Re: [PATCH v2 8/8] arm64: defconfig: enable interconnect and pinctrl for SM4450

[Tengfei Fan <quic_tengfan@quicinc.com>]

在 9/17/2023 3:35 PM, Krzysztof Kozlowski 写道:
> On 17/09/2023 00:55, Georgi Djakov wrote:
>> Hi Krzysztof,
>>
>> On 16.09.23 23:32, Krzysztof Kozlowski wrote:
>>> On 16/09/2023 17:13, Georgi Djakov wrote:
>>>>
>>>>>
>>>>> Complete list here, and it is inconsistent. Latest 8550 is also =y. Do we document
>>>>> the reasons somewhere on why they are added as =y?
>>>>>
>>>>> CONFIG_INTERCONNECT_QCOM=y
>>>>> CONFIG_INTERCONNECT_QCOM_MSM8916=m
>>>>> CONFIG_INTERCONNECT_QCOM_MSM8996=m
>>>>> CONFIG_INTERCONNECT_QCOM_OSM_L3=m
>>>>> CONFIG_INTERCONNECT_QCOM_QCM2290=m
>>>>> CONFIG_INTERCONNECT_QCOM_QCS404=m
>>>>> CONFIG_INTERCONNECT_QCOM_SA8775P=y
>>>>> CONFIG_INTERCONNECT_QCOM_SC7180=y
>>>>> CONFIG_INTERCONNECT_QCOM_SC7280=y
>>>>> CONFIG_INTERCONNECT_QCOM_SC8180X=y
>>>>> CONFIG_INTERCONNECT_QCOM_SC8280XP=y
>>>>> CONFIG_INTERCONNECT_QCOM_SDM845=y
>>>>> CONFIG_INTERCONNECT_QCOM_SM8150=m
>>>>> CONFIG_INTERCONNECT_QCOM_SM8250=m
>>>>> CONFIG_INTERCONNECT_QCOM_SM8350=m
>>>>> CONFIG_INTERCONNECT_QCOM_SM8450=y
>>>>> CONFIG_INTERCONNECT_QCOM_SM8550=y
>>>>
>>>> If the device can boot (to console/initramfs) with =m, we go with that.
>>>> But if something critical like the UART depends on the interconnect
>>>> provider, then we make it built-in.
>>>>
>>>> On SM8550 for example, we have enabled bandwidth scaling support for QUP
>>>> and that's why it needs to be =y.
>>>>
>>>> It looks like on SM4450 this should be =y too.
>>>
>>> I asked why SM4450 has to be =y and there was no answer. The argument
>>> that SM8450 is a module, is not applicable.
>>
>>   From the hardware description i see in patch 7, the serial engine depends
>> on some interconnect provider. But as the serial console driver is only
>> available as built-in, the interconnect provider also needs be built-in
>> for the UART device to probe and register the console.
>>
>> So the answer to your question should be that this is needed by the UART
>> device (at least).
>>
>> Such details of course deserve to be mentioned in the commit message of
>> this patch.
> 
> If you mean here the debug UART with console, then it is the same valid
> reason as in my change for others. This should be mentioned in commit msg.
> 
> Best regards,
> Krzysztof
> 
yes, will mention this valid reason in next patch commit message detailly.

-- 
Thx and BRs,
Tengfei Fan