> > NetBSD 1.3_BETA (HWDIRECT) #360: Tue Dec 30 18:42:20 PST 1997
> > 
> > I get scsi bug reporting like this: (in fact, the bug doesn't harm,
> > it seems to be self repairing).  REMARK: it is not connected to the
> > fact I use a jaz.
> 
> [snip]
> > 
> > At 24, it seems to repair itself.  But, what's wrong?
> 
> Well, I'm not quite familiar enough with SCSI to truly tell you what's
> wrong, but you might consider compiling your own kernel with HWDIRECT ADB
> support and the sbc SCSI driver instead of the ncrscsi driver; it works
> better for some people on some drives.

I've reported this problem on my IIvx, as have other IIvx users. the
problem is indeed self-repairing, but very syslog-cluttering. the problem
seems to have been evident since GENERIC #47, but it've only used that and
GENERIC#56. i'd rather not use SBC if i don't have to. i don't think it
has anything to do with the drive, and whether it likes ncr or sbc, just
that the kernel isn't doing something right with ncrscsi on some machines 
- i hope this doesn't cause errors in future kernels to the extent that we
are not able to use ncrscsi.  I feel this way because ncr has always
worked on my machine, whereas sbc has had a reputation to hang on
boot (SBC#55 booted successfully, however). i wonder if this was a known
bug that's been fixed...

later

  - a