M8207 anyone?

From: ajp166 <ajp166_at_bellatlantic.net>
Date: Fri Aug 11 17:58:54 2000

From: Eric Smith <eric_at_brouhaha.com>


>> Suffice to say IDE is not a protocal like SCSI it is a DEVICE
>> interface and is somewhat uncooked at that.
>
>That's a fairly disingenuous claim. Both SCSI and IDE have a "device
>interface" (physical layer) and a "protocol" (interpretation of bytes
>in various registers and phases).


In one sense your right and in another your making it look complex
where it's not.

For scsi I have to deal with the SCSI interface chip (really nasty if
it's
5380!). For IDE I write/read to registers at addresses, al having
specific function. there is some sense of sequence and if you want
it to work you need to talk to the right one in the right order or
garbage
ensues. I liken it more to talking to an Z80 SIO raw than making the
SIO do HDLC and at the end of that link getting something
to happen. IDE is much simpler. I've done SCSI it's not!

>IDE in fact has two different protocols, ATA and ATAPI; ATAPI protocol
>is by design very similar to SCSI protocol.


Already known.

>As of SCSI 3, there are *many* different SCSI physical layers and
>several different protocols.


The difference between IDE and SCSI(any) is that from the CPU side if
I want to write the cylinder address to IDE I address the correct
register
and write (assuming not LBA). For SCSI I first have to talk to the scsi
chip and tell it what to say to the disk. The later already is a full
layer
of protocal from what you want to do.

The point originally being that IDE is the simplest disk interface there
is of the available sets and from doing a driver for z80 I can say its
easier than floppy as the interface if buffered, not time critical reads
or writes as the data flys by.

Allison
Received on Fri Aug 11 2000 - 17:58:54 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:32:44 BST