Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220713214246.2545204-1-jevburton.kernel@gmail.com>
In-Reply-To: <20220713214246.2545204-1-jevburton.kernel@gmail.com>
From:   Stanislav Fomichev <sdf@google.com>
Date:   Wed, 13 Jul 2022 14:52:14 -0700
Message-ID: <CAKH8qBtkgsQ9snhno3aYnhyc8vG2a0xhgg_sCb4KFhcQt+gfqA@mail.gmail.com>
Subject: Re: [PATCH bpf-next] libbpf: Add bpf_map__set_name()
To:     Joe Burton <jevburton.kernel@gmail.com>
Cc:     Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Andrii Nakryiko <andrii@kernel.org>,
        Martin KaFai Lau <martin.lau@linux.dev>,
        Song Liu <song@kernel.org>, Yonghong Song <yhs@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@kernel.org>, Hao Luo <haoluo@google.com>,
        Jiri Olsa <jolsa@kernel.org>, bpf@vger.kernel.org,
        linux-kernel@vger.kernel.org, Joe Burton <jevburton@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Jul 13, 2022 at 2:43 PM Joe Burton <jevburton.kernel@gmail.com> wrote:
>
> From: Joe Burton <jevburton@google.com>
>
> Add the capability to set a `struct bpf_map` name.
>
> bpf_map__reuse_fd(struct bpf_map *map, int fd) does the following:
>
> 1. get the bpf_map_info of the passed-in fd
> 2. strdup the name from the bpf_map_info
> 3. assign that name to the map
> 4. and some other stuff
>
> While `map.name` may initially be arbitrarily long, this operation
> truncates it after 15 characters.
>
> We have some infrastructure that uses bpf_map__reuse_fd() to preserve
> maps across upgrades. Some of our users have long map names, and are
> seeing their maps 'disappear' after an upgrade, due to the name
> truncation.
>
> By invoking `bpf_map__set_name()` after `bpf_map__reuse_fd()`, we can
> trivially work around the issue.

Asked you internally, but not sure I follow. Can you share more on why
the following won't fix it for us:

https://lore.kernel.org/bpf/OSZP286MB1725CEA1C95C5CB8E7CCC53FB8869@OSZP286MB1725.JPNP286.PROD.OUTLOOK.COM/

?

The idea seems to be to get the supplied map name (from the obj)
instead of using pin name? So why is it not enough?

> Signed-off-by: Joe Burton <jevburton@google.com>
> ---
>  tools/lib/bpf/libbpf.c | 22 ++++++++++++++++++++++
>  tools/lib/bpf/libbpf.h |  3 ++-
>  2 files changed, 24 insertions(+), 1 deletion(-)
>
> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> index 72548798126b..725baf508e6f 100644
> --- a/tools/lib/bpf/libbpf.c
> +++ b/tools/lib/bpf/libbpf.c
> @@ -9089,6 +9089,28 @@ const char *bpf_map__name(const struct bpf_map *map)
>         return map->name;
>  }
>
> +int bpf_map__set_name(struct bpf_map *map, const char *name)
> +{
> +       char *new_name;
> +
> +       if (!map)
> +               return libbpf_err(-EINVAL);
> +
> +       new_name = strdup(name);
> +       if (!new_name)
> +               return libbpf_err(-ENOMEM);
> +
> +       if (map_uses_real_name(map)) {
> +               free(map->real_name);
> +               map->real_name = new_name;
> +       } else {
> +               free(map->name);
> +               map->name = new_name;
> +       }
> +
> +       return 0;
> +}
> +
>  enum bpf_map_type bpf_map__type(const struct bpf_map *map)
>  {
>         return map->def.type;
> diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
> index e4d5353f757b..e898c4cb514a 100644
> --- a/tools/lib/bpf/libbpf.h
> +++ b/tools/lib/bpf/libbpf.h
> @@ -731,8 +731,9 @@ LIBBPF_API bool bpf_map__autocreate(const struct bpf_map *map);
>   */
>  LIBBPF_API int bpf_map__fd(const struct bpf_map *map);
>  LIBBPF_API int bpf_map__reuse_fd(struct bpf_map *map, int fd);
> -/* get map name */
> +/* get/set map name */
>  LIBBPF_API const char *bpf_map__name(const struct bpf_map *map);
> +LIBBPF_API int bpf_map__set_name(struct bpf_map *map, const char *name);
>  /* get/set map type */
>  LIBBPF_API enum bpf_map_type bpf_map__type(const struct bpf_map *map);
>  LIBBPF_API int bpf_map__set_type(struct bpf_map *map, enum bpf_map_type type);
> --
> 2.37.0.144.g8ac04bfd2-goog
>