Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site cornell.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!ihnp4!houxm!vax135!cornell!doug
From: doug@cornell.UUCP (Douglas Campbell)
Newsgroups: net.unix,net.micro,net.micro.pc
Subject: Xenix-286 access to real addresses
Message-ID: <1690@cornell.UUCP>
Date: Tue, 29-Jan-85 15:34:13 EST
Article-I.D.: cornell.1690
Posted: Tue Jan 29 15:34:13 1985
Date-Received: Wed, 30-Jan-85 07:23:29 EST
Distribution: net
Organization: Cornell Univ. CS Dept.
Lines: 18
Xref: watmath net.unix:3483 net.micro:9222 net.micro.pc:3229

Here's an interesting question for those who know something about
Xenix-286.  How difficult will it be to manipulate *real* addresses,
such as bit-mapped graphics memory?  Assuming that some sort of
user-definable device drivers can be written, it *should* at least
be possible.  But what will the performance and design costs be?

In short, what's needed is some way to drop into real address mode
and then pop back out again.  Without an easy way to do this, memory-
mapped devices will be highly difficult to work with.

I'll post a summary of responses.  (Is Microsoft listening?)

						Thanks,
						Doug Campbell
						doug@cornell.{UUCP|ARPA}
P.S.  Please don't flame the 286 architecture.
      The 286 is still smoking from an earlier
      discussion on net.micro.