> The main problem of doing ordered inserts into the run queue is that
> the current scheduler code assumes that inserts can only happen at
> the head and so it leaves the queue unlocked while traversing the
> list. This only affects SMP performance though, and could be solved
> by having per-processor run queues. Its debatable whether its worth
> the extra complexity though.
Look closer. The run queue can have entries added to its head during
interrupts as well. So it's not true that it only affects SMP
performance. It would affect interrupt latency too.
Sacrificing interrupt latency is not on.
Regards,
Richard....
-
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/