can: j1939: prevent deadlock by moving j1939_sk_errqueue()

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

can: j1939: prevent deadlock by moving j1939_sk_errqueue()

This commit addresses a deadlock situation that can occur in certain

scenarios, such as when running data TP/ETP transfer and subscribing to

the error queue while receiving a net down event. The deadlock involves

locks in the following order:

3

j1939_session_list_lock -> active_session_list_lock

j1939_session_activate

...

j1939_sk_queue_activate_next -> sk_session_queue_lock

...

j1939_xtp_rx_eoma_one

2

j1939_sk_queue_drop_all -> sk_session_queue_lock

...

j1939_sk_netdev_event_netdown -> j1939_socks_lock

j1939_netdev_notify

1

j1939_sk_errqueue -> j1939_socks_lock

__j1939_session_cancel -> active_session_list_lock

j1939_tp_rxtimer

CPU0 CPU1

---- ----

lock(&priv->active_session_list_lock);

lock(&jsk->sk_session_queue_lock);

lock(&priv->active_session_list_lock);

lock(&priv->j1939_socks_lock);

The solution implemented in this commit is to move the

j1939_sk_errqueue() call out of the active_session_list_lock context,

thus preventing the deadlock situation.