The current location of cancel_work_sync() in hci_uart_tty_close()
before the call to hci_unregister_dev(hdev) may momentarily interfere
with the execution of hci_uart_tty_close() by waiting for an active call
to hci_uart_write_work() to finish. However, its current location is
ineffective at preventing a race condition described next.
cancel_work_sync() is needed to try to prevent hci_uart_write_work() from
running whilst the Data Link protocol layer is being unbound (closed).
However, it is weak as a race condition exists between
hci_uart_write_work() being scheduled via hci_uart_tx_wakeup() and
closure of the Data Link protocol layer. In particular, the Data Link
protocol layer may have a retransmission timer which will independently
call hci_uart_tx_wakeup().
To minimise the race condition, cancel_work_sync() needs to be as close
as possible to the code that closes down the Data Link protocol layer.
Therefore, move the cancel_work_sync() statement to be just before
the code which unbinds (closes) the Data Link protocol layer.
Signed-off-by: Dean Jenkins <Dean_Jenkins@xxxxxxxxxx>
---
drivers/bluetooth/hci_ldisc.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c
index acb08c6..61e03dd4 100644
--- a/drivers/bluetooth/hci_ldisc.c
+++ b/drivers/bluetooth/hci_ldisc.c
@@ -498,8 +498,6 @@ static void hci_uart_tty_close(struct tty_struct *tty)
hdev = hu->hdev;
- cancel_work_sync(&hu->write_work);
-
if (test_bit(HCI_UART_REGISTERED, &hu->flags))
/* Note hci_unregister_dev() may try to send a HCI RESET
* command. If the transmission fails then hci_unregister_dev()
@@ -509,6 +507,7 @@ static void hci_uart_tty_close(struct tty_struct *tty)
if (test_bit(HCI_UART_PROTO_READY, &hu->flags)) {
clear_bit(HCI_UART_PROTO_READY, &hu->flags);
+ cancel_work_sync(&hu->write_work);
hu->proto->close(hu);
}
--
2.7.4
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
