Re: [PATCH mptcp-next] Squash to "selftests: bpf: test bpf_skc_to_mptcp_sock"

[Mat Martineau <mathew.j.martineau@linux.intel.com>]

[-- Attachment #1: Type: text/plain, Size: 3127 bytes --]

On Wed, 27 Apr 2022, Geliang Tang wrote:

> Hi Mat,
>
> Mat Martineau <mathew.j.martineau@linux.intel.com> 于2022年4月27日周三 08:20写道：
>>
>> On Tue, 26 Apr 2022, Geliang Tang wrote:
>>
>>> Add CONFIG_MPTCP check.
>>>
>>> When CONFIG_MPTCP is not enabled, we'll get a clearer error message:
>>>
>>> libbpf: extern CONFIG_MPTCP (strong) not resolved
>>> libbpf: failed to load object './mptcp_sock.o'
>>>
>>> The message before is like this:
>>>
>>> libbpf: prog '_sockops': BPF program load failed: Invalid argument
>>> libbpf: prog '_sockops': -- BEGIN PROG LOAD LOG --
>>>
>>> Signed-off-by: Geliang Tang <geliang.tang@suse.com>
>>> ---
>>> tools/testing/selftests/bpf/progs/mptcp_sock.c | 4 ++++
>>> 1 file changed, 4 insertions(+)
>>>
>>
>> Looks like this also requires CONFIG_IKCONFIG and CONFIG_IKCONFIG_PROC,
>> can you also add those to tools/testing/selftests/bpf/config? Maybe that
>> should be a separate commit, since other bpf test progs appear to rely on
>> this undocumented config requirement.
>
> In my test, this bpf base test still works without CONFIG_IKCONFIG and
> CONFIG_IKCONFIG_PROC:
>
>> ls /proc/config.gz
> ls: cannot access '/proc/config.gz': No such file or directory
>> sudo ./test_progs -n 103
> #103 mptcp:OK
> Summary: 1/3 PASSED, 0 SKIPPED, 0 FAILED
>

Huh, I wonder what's going on? It fails for me:

$ sudo ./test_progs -v -a mptcp
bpf_testmod.ko is already unloaded.
Loading bpf_testmod.ko...
Successfully loaded bpf_testmod.ko.
libbpf: failed to open system Kconfig
libbpf: failed to load object './mptcp_sock.o'
run_test:FAIL:165
test_base:FAIL:227
libbpf: failed to open system Kconfig
libbpf: failed to load object './mptcp_sock.o'
run_test:FAIL:165
test_base:FAIL:244
#103/1 mptcp/base:FAIL
(cgroup_helpers.c:146: errno: Device or resource busy) Removing cgroup: /mnt/cgroup-test-work-dir831/mptcp
(cgroup_helpers.c:146: errno: Device or resource busy) Removing cgroup: /mnt/cgroup-test-work-dir831
#103 mptcp:FAIL

All error logs:
#103 mptcp:FAIL

Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED

This is with Fedora 35. I'll check the BPF CI to see if there's any 
information about their configuration that might explain the difference.


- Mat

>
>
>>
>> - Mat
>>
>>> diff --git a/tools/testing/selftests/bpf/progs/mptcp_sock.c b/tools/testing/selftests/bpf/progs/mptcp_sock.c
>>> index 5cfaec4e7245..7b6a25e37de8 100644
>>> --- a/tools/testing/selftests/bpf/progs/mptcp_sock.c
>>> +++ b/tools/testing/selftests/bpf/progs/mptcp_sock.c
>>> @@ -7,6 +7,7 @@
>>>
>>> char _license[] SEC("license") = "GPL";
>>> __u32 _version SEC("version") = 1;
>>> +extern bool CONFIG_MPTCP __kconfig;
>>>
>>> struct mptcp_storage {
>>>       __u32 invoked;
>>> @@ -46,6 +47,9 @@ int _sockops(struct bpf_sock_ops *ctx)
>>>               if (!storage)
>>>                       return 1;
>>>       } else {
>>> +             if (!CONFIG_MPTCP)
>>> +                     return 1;
>>> +
>>>               msk = bpf_skc_to_mptcp_sock(sk);
>>>               if (!msk)
>>>                       return 1;
>>> --
>>> 2.34.1
>>>
>>>
>>>
>>
>> --
>> Mat Martineau
>> Intel
>>
>

--
Mat Martineau
Intel