BGP-ID is insufficient to distinguish between BGP sessions, as it's
perfectly valid/reasonable to have multiple sessions to the same peer.
Or an operator can (incorrectly or otherwise) assign the same BGP-ID to
multiple routers.
This ensures that Session-A to Peer-A going down does not impact
Session-B to Peer-A.

Fixes: #241

Signed-off-by: Trey Aspelund <trey@oxidecomputer.com>

Signed-off-by: Trey Aspelund <trey@oxidecomputer.com>

Add test case to simulate multiple BGP sessions to same peer
Also refactor rdb's test_rib() to be a bit more clean and clear.

Signed-off-by: Trey Aspelund <trey@oxidecomputer.com>

Switch from SocketAddrPair to IpAddr for distinguishing between BGP
sessions. A BGP peer is identified by the IP we are connecting to,
and multiple TCP sessions via the same IP isn't something we're
looking to support. So instead of trying to handle all the cases
where the local SocketAddr may or may not be available, simplify
things by using the configured peer IP which will always be
present regardless of the peer's FSM State.

Signed-off-by: Trey Aspelund <trey@oxidecomputer.com>

move rdb over to using mg_common logging functions, rather than its own
re-implementation. Also add a .gitignore for *.log so nextest doesn't
pollute the repo.

Signed-off-by: Trey Aspelund <trey@oxidecomputer.com>