Merge branch 'rds-use-RCU-between-work-enqueue-and-connection-teardown'

Sowmini Varadhan says:

====================
rds: use RCU between work-enqueue and connection teardown

This patchset follows up on the root-cause mentioned in
https://www.spinics.net/lists/netdev/msg472849.html

Patch1 implements some code refactoring that was suggeseted
as an enhancement in http://patchwork.ozlabs.org/patch/843157/
It replaces the c_destroy_in_prog bit in rds_connection with
an atomically managed flag in rds_conn_path.

Patch2 builds on Patch1 and uses RCU to make sure that
work is only enqueued if the connection destroy is not already
in progress: the test-flag-and-enqueue is done under rcu_read_lock,
while destroy first sets the flag, uses synchronize_rcu to
wait for existing reader threads to complete, and then starts
all the work-cancellation.

Since I have not been able to reproduce the original stack traces
reported by syszbot, and these are fixes for a race condition that
are based on code-inspection I am not marking these as reported-by
at this time.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 6 insertions, 2 deletions

diff --git a/net/rds/tcp_recv.c b/net/rds/tcp_recv.c
index e006ef8e6d40..dd707b9e73e5 100644
--- a/net/rds/tcp_recv.c
+++ b/net/rds/tcp_recv.c
@@ -321,8 +321,12 @@ void rds_tcp_data_ready(struct sock *sk)
 	ready = tc->t_orig_data_ready;
 	rds_tcp_stats_inc(s_tcp_data_ready_calls);
 
-	if (rds_tcp_read_sock(cp, GFP_ATOMIC) == -ENOMEM)
-		queue_delayed_work(rds_wq, &cp->cp_recv_w, 0);
+	if (rds_tcp_read_sock(cp, GFP_ATOMIC) == -ENOMEM) {
+		rcu_read_lock();
+		if (!test_bit(RDS_DESTROY_PENDING, &cp->cp_flags))
+			queue_delayed_work(rds_wq, &cp->cp_recv_w, 0);
+		rcu_read_unlock();
+	}
 out:
 	read_unlock_bh(&sk->sk_callback_lock);
 	ready(sk);