Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!bu.edu!polygen!jerry
From: jerry@polygen.uucp (Jerry Shekhel)
Newsgroups: comp.sys.amiga.advocacy
Subject: Re: The Amiga's Future
Message-ID: <1134@stewart.UUCP>
Date: 15 Jun 91 17:16:51 GMT
References: <#g1H3+$o@cs.psu.edu> <12901@uwm.edu> <22308@cbmvax.commodore.com> <1991Jun11.134256.9465@cs.mcgill.ca>
Reply-To: jerry@stewart.UUCP (Jerry Shekhel)
Organization: Polygen Corporation, Waltham, MA
Lines: 28

genius@cs.mcgill.ca (Michel NGUYEN) writes:
>
>Theorically, the 386 and 486 can have more than 16000 segments of
>4GB each segment (48 bits). This feature is ideal for multiuser
>systems ---> memory protection. So far, no OS (that I know of) has
>taken that advantage. Not even Unix, maybe OS/2 4.0 ?? :-)
>

386-based UNIX has complete memory protection (as does OS/2 and even
286-based UNIX).  The processor maintains both a segment descriptor table and
a page table for each running task.  So no process can step on your memory,
be it another user's process or your own.

It is true that it is theoretically possible to develop a 386-based OS that
runs in the 386 equivalent of the 286's "large memory model".  This OS would
manage multiple 4GB code and data segments for each process.  My question is,
who needs this?  Also, you'd have to start worrying about near and far pointers
again!  A near pointer would be a 32-bit offset, and a far pointer would be a
48-bit segment/offset pair.  Virtual memory space would be 64 terabytes.  Wow.
--
+-------------------+----------------------+---------------------------------+
| JERRY J. SHEKHEL  | POLYGEN CORPORATION  | When I was young, I had to walk |
| Drummers do it... | Waltham, MA USA      | to school and back every day -- |
|    ... In rhythm! | (617) 890-2175       | 20 miles, uphill both ways.     |
+-------------------+----------------------+---------------------------------+
|           ...! [ princeton mit-eddie bu sunne ] !polygen!jerry             |
|                            jerry@polygen.com                               |
+----------------------------------------------------------------------------+