net: release dst entry in dev_hard_start_xmit()
One point of contention in high network loads is the dst_release() performed
when a transmited skb is freed. This is because NIC tx completion calls
dev_kree_skb() long after original call to dev_queue_xmit(skb).
CPU cache is cold and the atomic op in dst_release() stalls. On SMP, this is
quite visible if one CPU is 100% handling softirqs for a network device,
since dst_clone() is done by other cpus, involving cache line ping pongs.
It seems right place to release dst is in dev_hard_start_xmit(), for most
devices but ones that are virtual, and some exceptions.
David Miller suggested to define a new device flag, set in alloc_netdev_mq()
(so that most devices set it at init time), and carefuly unset in devices
which dont want a NULL skb->dst in their ndo_start_xmit().
List of devices that must clear this flag is :
- loopback device, because it calls netif_rx() and quoting Patrick :
"ip_route_input() doesn't accept loopback addresses, so loopback packets
already need to have a dst_entry attached."
- appletalk/ipddp.c : needs skb->dst in its xmit function
- And all devices that call again dev_queue_xmit() from their xmit function
(as some classifiers need skb->dst) : bonding, vlan, macvlan, eql, ifb, hdlc_fr
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/drivers/net/appletalk/ipddp.c b/drivers/net/appletalk/ipddp.c
index da64ba8..f939e92 100644
--- a/drivers/net/appletalk/ipddp.c
+++ b/drivers/net/appletalk/ipddp.c
@@ -71,6 +71,7 @@
if (!dev)
return ERR_PTR(-ENOMEM);
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
strcpy(dev->name, "ipddp%d");
if (version_printed++ == 0)
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 96d7689..92a9d69 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -5148,6 +5148,7 @@
goto out_rtnl;
}
+ bond_dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
if (!name) {
res = dev_alloc_name(bond_dev, "bond%d");
if (res < 0)
diff --git a/drivers/net/eql.c b/drivers/net/eql.c
index 5210bb1..19b7dd9 100644
--- a/drivers/net/eql.c
+++ b/drivers/net/eql.c
@@ -194,6 +194,7 @@
dev->type = ARPHRD_SLIP;
dev->tx_queue_len = 5; /* Hands them off fast */
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
}
static int eql_open(struct net_device *dev)
diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c
index 60a2630..96713ef 100644
--- a/drivers/net/ifb.c
+++ b/drivers/net/ifb.c
@@ -156,6 +156,7 @@
dev->flags |= IFF_NOARP;
dev->flags &= ~IFF_MULTICAST;
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
random_ether_addr(dev->dev_addr);
}
diff --git a/drivers/net/loopback.c b/drivers/net/loopback.c
index 6f71157..da472c6 100644
--- a/drivers/net/loopback.c
+++ b/drivers/net/loopback.c
@@ -170,6 +170,7 @@
dev->tx_queue_len = 0;
dev->type = ARPHRD_LOOPBACK; /* 0x0001*/
dev->flags = IFF_LOOPBACK;
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
dev->features = NETIF_F_SG | NETIF_F_FRAGLIST
| NETIF_F_TSO
| NETIF_F_NO_CSUM
diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index 329cd50..d5334b4 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -414,6 +414,7 @@
{
ether_setup(dev);
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
dev->netdev_ops = &macvlan_netdev_ops;
dev->destructor = free_netdev;
dev->header_ops = &macvlan_hard_header_ops,
diff --git a/drivers/net/wan/hdlc_fr.c b/drivers/net/wan/hdlc_fr.c
index 8005301..bfa0161 100644
--- a/drivers/net/wan/hdlc_fr.c
+++ b/drivers/net/wan/hdlc_fr.c
@@ -1054,6 +1054,7 @@
dev->flags = IFF_POINTOPOINT;
dev->hard_header_len = 10;
dev->addr_len = 2;
+ dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
}
static const struct net_device_ops pvc_ops = {