nfc: nci: fix circular locking dependency in nci_close_device

nci_close_device() flushes rx_wq and tx_wq while holding req_lock.
This causes a circular locking dependency because nci_rx_work()
running on rx_wq can end up taking req_lock too:

  nci_rx_work -> nci_rx_data_packet -> nci_data_exchange_complete
    -> __sk_destruct -> rawsock_destruct -> nfc_deactivate_target
    -> nci_deactivate_target -> nci_request -> mutex_lock(&ndev->req_lock)

Move the flush of rx_wq after req_lock has been released.
This should safe (I think) because NCI_UP has already been cleared
and the transport is closed, so the work will see it and return
-ENETDOWN.

NIPA has been hitting this running the nci selftest with a debug
kernel on roughly 4% of the runs.

Fixes: 6a2968a ("NFC: basic NCI protocol implementation")
Reviewed-by: Ian Ray <[email protected]>
Link: https://patch.msgid.link/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

Author: kuba-moo

net/nfc/nci/core.c

@@ -579,8 +579,7 @@ static int nci_close_device(struct nci_dev *ndev)
 	skb_queue_purge(&ndev->rx_q);
 	skb_queue_purge(&ndev->tx_q);
 
-	/* Flush RX and TX wq */
-	flush_workqueue(ndev->rx_wq);
+	/* Flush TX wq, RX wq flush can't be under the lock */
 	flush_workqueue(ndev->tx_wq);
 
 	/* Reset device */
@@ -592,13 +591,13 @@ static int nci_close_device(struct nci_dev *ndev)
 		      msecs_to_jiffies(NCI_RESET_TIMEOUT));
 
 	/* After this point our queues are empty
-	 * and no works are scheduled.
+	 * rx work may be running but will see that NCI_UP was cleared
 	 */
 	ndev->ops->close(ndev);
 
 	clear_bit(NCI_INIT, &ndev->flags);
 
-	/* Flush cmd wq */
+	/* Flush cmd and tx wq */
 	flush_workqueue(ndev->cmd_wq);
 
 	timer_delete_sync(&ndev->cmd_timer);
@@ -613,6 +612,9 @@ static int nci_close_device(struct nci_dev *ndev)
 
 	mutex_unlock(&ndev->req_lock);
 
+	/* rx_work may take req_lock via nci_deactivate_target */
+	flush_workqueue(ndev->rx_wq);
+
 	return 0;
 }