Re: [PATCH 1/2] dt-bindings: mmc: arasan,sdci: Add Xilinx Versal Net compatible

[Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>]

On 28/03/2023 09:31, Michal Simek wrote:
> 
> 
> On 3/28/23 09:14, Krzysztof Kozlowski wrote:
>> On 27/03/2023 11:58, Potthuri, Sai Krishna wrote:
>>> Hi Krzysztof,
>>>
>>>> -----Original Message-----
>>>> From: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>>> Sent: Friday, March 24, 2023 5:14 PM
>>>> To: Potthuri, Sai Krishna <sai.krishna.potthuri@amd.com>; Ulf Hansson
>>>> <ulf.hansson@linaro.org>; Rob Herring <robh+dt@kernel.org>; Krzysztof
>>>> Kozlowski <krzysztof.kozlowski+dt@linaro.org>; Michal Simek
>>>> <michal.simek@xilinx.com>; Adrian Hunter <adrian.hunter@intel.com>
>>>> Cc: linux-mmc@vger.kernel.org; linux-kernel@vger.kernel.org;
>>>> devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org; git (AMD-
>>>> Xilinx) <git@amd.com>; saikrishna12468@gmail.com
>>>> Subject: Re: [PATCH 1/2] dt-bindings: mmc: arasan,sdci: Add Xilinx Versal Net
>>>> compatible
>>>>
>>>> On 24/03/2023 08:36, Sai Krishna Potthuri wrote:
>>>>> Add Xilinx Versal Net compatible to support eMMC 5.1 PHY.
>>>>>
>>>>> Signed-off-by: Sai Krishna Potthuri <sai.krishna.potthuri@amd.com>
>>>>> ---
>>>>>   Documentation/devicetree/bindings/mmc/arasan,sdhci.yaml | 6 ++++++
>>>>>   1 file changed, 6 insertions(+)
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/mmc/arasan,sdhci.yaml
>>>>> b/Documentation/devicetree/bindings/mmc/arasan,sdhci.yaml
>>>>> index 8296c34cfa00..cf44a4b988a7 100644
>>>>> --- a/Documentation/devicetree/bindings/mmc/arasan,sdhci.yaml
>>>>> +++ b/Documentation/devicetree/bindings/mmc/arasan,sdhci.yaml
>>>>> @@ -27,6 +27,7 @@ allOf:
>>>>>               enum:
>>>>>                 - xlnx,zynqmp-8.9a
>>>>>                 - xlnx,versal-8.9a
>>>>> +              - xlnx,versal-net-5.1-emmc
>>>>
>>>> v5.1 is eMMC standard or Versal block version? If the first, it's not suitable for
>>>> compatibles.
>>>>
>>>> Also, what's the difference from xlnx,versal-8.9a?
>>> V5.1 is an eMMC standard and this compatible is defined based on sdhci arasan
>>> eMMC5.1 Host Controller(arasan,sdhci-5.1), where as in Versal, it’s a different
>>> controller and it is based on 4.51 Host Controller(arasan,sdhci-8.9a).
>>
>> Mixing IP block versions and eMMC spec versions in one binding is a
>> great way to confuse.
> 
> What do you suggest then?

Stick to IP block versions or code names. The eMMC spec version would
only make sense if you had such possibility:

xlnx,versal-net-emmc-5.0
xlnx,versal-net-emmc-5.1
xlnx,versal-net-emmc-x.y

So exactly one device with different blocks inside. This is very
uncommon, but there such SoC (SunPlus IIRC).

> 
>>
>>> Versal Net Compatible is defined it this way to make it inline with the other
>>> existing SoC compatibles like "intel,keembay-sdhci-5.1-emmc".
>>> Please suggest if the compatible need to be renamed to "xlnx,versal-net-emmc"?
>>
>> Is Versal Net uniquely identifying your SoC or IP block?
> 
> Yes. versal-net is unique identifier for specific silicon with fixed set if IPs.

Then I suggest xlnx,versal-net-emmc.

> Can you please refresh my mind if we can introduce specific compatible strings 
> for this SOC or should we used existing one if functionality is the same with 
> previous SOC family?

It's regular case and recommendation is always (for every SoC) the same:
https://elixir.bootlin.com/linux/v6.1-rc1/source/Documentation/devicetree/bindings/writing-bindings.rst#L42

You should add new SoC specific compatible followed by existing one
(fallback).

> There could be currently unknown issues related to SOC wiring out of specific IP 
> version.

Best regards,
Krzysztof