[BUG] firmware: imx: possible null-pointer dereferences and data-inconsistency due to data races in imx_scu_rx_callback()

[Tuo Li]

Hello,

Our static analysis tool finds some possible data races in the IMX
firmware in Linux 6.4.0.

In the function imx_scu_call_rpc(), the variable sc_ipc->msg is
accessed with holding the lock sc_ipc->lock:

   imx_scu_call_rpc()  --> Line 203
     mutex_lock(&sc_ipc->lock);  --> Line 212 (Lock sc_ipc->lock)
     sc_ipc->msg = msg;  --> Line 216 (Access sc_ipc->msg)
     sc_ipc->msg = NULL;  --> Line 251 (Access sc_ipc->msg)

However, in the function imx_scu_rx_callback(), the variable
sc_ipc->msg is accessed without holding the lock sc_ipc->lock:

   imx_scu_rx_callback()  --> Line 112
     sc_ipc->msg[0] = *data++;  --> Line 129 (Access sc_ipc->msg)
     sc_ipc->msg[sc_chan->idx] = *data;  --> Line 148 (Access sc_ipc->msg)

I am not sure whether the functions imx_scu_call_rpc() and
imx_scu_rx_callback() can be executed concurrently and whether
the variables sc_ipc in the two functions can be the same. If so,
harmful data races can occur because:
1. If the variable sc_ipc->msg is set to NULL after the if check at
Line 120, null-pointer dereferences may occur at Lines 129 and 148;
2. If sc_ipc->msg is changed by imx_scu_call_rpc() when updating
sc_ipc->msg with data. The values in data can be written to different
msgs and cause data-inconsistency.

I am not quite sure whether these possible data races are real and
how to fix them if they are real.

Any feedback would be appreciated, thanks!

Reported-by: BassCheck <bass at buaa.edu.cn>

Best wishes,
Tuo Li