Re: [Idr] WG LC for draft-ietf-idr-bgp-sendholdtimer-03 (3/23 to 4/12/2024)

[Claudio Jeker <[e-mail]>]

On Sat, Mar 23, 2024 at 02:24:59PM +0100, Claudio Jeker wrote:
> On Fri, Mar 22, 2024 at 08:53:51AM +0000, Susan Hares wrote:
> > This begins a 3-week WG LC for draft-ietf-idr-bgp-sendholdtimer (3/23/2024 to 4/12/2024).
> >  https://datatracker.ietf.org/doc/draft-ietf-idr-bgp-sendholdtimer/
> > 
> > The authors should reply to this message indicating if they know of any IPR relating to this draft.
> > 
> > This draft modifies RFC4271's finite state machine.  Please review it carefully.
> > 
> > Cheerily, Susan Hares
> 
> As one of the developers that implemented this in OpenBGPD I'm in full
> support of this draft. The changes to the BGP FSM are minimal, well
> contained and clearly formulated.

I decided to double check my patch to OpenBGPD and validate it against the
draft and noticed that the draft left out one bit regarding the special
case of a Hold Time interval of zero.

I suggest to alter the last sentence of Section 3.3 from
    Each time the local system sends a KEEPALIVE, UPDATE, and/or NOTIFICATION
    message, it restarts its SendHoldTimer.
to
    Each time the local system sends a KEEPALIVE, UPDATE, and/or NOTIFICATION
    message, it restarts its SendHoldTimer unless the negotiated HoldTime
    value is zero in which case the SendHoldTimer is stopped.

On top of this one could argue that in "Connect State:", the handling of
Event 20 and in "OpenSent", the handling of Event 19 need to be adjusted
as well. But IMO that is not really needed since at that point the
SendHoldTimer is not yet running.

-- 
:wq Claudio