Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
User-Agent: Cyrus-JMAP/3.5.0-alpha0-141-gf094924a34-fm-20210210.001-gf094924a
Mime-Version: 1.0
Message-Id: <1ad78446-4a40-4c3e-8680-6dbf19616515@www.fastmail.com>
In-Reply-To: <20210211195559.n2j4jnchl2ho54mg@pali>
References: <20210209225630.mdwnzkvnaz3r4blt@pali>
 <a86c5069-d423-44db-92dd-b3e406b7ec91@www.fastmail.com>
 <20210210002619.43104a9b@kernel.org>
 <ac03801e-87e2-4e57-b131-bff52f03579d@www.fastmail.com>
 <1cd0c2ee-aa3a-4da2-9c0c-57cc5a1dad49@www.fastmail.com>
 <a1277b1f-f829-4d02-9e54-68ab4faaa047@www.fastmail.com>
 <20210210092339.qy6wwuq6qr5m2ozr@pali>
 <d6971325-af71-4f71-91c2-7b661804c022@www.fastmail.com>
 <20210210180322.rlfxdussqhejqpo6@pali>
 <966f50f2-68b2-4d4f-85f0-396df112c0f4@www.fastmail.com>
 <20210211195559.n2j4jnchl2ho54mg@pali>
Date:   Thu, 11 Feb 2021 12:22:52 -0800
From:   nnet <nnet@fastmail.fm>
To:     =?UTF-8?Q?Pali_Roh=C3=A1r?= <pali@kernel.org>
Cc:     =?UTF-8?Q?Marek_Beh=C3=BAn?= <kabel@kernel.org>,
        a.heider@gmail.com, andrew@lunn.ch, gerald@gk2.net,
        gregory.clement@bootlin.com, kostap@marvell.com,
        linux-arm-kernel@lists.infradead.org, linux-clk@vger.kernel.org,
        linux-kernel@vger.kernel.org, luka.perkov@sartura.hr,
        miquel.raynal@bootlin.com, mturquette@baylibre.com,
        rmk+kernel@armlinux.org.uk, sboyd@kernel.org, tmn505@gmail.com,
        vladimir.vid@sartura.hr
Subject: =?UTF-8?Q?Re:_[PATCH_mvebu_v2_00/10]_Armada_37xx:_Fix_cpufreq_changing_b?=
 =?UTF-8?Q?ase_CPU_speed_to_800_MHz_from_1000_MHz?=
Content-Type: text/plain;charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Feb 11, 2021, at 11:55 AM, Pali Roh=C3=A1r wrote:
> On Wednesday 10 February 2021 11:08:59 nnet wrote:
> > On Wed, Feb 10, 2021, at 10:03 AM, Pali Roh=C3=A1r wrote:
> > > > > Hello! Could you please enable userspace governor during kerne=
l
> > > > > compilation?
> > > > >=20
> > > > >     CONFIG_CPU_FREQ_GOV_USERSPACE=3Dy
> > > > >=20
> > > > > It can be activated via command:
> > > > >=20
> > > > >     echo userspace > /sys/devices/system/cpu/cpufreq/policy0/s=
caling_governor
> > > > >=20
> > > > > After that you can "force" CPU frequency to specific value, e.=
g.:
> > > > >=20
> > > > >     echo 1000000 > /sys/devices/system/cpu/cpufreq/policy0/sca=
ling_setspeed
> > > > >=20
> > > > > I need to know which switch (from --> to freq) cause this syst=
em hang.
> > > > >=20
> > > > > This patch series (via MIN_VOLT_MV_FOR_L0_L1_1GHZ) is fixing o=
nly
> > > > > switching from 500 MHz to 1000 MHz on 1 GHz variant. As only t=
his switch
> > > > > is causing issue.
> > > > >=20
> > > > > I have used following simple bash script to check that switchi=
ng between
> > > > > 500 MHz and 1 GHz is stable:
> > > > >=20
> > > > >     while true; do
> > > > >         echo 1000000 > /sys/devices/system/cpu/cpufreq/policy0=
/scaling_setspeed;
> > > > >         echo 500000 > /sys/devices/system/cpu/cpufreq/policy0/=
scaling_setspeed;
> > > > >         echo 1000000 > /sys/devices/system/cpu/cpufreq/policy0=
/scaling_setspeed;
> > > > >         echo 500000 > /sys/devices/system/cpu/cpufreq/policy0/=
scaling_setspeed;
> > > > >     done
> > > >=20
> > > > echo userspace | tee /sys/devices/system/cpu/cpufreq/policy0/sca=
ling_governor
> > > > while true; do
> > > >   echo 1200000 | tee /sys/devices/system/cpu/cpufreq/policy0/sca=
ling_setspeed;
> > > >   echo 600000 | tee /sys/devices/system/cpu/cpufreq/policy0/scal=
ing_setspeed;
> > > > done
> > > >=20
> > > > >> +#define MIN_VOLT_MV_FOR_L0_L1_1GHZ 1108
> > > >=20
> > > > With 1108 I get a freeze within a minute. The last output to std=
out is 600000.
> > > >=20
> > > > With 1120 it takes a few minutes.
> > > >=20
> > > > With any of 1225, 1155, 1132 the device doesn't freeze over the =
full 5 minute load test.
> > > >=20
> > > > I'm using ondemand now with the above at 1132 without issue so f=
ar.
> > >=20
> > > Great, thank you for testing!
> > >=20
> > > Can you check if switching between any two lower frequencies 20000=
0
> > > 300000 600000 is stable?
> >=20
> > This is stable using 1132 mV for MIN_VOLT_MV_FOR_L0_L1_1GHZ:
> >=20
> > while true; do
> >   # down
> >   echo 1200000 | tee /sys/devices/system/cpu/cpufreq/policy0/scaling=
_setspeed;
> ...
>=20
> Hello!
>=20
> Could you please re-run test without tee, in form as I have shown abov=
e?
> UART is slow and printing something to console adds delay which decrea=
se
> probability that real issue is triggered as this is timing issue.

The test was done over SSH.

> Also please do tests just between two frequencies in loop as I observe=
d
> that switching between more decreased probability to hit issue.

> > > > echo userspace | tee /sys/devices/system/cpu/cpufreq/policy0/sca=
ling_governor
> > > > while true; do
> > > >   echo 1200000 | tee /sys/devices/system/cpu/cpufreq/policy0/sca=
ling_setspeed;
> > > >   echo 600000 | tee /sys/devices/system/cpu/cpufreq/policy0/scal=
ing_setspeed;
> > > > done

The first test ^ switched between 600 MHz and 1.2 GHz.

> The real issue for 1 GHz variant of A3720 is only when doing switch fr=
om
> 500 MHz to 1 GHz. So could you try to do some tests also without
> changing MIN_VOLT_MV_FOR_L0_L1_1GHZ and switching just between non-1.2=

> frequencies (to verify that on 1.2 GHz variant it is also from 600 MHz=

> to 1.2 GHz)?

With 1108 mV and switching between 600 MHz and 1.2GHz, I always saw a fr=
eeze within a minute.