On Fri, 2017-12-15 at 16:52 +0000, Mel Gorman wrote: > > It's a small improvement...
>From my log on corr.arch.suse.de (2x8 box) 4.15.0.g2db767d-default Throughput 2665.88 MB/sec 8 clients 8 procs max_latency=21.472 ms 4.15.0.g2db767d-default NO_WA_IDLE Throughput 3416.35 MB/sec 8 clients 8 procs max_latency=9.825 ms Not so small improvement. WA_IDLE ripped corr up pretty bad, turning it off restored 4.4 performance. > > Signed-off-by: Mel Gorman <[email protected]> > --- > kernel/sched/fair.c | 8 +++++++- > 1 file changed, 7 insertions(+), 1 deletion(-) > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index 2fe3aa853e4d..4a1f7d32ecf6 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -5701,7 +5701,13 @@ static bool > wake_affine_idle(struct sched_domain *sd, struct task_struct *p, > int this_cpu, int prev_cpu, int sync) > { > - if (idle_cpu(this_cpu)) > + /* > + * If this_cpu is idle, it implies the wakeup is from interrupt > + * context. Only allow the move if cache is shared. Otherwise an > + * interrupt intensive workload could force all tasks onto one > + * node depending on the IO topology or IRQ affinity settings. > + */ > + if (idle_cpu(this_cpu) && cpus_share_cache(this_cpu, prev_cpu)) > return true; > > if (sync && cpu_rq(this_cpu)->nr_running == 1)
