Re: [PATCH v4 2/3] regulator: max77686: Implement suspend disable for some LDOs

[Krzysztof Kozlowski]

On pon, 2014-10-27 at 11:08 +0100, Javier Martinez Canillas wrote:
> Hello Krzysztof,
> 
> On 10/27/2014 10:44 AM, Krzysztof Kozlowski wrote:
> > Some LDOs of Maxim 77686 PMIC support disabling during system suspend
> > (LDO{2,6,7,8,10,11,12,14,15,16}). This was already implemented as part
> > of set_suspend_mode function. In that case the mode was one of:
> >  - disable,
> >  - normal mode,
> >  - low power mode.
> > However there are no bindings for setting the mode during suspend.
> > 
> > Add suspend disable for LDO regulators supporting this to the existing
> > max77686_buck_set_suspend_disable() function. This helps reducing
> > energy consumption during system sleep.
> > 
> > Signed-off-by: Krzysztof Kozlowski <k.kozlowski@xxxxxxxxxxx>
> > ---
> >  drivers/regulator/max77686.c | 75 ++++++++++++++++++++++++++++++++++++++------
> >  1 file changed, 66 insertions(+), 9 deletions(-)
> > 
> > diff --git a/drivers/regulator/max77686.c b/drivers/regulator/max77686.c
> > index 3d0922051488..4b35c19c9286 100644
> > --- a/drivers/regulator/max77686.c
> > +++ b/drivers/regulator/max77686.c
> > @@ -87,18 +87,31 @@ struct max77686_data {
> >  	unsigned int opmode[MAX77686_REGULATORS];
> >  };
> >  
> > -/* Some BUCKS supports Normal[ON/OFF] mode during suspend */
> > -static int max77686_buck_set_suspend_disable(struct regulator_dev *rdev)
> > +/* Some BUCKs and LDOs supports Normal[ON/OFF] mode during suspend */
> > +static int max77686_set_suspend_disable(struct regulator_dev *rdev)
> >  {
> >  	unsigned int val;
> >  	struct max77686_data *max77686 = rdev_get_drvdata(rdev);
> >  	int ret, id = rdev_get_id(rdev);
> >  
> > -	if (id == MAX77686_BUCK1)
> > +	switch (id) {
> > +	case MAX77686_BUCK1:
> >  		val = MAX77686_BUCK_OFF_PWRREQ;
> > -	else
> > -		val = MAX77686_BUCK_OFF_PWRREQ
> > -			<< MAX77686_OPMODE_BUCK234_SHIFT;
> > +		break;
> > +	case MAX77686_BUCK2 ...	MAX77686_BUCK4:
> > +		val = MAX77686_BUCK_OFF_PWRREQ << MAX77686_OPMODE_BUCK234_SHIFT;
> > +		break;
> > +	case MAX77686_LDO2:
> > +	case MAX77686_LDO6 ... MAX77686_LDO8:
> > +	case MAX77686_LDO10 ... MAX77686_LDO12:
> > +	case MAX77686_LDO14 ... MAX77686_LDO16:
> > +		val = MAX77686_LDO_OFF_PWRREQ << MAX77686_OPMODE_SHIFT;
> > +		break;
> > +	default:
> > +		pr_warn("%s: regulator_suspend_disable not supported\n",
> > +			rdev->desc->name);
> > +		return -EINVAL;
> > +	}
> >  
> >  	ret = regmap_update_bits(rdev->regmap, rdev->desc->enable_reg,
> >  				 rdev->desc->enable_mask, val);
> > @@ -176,13 +189,53 @@ static int max77686_ldo_set_suspend_mode(struct regulator_dev *rdev,
> >  	return 0;
> >  }
> >  
> > +/*
> > + * For regulators supporting disabled in suspend it checks if opmode
> > + * is 'off_pwrreq' mode (disabled in suspend) and changes it to 'enable'.
> > + * For other regulators does nothing.
> > + */
> > +static void max77686_set_opmode_enable(struct max77686_data *max77686, int id)
> > +{
> > +	unsigned int opmode_shift;
> > +
> > +	/*
> > +	 * Same enable function is used for LDO and bucks. Assuming
> > +	 * the same values are used for enable registers.
> > +	 */
> > +	BUILD_BUG_ON(MAX77686_LDO_OFF_PWRREQ != MAX77686_BUCK_OFF_PWRREQ);
> > +	BUILD_BUG_ON(MAX77686_LDO_NORMAL != MAX77686_BUCK_NORMAL);
> > +
> > +	switch (id) {
> > +	case MAX77686_BUCK1:
> > +		opmode_shift = 0;
> > +		break;
> > +	case MAX77686_BUCK2 ...	MAX77686_BUCK4:
> > +		opmode_shift = MAX77686_OPMODE_BUCK234_SHIFT;
> > +		break;
> > +	case MAX77686_LDO2:
> > +	case MAX77686_LDO6 ... MAX77686_LDO8:
> > +	case MAX77686_LDO10 ... MAX77686_LDO12:
> > +	case MAX77686_LDO14 ... MAX77686_LDO16:
> > +		opmode_shift = MAX77686_OPMODE_SHIFT;
> > +		break;
> > +	default:
> > +		return;
> > +	}
> > +
> > +	if (max77686->opmode[id] == (MAX77686_LDO_OFF_PWRREQ << opmode_shift))
> > +		max77686->opmode[id] = MAX77686_LDO_NORMAL << opmode_shift;
> > +}
> > +
> >  static int max77686_enable(struct regulator_dev *rdev)
> >  {
> >  	struct max77686_data *max77686 = rdev_get_drvdata(rdev);
> > +	int id = rdev_get_id(rdev);
> > +
> > +	max77686_set_opmode_enable(max77686, id);
> >  
> >  	return regmap_update_bits(rdev->regmap, rdev->desc->enable_reg,
> >  				  rdev->desc->enable_mask,
> > -				  max77686->opmode[rdev_get_id(rdev)]);
> > +				  max77686->opmode[id]);
> >  }
> 
> the max77802 driver handles this slightly differently. It stores in opmode[id]
> just the value without the shift and there is a max77802_get_opmode_shift() to
> obtain the shift that is used before updating the register, e.g:
> 
> int shift = max77802_get_opmode_shift(id);
> return regmap_update_bits(..., max77802->opmode[id] << shift);
> 
> I think the max77802 driver approach is better since it avoids duplicating the
> switch statement to get the shift for each regulator. Just look how the max77802
> enable and disable functions are easier to read. Looking at the max77686 driver,
> I see that there are assumptions that opmode[id] has the value shifted so your
> changes are consistent with the rest of the driver but I would consider changing
> this.

Storing opmode in non-shifted form is intuitive also to me. That's way I
slipped the bug in previous version.

I'll change this to non-shifted opmode. The patch will grow bigger but
the code should be more readable.

Thanks!
Krzysztof


> Best regards,
> Javier
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/