Received-SPF: pass (google.com: domain of linux-kernel+bounces-119968-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Precedence: bulk
MIME-Version: 1.0
References: <20240325153356.46112-1-l.rubusch@gmail.com> <20240325153356.46112-2-l.rubusch@gmail.com>
 <20240325203155.23ddfe3d@jic23-huawei>
In-Reply-To: <20240325203155.23ddfe3d@jic23-huawei>
From: Lothar Rubusch <l.rubusch@gmail.com>
Date: Tue, 26 Mar 2024 21:59:34 +0100
Message-ID: <CAFXKEHaQLu9WFJe4r4+QaWO-wjM0hpYkWF_s8NOSh2Hoo5w8FQ@mail.gmail.com>
Subject: Re: [PATCH v4 1/7] iio: accel: adxl345: Make data_range obsolete
To: Jonathan Cameron <jic23@kernel.org>
Cc: lars@metafoo.de, Michael.Hennerich@analog.com, robh+dt@kernel.org, 
	krzysztof.kozlowski+dt@linaro.org, conor+dt@kernel.org, 
	linux-iio@vger.kernel.org, devicetree@vger.kernel.org, 
	linux-kernel@vger.kernel.org, eraretuya@gmail.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, Mar 25, 2024 at 9:32=E2=80=AFPM Jonathan Cameron <jic23@kernel.org>=
 wrote:
>
> On Mon, 25 Mar 2024 15:33:50 +0000
> Lothar Rubusch <l.rubusch@gmail.com> wrote:
>
> > Replace write() data_format by regmap_update_bits(), because
> > bus specific pre-configuration may have happened before on
> > the same register. Changes then need to be masked.
> >
> > Remove the data_range field from the struct adxl345_data,
> > because it is not used anymore.
> >
> > Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>
> > ---
> >  drivers/iio/accel/adxl345_core.c | 9 ++++-----
> >  1 file changed, 4 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/iio/accel/adxl345_core.c b/drivers/iio/accel/adxl3=
45_core.c
> > index 8bd30a23e..be6758015 100644
> > --- a/drivers/iio/accel/adxl345_core.c
> > +++ b/drivers/iio/accel/adxl345_core.c
> > @@ -42,13 +42,13 @@
> >  #define ADXL345_DATA_FORMAT_4G               1
> >  #define ADXL345_DATA_FORMAT_8G               2
> >  #define ADXL345_DATA_FORMAT_16G              3
> > +#define ADXL345_DATA_FORMAT_MSK              ~((u8) BIT(6)) /* ignore =
spi-3wire */
>
> I'm not keen on seeing masking of a bit we don't yet
> handle done by value.  Can we instead build this up by what we 'want' to
> write rather than don't. Will need a few more defines perhaps to cover
> the masks of SELF_TEST, INT_INVERT, FULL_RES, Justify and Range.
>

Good point. Anyway, there is also an input driver implementation for
the adxl345, mainly dealing with the interrupt feature as input
device. Thus, for the iio implementation I would suggest to reduce the
mask just to cover SELF_TEST and FULL_RES and leave INT_INVERT out. Is
this ok?

> >
> >  #define ADXL345_DEVID                        0xE5
> >
> >  struct adxl345_data {
> >       const struct adxl345_chip_info *info;
> >       struct regmap *regmap;
> > -     u8 data_range;
> >  };
> >
> >  #define ADXL345_CHANNEL(index, axis) {                                =
       \
> > @@ -219,14 +219,13 @@ int adxl345_core_probe(struct device *dev, struct=
 regmap *regmap)
> >       data =3D iio_priv(indio_dev);
> >       data->regmap =3D regmap;
> >       /* Enable full-resolution mode */
> > -     data->data_range =3D ADXL345_DATA_FORMAT_FULL_RES;
> >       data->info =3D device_get_match_data(dev);
> >       if (!data->info)
> >               return -ENODEV;
> >
> > -     ret =3D regmap_write(data->regmap, ADXL345_REG_DATA_FORMAT,
> > -                        data->data_range);
> > -     if (ret < 0)
> > +     ret =3D regmap_update_bits(regmap, ADXL345_REG_DATA_FORMAT,
> > +                              ADXL345_DATA_FORMAT_MSK, ADXL345_DATA_FO=
RMAT_FULL_RES);
> > +     if (ret)
> >               return dev_err_probe(dev, ret, "Failed to set data range\=
n");
> >
> >       indio_dev->name =3D data->info->name;
>