Re: Perf regression from scheduler load_balance rework in 5.5?

From: Vincent Guittot
Date: Mon Jun 27 2022 - 07:00:04 EST


Hi,

Le vendredi 24 juin 2022 à 21:16:05 (+0800), Zhang Qiao a écrit :
>
> Hi,
> 在 2022/6/24 16:22, Vincent Guittot 写道:
> > On Thu, 23 Jun 2022 at 21:50, David Chen <david.chen@xxxxxxxxxxx> wrote:
> >>
> >> Hi,
> >>
> >> I'm working on upgrading our kernel from 4.14 to 5.10
> >> However, I'm seeing performance regression when doing rand read from windows client through smbd
> >> with a well cached file.
> >>
> >> One thing I noticed is that on the new kernel, the smbd thread doing socket I/O tends to stay on
> >> the same cpu core as the net_rx softirq, where as in the old kernel it tends to be moved around
> >> more randomly. And when they are on the same cpu, it tends to saturate the cpu more and causes
> >> performance to drop.
> >>
> >> For example, here's the duration (ns) the thread spend on each cpu I captured using bpftrace
> >> On 4.14:
> >> @cputime[7]: 20741458382
> >> @cputime[0]: 25219285005
> >> @cputime[6]: 30892418441
> >> @cputime[5]: 31032404613
> >> @cputime[3]: 33511324691
> >> @cputime[1]: 35564174562
> >> @cputime[4]: 39313421965
> >> @cputime[2]: 55779811909 (net_rx cpu)
> >>
> >> On 5.10:
> >> @cputime[3]: 2150554823
> >> @cputime[5]: 3294276626
> >> @cputime[7]: 4277890448
> >> @cputime[4]: 5094586003
> >> @cputime[1]: 6058168291
> >> @cputime[0]: 14688093441
> >> @cputime[6]: 17578229533
> >> @cputime[2]: 223473400411 (net_rx cpu)
> >>
> >> I also tried setting the cpu affinity of the smbd thread away from the net_rx cpu and indeed that
> >> seems to bring the perf on par with old kernel.
>
> I observed the same problem for the past two weeks.
>
> >>
> >> I noticed that there's scheduler load_balance rework in 5.5, so I did the test on 5.4 and 5.5 and
> >> it did show the behavior changed between 5.4 and 5.5.
> >
> > Have you tested v5.18 ? several improvements happened since v5.5
> >
> >>
> >> Anyone know how to work around this?
> >
> > Have you enabled IRQ_TIME_ACCOUNTING ?
>
>
> CONFIG_IRQ_TIME_ACCOUNTING=y.
>
> >
> > When the time spent under interrupt becomes significant, scheduler
> > migrate task on another cpu
>
>
> My board has two cpus, and i used iperf3 to test upload bandwidth,then I saw the same situation,
> the iperf3 thread run on the same cpu as the NET_RX softirq.
>
> After debug in find_busiest_group(), i noticed when the cpu(env->idle is CPU_IDLE or CPU_NEWLY_IDLE) try to pull task,
> the busiest->group_type == group_fully_busy, busiest->sum_h_nr_running == 1, local->group_type==group_has_spare,
> and the loadbalance will failed at find_busiest_group(), as follows:
>
> find_busiest_group():
> ...
> if (busiest->group_type != group_overloaded) {
> ....
> if (busiest->sum_h_nr_running == 1)
> goto out_balanced; ----> loadbalance will returned at here.

Yes, you're right, we filter such case. Could you try the patch below ?
I use the misfit task state to detect cpu with reduced capacity and migrate_load
to check if it worth migration the task on the dst cpu.

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 6775a117f3c1..013dcd97472b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -8757,11 +8757,19 @@ static inline void update_sg_lb_stats(struct lb_env *env,
if (local_group)
continue;

- /* Check for a misfit task on the cpu */
- if (env->sd->flags & SD_ASYM_CPUCAPACITY &&
- sgs->group_misfit_task_load < rq->misfit_task_load) {
- sgs->group_misfit_task_load = rq->misfit_task_load;
- *sg_status |= SG_OVERLOAD;
+ if (env->sd->flags & SD_ASYM_CPUCAPACITY) {
+ /* Check for a misfit task on the cpu */
+ if (sgs->group_misfit_task_load < rq->misfit_task_load) {
+ sgs->group_misfit_task_load = rq->misfit_task_load;
+ *sg_status |= SG_OVERLOAD;
+ }
+ } else if ((env->idle != CPU_NOT_IDLE) &&
+ (group->group_weight == 1) &&
+ (rq->cfs.h_nr_running == 1) &&
+ check_cpu_capacity(rq, env->sd) &&
+ (sgs->group_misfit_task_load < cpu_load(rq))) {
+ /* Check for a task running on a CPU with reduced capacity */
+ sgs->group_misfit_task_load = cpu_load(rq);
}
}

@@ -8814,7 +8822,8 @@ static bool update_sd_pick_busiest(struct lb_env *env,
* CPUs in the group should either be possible to resolve
* internally or be covered by avg_load imbalance (eventually).
*/
- if (sgs->group_type == group_misfit_task &&
+ if ((env->sd->flags & SD_ASYM_CPUCAPACITY) &&
+ (sgs->group_type == group_misfit_task) &&
(!capacity_greater(capacity_of(env->dst_cpu), sg->sgc->max_capacity) ||
sds->local_stat.group_type != group_has_spare))
return false;
@@ -9360,9 +9369,15 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
busiest = &sds->busiest_stat;

if (busiest->group_type == group_misfit_task) {
- /* Set imbalance to allow misfit tasks to be balanced. */
- env->migration_type = migrate_misfit;
- env->imbalance = 1;
+ if (env->sd->flags & SD_ASYM_CPUCAPACITY) {
+ /* Set imbalance to allow misfit tasks to be balanced. */
+ env->migration_type = migrate_misfit;
+ env->imbalance = 1;
+ } else {
+ /* Set group overloaded as one cpu has reduced capacity */
+ env->migration_type = migrate_load;
+ env->imbalance = busiest->group_misfit_task_load;
+ }
return;
}


> ....
>
>
> Thanks,
> Qiao
>
>
> > Vincent>>
> >> Thanks,
> >> David
> > .
> >