2024-09-04 03:57 PM - edited 2024-09-04 04:14 PM
Hi!
I notice one thing when I build my STM32MP151 custom board and implemented OpenSTLinux with my own .dtb and Linux kernel.
I hade enormous issues to get a basic setup of STM32MP151 up and running with Linux. I found out that the Linux kernel is destroying the working queue, which results that the linux kernel halts.
I have posted the working custom board for STM32MP151 project here. Give it a star so other knows that this is a working board for STM32MP151.
https://github.com/DanielMartensson/STM32-Computer/
Question:
When I comment this line in the file kernel/trace/trace.c, my board just works like a charm. So why is it that? I need to disable the destruction of the working queue eval_map_wp in order to make the board lose it's critical errors?
static int __init trace_eval_sync(void)
{
/* Make sure the eval map updates are finished */
if (eval_map_wq){
/*destroy_workqueue(eval_map_wq); */
}
return 0;
}
Feel free to download the gerber files and build the board. The unique thing with the board, is that it has no SD-card slot. Only an eMMC that being flashed through the USB OTG. The manufacturing cost is about 100 USD for a total build. Compare that to STM32MP157-EV board which cost 470 USD.
Solved! Go to Solution.
2024-09-06 03:43 AM
Why? I bet a lot of people understand that this is more like a hack, rather than a proper solution.
The best way for ST is to understand why this is happening. I think it's the I2C4 device that being destroyed.
I'm not going to create another thread about the same topic.
If you want to debug, just build my board and try to boot the kernel :)
2024-09-06 02:31 AM
Hi @DMårt ,
Thanks for sharing this initiative.
I need to warn you and anyone reading this post and your github that the solution you expose seems not appropriate.
Can you please open a specific topic describing the original issue you faced.
Thanks
Olivier
2024-09-06 03:43 AM
Why? I bet a lot of people understand that this is more like a hack, rather than a proper solution.
The best way for ST is to understand why this is happening. I think it's the I2C4 device that being destroyed.
I'm not going to create another thread about the same topic.
If you want to debug, just build my board and try to boot the kernel :)
2024-09-10 11:15 AM
Can you give me the configuration to the working queue? I mean, where can I disable which device that should be included inside the working queue?
2024-09-23 01:57 AM
Hi! What's the status of this bug report?
Is it included in the tasks?
By the way! I have found a way to implement a Ubuntu-eco system om STM32MP1 processors.
It's available on my GitHub URL above.