On 20/12/2023 14:40, Jinlong Mao wrote:
>
>
> On 12/20/2023 9:21 PM, Krzysztof Kozlowski wrote:
>> On 20/12/2023 14:07, Jinlong Mao wrote:
>>>
>>>
>>> On 12/20/2023 8:46 PM, Krzysztof Kozlowski wrote:
>>>> On 20/12/2023 13:40, Mao Jinlong wrote:
>>>>> Add coresight components on Qualcomm SM8450 Soc. The components include
>>>>> TMC ETF/ETR, ETE, STM, TPDM, CTI.
>>>>>
>>>>> Signed-off-by: Mao Jinlong <quic_jinlmao@xxxxxxxxxxx>
>>>>> ---
>>>>> arch/arm64/boot/dts/qcom/sm8450.dtsi | 742 +++++++++++++++++++++++++++
>>>>> 1 file changed, 742 insertions(+)
>>>>>
>>>>> diff --git a/arch/arm64/boot/dts/qcom/sm8450.dtsi b/arch/arm64/boot/dts/qcom/sm8450.dtsi
>>>>> index 1783fa78bdbc..112b5a069c94 100644
>>>>> --- a/arch/arm64/boot/dts/qcom/sm8450.dtsi
>>>>> +++ b/arch/arm64/boot/dts/qcom/sm8450.dtsi
>>>>> @@ -285,6 +285,192 @@ CLUSTER_SLEEP_1: cluster-sleep-1 {
>>>>> };
>>>>> };
>>>>>
>>>>> + ete0 {
>>>>
>>>> ete-0
>>> Thanks for the review.
>>>
>>> In arm,embedded-trace-extension.yaml, the node name pattern is
>>> "^ete([0-9a-f]+)$".
>>
>> I don't understand why this binding requires ete name. It's not like it
>> is a generic name worth preserving. Also, the recommended suffix for
>> names is with '-'.
>>
> The number in the ete name should be the same as the number of the CPU.
My question was why "ete" is needed.
> So we can know which CPU this ete belongs to from the name.
Wait, what? Why? Node names should not be for that.
>
> I will update the binding in arm,embedded-trace-extension.yaml.
No, wait, please finish this discussion first. You need to give people
chance to respond...
Best regards,
Krzysztof
