Re: [PATCH 2/5] selftests/resctrl: Remove duplicate feature check from CMT test

[Reinette Chatre]

Hi Ilpo,

On 9/14/2023 2:58 AM, Ilpo Järvinen wrote:
> On Wed, 13 Sep 2023, Reinette Chatre wrote:
>> On 9/13/2023 4:11 AM, Ilpo Järvinen wrote:
>>> On Tue, 12 Sep 2023, Reinette Chatre wrote:
>>>> On 9/11/2023 4:19 AM, Ilpo Järvinen wrote:
>>>>> The test runner run_cmt_test() in resctrl_tests.c checks for CMT
>>>>> feature and does not run cmt_resctrl_val() if CMT is not supported.
>>>>> Then cmt_resctrl_val() also check is CMT is supported.
>>>>>
>>>>> Remove the duplicated feature check for CMT from cmt_resctrl_val().
>>>>>
>>>>> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
>>>>> Cc: <stable@xxxxxxxxxxxxxxx>
>>>>
>>>> This does not look like stable material to me. 
>>>
>>> I know but when constructing this series I had 2 options:
>>>
>>> Either convert also this when changing validate_resctrl_feature_request() 
>>> or remove this call entirely.
>>>
>>> Given it's duplicate of the other CMT check, I chose to just remove it 
>>> (which I'd do anyway). As patch 4/5 requires 3/5 which in turn requires 
>>> this, this has to go stable if 4/5 goes too.
>>>
>>
>> Understood. This makes it a dependency of an actual fix, which is addressed
>> in 4/5's sign-off area. This notation is new to me but it is not clear to me
>> that the dependency should also be tagged as stable material (without a 
>> fixes tag). Since it is not an actual fix by itself yet is sent to @stable
>> I think it may cause confusion. Is just listing it as a dependency of the
>> actual fix not sufficient (as you already do in 4/5)? Perhaps as compromise
>> this patch can also get a note to the stable team. Something like:
>>
>> 	Cc: <stable@xxxxxxxxxxxxxxx> # dependency of "selftests/resctrl: Fix feature checks"
>>
>> I am not sure though - I would like to avoid confusion and not burden
>> the stable team. If this is a flow you have used before successfully I'd
>> defer to your experience.
> 
> I came across that dependency format when Greg KH replied to somebody how 
> to deal with the cases where there isn't yet a commit id 
> (the cases mentioned in Documentation/process/stable-kernel-rules.rst 
> assumes there is already a commit id). Unfortunately it's long time ago 
> so I cannot easily find the link.

I see, thank you. I was not aware of this custom.

Reinette