linux/net/8021q
Eric Dumazet 93f154b594 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>
2009-05-18 22:19:19 -07:00
..
Kconfig vlan: Add GVRP support 2008-07-05 21:26:57 -07:00
Makefile vlan: uninline __vlan_hwaccel_rx 2008-07-08 03:23:36 -07:00
vlan_core.c Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6 2009-04-16 17:35:26 -07:00
vlan_dev.c net: release dst entry in dev_hard_start_xmit() 2009-05-18 22:19:19 -07:00
vlan_gvrp.c vlan: TCI related type and naming cleanups 2008-07-08 03:24:44 -07:00
vlan_netlink.c vlan: Add GVRP support 2008-07-05 21:26:57 -07:00
vlan.c vlan: update vlan carrier state for admin up/down 2009-04-25 18:03:35 -07:00
vlan.h vlan: propogate ethtool speed values 2008-10-28 23:02:34 -07:00
vlanproc.c vlan: use struct netdev_queue counters instead of dev->stats 2009-05-18 15:15:06 -07:00
vlanproc.h [VLAN]: Add a net argument to proc init and cleanup calls. 2008-04-16 00:51:12 -07:00