Re: [patch] sched: fix scheduling latencies for !PREEMPT kernels

From: Nick Piggin
Date: Tue Sep 14 2004 - 08:40:17 EST


Ingo Molnar wrote:
* Nick Piggin <nickpiggin@xxxxxxxxxxxx> wrote:


Could these ones go up a level? We break down scanning into 32 page
chunks, so I don't think it needs to be checked every page.


not really - we can occasionally get into high latencies even with a
single page - if a single page is mapped by alot of processes.


So doing it in the loop doesn't really give you a deterministic
maximum latency if somebody is out to cause trouble, does it?

OTOH, I guess libc or some shared memory segment may be mapped
into a lot of processes even on RT applictions.

Another thing, I don't mean this to sound like a rhetorical question,
but if we have a preemptible kernel, why is it a good idea to sprinkle
cond_rescheds everywhere? Isn't this now the worst of both worlds?
Why would someone who really cares about latency not enable preempt?
-
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/