Re: INFO: rcu detected stall in do_idle

[Juri Lelli <juri.lelli@redhat.com>]

On 18/10/18 13:08, luca abeni wrote:
> On Thu, 18 Oct 2018 12:47:13 +0200
> Juri Lelli <juri.lelli@redhat.com> wrote:
> 
> > Hi,
> > 
> > On 18/10/18 12:23, luca abeni wrote:
> > > Hi Juri,
> > > 
> > > On Thu, 18 Oct 2018 10:28:38 +0200
> > > Juri Lelli <juri.lelli@redhat.com> wrote:
> > > [...]  
> > > >  struct sched_attr {
> > > >     .size	= 0,
> > > >     .policy	= 6,
> > > >     .flags	= 0,
> > > >     .nice	= 0,
> > > >     .priority	= 0,
> > > >     .runtime	= 0x9917,
> > > >     .deadline	= 0xffff,
> > > >     .period	= 0,
> > > >  }
> > > > 
> > > > So, we seem to be correctly (in theory, see below) accepting the
> > > > task.
> > > > 
> > > > What seems to generate the problem here is that CONFIG_HZ=100 and
> > > > reproducer task has "tiny" runtime (~40us) and deadline (~66us)
> > > > parameters, combination that "bypasses" the enforcing mechanism
> > > > (performed at each tick).  
> > > 
> > > Ok, so the task can execute for at most 1 tick before being
> > > throttled... Which does not look too bad.
> > > 
> > > I missed the original emails, but maybe the issue is that the task
> > > blocks before the tick, and when it wakes up again something goes
> > > wrong with the deadline and runtime assignment? (maybe because the
> > > deadline is in the past?)  
> > 
> > No, the problem is that the task won't be throttled at all, because
> > its replenishing instant is always way in the past when tick
> > occurs. :-/
> 
> Ok, I see the issue now: the problem is that the "while (dl_se->runtime
> <= 0)" loop is executed at replenishment time, but the deadline should
> be postponed at enforcement time.
> 
> I mean: in update_curr_dl() we do:
> 	dl_se->runtime -= scaled_delta_exec;
> 	if (dl_runtime_exceeded(dl_se) || dl_se->dl_yielded) {
> 		...
> 		enqueue replenishment timer at dl_next_period(dl_se)
> But dl_next_period() is based on a "wrong" deadline!
> 
> 
> I think that inserting a
>         while (dl_se->runtime <= -pi_se->dl_runtime) {
>                 dl_se->deadline += pi_se->dl_period;
>                 dl_se->runtime += pi_se->dl_runtime;
>         }
> immediately after "dl_se->runtime -= scaled_delta_exec;" would fix the
> problem, no?

Mmm, I also thought of letting the task "pay back" its overrunning. But,
doesn't this get us quite far from what one would expect. I mean,
enforcement granularity will be way different from task period, no?