Re: [PATCH 2/3] dt-bindings: Document BCM6328 PCIe Host Controller

[Florian Fainelli <f.fainelli@xxxxxxxxx>]

On 6/17/2020 3:25 AM, Álvaro Fernández Rojas wrote:
> BCM6328 PCIe host controller is found on BCM6328, BCM6362 and BCM63268 SoCs.
> 
> Signed-off-by: Álvaro Fernández Rojas <noltari@xxxxxxxxx>
> ---
>  .../bindings/pci/brcm,bcm6328-pcie.yaml       | 109 ++++++++++++++++++
>  1 file changed, 109 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/pci/brcm,bcm6328-pcie.yaml
> 
> diff --git a/Documentation/devicetree/bindings/pci/brcm,bcm6328-pcie.yaml b/Documentation/devicetree/bindings/pci/brcm,bcm6328-pcie.yaml
> new file mode 100644
> index 000000000000..d2bd4933a5fa
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pci/brcm,bcm6328-pcie.yaml
> @@ -0,0 +1,109 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/pci/brcm,bcm6328-pcie.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: BCM6328 PCIe Host Controller Device Tree Bindings
> +
> +maintainers:
> +  - Álvaro Fernández Rojas <noltari@xxxxxxxxx>
> +
> +allOf:
> +  - $ref: /schemas/pci/pci-bus.yaml#
> +
> +properties:
> +  compatible:
> +    enum:
> +     - brcm,bcm6328-pcie
> +     - brcm,bcm6362-pcie
> +     - brcm,bcm63268-pcie
> +
> +  clocks:
> +    maxItems: 1
> +
> +  clock-names:
> +    items:
> +      - const: pcie
> +
> +  "#interrupt-cells":
> +    const: 1
> +
> +  interrupt-map-mask:
> +    maxItems: 1
> +
> +  interrupt-map:
> +    maxItems: 1
> +
> +  power-domains:
> +    maxItems: 1
> +
> +  ranges:
> +    maxItems: 1
> +
> +  reg:
> +    maxItems: 1
> +
> +  resets:
> +    maxItems: 4
> +
> +  reset-names:
> +    items:
> +      - const: pcie
> +      - const: pcie-core
> +      - const: pcie-ext
> +      - const: pcie-hard
> +
> +required:
> +  - brcm,serdes
> +  - clocks
> +  - clock-names
> +  - "#interrupt-cells"
> +  - interrupt-map-mask
> +  - interrupt-map
> +  - ranges
> +  - reg
> +  - resets
> +  - reset-names
> +
> +unevaluatedProperties: false
> +
> +examples:
> +  - |
> +    serdes_cntl: syscon@10001800 {
> +      compatible = "syscon";
> +      reg = <0x10001800 0x4>;
> +      native-endian;
> +    };

I believe you could be modelling the SerDes as generic PHY driver which
would be a little cleaner than the syscon approach. In newer chips like
6318 it looks like you should be able to use pcie-brcmstb.c since the
controller appears to be nearly the same and the PHY abstraction would
work nicely there.
-- 
Florian