Re: [PATCH bpf-next v2 07/13] selftests/bpf: Test __arena pointer members

["Kumar Kartikeya Dwivedi" <memxor@gmail.com>]

On Tue Oct 13, 2026 at 11:14 AM CEST, bot+bpf-ci wrote:
>> selftests/bpf: Test __arena pointer members
>
> TL;DR: Test issue in test_struct_ops_arena.c: if the bpf_testmod BTF lacks the
> "arena" type tag, the new arena member tests are silently dropped,
> test__skip() is never called, and the test still reports OK.
>
>> diff --git a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
>> --- a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
>> +++ b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
>
> [ ... ]
>
>> +/*
>> + * The tags on the members of struct bpf_testmod_arena_member_args reach the
>> + * module BTF only from a toolchain with btf_type_tag support.
>> + */
>> +static bool testmod_has_type_tag(const char *tag)
>> +{
>> +	struct btf *vmlinux_btf, *module_btf;
>> +	bool found = false;
>> +
>> +	vmlinux_btf = btf__load_vmlinux_btf();
>> +	if (!ASSERT_OK_PTR(vmlinux_btf, "vmlinux_btf"))
>> +		return false;
>> +	module_btf = btf__load_module_btf("bpf_testmod", vmlinux_btf);
>> +	if (ASSERT_OK_PTR(module_btf, "module_btf"))
>> +		found = btf__find_by_name_kind(module_btf, tag, BTF_KIND_TYPE_TAG) > 0;
>> +	if (!found)
>> +		printf("%s:SKIP: no %s type tag in the bpf_testmod BTF\n", __func__, tag);
>> +	btf__free(module_btf);
>> +	btf__free(vmlinux_btf);
>> +	return found;
>> +}
>
> [ ... ]
>
>> @@ -122,6 +201,18 @@ void serial_test_struct_ops_arena(void)
>>  		arena_arg_fail();
>>  	if (test__start_subtest("arena_arg_attach"))
>>  		arena_arg_attach();
>> +	if (testmod_has_type_tag("arena")) {
>> +		if (test__start_subtest("arena_member"))
>> +			arena_member();
>> +		if (test__start_subtest("arena_member_render"))
>> +			arena_member_render();
>> +		RUN_TESTS(struct_ops_arena_member_fail__union);
>> +		RUN_TESTS(struct_ops_arena_member_fail__load_acq);
>> +		RUN_TESTS(struct_ops_arena_member_fail__mismatch);
>> +		RUN_TESTS(struct_ops_arena_member_fail__mismatch_mem);
>> +		RUN_TESTS(struct_ops_arena_member_fail__write);
>> +		RUN_TESTS(struct_ops_arena_member_fail__no_arena);
>> +	}
>>  #else
>>  	test__skip();
>>  #endif
>
> When the bpf_testmod BTF has no "arena" type tag, testmod_has_type_tag()
> prints a SKIP line but never calls test__skip().  Is the skip reported
> anywhere?
>
> serial_test_struct_ops_arena() then leaves out the arena_member and
> arena_member_render subtests and all six RUN_TESTS() groups, and the parent
> test still reports OK.  test_progs only prints a test's captured stdout when
> the test fails or runs verbose (dump_test_log() uses print_test = verbose()
> || force_log || test_failed), so in a normal run the message is never shown
> and the skip is never counted.
>
> That seems realistic for kernels built with a toolchain that lacks
> btf_type_tag support, where this coverage would be silently dropped.
> prog_tests/btf_tag.c does the same check
> (btf__find_by_name_kind(*module_btf, "user", BTF_KIND_TYPE_TAG) <= 0) and
> pairs the printf with test__skip().
>
> Could the arena_member and arena_member_render subtests still be started
> unconditionally, calling test__skip() from inside each one when the tag is
> missing?  That would keep the arena_arg* results intact while making the
> skip visible.
>

It's fine.

>
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
>
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/38370395582