Re: [PATCH 1/2] dt-bindings: phy: imx8mq-usb: improve some tuning properties

[Xu Yang <xu.yang_2@nxp.com>]

On Mon, Apr 28, 2025 at 09:59:42AM +0200, Krzysztof Kozlowski wrote:
> On Wed, Apr 23, 2025 at 05:46:06PM GMT, Xu Yang wrote:
> > Commit b2e75563dc39 ("dt-bindings: phy: imx8mq-usb: add phy tuning
> > properties") add many tuning properties, but some parameter value doesn't
> > match the register description. It made some changes based on the original
> > value: add offset to a negative number so turn it to a non-negative number.
> > However, this conversion bring many pain when some properties need to be
> > improved.
> > 
> > Because device-tree supports negative parameter number, this will improve
> > some propertie's parameter.
> > 
> > Mainly include below properties:
> >  - fsl,phy-tx-vref-tune-percent
> >  - fsl,phy-tx-rise-tune-percent
> >  - fsl,phy-comp-dis-tune-percent
> > 
> > The parameter value of above 3 properties are USB PHY specific. i.MX8MP
> > and i.MX95 USB PHY has different meanings. So this add restrictions for
> > them.
> > 
> >  - fsl,phy-tx-vboost-level-microvolt
> > 
> > For this property, the parameter value is wrong in register description.
> > This will correct it according to true value.
> > 
> > For detailed info, please refer to i.MX8MP and i.MX95 latest reference
> > manual.
> > 
> > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > ---
> >  .../bindings/phy/fsl,imx8mq-usb-phy.yaml      | 43 +++++++++++++++----
> >  1 file changed, 35 insertions(+), 8 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/phy/fsl,imx8mq-usb-phy.yaml b/Documentation/devicetree/bindings/phy/fsl,imx8mq-usb-phy.yaml
> > index daee0c0fc915..b09e3dab4c59 100644
> > --- a/Documentation/devicetree/bindings/phy/fsl,imx8mq-usb-phy.yaml
> > +++ b/Documentation/devicetree/bindings/phy/fsl,imx8mq-usb-phy.yaml
> > @@ -43,15 +43,11 @@ properties:
> >    fsl,phy-tx-vref-tune-percent:
> >      description:
> >        Tunes the HS DC level relative to the nominal level
> > -    minimum: 94
> > -    maximum: 124
> 
> Widest constraints should always stay here.

If so, I will restore and modify them in v2.

> 
> >  
> >    fsl,phy-tx-rise-tune-percent:
> >      description:
> >        Adjusts the rise/fall time duration of the HS waveform relative to
> >        its nominal value
> > -    minimum: 97
> > -    maximum: 103
> >  
> >    fsl,phy-tx-preemp-amp-tune-microamp:
> >      description:
> > @@ -63,15 +59,12 @@ properties:
> >    fsl,phy-tx-vboost-level-microvolt:
> >      description:
> >        Adjust the boosted transmit launch pk-pk differential amplitude
> > -    minimum: 880
> > -    maximum: 1120
> > +    enum: [844, 1008, 1156]
> >  
> >    fsl,phy-comp-dis-tune-percent:
> >      description:
> >        Adjust the voltage level used to detect a disconnect event at the host
> >        relative to the nominal value
> > -    minimum: 91
> > -    maximum: 115
> >  
> >    fsl,phy-pcs-tx-deemph-3p5db-attenuation-db:
> >      description:
> > @@ -112,6 +105,40 @@ allOf:
> >          reg:
> >            maxItems: 1
> >  
> > +  - if:
> > +      properties:
> > +        compatible:
> > +          enum:
> > +            - fsl,imx8mq-usb-phy
> > +            - fsl,imx8mp-usb-phy
> > +    then:
> > +      properties:
> > +        fsl,phy-tx-vref-tune-percent:
> > +          minimum: -6
> > +          maximum: 24
> 
> That's ABI break with vague reason "some pain".

For these four properties, no DTS files use them now. So that's why I'm
going to improve them. I think no ABI break here.

 - fsl,phy-tx-vref-tune-percent
 - fsl,phy-tx-rise-tune-percent
 - fsl,phy-comp-dis-tune-percent
 - fsl,phy-tx-vboost-level-microvolt

Thanks,
Xu Yang

> 
> Best regards,
> Krzysztof
>