To: djgpp AT sun DOT soe DOT clarkson DOT edu
Subject: HD Errors
From: peters AT beltrix DOT guild DOT org (Peter Sleggs)
Date: 	Sun, 10 May 1992 16:24:17 -0400
Organization: Bellatrix Systems Corp., Mississauga, ONT Canada
Status: O

davidf AT algol DOT cs DOT umbc DOT edu (Mr. David W. Flater) writes:

> This is what I originally believed--but see below.
> 
> > Now this business of DOS reporting disk errors has only one explanation -- 
> > Somebody (either go32 or your DJGPP-compiled program) has been writing
> > through
> > a wild pointer into the DOS data areas.  That's the only time you see those
> > bogus hard disk error messages.  Believe me, I've seen that many times with
> 
> I would not claim for a minute that DJGPP does not corrupt memory.  Ctrl-
> Alt-Delete doesn't work on my machine after I run go32 on anything.  However,
> people have been saying that they get the same kind of crashes with
> lots of different software, in which the only common factor is heavy HD usage

There are conditions that show up under dos that cause problems like
"Drive not ready drive C:"  and a famous one from lotus agenda
"Drive door open on Drive C:".

The root cause of these problems is "supposed" to be a problem with maiy
systems BIOS code. I can make it happen on on of the systems I use here
IF
    1. I use QEMM for a memory manager and use the mapping of RAM
    2. I use a drive with 1:1 interleave.

The results are repeatable and consistent heavy disk usage and the
memory managers use of the memory management (which puts the processor
into _____ mode [I never remember which] ) causes a lost interrupt as I
understand the explanation I was given.

Could there be a similar interaction here ?

peter

--
peters AT beltrix DOT guild DOT org or torag!beltrix!peters