2.6.32-longterm review patch.  If anyone has any objections, please let us know.

------------------

From: Jiri Slaby <[email protected]>

commit cb174681a9ececa6702f114b85bdf82144b6a5af upstream.

[ Backport to .32.y by Antonio Ospite <[email protected]> ]

There is a window between hidraw_table check and its dereference.
In that window, the device may be unplugged and removed form the
system and we will then dereference NULL.

Lock that place properly so that either we get NULL and jump out or we
can work with real pointer.

Signed-off-by: Jiri Slaby <[email protected]>
Signed-off-by: Jiri Kosina <[email protected]>
Signed-off-by: Antonio Ospite <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

---
 drivers/hid/hidraw.c |   11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

--- a/drivers/hid/hidraw.c
+++ b/drivers/hid/hidraw.c
@@ -211,11 +211,14 @@ static int hidraw_release(struct inode *
        unsigned int minor = iminor(inode);
        struct hidraw *dev;
        struct hidraw_list *list = file->private_data;
+       int ret;
 
+       mutex_lock(&minors_lock);
        if (!hidraw_table[minor]) {
                printk(KERN_EMERG "hidraw device with minor %d doesn't exist\n",
                                minor);
-               return -ENODEV;
+               ret = -ENODEV;
+               goto unlock;
        }
 
        list_del(&list->node);
@@ -229,10 +232,12 @@ static int hidraw_release(struct inode *
                        kfree(list->hidraw);
                }
        }
-
        kfree(list);
+       ret = 0;
+unlock:
+       mutex_unlock(&minors_lock);
 
-       return 0;
+       return ret;
 }
 
 static long hidraw_ioctl(struct file *file, unsigned int cmd,


_______________________________________________
stable mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/stable

Reply via email to