Subject: Re: A1200 + > 2 IDE Devices
To: Phil Shepherd <philshepherd@enterprise.net>
From: Ignatios Souvatzis <is@jocelyn.rhein.de>
List: port-amiga
Date: 06/12/1998 20:17:31
On Wed, Jun 03, 1998 at 09:03:41PM +0200, Phil Shepherd wrote:
> I have a Tower A1200 with a buffered 4 x Device IDE interface running
> IDEFIX software in Amiga OS & was just wondering, having just got
> NetBSD 1.3 running, whether it supports more than 2 devices on the IDE
> bus as I have a 2 x IDE drives at ID 0 & 1, and a CD Drive at ID 2 ?
> 
> I Appreciate that normally the IDE bus supports 2 devices, but
> wondered if anyone has managed to overcome this in conjunction with,
> for example, an IDEFIX-type device ?

I always wondered about this myself... so, having bought a book about
SCSI and IDE, I finally dug out some AMInet freeware circuit board doing
2->4 interfacing and analyzed it.

As far as I can tell, those interfaces are abusing the 2nd I/O port range,
[which contains an alternative copy of the status register which doesn't
reset the interupt condition on read, and a writable control register with a 
reset and a interupt disable bit] to address the 2nd IDE bus.
This way, they disable this 2nd copy of the status register and the control
register.

This ONLY WORKS if you CAN GUARANTEE that no driver will ever use that
2nd status port.

It might work if we would like to stay with the special amiga driver we have
now and change it to not use that 2nd address range.

However, lots of people are crying for ATAPI support (for CD-ROM drivers),
and the way to get this before the millenium changes is to switch to the
machine independent "wdc" driver, which has "wd" and "atapibus" subdevices.

Btw, both drivers currently use the control register to reset the IDE 
devices. I have no idea what IDEFIX, or other changed-for-4-device
AmigaDOS drivers do; can anybody tell this?

Regards,
	Ignatios Souvatzis