From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <9610272040.AA12586@clio.rice.edu>
Subject: Re: app problem; don't lock null prot page
To: tony AT dictator DOT nt DOT tuwien DOT ac DOT at (Anton Helm)
Date: Sun, 27 Oct 1996 14:40:56 -0600 (CST)
Cc: djgpp AT delorie DOT com
In-Reply-To: <3.0b36.16.19961027211224.48ff26e4@dictator.nt.tuwien.ac.at> from "Anton Helm" at Oct 27, 96 10:12:15 pm
Content-Type: text

>   __dpmi_get_segment_base_address(_my_cs(), &address);
> 
>   info.address = (int) address + (int) __djgpp_ds_alias;

Almost certainly should be & __djgpp_ds_alias ....

>   info.size = 2;
>   errval = __dpmi_lock_linear_region(&info);
> 
> 
> Hmmm. That's the piece of code we are talking about. I have some ideas
> what all the other __dpmi_lock_linear_region() calls (not quoted here) 
> are supposed to do, but I don't understand this one. What are they trying 
> to lock here? Can I remove that code without damage?

This should be removed - djgpp_ds_alias is already locked anyway!  And
yes, this is buggy code, broken, and r3 catches it where r2 let it get by.