Re: [PATCH v5 13/16] iio: adc: xilinx-xadc: Add helper functions for the device setup

From: Nagavarapu, Dileep Kumar

Date: Tue Nov 17 2026 - 03:07:57 EST



Hi Andy,

On 11/10/2026 9:55 PM, Andy Shevchenko wrote:
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.


On Tue, Nov 10, 2026 at 10:31:58AM +0530, Dileep Kumar Nagavarapu wrote:
Refactor the platform driver probe function by extracting device
setup and configuration logic into reusable helper functions:
xadc_device_setup(): handles IIO device allocation and basic setup
xadc_device_configure(): handles device tree parsing and bipolar mask
configuration

Move the bipolar channel mask calculation immediately after
xadc_parse_dt(), as the channel information used to build
the mask is populated there and remains unchanged before it is consumed.

This refactoring reduces code duplication and prepares for sharing the
common setup logic between platform and I2C drivers.

...

+ if (irq <= 0)
+ return 0;

When == 0 is true?


irq == 0 is not expected to be a valid successful return in this path. I’ll update the code to use the suggested early-return structure

...

+static int xadc_device_configure(struct iio_dev *indio_dev, int irq,
+ unsigned int *conf0, unsigned int *bipolar_mask)
+{
+ int ret;
+
+ ret = xadc_parse_dt(indio_dev, conf0, irq);
+ if (ret)
+ return ret;
+
+ *bipolar_mask = 0;
+ for (unsigned int i = 0; i < indio_dev->num_channels; i++) {
+ if (indio_dev->channels[i].scan_type.sign == 's')

Wasn't that changed to format?


I'll switch to scan_type.format == IIO_SCAN_FORMAT_SIGNED_INT, is it ok?

+ *bipolar_mask |= BIT(indio_dev->channels[i].scan_index);
+ }
+
+ return 0;

Don't we have a specific for_each type of macro to iterate over all channels?


I couldn't find a for_each helper for iterating over all channels, so I kept the plain loop

+}

...

irq = platform_get_irq_optional(pdev, 0);
if (irq < 0 && (irq != -ENXIO || !(ops->flags & XADC_FLAGS_IRQ_OPTIONAL)))

Side note, this conditional is wrong. It doesn't handle properly deferred probe.

return irq;

--
With Best Regards,
Andy Shevchenko

Thanks,
Dileep