View Single Post
Old 23 April 2014, 14:27   #198
Toni Wilen
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 43
Posts: 20,643
Originally Posted by amilo3438 View Post
Well, regarding the "behavior of reading write-only custom register is not fully known" is maybe not quite true ... why ?!

On an real amiga is possible to check what values written in some write-only custom register will be returned after reading and thus f.e. can be created some table for every write-only custom register, right ?!

It is similar to "black box"; we dont know what is inside and how it function but we know what values we got.
It does not work. Result depends on previous state of system and not just last cycle. You get different results if you repeat same test multiple times.

Result only seems to be stable if previous cycle was idle cycle (no cpu and dma accesses), most likely result is FFFF in that case.

It seems there is some low level chip behavior which causes some kind of semi-random AND logic behavior. Result usually looks like it last bus value AND last value in buffer chip between CPU and Agnus.

It is most likely chip level undefined state, reading chip outputs when chip is in high-z mode or similar, it depends on internal design.

Thanks for fixing the demo, but also I never said it is "not working" ?! ... what I said is its "stop to work somewhere from 2710 beta 8", right ?

Yes you did. "not working" means problem was not described. It could mean it stopped booting, it crashes before it starts, it has bad colors, bad graphics, sound problems, anything.

Something like "it now crashes with graphics corruption before camera part" would have worked to get some information and to confirm I have duplicated exact same problem.
Toni Wilen is offline  
Page generated in 0.09676 seconds with 9 queries