Re: [PATCH v2 phy-next 13/15] dt-bindings: phy: lynx-10g: initial document

[Vladimir Oltean]

Hi Alexander,

On Mon, Jun 01, 2026 at 08:34:25AM +0200, Alexander Stein wrote:
> Hi,
> 
> Am Freitag, 29. Mai 2026, 19:15:07 CEST schrieb Vladimir Oltean:
> > Add a schema for the 10G Lynx SerDes. This is very similar to the modern
> > form of the 28G Lynx SerDes, which is very much the intention.
> > 
> > We allow both forms of #phy-cells = <1> in the top-level provider
> > and #phy-cells = <0> in the per-lane provider for more flexibility to
> > consumers, and because the kernel code is shared with the 28G Lynx which
> > already has that support for compatibility reasons.
> > 
> > Signed-off-by: Vladimir Oltean <vladimir.oltean@xxxxxxx>
> > ---
> > Cc: devicetree@xxxxxxxxxxxxxxx
> > Cc: Conor Dooley <conor+dt@xxxxxxxxxx>
> > Cc: Krzysztof Kozlowski <krzk+dt@xxxxxxxxxx>
> > Cc: Rob Herring <robh@xxxxxxxxxx>
> > 
> > v1->v2:
> > - move patch later in series, right before driver
> > - deliberately ignoring this Sashiko feedback:
> >   https://lore.kernel.org/linux-phy/20260529125017.ifqunh52gdzhthdg@skbuf/
> > ---
> >  .../devicetree/bindings/phy/fsl,lynx-10g.yaml | 131 ++++++++++++++++++
> >  1 file changed, 131 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/phy/fsl,lynx-10g.yaml
> > 
> > diff --git a/Documentation/devicetree/bindings/phy/fsl,lynx-10g.yaml b/Documentation/devicetree/bindings/phy/fsl,lynx-10g.yaml
> > new file mode 100644
> > index 000000000000..993f076bba4e
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/phy/fsl,lynx-10g.yaml
> > @@ -0,0 +1,131 @@
> > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/phy/fsl,lynx-10g.yaml
> > +$schema: http://devicetree.org/meta-schemas/core.yaml
> > +
> > +title: Freescale Lynx 10G SerDes PHY
> > +
> > +maintainers:
> > +  - Vladimir Oltean <vladimir.oltean@xxxxxxx>
> > +
> > +description:
> > +  The 10G Lynx is a multi-protocol SerDes block which handles networking, PCIe,
> > +  SATA and other high-speed interfaces. It is present on most QorIQ and
> > +  Layerscape SoCs. The register map is common, but the integration is
> > +  SoC-specific, with the differences consisting in register endianness, the
> > +  number of lanes, protocol converters available per lane and their location in
> > +  the PCCR registers. Some SoCs have multiple SerDes blocks and those differ in
> > +  their protocol capabilities per lane.
> > +
> > +properties:
> > +  compatible:
> > +    description:
> > +      There is intentionally no generic fsl,lynx-10g compatible string due to
> > +      the hardware inability to report its capabilities, despite having a
> > +      common register map.
> > +    enum:
> > +      - fsl,ls1028a-serdes
> > +      - fsl,ls1046a-serdes1
> > +      - fsl,ls1046a-serdes2
> > +      - fsl,ls1088a-serdes1
> > +      - fsl,ls1088a-serdes2
> > +      - fsl,ls2088a-serdes1
> > +      - fsl,ls2088a-serdes2
> 
> Silly question: What about LS1043A? AFAIK it has a single serdes block.
> 
> Best regards
> Alexander

My understanding is that hardware validation for LS1043A was not
budgeted for the main two features why the lynx-10g driver is necessary:
RCW override for 1G <-> 10G dynamic protocol switching and KR link
training. As such, this SoC isn't supported by the SerDes driver in the
NXP BSP either. With the exception of 1G <-> 2.5G minor protocol
switching, having a lynx-10g driver would not be very useful for the
LS1043A as is, without a procedure from h/w validation to do RCW
override.