Message-ID: <32A373A3.6C6@ananke.amu.edu.pl>
Date: Tue, 03 Dec 1996 01:26:11 +0100
From: Mark Habersack <grendel AT ananke DOT amu DOT edu DOT pl>
Reply-To: grendel AT ananke DOT amu DOT edu DOT pl
Organization: Home, sweet home
MIME-Version: 1.0
To: sandmann AT clio DOT rice DOT edu
CC: djgpp AT delorie DOT com
Subject: Re: Problem raising exceptions in tight loops
References: <849468187 DOT 25048 DOT 0 AT abwillms DOT demon DOT co DOT uk> <32a21f73 DOT sandmann AT clio DOT rice DOT edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Charles Sandmann wrote:
 
> No, no, no.  In a hardware interrupt you have no information about the state
> of the DJGPP program or any way to change it.  So you don't know the EIP (if
> you did, you could use lots of tricks to transfer control) and making
> the app CS illegal will puke the DPMI.  You have to be very careful messing
> with the app DS also - which is why the DS alias exists.
I was thinking about using a TSS as a source of information about
CS:EIP. Wouldn't it be possible to store a pointer to the last tasks's
TSS in some place accessible for the HW handler, and then take all the
data you need from the TSS? In a non-multitasking environment like DOS
it should be quite possible.

-- 
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
>From the delequeue to the regiment, a profession in a flash. But
remember Monday signings when from door to door you dash... On the
news the nation mourns you - Unknown Soldier count the cost: for a
second you'll be famous but labeled posthumous... Forgotten Sons...
<A HREF="http://ananke.amu.edu.pl/~grendel">Home page</A>_____________