mptcp: cope racing subflow creation in mptcp_rcv_space_adjust

jira VULN-46443
cve CVE-2024-53122
commit-author Paolo Abeni <[email protected]>
commit ce7356a
upstream-diff A recent change f410cbe introduced in v6.10-rc1
              tcp annotate data-races around tp->window_clamp
	      had some fuzz due to the WRITE_ONCE and keep the
	      original code.  No conflicts in merged in code.

Additional active subflows - i.e. created by the in kernel path
manager - are included into the subflow list before starting the
3whs.

A racing recvmsg() spooling data received on an already established
subflow would unconditionally call tcp_cleanup_rbuf() on all the
current subflows, potentially hitting a divide by zero error on
the newly created ones.

Explicitly check that the subflow is in a suitable state before
invoking tcp_cleanup_rbuf().

Fixes: c76c695 ("mptcp: call tcp_cleanup_rbuf on subflows")
	Signed-off-by: Paolo Abeni <[email protected]>
	Reviewed-by: Matthieu Baerts (NGI0) <[email protected]>
Link: https://patch.msgid.link/02374660836e1b52afc91966b7535c8c5f7bafb0.1731060874.git.pabeni@redhat.com
	Signed-off-by: Jakub Kicinski <[email protected]>
(cherry picked from commit ce7356a)
	Signed-off-by: Jonathan Maple <[email protected]>

Author: PlaidCat

net/mptcp/protocol.c

@@ -1980,7 +1980,8 @@ static void mptcp_rcv_space_adjust(struct mptcp_sock *msk, int copied)
 				slow = lock_sock_fast(ssk);
 				WRITE_ONCE(ssk->sk_rcvbuf, rcvbuf);
 				tcp_sk(ssk)->window_clamp = window_clamp;
-				tcp_cleanup_rbuf(ssk, 1);
+				if (tcp_can_send_ack(ssk))
+					tcp_cleanup_rbuf(ssk, 1);
 				unlock_sock_fast(ssk, slow);
 			}
 		}