Re: [PATCH v3 2/2] hung_task: Enable runtime reset of hung_task_detect_count

[Aaron Tomlin]

On Fri, Dec 19, 2025 at 10:15:14PM +0800, Lance Yang wrote:
> Clever! If it's a real persistent hang, we'll catch it next
> time anyway.
> 
> But yeah, add a comment explaining that we're ok with
> undercounting one scan if counter gets reset mid-flight.
> 
> Future readers will thank you ;)

Hi Petr, Lance,

Thank you for the feedback.

> > It should be acceptable. The panic() is needed when the lockup looks
> > permanent. And if the lockup is permanent then the right
> > total_hung_task number will be computed by the next check.
> > 
> > Please, add a comment into the code explaining why it is done this way
> > and why it is OK.

Acknowledged.

> Right, memory ordering matters!

Agreed.

> Explicit is better. Echo 0 to reset, anything else is an error :)

Acknowledged.

> Yeah, I think this is the way to go. Much better than adding a
> lock to hung task detector ...

Agreed.

I will prepare a new series.


Kind regards,
-- 
Aaron Tomlin

Attachment: signature.asc
Description: PGP signature