Date: Mon, 13 Aug 2001 21:28:53 +0300
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: "Andrew Cottrell" <acottrel AT ihug DOT com DOT au>
Message-Id: <5832-Mon13Aug2001212853+0300-eliz@is.elta.co.il>
X-Mailer: Emacs 20.6 (via feedmail 8.3.emacs20_6 I) and Blat ver 1.8.9
CC: djgpp-workers AT delorie DOT com, pavenis AT lanet DOT lv, sandmann AT clio DOT rice DOT edu
In-reply-to: <026201c12403$e0d609f0$0a02a8c0@acceleron> (acottrel@ihug.com.au)
Subject: Re: Selector Exhaustion
References: <10108100454 DOT AA13597 AT clio DOT rice DOT edu> <001901c12239$16e71dd0$0a02a8c0 AT acceleron> <2427-Mon13Aug2001123310+0300-eliz AT is DOT elta DOT co DOT il> <026201c12403$e0d609f0$0a02a8c0 AT acceleron>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> From: "Andrew Cottrell" <acottrel AT ihug DOT com DOT au>
> Date: Tue, 14 Aug 2001 00:26:00 +1000
> 
> > > I slightly modifed Charles patch so that the changes were only executed
> if
> > > on a Win NT / 2000 / XP machine, this way I can also use the same LIBC
> and
> > > code on the Win 98 box.
> >
> > ??? Why would the original code prevent you from using the same libc
> > on all systems?
> The original code leaked DPMI selector like a sive on Win2K when building
> LIBC on Win 2K, it was a night mare to have to restart every minute or two.

By ``original code'' I meant the patch posted by Charles.

> The issue I had was that Charles patch was for Win 2K and Andris patch was
> for Win 98. I wanted to have the same source and executables for both Win 98
> and 2K.  Between the patch from Charles that was for Win 2K issues and
> Andris and your comments about this last week I thought it best to only
> enable the code in Charles's patch on my Win 2K box and don't enable the
> function calls when running on my Win 98 box. I thought it was better safe
> than sorry.

If there's a conflict between different patches, I suggest to discuss
that and arrive at a unified solution.  Rushing into rebuilding
packages before we have an agreed solutioin that should be good for
all platforms might be a waste of resources.

> > Looks like the stack is smashed (EBP actually looks like ASCII text).
> > Did you try to stubedit gcc.exe to a larger stack?
> I need to read and try to understand the FAQ section 12.2. Is there any
> further pointers on what to look for in the registers or should I just keep
> on sending the crash info?

Section 12.2 is the starting point; after that, ask here.  If, after
reading the FAQ, you don't understand how did I arrive at the above
conclusion, I can explain.

> > Also, the EIP value seems right at the program start.  Can you see
> > where it is, exactly?
> I need to do some background reading in the FAQ and GDB etc on this so I can
> give the info next time the crash occurs or is it too late once the crash
> occurs?

The mapping between the EIP and the code doesn't change unless you
rebuild the program.  So, if you need to know where in the program is
a specific EIP value, run "gdb program", and then try either one of
the two commands "list *0xNNNN" and "disassemble 0xNNNN", where NNNN
is the address you are after (13f5, in this case).  Note that `list'
needs an asterisk before the address, while `disassemble' does not.