Received-SPF: pass (google.com: domain of linux-kernel+bounces-80754-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Message-ID: <2dbc4a0faaef7f4b7d1cd7d46c75e30fb563b227.camel@gmail.com>
Subject: Re: [PATCH v5 5/5] iio: amplifiers: hmc425a: add support for
 LTC6373 Instrumentation Amplifier
From: Nuno =?ISO-8859-1?Q?S=E1?= <noname.nuno@gmail.com>
To: Jonathan Cameron <jic23@kernel.org>
Cc: Dumitru Ceclan <mitrutzceclan@gmail.com>, Lars-Peter Clausen
	 <lars@metafoo.de>, Rob Herring <robh+dt@kernel.org>, Krzysztof Kozlowski
	 <krzysztof.kozlowski+dt@linaro.org>, Conor Dooley <conor+dt@kernel.org>, 
	linux-iio@vger.kernel.org, devicetree@vger.kernel.org, 
	linux-kernel@vger.kernel.org, Ceclan Dumitru <dumitru.ceclan@analog.com>
Date: Mon, 26 Feb 2024 09:25:59 +0100
In-Reply-To: <20240224175448.484ccd10@jic23-huawei>
References: <20240220153553.2432-1-mitrutzceclan@gmail.com>
	 <20240220153553.2432-6-mitrutzceclan@gmail.com>
	 <c92c4c71f433c7375588e832d3c084a50190a1b5.camel@gmail.com>
	 <20240224175448.484ccd10@jic23-huawei>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.50.4 (3.50.4-1.fc39) 
Precedence: bulk
MIME-Version: 1.0

On Sat, 2024-02-24 at 17:54 +0000, Jonathan Cameron wrote:
> On Wed, 21 Feb 2024 14:23:51 +0100
> Nuno S=C3=A1 <noname.nuno@gmail.com> wrote:
>=20
> > On Tue, 2024-02-20 at 17:34 +0200, Dumitru Ceclan wrote:
> > > This adds support for LTC6373 36 V Fully-Differential Programmable-Ga=
in
> > > Instrumentation Amplifier with 25 pA Input Bias Current.
> > > The user can program the gain to one of seven available settings thro=
ugh
> > > a 3-bit parallel interface (A2 to A0).
> > >=20
> > > Signed-off-by: Dumitru Ceclan <mitrutzceclan@gmail.com>
> > > ---=C2=A0=20
> >=20
> > Just one minor comment. With that:
> >=20
> > Reviewed-by: Nuno Sa <nuno.sa@analog.com>
> >=20
> > > =C2=A0drivers/iio/amplifiers/hmc425a.c | 124 ++++++++++++++++++++++++=
++++++-
> > > =C2=A01 file changed, 120 insertions(+), 4 deletions(-)
> > >=20
> > > diff --git a/drivers/iio/amplifiers/hmc425a.c
> > > b/drivers/iio/amplifiers/hmc425a.c
> > > index 77872e2dfdfe..50c86c2d28d7 100644
> > > --- a/drivers/iio/amplifiers/hmc425a.c
> > > +++ b/drivers/iio/amplifiers/hmc425a.c
> > > @@ -2,9 +2,10 @@
> > > =C2=A0/*
> > > =C2=A0 * HMC425A and similar Gain Amplifiers
> > > =C2=A0 *
> > > - * Copyright 2020 Analog Devices Inc.
> > > + * Copyright 2020, 2024 Analog Devices Inc.
> > > =C2=A0 */=C2=A0=20
> >=20
> > ...
> >=20
> > >=20
> > > =C2=A0
> > > +static ssize_t ltc6373_read_powerdown(struct iio_dev *indio_dev,
> > > +				=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 uintptr_t private,
> > > +				=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 const struct iio_chan_spec *chan,
> > > +				=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 char *buf)
> > > +{
> > > +	struct hmc425a_state *st =3D iio_priv(indio_dev);
> > > +
> > > +	return sysfs_emit(buf, "%d\n", st->powerdown);=C2=A0=20
> >=20
> > Well, in theory the read should also be protected with the lock...
>=20
> Only reason I can think of for that is potential read tearing.
> If that happens on a bool we are going to be in a mess so I think this
> is in practice fine without, though paranoia might suggest locking.

Yeah, also mentioned it for correctness. I mean, in theory, read_once() sho=
uld be
more that enough in here but I often find that too much for using in "simpl=
e" drivers
where a lock is surely easier to understand for someone reading the code.

Now, about your bool comment, I used to think like that until I saw the LF =
rcu
mentorship video from Paul. I'm fairly sure he comes up with some "crazy" p=
ossibility
about the CPU/compiler screwing you even on a char (IIRC, he was also argui=
ng about
not using read_once() on a bool).

Now, practically speaking, I tend to agree that for the archs we care this =
will
likely never be an issue but yeah, not 100% correct kernel code IMO.

- Nuno S=C3=A1