drm: correctly restore mappings if drm_open fails
authorIlija Hadzic <ihadzic@research.bell-labs.com>
Tue, 2 Apr 2013 13:56:10 +0000 (09:56 -0400)
committerDave Airlie <airlied@redhat.com>
Tue, 2 Apr 2013 20:44:38 +0000 (06:44 +1000)
If first drm_open fails, the error-handling path will
incorrectly restore inode's mapping to NULL. This can
cause the crash later on. Fix by separately storing
away mapping pointers that drm_open can touch and
restore each from its own respective variable if the
call fails.

Fixes: https://bugzilla.novell.com/show_bug.cgi?id=807850
(thanks to Michal Hocko for investigating investigating and
finding the root cause of the bug)

Reference:
http://lists.freedesktop.org/archives/dri-devel/2013-March/036564.html

v2: Use one variable to store file and inode mapping
    since they are the same at the function entry.
    Fix spelling mistakes in commit message.

v3: Add reference to the original bug report.

Reported-by: Marco Munderloh <munderl@tnt.uni-hannover.de>
Tested-by: Marco Munderloh <munderl@tnt.uni-hannover.de>
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
drivers/gpu/drm/drm_fops.c

index 13fdcd10a6051ddd9359af281390748528a6885e..429e07d0b0f147f2c9f672d65a18849f0c57f0f4 100644 (file)
@@ -123,6 +123,7 @@ int drm_open(struct inode *inode, struct file *filp)
        int retcode = 0;
        int need_setup = 0;
        struct address_space *old_mapping;
+       struct address_space *old_imapping;
 
        minor = idr_find(&drm_minors_idr, minor_id);
        if (!minor)
@@ -137,6 +138,7 @@ int drm_open(struct inode *inode, struct file *filp)
        if (!dev->open_count++)
                need_setup = 1;
        mutex_lock(&dev->struct_mutex);
+       old_imapping = inode->i_mapping;
        old_mapping = dev->dev_mapping;
        if (old_mapping == NULL)
                dev->dev_mapping = &inode->i_data;
@@ -159,8 +161,8 @@ int drm_open(struct inode *inode, struct file *filp)
 
 err_undo:
        mutex_lock(&dev->struct_mutex);
-       filp->f_mapping = old_mapping;
-       inode->i_mapping = old_mapping;
+       filp->f_mapping = old_imapping;
+       inode->i_mapping = old_imapping;
        iput(container_of(dev->dev_mapping, struct inode, i_data));
        dev->dev_mapping = old_mapping;
        mutex_unlock(&dev->struct_mutex);