Re: [announcement, patch] real-time interrupts for the Linux kernel

[Bernhard Kuhn]

Martin Schlemmer wrote:

> Do you have actual benchmarks,

The example kernel module and application are doing
some simple benchmarking: the local APIC timer
runs in period mode (1 kHz), and the interrupt
handler then reads out the current value of the timer.
Then you know how long it took from the occurence of the
interrupt to the moment the handler has been invoked
(including context switching, do_IRQ and handle_IRQ_event).

Number is my setup under heavy load (ping-flood, glxgears):

                    99.9%     worst case
high priority        3.1 µs       5.2 µs
standard kernel    486.1 µs     500.8 µs

"99.9%" means that 99.9% of the interrupts or handled
in withing the given time.


> and what 2.4 rather than 2.6 ?

I started with 2.4 becuase that's the devil i know, but from
what i have inspected so far, adopting this scheme for 2.6
should be doable within a day - i just started to work in it.

best regards

Bernhard

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/