Re: [tip: locking/core] locking/futex/selftests: Remove duplicate ABI defines

From: Ingo Molnar
Date: Fri Oct 06 2023 - 07:04:30 EST



* Ingo Molnar <mingo@xxxxxxxxxx> wrote:

>
> * Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>
> > On Fri, Oct 06, 2023 at 10:32:20AM -0000, tip-bot2 for Muhammad Usama Anjum wrote:
> > > The following commit has been merged into the locking/core branch of tip:
> > >
> > > Commit-ID: d351a9e56cc90a9ff694550e4b3bcaf51a391525
> > > Gitweb: https://git.kernel.org/tip/d351a9e56cc90a9ff694550e4b3bcaf51a391525
> > > Author: Muhammad Usama Anjum <usama.anjum@xxxxxxxxxxxxx>
> > > AuthorDate: Fri, 06 Oct 2023 14:55:37 +05:00
> > > Committer: Ingo Molnar <mingo@xxxxxxxxxx>
> > > CommitterDate: Fri, 06 Oct 2023 12:29:45 +02:00
> > >
> > > locking/futex/selftests: Remove duplicate ABI defines
> > >
> > > Kselftests are kernel tests that are built with kernel headers
> > > from the same source version. The kernel headers, which includes
> > > current ABI definitions, are already being included correctly
> > > in the futex selftest Makefile with the help of KHDR_INCLUDE,
> > > no need to define them again.
> > >
> > > Remove duplicate ABI definitions, which is effectively dead code.
> > >
> > > No functional changes intended.
> >
> > so.. as it happens I recently built these things as stand-alone, and
> > then you ver much end up using the system headers.
> >
> > Also see 20230922205449.808782861@xxxxxxxxxxxxx where I add more of
> > this.
> >
> > Specifically, if one does:
> >
> > cd tools/testing/selftests/futex/functional; make
> >
> > You don't get kernel headers and stuff does not build.
>
> Hm, I did this after applying the patch, and it does work,
> but maybe I missed that those definitions were picked up
> from system headers...
>
> So how about we make sure current kernel headers are applied
> correctly in a 'standalone' build? There's no reason they
> shouldn't be.

Anyway, I've removed this patch from tip:locking/core until
this is cleared up, as your usecase is obviously a valid one ...

Thanks,

Ingo