Re: [PATCH 3/4] memcg: manipulate memcg private ID references by ID

[Bingfang Guo]

On Fri, Sep 18, 2026 at 01:24:01PM -0700, Shakeel Butt wrote:
> On Sat, Sep 19, 2026 at 02:46:52AM +0800, Bingfang Guo wrote:
>> Thanks for pointing out this.
>> 
>> Agreed.  Both two callers are already holding the rcu lock so
>> taking the lock here is unnecessary. So I will drop the
>> rcu_read_lock() and use that in the next version!
>> 
>> My concern is that: mem_cgroup_private_id_put() looks like a
>> universal put function, requiring rcu held (which is true today)
>> is not that obvious to the users.  So I think adding a short kdoc
>> comment to make it clear later might be a good idea.
> 
> This function is internal to memcg code, so whoever is going to call it better
> know to call it with rcu held because of lockdep assert. No need to have a
> comment.

Okay, I'll keep that in mind!