Subject: Re: ring overflows (was Question)
To: <>
From: Bill Studenmund <wrstuden@loki.stanford.edu>
List: port-mac68k
Date: 05/19/1997 11:42:12
Whoever first wrote this note needs to CORRECTLY configure their
EMail. root@myhost.my.domain wont' work!

> [pppd messages]
> 
> I changed /etc/syslog.conf (or whatever it's called) mostly to eliminate
> irritating *.notice messages from fwtk, and it appears to have ended
> the pppd messages as well.
> 
> 57600 port + 28.8 modem didn't upset my machine much at all; the ring
> overruns were few and far between.  It was only when I was ftping from
> the PBase _and_ ftping from the 'Net at the same time that I had a
> plethora of ring overruns.

Ack! Two 5 kB/s streams at once! Your poor kernel!

I think I should remind everyone that receiving at 57.6k can melt
your kernel. Certainly two 57.6k streams at once.

> I'd _love_ to compile my own kernel - but I lack the HD space at present.
> Thus, I'm using GENERICSBC-26 right now.  Given that more than one of
> us is having some problems with ring overruns, I'd hope that Scott (or
> whoever is compiling the GENERIC series) would consider bumping up the
> zstty buffer for future releases  :)

We might bump it to 4k (it's at 2k now), but that'll only help if the
problem's that the kernel gets delayed from dealing with the input.
If the problem's that the kenrel can't keep up, no.

At 57600 baud, it takes 354 ms to overflow a 2k ring buffer. Something
is massively keeping interrupts blocked for way to long (data come out
of the ring at softinterrupt level, which is at about spl 0.5.

I think upping the ring buffer's not the solution, We need to figure out
what's really wrong.

Take care,

Bill