[PATCH 1/2 T] ALSA: timer: Fix race between read and ioctl

Stefan Bader stefan.bader at canonical.com
Wed Jun 21 16:56:00 UTC 2017


From: Takashi Iwai <tiwai at suse.de>

The read from ALSA timer device, the function snd_timer_user_tread(),
may access to an uninitialized struct snd_timer_user fields when the
read is concurrently performed while the ioctl like
snd_timer_user_tselect() is invoked.  We have already fixed the races
among ioctls via a mutex, but we seem to have forgotten the race
between read vs ioctl.

This patch simply applies (more exactly extends the already applied
range of) tu->ioctl_lock in snd_timer_user_tread() for closing the
race window.

Reported-by: Alexander Potapenko <glider at google.com>
Tested-by: Alexander Potapenko <glider at google.com>
Cc: <stable at vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai at suse.de>

CVE-2017-1000380

(cherry-picked from commit d11662f4f798b50d8c8743f433842c3e40fe3378)
Signed-off-by: Stefan Bader <stefan.bader at canonical.com>
---
 sound/core/timer.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sound/core/timer.c b/sound/core/timer.c
index 392aabe..bf97d66 100644
--- a/sound/core/timer.c
+++ b/sound/core/timer.c
@@ -1899,6 +1899,7 @@ static ssize_t snd_timer_user_read(struct file *file, char __user *buffer,
 
 	tu = file->private_data;
 	unit = tu->tread ? sizeof(struct snd_timer_tread) : sizeof(struct snd_timer_read);
+	mutex_lock(&tu->ioctl_lock);
 	spin_lock_irq(&tu->qlock);
 	while ((long)count - result >= unit) {
 		while (!tu->qused) {
@@ -1914,7 +1915,9 @@ static ssize_t snd_timer_user_read(struct file *file, char __user *buffer,
 			add_wait_queue(&tu->qchange_sleep, &wait);
 
 			spin_unlock_irq(&tu->qlock);
+			mutex_unlock(&tu->ioctl_lock);
 			schedule();
+			mutex_lock(&tu->ioctl_lock);
 			spin_lock_irq(&tu->qlock);
 
 			remove_wait_queue(&tu->qchange_sleep, &wait);
@@ -1948,6 +1951,7 @@ static ssize_t snd_timer_user_read(struct file *file, char __user *buffer,
 	}
  _error:
 	spin_unlock_irq(&tu->qlock);
+	mutex_unlock(&tu->ioctl_lock);
 	return result > 0 ? result : err;
 }
 
-- 
2.7.4





More information about the kernel-team mailing list