Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230726002904.655377-1-yosryahmed@google.com>
 <20230726002904.655377-2-yosryahmed@google.com> <CAJD7tkZK2T2ebOPw6K0M+YWyKUtx9bE2uyFj4VOehhd+fYnk8w@mail.gmail.com>
 <20230726151315.GB1365610@cmpxchg.org>
In-Reply-To: <20230726151315.GB1365610@cmpxchg.org>
From:   Yosry Ahmed <yosryahmed@google.com>
Date:   Wed, 26 Jul 2023 08:14:35 -0700
Message-ID: <CAJD7tkYS1iSo94WhtDE2GbpTkukaKY3aAKTSaUkGj0K5ifek+A@mail.gmail.com>
Subject: Re: [PATCH v2] mm: memcg: use rstat for non-hierarchical stats
To:     Johannes Weiner <hannes@cmpxchg.org>
Cc:     linux-kernel@vger.kernel.org, Michal Hocko <mhocko@kernel.org>,
        cgroups@vger.kernel.org, linux-mm@kvack.org,
        Muchun Song <muchun.song@linux.dev>,
        Shakeel Butt <shakeelb@google.com>,
        Roman Gushchin <roman.gushchin@linux.dev>,
        Andrew Morton <akpm@linux-foundation.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Jul 26, 2023 at 8:13=E2=80=AFAM Johannes Weiner <hannes@cmpxchg.org=
> wrote:
>
> On Tue, Jul 25, 2023 at 05:36:45PM -0700, Yosry Ahmed wrote:
> > On Tue, Jul 25, 2023 at 5:29=E2=80=AFPM Yosry Ahmed <yosryahmed@google.=
com> wrote:
> > > - Fix a subtle bug where updating a local counter would be missed if =
it
> > >   was cancelled out by a pending update from child memcgs.
> >
> >
> > Johannes, I fixed a subtle bug here and I kept your Ack, I wasn't sure
> > what the Ack retention policy should be here. A quick look at the fix
> > would be great.
>
> Ah, I found it:
>
> > > @@ -5542,19 +5539,23 @@ static void mem_cgroup_css_rstat_flush(struct=
 cgroup_subsys_state *css, int cpu)
> > >                         memcg->vmstats->state_pending[i] =3D 0;
> > >
> > >                 /* Add CPU changes on this level since the last flush=
 */
> > > +               delta_cpu =3D 0;
> > >                 v =3D READ_ONCE(statc->state[i]);
> > >                 if (v !=3D statc->state_prev[i]) {
> > > -                       delta +=3D v - statc->state_prev[i];
> > > +                       delta_cpu =3D v - statc->state_prev[i];
> > > +                       delta +=3D delta_cpu;
> > >                         statc->state_prev[i] =3D v;
> > >                 }
> > >
> > > -               if (!delta)
> > > -                       continue;
> > > -
> > >                 /* Aggregate counts on this level and propagate upwar=
ds */
> > > -               memcg->vmstats->state[i] +=3D delta;
> > > -               if (parent)
> > > -                       parent->vmstats->state_pending[i] +=3D delta;
> > > +               if (delta_cpu)
> > > +                       memcg->vmstats->state_local[i] +=3D delta_cpu=
;
>
> When delta nulls out, but delta_cpu is non-zero... subtle.

Yup, led to a few missed updates, was not trivial to track down :)

>
> This fixed version looks good, please keep my ack :)

Thanks!