Re: mm: deadlock between get_online_cpus/pcpu_alloc

From: Michal Hocko
Date: Wed Feb 08 2017 - 07:23:55 EST

On Wed 08-02-17 13:02:07, Thomas Gleixner wrote:
> On Wed, 8 Feb 2017, Michal Hocko wrote:
> > [1]
> Well, yes. It's simple, but from an RT point of view I really don't like
> it as we have to fix it up again.

I thought that preempt_disable would turn into migrate_disable or
something like that which shouldn't cause too much trouble. Or am I
missing something? Which part of the patch is so RT unfriendly?
Michal Hocko