net: better IFF_XMIT_DST_RELEASE support

Testing xmit_more support with netperf and connected UDP sockets,
I found strange dst refcount false sharing.

Current handling of IFF_XMIT_DST_RELEASE is not optimal.

Dropping dst in validate_xmit_skb() is certainly too late in case
packet was queued by cpu X but dequeued by cpu Y

The logical point to take care of drop/force is in __dev_queue_xmit()
before even taking qdisc lock.

As Julian Anastasov pointed out, need for skb_dst() might come from some
packet schedulers or classifiers.

This patch adds new helper to cleanly express needs of various drivers
or qdiscs/classifiers.

Drivers that need skb_dst() in their ndo_start_xmit() should call
following helper in their setup instead of the prior :

	dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
->
	netif_keep_dst(dev);

Instead of using a single bit, we use two bits, one being
eventually rebuilt in bonding/team drivers.

The other one, is permanent and blocks IFF_XMIT_DST_RELEASE being
rebuilt in bonding/team. Eventually, we could add something
smarter later.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 9 insertions, 10 deletions

diff --git a/net/core/dev.c b/net/core/dev.c
index a63b8c43c1b6..3c5bdaa44486 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2665,12 +2665,6 @@ static struct sk_buff *validate_xmit_skb(struct sk_buff *skb, struct net_device
         if (skb->next)
                 return skb;
 
-        /* If device doesn't need skb->dst, release it right now while
-         * its hot in this cpu cache
-         */
-        if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
-                skb_dst_drop(skb);
-
         features = netif_skb_features(skb);
         skb = validate_xmit_vlan(skb, features);
         if (unlikely(!skb))
@@ -2811,8 +2805,6 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
                  * waiting to be sent out; and the qdisc is not running -
                  * xmit the skb directly.
                  */
-                if (!(dev->priv_flags & IFF_XMIT_DST_RELEASE))
-                        skb_dst_force(skb);
 
                 qdisc_bstats_update(q, skb);
 
@@ -2827,7 +2819,6 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
 
                 rc = NET_XMIT_SUCCESS;
         } else {
-                skb_dst_force(skb);
                 rc = q->enqueue(skb, q) & NET_XMIT_MASK;
                 if (qdisc_run_begin(q)) {
                         if (unlikely(contended)) {
@@ -2924,6 +2915,14 @@ static int __dev_queue_xmit(struct sk_buff *skb, void *accel_priv)
 
         skb_update_prio(skb);
 
+        /* If device/qdisc don't need skb->dst, release it right now while
+         * its hot in this cpu cache.
+         */
+        if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
+                skb_dst_drop(skb);
+        else
+                skb_dst_force(skb);
+
         txq = netdev_pick_tx(dev, skb, accel_priv);
         q = rcu_dereference_bh(txq->qdisc);
 
@@ -6674,7 +6673,7 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name,
         INIT_LIST_HEAD(&dev->adj_list.lower);
         INIT_LIST_HEAD(&dev->all_adj_list.upper);
         INIT_LIST_HEAD(&dev->all_adj_list.lower);
-        dev->priv_flags = IFF_XMIT_DST_RELEASE;
+        dev->priv_flags = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM;
         setup(dev);
 
         dev->num_tx_queues = txqs;