48 bit compiler for Intel 386?

dxdpsc@login.dknet.dk (Peter Skov Christensen)
Mon, 10 Oct 1994 10:34:01 GMT

          From comp.compilers

Related articles
48 bit compiler for Intel 386? dxdpsc@login.dknet.dk (1994-10-10)
| List of all articles for this month |

Newsgroups: comp.compilers
From: dxdpsc@login.dknet.dk (Peter Skov Christensen)
Summary: Looking for a compiler capable of handling 48 bit pointers
Keywords: 386, question, design, comment
Organization: DKnet
Date: Mon, 10 Oct 1994 10:34:01 GMT

Does anybody know of a compiler for the 386 processor capable of handling
48 bit pointers


By 48 bit pointers I mean the usage og the 16 bit segment plus the 32 bit
offset. Normal compilers and operating systems does only use the offset
and sets all segment to the same value for the running processes.


The reason I need this is the unique protection which the 386 MMU offers,
making it possible to protect memory down to 1 byte. If only the the 32 bit
offset is used as pointers, however this protection will be 4 Kb, which is
very much, because I want to protect more than 2000 processes, which then will
require at least 8 Mb of memory, where propably more that 50% will be waste.


I will be using this for an embedded system, which gives me a restriction
on memory available.


If anyone have knowlegde of such a compiler, or have experience with
complex systems with many processes and need for protection, please
contact me.


              Kjeld Flarup - DIAX Telecommunicatins, Denmark
              dxdpsc@dkuug.login.dk
[Aren't there versions of PL/M that does that? It is certainly the case that
one of the major reasons that programmers dislike Intel-style segementation
is that it doesn't match the storage model used in programming languages that
people use. It's barely tolerable in Pascal, and hopeless in C or Fortran.
-John]
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.