Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 30 May 2023 07:56:57 +0100
From:   Conor Dooley <conor.dooley@microchip.com>
To:     Stanley =?utf-8?B?Q2hhbmdb5piM6IKy5b63XQ==?= 
        <stanley_chang@realtek.com>
CC:     Conor Dooley <conor@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Vinod Koul <vkoul@kernel.org>,
        Kishon Vijay Abraham I <kishon@kernel.org>,
        Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Conor Dooley <conor+dt@kernel.org>,
        Alan Stern <stern@rowland.harvard.edu>,
        Flavio Suligoi <f.suligoi@asem.it>,
        Mathias Nyman <mathias.nyman@linux.intel.com>,
        Douglas Anderson <dianders@chromium.org>,
        Matthias Kaehlcke <mka@chromium.org>,
        Ray Chi <raychi@google.com>,
        Michael Grzeschik <m.grzeschik@pengutronix.de>,
        "linux-phy@lists.infradead.org" <linux-phy@lists.infradead.org>,
        "devicetree@vger.kernel.org" <devicetree@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>
Subject: Re: [PATCH v2 3/3] dt-bindings: phy: realtek: Add the doc about the
 Realtek SoC USB 2.0/3.0 PHY
Message-ID: <20230530-monsieur-dumpling-32c102f78c56@wendy>
References: <20230525022617.30537-1-stanley_chang@realtek.com>
 <20230525022617.30537-3-stanley_chang@realtek.com>
 <20230529-impurity-dismount-bca5c9100c9b@spud>
 <44015844858c42a79e0e7f9207d01496@realtek.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
        protocol="application/pgp-signature"; boundary="W3NXbn335O8/41NE"
Content-Disposition: inline
In-Reply-To: <44015844858c42a79e0e7f9207d01496@realtek.com>
Precedence: bulk

--W3NXbn335O8/41NE
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, May 30, 2023 at 03:08:29AM +0000, Stanley Chang[=E6=98=8C=E8=82=B2=
=E5=BE=B7] wrote:
> Hi Conor,
>=20
> > > +properties:
> > > +  compatible:
> > > +    enum:
> > > +      - realtek,usb2phy
> > > +      - realtek,rtd-usb2phy
> > > +      - realtek,rtd1295-usb2phy
> > > +      - realtek,rtd1395-usb2phy
> > > +      - realtek,rtd1619-usb2phy
> > > +      - realtek,rtd1319-usb2phy
> > > +      - realtek,rtd1619b-usb2phy
> > > +      - realtek,rtd1312c-usb2phy
> > > +      - realtek,rtd1319d-usb2phy
> > > +      - realtek,rtd1315e-usb2phy
>=20
> > > +properties:
> > > +  compatible:
> > > +    enum:
> > > +      - realtek,usb3phy
> > > +      - realtek,rtd-usb3phy
> > > +      - realtek,rtd1295-usb3phy
> > > +      - realtek,rtd1619-usb3phy
> > > +      - realtek,rtd1319-usb3phy
> > > +      - realtek,rtd1619b-usb3phy
> > > +      - realtek,rtd1319d-usb3phy
>=20
> > Ignoring everything else, because I really want Krzysztof or Rob to
> > review this rather than me, but what's going on here with the
> > compatibles?
> > What hardware do "usbNphy" and "rtd-usbNphy" represent?
> >=20
> > You have device-specific compatibles, which is great, but you also allow
> > only those two generic ones. I had a _brief_ look at the driver, and it
> > seems like there is no decision making done based on the compatibles,
> > only on the properties. Is that correct?
> > If it is, I would understand having "realtek,usb3phy" as a fallback
> > compatible for "realtek,rtd1619-usb3phy", but I do not get the current
> > setup.
>=20
> This driver is compatible with all Realtek RTD SoCs without specifying di=
fferent settings.
> So use "realtek,usb3phy" as fallback compatible for all SoCs.
> This is the compatible name we use.
> Other compatible names simply indicate that the driver supports the SoCs.

Then you should write the binding such that having fallback compatibles
is permitted. Try plugging
compatible =3D "realtek,rtd1295-usb2phy", "realtek,rtd-usb2phy", "realtek,u=
sb2phy";
into your example below and see what happens.

> The name "usbNphy" and "rtd-usbNphy" seem to be more generic for all RTD =
SoCs,
> but they are not device-specific compatible.
> Do you have a better suggestion?

Write the binding so that having fallback compatibles in the DT actually
works, don't add the SoC-specific ones merely as indicators that those
SoCs are supported and don't permit "realtek,usbNphy" or
"realtek,rtd-usbNphy" in isolation ;)

Cheers,
Conor.

--W3NXbn335O8/41NE
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----

iHUEABYIAB0WIQRh246EGq/8RLhDjO14tDGHoIJi0gUCZHWeKgAKCRB4tDGHoIJi
0gUMAQDaAWRa95AFdcTrun9KJrWgJmpk4c2Jh03zukk+uTrWkQEAp5S58xu4+voq
u38k5+Xn+V6Fiovzr8Qywo1BJggf1Q0=
=a78/
-----END PGP SIGNATURE-----

--W3NXbn335O8/41NE--