Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!bionet!ames!amdahl!pyramid!leadsv!zech
From: zech@leadsv.UUCP (Bill Zech)
Newsgroups: comp.sys.ibm.pc
Subject: Re: 640K limit
Summary: You MUST have ROM at FFFF:0
Message-ID: <9484@leadsv.UUCP>
Date: 19 Jan 90 22:22:55 GMT
References: <4668.25aed7f2@uwovax.uwo.ca> <1468@blackbird.afit.af.mil> <10344@saturn.ucsc.edu>
Organization: LOCKHEED, Sunnyvale, CA
Lines: 27

In article <10344@saturn.ucsc.edu>, ted@helios.ucsc.edu (Ted Cantrall) writes:s
.> In article <1990Jan17.031934.3374@esegue.segue.boston.ma.us> johnl@esegue.segue.boston.ma.us (John R. Levine) writes:
.> >In article <4308@brazos.Rice.edu> solomon@screech.rice.edu (Richard L. Solomon) writes:
.> >>>Simple.  IBM/MicroSoft *should* have used soft pointers to the I/O memory
.> >>>areas.
.> >>	NO....they SHOULD have mapped the I/O in the I/O ADDRESS SPACE where
.> --
.> How would things have worked out if IBM had put this 384k block at
.> the bottom
.> of the memory (0-384). That would have left no constrictions on upward
.> expansion except the 8088. And that problem would have been remedied by
.> the 80286.		-ted-
.> 

When the RESET line is asserted to the 8088, CS is forced to FFFF and
IP is cleared to zero.  So you must have some kind of read/only
memory at FFFF:0.

Some computers (the old Motorola EXORmacs comes to mind) had some
logic that modified the memory map temporarily during reset, so 
that ROM would appear where needed during bootup, then get out
of the way.  

The PC was designed as a simple machine, and most everything on
it is the minimum needed to get the job done.

- Bill