> 
> Ken,
>   I am following the same path as you with my display driver. I think
> I finally got the driver under control. I can do open, reset, switch
> modes. I do not see any more problems with all other calls, I will
> implement then as soon as I find time. Yes, I implemented those system
> calls (traps) and also some of the slot manager. I had also to map in
> 0xF900000 space , as the driver uses hard-wired adresses like
> 0xF920000. I did not mess around SIntInstall and SIntRemove, I used
> stub routines. I guess you did not come yet to control calls, but they
> have the following problem: if I call them asynchronously, they jump
> to the address specified in 0x8FC, which is zero. What is it, queue
> handler? Anyway, there is no sign of thge true interrupt handler, so
> this part of the code should be left alone (I guess).

Yea! Two people making progress on this front! :-) :-)

I seem to recall from my driver docs (which are at home) that all
drivers are supposed to JUMP to an Apple IO completion routine. I think
it's the IODone routine. In fact, I just snooped in the CodeWarior
headers, and $8FC is named "JIODone". As I recall this address points
to the routine which will de-queue the IO request, and make any
requesit "IO Done" callbacks.

Take care,

Bill