Re: [PATCH v2 2/2] iio: adc: mcp3911: add support for the whole MCP39xx family

From: Andy Shevchenko
Date: Fri Aug 04 2023 - 15:37:41 EST


On Fri, Aug 04, 2023 at 12:02:48PM +0200, Marcus Folkesson wrote:
> Microchip does have many similar chips, add support for those.
>
> The new supported chips are:
> - microchip,mcp3910
> - microchip,mcp3912
> - microchip,mcp3913
> - microchip,mcp3914
> - microchip,mcp3918
> - microchip,mcp3919

...

> help
> - Say yes here to build support for Microchip Technology's MCP3911
> - analog to digital converter.
> + Say yes here to build support for Microchip Technology's MCP3910,

> + MCP3911, MCP3912, MCP3913, MCP3914, MCP3918 and MCP3919

This line is misindented. Should be <TAB><space><space>.

> + analog to digital converters.

...

> +#define MCP3910_REG_OFFCAL_CH0 0x0f
> +#define MCP3910_OFFCAL(x) (MCP3910_REG_OFFCAL_CH0 + x * 6)
> +
> +

Single blank line is enough.

...

> +static int mcp3910_get_offset(struct mcp3911 *adc, int channel, int *val)
> +{
> + return mcp3911_read(adc, MCP3910_OFFCAL(channel), val, 3);

Just to be sure, the proper endianess conversion is done in mcp3911_read()
and mcp3911_write() calls?

This question applies to all calls to that APIs.

> +}

...

> + int ret = mcp3911_write(adc, MCP3910_OFFCAL(channel), val,
> + 3);

This looks weird not being on a single line. Moreover it fits even
80 characters. Same applies to other similar cases.

Also, please use better approach, i.e.

int ret;

ret = ...(...);
if (ret)
...

Also applies to several places.

> + if (ret)
> + return ret;

...

> + /* Enable offset*/

Missing space.

...

> +static int mcp3910_get_osr(struct mcp3911 *adc, int *val)
> +{
> + int ret = mcp3911_read(adc, MCP3910_REG_CONFIG0, val, 3);

Have you run checkpatch? Here should be a blank line. Same in other several
places.

> + *val = FIELD_GET(MCP3910_CONFIG0_OSR, *val);
> + *val = 32 << *val;

Please, use a temporary variable and assign the result only once.
It will be a better code.

> + return ret;
> +}

...

> +static int mcp3911_set_osr(struct mcp3911 *adc, int val)
> +{
> + val = FIELD_PREP(MCP3911_CONFIG_OSR, val);

As per above.

> + return mcp3911_update(adc, MCP3911_REG_CONFIG,
> + MCP3911_CONFIG_OSR, val, 2);

> +

Redundant blank line.

> +}

...

> +static int mcp3911_get_osr(struct mcp3911 *adc, int *val)
> +{

As per above.

> +}

...

> + /* Set gain to 1 for all channels */

Again, wrong indentation. Can you check all your lines for the proper
indentation. (Here should be just <TAB>.)

...

> + for (int i = 0; i < adc->chip->num_channels - 1; i++) {
> + adc->gain[i] = 1;
> + regval &= ~MCP3911_GAIN_MASK(i);
> + }

Missing blank line.

> + return mcp3911_write(adc, MCP3911_REG_GAIN, regval, 1);
> +
> +

Too many redundant blank lines.

> +}

...

> + /* Set gain to 1 for all channels */

Mind indentation.

...

> + /* Disable offset to ignore any old values in offset register*/

Missing space.

...

> + adc->chip = (struct mcp3911_chip_info *)spi_get_device_id(spi)->driver_data;

Can't you use spi_get_device_match_data()?

...

> + /*
> + * Fallback to "device-addr" due to historical mismatch between
> + * dt-bindings and implementation

Missing grammatical period at the end.

> + */

...

> + device_property_read_u32(&adc->spi->dev, "device-addr", &adc->dev_addr);

With

struct device *dev = &adc->spi->dev;

lines like this will be neater.

--
With Best Regards,
Andy Shevchenko