On Thu, Oct 20, 2022 at 09:29:06PM +0200, netdev@xxxxxxxxxxxxxxxxxxxx wrote: > On 2022-10-20 14:55, Vladimir Oltean wrote: > > On Tue, Oct 18, 2022 at 06:56:10PM +0200, Hans J. Schultz wrote: > > > diff --git a/net/bridge/br_switchdev.c b/net/bridge/br_switchdev.c > > > index 8f3d76c751dd..c6b938c01a74 100644 > > > --- a/net/bridge/br_switchdev.c > > > +++ b/net/bridge/br_switchdev.c > > > @@ -136,6 +136,7 @@ static void br_switchdev_fdb_populate(struct > > > net_bridge *br, > > > item->added_by_user = test_bit(BR_FDB_ADDED_BY_USER, &fdb->flags); > > > item->offloaded = test_bit(BR_FDB_OFFLOADED, &fdb->flags); > > > item->is_local = test_bit(BR_FDB_LOCAL, &fdb->flags); > > > + item->locked = test_bit(BR_FDB_LOCKED, &fdb->flags); > > > > Shouldn't this be set to 0 here, since it is the bridge->driver > > direction? > > Wouldn't it be a good idea to allow drivers to add what corresponds to a blackhole > entry when using the bridge input chain to activate the MAB feature, or in general > to leave the decision of what to do to the driver implementation? The patch doesn't propose that. It proposes: | net: bridge: enable bridge to install locked fdb entries from drivers | | The bridge will be able to install locked entries when receiving | SWITCHDEV_FDB_ADD_TO_BRIDGE notifications from drivers. Please write patches which make just one logical change, and explain the justification for that change and precisely that change in the commit message.
