Re: [PATCH v3 07/10] regulator: mt6358: Add support for MT6358 regulator

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1918 bytes --]

On Fri, May 03, 2019 at 05:31:14PM +0800, Hsin-Hsiung Wang wrote:

A few fairly minor things but mostly this looks good.

> +static const u32 vcama1_voltages[] = {
> +	1800000, 2500000, 2700000,
> +	2800000, 2900000, 3000000,
> +};

> +static const u32 vcama2_voltages[] = {
> +	1800000, 2500000, 2700000,
> +	2800000, 2900000, 3000000,
> +};
> +

These two tables look the same?

> +static int mt6358_get_voltage_sel(struct regulator_dev *rdev)
> +{
> +	int idx, ret;
> +	u32 selector;
> +	struct mt6358_regulator_info *info = rdev_get_drvdata(rdev);
> +	const u32 *pvol;
> +
> +	ret = regmap_read(rdev->regmap, info->desc.vsel_reg, &selector);
> +	if (ret != 0) {
> +		dev_info(&rdev->dev,
> +			 "Failed to get mt6358 %s vsel reg: %d\n",
> +			 info->desc.name, ret);
> +		return ret;
> +	}
> +
> +	selector = (selector & info->desc.vsel_mask) >> info->vsel_shift;
> +	pvol = (const u32 *)info->index_table;
> +	ret = -1;
> +	for (idx = 0; idx < info->desc.n_voltages; idx++) {
> +		if (pvol[idx] == selector) {
> +			ret = idx;
> +			break;
> +		}
> +	}
> +
> +	return ret;
> +}

I don't entirely understand what this is intended to do which suggests
it could use some comments.  As far as I can tell it's looking up the
hardware selector in a table and then using the index of the entry in
that table as the selector in order to deal with a sparse set of
selectors, is that right?  You should be able to handle this by having
_list_voltage() return an error code for the invalid selectors though we
don't have helpers for that yet as it's an unusual design.

Also -1 is not a good return value in the case where you fail to do a
mapping, use -EINVAL.

> +	ret = regmap_read(rdev->regmap, info->da_vsel_reg, &regval);
> +	if (ret != 0) {
> +		dev_info(&rdev->dev,
> +			 "Failed to get mt6358 Buck %s vsel reg: %d\n",
> +			 info->desc.name, ret);
> +		return ret;
> +	}

These error messages should be dev_err().

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]