2005-04-17 02:20:36 +04:00
/*
* INET An implementation of the TCP / IP protocol suite for the LINUX
* operating system . INET is implemented using the BSD Socket
* interface as the means of communication with the user level .
*
* Pseudo - driver for the loopback interface .
*
* Version : @ ( # ) loopback . c 1.0 .4 b 08 / 16 / 93
*
2005-05-06 03:16:16 +04:00
* Authors : Ross Biro
2005-04-17 02:20:36 +04:00
* Fred N . van Kempen , < waltje @ uWalt . NL . Mugnet . ORG >
* Donald Becker , < becker @ scyld . com >
*
* Alan Cox : Fixed oddments for NET3 .014
* Alan Cox : Rejig for NET3 .029 snap # 3
* Alan Cox : Fixed NET3 .029 bugs and sped up
* Larry McVoy : Tiny tweak to double performance
* Alan Cox : Backed out LMV ' s tweak - the linux mm
* can ' t take it . . .
* Michael Griffith : Don ' t bother computing the checksums
* on packets received on the loopback
* interface .
* Alexey Kuznetsov : Potential hang under some extreme
* cases removed .
*
* 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 .
*/
# include <linux/kernel.h>
# include <linux/jiffies.h>
# include <linux/module.h>
# include <linux/interrupt.h>
# include <linux/fs.h>
# include <linux/types.h>
# include <linux/string.h>
# include <linux/socket.h>
# include <linux/errno.h>
# include <linux/fcntl.h>
# include <linux/in.h>
# include <linux/init.h>
# include <asm/system.h>
# include <asm/uaccess.h>
# include <asm/io.h>
# include <linux/inet.h>
# include <linux/netdevice.h>
# include <linux/etherdevice.h>
# include <linux/skbuff.h>
# include <linux/ethtool.h>
# include <net/sock.h>
# include <net/checksum.h>
# include <linux/if_ether.h> /* For the statistics structure. */
# include <linux/if_arp.h> /* For ARPHRD_ETHER */
# include <linux/ip.h>
# include <linux/tcp.h>
# include <linux/percpu.h>
2007-09-27 09:10:56 +04:00
# include <net/net_namespace.h>
2010-06-22 16:44:11 +04:00
# include <linux/u64_stats_sync.h>
2005-04-17 02:20:36 +04:00
2006-10-19 07:51:57 +04:00
struct pcpu_lstats {
2010-06-22 16:44:11 +04:00
u64 packets ;
u64 bytes ;
struct u64_stats_sync syncp ;
2006-10-19 07:51:57 +04:00
} ;
2005-04-17 02:20:36 +04:00
/*
* The higher levels take care of making this non - reentrant ( it ' s
* called with bh ' s disabled ) .
*/
2009-08-31 23:50:58 +04:00
static netdev_tx_t loopback_xmit ( struct sk_buff * skb ,
struct net_device * dev )
2005-04-17 02:20:36 +04:00
{
2010-02-16 18:21:08 +03:00
struct pcpu_lstats * lb_stats ;
2009-04-18 02:03:10 +04:00
int len ;
2005-04-17 02:20:36 +04:00
skb_orphan ( skb ) ;
2009-04-18 02:03:10 +04:00
skb - > protocol = eth_type_trans ( skb , dev ) ;
2005-04-17 02:20:36 +04:00
2007-09-28 04:09:39 +04:00
/* it's OK to use per_cpu_ptr() because BHs are off */
2010-09-24 03:51:51 +04:00
lb_stats = this_cpu_ptr ( dev - > lstats ) ;
2005-04-17 02:20:36 +04:00
2009-04-18 02:03:10 +04:00
len = skb - > len ;
if ( likely ( netif_rx ( skb ) = = NET_RX_SUCCESS ) ) {
2010-06-22 16:44:11 +04:00
u64_stats_update_begin ( & lb_stats - > syncp ) ;
2009-04-18 02:03:10 +04:00
lb_stats - > bytes + = len ;
lb_stats - > packets + + ;
2010-06-22 16:44:11 +04:00
u64_stats_update_end ( & lb_stats - > syncp ) ;
2010-10-01 01:06:55 +04:00
}
2005-04-17 02:20:36 +04:00
2009-06-23 10:03:08 +04:00
return NETDEV_TX_OK ;
2005-04-17 02:20:36 +04:00
}
2010-07-08 01:58:56 +04:00
static struct rtnl_link_stats64 * loopback_get_stats64 ( struct net_device * dev ,
struct rtnl_link_stats64 * stats )
2005-04-17 02:20:36 +04:00
{
2010-06-14 09:59:22 +04:00
u64 bytes = 0 ;
u64 packets = 0 ;
2005-04-17 02:20:36 +04:00
int i ;
2006-03-28 13:56:37 +04:00
for_each_possible_cpu ( i ) {
2006-10-19 07:51:57 +04:00
const struct pcpu_lstats * lb_stats ;
2010-06-22 16:44:11 +04:00
u64 tbytes , tpackets ;
unsigned int start ;
2005-04-17 02:20:36 +04:00
2010-09-24 03:51:51 +04:00
lb_stats = per_cpu_ptr ( dev - > lstats , i ) ;
2010-06-22 16:44:11 +04:00
do {
start = u64_stats_fetch_begin ( & lb_stats - > syncp ) ;
tbytes = lb_stats - > bytes ;
tpackets = lb_stats - > packets ;
} while ( u64_stats_fetch_retry ( & lb_stats - > syncp , start ) ) ;
bytes + = tbytes ;
packets + = tpackets ;
2005-04-17 02:20:36 +04:00
}
2006-10-19 07:51:57 +04:00
stats - > rx_packets = packets ;
stats - > tx_packets = packets ;
2009-04-18 02:03:10 +04:00
stats - > rx_bytes = bytes ;
stats - > tx_bytes = bytes ;
2005-04-17 02:20:36 +04:00
return stats ;
}
2006-09-28 07:33:34 +04:00
static u32 always_on ( struct net_device * dev )
2005-04-17 02:20:36 +04:00
{
return 1 ;
}
2006-09-13 22:30:00 +04:00
static const struct ethtool_ops loopback_ethtool_ops = {
2006-09-28 07:33:34 +04:00
. get_link = always_on ,
2005-04-17 02:20:36 +04:00
. set_tso = ethtool_op_set_tso ,
2006-09-28 07:33:34 +04:00
. get_tx_csum = always_on ,
. get_sg = always_on ,
. get_rx_csum = always_on ,
2005-04-17 02:20:36 +04:00
} ;
2007-09-27 09:08:12 +04:00
static int loopback_dev_init ( struct net_device * dev )
{
2010-09-24 03:51:51 +04:00
dev - > lstats = alloc_percpu ( struct pcpu_lstats ) ;
if ( ! dev - > lstats )
2007-09-27 09:08:12 +04:00
return - ENOMEM ;
return 0 ;
}
static void loopback_dev_free ( struct net_device * dev )
{
2010-09-24 03:51:51 +04:00
free_percpu ( dev - > lstats ) ;
2007-09-27 09:08:12 +04:00
free_netdev ( dev ) ;
}
2008-11-20 08:46:18 +03:00
static const struct net_device_ops loopback_ops = {
. ndo_init = loopback_dev_init ,
2008-11-21 07:14:53 +03:00
. ndo_start_xmit = loopback_xmit ,
2010-06-14 09:59:22 +04:00
. ndo_get_stats64 = loopback_get_stats64 ,
2008-11-20 08:46:18 +03:00
} ;
2006-09-28 07:33:34 +04:00
/*
2007-09-28 04:09:39 +04:00
* The loopback device is special . There is only one instance
* per network namespace .
2006-09-28 07:33:34 +04:00
*/
2007-09-26 06:18:04 +04:00
static void loopback_setup ( struct net_device * dev )
{
dev - > mtu = ( 16 * 1024 ) + 20 + 20 + 12 ;
dev - > hard_header_len = ETH_HLEN ; /* 14 */
dev - > addr_len = ETH_ALEN ; /* 6 */
dev - > tx_queue_len = 0 ;
dev - > type = ARPHRD_LOOPBACK ; /* 0x0001*/
dev - > flags = IFF_LOOPBACK ;
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>
2009-05-19 09:19:19 +04:00
dev - > priv_flags & = ~ IFF_XMIT_DST_RELEASE ;
2007-09-26 06:18:04 +04:00
dev - > features = NETIF_F_SG | NETIF_F_FRAGLIST
| NETIF_F_TSO
| NETIF_F_NO_CSUM
| NETIF_F_HIGHDMA
| NETIF_F_LLTX
2007-12-27 17:17:22 +03:00
| NETIF_F_NETNS_LOCAL ;
2007-09-26 06:18:04 +04:00
dev - > ethtool_ops = & loopback_ethtool_ops ;
2007-10-09 12:40:57 +04:00
dev - > header_ops = & eth_header_ops ;
2008-11-20 08:46:18 +03:00
dev - > netdev_ops = & loopback_ops ;
dev - > destructor = loopback_dev_free ;
2007-09-26 06:18:04 +04:00
}
2007-09-26 06:16:28 +04:00
2005-08-19 01:05:18 +04:00
/* Setup and register the loopback device. */
2007-10-09 07:38:39 +04:00
static __net_init int loopback_net_init ( struct net * net )
2005-04-17 02:20:36 +04:00
{
2007-09-26 06:18:04 +04:00
struct net_device * dev ;
int err ;
err = - ENOMEM ;
dev = alloc_netdev ( 0 , " lo " , loopback_setup ) ;
if ( ! dev )
goto out ;
2007-07-31 03:37:19 +04:00
2008-03-25 15:47:49 +03:00
dev_net_set ( dev , net ) ;
2007-09-26 06:18:04 +04:00
err = register_netdev ( dev ) ;
2007-07-31 03:37:19 +04:00
if ( err )
2007-09-26 06:18:04 +04:00
goto out_free_netdev ;
2007-07-31 03:37:19 +04:00
2007-09-27 09:10:56 +04:00
net - > loopback_dev = dev ;
2007-10-15 23:55:33 +04:00
return 0 ;
2007-09-26 06:18:04 +04:00
2005-04-17 02:20:36 +04:00
2007-09-26 06:18:04 +04:00
out_free_netdev :
free_netdev ( dev ) ;
2007-10-15 23:55:33 +04:00
out :
2009-11-26 02:14:13 +03:00
if ( net_eq ( net , & init_net ) )
2007-10-15 23:55:33 +04:00
panic ( " loopback: Failed to register netdevice: %d \n " , err ) ;
return err ;
2007-09-26 06:18:04 +04:00
}
2008-11-08 09:54:20 +03:00
/* Registered in net/core/dev.c */
struct pernet_operations __net_initdata loopback_net_ops = {
2007-09-27 09:10:56 +04:00
. init = loopback_net_init ,
} ;