[PATCH v3 5/8] dt-bindings: memory: lpddr3: deprecate manufacturer ID

[Alim Akhtar]

>-----Original Message-----
>From: Krzysztof Kozlowski [mailto:krzysztof.kozlowski at canonical.com]
>Sent: Monday, February 7, 2022 2:27 PM
>To: Alim Akhtar <alim.akhtar at samsung.com>; 'Rob Herring'
><robh+dt at kernel.org>; 'Lukasz Luba' <lukasz.luba at arm.com>; 'Dmitry
>Osipenko' <digetx at gmail.com>; linux-kernel at vger.kernel.org;
>devicetree at vger.kernel.org; linux-pm at vger.kernel.org; linux-samsung-
>soc at vger.kernel.org; linux-arm-kernel at lists.infradead.org
>Subject: Re: [PATCH v3 5/8] dt-bindings: memory: lpddr3: deprecate
>manufacturer ID
>
>On 07/02/2022 05:14, Alim Akhtar wrote:
>> Hi Krzysztof
>>
>>> -----Original Message-----
>>> From: Krzysztof Kozlowski [mailto:krzysztof.kozlowski at canonical.com]
>>> Sent: Sunday, February 6, 2022 7:28 PM
>>> To: Krzysztof Kozlowski <krzysztof.kozlowski at canonical.com>; Rob
>>> Herring <robh+dt at kernel.org>; Lukasz Luba <lukasz.luba at arm.com>; Alim
>>> Akhtar <alim.akhtar at samsung.com>; Dmitry Osipenko
><digetx at gmail.com>;
>>> linux- kernel at vger.kernel.org; devicetree at vger.kernel.org; linux-
>>> pm at vger.kernel.org; linux-samsung-soc at vger.kernel.org; linux-arm-
>>> kernel at lists.infradead.org
>>> Subject: [PATCH v3 5/8] dt-bindings: memory: lpddr3: deprecate
>>> manufacturer ID
>>>
>>> The memory manufacturer should be described in vendor part of
>>> compatible, so there is no need to duplicate it in a separate property.
>>> Similarly is done in LPDDR2 bindings.
>>>
>>> Signed-off-by: Krzysztof Kozlowski
>>> <krzysztof.kozlowski at canonical.com>
>>> ---
>>> .../bindings/memory-controllers/ddr/jedec,lpddr3.yaml         | 4 +++-
>>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/memory-
>>> controllers/ddr/jedec,lpddr3.yaml
>>> b/Documentation/devicetree/bindings/memory-
>>> controllers/ddr/jedec,lpddr3.yaml
>>> index d6787b5190ee..3bcba15098ea 100644
>>> --- a/Documentation/devicetree/bindings/memory-
>>> controllers/ddr/jedec,lpddr3.yaml
>>> +++ b/Documentation/devicetree/bindings/memory-
>>> controllers/ddr/jedec,lpd
>>> +++ dr3.yaml
>>> @@ -40,7 +40,9 @@ properties:
>>>   manufacturer-id:
>>>     $ref: /schemas/types.yaml#/definitions/uint32
>>>     description: |
>>> -      Manufacturer ID value read from Mode Register 5.
>>> +      Manufacturer ID value read from Mode Register 5.  The property is
>>> +      deprecated, manufacturer should be derived from the compatible.
>>> +    deprecated: true
>>>
>>
>> Shouldn't it be the other way? As DT describes hardware and MR5 does
>> contain the Manufacturer ID, so getting Manufacturer ID from MR5 makes
>> aligned to hardware description.
>
>The code/driver can read MR5 and report it to user-space in case for example
>we have a device compatible with different vendor and same compatible is
>used. So compatible is re-used but we want real manufacturer ID from the
>hardware.
>
>But storing a fixed MR5 value in DT does not fit here. If someone takes effort
>to encode manufacturer ID in the DTS, then he/she should take effort to
>actually document the compatible.
>
>Basically having MR5 in DT is equal to having a compatible in DTS. I prefer the
>latter - simpler, less properties, using existing property from DT spec instead
>of custom one.
>
Ok, so there are two part of it, first one to get the compatible and bind the device 
and second one to get the manufacturer ID from MR5 for say user application.
Second one can still be handled at driver side, so

Reviewed-by: Alim Akhtar <alim.akhtar at samsung.com>


>Best regards,
>Krzysztof