Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:3 as permitted sender) client-ip=2620:137:e000::3:3;
Message-ID: <9b11a42d83dfe77215bafae1d116375ee2398ae6.camel@gmail.com>
Subject: Re: [PATCH 04/12] iio: adc: ad9467: fix reset gpio handling
From:   Nuno =?ISO-8859-1?Q?S=E1?= <noname.nuno@gmail.com>
To:     Jonathan Cameron <jic23@kernel.org>
Cc:     David Lechner <dlechner@baylibre.com>, nuno.sa@analog.com,
        linux-kernel@vger.kernel.org, devicetree@vger.kernel.org,
        linux-iio@vger.kernel.org,
        Olivier MOYSAN <olivier.moysan@foss.st.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Rob Herring <robh+dt@kernel.org>,
        Frank Rowand <frowand.list@gmail.com>,
        Lars-Peter Clausen <lars@metafoo.de>,
        Michael Hennerich <Michael.Hennerich@analog.com>
Date:   Mon, 04 Dec 2023 17:41:48 +0100
In-Reply-To: <20231204151514.4e2c8ada@jic23-huawei>
References: <20231121-dev-iio-backend-v1-0-6a3d542eba35@analog.com>
         <20231121-dev-iio-backend-v1-4-6a3d542eba35@analog.com>
         <CAMknhBGCqnzCp6vQ+59Z-SybScvbtU7aWdAD6KnP1e6=q60gVQ@mail.gmail.com>
         <d534c3323c32d4ed2aedae19a9f101be90ef0cc7.camel@gmail.com>
         <CAMknhBGjm2ja9HOenOWi9O5Ao8qUg=gT=_Vz8CyxQ=pfNX2EJQ@mail.gmail.com>
         <3925cb4b6453644c889675c20329b3477a06fcd5.camel@gmail.com>
         <20231204151514.4e2c8ada@jic23-huawei>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
MIME-Version: 1.0
Precedence: bulk

On Mon, 2023-12-04 at 15:15 +0000, Jonathan Cameron wrote:
> On Sat, 02 Dec 2023 09:36:47 +0100
> Nuno S=C3=A1 <noname.nuno@gmail.com> wrote:
>=20
> > On Fri, 2023-12-01 at 11:01 -0600, David Lechner wrote:
> > > On Fri, Dec 1, 2023 at 2:47=E2=80=AFAM Nuno S=C3=A1 <noname.nuno@gmai=
l.com> wrote:=C2=A0=20
> > > >=20
> > > > On Thu, 2023-11-30 at 15:41 -0600, David Lechner wrote:=C2=A0=20
> > > > > On Tue, Nov 21, 2023 at 4:17=E2=80=AFAM Nuno Sa via B4 Relay
> > > > > <devnull+nuno.sa.analog.com@kernel.org> wrote:=C2=A0=20
> > > > > >=20
> > > > > > From: Nuno Sa <nuno.sa@analog.com>
> > > > > >=20
> > > > > > The reset gpio was being requested with GPIOD_OUT_LOW which mea=
ns, not
> > > > > > asserted. Then it was being asserted but never de-asserted whic=
h means
> > > > > > the devices was left in reset. Fix it by de-asserting the gpio.=
=C2=A0=20
> > > > >=20
> > > > > It could be helpful to update the devicetree bindings to state th=
e
> > > > > expected active-high or active-low setting for this gpio so it is
> > > > > clear which state means asserted.
> > > > > =C2=A0=20
> > > >=20
> > > > You could state that the chip is active low but I don't see that ch=
ange that
> > > > important for now. Not sure if this is clear and maybe that's why y=
our
> > > > comment.
> > > > GPIOD_OUT_HIGH has nothing to do with active high or low. It just m=
eans, "get
> > > > me
> > > > the
> > > > pin in the asserted state".
> > > > =C2=A0=20
> > >=20
> > > I would assume that this bug happened in the first place because
> > > someone forgot GPIOD_OUT_LOW in the devicetree when they were
> > > developing the driver. So this is why I suggested that updating the
> > > devicetree binding docs so that future users are less likely to make
> > > the same mistake. Currently, the bindings don't even have reset-gpios
> > > in the examples.=C2=A0=20
> >=20
> > Hmm, I think you're missing the point... The bug has nothing to do with
> > devicetree.
> > This is what was happening:
> >=20
> > 1) We were calling devm_gpiod_get_optional() with GPIOD_OUT_LOW. What t=
his means
> > is
> > that you get an output gpio deasserted. Hence the device is out of rese=
t. And
> > here is
> > the important part... what you have in dts does not matter. If you have=
 active
> > low,
> > it means the pin level will be 1. If you have high, the pin level is 0.=
 And this
> > is
> > all handled by gpiolib for you.=20
> >=20
> > 2) Then, we called gpiod_direction_output(..., 1), which means set the =
direction
> > out
> > (which is actually not needed since it was already done when getting th=
e pin) and
> > assert the pin. Hence, reset the device. And we were never de-asserting=
 the pin
> > so
> > the device would be left in reset.
>=20
> Functionally I believe David is correct.=C2=A0=C2=A0 Flipping the DT woul=
d 'fix' this.
> It's all down to a nreset vs reset pin description.
>=20

Ahh I see. Well would not really be a fix :)

- Nuno S=C3=A1