From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <9906031607.AA17181@clio.rice.edu>
Subject: Re: cwsdpmi r5
To: Martin DOT Stromberg AT lu DOT erisoft DOT se (Martin Stromberg)
Date: Thu, 3 Jun 1999 11:07:22 -0600 (CDT)
Cc: djgpp-workers AT delorie DOT com
In-Reply-To: <199906020704.JAA06580@mars.lu.erisoft.se> from "Martin Stromberg" at Jun 2, 99 09:04:52 am
X-Mailer: ELM [version 2.4 PL20]
Content-Type: text
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> Please make sure to put in the source code that has been done for DPMI 1.0 
> support (#if 0ed out, if deemed necessary). Thanks.

The current code has no DPMI 1.0 source which is not active.  I have been
sent some patches for small pieces of DPMI 1.0 support over the last 5 years
but I can't find them.  In any case, the features added were extensions
which had simple dpmi 0.9 workarounds.

The real DPMI 1.0 features which add value are:
 1) Enhanced exception stacks (getting page fault locations so user written
    page fault handlers are possible/useful), and
 2) Complete control of your own address space (where to put dpmi memory
    blocks) and separate page tables for each process, and
 3) Separate IDT for each process (allows 16-bit and 32-bit dpmi at same time).

All of these are huge efforts in complicated code, and unlikely to be added.
These would be a major re-write, and not just a patch.  Things which are 
easily patchable you can currently do today with DPMI 0.9 ... which is why
the features haven't been increased recently.