[comp.sys.nsc.32k] SCSI questions

garyj@neptun.pcs.com (Gary Jennejohn) (04/26/91)

I assume that Randy is working on a driver for the AIC6250. I'm also in
the process of writing one. Since my attempts to mail Randy directly
always get bounced I'll use the mail group to discuss it.

I wrote my driver so that it uses the AUTOMATIC PIO (APIO) feature of the chip
for transferring commands and messages (actually, at the moment ALL
data transfers use the APIO, I haven't gotten around to using DMA yet).
I did this because I didn't want to have to handle SCSI phase changes in the
middle of transferring messages with DMA. APIO allows me to read/write a byte,
check if the phase changed, and then continue accordingly. It's also much
easier to implement APIO in a polled driver. Once I get interrupts going
I'll probably switch to DMA for all transfers.

Using APIO I have gotten the (polled) driver to the point where read/write
works without problems. The only difficulties I have had come immediately
after power on. The chip in my system always asserts ATN which causes my
disk drive to go bonkers. I see the weirdest phase changes on the bus
when I try to select the disk the first time. Once I've managed to do a
select everything works fine. The strange thing is that Jordan Hubbard
didn't see any of these weirdnesses when he tried out the driver. Must be
something peculiar with my controller/drive combination.

Randy, I'd be glad to exchange notes with you on this. As the Germans say,
"geteiltes Leid ist halbes Leid" (shared misery is half misery).

Gary