On 2016/06/04 2:35, Nikolay Aleksandrov wrote:
...
>>> void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr)
>>> @@ -288,6 +296,95 @@ out:
>>> spin_unlock_bh(&br->hash_lock);
>>> }
>>>
>>> +void br_fdb_sync_uc(struct net_bridge *br)
>>> +{
>>> + struct net_bridge_vlan_group *vg;
>>> + struct netdev_hw_addr *ha;
>>> + int i;
>>> +
>>> + spin_lock_bh(&br->hash_lock);
>>> +
>>> + for (i = 0; i < BR_HASH_SIZE; i++) {
>>> + struct hlist_node *h;
>>> +
>>> + hlist_for_each(h, &br->hash[i]) {
>>> + struct net_bridge_fdb_entry *f;
>>> +
>>> + f = hlist_entry(h, struct net_bridge_fdb_entry, hlist);
>>> + if (!f->dst && f->is_local && !f->added_by_user &&
>>> + !ether_addr_equal(f->addr.addr, br->dev->dev_addr)) {
>>> + /* delete old one */
>>> + fdb_delete_local(br, NULL, f);
>>> + }
>>> + }
>>> + }
>>> +
>>> + vg = br_vlan_group(br);
>>> +
>>> + /* insert new address, may fail if invalid address or dup. */
>>> + netdev_for_each_uc_addr(ha, br->dev) {
>>> + struct net_bridge_vlan *v;
>>> +
>>> + fdb_insert(br, NULL, ha->addr, 0);
>>> +
>>> + if (!vg || !vg->num_vlans)
>>> + continue;
>>> +
>>> + list_for_each_entry(v, &vg->vlan_list, vlist)
>>> + fdb_insert(br, NULL, ha->addr, v->vid);
>>
>> Since here you’re walking over the bridge’s vlan list, you should test the vlans with br_vlan_should_use()
>> because it can be a global context holder if the vlan was configured only on ports.
Thank you for your feedback.
will fix in v2.
I actually thought that this is the same logic as
br_fdb_change_mac_address() so assumed it should be all right.
Does br_fdb_change_mac_address() need br_vlan_should_use() as well?
Toshiaki Makita
