netdev - Re: [PATCH bpf-next v2 0/6] skip verifier/map tests if kernel support is missing

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: <20181218212501.bw5qn3fae6snubze@ast-mbp>
Date:   Tue, 18 Dec 2018 13:25:02 -0800
From:   Alexei Starovoitov <alexei.starovoitov@...il.com>
To:     Stanislav Fomichev <sdf@...gle.com>
Cc:     netdev@...r.kernel.org, ast@...nel.org, davem@...emloft.net,
        daniel@...earbox.net, ecree@...arflare.com,
        quentin.monnet@...ronome.com
Subject: Re: [PATCH bpf-next v2 0/6] skip verifier/map tests if kernel
 support is missing

On Mon, Dec 17, 2018 at 10:25:48AM -0800, Stanislav Fomichev wrote:
> If test_maps/test_verifier is running against the kernel which doesn't
> have _all_ BPF features enabled, it fails with an error. This patch
> series tries to probe kernel support for each failed test and skip
> it instead. This lets users run BPF selftests in the not-all-bpf-yes
> environments and received correct PASS/NON-PASS result.
> 
> See https://www.spinics.net/lists/netdev/msg539331.html for more
> context.
> 
> The series goes like this:
> 
> * patch #1 adds bpf_prog_type_supported() and
>   bpf_map_type_supported() which query the kernel (insert 'return 0'
>   program or try to create empty map with correct key/value sizes) and
>   return supported/unsupported.
>   Note: this functionality can later be reimplemented on top of Quentin's
>   recent 'bpftool feature' patchset if he decides to move the probes
>   into libbpf.
> * patch #2 skips sockmap tests in test_maps.c if BPF_MAP_TYPE_SOCKMAP
>   map is not supported (if bpf_create_map fails, we probe the kernel
>   for support)
> * patch #3 skips verifier tests if test->prog_type is not supported (if
>   bpf_verify_program fails, we probe the kernel for support)
> * patch #4 skips verifier tests if test fixup map is not supported (if
>   create_map fails, we probe the kernel for support)
>   Note: we can probably move this probe into create_map helper and
>   return some argument instead of adding skip_unsupported_map()
>   to each fixup; but I'm not sure it's better.
>   Also note: in current implementation we still print 'Failed to
>   create hash map' from the create_map, but still skip the test.
> * next patches fix various small issues that arise from the first four:
>   * patch #5 sets "unknown func bpf_trace_printk#6" prog_type to
>     BPF_PROG_TYPE_TRACEPOINT so it is correctly skipped in
>     CONFIG_BPF_EVENTS=n case
>   * patch #6 exposes BPF_PROG_TYPE_CGROUP_{SKB,SOCK,SOCK_ADDR} only when
>     CONFIG_CGROUP_BPF=y, this makes verifier correctly skip appropriate
>     tests
> 
> v2 changes:
> 
> * don't sprinkle "ifdef CONFIG_CGROUP_BPF" all around net/core/filter.c,
>   doing it only in the bpf_types.h is enough to disable
>   BPF_PROG_TYPE_CGROUP_{SKB,SOCK,SOCK_ADDR} prog types for non-cgroup
>   enabled kernels

the patches look good to me.
I think it's ok to proceed this way though long term we probably
want to have such bpf_prog_type_supported() to be part of libbpf
and reused in test_verifier.c and in bpftool.
Daniel, thoughts?