Felix von Leitner wrote:
>
> Thus spake Paul Barton-Davis (pbd@Op.Net):
> > All we need is guaranteed scheduling response. We don't need QOS
> > guarantees for any other subsystems, for example (it would be nice,
> > but its not necessary).
>
> Oh yes, we do!
>
> What good is a low latency response if there is no QoS to guarantee
> enough CPU and memory? I don't consider locking memory pages and the
> FIFO scheduler a valid solution to this, because a single misbehaving
> process can cause a complete denial of service.
>
> I find guaranteed disk and SCSI bus throughput much more important than
> your MIDI stuff, because more people are burning CDs under Linux than
> there are MIDI users on all operating systems combined ;-)
But the problem is the same - what good will your throughput be if your
burner process is not run for 500ms? You need A LOT of throughput to be
able to catch up several latency hits...
[ok, There is a lot of buffer on cd writers - but it will prevent going
faster]
The problem is exactly the same - you stream data to a device.
The device interrupts when it needs more data.
The actual program is not run.....
You get a drop out.
With audio you get audible disturbance, with CD writing you get a
destroyed CD...
/RogerL
-- Home page: http://www.norran.net/nra02596/- 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/
This archive was generated by hypermail 2b29 : Fri Jul 07 2000 - 21:00:13 EST