Re: [PATCH v3 7/9] soc: mediatek: pwrap: add MediaTek MT6380 as one slave of pwrap

[Sean Wang]

On Tue, 2017-10-10 at 12:02 +0200, Matthias Brugger wrote:
> 
> On 08/15/2017 11:09 AM, sean.wang@xxxxxxxxxxxx wrote:
> > From: Sean Wang <sean.wang@xxxxxxxxxxxx>
> > 
> > Add MediaTek MT6380 regulator becoming one of PMIC wrapper slave
> > and also add extra new regmap_config of 32-bit mode for MT6380
> > since old regmap_config of 16-bit mode can't be fit into the need.
> > 
> > Signed-off-by: Chenglin Xu <chenglin.xu@xxxxxxxxxxxx>
> > Signed-off-by: Chen Zhong <chen.zhong@xxxxxxxxxxxx>
> > Signed-off-by: Sean Wang <sean.wang@xxxxxxxxxxxx>
> > ---
> >   drivers/soc/mediatek/mtk-pmic-wrap.c | 25 ++++++++++++++++++++++---
> >   1 file changed, 22 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/soc/mediatek/mtk-pmic-wrap.c b/drivers/soc/mediatek/mtk-pmic-wrap.c
> > index 1f8b69a..047e3d9 100644
> > --- a/drivers/soc/mediatek/mtk-pmic-wrap.c
> > +++ b/drivers/soc/mediatek/mtk-pmic-wrap.c
> > @@ -501,6 +501,7 @@ struct pmic_wrapper;
> >   struct pwrap_slv_type {
> >   	const u32 *dew_regs;
> >   	enum pmic_type type;
> > +	const struct regmap_config *regmap;
> >   	/* pwrap operations are highly associated with the PMIC types,
> >   	 * so the pointers added increases flexibility allowing determination
> >   	 * which type is used by the detection through device tree.
> > @@ -1109,7 +1110,7 @@ static irqreturn_t pwrap_interrupt(int irqno, void *dev_id)
> >   	return IRQ_HANDLED;
> >   }
> >   
> > -static const struct regmap_config pwrap_regmap_config = {
> > +static const struct regmap_config pwrap_regmap_config16 = {
> >   	.reg_bits = 16,
> >   	.val_bits = 16,
> >   	.reg_stride = 2,
> > @@ -1118,9 +1119,19 @@ static const struct regmap_config pwrap_regmap_config = {
> >   	.max_register = 0xffff,
> >   };
> >   
> > +static const struct regmap_config pwrap_regmap_config32 = {
> > +	.reg_bits = 32,
> > +	.val_bits = 32,
> > +	.reg_stride = 4,
> > +	.reg_read = pwrap_regmap_read,
> > +	.reg_write = pwrap_regmap_write,
> > +	.max_register = 0xffff,
> > +};
> > +
> >   static const struct pwrap_slv_type pmic_mt6323 = {
> >   	.dew_regs = mt6323_regs,
> >   	.type = PMIC_MT6323,
> > +	.regmap = &pwrap_regmap_config16,
> >   	.pwrap_read = pwrap_read16,
> >   	.pwrap_write = pwrap_write16,
> >   };
> > @@ -1128,6 +1139,7 @@ static const struct pwrap_slv_type pmic_mt6323 = {
> >   static const struct pwrap_slv_type pmic_mt6380 = {
> >   	.dew_regs = NULL,
> >   	.type = PMIC_MT6380,
> > +	.regmap = &pwrap_regmap_config32,
> >   	.pwrap_read = pwrap_read32,
> >   	.pwrap_write = pwrap_write32,
> >   };
> > @@ -1135,6 +1147,7 @@ static const struct pwrap_slv_type pmic_mt6380 = {
> >   static const struct pwrap_slv_type pmic_mt6397 = {
> >   	.dew_regs = mt6397_regs,
> >   	.type = PMIC_MT6397,
> > +	.regmap = &pwrap_regmap_config16,
> >   	.pwrap_read = pwrap_read16,
> >   	.pwrap_write = pwrap_write16,
> >   };
> > @@ -1144,9 +1157,15 @@ static const struct of_device_id of_slave_match_tbl[] = {
> >   		.compatible = "mediatek,mt6323",
> >   		.data = &pmic_mt6323,
> >   	}, {
> > +		/* The MT6380 slave device is directly pointed to the regulator
> > +		 * device which is different from the cases MT6323 and MT6397
> > +		 * where they're one kind of MFDs.
> > +		 */
> > +		.compatible = "mediatek,mt6380-regulator",
> > +		.data = &pmic_mt6380,
> 
> I understand that mt6380 only provides a regulator and no other function other 
> PMICs provide, right?

> Then maybe write a comment like:
> The MT6380 PMIC only implements a regulator, so we bind it directly instead of 
> using a MFD. If so, we should state that in the pwrap bindings document, I think.
> 

You're right. It is worth making them better in both comments and the
bindings document. I'll do it in the next version.


> Regards,
> Matthias