[PATCH v5] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for audio blocks

Alexey Klimov posted 1 patch 3 months, 3 weeks ago
Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
[PATCH v5] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for audio blocks
Posted by Alexey Klimov 3 months, 3 weeks ago
If/when pm4125 audio codec will be added to a device tree file, then dtbs
check will emit messages that pmic audio-codec@f000 doesn't match any
of the regexes: '^pinctrl-[0-9]+$'.

Add the compatibles for two possible audio codecs so the devicetree for
such audio blocks of PMIC can be validated properly while also
removing reference to qcom,pm8916-wcd-analog-codec schema file.

Suggested-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>
---

v5: implemented Krzysztof's suggestion, completely rewrote commit messasge

Previous version:
https://lore.kernel.org/linux-arm-msm/20250915-pm4125_audio_codec_v1-v4-2-b247b64eec52@linaro.org/

 Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml b/Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml
index 078a6886f8b1..2a7a92371b55 100644
--- a/Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml
+++ b/Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml
@@ -137,7 +137,11 @@ patternProperties:
 
   "^audio-codec@[0-9a-f]+$":
     type: object
-    $ref: /schemas/sound/qcom,pm8916-wcd-analog-codec.yaml#
+    properties:
+      compatible:
+        enum:
+          - qcom,pm4125-codec
+          - qcom,pm8916-wcd-analog-codec
 
   "^battery@[0-9a-f]+$":
     type: object
-- 
2.47.3
Re: [PATCH v5] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for audio blocks
Posted by Krzysztof Kozlowski 3 months, 3 weeks ago
On 17/10/2025 08:13, Alexey Klimov wrote:
> If/when pm4125 audio codec will be added to a device tree file, then dtbs
> check will emit messages that pmic audio-codec@f000 doesn't match any
> of the regexes: '^pinctrl-[0-9]+$'.


Future errors because of present mistakes are not a reason to do
something. This makes no sense because there is no DTBs with that
compatible, so drop this sentence. We never document compatibles,
because in the future they will be errors (if I get it right?).

> 
> Add the compatibles for two possible audio codecs so the devicetree for
> such audio blocks of PMIC can be validated properly while also
> removing reference to qcom,pm8916-wcd-analog-codec schema file.

And that's now incomplete. You add new device here and because preferred
and sufficient is to list compatibles, you change existing audio codec
child schema reference into just list of compatibles.

> 
> Suggested-by: Krzysztof Kozlowski <krzk@kernel.org>


I don't think I suggested this patch. What's more, it wasn't here at v4.


> Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>


Best regards,
Krzysztof
Re: [PATCH v5] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for audio blocks
Posted by Alexey Klimov 3 months, 3 weeks ago
On Fri Oct 17, 2025 at 7:25 AM BST, Krzysztof Kozlowski wrote:
> On 17/10/2025 08:13, Alexey Klimov wrote:
>> If/when pm4125 audio codec will be added to a device tree file, then dtbs
>> check will emit messages that pmic audio-codec@f000 doesn't match any
>> of the regexes: '^pinctrl-[0-9]+$'.
>
>
> Future errors because of present mistakes are not a reason to do
> something. This makes no sense because there is no DTBs with that
> compatible, so drop this sentence. We never document compatibles,
> because in the future they will be errors (if I get it right?).

Ok. I can hold it off till it will be started to be used then.

>> Add the compatibles for two possible audio codecs so the devicetree for
>> such audio blocks of PMIC can be validated properly while also
>> removing reference to qcom,pm8916-wcd-analog-codec schema file.
>
> And that's now incomplete. You add new device here and because preferred
> and sufficient is to list compatibles, you change existing audio codec
> child schema reference into just list of compatibles.

So the way I understand this is that commit description is incomplete.
I can change it to your liking, okay.
FWIW, "add new device here" is said as
"Add the compatibles for two possible audio codecs" and removal of
child schema reference is also mentioned as "while also                                        
removing reference to qcom,pm8916-wcd-analog-codec schema file".
But I can change it, okay.

>> Suggested-by: Krzysztof Kozlowski <krzk@kernel.org>
>
>
> I don't think I suggested this patch. What's more, it wasn't here at v4.

The original idea was to fix the warning or error emitted by dtbs check
but now the whole body os the change is your suggestion. Now it seems
it was not even needed at that point earlier which is a new finding.
Your prefference -- drop it or keep it.

>> Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>

Best regards,
Alexey
Re: [PATCH v5] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for audio blocks
Posted by Krzysztof Kozlowski 3 months, 3 weeks ago
On 17/10/2025 08:45, Alexey Klimov wrote:
> On Fri Oct 17, 2025 at 7:25 AM BST, Krzysztof Kozlowski wrote:
>> On 17/10/2025 08:13, Alexey Klimov wrote:
>>> If/when pm4125 audio codec will be added to a device tree file, then dtbs
>>> check will emit messages that pmic audio-codec@f000 doesn't match any
>>> of the regexes: '^pinctrl-[0-9]+$'.
>>
>>
>> Future errors because of present mistakes are not a reason to do
>> something. This makes no sense because there is no DTBs with that
>> compatible, so drop this sentence. We never document compatibles,
>> because in the future they will be errors (if I get it right?).
> 
> Ok. I can hold it off till it will be started to be used then.

No, you don't get the point. You invented fake (future) error as
rationale... and want to wait till there is real error to commit this?
This makes no sense.

You need to provide rationale for this patch - look at every other commit.

Best regards,
Krzysztof