Re: [PATCH 00/19] Add initial support for QCS8300

[Konrad Dybcio <konradybcio@kernel.org>]

On 5.09.2024 7:08 AM, Jingyi Wang wrote:
> Hi Krzysztof,
> 
> On 9/4/2024 6:19 PM, Krzysztof Kozlowski wrote:
>> On 04/09/2024 11:34, Krzysztof Kozlowski wrote:
>>> On 04/09/2024 10:33, Jingyi Wang wrote:
>>>> Add initial support for QCS8300 SoC and QCS8300 RIDE board.
>>>>
>>>> This revision brings support for:
>>>> - CPUs with cpu idle
>>>> - interrupt-controller with PDC wakeup support
>>>> - gcc
>>>> - TLMM
>>>> - interconnect
>>>> - qup with uart
>>>> - smmu
>>>> - pmic
>>>> - ufs
>>>> - ipcc
>>>> - sram
>>>> - remoteprocs including ADSP,CDSP and GPDSP
>>>>
>>>> Signed-off-by: Jingyi Wang <quic_jingyw@quicinc.com>
>>>> ---
>>>> patch series organized as:
>>>> - 1-2: remoteproc binding and driver
>>>> - 3-5: ufs binding and driver
>>>> - 6-7: rpmhpd binding and driver
>>>> - 8-15: bindings for other components found on the SoC
>>>
>>> Limit your CC list. I found like 8 unnecessary addresses for already
>>> huge Cc list. Or organize your patches per subsystem, as we usually expect.
>>>
>>>> - 16-19: changes to support the device tree
>>>>
>>>> dependencies:
>>>> tlmm: https://lore.kernel.org/linux-arm-msm/20240819064933.1778204-1-quic_jingyw@quicinc.com/
>>>> gcc: https://lore.kernel.org/all/20240820-qcs8300-gcc-v1-0-d81720517a82@quicinc.com/
>>>> interconnect: https://lore.kernel.org/linux-arm-msm/20240827151622.305-1-quic_rlaggysh@quicinc.com/
>>>
>>> Why? UFS cannot depend on pinctrl for example.
>>>
>>> This blocks testing and merging.
>>>
>>> Please organize properly (so decouple) your patches, so that there is no
>>> fake dependency.
>>
>> Let me also add here one more thought. That's like fourth or fifth
>> QCS/SA patchset last two weeks from Qualcomm and they repeat the same
>> mistakes. Not correctly organized, huge cc list, same problems with
>> bindings or drivers.
>>
>> I am giving much more comments to fix than review/ack tags.
>>
>> I am not going to review this. I will also slow down with reviewing
>> other Qualcomm patches. Why? Because you post simultaneously, apparently
>> you do not learn from other review, so I have to keep repeating the same.
>>
>> I am overwhelmed with this, so please expect two week review time from me.
>>
>> Best regards,
>> Krzysztof
>>
> The CC list is generated from B4 tool, however, thanks for your advice and we
> will decouple the changes to avoid this. And could you please help us to confirm
> the better way to handle binding changes which just add one compatible, should
> it be submitted as a single patch or submmitted together with dts patch series?

The tool did its job here, it's just that this series is very long and a ton
of people ended up being involved due to bindings oneliners

Konrad