On 2017/06/24 14:31, Harald Dunkel wrote: > On 06/24/17 11:57, Stuart Henderson wrote: > > > > Since you have a 6.0 system handy, are you able to apply > > https://github.com/openbsd/src/commit/3e4bd4b41d124df8849a234977327b543297783e.patch > > to the kernel to see if this is what introduces the breakage? > > > > No, it doesn't: > > # tcpdump -i em0 -env icmp6 > tcpdump: listening on em0, link-type EN10MB > 14:18:35.921888 d6:19:0d:b4:0f:cb 52:54:00:12:34:56 86dd 118: > fe80::f66d:4ff:fe73:ab4e > fe80::5054:ff:fe12:3456: icmp6: echo request > (id:03f9 seq:1) [flowlabel 0x2fecc] (len 64, hlim 255) > 14:18:35.922092 52:54:00:12:34:56 d6:19:0d:b4:0f:cb 86dd 118: > fe80::5054:ff:fe12:3456 > fe80::f66d:4ff:fe73:ab4e: icmp6: echo reply > (id:03f9 seq:1) (len 64, hlim 64) > 14:18:36.941135 d6:19:0d:b4:0f:cb 52:54:00:12:34:56 86dd 118: > fe80::f66d:4ff:fe73:ab4e > fe80::5054:ff:fe12:3456: icmp6: echo request > (id:03f9 seq:2) [flowlabel 0x2fecc] (len 64, hlim 255) > 14:18:36.941294 52:54:00:12:34:56 d6:19:0d:b4:0f:cb 86dd 118: > fe80::5054:ff:fe12:3456 > fe80::f66d:4ff:fe73:ab4e: icmp6: echo reply > (id:03f9 seq:2) (len 64, hlim 64) > 14:18:37.960175 d6:19:0d:b4:0f:cb 52:54:00:12:34:56 86dd 118: > fe80::f66d:4ff:fe73:ab4e > fe80::5054:ff:fe12:3456: icmp6: echo request > (id:03f9 seq:3) [flowlabel 0x2fecc] (len 64, hlim 255) > 14:18:37.960277 52:54:00:12:34:56 d6:19:0d:b4:0f:cb 86dd 118: > fe80::5054:ff:fe12:3456 > fe80::f66d:4ff:fe73:ab4e: icmp6: echo reply > (id:03f9 seq:3) (len 64, hlim 64) > > That was 6.0 stable plus your patch.
Hmm.. Are you able to go through some date-based checkouts of sys/netinet6 between say mid-august and mid-november to try to find when it got introduced?
