Re: [PATCH 5.15 0/6] Backporting for 5.15 test_verifier failed

[Pu Lehui <pulehui@huawei.com>]

On 2023/8/9 22:59, Luiz Capitulino wrote:
> 
> 
> On 2023-08-04 12:43, Luiz Capitulino wrote:
>>
>>
>> On 2023-08-04 11:30, Pu Lehui wrote:
>>
>>>
>>>
>>>
>>> Hi Luiz,
>>>
>>> My local 5.15 environment is a little bit weird, could you help me to
>>> test it?
>>
>> I'll give this a try, but unfortunately I'm not sure I'll be able to
>> get back to you before mid next week.
> 
> I was finally able to test this, it does fix the original failure but
> I'm also getting:
> 
> """
> #150/p calls: trigger reg2btf_ids[reg->type] for reg->type > 
> __BPF_REG_TYPE_MAX FAIL
> FAIL
> Summary: 1236 PASSED, 582 SKIPPED, 1 FAILED
> """
> 
> Also, some bpf tests don't build for me which causes the bpf tests
> not to be installed. I'm attaching the build errors, although this
> could be because my user-space is old or some misconfiguration
> on my part.
> 
> Since the original issue is fixed:
> 
> Tested-by: Luiz Capitulino <luizcap@amazon.com>

Thanks Luiz, will take time to address this new issue

> 
>>
>> - Luiz
>>
>>>
>>> On 2023/8/4 23:24, Pu Lehui wrote:
>>>> Luiz Capitulino reported the test_verifier test failed:
>>>> "precise: ST insn causing spi > allocated_stack".
>>>> And it was introduced by the following upstream commit:
>>>> ecdf985d7615 ("bpf: track immediate values written to stack by 
>>>> BPF_ST instruction")
>>>>
>>>> Eduard's investigation [4] shows that test failure is not a bug, but a
>>>> difference in BPF verifier behavior between upstream, where commits
>>>> [1,2,3] by Andrii are present, and 5.15, where these commits are 
>>>> absent.
>>>>
>>>> Backporting strategy is consistent with Eduard in kernel version 6.1 
>>>> [5],
>>>> but with some conflicts in patch #1, #4 and #6 due to the bpf of 5.15
>>>> doesn't support more features.
>>>>
>>>> Commits of Andrii:
>>>> [1] be2ef8161572 ("bpf: allow precision tracking for programs with 
>>>> subprogs")
>>>> [2] f63181b6ae79 ("bpf: stop setting precise in current state")
>>>> [3] 7a830b53c17b ("bpf: aggressively forget precise markings during 
>>>> state checkpointing")
>>>>
>>>> Links:
>>>> [4] 
>>>> https://lore.kernel.org/stable/c9b10a8a551edafdfec855fbd35757c6238ad258.camel@gmail.com/
>>>> [5] 
>>>> https://lore.kernel.org/all/20230724124223.1176479-2-eddyz87@gmail.com/
>>>>
>>>> Andrii Nakryiko (4):
>>>>    bpf: allow precision tracking for programs with subprogs
>>>>    bpf: stop setting precise in current state
>>>>    bpf: aggressively forget precise markings during state checkpointing
>>>>    selftests/bpf: make test_align selftest more robust
>>>>
>>>> Ilya Leoshkevich (1):
>>>>    selftests/bpf: Fix sk_assign on s390x
>>>>
>>>> Yonghong Song (1):
>>>>    selftests/bpf: Workaround verification failure for
>>>>      fexit_bpf2bpf/func_replace_return_code
>>>>
>>>>   kernel/bpf/verifier.c                         | 199 
>>>> ++++++++++++++++--
>>>>   .../testing/selftests/bpf/prog_tests/align.c  |  36 ++--
>>>>   .../selftests/bpf/prog_tests/sk_assign.c      |  25 ++-
>>>>   .../selftests/bpf/progs/connect4_prog.c       |   2 +-
>>>>   .../selftests/bpf/progs/test_sk_assign.c      |  11 +
>>>>   .../bpf/progs/test_sk_assign_libbpf.c         |   3 +
>>>>   6 files changed, 243 insertions(+), 33 deletions(-)
>>>>   create mode 100644 
>>>> tools/testing/selftests/bpf/progs/test_sk_assign_libbpf.c
>>>>