Bluetooth: hci_ldisc: Fix another race when closing the tty.

[ Upstream commit 0338b1b393ec7910898e8f7b25b3bf31a7282e16 ]

The following race condition still existed:

         P1                                P2
  cancel_work_sync()
                                     hci_uart_tx_wakeup()
                                     hci_uart_write_work()
                                     hci_uart_dequeue()
  clear_bit(HCI_UART_PROTO_READY)
  hci_unregister_dev(hdev)
  hci_free_dev(hdev)
  hu->proto->close(hu)
  kfree(hu)
                                     access to hdev and hu

Cancelling the work after clearing the HCI_UART_PROTO_READY bit avoids
this as any hci_uart_tx_wakeup() issued after the flag is cleared will
detect that and not schedule further work.

Signed-off-by: Ronald Tschalär <ronald@innovation.ch>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c
index a746627..6e24038 100644 (file)
--- a/drivers/bluetooth/hci_ldisc.c
+++ b/drivers/bluetooth/hci_ldisc.c
@@ -510,13 +510,13 @@ static void hci_uart_tty_close(struct tty_struct *tty)
        if (hdev)
                hci_uart_close(hdev);
 
-       cancel_work_sync(&hu->write_work);
-
        if (test_bit(HCI_UART_PROTO_READY, &hu->flags)) {
                write_lock_irqsave(&hu->proto_lock, flags);
                clear_bit(HCI_UART_PROTO_READY, &hu->flags);
                write_unlock_irqrestore(&hu->proto_lock, flags);
 
+               cancel_work_sync(&hu->write_work);
+
                if (hdev) {
                        if (test_bit(HCI_UART_REGISTERED, &hu->flags))
                                hci_unregister_dev(hdev);