netdev - Re: [PATCH bpf-next v3 4/8] libbpf: implement bpf_prog_find

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 for Android: free password hash cracker in your pocket

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

Message-ID: <87mu27hnji.fsf@toke.dk>
Date:   Wed, 02 Sep 2020 23:33:05 +0200
From:   Toke Høiland-Jørgensen <toke@...hat.com>
To:     Alexei Starovoitov <alexei.starovoitov@...il.com>
Cc:     sdf@...gle.com, netdev@...r.kernel.org, bpf@...r.kernel.org,
        davem@...emloft.net, ast@...nel.org, daniel@...earbox.net,
        YiFei Zhu <zhuyifei1999@...il.com>, andriin@...com
Subject: Re: [PATCH bpf-next v3 4/8] libbpf: implement bpf_prog_find_metadata

Alexei Starovoitov <alexei.starovoitov@...il.com> writes:

> On Wed, Sep 02, 2020 at 11:43:26AM +0200, Toke HÃƒÂ¸iland-JÃƒÂ¸rgensen wrote:
>> >
>> > I don't feel great about this libbpf api. bpftool already does
>> > bpf_obj_get_info_by_fd() for progs and for maps.
>> > This extra step and extra set of syscalls is redundant work.
>> > I think it's better to be done as part of bpftool.
>> > It doesn't quite fit as generic api.
>> 
>> Why not? 
>
> It's a helper function on top of already provided api and implemented
> in the most brute force and inefficient way.
> bpftool implementation of the same will be more efficient.

Right, certainly wouldn't mind something more efficient. But to me, the
inefficiency is outweighed by convenience of having this canonical
reference for 'this is the metadata map'.

>> so. If we don't have it, people will have to go look at bpftool code,
>> and we'll end up with copied code snippets, which seems less than ideal.
>
> I'd like to see the real use case first before hypothesising.

For me, that would be incorporating support for this into
libxdp/xdp-tools; which was the reason I asked for this to be split into
a separate API in the first place. But okay, not going to keep arguing
about this, I can copy-paste code as well as the next person.

-Toke