🚀 go-pugleaf

In <Cr6r0J.DDC@news.atitech.ca> phsieh@atitech.ca (Paul Hsieh) writes:

>Hello out there.  I realize that I'm about to be flamed for being such an
>advocate of WC and end up asking a question about it, however ...

>I have been trying to call my VESA driver from a WC 9.0 program.  The 
>DOS4GW version is 1.8 (which is pretty ancient, but it mostly works well
>for me).  The scenario is pretty simply.  I need to set AX=4F00, and
>ES:DI to my _far address (i.e., below the first meg) then call int 10
>and watch ES:DI fill up with lotsa goodly info.

>In real mode this is not a problem (I wrote a little .com file which
>functions perfectly).  To do this in WC, I first allocated some low
>memory using DPMI function 0x0400.  I got back a segment and selector
>that I verified were reasonable values.  Then to use it, I simply
>calculated the flat address equivalent (i.e., segment*0x10) in a
>variable I called "VIS".  Then, before calling int 10 I set ES to 
>(VIS & 0xF0000) >> 4 and DI to (VIS & 0xFFFF), AX=0x4F00.  Upon return
>from the int 10, I got the required 0x004F.  But the information in
>VIS was untouched.  I verified this with 3 different VESA drivers, so
>I am pretty sure that the VESA drivers are not at fault.

>Does anyone know exactly what's going wrong?  I might be mixing up the
>definition of selector and segment, but in most cases ES becomes 0,
>which I assume was translated to 0 either as a selector or a segment.
>Is version 1.8 of DOS4GW just too old or buggy?  Have I pulled a major
>braino here?  Any help on this problem would be greatly appreciated.

This has nothing to do with Linux, so please keep it off that group!

Rob
-- 
-------------------------------------------------------------------------
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
-------------------------------------------------------------------------