How to set GPIO pins as Input or Output on Nvidia Jetson Orin Nano

Issue Overview

Users are experiencing difficulties setting GPIO pins as inputs or outputs on the Nvidia Jetson Orin Nano development board. The main symptoms include:

• Out of the box, using libgpiod and Python, only GPIO pins predefined as inputs can be used as inputs, and only pins predefined as outputs can be used as outputs
• This is different behavior compared to the older Jetson Nano, where input signals could be read on GPIO pins without making changes to the device tree
• Applying the desired input/output function seems to require building the kernel from source according to some guides, which is confusing and cumbersome

The issue impacts the user experience and makes it difficult to use GPIO functionality on the Jetson Orin Nano compared to other embedded platforms. Users need a clear, straightforward method to configure GPIO pins without having to compile the entire kernel.

Possible Causes

There are a few potential causes for this GPIO configuration issue on the Jetson Orin Nano:

1. Software limitations: The default software setup, drivers, or libraries on the Jetson Orin Nano may have limitations or bugs preventing easy GPIO configuration, unlike the previous Jetson Nano.

2. Device tree configuration: The predefined input/output states of GPIO pins are likely controlled by the device tree. Modifying the device tree seems necessary to change pin functions, but the process is unclear.

3. Incomplete or confusing documentation: The available guides and resources for configuring GPIO on the Jetson Orin Nano may be incomplete, contradictory, or not applicable to the latest software versions, leading to user confusion.

4. User misconfiguration: It’s possible some users are not following the correct steps or are misinterpreting the guides, leading to the perception that kernel compilation is necessary.

Troubleshooting Steps, Solutions & Fixes

To resolve the GPIO input/output configuration issue on the Nvidia Jetson Orin Nano, follow these steps:

1. Use the Pinmux spreadsheet to configure GPIO functions:

   • Open the Pinmux spreadsheet provided by Nvidia for the Jetson Orin Nano
   • Locate the desired GPIO pins and set their function to either "Input" or "Output" as needed
   • Generate the .dtsi (Device Tree Source Include) files from the spreadsheet

2. Apply the Pinmux changes without building the kernel:

   • Contrary to some guides, building the kernel from source is not required to apply Pinmux changes
   • Instead, follow these steps to flash the changes:

     $ tar xf ${L4T_RELEASE_PACKAGE}
     $ sudo tar xpf ${SAMPLE_FS_PACKAGE} -C Linux_for_Tegra/rootfs/
     $ cd Linux_for_Tegra/
     $ sudo ./tools/l4t_flash_prerequisites.sh
     $ sudo ./apply_binaries.sh
     

   • Place the generated .dtsi files in the correct directories within the extracted Linux_for_Tegra package before running the flash commands

3. Flash from a Linux host PC or Docker container:

   • Flashing should be done from an x86 Ubuntu host PC, not from the Jetson device itself
   • If your host PC runs a different Linux distribution, use an Ubuntu Docker container to avoid compatibility issues
   • Flashing from a virtual machine is not recommended and can lead to problems

4. Verify GPIO configuration after flashing:

   • After flashing the Jetson Orin Nano with the updated Pinmux configuration, verify the GPIO pins are set to the desired input/output functions
   • Use the gpioinfo command or check the pin states programmatically to confirm the changes were applied successfully

By following the Pinmux spreadsheet method and flashing the changes from a compatible Linux environment, users should be able to configure GPIO pins as inputs or outputs on the Jetson Orin Nano without needing to build the kernel from source. If issues persist, consult the Nvidia documentation, forums, or support channels for further assistance.