Re: [RFC PATCH] mm: memcg/slab: Stop reparented obj_cgroups from charging root

[Richard Palethorpe]

Hello,

Michal Koutný <mkoutny@xxxxxxxx> writes:

> On Fri, Oct 16, 2020 at 10:53:08AM -0400, Johannes Weiner <hannes@xxxxxxxxxxx> wrote:
>> The central try_charge() function charges recursively all the way up
>> to and including the root.
> Except for use_hiearchy=0 (which is the case here as Richard
> wrote). The reparenting is hence somewhat incompatible with
> new_parent.use_hiearchy=0 :-/
>

Yes and it also seems

new_parent.use_hierarch=0 -> new_child.use_hierarchy=0

and

new_parent.use_hierarch=0 -> new_child.use_hierarchy=1

are considered valid on cgroupsV1. The kernel will also allow more
descendants on new_child.use_hierarchy=0, but sets
broken_hierarchy=1. However this will not stop the stack trace occuring
(AFAICT) when the reparenting happens between two descendants.

>> We should clean this up one way or another: either charge the root or
>> don't, but do it consistently.
> I agree this'd be good to unify. One upside of excluding root memcg from
> charging is that users are spared from the charging overhead when memcg
> tree is not created.  (Actually, I thought that was the reason for this
> exception.)
>
> Michal


-- 
Thank you,
Richard.