Sat, Dec 07, 2013 at 06:42:35PM CET, stephen@xxxxxxxxxxxxxxxxxx wrote: >On Sat, 7 Dec 2013 09:51:05 +0100 >Jiri Pirko <jiri@xxxxxxxxxxx> wrote: > >> Fri, Dec 06, 2013 at 10:10:28PM CET, stephen@xxxxxxxxxxxxxxxxxx wrote: >> >On Fri, 06 Dec 2013 15:43:21 -0500 (EST) >> >David Miller <davem@xxxxxxxxxxxxx> wrote: >> > >> >> From: Jiri Pirko <jiri@xxxxxxxxxxx> >> >> Date: Thu, 5 Dec 2013 16:27:37 +0100 >> >> >> >> > br_stp_rcv() is reached by non-rx_handler path. That means there is no >> >> > guarantee that dev is bridge port and therefore simple NULL check of >> >> > ->rx_handler_data is not enough. There is need to check if dev is really >> >> > bridge port and since only rcu read lock is held here, do it by checking >> >> > ->rx_handler pointer. >> >> > >> >> > Note that synchronize_net() in netdev_rx_handler_unregister() ensures >> >> > this approach as valid. >> >> > >> >> > Introduced originally by: >> >> > commit f350a0a87374418635689471606454abc7beaa3a >> >> > "bridge: use rx_handler_data pointer to store net_bridge_port pointer" >> >> > >> >> > Fixed but not in the best way by: >> >> > commit b5ed54e94d324f17c97852296d61a143f01b227a >> >> > "bridge: fix RCU races with bridge port" >> >> > >> >> > Reintroduced by: >> >> > commit 716ec052d2280d511e10e90ad54a86f5b5d4dcc2 >> >> > "bridge: fix NULL pointer deref of br_port_get_rcu" >> >> > >> >> > Please apply to stable trees as well. Thanks. >> >> > >> >> > RH bugzilla reference: https://bugzilla.redhat.com/show_bug.cgi?id=1025770 >> >> > >> >> > Reported-by: Laine Stump <laine@xxxxxxxxxx> >> >> > Debugged-by: Michael S. Tsirkin <mst@xxxxxxxxxx> >> >> > Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx> >> >> > Signed-off-by: Jiri Pirko <jiri@xxxxxxxxxxx> >> >> > --- >> >> > v1->v2: moved br_port_get_check_rcu definition below br_handle_frame definition >> >> >> >> Applied and queued up for -stable, thanks Jiri. >> > >> >How come you ignored my simpler fix, that used the existing logic. >> >I don't like introducing this especially into the stable; much prefer >> >to go back to testing the flag as was being done before. >> >> Although your patch is technically sane, it depends on rtnl indirectly. >> My patch depends on rcu locking and synchronize_rcu which is direct. >> Therefore I think it is more appropriate. > >After more review and thought I agree. But could we put some comments >in br_private.h to describe the dependency on ordering (synchronize_net). Sure. I will send follow-up patch addressing it. Thanks. > >Acked-by: Stephen Hemminger <stephen@xxxxxxxxxxxxxxxxxx>
