On Thu, 2026-05-21 at 07:42 -0700, Jakub Kicinski wrote:
> On Wed, 20 May 2026 11:34:43 +1000 Matthieu Baerts (NGI0) wrote:
> > The run.sh script explicitly checks that CONFIG_MODULES is disabled.
> >
> > By default, this config option is enabled. Explicitly disable it to be
> > able to run the RDS tests.
> >
> > Note that writing '# CONFIG_(...) is not set' is usually recommended to
> > disable an option in the .config, but it looks like selftests usually
> > set 'CONFIG_(...)=n', which looks clearer.
> >
> > Fixes: 0f5d68004780 ("selftests: rds: add
> > tools/testing/selftests/net/rds/config")
> > Signed-off-by: Matthieu Baerts (NGI0) <[email protected]>
> > ---
> > tools/testing/selftests/net/rds/config | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/tools/testing/selftests/net/rds/config
> > b/tools/testing/selftests/net/rds/config
> > index 97db7ecb892a..3d62d0c750a8 100644
> > --- a/tools/testing/selftests/net/rds/config
> > +++ b/tools/testing/selftests/net/rds/config
> > @@ -1,3 +1,4 @@
> > +CONFIG_MODULES=n
> > CONFIG_NET_NS=y
> > CONFIG_NET_SCH_NETEM=y
> > CONFIG_RDS=y
>
> Hm, okay, if it works it works, but IIUC disabling modules turns all =m
> from the default config into =n (not =y as one would naively hope?) so
> this may come back to bite us. Unless there's a strong reason to not use
> modules it may be good to follow up in net-next and life this
> requirement.
Ok, thanks for the feedback. I will work on a follow up set to rework the
module configs as well as rename the scripts.
I think the initial motivation for CONFIG_MODULES=n was to simplify gcov
collection, but it should work either way. I
will try to get a patch set out later this week. Thank you!
Allison
