In tcp_input.c, function tcp_rcv_state_process(),
when doing ack processing of state TCP_FIN_WAIT1,
there are statements (kernel 2.4.21):
if (!sk->dead)
......
else {
......
if (tmo > TCP_TIMEWAIT_LEN) {
......
--->> } else if (th->fin || sk->lock.users) {
/* Bad case. We could lose such FIN otherwise.
* It is not a big problem, but it looks confusing
* and not so rare event. We still can lose it now,
* if it spins in bh_lock_sock(), but it is really
* marginal case.
*/
tcp_reset_keepalive_timer(sk, tmo);
} else {
......
}
but as far as I understand,
1)if sk is dead, there would be no process context reference to it.
2)if sk->lock.users==1, it means that some process context is
referencing it, and most likely it is in backlog processing when
doing release_sock().
1) and 2) seems to be a bit conflict to me,
Would you please tell me what is the important points that I have missed?
Thank you for your kind help in advance.
http://goto.sohu.com/goto.php3?code=libo-gz096mailzhujiao ???????????????
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
