Dnia 00-08-02 Andrew Gale pisze: >> The easiest way to avoid this problem is to latch the NMI >> request from the button and then wait for the CPU to take opcode >> (/M1=0) dealing with the stack. > That's a very neat idea.... And perhaps simplest to do. The possible delay have no meaning, as nobody is able to press NMI button with precision greater than 100ms (unless light pen or a condition from bus is used to generate NMI, but I doubt someone will use them).
> That had been my thinking - disable the SAM's z80 and hang a > different z80 off the back... then I could trap and memory > writes before they reach the ASIC. The problem could be that the NMI isn't acknowledged so you have to distinguish stacking the PC from opcode writing to RAM. > Does a write to the ROM region not write to the underlying RAM? Yes, it doesn't. It happens so in Amstrad CPC, but not in SAM. I've expected so, but now I checked it. -- Yarek.
