Date: Mon, 16 May 2016 01:41:40 +0200 From: Ruben <mail@osfux.nl> To: freebsd-net@freebsd.org Subject: Re: bhyve with vlans - host and vm can't pass traffic Message-ID: <8fae9862-90e5-9098-d30c-a2f0a4b01aa1@osfux.nl> In-Reply-To: <9c759cf4-01f0-6f21-c18a-27b3d491dcbf@osfux.nl> References: <553A8DF0.2060901@sds.com> <9c759cf4-01f0-6f21-c18a-27b3d491dcbf@osfux.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Scott, It turns out that my predicament (incorrect _tcp_ connectivity between guest and host, but correct tcp connectivity between guest and rest of network) was probably being caused by a combination of both using the lagg driver, the intel em driver, and probably vlan traffic (an old post (finally) pointed me towards a solution : https://forums.freebsd.org/threads/7668 ). tcpdump -vvv on the related tap / vtnet interfacs showed a lot of checksum errors specifically related to tcp traffic (icmp was fine) and after disabling tso4, lro and hardware vlan tagging for both the lagg and the nics making up the lagg interface, my ordeal was over. I just added these directives to the server's specific rc.conf ifconfig_ lines : -vlanhwtag -lro -tso eg: ifconfig_em0="up media 1000baseTX mediaopt full-duplex -tso4 -lro -vlanhwtag" ifconfig_lagg0="laggproto lacp laggport em0 -tso4 -lro -vlanhwtag" And (after a reboot) everything started working like a charm. So: my problem was not bhyve related, but rather to the networking stack / specific drivers (and my "combination" of using them ofc). Kind regards / sorry for polluting this bhyve thread, Ruben On 13/05/16 02:37, Ruben wrote: > Hi Scott, > > Since I'm currently facing a somewhat similar problem (to the best of my > understanding): where you ever able to solve this? I'd be more than > happy to hear of any solution for the predicament you were in :P > > Kind regards, > > Ruben
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8fae9862-90e5-9098-d30c-a2f0a4b01aa1>