[PATCH v1 8/8] ARM: dts: aspeed: System1: PS, sensor and more

[Krzysztof Kozlowski]

On 13/12/2023 20:49, Ninad Palsule wrote:
> Hello Krzysztof,
> 
> On 12/13/23 13:37, Krzysztof Kozlowski wrote:
>> On 13/12/2023 20:02, Ninad Palsule wrote:
>>> Hello Krzysztof,
>>>
>>> On 12/12/23 14:26, Krzysztof Kozlowski wrote:
>>>> On 12/12/2023 17:40, Ninad Palsule wrote:
>>>>> This drop adds following devices in the device tree.
>>>>> - EEPROM/VPD
>>>>> - Power supplies
>>>>> - Humidity, pressure and temperature sensors.
>>>>> - Trusted platform module(TPM) chip
>>>>>
>>>>> Tested:
>>>>>       This board is tested using the simics simulator.
>>>>>
>>>>> Signed-off-by: Ninad Palsule <ninad at linux.ibm.com>
>>>>> ---
>>>> Don't mix DTS with drivers. DTS and drivers go via different subsystems
>>>> and cannot have dependencies, so why DTS is patch #6, then driver #7 and
>>>> now again DTS #7?
>>> There is a dependency on driver code as patch #8 uses the compatibility
>>> string added in driver patch #7.  I have now moved driver patch at the
>>> start. Is that ok? OR you are suggesting something else?
>> First, there is no dependency. Second, except confusing order anyway DTS
>> will go via separate trees. Third, again, there is no dependency. If
>> there is, your patchset is broken and this needs to be fixed. Although I
>> don't understand how new hardware can depend on driver... it's really odd.
> 
> Thanks for the quick response.
> 
> This board uses the nuvoton TPM device. The tpm devices uses 
> "nuvoton,npct75x" driver hence we added it in the device tree. If the 
> driver doesn't have this compatibility string then it won't load. So if 
> someone tries to use this board then tpm won't work unless the 

... and if there is no board it also fails to load.

> compatibility string is added in the driver. That is the dependency I am 
> talking about.

This is not a dependency! It's unrelated.

Best regards,
Krzysztof