[Top] [All Lists]

Re: netif_rx packet dumping

To: Stephen Hemminger <shemminger@xxxxxxxx>
Subject: Re: netif_rx packet dumping
From: John Heffner <jheffner@xxxxxxx>
Date: Thu, 3 Mar 2005 18:48:50 -0500 (EST)
Cc: hadi@xxxxxxxxxx, "David S. Miller" <davem@xxxxxxxxxxxxx>, rhee@xxxxxxxxxxxx, Yee-Ting.Li@xxxxxxx, baruch@xxxxxxxxx, netdev@xxxxxxxxxxx
In-reply-to: <20050303151606.3587394f@xxxxxxxxxxxxxxxxx>
References: <20050303123811.4d934249@xxxxxxxxxxxxxxxxx> <20050303125556.6850cfe5.davem@xxxxxxxxxxxxx> <1109884688.1090.282.camel@xxxxxxxxxxxxxxxx> <20050303132143.7eef517c@xxxxxxxxxxxxxxxxx> <1109885065.1098.285.camel@xxxxxxxxxxxxxxxx> <20050303133237.5d64578f.davem@xxxxxxxxxxxxx> <20050303135416.0d6e7708@xxxxxxxxxxxxxxxxx> <Pine.LNX.4.58.0503031657300.22311@xxxxxxxxxxxxx> <1109888811.1092.352.camel@xxxxxxxxxxxxxxxx> <20050303151606.3587394f@xxxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
On Thu, 3 Mar 2005, Stephen Hemminger wrote:

> Another alternative would be some form of adaptive threshold,
> something like adaptive drop tail described in this paper.
> Since netif_rx is running at interrupt time, it has to be simple/quick.

All these AQM schemes are trying to solve a fundamentally different
problem.  With TCP at least, the only congestion experienced at this point
will be transient, so you do not want to send any congestion signals (drop
packets) if you can avoid it at all.  Making the limit as high as you can
tolerate seems like the best thing to me.

I am a bit worried that removing the throttling *is* a DOS risk though.


<Prev in Thread] Current Thread [Next in Thread>