On 30/05/2023 12:13, Francesco Dolcini wrote:
> On Sat, May 27, 2023 at 11:49:17AM +0530, Vignesh Raghavendra wrote:
>> On 19/05/23 9:39 pm, Francesco Dolcini wrote:
>>> Hello,
>>> while writing a new DT file I stumbled across this warning
>>>
>>> .../arch/arm64/boot/dts/ti/k3-am625-verdin-wifi-dev.dtb: serial@2810000: Unevaluated properties are not allowed ('rs485-rts-active-high' was unexpected)
>>> From schema: .../Documentation/devicetree/bindings/serial/8250_omap.yaml
>>>
>>> The property is currently used in the OMAP serial driver
>>>
>>> drivers/tty/serial/omap-serial.c
>>> 1511: if (of_property_read_bool(np, "rs485-rts-active-high")) {
>>>
>>
>> Would be it possible to update driver to imply rs485-rts-active-high"
>> this by lack of rs485-rts-active-low property in DT instead?
>
> What about backward compatibility? This is what is done in all drivers
> apart omap-serial if I'm not wrong.
Uh, I removed too much in that commit. We need to keep backwards
compatibility because driver used default (missing property) as active low.
>
>>> I would be inclined to send the following patch, do you agree?
>>>
>>> diff --git a/Documentation/devicetree/bindings/serial/8250_omap.yaml b/Documentation/devicetree/bindings/serial/8250_omap.yaml
>>> index eb3488d8f9ee..e634e98aa994 100644
>>> --- a/Documentation/devicetree/bindings/serial/8250_omap.yaml
>>> +++ b/Documentation/devicetree/bindings/serial/8250_omap.yaml
>>> @@ -70,6 +70,7 @@ properties:
>>> dsr-gpios: true
>>> rng-gpios: true
>>> dcd-gpios: true
>>> + rs485-rts-active-low: true
>>
>> I believe you mean rs485-rts-active-high here
> whoops, yes of course.
Yes, with a Fixes tag.
>
>>
>>> rts-gpio: true
>>> power-domains: true
>>> clock-frequency: true
>>>
>>> [1] https://lore.kernel.org/all/ZBItlBhzo+YETcJO@xxxxxxxxxxxxxxxxxxxxxxxxxxxx/
>>>
>>
>> Also, I hope you are using 8250_ompa.c and not omap-serial.c for newer
>> designs. omap-serial.c is mostly there to support legacy SoCs and not to
>> be used with K3 SoCs.
>
> Thanks for this head-up. This confused myself, while the issue I
> reported here is real, it does affect only omap-serial.c. In my case I
> am using 8250_omap.c and I can just omit the property from the DTS as
> generally expected!
OK. Are you going send a fix for original issue?
Best regards,
Krzysztof
