[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20161129144437.GA9796@dhcp22.suse.cz>
Date: Tue, 29 Nov 2016 15:44:37 +0100
From: Michal Hocko <mhocko@...nel.org>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
linux-kernel@...r.kernel.org, rt@...utronix.de,
Andrew Morton <akpm@...ux-foundation.org>,
Vlastimil Babka <vbabka@...e.cz>,
Mel Gorman <mgorman@...hsingularity.net>,
Johannes Weiner <hannes@...xchg.org>, linux-mm@...ck.org
Subject: Re: [PATCH 08/22] mm/vmstat: Avoid on each online CPU loops
On Tue 29-11-16 15:08:10, Thomas Gleixner wrote:
> On Mon, 28 Nov 2016, Michal Hocko wrote:
> > On Sun 27-11-16 00:13:36, Sebastian Andrzej Siewior wrote:
> > [...]
> > > static void __init init_cpu_node_state(void)
> > > {
> > > - int cpu;
> > > + int node;
> > >
> > > - for_each_online_cpu(cpu)
> > > - node_set_state(cpu_to_node(cpu), N_CPU);
> > > + for_each_online_node(node)
> > > + node_set_state(node, N_CPU);
> >
> > Is this really correct? The point of the original code was to mark only
> > those nodes which have at least one CPU. Or am I missing something?
>
> You're right. An online node does not necessarily have an online CPU.
>
> for_each_online_node(node) {
> if (cpumask_weight(cpumask_of_node(node)) > 0)
> node_set_state(node, N_CPU);
> }
>
> is probably more correct.
Yes, this looks correct. Considering that the same cpumask_weight is
used in another function I guess a small helper would be nice. E.g.
bool node_has_cpus(int node)
{
return cpumask_weight(cpumask_of_node(node)) > 0;
}
--
Michal Hocko
SUSE Labs
Powered by blists - more mailing lists