net: socket: implement SO_DESCRIPTION

This command attaches the zero terminated string in optval to the
socket for troubleshooting purposes. The free string is displayed in the
process fdinfo file for that fd (/proc/<pid>/fdinfo/<fd>).

One intended usage is to allow processes to self-document sockets
for netstat and friends to report

We ignore optlen and constrain the string to a static max size

Signed-off-by: Pascal Bouchareine <kalou@tfz.net>

include/net/sock.h

@@ -342,6 +342,7 @@ struct bpf_sk_storage;
   *	@sk_txtime_deadline_mode: set deadline mode for SO_TXTIME
   *	@sk_txtime_report_errors: set report errors mode for SO_TXTIME
   *	@sk_txtime_unused: unused txtime flags
+  *	@sk_description: user supplied with SO_DESCRIPTION
   */
 struct sock {
 	/*
@@ -519,6 +520,9 @@ struct sock {
 	struct bpf_sk_storage __rcu	*sk_bpf_storage;
 #endif
 	struct rcu_head		sk_rcu;
+
+#define	SK_MAX_DESC_SIZE	256
+	char			*sk_description;
 };
 
 enum sk_pacing {

include/uapi/asm-generic/socket.h

@@ -119,6 +119,8 @@
 
 #define SO_DETACH_REUSEPORT_BPF 68
 
+#define SO_DESCRIPTION		69
+
 #if !defined(__KERNEL__)
 
 #if __BITS_PER_LONG == 64 || (defined(__x86_64__) && defined(__ILP32__))

net/core/sock.c

@@ -828,6 +828,49 @@ void sock_set_rcvbuf(struct sock *sk, int val)
 }
 EXPORT_SYMBOL(sock_set_rcvbuf);
 
+static int sock_set_description(struct sock *sk, char __user *user_desc)
+{
+	char *old, *desc;
+
+	desc = strndup_user(user_desc, SK_MAX_DESC_SIZE, GFP_KERNEL_ACCOUNT);
+	if (IS_ERR(desc))
+		return PTR_ERR(desc);
+
+	lock_sock(sk);
+	old = sk->sk_description;
+	sk->sk_description = desc;
+	release_sock(sk);
+
+	kfree(old);
+
+	return 0;
+}
+
+static int sock_get_description(struct sock *sk, char __user *optval,
+				int __user *optlen, int len)
+{
+	char desc[SK_MAX_DESC_SIZE];
+
+	lock_sock(sk);
+	if (sk->sk_description) {
+		/* strndup_user: len(desc + nul) <= SK_MAX_DESC_SIZE */
+		len = min_t(unsigned int, len,
+			    strlen(sk->sk_description) + 1);
+		memcpy(desc, sk->sk_description, len);
+	} else {
+		len = 0;
+	}
+	release_sock(sk);
+
+	if (copy_to_user(optval, desc, len))
+		return -EFAULT;
+
+	if (put_user(len, optlen))
+		return -EFAULT;
+
+	return 0;
+}
+
 /*
  *	This is meant for all protocols to use and covers goings on
  *	at the socket level. Everything here is generic.
@@ -850,6 +893,9 @@ int sock_setsockopt(struct socket *sock, int level, int optname,
 	if (optname == SO_BINDTODEVICE)
 		return sock_setbindtodevice(sk, optval, optlen);
 
+	if (optname == SO_DESCRIPTION)
+		return sock_set_description(sk, optval);
+
 	if (optlen < sizeof(int))
 		return -EINVAL;
 
@@ -1614,6 +1660,9 @@ int sock_getsockopt(struct socket *sock, int level, int optname,
 		v.val = sk->sk_bound_dev_if;
 		break;
 
+	case SO_DESCRIPTION:
+		return sock_get_description(sk, optval, optlen, len);
+
 	default:
 		/* We implement the SO_SNDLOWAT etc to not be settable
 		 * (1003.1g 7).
@@ -1792,6 +1841,8 @@ static void __sk_destruct(struct rcu_head *head)
 		RCU_INIT_POINTER(sk->sk_filter, NULL);
 	}
 
+	kfree(sk->sk_description);
+
 	sock_disable_timestamp(sk, SK_FLAGS_TIMESTAMP);
 
 #ifdef CONFIG_BPF_SYSCALL
@@ -1964,6 +2015,8 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 		if (sk_user_data_is_nocopy(newsk))
 			newsk->sk_user_data = NULL;
 
+		newsk->sk_description = NULL;
+
 		newsk->sk_err	   = 0;
 		newsk->sk_err_soft = 0;
 		newsk->sk_priority = 0;

net/socket.c

@@ -134,6 +134,11 @@ static void sock_show_fdinfo(struct seq_file *m, struct file *f)
 {
 	struct socket *sock = f->private_data;
 
+	lock_sock(sock->sk);
+	if (sock->sk->sk_description)
+		seq_printf(m, "desc:\t%s\n", sock->sk->sk_description);
+	release_sock(sock->sk);
+
 	if (sock->ops->show_fdinfo)
 		sock->ops->show_fdinfo(m, sock);
 }