michaelh/zephyr
1
0
Fork 0
Code Releases Activity

net: tcp: Set timer for last ACK

Browse source 
If we do not receive last ACK when the connection is tore down,
then do not wait forever as that would eat all the resources in
the network stack. So when we enter the LAST_ACK state, we setup
a timer that will unref the connection if the last ACK is not
received.

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit is contained in:
Jukka Rissanen 2017-05-17 16:15:47 +03:00
commit 27b5540118
1 changed files with 22 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

26
subsys/net/ip/net_context.c
View file

@ -36,6 +36,8 @@
#define NET_MAX_CONTEXT CONFIG_NET_MAX_CONTEXTS
#define ACK_TIMEOUT MSEC_PER_SEC
/* Declares a wrapper function for a net_conn callback that refs the
* context around the invocation (to protect it from premature
* deletion). Long term would be nice to see this feature be part of
@ -884,6 +886,14 @@ NET_CONN_CB(tcp_established)
context->recv_cb(context, NULL, 0,
context->tcp->recv_user_data);
}
/* We should receive ACK next in order to get rid of LAST_ACK
* state that we are entering in a short while. But we need to
* be prepared to NOT to receive it as otherwise the connection
* would be stuck forever.
*/
k_delayed_work_submit(&context->tcp->ack_timer, ACK_TIMEOUT);
context->tcp->ack_timer_cancelled = false;
}
send_ack(context, &conn->remote_addr, false);
@ -1219,8 +1229,6 @@ int net_context_connect(struct net_context *context,
#if defined(CONFIG_NET_TCP)
#define ACK_TIMEOUT MSEC_PER_SEC
static void ack_timer_cancel(struct net_tcp *tcp)
{
tcp->ack_timer_cancelled = true;
@ -1238,9 +1246,19 @@ static void ack_timeout(struct k_work *work)
NET_DBG("Did not receive ACK in %dms", ACK_TIMEOUT);
send_reset(tcp->context, &tcp->context->remote);
if (net_tcp_get_state(tcp) == NET_TCP_LAST_ACK) {
/* We did not receive the last ACK on time. We can only
* close the connection at time point. We also do not send
* anything in this last state but will go to to CLOSED state.
*/
net_tcp_change_state(tcp, NET_TCP_CLOSED);
net_tcp_change_state(tcp, NET_TCP_LISTEN);
net_context_unref(tcp->context);
} else {
send_reset(tcp->context, &tcp->context->remote);
net_tcp_change_state(tcp, NET_TCP_LISTEN);
}
}
static void pkt_get_sockaddr(sa_family_t family, struct net_pkt *pkt,