Re: [PATCH bpf-next v4 6/8] selftests/bpf: Add weak/typeless ksym test for light skeleton

[Andrii Nakryiko <andrii.nakryiko@gmail.com>]

On Wed, Oct 20, 2021 at 12:15 PM Kumar Kartikeya Dwivedi
<memxor@gmail.com> wrote:
>
> Also, avoid using CO-RE features, as lskel doesn't support CO-RE, yet.
> Include both light and libbpf skeleton in same file to test both of them
> together.
>
> In c48e51c8b07a ("bpf: selftests: Add selftests for module kfunc support"),
> I added support for generating both lskel and libbpf skel for a BPF
> object, however the name parameter for bpftool caused collisions when
> included in same file together. This meant that every test needed a
> separate file for a libbpf/light skeleton separation instead of
> subtests.
>
> Change that by appending a "_light" suffix to the name for files listed
> in LSKELS_EXTRA, such that both light and libbpf skeleton can be used in
> the same file for subtests, leading to better code sharing.
>
> While at it, improve the build output by saying GEN-LSKEL instead of
> GEN-SKEL for light skeleton generation recipe.
>
> Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
> ---
>  tools/testing/selftests/bpf/Makefile          |  7 ++--
>  .../selftests/bpf/prog_tests/ksyms_btf.c      | 35 +++++++++++++++-
>  .../selftests/bpf/prog_tests/ksyms_module.c   | 40 +++++++++++++++++--
>  .../bpf/prog_tests/ksyms_module_libbpf.c      | 28 -------------
>  .../selftests/bpf/progs/test_ksyms_weak.c     |  3 +-
>  5 files changed, 74 insertions(+), 39 deletions(-)
>  delete mode 100644 tools/testing/selftests/bpf/prog_tests/ksyms_module_libbpf.c
>
> diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile
> index 498222543c37..1c3c8befc249 100644
> --- a/tools/testing/selftests/bpf/Makefile
> +++ b/tools/testing/selftests/bpf/Makefile
> @@ -325,7 +325,7 @@ LINKED_SKELS := test_static_linked.skel.h linked_funcs.skel.h               \
>  LSKELS := kfunc_call_test.c fentry_test.c fexit_test.c fexit_sleep.c \
>         test_ringbuf.c atomics.c trace_printk.c trace_vprintk.c
>  # Generate both light skeleton and libbpf skeleton for these
> -LSKELS_EXTRA := test_ksyms_module.c
> +LSKELS_EXTRA := test_ksyms_module.c test_ksyms_weak.c
>  SKEL_BLACKLIST += $$(LSKELS)
>
>  test_static_linked.skel.h-deps := test_static_linked1.o test_static_linked2.o
> @@ -399,12 +399,13 @@ $(TRUNNER_BPF_SKELS): %.skel.h: %.o $(BPFTOOL) | $(TRUNNER_OUTPUT)
>         $(Q)$$(BPFTOOL) gen skeleton $$(<:.o=.linked3.o) name $$(notdir $$(<:.o=)) > $$@
>
>  $(TRUNNER_BPF_LSKELS): %.lskel.h: %.o $(BPFTOOL) | $(TRUNNER_OUTPUT)
> -       $$(call msg,GEN-SKEL,$(TRUNNER_BINARY),$$@)
> +       $$(call msg,GEN-LSKEL,$(TRUNNER_BINARY),$$@)

This breaks nice output alignment:

  GEN-SKEL [test_progs-no_alu32] bpf_iter_tcp4.skel.h
  GEN-LSKEL [test_progs-no_alu32] trace_vprintk.lskel.h

Isn't ".lskel.h" suffix enough to distinguish them?

>         $(Q)$$(BPFTOOL) gen object $$(<:.o=.linked1.o) $$<
>         $(Q)$$(BPFTOOL) gen object $$(<:.o=.linked2.o) $$(<:.o=.linked1.o)
>         $(Q)$$(BPFTOOL) gen object $$(<:.o=.linked3.o) $$(<:.o=.linked2.o)
>         $(Q)diff $$(<:.o=.linked2.o) $$(<:.o=.linked3.o)
> -       $(Q)$$(BPFTOOL) gen skeleton -L $$(<:.o=.linked3.o) name $$(notdir $$(<:.o=)) > $$@
> +       $$(eval LSKEL_NAME := $$(notdir $$(<:.o=$$(if $$(filter $$(notdir $$(<:.o=.c)),$(LSKELS_EXTRA)),_light,))))

eval inside eval?.. Wow, do we really need that? If you just want to
add _light (I suggest _lskel though, it will make for a more
meaningful and recognizable names in user-space code) suffix, do it
for all light skeletons unconditionally and keep it simple?

> +       $(Q)$$(BPFTOOL) gen skeleton -L $$(<:.o=.linked3.o) name $$(LSKEL_NAME) > $$@
>
>  $(TRUNNER_BPF_SKELS_LINKED): $(TRUNNER_BPF_OBJS) $(BPFTOOL) | $(TRUNNER_OUTPUT)
>         $$(call msg,LINK-BPF,$(TRUNNER_BINARY),$$(@:.skel.h=.o))

[...]