by Francesco Dolcini

[permalink] [raw]

Subject: Re: [PATCH v2 3/5] arm64: dts: ti: add verdin am62

On Thu, Jun 15, 2023 at 01:08:11PM +0530, Vignesh Raghavendra wrote:
>
>
> On 15/06/23 12:21, Francesco Dolcini wrote:
> > On Thu, Jun 15, 2023 at 11:38:00AM +0530, Vignesh Raghavendra wrote:
> >>
> >>
> >> On 01/06/23 18:43, Francesco Dolcini wrote:
> >>
> >> [...]
> >>
> >>> +/* Verdin I2C_1 */
> >>> +&main_i2c1 {
> >>> + status = "okay";
> >>> +
> >>> + /* Audio Codec */
> >>> + nau8822_1a: audio-codec@1a {
> >>> + compatible = "nuvoton,nau8822";
> >>> + reg = <0x1a>;
> >>> + pinctrl-names = "default";
> >>> + pinctrl-0 = <&pinctrl_i2s1_mclk>; /* Configure AUDIO_EXT_REFCLK1 pin as output */
> >>> + #sound-dai-cells = <0>;
> >>> +
> >>> + /* POSTDIV4_16FF_MAIN_2_HSDIVOUT8_CLK -> AUDIO_EXT_REFCLK1 */
> >>> + assigned-clocks = <&k3_clks 157 10>;
> >>> + assigned-clock-parents = <&k3_clks 157 18>;
> >>> + assigned-clock-rates = <25000000>; /* for 48KHz ? 1.7% */
> >>> + };
> >>> +
> >>
> >> Oops, this node fails dtbs_check
> >>
> >> make ARCH=arm64 CROSS_COMPILE=$V8_CROSS CHECK_DTBS=y ti/k3-am625-verdin-nonwifi-dev.dtb
> >>
> >> /work/linux/arch/arm64/boot/dts/ti/k3-am625-verdin-nonwifi-dev.dtb: audio-codec@1a: '#sound-dai-cells', 'assigned-clock-parents', 'assigned-clock-rates', 'assigned-clocks' do not match any of the regexes: 'pinctrl-[0-9]+'
> >> From schema: /home/a0132425/workspace/k3-next/Documentation/devicetree/bindings/sound/nuvoton,nau8822.yaml
> >> /work/linux/arch/arm64/boot/dts/ti/k3-am625-verdin-nonwifi-dev.dtb: audio-codec@1a: 'anyOf' conditional failed, one must be fixed:
> >> 'clocks' is a required property
> >> '#clock-cells' is a required property
> >> From schema: /work/.local/lib/python3.10/site-packages/dtschema/schemas/clock/clock.yaml
> >
> > This properties are needed here, it will not work without.
>
> Agree on the need. But, I cannot take the patch w/o binding update. We
> will end up with messy dts files that would have deviated from bindings.

Fine, however, one question, is this a new policy? Or a specific TI
branch policy? From what I can tell so far there was nothing mandatory
while the DT binding conversion from txt to yaml is in progress.

> Could you please post an update to bindings yaml? I suggest to drop
> audio codec support and add it back once bindings are updated/fixed.
I will proceed like that, thanks. I would be happy to see the
verdin-am62 added in the next merge window.

Thanks for the review,
Francesco

2023-06-15 10:29:31

by Vignesh Raghavendra

[permalink] [raw]

Subject: Re: [PATCH v2 3/5] arm64: dts: ti: add verdin am62

On 15/06/23 13:18, Francesco Dolcini wrote:
> On Thu, Jun 15, 2023 at 01:08:11PM +0530, Vignesh Raghavendra wrote:
>>
>>
>> On 15/06/23 12:21, Francesco Dolcini wrote:
>>> On Thu, Jun 15, 2023 at 11:38:00AM +0530, Vignesh Raghavendra wrote:
>>>>
>>>>
>>>> On 01/06/23 18:43, Francesco Dolcini wrote:
>>>>
>>>> [...]
>>>>
>>>>> +/* Verdin I2C_1 */
>>>>> +&main_i2c1 {
>>>>> + status = "okay";
>>>>> +
>>>>> + /* Audio Codec */
>>>>> + nau8822_1a: audio-codec@1a {
>>>>> + compatible = "nuvoton,nau8822";
>>>>> + reg = <0x1a>;
>>>>> + pinctrl-names = "default";
>>>>> + pinctrl-0 = <&pinctrl_i2s1_mclk>; /* Configure AUDIO_EXT_REFCLK1 pin as output */
>>>>> + #sound-dai-cells = <0>;
>>>>> +
>>>>> + /* POSTDIV4_16FF_MAIN_2_HSDIVOUT8_CLK -> AUDIO_EXT_REFCLK1 */
>>>>> + assigned-clocks = <&k3_clks 157 10>;
>>>>> + assigned-clock-parents = <&k3_clks 157 18>;
>>>>> + assigned-clock-rates = <25000000>; /* for 48KHz ± 1.7% */
>>>>> + };
>>>>> +
>>>>
>>>> Oops, this node fails dtbs_check
>>>>
>>>> make ARCH=arm64 CROSS_COMPILE=$V8_CROSS CHECK_DTBS=y ti/k3-am625-verdin-nonwifi-dev.dtb
>>>>
>>>> /work/linux/arch/arm64/boot/dts/ti/k3-am625-verdin-nonwifi-dev.dtb: audio-codec@1a: '#sound-dai-cells', 'assigned-clock-parents', 'assigned-clock-rates', 'assigned-clocks' do not match any of the regexes: 'pinctrl-[0-9]+'
>>>> From schema: /home/a0132425/workspace/k3-next/Documentation/devicetree/bindings/sound/nuvoton,nau8822.yaml
>>>> /work/linux/arch/arm64/boot/dts/ti/k3-am625-verdin-nonwifi-dev.dtb: audio-codec@1a: 'anyOf' conditional failed, one must be fixed:
>>>> 'clocks' is a required property
>>>> '#clock-cells' is a required property
>>>> From schema: /work/.local/lib/python3.10/site-packages/dtschema/schemas/clock/clock.yaml
>>>
>>> This properties are needed here, it will not work without.
>>
>> Agree on the need. But, I cannot take the patch w/o binding update. We
>> will end up with messy dts files that would have deviated from bindings.
>
> Fine, however, one question, is this a new policy? Or a specific TI
> branch policy?

No, this isn't TI specific. In general if YAML binding exists, then the
DT nodes need to confirm to that schema.

> From what I can tell so far there was nothing mandatory
> while the DT binding conversion from txt to yaml is in progress.
>

I understand if the bindings are still in .txt format. But, in this
case, bindings are converted to YAML and prohibit other properties
unfortunately.

>> Could you please post an update to bindings yaml? I suggest to drop
>> audio codec support and add it back once bindings are updated/fixed.
> I will proceed like that, thanks. I would be happy to see the
> verdin-am62 added in the next merge window.
>

Great, thanks!

> Thanks for the review,
> Francesco

--
Regards
Vignesh