Re: [PATCH v3 0/1] iio/scmi: Add reading "raw" attribute.

[Jyoti Bhayana <jbhayana@google.com>]

Hi Vasyl,

Regarding below question, yes reading raw attribute should be blocked
if buffer is enabled for that sensor.

> 1. Should we block reading raw attribute and IIO buffer enabled, for for
> SCMI sensor it can coexist?


PLease see  https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/inv_mpu6050/inv_mpu_core.c#L667
as well. It has

case IIO_CHAN_INFO_RAW:
ret = iio_device_claim_direct_mode(indio_dev);
if (ret)
return ret;
mutex_lock(&st->lock);
ret = inv_mpu6050_read_channel_data(indio_dev, chan, val);
mutex_unlock(&st->lock);
iio_device_release_direct_mode(indio_dev);
return ret;

Regarding the question below, the answer is yes.

> 2. Should we wrap reading raw attribute implementation in iio_dev->mlock
> mutex?


Thanks, Jyoti





On Tue, Oct 5, 2021 at 5:52 AM Vasyl Vavrychuk
<vasyl.vavrychuk@opensynergy.com> wrote:
>
> Hi, Jyoti,
>
> > In the code below, why is the logic of enabling and disabling the
> > sensor in this function?  Generally the function to read the sensor
> > value is just used for the code to read the sensor values ? and not
> > enable/disable the sensor
>
> But to read sensor value we have to enable it first. Other way to enable
> sensor we found is, for example:
>
>    echo 1 > /sys/bus/iio/devices/.../scan_elements/in_anglvel_x_en
>
> But, this command is related to IIO buffers use.
>
> Other sensors drivers enable/disable sensor in read raw too, for
> example, drivers/iio/accel/kxcjk-1013.c has:
>
>    case IIO_CHAN_INFO_RAW:
>            mutex_lock(&data->mutex);
>            if (iio_buffer_enabled(indio_dev))
>                    ret = -EBUSY;
>            else {
>                    ret = kxcjk1013_set_power_state(data, true);
>                    ... reading ...
>                    ret = kxcjk1013_set_power_state(data, false);
>            }
>            mutex_unlock(&data->mutex);
>
> But, after looking on this code I have some questions:
>
> 1. Should we block reading raw attribute and IIO buffer enabled, for for
> SCMI sensor it can coexist?
> 2. Should we wrap reading raw attribute implementation in iio_dev->mlock
> mutex?
>
> >>     case IIO_CHAN_INFO_RAW:
> >> +               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 enabling sensor %s err %d",
> >> +                               sensor->sensor_info->name, err);
> >> +                       return err;
> >> +               }
> >> +               /* Check if raw value fits 32 bits */
> >> +               if (readings[ch->scan_index].value < INT_MIN ||
> >> +                   readings[ch->scan_index].value > INT_MAX)
> >> +                       return -ERANGE;
> >> +               /* Use 32-bit value, since practically there is no need in 64 bits */
> >> +               *val = (int)readings[ch->scan_index].value;
> >>
> >> +               return IIO_VAL_INT;