Message-ID: <19990915225715.40078.qmail@hotmail.com>
X-Originating-IP: [134.134.248.22]
From: "Jay Warling" <bitheads AT hotmail DOT com>
To: eliz AT is DOT elta DOT co DOT il
Cc: djgpp AT delorie DOT com
Subject: Re: Problem accessing absolute physical address -1MB
Date: Wed, 15 Sep 1999 22:57:15 GMT
Mime-Version: 1.0
Content-Type: text/plain; format=flowed
Reply-To: djgpp AT delorie DOT com

Eli,
It turns out my main problem was that in the _farpokel and _farpeekl calls I 
was passing the linear address of the start of the descriptor (DUH!), rather 
than an offset into the segment referenced by the descriptor. There was also 
a problem with the descriptor buffer for the _dpmi_set_descriptor call.

Thanks for the help.
Jay


On Wed, 15 Sep 1999 Eli Zaretskii wrote:
>
>Please always make a point of posting the complete crash traceback and
>registers' dump when you report problems that cause programs to crash.
>The crash message includes some crucial info that makes our guesswork
>much easier.  Please run SYMIFY and post the crash message with the
>information added by SYMIFY.

My apologies for not showing the register dump for the SIGSEGV. I looked 
into running SYMIFY but I didn't see a core dump to use with it.

>
>What exactly is ``bizarre'' with this, and what special problems do
>you expect?  DJGPP gives you a flat 2GB address space, and the above
>numbers are well within it...

I didn't realize it could handle a 2GB address space. As you point out it 
shouldn't be a problem.

>
>Why did you need to copy the segment descriptor?  The code in the FAQ
>doesn't tell you should do so.

I had no real reason to do this. It was simply something I did during the 
debug process. As you point out it probably is wrong to do this for this 
example.


______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com