Some time ago I wrote:
> On Mar 4, 2017, at 9:05 PM, Viktor Dukhovni <free...@dukhovni.org> wrote:
>
> In trying to set up a 6to4 node under 11.0-RELEASE-p8, and the stf0
> interface is constantly flipping in and out of "tentative" state, at
> seemingly 1 second intervals. Is that to be expected? Am I missing
> some non-obvious configuration setting?
Today, I looked closely at the kernel sources and found the issue.
Patch below. Interfaces found to not have IFF_DRV_RUNNING in
if_drv_flags by the periodic interface scan are marked tentative.
Since the "stf" driver did not set that flag, the "stf" interface
flaps.
Index: sys/net/if_stf.c
===================================================================
--- sys/net/if_stf.c (revision 325169)
+++ sys/net/if_stf.c (working copy)
@@ -722,6 +722,7 @@
}
ifp->if_flags |= IFF_UP;
+ ifp->if_drv_flags |= IFF_DRV_RUNNING;
break;
case SIOCADDMULTI:
While I was at it, I also changed the below because for "stf"
the only effect of "break" was the misleading logging:
stf0 is not multicast caple, IPv6 not enabled
Perhaps there is some configuration I am not taking into account
in which something more useful happens... Feeback appreciated.
Index: sys/netinet6/in6_ifattach.c
===================================================================
--- sys/netinet6/in6_ifattach.c (revision 325169)
+++ sys/netinet6/in6_ifattach.c (working copy)
@@ -696,25 +696,25 @@
/*
* quirks based on interface type
*/
switch (ifp->if_type) {
case IFT_STF:
/*
* 6to4 interface is a very special kind of beast.
* no multicast, no linklocal. RFC2529 specifies how to make
* linklocals for 6to4 interface, but there's no use and
* it is rather harmful to have one.
*/
ND_IFINFO(ifp)->flags &= ~ND6_IFF_AUTO_LINKLOCAL;
- break;
+ return;
default:
break;
}
/*
* usually, we require multicast capability to the interface
*/
if ((ifp->if_flags & IFF_MULTICAST) == 0) {
nd6log((LOG_INFO, "in6_ifattach: "
"%s is not multicast capable, IPv6 not enabled\n",
if_name(ifp)));
return;
--
Viktor.
_______________________________________________
freebsd-net@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
