[Coco] Coco virus/worms are possible!

[Arthur Flexser]

On Sun, 12 Dec 2004, Robert Gault wrote:

> A topic that comes up occasionally is "can a Coco be infected in a 
> manner similar to current PCs". Generally the answer is no but I've just 
> come across a method that could make infection a possibility.
> 
> I was asked about a Coco game that was stated to start running when the 
> DIR command was issued. Now how can that be possible as DIR just prints 
> text to the screen. Well, DIR does much more in order to get that text 
> for printing.
> 
> One of the things done by DIR is to load the disk FAT table into a 
> buffer at $600. As this table is no larger than a 256 byte sector, this 
> should be safe. However, the game formats track 17 with six 1024 byte 
> sectors. That means when DIR loads T17S2 (the FAT) it gets more than 
> bargained for. The buffer is flooded way past the NMI vector at $983.
> 
> Since all disk I/O makes use of the NMI, this game probably puts an 
> address into $982 that vectors to some code loaded from T17S2. Nothing 
> really bad can happen on a floppy system but on a hard drive system a 
> malicious program could erase the hard drive.
> 
> 

This sounds like an urban legend to me, at least in the absence of information
about what game supposedly does this.  Steve Bjork wrote a CoCo virus of this
sort, he said, but never released it, just to see if it could be done.  But it
seems highly unlikely to me that there would have been a game that used this
technique without anyone prior to just now having remarked upon it.

Art