From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Sun,  6 Mar 1994 23:35:45 -0500
From: Jerry W Johnson jwjohn@ecst.csuchico.edu
Subject: Lance Error arrgghhh!
Topicbox-Message-UUID: 016bda38-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <19940307043545.UPlOeYdbGgu-QfIkMzWgcryLyGuSFOljRRPYiSNKXiI@z>

Hi 9'ers,
We've got two sparc2's that serve as our main fs and cpu/auth
server (with various ipx's as cpuservers and terms and pc's as terms).
I'm hoping there might be someone a little more hardware oriented who 
might offer some insight. With increasing regularity (say four times in
a day lately) the message "lance error #88c3" appears on the fs and
essentially brings the whole system down. The equipment all responds,
"connection timed out", and requires a reboot of everything!

We've tracked the error to lance.c in /sys/src/fs/ss and have debugged
it enough to say that the MERR bit is the culprit. According to "Ethernet/
IEE 802.3 Family; 1992 World Network Data Book/Handbook", pg 1-22;

MERR	MEMORY ERROR is set when the LANCE is the Bus Master and has not
       		 _____
	received READY within 25.6 us after asserting the address on the 
	DAL lines.  When a memory error is detected, the receiver and 
	transmitter are turned off and an interrupt is generated if NEA=1.
	MERR is READ/CLEAR ONLY, and is set by the LANCE and cleared by
	writing a "1" into the bit. Writing a "0" has not effect. It is
		   _____ 
	cleared by RESET or by setting the STOP bit.

Has anyone had a similar problem?

What might be possible solutions?

Is the problem more likely to be hardware or software?

Our best guess now would be to swap the fs and cpu server in the hopes
that if indeed it's a bad lance on the fs, at least we wouldn't continue
to lose the whole net. Comments? Suggestions?

	TIA. Cheers, --Jerry