> For starters, it seems that the latest snapshot (don't know about the ones
> before it) doesn't like certain drives.  For example, hooking up an
> old but functional RZ23-E in an external enclosure, and then trying
> to utilize that drive (extract a tar.gz file to it) would cause the
> kernel to panic with a 'bad dir' error message.  If this needs to
> be replicated so I can get the real error message, let me know. (But
> it did newfs just fine.. *sigh*)

You don't even say what kind of machine you're using, and hence what
SCSI driver you're using...

I assume you're using a TC machine.  Yes, I could very well believe
that the SCSI driver doesn't like certain drives, but i've never seen
it fail in _that_ way...

An error message, and a list of commands that you run to trigger it,
would be appreciated.


> Also, I took my trusty old config file, which doesn't have GENERIC defined,
> and does tell it where to look for the root device:
> 
> config netbsd root on sd2 swap on sd2
> 
> But it still asks me for the root device when it boots...
> Ideas?  What'd I'd overlook?

Sounds like you didn't 'make clean' before recompiling your kernel.

if you don't have GENERIC defined and aren't using 'swap generic', it
shouldn't ever ask you for the root device.

The root device selection code is ... very bogus, but that's the way
it works right now...


chris