[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <9ae25267-c4de-8f14-df62-d9d0d1f1420c@linaro.org>
Date: Tue, 16 May 2023 10:19:33 +0200
From: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To: Marek Vasut <marex@...x.de>, mike.looijmans@...ic.nl,
Richard Leitner <richard.leitner@...ux.dev>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Conor Dooley <conor+dt@...nel.org>, linux-usb@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dt-bindings: usb: usb251xb: correct swap-dx-lanes type to
uint32
On 15/05/2023 15:47, Marek Vasut wrote:
>> Please consider the environment before printing this e-mail
>> On 15-05-2023 14:55, Marek Vasut wrote:
>>> On 5/15/23 12:33, Krzysztof Kozlowski wrote: diff --git
>>> a/Documentation/devicetree/bindings/usb/usb251xb.yaml
>>> b/Documentation/devicetree/bindings/usb/usb251xb.yaml
>>>> index 4d1530816817..ac5b99710332 100644
>>>> --- a/Documentation/devicetree/bindings/usb/usb251xb.yaml
>>>> +++ b/Documentation/devicetree/bindings/usb/usb251xb.yaml
>>>> @@ -231,7 +231,7 @@ properties:
>>>> power-on sequence to a port until the port has adequate power.
>>>> swap-dx-lanes:
>>>> - $ref: /schemas/types.yaml#/definitions/uint8-array
>>>> + $ref: /schemas/types.yaml#/definitions/uint32-array
>>>> description: |
>>>> Specifies the ports which will swap the differential-pair
>>>> (D+/D-),
>>>> default is not-swapped.
>>>
>>> Would it make more sense to update the driver instead ? I doubt you
>>> could have more than 256 ports on this device after all.
>>
>>
>> I guess there's a bunch of devicetrees already out there using the
>> (misdocumented) 32-bit array binding, they'd break in a bad way...
>
> I think it is the other way around -- if the binding was documented as
> u8, then the existing DTs should use the u8 type if they are compliant
> to the binding document.
>
> I see one board in next which uses this property and sets it to 0 , so
> this one is not affected either way:
> arch/arm64/boot/dts/freescale/imx8mq-zii-ultra-rmb3.dts:
> swap-dx-lanes = <0>;
First of all, the original binding did not define it as u8. It actually
skipped the type entirely but:
- Example shown u32,
- Driver used u32,
- In-tree user uses u32 (although as pointed - as 0 so not really
relevant).
Thus the ABI is rather defined by not-breaking users here, so I would
stick to fixing it to u32.
Best regards,
Krzysztof
Powered by blists - more mailing lists