From:	"John M. Aldrich" <fighteer AT cs DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Polling busy video memory
Date:	Sat, 07 Feb 1998 14:31:14 -0500
Organization:	Two pounds of chaos and a pinch of salt.
Lines:	47
Message-ID:	<34DCB682.1D3F@cs.com>
References:	<6bibs8$l0i$1 AT herald DOT Mines DOT EDU>
NNTP-Posting-Host:	ppp209.cs.com
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Jean-Luc Romano wrote:
> 
>    Does there a function in DJGPP or Allegro that returns whether
> the video memory is busy or not?
> 
>    Something along the lines of:
> 
>    if ( is_video_memory_busy() )
>       printf("Still writing to video memory.\n");
>    else
>       printf("Video memory is available for use.\n");
> 
> 
>    I figure something like this would be useful in writing the
> screen buffer to the screen only when the screen is available, so
> the program doesn't have to halt until it can blit the video buffer
> again (if that made any sense).

AFAIK, unless you are talking about a multithreaded system wherein one
thread handles writing to video memory and another handles generating
the graphics, then your question is meaningless.  The basic task of
blitting is a loop that copies data from your program into the video
memory; it is part of your program like any other.

If you are discussing some sort of interrupt code, such as a timer or
mouse handler, which might be called at any time during your program,
then you must remember that interrupt handling code should be as small
and as fast as possible, to avoid disrupting the system for any length
of time.  Only very rarely should interrupt code draw anything to the
screen.  For example, if you are writing a mouse handler, the easiest
thing to do is have the handler update a counter in your program that
represents the mouse position, and have your graphics code read the
counter and draw the mouse pointer along with the rest of the screen.

With the above points in mind, it's trivial to create a global variable
in your code that holds a flag set by the blitting code when it starts,
and reset when it finishes.  There's no need, nor is there any
convenient way to detect such a situation in hardware.  To the CPU, one
memory address is like any other.

-- 
---------------------------------------------------------------------
|      John M. Aldrich       | "Always listen to experts.  They'll  |
|       aka Fighteer I       | tell you what can't be done, and why.|
|   mailto:fighteer AT cs DOT com   | Then do it."                         |
| http://www.cs.com/fighteer/|                - Lazarus Long        |
---------------------------------------------------------------------

- Raw text -