Re: [PATCH] dt-bindings: serial: 8250: add marvell,armada-38x-uart compatible

[Krzysztof Kozlowski]

On Mon, Aug 31, 2026 at 05:42:03PM -0700, Rosen Penev wrote:
> The Armada 38x and 98DX3236 SoCs use this compatible, either paired
> with "ns16550a" (armada-38x.dtsi) or on its own
> (armada-xp-98dx3236.dtsi). Neither form is covered by the 8250
> binding, so dtbs_check reports:
> 
>   serial@12000: failed to match any schema with compatible:
>   ['marvell,armada-38x-uart', 'ns16550a']
> 
> Add a standalone const entry and a two-string variant paired with
> "ns16550a".
> 
> Validated with make dt_binding_check (8250.yaml) and dt-validate on
> armada-38x and armada-xp 98DX3236 boards: the UART schema errors are
> gone.

Drop last paragraph. If AI tool added it, then NEVER DO THAT. We don't
need to read LLM hallucinations.

Commit msgs do not have obvious build process outputs as confirmation
that you did absolutely basic build process step.

The same you do not say:
1. Code compiles
2. Code passes checkpatch
3. Code passes sparse, smatch etc.

> 
> Assisted-by: opencode:big-pickle
> Signed-off-by: Rosen Penev <rosenp@xxxxxxxxx>
> ---
>  Documentation/devicetree/bindings/serial/8250.yaml | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/serial/8250.yaml b/Documentation/devicetree/bindings/serial/8250.yaml
> index 8c2c177faf89..f71c6afe02f2 100644
> --- a/Documentation/devicetree/bindings/serial/8250.yaml
> +++ b/Documentation/devicetree/bindings/serial/8250.yaml
> @@ -109,6 +109,7 @@ properties:
>        - const: aspeed,ast2500-vuart
>        - const: intel,xscale-uart
>        - const: mrvl,pxa-uart
> +      - const: marvell,armada-38x-uart

ma < mr, keep list ordered.

>        - const: nuvoton,wpcm450-uart
>        - const: nuvoton,npcm750-uart
>        - const: nvidia,tegra20-uart
> @@ -119,6 +120,9 @@ properties:
>                - exar,xr16l2551
>                - exar,xr16l2550
>            - const: ns8250
> +      - items:
> +          - const: marvell,armada-38x-uart

And what if the DTS is just wrong? So warning about wrong DTS is now
suppressed and made documented ABI change?

Best regards,
Krzysztof