Re: [PATCH v15 1/8] dt-bindings: mfd: Initial commit of silergy,sy7636a.yaml

[Alistair Francis]

On Fri, Nov 26, 2021 at 8:59 AM Andreas Kemnade <andreas@xxxxxxxxxxxx> wrote:
>
> On Tue, 23 Nov 2021 23:29:26 +1000
> Alistair Francis <alistair23@xxxxxxxxx> wrote:
>
> > On Thu, Nov 18, 2021 at 7:40 AM Andreas Kemnade <andreas@xxxxxxxxxxxx> wrote:
> > >
> > > On Wed, 10 Nov 2021 22:29:41 +1000
> > > Alistair Francis <alistair@xxxxxxxxxxxxx> wrote:
> > >
> > > > Initial support for the Silergy SY7636A Power Management chip
> > > > and regulator.
> > > >
> > > > Signed-off-by: Alistair Francis <alistair@xxxxxxxxxxxxx>
> > > > Reviewed-by: Rob Herring <robh@xxxxxxxxxx>
> > > > Acked-for-MFD-by: Lee Jones <lee.jones@xxxxxxxxxx>
> > > > ---
> > > >  .../bindings/mfd/silergy,sy7636a.yaml         | 79 +++++++++++++++++++
> > > >  1 file changed, 79 insertions(+)
> > > >  create mode 100644 Documentation/devicetree/bindings/mfd/silergy,sy7636a.yaml
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/mfd/silergy,sy7636a.yaml b/Documentation/devicetree/bindings/mfd/silergy,sy7636a.yaml
> > > > new file mode 100644
> > > > index 000000000000..0566f9498e2f
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/mfd/silergy,sy7636a.yaml
> > > [...]
> > > > +  regulators:
> > > > +    type: object
> > > > +
> > > > +    properties:
> > > > +      compatible:
> > > > +        const: silergy,sy7636a-regulator
> > > > +
> > > > +      vcom:
> > > > +        type: object
> > > > +        $ref: /schemas/regulator/regulator.yaml#
> > > > +        properties:
> > > > +          regulator-name:
> > > > +            const: vcom
> > > > +
> > > hmm, this is what? If I understand it correctly, vcom means some
> > > voltage for compensation. On other comparable pmics (e.g. TPS65185
> > > which has also a sane public datasheet, MAX17135) I have seen some
> > > methods to measure a voltage while the display is doing something
> > > defined and then program this voltage non-volatile for compensation
> > > during manufacturing.
> > >
> > > If I understand the code correctly all the bunch of voltages are
> > > powered up if this one is enabled.
> > > So at least a description should be suitable.
> > >
> > > The other comparable PMICs have at least regulators named VCOM, DISPLAY
> > > (controls several regulators, started with delays configured via
> > > registers) and V3P3. MAX17135 source can be found in NXP kernels,
> > > TPS65185 in Kobo vendor kernels.
> > >
> > > So I would expect to see something similar here and a description or at
> > > least not such a misleading name as vcom if it is for some reason not
> > > feasible to separate the regulators.
> >
> > This is a vcom in the sense of voltage for compensation. We just
> > currently don't support setting the vcom.
> >
> > I had a look at the Kobo code and this is similar to
> > https://github.com/akemnade/linux/blob/kobo/epdc-pmic-5.15/drivers/regulator/sy7636-regulator.c#L614
> >
> > So I think that vcom is still the appropriate name for this.
> >
> seems that you did not get me. If I understand the code behind it
> correctly, it turns on all power rails (the +-15V stuff, VEE and so on)
> with the defined delays, not just vcom because it sets
> SY7636A_OPERATION_MODE_CRL_ONOFF. Controlling VCOM separately is possible
> by using SY7636A_OPERATION_MODE_CRL_VCOMCTL in combintion with a
> vcom_en gpio.
>
> I do not see a reason to turn on vcom only without the other higher
> voltage rails, so the behaviour is not necessarily wrong but if I read
> the binding documentation I would expect that just vcom is turned on.
> That is the mismatch I am talking about.

Ah! Ok I understand. I'll rename it to vdd then.

Alistair

>
> If we agree on this idea that one regulator is enabling everything, I
> would adapt my EPDC drm driver and tps65185 driver (which are both in
> my clean up to be upstreamable-queue).
>
> Just another thing to compare with:
> https://github.com/Freescale/linux-fslc/blob/4.1-2.0.x-imx/drivers/regulator/max17135-regulator.c
> that seems to be the starting point for kobo vendor kernel epd pmics.
> They seem to have taken the source and modified things. There we have
> regulators with empty ops for that step-up converted stuff, a separate
> vcom and a display regulator which really controls all of these step-up
> things.
>
> Regards,
> Andreas