Duplicated Event on GPIO Rising Edge with Nvidia Jetson Orin Nano

Issue Overview

Users are experiencing an issue with the Nvidia Jetson Orin Nano Dev board where the General Timestamp Engine (GTE) generates two timestamps for each rising edge of a GPIO input, instead of the expected single timestamp. This problem occurs during the setup of the GTE on the Linux side, specifically when attempting to timestamp GPIO events.

Specific Symptoms

• Users observe two timestamps generated for each rising edge of GPIO Z.04, which is managed by GPIO controller 1.
• The intervals between timestamps are inconsistent, with some intervals being very small (around 0.0002 seconds) while others align with the expected 10-second interval.

Context

• The setup involves a Jetson Orin Nano 4GB module paired with a custom carrier board, running Jetpack 5.1.1 / L4T 35.3.1.
• The user is utilizing the LIC GTE since the AON GTE is already in use by another component (SPE).
• The issue persists despite confirming that only GPIO Z.04 is reporting interrupts and that no other GPIOs are causing interference.

Impact

This issue significantly affects the reliability of timestamping events, which is crucial for applications relying on accurate timing for GPIO interactions. The presence of duplicate timestamps can lead to confusion in data processing and application logic.

Possible Causes

• Hardware Incompatibilities or Defects: There may be issues with the custom carrier board or its configuration that could lead to unexpected behavior in GPIO handling.

• Software Bugs or Conflicts: There could be bugs in the kernel module or conflicts with existing drivers that handle GPIO events improperly.

• Configuration Errors: Incorrect settings in the kernel module or improper initialization of the GTE could result in duplicate event handling.

• Driver Issues: The driver responsible for managing GPIO interrupts may not be functioning correctly, leading to multiple interrupts being fired for a single event.

• Environmental Factors: External factors like noise on the GPIO lines could potentially trigger multiple interrupts.

• User Errors or Misconfigurations: Misconfigurations during setup or incorrect parameters passed to kernel modules may lead to this issue.

Troubleshooting Steps, Solutions & Fixes

Step-by-Step Instructions

1. Verify Hardware Setup:

   • Ensure that the custom carrier board is functioning correctly and that GPIO Z.04 is properly connected and configured.

2. Check Kernel Module Source:

   • Use the tegra194_gte_test.c kernel module as a base for testing. Apply any necessary patches as indicated in user discussions.

3. Build and Load Kernel Module:

   • Compile the kernel module and load it using:

     sudo insmod tegra194_gte_test.ko gpio_in=482 gpio_out=483 lic_irq=296
     

4. Enable GTE Timestamping:

   • Enable timestamping by executing:

     sudo su -c "echo 1 > /sys/kernel/tegra_gte_test/gpio_en_dis"
     

5. Monitor Interrupts:

   • Check /proc/interrupts to verify which GPIOs are generating interrupts when input and output pins are connected together.

6. Debugging Output:

   • Use dmesg to capture kernel logs after running tests to identify any anomalies or repeated messages related to GPIO interrupts.

7. Debounce Configuration:

   • If applicable, try adding debounce time on the input GPIO to filter out noise that might cause multiple interrupts.

8. Isolation Tests:

   • Test with different input/output GPIO pairs to see if the issue persists across different configurations.

9. Disable Unused Interrupts:

   • Consider disabling other interrupt sources (like tegra-hsp) in the device tree if they are suspected of causing interference.

Recommended Fixes

• If multiple users reported success with specific configurations, prioritize those setups for further testing.

• Consider updating drivers or firmware if newer versions are available that address known issues with GPIO handling.

Best Practices

• Regularly check for updates from Nvidia regarding Jetpack and L4T releases as they may include fixes for similar issues.

• Maintain clear documentation of any changes made during troubleshooting to assist in future investigations or support requests.

Unresolved Aspects

Further investigation may be needed into how other components (like tegra-hsp) interact with GPIO events and whether additional configuration changes could mitigate this issue. Users should remain open to community feedback and updates regarding potential patches or solutions from Nvidia’s support channels.