net: core: explicitly select a txq before doing l2 forwarding

Currently, the tx queue were selected implicitly in ndo_dfwd_start_xmit(). The
will cause several issues:

- NETIF_F_LLTX were removed for macvlan, so txq lock were done for macvlan
  instead of lower device which misses the necessary txq synchronization for
  lower device such as txq stopping or frozen required by dev watchdog or
  control path.
- dev_hard_start_xmit() was called with NULL txq which bypasses the net device
  watchdog.
- dev_hard_start_xmit() does not check txq everywhere which will lead a crash
  when tso is disabled for lower device.

Fix this by explicitly introducing a new param for .ndo_select_queue() for just
selecting queues in the case of l2 forwarding offload. netdev_pick_tx() was also
extended to accept this parameter and dev_queue_xmit_accel() was used to do l2
forwarding transmission.

With this fixes, NETIF_F_LLTX could be preserved for macvlan and there's no need
to check txq against NULL in dev_hard_start_xmit(). Also there's no need to keep
a dedicated ndo_dfwd_start_xmit() and we can just reuse the code of
dev_queue_xmit() to do the transmission.

In the future, it was also required for macvtap l2 forwarding support since it
provides a necessary synchronization method.

Cc: John Fastabend <john.r.fastabend@intel.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: e1000-devel@lists.sourceforge.net
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 25 insertions, 16 deletions

diff --git a/net/core/dev.c b/net/core/dev.c
index 4fc17221545d..0ce469e5ec80 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2539,7 +2539,7 @@ static inline int skb_needs_linearize(struct sk_buff *skb,
 }
 
 int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev,
-			struct netdev_queue *txq, void *accel_priv)
+			struct netdev_queue *txq)
 {
 	const struct net_device_ops *ops = dev->netdev_ops;
 	int rc = NETDEV_TX_OK;
@@ -2605,13 +2605,10 @@ int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev,
 			dev_queue_xmit_nit(skb, dev);
 
 		skb_len = skb->len;
-		if (accel_priv)
-			rc = ops->ndo_dfwd_start_xmit(skb, dev, accel_priv);
-		else
 			rc = ops->ndo_start_xmit(skb, dev);
 
 		trace_net_dev_xmit(skb, rc, dev, skb_len);
-		if (rc == NETDEV_TX_OK && txq)
+		if (rc == NETDEV_TX_OK)
 			txq_trans_update(txq);
 		return rc;
 	}
@@ -2627,10 +2624,7 @@ gso:
 			dev_queue_xmit_nit(nskb, dev);
 
 		skb_len = nskb->len;
-		if (accel_priv)
-			rc = ops->ndo_dfwd_start_xmit(nskb, dev, accel_priv);
-		else
-			rc = ops->ndo_start_xmit(nskb, dev);
+		rc = ops->ndo_start_xmit(nskb, dev);
 		trace_net_dev_xmit(nskb, rc, dev, skb_len);
 		if (unlikely(rc != NETDEV_TX_OK)) {
 			if (rc & ~NETDEV_TX_MASK)
@@ -2811,7 +2805,7 @@ EXPORT_SYMBOL(dev_loopback_xmit);
  *      the BH enable code must have IRQs enabled so that it will not deadlock.
  *          --BLG
  */
-int dev_queue_xmit(struct sk_buff *skb)
+int __dev_queue_xmit(struct sk_buff *skb, void *accel_priv)
 {
 	struct net_device *dev = skb->dev;
 	struct netdev_queue *txq;
@@ -2827,7 +2821,7 @@ int dev_queue_xmit(struct sk_buff *skb)
 
 	skb_update_prio(skb);
 
-	txq = netdev_pick_tx(dev, skb);
+	txq = netdev_pick_tx(dev, skb, accel_priv);
 	q = rcu_dereference_bh(txq->qdisc);
 
 #ifdef CONFIG_NET_CLS_ACT
@@ -2863,7 +2857,7 @@ int dev_queue_xmit(struct sk_buff *skb)
 
 			if (!netif_xmit_stopped(txq)) {
 				__this_cpu_inc(xmit_recursion);
-				rc = dev_hard_start_xmit(skb, dev, txq, NULL);
+				rc = dev_hard_start_xmit(skb, dev, txq);
 				__this_cpu_dec(xmit_recursion);
 				if (dev_xmit_complete(rc)) {
 					HARD_TX_UNLOCK(dev, txq);
@@ -2892,8 +2886,19 @@ out:
 	rcu_read_unlock_bh();
 	return rc;
 }
+
+int dev_queue_xmit(struct sk_buff *skb)
+{
+	return __dev_queue_xmit(skb, NULL);
+}
 EXPORT_SYMBOL(dev_queue_xmit);
 
+int dev_queue_xmit_accel(struct sk_buff *skb, void *accel_priv)
+{
+	return __dev_queue_xmit(skb, accel_priv);
+}
+EXPORT_SYMBOL(dev_queue_xmit_accel);
+
 
 /*=======================================================================
 			Receiver routines
diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index d6ef17322500..2fc5beaf5783 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -395,17 +395,21 @@ u16 __netdev_pick_tx(struct net_device *dev, struct sk_buff *skb)
 EXPORT_SYMBOL(__netdev_pick_tx);
 
 struct netdev_queue *netdev_pick_tx(struct net_device *dev,
-				    struct sk_buff *skb)
+				    struct sk_buff *skb,
+				    void *accel_priv)
 {
 	int queue_index = 0;
 
 	if (dev->real_num_tx_queues != 1) {
 		const struct net_device_ops *ops = dev->netdev_ops;
 		if (ops->ndo_select_queue)
-			queue_index = ops->ndo_select_queue(dev, skb);
+			queue_index = ops->ndo_select_queue(dev, skb,
+							    accel_priv);
 		else
 			queue_index = __netdev_pick_tx(dev, skb);
-		queue_index = dev_cap_txqueue(dev, queue_index);
+
+		if (!accel_priv)
+			queue_index = dev_cap_txqueue(dev, queue_index);
 	}
 
 	skb_set_queue_mapping(skb, queue_index);
diff --git a/net/core/netpoll.c b/net/core/netpoll.c
index 303097874633..19fe9c717ced 100644
--- a/net/core/netpoll.c
+++ b/net/core/netpoll.c
@@ -375,7 +375,7 @@ void netpoll_send_skb_on_dev(struct netpoll *np, struct sk_buff *skb,
 	if (skb_queue_len(&npinfo->txq) == 0 && !netpoll_owner_active(dev)) {
 		struct netdev_queue *txq;
 
-		txq = netdev_pick_tx(dev, skb);
+		txq = netdev_pick_tx(dev, skb, NULL);
 
 		/* try until next clock tick */
 		for (tries = jiffies_to_usecs(1)/USEC_PER_POLL;