Baremetal Amiga Emulator

[mschulz]

Originally Posted by bloodline

I have uploaded a few MFM ADFs to the GitHub, one has SYSInfo on it.

Now this is awesome! :-D Congrats!

[bloodline]

Originally Posted by mschulz

Now this is awesome! :-D Congrats!

Cheers Michal!

Now I need to figure out how to swap in your Emu68... we may need a long chat about interrupts

[Pyromania]

Baremetal is an interesting and exciting project.

[Higgy]

Excellent project. I came across this via looking for C64 Bare Metal emulators.
Sorry I can't help in any technical capacity.
Power to the Pi!

[bloodline]

Ok, CIA question time!!

When a Key is pressed a raw key value (0x20 for the "A" key for example) is placed into CIA A sdr register. The Interrupt bit SP is set in the icr, and the CIA generates an interrupt on paula.

when a key is released, a raw key value is placed into CIA A sdr register (and bit 7, the topmost bit is set (0xA0 for the "A" key in this example) )... SP bit set, interrupt is again generated.

This is the model I am using... But I'm getting spurious keyboard inputs, have I missed a step?

-Edit- Reading the HRM again... Perhaps the rawkey value is shifted left 1 bit and the key up flag is added at bit position 0... I'll need to wait until later to test.

-edit2- and now I realise I need to invert the bits... ok it works. Omega has keyboard input

[aros-sg]

Originally Posted by bloodline

Running DPaint seems to give a clear reproducible blitter error.

If adat was masked (afwm/alwm) previousA may need to be set to that masked adat, not to the original unmasked adat.

[Toni Wilen]

Keycode is rotated and inverted. Keyboard handshake is also required, keyboard won't send new key codes (*) until handshake pulse.

*) not 100% true but lets not make this unnecessarily complex.

[bloodline]

Originally Posted by Toni Wilen

Keycode is rotated and inverted. Keyboard handshake is also required, keyboard won't send new key codes (*) until handshake pulse.

*) not 100% true but lets not make this unnecessarily complex.

Thanks Toni.

Yes, I was missing that inversion step. I now have basic keyboard support. I haven't implemented any handshake, which will probably cause a problem later on... but for now I have it working enough to test more critical parts of the emulator.

[bloodline]

Originally Posted by aros-sg

If adat was masked (afwm/alwm) previousA may need to be set to that masked adat, not to the original unmasked adat.

Good spot! I will test this later! Many thanks.

-Edit- Yes, that was a significant bug. Fixed now! I can't believe I managed to get as far as I did with it

[bloodline]

Lores screens work fine up to 5 bit planes.
Hires screens work fine up to 3 bit planes... When the forth bit plane is active this happens:

[ Show youtube player ]



Probably related, but my Copper is not coping well either:

[ Show youtube player ]