Re: [PATCH AUTOSEL 5.10 11/11] firmware: include drivers/firmware/Kconfig unconditionally

From: Arnd Bergmann
Date: Wed Oct 13 2021 - 04:50:41 EST


On Wed, Oct 13, 2021 at 10:45 AM Uwe Kleine-König
<u.kleine-koenig@xxxxxxxxxxxxxx> wrote:
> On Wed, Oct 13, 2021 at 10:38:01AM +0200, Arnd Bergmann wrote:
> > On Wed, Oct 13, 2021 at 7:58 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > On Tue, Oct 12, 2021 at 08:55:31PM -0400, Sasha Levin wrote:
> > >
> > > This isn't for stable kernels, it should be dropped from all of your
> > > AUTOSEL queues.
> >
> > Agreed. The second patch that depends on this does fix a (randconfig)
> > build issue in stable kernels as well, but that patch is currently broken,
>
> Fixing randconfig issues isn't important for stable, is it? The target
> audience for 5.10.74 are people running a kernel between 5.10 and
> 5.10.73, and those don't suffer from this type of build problem, right?

In general, I think randconfig build testing is useful for validating stable
kernels, to help avoid regressions, but there are a number of known
randconfig problems that you hit much more frequently than this one,
which only breaks once every few hundred kernel builds.

Arnd