DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 88751208E7
MIME-Version: 1.0
References: <20181006173507.7316-1-nbd@nbd.name> <20181006173507.7316-9-nbd@nbd.name>
 <CAA93jw48KRNTcGy_RecY8u-1wqL9uWzeP4JK4-pgf9LV9ihzDQ@mail.gmail.com> <43da677d-60b8-b940-4d89-d27578f9c20b@nbd.name>
In-Reply-To: <43da677d-60b8-b940-4d89-d27578f9c20b@nbd.name>
From:   Dave Taht <dave.taht@gmail.com>
Date:   Sat, 6 Oct 2018 11:23:34 -0700
Message-ID: <CAA93jw5T40FP0O8dHbS5r2QLpYVDbF9B6NHrc48jceBubENC6w@mail.gmail.com>
Subject: Re: [PATCH 9/9] mac80211: rc80211_minstrel: remove variance / stddev calculation
To:     Felix Fietkau <nbd@nbd.name>
Cc:     linux-wireless <linux-wireless@vger.kernel.org>,
        Johannes Berg <johannes@sipsolutions.net>,
        Andrew McGregor <andrewmcgr@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On Sat, Oct 6, 2018 at 11:18 AM Felix Fietkau <nbd@nbd.name> wrote:
>
> On 2018-10-06 19:59, Dave Taht wrote:
> > On Sat, Oct 6, 2018 at 10:37 AM Felix Fietkau <nbd@nbd.name> wrote:
> >>
> >> When there are few packets (e.g. for sampling attempts), the exponenti=
ally
> >> weighted variance is usually vastly overestimated, making the resultin=
g data
> >> essentially useless. As far as I know, there has not been any practica=
l use
> >> for this, so let's not waste any cycles on it.
> >>
> >> Signed-off-by: Felix Fietkau <nbd@nbd.name>
> >> ---
> >>  net/mac80211/rc80211_minstrel.c            |  6 -----
> >>  net/mac80211/rc80211_minstrel.h            | 26 +--------------------=
-
> >>  net/mac80211/rc80211_minstrel_debugfs.c    | 14 ++++--------
> >>  net/mac80211/rc80211_minstrel_ht_debugfs.c | 14 ++++--------
> >>  4 files changed, 9 insertions(+), 51 deletions(-)
> >>
> >> diff --git a/net/mac80211/rc80211_minstrel.c b/net/mac80211/rc80211_mi=
nstrel.c
> >> index dead57ba9eac..a34e9c2ca626 100644
> >> --- a/net/mac80211/rc80211_minstrel.c
> >> +++ b/net/mac80211/rc80211_minstrel.c
> >> @@ -167,12 +167,6 @@ minstrel_calc_rate_stats(struct minstrel_rate_sta=
ts *mrs)
> >>                 if (unlikely(!mrs->att_hist)) {
> >>                         mrs->prob_ewma =3D cur_prob;
> >>                 } else {
> >> -                       /* update exponential weighted moving variance=
 */
> >> -                       mrs->prob_ewmv =3D minstrel_ewmv(mrs->prob_ewm=
v,
> >> -                                                       cur_prob,
> >> -                                                       mrs->prob_ewma=
,
> >> -                                                       EWMA_LEVEL);
> >> -
> >>                         /*update exponential weighted moving avarage *=
/
> >>                         mrs->prob_ewma =3D minstrel_ewma(mrs->prob_ewm=
a,
> >>                                                        cur_prob,
> >> diff --git a/net/mac80211/rc80211_minstrel.h b/net/mac80211/rc80211_mi=
nstrel.h
> >> index 54b2b2c3e10a..23ec953e3a24 100644
> >> --- a/net/mac80211/rc80211_minstrel.h
> >> +++ b/net/mac80211/rc80211_minstrel.h
> >> @@ -35,19 +35,6 @@ minstrel_ewma(int old, int new, int weight)
> >>         return old + incr;
> >>  }
> >>
> >> -/*
> >> - * Perform EWMV (Exponentially Weighted Moving Variance) calculation
> >> - */
> >
> > I worry about this one. where are you getting your proof from?
> I've done quite a few measurements myself to see if this can be usable
> for further rate control improvements or for the upcoming TPC work.
> The data this generates simply fluctuates wildly and incoherently based
> on the sampling behavior, making it completely useless.
> Together with Thomas (who introduced this code), I tried a few times to
> fix this, but couldn't find any way to make it coherent and usable.
>
> Thomas and I both agreed that it's better to just remove it until
> somebody has a better idea what to do.
>
> Also, this was only used for debugfs statistics, not for any actual rate
> control behavior.

OK, thanks. I'm totally delighted to see this patchset otherwise.

> - Felix


--=20

Dave T=C3=A4ht
CTO, TekLibre, LLC
http://www.teklibre.com
Tel: 1-831-205-9740