Date: Tue, 21 Jul 2015 18:18:46 +0800
From: Boqun Feng <boqun.feng@gmail.com>
To: Yuyang Du <yuyang.du@intel.com>
Cc: mingo@kernel.org, peterz@infradead.org, linux-kernel@vger.kernel.org,
        pjt@google.com, bsegall@google.com, morten.rasmussen@arm.com,
        vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
        umgwanakikbuti@gmail.com, len.brown@intel.com,
        rafael.j.wysocki@intel.com, arjan@linux.intel.com,
        fengguang.wu@intel.com
Subject: Re: [PATCH v10 6/7] sched: Provide runnable_load_avg back to cfs_rq
Message-ID: <20150721101845.GB2882@fixme-laptop.cn.ibm.com>
References: <1436918682-4971-1-git-send-email-yuyang.du@intel.com>
 <1436918682-4971-7-git-send-email-yuyang.du@intel.com>
 <20150721010346.GA2882@fixme-laptop.cn.ibm.com>
 <20150721004400.GA28512@intel.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
	protocol="application/pgp-signature"; boundary="tKW2IUtsqtDRztdT"
Content-Disposition: inline
In-Reply-To: <20150721004400.GA28512@intel.com>
User-Agent: Mutt/1.5.23+102 (2ca89bed6448) (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3049
Lines: 88


--tKW2IUtsqtDRztdT
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Jul 21, 2015 at 08:44:01AM +0800, Yuyang Du wrote:
> On Tue, Jul 21, 2015 at 09:08:07AM +0800, Boqun Feng wrote:
> > Hi Yuyang,
> >=20
> > On Wed, Jul 15, 2015 at 08:04:41AM +0800, Yuyang Du wrote:
> > > The cfs_rq's load_avg is composed of runnable_load_avg and blocked_lo=
ad_avg.
> > > Before this series, sometimes the runnable_load_avg is used, and some=
times
> > > the load_avg is used. Completely replacing all uses of runnable_load_=
avg
> > > with load_avg may be too big a leap, i.e., the blocked_load_avg is co=
ncerned
> > > to result in overrated load. Therefore, we get runnable_load_avg back.
> > >=20
> > > The new cfs_rq's runnable_load_avg is improved to be updated with all=
 of the
> > > runnable sched_eneities at the same time, so the one sched_entity upd=
ated and
> > > the others stale problem is solved.
> > >=20
> >=20
> > How about tracking cfs_rq's blocked_load_avg instead of
> > runnable_load_avg, because, AFAICS:
> >=20
> > cfs_rq->runnable_load_avg =3D se->avg.load_avg - cfs_rq->blocked_load_a=
vg.
>=20
> No, cfs_rq->runnable_load_avg =3D cfs_rq->avg.load_avg - cfs_rq->blocked_=
load_avg,
> without rounding errors and the like.
> =20

Oh, sorry.. yeah, you're right here.

> > se is the corresponding sched_entity of cfs_rq. And when we need the
> > runnable_load_avg, we just calculate by the expression above.
> >=20
> > This can be thought as a lazy way to update runnable_load_avg, and we
> > don't need to modify __update_load_avg any more.
>=20
> Not lazy at all, but adding (as of now) useless blocked_load_avg and an
> extra subtraction.

but we can remove runnable_load_avg tracking code in __update_load_avg,
as you do in this patch, right?

> Or did you forget blocked_load_avg also needs to be updated/decayed as
> time elapses?

I know we need to update or decay the blocked_load_avg, but we only need
to update and decay when 1) entity dequeued/enqueued 2) entity migrated
or 3) we need the runnable_load_avg value calcuated by blocked_load_avg,
right?

These are more rare than __update_load_avg called, right?

Regards,
Boqun


--tKW2IUtsqtDRztdT
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAABCAAGBQJVrhyCAAoJEEl56MO1B/q4iy0H/3mIMg7A4XsSzczmy2f/2/yK
CBZ2Iuc6K8gCrUeT4y9YGXaB2ws5vPoRWeHSJuiWvn0FuQD+UEFsSuogTbu4QHKY
Akdm6sNsE4U+fpEUdBn/rMjmwGC2HiPNoZTuxT4dUO50vW2p4bmgaOOVpoItDdr4
NWEJVVfu/huw9xJkLC8a5Z2bFOy0iP1yg+2lzsw5Q7HDYIU2M5tQug3ehM0XtArC
bwAMh4YmRJEv5XwUir8JyPCnhxx6VEaa7XO0XOIy27kOy9n6IY4qRj+eXOGFPxMR
emt0VKmsj4/J2e0dZwQW6kWmFgo7bS78r3lEMefZd7CnMu44joAC01Ux6p2pz8k=
=P6DH
-----END PGP SIGNATURE-----

--tKW2IUtsqtDRztdT--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/