mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* Subject: [PATCH 03/91] um: fix oopsable race in line_close()
@ 2011-08-18 18:59 Al Viro
  0 siblings, 0 replies; only message in thread
From: Al Viro @ 2011-08-18 18:59 UTC (permalink / raw)
  To: richard; +Cc: user-mode-linux-devel, linux-kernel


tty->count is decremented only after ->close() had been called and several
tasks can hit it in parallel.  As the result, using tty->count to check if
you are the last one is broken.  We end up leaving line->tty not reset to
NULL and the next IRQ on that sucker will blow up trying to dereference
pointers from kfree'd struct tty.

Fix is obvious: we need to use a counter of our own.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/um/drivers/line.c        |   25 ++++++++++++-------------
 arch/um/include/shared/line.h |    1 +
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index d51c404..c5bff1d 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -399,8 +399,8 @@ int line_setup_irq(int fd, int input, int output, struct line *line, void *data)
  * is done under a spinlock.  Checking whether the device is in use is
  * line->tty->count > 1, also under the spinlock.
  *
- * tty->count serves to decide whether the device should be enabled or
- * disabled on the host.  If it's equal to 1, then we are doing the
+ * line->count serves to decide whether the device should be enabled or
+ * disabled on the host.  If it's equal to 0, then we are doing the
  * first open or last close.  Otherwise, open and close just return.
  */
 
@@ -414,16 +414,16 @@ int line_open(struct line *lines, struct tty_struct *tty)
 		goto out_unlock;
 
 	err = 0;
-	if (tty->count > 1)
+	if (line->count++)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
+	BUG_ON(tty->driver_data);
 	tty->driver_data = line;
 	line->tty = tty;
 
+	spin_unlock(&line->count_lock);
 	err = enable_chan(line);
-	if (err)
+	if (err) /* line_close() will be called by our caller */
 		return err;
 
 	INIT_DELAYED_WORK(&line->task, line_timer_cb);
@@ -436,7 +436,7 @@ int line_open(struct line *lines, struct tty_struct *tty)
 	chan_window_size(&line->chan_list, &tty->winsize.ws_row,
 			 &tty->winsize.ws_col);
 
-	return err;
+	return 0;
 
 out_unlock:
 	spin_unlock(&line->count_lock);
@@ -460,17 +460,16 @@ void line_close(struct tty_struct *tty, struct file * filp)
 	flush_buffer(line);
 
 	spin_lock(&line->count_lock);
-	if (!line->valid)
-		goto out_unlock;
+	BUG_ON(!line->valid);
 
-	if (tty->count > 1)
+	if (--line->count)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
 	line->tty = NULL;
 	tty->driver_data = NULL;
 
+	spin_unlock(&line->count_lock);
+
 	if (line->sigio) {
 		unregister_winch(tty);
 		line->sigio = 0;
@@ -498,7 +497,7 @@ static int setup_one_line(struct line *lines, int n, char *init, int init_prio,
 
 	spin_lock(&line->count_lock);
 
-	if (line->tty != NULL) {
+	if (line->count) {
 		*error_out = "Device is already open";
 		goto out;
 	}
diff --git a/arch/um/include/shared/line.h b/arch/um/include/shared/line.h
index 72f4f25..63df3ca 100644
--- a/arch/um/include/shared/line.h
+++ b/arch/um/include/shared/line.h
@@ -33,6 +33,7 @@ struct line_driver {
 struct line {
 	struct tty_struct *tty;
 	spinlock_t count_lock;
+	unsigned long count;
 	int valid;
 
 	char *init_str;
-- 
1.7.2.5



^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2011-08-18 18:59 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-08-18 18:59 Subject: [PATCH 03/91] um: fix oopsable race in line_close() Al Viro

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome