| ??? 09/09/04 14:59 Read: times |
#77137 - RE: Proposal: Internal Data Bus Responding to: ???'s previous message |
Craig Steiner wrote:
------------------------------- That too. I suspect mass storage can be handled by that same 8052 though. If not it could be broken out into a fourth slave. -even outside the basic "address space" being a slave to the main I/O chip, not a slave to the CPU. Wow, looking to play movies and MP3s Movies - not really, but Amiga .mod music - if we can only fit all the samples in RAM, playing them should be easy. Otherwise we're going to have to constantly poll for mouse movement, keypresses, network traffic, etc. We could poll one "activity bit" only. Then if set, just find the source. But true, interrupts and buffers are a prettier way to do it. I wouldn't think video would need to interrupt the master since it is output only. I'd like to have a maskable vblank interrupt. Allows for making smooth animation and a hundred of other valuable applications. vblank seemed to be the most used and abused interrupt both in Amiga and in Atari, and with 3 INT lines of the CPU we can afford it I think. I doubt the audio 8052 would need to interrupt the master--unless we're considering a soundcard-like "Audio In" functionality to digitize incoming audio?? Even that doesn't need an interrupt, if given DMA ability. But the audio chip should have an incoming interrupt line, so it performs its standard operation (generating a wave, playback from RAM etc) all the time and accepts commands only on interrupt. |
