Subject: NCR problems
To: None <port-i386@NetBSD.ORG>
From: Andreas Gustafsson <gson@araneus.fi>
List: port-i386
Date: 11/12/1995 21:29:30
I'm trying to hook up a WangDAT 3200 to a PC clone with a 486DX4-120,
a brand new "MG" PCI-bus motherbord built around a UMC chipset, a
generic NCR 53C810 -based SCSI card, and NetBSD-current of November 5,
1995.  It's not working.  Here is what gets logged at boot time:

Nov  9 18:32:24 guava /netbsd: pci0 (root): configuration mode 1
Nov  9 18:32:24 guava /netbsd: ncr0 at pci0 bus 0 device 13
Nov  9 18:32:24 guava /netbsd: pci_map_mem: mapping memory at virtual f9683f00, physical ffbdff00
Nov  9 18:32:25 guava /netbsd: pci_map_int: pin A mapped to line 10
Nov  9 18:32:25 guava /netbsd: ncr0: restart (scsi reset).
Nov  9 18:32:25 guava /netbsd: scsibus0 at ncr0
Nov  9 18:32:25 guava /netbsd: ncr0 targ 4 lun 0: <WangDAT, Model 3200, 02.0> SCSI2 1/sequential removable
Nov  9 18:32:25 guava /netbsd: st0 at scsibus0: st0(ncr0:4:0): 200ns (5 Mb/sec) offset 8.
Nov  9 18:32:25 guava /netbsd: st0(ncr0:4:0): COMMAND FAILED (4 88) @f86c8400.
Nov  9 18:32:25 guava /netbsd: ncr0: SCSI phase error fixup: CCB already dequeued (0xf86c8400)
Nov  9 18:32:25 guava /netbsd: drive empty
Nov  9 18:32:25 guava /netbsd: ncr0:4: ERROR (80:100) (e-ab-2) (88/13) @ (1214:0e000000).
Nov  9 18:32:25 guava /netbsd: 	script cmd = c0000001
Nov  9 18:32:25 guava /netbsd: 	reg:	 da 10 00 13 47 88 04 1f 00 0e 84 ab 80 00 03 00.
Nov  9 18:32:25 guava /netbsd: ncr0: handshake timeout
Nov  9 18:32:25 guava /netbsd: ncr0: restart (scsi reset).
Nov  9 18:32:25 guava /netbsd: probe(ncr0:5:0): COMMAND FAILED (6 ff) @f86d192c.

The SCSI bus is about 50 cm long and properly terminated.  I have already tried
disabling all caches and obscure-looking PCI options in the BIOS setup, as well
as different PCI slots, but I have not yet tried any other SCSI devices. Any hints?
-- 
Andreas Gustafsson, gson@araneus.fi