RPMGS on imx8mm. Hard Fault on rpmsg_lite_send_nocopy

[ToradexAI]

Hi, I’m the Toradex AI Assistant. The Toradex Support team will follow up with you shortly!

Did you like my answer?

---

Hello,

It sounds like you’re encountering a challenging issue with the rpmsg_lite_send_nocopy function on your i.MX8MM platform, resulting in a hard fault or an ERROR_VRING_NO_BUFF error. This typically points to a problem with buffer management or memory allocation. Here are some steps to help you troubleshoot and resolve this issue:

1. Memory Allocation and Device Tree Configuration:

   • Verify that the memory regions used by RPMsg are correctly defined in your device tree. Ensure that the reserved memory regions for RPMsg are properly allocated and do not overlap with other memory regions.
   • Check the device tree overlays provided by Toradex for RPMsg and ensure they are correctly applied.

2. Buffer Management:

   • The incorrect head_idx value suggests there might be an issue with buffer management. Ensure that buffer indices are correctly tracked and updated.
   • Confirm that the buffer sizes and counts are correctly configured in your RPMsg setup. The number of buffers and their sizes should match between the A53 and M4 cores.

3. Synchronization and Timing:

   • Ensure proper synchronization between the A53 and M4 cores. Timing issues can lead to incorrect buffer indices if one core is reading or writing to a buffer that the other core is not ready for.

4. Code Review and Debugging:

   • Review your implementation of the RPMsg functions, especially around rpmsg_lite_send_nocopy and rpmsg_queue_recv_nocopy. Ensure all parameters and return values are handled correctly.
   • Use debugging tools to step through the code and monitor variable values, particularly around buffer management functions.

5. Example Code Comparison:

   • Compare your implementation with the rpmsg_lite_str_echo_rtos example to ensure all necessary steps and configurations are correctly followed.

By carefully reviewing these aspects, you should be able to identify and resolve the issue causing the hard fault or buffer error. If the problem persists, consider simplifying the setup to isolate the issue further.

I hope this helps you resolve the issue. If you have any more questions, feel free to ask.

Best regards,
Toradex AI Assistant

Please, also check if these links can help you solve your problem:

Verdin M4 core rpmsg example problem

Hello, I started playing around with Verdin M4 core. I have been following AN article about RUNNING freeRTOS on M4 core. I have successfully run hello_world demo app and next thing I wanted to get familiar with is RPMSG to get an idea how to communicate between A and M core. I built rpmsg pingpong example, copied debug binary to the SD card and run it in uboot: fatload mmc 1 0x7e0000 rpmsg_pingpong.bin && dcache flush && bootaux 0x7e0000 The program starts and prints: RPMSG Ping-Pong FreeRTO…

Bare metal RPMSG on Verdin

Hi, I am trying to run the RPMSG-Lite on bare metal M4. The two rpmsg examples in the MCUXpresso SDK are built with freeRTOS. Extracting the code is not the problem, but the build process is, because all the cmake settings and includes are set for FreeRTOS and not bare-metal. Only example I could find using bare metal was this one, but it uses a different (older) template of cmake files and older rpmsg version, so it is hard to translate. From the RPMsg-Lite Users’s Guide it is explained that r…

https://community.nxp.com/t5/i-MX-Processors/Where-is-the-RPMsg-document-for-IMX8MM/td-p/894428/?profile.language=zh-CN.