On Tue, 2024-03-05 at 11:29 -0800, Andrii Nakryiko wrote:
> On Mon, Mar 4, 2024 at 2:52 PM Eduard Zingerman <eddyz87@xxxxxxxxx> wrote:
> >
> > When loading struct_ops programs kernel requires BTF id of the
> > struct_ops type and member index for attachment point inside that
> > type. This makes it not possible to have same BPF program used in
> > struct_ops maps that have different struct_ops type.
> > Check if libbpf rejects such BPF objects files.
> >
> > Signed-off-by: Eduard Zingerman <eddyz87@xxxxxxxxx>
> > ---
> > .../selftests/bpf/bpf_testmod/bpf_testmod.c | 24 +++++++++++++
> > .../selftests/bpf/bpf_testmod/bpf_testmod.h | 4 +++
> > .../selftests/bpf/prog_tests/bad_struct_ops.c | 35 +++++++++++++++++++
> > .../selftests/bpf/progs/bad_struct_ops.c | 25 +++++++++++++
> > 4 files changed, 88 insertions(+)
> > create mode 100644 tools/testing/selftests/bpf/prog_tests/bad_struct_ops.c
> > create mode 100644 tools/testing/selftests/bpf/progs/bad_struct_ops.c
> >
>
> CI reports kernel crashes, so please check that, but overall things look good:
>
> Acked-by: Andrii Nakryiko <andrii@xxxxxxxxxx>
Newly added test case struct_ops_multi_pages does not define
bpf_testmod_ops->test_1 field for struct_ops map,
so the following code from my patch caused null pointer dereference:
563 static int bpf_dummy_reg(void *kdata)
564 {
565 struct bpf_testmod_ops *ops = kdata;
566
567 ops->test_1();
Fixed by adding "if (ops)" check.
[...]
