> I thought that the point of RT sched. was that in a soft real time
> environment, given an expected amount of faults and a certain sys.
> load you can expect a certain failure rate, and in a hard real time
> environment given an expected number of faults and a sys. load you
> can guarantee that if you add another task to the current set of tasks
> it will (under most circumstances) succeed.
Although there are many different genre of 'realtime', the fundamental
factor is deterministic behavior such that a running thread or process
shall not be pre-empted except by an interrupt, or a higher priority
thread/process being marked 'runnable' (generally by an ISR via
semaphore/mutex release). This is vital to ensure a system can keep up
with hardware and data flow that cannot be allowed to 'miss' a critical
timing deadline.
Everything else just boils down to implementation differencess.
Fault tolerance is another issue entirely...
bye-
ron
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/