netdev - Re: [PATCH bpf-next v4 2/3] bpf: selftests: remove percpu macros from bpf

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <e9c5baa2-62e1-86eb-6cde-a6ceec8f05dc@iogearbox.net>
Date:   Tue, 20 Apr 2021 17:58:27 +0200
From:   Daniel Borkmann <daniel@...earbox.net>
To:     Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Pedro Tammela <pctammela@...il.com>
Cc:     Alexei Starovoitov <ast@...nel.org>,
        Andrii Nakryiko <andrii@...nel.org>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        John Fastabend <john.fastabend@...il.com>,
        KP Singh <kpsingh@...nel.org>, Shuah Khan <shuah@...nel.org>,
        Pedro Tammela <pctammela@...atatu.com>,
        David Verbeiren <david.verbeiren@...sares.net>,
        Matthieu Baerts <matthieu.baerts@...sares.net>,
        Network Development <netdev@...r.kernel.org>,
        bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH bpf-next v4 2/3] bpf: selftests: remove percpu macros from
 bpf_util.h

On 4/20/21 3:17 AM, Alexei Starovoitov wrote:
> On Thu, Apr 15, 2021 at 10:47 AM Pedro Tammela <pctammela@...il.com> wrote:
>>
>> Andrii suggested to remove this abstraction layer and have the percpu
>> handling more explicit[1].
>>
>> This patch also updates the tests that relied on the macros.
>>
>> [1] https://lore.kernel.org/bpf/CAEf4BzYmj_ZPDq8Zi4dbntboJKRPU2TVopysBNrdd9foHTfLZw@mail.gmail.com/
>>
>> Suggested-by: Andrii Nakryiko <andrii@...nel.org>
>> Signed-off-by: Pedro Tammela <pctammela@...atatu.com>
>> ---
>>   tools/testing/selftests/bpf/bpf_util.h        |  7 --
>>   .../bpf/map_tests/htab_map_batch_ops.c        | 87 +++++++++----------
>>   .../selftests/bpf/prog_tests/map_init.c       |  9 +-
>>   tools/testing/selftests/bpf/test_maps.c       | 84 +++++++++++-------
>>   4 files changed, 96 insertions(+), 91 deletions(-)
>>
>> diff --git a/tools/testing/selftests/bpf/bpf_util.h b/tools/testing/selftests/bpf/bpf_util.h
>> index a3352a64c067..105db3120ab4 100644
>> --- a/tools/testing/selftests/bpf/bpf_util.h
>> +++ b/tools/testing/selftests/bpf/bpf_util.h
>> @@ -20,13 +20,6 @@ static inline unsigned int bpf_num_possible_cpus(void)
>>          return possible_cpus;
>>   }
>>
>> -#define __bpf_percpu_val_align __attribute__((__aligned__(8)))
>> -
>> -#define BPF_DECLARE_PERCPU(type, name)                         \
>> -       struct { type v; /* padding */ } __bpf_percpu_val_align \
>> -               name[bpf_num_possible_cpus()]
>> -#define bpf_percpu(name, cpu) name[(cpu)].v
>> -
> 
> Hmm. I wonder what Daniel has to say about it, since he
> introduced it in commit f3515b5d0b71 ("bpf: provide a generic macro
> for percpu values for selftests")
> to address a class of bugs.

I would probably even move those into libbpf instead. ;-) The problem is that this can
be missed easily and innocent changes would lead to corruption of the applications
memory if there's a map lookup. Having this at least in selftest code or even in libbpf
would document code-wise that care needs to be taken on per cpu maps. Even if we'd put
a note under Documentation/bpf/ or such, this might get missed easily and finding such
bugs is like looking for a needle in a haystack.. so I don't think this should be removed.

Thanks,
Daniel