Re: [PATCH bpf-next] selftests/bpf: Fix error undeclared identifier 'NF_NAT_MANIP_SRC'

[Andrii Nakryiko]

On Fri, Nov 11, 2022 at 10:55 AM Stanislav Fomichev <sdf@xxxxxxxxxx> wrote:
>
> On Fri, Nov 11, 2022 at 10:12 AM Lorenzo Bianconi <lorenzo@xxxxxxxxxx> wrote:
> >
> > > On Fri, Nov 11, 2022 at 9:54 AM Lorenzo Bianconi <lorenzo@xxxxxxxxxx> wrote:
> > > >
> > > > > On 11/11, Rong Tao wrote:
> > > > > > From: Rong Tao <rongtao@xxxxxxxx>
> > > > >
> > > > > > commit 472caa69183f("netfilter: nat: un-export nf_nat_used_tuple")
> > > > > > introduce NF_NAT_MANIP_SRC/DST enum in include/net/netfilter/nf_nat.h,
> > > > > > and commit b06b45e82b59("selftests/bpf: add tests for bpf_ct_set_nat_info
> > > > > > kfunc") use NF_NAT_MANIP_SRC/DST in test_bpf_nf.c. We copy enum
> > > > > > nf_nat_manip_type to test_bpf_nf.c fix this error.
> > > > >
> > > > > > How to reproduce the error:
> > > > >
> > > > > >      $ make -C tools/testing/selftests/bpf/
> > > > > >      ...
> > > > > >        CLNG-BPF [test_maps] test_bpf_nf.bpf.o
> > > > > >        error: use of undeclared identifier 'NF_NAT_MANIP_SRC'
> > > > > >              bpf_ct_set_nat_info(ct, &saddr, sport, NF_NAT_MANIP_SRC);
> > > > > >                                                             ^
> > > > > >        error: use of undeclared identifier 'NF_NAT_MANIP_DST'
> > > > > >              bpf_ct_set_nat_info(ct, &daddr, dport, NF_NAT_MANIP_DST);
> > > > > >                                                             ^
> > > > > >      2 errors generated.
> > > > >
> > > > > $ grep NF_NAT_MANIP_SRC
> > > > > ./tools/testing/selftests/bpf/tools/include/vmlinux.h
> > > > >         NF_NAT_MANIP_SRC = 0,
> > > > >
> > > > > Doesn't look like your kernel config compiles netfilter nat modules?
> > > >
> > > > yes, in bpf kself-test config (tools/testing/selftests/bpf/config) nf_nat
> > > > is compiled as built-in. This issue occurs just if it is compiled as module.
> > >
> > > Right, but if we unconditionally define this enum, I think you'll
> > > break the case where it's compiled as a built-in?
> > > Since at least in my vmlinux.h I have all the defines and this test
> > > includes vmlinux.h...
> >
> > yes, it is correct.
>
> And it will break the CI:
>

It does break CI ([0]). We could use BPF CO-RE and ___suffix rule to
avoid this. But we can also say that selftests/bpf/config{,<arch>} has
to be used by bots that want to build BPF selftests.

  [0] https://github.com/kernel-patches/bpf/actions/runs/3446651033/jobs/5752592868

> $ grep NETFILTER tools/testing/selftests/bpf/config
> CONFIG_NETFILTER=y
>
> So yeah, not sure what to do here. The selftests expect "sane" configs
> (see that bpf/config above) which is not what the bot seems to be
> doing...
>
> > > > Regards,
> > > > Lorenzo
> > > >
> > > > >
> > > > > > Link: https://lore.kernel.org/lkml/202210280447.STsT1gvq-lkp@xxxxxxxxx/
> > > > > > Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > > > > > Signed-off-by: Rong Tao <rongtao@xxxxxxxx>
> > > > > > ---
> > > > > >   tools/testing/selftests/bpf/progs/test_bpf_nf.c | 5 +++++
> > > > > >   1 file changed, 5 insertions(+)
> > > > >
> > > > > > diff --git a/tools/testing/selftests/bpf/progs/test_bpf_nf.c
> > > > > > b/tools/testing/selftests/bpf/progs/test_bpf_nf.c
> > > > > > index 227e85e85dda..307ca166ff34 100644
> > > > > > --- a/tools/testing/selftests/bpf/progs/test_bpf_nf.c
> > > > > > +++ b/tools/testing/selftests/bpf/progs/test_bpf_nf.c
> > > > > > @@ -3,6 +3,11 @@
> > > > > >   #include <bpf/bpf_helpers.h>
> > > > > >   #include <bpf/bpf_endian.h>
> > > > >
> > > > > > +enum nf_nat_manip_type {
> > > > > > +   NF_NAT_MANIP_SRC,
> > > > > > +   NF_NAT_MANIP_DST
> > > > > > +};
> > > > > > +
> > > > > >   #define EAFNOSUPPORT 97
> > > > > >   #define EPROTO 71
> > > > > >   #define ENONET 64
> > > > > > --
> > > > > > 2.31.1
> > > > >