On Mon, 17 May 1999 00:20:16 +0200 (MET DST) 
 Joerg Czeranski <jc@joerch.org> wrote:

 > > After adding "options SCSI_NCR_MAX_WIDE=0" to the kernel config file,
 > > it detects and uses the DDRS without problems:

 > rz0 at scsi0 target 0 lun 0 (LID=0) (IBM     DDRS-34560D      DC1B) (Wide16) 

Okay...

So, you're saying that NetBSD negotiated Wide on the disk, then failed because
it was connected via a narrow cable.

...yet Digital UNIX _also_ negotiated Wide, yet can use it even though it's
connected via a narrow cable?

FWIW, I have the exact same disk on one of my Alphas at home:

sd1 at scsibus0 targ 1 lun 0: <IBM, DDRS-34560D, DC1B> SCSI2 0/direct fixed
sd1(ncr0:1:0): 10.0 MB/s (100 ns, offset 8)
sd1: 4357MB, 8387 cyl, 5 head, 212 sec, 512 bytes/sect x 8925000 sectors

...albeit connected to the internal SCSI on the machine (a 53c810; it's
currently in a Multia, and is going to be in an AlphaStation 200 by
next Saturday).

A couple of comments...

In order to make the disk work properly, I had to do some frobbing... First
of all, I needed to set the "single-ended mode" jumper on the disk; I'm
assuming you've already done that, since it talks to the host _at all_.

Second, I needed to get a wide->narrow adapter that did the right thing
with the unused signal lines; namely, pull them (and only them) to an
active terminator.  I think there's also a jumper that allows you to
disable Wide negotiation.

Perhaps if you do the latter, the disk will (perperly) refuse to negotiate
Wide transfers.  In my case, obviously, the driver never tried.

Now, it could be that Digital UNIX, when it detected a failed transfer
in Wide mode, suppressed the error message, and fell back on Narrow.
I don't know.  And Digital UNIX drivers are pretty awful reading :-)

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>