Re: [PATCH v5 2/2] dt-bindings: mtd: davinci: convert to yaml

[Krzysztof Kozlowski]

On 14/10/2024 14:04, Marcus Folkesson wrote:
> 
> Krzysztof,
> 
> I'm sorry for all these iterations, it wouldn't have been necessary if I had
> done my homework better. I'm not too familiar with writing these and I do
> often find the descriptions unclear and not obvious.
> Anyway, thank you for your patience, reviews and help.
> 
> 
> On Tue, Oct 08, 2024 at 03:28:33PM +0200, Krzysztof Kozlowski wrote:
>> On Tue, Oct 08, 2024 at 09:02:45AM +0200, Marcus Folkesson wrote:
>>> Convert the bindings to yaml format.
>>>
>>> Signed-off-by: Marcus Folkesson <marcus.folkesson@xxxxxxxxx>
>>> ---
>>>  .../devicetree/bindings/mtd/davinci-nand.txt       |  94 -----------------
>>>  .../devicetree/bindings/mtd/ti,davinci-nand.yaml   | 115 +++++++++++++++++++++
>>>  2 files changed, 115 insertions(+), 94 deletions(-)
>>>
>>
>>
>>> diff --git a/Documentation/devicetree/bindings/mtd/ti,davinci-nand.yaml b/Documentation/devicetree/bindings/mtd/ti,davinci-nand.yaml
>>> new file mode 100644
>>> index 0000000000000000000000000000000000000000..1263616593532e8483d556b4242b004a16620ddf
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/mtd/ti,davinci-nand.yaml
>>> @@ -0,0 +1,115 @@
>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>> +%YAML 1.2
>>> +---
>>> +$id: http://devicetree.org/schemas/mtd/ti,davinci-nand.yaml#
>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>> +
>>> +title: TI DaVinci NAND controller
>>> +
>>> +maintainers:
>>> +  - Marcus Folkesson <marcus.folkesson@xxxxxxxxx>
>>> +
>>> +allOf:
>>> +  - $ref: nand-controller.yaml#
>>> +
>>> +properties:
>>> +  compatible:
>>> +    enum:
>>> +      - ti,davinci-nand
>>> +      - ti,keystone-nand
>>> +
>>> +  reg:
>>> +    maxItems: 1
>>
>> This was different in original binding and commit msg does not explain
>> changes.  Be sure any change from pure conversion is explained in the
>> commit msg.
> 
> Hm. Another misinterpretation from my side.
> Should I use items instead? E.g.
> 
>   reg:
>     items:
>       - description: |
>         Contains 2 offset/length values:
>         - offset and length for the access window.
>         - offset and length for accessing the AEMIF
>         control registers.

You need to list two items, so two times "- description:"
(and drop redundant parts like "offset and length" because this cannot
be anything else)

and this should be pointed out by testing, so open the example and try
to explain why it is so different than expected.


Unless, it is supposed to be one item... but how do I know? You are
looking at the device and its driver, not  me.

> 
>>
>>> +
>>> +  partitions:
>>> +    $ref: /schemas/mtd/partitions/partitions.yaml
>>> +
>>> +  ti,davinci-chipselect:
>>> +    description:
>>> +      Number of chipselect. Indicate on the davinci_nand driver which
>>> +      chipselect is used for accessing the nand.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    enum: [0, 1, 2, 3]
>>> +
>>> +  ti,davinci-mask-ale:
>>> +    description:
>>> +      Mask for ALE. Needed for executing address phase. These offset will be
>>> +      added to the base address for the chip select space the NAND Flash
>>> +      device is connected to.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    default: 0x08
>>> +
>>> +  ti,davinci-mask-cle:
>>> +    description:
>>> +      Mask for CLE. Needed for executing command phase. These offset will be
>>> +      added to the base address for the chip select space the NAND Flash device
>>> +      is connected to.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    default: 0x10
>>> +
>>> +  ti,davinci-mask-chipsel:
>>> +    description:
>>> +      Mask for chipselect address. Needed to mask addresses for given
>>> +      chipselect.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    default: 0
>>> +
>>> +  ti,davinci-ecc-bits:
>>> +    description: Used ECC bits.
>>> +    enum: [1, 4]
>>> +
>>> +  ti,davinci-ecc-mode:
>>> +    description: Operation mode of the NAND ECC mode.
>>> +    $ref: /schemas/types.yaml#/definitions/string
>>> +    enum: [none, soft, hw, on-die]
>>> +    deprecated: true
>>> +
>>> +  ti,davinci-nand-buswidth:
>>> +    description: Bus width to the NAND chip
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    enum: [8, 16]
>>> +    default: 8
>>> +    deprecated: true
>>> +
>>> +  ti,davinci-nand-use-bbt:
>>> +    type: boolean
>>> +    description:
>>> +      Use flash based bad block table support. OOB identifier is saved in OOB
>>> +      area.
>>> +    deprecated: true
>>> +
>>> +required:
>>> +  - compatible
>>> +  - reg
>>> +  - ti,davinci-chipselect
>>> +
>>> +unevaluatedProperties: false
>>> +
>>> +examples:
>>> +  - |
>>> +    nand-controller@2000000 {
>>> +      compatible = "ti,davinci-nand";
>>> +      #address-cells = <1>;
>>> +      #size-cells = <0>;
>>
>> I did not notice it last time.... but what is this? How could you have
>> no sizes?
>>
>>> +
>>> +      reg = <0 0x02000000>;
>>
>> This is odd. Address is not 0... and size should be 0.
>>
>> I don't get how it even works. For sure it is not correct.
> 
> Outch. It slipped through when I was laborating.
> 
> This was the example I wanted to get working:
> 
> 
> ```
> examples:
>   - |
>     nand-controller@2000000,0 {
>       compatible = "ti,davinci-nand";
>       #address-cells = <1>;
>       #size-cells = <1>;
>       reg = <0 0x02000000 0x02000000
>       1 0x00000000 0x00008000>;

But that's not what is being used. Open the DTS and look how it is
encoded there. You have unnecessary 64-bit addressing here, plus you
still have technically one item, not two. Maybe existing DTS uses it,
then it's fine, but then take DTS (and correct it if it does not define
proper tuples/items).

> 
>       ti,davinci-chipselect = <1>;
>       ti,davinci-mask-ale = <0>;
>       ti,davinci-mask-cle = <0>;
>       ti,davinci-mask-chipsel = <0>;
> 
>       ti,davinci-nand-buswidth = <16>;
>       ti,davinci-ecc-mode = "hw";
>       ti,davinci-ecc-bits = <4>;
>       ti,davinci-nand-use-bbt;
> 
>       partitions {
>         compatible = "fixed-partitions";
>         #address-cells = <1>;
>         #size-cells = <1>;
> 
>         partition@0 {
>           label = "u-boot env";
>           reg = <0 0x020000>;
>         };
>       };
>     };
> ```
> 
> 
> But I'm getting the following errors:
> 
> ```
> .../ti,davinci-nand.example.dtb: nand-controller@2000000,0: #size-cells: 0 was expected
>         from schema $id: http://devicetree.org/schemas/mtd/ti,davinci-nand.yaml#
> .../ti,davinci-nand.example.dtb: nand-controller@2000000,0: reg: [[0, 33554432], [33554432, 1], [0, 32768]] is too long
>         from schema $id: http://devicetree.org/schemas/mtd/ti,davinci-nand.yaml#
> .../ti,davinci-nand.example.dtb: nand-controller@2000000,0: Unevaluated properties are not allowed ('reg' was unexpected)
>         from schema $id: http://devicetree.org/schemas/mtd/ti,davinci-nand.yaml#
> ```

Well, obviously. That's the point of the schema. The example part is
embedded in node with some address/size cells, which might fit your case
or might not.

> 
> The resuling 'ti,davinci-nand.example.dts' contains the following:
> 
> ```
>     example-0 {
>         #address-cells = <1>;
>         #size-cells = <1>;
> 
>         nand-controller@2000000,0 {
>           compatible = "ti,davinci-nand";
>           #address-cells = <1>;
>           #size-cells = <1>;
>           reg = <0 0x02000000 0x02000000
>           1 0x00000000 0x00008000>;
> ```
>         
> 
> How do I set #address-cells in example-0 to 2?
> I guess that is the problem.



You just set it... There are dozens/hundreds of schemas doing it, what
is exactly a problem here?

Best regards,
Krzysztof