msg/async: add timeout for connections which are not yet ready
It turns out the new keep_alive mechanism introduced in
9499e85b16ae865e16c6f6f83fa90349d572e322 does not make a lot of
sense if the corresponding connection is not yet ready, as we
always require the underlying connection to be **ready** in order
to send out a keep_alive message.
Since there could be various corner cases that may cause an async
connection stuck in the connecting stage, I think introducing a
timeout strategy should generally make more much sense...
Master PR: https://github.com/ceph/ceph/pull/27337
This does not follow the normal backport process because the protocol
of transport layer of Ceph has been greatly changed for Nautilus.
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>