Re: [PATCH 3/3] iio: magnetometer: ak8975: Add gpio reset support

[Andy Shevchenko <andy.shevchenko@gmail.com>]

On Mon, May 18, 2020 at 8:43 PM Jonathan Albrieux
<jonathan.albrieux@gmail.com> wrote:
>
> On Mon, May 18, 2020 at 07:43:17PM +0300, Andy Shevchenko wrote:
> > On Mon, May 18, 2020 at 06:01:20PM +0200, Jonathan Albrieux wrote:
> > > On Mon, May 18, 2020 at 05:55:51PM +0300, Andy Shevchenko wrote:
> > > > On Mon, May 18, 2020 at 4:38 PM Jonathan Albrieux
> > > > <jonathan.albrieux@gmail.com> wrote:
> > > >
> > > > > +       gpiod_set_value_cansleep(data->reset_gpiod, 1);
> > > >
> > > > (1)
> > > >
> > > > ...
> > > >
> > > > > +       /*
> > > > > +        * If reset pin is provided then will be set to high on power on
> > > > > +        * and to low on power off according to AK09911 datasheet
> > > > > +        */
> > > >
> > > > Wording is confusing, perhaps you have to use 'asserted / deasserted'.
> > >
> > > Thank you for the suggestion, I'll be working on rewording as soon as
> > > possible.
> > >
> > > > Btw, in (1) it's also "high" (asserted). I barely understand how it's
> > > > supposed to work in all cases?
> > > >
> > > > > +       reset_gpiod = devm_gpiod_get_optional(&client->dev,
> > > > > +                                             "reset", GPIOD_OUT_HIGH);
> > > > > +       if (IS_ERR(reset_gpiod))
> > > > > +               return PTR_ERR(reset_gpiod);
> > > >
> > >
> > > I'm sorry but I'm not sure about what you mean by saying all cases.
> > > Currently  I'm testing this driver on a msm8916 device having AK09911
> > > magnetometer. At the current stage the driver is failing on probe
> > > because reset pin is not connected to VID (as datasheet requires in case
> > > of pin not being used). In case of reset pin not asserted, register's
> > > reset is triggered resulting in empty registers, leading to probe fail.
> > > For this reason pin is asserted during power on in order to have
> > > informations in registers and deasserted before power off triggering
> > > a reset.
> > >
> > > A workaround that gets AK09911 working on device is by setting the
> > > reset pin always high on device tree. This way registers gets reset by
> > > a Power On Reset circuit autonomously and reset pin never triggers the
> > > reset.
> >
> > You need to distinguish electrical level from logical (GPIO flag defines
> > logical). So, I'm talking about active-high vs. active-low case.
> >
> > Now I re-read above, and see that here you assert the reset signal. But where
> > is desertion?
>
> Oh I see, I'll try explaining by points the proposed approach:
> - reset pin is active low
> - during power on gpio is set to 0 so the reset pin is high, thus no reset

deasserted

> - during power off gpio is set to 1 so the reset pin becomes low, thus resetting

asserted

> this is a possible solution but maybe there are other ways to achieve that,
> do you have suggestions on how to get a better approach for solving this issue?

I see now, that at requesting reset you wanted to chip be in reset
state (asserted) till driver calls power_on().

Seems everything you done is correct. Just correct terminology, please.

-- 
With Best Regards,
Andy Shevchenko