Forwarding this to netdev since the bug is still present in Linux 3.7.1. For those joining us, this thread is archive at <http://bugs.debian.org/697357>.
On Fri, 2013-01-04 at 13:40 +0100, Peter Palfrader wrote: > On Fri, 04 Jan 2013, Bastian Blank wrote: > > > On Fri, Jan 04, 2013 at 12:17:12PM +0100, Peter Palfrader wrote: > > > I tried to set up a new KVM host in the usual way: > > > - have a bond0 interface over all physcial eth* interfaces > > > - create a br0 over that bond0 and any vnet* interfaces for the > > > guests. > > Also, arp works through the bridge, forgot to mention that. > > > > * Manually switching the eth* to promiscious mode makes stuff work. > > > > Okay, so a workaround is available. > > > > When did it last work? Does it work with the kernel from experimental? > > It works nicely on stable, e.g. pasquini.debian.org: > > | weasel@pasquini:~$ sudo brctl show > | bridge name bridge id STP enabled interfaces > | br0 8000.e83935a9ec10 no bond0 > | tap0 > | tap1 > | tap2 > | br1 8000.e83935a9ec10 no bond0.221 > | br2 8000.da1343affe92 no bond0.3301 > | tap3 > > And there the bond interface is promisc automatically: > weasel@pasquini:~$ ip a | grep -i bond > | 2: eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast > master bond0 state UP qlen 1000 > | 3: eth1: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast > master bond0 state UP qlen 1000 > | 4: bond0: <BROADCAST,MULTICAST,PROMISC,MASTER,UP,LOWER_UP> mtu 1500 qdisc > noqueue state UP > | 7: bond0.221@bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc > noqueue state UP > | 9: bond0.3301@bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 > qdisc noqueue state UP > > Likewise it works as expected when one does br directly on eth0, without any > bonding. > > It's also broken with > linux-image-3.7-trunk-amd64_3.7.1-1~experimental.1_amd64.deb. -- Ben Hutchings If you seem to know what you are doing, you'll be given more to do.
signature.asc
Description: This is a digitally signed message part