Re: [PATCH v2 03/12] iio: adc: at91-sama5d2_adc: adapt the driver for sama7d65

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] Uninitialized touch channel indices incorrectly configure regular channel 0 as a touchscreen channel, breaking IIO buffered reads.
- [High] Temperature channel scan_index (31) exceeds the defined max_index (25), breaking IIO bitmap validations and hardware triggers.
--

commit 2d567bf99b76168b7ea132082c84ddd5029c1706
Author: Varshini Rajendran <varshini.rajendran@microchip.com>

iio: adc: at91-sama5d2_adc: adapt the driver for sama7d65

Add support for sama7d65 ADC. The differences are highlighted with the
compatible. The calibration data layout is the main difference.

> diff --git a/drivers/iio/adc/at91-sama5d2_adc.c b/drivers/iio/adc/at91-sama5d2_adc.c
> index b569d175f4c3d..237d339f342a1 100644
> --- a/drivers/iio/adc/at91-sama5d2_adc.c
> +++ b/drivers/iio/adc/at91-sama5d2_adc.c
[ ... ]
> @@ -768,6 +778,24 @@ static const struct at91_adc_platform sama7g5_platform = {
>  	.temp_calib_layout = &sama7g5_temp_calib,
>  };
>  
> +static const struct at91_adc_platform sama7d65_platform = {
> +	.layout = &sama7g5_layout,
> +	.adc_channels = &at91_sama7g5_adc_channels,
> +	.nr_channels = AT91_SAMA7G5_SINGLE_CHAN_CNT +
> +		       AT91_SAMA7G5_DIFF_CHAN_CNT +
> +		       AT91_SAMA7G5_TEMP_CHAN_CNT,

[Severity: High]
This isn't a bug introduced by this patch, but inherited from the SAMA7G5
implementation. Does omitting the touch channel indices cause an issue for
channel 0?

By omitting touch_chan_x and similar fields, they default to 0. Later,
during probe, this zero-initialized value is used:

at91_adc_probe() {
    ...
    bitmap_set(&st->touch_st.channels_bitmask,
               st->soc_info.platform->touch_chan_x, 1);
    ...
}

Will this cause the driver to incorrectly set bit 0 in the touchscreen
bitmask, diverting regular channel 0 captures into the touchscreen logic
and breaking IIO buffered reads for that channel?

> +	.max_channels = ARRAY_SIZE(at91_sama7g5_adc_channels),
> +	.max_index = AT91_SAMA7G5_MAX_CHAN_IDX,

[Severity: High]
This is also a pre-existing issue inherited from the SAMA7G5 implementation,
but does this max_index configuration conflict with the temperature channel?

The max_index is configured to 25 (AT91_SAMA7G5_MAX_CHAN_IDX), but the
temperature channel has a scan_index of 31. This mismatch appears to break
IIO bitmap validations and hardware triggers.

For instance, in the hardware enablement path:

at91_adc_active_scan_mask_to_reg() {
    ...
    for_each_set_bit(bit, indio_dev->active_scan_mask,
                     indio_dev->num_channels) {
    ...
    return mask & GENMASK(st->soc_info.platform->nr_channels, 0);
}

This applies a 26-bit mask, which permanently ignores the temperature channel
at index 31. Does this structurally break the temperature channel for IIO
buffered reads by causing the core to expect data that the driver never
fetches?

> +	.hw_trig_cnt = AT91_SAMA7G5_HW_TRIG_CNT,
> +	.osr_mask = GENMASK(18, 16),
> +	.oversampling_avail = { 1, 4, 16, 64, 256, },
> +	.oversampling_avail_no = 5,
> +	.chan_realbits = 16,
> +	.temp_sensor = true,
> +	.temp_chan = AT91_SAMA7G5_ADC_TEMP_CHANNEL,
> +	.temp_calib_layout = &sama7d65_temp_calib,
> +};

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260623105944.128840-1-varshini.rajendran@microchip.com?part=3