On Wed, 15 Sep 2021 11:27:15 +0000 "Ananyev, Konstantin" <konstantin.anan...@intel.com> wrote:
> > The BPF autotest is defined but not run automatically. > > Since it is short, it should be added to the autotest suite. > > > > Signed-off-by: Stephen Hemminger <step...@networkplumber.org> > > --- > > app/test/meson.build | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/app/test/meson.build b/app/test/meson.build > > index 0d551ac9c2b2..cd18484bb73a 100644 > > --- a/app/test/meson.build > > +++ b/app/test/meson.build > > @@ -194,6 +194,8 @@ test_deps = [ > > fast_tests = [ > > ['acl_autotest', true], > > ['atomic_autotest', false], > > + ['bpf_autotest', true], > > + ['bpf_convert_autotest', true], > > ['bitops_autotest', true], > > ['byteorder_autotest', true], > > ['cksum_autotest', true], > > -- > > Acked-by: Konstantin Ananyev <konstantin.anan...@intel.com> > > > 2.30.2 > One oddity of original BPF test is that it constructs an mbuf where there is no Ethernet header. Didn't want to change that since not only would the constructor have to be changed but also the hand written BPF programs as well. Not necessarily a bug, but the test is not doing what any application using BPF on incoming mbuf's would expect to do.