Re: [PATCH v3 bpf-next/net 2/5] bpf: Support bpf_setsockopt() for BPF_CGROUP_INET_SOCK_CREATE.

On Wed, Aug 27, 2025 at 3:23 PM Martin KaFai Lau [off-list ref] wrote:

On 8/26/25 11:38 AM, Kuniyuki Iwashima wrote:

quoted

We will store a flag in sk->sk_memcg by bpf_setsockopt() during
socket() or before sk->sk_memcg is set in accept().

BPF_CGROUP_INET_SOCK_CREATE is invoked by __cgroup_bpf_run_filter_sk()
that passes a pointer to struct sock to the bpf prog as void *ctx.

But there are no bpf_func_proto for bpf_setsockopt() that receives
the ctx as a pointer to struct sock.

Let's add a new bpf_setsockopt() variant for BPF_CGROUP_INET_SOCK_CREATE.

Note that inet_create() is not under lock_sock().

Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
---
v3: Remove bpf_func_proto for accept()
v2: Make 2 new bpf_func_proto static
---
  net/core/filter.c | 24 ++++++++++++++++++++++++
  1 file changed, 24 insertions(+)

diff --git a/net/core/filter.c b/net/core/filter.c
index 63f3baee2daf..443d12b7d3b2 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c

@@ -5743,6 +5743,23 @@ static const struct bpf_func_proto bpf_sock_ops_setsockopt_proto = {
      .arg5_type      = ARG_CONST_SIZE,
  };

+BPF_CALL_5(bpf_unlocked_sock_setsockopt, struct sock *, sk, int, level,
+        int, optname, char *, optval, int, optlen)
+{
+     return _bpf_setsockopt(sk, level, optname, optval, optlen);

The sock_owned_by_me() will warn.

 From CI:
WARNING: CPU: 0 PID: 102 at include/net/sock.h:1756 bpf_unlocked_sock_setsockopt+0xc7/0x110

Oh sorry, I copied from a wrong place.. will fix it.

BTW, I'm thinking I should inherit flags from the listener
in sk_clone_lock() and disallow other bpf hooks.

Given the listener's flag and bpf hooks come from the
same cgroup, there is no point having other hooks.

quoted

+}
+
+static const struct bpf_func_proto bpf_unlocked_sock_setsockopt_proto = {
+     .func           = bpf_unlocked_sock_setsockopt,
+     .gpl_only       = false,
+     .ret_type       = RET_INTEGER,
+     .arg1_type      = ARG_PTR_TO_CTX,
+     .arg2_type      = ARG_ANYTHING,
+     .arg3_type      = ARG_ANYTHING,
+     .arg4_type      = ARG_PTR_TO_MEM | MEM_RDONLY,
+     .arg5_type      = ARG_CONST_SIZE,
+};
+
  static int bpf_sock_ops_get_syn(struct bpf_sock_ops_kern *bpf_sock,
                              int optname, const u8 **start)
  {

@@ -8051,6 +8068,13 @@ sock_filter_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
              return &bpf_sk_storage_get_cg_sock_proto;
      case BPF_FUNC_ktime_get_coarse_ns:
              return &bpf_ktime_get_coarse_ns_proto;
+     case BPF_FUNC_setsockopt:
+             switch (prog->expected_attach_type) {
+             case BPF_CGROUP_INET_SOCK_CREATE:
+                     return &bpf_unlocked_sock_setsockopt_proto;
+             default:
+                     return NULL;
+             }
      default:
              return bpf_base_func_proto(func_id, prog);
      }