| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <aGBKMcZGYOcXmKdB@dev-linux>
Date: Sat, 28 Jun 2025 13:01:53 -0700
From: Sukrut Bellary <sbellary@...libre.com>
To: Jonathan Cameron <jic23@...nel.org>
Cc: David Lechner <dlechner@...libre.com>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley <conor+dt@...nel.org>,
Matti Vaittinen <mazziesaccount@...il.com>,
Nuno Sá <nuno.sa@...log.com>,
Andy Shevchenko <andy@...nel.org>,
Angelo Compagnucci <angelo.compagnucci@...il.com>,
Nishanth Menon <nm@...com>, linux-iio@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 2/5] iio: adc: ti-adc128s052: Use shift and realbits
On Sat, Jun 14, 2025 at 02:27:43PM +0100, Jonathan Cameron wrote:
> On Sat, 14 Jun 2025 02:15:01 -0700
> Sukrut Bellary <sbellary@...libre.com> wrote:
>
> > This adcxx communicates with a host processor via an SPI/Microwire Bus
> > interface. The device family responds with 12-bit data, of which the LSB bits
> > are transmitted by the lower resolution devices as 0. The unavailable bits are
> > 0 in LSB. Shift is calculated per resolution and used in scaling and raw data
> > read.
> >
> > Create a separate structure for each device type instead of an array.
> > These changes help to reuse the driver to support the family of devices with
> > name ADC<bb><c>S<sss>, where
> > * bb is the resolution in number of bits (8, 10, 12)
> > * c is the number of channels (1, 2, 4, 8)
> > * sss is the maximum conversion speed (021 for 200 kSPS, 051 for 500 kSPS
> > and 101 for 1 MSPS)
> >
> > Complete datasheets are available at TI's website here:
> > https://www.ti.com/lit/ds/symlink/adc<bb><c>s<sss>.pdf
> >
> > Co-developed-by: Nishanth Menon <nm@...com>
> > Signed-off-by: Nishanth Menon <nm@...com>
> > Signed-off-by: Sukrut Bellary <sbellary@...libre.com>
> > ---
> > drivers/iio/adc/ti-adc128s052.c | 115 ++++++++++++++++++--------------
> > 1 file changed, 66 insertions(+), 49 deletions(-)
> >
> > diff --git a/drivers/iio/adc/ti-adc128s052.c b/drivers/iio/adc/ti-adc128s052.c
> > index 1b46a8155803..2b206745e53d 100644
> > --- a/drivers/iio/adc/ti-adc128s052.c
> > +++ b/drivers/iio/adc/ti-adc128s052.c
> > @@ -41,13 +41,14 @@ struct adc128 {
> > } __aligned(IIO_DMA_MINALIGN);
> > };
> >
> > -static int adc128_adc_conversion(struct adc128 *adc, u8 channel)
> > +static int adc128_adc_conversion(struct adc128 *adc,
> > + struct iio_chan_spec const *channel)
> > {
> > int ret;
> >
> > guard(mutex)(&adc->lock);
> >
> > - adc->buffer[0] = channel << 3;
> > + adc->buffer[0] = channel->channel << 3;
> > adc->buffer[1] = 0;
> >
> > ret = spi_write(adc->spi, &adc->buffer, sizeof(adc->buffer));
> > @@ -58,7 +59,10 @@ static int adc128_adc_conversion(struct adc128 *adc, u8 channel)
> > if (ret < 0)
> > return ret;
> >
> > - return be16_to_cpu(adc->buffer16) & 0xFFF;
> > + ret = (be16_to_cpu(adc->buffer16) >> channel->scan_type.shift) &
> > + GENMASK(channel->scan_type.realbits - 1, 0);
> > +
> Even though it is a bit long I'd go with
>
> return (be16_to_cpu(adc->buffer16) >> channel->scan_type.shift) &
> GENMASK();
>
Thanks for the review.
I will fix this in v5.
> > + return ret;
> > }
> >
> > static int adc128_read_raw(struct iio_dev *indio_dev,
> > @@ -71,7 +75,7 @@ static int adc128_read_raw(struct iio_dev *indio_dev,
> > switch (mask) {
> > case IIO_CHAN_INFO_RAW:
> >
> > - ret = adc128_adc_conversion(adc, channel->channel);
> > + ret = adc128_adc_conversion(adc, channel);
> > if (ret < 0)
> > return ret;
> >
> > @@ -81,7 +85,7 @@ static int adc128_read_raw(struct iio_dev *indio_dev,
> > case IIO_CHAN_INFO_SCALE:
> >
> > *val = adc->vref_mv;
> > - *val2 = 12;
> > + *val2 = channel->scan_type.realbits;
> > return IIO_VAL_FRACTIONAL_LOG2;
> >
> > default:
> > @@ -90,15 +94,24 @@ static int adc128_read_raw(struct iio_dev *indio_dev,
> >
> > }
> >
> > -#define ADC128_VOLTAGE_CHANNEL(num) \
> > - { \
> > - .type = IIO_VOLTAGE, \
> > - .indexed = 1, \
> > - .channel = (num), \
> > - .info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \
> > - .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE) \
> > +#define _ADC128_VOLTAGE_CHANNEL(num, real_bits) \
> > + { \
>
> I would minimise the churn and stick to existing style of one space then \
> I don't think we have any specific style guidance around this.
>
I will fix this in v5.
> > + .type = IIO_VOLTAGE, \
> > + .indexed = 1, \
> > + .channel = (num), \
> > + .info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \
> > + .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE), \
> > + .scan_index = (num), \
> > + .scan_type = { \
> > + .sign = 'u', \
> > + .realbits = (real_bits), \
> > + .storagebits = 16, \
> > + .shift = (12 - real_bits), \
> > + }, \
> > }
> >
> > +#define ADC128_VOLTAGE_CHANNEL(num) _ADC128_VOLTAGE_CHANNEL(num, 12)
>
> I wonder if it would be clearer to just have the 12 explicit in each entry
> and skip this two levels of macro thing?
>
Do you mean to pass realbits to
ADC128_VOLTAGE_CHANNEL/_ADC128_VOLTAGE_CHANNEL as e.g.,
static const struct iio_chan_spec adc122s021_channels[] = {
ADC128_VOLTAGE_CHANNEL(0, 12),
ADC128_VOLTAGE_CHANNEL(1, 12),
};
I think we added 2nd level macros as ADC082_VOLTAGE_CHANNEL,
ADC102_VOLTAGE_CHANNEL, etc., to have a visual distinction for a different
part nos.
But I am ok if you prefer ADC128_VOLTAGE_CHANNEL with a second parameter
as real_bits.
> > +
> > static const struct iio_chan_spec adc128s052_channels[] = {
> > ADC128_VOLTAGE_CHANNEL(0),
> > ADC128_VOLTAGE_CHANNEL(1),
> > @@ -124,26 +137,30 @@ static const struct iio_chan_spec adc124s021_channels[] = {
> >
> > static const char * const bd79104_regulators[] = { "iovdd" };
> >
> > -static const struct adc128_configuration adc128_config[] = {
> > - {
> > - .channels = adc128s052_channels,
> > - .num_channels = ARRAY_SIZE(adc128s052_channels),
> > - .refname = "vref",
> > - }, {
> > - .channels = adc122s021_channels,
> > - .num_channels = ARRAY_SIZE(adc122s021_channels),
> > - .refname = "vref",
> > - }, {
> > - .channels = adc124s021_channels,
> > - .num_channels = ARRAY_SIZE(adc124s021_channels),
> > - .refname = "vref",
> > - }, {
> > - .channels = adc128s052_channels,
> > - .num_channels = ARRAY_SIZE(adc128s052_channels),
> > - .refname = "vdd",
> > - .other_regulators = &bd79104_regulators,
> > - .num_other_regulators = 1,
> > - },
> > +static const struct adc128_configuration adc122s021_config = {
> > + .channels = adc122s021_channels,
> > + .num_channels = ARRAY_SIZE(adc122s021_channels),
> > + .refname = "vref",
> > +};
>
> Ideal would be to have this as a precursor patch rather than adding complexity
> to this one which is focused on the bits related stuff.
>
> It's a good change to have but does make it harder to spot the main
> content in here.
>
>
I will split this in v5.
> > +
> > +static const struct adc128_configuration adc124s021_config = {
> > + .channels = adc124s021_channels,
> > + .num_channels = ARRAY_SIZE(adc124s021_channels),
> > + .refname = "vref",
> > +};
>
Powered by blists - more mailing lists