thunderbolt: Fix NULL dereference in tb_remove_work()grafted

There is a slight race between tb_remove_work() and tb_domain_remove()
which leads to dereferencing a NULL tb->root_switch pointer inside
tb_free_unplugged_xdomains():

        Thread A                            Thread B

  tb_remove_work()
                                    tb_domain_remove()
                                      mutex_lock(&tb->lock)
                                      tb_stop()
                                        /* doesn't cancel a running callback */
                                        cancel_delayed_work(&tcm->remove_work)
                                        ...
                                        tb_switch_remove(tb->root_switch)
                                        tb->root_switch = NULL
                                      mutex_unlock(&tb->lock)
    mutex_lock(&tb->lock)
    ...
    /* without checking ->root_switch */
    tb_free_unplugged_xdomains(tb->root_switch)
    mutex_unlock(&tb->lock)

Commit a8937f35cf39 ("thunderbolt: Remove XDomain from the bus without
holding tb->lock") doesn't seem right to move tb_free_unplugged_xdomains()
out of the &tb->lock section and the check for tb->root_switch, in
particular.  It states:

  For this reason separate removing the XDomain from the topology data
  structures (where we need the lock) from unregistering the device from
  the bus (where remove callbacks of the drivers are being called).

tb_free_unplugged_xdomains() belongs to the former group of functions
requiring the lock.  And it also calls tb_xdomain_remove() which should
only be called with &tb->lock held.

Found by Linux Verification Center (linuxtesting.org) with Svace static
analysis tool.

Fixes: a8937f35cf39 ("thunderbolt: Remove XDomain from the bus without holding tb->lock")
Cc: stable@vger.kernel.org
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>

1 files changed, 129 insertions, 0 deletions

diff --git a/Documentation/usb/authorization.rst b/Documentation/usb/authorization.rst
new file mode 100644
index 000000000..150a14970
--- /dev/null
+++ b/Documentation/usb/authorization.rst
@@ -0,0 +1,129 @@
+==============================================================
+Authorizing (or not) your USB devices to connect to the system
+==============================================================
+
+Copyright (C) 2007 Inaky Perez-Gonzalez <inaky@linux.intel.com> Intel Corporation
+
+This feature allows you to control if a USB device can be used (or
+not) in a system. This feature will allow you to implement a lock-down
+of USB devices, fully controlled by user space.
+
+As of now, when a USB device is connected it is configured and
+its interfaces are immediately made available to the users.  With this
+modification, only if root authorizes the device to be configured will
+then it be possible to use it.
+
+Usage
+=====
+
+Authorize a device to connect::
+
+	$ echo 1 > /sys/bus/usb/devices/DEVICE/authorized
+
+De-authorize a device::
+
+	$ echo 0 > /sys/bus/usb/devices/DEVICE/authorized
+
+Set new devices connected to hostX to be deauthorized by default (ie:
+lock down)::
+
+	$ echo 0 > /sys/bus/usb/devices/usbX/authorized_default
+
+Remove the lock down::
+
+	$ echo 1 > /sys/bus/usb/devices/usbX/authorized_default
+
+By default, all USB devices are authorized.  Writing "2" to the
+authorized_default attribute causes the kernel to authorize by default
+only devices connected to internal USB ports.
+
+
+Example system lockdown (lame)
+------------------------------
+
+Imagine you want to implement a lockdown so only devices of type XYZ
+can be connected (for example, it is a kiosk machine with a visible
+USB port)::
+
+  boot up
+  rc.local ->
+
+   for host in /sys/bus/usb/devices/usb*
+   do
+      echo 0 > $host/authorized_default
+   done
+
+Hookup an script to udev, for new USB devices::
+
+ if device_is_my_type $DEV
+ then
+   echo 1 > $device_path/authorized
+ done
+
+
+Now, device_is_my_type() is where the juice for a lockdown is. Just
+checking if the class, type and protocol match something is the worse
+security verification you can make (or the best, for someone willing
+to break it). If you need something secure, use crypto and Certificate
+Authentication or stuff like that. Something simple for an storage key
+could be::
+
+ function device_is_my_type()
+ {
+   echo 1 > authorized		# temporarily authorize it
+                                # FIXME: make sure none can mount it
+   mount DEVICENODE /mntpoint
+   sum=$(md5sum /mntpoint/.signature)
+   if [ $sum = $(cat /etc/lockdown/keysum) ]
+   then
+        echo "We are good, connected"
+        umount /mntpoint
+        # Other stuff so others can use it
+   else
+        echo 0 > authorized
+   fi
+ }
+
+
+Of course, this is lame, you'd want to do a real certificate
+verification stuff with PKI, so you don't depend on a shared secret,
+etc, but you get the idea. Anybody with access to a device gadget kit
+can fake descriptors and device info. Don't trust that. You are
+welcome.
+
+
+Interface authorization
+-----------------------
+
+There is a similar approach to allow or deny specific USB interfaces.
+That allows to block only a subset of an USB device.
+
+Authorize an interface::
+
+	$ echo 1 > /sys/bus/usb/devices/INTERFACE/authorized
+
+Deauthorize an interface::
+
+	$ echo 0 > /sys/bus/usb/devices/INTERFACE/authorized
+
+The default value for new interfaces
+on a particular USB bus can be changed, too.
+
+Allow interfaces per default::
+
+	$ echo 1 > /sys/bus/usb/devices/usbX/interface_authorized_default
+
+Deny interfaces per default::
+
+	$ echo 0 > /sys/bus/usb/devices/usbX/interface_authorized_default
+
+Per default the interface_authorized_default bit is 1.
+So all interfaces would authorized per default.
+
+Note:
+  If a deauthorized interface will be authorized so the driver probing must
+  be triggered manually by writing INTERFACE to /sys/bus/usb/drivers_probe
+
+For drivers that need multiple interfaces all needed interfaces should be
+authorized first. After that the drivers should be probed.
+This avoids side effects.