On 24/08/2017 04:31, Alexei Starovoitov wrote: > On Mon, Aug 21, 2017 at 02:09:24AM +0200, Mickaël Salaün wrote: >> This step mechanism may be useful to return an information about the >> error without being able to write to TH_LOG_STREAM. >> >> Set _metadata->no_print to true to print this counter. >> >> Signed-off-by: Mickaël Salaün >> Cc: Andy Lutomirski >> Cc: Arnaldo Carvalho de Melo >> Cc: Kees Cook >> Cc: Shuah Khan >> Cc: Will Drewry >> Link: https://lkml.kernel.org/r/CAGXu5j+D-FP8Kt9unNOqKrQJP4DYTpmgkJxWykZyrYiVPz3Y3Q@mail.gmail.com >> --- >> >> This patch is intended to the kselftest tree: >> https://lkml.kernel.org/r/20170806232337.4191-1-mic@digikod.net >> >> Changes since v6: >> * add the step counter in assert/expect macros and use _metadata to >> enable the counter (suggested by Kees Cook) >> --- >> tools/testing/selftests/kselftest_harness.h | 31 ++++++++++++++++++++++----- >> tools/testing/selftests/seccomp/seccomp_bpf.c | 2 +- >> 2 files changed, 27 insertions(+), 6 deletions(-) > > is there a dependency on this in patches 2+ ? > if not, I would send this patch to selftests right away. > > The Landlock tests [patch 9/10] rely on it for now. I sent it three weeks ago: https://lkml.kernel.org/r/20170806232337.4191-1-mic@digikod.net Anyway, until this patch is merged in the kselftest tree and then available to net-next, I'll have to keep it here.