Re: [RFC v5 2/9] sched/deadline: improve the tracking of active utilization

From: Juri Lelli <juri.lelli@arm.com>
To: luca abeni <luca.abeni@santannapisa.it>
Cc: Peter Zijlstra <peterz@infradead.org>,
	linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
	Claudio Scordino <claudio@evidence.eu.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	Tommaso Cucinotta <tommaso.cucinotta@sssup.it>,
	Daniel Bristot de Oliveira <bristot@redhat.com>,
	Joel Fernandes <joelaf@google.com>,
	Mathieu Poirier <mathieu.poirier@linaro.org>
Subject: Re: [RFC v5 2/9] sched/deadline: improve the tracking of active utilization
Date: Mon, 27 Mar 2017 08:17:45 +0100	[thread overview]
Message-ID: <20170327071745.GA10289@e106622-lin> (raw)
In-Reply-To: <20170324224715.4098dbfb@nowhere>

On 24/03/17 22:47, Luca Abeni wrote:
> Hi Peter,
> 
> On Fri, 24 Mar 2017 14:20:41 +0100
> Peter Zijlstra <peterz@infradead.org> wrote:
> 
> > On Fri, Mar 24, 2017 at 04:52:55AM +0100, luca abeni wrote:
> > 

[...]

> > 
> > In general I feel it would be nice to have a state diagram included
> > somewhere near these two functions. It would be nice to not have to
> > dig out the PDF every time.
> 
> Ok... Since I am not good at ascii art, would it be ok to add a textual
> description? If yes, I'll add a comment like:
> "
> The utilization of a task is added to the runqueue's active utilization
> when the task becomes active (is enqueued in the runqueue), and is

Is enqueued for the first time on a new period, maybe? It seems to be
contradictory w.r.t. what below (if wakeup before 0 lag time) otherwise.

> removed when the task becomes inactive. A task does not become
> immediately inactive when it blocks, but becomes inactive at the so
> called "0 lag time"; so, we setup the "inactive timer" to fire at the
> "0 lag time". When the "inactive timer" fires, the task utilization is
> removed from the runqueue's active utilization. If the task wakes up
> again on the same runqueue before the "0 lag time", the active
> utilization must not be changed and the "inactive timer" must be
> cancelled. If the task wakes up again on a different runqueue before
> the "0 lag time", then the task's utilization must be removed from the
> previous runqueue's active utilization and must be added to the new
> runqueue's active utilization.
> In order to avoid races between a task waking up on a runqueue while the
> "inactive timer" is running on a different CPU, the "dl_non_contending"
> flag is used to indicate that a task is not on a runqueue but is active
> (so, the flag is set when the task blocks and is cleared when the
> "inactive timer" fires or when the task  wakes up).
> "
> (if this is ok, where can I add this comment?)
> 

Thanks for this Luca. Not sure it adds much to your text above, but we
might want to consider adding something like below?

--->8---
           1st enqueue       +------------------+
                             |                  |
           +---------------->+ ACTIVEcontending |
           |                 |                  |
           |                 +----+------+------+
           |                      |      ^
           |                      |      |
  +--------+-------+              |      |
  |                |     dequeue  |      |  wakeup before
  |    INACTIVE    |              |      |  0 lag time
  |                |              |      |
  +--------+-------+              |      |
           ^                      |      |
           |                      V      |
           |                 +----+------+------+
           |                 |                  |
           +-----------------+ ACTIVEnonCONTEND |
                             |                  |
            0 lag time       +------------------+
            elapsed
--->8---

Thanks,

- Juri