On Thu, Feb 13, 2003 at 09:23:40AM -0700, Justin T. Gibbs wrote:
> Can you be more specific about the nature of the failure?  Are there any
> driver diagnostics produced during the failure?

Ok, I booted a compiled kernel with the ported driver I had lying around,
and easily reproduced the problem. I noticed that you had made some
changes in the FreeBSD tree, with one entry saying "perhaps solves
spurious parity errors people are seeing on U160 controllers". So,
I incorporated those changes, and tried again. Unfortunately it still
fails on this particular system. It works fine with a 29160 in
another system (with only one disk attached).

The command I use is

dd if=/dev/zero of=out bs=1024k count=256

..on a filesystem on disk sd1.

Below are some details. They're not as verbose as they could be,
but I ifdefed out some of the print_registers code because it
used the return value of printf. The kernel printf is void
on NetBSD. It should probably be changed, but it required more
changes elsewhere in the kernel than I wanted to make at the
time.

The system is a dual Athlon 1.2Ghz, Tyan board, 1G of memory,
on-board 7899.

===========================================================================

ahc0 at pci0 dev 13 function 0
ahc0: interrupting at apic 2 int 5 (irq 5)
ahc0: aic7899: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
scsibus0 at ahc0: 16 targets, 8 luns per target

ahc1 at pci0 dev 13 function 1
ahc1: interrupting at apic 2 int 10 (irq 10)
ahc1: aic7899: Ultra160 Wide Channel B, SCSI Id=0, 32/253 SCBs
scsibus1 at ahc1: 16 targets, 8 luns per target

scsibus0: waiting 2 seconds for devices to settle...
sd0 at scsibus0 target 0 lun 0: <IBM, DDYS-T36950N, S93E> disk fixed
sd0: 35003 MB, 15110 cyl, 12 head, 395 sec, 512 bytes/sect x 71687340 sectors
sd0: sync (12.5ns offset 63), 16-bit (160.000MB/s) transfers, tagged queueing
sd1 at scsibus0 target 6 lun 0: <QUANTUM, ATLAS IV 9 WLS, 0A0A> disk fixed
sd1: 8761 MB, 13816 cyl, 4 head, 324 sec, 512 bytes/sect x 17942584 sectors
sd1: sync (12.5ns offset 31), 16-bit (160.000MB/s) transfers, tagged queueing

scsibus1: waiting 2 seconds for devices to settle...
st0 at scsibus1 target 3 lun 0: <HP, C1537A, L708> tape removable
st0: drive empty
st0: sync (100.0ns offset 32), 8-bit (10.000MB/s) transfers
cd1 at scsibus1 target 4 lun 0: <PLEXTOR, CD-R   PX-W124TS, 1.07> cdrom removabe
cd1: sync (50.0ns offset 8), 8-bit (20.000MB/s) transfers
ss0 at scsibus1 target 5 lun 0: <UMAX, Astra 1220S, V1.2> scanner fixed
ss0: async, 8-bit transfers

ahc0: brkadrint, Scratch or SCB Memory Parity Error at seqaddr = 0x60
>>>>>>>>>>>>>>>>>> Dump Card State Begins <<<<<<<<<<<<<<<<<
ahc0: Dumping Card State in Data-out phase, at SEQADDR 0x60
Card was paused
ACCUM = 0x0, SINDEX = 0x68, DINDEX = 0xe4, ARG_2 = 0x3
HCNT = 0x0 SCBPTR = 0xc
STACK: 0x89 0x0 0x16b 0x63
SCB count = 32
Kernel NEXTQSCB = 13
Card NEXTQSCB = 12
QINFIFO entries: 12 4 
Waiting Queue entries: 
Disconnected Queue entries: 2:11 
QOUTFIFO entries: 
Sequencer Free SCB List: 9 1 14 8 4 11 3 0 15 7 6 5 10 13 16 17 18 19 20 21 22  
Kernel Free SCB list: 14 5 1 2 8 31 6 7 0 10 3 9 30 29 28 27 26 25 24 23 22 21  

<<<<<<<<<<<<<<<<< Dump Card State Ends >>>>>>>>>>>>>>>>>>

-- 
Frank van der Linden                                    fvdl@wasabisystems.com
==============================================================================
Quality NetBSD Development, Support & Service.   http://www.wasabisystems.com/