| Message ID | BN6PR04MB066058A68D6471E7F6AFCFF7A3A20@BN6PR04MB0660.namprd04.prod.outlook.com (mailing list archive) |
|---|---|
| State | Mainlined |
| Commit | 754718a5b43c851546c3bb70e8d41bf81cb42b30 |
| Headers | show |
| Series | iio: adc: Add scaling support to exynos adc driver | expand |
On Fri, 8 May 2020 14:14:00 -0700 Jonathan Bakker <xc-racer2@live.ca> wrote: > Currently the driver only exposes the raw counts. As we > have the regulator voltage and the maximum value (stored in > the data mask), we can trivially produce a scaling fraction > of voltage / max value. > > This assumes that the regulator voltage is in fact the max > voltage, which appears to be the case for all mainline dts > and cross referenced with the public Exynos4412 and S5PV210 > datasheets. > > Signed-off-by: Jonathan Bakker <xc-racer2@live.ca> Seems reasonable to me. I'd like an exynos Ack though before applying. thanks, Jonathan > --- > drivers/iio/adc/exynos_adc.c | 14 +++++++++++++- > 1 file changed, 13 insertions(+), 1 deletion(-) > > diff --git a/drivers/iio/adc/exynos_adc.c b/drivers/iio/adc/exynos_adc.c > index 22131a677445..9d29b56805cd 100644 > --- a/drivers/iio/adc/exynos_adc.c > +++ b/drivers/iio/adc/exynos_adc.c > @@ -531,8 +531,19 @@ static int exynos_read_raw(struct iio_dev *indio_dev, > unsigned long timeout; > int ret; > > - if (mask != IIO_CHAN_INFO_RAW) > + if (mask == IIO_CHAN_INFO_SCALE) { > + ret = regulator_get_voltage(info->vdd); > + if (ret < 0) > + return ret; > + > + /* Regulator voltage is in uV, but need mV */ > + *val = ret / 1000; > + *val2 = info->data->mask; > + > + return IIO_VAL_FRACTIONAL; > + } else if (mask != IIO_CHAN_INFO_RAW) { > return -EINVAL; > + } > > mutex_lock(&indio_dev->mlock); > reinit_completion(&info->completion); > @@ -683,6 +694,7 @@ static const struct iio_info exynos_adc_iio_info = { > .channel = _index, \ > .address = _index, \ > .info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \ > + .info_mask_shared_by_all = BIT(IIO_CHAN_INFO_SCALE), \ > .datasheet_name = _id, \ > } >
On Sun, May 10, 2020 at 11:24:17AM +0100, Jonathan Cameron wrote: > On Fri, 8 May 2020 14:14:00 -0700 > Jonathan Bakker <xc-racer2@live.ca> wrote: > > > Currently the driver only exposes the raw counts. As we > > have the regulator voltage and the maximum value (stored in > > the data mask), we can trivially produce a scaling fraction > > of voltage / max value. > > > > This assumes that the regulator voltage is in fact the max > > voltage, which appears to be the case for all mainline dts > > and cross referenced with the public Exynos4412 and S5PV210 > > datasheets. > > > > Signed-off-by: Jonathan Bakker <xc-racer2@live.ca> > > Seems reasonable to me. I'd like an exynos Ack though before applying. It's correct, at least with ARMv7 Exynos datasheets The few ARMv8 Exynos chips are silent about the voltage levels. The Exynos 7 DTS board in mainline kernel does not have regulator but it looks clearly like mistake. I think they behave the same, so for Exynos: Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org> Best regards, Krzysztof > thanks, > > Jonathan > >
On Mon, 11 May 2020 09:42:32 +0200 Krzysztof Kozlowski <krzk@kernel.org> wrote: > On Sun, May 10, 2020 at 11:24:17AM +0100, Jonathan Cameron wrote: > > On Fri, 8 May 2020 14:14:00 -0700 > > Jonathan Bakker <xc-racer2@live.ca> wrote: > > > > > Currently the driver only exposes the raw counts. As we > > > have the regulator voltage and the maximum value (stored in > > > the data mask), we can trivially produce a scaling fraction > > > of voltage / max value. > > > > > > This assumes that the regulator voltage is in fact the max > > > voltage, which appears to be the case for all mainline dts > > > and cross referenced with the public Exynos4412 and S5PV210 > > > datasheets. > > > > > > Signed-off-by: Jonathan Bakker <xc-racer2@live.ca> > > > > Seems reasonable to me. I'd like an exynos Ack though before applying. > > > It's correct, at least with ARMv7 Exynos datasheets > > The few ARMv8 Exynos chips are silent about the voltage levels. The > Exynos 7 DTS board in mainline kernel does not have regulator but it > looks clearly like mistake. > > I think they behave the same, so for Exynos: > Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org> Applied to the togreg branch of iio.git and pushed out as testing for the autobuilders to poke at it. Thanks, Jonathan > > Best regards, > Krzysztof > > > thanks, > > > > Jonathan > > > >
diff --git a/drivers/iio/adc/exynos_adc.c b/drivers/iio/adc/exynos_adc.c index 22131a677445..9d29b56805cd 100644 --- a/drivers/iio/adc/exynos_adc.c +++ b/drivers/iio/adc/exynos_adc.c @@ -531,8 +531,19 @@ static int exynos_read_raw(struct iio_dev *indio_dev, unsigned long timeout; int ret; - if (mask != IIO_CHAN_INFO_RAW) + if (mask == IIO_CHAN_INFO_SCALE) { + ret = regulator_get_voltage(info->vdd); + if (ret < 0) + return ret; + + /* Regulator voltage is in uV, but need mV */ + *val = ret / 1000; + *val2 = info->data->mask; + + return IIO_VAL_FRACTIONAL; + } else if (mask != IIO_CHAN_INFO_RAW) { return -EINVAL; + } mutex_lock(&indio_dev->mlock); reinit_completion(&info->completion); @@ -683,6 +694,7 @@ static const struct iio_info exynos_adc_iio_info = { .channel = _index, \ .address = _index, \ .info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \ + .info_mask_shared_by_all = BIT(IIO_CHAN_INFO_SCALE), \ .datasheet_name = _id, \ }
Currently the driver only exposes the raw counts. As we have the regulator voltage and the maximum value (stored in the data mask), we can trivially produce a scaling fraction of voltage / max value. This assumes that the regulator voltage is in fact the max voltage, which appears to be the case for all mainline dts and cross referenced with the public Exynos4412 and S5PV210 datasheets. Signed-off-by: Jonathan Bakker <xc-racer2@live.ca> --- drivers/iio/adc/exynos_adc.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-)