Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 06 Dec 1996 20:33:02 +0100
From:      Poul-Henning Kamp <phk@critter.tfs.com>
To:        Greg Lehey <grog@lemis.de>
Cc:        FreeBSD-current@freebsd.org (FreeBSD current users)
Subject:   Re: Problems with PLIP driver? 
Message-ID:  <9851.849900782@critter.tfs.com>
In-Reply-To: Your message of "Fri, 06 Dec 1996 18:32:46 %2B0100." <199612061732.SAA02168@freebie.lemis.de> 

next in thread | previous in thread | raw e-mail | index | archive | help

In message <199612061732.SAA02168@freebie.lemis.de>, Greg Lehey writes:
>Some of you will know that I've been having troubles connecting up the
>Ethernet board on my new notebook.  One of the consequences is that I
>have currently connected it to the net via PLIP.  On the up side, it
>worked like a dream: two ifconfigs (and a LapLink cable), and it was
>up.  However, there seem to be a few problems:
>
>1.  There seem to be an abnormal number of errors on the line.  In the
>    following, freebie is my FreeBSD machine, and papillon is the
>    laptop:

Well, if there is a significant speed difference between the machines,
you will have a problem.

>2.  There's definitely a bug in the bpf code, causing 100% of all
>    messages to be reported as invalid.  I think I can find that one
>    by myself.
please.  I've never got around to licking it definitively.

>3.  NFS across the link is very poor, and keeps hanging.  With TCP
>    transport, it's bearable.

Try either reducing your request size to 1024 or increasing the
MTU to 9180.  The former is probably the least unhealty (see below).

>4.  These problems seem to affect the Ethernet link (ep0) as well.
>    I've had to take it down and up again to get any kind of response,
>    and ping shows a strange cyclic behaviour.  Succesive pings seem
>    to complete in pairs, with a difference in time of about 1
>    second.  This only happens after errors on the PLIP interface:

The problem is that your speed is probably around 70Kbyte/sec, that
means that the cpus are running for 22 msec with irqs disabled.

If you increase the MTU, this will increase in a linear fashion of
course.  I usually use wsize=1024,rsize=1024 on NFS for that reason.

>I'm not asking anybody to jump in and fix these problems, but if this
>rings a bell with anybody, I'd be grateful for some pointers.

You know, it's really just a hack for emergencies and weird 
circumstances (which I will admit, is a permanent condition for 
some of us).  And as far as I know from you email, you got through
on it :-)

It works sufficiently well that it is usable, but it is impossible
to get it to really work well and preserve a decent transfer-rate. 


--
Poul-Henning Kamp           | phk@FreeBSD.ORG       FreeBSD Core-team.
http://www.freebsd.org/~phk | phk@login.dknet.dk    Private mailbox.
whois: [PHK]                | phk@tfs.com           TRW Financial Systems, Inc.
Power and ignorance is a disgusting cocktail.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9851.849900782>