why not use the tcp_set_state() in inet_csk_listen_start()?

why not use the tcp_set_state() in inet_csk_listen_start()?

[杨硕]

why use "sk->sk_state = TCP_LISTEN" rather than "tcp_set_state(sk, TCP_LISTEN)"?

TIA :)

int inet_csk_listen_start(struct sock *sk, const int nr_table_entries)
{
	struct inet_sock *inet = inet_sk(sk);
	struct inet_connection_sock *icsk = inet_csk(sk);
	int rc = reqsk_queue_alloc(&icsk->icsk_accept_queue, nr_table_entries);

	if (rc != 0)
		return rc;

	sk->sk_max_ack_backlog = 0;
	sk->sk_ack_backlog = 0;
	inet_csk_delack_init(sk);

	/* There is race window here: we announce ourselves listening,
	 * but this transition is still not validated by get_port().
	 * It is OK, because this socket enters to hash table only
	 * after validation is complete.
	 */
	sk->sk_state = TCP_LISTEN;
	if (!sk->sk_prot->get_port(sk, inet->num)) {
		inet->sport = htons(inet->num);

		sk_dst_reset(sk);
		sk->sk_prot->hash(sk);

		return 0;
	}

	sk->sk_state = TCP_CLOSE;
	__reqsk_queue_destroy(&icsk->icsk_accept_queue);
	return -EADDRINUSE;
}

Re: why not use the tcp_set_state() in inet_csk_listen_start()?

[Eric Dumazet]

Le mercredi 24 mars 2010 à 16:13 +0800, 杨硕 a écrit :
> why use "sk->sk_state = TCP_LISTEN" rather than "tcp_set_state(sk, TCP_LISTEN)"?
> 
> TIA :)
> 
> int inet_csk_listen_start(struct sock *sk, const int nr_table_entries)
> {
> 	struct inet_sock *inet = inet_sk(sk);
> 	struct inet_connection_sock *icsk = inet_csk(sk);
> 	int rc = reqsk_queue_alloc(&icsk->icsk_accept_queue, nr_table_entries);
> 
> 	if (rc != 0)
> 		return rc;
> 
> 	sk->sk_max_ack_backlog = 0;
> 	sk->sk_ack_backlog = 0;
> 	inet_csk_delack_init(sk);
> 
> 	/* There is race window here: we announce ourselves listening,
> 	 * but this transition is still not validated by get_port().
> 	 * It is OK, because this socket enters to hash table only
> 	 * after validation is complete.
> 	 */
> 	sk->sk_state = TCP_LISTEN;

Because its not necessary ?

tcp_set_state() takes care of particular state transitions.

In case of inet_csk_listen_start(), old state is TCP_CLOSE, so there is
nothing special to do in tcp_set_state().

Re: why not use the tcp_set_state() in inet_csk_listen_start()?

[杨硕]

But if we use "STATE_TRACE", maybe the state will be incomplete.

> Because its not necessary ?
>
> tcp_set_state() takes care of particular state transitions.
>
> In case of inet_csk_listen_start(), old state is TCP_CLOSE, so there is
> nothing special to do in tcp_set_state().
>