From: luca abeni <luca.abeni@unitn.it>
To: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
Juri Lelli <juri.lelli@arm.com>,
Claudio Scordino <claudio@evidence.eu.com>,
Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [RFC v3 2/6] Improve the tracking of active utilisation
Date: Fri, 18 Nov 2016 16:56:47 +0100 [thread overview]
Message-ID: <20161118165647.46f8c867@sweethome> (raw)
In-Reply-To: <20161118153615.GM3142@twins.programming.kicks-ass.net>
On Fri, 18 Nov 2016 16:36:15 +0100
Peter Zijlstra <peterz@infradead.org> wrote:
> On Mon, Oct 24, 2016 at 04:06:34PM +0200, Luca Abeni wrote:
> > @@ -514,7 +556,20 @@ static void update_dl_entity(struct
> > sched_dl_entity *dl_se, struct dl_rq *dl_rq = dl_rq_of_se(dl_se);
> > struct rq *rq = rq_of_dl_rq(dl_rq);
> >
> > + if (hrtimer_is_queued(&dl_se->inactive_timer)) {
> > + hrtimer_try_to_cancel(&dl_se->inactive_timer);
> > + WARN_ON(dl_task_of(dl_se)->nr_cpus_allowed > 1);
>
> Isn't that always so? That is, DL tasks cannot be but 'global', right?
Well, if I understand well in general (that is, if admission control is
enabled) nr_cpus_allowed is equal to the number of CPUs in the
cpuset...
This is generally > 1 (and in this case select_task_rq_dl() is invoked
first, and tries to cancel the timer - so I think the timer cannot be
queued), or can be = 1 if we do partitioned scheduling (cpusets
containing only 1 CPU, or disabled admission control). If
nr_cpus_allowed is 1, then select_task_rq_dl() is not invoked, so the
timer can be queued.
In some of my tests I used partitioned scheduling; in some other tests
I disabled admission control to mix tasks with different affinities, so
I made the warning conditional to the number of CPUs being > 1.
> Also, you could use the return value of hrtimer_try_to_cancel() to
> determine hrtimer_is_queued() I suppose.
Ah, ok... I was under the impression that
"if (hrtimer_is_queued()) hrtimer_try_to_cancel()"
is less overhead than a simple "hrtimer_try_to_cancel()", but this was
just an uneducated guess... I'll change the code to avoid the check on
hrtimer_is_queued().
>
> > + } else {
> > + /*
> > + * The "inactive timer" has been cancelled in
> > + * select_task_rq_dl() (and the acvive utilisation
> > has
> > + * been decreased). So, increase the active
> > utilisation.
> > + * If select_task_rq_dl() could not cancel the
> > timer,
> > + * inactive_task_timer() will * find the task
> > state as
> ^^^
> superfluous '*'?
Yes, sorry... Something went wrong when I re-indented the comment :(
Thanks,
Luca
next prev parent reply other threads:[~2016-11-18 15:56 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-24 14:06 [RFC v3 0/6] CPU reclaiming for SCHED_DEADLINE Luca Abeni
2016-10-24 14:06 ` [RFC v3 1/6] Track the active utilisation Luca Abeni
2016-10-25 9:09 ` Daniel Bristot de Oliveira
2016-10-25 9:29 ` luca abeni
2016-10-25 13:58 ` Steven Rostedt
2016-10-25 18:04 ` Luca Abeni
2016-11-18 14:23 ` Peter Zijlstra
2016-11-18 15:10 ` luca abeni
2016-11-18 15:28 ` Peter Zijlstra
2016-11-18 16:42 ` Steven Rostedt
2016-12-05 22:30 ` luca abeni
2016-12-06 8:35 ` Peter Zijlstra
2016-12-06 8:57 ` luca abeni
2016-12-06 13:47 ` luca abeni
2016-11-01 16:45 ` Juri Lelli
2016-11-01 21:10 ` luca abeni
2016-11-08 17:56 ` Juri Lelli
2016-11-08 18:17 ` Luca Abeni
2016-11-08 18:53 ` Juri Lelli
2016-11-08 19:09 ` Luca Abeni
2016-11-08 20:02 ` Juri Lelli
2016-11-09 15:25 ` luca abeni
2016-11-09 16:29 ` luca abeni
2016-11-18 14:55 ` Peter Zijlstra
2016-11-18 13:55 ` Peter Zijlstra
2016-11-18 15:06 ` luca abeni
2016-10-24 14:06 ` [RFC v3 2/6] Improve the tracking of " Luca Abeni
2016-11-01 16:46 ` Juri Lelli
2016-11-01 21:46 ` luca abeni
2016-11-02 2:35 ` luca abeni
2016-11-10 10:04 ` Juri Lelli
2016-11-10 11:56 ` Juri Lelli
2016-11-10 12:15 ` luca abeni
2016-11-10 12:34 ` Juri Lelli
2016-11-10 12:45 ` luca abeni
2016-11-02 2:41 ` luca abeni
2016-11-18 15:36 ` Peter Zijlstra
2016-11-18 15:56 ` luca abeni [this message]
2016-11-18 15:47 ` Peter Zijlstra
2016-11-18 16:06 ` luca abeni
2016-11-18 18:49 ` Peter Zijlstra
2016-10-24 14:06 ` [RFC v3 3/6] Fix the update of the total -deadline utilization Luca Abeni
2016-10-24 14:06 ` [RFC v3 4/6] GRUB accounting Luca Abeni
2016-10-24 14:06 ` [RFC v3 5/6] Do not reclaim the whole CPU bandwidth Luca Abeni
2016-10-24 14:06 ` [RFC v3 6/6] Make GRUB a task's flag Luca Abeni
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20161118165647.46f8c867@sweethome \
--to=luca.abeni@unitn.it \
--cc=claudio@evidence.eu.com \
--cc=juri.lelli@arm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).