Sender: rich AT phekda DOT freeserve DOT co DOT uk
Message-ID: <3D396078.11E91027@phekda.freeserve.co.uk>
Date: Sat, 20 Jul 2002 14:07:04 +0100
From: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.2.19 i586)
X-Accept-Language: de,fr
MIME-Version: 1.0
To: djgpp-workers AT delorie DOT com
Subject: Re: unixy sbrk and win2k
References: <10206100232 DOT AA17839 AT clio DOT rice DOT edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com

Hello.

Charles Sandmann wrote:
> The following patch to crt0.S seems to have fixed the problem:
> 
> *** crt0.BAK    Sun Jun  9 20:14:46 2002
> --- crt0.S      Sun Jun  9 20:21:48 2002
> *************** brk_common:
> *** 475,478 ****
> --- 475,479 ----
>         movw    $0x0900, %ax                                    /* disable interrupts */
>         int     $0x31
> +       cli
>         pushl   %eax
> 
> 
> (Line numbers are vs 2.03, but it's just a single line insertion).  This
> should be a no op - since the previous line disables and stores the
> previous state in AX; we then restore (enable) them with an equivalent call.
> 
> Under Win2K the get interrupt state DPMI call shows the program starts
> with interrupts disabled (but they aren't) - the cli above seems to
> fix win2K's bad behavior when moving the arena but does not break
> win95 for example.
> 
> Any comments?  We ought to try and build something substantial (like
> emacs) with this fix above to crt0 and see if there is joy in W2K land
> again.
> 
> I can't see how this above could break anything.

Unfortunately Emacs seems to bomb out (itself + bash session) under Windows NT
4, when executing this 'cli' instruction. I'm using month-old(ish) Emacs CVS
built against month-old(ish) DJGPP CVS with the above patch. I'm running
Windows NT 4 + SP5 under VMware 3.1.

Bye, Rich =]

-- 
Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]