sched/rt: Fix rt task's sched latency statistics in sched_stat_wait trace_point
PATCH 1/1 sched/rt: Fix rt task's sched latency statistics in sched_stat_wait trace_point
When enable sched_stat_wait trace_point, some rt tasks sched latency so long, like this:sched_stat_wait: comm=rcu_preempt pid=14 delay=4936139545261 [ns] Rt task has low latency, it must have a bug. I found the reason is When rt task balance off source cpu, dequeue operation not update the sched_statistics, so follow update_stats_wait_end_fair update method.
--- kernel/sched/rt.c | 3 +++ 1 file changed, 3 insertions(+)
kernel/sched/rt.c +3 -0
@@ -1360,12 +1360,15 @@ update_stats_dequeue_rt(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, int flags) {struct task_struct *p =NULL ; +struct rq *rq =rq_of_rt_se (rt_se);if (!schedstat_enabled ())return ;if (rt_entity_is_task (rt_se)) p =rt_task_of (rt_se); +if (p != rq->curr) +update_stats_wait_end_rt (rt_rq, rt_se);if ((flags &DEQUEUE_SLEEP ) && p) {unsigned int state;
Signature · 1 line
2.34.1
Quoted 8 lines · Junwen Wu
When enable
sched_stat_waittrace_point, some rt tasks sched latency so long, like this:sched_stat_wait: comm=rcu_preempt pid=14 delay=4936139545261 [ns] Rt task has low latency, it must have a bug. I found the reason is When rt task balance off source cpu, dequeue operation not update thesched_statistics, so followupdate_stats_wait_end_fairupdate method.
Thanks for your report!
Quoted 12 lines
Signed-off-by Junwen Wu <[email protected]>--- kernel/sched/rt.c | 3 +++ 1 file changed, 3 insertions(+)kernel/sched/rt.c +0
-0@@ -1360,12 +1360,15 @@ update_stats_dequeue_rt(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, int flags)
kernel/sched/rt.c update_stats_dequeue_rt(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, · line 1370
Quoted 10 lines
kernel/sched/rt.c +2
-0{struct task_struct *p =NULL ; +struct rq *rq =rq_of_rt_se (rt_se);if (!schedstat_enabled ())return ;if (rt_entity_is_task (rt_se)) p =rt_task_of (rt_se); +if (p != rq->curr)
This seems wrong, the rt_se might not be a task.
I think we'd better use if (!on_rt_rq(rt_se))
Quoted 7 lines
kernel/sched/rt.c +1
-1+update_stats_wait_end_rt (rt_rq, rt_se);if ((flags &DEQUEUE_SLEEP ) && p) {unsigned int state; --2.34.1
--
Regards
Yafang
ok, I will take your advise, and test my platform.
--
Regards
Junwen
Re: [PATCH v1] sched/rt: Fix rt task's sched latency statistics error in sched_stat_wait trace_point
Quoted 4 lines
Quoted 4 lines
I think we'd better use if (!on_rt_rq(rt_se))
+
update_stats_wait_end_rt(rt_rq, rt_se);
hi, Yafang.when execute update_stats_dequeue_rt, rt_se->on_rq is still 0, util dequeue_rt_stack, the method is not effect.
I this we can use if (p && p != rq->curr) /*Mark the end of the wait period if dequeueing task*/ because schedstats is not supported for rt group,we only need to update rt_se that is realy task. --
Best regards
Re: [PATCH v1] sched/rt: Fix rt task's sched latency statistics error in sched_stat_wait trace_point
sorry, type error. rt_se->on_rq is still 1
--
Best regards
Re: [PATCH v1] sched/rt: Fix rt task's sched latency statistics error in sched_stat_wait trace_point
In reply to Junwen WuQuoted 7 lines · Junwen Wu
Quoted 5 lines
Quoted 5 lines
I think we'd better use if (!on_rt_rq(rt_se))
+
update_stats_wait_end_rt(rt_rq, rt_se);hi, Yafang.when execute
update_stats_dequeue_rt,rt_se->on_rqis still 0, utildequeue_rt_stack, the method is not effect.
Ah, you are right.
Quoted 1 line
I this we can use if (p && p !=
rq->curr) /*Mark the end of the wait period if dequeueing task*/
Agreed.
Quoted 4 lines
because schedstats is not supported for rt group,we only need to update
rt_sethat is realy task. --
Best regards
Signature · 2 lines
Regards Yafang
Re: [PATCH v1] sched/rt: Fix rt task's sched latency statistics error in sched_stat_wait trace_point
Quoted 10 lines
Ah, you are right.
Quoted 1 line
I this we can use if (p && p !=
rq->curr) /*Mark the end of the wait period if dequeueing task*/Agreed.
Quoted 4 lines
because schedstats is not supported for rt group,we only need to update
rt_sethat is realy task. --
Best regards
ok, I send PATCH v2 later
--
Best regards