Messages in this thread | | | Date | Fri, 23 Aug 2013 13:51:54 +0900 | From | Joonsoo Kim <> | Subject | Re: [PATCH 02/10] sched: Factor out code to should_we_balance() |
| |
On Thu, Aug 22, 2013 at 12:42:57PM +0200, Peter Zijlstra wrote: > > > > > > +redo: > > > > One behavioral change worth noting here is that in the redo case if a > > CPU has become idle we'll continue trying to load-balance in the > > !new-idle case. > > > > This could be unpleasant in the case where a package has a pinned busy > > core allowing this and a newly idle cpu to start dueling for load. > > > > While more deterministically bad in this case now, it could racily do > > this before anyway so perhaps not worth worrying about immediately. > > Ah, because the old code would effectively redo the check and find the > idle cpu and thereby our cpu would no longer be the balance_cpu. > > Indeed. And I don't think this was an intentional change. I'll go put > the redo back before should_we_balance().
Ah, yes. It isn't my intention. Please fix it.
Thanks.
| |