2006-09-13 11:24:59 -06:00
|
|
|
/* drivers/net/ifb.c:
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
The purpose of this driver is to provide a device that allows
|
|
|
|
for sharing of resources:
|
|
|
|
|
|
|
|
1) qdiscs/policies that are per device as opposed to system wide.
|
|
|
|
ifb allows for a device which can be redirected to thus providing
|
|
|
|
an impression of sharing.
|
|
|
|
|
|
|
|
2) Allows for queueing incoming traffic for shaping instead of
|
2006-09-13 11:24:59 -06:00
|
|
|
dropping.
|
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
The original concept is based on what is known as the IMQ
|
|
|
|
driver initially written by Martin Devera, later rewritten
|
|
|
|
by Patrick McHardy and then maintained by Andre Correa.
|
|
|
|
|
|
|
|
You need the tc action mirror or redirect to feed this device
|
|
|
|
packets.
|
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or
|
|
|
|
modify it under the terms of the GNU General Public License
|
|
|
|
as published by the Free Software Foundation; either version
|
|
|
|
2 of the License, or (at your option) any later version.
|
2006-09-13 11:24:59 -06:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
Authors: Jamal Hadi Salim (2005)
|
2006-09-13 11:24:59 -06:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/etherdevice.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/moduleparam.h>
|
2006-09-13 11:24:59 -06:00
|
|
|
#include <net/pkt_sched.h>
|
2007-09-17 12:56:21 -06:00
|
|
|
#include <net/net_namespace.h>
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
#define TX_TIMEOUT (2*HZ)
|
2006-09-13 11:24:59 -06:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
#define TX_Q_LIMIT 32
|
|
|
|
struct ifb_private {
|
|
|
|
struct tasklet_struct ifb_tasklet;
|
|
|
|
int tasklet_pending;
|
|
|
|
/* mostly debug stats leave in for now */
|
|
|
|
unsigned long st_task_enter; /* tasklet entered */
|
|
|
|
unsigned long st_txq_refl_try; /* transmit queue refill attempt */
|
|
|
|
unsigned long st_rxq_enter; /* receive queue entered */
|
|
|
|
unsigned long st_rx2tx_tran; /* receive to trasmit transfers */
|
|
|
|
unsigned long st_rxq_notenter; /*receiveQ not entered, resched */
|
|
|
|
unsigned long st_rx_frm_egr; /* received from egress path */
|
|
|
|
unsigned long st_rx_frm_ing; /* received from ingress path */
|
|
|
|
unsigned long st_rxq_check;
|
|
|
|
unsigned long st_rxq_rsch;
|
|
|
|
struct sk_buff_head rq;
|
|
|
|
struct sk_buff_head tq;
|
|
|
|
};
|
|
|
|
|
2006-02-23 17:23:51 -07:00
|
|
|
static int numifbs = 2;
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
static void ri_tasklet(unsigned long dev);
|
2009-08-31 13:50:51 -06:00
|
|
|
static netdev_tx_t ifb_xmit(struct sk_buff *skb, struct net_device *dev);
|
2006-01-08 23:34:25 -07:00
|
|
|
static int ifb_open(struct net_device *dev);
|
|
|
|
static int ifb_close(struct net_device *dev);
|
|
|
|
|
2006-09-13 11:24:59 -06:00
|
|
|
static void ri_tasklet(unsigned long dev)
|
2006-01-08 23:34:25 -07:00
|
|
|
{
|
|
|
|
|
|
|
|
struct net_device *_dev = (struct net_device *)dev;
|
|
|
|
struct ifb_private *dp = netdev_priv(_dev);
|
2007-10-03 18:41:50 -06:00
|
|
|
struct net_device_stats *stats = &_dev->stats;
|
2008-07-31 17:58:50 -06:00
|
|
|
struct netdev_queue *txq;
|
2006-01-08 23:34:25 -07:00
|
|
|
struct sk_buff *skb;
|
|
|
|
|
2008-07-31 17:58:50 -06:00
|
|
|
txq = netdev_get_tx_queue(_dev, 0);
|
2006-01-08 23:34:25 -07:00
|
|
|
dp->st_task_enter++;
|
|
|
|
if ((skb = skb_peek(&dp->tq)) == NULL) {
|
|
|
|
dp->st_txq_refl_try++;
|
2008-07-31 17:58:50 -06:00
|
|
|
if (__netif_tx_trylock(txq)) {
|
2006-01-08 23:34:25 -07:00
|
|
|
dp->st_rxq_enter++;
|
|
|
|
while ((skb = skb_dequeue(&dp->rq)) != NULL) {
|
|
|
|
skb_queue_tail(&dp->tq, skb);
|
|
|
|
dp->st_rx2tx_tran++;
|
|
|
|
}
|
2008-07-31 17:58:50 -06:00
|
|
|
__netif_tx_unlock(txq);
|
2006-01-08 23:34:25 -07:00
|
|
|
} else {
|
|
|
|
/* reschedule */
|
|
|
|
dp->st_rxq_notenter++;
|
|
|
|
goto resched;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
while ((skb = skb_dequeue(&dp->tq)) != NULL) {
|
|
|
|
u32 from = G_TC_FROM(skb->tc_verd);
|
|
|
|
|
|
|
|
skb->tc_verd = 0;
|
|
|
|
skb->tc_verd = SET_TC_NCLS(skb->tc_verd);
|
|
|
|
stats->tx_packets++;
|
|
|
|
stats->tx_bytes +=skb->len;
|
2007-03-29 12:46:52 -06:00
|
|
|
|
2009-11-01 12:45:16 -07:00
|
|
|
rcu_read_lock();
|
2009-11-20 16:35:04 -07:00
|
|
|
skb->dev = dev_get_by_index_rcu(&init_net, skb->skb_iif);
|
2007-03-29 12:46:52 -06:00
|
|
|
if (!skb->dev) {
|
2009-11-01 12:45:16 -07:00
|
|
|
rcu_read_unlock();
|
2007-03-29 12:46:52 -06:00
|
|
|
dev_kfree_skb(skb);
|
|
|
|
stats->tx_dropped++;
|
|
|
|
break;
|
|
|
|
}
|
2009-11-01 12:45:16 -07:00
|
|
|
rcu_read_unlock();
|
2009-11-20 16:35:04 -07:00
|
|
|
skb->skb_iif = _dev->ifindex;
|
2007-03-29 12:46:52 -06:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
if (from & AT_EGRESS) {
|
|
|
|
dp->st_rx_frm_egr++;
|
|
|
|
dev_queue_xmit(skb);
|
|
|
|
} else if (from & AT_INGRESS) {
|
|
|
|
dp->st_rx_frm_ing++;
|
2007-03-29 12:46:52 -06:00
|
|
|
skb_pull(skb, skb->dev->hard_header_len);
|
2006-01-08 23:34:25 -07:00
|
|
|
netif_rx(skb);
|
2007-03-29 12:46:52 -06:00
|
|
|
} else
|
|
|
|
BUG();
|
2006-01-08 23:34:25 -07:00
|
|
|
}
|
|
|
|
|
2008-07-31 17:58:50 -06:00
|
|
|
if (__netif_tx_trylock(txq)) {
|
2006-01-08 23:34:25 -07:00
|
|
|
dp->st_rxq_check++;
|
|
|
|
if ((skb = skb_peek(&dp->rq)) == NULL) {
|
|
|
|
dp->tasklet_pending = 0;
|
|
|
|
if (netif_queue_stopped(_dev))
|
|
|
|
netif_wake_queue(_dev);
|
|
|
|
} else {
|
|
|
|
dp->st_rxq_rsch++;
|
2008-07-31 17:58:50 -06:00
|
|
|
__netif_tx_unlock(txq);
|
2006-01-08 23:34:25 -07:00
|
|
|
goto resched;
|
|
|
|
}
|
2008-07-31 17:58:50 -06:00
|
|
|
__netif_tx_unlock(txq);
|
2006-01-08 23:34:25 -07:00
|
|
|
} else {
|
|
|
|
resched:
|
|
|
|
dp->tasklet_pending = 1;
|
|
|
|
tasklet_schedule(&dp->ifb_tasklet);
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2008-11-19 22:47:07 -07:00
|
|
|
static const struct net_device_ops ifb_netdev_ops = {
|
|
|
|
.ndo_open = ifb_open,
|
|
|
|
.ndo_stop = ifb_close,
|
2008-11-20 21:14:53 -07:00
|
|
|
.ndo_start_xmit = ifb_xmit,
|
|
|
|
.ndo_validate_addr = eth_validate_addr,
|
2008-11-19 22:47:07 -07:00
|
|
|
};
|
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
static void ifb_setup(struct net_device *dev)
|
2006-01-08 23:34:25 -07:00
|
|
|
{
|
|
|
|
/* Initialize the device structure. */
|
2007-06-13 13:05:06 -06:00
|
|
|
dev->destructor = free_netdev;
|
2008-11-19 22:47:07 -07:00
|
|
|
dev->netdev_ops = &ifb_netdev_ops;
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
/* Fill in device structure with ethernet-generic values. */
|
|
|
|
ether_setup(dev);
|
|
|
|
dev->tx_queue_len = TX_Q_LIMIT;
|
2008-11-19 22:47:07 -07:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
dev->flags |= IFF_NOARP;
|
|
|
|
dev->flags &= ~IFF_MULTICAST;
|
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 23:19:19 -06:00
|
|
|
dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
|
2006-01-08 23:34:25 -07:00
|
|
|
random_ether_addr(dev->dev_addr);
|
|
|
|
}
|
|
|
|
|
2009-08-31 13:50:51 -06:00
|
|
|
static netdev_tx_t ifb_xmit(struct sk_buff *skb, struct net_device *dev)
|
2006-01-08 23:34:25 -07:00
|
|
|
{
|
|
|
|
struct ifb_private *dp = netdev_priv(dev);
|
2007-10-03 18:41:50 -06:00
|
|
|
struct net_device_stats *stats = &dev->stats;
|
2006-01-08 23:34:25 -07:00
|
|
|
u32 from = G_TC_FROM(skb->tc_verd);
|
|
|
|
|
2007-01-01 20:39:09 -07:00
|
|
|
stats->rx_packets++;
|
|
|
|
stats->rx_bytes+=skb->len;
|
2006-01-08 23:34:25 -07:00
|
|
|
|
2009-11-20 16:35:04 -07:00
|
|
|
if (!(from & (AT_INGRESS|AT_EGRESS)) || !skb->skb_iif) {
|
2006-01-08 23:34:25 -07:00
|
|
|
dev_kfree_skb(skb);
|
|
|
|
stats->rx_dropped++;
|
2009-08-31 13:50:51 -06:00
|
|
|
return NETDEV_TX_OK;
|
2006-01-08 23:34:25 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (skb_queue_len(&dp->rq) >= dev->tx_queue_len) {
|
|
|
|
netif_stop_queue(dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
skb_queue_tail(&dp->rq, skb);
|
|
|
|
if (!dp->tasklet_pending) {
|
|
|
|
dp->tasklet_pending = 1;
|
|
|
|
tasklet_schedule(&dp->ifb_tasklet);
|
|
|
|
}
|
|
|
|
|
2009-08-31 13:50:51 -06:00
|
|
|
return NETDEV_TX_OK;
|
2006-01-08 23:34:25 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int ifb_close(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct ifb_private *dp = netdev_priv(dev);
|
|
|
|
|
|
|
|
tasklet_kill(&dp->ifb_tasklet);
|
|
|
|
netif_stop_queue(dev);
|
|
|
|
skb_queue_purge(&dp->rq);
|
|
|
|
skb_queue_purge(&dp->tq);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ifb_open(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct ifb_private *dp = netdev_priv(dev);
|
|
|
|
|
|
|
|
tasklet_init(&dp->ifb_tasklet, ri_tasklet, (unsigned long)dev);
|
|
|
|
skb_queue_head_init(&dp->rq);
|
|
|
|
skb_queue_head_init(&dp->tq);
|
|
|
|
netif_start_queue(dev);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-07-11 20:42:31 -06:00
|
|
|
static int ifb_validate(struct nlattr *tb[], struct nlattr *data[])
|
|
|
|
{
|
|
|
|
if (tb[IFLA_ADDRESS]) {
|
|
|
|
if (nla_len(tb[IFLA_ADDRESS]) != ETH_ALEN)
|
|
|
|
return -EINVAL;
|
|
|
|
if (!is_valid_ether_addr(nla_data(tb[IFLA_ADDRESS])))
|
|
|
|
return -EADDRNOTAVAIL;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
static struct rtnl_link_ops ifb_link_ops __read_mostly = {
|
|
|
|
.kind = "ifb",
|
|
|
|
.priv_size = sizeof(struct ifb_private),
|
|
|
|
.setup = ifb_setup,
|
2007-07-11 20:42:31 -06:00
|
|
|
.validate = ifb_validate,
|
2007-06-13 13:05:06 -06:00
|
|
|
};
|
|
|
|
|
2007-07-11 20:42:13 -06:00
|
|
|
/* Number of ifb devices to be set up by this module. */
|
|
|
|
module_param(numifbs, int, 0);
|
|
|
|
MODULE_PARM_DESC(numifbs, "Number of ifb devices");
|
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
static int __init ifb_init_one(int index)
|
|
|
|
{
|
|
|
|
struct net_device *dev_ifb;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
dev_ifb = alloc_netdev(sizeof(struct ifb_private),
|
|
|
|
"ifb%d", ifb_setup);
|
|
|
|
|
|
|
|
if (!dev_ifb)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
err = dev_alloc_name(dev_ifb, dev_ifb->name);
|
|
|
|
if (err < 0)
|
|
|
|
goto err;
|
2006-01-08 23:34:25 -07:00
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
dev_ifb->rtnl_link_ops = &ifb_link_ops;
|
|
|
|
err = register_netdevice(dev_ifb);
|
|
|
|
if (err < 0)
|
|
|
|
goto err;
|
2008-03-20 18:05:13 -06:00
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
return 0;
|
2007-06-13 13:04:51 -06:00
|
|
|
|
2007-06-13 13:05:06 -06:00
|
|
|
err:
|
|
|
|
free_netdev(dev_ifb);
|
|
|
|
return err;
|
2006-09-13 11:24:59 -06:00
|
|
|
}
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
static int __init ifb_init_module(void)
|
2006-09-13 11:24:59 -06:00
|
|
|
{
|
2007-06-13 13:05:06 -06:00
|
|
|
int i, err;
|
|
|
|
|
|
|
|
rtnl_lock();
|
|
|
|
err = __rtnl_link_register(&ifb_link_ops);
|
2007-06-13 13:04:51 -06:00
|
|
|
|
2006-01-08 23:34:25 -07:00
|
|
|
for (i = 0; i < numifbs && !err; i++)
|
2006-09-13 11:24:59 -06:00
|
|
|
err = ifb_init_one(i);
|
2007-07-11 20:42:13 -06:00
|
|
|
if (err)
|
2007-06-13 13:05:06 -06:00
|
|
|
__rtnl_link_unregister(&ifb_link_ops);
|
|
|
|
rtnl_unlock();
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
return err;
|
2006-09-13 11:24:59 -06:00
|
|
|
}
|
2006-01-08 23:34:25 -07:00
|
|
|
|
|
|
|
static void __exit ifb_cleanup_module(void)
|
|
|
|
{
|
2007-07-11 20:42:13 -06:00
|
|
|
rtnl_link_unregister(&ifb_link_ops);
|
2006-01-08 23:34:25 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
module_init(ifb_init_module);
|
|
|
|
module_exit(ifb_cleanup_module);
|
|
|
|
MODULE_LICENSE("GPL");
|
|
|
|
MODULE_AUTHOR("Jamal Hadi Salim");
|
2007-06-13 13:05:06 -06:00
|
|
|
MODULE_ALIAS_RTNL_LINK("ifb");
|