>Number:         9216
>Category:       kern
>Synopsis:       IBM U3 LVD disks terribly hosed on Tekram 390F/U2W controllers
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jan 17 15:18:01 2000
>Last-Modified:
>Originator:     Tom Spindler
>Organization:
Redback Networks
	
>Release:        <NetBSD-current source date>1.4.1 and 1.4.2ALPHA (as of 15 Jan 2000)
>Environment:
	
System: NetBSD fartmangler.redback.com 1.4.1 NetBSD 1.4.1 (EQUITY) #5: Thu Jan 13 20:22:34 PST 2000 dogcow@fartmangler.redback.com:/amd/elvis/build/NetBSD-1.4.1/src/sys/arch/i386/compile/EQUITY i386

>Description:
With Tekram 390F and 390U2W controllers (53c875/53c895 chips, respectively)
it appears that the IBM LVD DMVS18V drives are not probed and/or set up
properly. Note that the 390F controller does not even support LVD, and
that the drives ought to fall back to U2 SE operation.

This fails on both 1.4.1 and 1.4.2 current as of 15 Jan 2000.

	
>How-To-Repeat:
from dmesg:
ncr1 at pci0 dev 10 function 0: ncr 53c895 fast40 wide scsi
ncr1: interrupting at irq 11
ncr1: minsync=10, maxsync=137, maxoffs=31, 128 dwords burst, large dma fifo
ncr1: single-ended, open drain IRQ driver, using on-chip SRAM
ncr1: restart (scsi reset).
scsibus1 at ncr1: 16 targets, 8 luns per target
probe(ncr1:2:0): WIDE SCSI (16 bit) enabled
probe(ncr1:2:0): 10.0 MB/s (200 ns, offset 31)
sd2 at scsibus1 targ 2 lun 0: <IBM, DMVS18V, 0100> SCSI3 0/direct fixed
sd2(ncr1:2:0): WIDE SCSI (16 bit) enabled
sd2(ncr1:2:0): 40.0 MB/s (50 ns, offset 31)
sd2: 17501MB, 11736 cyl, 10 head, 305 sec, 512 bytes/sect x 35843670 sectors
ncr1:2: ERROR (0:4) (8-0-0) (1f/bf) @ (scripth 6d0:878b0000).
ncr1: script cmd = 80080000
ncr1: regdump: ca 00 80 bf 47 1f 02 03 02 08 82 00 80 00 0f 02.
ncr1: have to clear fifos.
ncr1:2: ERROR (0:1) (8-0-0) (1f/bf) @ (script 3f0:878b0000).
ncr1: script cmd = 72100000
ncr1: regdump: ca 00 80 bf 47 1f 02 03 02 08 82 00 80 00 0f 02.
ncr1: have to clear fifos.
ncr1: restart (fatal error).
probe(ncr1:2:1): COMMAND FAILED (9 2) @0xf0a77400.
probe(ncr1:2:1): WIDE SCSI (16 bit) enabled
ncr1:2: ERROR (0:4) (8-0-0) (0/7) @ (scripth 6d0:878b0000).
ncr1: script cmd = 80080000
ncr1: regdump: ca 00 80 07 47 00 02 03 02 08 00 00 80 00 0f 02.
ncr1: have to clear fifos.
ncr1:2: ERROR (0:1) (8-0-0) (0/7) @ (script 3f0:878b0000).
ncr1: script cmd = 72100000
ncr1: regdump: ca 00 80 07 47 00 02 03 02 08 00 00 80 00 0f 02.
ncr1: have to clear fifos.
ncr1: restart (fatal error).
probe(ncr1:2:1): COMMAND FAILED (9 2) @0xf0a77400.
probe(ncr1:2:1): WIDE SCSI (16 bit) enabled

and repeat the same two errors many many many other times.


	
>Fix:
	
>Audit-Trail:
>Unformatted: