This problem has been outstanding for a while, but I didn't have the
information to ask the question properly before...

i386 - PII-200, 2940UW, Seagate 9Gig drives, HPIIc scanners, IDE boot disk
64MB Ram, 256 Swap

On an HP-IIIc scanner on a 1.3 system (I know I crossposted -current, getting
to that...) intermittently, the scanner will decide to only send 0s to the
system, at which point a reboot is required to fix the problem.

This system was originally installed with 1.0, and has had this trouble
from day 1. I brought the box up to 1.3, hoping that many of the changes
made to the scsi code might help with this. It hasn't. Has anything changed
since 1.3 that might have an effect?

I've built a kernel with SCSIDEBUG, and posted the clips from /var/log/messages
on a web page - http://user.fundy.net/david - The clips include two successfully
scanned pages, and two scans where the scanner simply sent (black) 0s to the
machine. In the bad scans, some of the SCSIDEBUG messages seem to be corrupted,
but I don't know if that's relevant or not. One of my next thoughts is to
disconnect the scanner from the scsi bus after it is in the hung state, and reboot,
while the 2940 does it's thing at bootup I'll leave it disconnected, and pause
at the kernel selection prompt, reconnect the scanner and see if the
initialization done by NetBSD at boot time is enough to clear the problem.

Anyone who might have a suggestion, please take a look at the page, I've
pointed out some specifics there, and the can retrieve the messages excerpts.
(32k mes1good, 26k mes2good, 12K mes3bad, 5K mes4bad)

I'd really like to clear this one up - it's been nagging me for a while. 
Thanks for the help in advance...

							David Maxwell
							david@fundy.ca