Re: [PATCH bpf-next v2 3/6] libbpf: Add bpf_link support for BPF_PROG_TYPE_SK_{MSG,SKB}

[Yonghong Song <yonghong.song@linux.dev>]

On 3/22/24 11:48 AM, Andrii Nakryiko wrote:
> On Tue, Mar 19, 2024 at 10:54 AM Yonghong Song <yonghong.song@linux.dev> wrote:
>> Introduce two libbpf API functions bpf_program__attach_sk_msg()
>> and bpf_program__attach_sk_skb() which allow user to get a bpf_link
>> for their corresponding programs.
>>
>> Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
>> ---
>>   tools/lib/bpf/libbpf.c   | 14 ++++++++++++++
>>   tools/lib/bpf/libbpf.h   |  4 ++++
>>   tools/lib/bpf/libbpf.map |  2 ++
>>   3 files changed, 20 insertions(+)
>>
>> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
>> index 7d413415d0d5..1b8e1f47a5e6 100644
>> --- a/tools/lib/bpf/libbpf.c
>> +++ b/tools/lib/bpf/libbpf.c
>> @@ -149,6 +149,8 @@ static const char * const link_type_name[] = {
>>          [BPF_LINK_TYPE_TCX]                     = "tcx",
>>          [BPF_LINK_TYPE_UPROBE_MULTI]            = "uprobe_multi",
>>          [BPF_LINK_TYPE_NETKIT]                  = "netkit",
>> +       [BPF_LINK_TYPE_SK_MSG]                  = "sk_msg",
>> +       [BPF_LINK_TYPE_SK_SKB]                  = "sk_skb",
>>   };
>>
>>   static const char * const map_type_name[] = {
>> @@ -12493,6 +12495,18 @@ bpf_program__attach_netns(const struct bpf_program *prog, int netns_fd)
>>          return bpf_program_attach_fd(prog, netns_fd, "netns", NULL);
>>   }
>>
>> +struct bpf_link *
>> +bpf_program__attach_sk_msg(const struct bpf_program *prog, int map_fd)
>> +{
>> +       return bpf_program_attach_fd(prog, map_fd, "sk_msg", NULL);
>> +}
>> +
>> +struct bpf_link *
>> +bpf_program__attach_sk_skb(const struct bpf_program *prog, int map_fd)
>> +{
>> +       return bpf_program_attach_fd(prog, map_fd, "sk_skb", NULL);
>> +}
>> +
>>   struct bpf_link *bpf_program__attach_xdp(const struct bpf_program *prog, int ifindex)
>>   {
>>          /* target_fd/target_ifindex use the same field in LINK_CREATE */
>> diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
>> index 7b510761f545..4a2c04a9c6c3 100644
>> --- a/tools/lib/bpf/libbpf.h
>> +++ b/tools/lib/bpf/libbpf.h
>> @@ -784,6 +784,10 @@ bpf_program__attach_cgroup(const struct bpf_program *prog, int cgroup_fd);
>>   LIBBPF_API struct bpf_link *
>>   bpf_program__attach_netns(const struct bpf_program *prog, int netns_fd);
>>   LIBBPF_API struct bpf_link *
>> +bpf_program__attach_sk_msg(const struct bpf_program *prog, int map_fd);
>> +LIBBPF_API struct bpf_link *
>> +bpf_program__attach_sk_skb(const struct bpf_program *prog, int map_fd);
>> +LIBBPF_API struct bpf_link *
>>   bpf_program__attach_xdp(const struct bpf_program *prog, int ifindex);
>>   LIBBPF_API struct bpf_link *
>>   bpf_program__attach_freplace(const struct bpf_program *prog,
>> diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map
>> index 86804fd90dd1..11a1ad798129 100644
>> --- a/tools/lib/bpf/libbpf.map
>> +++ b/tools/lib/bpf/libbpf.map
>> @@ -410,6 +410,8 @@ LIBBPF_1.3.0 {
>>
>>   LIBBPF_1.4.0 {
>>          global:
>> +               bpf_program__attach_sk_msg;
>> +               bpf_program__attach_sk_skb;
> same suggestion, it seems like having a more generic
> bpf_program__attach_sockmap() or something along those lines would be
> better?

If we will have one link type, I think we should have one attach API
as well.

>
> please also see what changes need to be done in bpf_link_create() API

Will do.

>
>>                  bpf_token_create;
>>                  btf__new_split;
>>                  btf_ext__raw_data;
>> --
>> 2.43.0
>>