> On Fri, 1 Oct 1999, B. James Phillippe wrote:
>
> >do_bottom_half functionality into a kernel thread. The reason is to allow
>
> it would be too slow. You want bh processing after incoming network packet
> is been queued from the irq.
That's probably true. The latency would suck. But I think if you were in
a situation where the bottom-half thread was continually ready to run
anyway, it might not be so slow. Anyway, throttling of the bh code was the
main intent of the wacky idea in the first place. :-)
> Also you can't preemt kernel code. no-way. You could only add a
> conditional schedule.
I assumed that kernel threads were scheduled as normal processes, from a
cursory glance at some of the places they are used (init, nfs, kmod). If
this true only after exec or ? Where did I go wrong?
> Currently if the bh is too long (as the ppa one) you can simply use a
> state machine, requeue the bh handler and exit from the bh. The next time
Actually, the specific case that caused me to want to try this is
maximum-throughput networking on multiple interfaces.
-bp
-- # bryan at terran dot org # http://www.terran.org/~bryan
- 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/