[Bug 280132] EHOSTUNREACH when using IPv4 over IPv6 nexthop
- Reply: bugzilla-noreply_a_freebsd.org: "[Bug 280132] EHOSTUNREACH when using IPv4 over IPv6 nexthop"
- Reply: bugzilla-noreply_a_freebsd.org: "[Bug 280132] EHOSTUNREACH when using IPv4 over IPv6 nexthop"
- Reply: bugzilla-noreply_a_freebsd.org: "[Bug 280132] EHOSTUNREACH when using IPv4 over IPv6 nexthop"
- Reply: bugzilla-noreply_a_freebsd.org: "[Bug 280132] EHOSTUNREACH when using IPv4 over IPv6 nexthop"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Thu, 04 Jul 2024 22:23:16 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=280132
Bug ID: 280132
Summary: EHOSTUNREACH when using IPv4 over IPv6 nexthop
Product: Base System
Version: 14.1-RELEASE
Hardware: amd64
OS: Any
Status: New
Severity: Affects Only Me
Priority: ---
Component: kern
Assignee: bugs@FreeBSD.org
Reporter: email@luiz.eng.br
After upgrading one of my hosts from FreeBSD 13.2-RELEASE to 14.1-RELEASE, I
started running into issues when originating IPv4 traffic from the host.
On my public interface, I have a single IPv6 GUA address. No IPv4.
I have a loopback interface that has a public IPv4 address configured on it,
with a /32 netmask.
I then install an IPv4 default route with: route add -net default -inet6
2a12:6e40:0::xx
This setup works fine on the 13.2 host, but on the 14.1 host, it results in a
EHOSTUNREACH. I first noticed the issue because unbound was having issues with
domains that had only IPv4 nameservers, but I was also able to reproduce it
when originating any type of IPv4 traffic from the host: mtr, telnet, ping all
fail.
I compared the netstat -no4 and netstat -nr4 outputs between the two systems
and they are similar, changing only the last octet of the IPs between the
systems.
I don't have a lot of experience with dtrace, but I put together a small script
to check the nhop_object returned by fib4_lookup and I noticed that on FreeBSD
14.1 nh_ifa has an IPv6 link-local address, while on 13.2 it has an IPv4 (which
interestingly is not the public one).
I'm glad to provide more information and run more tests on request.
--
You are receiving this mail because:
You are the assignee for the bug.