Xref:	news-dnh.mv.net comp.os.msdos.djgpp:4310
From:	Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: cwsdpmi beta9 - Disk protection!
Date:	Fri, 12 Jan 1996 20:17:07 CST
Organization:	Rice University, Houston, Texas
Lines:	23
Message-ID:	<30f71623.sandmann@clio.rice.edu>
References:	<DL3739 DOT Ct9 AT lut DOT ac DOT uk>
Reply-To:	sandmann AT clio DOT rice DOT edu
NNTP-Posting-Host:	clio.rice.edu
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

> In the latest version of the cwsdpmi program (beta9), what is the correct
> method of redirection the int24 interrupt vector. This is required so that
> the program prints up a user friendly message box if the user tries to 
> access the drive, when there is no disk within the drive. The method I was
> using, worked for the previous version, but now issues a page fault with
> error=0006. 

The previous version did not force you to lock the code called in a
real mode call back.  This left the potential of destroying the structure
of the hard disk in certain situations if the machine was tight on memory.

Using the wrappers should lock most of the memory needed, but you will
still need to lock the code, data (and register structure).  The error
message from CWSDPMI tells the EIP of the page fault (so you can track
down the executable line) and the CR2 value (typically drop the first one
and use a debugger to convert the address to a symbolic value to see
what wasn't locked).

This may be painful, but it's the only way to make sure code doesn't get
distributed with my name on it that destroys someone's hard disk.

You can either use the new crt0.h feature to lock all memory if you don't need 
virtual memory, or CWSDPR0 if ring 0 and no VM is OK.

- Raw text -