AM625: rcu_preempt detected stalls with SDK10.00.07 after patched

Hi Divyansh，

Want to know if you accepted the report.

Tony 

Divyansh is currently on business travel till March 17th. What is the action here?

This issue did get resolved per other customers with these patches. So is your customer still facing the issue even though it is drastically reduced?

Hi Tony,

Is it possible for your customer to conduct tests with SDK10.01.10.04? It has the DSS fix implemented and dev team was wondering if you can test with 10.01 baseline.

Additionally, any logs your customer can share? The following ticket has debug statements your customer can add as well: e2e.ti.com/.../5450993. It would be helpful if they can provide the information. 

Regards,
Krunal

Hi Krunal,

I see all change under drivers/gpu/drm/tidss folder, so does the error only relate to GPU? if did not use GPU, but used display, is it the same issue?

Hi Tony,
The patches are on tidss driver and not the gpu (even though it is placed within that folder). So yes the issue is the same.
If possible, please check on 11.01 SDK now, there has been one additional patch that may be related to your case as well.

Hi Divyansh,

Is this the latest patch?  some customer can't upgrade to SDK11.1 due to project schedule, a patch is convenient. 

 RE: AM625: Issue about tidss rcu_preempt self-detected stall on CPU 

Hi Tony,
Yes, https://lore.kernel.org/lkml/20241021-tidss-irq-fix-v1-0-82ddaec94e4a@ideasonboard.com/#t contain the patch series implemented in 10.x.

In case the above does not work, try this patch on top of it: https://lore.kernel.org/all/20250228140420.379498-1-andreas@kemnade.info/