The addition of two dummy fields is a temporary measure to
satisfy the logic in o2net_check_handshake() and will be
rectified in a future version of this patch
Signed-off-by: Andrew Beekhof <[EMAIL PROTECTED]>
---
fs/ocfs2/cluster/tcp.c | 44 ++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/cluster/tcp.h | 2 +
fs/ocfs2/cluster/tcp_internal.h | 6 ++++-
3 files changed, 51 insertions(+), 1 deletion(-)
Index: fs/ocfs2/cluster/tcp.c
===================================================================
--- fs/ocfs2/cluster/tcp.c.orig 2006-11-17 16:26:15.000000000 +0100
+++ fs/ocfs2/cluster/tcp.c 2006-11-17 16:26:27.000000000 +0100
@@ -1121,6 +1121,33 @@ static int o2net_check_handshake(struct
return -1;
}
+ /*
+ * Ensure timeouts are consistent with other nodes, otherwise
+ * we can end up with one node thinking that the other must be down,
+ * but isn't. This can ultimately cause corruption.
+ */
+ if (be32_to_cpu(hand->o2net_idle_timeout_ms) !=
+ o2net_idle_timeout(sc->sc_node)) {
+ mlog(ML_NOTICE, SC_NODEF_FMT " uses network idle timeout of "
+ "%u ms, but we use %u ms locally. disconnecting\n",
+ SC_NODEF_ARGS(sc),
+ be32_to_cpu(hand->o2net_idle_timeout_ms),
+ o2net_idle_timeout(sc->sc_node));
+ o2net_ensure_shutdown(nn, sc, -ENOTCONN);
+ return -1;
+ }
+
+ if (be32_to_cpu(hand->o2net_keepalive_delay_ms) !=
+ o2net_keepalive_delay(sc->sc_node)) {
+ mlog(ML_NOTICE, SC_NODEF_FMT " uses keepalive delay of "
+ "%u ms, but we use %u ms locally. disconnecting\n",
+ SC_NODEF_ARGS(sc),
+ be32_to_cpu(hand->o2net_keepalive_delay_ms),
+ o2net_keepalive_delay(sc->sc_node));
+ o2net_ensure_shutdown(nn, sc, -ENOTCONN);
+ return -1;
+ }
+
sc->sc_handshake_ok = 1;
spin_lock(&nn->nn_lock);
@@ -1269,6 +1296,21 @@ static int o2net_set_nodelay(struct sock
return ret;
}
+static void o2net_initialize_handshake(void)
+{
+ static int initialized = 0;
+ if(initialized)
+ return;
+
+ initialized = 1;
+ o2net_hand->o2net_idle_timeout_ms = cpu_to_be32(
+ o2net_idle_timeout(NULL));
+ o2net_hand->o2net_keepalive_delay_ms = cpu_to_be32(
+ o2net_keepalive_delay(NULL));
+ o2net_hand->o2net_reconnect_delay_ms = cpu_to_be32(
+ o2net_reconnect_delay(NULL));
+}
+
/* ------------------------------------------------------------ */
/* called when a connect completes and after a sock is accepted. the
@@ -1281,6 +1323,7 @@ static void o2net_sc_connect_completed(v
(unsigned long long)O2NET_PROTOCOL_VERSION,
(unsigned long long)be64_to_cpu(o2net_hand->connector_id));
+ o2net_initialize_handshake();
o2net_sendpage(sc, o2net_hand, sizeof(*o2net_hand));
sc_put(sc);
}
@@ -1668,6 +1711,7 @@ static int o2net_accept_one(struct socke
o2net_register_callbacks(sc->sc_sock->sk, sc);
o2net_sc_queue_work(sc, &sc->sc_rx_work);
+ o2net_initialize_handshake();
o2net_sendpage(sc, o2net_hand, sizeof(*o2net_hand));
out:
Index: fs/ocfs2/cluster/tcp.h
===================================================================
--- fs/ocfs2/cluster/tcp.h.orig 2006-11-17 16:26:15.000000000 +0100
+++ fs/ocfs2/cluster/tcp.h 2006-11-17 16:26:27.000000000 +0100
@@ -47,6 +47,8 @@ struct o2net_msg
__be32 status;
__be32 key;
__be32 msg_num;
+ __be64 dummy;
+ __be32 dummy2;
__u8 buf[0];
};
Index: fs/ocfs2/cluster/tcp_internal.h
===================================================================
--- fs/ocfs2/cluster/tcp_internal.h.orig 2006-11-17 16:26:15.000000000
+0100
+++ fs/ocfs2/cluster/tcp_internal.h 2006-11-17 16:26:54.000000000 +0100
@@ -48,10 +48,14 @@
* - full 64 bit i_size in the metadata lock lvbs
* - introduction of "rw" lock and pushing meta/data locking down
*/
-#define O2NET_PROTOCOL_VERSION 4ULL
+#define O2NET_PROTOCOL_VERSION 5ULL
struct o2net_handshake {
__be64 protocol_version;
__be64 connector_id;
+ __be32 o2hb_heartbeat_timeout_ms;
+ __be32 o2net_idle_timeout_ms;
+ __be32 o2net_keepalive_delay_ms;
+ __be32 o2net_reconnect_delay_ms;
};
struct o2net_node {
--
_______________________________________________
Ocfs2-devel mailing list
[email protected]
http://oss.oracle.com/mailman/listinfo/ocfs2-devel
