Re: [PATCH v3] iio: dac: mcp47feb02: Fix passing uninitialized vref1_uV for no Vref1 case

From: Andy Shevchenko

Date: Thu Jun 04 2026 - 04:00:55 EST


On Wed, Jun 03, 2026 at 05:44:48PM +0300, Ariana Lazar wrote:
> Ensure that if a device has Vref1 but reading the regulator returns an
> error, mcp47feb02_init_ctrl_regs() is not called with an uninitialized
> vref1_uV value.
>
> Also add a device_property_present() check for the Vref1 supply before
> reading the regulator.

...

> +static void mcp47feb02_regulator_disable(void *d)
> +{
> + struct regulator *reg = (struct regulator *)d;

Unneeded casting, then for the maintenance it's better to have assignment and
definition to be split as we have a validation check.

> + if (reg)
> + regulator_disable(reg);
> +}

With the above being said, there are two alternatives:

static void mcp47feb02_regulator_disable(void *d)
{
struct regulator *reg;

reg = d;
if (reg)
regulator_disable(reg);
}

OR (my preference)

static void mcp47feb02_regulator_disable(void *reg)
{
if (reg)
regulator_disable(reg);
}

...

> +static bool mcp47feb02_ref_mismatch(struct mcp47feb02_data *data, unsigned int ch)
> +{
> + bool use_vref, use_bandgap;
> +
> + if (data->chdata[ch].ref_mode == MCP47FEB02_VREF_VDD)
> + return false;
> +
> + use_vref = (data->phys_channels >= 4 && (ch % 2)) ? data->use_vref1 : data->use_vref;

Too many spaces.

> + use_bandgap = (data->chdata[ch].ref_mode == MCP47FEB02_INTERNAL_BAND_GAP);

(Unneeded parentheses.)

> + return use_vref == use_bandgap;

I would rewrite this as

bool use_bandgap;

if (data->chdata[ch].ref_mode == MCP47FEB02_VREF_VDD)
return false;

use_bandgap = data->chdata[ch].ref_mode == MCP47FEB02_INTERNAL_BAND_GAP;

if (data->phys_channels >= 4 && (ch % 2))
return data->use_vref1 == use_bandgap;
else // redundant, but left for the formatting purposes
return data->use_vref == use_bandgap;

> +}

...

> + data->phys_channels >= 4 && (i % 2)) {

This idiom is repeated so many times that I would rather see

static inline bool is_...(data, ch)
{
return data->phys_channels >= 4 && (ch % 2);
}

and be used everywhere. If required, add also a comment on top to explain the logic,
why we need it to be done this way.

--
With Best Regards,
Andy Shevchenko