Newsgroups: comp.sys.amiga.programmer
Path: utzoo!utgpu!watserv1!watdragon!rose!ccplumb
From: ccplumb@rose.uwaterloo.ca (Colin Plumb)
Subject: Re: Low level Hard disk drivers
Message-ID: <1991Apr1.063939.17518@watdragon.waterloo.edu>
Sender: daemon@watdragon.waterloo.edu (Owner of Many System Processes)
Organization: University of Waterloo
References: <1991Mar27.175514.25590@cunixf.cc.columbia.edu> <00670283123@elgamy.RAIDERNET.COM> <mykes.0866@amiga0.SF-Bay.ORG>
Date: Mon, 1 Apr 1991 06:39:39 GMT
Lines: 26

mykes@amiga0.SF-Bay.ORG (Mike Schwartz) wrote:
>If EXEC is needed to work a hard disk, then how does AMAX do it?  Exec
>does not run at the same time as AMax.  You show less understanding than you
>claim I do.  Ever see DiskDoctor?  Do you think it uses AmigaDos packets to
>deal with corrupted sectors on disks?  Ever see SectorX or QB Tools?  These
>programs don't need nearly as much of the OS as you indicate.  These programs
>could just as easily use low-level hard disk access routines as any device.

Actually, DiskDoctor might... there are obsolete AmigaDOS raw sector
read/write packets it might use.  But a good program is more likely
to dig through the Dos device list for the exec device name and stuff
and use that.  And the trackdisk.device, at least, uses the exec.library
and the disk.resource.  Another device driver could use anything
it damn well felt like... like SLIP to run Sun's Network Disk protocol
if it liked.  If you are as nice as possible, you can run your fancy
disk editor on that network disk.  If you hit the hardware, well, it ain't
there.

If I wanted to do what Amax does, I'd somehow make a new exec.library
(maybe stealing from the ROMs, maybe not) that let a device driver
access various utility functions, install interrupt handlers, and
access the timer.device.  As much of necessary of the above would be
rewritten not to stomp on the Mac part of memory.  Then fire up the
device driver and lie to it about what OS it's running under.
-- 
	-Colin
