Re: [RFC] sched_tick with interrupts enabled

[Nick Piggin]

Christoph Lameter wrote:
> > After that, it might be acceptable to call rebalance_tick from a tasklet,
> > although it would be uneeded overhead on small systems. It might be better
>
>
> Here is a patch that only runs rebalance tick from a tasklet in the 
> scheduler. For UP there will be no tasklet.
>
> [RFC] sched_rebalance_tick with interrupts enabled
>
> scheduler_tick() has the potential of running for some time if f.e.
> sched_domains for a system with 1024 processors have to be balanced.
> We currently do all of that with interrupts disabled. So we may be unable
> to service interrupts for some time. Most of that time is potentially
> spend in rebalance_tick.
>
> This patch splits off rebalance_tick from scheduler_tick and schedules
> it via a tasklet.
>
> Signed-off-by: Christoph Lameter <clameter@xxxxxxx>

wake_priority_sleeper should not be called from rebalance_tick. That
code was OK where it was before, I think.

And you need to now turn off interrupts when doing the locking in
load_balance.

--
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com 
-
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/