Re: [PATCH 1/1] mm: memcontrol: fix vmstats_percpu state incorrect subtraction after reparent

[Cai Xinchen]

Yes, only cgroup v1.

On 2023/3/25 1:11, Michal Koutný wrote:
> Hello.
>
> On Mon, Mar 20, 2023 at 03:06:48AM +0000, Cai Xinchen <caixinchen1@xxxxxxxxxx> wrote:
> > When memcg C is offllined, its pages are reparented to memcg P,
> > so far P->vmstats (hierarchical) have those pages, and
> > P->vmstats_percpu (non-hierarchical) don't. When those pages get
> > uncharged, P->vmstats (hierachical) decreases, which is correct,
> > but P->vmstats_percpu (non-hierarchical) also decreases, which
> > is wrong, as those stats were never added to P->vmstats_percpu to
> > begin with.
> I was wondering why ->vmstats_percpu matters (in the end all is summed
> in ->vmstats) -- do you mean this is a cgroup v1 only issue? As only
> that exposes the non-hieararchical stats.
>
> Thanks,
> Michal