tipc: increase timeout in tipc_sk_enqueue()

commit f4bb62e upstream.

In tipc_sk_enqueue() we use hardcoded 2 jiffies to extract
socket buffer from generic queue to particular socket.
The 2 jiffies is too short in case there are other high priority
tasks get CPU cycles for multiple jiffies update. As result, no
buffer could be enqueued to particular socket.

To solve this, we switch to use constant timeout 20msecs.
Then, the function will be expired between 2 jiffies (CONFIG_100HZ)
and 20 jiffies (CONFIG_1000HZ).

Fixes: c637c10 ("tipc: resolve race problem at unicast message reception")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Author: Hoang Le

net/tipc/socket.c

@@ -2422,7 +2422,7 @@ static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
 static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk,
 			    u32 dport, struct sk_buff_head *xmitq)
 {
-	unsigned long time_limit = jiffies + 2;
+	unsigned long time_limit = jiffies + usecs_to_jiffies(20000);
 	struct sk_buff *skb;
 	unsigned int lim;
 	atomic_t *dcnt;