Bug 231144 - fragmented IPv6 drops
Summary: fragmented IPv6 drops
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 11.2-RELEASE
Hardware: amd64 Any
: --- Affects Only Me
Assignee: freebsd-net mailing list
Depends on:
Reported: 2018-09-04 08:12 UTC by Igor
Modified: 2019-02-02 10:27 UTC (History)
2 users (show)

See Also:


Note You need to log in before you can comment on or make changes to this bug.
Description Igor 2018-09-04 08:12:07 UTC

I have a problem with handling fragmented IPv6 packets.

- My system is FreeBSD 11.2-RELEASE-p2 with GENERIC kernel
- I have configured IPv6-only authoritative DNS server
And when the oid net.inet6.ip6.forwarding is set to 1, the system stops to receive fragmented IPv6 packets. If i set net.inet6.ip6.forwarding to 0, the behavior is returning to normal. As i know IPv6 packets MUST NO be fragmented by transit gateways itself. But there is no anything said about incoming packets that are ALREADY fragmented. So could you please help me to solve this issue?
For comparesion i tried to reproduce the problem under FreeBSD-10.3. There is no any issues with fragmented IPv6, all the packets were handled by system as well, regardless of meaning net.inet6.ip6.forwarding sysctl.

Thank you in advance!
Comment 1 Kristof Provost freebsd_committer 2018-09-04 08:26:15 UTC
I suspect you may be hitting 231045.
The fix is in stable/11, and will hopefully get an EN update soon.
Comment 2 Igor 2018-09-04 09:26:36 UTC
(In reply to Kristof Provost from comment #1)

And as a dirty hack, can i just copy frag6.c from 11/STABLE to RELENG and then recompile the kernel?
Comment 3 Kristof Provost freebsd_committer 2018-09-04 09:30:25 UTC
(In reply to Igor from comment #2)
That should work, yes.
Comment 4 Igor 2018-09-04 09:36:02 UTC
(In reply to Kristof Provost from comment #3)
Ok. Thank you very much for your help!
Comment 5 Tom Jones freebsd_committer 2019-02-02 10:27:47 UTC
kp@ suggests this was resolved after r231045. If you can still reproduce this issue, please reopen this bug with a reproduction case.