| |
6R6
Registered: Feb 2002
Posts: 245 |
Action Replay & Retro Replay Bugs
Hi again,
I'll begin with the known ones first.
Please fill in if you know anything else.
AR (all versions ?):
--------------------
The infaomus $dexx read bug, only happens on real hardware.
Doesnt happen in emulator afaik.
All sort of read attempts from registers between $de00-$df00
will make your computer crash.
AR 4.2 + AR 6.0 plus (not to be mistaken by the normal AR 6.0)
--------------------------------------------------
Certain games using NMI interrupts will crash because of this cartridge. If you have such a cart then check the game 720 usa (Don't check the nostalgia version, because they fixed it.).
RR:
---
The RR+RRnet doesnt work on my C128D - Dont know why.
AR 6.0 + RR:
------------
When fixing the problems I had with N0SD0S(Drive Code C128D N0SD0S Scanner bug.)
I came across another interesting bug with both Action Replay and Retro Replay.
It has to do with the $1802 register in a 1541 drive using device number #9, #10 or #11.
AR & RR set this register to $7a when using fastload.
When fastload finishes it does not reset $1802 back to it's default value ($1a).
When you load and run a program from device #9,10 or 11 that makes use of the "UI" dos command, the following will happen:
The device number of your drive (9, 10 or 11) will be soft-changed to device #8. And you loose the connection, and
your program is stuck trying to communicate with a device that doesnt exist anymore. Nice!! :)
If you turn your drive off and on before running that
program it will work properly. But we dont want to do that everytime so the method to fix it is to memory-write $1a to the $1802 register.
I think RR should do that.
Many thanks to Soci for helping debugging this one.
GRG
|
|
| |
Frantic
Registered: Mar 2003
Posts: 1648 |
I think it would be a good idea to make a codebase wiki article out of this, since it could be useful for others. Is it OK if I put this stuff on there?
http://codebase64.org |
| |
6R6
Registered: Feb 2002
Posts: 245 |
@frantic: yes. |
| |
Frantic
Registered: Mar 2003
Posts: 1648 |
Alright.. Nice. It can now be found on this place:
http://codebase64.org/doku.php?id=base:cartridge_bugs
In case people have additional information on this topic, please add this to the wiki page. :) |
| |
Devia
Registered: Oct 2004
Posts: 401 |
At some point I discovered some loading bug in the Retro Replay ROM which I never did get to the bottom of..
Details are over at: http://retrohackers.org/forum/viewtopic.php?t=195
Anybody know of this or encountered it aswell?
.hmm.. maybe I should find a couple of minutes to answer RaveGuru in that post ;-) |
| |
Devia
Registered: Oct 2004
Posts: 401 |
@GRG: Rearding your RR+RRNet problems on C128D (Plastic i presume?), I have heard of several people reporting that it doesn't work. However it works on the 128Ds I have testet it on. There are similar symptoms on the SX-64.. For some it works, for some it don't. And it doesn't work on most of my SX-64s so I have tried a few things out.
For the record, I have the same problems when the RRNet is connected to an MMC64.
Please read the second post of this thread: http://retrohackers.org/forum/viewtopic.php?t=193
You could try out that little test code and report the results back.
I did manage to create a WORKING RRNet for my SX by building a custom ClockPort adapter. How I did that is somewhere in this thread: http://retrohackers.org/forum/viewtopic.php?t=132
|
| |
REBEL 1
Registered: May 2023
Posts: 5 |
With my AR5 (real hardware), the situation is as follows:
1) DF25 is in vector 0316/17. A “g DF25” does not open the freeze mode.
2) If you read the vectors (with LDA$0316 and 0317, “36” and “01” are then in the accumulator.
With AR6 (real hardware) everything works as expected: DFCD on 0316/17. “G DFCD” opens freeze mode; reading out 0316/0317 produces valid values. |
