Re: [RFC PATCH v2 3/3] mmc: renesas_sdhi: Add support for RZ/V2H(P) SoC

["Lad, Prabhakar" <prabhakar.csengg@gmail.com>]

Hi Biju,

On Thu, Jun 20, 2024 at 10:49 AM Biju Das <biju.das.jz@bp.renesas.com> wrote:
>
> Hi Prabhakar,
>
> > -----Original Message-----
> > From: Lad, Prabhakar <prabhakar.csengg@gmail.com>
> > Subject: Re: [RFC PATCH v2 3/3] mmc: renesas_sdhi: Add support for RZ/V2H(P) SoC
> >
> > Hi Biju,
> >
> > On Thu, Jun 20, 2024 at 10:30 AM Biju Das <biju.das.jz@bp.renesas.com> wrote:
> > >
> > > Hi Wolfram, Prabhakar,
> > >
> > > > -----Original Message-----
> > > > From: Wolfram Sang <wsa+renesas@sang-engineering.com>
> > > > Sent: Thursday, June 20, 2024 8:40 AM
> > > > Subject: Re: [RFC PATCH v2 3/3] mmc: renesas_sdhi: Add support for
> > > > RZ/V2H(P) SoC
> > > >
> > > > Hi Prabhakar,
> > > >
> > > > > I did give it a try with platform_driver_probe() and failed.
> > > >
> > > > Ok, thanks for trying nonetheless!
> > > >
> > > > > - Firstly I had to move the regulator node outside the SDHI node
> > > > > for
> > > > > platform_driver_probe() to succeed or else it failed with -ENODEV
> > > > > (at
> > > > > https://elixir.bootlin.com/linux/latest/source/drivers/base/platfo
> > > > > rm.c
> > > > > #L953)
> > > >
> > > > This makes sense to me because it is just a "regular" regulator.
> > > >
> > > > > - In Renesas SoCs we have multiple instances of SDHI, the problem
> > > > > being for each instance we are calling platform_driver_probe().
> > > > > Which causes a problem as the regulator node will use the first device.
> > > >
> > > > I see... we would need a reg property to differentiate between the
> > > > internal regulators but that is already used by the parent SDHI node.
> > > >
> > > > Okay, so let's scrap that idea. However, we need to ensure that we
> > > > can still have an external regulator. Seeing the bindings, it looks
> > > > like you enable the internal regulator with the "vqmmc- r9a09g057-regulator"
> > > > property? I wonder now if we can simplify this to an
> > > > "use-internal-regulator" property because we have 'compatible' already to differentiate? Needs
> > advice from DT maintainers, probably.
> > >
> > > Why this cannot be modelled as a regular "regulator" as a child device of SDHI device?
> > >
> > The current implementation does implement the regulator as a child device of the sdhi node [0]
> > itself.
> >
> > Wolfram was suggesting to have the regulator outside and use platform_driver_probe(), which caused
> > an issue as mentioned above.
>
> You, mean standalone node with a device compatible for each SDHI device nodes(Assuming 3 sdhi devices)?
>
Yep.

> 3 SDHI devices nodes(stand alone) + 3 regulator device nodes (stand alone) ?
>
This one (since as per the HW we have three SDHI instances and 3
internal regulators) so we need to describe the same in DT.

Cheers,
Prabhakar