[PATCH v2 0/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[PATCH v2 0/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Allen-KH Cheng]

Convert MediaTek eFuse devicetree binding to YAML.

Based on tag: next-20220422, linux-next/master

make DT_CHECKER_FLAGS=-m dt_binding_check
DT_SCHEMA_FILES=Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml

make ARCH=arm64
dtbs_check DT_SCHEMA_FILES=Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml

changes since v1:
- change file name from mtk,efuse-yaml to mtk,efuse.yaml
- add mtk-efuse in commit title
- change compatible entries from const to enum

Allen-KH Cheng (1):
  dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

 .../devicetree/bindings/nvmem/mtk,efuse.yaml  | 70 +++++++++++++++++++
 .../devicetree/bindings/nvmem/mtk-efuse.txt   | 43 ------------
 2 files changed, 70 insertions(+), 43 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
 delete mode 100644 Documentation/devicetree/bindings/nvmem/mtk-efuse.txt

-- 
2.18.0

[PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Allen-KH Cheng]

Convert MediaTek eFuse devicetree binding to YAML.

Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
---
 .../devicetree/bindings/nvmem/mtk,efuse.yaml  | 70 +++++++++++++++++++
 .../devicetree/bindings/nvmem/mtk-efuse.txt   | 43 ------------
 2 files changed, 70 insertions(+), 43 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
 delete mode 100644 Documentation/devicetree/bindings/nvmem/mtk-efuse.txt

diff --git a/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml b/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
new file mode 100644
index 000000000000..d056bc61dd5b
--- /dev/null
+++ b/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
@@ -0,0 +1,70 @@
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/nvmem/mtk,efuse.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: MediaTek eFuse device tree bindings
+
+maintainers:
+  - Lala Lin <lala.lin@mediatek.com>
+  - Allen-KH Cheng <allen-kh.cheng@mediatek.com>
+
+allOf:
+  - $ref: "nvmem.yaml#"
+
+properties:
+  compatible:
+    oneOf:
+      - enum:
+          - mediatek,mt8173-efuse
+          - mediatek,efuse
+      - items:
+          - enum:
+              - mediatek,mt7622-efuse
+              - mediatek,mt7623-efuse
+              - mediatek,mt8183-efuse
+              - mediatek,mt8192-efuse
+              - mediatek,mt8195-efuse
+              - mediatek,mt8516-efuse
+          - const: mediatek,efuse
+
+  reg:
+    maxItems: 1
+
+patternProperties:
+  "^.*@[0-9a-f]+$":
+    type: object
+
+    properties:
+      reg:
+        maxItems: 1
+        description:
+          Offset and size in bytes within the storage device.
+
+    required:
+      - reg
+
+    additionalProperties: false
+
+required:
+  - compatible
+  - reg
+
+unevaluatedProperties: false
+
+examples:
+  - |
+
+    efuse: efuse@10206000 {
+            compatible = "mediatek,mt8173-efuse";
+            reg	   = <0x10206000 0x1000>;
+            #address-cells = <1>;
+            #size-cells = <1>;
+
+            /* Data cells */
+            thermal_calibration: calib@528 {
+                reg = <0x528 0xc>;
+            };
+    };
+...
diff --git a/Documentation/devicetree/bindings/nvmem/mtk-efuse.txt b/Documentation/devicetree/bindings/nvmem/mtk-efuse.txt
deleted file mode 100644
index 39d529599444..000000000000
--- a/Documentation/devicetree/bindings/nvmem/mtk-efuse.txt
+++ /dev/null
@@ -1,43 +0,0 @@
-= Mediatek MTK-EFUSE device tree bindings =
-
-This binding is intended to represent MTK-EFUSE which is found in most Mediatek SOCs.
-
-Required properties:
-- compatible: should be
-	      "mediatek,mt7622-efuse", "mediatek,efuse": for MT7622
-	      "mediatek,mt7623-efuse", "mediatek,efuse": for MT7623
-	      "mediatek,mt8173-efuse" or "mediatek,efuse": for MT8173
-	      "mediatek,mt8192-efuse", "mediatek,efuse": for MT8192
-	      "mediatek,mt8195-efuse", "mediatek,efuse": for MT8195
-	      "mediatek,mt8516-efuse", "mediatek,efuse": for MT8516
-- reg: Should contain registers location and length
-- bits: contain the bits range by offset and size
-
-= Data cells =
-Are child nodes of MTK-EFUSE, bindings of which as described in
-bindings/nvmem/nvmem.txt
-
-Example:
-
-	efuse: efuse@10206000 {
-		compatible = "mediatek,mt8173-efuse";
-		reg	   = <0 0x10206000 0 0x1000>;
-		#address-cells = <1>;
-		#size-cells = <1>;
-
-		/* Data cells */
-		thermal_calibration: calib@528 {
-			reg = <0x528 0xc>;
-		};
-	};
-
-= Data consumers =
-Are device nodes which consume nvmem data cells.
-
-For example:
-
-	thermal {
-		...
-		nvmem-cells = <&thermal_calibration>;
-		nvmem-cell-names = "calibration";
-	};
-- 
2.18.0

Re: [PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Krzysztof Kozlowski]

On 25/04/2022 10:48, Allen-KH Cheng wrote:
> Convert MediaTek eFuse devicetree binding to YAML.
> 
> Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
> ---
>  .../devicetree/bindings/nvmem/mtk,efuse.yaml  | 70 +++++++++++++++++++

The vendor prefix is mediatek. Quoting my previous reply:

  Same comments as usual, so "vendor,device-name", e.g. "mediatek,efuse"
  if this is going to match all possible future MediaTek chips or
  "mediatek,mt7622-efuse"

so keep it mediatek,efuse.yaml.

>  .../devicetree/bindings/nvmem/mtk-efuse.txt   | 43 ------------
>  2 files changed, 70 insertions(+), 43 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
>  delete mode 100644 Documentation/devicetree/bindings/nvmem/mtk-efuse.txt
> 
> diff --git a/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml b/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
> new file mode 100644
> index 000000000000..d056bc61dd5b
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
> @@ -0,0 +1,70 @@
> +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/nvmem/mtk,efuse.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: MediaTek eFuse device tree bindings

No changes here. Please read my comments from your v1, don't ignore them.

> +
> +maintainers:
> +  - Lala Lin <lala.lin@mediatek.com>
> +  - Allen-KH Cheng <allen-kh.cheng@mediatek.com>
> +
> +allOf:
> +  - $ref: "nvmem.yaml#"
> +
> +properties:
> +  compatible:
> +    oneOf:
> +      - enum:
> +          - mediatek,mt8173-efuse
> +          - mediatek,efuse

Still no changes...

> +      - items:
> +          - enum:
> +              - mediatek,mt7622-efuse
> +              - mediatek,mt7623-efuse
> +              - mediatek,mt8183-efuse
> +              - mediatek,mt8192-efuse
> +              - mediatek,mt8195-efuse
> +              - mediatek,mt8516-efuse
> +          - const: mediatek,efuse
> +
> +  reg:
> +    maxItems: 1
> +
> +patternProperties:
> +  "^.*@[0-9a-f]+$":
> +    type: object
> +
> +    properties:
> +      reg:
> +        maxItems: 1
> +        description:
> +          Offset and size in bytes within the storage device.
> +
> +    required:
> +      - reg
> +
> +    additionalProperties: false

Still no changes.

> +
> +required:
> +  - compatible
> +  - reg
> +
> +unevaluatedProperties: false
> +
> +examples:
> +  - |
> +

Still no changes.


Best regards,
Krzysztof

Re: [PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Krzysztof Kozlowski]

On 26/04/2022 08:23, allen-kh.cheng wrote:
>>> +properties:
>>> +  compatible:
>>> +    oneOf:
>>> +      - enum:
>>> +          - mediatek,mt8173-efuse
>>> +          - mediatek,efuse
>>
>> Still no changes...
>>
> 
> I just want to confirm again.
> 
> "Generic compatibles should not be used standalone"
> 
> It seems we should remove mediatek,efuse and keep "mediatek,mt8173-
> efuse"in binding. have I got that right?

You should comment for which chipsets this compatible is and add a
deprecated:true. In such case it cannot be part of enum but separate
item in this oneOf.


Best regards,
Krzysztof

Re: [PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Krzysztof Kozlowski]

On 26/04/2022 12:02, allen-kh.cheng wrote:
> Hi Krzysztof,
> 
> On Tue, 2022-04-26 at 08:31 +0200, Krzysztof Kozlowski wrote:
>> On 26/04/2022 08:23, allen-kh.cheng wrote:
>>>>> +properties:
>>>>> +  compatible:
>>>>> +    oneOf:
>>>>> +      - enum:
>>>>> +          - mediatek,mt8173-efuse
>>>>> +          - mediatek,efuse
>>>>
>>>> Still no changes...
>>>>
>>>
>>> I just want to confirm again.
>>>
>>> "Generic compatibles should not be used standalone"
>>>
>>> It seems we should remove mediatek,efuse and keep "mediatek,mt8173-
>>> efuse"in binding. have I got that right?
>>
>> You should comment for which chipsets this compatible is and add a
>> deprecated:true. In such case it cannot be part of enum but separate
>> item in this oneOf.
>>
>>
>> Best regards,
>> Krzysztof
> 
> Thanks for your suggestions, I would plan to send PATCHs as below,
> 
> We have a PATCH 01 for current accepted dts
> 
> properties:
>   compatible:
>     oneOf:
>       - enum:
>         - mediatek,efuse
> 
>         - mediatek,mt8173-efuse
>         description: Only mt8173-efuse
> with generic fallback should be used
>       - items:
>           - enum:
>     
> - mediatek,mt7622-efuse
> 			  ...
>           - const: mediatek,efuse
> 
> Then add PATCH 02 to deprecate it
> 
> properties:
>   compatible:
>     oneOf:
>       - enum:
>         - mediatek,efuse
>         - mediatek,mt8173-efuse
>         deprecated: true
>         description: The mediatek,efuse is a generic fallback for other
> Chipset. Do not use the single compatible such as mediatek,efuse
> or mediatek,mt8173-efuse. It is deprecated.
>       - items:
>           - enum:
>               - mediatek,mt7622-efuse
>               ...
>           - const: mediatek,efuse
> 
> 
> PATCH 03 for 8173
> 
> update mt8173.dtsi 
> 
> change compatible from "mediatek,mt8173-efuse" to "mediatek,mt8173-
> efuse", "mediatek,efuse";
> 
> 
> Do you think it'd be okay ?

The idea is correct, but as I said it cannot be part of enum, but
separate item in oneOf. You should see an error when testing your patch.


Best regards,
Krzysztof

Re: [PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Krzysztof Kozlowski]

On 27/04/2022 11:28, allen-kh.cheng wrote:
> Hi Krzysztof,
> 
> On Tue, 2022-04-26 at 12:14 +0200, Krzysztof Kozlowski wrote:
>> On 26/04/2022 12:02, allen-kh.cheng wrote:
>>> Hi Krzysztof,
>>>
>>> On Tue, 2022-04-26 at 08:31 +0200, Krzysztof Kozlowski wrote:
>>>> On 26/04/2022 08:23, allen-kh.cheng wrote:
>>>>>>> +properties:
>>>>>>> +  compatible:
>>>>>>> +    oneOf:
>>>>>>> +      - enum:
>>>>>>> +          - mediatek,mt8173-efuse
>>>>>>> +          - mediatek,efuse
>>>>>>
>>>>>> Still no changes...
>>>>>>
>>>>>
>>>>> I just want to confirm again.
>>>>>
>>>>> "Generic compatibles should not be used standalone"
>>>>>
>>>>> It seems we should remove mediatek,efuse and keep
>>>>> "mediatek,mt8173-
>>>>> efuse"in binding. have I got that right?
>>>>
>>>> You should comment for which chipsets this compatible is and add
>>>> a
>>>> deprecated:true. In such case it cannot be part of enum but
>>>> separate
>>>> item in this oneOf.
>>>>
>>>>
>>>> Best regards,
>>>> Krzysztof
>>>
>>> Thanks for your suggestions, I would plan to send PATCHs as below,
>>>
>>> We have a PATCH 01 for current accepted dts
>>>
>>> properties:
>>>   compatible:
>>>     oneOf:
>>>       - enum:
>>>         - mediatek,efuse
>>>
>>>         - mediatek,mt8173-efuse
>>>         description: Only mt8173-efuse
>>> with generic fallback should be used
>>>       - items:
>>>           - enum:
>>>     
>>> - mediatek,mt7622-efuse
>>> 			  ...
>>>           - const: mediatek,efuse
>>>
>>> Then add PATCH 02 to deprecate it
>>>
>>> properties:
>>>   compatible:
>>>     oneOf:
>>>       - enum:
>>>         - mediatek,efuse
>>>         - mediatek,mt8173-efuse
>>>         deprecated: true
>>>         description: The mediatek,efuse is a generic fallback for
>>> other
>>> Chipset. Do not use the single compatible such as mediatek,efuse
>>> or mediatek,mt8173-efuse. It is deprecated.
>>>       - items:
>>>           - enum:
>>>               - mediatek,mt7622-efuse
>>>               ...
>>>           - const: mediatek,efuse
>>>
>>>
>>> PATCH 03 for 8173
>>>
>>> update mt8173.dtsi 
>>>
>>> change compatible from "mediatek,mt8173-efuse" to "mediatek,mt8173-
>>> efuse", "mediatek,efuse";
>>>
>>>
>>> Do you think it'd be okay ?
>>
>> The idea is correct, but as I said it cannot be part of enum, but
>> separate item in oneOf. You should see an error when testing your
>> patch.
>>
>>
>> Best regards,
>> Krzysztof
> 
> I have tested 
> make DT_CHECKER_FLAGS=-m dt_binding_check
> DT_SCHEMA_FILES=Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
> 
> make ARCH=arm64 dtbs_check
> DT_SCHEMA_FILES=Documentation/devicetree/bindings/nvmem/mtk,efuse.yaml
> 
> Is the following correct as final version ?

Almost :)

> 
> properties:
>   compatible:
>     oneOf:
>       - const: mediatek,mt8173-efuse
> #Don't use this in new dts files

This compatible above is correct for mt8173, isn't it?

>         deprecated: true
>       - const:
> mediatek,efuse
>         deprecated: true
>         description:
>          
> Please use mediatek,efuse as fallback.

Description does not match. This should be something like:
"MediaTek efuse for MT8173. Deprecated, use mediatek,mt8173-efuse instead"



Best regards,
Krzysztof

Re: [PATCH v2 1/1] dt-bindings: nvmem: mediatek: Convert mtk-efuse binding to YAML

[Krzysztof Kozlowski]

On 27/04/2022 12:00, allen-kh.cheng wrote:
> I think there are two cases in mediatek efuse dirver now.
> 
> Case 1, 
> const: mediatek,efuse is deprecated.
> const: mediatek,mt8173-efuse is remained. All mediatek chipsets will
> use mediatek,mt8173-efuse as fallback.
> 
> Case 2,
> const: mediatek,efuse is deprecated.
> const: mediatek,mt8173-efuse is deprecated.
> 
> All mediatek chipsets(include ediatek,mt8173-efuse) will use
> mediatek,efuse as fallback.
> 
> Which one do you think is better?

Indeed, I forgot that mt8173 would also fallback to generic efuse.
Indeed let's go with case 2, so your proposal before was correct.


Best regards,
Krzysztof