[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7hk1zvx36t.fsf@baylibre.com>
Date: Mon, 16 Oct 2017 03:59:54 +0200
From: Kevin Hilman <khilman@...libre.com>
To: Viresh Kumar <viresh.kumar@...aro.org>
Cc: Rafael Wysocki <rjw@...ysocki.net>, ulf.hansson@...aro.org,
linux-pm@...r.kernel.org,
Vincent Guittot <vincent.guittot@...aro.org>,
Stephen Boyd <sboyd@...eaurora.org>,
Nishanth Menon <nm@...com>, robh+dt@...nel.org,
lina.iyer@...aro.org, rnayak@...eaurora.org, sudeep.holla@....com,
linux-kernel@...r.kernel.org, Len Brown <len.brown@...el.com>,
Pavel Machek <pavel@....cz>,
Andy Gross <andy.gross@...aro.org>,
David Brown <david.brown@...aro.org>
Subject: Re: [PATCH V13 1/7] PM / Domains: Add support to select performance-state of domains
Viresh Kumar <viresh.kumar@...aro.org> writes:
> Some platforms have the capability to configure the performance state of
> PM domains. This patch enhances the genpd core to support such
> platforms.
>
> The performance levels (within the genpd core) are identified by
> positive integer values, a lower value represents lower performance
> state.
>
> This patch adds a new genpd API, which is called by user drivers (like
> OPP framework):
>
> - int dev_pm_genpd_set_performance_state(struct device *dev,
> unsigned int state);
>
> This updates the performance state constraint of the device on its PM
> domain. On success, the genpd will have its performance state set to a
> value which is >= "state" passed to this routine. The genpd core calls
> the genpd->set_performance_state() callback, if implemented,
> else -ENODEV is returned to the caller.
>
> The PM domain drivers need to implement the following callback if they
> want to support performance states.
>
> - int (*set_performance_state)(struct generic_pm_domain *genpd,
> unsigned int state);
>
> This is called internally by the genpd core on several occasions. The
> genpd core passes the genpd pointer and the aggregate of the
> performance states of the devices supported by that genpd to this
> callback. This callback must update the performance state of the genpd
> (in a platform dependent way).
>
> The power domains can avoid supplying above callback, if they don't
> support setting performance-states.
>
> Currently we aren't propagating performance state changes of a subdomain
> to its masters as we don't have hardware that needs it right now. Over
> that, the performance states of subdomain and its masters may not have
> one-to-one mapping and would require additional information. We can get
> back to this once we have hardware that needs it.
>
> Tested-by: Rajendra Nayak <rnayak@...eaurora.org>
> Signed-off-by: Viresh Kumar <viresh.kumar@...aro.org>
OK, I really like the OPP-related changes suggested by Ulf, and you've
removed a lot of the complexity that made it a bit confusing to follow.
It's definitley cleaned up and much easier to follow.
Thanks for your persistence. This is definitely a needed feature.
I have some usecases in mind where the performance state might need to
be selected based on OPP voltage, but that's now a change that can be
added later when that feature is needed.
Reviewed-by: Kevin Hilman <khilman@...libre.com>
Powered by blists - more mailing lists