Re: [RFC PATCH net-next 5/5] bridge: vlan lwt dst_metadata hooks in ingress and egress paths

On 1/22/17, 4:15 AM, Nikolay Aleksandrov wrote:

On 21/01/17 06:46, Roopa Prabhu wrote:

quoted

From: Roopa Prabhu <redacted>

- ingress hook:
    - if port is a lwt tunnel port, use tunnel info in
      attached dst_metadata to map it to a local vlan
- egress hook:
    - if port is a lwt tunnel port, use tunnel info attached to
      vlan to set dst_metadata on the skb

CC: Nikolay Aleksandrov <redacted>
Signed-off-by: Roopa Prabhu <redacted>
---
CC'ing Nikolay for some more eyes as he has been trying to keep the
bridge driver fast path lite.

 net/bridge/br_input.c   |    4 ++++
 net/bridge/br_private.h |    4 ++++
 net/bridge/br_vlan.c    |   55 +++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 63 insertions(+)

diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
index 83f356f..96602a1 100644
--- a/net/bridge/br_input.c
+++ b/net/bridge/br_input.c

@@ -262,6 +262,10 @@ rx_handler_result_t br_handle_frame(struct sk_buff **pskb)
 		return RX_HANDLER_CONSUMED;
 
 	p = br_port_get_rcu(skb->dev);
+	if (p->flags & BR_LWT_VLAN) {
+		if (br_handle_ingress_vlan_tunnel(skb, p, nbp_vlan_group_rcu(p)))
+			goto drop;
+	}

Is there any reason to do this so early (perhaps netfilter?) ? If not, you can push it to the vlan __allowed_ingress
(and rename that function to something else, it does a hundred additional things)
and avoid this check for all packets if vlans are disabled, thus people using non-vlan filtering
bridge won't have an additional test in their fast path

yes, forgot to mention it in the commit log. I had it close to __allowed_ingress in my first version...had to move it up here
because br_nf_pre_routing/br_nf_pre_routing_finish reset the dst...and hence already late..