Re: Interrupt Processing
Posted: Tue Jul 06, 2010 10:01 pm
That's my point. For every single "generic" method/template that I know of for creating drivers and handling interrupts, you will run into problems. Each of those problems needs to have a workaround. Hopefully, the path that you choose for creating your interrupt handlers/device drivers ends up being flexible enough that you can find a workaround for every single incompatibility that you run into. You are guaranteed to find many. Sometimes the workaround (as in this case) will be to refuse to support some operation mode of some device. You always have to play a game of supporting as many devices and modes as possible, with as little CPU/system overhead as possible, with as few bugs as possible. It's not a trivial exercise. If you ever run into a significant hardware incompatibility which you can't find a workaround for, then you need to start all over and redesign your whole interrupt system.