Re: Linux TCP Fixing everyones problems?

Peter Benie (pjb1008@cam.ac.uk)
Fri, 16 Apr 1999 11:42:44 +0100


alex.buell@tahallah.demon.co.uk writes ("Re: Linux TCP Fixing everyones problems?"):
>
> True. but perhaps we can embarrass these people into fixing *their* bugs
> by logging such buggy interactions via the system's logs. Something like
> printk("Barf! host a.b.c.d <insert your most hated operating system here>
> sending broken packets"). If enough people notice the complaints in the
> system logs, and scream... well... one can only live in hope. 8)

Unless you want a stream of 'I upgrade to linux version 2.2.whatever
and now I get these messages - what am I doing wrong?' queries, this
is a really bad idea. It also reports the problem where it is least
likely to be acted upon, and once the vendor had fixed their TCP
implementation, the fix would have to deployed on all broken machines.

This whole thread is a lot of fuss over nothing. The latency in
Linux's IP implementation is orders of magnitude larger than the time
wasted by the TCP workarounds (especially true for SMP machines), and
with many network services (eg. the WWW), you don't get to choose the
TCP implementation used on the server anyway - you have to
interoperate well with all of them.

Peter

-
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/