[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YxHClIhy1RPLT8P4@hirez.programming.kicks-ass.net>
Date: Fri, 2 Sep 2022 10:45:08 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Dietmar Eggemann <dietmar.eggemann@....com>
Cc: Ingo Molnar <mingo@...nel.org>,
Vincent Guittot <vincent.guittot@...aro.org>,
Morten Rasmussen <morten.rasmussen@....com>,
Vincent Donnefort <vdonnefort@...gle.com>,
Quentin Perret <qperret@...gle.com>,
Patrick Bellasi <patrick.bellasi@...bug.net>,
Abhijeet Dharmapurikar <adharmap@...cinc.com>,
Jian-Min <Jian-Min.Liu@...iatek.com>,
Qais Yousef <qais.yousef@....com>, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH 1/1] sched/pelt: Introduce PELT multiplier
On Fri, Sep 02, 2022 at 09:53:22AM +0200, Dietmar Eggemann wrote:
> > void update_rq_clock(struct rq *rq)
> > --- a/kernel/sched/features.h
> > +++ b/kernel/sched/features.h
> > @@ -101,3 +101,6 @@ SCHED_FEAT(LATENCY_WARN, false)
> >
> > SCHED_FEAT(ALT_PERIOD, true)
> > SCHED_FEAT(BASE_SLICE, true)
> > +
> > +SCHED_FEAT(PELT_M2, false)
> > +SCHED_FEAT(PELT_M4, false)
>
> The sched features interface would definitely be much less official but
> I think it's not useful for Android since they don't mount debugfs anymore.
Well, they can haz a small patch that sets either one or both to true,
no?
Powered by blists - more mailing lists