On 7/10/2020 9:45 AM, Rob Herring wrote: > On Fri, Jul 10, 2020 at 11:06:18AM +0200, Kurt Kanzenbach wrote: >> For future DSA drivers it makes sense to add a generic DSA yaml binding which >> can be used then. This was created using the properties from dsa.txt. It >> includes the ports and the dsa,member property. >> >> Suggested-by: Florian Fainelli <f.fainelli@xxxxxxxxx> >> Signed-off-by: Kurt Kanzenbach <kurt@xxxxxxxxxxxxx> >> --- >> .../devicetree/bindings/net/dsa/dsa.yaml | 80 +++++++++++++++++++ >> 1 file changed, 80 insertions(+) >> create mode 100644 Documentation/devicetree/bindings/net/dsa/dsa.yaml >> >> diff --git a/Documentation/devicetree/bindings/net/dsa/dsa.yaml b/Documentation/devicetree/bindings/net/dsa/dsa.yaml >> new file mode 100644 >> index 000000000000..bec257231bf8 >> --- /dev/null >> +++ b/Documentation/devicetree/bindings/net/dsa/dsa.yaml >> @@ -0,0 +1,80 @@ >> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) >> +%YAML 1.2 >> +--- >> +$id: http://devicetree.org/schemas/net/dsa/dsa.yaml# >> +$schema: http://devicetree.org/meta-schemas/core.yaml# >> + >> +title: Distributed Switch Architecture Device Tree Bindings > > DSA is a Linuxism, right? Not really, it is a Marvell term that describes their proprietary switching protocol. Since then DSA within Linux expands well beyond just Marvell switches, so the terms have been blurred a little bit. > >> + >> +maintainers: >> + - Andrew Lunn <andrew@xxxxxxx> >> + - Florian Fainelli <f.fainelli@xxxxxxxxx> >> + - Vivien Didelot <vivien.didelot@xxxxxxxxx> >> + >> +description: >> + Switches are true Linux devices and can be probed by any means. Once probed, > > Bindings are OS independent. > >> + they register to the DSA framework, passing a node pointer. This node is >> + expected to fulfil the following binding, and may contain additional >> + properties as required by the device it is embedded within. > > Describe what type of h/w should use this binding. > >> + >> +properties: >> + $nodename: >> + pattern: "^switch(@.*)?$" >> + >> + dsa,member: >> + minItems: 2 >> + maxItems: 2 >> + description: >> + A two element list indicates which DSA cluster, and position within the >> + cluster a switch takes. <0 0> is cluster 0, switch 0. <0 1> is cluster 0, >> + switch 1. <1 0> is cluster 1, switch 0. A switch not part of any cluster >> + (single device hanging off a CPU port) must not specify this property >> + $ref: /schemas/types.yaml#/definitions/uint32-array >> + >> + ports: >> + type: object >> + properties: >> + '#address-cells': >> + const: 1 >> + '#size-cells': >> + const: 0 >> + >> + patternProperties: >> + "^port@[0-9]+$": > > As ports and port are OF graph nodes, it would be better if we > standardized on a different name for these. I think we've used > 'ethernet-port' some. Yes we did talk about that before, however when the original DSA binding was introduced about 7 years ago (or maybe more recently, my memory fails me now), "ports" was chosen as the encapsulating node. We should be accepting both ethernet-ports and ports. > >> + type: object >> + description: DSA switch ports >> + >> + allOf: >> + - $ref: ../ethernet-controller.yaml# > > How does this and 'ethernet' both apply? I think the intent here was to mean that some of the properties from the Ethernet controller such as phy-mode, phy-handle, fixed-link also apply here since the switch port is a simplified Ethernet MAC on a number of counts. -- Florian
