With a lot of small tasks, the softirq sched is nearly never called when no_hz is enable. In this case the load_balance is mainly called with the newly_idle mode which doesn't update the cpu_power. Add a next_update field which ensure a maximum update period when there is short activity
Signed-off-by: Vincent Guittot vincent.guittot@linaro.org --- include/linux/sched.h | 1 + kernel/sched/fair.c | 24 ++++++++++++++++-------- 2 files changed, 17 insertions(+), 8 deletions(-)
diff --git a/include/linux/sched.h b/include/linux/sched.h index 64527c4..7178446 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -903,6 +903,7 @@ struct sched_group_power { * single CPU. */ unsigned int power, power_orig; + unsigned long next_update; /* * Number of busy cpus in this group. */ diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index a4d2b7a..809f484 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -215,6 +215,8 @@ calc_delta_mine(unsigned long delta_exec, unsigned long weight,
const struct sched_class fair_sched_class;
+static unsigned long __read_mostly max_load_balance_interval = HZ/10; + /************************************************************** * CFS operations on generic schedulable entities: */ @@ -3786,6 +3788,11 @@ void update_group_power(struct sched_domain *sd, int cpu) struct sched_domain *child = sd->child; struct sched_group *group, *sdg = sd->groups; unsigned long power; + unsigned long interval; + + interval = msecs_to_jiffies(sd->balance_interval); + interval = clamp(interval, 1UL, max_load_balance_interval); + sdg->sgp->next_update = jiffies + interval;
if (!child) { update_cpu_power(sd, cpu); @@ -3893,12 +3900,15 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, * domains. In the newly idle case, we will allow all the cpu's * to do the newly idle load balance. */ - if (idle != CPU_NEWLY_IDLE && local_group) { - if (balance_cpu != this_cpu) { - *balance = 0; - return; - } - update_group_power(sd, this_cpu); + if (local_group) { + if (idle != CPU_NEWLY_IDLE) { + if (balance_cpu != this_cpu) { + *balance = 0; + return; + } + update_group_power(sd, this_cpu); + } else if (time_after_eq(jiffies, group->sgp->next_update)) + update_group_power(sd, this_cpu); }
/* Adjust by relative CPU power of the group */ @@ -4917,8 +4927,6 @@ void select_nohz_load_balancer(int stop_tick)
static DEFINE_SPINLOCK(balancing);
-static unsigned long __read_mostly max_load_balance_interval = HZ/10; - /* * Scale the max load_balance interval with the number of CPUs in the system. * This trades load-balance latency on larger machines for less cross talk.