Re: [PATCH net v4 3/4] selftests: net: Test UDP length overflow with PMTU discover and big MTU

["Alice Mikityanska" <alice.kernel@fastmail.im>]

On Thu, Aug 27, 2026, at 22:28, Jakub Kicinski wrote:
> On Wed, 26 Aug 2026 13:18:18 -0400 Willem de Bruijn wrote:
>> > +def check_dmesg_clean(func: str) -> bool:
>> > +    with subprocess.Popen(['dmesg'], stdout=subprocess.PIPE) as dmesg:
>> > +        res = subprocess.run(['grep', '-q', f'WARNING:.*{func}'], stdin=dmesg.stdout, check=False)
>> > +    return res.returncode != 0 and dmesg.returncode == 0
>> > +
>> > +
>> > +@contextmanager  
>> 
>> Interesting approach. In ksft the built-in defer() mechanism may be
>> preferred. Not sure.
>
> +1 small preference towards defer() if Alice is willing.
> It's be more idiomatic for our selftests

No problem, I can switch to defer, thanks for the suggestion.

> More importantly ruff check and pylink --disable=R have to be clean
> for new tests. The test cases need a docstring explaining what they are
> trying to test. The helpers can probably get a _ prefix to mark them
> as private and silence the warning.

Got it, will do. I thought it was OK, because all tests I saw didn't
have docstrings.

> -- 
> pw-bot: cr