trunk/src/mame/audio/hng64.c
| r244986 | r244987 | |
|---|
| 21 | 21 | |
| 22 | 22 | IRQ mask register on the internal interrupt controller is set to 0xd8 |
| 23 | 23 | |
| 24 | | so levels 0,1,2,5 are unmasked |
| 24 | so levels 0,1,2,5 are unmasked, vectors get set during the sound CPU init code. |
| 25 | 25 | |
| 26 | | returning random values / triggering random interrupts eventually results in a situation |
| 27 | | where the CPU stops writing to the sound related addresses and starts reading / masking the |
| 28 | | serial comms register. |
| 26 | level 0/1 irq (fatfurwa) starts at 0xd277 (both the same vector) |
| 27 | serial comms related, maybe to get commands from main CPU if not done with shared ram? |
| 29 | 28 | |
| 29 | level 2 irq (fatfurwa) 0xdd20 |
| 30 | simple routine increases counter in RAM, maybe hooked to one / all of the timer irqs |
| 31 | |
| 32 | level 5 irq: (fatfurwa) starts at 0xc1e1 |
| 33 | largest irq, does things with ports 100 / 102 / 104 / 106, 10a (not 108 directly tho) |
| 34 | |
| 35 | no other irqs (or the NMI) are valid. |
| 36 | |
| 30 | 37 | */ |
| 31 | 38 | |
| 32 | 39 | |