Re: [PATCH v2 5/7] regulator: max77802: Document regulator opmode DT properties

[Javier Martinez Canillas <javier.martinez@collabora.co.uk>]

Hello Mark,

On 10/17/2014 01:57 PM, Mark Brown wrote:
> On Thu, Oct 16, 2014 at 06:48:51PM +0200, Javier Martinez Canillas wrote:
> 
>> +- maxim,regulator-initial-mode: initial operating mode.
>> +  This property can only be used on regulators that support changing their mode
>> +  during normal operation. These regulators are LDO1, LDO3, LDO20 and LDO21.
>> +- maxim,regulator-disk-mode: operating mode for the regulator when the system
>> +  enters in the Suspend-to-Disk state.
>> +- maxim,regulator-mem-mode: operating mode for the regulator when the system
>> +  enters in the Suspend-to-RAM state.
> 
> This seems pretty ugly since it's not integrated with the suspend state
> binding at all - adding new suspend modes is going to involve changing
> the binding which seems icky.  Adding a standard property to set modes
> doesn't seem so bad, I think a translation function to parse device
> specific mode bindings in properties might be the way forwards.
> 

Just to be sure I understood correctly, are you suggesting something like this?

	ldo1_reg: LDO1 {
		regulator-name = "vdd_1v0";
		regulator-min-microvolt = <1000000>;
		regulator-max-microvolt = <1000000>;
		regulator-state-mem {
			regulator-on-in-suspend;
			regulator-mode = <MAX77802_OPMODE_LP>;
		};
	};

In other words, extending Chanwoo Choi's original suspend state binding to add
the regulator-mode property that was present in his v3 [0] but instead trying
to use the standard REGULATOR_MODE_*, say that each regulator driver should
define it's own device-specific set of modes and a do the translation to fill
standard modes in the struct regulation_constraints {initial,disk,mem} mode?

That way adding new suspend states, will only require changing the generic
regulator binding but not the regulator driver specific bindings.

Best regards,
Javier

[0]: https://lkml.org/lkml/2014/8/13/768