On Mon, Jul 15, 2002 at 02:58:20AM +0200, Klaus Heinz wrote:
> Ignatios Souvatzis wrote:
> 
> > Get the docs and write support. But as far as I know, it is a polled-only 
> > programmed-i/o-only interface, so I guess using the Amiga 4000 IDE port
> > won't be much slower. And I am NOT sure that the known timing problems of
> 
> Does this mean, it is possible to write a NetBSD driver for a hardware
> device without an interrupt (== polled-only)? This sounds very much
> like the old harddisk interfaces from way back in 1987, e.g. the OMTI 5520
> I had on my A1000. You asked for a certain disk block, did busy-wait
> until the status bit indicated the arrival of the disk block and then
> had some unrolled loop to read the data from the data register.

It is possible, but will - so I guess - make your system feel very un-
multiprocessish. Think of it as the Amiga equivalent of so-called 
Win-"MoDem"s.

The situation will be a bit better if it can somehow generate an interupt
on SCSI phase transitions. But I repeat, I have no hardware specifications,
only the rumour that it is polled-only.

Regards,
	-is
-- 
seal your e-mail: http://www.gnupg.org/