net: fix packet socket delivery in rx irq handler

The changes to deliver hardware accelerated VLAN packets to packet
sockets (commit bc1d0411) caused a warning for non-NAPI drivers.
The __vlan_hwaccel_rx() function is called directly from the drivers
RX function, for non-NAPI drivers that means its still in RX IRQ
context:

[   27.779463] ------------[ cut here ]------------
[   27.779509] WARNING: at kernel/softirq.c:136 local_bh_enable+0x37/0x81()
...
[   27.782520]  [<c0264755>] netif_nit_deliver+0x5b/0x75
[   27.782590]  [<c02bba83>] __vlan_hwaccel_rx+0x79/0x162
[   27.782664]  [<f8851c1d>] atl1_intr+0x9a9/0xa7c [atl1]
[   27.782738]  [<c0155b17>] handle_IRQ_event+0x23/0x51
[   27.782808]  [<c015692e>] handle_edge_irq+0xc2/0x102
[   27.782878]  [<c0105fd5>] do_IRQ+0x4d/0x64

Split hardware accelerated VLAN reception into two parts to fix this:

- __vlan_hwaccel_rx just stores the VLAN TCI and performs the VLAN
  device lookup, then calls netif_receive_skb()/netif_rx()

- vlan_hwaccel_do_receive(), which is invoked by netif_receive_skb()
  in softirq context, performs the real reception and delivery to
  packet sockets.

Reported-and-tested-by: Ramon Casellas <ramon.casellas@cttc.es>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c
index 916061f..68ced4b 100644
--- a/net/8021q/vlan_core.c
+++ b/net/8021q/vlan_core.c
@@ -3,11 +3,20 @@
 #include <linux/if_vlan.h>
 #include "vlan.h"
 
+struct vlan_hwaccel_cb {
+	struct net_device	*dev;
+};
+
+static inline struct vlan_hwaccel_cb *vlan_hwaccel_cb(struct sk_buff *skb)
+{
+	return (struct vlan_hwaccel_cb *)skb->cb;
+}
+
 /* VLAN rx hw acceleration helper.  This acts like netif_{rx,receive_skb}(). */
 int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp,
 		      u16 vlan_tci, int polling)
 {
-	struct net_device_stats *stats;
+	struct vlan_hwaccel_cb *cb = vlan_hwaccel_cb(skb);
 
 	if (skb_bond_should_drop(skb)) {
 		dev_kfree_skb_any(skb);
@@ -15,23 +24,35 @@
 	}
 
 	skb->vlan_tci = vlan_tci;
+	cb->dev = vlan_group_get_device(grp, vlan_tci & VLAN_VID_MASK);
+
+	return (polling ? netif_receive_skb(skb) : netif_rx(skb));
+}
+EXPORT_SYMBOL(__vlan_hwaccel_rx);
+
+int vlan_hwaccel_do_receive(struct sk_buff *skb)
+{
+	struct vlan_hwaccel_cb *cb = vlan_hwaccel_cb(skb);
+	struct net_device *dev = cb->dev;
+	struct net_device_stats *stats;
+
 	netif_nit_deliver(skb);
 
-	skb->dev = vlan_group_get_device(grp, vlan_tci & VLAN_VID_MASK);
-	if (skb->dev == NULL) {
-		dev_kfree_skb_any(skb);
-		/* Not NET_RX_DROP, this is not being dropped
-		 * due to congestion. */
-		return NET_RX_SUCCESS;
+	if (dev == NULL) {
+		kfree_skb(skb);
+		return -1;
 	}
-	skb->dev->last_rx = jiffies;
+
+	skb->dev = dev;
+	skb->priority = vlan_get_ingress_priority(dev, skb->vlan_tci);
 	skb->vlan_tci = 0;
 
-	stats = &skb->dev->stats;
+	dev->last_rx = jiffies;
+
+	stats = &dev->stats;
 	stats->rx_packets++;
 	stats->rx_bytes += skb->len;
 
-	skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tci);
 	switch (skb->pkt_type) {
 	case PACKET_BROADCAST:
 		break;
@@ -43,13 +64,12 @@
 		 * This allows the VLAN to have a different MAC than the
 		 * underlying device, and still route correctly. */
 		if (!compare_ether_addr(eth_hdr(skb)->h_dest,
-					skb->dev->dev_addr))
+					dev->dev_addr))
 			skb->pkt_type = PACKET_HOST;
 		break;
 	};
-	return (polling ? netif_receive_skb(skb) : netif_rx(skb));
+	return 0;
 }
-EXPORT_SYMBOL(__vlan_hwaccel_rx);
 
 struct net_device *vlan_dev_real_dev(const struct net_device *dev)
 {