Date: Tue, 25 Jan 2000 09:48:02 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Erik Berglund <erik2 DOT berglund AT telia DOT com>
cc: djgpp AT delorie DOT com
Subject: Re: gcc page fault
In-Reply-To: <Z_0j4.1339$dT4.192511488@newsb.telia.net>
Message-ID: <Pine.SUN.3.91.1000125094648.23222I-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com
Errors-To: dj-admin AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


On Mon, 24 Jan 2000, Erik Berglund wrote:

> I was wondering if anyone could help me find
> a mysterious problem in gcc 2.95.2:

Isn't this from the same series of crashes on Windows 3.X that you
reported several months ago?  If not, how is this different?

> for (bp=freelist[bu]; bp; bpp=&(bp->next), bp=bp->next)   <---page fault
> 
> bp=freelist[bu] gives page fault because "bu"
> has a very high value (bu=edi=0x00391610).

I don't see what's wrong with this value: it's well within the limit
of your DS selector:

> ds: sel=00af  base=81043000  limit=fffeffff

So why do you think EDI is wrong?  It strikes me that the Windows DPMI
host might actuallly be the culprit here.  Or perhaps some GCC code 
expects all allocated memory to be contiguous.

> Now, I have a work-around for this problem and 
> that is to disable Virtual Memory either in Windows 
> or with _CRT0_FLAG_LOCK_MEMORY in gcc.
> Then everything works fine again.

How about the Unixy-sbrk flag you used earlier: does that solve the
problem as well?