diff options
| author | Peter Hill <peter.j.hill@oracle.com> | 2021-08-10 17:07:32 -0700 |
|---|---|---|
| committer | FUJITA Tomonori <fujita.tomonori@gmail.com> | 2021-08-13 07:43:18 +0900 |
| commit | 0b449fda7d993575cc141a2d822721fde4fcb0a4 (patch) | |
| tree | cc0dab2bc18892ec726aa125213d0d2359a992cc | |
| parent | 696ff4de8ec2906d00de2a21df40e261af59803a (diff) | |
Update the hold timer when a BGP Update message is received
According to RFC4271,
The calculated value indicates the maximum number of
seconds that may elapse between the receipt of successive
KEEPALIVE and/or UPDATE messages from the sender.
This change will reset the hold timer in the update FSM case.
(note, I am super new with Go. I'm not quite sure how to craft a good test case for this.)
| -rw-r--r-- | pkg/server/fsm.go | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/pkg/server/fsm.go b/pkg/server/fsm.go index abd94ed9..fedf3018 100644 --- a/pkg/server/fsm.go +++ b/pkg/server/fsm.go @@ -1019,6 +1019,13 @@ func (h *fsmHandler) recvMessageWithError() (*fsmMsg, error) { case bgp.BGP_MSG_ROUTE_REFRESH: fmsg.MsgType = fsmMsgRouteRefresh case bgp.BGP_MSG_UPDATE: + // if the length of h.holdTimerResetCh + // isn't zero, the timer will be reset + // soon anyway. + select { + case h.holdTimerResetCh <- true: + default: + } body := m.Body.(*bgp.BGPUpdate) isEBGP := h.fsm.pConf.IsEBGPPeer(h.fsm.gConf) isConfed := h.fsm.pConf.IsConfederationMember(h.fsm.gConf) |