Re: tulip driver in 2.1.11* - 2.1.21 is broken - new driver

Alan Cox (alan@lxorguk.ukuu.org.uk)
Sun, 20 Sep 1998 04:47:40 +0100 (BST)


> To optimize for this behaviour it might be worth it to add a output flush
> check into the net_bh loop. The XMIT_EVERY define in 2.0 suggests that you
> once tried that, did you try it with routing and SMP again?
>
> Or do I miss something here?

I tried it but it doesnt help. The real problems are much more fundamental
but only really come into play with SMP and very fast networks. Networking
gets faster at a different speed to memory, and machine properties change
dramatically over time.

For 2.3.x we realistically have to be forwarding packets off interrupts,
doing intelligent interrupt reduction (something 1Gbit cards are actually
normally good at without much help) and making net_bh a per cpu handler.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/