rxrpc: Fix lockup due to no error backoff after ack transmit error

If the network becomes (partially) unavailable, say by disabling IPv6, the
background ACK transmission routine can get itself into a tizzy by
proposing immediate ACK retransmission.  Since we're in the call event
processor, that happens immediately without returning to the workqueue
manager.

The condition should clear after a while when either the network comes back
or the call times out.

Fix this by:

 (1) When re-proposing an ACK on failed Tx, don't schedule it immediately.
     This will allow a certain amount of time to elapse before we try
     again.

 (2) Enforce a return to the workqueue manager after a certain number of
     iterations of the call processing loop.

 (3) Add a backoff delay that increases the delay on deferred ACKs by a
     jiffy per failed transmission to a limit of HZ.  The backoff delay is
     cleared on a successful return from kernel_sendmsg().

 (4) Cancel calls immediately if the opening sendmsg fails.  The layer
     above can arrange retransmission or rotate to another server.

Fixes: 248f219cb8bc ("rxrpc: Rewrite the data and ack handling code")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/rxrpc/call_event.c b/net/rxrpc/call_event.c
index 8e7434e..468efc3 100644
--- a/net/rxrpc/call_event.c
+++ b/net/rxrpc/call_event.c
@@ -123,6 +123,7 @@ static void __rxrpc_propose_ACK(struct rxrpc_call *call, u8 ack_reason,
 		else
 			ack_at = expiry;
 
+		ack_at += READ_ONCE(call->tx_backoff);
 		ack_at += now;
 		if (time_before(ack_at, call->ack_at)) {
 			WRITE_ONCE(call->ack_at, ack_at);
@@ -311,6 +312,7 @@ void rxrpc_process_call(struct work_struct *work)
 		container_of(work, struct rxrpc_call, processor);
 	rxrpc_serial_t *send_ack;
 	unsigned long now, next, t;
+	unsigned int iterations = 0;
 
 	rxrpc_see_call(call);
 
@@ -319,6 +321,11 @@ void rxrpc_process_call(struct work_struct *work)
 	       call->debug_id, rxrpc_call_states[call->state], call->events);
 
 recheck_state:
+	/* Limit the number of times we do this before returning to the manager */
+	iterations++;
+	if (iterations > 5)
+		goto requeue;
+
 	if (test_and_clear_bit(RXRPC_CALL_EV_ABORT, &call->events)) {
 		rxrpc_send_abort_packet(call);
 		goto recheck_state;
@@ -447,13 +454,16 @@ void rxrpc_process_call(struct work_struct *work)
 	rxrpc_reduce_call_timer(call, next, now, rxrpc_timer_restart);
 
 	/* other events may have been raised since we started checking */
-	if (call->events && call->state < RXRPC_CALL_COMPLETE) {
-		__rxrpc_queue_call(call);
-		goto out;
-	}
+	if (call->events && call->state < RXRPC_CALL_COMPLETE)
+		goto requeue;
 
 out_put:
 	rxrpc_put_call(call, rxrpc_call_put);
 out:
 	_leave("");
+	return;
+
+requeue:
+	__rxrpc_queue_call(call);
+	goto out;
 }