Re: [PATCH] sched/rt: Print curr when RT throttling activated

[Steven Rostedt]

Peter?

-- Steve


On Tue, 8 Dec 2020 07:58:54 +0000
Tianxianting <tian.xianting@xxxxxxx> wrote:

> Thanks,
> We met an issue that a normal thread can't get cpu, 
> And at this moment, we found 'sched: RT throttling activated' log.
> 
> So I think this patch is useful for such issue.
> 
> Could I get more comments?  Thanks in advance
> -----Original Message-----
> From: Steven Rostedt [mailto:rostedt@xxxxxxxxxxx] 
> Sent: Thursday, December 03, 2020 10:40 PM
> To: tianxianting (RD) <tian.xianting@xxxxxxx>
> Cc: mingo@xxxxxxxxxx; peterz@xxxxxxxxxxxxx; juri.lelli@xxxxxxxxxx; vincent.guittot@xxxxxxxxxx; dietmar.eggemann@xxxxxxx; bsegall@xxxxxxxxxx; mgorman@xxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH] sched/rt: Print curr when RT throttling activated
> 
> On Thu, 3 Dec 2020 15:51:29 +0800
> Xianting Tian <tian.xianting@xxxxxxx> wrote:
> 
> > We may meet the issue, that one RT thread occupied the cpu by 
> > 950ms/1s, The RT thread maybe is a business thread or other unknown thread.
> > 
> > Currently, it only outputs the print "sched: RT throttling activated"
> > when RT throttling happen. It is hard to know what is the RT thread, 
> > For further analysis, we need add more prints.
> > 
> > This patch is to print current RT task when RT throttling activated, 
> > It help us to know what is the RT thread in the first time.  
> 
> I think this can be useful information to include.
> 
> Acked-by: Steven Rostedt (VMware) <rostedt@xxxxxxxxxxx>
> 
> -- Steve
> 
> > 
> > Signed-off-by: Xianting Tian <tian.xianting@xxxxxxx>
> > ---
> >  kernel/sched/rt.c | 7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> > 
> > diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 
> > f215eea6a..8913f38cb 100644
> > --- a/kernel/sched/rt.c
> > +++ b/kernel/sched/rt.c
> > @@ -946,7 +946,7 @@ static inline int rt_se_prio(struct sched_rt_entity *rt_se)
> >  	return rt_task_of(rt_se)->prio;
> >  }
> >  
> > -static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq)
> > +static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq, struct 
> > +task_struct *curr)
> >  {
> >  	u64 runtime = sched_rt_runtime(rt_rq);
> >  
> > @@ -970,7 +970,8 @@ static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq)
> >  		 */
> >  		if (likely(rt_b->rt_runtime)) {
> >  			rt_rq->rt_throttled = 1;
> > -			printk_deferred_once("sched: RT throttling activated\n");
> > +			printk_deferred_once("sched: RT throttling activated (curr: pid %d, comm %s)\n",
> > +						curr->pid, curr->comm);
> >  		} else {
> >  			/*
> >  			 * In case we did anyway, make it go away, @@ -1026,7 +1027,7 @@ 
> > static void update_curr_rt(struct rq *rq)
> >  		if (sched_rt_runtime(rt_rq) != RUNTIME_INF) {
> >  			raw_spin_lock(&rt_rq->rt_runtime_lock);
> >  			rt_rq->rt_time += delta_exec;
> > -			if (sched_rt_runtime_exceeded(rt_rq))
> > +			if (sched_rt_runtime_exceeded(rt_rq, curr))
> >  				resched_curr(rq);
> >  			raw_spin_unlock(&rt_rq->rt_runtime_lock);
> >  		}