Subject: Re: fifo overruns
To: Boudreault Benoit <boudreab@ERE.UMontreal.CA>
From: Bill Studenmund <wrstuden@loki.stanford.edu>
List: port-mac68k
Date: 01/16/1996 14:15:29
> 
> I have been running various Kernels recently
> I am currently running 
> NetBSD 1.1A (GENERIC) #4: Mon Jan  8 16:33:09 EST 1996     briggs@
> It sends me fifo overruns with stunning speed, for a yay or a nay it
> sends those to me, i am wondering what could cause this because my link
> isn't fast enough to even overload a 2k buffer, verified netstat -m
> and also verified vmstat -m and there is still buffers left
> Am I looking in the wrong places to see why it does that?

What device are they on? If they're on one of the zs's (zsa or zsb),
then they're serial port overflows. They happen before the characters
get to either slip or ppp.

We've recently found the problem, and the current source on ftp.netbsd.org
reflects the correct code. I could release a kernel, but my kernel's
a bit customized. I'm not certain that I have all the scsi fixes, etc.
Also, we seem to have a few too many kernels lying around at the moment.

I'd prefer it if one of the more-traditional-maker-of-kernels makes a
fixed one. I'm happy to give suggestions on what's necessary to make it
all worke fine. Oh, I should thank Steve Allen for his distributions. :-)

Take care,

Bill