Re: [PATCH 1/3] cgroup/cpuset: Make cpuset_fork() handle CLONE_INTO_CGROUP properly

[Waiman Long <longman@redhat.com>]

On 4/4/23 09:52, Waiman Long wrote:
>
> On 4/4/23 05:19, Michal Koutný wrote:
>> On Mon, Apr 03, 2023 at 01:18:42PM -0400, Waiman Long 
>> <longman@redhat.com> wrote:
>>> 1) PF_NO_SETAFFINITY flag - which won't be set in the case of fork() 
>>> as it
>>> is for kthread only.
>>> 2) DL bandwidth - Juri has a cpuset outstanding to modify the way 
>>> this check
>>> is being done. I want to wait until it is settled before tackling 
>>> this, if
>>> necessary.
>> BTW what about CLONE_INTO_CGROUP where the target cpuset has empty
>> effective cpuset?
> Good point. That will require a can_fork() method then. I will look 
> into that.
>>
>>> 3) security_task_setscheduler() - the CLONE_INTO_CGROUP code has 
>>> already
>>> checked that, we don't need to duplicate the check.
>> Not sure what this refers to.
> It is just checking of the task the has privilege of running into that 
> cgroup.
>>
>>> So we don't need a can_fork() check for now.
>> Anyway, good breakdown. Could you please add it to the commit message
>> too?
>
> Yes, I can put that into the commit log.

I decide to add the can_fork method to do a pre-check. So I don't think 
I need to update the commit log of this patch.

Cheers,
Longman