netdev - Re: [PATCH v5 bpf-next 0/4] capture integers in BTF type info for map defs

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: <CAEf4BzbkAq=C5NF1_XRc1WNE-i7pavt48k9D7kEM_oTS3pWADA@mail.gmail.com>
Date:   Fri, 5 Jul 2019 20:49:23 -0700
From:   Andrii Nakryiko <andrii.nakryiko@...il.com>
To:     Daniel Borkmann <daniel@...earbox.net>
Cc:     Andrii Nakryiko <andriin@...com>, Kernel Team <kernel-team@...com>,
        Alexei Starovoitov <ast@...com>,
        Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH v5 bpf-next 0/4] capture integers in BTF type info for map defs

On Fri, Jul 5, 2019 at 2:15 PM Daniel Borkmann <daniel@...earbox.net> wrote:
>
> On 07/05/2019 05:50 PM, Andrii Nakryiko wrote:
> > This patch set implements an update to how BTF-defined maps are specified. The
> > change is in how integer attributes, e.g., type, max_entries, map_flags, are
> > specified: now they are captured as part of map definition struct's BTF type
> > information (using array dimension), eliminating the need for compile-time
> > data initialization and keeping all the metadata in one place.
> >
> > All existing selftests that were using BTF-defined maps are updated, along
> > with some other selftests, that were switched to new syntax.
> >
> > v4->v5:
> > - revert sample_map_ret0.c, which is loaded with iproute2 (kernel test robot);
> > v3->v4:
> > - add acks;
> > - fix int -> uint type in commit message;
> > v2->v3:
> > - rename __int into __uint (Yonghong);
> > v1->v2:
> > - split bpf_helpers.h change from libbpf change (Song).
> >
> > Andrii Nakryiko (4):
> >   libbpf: capture value in BTF type info for BTF-defined map defs
> >   selftests/bpf: add __uint and __type macro for BTF-defined maps
> >   selftests/bpf: convert selftests using BTF-defined maps to new syntax
> >   selftests/bpf: convert legacy BPF maps to BTF-defined ones
> >
> >  tools/lib/bpf/libbpf.c                        |  58 +++++----
> >  tools/testing/selftests/bpf/bpf_helpers.h     |   3 +
> >  tools/testing/selftests/bpf/progs/bpf_flow.c  |  28 ++---
> >  .../selftests/bpf/progs/get_cgroup_id_kern.c  |  26 ++---
> >  .../testing/selftests/bpf/progs/netcnt_prog.c |  20 ++--
> >  tools/testing/selftests/bpf/progs/pyperf.h    |  90 +++++++-------
> >  .../selftests/bpf/progs/socket_cookie_prog.c  |  13 +--
> >  .../bpf/progs/sockmap_verdict_prog.c          |  48 ++++----
> >  .../testing/selftests/bpf/progs/strobemeta.h  |  68 +++++------
> >  .../selftests/bpf/progs/test_btf_newkv.c      |  13 +--
> >  .../bpf/progs/test_get_stack_rawtp.c          |  39 +++----
> >  .../selftests/bpf/progs/test_global_data.c    |  37 +++---
> >  tools/testing/selftests/bpf/progs/test_l4lb.c |  65 ++++-------
> >  .../selftests/bpf/progs/test_l4lb_noinline.c  |  65 ++++-------
> >  .../selftests/bpf/progs/test_map_in_map.c     |  30 ++---
> >  .../selftests/bpf/progs/test_map_lock.c       |  26 ++---
> >  .../testing/selftests/bpf/progs/test_obj_id.c |  12 +-
> >  .../bpf/progs/test_select_reuseport_kern.c    |  67 ++++-------
> >  .../bpf/progs/test_send_signal_kern.c         |  26 ++---
> >  .../bpf/progs/test_sock_fields_kern.c         |  78 +++++--------
> >  .../selftests/bpf/progs/test_spin_lock.c      |  36 +++---
> >  .../bpf/progs/test_stacktrace_build_id.c      |  55 ++++-----
> >  .../selftests/bpf/progs/test_stacktrace_map.c |  52 +++------
> >  .../selftests/bpf/progs/test_tcp_estats.c     |  13 +--
> >  .../selftests/bpf/progs/test_tcpbpf_kern.c    |  26 ++---
> >  .../selftests/bpf/progs/test_tcpnotify_kern.c |  28 ++---
> >  tools/testing/selftests/bpf/progs/test_xdp.c  |  26 ++---
> >  .../selftests/bpf/progs/test_xdp_loop.c       |  26 ++---
> >  .../selftests/bpf/progs/test_xdp_noinline.c   |  81 +++++--------
> >  .../selftests/bpf/progs/xdp_redirect_map.c    |  12 +-
> >  .../testing/selftests/bpf/progs/xdping_kern.c |  12 +-
> >  .../selftests/bpf/test_queue_stack_map.h      |  30 ++---
> >  .../testing/selftests/bpf/test_sockmap_kern.h | 110 +++++++++---------
> >  33 files changed, 559 insertions(+), 760 deletions(-)
>
> LGTM, applied, thanks! Shouldn't we also move __uint and __type macros
> into libbpf as otherwise people tend to redefine this over and over?


Yes, we need something like bpf_helpers.h as part of libbpf for
inclusion into BPF programs. It's on my todo list as part of BPF CO-RE
project as well.