netdev - Re: [PATCH net-next 04/14] net_sched: sch_choke: implement lockless choke

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: <20240417131404.GX2320920@kernel.org>
Date: Wed, 17 Apr 2024 14:14:04 +0100
From: Simon Horman <horms@...nel.org>
To: Eric Dumazet <edumazet@...gle.com>
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 04/14] net_sched: sch_choke: implement lockless
 choke_dump()

On Mon, Apr 15, 2024 at 01:20:44PM +0000, Eric Dumazet wrote:
> Instead of relying on RTNL, choke_dump() can use READ_ONCE()
> annotations, paired with WRITE_ONCE() ones in choke_change().
> 
> Also use READ_ONCE(q->limit) in choke_enqueue() as the value
> could change from choke_change().

Hi Eric,

I'm wondering if you could expand on why q->limit needs this treatment
but not other fields, f.e. q->parms.qth_min (aka p->qth_min).

> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> ---
>  include/net/red.h     | 10 +++++-----
>  net/sched/sch_choke.c | 23 ++++++++++++-----------
>  2 files changed, 17 insertions(+), 16 deletions(-)
> 
> diff --git a/include/net/red.h b/include/net/red.h

...

> @@ -244,7 +244,7 @@ static inline void red_set_parms(struct red_parms *p,
>  		max_P = red_maxp(Plog);
>  		max_P *= delta; /* max_P = (qth_max - qth_min)/2^Plog */
>  	}
> -	p->max_P = max_P;
> +	WRITE_ONCE(p->max_P, max_P);
>  	max_p_delta = max_P / delta;
>  	max_p_delta = max(max_p_delta, 1U);
>  	p->max_P_reciprocal  = reciprocal_value(max_p_delta);

A little further down in this function p->Scell_log is set.
I think it also needs the WRITE_ONCE() treatment as it
is read in choke_dump().

> diff --git a/net/sched/sch_choke.c b/net/sched/sch_choke.c

...

> @@ -431,15 +431,16 @@ static int choke_init(struct Qdisc *sch, struct nlattr *opt,
>  static int choke_dump(struct Qdisc *sch, struct sk_buff *skb)
>  {
>  	struct choke_sched_data *q = qdisc_priv(sch);
> +	u8 Wlog = READ_ONCE(q->parms.Wlog);
>  	struct nlattr *opts = NULL;
>  	struct tc_red_qopt opt = {
> -		.limit		= q->limit,
> -		.flags		= q->flags,
> -		.qth_min	= q->parms.qth_min >> q->parms.Wlog,
> -		.qth_max	= q->parms.qth_max >> q->parms.Wlog,
> -		.Wlog		= q->parms.Wlog,
> -		.Plog		= q->parms.Plog,
> -		.Scell_log	= q->parms.Scell_log,
> +		.limit		= READ_ONCE(q->limit),
> +		.flags		= READ_ONCE(q->flags),
> +		.qth_min	= READ_ONCE(q->parms.qth_min) >> Wlog,
> +		.qth_max	= READ_ONCE(q->parms.qth_max) >> Wlog,
> +		.Wlog		= Wlog,
> +		.Plog		= READ_ONCE(q->parms.Plog),
> +		.Scell_log	= READ_ONCE(q->parms.Scell_log),
>  	};
>  
>  	opts = nla_nest_start_noflag(skb, TCA_OPTIONS);