I've pondered this for a while. I know that this behaviour is discrepant between Amiga models (Linux on my A1200 used to moan about it when ctrl-a-a was pressed). I don't actually know if the Amiga handles this; I've never probed the Kickstart ROM to work out it out. Can the 6502 code be read off of the keyboard MCU? That would be useful but I doubt it (most microcontrollers employ anti-theft mechanisms to prevent code being lifted from a device once programmed). I also don't have the kit for that.
However.
Anecdotally, dev docs refer to sending a reset warning keycode to the Amiga in order to get it to shore up its business before reset is asserted. If the reset warning isn't acknowledged, it's assumed the Amiga has gone to lunch and the reset pin (or clock low hold) is employed after a delay to ensure that the reset actually happens. This should give the Amiga time to sync any filesystem buffers to disk before a reset occurs.
I don't actually know if in AmigaOS this ever occurs.
I've pondered this for a while. I know that this behaviour is discrepant between Amiga models (Linux on my A1200 used to moan about it when ctrl-a-a was pressed). I don't actually know if the Amiga handles this; I've never probed the Kickstart ROM to work out it out. Can the 6502 code be read off of the keyboard MCU? That would be useful but I doubt it (most microcontrollers employ anti-theft mechanisms to prevent code being lifted from a device once programmed). I also don't have the kit for that.
However.
Anecdotally, dev docs refer to sending a reset warning keycode to the Amiga in order to get it to shore up its business before reset is asserted. If the reset warning isn't acknowledged, it's assumed the Amiga has gone to lunch and the reset pin (or clock low hold) is employed after a delay to ensure that the reset actually happens. This should give the Amiga time to sync any filesystem buffers to disk before a reset occurs.
I don't actually know if in AmigaOS this ever occurs.