FRRouting: bgpd keeps dead BGP sessions up when neighbor stops reading

In FRRouting 10.1 and later, bgpd does not correctly close a BGP session when the Send Hold Timer (RFC 9687) expires, i.e. when a neighbor has stopped accepting data from the router for twice the hold time. The connection is closed, but bgpd still considers the session to be up.

A BGP neighbor that keeps sending KEEPALIVEs but stops reading from its connection, at a time when the router has a large amount of data to send to it (for example while sending it a full routing table), can cause one of the following:

• If the neighbor announced Graceful Restart support: bgpd crashes shortly after the session is closed (after the neighbor's announced restart time plus one keepalive interval), taking down all BGP sessions on the router. The neighbor can repeat this each time bgpd restarts.
• Otherwise: the session stays stuck in the Established state. The neighbor's routes stay installed and keep being advertised to other peers, and the neighbor cannot reconnect until an operator runs "clear bgp <neighbor>".

The neighbor does not need to be malicious; an overloaded or stuck peer can trigger the same behavior.

Required configuration: no special configuration is needed. Any BGP neighbor with an established session can trigger the issue under default settings, for example:

router bgp 65001 neighbor 192.0.2.2 remote-as 65002

Graceful Restart helper mode, which leads to the crash, is enabled by default. With default timers (keepalive 60s, hold time 180s), the session is closed about 6 minutes after the neighbor stops reading.