Re: [PATCH v2 00/11] i2c: npcm: Bug fixes timeout, spurious interrupts

[Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>]

On 21/02/2022 09:16, Tyrone Ting wrote:
> Hi Krzysztof:
> 
> Thank you for your comments and please find my reply next to your comments.
> 
> Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com> 於 2022年2月20日
> 週日 下午5:30寫道：
>>
>> On 20/02/2022 04:53, Tyrone Ting wrote:
>>> From: Tyrone Ting <kfting@nuvoton.com>
>>>
>>> This patchset includes the following fixes:
>>>
>>> - Add dt-bindings description for NPCM845.
>>> - Bug fix for timeout calculation.
>>> - Better handling of spurious interrupts.
>>> - Fix for event type in slave mode.
>>> - Removal of own slave addresses [2:10].
>>> - Support for next gen BMC (NPCM845).
>>>
>>> The NPCM I2C driver is tested on NPCM750 and NPCM845 evaluation boards.
>>>
>>> Addressed comments from:
>>>  - Jonathan Neuschäfer : https://lkml.org/lkml/2022/2/7/670
>>>  - Krzysztof Kozlowski : https://lkml.org/lkml/2022/2/7/760
>>
>> How did you address the ABI change comment? I still see you break the
>> ABI with the introduction of a new, required property.
>>
> 
> I add the new, required property "nuvoton,sys-mgr" in the file
> nuvoton-common-npcm7xx.dtsi.
> The file nuvoton-common-npcm7xx.dtsi is required by the existing
> upstream NPCM devicetree files.
> It is also updated and committed in this patch set [PATCH v2 01/11]
> arm: dts: add new property for NPCM i2c module.
> Please let me know if I misunderstand the meaning of "breaking the ABI".
> Thank you again.

Breaking the ABI means that old DTS stop working with new kernel. Your
change breaks old (and out-of-tree) DTS.

What is more, your change is not bisectable because DTS goes via
separate branch or tree than driver change.

You need to keep old code as fallback, if getting nuvoton,sys-mgr fails.

Best regards,
Krzysztof