Re: 264 series questions

From: Tibor Biczo (crown_at_mail.matav.hu)
Date: 1999-10-04 20:35:40

Hi

> > One interesting thing, found by Levente (I think), is that memory
> > refreshes are like reads from $FF00-$FFFF.  When the refresh was at
> > $FF00-$FF3F, the program recorded the corresponding TED register values
> > (although it was reading from the same constant address in open address
> > space all the time).
>
> Ugh... Sure I did prove this refresh thing someway. But I don't remember
> how.
I'm sure you could have, but actually what Marko was referring to, happened
on
the Plus4 mailing list, and has been proved by me....
The method was a very simple one, as Marko already described in his mail.
Just have a read cycle reading from a not-connected address space right
after
the refresh-cycle, and you will get the value what was on the bus in the
refresh
cycle. Then just do a mapping loop which does this for all refresh-cycle
position,
and the received data is easily analysable. It also proved, that the refresh
cycle
counter counts upwards, unlike the C64 where it does in the opposite
direction.

It happened around when we were talking about writing a program which
runs in unconnected address space on the Plus4.


> The same goes for the Plus/4. Just remove the vertical border with the
> same trick you'd use on C64; if you switch the computer to single clock
> mode, you can set the pattern the same way as in C64. But when enable
> twiee clock, the borders will display the data fetched by the proc.
And there is also a way, to have both TED and CPU drive the address bus
at the same time, so both are active. (Actually they are active all the
time,
just when the cycle is not theirs they put $ffff on their address bus, to
let
the other device have it.)
This happened to me by accident, and actually haven't got the time, yet to
investigate, how it's reproducable....


Tibor Biczo



-
This message was sent through the cbm-hackers mailing list.
To unsubscribe: echo unsubscribe | mail cbm-hackers-request@dot.tcm.hut.fi.

Archive generated by hypermail 2.1.1.