Re: [PATCH v1] usb: chipidea: tegra: Fix missed ci_hdrc_remove_device()

[Dmitry Osipenko]

26.02.2019 18:21, Greg Kroah-Hartman ÐÐÑÐÑ:
> On Tue, Feb 26, 2019 at 06:08:17PM +0300, Dmitry Osipenko wrote:
>> 26.02.2019 17:58, Greg Kroah-Hartman ÐÐÑÐÑ:
>>> On Tue, Feb 26, 2019 at 05:33:05PM +0300, Dmitry Osipenko wrote:
>>>> 26.02.2019 13:56, Greg Kroah-Hartman ÐÐÑÐÑ:
>>>>> On Mon, Feb 25, 2019 at 08:07:15AM +0300, Dmitry Osipenko wrote:
>>>>>> Ð Mon, 25 Feb 2019 02:27:19 +0000
>>>>>> Peter Chen <peter.chen@xxxxxxx> ÐÐÑÐÑ:
>>>>>>
>>>>>>>  
>>>>>>>>
>>>>>>>> Fixes: dfebb5f43a78827a ("usb: chipidea: Add support for
>>>>>>>> Tegra20/30/114/124")  
>>>>>>>
>>>>>>> I suppose you need to apply at stable tree too, right?
>>>>>>>
>>>>>>
>>>>>> It is enough to have the "Fixes" tag to get patch backported into all
>>>>>> relevant kernel versions.
>>>>>
>>>>> No it is not.  My scripts do NOT trigger off of the fixes: tag, please
>>>>> read:
>>>>>     https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
>>>>> for how to do this properly.
>>>>
>>>> Okay, my bad then. Maybe this is something that could warned by checkpatch.. adding Joe and Andy to the thread.
>>>
>>> Why?  It's allowed to put fixes: tags for a patch that does not belong
>>> in a stable tree.  That happens all the time, and is encouraged.  Look
>>> at some of the stuff in linux-next now, we have Fixes: for commits that
>>> are still in linux-next as well, because we do not rebase our trees.
>>> When they all merge into Linus's tree, all is good.
>>>
>>> So this is not something that checkpatch needs to do anything about.
>>
>> At least that might help in cases like this if maintainer is also oblivious.
> 
> If the maintainer is "oblivious", they are not going to be running
> checkpatch :)
> 
> Remember, the "Fixes:" tag is a relatively new thing compared to the cc:
> stable tag, which has been a documented requirement for over a decade.
> Yes, some subsystems do not even do cc: stable, but that is because
> those subsystem maintainers do not want to do it, or do not care.
> Again, checkpatch is not going to help them.
> 
> checkpatch is not a panacea, people still have to use their brains.

My point that either submitter or maintainers may catch up the problem if at least there are means to report the problem. So I'm still thinking it's not a bad idea to add the warning message, maybe even for the "checkpatch --strict" option only. I'm not insisting that it is something super-necessary and it's up to you guys to decide in the end.