[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAEf4BzZZXx28w1y_6xfsue91c_7whvHzMhKvbSnsQRU4yA+RwA@mail.gmail.com>
Date: Thu, 22 Jul 2021 19:45:32 -0700
From: Andrii Nakryiko <andrii.nakryiko@...il.com>
To: Quentin Monnet <quentin@...valent.com>
Cc: Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Andrii Nakryiko <andrii@...nel.org>,
Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH bpf-next v2 0/5] libbpf: rename btf__get_from_id() and
btf__load() APIs, support split BTF
On Thu, Jul 22, 2021 at 5:58 PM Andrii Nakryiko
<andrii.nakryiko@...il.com> wrote:
>
> On Wed, Jul 21, 2021 at 8:38 AM Quentin Monnet <quentin@...valent.com> wrote:
> >
> > As part of the effort to move towards a v1.0 for libbpf [0], this set
> > improves some confusing function names related to BTF loading from and to
> > the kernel:
> >
> > - btf__load() becomes btf__load_into_kernel().
> > - btf__get_from_id becomes btf__load_from_kernel_by_id().
> > - A new version btf__load_from_kernel_by_id_split() extends the former to
> > add support for split BTF.
> >
> > The old functions are not removed or marked as deprecated yet, there
> > should be in a future libbpf version.
>
> Oh, and I was thinking about this whole deprecation having to be done
> in two steps. It's super annoying to keep track of that. Ideally, we'd
> have some macro that can mark API deprecated "in the future", when
> actual libbpf version is >= to defined version. So something like
> this:
>
> LIBBPF_DEPRECATED_AFTER(V(0,5), "API that will be marked deprecated in v0.6")
Better:
LIBBPF_DEPRECATED_SINCE(0, 6, "API that will be marked deprecated in v0.6")
>
>
> We'd need to make sure that during the build time we have some
> LIBBPF_VERSION macro available against which we compare the expected
> version and add or not the __attribute__((deprecated)).
>
> Does this make sense? WDYT? I haven't looked into how hard it would be
> to implement this, but it should be easy enough, so if you'd like some
> macro challenge, please take a stab at it.
>
> Having this it would be possible to make all the deprecations at the
> same time that we add replacement APIs and not ask anyone to follow-up
> potentially a month or two later, right?
>
> >
> > The last patch is a trivial change to bpftool to add support for dumping
> > split BTF objects by referencing them by their id (and not only by their
> > BTF path).
> >
> > [0] https://github.com/libbpf/libbpf/wiki/Libbpf:-the-road-to-v1.0#btfh-apis
> >
> > v2:
> > - Remove deprecation marking of legacy functions (patch 4/6 from v1).
> > - Make btf__load_from_kernel_by_id{,_split}() return the btf struct.
> > - Add new functions to v0.5.0 API (and not v0.6.0).
> >
> > Quentin Monnet (5):
> > libbpf: rename btf__load() as btf__load_into_kernel()
> > libbpf: rename btf__get_from_id() as btf__load_from_kernel_by_id()
> > tools: replace btf__get_from_id() with btf__load_from_kernel_by_id()
> > libbpf: add split BTF support for btf__load_from_kernel_by_id()
> > tools: bpftool: support dumping split BTF by id
> >
> > tools/bpf/bpftool/btf.c | 8 ++---
> > tools/bpf/bpftool/btf_dumper.c | 6 ++--
> > tools/bpf/bpftool/map.c | 16 +++++-----
> > tools/bpf/bpftool/prog.c | 29 +++++++++++------
> > tools/lib/bpf/btf.c | 33 ++++++++++++++------
> > tools/lib/bpf/btf.h | 4 +++
> > tools/lib/bpf/libbpf.c | 7 +++--
> > tools/lib/bpf/libbpf.map | 3 ++
> > tools/perf/util/bpf-event.c | 11 ++++---
> > tools/perf/util/bpf_counter.c | 12 +++++--
> > tools/testing/selftests/bpf/prog_tests/btf.c | 4 ++-
> > 11 files changed, 86 insertions(+), 47 deletions(-)
> >
> > --
> > 2.30.2
> >
Powered by blists - more mailing lists