-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello everybody,

I recently got a scsi-host based on AMD's Am53C947-chipset. The=20
controller was built by DawiControl.
Since I also recently got an internal 640MB-MO-drive, I happily=20
installed the host, the MO and a CD-ROM.
I compiled a new kernel with the following lines included:
- ---
pcscp0          at pci? dev ? function ?        # AMD 53c947 PCscsi-PCI
scsibus0        at pcscp0

sd*             at scsibus0 target ? lun ?
cd*             at scsibus0 target ? lun ?
uk*             at scsibus0 target ? lun ?
- ---

I rebooted, but soon after the kernel detected the scsi-host I got this=20
message:
- ---
pcscp0: reset scsi bus
- ---

The message repeated endlessly, startup stopped.=20
Since my experience with scsi is pretty limited (equals zero), I am at a=20
loss here. What's causing this problem?

To make sure I hadn't configured the kernel incorrectly, I booted a=20
GENERIC-kernel, same problem. I had a NetBSD-Live-CD handy, same=20
problem.=20
To make sure it wasn't NetBSD's fault, I tried to boot a knoppix 3.1 and=20
a FreeBSD 5.0 Live-CD (CD 2 from the two-cd-set), both hung at=20
detecting the scsi-host, but didn't give me any messages.

Since three operating systems fail to boot with the scsi host installed,=20
I am pretty sure it's a hardware-problem, so maybe this is the wrong=20
place to ask...=20
But I found my problem mentioned in an ancient posting to port-sun3 or=20
something similar, where someone mentioned the exact problem I have,=20
but I did not find a solution there.
My guess is that a) the devices are fighting for a scsi-id or b) the=20
scsi-bus isn't terminated correctly.=20

A friend who's more familiar with scsi-hardware told me the controller=20
usually terminates its end of the bus. At the other hand there's the=20
cd-rom. It has a jumper at its back that says "Terminate", and it's=20
set. So, if I am not mistaken, the scsi bus *should* be terminated=20
correctly.=20
The ID of the cd-rom is set by jumpers as well, it's got three jumpers=20
functioning as a binary counter (1-2-4) which is set to 1 (1-0-0).
As for the MO, I am not sure. It does not have any jumpers on it, only=20
little switches on its belly. Unfortunately, I do not have a manual to=20
it, so I can only guess that these set the ID, too. Since it's in the=20
computer now, I cannot look up what the switches are set to, but I=20
think it was 2 (0-1-0-0).=20
So again, there *should* not be any problem.=20
Since there is, I hope someone had the same problem and found a=20
solution.=20

Kind regards,

Benjamin Walkenhorst

- --=20
Benjamin Walkenhorst
eMail: krylon@gmx.net
http://www.krylon.de
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQE/smA8/JWwsvZUqOwRAs5/AJwOTG+NqNXXKtX86SXG9tSivZMo9wCdEFqg
hDDk8kmrweK2mzp85vKb4HQ=3D
=3DbFl5
-----END PGP SIGNATURE-----