I have been playing around with a couple of 1542B controllers I picked
up.  The symptoms are pretty strange.

I had a Seagate hard drive that I was using, and gave it up.  I still
have a couple of SCSI devices.  Here is the appropriate part of the
boot-up dmesg:

#
#  Actually, it is a TI 486DLC.  
#
CPU: Cx486DLC (Cyrix 486-class CPU)
WARNING: CYRIX 486DLC CACHE UNCHANGED.
real mem  = 16384000
avail mem = 13557760
using 225 buffers containing 921600 bytes of memory
...
aha0 at isa0 port 0x330-0x333 irq 12 drq 6
scsibus0 at aha0
#
# Target 3 is a Sony CDR-55JD Double Spin CD-ROM drive
#
aha0 targ 3 lun 0: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 1: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 2: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 3: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 4: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 5: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 6: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 3 lun 7: <\\005\\005, , > SCSI5 5/cdrom fixed
aha0 targ 4 lun 0: <WANGTEK, KS23465, F> SCSI1 1/sequential removable
st0 at scsibus0: drive empty
#
#  The tape drive info is right.
#
aha0 targ 5 lun 0: <TOSHIBA, CD-ROM DRIVE:XM, 31CS> SCSI1 5/cdrom removable
cd0 at scsibus0: drive empty
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
aha0: host_stat 13
#
#  The Toshiba Single Spin works OK, except as noted below.
#
mcd0 at isa0 port 0x320-0x323 irq 10: model LU005S
#
#  On to the Mitsumi, etc.
#

...

There was more, but that was all that seemed to be important.

The problem is that whenever I do anything in the SCSI devices, the least 
bit intensive, the machine wedges tight.  I can mount a CD or mount a hard
drive, and maybe even get a couple of 'ls' commands to work.  I can write a
compressed tar tape, and it will fo as long as I want to let it run; as
soon as I try and read it, the system will hang.

These symptoms never change, regardless of which drives are in or out of
the list.  I have tried every possible combination and it doesn't seem
to make any difference.  I have tried to make sure the chain it
terminated correctly (see request below for caveats on that:-)

On a possibly related note, does anyone know how to terminate (or
unterminate) a CDR-55JD or the Toshiba drive list above.  I haven't got
docs for these drives and I would like to figure out if I am over
terminating (or underterminating) the SCSI line.

While the problem is producable, it is not predictable.  It comes and
goes seemingly at will.  It never works for very long, but it seems to
work adequately from time to time, if only for a minute.

I've talked to Julian Elisher about it in general terms, and he
suggested I recompile the kernel to initialize the DMA transfers at a
slower (them 10 MBS) speed.  Instead of trying to guess a good speed, or
pick one that is TOO slow, I am going to try "options TUNE_1542" as a
temporary measure.  I understand some of the risks, but it is worth a
shot.

Now for the really strange part - This equipment all works without 
complaint on a 1522 controller.

I have the docs for the 1542, and they seem to do exactly the same
thing every other 1542 does.  I have two controllers and two machines.
I have tried this with other 1542s and it consistently seems to fail.

I have tried seveeral different IRQ/DRQ combinations and all of them
produce the same results.
-- 
Dave Burgess  (The man of a thousand E-Mail addresses)
386bsd FAQ Maintainer / SysAdmin for the NetBSD system in my spare bedroom
"Just because something is stupid doesn't mean that there isn't someone
that wants to do it...."