From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: sendmailtogoran@gmail.com Received: from krantz.zx2c4.com (localhost [127.0.0.1]) by krantz.zx2c4.com (ZX2C4 Mail Server) with ESMTP id b3f41ace for ; Sun, 6 May 2018 11:26:57 +0000 (UTC) Received: from mail-it0-x231.google.com (mail-it0-x231.google.com [IPv6:2607:f8b0:4001:c0b::231]) by krantz.zx2c4.com (ZX2C4 Mail Server) with ESMTP id 82d91f29 for ; Sun, 6 May 2018 11:26:57 +0000 (UTC) Received: by mail-it0-x231.google.com with SMTP id n64-v6so2557103itb.3 for ; Sun, 06 May 2018 04:29:10 -0700 (PDT) MIME-Version: 1.0 From: Goran Date: Sun, 6 May 2018 13:29:09 +0200 Message-ID: Subject: vxlan on top of wireguard To: wireguard@lists.zx2c4.com Content-Type: text/plain; charset="UTF-8" List-Id: Development discussion of WireGuard List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , I have configured a distributed swith (vxlan) on top of a wireguard tunnel. I learned it from https://insom.github.io/journal/2017/04/02/ but did not tried the workaround described there. The workaround states that wireguard has some sort of bug and the hardware-accelerated checksums must be disabled: ethtool -K $i tx off . As the article is from last year, I tried it first without this workaround. With that I could send icmp packets but no tcp. Sometimes I could send a tcp packet with crazy MTU like 64 or so. Does someone knows why this strange behaviour exists? With workaround all works fine.