netdev
[Top] [All Lists]

Re: [Fwd: Re: possible bug x86 2.4.2 SMP in IP receive stack]

To: hadi@xxxxxxxxxx (jamal)
Subject: Re: [Fwd: Re: possible bug x86 2.4.2 SMP in IP receive stack]
From: kuznet@xxxxxxxxxxxxx
Date: Fri, 9 Mar 2001 21:01:28 +0300 (MSK)
Cc: ak@xxxxxx, feldy@xxxxxxxx, andrewm@xxxxxxxxxx, davem@xxxxxxxxxx, netdev@xxxxxxxxxxx, pp@xxxxxxxxxxxxxx
In-reply-to: <Pine.GSO.4.30.0103090903440.6867-100000@xxxxxxxxxxxxxxxx> from "jamal" at Mar 9, 1 09:13:16 am
Sender: owner-netdev@xxxxxxxxxxx
Hello!

> Bob, One of the things that could be done is have the driver stop sending
> packets after the toplayer gets congested (return is NET_RX_CN_HIGH).
> linux/include/linux/netdevice.h has the NET_RX_* which are returned
> by netif_rx()

Bulk jumbo transfers at giga rates is one of the cases, when the engine
should and can be tuned to _never_ slowdown or to drop in backlog.
We do not make any work at BHs, data is not accessed. It must fly.

Alexey

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