I've been using the PINT scsi scanner driver with an HP3C scanner, and I'm
having a variety of problems, which generally end up in a crash.

I need some suggesstions about how to debug this - or even how to get 
enough information so someone could suggest a source of action.

The HP3C has a sheet feeder on it - if the sheet feeder jams, the scanner
device (/dev/scan0) can't be read until I cycle the power on the scanner.
It would be nice if there was a way to handle this in software. There is
a reset_scanner script provided with PINT, but it is only for AIX, and
essentially does a remove and re-add on the scanner device.

Worse yet, sometimes even after power cycling the scanner the driver
doesn't behave correctly anymore. Sometimes the driver feeds back blank
pages without the scanner ever doing a pass, and sometimes I get device
already open errors, but I can't find any process alive that  would have the
scanner open.

Finally, sometimes the system ends up dropping into the kernel debugger. 
Usually, syslog lists a number of /dev/ss0 device already open errors 
just before the crash.

This is on a NetBSD-1.1 release system, 486-100 with a 1542 controller.

Any ideas about what to try at the DB> prompt (or elsewhere ) would be
greatly appreciated I need to make this system more reliable as I'm
getting dumb suggestion like 'use Linux instead' when I don't think NetBSD
is at fault here, but I'd like to be able to back up my opinion ;-)

						David Maxwell
						david@mobtech.com