[PATCH v11 11/12] samples/landlock: Add network demo

[Mickaël Salaün]

On 04/07/2023 14:33, Konstantin Meskhidze (A) wrote:
> 
> 
> 7/3/2023 8:09 PM, Mickaël Salaün пишет:
>>
>> On 03/07/2023 14:50, Konstantin Meskhidze (A) wrote:
>>>
>>>
>>> 6/22/2023 1:18 PM, Mickaël Salaün пишет:
>>>>
>>>> On 22/06/2023 10:00, Konstantin Meskhidze (A) wrote:
>>>>>
>>>>>
>>>>> 6/19/2023 9:19 PM, Mickaël Salaün пишет:
>>>>>>
>>>>>> On 19/06/2023 16:24, Konstantin Meskhidze (A) wrote:
>>>>>>>
>>>>>>>
>>>>>>> 6/13/2023 11:38 PM, Mickaël Salaün пишет:
>>>>>>>>
>>>>>>>> On 13/06/2023 12:54, Konstantin Meskhidze (A) wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> 6/6/2023 6:17 PM, Günther Noack пишет:


[...]

>>>>>>>>>         Thanks for a tip. I think it's a better solution here. Now this
>>>>>>>>> commit is in Mickaёl's -next branch. I could send a one-commit patch later.
>>>>>>>>> Mickaёl, what do you think?
>>>>>>>>
>>>>>>>> I removed this series from -next because there is some issues (see the
>>>>>>>> bot's emails), but anyway, this doesn't mean these patches don't need to
>>>>>>>> be changed, they do. The goal of -next is to test more widely a patch
>>>>>>>> series and get more feedbacks, especially from bots. When this series
>>>>>>>> will be fully ready (and fuzzed with syzkaller), I'll push it to Linus
>>>>>>>> Torvalds.
>>>>>>>>
>>>>>>>> I'll review the remaining tests and sample code this week, but you can
>>>>>>>> still take into account the documentation review.
>>>>>>>
>>>>>>>       Hi, Mickaёl.
>>>>>>>
>>>>>>>       I have a few quetions?
>>>>>>>        - Are you going to fix warnings for bots, meanwhile I run syzcaller?
>>>>>>
>>>>>> No, you need to fix that with the next series (except the Signed-off-by
>>>>>> warnings).
>>>>>
>>>>>      Hi, Mickaёl.
>>>>>       As I understand its possible to check bots warnings just after you
>>>>> push the next V12 series again into your -next branch???
>>>>
>>>> Yes, we get bot warnings on the -next tree, but the command that
>>>> generate it should be reproducible.
>>>
>>>      Stephen Rothwell sent a few warnings he got with powerpc
>>> pseries_le_defconfig. Do I need to fix it in V12 patch? How can I handle
>>> it cause no warnings in current .config?
>>
>> Yes, this need to be fixed in the next series. Could you point to the
>> message?
>>
>     Here you are please:
>        1.
> https://lore.kernel.org/linux-next/20230607141044.1df56246@canb.auug.org.au

This issue is because the WARN_ON_ONCE() is triggered by any 
non-landlocked process, so removing the WARN_ON_ONCE() will fix that.


> 
>        2.
> https://lore.kernel.org/linux-next/20230607135229.1f1e5c91@canb.auug.org.au/

Wrong printf format.


>        3.
> https://lore.kernel.org/linux-next/20230607124940.44af88bb@canb.auug.org.au/

It looks like htmldocs doesn't like #if in enum definition. Anyway, I 
think it should be better to not conditionally define an enum. I've 
pushed this change here: https://git.kernel.org/mic/c/8c96c7eee3ff 
(landlock-net-v11 branch)


> 
>> I'm almost done with the test, I revamped code and I'll send that tomorrow.
>>
>     Ok.Thanks you. Please take your time. I will wait.

[...]