rds: rds_cong_queue_updates needs to defer the congestion update transmission

[uclinux-h8/linux.git] / net / rds / cong.c

diff --git a/net/rds/cong.c b/net/rds/cong.c
index e5b65ac..e6144b8 100644 (file)
--- a/net/rds/cong.c
+++ b/net/rds/cong.c
@@ -221,7 +221,21 @@ void rds_cong_queue_updates(struct rds_cong_map *map)
        list_for_each_entry(conn, &map->m_conn_list, c_map_item) {
                if (!test_and_set_bit(0, &conn->c_map_queued)) {
                        rds_stats_inc(s_cong_update_queued);
-                       rds_send_xmit(conn);
+                       /* We cannot inline the call to rds_send_xmit() here
+                        * for two reasons (both pertaining to a TCP transport):
+                        * 1. When we get here from the receive path, we
+                        *    are already holding the sock_lock (held by
+                        *    tcp_v4_rcv()). So inlining calls to
+                        *    tcp_setsockopt and/or tcp_sendmsg will deadlock
+                        *    when it tries to get the sock_lock())
+                        * 2. Interrupts are masked so that we can mark the
+                        *    the port congested from both send and recv paths.
+                        *    (See comment around declaration of rdc_cong_lock).
+                        *    An attempt to get the sock_lock() here will
+                        *    therefore trigger warnings.
+                        * Defer the xmit to rds_send_worker() instead.
+                        */
+                       queue_delayed_work(rds_wq, &conn->c_send_w, 0);
                }
        }