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: <20220829154805.1205507-1-mlombard@redhat.com> <Yw3lXaDTfvJcEM4Q@linutronix.de>
In-Reply-To: <Yw3lXaDTfvJcEM4Q@linutronix.de>
From:   Maurizio Lombardi <mlombard@redhat.com>
Date:   Tue, 30 Aug 2022 17:48:06 +0200
Message-ID: <CAFL455=4SBhJ4LpiPQr+PmL2ShuTpR=bAo8JVfe-2x3cg85cbQ@mail.gmail.com>
Subject: Re: [PATCH RFC] mm: slub: fix flush_cpu_slab()/__free_slab()
 invocations in task context.
To:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc:     Vlastimil Babka <vbabka@suse.cz>, linux-mm <linux-mm@kvack.org>,
        David Rientjes <rientjes@google.com>,
        Pekka Enberg <penberg@kernel.org>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        LKML <linux-kernel@vger.kernel.org>, tglx@linutronix.de
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

=C3=BAt 30. 8. 2022 v 12:24 odes=C3=ADlatel Sebastian Andrzej Siewior
<bigeasy@linutronix.de> napsal:
>
> On 2022-08-29 17:48:05 [+0200], Maurizio Lombardi wrote:
> > diff --git a/mm/slub.c b/mm/slub.c
> > index 862dbd9af4f5..d46ee90651d2 100644
> > --- a/mm/slub.c
> > +++ b/mm/slub.c
> > @@ -2681,30 +2681,34 @@ struct slub_flush_work {
> >       bool skip;
> >  };
> >
> > +static void flush_cpu_slab(void *d)
> > +{
> > +     struct kmem_cache *s =3D d;
> > +     struct kmem_cache_cpu *c =3D this_cpu_ptr(s->cpu_slab);
> > +
> > +     if (c->slab)
> > +             flush_slab(s, c);
> > +
> > +     unfreeze_partials(s);
> > +}
> =E2=80=A6
> > @@ -2721,13 +2725,18 @@ static void flush_all_cpus_locked(struct kmem_c=
ache *s)
> >       lockdep_assert_cpus_held();
> >       mutex_lock(&flush_lock);
> >
> > +     if (in_task()) {
> > +             on_each_cpu_cond(has_cpu_slab, flush_cpu_slab, s, 1);
>
> This blocks with disabled preemption until it completes flush_cpu_slab()
> on all CPUs.
> That function acquires a local_lock_t which can not be
> acquired from in-IRQ which is where this function will be invoked due to
> on_each_cpu_cond().

Hmm, this is not good indeed. I guess I should have used for_each_online_cp=
u()
instead of on_each_cpu_cond().

>
> Couldn't we instead use a workqueue with that WQ_MEM_RECLAIM bit? It may
> reclaim memory after all ;)

That should also fix it, do you think it would be ok to allocate a workqueu=
e in
in kmem_cache_init() ?

Thanks,
Maurizio