Re: [PATCH 0/6] Optimize the cpu hotplug locking -v2

[Peter Zijlstra <peterz@infradead.org>]

On Thu, Oct 10, 2013 at 10:48:56AM -0700, Andrew Morton wrote:
> > > Very much agreed; now stop_machine() wouldn't actually work for hotplug
> > > because it will instantly preempt everybody, including someone who might
> > > be in the middle of using per-cpu state of the cpu we're about to
> > > remove.
> > 
> > Well, stop machine doesn't instantly preempt everybody. Only those that
> > don't have preemption disabled. Using per_cpu without preemption
> > disabled can be dangerous.

Can be, but there's more of that around than you want. Also that's not
the entire issue.

> Yes, I'd have thought that the cases where a CPU is fiddling with
> another CPU's percpu data with preemption enabled would be rather rare.

Look at kernel/events/core.c:swevent_hlist_get() it does something like:

  get_online_cpus();
  for_each_online_cpu() {
  	allocate_hash_table();
  }
  put_online_cpus();

Surely you don't want to do that with preemption disabled?

But my point is that even though there aren't many of these today; with
the growing number of cpus in 'commodity' hardware you want to move away
from using preempt_disable() as hotplug lock.