Re: [PATCH v2 -next] cgroup: remove offline draining in root destruction to avoid hung_tasks

[Chen Ridong]

On 2025/8/14 23:17, Michal Koutný wrote:
> Hi Ridong.
> 
> On Thu, Jul 31, 2025 at 07:53:02PM +0800, Chen Ridong <chenridong@xxxxxxxxxxxxxxx> wrote:
>> Have you come up with a better solution for this?
>> Would appreciate your thoughts when you have time.
> 
> Sorry for taking so long. (Also expect my next response here may be
> slow.)
> I tried reproducing it with the described LTP tests [1] (to get a better
> idea about what and why needs to be offlined) but I cannot bring it to a
> hang nor lockdep report. How do you launch the particular LTP tests to
> trigger this?
> 
> Thanks,
> Michal
> 
> [1]
> while true ; do
> 	/opt/ltp/testcases/bin/cgroup_fj_function.sh net_cls $pp
> 	/opt/ltp/testcases/bin/cgroup_fj_function.sh perf_event
> done
> (with pp both `;` or `&` for concurrent runs, two vCPUs)

Hi Michal,

I’ve provided details on how to reproduce the issue—it’s quite straightforward. For your reference,
here’s the link to the discussion:

https://lore.kernel.org/cgroups/btaaerpdl3bolxbysbqcig6kiccdgsoo32td64sk6yo4m5l5zy@nds6s35p6e6w/T/#m01f1229f2e84e1186abaf04378b4c0f47151f7e4

Let me know if you need further clarification.

-- 
Best regards,
Ridong