I corresponded with Tilmann Reh, the author of the GIDE, and from his
reaction, which gave no indication that he'd seen the 8-bit mode in his
spec, though I used the same version as he, suggests that while it might
exist in the spec, it may not have seem many real implementations.
The coding, unfortunately, is not similar enough to the 1003-wah to make the
code directly portable, but there are simply more registers in the IDE.
I hate to face the fact, but I'll probably end up latching the data and
build the words . . .
Dick
----- Original Message -----
From: allisonp <allisonp_at_world.std.com>
To: <classiccmp_at_classiccmp.org>
Sent: Monday, April 17, 2000 3:48 PM
Subject: Re: 8-bit IDE
> >I seem to remember that the IDE standard that I read said that the
> >command/status registers were all 8 bits (and mapped to the lower 8 data
> >lines). Only the data register was 16 bits, and the drive would assert
> >I/OCS16 when that was accessed and at no other time. Of course I have no
> >idea how modern drives handle this...
>
>
> This is still true for all the ones I've worked with (up to 528mb). Based
> on how the larger ones work in older systems they should be identical.
The
> emulation is supposed to be WD1003 controller and based on the work I've
> done with them
> it rings true. This is why every 8bit system example you see the data
path
> is folded somehow to 8bits.
>
> While reading the spec I discovered that Set_IO_8bit command and did a,
> Wow this will help. Never could make it do what is written in the spec.
I
> suspect the drive never read the spec. If it worked I'm sure the 8bit
> community
> like those that did the GIDE, COCO IDE and others would have jumped on
that.
>
> Allison
>
>
Received on Mon Apr 17 2000 - 18:24:12 BST