bpf_skc_to_ssk_sock

Signed-off-by: Geliang Tang <[email protected]>

include/linux/bpf.h

@@ -3259,6 +3259,7 @@ extern const struct bpf_func_proto bpf_skc_to_tcp_request_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_udp6_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_unix_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_mptcp_sock_proto;
+extern const struct bpf_func_proto bpf_skc_to_ssk_sock_proto;
 extern const struct bpf_func_proto bpf_copy_from_user_proto;
 extern const struct bpf_func_proto bpf_snprintf_btf_proto;
 extern const struct bpf_func_proto bpf_snprintf_proto;

include/uapi/linux/bpf.h

@@ -5796,6 +5796,12 @@ union bpf_attr {
  *		0 on success.
  *
  *		**-ENOENT** if the bpf_local_storage cannot be found.
+ *
+ * void *bpf_skc_to_ssk_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or **NULL** otherwise.
  */
 #define ___BPF_FUNC_MAPPER(FN, ctx...)			\
 	FN(unspec, 0, ##ctx)				\
@@ -6010,6 +6016,7 @@ union bpf_attr {
 	FN(user_ringbuf_drain, 209, ##ctx)		\
 	FN(cgrp_storage_get, 210, ##ctx)		\
 	FN(cgrp_storage_delete, 211, ##ctx)		\
+	FN(skc_to_ssk_sock, 212, ##ctx)			\
 	/* */
 
 /* backwards-compatibility macros for users of __BPF_FUNC_MAPPER that don't

net/core/filter.c

@@ -11844,6 +11844,20 @@ const struct bpf_func_proto bpf_skc_to_mptcp_sock_proto = {
 	.ret_btf_id	= &btf_sock_ids[BTF_SOCK_TYPE_MPTCP],
 };
 
+BPF_CALL_1(bpf_skc_to_ssk_sock, struct sock *, sk)
+{
+	BTF_TYPE_EMIT(struct sock);
+	return (unsigned long)sk;
+}
+
+const struct bpf_func_proto bpf_skc_to_ssk_sock_proto = {
+	.func		= bpf_skc_to_ssk_sock,
+	.gpl_only	= false,
+	.ret_type	= RET_PTR_TO_BTF_ID_OR_NULL,
+	.arg1_type	= ARG_ANYTHING,
+	.ret_btf_id	= &btf_sock_ids[BTF_SOCK_TYPE_SOCK],
+};
+
 BPF_CALL_1(bpf_sock_from_file, struct file *, file)
 {
 	return (unsigned long)sock_from_file(file);
@@ -11889,6 +11903,9 @@ bpf_sk_base_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_skc_to_mptcp_sock:
 		func = &bpf_skc_to_mptcp_sock_proto;
 		break;
+	case BPF_FUNC_skc_to_ssk_sock:
+		func = &bpf_skc_to_ssk_sock_proto;
+		break;
 	case BPF_FUNC_ktime_get_coarse_ns:
 		return &bpf_ktime_get_coarse_ns_proto;
 	default:

net/mptcp/bpf.c

@@ -294,6 +294,8 @@ bpf_mptcp_sched_get_func_proto(enum bpf_func_id func_id,
 		return &bpf_skc_to_tcp6_sock_proto;
 	case BPF_FUNC_skc_to_tcp_sock:
 		return &bpf_skc_to_tcp_sock_proto;
+	case BPF_FUNC_skc_to_ssk_sock:
+		return &bpf_skc_to_ssk_sock_proto;
 	default:
 		return bpf_base_func_proto(func_id, prog);
 	}

tools/include/uapi/linux/bpf.h

@@ -5796,6 +5796,12 @@ union bpf_attr {
  *		0 on success.
  *
  *		**-ENOENT** if the bpf_local_storage cannot be found.
+ *
+ * void *bpf_skc_to_ssk_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or **NULL** otherwise.
  */
 #define ___BPF_FUNC_MAPPER(FN, ctx...)			\
 	FN(unspec, 0, ##ctx)				\
@@ -6010,6 +6016,7 @@ union bpf_attr {
 	FN(user_ringbuf_drain, 209, ##ctx)		\
 	FN(cgrp_storage_get, 210, ##ctx)		\
 	FN(cgrp_storage_delete, 211, ##ctx)		\
+	FN(skc_to_ssk_sock, 212, ##ctx)			\
 	/* */
 
 /* backwards-compatibility macros for users of __BPF_FUNC_MAPPER that don't

tools/testing/selftests/bpf/progs/mptcp_bpf_burst.c

@@ -15,11 +15,6 @@ char _license[] SEC("license") = "GPL";
 
 #define min(a, b) ((a) < (b) ? (a) : (b))
 
-struct bpf_subflow_send_info {
-	struct mptcp_subflow_context *subflow;
-	__u64 linger_time;
-};
-
 extern bool mptcp_subflow_active(struct mptcp_subflow_context *subflow) __ksym;
 extern void mptcp_set_timeout(struct sock *sk) __ksym;
 extern __u64 mptcp_wnd_end(const struct mptcp_sock *msk) __ksym;
@@ -70,7 +65,7 @@ void BPF_PROG(mptcp_sched_burst_release, struct mptcp_sock *msk)
 SEC("struct_ops")
 int BPF_PROG(bpf_burst_get_send, struct mptcp_sock *msk)
 {
-	struct bpf_subflow_send_info send_info[SSK_MODE_MAX];
+	struct subflow_send_info send_info[SSK_MODE_MAX];
 	struct mptcp_subflow_context *subflow;
 	struct sock *sk = (struct sock *)msk;
 	__u32 pace, burst, wmem;
@@ -80,7 +75,7 @@ int BPF_PROG(bpf_burst_get_send, struct mptcp_sock *msk)
 
 	/* pick the subflow with the lower wmem/wspace ratio */
 	for (i = 0; i < SSK_MODE_MAX; ++i) {
-		send_info[i].subflow = NULL;
+		send_info[i].ssk = NULL;
 		send_info[i].linger_time = -1;
 	}
 
@@ -103,24 +98,22 @@ int BPF_PROG(bpf_burst_get_send, struct mptcp_sock *msk)
 
 		linger_time = div_u64((__u64)ssk->sk_wmem_queued << 32, pace);
 		if (linger_time < send_info[backup].linger_time) {
-			if (!sk_stream_memory_free(ssk))
-				continue;
-			send_info[backup].subflow = subflow;
+			send_info[backup].ssk = ssk;
 			send_info[backup].linger_time = linger_time;
 		}
 	}
 	mptcp_set_timeout(sk);
 
 	/* pick the best backup if no other subflow is active */
 	if (!nr_active)
-		send_info[SSK_MODE_ACTIVE].subflow = send_info[SSK_MODE_BACKUP].subflow;
+		send_info[SSK_MODE_ACTIVE].ssk = send_info[SSK_MODE_BACKUP].ssk;
 
-	subflow = bpf_core_cast(send_info[SSK_MODE_ACTIVE].subflow,
-				struct mptcp_subflow_context);
-	if (!subflow)
+	ssk = bpf_skc_to_ssk_sock(send_info[SSK_MODE_ACTIVE].ssk);
+	if (!ssk || !sk_stream_memory_free(ssk))
 		return -1;
-	ssk = mptcp_subflow_tcp_sock(subflow);
-	if (!ssk)
+
+	subflow = bpf_mptcp_subflow_ctx(ssk);
+	if (!subflow)
 		return -1;
 
 	burst = min(MPTCP_SEND_BURST_SIZE, mptcp_wnd_end(msk) - msk->snd_nxt);

tools/testing/selftests/bpf/progs/mptcp_bpf_rr.c

@@ -59,7 +59,6 @@ static int bpf_rr_get_subflow(struct mptcp_sock *msk)
 	}
 
 out:
-	next = bpf_core_cast(next, struct mptcp_subflow_context);
 	mptcp_subflow_set_scheduled(next, true);
 	//bpf_printk("rr subflow=%u/%u", next->subflow_id, msk->pm.subflows + 1);
 	ptr->last_snd = mptcp_subflow_tcp_sock(next);