Date: Tue, 9 Feb 1999 15:30:16 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Pierre Muller <muller AT cerbere DOT u-strasbg DOT fr>
cc: djgpp-workers AT delorie DOT com
Subject: Re: Windows 95 proposal !
In-Reply-To: <3.0.5.32.19990209132123.0091b930@ics.u-strasbg.fr>
Message-ID: <Pine.SUN.3.91.990209152729.11823F-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com


On Tue, 9 Feb 1999, Pierre Muller wrote:

>   Lots of bugs are not caught by Windows 95 because it does not allow zero
> page protection !
> 
>   Why don't we simply fill this page with NOP
> and a call to a fake exception at the very end of it ?

The problem is not with accessing the null page as code, it's with 
accessing it as data.  NULL-pointer derefencing usually happens with data 
pointers, not with code pointers.

>   Of course reading data in this region will still not be caught but
> reading a longint or a pointer would give 0x90909090
> that has big chances to be invalid !

There's nothing invalid with an int with such a pattern.  And most NULL 
pointers are pointers to data, not to other pointers.