jhujhiti_adjectivism.org added a comment.
Forgot to change the other tests - this ought to do it.
One annoying detail: while running these tests, I noticed intermittent
failures on the SLAAC test. I was reminded of a similar intermittent issue I
had while developing this: for reasons I can not explain, the epair0a interface
(in this example) oftentimes gets correctly configured, but the IFDISABLED flag
is never removed, like so:
epair0a: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu
1500
options=8<VLAN_MTU>
ether 02:ff:e0:00:04:0a
inet6 2001:db8::2 prefixlen 64 tentative
inet6 fe80::ff:e0ff:fe00:40a%epair0a prefixlen 64 tentative scopeid
0x4
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
media: Ethernet 10Gbase-T (10Gbase-T <full-duplex>)
status: active
fib: 1
groups: epair
Inserting a sleep between creation of the epair and configuration solves it.
The good news is that this behaves identically on my system without the patch,
so it's not a new regression.
INLINE COMMENTS
> jhujhiti_adjectivism.org wrote in icmp6.c:2147
> > At line 2159, do we know the fib of the receiving interface?
>
> It looks like the mbuf passed to icmp6_reflect() will always have the FIB set
> based on the receiving interface, so we can just use that! Can a more
> experienced set of eyes confirm?
@asomers, can you confirm that M_GETFIB(m) is always correctly set to the FIB
of the receiving interface?
REPOSITORY
rS FreeBSD src repository
REVISION DETAIL
https://reviews.freebsd.org/D9451
EMAIL PREFERENCES
https://reviews.freebsd.org/settings/panel/emailpreferences/
To: jhujhiti_adjectivism.org, #network, bz, asomers
Cc: jch, bz, imp, ae, freebsd-net-list
_______________________________________________
[email protected] mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "[email protected]"
