Subject: 23/9 kernel unable to recognise drive?
To: None <port-mac68k@NetBSD.ORG>
From: David Johnston <DAVIDJ@info.wh.su.edu.au>
List: port-mac68k
Date: 09/27/1995 11:49:46
Grr... 
Ignore that last message, I pressed the wrong key and the stupid 
PC mailer sent it half finished.

The full message is:


Anybody seen something like this?

I installed a new kernel yesterday compiled from the 23/9 current tar 
files. When I booted with the new kernel I got the message (after the 
message complaining that floppy0 is not configured):

sd0: mode sense (4) returned nonsense; using fictitious geometry

This message does not appear with a kernel compiled from the 16/9 tar 
files.

My machine is a IIcx with three disks attached. sd0 is an old 100MB 
PLUS drive, sd1 is a 540MB Quantum and sd2 is a 730MB Quantum drive.

Unfortunately I didn't notice the above message and went on compiling 
and installing more of the 23/9 tar files. About 2 hours later, while 
the system was under reasonably heavy load, the machine froze with the
drive light of the 730MB drive hard on. After waiting for a while to see
if it would unfreeze (it didn't) I pressed the programmers switch and got
the following stack trace from the debugger:

db> trace
_Debugger(8922e, 8392e, 7fef40, 1754, 0) + 6
_nmihand(0, 2204, 49281, 94200, 15ff2e) + 26
_lev7intr(?)
_pdma_cleanup(ab5fc, ab5e0, 49281, 94200, 15ff2e) + 12
_via2_intr(7fefb4) + 3e
_lev2intr() + 14

The file system damage was interesting...

Any ideas? The PLUS drive which produced the boot-time warning 
message is mainly used for swap space (among other things), so I guess
the system fell apart while paging. My root file system is on the 
730MB drive.

        david

David Johnston
DAVIDJ@info.wh.su.edu.au