Re: [PATCH 1/2] dt-bindings: interconnect: document the RPMh Network-On-Chip interconnect in Kaanapali SoC

[Raviteja Laggyshetty <raviteja.laggyshetty@oss.qualcomm.com>]

On 9/30/2025 3:27 PM, Eugen Hristev wrote:
> 
> 
> On 9/30/25 06:06, Raviteja Laggyshetty wrote:
>>
>>
>> On 9/25/2025 6:10 PM, Konrad Dybcio wrote:
>>> On 9/25/25 10:57 AM, Eugen Hristev wrote:
>>>>
>>>>
>>>> On 9/25/25 02:02, Jingyi Wang wrote:
>>>>> From: Raviteja Laggyshetty <raviteja.laggyshetty@oss.qualcomm.com>
>>>>>
>>>>> Document the RPMh Network-On-Chip Interconnect of the Kaanapali platform.
>>>>>
>>>>> Signed-off-by: Raviteja Laggyshetty <raviteja.laggyshetty@oss.qualcomm.com>
>>>>> Signed-off-by: Jingyi Wang <jingyi.wang@oss.qualcomm.com>
>>>>> ---
>>>
>>> [...]
>>>
>>>>> +properties:
>>>>> +  compatible:
>>>>> +    enum:
>>>>> +      - qcom,kaanapali-aggre-noc
>>>>
>>>> Hi,
>>>>
>>>> Does Kaanapali have a single aggre node, or there are several ?
>>>> On previous SoC, I see there are two (aggre1 and aggre2).
>>>> Also in your driver (second patch), I notice aggre1_noc and aggre2_noc .
>>>> It would make sense to accurately describe here the hardware.
>>>
>>> They're physically separate
>>>
>> Yes, they are physically separate but the topology treats them as a single noc
>> with two slave connections to system noc which you have noticed in the topology file.
> 
> Is it any difference from previous sm8750 where there are two nodes ?
> If yes, can you mention the difference and reasoning for merging them
> into a single node.
> If no, can you reason why the decision to merge them into a single node
> this time instead of keeping it consistent ?
>This is due to HW design issue.
qxm_crypto port is on aggre1_noc and it's qos config is in aggre2_noc. Since
aggre1_noc and aggre2_noc are separate devices they can not access intra address
space. To access qxm_crypto qos combined agree1_noc, aggre2_noc into aggre_noc.

Thanks,
Raviteja.

>>
>> Thanks,
>> Raviteja.
>>
>>> Konrad
>