With the upstream Linux version, the oversampling, oversampling_ratio
attributes are available at device level.

_get_iio_attr() is used for accessing channel attributes.

Signed-off-by: Alexandru Ardelean <[email protected]>

There seems to be a confusion between scale and range.
In the case of AD7606, the range (specified in the datasheet) determines
the scale of the device (or the channel in some cases).

Signed-off-by: Alexandru Ardelean <[email protected]>

Signed-off-by: Alexandru Ardelean <[email protected]>

The AD7606C-16 and AD7606C-18 support per-channel 'scale_availble' in SW
mode.
That is because each channel can be configured individually, to be
bipolar-differential, bipolar-single-ended or unipolar-single-ended.

Depending on the above configuration, the available scales change.
So, we need to tweak this per-channel.

Since there isn't a way to detect SW mode, we use some 'try' blocks in
Python.

Signed-off-by: Alexandru Ardelean <[email protected]>

For AD7606C-18 especially, there will be 32-bit samples getting returned
from the ADC.

Signed-off-by: Alexandru Ardelean <[email protected]>

The soft timestamp channel is a special IIO channel.
It's not an RX channel, so it hasn't any of the typically attributes (raw,
scale, etc) like the other RX channels.

So, just ignore it.

Signed-off-by: Alexandru Ardelean <[email protected]>

At least the 'raw' channel attribute returns int.
So,  'ad7606.to_volts(0, channel0.raw)' won't work with checking for numpy
types.

Signed-off-by: Alexandru Ardelean <[email protected]>