For some reason my summary did not make it to the net, so I am sending
it again :
I posted a problem some time ago about an Ethernet card not working
when moved from a Sun-3/160 to a Sun-4/370 (my complete message is
appended at the end).
The problem came from wrong dip-switches. The interrupt vector for a
VME-Multibus Adapter with Sun-2 Ethernet is 0x75 for SunOS versions
3.0 and above, it is 0x1B for earlier versions.
I had to change the switch bank U12 on the Adapter board from
1 2 3 4 5 6 7 8
OFF OFF ON OFF OFF ON ON ON
to
1 2 3 4 5 6 7 8
OFF ON OFF ON OFF OFF OFF ON
It works on the 3/XXX as 0x1b is a generic i/o device vector, and the CPU
polls to see who is interrupting. On the 4/XXX machines, 0x1b is not defined
and the system hangs at interrupt level.
Below are a couple other suggestions that I received and that may be
of interest for some of you :
> From: Jeff Nieusma <nieusma@cs.Colorado.EDU>
>
> did you set the jumpers correctly on the backplane when you put the
> card in? You have to take the bottom two jumpers off to allow for the
> correct operation of bus-request and bus-grant on the VME bus.
>
> Since the ethernet card can be the bus master, it has to be able to
> interrupt the bus_grant line. If it does NOT have this capability,
> any cards above (farther away from the CPU) can cause collisions on
> the bus.
>
> The bottom two jumpers pass the bus_request and bus_grant signals. So,
> if they are still in and you have another card above the ethernet card,
> here is a bad situation:
>
> 1. the ethernet card requests the bus
> 2. the higher card requests the bus
> 3. the CPU asserts the bus_grant line
> 4. the ethernet card sees the bus_grant line and assumes it has the bus
> 5. the jumper passes the bus_grant signal
> 6. the higher card sees the bus_grant line and assumes it has the bus
> 7. both cards try to use the bus...
>
> If you have any cards above (farther away from the CPU) your ethernet card
> this could be your problem...
> From: "Manavendra K. Thakur" <thakur@zerkalo.harvard.edu>
>
> Assuming that your hardware and software are configured correctly,
> probably what's happening is that some machine on the ie1 network is
> pounding your 4/300 with a batzillion packets. You machine then pegs
> its cpu to the limit trying to cope with all these packets.
>
> This situation arises most likely due to client NFS requests. What
> you should do is monitor the traffic on the ethernet that is hooked up
> to bullwinkle's ie1 interface. You can do this by logging into
> another machine that is attached to the same ethernet and typing the
> command:
>
> etherfind -r -i ie1 -host bullwinkle-1
>
> This must be run as root, and you should replace "ie1" with whatever
> network interface that hooks up to bullwinkle's ie1 network.
>
> You should look for machines that are repeatedly sending packets to
> bullwinkle-1 - not just isolated bursts but a deluge of packets. Once
> you've identified the culprit client machine(s), halt them
> individually and then reboot the bullwinkle. This should allow
> bullwinkle to boot normally. Then you can boot the clients up again.
And my original message was :
----- Begin Included Message -----
Date: Fri, 8 Mar 91 10:54:05 CST
From: Fabrice Le Metayer <fabrice@sj.ate.slb.com>
Subject: ifconfig hangs on second Ethernet card
Here is the situation :
I have an Ethernet card inserted in a Sun-3/160. It is recognized as 'ie1'
and works fine. Now the problem starts when I want to move the card to a
Sun-4/370 (actually, a 3/260 upgraded to a 4/300-series CPU).
The device 'ie1' is seen correctly at boot time, but when the command :
ifconfig ie1 bullwinkle-1 -trailers up
is issued in /etc/rc.boot (or even at the console prompt, for that matter),
the machine hangs. In fact, I am not sure it really hangs, but the LED's
on the CPU start moving VERY slowly, may be indicating the CPU cycles
shooting up.
Oh, I know you would ask. The /etc/hosts file contains :
192.23.6.1 bullwinkle
129.87.112.2 bullwinkle-1
Can anybody help ?
----- End Included Message -----
Many thanks to :
Curt Freeland <curt@ecn.purdue.edu> [who gave the right answer]
David Robinson <david@elroy.Jpl.Nasa.Gov>
Jeff Nieusma <nieusma@cs.Colorado.EDU>
Leif Andersson <leif@control.lth.se>
Manavendra K. Thakur <thakur@zerkalo.harvard.edu>
Regards,
--
Fabrice
-- ,
Fabrice Le Metayer DOMAIN : fabrice@sj.ate.slb.com
Schlumberger Technologies - ATE UUCP : {amdahl,decwrl,uunet}!sjsca4!fabrice
San Jose, CA 95110 BELL : (408) 437-5114
"If you wonder whether your mission on Earth is finished,
If you are still alive, it isn't." -- Richard Bach
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:12 CDT