[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20071128.015704.113039455.takano@axe-inc.co.jp>
Date: Wed, 28 Nov 2007 01:57:04 +0900 (JST)
From: Ryousei Takano <takano-ryousei@...t.go.jp>
To: kaber@...sh.net
Cc: netdev@...r.kernel.org, shemminger@...ux-foundation.org,
dada1@...mosbay.com, t.kudoh@...t.go.jp
Subject: Re: [PATCHv2 1/3] NET_SCHED: PSPacer qdisc module
> One more thing: your qdisc can only be used as root qdisc since it
> produces packets itself and thereby violates the rule that a qdisc
> can only hand out packets that were previously enqueued to it.
> Using it as a leaf qdisc can make the upper qdiscs qlen counter
> go negative, causing infinite dequeue-loops, so you should make
> sure that its only possibly to use as root qdisc by checking the
> parent. It would also be better to do something like netem
> (enqueue produced packets at the root) to make sure the qlen
> counter is always accurate.
>
I agree with you.
PSPacer should not use with other rate regulation qdiscs. But,
I think that a combination of netem and PSPacer is a useful for
emulating networks. The following paper describes experimental
results using PSPacer with netem:
"Large Scale Gigabit Emulated Testbed for Grid Transport
Evaluation", PFLDnet 2006.
http://www.hpcc.jp/pfldnet2006/paper/s1_02.pdf
Best regards,
Ryousei Takano
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists