[PATCH 2/9] dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy
Varadarajan Narayanan
quic_varada at quicinc.com
Wed Jun 14 22:27:47 PDT 2023
On Wed, Jun 07, 2023 at 08:31:50PM +0200, Krzysztof Kozlowski wrote:
> On 07/06/2023 12:56, Varadarajan Narayanan wrote:
> > Document the M31 USB2 phy present in IPQ5332
>
> Full stop.
Ok.
> > Signed-off-by: Sricharan Ramabadhran <quic_srichara at quicinc.com>
> > Signed-off-by: Varadarajan Narayanan <quic_varada at quicinc.com>
> > ---
> > .../devicetree/bindings/phy/qcom,m31.yaml | 69 ++++++++++++++++++++++
> > 1 file changed, 69 insertions(+)
> > create mode 100644 Documentation/devicetree/bindings/phy/qcom,m31.yaml
> >
> > diff --git a/Documentation/devicetree/bindings/phy/qcom,m31.yaml b/Documentation/devicetree/bindings/phy/qcom,m31.yaml
> > new file mode 100644
> > index 0000000..8ad4ba4
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/phy/qcom,m31.yaml
> > @@ -0,0 +1,69 @@
> > +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
> > +
>
> Drop stray blank lines.
Ok.
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/phy/qcom,m31.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: Qualcomm M31 USB PHY
>
> What is M31?
M31 is an external IP integrated into IPQ5332 SoC.
> > +
> > +maintainers:
> > + - Sricharan Ramabadhran <quic_srichara at quicinc.com>
> > + - Varadarajan Narayanan <quic_varada at quicinc.org>
> > +
> > +description:
> > + This file contains documentation for the USB M31 PHY found in Qualcomm
>
> Drop redundant parts ("This file contains documentation for the").
Ok.
> > + IPQ5018, IPQ5332 SoCs.
> > +
> > +properties:
> > + compatible:
> > + oneOf:
>
> That's not oneOf.
Ok.
> > + - items:
>
> No items.
Ok.
> > + - enum:
> > + - qcom,m31-usb-hsphy
>
> I am confused what's this. If m31 is coming from some IP block provider,
> then you are using wrong vendor prefix.
> https://www.m31tech.com/download_file/M31_USB.pdf
>
>
> > + - qcom,ipq5332-m31-usb-hsphy
>
> This confuses me even more. IPQ m31?
Will change this to m31,usb-hsphy and m31,ipq5332-usb-hsphy respectively.
Will that be acceptable?
> > +
> > + reg:
> > + description:
> > + Offset and length of the M31 PHY register set
>
> Drop description, obvious.
Ok.
> > + maxItems: 2
> > +
> > + reg-names:
> > + items:
> > + - const: m31usb_phy_base
> > + - const: qscratch_base
>
> Drop "_base" from both.
Ok. Will drop qscratch_base. This is in the controller space.
Should not come here.
> > +
> > + phy_type:
> > + oneOf:
> > + - items:
> > + - enum:
> > + - utmi
> > + - ulpi
>
> This does not belong to phy, but to USB node.
This is used by the driver to set a bit during phy init. Hence
have it as a replication of the USB node's entry. If this is not
permissible, is there some way to get this from the USB node,
or any other alternative mechanism?
> > +
> > + resets:
> > + maxItems: 1
> > + description:
> > + List of phandles and reset pairs, one for each entry in reset-names.
>
> Drop useless description.
Ok.
> > +
> > + reset-names:
> > + items:
> > + - const:
> > + usb2_phy_reset
>
> Drop entire reset-names.
Ok.
> > +
> > +additionalProperties: false
> > +
> > +examples:
> > + - |
> > + #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
> > + hs_m31phy_0: hs_m31phy at 5b00 {
>
> Node names should be generic. See also explanation and list of examples
> in DT specification:
> https://devicetree-specification.readthedocs.io/en/latest/chapter2-devicetree-basics.html#generic-names-recommendation
>
> Also, no underscores in node names.
Will change this as usbphy0:hs_m31phy at 7b000
> > + compatible = "qcom,m31-usb-hsphy";
> > + reg = <0x5b000 0x3000>,
>
> This is not what your unit address is saying.
Will change to 0x0007b000.
> > + <0x08af8800 0x400>;
>
> Align it.
Will drop this. This is in the controller space.
> > + reg-names = "m31usb_phy_base",
> > + "qscratch_base";
>
> Align it.
Ok.
> > + phy_type = "utmi";
> > + resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
> > + reset-names = "usb2_phy_reset";
> > +
> > + status = "ok";
>
> Drop.
Ok. Hopefully will get an alternative for the phy_type.
Thanks
Varada
> > + };
>
> Best regards,
> Krzysztof
>
More information about the linux-phy
mailing list