From: luisllo AT alumni DOT uv DOT es (Luis)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Urgent help, please
Date: 27 Aug 2002 03:17:57 -0700
Organization: http://groups.google.com/
Lines: 40
Message-ID: <6124b64f.0208270217.5eb4a2e2@posting.google.com>
References: <6124b64f DOT 0208261541 DOT 4592f51c AT posting DOT google DOT com> <3d6b06cd DOT sandmann AT clio DOT rice DOT edu>
NNTP-Posting-Host: 80.25.146.7
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 1030443477 7896 127.0.0.1 (27 Aug 2002 10:17:57 GMT)
X-Complaints-To: groups-abuse AT google DOT com
NNTP-Posting-Date: 27 Aug 2002 10:17:57 GMT
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Charles Sandmann <sandmann AT clio DOT rice DOT edu> wrote in message news:<3d6b06cd DOT sandmann AT clio DOT rice DOT edu>...
> > I'm using DJGPP and dpmi libray in a machine with Win2000. I
> > previously had it running under win98 Second Edition. When running
> > under win98 everytime I called the dpmi function:
>  
> > __dpmi_meminfo mem;
> > __dpmi_allocate_memory(&mem)
>  
> > It always gave as a result same value for mem.handle and mem.address.
> > I wrote a multi-threaded executive with a different data segment for
> > executive structures, being confident about this being the normal
> > behaviour of the function.
> 
> This was a bad assumption - different DPMI providers behave 
> differently.  Windows 2000 and Windows XP behave like Windows NT, 
> not like Win 9x.  (And Win 9x is dead as far as MS is concerned).
> Even under Win9x if you do some heavy multi-tasking in different
> windows you can see different behavior.
> 
> > And now under Win2000 everytime I call the function I get different
> > values for mem.handle and mem.address, and this implies making a lot
> > of changes in my code. 
> 
> You need to make those changes and not make assumptions like this in
> the future, unless it's specifically stated to be that way in the
> DPMI specification.
> 
> Sorry, maybe there is an easy way to fix this in your code if you look
> hard enough.

Thank's for your help. Just another question, is there anyway to
obtain the handle of a memory block when you only have the selector
that points to the ldt table?. I mean:
I can obtain the segment base address by using
__dpmi_get_segment_base_address(selector, &base_addr)
But I need to obtain the handle of memory allocated for that segment,
in order to call __dpmi_free_memory(handle). And I don't have that
info available (I know it's given to you when you call
__dpmi_allocate_memory(&mem)).
Thank's a lot for you help