On Thu, Nov 13, 2025 at 05:57:20PM +0100, Claudio Jeker wrote:
> The comment about NEXTHOP_NOMODIFY makes no sense, in this case of
> up_get_nexthop() there is nothing checking for NEXTHOP_NOMODIFY.
> Just remove this comment.

Not sure it makes no sense. You're modifying the nexthop and ignore a
possible NOMODIFY. But sure, kill it.

> 
> -- 
> :wq Claudio
> 
> Index: rde_update.c
> ===================================================================
> RCS file: /cvs/src/usr.sbin/bgpd/rde_update.c,v
> diff -u -p -r1.178 rde_update.c
> --- rde_update.c	13 Nov 2025 09:45:16 -0000	1.178
> +++ rde_update.c	13 Nov 2025 15:56:28 -0000
> @@ -488,7 +488,7 @@ up_get_nexthop(struct rde_peer *peer, st
>  		/*
>  		 * per RFC: if remote peer address is equal to the nexthop set
>  		 * the nexthop to our local address. This reduces the risk of
> -		 * routing loops. This overrides NEXTHOP_NOMODIFY.
> +		 * routing loops.
>  		 */
>  		if (memcmp(&state->nexthop->exit_nexthop,
>  		    &peer->remote_addr, sizeof(peer->remote_addr)) == 0) {
>