tty: Change tty lock order to master->slave

When releasing the master pty, the slave pty also needs to be locked
to prevent concurrent tty count changes for the slave pty and to
ensure that only one parallel master and slave release observe the
final close, and proceed to destruct the pty pair. Conversely, when
releasing the slave pty, locking the master pty is not necessary
(since the master's state can be inferred by the slave tty count).

Introduce tty_lock_slave()/tty_unlock_slave() which acquires/releases
the tty lock of the slave pty. Remove tty_lock_pair()/tty_unlock_pair().

Dropping the tty_lock is no longer required to re-establish a stable
lock order.

Reviewed-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

1 files changed, 6 insertions, 4 deletions

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index bd7cde3..4ecee28 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -1790,7 +1790,9 @@ int tty_release(struct inode *inode, struct file *filp)
 	if (tty->ops->close)
 		tty->ops->close(tty, filp);
 
-	tty_unlock(tty);
+	/* If tty is pty master, lock the slave pty (stable lock order) */
+	tty_lock_slave(o_tty);
+
 	/*
 	 * Sanity check: if tty->count is going to zero, there shouldn't be
 	 * any waiters on tty->read_wait or tty->write_wait.  We test the
@@ -1804,8 +1806,6 @@ int tty_release(struct inode *inode, struct file *filp)
 	 * Thus this test wouldn't be triggered at the time the slave closed,
 	 * so we do it now.
 	 */
-	tty_lock_pair(tty, o_tty);
-
 	while (1) {
 		do_sleep = 0;
 
@@ -1879,7 +1879,9 @@ int tty_release(struct inode *inode, struct file *filp)
 	/* check whether both sides are closing ... */
 	final = !tty->count && !(o_tty && o_tty->count);
 
-	tty_unlock_pair(tty, o_tty);
+	tty_unlock_slave(o_tty);
+	tty_unlock(tty);
+
 	/* At this point, the tty->count == 0 should ensure a dead tty
 	   cannot be re-opened by a racing opener */