Re: [PATCH v3 1/3] dt-bindings: hwlock: Adding brcmstb-hwspinlock support

[Florian Fainelli <florian.fainelli@broadcom.com>]

On 10/14/25 14:46, Krzysztof Kozlowski wrote:
> On 14/10/2025 17:40, Kamal Dasu wrote:
>> On Thu, Oct 9, 2025 at 10:00 PM Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>>
>>> On 09/10/2025 23:20, Kamal Dasu wrote:
>>>> The broadcom settop SoCs have hardware semaphores as part of "sundry"
>>>> IP block which has other controls that do not belong anywhere else.
>>>> e.g. pin/mux controls, SoC identification, drive strength, reset controls,
>>>> and other misc bits are part of this block.
>>>
>>> This does not explain why this has to be a separate block. Rob comment
>>> stays valid - this belongs to parent node.
>>>
>>
>> We do not have a separate parent node where this will fit in. We have
> 
> 
> So what is the parent? simple bus?

Ultimately yes, there is a parent node which is a "simple-bus" but what 
we effectively did is break up the "sun-top-ctrl" node into register 
ranges that are on 4 byte boundary, because the functionality offered in 
that aggregate is partitioned on a 4 byte boundary.

Now, we could utilize the fact that the node has been exposed in our 
Device Trees using a syson:

                 sun_top_ctrl: syscon@8404000 {
                         compatible = "brcm,brcmstb-sun-top-ctrl",
                                 "syscon";
                         reg = <0x8404000 0x708>;
                 };


> 
>> the DWORD aligned  register defined here as organized in the SUNDRY
>> block.
> 
> 
> I don't know what SUNDRY is.

https://www.merriam-webster.com/dictionary/sundry
-- 
Florian