channel, callback_one)
GPIO.add_event_callback(channel, callback_two)

The two callbacks in this case are run sequentially, not concurrently since there is only thread running all callback functions.

In order to prevent multiple calls to the callback functions by collapsing multiple events in to a single one, a debounce time can be optionally set:

# bouncetime set in milliseconds
GPIO.add_event_detect(channel, GPIO.RISING, callback=callback_fn,
bouncetime=200)

If the edge detection is not longer required it can be removed as follows:

GPIO.remove_event_detect(channel)

This feature allows you to check the function of the provided GPIO channel:

GPIO.gpio_function(channel)

The function returns either GPIO.IN or GPIO.OUT.

See samples/simple_pwm.py for details on how to use PWM channels.

The Jetson.GPIO library supports PWM only on pins with attached hardware PWM controllers. Unlike the RPi.GPIO library, the Jetson.GPIO library does not implement Software emulated PWM. Jetson Nano supports 2 PWM channels, and Jetson AGX Xavier supports 3 PWM channels. Jetson TX1 and TX2 do not support any PWM channels.

The system pinmux must be configured to connect the hardware PWM controlller(s) to the relevant pins. If the pinmux is not configured, PWM signals will not reach the pins! The Jetson.GPIO library does not dynamically modify the pinmux configuration to achieve this. Read the L4T documentation for details on how to configure the pinmux.

The following describes how to use the Jetson GPIO library from a docker container.

samples/docker/Dockerfile is a sample Dockerfile for the Jetson GPIO library. The following command will build a docker image named testimg from it.

sudo docker image build -f samples/docker/Dockerfile -t testimg .

You should map /sys/devices, /sys/class/gpio into the container to access to the GPIO pins. So you need to add these options to docker container run command.

-v /sys/devices/:/sys/devices/ \
-v /sys/class/gpio:/sys/class/gpi

and if you want to use GPU from the container you also need to add these options:

--runtime=nvidia --gpus all

The library determines the jetson model by checking /proc/device-tree/compatible and /proc/device-tree/chosen by default. These paths only can be mapped into the container in privilleged mode.

The following example will run /bin/bash from the container in privilleged mode.

sudo docker container run -it --rm \
--runtime=nvidia --gpus all \
--privileged \
-v /proc/device-tree/compatible:/proc/device-tree/compatible \
-v /proc/device-tree/chosen:/proc/device-tree/chosen \
-v /sys/devices/:/sys/devices/ \
-v /sys/class/gpio:/sys/class/gpio \
testimg /bin/bash

If you don't want to run the container in privilleged mode, you can directly provide your jetson model name to the library through the environment variable JETSON_MODEL_NAME:

# ex> -e JETSON_MODEL_NAME=JETSON_NANO
-e JETSON_MODEL_NAME=[PUT_YOUR_JETSON_MODEL_NAME_HERE]

You can get the proper value for this variable by running samples/jetson_model.py on the host or in previlleged mode.

# run on the host or in previlleged mode
sudo python3 samples/jetson_model.py

The following example will run /bin/bash from the container in non-privilleged mode.

sudo docker container run -it --rm \
--runtime=nvidia --gpus all \
-v /sys/devices/:/sys/devices/ \
-v /sys/class/gpio:/sys/class/gpio \
-e JETSON_MODEL_NAME=[PUT_YOUR_JETSON_MODEL_NAME_HERE] \
testimg /bin/bash

The L4T documentation may be available in the following locations:

• Jetson Download Center; search for the "L4T Documentation" package.
• docs.nvidia.com.

Within the documentation, relevant topics may be found by searching for e.g.:

• Hardware Setup.
• Configuring the 40-Pin Expansion Header.
• Jetson-IO.
• Platform Adaptation and Bring-Up.
• Pinmux Changes.