Re: max1363 : Warnings from iio_sanity_check_avail_scan_masks()

[Matti Vaittinen <mazziesaccount@gmail.com>]

On 28/04/2025 22:04, Fabio Estevam wrote:
> Hi Matti,
> 
> On Mon, Apr 28, 2025 at 2:38 AM Matti Vaittinen
> <mazziesaccount@gmail.com> wrote:
> 
>> I believe it'd be more informative if we printed the actual bit masks -
>> but I am not sure if we know the size of the mask. If we keep printing
>> the indexes, then we might want to add brackets [] to indicate the value
>> is an index? How about:
>>
>> "available_scan_masks: mask[8] subset of mask[0]. Never used"
>>
>> instead of:
>> available_scan_mask 8 subset of 0. Never used
>>
>> Would that be more obvious?
> 
> I still find the "Never used" part a bit cryptic. What exactly is never used?

When user wants to enable some channels, the IIO core scans the 
available_scan_masks -array in order to find a supported configuration 
which meets the requirements. IIO core does this scanning in the order, 
and stops when first suitable mask is found from the available_scan_masks.

This means that if we have a mask which is a subset of a mask which is 
earlier in the array, this latter mask will never be used because the 
earlier one (the superset) will always be found first and picked by the 
IIO core.

Eg, if mask[0] consists of bits
0,3,5,7

and mask[1] consists of bits
0,5

- then if user wants to enable channels 0 or 5 (or 0 and 5) - the 
mask[0] will always be picked by the IIO core because it fulfills the 
needs. This means IIO will always enable channels 0,3,5 and 7 - and 
never use the smaller configuration of only channels 0 and 5.

Hence, the ordering of the masks matters, and this check (and print) was 
added.

I hope this clarifies.

Yours,
	-- Matti