From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [patch net-next v3 08/11] tbf: fix value set for q->ptokens
Date: Sat, 09 Feb 2013 17:30:46 -0800
Message-ID: <1360459846.20362.2.camel@edumazet-glaptop>
References: <1360428312-1277-1-git-send-email-jiri@resnulli.us>
	 <1360428312-1277-9-git-send-email-jiri@resnulli.us>
Mime-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Cc: netdev@vger.kernel.org, davem@davemloft.net, edumazet@google.com,
	jhs@mojatatu.com, kuznet@ms2.inr.ac.ru, j.vimal@gmail.com
To: Jiri Pirko <jiri@resnulli.us>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-da0-f47.google.com ([209.85.210.47]:34486 "EHLO
	mail-da0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1760897Ab3BJBau (ORCPT
	<rfc822;netdev@vger.kernel.org>); Sat, 9 Feb 2013 20:30:50 -0500
Received: by mail-da0-f47.google.com with SMTP id s35so2289242dak.6
        for <netdev@vger.kernel.org>; Sat, 09 Feb 2013 17:30:49 -0800 (PST)
In-Reply-To: <1360428312-1277-9-git-send-email-jiri@resnulli.us>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Sat, 2013-02-09 at 17:45 +0100, Jiri Pirko wrote:
> q->ptokens is in ns and we are assigning q->mtu directly to it. That is
> wrong. psched_l2t_ns() should be used to compute correct value.
> 


Not clear why its a separate patch, and not folded in the 6th


> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
> ---
>  net/sched/sch_tbf.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c
> index dc562a8..6e8b670 100644
> --- a/net/sched/sch_tbf.c
> +++ b/net/sched/sch_tbf.c
> @@ -165,8 +165,8 @@ static struct sk_buff *tbf_dequeue(struct Qdisc *sch)
>  
>  		if (q->peak_present) {
>  			ptoks = toks + q->ptokens;
> -			if (ptoks > (long)q->mtu)
> -				ptoks = q->mtu;
> +			if (ptoks > (s64) psched_l2t_ns(&q->peak, q->mtu))
> +				ptoks = (s64) psched_l2t_ns(&q->peak, q->mtu);

It seems a bit expensive to perform this in fast path.

Please add a variable to cache psched_l2t_ns(&q->peak, q->mtu)


>  			ptoks -= (s64) psched_l2t_ns(&q->peak, len);
>  		}
>  		toks += q->tokens;
> @@ -215,7 +215,8 @@ static void tbf_reset(struct Qdisc *sch)
>  	sch->q.qlen = 0;
>  	q->t_c = ktime_to_ns(ktime_get());
>  	q->tokens = q->buffer;
> -	q->ptokens = q->mtu;
> +	if (q->peak_present)
> +		q->ptokens = (s64) psched_l2t_ns(&q->peak, q->mtu);
>  	qdisc_watchdog_cancel(&q->watchdog);
>  }
>  
> @@ -296,11 +297,11 @@ static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
>  	q->max_size = max_size;
>  	q->buffer = PSCHED_TICKS2NS(qopt->buffer);
>  	q->tokens = q->buffer;
> -	q->ptokens = q->mtu;
>  
>  	psched_ratecfg_precompute(&q->rate, rtab->rate.rate);
>  	if (ptab) {
>  		psched_ratecfg_precompute(&q->peak, ptab->rate.rate);
> +		q->ptokens = (s64) psched_l2t_ns(&q->peak, q->mtu);

Here probably.