Re: 回复: 回复: [PATCH v2 2/3] i2c: designware: Export symbols and add __weak for Starfive I2C driver

[Krzysztof Kozlowski <krzk@kernel.org>]

On 28/05/2026 10:28, Lianfeng Ouyang wrote:
> 
> 
>> -----邮件原件-----
>> 发件人: Krzysztof Kozlowski <krzk@kernel.org>
>> 发送时间: 2026年5月28日 15:46
>> 收件人: Lianfeng Ouyang <lianfeng.ouyang@starfivetech.com>; Andi Shyti
>> <andi.shyti@kernel.org>; Rob Herring <robh@kernel.org>; Krzysztof Kozlowski
>> <krzk+dt@kernel.org>; Conor Dooley <conor+dt@kernel.org>; Mika
>> Westerberg <mika.westerberg@linux.intel.com>; Andy Shevchenko
>> <andriy.shevchenko@linux.intel.com>
>> 抄送: linux-i2c@vger.kernel.org; devicetree@vger.kernel.org;
>> linux-kernel@vger.kernel.org
>> 主题: Re: 回复: [PATCH v2 2/3] i2c: designware: Export symbols and add
>> __weak for Starfive I2C driver
>>
>> On 28/05/2026 04:36, Lianfeng Ouyang wrote:
>>>>>  	{ .compatible = "mobileye,eyeq6lplus-i2c" },
>>>>>  	{ .compatible = "mscc,ocelot-i2c" },
>>>>>  	{ .compatible = "snps,designware-i2c" },
>>>>> +	{ .compatible = "starfive,jhb100-i2c-master" },
>>>>> +	{ .compatible = "starfive,jhb100-i2c-slave" },
>>>>
>>>> So both devices are the same... Or you miss proper entry in ID table.
>>>> Don't sprinkle compatibles around the code - ID table match data is for
>>>> variant customization.
>>>>
>>>> Best regards,
>>>> Krzysztof
>>>
>>> Ok, Master and slave have different base addresses and interrupt numbers, but
>>> the register layout is the same, so in the next version, I will only use the same
>>> compatible ”starfive, jhb100-i2c“, and then add starfive, i2c-slave to the dts
>> node
>>> of the slave to distinguish them
>>
>> You do not need to distinguish same devices usually, but I still do not
>> understand why do you need all these different compatibles and your
>> commit msgs or binding text do not help, so I don't know what to advise.
>>
>>
>> Best regards,
>> Krzysztof
> 
> The IP corresponding to the i2c driver submitted this time is quite unique, 
> so it should be distinguished. The master-slave of i2c designware are at the
> same base address and switch according to the scene, so only one is initialized
> in probe(). But i2c starfive has master-slave at different base addresses,
> and it can work simultaneously. Therefore, both roles need to be initialized in 
> their respective probes. If we don't distinguish, we don't know if this probe is 
> initializing the master or the slave
> 

I don't think you read the comments. Again, you claim devices are
incompatible but have exactly the same programming interface.

Best regards,
Krzysztof