netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jarek Poplawski <jarkao2@gmail.com>
To: xiaosuo@gmail.com
Cc: Jamal Hadi Salim <hadi@cyberus.ca>, devik@cdi.cz, netdev@vger.kernel.org
Subject: Re: [PATCH] sch_htb.c consume the classes's tokens bellow the HTB_CAN_SEND level
Date: Tue, 3 Nov 2009 08:00:22 +0000	[thread overview]
Message-ID: <20091103080022.GA6718@ff.dom.local> (raw)
In-Reply-To: <4AEF9862.8040404@gmail.com>

On 03-11-2009 03:41, Changli Gao wrote:
> sch_htb.c consume the classes's tokens bellow the HTB_CAN_SEND level.
> 
> When a class enters HTB_MAY_BORROW state, it relies on its parents to
> sent packets. The parent class in HTB_CAN_SEND state only consumes
> itself and its parents's tokens, but ADD tokens to the classes under its
> level. It is totally wrong.

Current code is OK.

> It means that a class, which sends packets
> in ceil rate, can also enter HTB_CAN_SEND state now and then.

Yes, a class is entitled to send on it's own then with it's guaranteed
rate, without depending on borrowing.

Jarek P.

> 
> Signed-off-by: Changli Gao <xiaosuo@gmail.com>
> 
> ----
> net/sched/sch_htb.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c
> index 85acab9..2705702 100644
> --- a/net/sched/sch_htb.c
> +++ b/net/sched/sch_htb.c
> @@ -629,11 +629,10 @@ static void htb_charge_class(struct htb_sched *q, struct htb_class *cl,
>  		if (cl->level >= level) {
>  			if (cl->level == level)
>  				cl->xstats.lends++;
> -			htb_accnt_tokens(cl, bytes, diff);
>  		} else {
>  			cl->xstats.borrows++;
> -			cl->tokens += diff;	/* we moved t_c; update tokens */
>  		}
> +		htb_accnt_tokens(cl, bytes, diff);
>  		htb_accnt_ctokens(cl, bytes, diff);
>  		cl->t_c = q->now;
> 

  reply	other threads:[~2009-11-03  8:00 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-11-03  2:41 [PATCH] sch_htb.c consume the classes's tokens bellow the HTB_CAN_SEND level Changli Gao
2009-11-03  8:00 ` Jarek Poplawski [this message]
2009-11-03  9:47   ` Changli Gao
2009-11-03 10:05     ` Jarek Poplawski
2009-11-03 13:18       ` Changli Gao
2009-11-03 23:00         ` Jarek Poplawski
2009-11-04  1:53           ` Changli Gao
2009-11-04  8:28             ` Jarek Poplawski
2009-11-04  9:16               ` Changli Gao
2009-11-04 10:42                 ` Jarek Poplawski
2009-11-04 10:58                   ` Martin Devera
2009-11-04 11:30                     ` Changli Gao
2009-11-04 11:51                       ` Martin Devera
2009-11-04 11:58                       ` Martin Devera
2009-11-04 12:08                         ` Changli Gao
2009-11-04 12:15                           ` Martin Devera
2009-11-05  5:44                             ` Changli Gao
2009-11-04 11:21                   ` Changli Gao
2009-11-04 11:49                     ` Jarek Poplawski
2009-11-04 12:01                       ` Changli Gao
2009-11-04 12:11                         ` Jarek Poplawski

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20091103080022.GA6718@ff.dom.local \
    --to=jarkao2@gmail.com \
    --cc=devik@cdi.cz \
    --cc=hadi@cyberus.ca \
    --cc=netdev@vger.kernel.org \
    --cc=xiaosuo@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).