Re: [PATCH v2 1/2] dt-bindings: soc: ti: bist: Add BIST for K3 devices

[Neha Malcom Francis <n-francis@ti.com>]

On 29/03/25 10:08, Krzysztof Kozlowski wrote:
> On 28/03/2025 13:42, Neha Malcom Francis wrote:
>> On 28/03/25 17:18, Krzysztof Kozlowski wrote:
>>> On 28/03/2025 12:14, Neha Malcom Francis wrote:
>>>> +properties:
>>>> +  compatible:
>>>> +    const: ti,j784s4-bist
>>>> +
>>>> +  reg:
>>>> +    maxItems: 2
>>>> +
>>>> +  reg-names:
>>>> +    items:
>>>> +      - const: cfg
>>>> +      - const: ctrl_mmr
>>>> +
>>>> +  clocks:
>>>> +    maxItems: 1
>>>> +
>>>> +  power-domains:
>>>> +    maxItems: 1
>>>> +
>>>> +  ti,bist-under-test:
>>>> +    $ref: /schemas/types.yaml#/definitions/uint32-array
>>>> +    description:
>>>> +      the device IDs of the devices under test control of the BIST device, the
>>>
>>> Still not phandle... What is a "device ID"?
>>
>> I took a shot at working with the phandle, however the test devices may
>> or may not be present in the devicetree at bootloader stage which is the
> 
> If the nodes are not in DT, then you should not reference them here.
> Bootloader is supposed to receive all the nodes it is expected to work on.

Understood.

> 
>> only place this BIST driver can execute (I know I shouldn't be bringing
>> up the driver here but it's crucial to how I can model this property).
>> HW mandates you run it as early as possible before any other software
>> executes on the test device.
>>
>> So now thinking of other possible ways to define the test devices, we
>> have unique HW identifiers [1] for each of the device which is what I've
>> used here...
>>
>> [1]
>> https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/arm/keystone/ti%2Ck3-sci-common.yaml#L31
> 
> Then do not redefine properties, but use one common definition.

Right, I'll respin the driver and the dt-binding in this direction, thanks!

> 
> 
> Best regards,
> Krzysztof

-- 
Thanking You
Neha Malcom Francis