Date: Fri, 17 Mar 2017 11:50:34 +0100 From: Harry Schmalzbauer <freebsd@omnilan.de> To: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org> Subject: Re: Are ./valte-ctl and ./bridge friends or competitors? Message-ID: <58CBBF7A.8050604@omnilan.de> In-Reply-To: <58CBA727.3040108@omnilan.de> References: <58CBA727.3040108@omnilan.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Bezüglich Harry Schmalzbauer's Nachricht vom 17.03.2017 10:06 (localtime): > Hello, > > I'm still having problems understanding netmap(4) and would highly > appreciate brief help. > > I'm running stable/11. I'd like to replace if_bridge(4) with netmap(4), > because virtio-net chops jumbu frames > (https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215737) and > if_bridge(4) requires members to have the same mtu. > > I'm creating vmnet0 (tap(4)) and vale0 with an physical (keeping host > stack usage) port: > ifconfig create vmnet0 > ./vale-ctl -h vale0:igb0 > ./vale-ctl -a vale0:vmnet0 > which results in > bdg_ctl [149] bridge:0 port:0 vale0:vlegn > bdg_ctl [149] bridge:0 port:1 vale0:vlegn^ > bdg_ctl [149] bridge:0 port:2 vale0:vmnet0 The above is wrong regarding bhyve(8) usage. For the records, and please correct me somone if I'm wrong: byhve(8) has some kind of native netmap(4) implementation, so one mustn't use if_tap(4) (vmnet), but simply a vale name (e.g. 'bhyve … -s 5,virtio-net,vale0:guest1 …' instead of '… -s 5,e1000,vmnet0 …') The physical interface connected to the vale switch must be put into promisc mode! No idea about the need to disable any offloading functions, since I don't know how it's implemented… Thanks, -harry P.S.: Still don't understand the basic difference between ./bridge and ./vale-ctl
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?58CBBF7A.8050604>