Re: [PATCH] bonding: allow bond in mode balance-alb to work properly in bridge -try3

[Jay Vosburgh <fubar@us.ibm.com>]

Jiri Pirko <jpirko@redhat.com> wrote:

>Wed, Mar 25, 2009 at 05:31:53PM CET, fubar@us.ibm.com wrote:
[...]
>>> #if defined(CONFIG_BRIDGE) || defined (CONFIG_BRIDGE_MODULE)
>>> /* These hooks defined here for ATM */
>>> struct net_bridge;
>>>@@ -2251,10 +2258,12 @@ int netif_receive_skb(struct sk_buff *skb)
>>> 	null_or_orig = NULL;
>>> 	orig_dev = skb->dev;
>>> 	if (orig_dev->master) {
>>>-		if (skb_bond_should_drop(skb))
>>>+		if (skb_bond_should_drop(skb)) {
>>> 			null_or_orig = orig_dev; /* deliver only exact match */
>>>-		else
>>>+		} else {
>>> 			skb->dev = orig_dev->master;
>>>+			bond_change_dest_hook(skb);
>>
>>	Since you put the hook outside of the skb_bond_should_drop
>>function, does the VLAN accelerated receive path do the right thing if,
>>e.g., there's a VLAN on top of bonding and that VLAN is part of the
>>bridge?

	Jiri: not trying to be pushy, but you didn't address the above
question about the VLAN path, and I just want to make sure that you saw
it (it was at the bottom of a long email, so I fear you may not have
seen it).

	-J

---
	-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com