mptcp: plug races between subflow fail and subflow creation

commit def5b7b2 upstream.

We have races similar to the one addressed by the previous patch between
subflow failing and additional subflow creation. They are just harder to
trigger.

The solution is similar. Use a separate flag to track the condition
'socket state prevent any additional subflow creation' protected by the
fallback lock.

The socket fallback makes such flag true, and also receiving or sending
an MP_FAIL option.

The field 'allow_infinite_fallback' is now always touched under the
relevant lock, we can drop the ONCE annotation on write.

Fixes: 478d7700 ("mptcp: send out MP_FAIL when data checksum fails")
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20250714-net-mptcp-fallback-races-v1-2-391aff963322@kernel.org


Signed-off-by: Jakub Kicinski <kuba@kernel.org>
[ Conflicts in subflow.c, because commit f1f26512 ("mptcp: use plain
  bool instead of custom binary enum") and commit 46a5d3ab
  ("mptcp: fix typos in comments") are not in this version. Both are
  causing conflicts in the context, and the same modifications can still
  be applied. Same in protocol.h with commit b8dc6d6c ("mptcp: fix
  rcv buffer auto-tuning"). Conflicts in protocol.c because commit
  ee2708ae ("mptcp: use get_retrans wrapper") is not in this version
  and refactor the code in __mptcp_retrans(), but the modification can
  still be applied, just not at the same indentation level. ]
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>