[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20181219044219.21209-1-hoang.h.le@dektech.com.au>
Date: Wed, 19 Dec 2018 11:42:19 +0700
From: Hoang Le <hoang.h.le@...tech.com.au>
To: tipc-discussion@...ts.sourceforge.net, jon.maloy@...csson.com,
maloy@...jonn.com, ying.xue@...driver.com, netdev@...r.kernel.org
Subject: [net-next] tipc: fix uninitialized value for broadcast retransmission
When sending broadcast message on high load system, there are a lot of
unnecessary packets restranmission. That issue was caused by missing in
initial criteria for retransmission.
To prevent this happen, just initialize this criteria for retransmission
in next 10 milliseconds.
Fixes: 31c4f4cc32f7 tipc: improve broadcast retransmission algorithm
Acked-by: Ying Xue <ying.xue@...driver.com>
Acked-by: Jon Maloy <jon.maloy@...csson.com>
Signed-off-by: Hoang Le <hoang.h.le@...tech.com.au>
---
net/tipc/link.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/net/tipc/link.c b/net/tipc/link.c
index 9e265eb89726..d6a4b6ae5504 100644
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -945,6 +945,10 @@ int tipc_link_xmit(struct tipc_link *l, struct sk_buff_head *list,
}
__skb_dequeue(list);
__skb_queue_tail(transmq, skb);
+ /* next retransmit attempt */
+ if (link_is_bc_sndlink(l))
+ TIPC_SKB_CB(skb)->nxt_retr =
+ jiffies + TIPC_BC_RETR_LIM;
__skb_queue_tail(xmitq, _skb);
TIPC_SKB_CB(skb)->ackers = l->ackers;
l->rcv_unacked = 0;
@@ -992,6 +996,10 @@ static void tipc_link_advance_backlog(struct tipc_link *l,
hdr = buf_msg(skb);
l->backlog[msg_importance(hdr)].len--;
__skb_queue_tail(&l->transmq, skb);
+ /* next retransmit attempt */
+ if (link_is_bc_sndlink(l))
+ TIPC_SKB_CB(skb)->nxt_retr = jiffies + TIPC_BC_RETR_LIM;
+
__skb_queue_tail(xmitq, _skb);
TIPC_SKB_CB(skb)->ackers = l->ackers;
msg_set_seqno(hdr, seqno);
--
2.17.1
Powered by blists - more mailing lists