Re: [PATCH 5/6] mfd: TPS65910: Fix tps65910_set_voltage

[Mark Brown <broonie@opensource.wolfsonmicro.com>]

On Tue, Oct 18, 2011 at 01:26:27PM -0500, Kyle Manna wrote:

*Always* CC maintainers on patches.

> Previously tps65910_set_voltage() only selected from a fixed number of
> voltages.  Rename that function to tps65910_set_voltage_sel(). Do the
> same for tps65911_set_voltage().

What is the issue being fixed here?  This looks like a stylistic change
rateher than a bug fix.

> Also add a tps65910_set_voltage that works with the regulator framework
> and applies the correct voltage with apply_uv is set in the regulator's
> constraints.

So this is adding support for a new chip?  Whatever the answer it's
clearly a distinct change from the above and should therefore be a
separate patch.

> +	/* Pick the nearest selector */
> +	for (i = 0; i < tps65910_regs[id].table_len; i++) {
> +		new_uV = tps65910_regs[id].table[i] * 1000;
> +
> +		if (new_uV >= min_uV && new_uV <= max_uV &&
> +		   (abs(new_uV - midpoint) < abs(selected_uV - midpoint))) {
> +			*selector = i;
> +			selected_uV = tps65910_regs[id].table[i] * 1000;
> +		}
> +	}

This looks wrong, the expected behaviour for the regulator API is that
the driver will pick the minimum voltage within the range.  Why is this
being done?