On 13-03-17 16:28, Eric Dumazet wrote:
On Mon, 2017-03-13 at 15:36 +0100, Thierry Du Tre wrote:
+void vlan_dev_flush_egress_priority(const struct net_device *dev)
+{
+ struct vlan_dev_priv *vlan = vlan_dev_priv(dev);
+ struct vlan_priority_tci_mapping *mp;
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(vlan->egress_priority_map); i++) {
+ while ((mp = vlan->egress_priority_map[i]) != NULL) {
+ vlan->egress_priority_map[i] = mp->next;
+ kfree(mp);
+ }
+ }
+ vlan->nr_egress_mappings = 0;
+}
Are you doing this on a live device ?
Looks you'll need proper RCU support then.
Otherwise, crashes will happen, say in egress_priority_map()
smp_rmb() wont be enough if items can now be destroyed.
Yes, good point about the smp concurrency.
I wrongly assumed there was already some locking mechanism in place to
protect these structures.
Let me take a look at this and I'll come back with a new proposal.
--
Thierry