quit key on 68000 (non-vbr) systems

[Yulquen74]

Are there any special code in WHDLoad to support quit key on systems that
lack VBR? The ACA500, Minimig and the upcoming Zeus supports standard quit key afaik.

Please correct me if Im wrong:

The end of the ROM contains some index bytes that is used to generate interrupt vector addresses, and with VBR it is possible to alter the location of those vectors, again making it possible to insert custom interrupt code to detect stuff like keypresses.

Is regular 68000 systems able to use quit key by the means of having ROM in same address space in RAM, and then changing the content of the "ROM" index bytes?

The reason for asking is a desire to implement the feature on my own simple
accelerator, if it is not too complicated.

Thanks in advance for any replies.

[Toni Wilen]

ACA500 method is described in ACA500 hardware documentation (http://wiki.icomp.de/wiki/ACA500)

WHDLoad uses it via aca.resource, there is no standard how to enable/disable it.

[Wepl]

Originally Posted by Yulquen74

Are there any special code in WHDLoad to support quit key on systems that
lack VBR? The ACA500, Minimig and the upcoming Zeus supports standard quit key afaik.

WHDLoad supports the ACA500 as Toni told and the unreleased Zeus in a similar way.

Originally Posted by Yulquen74

The end of the ROM contains some index bytes that is used to generate interrupt vector addresses, and with VBR it is possible to alter the location of those vectors, again making it possible to insert custom interrupt code to detect stuff like keypresses.

AFAIK the bytes at the end of ROM (end of address space of the cpu) are only used on 68000. The hardware tells the way how to obtain the vectors. The vbr specifies the base address for all interrupt and exception vectors (0 on 68000).

Originally Posted by Yulquen74

Is regular 68000 systems able to use quit key by the means of having ROM in same address space in RAM, and then changing the content of the "ROM" index bytes?

I rather think it snoops the bus during vector fetch and redirects the memory read.

[Yulquen74]

Ok, thanks for the clarification.

I have looked at the ACA500 manual, and the implementation seems easy enough to
understand, but if WHDLoad uses specially written software to access the actual config bit to enable the VBR move, even via an ACA spesific software layer, theres no point of pursuing this any further.

[jotd]

Bert, would it be possible to install VBLANK & KBINT on 68000 when calling resload functions?
If not already installed, then install VBLANK & KBINT which check for quitkey and then call the original value.
May work in a lot of slaves.

[Wepl]

Yes, that would work for some installs.
Not for installs which do not call a resload function after init the interrupt vectors. I think there are a bunch of games which load and then run the code without further resload calls.
Another problem may arise if the vectors are saved and restored multiple times. This may cause whdload saved vector to be wrong (game has saved the by whdload overwritten vector and later restores it instead the vector previous active by the game itself).