On Fri, 2005-04-08 at 09:24 -0500, Jan Depner wrote: > > Personally I think you're wasting your time. Ingo's RT preempt patches > > let you do hard realtime with Linux using the existing driver base. > > > > I'm not sure why he's trying to do this because for audio the latest > patched kernels appear to be more than adequate. That said, I don't > think you can get a guaranteed 15 microsecond interrupt response with > anyone's patches to the standard kernel. I wouldn't call what we have > now hard real-time. More like soft real-time. Maybe when MontaVista > gets done...
Hard realtime has nothing to do with how short the deadlines are, it's about being able to *guarantee* that you can meet some deadline, IOW the ability to guarantee that every code path in the kernel is deterministic. And the RT preempt kernel absolutely can do hard realtime, down to about 50 usecs (which is the length of time the timer interrupt, the longest non preemptible path in the kernel, takes to run). Right now all the work that various people have done to enable hard realtime on Linux is being ported to the RT preempt kernel. When MontaVista and others get all their patches in we should be able to meet even smaller RT constraints. Lee
