netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Paul Gortmaker <paul.gortmaker@windriver.com>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, allan.stephens@windriver.com,
	ying.xue@windriver.com,
	Paul Gortmaker <paul.gortmaker@windriver.com>
Subject: [PATCH net-next 13/15] tipc: Prevent loss of fragmented messages over broadcast link
Date: Mon,  6 Feb 2012 19:52:46 -0500	[thread overview]
Message-ID: <1328575968-20643-14-git-send-email-paul.gortmaker@windriver.com> (raw)
In-Reply-To: <1328575968-20643-1-git-send-email-paul.gortmaker@windriver.com>

From: Allan Stephens <allan.stephens@windriver.com>

Modifies broadcast link so that an incoming fragmented message is not
lost if reassembly cannot begin because there currently is no buffer
big enough to hold the entire reassembled message. The broadcast link
now ignores the first fragment completely, which causes the sending node
to retransmit the first fragment so that reassembly can be re-attempted.

Previously, the sender would have had no reason to retransmit the 1st
fragment, so we would never have a chance to re-try the allocation.

To do this cleanly without duplicaton, a new bclink_accept_pkt()
function is introduced.

Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 net/tipc/bcast.c |   64 +++++++++++++++++++++++++++++++++++------------------
 1 files changed, 42 insertions(+), 22 deletions(-)

diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c
index a9b7132..41ecf31 100644
--- a/net/tipc/bcast.c
+++ b/net/tipc/bcast.c
@@ -389,7 +389,33 @@ exit:
 	return res;
 }
 
-/**
+/*
+ * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet
+ *
+ * Called with both sending node's lock and bc_lock taken.
+ */
+
+static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
+{
+	bclink_update_last_sent(node, seqno);
+	node->bclink.last_in = seqno;
+	node->bclink.oos_state = 0;
+	bcl->stats.recv_info++;
+
+	/*
+	 * Unicast an ACK periodically, ensuring that
+	 * all nodes in the cluster don't ACK at the same time
+	 */
+
+	if (((seqno - tipc_own_addr) % TIPC_MIN_LINK_WIN) == 0) {
+		tipc_link_send_proto_msg(
+			node->active_links[node->addr & 1],
+			STATE_MSG, 0, 0, 0, 0, 0);
+		bcl->stats.sent_acks++;
+	}
+}
+
+/*
  * tipc_bclink_recv_pkt - receive a broadcast packet, and deliver upwards
  *
  * tipc_net_lock is read_locked, no other locks set
@@ -443,29 +469,12 @@ void tipc_bclink_recv_pkt(struct sk_buff *buf)
 	next_in = mod(node->bclink.last_in + 1);
 
 	if (likely(seqno == next_in)) {
-		bclink_update_last_sent(node, seqno);
 receive:
-		node->bclink.last_in = seqno;
-		node->bclink.oos_state = 0;
-
-		spin_lock_bh(&bc_lock);
-		bcl->stats.recv_info++;
-
-		/*
-		 * Unicast an ACK periodically, ensuring that
-		 * all nodes in the cluster don't ACK at the same time
-		 */
-
-		if (((seqno - tipc_own_addr) % TIPC_MIN_LINK_WIN) == 0) {
-			tipc_link_send_proto_msg(
-				node->active_links[node->addr & 1],
-				STATE_MSG, 0, 0, 0, 0, 0);
-			bcl->stats.sent_acks++;
-		}
-
 		/* Deliver message to destination */
 
 		if (likely(msg_isdata(msg))) {
+			spin_lock_bh(&bc_lock);
+			bclink_accept_pkt(node, seqno);
 			spin_unlock_bh(&bc_lock);
 			tipc_node_unlock(node);
 			if (likely(msg_mcast(msg)))
@@ -473,24 +482,35 @@ receive:
 			else
 				buf_discard(buf);
 		} else if (msg_user(msg) == MSG_BUNDLER) {
+			spin_lock_bh(&bc_lock);
+			bclink_accept_pkt(node, seqno);
 			bcl->stats.recv_bundles++;
 			bcl->stats.recv_bundled += msg_msgcnt(msg);
 			spin_unlock_bh(&bc_lock);
 			tipc_node_unlock(node);
 			tipc_link_recv_bundle(buf);
 		} else if (msg_user(msg) == MSG_FRAGMENTER) {
+			int ret = tipc_link_recv_fragment(&node->bclink.defragm,
+						      &buf, &msg);
+			if (ret < 0)
+				goto unlock;
+			spin_lock_bh(&bc_lock);
+			bclink_accept_pkt(node, seqno);
 			bcl->stats.recv_fragments++;
-			if (tipc_link_recv_fragment(&node->bclink.defragm,
-						    &buf, &msg))
+			if (ret > 0)
 				bcl->stats.recv_fragmented++;
 			spin_unlock_bh(&bc_lock);
 			tipc_node_unlock(node);
 			tipc_net_route_msg(buf);
 		} else if (msg_user(msg) == NAME_DISTRIBUTOR) {
+			spin_lock_bh(&bc_lock);
+			bclink_accept_pkt(node, seqno);
 			spin_unlock_bh(&bc_lock);
 			tipc_node_unlock(node);
 			tipc_named_recv(buf);
 		} else {
+			spin_lock_bh(&bc_lock);
+			bclink_accept_pkt(node, seqno);
 			spin_unlock_bh(&bc_lock);
 			tipc_node_unlock(node);
 			buf_discard(buf);
-- 
1.7.9

  parent reply	other threads:[~2012-02-07  0:53 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-07  0:52 [PATCH net-next 00/15] TIPC updates for upcoming v3.4 Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 01/15] tipc: improve the link deferred queue insertion algorithm Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 02/15] tipc: Prevent transmission of outdated link protocol messages Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 03/15] tipc: Prevent broadcast link stalling in dual LAN environments Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 04/15] tipc: Ensure broadcast link re-acquires node after link failure Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 05/15] tipc: Fix problem with broadcast link synchronization between nodes Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 06/15] tipc: Add missing broadcast link lock when sending NACK Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 07/15] tipc: Fix node lock reclamation issues in broadcast link reception Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 08/15] tipc: Fix bug in broadcast link duplicate message statistics Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 09/15] tipc: Add missing locks in broadcast link statistics accumulation Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 10/15] tipc: Major redesign of broadcast link ACK/NACK algorithms Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 11/15] tipc: Remove obsolete broadcast tag capability Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 12/15] tipc: Prevent loss of fragmented messages over unicast links Paul Gortmaker
2012-02-07  0:52 ` Paul Gortmaker [this message]
2012-02-07  0:52 ` [PATCH net-next 14/15] tipc: Eliminate alteration of publication key during name table purging Paul Gortmaker
2012-02-07  0:52 ` [PATCH net-next 15/15] tipc: Minor optimization to rejection of connection-based messages Paul Gortmaker
2012-02-07 17:33 ` [PATCH net-next 00/15] TIPC updates for upcoming v3.4 David Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1328575968-20643-14-git-send-email-paul.gortmaker@windriver.com \
    --to=paul.gortmaker@windriver.com \
    --cc=allan.stephens@windriver.com \
    --cc=davem@davemloft.net \
    --cc=netdev@vger.kernel.org \
    --cc=ying.xue@windriver.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).