CVE-2023-53523

In the Linux kernel, the following vulnerability has been resolved:

can: gs_usb: fix time stamp counter initialization

If the gs_usb device driver is unloaded (or unbound) before the interface is shut down, the USB stack first calls the struct usb_driver::disconnect and then the struct net_device_ops::ndo_stop callback.

In gs_usb_disconnect() all pending bulk URBs are killed, i.e. no more RX'ed CAN frames are send from the USB device to the host. Later in gs_can_close() a reset control message is send to each CAN channel to remove the controller from the CAN bus. In this race window the USB device can still receive CAN frames from the bus and internally queue them to be send to the host.

At least in the current version of the candlelight firmware, the queue of received CAN frames is not emptied during the reset command. After loading (or binding) the gs_usb driver, new URBs are submitted during the struct net_device_ops::ndo_open callback and the candlelight firmware starts sending its already queued CAN frames to the host.

However, this scenario was not considered when implementing the hardware timestamp function. The cycle counter/time counter infrastructure is set up (gs_usb_timestamp_init()) after the USBs are submitted, resulting in a NULL pointer dereference if timecounter_cyc2time() (via the call chain: gs_usb_receive_bulk_callback() -> gs_usb_set_timestamp() -> gs_usb_skb_set_timestamp()) is called too early.

Move the gs_usb_timestamp_init() function before the URBs are submitted to fix this problem.

For a comprehensive solution, we need to consider gs_usb devices with more than 1 channel. The cycle counter/time counter infrastructure is setup per channel, but the RX URBs are per device. Once gs_can_open() of a channel has been called, and URBs have been submitted, the gs_usb_receive_bulk_callback() can be called for all available channels, even for channels that are not running, yet. As cycle counter/time counter has not set up, this will again lead to a NULL pointer dereference.

Convert the cycle counter/time counter from a "per channel" to a "per device" functionality. Also set it up, before submitting any URBs to the device.

Further in gs_usb_receive_bulk_callback(), don't process any URBs for not started CAN channels, only resubmit the URB.

A race in the gs_usb driver allowed RX URBs to deliver queued CAN frames before the timecounter was initialized, causing a NULL-pointer dereference and kernel crash. The issue requires a non-default setup (USB-CAN “candlelight” class device and driver unbind/bind timing). For the CVSS the Privileges Required is High (admin actions to rebind/unbind or manage the netdevice).

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.