Date: Thu, 25 May 1995 14:00:22 -0400
From: Dong Liu <dliu AT faraday-gw DOT njit DOT edu>
To: "Junaid A. Walker" <junaid AT dino DOT eng DOT monash DOT edu DOT AU>
Cc: Charles Sandmann <sandmann AT praline DOT no DOT NeoSoft DOT com>,
        DJGPP Mailing List <djgpp AT sun DOT soe DOT clarkson DOT edu>, mat AT ardi DOT com
Subject: Re: Compiling gdb, dpmi DS:VRAM hack.
References: <199505242011 DOT PAA22589 AT praline DOT no DOT NeoSoft DOT com> <Pine DOT 3 DOT 87 DOT 9505251226 DOT A7036-0100000 AT dino DOT eng DOT monash DOT edu DOT au>
Mailer: VM 5.32 (beta) for GNU Emacs 18.59.5

>>>>> On Thu, 25 May 1995 12:27:27 +1000 (EST), "Junaid A. Walker" <junaid AT dino DOT eng DOT monash DOT edu DOT AU> said:


Junaid> 	BTW the hack seems to work with OS/2, qdpmi101 (a wonder
Junaid> anything works under qdpmi), and most probably everything else.
Junaid> Would people please test the program under different DPMI
Junaid> providers? Eg 386Max, Win NT, OS/2 2.x, OS/2 Warp, Linux DOS
Junaid> emulator, etc.  	Presumably all dpmi providers add 1 to
Junaid> the segment limit, to get the segment size, and with 32-bit ints
Junaid> this invariably wraps around to 0 which is permissible
Junaid> (=>disable segment access?). So the hack is going to work for a
Junaid> long time to come, provided dpmi servers dont get smart and test
Junaid> for a segment limit of -1.

It won't work under Linux, DOSEMU of Linux don't allow a selector to be
set if (base+limit < base) || (base + limit >= 0xc0000000). I guess this
make DOSEMU the only smart DPMI host around :=).

---
Dong Liu  <dliu AT ace DOT njit DOT edu>
Electrical and Computer Engineering Department
New Jersey Institute of Technology, Newark, NJ 07102