[3208] in Release_7.7_team
Progress on 3com bug!
daemon@ATHENA.MIT.EDU (Bill Cattey)
Thu Mar 28 17:17:45 2002
From: Bill Cattey <wdc@MIT.EDU>
To: release-team@mit.edu, ops@mit.edu, hotline@mit.edu, alex_c@mit.edu,
csnowden@mit.edu
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: 28 Mar 2002 17:17:43 -0500
Message-Id: <1017353863.31009.163.camel@m1-142-14.mit.edu>
Mime-Version: 1.0
My current theory is that etherboot trashed TWO things, one of which
we fixed with the vortex-diag hack, the other one of which gets fixed
when I run a different utility
mii-diag -R
I have sent detailed state logs to the author of the driver to get
an explanation of what bits are being diddled.
Meanwhile you can fix a dead machine as follows:
boot: linux single
insmod 3c59x options=2
^D
login as yourself
attach wdc
become root
cp /mit/wdc/3com/mii-diag /var/tmp
reboot
boot: linux single
insmod 3c59x
/etc/init.d/network start
mii-diag -r (this step may be unnecessary but is what I did)
mii-diag -R
mii-diag -v (this step may be unnecessary, but will confirm happiness)
^D
All will be well.
The first person to try this on a newly found dead system
should eliminate the -r and the -v.
NOTE: The negotiation heartbeat is 60 seconds and it may take
that long for the reset and autonegotiate to actually happen.
When I get back the explanation from Bogdan, I'll suggest a roll-out
strategy.
-wdc
