Re: [PATCH bpf-next] selftests/bpf: test_progs option for listing test names

[Jesper Dangaard Brouer <brouer@redhat.com>]

On Wed, 1 Jul 2020 08:36:08 -0700
Alexei Starovoitov <alexei.starovoitov@gmail.com> wrote:

> On Tue, Jun 30, 2020 at 2:19 PM Andrii Nakryiko
> <andrii.nakryiko@gmail.com> wrote:
> >
> > On Tue, Jun 30, 2020 at 1:32 PM Jesper Dangaard Brouer
> > <brouer@redhat.com> wrote:  
> > >
> > > On Tue, 30 Jun 2020 08:46:01 -0700
> > > Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote:
> > >  
> > > > > @@ -688,9 +700,17 @@ int main(int argc, char **argv)
> > > > >                         cleanup_cgroup_environment();
> > > > >         }
> > > > >         stdio_restore();
> > > > > +
> > > > > +       if (env.list_test_names) {
> > > > > +               if (env.succ_cnt == 0)
> > > > > +                       env.fail_cnt = 1;
> > > > > +               goto out;
> > > > > +       }
> > > > > +  
> > > >
> > > > Why failure if no test matched? Is that to catch bugs in whitelisting?  
> > >
> > > I would not call it catch bugs, but sort of.  The purpose is to know if
> > > requested test is valid.  This can be used to e.g. run through all the
> > > tests numbers, and stopping when a test number (-n) is no-longer valid,
> > > by using this shell exit value as a test, like:
> > >
> > >  n=1;
> > >  while [ $(./test_progs --list -n $n) ] ; do \
> > >    echo "./test_progs -n $n" ; n=$(( n+1 )); \
> > >  done
> > >
> > > Notice that this features that be used for looking up a test number,
> > > and returning a testname, which was the original request from CI.  I
> > > choose this implementation as it more generic and generally useful.
> > >
> > >  $ ./test_progs --list -n 89
> > >  xdp_adjust_tail
> > >  
> >
> > Yeah, it has a nice querying effect. Makes sense.
> >
> > Acked-by: Andrii Nakryiko <andriin@fb.com>  
> 
> hmm. it doesn't apply.
> Applying: selftests/bpf: Test_progs option for listing test names
> error: sha1 information is lacking or useless
> (tools/testing/selftests/bpf/test_progs.c).
> error: could not build fake ancestor
> Patch failed at 0001 selftests/bpf: Test_progs option for listing test names

It doesn't apply because it depend on my previous changes, that Daniel
said he applied:

 https://lore.kernel.org/bpf/6e7543fa-f496-a6d2-a6d5-70dff9f84090@iogearbox.net/

But I can see that it is not in the net-next git tree.
 
> Could you please respin.

I will respin together with the other unapplied patch.  Which is
actually fine, as I have an improvement for the previous patch, that I
can squash.

-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer