Sender: nate AT cartsys DOT com
Message-ID: <36E330B3.50D22056@cartsys.com>
Date: Sun, 07 Mar 1999 18:06:43 -0800
From: Nate Eldredge <nate AT cartsys DOT com>
X-Mailer: Mozilla 4.08 [en] (X11; I; Linux 2.2.1 i586)
MIME-Version: 1.0
To: djgpp AT delorie DOT com
Subject: Re: Getting cr2 in exception handler
References: <Pine DOT SUN DOT 3 DOT 91 DOT 990307135512 DOT 26544m-100000 AT is>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp AT delorie DOT com

Eli Zaretskii wrote:
> 
> On Sat, 6 Mar 1999, Nate Eldredge wrote:
> 
> > I have an application where I need to set up an exception handler for
> > page faults and have access to the value of cr2 (linear address to which
> > access faulted).
> 
> Why couldn't you look at the instruction that faulted and use some
> heuristics to find out the address, without accessing cr2?
> 
> For example, if the faulting instruction is "movl (%edx), %eax", then
> the offending address is in EDX.

I could, but I'd much rather not bloat my project by writing a
disassembler into it.  Especially with Intel's infinitely hairy
addressing modes :-\

Using CWSDPR0 is starting to look better.  The cr2 read is not critical,
just a very nice feature, and so I can enable it only when CWSDPR0 is in
use.  This will of course require that the user has enough memory, and
is okay with their buggy program (since my project is for debugging)
running in ring 0... :)

Still open to other suggestions of course.
-- 

Nate Eldredge
nate AT cartsys DOT com