Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!mips!apple!vsi1!daver!intersil!hamilton
From: hamilton@intersil.uucp (Fred Hamilton)
Newsgroups: comp.sys.amiga.tech
Subject: Reading absolute memory locations
Message-ID: <75.25cde84e@intersil.uucp>
Date: 5 Feb 90 20:55:42 GMT
Organization: Harris Semiconductor, Santa Clara CA
Lines: 42

Hello World.

I'm rolling my own real time clock for my A1000.  To read the registers, I
have to address specific memory locations.  I believe the correct way to
do this is along the lines of "value = *(long *)address;" (stop me now
if I've already blown it).  So I wrote a program to read some addresses
in the $00D80000-$00D80044 range (the same range the 500/2000's clock 
resides at), and I found I got this bizzare screen glitch when I read 
location $00D80028.  Here is the program (compiled with Lattice 5.04a
as "lc -L strange.c" ):

main()
{
long i, value;
long address = 0x00D80028;
/* for(i=0;i<1000000;i++) */
value = *(long *)address;
}

If you run the program once the screen you run it on will "bounce".  If you
run it several times you will notice that sometimes the screen ends up a
pixel higher than before you ran it, and sometimes it's a pixel lower.
If you take out the commented loop, the screen will go berserk for a few
seconds.

So, what is it?

	A)  I don't know how to program in C.  (typical)
	B)  My modified 1000 (LUCAS/FRANCES, CLtd HD controller) has
	    become rather weird.
	C)  The stock A1000 has something against people trying to read
	    location $D80028.
or	D)  Something else.

All I was trying to do was read the value at location $D80028 :-(

Any help will be appreciated.  Even if you can't help, please let me know
what the program does on your system if you try it.  Thanks.
-- 
Fred Hamilton                  Any views, comments, or ideas expressed here
Harris Semiconductor           are entirely my own.  Even good ones.
Santa Clara, CA