Re: [PATCH v2 4/4] pintcrl: support bias-disable of generic and special pins simultaneously

[Sean Wang <sean.wang@mediatek.com>]

On Thu, 2018-03-01 at 11:47 +0800, Zhiyong Tao wrote:
> On Wed, 2018-02-28 at 15:49 +0800, Zhiyong Tao wrote:
> > On Wed, 2018-02-28 at 15:33 +0800, Sean Wang wrote:
> > > On Mon, 2018-02-26 at 16:34 +0800, Zhiyong Tao wrote:
> > > > For generic pins, parameter "arg" is 0 or 1.
> > > > For special pins, bias-disable is set by R0R1,
> > > > so we need transmited "00" to set bias-disable
> > > > When we set "bias-disable" as high-z property,
> > > > the parameter should be "MTK_PUPD_SET_R1R0_00".
> > > > 
> > > > Signed-off-by: Zhiyong Tao <zhiyong.tao@mediatek.com>
> > > > ---
> > > >  drivers/pinctrl/mediatek/pinctrl-mtk-common.c | 13 +++++++++++--
> > > >  1 file changed, 11 insertions(+), 2 deletions(-)
> > > > 
> > > > diff --git a/drivers/pinctrl/mediatek/pinctrl-mtk-common.c b/drivers/pinctrl/mediatek/pinctrl-mtk-common.c
> > > > index 3cf384f..e88ba04 100644
> > > > --- a/drivers/pinctrl/mediatek/pinctrl-mtk-common.c
> > > > +++ b/drivers/pinctrl/mediatek/pinctrl-mtk-common.c
> > > > @@ -301,8 +301,17 @@ static int mtk_pconf_set_pull_select(struct mtk_pinctrl *pctl,
> > > >  	 * resistor bit, so we need this special handle.
> > > >  	 */
> > > >  	if (pctl->devdata->spec_pull_set) {
> > > > -		ret = pctl->devdata->spec_pull_set(mtk_get_regmap(pctl, pin),
> > > > -			pin, pctl->devdata->port_align, isup, arg);
> > > > +		if (enable) {
> > > > +			ret = pctl->devdata->spec_pull_set(
> > > > +				mtk_get_regmap(pctl, pin), pin,
> > > > +				pctl->devdata->port_align, isup,
> > > > +				arg);
> > > > +		} else {
> > > > +			ret = pctl->devdata->spec_pull_set(
> > > > +				mtk_get_regmap(pctl, pin), pin,
> > > > +				pctl->devdata->port_align, isup,
> > > > +				MTK_PUPD_SET_R1R0_00);
> > > > +		}
> > > 
> > > 
> > > it looks like you can use as following snippet with reusing original
> > > logic and less effort on maintaining these common code (?)
> > ==> Thanks for your suggestion very much. It seems better.
> > I will change it in v3.
> 
> ==>
>  Hi sean,
> 
> we try it in our site. we can't change like this.
> Because we will check "arg" after setting "arg = MTK_PUPD_SET_R1R0_00;".
> following snippet which check "arg" value is o or 1. If it not, which
> will print error. so we can't change like this.
> /* For generic pull config, default arg value should be 0 or 1. */
> if (arg != 0 && arg != 1) {
> 	dev_err(pctl->dev, "invalid pull-up argument %d on pin %d .\n", arg,
> pin);
> 	return -EINVAL;
> }
> 
> Thanks.

why not rewrite it with

@@ -301,8 +301,10 @@ static int mtk_pconf_set_pull_select(struct
mtk_pinctrl *pctl,
         * resistor bit, so we need this special handle.
         */
        if (pctl->devdata->spec_pull_set) {
+               unsigned int r1r0 = enable ? arg : MTK_PUPD_SET_R1R0_00;
+
                ret = pctl->devdata->spec_pull_set(mtk_get_regmap(pctl,
pin),
-                       pin, pctl->devdata->port_align, isup, arg);
+                       pin, pctl->devdata->port_align, isup, r1r0);
                if (!ret)
                        return 0;
        }

and also I appeared to find another bug in the original code when I had
a close look at the code path, but it's not related to the patch: 

a special pin would still go through the code path to handle
generic pin when the special pin uses an invalid argument.

It looks like err-code from pctl->devdata->spec_pull_set cannot be
used directly to judge its pin type. It's also possible to use it as an
indication of the real failure during special pin is being set up.

> > > 
> > > 
> > > if (pctl->devdata->spec_pull_set) {
> > > 	+ if (!enable)
> > > 	+ 	arg = MTK_PUPD_SET_R1R0_00;
> > > 
> > > 	ret = pctl->devdata->spec_pull_set(mtk_get_regmap(pctl, pin),
> > > 		pin, pctl->devdata->port_align, isup, arg);
> > > 
> > > 
> > > Also, it's better to add more comments to such kind of special path for
> > > allowing future SoCs to follow and extend more easily.
> > 
> > ==> Thanks for your suggestion very much. I will add more comments here.
> > > 
> > > >  		if (!ret)
> > > >  			return 0;
> > > >  	}
> > > 
> > > 
> > 
> 
>