Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.16 $; site haddock.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!haddock!dan
From: dan@haddock.UUCP
Newsgroups: net.micro.68k
Subject: Re: Orphaned Response
Message-ID: <13300001@haddock.UUCP>
Date: Sun, 26-May-85 20:17:00 EDT
Article-I.D.: haddock.13300001
Posted: Sun May 26 20:17:00 1985
Date-Received: Thu, 30-May-85 01:38:48 EDT
References: <415@oakhill.UUCP>
Lines: 19
Nf-ID: #R:oakhill:-41500:haddock:13300001:177600:1005
Nf-From: haddock!dan    May 26 20:17:00 1985


I would certainly approve if a future 68000 supported writes into PC
space.  Awhile back I was designing a dynamic linking facility for use on
several different architectures.  To load a new module into an existing
address space, I wanted to be able to put the text (PIC, of course) at any
point in the address space, and put the data a fixed distance after it (said
distance determined by the MMU, of course).  Then I wanted to address the
data using PC-relative addressing.  The deliberate prohibition of PC-relative
addressing on writes on the 68000 got in the way, forcing the use of
a dedicated register for the data.

Another thing that got in the way was the absence of a rather peculiar
addressing mode: PC relative indirect.  What I wanted to do on external
calls was call through a pointer located a known distance from the call
instruction.  I have a vague impression that the 68020 may have solved
this problem, but I haven't seen any recent 68020 information so I'm not sure.


	Dan Franklin