rxrpc: Fix locking issue

There's a locking issue with the per-netns list of calls in rxrpc.  The
pieces of code that add and remove a call from the list use write_lock()
and the calls procfile uses read_lock() to access it.  However, the timer
callback function may trigger a removal by trying to queue a call for
processing and finding that it's already queued - at which point it has a
spare refcount that it has to do something with.  Unfortunately, if it puts
the call and this reduces the refcount to 0, the call will be removed from
the list.  Unfortunately, since the _bh variants of the locking functions
aren't used, this can deadlock.

================================
WARNING: inconsistent lock state
5.18.0-rc3-build4+ #10 Not tainted
--------------------------------
inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
ksoftirqd/2/25 [HC0[0]:SC1[1]:HE1:SE0] takes:
ffff888107ac4038 (&rxnet->call_lock){+.?.}-{2:2}, at: rxrpc_put_call+0x103/0x14b
{SOFTIRQ-ON-W} state was registered at:
...
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(&rxnet->call_lock);
  <Interrupt>
    lock(&rxnet->call_lock);

 *** DEADLOCK ***

1 lock held by ksoftirqd/2/25:
 #0: ffff8881008ffdb0 ((&call->timer)){+.-.}-{0:0}, at: call_timer_fn+0x5/0x23d

Changes
=======
ver #2)
 - Changed to using list_next_rcu() rather than rcu_dereference() directly.

Fixes: 17926a79320a ("[AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 16 insertions, 22 deletions

diff --git a/net/rxrpc/ar-internal.h b/net/rxrpc/ar-internal.h
index 52a23d03d694..dcc0ec0bf3de 100644
--- a/net/rxrpc/ar-internal.h
+++ b/net/rxrpc/ar-internal.h
@@ -60,7 +60,7 @@ struct rxrpc_net {
 	struct proc_dir_entry	*proc_net;	/* Subdir in /proc/net */
 	u32			epoch;		/* Local epoch for detecting local-end reset */
 	struct list_head	calls;		/* List of calls active in this namespace */
-	rwlock_t		call_lock;	/* Lock for ->calls */
+	spinlock_t		call_lock;	/* Lock for ->calls */
 	atomic_t		nr_calls;	/* Count of allocated calls */
 
 	atomic_t		nr_conns;
diff --git a/net/rxrpc/call_accept.c b/net/rxrpc/call_accept.c
index 8ae59edc2551..99e10eea3732 100644
--- a/net/rxrpc/call_accept.c
+++ b/net/rxrpc/call_accept.c
@@ -140,9 +140,9 @@ static int rxrpc_service_prealloc_one(struct rxrpc_sock *rx,
 	write_unlock(&rx->call_lock);
 
 	rxnet = call->rxnet;
-	write_lock(&rxnet->call_lock);
-	list_add_tail(&call->link, &rxnet->calls);
-	write_unlock(&rxnet->call_lock);
+	spin_lock_bh(&rxnet->call_lock);
+	list_add_tail_rcu(&call->link, &rxnet->calls);
+	spin_unlock_bh(&rxnet->call_lock);
 
 	b->call_backlog[call_head] = call;
 	smp_store_release(&b->call_backlog_head, (call_head + 1) & (size - 1));
diff --git a/net/rxrpc/call_object.c b/net/rxrpc/call_object.c
index 8764a4f19c03..84d0a4109645 100644
--- a/net/rxrpc/call_object.c
+++ b/net/rxrpc/call_object.c
@@ -337,9 +337,9 @@ struct rxrpc_call *rxrpc_new_client_call(struct rxrpc_sock *rx,
 	write_unlock(&rx->call_lock);
 
 	rxnet = call->rxnet;
-	write_lock(&rxnet->call_lock);
-	list_add_tail(&call->link, &rxnet->calls);
-	write_unlock(&rxnet->call_lock);
+	spin_lock_bh(&rxnet->call_lock);
+	list_add_tail_rcu(&call->link, &rxnet->calls);
+	spin_unlock_bh(&rxnet->call_lock);
 
 	/* From this point on, the call is protected by its own lock. */
 	release_sock(&rx->sk);
@@ -633,9 +633,9 @@ void rxrpc_put_call(struct rxrpc_call *call, enum rxrpc_call_trace op)
 		ASSERTCMP(call->state, ==, RXRPC_CALL_COMPLETE);
 
 		if (!list_empty(&call->link)) {
-			write_lock(&rxnet->call_lock);
+			spin_lock_bh(&rxnet->call_lock);
 			list_del_init(&call->link);
-			write_unlock(&rxnet->call_lock);
+			spin_unlock_bh(&rxnet->call_lock);
 		}
 
 		rxrpc_cleanup_call(call);
@@ -707,7 +707,7 @@ void rxrpc_destroy_all_calls(struct rxrpc_net *rxnet)
 	_enter("");
 
 	if (!list_empty(&rxnet->calls)) {
-		write_lock(&rxnet->call_lock);
+		spin_lock_bh(&rxnet->call_lock);
 
 		while (!list_empty(&rxnet->calls)) {
 			call = list_entry(rxnet->calls.next,
@@ -722,12 +722,12 @@ void rxrpc_destroy_all_calls(struct rxrpc_net *rxnet)
 			       rxrpc_call_states[call->state],
 			       call->flags, call->events);
 
-			write_unlock(&rxnet->call_lock);
+			spin_unlock_bh(&rxnet->call_lock);
 			cond_resched();
-			write_lock(&rxnet->call_lock);
+			spin_lock_bh(&rxnet->call_lock);
 		}
 
-		write_unlock(&rxnet->call_lock);
+		spin_unlock_bh(&rxnet->call_lock);
 	}
 
 	atomic_dec(&rxnet->nr_calls);
diff --git a/net/rxrpc/net_ns.c b/net/rxrpc/net_ns.c
index 34f389975a7d..bb4c25d6df64 100644
--- a/net/rxrpc/net_ns.c
+++ b/net/rxrpc/net_ns.c
@@ -50,7 +50,7 @@ static __net_init int rxrpc_init_net(struct net *net)
 	rxnet->epoch |= RXRPC_RANDOM_EPOCH;
 
 	INIT_LIST_HEAD(&rxnet->calls);
-	rwlock_init(&rxnet->call_lock);
+	spin_lock_init(&rxnet->call_lock);
 	atomic_set(&rxnet->nr_calls, 1);
 
 	atomic_set(&rxnet->nr_conns, 1);
diff --git a/net/rxrpc/proc.c b/net/rxrpc/proc.c
index 8967201fd8e5..245418943e01 100644
--- a/net/rxrpc/proc.c
+++ b/net/rxrpc/proc.c
@@ -26,29 +26,23 @@ static const char *const rxrpc_conn_states[RXRPC_CONN__NR_STATES] = {
  */
 static void *rxrpc_call_seq_start(struct seq_file *seq, loff_t *_pos)
 	__acquires(rcu)
-	__acquires(rxnet->call_lock)
 {
 	struct rxrpc_net *rxnet = rxrpc_net(seq_file_net(seq));
 
 	rcu_read_lock();
-	read_lock(&rxnet->call_lock);
-	return seq_list_start_head(&rxnet->calls, *_pos);
+	return seq_list_start_head_rcu(&rxnet->calls, *_pos);
 }
 
 static void *rxrpc_call_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 {
 	struct rxrpc_net *rxnet = rxrpc_net(seq_file_net(seq));
 
-	return seq_list_next(v, &rxnet->calls, pos);
+	return seq_list_next_rcu(v, &rxnet->calls, pos);
 }
 
 static void rxrpc_call_seq_stop(struct seq_file *seq, void *v)
-	__releases(rxnet->call_lock)
 	__releases(rcu)
 {
-	struct rxrpc_net *rxnet = rxrpc_net(seq_file_net(seq));
-
-	read_unlock(&rxnet->call_lock);
 	rcu_read_unlock();
 }