tty: Destroy ldisc instance on hangup

Currently, when the tty is hungup, the ldisc is re-instanced; ie., the
current instance is destroyed and a new instance is created. The purpose
of this design was to guarantee a valid, open ldisc for the lifetime of
the tty.

However, now that tty buffers are owned by and have lifetime equivalent
to the tty_port (since v3.10), any data received immediately after the
ldisc is re-instanced may cause continued driver i/o operations
concurrently with the driver's hangup() operation. For drivers that
shutdown h/w on hangup, this is unexpected and usually bad. For example,
the serial core may free the xmit buffer page concurrently with an
in-progress write() operation (triggered by echo).

With the existing stable and robust ldisc reference handling, the
cleaned-up tty_reopen(), the straggling unsafe ldisc use cleaned up, and
the preparation to properly handle a NULL tty->ldisc, the ldisc instance
can be destroyed and only re-instanced when the tty is re-opened.

If the tty was opened as /dev/console or /dev/tty0, the original behavior
of re-instancing the ldisc is retained (the 'reinit' parameter to
tty_ldisc_hangup() is true). This is required since those file descriptors
are never hungup.

This patch has neglible impact on userspace; the tty file_operations ptr
is changed to point to the hungup file operations _before_ the ldisc
instance is destroyed, so only racing file operations might now retrieve
a NULL ldisc reference (which is simply handled as if the hungup file
operation had been called instead -- see "tty: Prepare for destroying
line discipline on hangup").

This resolves a long-standing FIXME and several crash reports.

Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

2 files changed, 23 insertions, 29 deletions

diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 7f556e3c1515..d3ecbb513fa1 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -727,7 +727,7 @@ static void __tty_hangup(struct tty_struct *tty, int exit_session)
 	while (refs--)
 		tty_kref_put(tty);
 
-	tty_ldisc_hangup(tty);
+	tty_ldisc_hangup(tty, cons_filp != NULL);
 
 	spin_lock_irq(&tty->ctrl_lock);
 	clear_bit(TTY_THROTTLED, &tty->flags);
@@ -752,10 +752,9 @@ static void __tty_hangup(struct tty_struct *tty, int exit_session)
 	} else if (tty->ops->hangup)
 		tty->ops->hangup(tty);
 	/*
-	 * We don't want to have driver/ldisc interactions beyond
-	 * the ones we did here. The driver layer expects no
-	 * calls after ->hangup() from the ldisc side. However we
-	 * can't yet guarantee all that.
+	 * We don't want to have driver/ldisc interactions beyond the ones
+	 * we did here. The driver layer expects no calls after ->hangup()
+	 * from the ldisc side, which is now guaranteed.
 	 */
 	set_bit(TTY_HUPPED, &tty->flags);
 	tty_unlock(tty);
@@ -1475,7 +1474,8 @@ static int tty_reopen(struct tty_struct *tty)
 
 	tty->count++;
 
-	WARN_ON(!tty->ldisc);
+	if (!tty->ldisc)
+		return tty_ldisc_reinit(tty, tty->termios.c_line);
 
 	return 0;
 }
diff --git a/drivers/tty/tty_ldisc.c b/drivers/tty/tty_ldisc.c
index e18c8e864110..c6f970d63060 100644
--- a/drivers/tty/tty_ldisc.c
+++ b/drivers/tty/tty_ldisc.c
@@ -257,6 +257,9 @@ const struct file_operations tty_ldiscs_proc_fops = {
  *	reference to it. If the line discipline is in flux then
  *	wait patiently until it changes.
  *
+ *	Returns: NULL if the tty has been hungup and not re-opened with
+ *		 a new file descriptor, otherwise valid ldisc reference
+ *
  *	Note: Must not be called from an IRQ/timer context. The caller
  *	must also be careful not to hold other locks that will deadlock
  *	against a discipline change, such as an existing ldisc reference
@@ -642,14 +645,15 @@ static void tty_reset_termios(struct tty_struct *tty)
  *	@disc: line discipline to reinitialize
  *
  *	Completely reinitialize the line discipline state, by closing the
- *	current instance and opening a new instance. If an error occurs opening
- *	the new non-N_TTY instance, the instance is dropped and tty->ldisc reset
- *	to NULL. The caller can then retry with N_TTY instead.
+ *	current instance, if there is one, and opening a new instance. If
+ *	an error occurs opening the new non-N_TTY instance, the instance
+ *	is dropped and tty->ldisc reset to NULL. The caller can then retry
+ *	with N_TTY instead.
  *
  *	Returns 0 if successful, otherwise error code < 0
  */
 
-static int tty_ldisc_reinit(struct tty_struct *tty, int disc)
+int tty_ldisc_reinit(struct tty_struct *tty, int disc)
 {
 	struct tty_ldisc *ld;
 	int retval;
@@ -693,11 +697,9 @@ static int tty_ldisc_reinit(struct tty_struct *tty, int disc)
  *	tty itself so we must be careful about locking rules.
  */
 
-void tty_ldisc_hangup(struct tty_struct *tty)
+void tty_ldisc_hangup(struct tty_struct *tty, bool reinit)
 {
 	struct tty_ldisc *ld;
-	int reset = tty->driver->flags & TTY_DRIVER_RESET_TERMIOS;
-	int err = 0;
 
 	tty_ldisc_debug(tty, "%p: hangup\n", tty->ldisc);
 
@@ -725,25 +727,17 @@ void tty_ldisc_hangup(struct tty_struct *tty)
 	 */
 	tty_ldisc_lock(tty, MAX_SCHEDULE_TIMEOUT);
 
-	if (tty->ldisc) {
+	if (tty->driver->flags & TTY_DRIVER_RESET_TERMIOS)
+		tty_reset_termios(tty);
 
-		/* At this point we have a halted ldisc; we want to close it and
-		   reopen a new ldisc. We could defer the reopen to the next
-		   open but it means auditing a lot of other paths so this is
-		   a FIXME */
-		if (reset == 0)
-			err = tty_ldisc_reinit(tty, tty->termios.c_line);
-
-		/* If the re-open fails or we reset then go to N_TTY. The
-		   N_TTY open cannot fail */
-		if (reset || err < 0)
-			tty_ldisc_reinit(tty, N_TTY);
+	if (tty->ldisc) {
+		if (reinit) {
+			if (tty_ldisc_reinit(tty, tty->termios.c_line) < 0)
+				tty_ldisc_reinit(tty, N_TTY);
+		} else
+			tty_ldisc_kill(tty);
 	}
 	tty_ldisc_unlock(tty);
-	if (reset)
-		tty_reset_termios(tty);
-
-	tty_ldisc_debug(tty, "%p: re-opened\n", tty->ldisc);
 }
 
 /**