| Message ID | 4ebd4e68dec83863c51a9114e6507524c8feafb7.1675698070.git.fmaurer@redhat.com (mailing list archive) |
|---|---|
| State | Rejected |
| Delegated to: | BPF |
| Headers | show |
| Series | [bpf-next] selftests: bpf: Use BTF map in sk_assign | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Clearly marked for bpf-next |
| netdev/apply | fail | Patch does not apply to bpf-next |
On Mon, Feb 6, 2023 at 8:07 AM Felix Maurer <fmaurer@redhat.com> wrote: > > The sk_assign selftest uses tc to load the BPF object file for the test. If > tc is linked against libbpf 1.0+, this test failed, because the BPF file > used the legacy maps section. This approach is considered legacy by libbpf > and tc (see examples/bpf/README in the iproute2 repo). > > Therefore, switch to the approach recommended by iproute2 and use a BTF > defined map. This is also well supported by libbpf. > > Signed-off-by: Felix Maurer <fmaurer@redhat.com> > --- This test was updated (see [0]) to support iproute2 version with and without libbpf support. Please check the latest bpf-next/master. [0] 7ce878ca81bc ("selftests/bpf: Fix sk_assign on s390x") > .../selftests/bpf/progs/test_sk_assign.c | 24 +++++-------------- > 1 file changed, 6 insertions(+), 18 deletions(-) > > diff --git a/tools/testing/selftests/bpf/progs/test_sk_assign.c b/tools/testing/selftests/bpf/progs/test_sk_assign.c > index 98c6493d9b91..b0536bdc002b 100644 > --- a/tools/testing/selftests/bpf/progs/test_sk_assign.c > +++ b/tools/testing/selftests/bpf/progs/test_sk_assign.c > @@ -16,25 +16,13 @@ > #include <bpf/bpf_helpers.h> > #include <bpf/bpf_endian.h> > > -/* Pin map under /sys/fs/bpf/tc/globals/<map name> */ > -#define PIN_GLOBAL_NS 2 > - > -/* Must match struct bpf_elf_map layout from iproute2 */ > struct { > - __u32 type; > - __u32 size_key; > - __u32 size_value; > - __u32 max_elem; > - __u32 flags; > - __u32 id; > - __u32 pinning; > -} server_map SEC("maps") = { > - .type = BPF_MAP_TYPE_SOCKMAP, > - .size_key = sizeof(int), > - .size_value = sizeof(__u64), > - .max_elem = 1, > - .pinning = PIN_GLOBAL_NS, > -}; > + __uint(type, BPF_MAP_TYPE_SOCKMAP); > + __uint(key_size, sizeof(int)); > + __uint(value_size, sizeof(__u64)); > + __uint(max_entries, 1); > + __uint(pinning, LIBBPF_PIN_BY_NAME); > +} server_map SEC(".maps"); > > char _license[] SEC("license") = "GPL"; > > -- > 2.39.1 >
On 06.02.23 23:58, Andrii Nakryiko wrote: > On Mon, Feb 6, 2023 at 8:07 AM Felix Maurer <fmaurer@redhat.com> wrote: >> >> The sk_assign selftest uses tc to load the BPF object file for the test. If >> tc is linked against libbpf 1.0+, this test failed, because the BPF file >> used the legacy maps section. This approach is considered legacy by libbpf >> and tc (see examples/bpf/README in the iproute2 repo). >> >> Therefore, switch to the approach recommended by iproute2 and use a BTF >> defined map. This is also well supported by libbpf. >> >> Signed-off-by: Felix Maurer <fmaurer@redhat.com> >> --- > > This test was updated (see [0]) to support iproute2 version with and > without libbpf support. Please check the latest bpf-next/master. > > [0] 7ce878ca81bc ("selftests/bpf: Fix sk_assign on s390x") Sorry for the noise, I missed that commit. Forget about my patch then. Felix
diff --git a/tools/testing/selftests/bpf/progs/test_sk_assign.c b/tools/testing/selftests/bpf/progs/test_sk_assign.c index 98c6493d9b91..b0536bdc002b 100644 --- a/tools/testing/selftests/bpf/progs/test_sk_assign.c +++ b/tools/testing/selftests/bpf/progs/test_sk_assign.c @@ -16,25 +16,13 @@ #include <bpf/bpf_helpers.h> #include <bpf/bpf_endian.h> -/* Pin map under /sys/fs/bpf/tc/globals/<map name> */ -#define PIN_GLOBAL_NS 2 - -/* Must match struct bpf_elf_map layout from iproute2 */ struct { - __u32 type; - __u32 size_key; - __u32 size_value; - __u32 max_elem; - __u32 flags; - __u32 id; - __u32 pinning; -} server_map SEC("maps") = { - .type = BPF_MAP_TYPE_SOCKMAP, - .size_key = sizeof(int), - .size_value = sizeof(__u64), - .max_elem = 1, - .pinning = PIN_GLOBAL_NS, -}; + __uint(type, BPF_MAP_TYPE_SOCKMAP); + __uint(key_size, sizeof(int)); + __uint(value_size, sizeof(__u64)); + __uint(max_entries, 1); + __uint(pinning, LIBBPF_PIN_BY_NAME); +} server_map SEC(".maps"); char _license[] SEC("license") = "GPL";
The sk_assign selftest uses tc to load the BPF object file for the test. If tc is linked against libbpf 1.0+, this test failed, because the BPF file used the legacy maps section. This approach is considered legacy by libbpf and tc (see examples/bpf/README in the iproute2 repo). Therefore, switch to the approach recommended by iproute2 and use a BTF defined map. This is also well supported by libbpf. Signed-off-by: Felix Maurer <fmaurer@redhat.com> --- .../selftests/bpf/progs/test_sk_assign.c | 24 +++++-------------- 1 file changed, 6 insertions(+), 18 deletions(-)