DSRs and VDP RAM observation

[Kchula-Rrit]

When I was working on my "L23Dump" package I noticed some strange behavior.  I placed my PABs at the top of free VDP RAM, to avoid interfering with any PABs that the calling programs had set-up.  For example, C99 sets theirs at 0x1B00-1FFF, if I recall, and Tom Bentley's TCIO uses 0x2000-2500.

 

Anyway, I took the value at CPU RAM 8370 and subtracted the amount I used for buffers and PABs, then saved these VDP pointers in my program.  Upon exit I restored the original top-of-free-VDP-RAM pointer to CPU RAM 8370.

 

Sometimes the program would crash for no reason I could tell.  Everything looked OK, until I realized that I had not set 8370 to the new top-of-free value.  After that everything was okay.

 

Naive me, I figured that the disk DSR would not stray outside the addresses that Thierry Nouspikel said they used.  Obviously, I had not "reserved" the region of VDP RAM I had used for my buffers, and so I can see it getting nuked at some point.

 

K-R.