netdev - Re: [PATCH net-next 09/14] net_sched: sch_fq_codel: implement lockless fq_codel

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CANn89i+jt2dk4yF-k2U3iDcDgwBp-hOs9EGpdHU9F4cEcQteSw@mail.gmail.com>
Date: Wed, 17 Apr 2024 19:14:10 +0200
From: Eric Dumazet <edumazet@...gle.com>
To: Simon Horman <horms@...nel.org>
Cc: "David S. Miller" <davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>, 
	Paolo Abeni <pabeni@...hat.com>, Jamal Hadi Salim <jhs@...atatu.com>, 
	Cong Wang <xiyou.wangcong@...il.com>, Jiri Pirko <jiri@...nulli.us>, netdev@...r.kernel.org, 
	eric.dumazet@...il.com
Subject: Re: [PATCH net-next 09/14] net_sched: sch_fq_codel: implement
 lockless fq_codel_dump()

On Wed, Apr 17, 2024 at 7:07 PM Simon Horman <horms@...nel.org> wrote:
>
> On Mon, Apr 15, 2024 at 01:20:49PM +0000, Eric Dumazet wrote:
> > Instead of relying on RTNL, fq_codel_dump() can use READ_ONCE()
> > annotations, paired with WRITE_ONCE() ones in fq_codel_change().
> >
> > Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> > ---
> >  net/sched/sch_fq_codel.c | 57 ++++++++++++++++++++++++----------------
> >  1 file changed, 35 insertions(+), 22 deletions(-)
> >
> > diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c
>
> ...
>
> > @@ -529,30 +539,33 @@ static int fq_codel_dump(struct Qdisc *sch, struct sk_buff *skb)
> >               goto nla_put_failure;
> >
> >       if (nla_put_u32(skb, TCA_FQ_CODEL_TARGET,
> > -                     codel_time_to_us(q->cparams.target)) ||
> > +                     codel_time_to_us(READ_ONCE(q->cparams.target))) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_LIMIT,
> > -                     sch->limit) ||
> > +                     READ_ONCE(sch->limit)) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_INTERVAL,
> > -                     codel_time_to_us(q->cparams.interval)) ||
> > +                     codel_time_to_us(READ_ONCE(q->cparams.interval))) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_ECN,
> > -                     q->cparams.ecn) ||
> > +                     READ_ONCE(q->cparams.ecn)) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_QUANTUM,
> > -                     q->quantum) ||
> > +                     READ_ONCE(q->quantum)) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_DROP_BATCH_SIZE,
> > -                     q->drop_batch_size) ||
> > +                     READ_ONCE(q->drop_batch_size)) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_MEMORY_LIMIT,
> > -                     q->memory_limit) ||
> > +                     READ_ONCE(q->memory_limit)) ||
> >           nla_put_u32(skb, TCA_FQ_CODEL_FLOWS,
> > -                     q->flows_cnt))
> > +                     READ_ONCE(q->flows_cnt)))
>
> Hi Eric,
>
> I think you missed the corresponding update for q->flows_cnt
> in fq_codel_change().

q->flows_cnt is set at init time only, it can not change yet.

Blindly using READ_ONCE() in a dump seems good hygiene,
it is not needed yet, but does no harm.