Re: [PATCH bpf-next 1/2] libbpf: Implement btf__save_to_file()

[Andrii Nakryiko <andrii.nakryiko@gmail.com>]

On Wed, Oct 27, 2021 at 1:37 PM Mauricio Vásquez <mauricio@kinvolk.io> wrote:
>
> Implement helper function to save the contents of a BTF object to a
> file.
>
> Signed-off-by: Mauricio Vásquez <mauricio@kinvolk.io>
> Signed-off-by: Rafael David Tinoco <rafael.tinoco@aquasec.com>
> Signed-off-by: Lorenzo Fontana <lorenzo.fontana@elastic.co>
> ---
>  tools/lib/bpf/btf.c      | 22 ++++++++++++++++++++++
>  tools/lib/bpf/btf.h      |  2 ++
>  tools/lib/bpf/libbpf.map |  1 +
>  3 files changed, 25 insertions(+)
>
> diff --git a/tools/lib/bpf/btf.c b/tools/lib/bpf/btf.c
> index 0c628c33e23b..087035574dba 100644
> --- a/tools/lib/bpf/btf.c
> +++ b/tools/lib/bpf/btf.c
> @@ -4773,3 +4773,25 @@ int btf_ext_visit_str_offs(struct btf_ext *btf_ext, str_off_visit_fn visit, void
>
>         return 0;
>  }
> +
> +int btf__save_to_file(struct btf *btf, const char *path)

given we have its loading counterpart as btf__parse_raw(), let's call
this one btf__save_raw()?

> +{
> +       const void *data;
> +       __u32 data_sz;
> +       FILE *file;
> +
> +       data = btf_get_raw_data(btf, &data_sz, btf->swapped_endian);

use btf__raw_data() instead? no need to think about btf->swapped_endian here

> +       if (!data)
> +               return -ENOMEM;

please use libbpf_err() helper for returning errors, see other use cases

> +
> +       file = fopen(path, "wb");
> +       if (!file)
> +               return -errno;
> +
> +       if (fwrite(data, 1, data_sz, file) != data_sz) {
> +               fclose(file);
> +               return -EIO;

why not propagating original errno? make sure you save it before
fclose(), though


> +       }
> +
> +       return fclose(file);

hm... I'd just do fclose(file) separately and return 0 (because
success). If file closing fails, there isn't anything that can be done
(but it also shouldn't fail in any normal situation).

> +}
> diff --git a/tools/lib/bpf/btf.h b/tools/lib/bpf/btf.h
> index bc005ba3ceec..300ad498c615 100644
> --- a/tools/lib/bpf/btf.h
> +++ b/tools/lib/bpf/btf.h
> @@ -114,6 +114,8 @@ LIBBPF_API struct btf *btf__parse_elf_split(const char *path, struct btf *base_b
>  LIBBPF_API struct btf *btf__parse_raw(const char *path);
>  LIBBPF_API struct btf *btf__parse_raw_split(const char *path, struct btf *base_btf);
>
> +LIBBPF_API int btf__save_to_file(struct btf *btf, const char *path);

const struct btf? btf__raw_data() (even though it internally modifies
btf) accepts `const struct btf*`, because this is conceptually
read-only operation

> +
>  LIBBPF_API struct btf *btf__load_vmlinux_btf(void);
>  LIBBPF_API struct btf *btf__load_module_btf(const char *module_name, struct btf *vmlinux_btf);
>  LIBBPF_API struct btf *libbpf_find_kernel_btf(void);
> diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map
> index 15239c05659c..0e9bed7c9b9e 100644
> --- a/tools/lib/bpf/libbpf.map
> +++ b/tools/lib/bpf/libbpf.map
> @@ -399,4 +399,5 @@ LIBBPF_0.6.0 {
>                 btf__add_decl_tag;
>                 btf__raw_data;
>                 btf__type_cnt;
> +               btf__save_to_file;
>  } LIBBPF_0.5.0;
> --
> 2.25.1
>