[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <1518021487.2100.2.camel@debian.org>
Date: Wed, 07 Feb 2018 17:38:07 +0100
From: Yves-Alexis Perez <corsac@...ian.org>
To: Mike Maloney <maloney@...gle.com>
Cc: "David S. Miller" <davem@...emloft.net>,
Alexey Kuznetsov <kuznet@....inr.ac.ru>,
Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
Eric Dumazet <edumazet@...gle.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
stable@...r.kernel.org, debian-kernel@...ts.debian.org
Subject: Regression for ip6-in-ip4 IPsec tunnel in 4.14.16
Hi Mike,
I noticed a regression in 4.14.16 stable kernel (I assume it's also
present in mainline).
I'm using an IPsec setup where I tunnel all my IP trafic to a gateway.
The tunnel can use either IPv6 or IPv4 (depending on what's available
locally) and will route both IPv4 and IPv6 (my gateway will assign both
family addresses).
The tunnel doesn't use ESP directly but rather encapsulates in UDP.
Starting with 4.14.16, IPv6 traffic is broken. When trying a simple ping
to an IPv6 address I get:
ping: sendmsg: Invalid argument
I bisected 4.14.15 to 4.14.16 and got the attached bisect log, which
ends with 8278804e05f6bcfe3fdfea4a404020752ead15a6 as the offending
commit. The -EINVAL is consistent with the “Invalid argument” return
from ping. I didn't try yet on a pure IPv6 setup (without IPsec
tunneling) but will followup when I have a chance to test it.
Reverting that commit on top of 4.14.17 fixes the problem.
If you need more info, please ask.
Regards,
--
Yves-Alexis
View attachment "ip6-bisect.log" of type "text/x-log" (1308 bytes)
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists