Re: [PATCH v6 1/1] iio/scmi: Add reading "raw" attribute.

From: Jonathan Cameron
Date: Wed Oct 20 2021 - 13:47:27 EST


On Tue, 19 Oct 2021 10:59:49 +0300
Andriy Tryshnivskyy <andriy.tryshnivskyy@xxxxxxxxxxxxxxx> wrote:

> Add IIO_CHAN_INFO_RAW to the mask and implement corresponding
> reading "raw" attribute in scmi_iio_read_raw.
> Introduce new type IIO_VAL_INT_64 to read 64-bit value
> for "raw" attribute.
>
Change log needs to be below the --- otherwise we'll store it forever
in git. A linked tag (which will be generated when I apply)
is sufficient for this sort of historical info.

> Changes comparing v5 -> v6:
> * revert v5 changes since with scmi_iio_read_raw() the channel
> can't be used by in kernel users (iio-hwmon)
> * returned to v3 with direct mode
> * introduce new type IIO_VAL_INT_64 to read 64-bit value
>
> Changes comparing v4 -> v5:
> * call iio_device_release_direct_mode() on error
> * code cleanup, fix typo
>
> Changes comparing v3 -> v4:
> * do not use scmi_iio_get_raw() for reading raw attribute due to 32-bit
> return value limitation (actually I reverted the previous v3)
> * introduce scmi_iio_read_raw to scmi_iio_ext_info[] which can return
> 64-bit value
> * enabling/disabling and reading raw attribute is done in direct mode
>
> Changes comparing v2 -> v3:
> * adaptation for changes in structure scmi_iio_priv (no member
> named 'handle')
>
> Changes comparing v0 -> v2:
> * added an error return when the error happened during config_set
> * removed redundant cast for "readings"
> * added check if raw value fits 32 bits
>
> Signed-off-by: Andriy Tryshnivskyy <andriy.tryshnivskyy@xxxxxxxxxxxxxxx>
> ---
> drivers/iio/common/scmi_sensors/scmi_iio.c | 57 +++++++++++++++++++++-
> drivers/iio/industrialio-core.c | 3 ++
> include/linux/iio/types.h | 1 +

Two patches needed. One to introduce the new core functionality then
a second to use it in the driver.

Actual code looks good to me though I think I'd like a comment next to
the #define as not obvious which way around the two parts will go.

There are some other places we will probably need to ultimately handle this
to allow for in kernel consumers but those can come when someone needs them.

Will need an ack from Jyoti on this one though as driver author.
Thanks,

Jonathan


> 3 files changed, 60 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iio/common/scmi_sensors/scmi_iio.c b/drivers/iio/common/scmi_sensors/scmi_iio.c
> index 7cf2bf282cef..2c1aec0fd5ff 100644
> --- a/drivers/iio/common/scmi_sensors/scmi_iio.c
> +++ b/drivers/iio/common/scmi_sensors/scmi_iio.c
> @@ -279,6 +279,52 @@ static int scmi_iio_get_odr_val(struct iio_dev *iio_dev, int *val, int *val2)
> return 0;
> }
>
> +static int scmi_iio_read_channel_data(struct iio_dev *iio_dev,
> + struct iio_chan_spec const *ch, int *val, int *val2)
> +{
> + struct scmi_iio_priv *sensor = iio_priv(iio_dev);
> + u32 sensor_config;
> + struct scmi_sensor_reading readings[SCMI_IIO_NUM_OF_AXIS];
> + int err;
> +
> + sensor_config = FIELD_PREP(SCMI_SENS_CFG_SENSOR_ENABLED_MASK,
> + SCMI_SENS_CFG_SENSOR_ENABLE);
> + err = sensor->sensor_ops->config_set(
> + sensor->ph, sensor->sensor_info->id, sensor_config);
> + if (err) {
> + dev_err(&iio_dev->dev,
> + "Error in enabling sensor %s err %d",
> + sensor->sensor_info->name, err);
> + return err;
> + }
> +
> + err = sensor->sensor_ops->reading_get_timestamped(
> + sensor->ph, sensor->sensor_info->id,
> + sensor->sensor_info->num_axis, readings);
> + if (err) {
> + dev_err(&iio_dev->dev,
> + "Error in reading raw attribute for sensor %s err %d",
> + sensor->sensor_info->name, err);
> + return err;
> + }
> +
> + sensor_config = FIELD_PREP(SCMI_SENS_CFG_SENSOR_ENABLED_MASK,
> + SCMI_SENS_CFG_SENSOR_DISABLE);
> + err = sensor->sensor_ops->config_set(
> + sensor->ph, sensor->sensor_info->id, sensor_config);
> + if (err) {
> + dev_err(&iio_dev->dev,
> + "Error in disabling sensor %s err %d",
> + sensor->sensor_info->name, err);
> + return err;
> + }
> +
> + *val = (u32)readings[ch->scan_index].value;
> + *val2 = (u32)(readings[ch->scan_index].value >> 32)
> +
> + return IIO_VAL_INT_64;
> +}
> +
> static int scmi_iio_read_raw(struct iio_dev *iio_dev,
> struct iio_chan_spec const *ch, int *val,
> int *val2, long mask)
> @@ -300,6 +346,14 @@ static int scmi_iio_read_raw(struct iio_dev *iio_dev,
> case IIO_CHAN_INFO_SAMP_FREQ:
> ret = scmi_iio_get_odr_val(iio_dev, val, val2);
> return ret ? ret : IIO_VAL_INT_PLUS_MICRO;
> + case IIO_CHAN_INFO_RAW:
> + ret = iio_device_claim_direct_mode(iio_dev);
> + if (ret)
> + return ret;
> +
> + ret = scmi_iio_read_channel_data(iio_dev, ch, val, val2);
> + iio_device_release_direct_mode(iio_dev);
> + return ret;
> default:
> return -EINVAL;
> }
> @@ -381,7 +435,8 @@ static void scmi_iio_set_data_channel(struct iio_chan_spec *iio_chan,
> iio_chan->type = type;
> iio_chan->modified = 1;
> iio_chan->channel2 = mod;
> - iio_chan->info_mask_separate = BIT(IIO_CHAN_INFO_SCALE);
> + iio_chan->info_mask_separate =
> + BIT(IIO_CHAN_INFO_SCALE) | BIT(IIO_CHAN_INFO_RAW);
> iio_chan->info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SAMP_FREQ);
> iio_chan->info_mask_shared_by_type_available =
> BIT(IIO_CHAN_INFO_SAMP_FREQ);
> diff --git a/drivers/iio/industrialio-core.c b/drivers/iio/industrialio-core.c
> index 6d2175eb7af2..49e42d04ea16 100644
> --- a/drivers/iio/industrialio-core.c
> +++ b/drivers/iio/industrialio-core.c
> @@ -702,6 +702,9 @@ static ssize_t __iio_format_value(char *buf, size_t offset, unsigned int type,
> }
> case IIO_VAL_CHAR:
> return sysfs_emit_at(buf, offset, "%c", (char)vals[0]);
> + case IIO_VAL_INT_64:
> + tmp2 = (s64)((((u64)vals[1]) << 32) | (u32)vals[0]);
> + return sysfs_emit_at(buf, offset, "%lld", tmp2);
> default:
> return 0;
> }
> diff --git a/include/linux/iio/types.h b/include/linux/iio/types.h
> index 84b3f8175cc6..e148fe11a3dc 100644
> --- a/include/linux/iio/types.h
> +++ b/include/linux/iio/types.h
> @@ -24,6 +24,7 @@ enum iio_event_info {
> #define IIO_VAL_INT_PLUS_NANO 3
> #define IIO_VAL_INT_PLUS_MICRO_DB 4
> #define IIO_VAL_INT_MULTIPLE 5
> +#define IIO_VAL_INT_64 6

Possibly worth a descriptive comment. The other
types tend to make it easy to assume the role
of val and that of val2, in this case, val being
the lower 32 bits isn't obvious...

> #define IIO_VAL_FRACTIONAL 10
> #define IIO_VAL_FRACTIONAL_LOG2 11
> #define IIO_VAL_CHAR 12