On Thu, 2014-04-24 at 18:04 +0200, Tom Gundersen wrote: > On Thu, Apr 24, 2014 at 2:16 PM, Toshiaki Makita > <makita.toshiaki@xxxxxxxxxxxxx> wrote: > > When bridge device is created with IFLA_ADDRESS, we are not calling > > br_stp_change_bridge_id(), which leads to incorrect local fdb > > management and bridge id calculation, and prevents us from receiving > > frames on the bridge device. > > > > Reported-by: Tom Gundersen <teg@xxxxxxx> > > Thanks. That looks correct to me (not able to test at the moment > though). Would this be appropriate for stable if it goes in? This can apply 3.14 as is. For 3.12 or earlier tree, this needs modification to insert the fdb entry (or backport fdb fix patchset (684bd2e1)). And for 3.2, another patch needs to be backported before this because br_fdb_change_mac_address() doesn't exist. Thanks, Toshiaki Makita