From:	Shawn Hargreaves <Shawn AT talula DOT demon DOT co DOT uk>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Can I use __dpmi_int() in real-mode-callback subroutine ?
Date:	Wed, 19 Feb 1997 19:39:42 +0000
Organization:	None
Distribution:	world
Message-ID:	<HLgylDA+b1CzEwV5@talula.demon.co.uk>
References:	<5edgrt$25d AT ccnews DOT nctu DOT edu DOT tw>
NNTP-Posting-Host:	talula.demon.co.uk
MIME-Version:	1.0
Lines:	29
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Jih-Shin Ho writes:
>For some reason, I want to use __dpmi_int() in real-mode-callback function.
>Can it work ?

I'm not 100% sure, but I doubt it. The only reason I can think of that
you would need to call __dpmi_int() is for accessing OS functions, and
given DOS's lack of reentrancy that is likely to cause trouble.

From personal experience, the only time I tried this was when I was
writing the mouse handler for Allegro, and it didn't work then. I wanted
to draw the pointer (which involves calling VESA interrupt functions),
inside the mouse driver callback, and that caused some spectacular
crashes :-) I later moved the pointer display code to my timer interrupt
handler, and just set a flag in the real mode callback which causes the
timer routine to redraw the pointer, and that works fine. So calling the
VESA functions in a timer interrupt seems to be ok, but not in a real
mode callback... Disclaimer: this knowledge is empirical in origin, and
could be completely wrong :-)

>This RMCB is called by a hardware interrupt service in real-mode.

Why? Wouldn't it be simpler to trap the hardware interrupt directly in
protected mode? If you install a protected mode handler, DPMI
automatically reflects real mode interrupts to it...

/*
 *  Shawn Hargreaves - shawn AT talula DOT demon DOT co DOT uk - http://www.talula.demon.co.uk/
 *  Ghoti: 'gh' as in 'enough', 'o' as in 'women', and 'ti' as in 'nation'.
 */

- Raw text -