rxrpc: Calls shouldn't hold socket refs

rxrpc calls shouldn't hold refs on the sock struct.  This was done so that
the socket wouldn't go away whilst the call was in progress, such that the
call could reach the socket's queues.

However, we can mark the socket as requiring an RCU release and rely on the
RCU read lock.

To make this work, we do:

 (1) rxrpc_release_call() removes the call's call user ID.  This is now
     only called from socket operations and not from the call processor:

	rxrpc_accept_call() / rxrpc_kernel_accept_call()
	rxrpc_reject_call() / rxrpc_kernel_reject_call()
	rxrpc_kernel_end_call()
	rxrpc_release_calls_on_socket()
	rxrpc_recvmsg()

     Though it is also called in the cleanup path of
     rxrpc_accept_incoming_call() before we assign a user ID.

 (2) Pass the socket pointer into rxrpc_release_call() rather than getting
     it from the call so that we can get rid of uninitialised calls.

 (3) Fix call processor queueing to pass a ref to the work queue and to
     release that ref at the end of the processor function (or to pass it
     back to the work queue if we have to requeue).

 (4) Skip out of the call processor function asap if the call is complete
     and don't requeue it if the call is complete.

 (5) Clean up the call immediately that the refcount reaches 0 rather than
     trying to defer it.  Actual deallocation is deferred to RCU, however.

 (6) Don't hold socket refs for allocated calls.

 (7) Use the RCU read lock when queueing a message on a socket and treat
     the call's socket pointer according to RCU rules and check it for
     NULL.

     We also need to use the RCU read lock when viewing a call through
     procfs.

 (8) Transmit the final ACK/ABORT to a client call in rxrpc_release_call()
     if this hasn't been done yet so that we can then disconnect the call.
     Once the call is disconnected, it won't have any access to the
     connection struct and the UDP socket for the call work processor to be
     able to send the ACK.  Terminal retransmission will be handled by the
     connection processor.

 (9) Release all calls immediately on the closing of a socket rather than
     trying to defer this.  Incomplete calls will be aborted.

The call refcount model is much simplified.  Refs are held on the call by:

 (1) A socket's user ID tree.

 (2) A socket's incoming call secureq and acceptq.

 (3) A kernel service that has a call in progress.

 (4) A queued call work processor.  We have to take care to put any call
     that we failed to queue.

 (5) sk_buffs on a socket's receive queue.  A future patch will get rid of
     this.

Whilst we're at it, we can do:

 (1) Get rid of the RXRPC_CALL_EV_RELEASE event.  Release is now done
     entirely from the socket routines and never from the call's processor.

 (2) Get rid of the RXRPC_CALL_DEAD state.  Calls now end in the
     RXRPC_CALL_COMPLETE state.

 (3) Get rid of the rxrpc_call::destroyer work item.  Calls are now torn
     down when their refcount reaches 0 and then handed over to RCU for
     final cleanup.

 (4) Get rid of the rxrpc_call::deadspan timer.  Calls are cleaned up
     immediately they're finished with and don't hang around.
     Post-completion retransmission is handled by the connection processor
     once the call is disconnected.

 (5) Get rid of the dead call expiry setting as there's no longer a timer
     to set.

 (6) rxrpc_destroy_all_calls() can just check that the call list is empty.

Signed-off-by: David Howells <dhowells@redhat.com>

1 files changed, 30 insertions, 44 deletions

diff --git a/net/rxrpc/call_event.c b/net/rxrpc/call_event.c
index fee8b6d..8365d33 100644
--- a/net/rxrpc/call_event.c
+++ b/net/rxrpc/call_event.c
@@ -811,8 +811,9 @@ static int rxrpc_post_message(struct rxrpc_call *call, u32 mark, u32 error,
 }
 
 /*
- * handle background processing of incoming call packets and ACK / abort
- * generation
+ * Handle background processing of incoming call packets and ACK / abort
+ * generation.  A ref on the call is donated to us by whoever queued the work
+ * item.
  */
 void rxrpc_process_call(struct work_struct *work)
 {
@@ -827,6 +828,7 @@ void rxrpc_process_call(struct work_struct *work)
 	unsigned long bits;
 	__be32 data, pad;
 	size_t len;
+	bool requeue = false;
 	int loop, nbit, ioc, ret, mtu;
 	u32 serial, abort_code = RX_PROTOCOL_ERROR;
 	u8 *acks = NULL;
@@ -838,6 +840,11 @@ void rxrpc_process_call(struct work_struct *work)
 	       call->debug_id, rxrpc_call_states[call->state], call->events,
 	       (jiffies - call->creation_jif) / (HZ / 10));
 
+	if (call->state >= RXRPC_CALL_COMPLETE) {
+		rxrpc_put_call(call, rxrpc_call_put);
+		return;
+	}
+
 	if (!call->conn)
 		goto skip_msg_init;
 
@@ -1088,16 +1095,21 @@ skip_msg_init:
 		spin_lock_bh(&call->lock);
 
 		if (call->state == RXRPC_CALL_SERVER_SECURING) {
+			struct rxrpc_sock *rx;
 			_debug("securing");
-			write_lock(&call->socket->call_lock);
-			if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
-			    !test_bit(RXRPC_CALL_EV_RELEASE, &call->events)) {
-				_debug("not released");
-				call->state = RXRPC_CALL_SERVER_ACCEPTING;
-				list_move_tail(&call->accept_link,
-					       &call->socket->acceptq);
+			rcu_read_lock();
+			rx = rcu_dereference(call->socket);
+			if (rx) {
+				write_lock(&rx->call_lock);
+				if (!test_bit(RXRPC_CALL_RELEASED, &call->flags)) {
+					_debug("not released");
+					call->state = RXRPC_CALL_SERVER_ACCEPTING;
+					list_move_tail(&call->accept_link,
+						       &rx->acceptq);
+				}
+				write_unlock(&rx->call_lock);
 			}
-			write_unlock(&call->socket->call_lock);
+			rcu_read_unlock();
 			read_lock(&call->state_lock);
 			if (call->state < RXRPC_CALL_COMPLETE)
 				set_bit(RXRPC_CALL_EV_POST_ACCEPT, &call->events);
@@ -1139,11 +1151,6 @@ skip_msg_init:
 		goto maybe_reschedule;
 	}
 
-	if (test_bit(RXRPC_CALL_EV_RELEASE, &call->events)) {
-		rxrpc_release_call(call);
-		clear_bit(RXRPC_CALL_EV_RELEASE, &call->events);
-	}
-
 	/* other events may have been raised since we started checking */
 	goto maybe_reschedule;
 
@@ -1209,10 +1216,8 @@ send_message_2:
 			     &msg, iov, ioc, len);
 	if (ret < 0) {
 		_debug("sendmsg failed: %d", ret);
-		read_lock_bh(&call->state_lock);
-		if (call->state < RXRPC_CALL_DEAD)
-			rxrpc_queue_call(call);
-		read_unlock_bh(&call->state_lock);
+		if (call->state < RXRPC_CALL_COMPLETE)
+			requeue = true;
 		goto error;
 	}
 
@@ -1245,41 +1250,22 @@ send_message_2:
 
 kill_ACKs:
 	del_timer_sync(&call->ack_timer);
-	if (test_and_clear_bit(RXRPC_CALL_EV_ACK_FINAL, &call->events))
-		rxrpc_put_call(call, rxrpc_call_put);
 	clear_bit(RXRPC_CALL_EV_ACK, &call->events);
 
 maybe_reschedule:
 	if (call->events || !skb_queue_empty(&call->rx_queue)) {
-		read_lock_bh(&call->state_lock);
-		if (call->state < RXRPC_CALL_DEAD)
-			rxrpc_queue_call(call);
-		read_unlock_bh(&call->state_lock);
-	}
-
-	/* don't leave aborted connections on the accept queue */
-	if (call->state >= RXRPC_CALL_COMPLETE &&
-	    !list_empty(&call->accept_link)) {
-		_debug("X unlinking once-pending call %p { e=%lx f=%lx c=%x }",
-		       call, call->events, call->flags, call->conn->proto.cid);
-
-		read_lock_bh(&call->state_lock);
-		if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
-		    !test_and_set_bit(RXRPC_CALL_EV_RELEASE, &call->events))
-			rxrpc_queue_call(call);
-		read_unlock_bh(&call->state_lock);
+		if (call->state < RXRPC_CALL_COMPLETE)
+			requeue = true;
 	}
 
 error:
 	kfree(acks);
 
-	/* because we don't want two CPUs both processing the work item for one
-	 * call at the same time, we use a flag to note when it's busy; however
-	 * this means there's a race between clearing the flag and setting the
-	 * work pending bit and the work item being processed again */
-	if (call->events && !work_pending(&call->processor)) {
+	if ((requeue || call->events) && !work_pending(&call->processor)) {
 		_debug("jumpstart %x", call->conn->proto.cid);
-		rxrpc_queue_call(call);
+		__rxrpc_queue_call(call);
+	} else {
+		rxrpc_put_call(call, rxrpc_call_put);
 	}
 
 	_leave("");