In the Linux kernel, the following vulnerability has been resolved:
mptcp: pm: Fix uaf in _timerdelete_sync
There are two paths to access mptcppmdeladdtimer, result in a race condition:
CPU1 CPU2
==== ====
net_rx_action
napi_poll netlink_sendmsg
__napi_poll netlink_unicast
process_backlog netlink_unicast_kernel
__netif_receive_skb genl_rcv
__netif_receive_skb_one_core netlink_rcv_skb
NF_HOOK genl_rcv_msg
ip_local_deliver_finish genl_family_rcv_msg
ip_protocol_deliver_rcu genl_family_rcv_msg_doit
tcp_v4_rcv mptcp_pm_nl_flush_addrs_doit
tcp_v4_do_rcv mptcp_nl_remove_addrs_list
tcp_rcv_established mptcp_pm_remove_addrs_and_subflows
tcp_data_queue remove_anno_list_by_saddr
mptcp_incoming_options mptcp_pm_del_add_timer
mptcp_pm_del_add_timer kfree(entry)
In removeannolistbysaddr(running on CPU2), after leaving the critical zone protected by "pm.lock", the entry will be released, which leads to the occurrence of uaf in the mptcppmdeladdtimer(running on CPU1).
Keeping a reference to addtimer inside the lock, and calling skstoptimersync() with this reference, instead of "entry->add_timer".
Move listdel(&entry->list) to mptcppmdeladd_timer and inside the pm lock, do not directly access any members of the entry outside the pm lock, which can avoid similar "entry->x" uaf.