Re: [PATCH v6 1/3] dt-bindings: phy: Add DT bindings for Xilinx ZynqMP PSGTR PHY

From: Rob Herring
Date: Fri Mar 20 2020 - 12:53:38 EST


On Fri, Mar 20, 2020 at 3:50 AM Laurent Pinchart
<laurent.pinchart@xxxxxxxxxxxxxxxx> wrote:
>
> Hi Rob,
>
> On Thu, Mar 19, 2020 at 08:35:20PM -0600, Rob Herring wrote:
> > On Wed, Mar 11, 2020 at 12:32:50PM +0200, Laurent Pinchart wrote:
> > > From: Anurag Kumar Vulisha <anurag.kumar.vulisha@xxxxxxxxxx>
> > >
> > > Add DT bindings for the Xilinx ZynqMP PHY. ZynqMP SoCs have a High Speed
> > > Processing System Gigabit Transceiver which provides PHY capabilities to
> > > USB, SATA, PCIE, Display Port and Ehernet SGMII controllers.
> > >
> > > Signed-off-by: Anurag Kumar Vulisha <anurag.kumar.vulisha@xxxxxxxxxx>
> > > Signed-off-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> > > ---
> > > Changes since v5:
> > >
> > > - Document clocks and clock-names properties
> > > - Document resets and reset-names properties
> > > - Replace subnodes with an additional entry in the PHY cells
> > > - Drop lane frequency PHY cell, replaced by reference clock phandle
> > > - Convert bindings to YAML
> > > - Reword the subject line
> > > - Drop Rob's R-b as the bindings have significantly changed
> > > - Drop resets and reset-names properties
> > > ---
> > > .../bindings/phy/xlnx,zynqmp-psgtr.yaml | 104 ++++++++++++++++++
> > > include/dt-bindings/phy/phy.h | 1 +
> > > 2 files changed, 105 insertions(+)
> > > create mode 100644 Documentation/devicetree/bindings/phy/xlnx,zynqmp-psgtr.yaml
> > >
> > > diff --git a/Documentation/devicetree/bindings/phy/xlnx,zynqmp-psgtr.yaml b/Documentation/devicetree/bindings/phy/xlnx,zynqmp-psgtr.yaml
> > > new file mode 100644
> > > index 000000000000..9948e4a60e45
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/phy/xlnx,zynqmp-psgtr.yaml
> > > @@ -0,0 +1,104 @@
> > > +# SPDX-License-Identifier: GPL-2.0
> >
> > For new bindings:
> >
> > (GPL-2.0-only OR BSD-2-Clause)
> >
> > Though I guess Anurag needs to agree.
>
> There's an ongoing similar discussion regarding the DPSUB (DRM/KMS)
> bindings. Hyun is checking with the Xilinx legal department. If they
> agree, I'll change the license here, otherwise I'll keep it as-is.

TBC, the choice is change it or take your toys elsewhere and play by
yourself. I don't really want to end up with whatever each submitter
desires. I don't expect there's many companies that object to a
permissive license.

> > > +%YAML 1.2
> > > +---
> > > +$id: http://devicetree.org/schemas/phy/xlnx,zynqmp-psgtr.yaml#
> > > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > > +
> > > +title: Xilinx ZynqMP Gigabit Transceiver PHY Device Tree Bindings
> > > +
> > > +maintainers:
> > > + - Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> > > +
> > > +description: |
> > > + This binding describes the Xilinx ZynqMP Gigabit Transceiver (GTR) PHY. The
> > > + GTR provides four lanes and is used by USB, SATA, PCIE, Display port and
> > > + Ethernet SGMII controllers.
> > > +
> > > +properties:
> > > + "#phy-cells":
> > > + const: 4
> > > + description: |
> > > + The cells contain the following arguments.
> > > +
> > > + - description: The GTR lane
> > > + minimum: 0
> > > + maximum: 3
> > > + - description: The PHY type
> > > + enum:
> > > + - PHY_TYPE_DP
> > > + - PHY_TYPE_PCIE
> > > + - PHY_TYPE_SATA
> > > + - PHY_TYPE_SGMII
> > > + - PHY_TYPE_USB
> > > + - description: The PHY instance
> > > + minimum: 0
> > > + maximum: 1 # for DP, SATA or USB
> > > + maximum: 3 # for PCIE or SGMII
> > > + - description: The reference clock number
> > > + minimum: 0
> > > + maximum: 3
> >
> > Humm, interesting almost json-schema. I guess it's fine as-is.
> >
> > I would like to figure out how to apply a schema like this to the
> > consumer nodes. We'd have to look up the phandle, get that node's
> > compatible, find the provider's schema, find #.*-cells property, and
> > extract a schema from it. Actually, doesn't sound too hard.
>
> That would be nice :-)
>
> > > +
> > > + compatible:
> > > + enum:
> > > + - xlnx,zynqmp-psgtr-v1.1
> > > + - xlnx,zynqmp-psgtr
> > > +
> > > + clocks:
> > > + minItems: 1
> > > + maxItems: 4
> > > + description: |
> > > + Clock for each PS_MGTREFCLK[0-3] reference clock input. Unconnected
> > > + inputs shall not have an entry.
> > > +
> > > + clock-names:
> > > + minItems: 1
> > > + maxItems: 4
> > > + items:
> > > + pattern: "^ref[0-3]$"
> > > +
> > > + reg:
> > > + items:
> > > + - description: SERDES registers block
> > > + - description: SIOU registers block
> > > +
> > > + reg-names:
> > > + items:
> > > + - const: serdes
> > > + - const: siou
> > > +
> > > +required:
> > > + - "#phy-cells"
> > > + - compatible
> > > + - reg
> > > + - reg-names
> > > +
> > > +if:
> > > + properties:
> > > + compatible:
> > > + const: xlnx,zynqmp-psgtr
> > > +
> > > +then:
> > > + properties:
> > > + xlnx,tx-termination-fix:
> > > + description: |
> > > + Include this for fixing functional issue with the TX termination
> > > + resistance in GT, which can be out of spec for the XCZU9EG silicon
> > > + version.
> > > + type: boolean
> > > +
> > > +additionalProperties: false
> >
> > This won't work with 'xlnx,tx-termination-fix'. You need to move it to
> > the main properties section and then do:
> >
> > if:
> > properties:
> > compatible:
> > const: xlnx,zynqmp-psgtr-v1.1
>
> It doesn't make a big difference as only two compatible values are
> allowed, but is there a way to express the condition the other way
> around, if (compatible != "xlnx,zynqmp-psgtr") ?

if:
properties:
compatible:
not:
const: xlnx,zynqmp-psgtr

I think if: { not: ... } would also work. You'll have to test them out.

> > then:
> > properties:
> > xlnx,tx-termination-fix: false
>
> This works.
>
> > I think this would also work:
> >
> > not:
> > required:
> > - xlnx,tx-termination-fix
>
> I've tested it and it works, but I'm not sure why, given that the
> property isn't required required in the first place. Could you enlighten
> me ?

'required' is true or false based on presence or absence of properties
in the list. If 'xlnx,tx-termination-fix' is present, then 'required'
evaluates to true. And the inverse is true. Then we take the inverse
of of that with 'not'.

The first case is what trips me up more because a property not present
evaluates to true. So if you look at 'select' schemas, we have to make
any properties we list (compatible typically) required.

Rob