Worked fine on my dual-ix(4) setup. In loadbalance mode with 9k frames it gave me more Gbits than with a single ix(4) connection, even if I didn't go ape on all possible combinations of mtu, trunkprotos and whatnot. Quick test with tcpbench on Dell 320s without pf and 9k mtu and two ix(4) showed 6.9Gbit/s over trunk with this patch as opposed to ~5.0 with just one single ix(4) link.
2013/2/22 Stuart Henderson <s...@spacehopper.org>: > I thought we already had something for this after the misc@ thread > a few months ago, but clearly not. > > Adapted from FreeBSD if_lagg.c r171661 (which includes capability > setting which we already do). > http://svnweb.freebsd.org/base/head/sys/net/if_lagg.c?r1=171603&r2=171661 > > Index: if_trunk.c > =================================================================== > RCS file: /cvs/src/sys/net/if_trunk.c,v > retrieving revision 1.78 > diff -u -p -r1.78 if_trunk.c > --- if_trunk.c 28 Oct 2011 12:49:43 -0000 1.78 > +++ if_trunk.c 22 Feb 2013 12:27:18 -0000 > @@ -315,6 +315,20 @@ trunk_port_create(struct trunk_softc *tr > if (ifp->if_type != IFT_ETHER) > return (EPROTONOSUPPORT); > > + /* Take MTU from the first member port */ > + if (SLIST_EMPTY(&tr->tr_ports)) { > + if (tr->tr_ifflags & IFF_DEBUG) > + printf("%s: first port, setting trunk mtu %u\n", > + tr->tr_ifname, ifp->if_mtu); > + tr->tr_ac.ac_if.if_mtu = ifp->if_mtu; > + tr->tr_ac.ac_if.if_hardmtu = ifp->if_mtu; > + } > + else if (tr->tr_ac.ac_if.if_mtu != ifp->if_mtu) { > + printf("%s: adding %s failed, MTU %u != %u\n", tr->tr_ifname, > + ifp->if_xname, ifp->if_mtu, tr->tr_ac.ac_if.if_mtu); > + return (EINVAL); > + } > + > if ((error = ifpromisc(ifp, 1)) != 0) > return (error); > > @@ -510,6 +524,10 @@ trunk_port_ioctl(struct ifnet *ifp, u_lo > } > > trunk_port2req(tp, rp); > + break; > + case SIOCSIFMTU: > + /* Do not allow the MTU to be changed once joined */ > + error = EINVAL; > break; > default: > error = ENOTTY; > -- May the most significant bit of your life be positive.
